Springer Series in Materials Science 154

Aline Auroux Editor

Calorimetry and
Thermal Methods
in Catalysis
Springer Series in Materials Science

Volume 154

Series Editors
Robert Hull, Charlottesville, VA, USA
Chennupati Jagadish, Canberra, ACT, Australia
Richard M. Osgood, New York, NY, USA
Jürgen Parisi, Oldenburg, Germany
Zhiming M. Wang, Fayetteville, AR, USA

For further volumes:

http://www.springer.com/series/856
The Springer Series in Materials Science covers the complete spectrum of
materials physics, including fundamental principles, physical properties, materials
theory and design. Recognizing the increasing importance of materials science in
future device technologies, the book titles in this series reflect the state-of-the-art
in understanding and controlling the structure and properties of all important
classes of materials.
Aline Auroux
Editor

Calorimetry and Thermal

Methods in Catalysis

123
Editor
Aline Auroux
Institut de Recherches sur la
Catalyse et l’Environnement de Lyon
Villeurbanne
France

ISSN 0933-033X ISSN 2196-2812 (electronic)

ISBN 978-3-642-11953-8 ISBN 978-3-642-11954-5 (eBook)
DOI 10.1007/978-3-642-11954-5
Springer Heidelberg New York Dordrecht London

Library of Congress Control Number: 2013944530

Ó Springer-Verlag Berlin Heidelberg 2013

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or
information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed. Exempted from this legal reservation are brief
excerpts in connection with reviews or scholarly analysis or material supplied specifically for the
purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the
work. Duplication of this publication or parts thereof is permitted only under the provisions of
the Copyright Law of the Publisher’s location, in its current version, and permission for use must
always be obtained from Springer. Permissions for use may be obtained through RightsLink at the
Copyright Clearance Center. Violations are liable to prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for
any errors or omissions that may be made. The publisher makes no warranty, express or implied, with
respect to the material contained herein.

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

Preface

Calorimetry and other thermal methods play an increasingly important role as

tools for the study of catalysts, supports, adsorbents, and for the characterization of
their surfaces. This makes it particularly timely to collect in a single volume a set
of texts on the fundamentals of adsorption and the bases of thermal analysis
techniques such as microcalorimetry, differential scanning calorimetry, thermo-
gravimetry, temperature-programmed desorption, temperature-programmed
reduction/oxidation, inverse gas chromatography, etc. The use of many of these
techniques is now fairly routine, but their application in the domain of catalysis
often requires coupling them to other methods such as volumetry, gas chroma-
tography, mass spectrometry, infrared spectroscopy, etc., in order to allow for an
in-depth study of the successive stages of the life cycle of a catalyst, from its
preparation to its use in the catalytic reaction and finally its regeneration. Tools
that allow the measurement of the heats evolved during the various steps of a
chemical reaction are particularly useful given the current emphasis on energy
efficiency. However, among the various techniques mentioned above, used alone,
or in couplings, each have their own strengths and weaknesses that need to be
carefully discussed.
This book, based on a series of summer schools held in Lyon every year since
2007, aims to provide students, engineers, and confirmed researchers alike with an
introduction to the major thermal analysis techniques used to characterize solid
materials and investigate their surface reactivity, including both physical and
chemical processes occurring at gas–solid or liquid–solid interfaces. The main
topics covered include:
– the basic phenomena (adsorption, competitive adsorption, desorption, thermo-
dynamics, and kinetics) involved at the solid–gas and solid–liquid interfaces;
– the main thermal analysis and calorimetry techniques used to investigate cata-
lytic materials, alone or linked to other techniques, and their relative advantages;
including the main types of calorimetric techniques (adsorption calorimetry, flow
calorimetry, titration calorimetry, immersion calorimetry, etc.), as well as tem-
perature-programmed desorption, reduction and oxidation, along with a number
of examples and a discussion of experimental considerations and constraints;

v
vi Preface

– applications of these techniques, such as the study of competitive or selective

adsorption processes, the characterization of acid/base sites in oxides and zeo-
lites, the adsorption or capture of gas or liquid pollutants (CO, CO2, VOCs,
nicotine, phenol, etc.), and processes for new energies (biodiesel production,
hydrogen production and storage, etc.).
We hope that this book will thus serve as a practical guide to end users about
selecting and implementing the most appropriate thermal analysis techniques for
solving a specific problem.
In closing, I would like to express my sincere gratitude to my colleagues who
kindly agreed to write up the various lecture notes and contributions that make up
this volume, as well as to all those who helped with the design, redaction, and
editing of this book.

Villeurbanne, France Aline Auroux

Contents

Part I Fundamentals and Techniques

1 Fundamentals in Adsorption at the Solid–Gas Interface.

Concepts and Thermodynamics . . . . . . . . . . . . . . . . . . . . . . .... 3
Vera Bolis
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 The Solid Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 Porous Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Adsorption Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.1 Adsorption Isotherms. . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4 Adsorption Microcalorimetry . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.4.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.4.2 Equilibrium Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.5 Thermodynamics of Adsorption. . . . . . . . . . . . . . . . . . . . . . . . 29
1.5.1 Heat of Adsorption from Direct Calorimetric
Methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 30
1.5.2 Heat of Adsorption from Indirect Non-Calorimetric
Methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 32
1.5.3 Entropy of Adsorption . . . . . . . . . . . . . . . . . . . . . .... 33
1.6 Adsorption of a Single Component: Physisorption Versus
Chemisorption. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 38
1.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 44
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 45

2 Thermal Analysis and Calorimetry Techniques for Catalytic

Investigations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 51
Pierre Le Parlouër
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 51
2.2 Thermal Analysis and Calorimetry: Techniques
and Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.3 The Differential Thermal Analysis Technique . . . . . . . . . . . . . . 53
2.3.1 Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.3.2 Detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

vii
viii Contents

2.3.3 Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.3.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.4 The Differential Scanning Calorimetry Technique . . . . . . . . . . . 58
2.4.1 Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.4.2 Detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.4.3 Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.4.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.5 The Calorimetric Techniques . . . . . . . . . . . . . . . . . . . . . . . . . 70
2.5.1 Calorimetric Principles . . . . . . . . . . . . . . . . . . . . . . . . 70
2.5.2 Isothermal Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . 72
2.5.3 Isothermal Titration Calorimetry . . . . . . . . . . . . . . . . . . 80
2.6 The Thermogravimetric Technique . . . . . . . . . . . . . . . . . . . . . 81
2.6.1 Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
2.6.2 Detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.6.3 Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
2.6.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
2.7 The Simultaneous Techniques . . . . . . . . . . . . . . . . . . . . . . . . . 89
2.8 Evolved Gas Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
2.8.1 TG-MS Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
2.8.2 TG-FTIR Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
2.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

3 Couplings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 103

Dušan Stošić and Aline Auroux
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
3.2 Coupled Calorimetry–Volumetry . . . . . . . . . . . . . . . . . . . . . . . 104
3.3 Coupled Calorimetry–Gravimetry . . . . . . . . . . . . . . . . . . . . . . 112
3.4 Temperature Programmed Desorption Technique. . . . . . . . . . . . 113
3.5 Temperature Programmed Reduction . . . . . . . . . . . . . . . . . . . . 115
3.6 Calorimetry–Gas Chromatography/Mass Spectrometry . . . . . . . . 116
3.7 Calorimetry-Syringe Pump-UV-Vis Fluorescence
Spectrometry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 117
3.8 Limitations of Technique . . . . . . . . . . . . . . . . . . . . . . . . .... 121
3.9 Influence of the Adsorption Temperature on the Acid/Base
Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
3.10 Probe Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
3.10.1 Probing Surface Acidic Properties. . . . . . . . . . . . . . . . . 123
3.10.2 Probing Surface Basic Properties . . . . . . . . . . . . . . . . . 125
3.10.3 Probing Surface Redox Properties . . . . . . . . . . . . . . . . . 126
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
Contents ix

4 Temperature-Programmed Desorption (TPD) Methods . . . . . . . . . 131

Vesna Rakić and Ljiljana Damjanović
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
4.2 Adsorption–Desorption; Fundamental Principles . . . . . . . . . . . . 134
4.2.1 Thermodynamic View . . . . . . . . . . . . . . . . . . . . . . . . . 135
4.2.2 Kinetics of Adsorption and Desorption . . . . . . . . . . . . . 136
4.3 Experimental Setups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
4.4 The Design of Temperature-Programmed Experiment; Obtained
Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
4.4.1 The Design of TPR/TPO Experiments; Obtained Data. . . 144
4.5 The Interpretation of Results Obtained
from Temperature-Programmed Desorption Experiments . . . . . . 145
4.5.1 The Application of Temperature-Programmed
Desorption in Active Sites Characterisation . . . . . . . . . . 146
4.5.2 The Application of TPD in the Determination
of Kinetic and Thermodynamic Parameters
of Desorption Processes . . . . . . . . . . . . . . . . . . . . . . . . 154
4.6 The Examples of TPD Application; the Comparison
with Data Obtained by Adsorption Calorimetry. . . . . . . . . . . . . 162
4.6.1 Zeolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
4.6.2 Metal Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
4.6.3 Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
4.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

5 Temperature Programmed Reduction/Oxidation (TPR/TPO)

Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 175
Antonella Gervasini
5.1 Redox Properties of Metal Oxides and Catalytic Implications . . . 175
5.2 Temperature-Programmed Reduction/Oxidation Technique . . . . . 180
5.2.1 General Operative Procedure . . . . . . . . . . . . . . . . . . . . 181
5.2.2 Analytical Parameters . . . . . . . . . . . . . . . . . . . . . . . . . 183
5.2.3 Selection of Operating Parameters. . . . . . . . . . . . . . . . . 185
5.3 Kinetics and Reduction Mechanisms . . . . . . . . . . . . . . . . . . . . 186
5.3.1 Nucleation Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
5.3.2 Contracting Sphere Model . . . . . . . . . . . . . . . . . . . . . . 188
5.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

6 Calorimetry at the Solid–Liquid Interface. . . . . . . . . . . . . . . . . . . 197

Jerzy Jozef Zajac
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
6.2 Thermodynamic Treatment of the Solid–Liquid Interface
and the Related Interfacial Phenomena. . . . . . . . . . . . . . . . . . . 199
x Contents

6.2.1 Surface Excess Functions and Surface Phase Model . . . . 199

6.2.2 Adhesion and Cohesion . . . . . . . . . . . . . . . . . . . . . . . . 203
6.2.3 Wetting in Solid–Liquid Systems . . . . . . . . . . . . . . . . . 208
6.2.4 Hydrophobic and Hydrophilic Substances . . . . . . . . . . . 210
6.3 Calorimetry Applied to Evaluate Surface Properties of Solids . . . 210
6.3.1 Enthalpy Changes in the Thermodynamic Cycle of
Immersion-Adsorption–Wetting . . . . . . . . . . . . . . . . .. 212
6.3.2 Immersional and Wetting Calorimetry Experiments. . . .. 214
6.3.3 Hydrophilic-Hydrophobic Series and Harkins-Jura
Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 219
6.4 Enthalpy Changes Accompanying Competitive Adsorption
from Dilute Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
6.4.1 Thermal Properties of Dilute Solutions . . . . . . . . . . . . . 225
6.4.2 Macroscopic Description of Competitive Adsorption . . . . 229
6.4.3 Competitive Adsorption Measurements . . . . . . . . . . . . . 231
6.4.4 Immersion in Dilute Solutions . . . . . . . . . . . . . . . . . . . 234
6.4.5 Model of Flow Calorimetry Experiment. . . . . . . . . . . . . 236
6.4.6 Model of Batch Calorimetry Experiment . . . . . . . . . . . . 240
6.5 Calorimetry Applied to Study Competitive Adsorption
from Dilute Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
6.5.1 Flow Calorimetry System. . . . . . . . . . . . . . . . . . . . . . . 247
6.5.2 Measurements of Integral Enthalpy of Displacement . . . . 250
6.5.3 Titration Calorimetry System . . . . . . . . . . . . . . . . . . . . 254
6.5.4 Scanning of Surfactant Aggregation by Titration
Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 258
6.6 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 263
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 264

Part II Applications and Case Studies

7 Study of Selective Adsorption of Gases by Calorimetry . . . . . . . .. 273

Jean-Pierre Bellat
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
7.2 Definition of Selective Adsorption. . . . . . . . . . . . . . . . . . . . . . 274
7.2.1 Adsorption of Single Component . . . . . . . . . . . . . . . . . 274
7.2.2 Adsorption of Gas Mixtures . . . . . . . . . . . . . . . . . . . . . 278
7.3 Adsorption Enthalpies and Entropies . . . . . . . . . . . . . . . . . . . . 282
7.3.1 Adsorption Enthalpy . . . . . . . . . . . . . . . . . . . . . . . . . . 283
7.3.2 Adsorption Entropy and Molar Adsorbate Entropy . . . . . 283
7.4 Calculation of Adsorption Enthalpy and Entropy from Single
Adsorption Isotherms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 284
7.4.1 Van’t Hoff Method . . . . . . . . . . . . . . . . . . . . . . . . . .. 285
7.4.2 Isosteric Method So-called Clausius-Clapeyron
Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 286
Contents xi

7.5 Determination of Coadsorption Enthalpy and Entropy

by Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
7.5.1 Experimental Calorimetric Technique . . . . . . . . . . . . . . 288
7.5.2 Measurement of Adsorbed Amounts . . . . . . . . . . . . . . . 289
7.5.3 Measurement of Differential Adsorption Enthalpy. . . . . . 292
7.5.4 Adsorption Gibbs Energy . . . . . . . . . . . . . . . . . . . . . . . 297
7.5.5 Differential Adsorption Entropy and Molar Entropy
of Adsorbate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302
7.5.6 Partial Adsorption Enthalpy and Entropy . . . . . . . . . . . . 303
7.6 Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
7.6.1 Separation of Xylenes Isomers by Selective Adsorption
on FAU Type Zeolite . . . . . . . . . . . . . . . . . . . . . . . . . 305
7.6.2 Desulphurization of Natural Gas by Selective Adsorption
on FAU Type Zeolite . . . . . . . . . . . . . . . . . . . . . . . . . 311
7.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318

8 Characterization of Acid–Base Sites in Oxides . . . . . . . . . . . . . . . 319

Antonella Gervasini
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
8.2 The Surface Acido–Basicity of Metal Oxides . . . . . . . . . . . . . . 320
8.3 Acid, Basic, and Amphoteric Oxides . . . . . . . . . . . . . . . . . . . . 323
8.4 Heterogeneous Character of Oxides . . . . . . . . . . . . . . . . . . . . . 325
8.5 Single Oxides, Doped and Modified Oxides, Supported Oxides,
Mixed Oxides, and Complex Oxides . . . . . . . . . . . . . . . . . . . . 330
8.6 Acidity Prediction from Composition . . . . . . . . . . . . . . . . . . . . 343
8.7 Intrinsic and Effective Acidity of Oxide Surfaces . . . . . . . . . . . 346
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349

9 Characterization of Acid–Base Sites in Zeolites . . . . . . . . . . . . ... 353

Dušan Stošić and Aline Auroux
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
9.2 Factors Influencing the Acid Properties of Zeolites . . . . . . . . . . 356
9.2.1 Influence of the Zeolite Topology . . . . . . . . . . . . . . . . . 357
9.2.2 Influence of the Si/Al Ratio . . . . . . . . . . . . . . . . . . . . . 359
9.2.3 Influence of the Pre-treatment Parameters . . . . . . . . . . . 365
9.2.4 The Effect of Proton (Cation) Exchange Level . . . . . . . . 368
9.2.5 The Influence of the Framework T Atom. . . . . . . . . . . . 370
9.3 Correlation Between Adsorption Heat and Catalytic Activity . . . 376
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
xii Contents

10 Adsorption/Desorption of Simple Pollutants . . . . . . . . . . . . . . . .. 385

Vesna Rakić
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 385
10.1.1 Possible Abatement Procedures. . . . . . . . . . . . . . . . . .. 387
10.2 The Application of Calorimetry in Environment Protection . . .. 389
10.2.1 Calorimeters Can be Applied for Direct Investigation
of Some Event that Includes Specific Pollutant(s) . . . . .. 390
10.2.2 Calorimeters Can be Applied for the Characterization
of Solid Materials . . . . . . . . . . . . . . . . . . . . . . . . . . .. 398
10.3 The Application of Temperature-Programmed
Techniques in Environment Protection . . . . . . . . . . . . . . . . . .. 400
10.4 The Application of Thermo-analytical Methods
in Environment Protection . . . . . . . . . . . . . . . . . . . . . . . . . .. 402
10.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 405
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 405

11 Hydrogen and Calorimetry: Case Studies . . . . . . . . . . . . . . ..... 409

Simona Bennici and Aline Auroux
11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
11.2 First Case Study: Irreversible H2 Storage (Borohydrides) . . . . . . 411
11.2.1 Hydrolysis of NaBH4 Stabilized Solutions . . . . . . . . . . . 411
11.2.2 Hydrolysis of NaBH4 and KBH4 Powders . . . . . . . . . . . 416
11.3 Second Case Study: Reversible H2 Storage
(Mg-Based Materials) . . . . . . . . . . . . . . . . . . . . . . . . . ..... 422
11.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 426
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 426

12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt

Interaction for PEMFC Applications: A Case Study . . . . . . . .... 429
Georgeta Postole and Aline Auroux
12.1 Evolution and Types of Fuel Cell . . . . . . . . . . . . . . . . . . .... 429
12.2 Proton Exchange Membrane Fuel Cells . . . . . . . . . . . . . . .... 433
12.3 CO Adsorption Microcalorimetry on Pt-Based Materials:
Literature Survey. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 437
12.4 The CO Poisoning Effects on Pt/C Studied by Adsorption
Microcalorimetry: A Case Study . . . . . . . . . . . . . . . . . . . .... 441
12.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 450
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 450

13 Biodiesel: Characterization by DSC and P-DSC . . . . . . . . . . . . . . 455

Rodica Chiriac, François Toche and Christian Brylinski
13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455
13.2 Study of the Cold Flow Behavior of Biodiesel by DSC
and Thermomicroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
Contents xiii

13.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . .......... 458

13.2.2 Experimental Procedures . . . . . . . . . . . . . .......... 460
13.2.3 DSC and Thermomicroscopy for the Study
of Biodiesel and Biodiesel Blends. . . . . . . . . . . . . . . . . 461
13.2.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 468
13.3 Oxidative Stability of Biodiesel by P-DSC . . . . . . . . . . . . . . . . 469
13.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 469
13.3.2 Experimental Procedure . . . . . . . . . . . . . . . . . . . . . . . . 472
13.3.3 Results and Discussions . . . . . . . . . . . . . . . . . . . . . . . . 473
13.3.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476
13.4 Overall Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 477

14 CO2 Capture in Industrial Effluents. Calorimetric Studies . . . . . .. 481

Jean-Yves Coxam and Karine Ballerat-Busserolles
14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 481
14.2 Presentation of Techniques for CO2 Separation from Gaseous
Effluents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 482
14.3 Industrial Processes Proposed for CO2 Capture in Post
Combustion Effluents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 484
14.4 Thermodynamic Approach of CO2 Dissolution in Aqueous
Solutions of Amine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485
14.4.1 Mechanism of CO2 Dissolution . . . . . . . . . . . . . . . . . . 485
14.4.2 Selection of Amines for CO2 Capture Processes . . . . . . . 486
14.4.3 Calorimetric Experimental Data Required . . . . . . . . . . . 487
14.5 Calorimetric Studies of CO2 Dissolution in Amine Solutions . . . 490
14.5.1 Calorimetric Techniques for Measuring
Heat of Mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 490
14.5.2 Calorimetric Investigations. . . . . . . . . . . . . . . . . . . . . . 494
14.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500

15 Adsorption Microcalorimetry, IR Spectroscopy and Molecular

Modelling in Surface Studies . . . . . . . . . . . . . . . . . . . . . . . . . ... 505
Vera Bolis
15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 505
15.2 CO Adsorbed on Coordinatively Unsaturated Metal Cations . ... 506
15.3 NH3 Adsorbed on All-Silica MFI Zeolites (Silicalite) . . . . . ... 512
15.4 H2O Vapor Adsorbed on Crystalline and on Amorphous
Alumino-Silicates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 514
15.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 516
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 517
xiv Contents

16 Characterisation of Catalysts and Adsorbents by Inverse Gas

Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 521
Eva Díaz and Salvador Ordóñez
16.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521
16.2 Experimental. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 522
16.3 Adsorption Isotherms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526
16.4 Thermodynamic Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . 529
16.4.1 Retention Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529
16.4.2 Free Energy of Adsorption . . . . . . . . . . . . . . . . . . . . . . 531
16.4.3 Enthalpy and Entropy of Adsorption . . . . . . . . . . . . . . . 531
16.4.4 Work of Adhesion: Dispersive and Specific
Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
16.4.5 Surface Heterogeneity . . . . . . . . . . . . . . . . . . . . . . . . . 537
16.5 Applications and Comparison to Other Techniques . . . . . . . . . . 538
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 540

17 Liquid–Solid Adsorption Properties: Measurement of the Effective

Surface Acidity of Solid Catalysts . . . . . . . . . . . . . . . . . . ....... 543
Paolo Carniti and Antonella Gervasini
17.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 543
17.2 Pulse Liquid Chromatographic Method . . . . . . . . . . . . . . . . . . 545
17.3 Liquid Recirculation Chromatographic Method . . . . . . . . . . . . . 547
17.4 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 550
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 551

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553
Contributors

Aline Auroux Institut de Recherches sur la Catalyse et l’Environnement de Lyon,

UMR 5256 CNRS/Université Lyon1, 2 avenue Einstein, 69626 Villeurbanne,
France, e-mail: [email protected]
Karine Ballerat-Busserolles Institut de Chimie de Clermont-Ferrand, Clermont
Université, Université Blaise Pascal, BP 10048, 63000 Clermont-Ferrand, France,
e-mail: [email protected]
Jean-Pierre Bellat Laboratoire Interdisciplinaire Carnot de Bourgogne, Univer-
sité de Bourgogne, UMR 6303 CNRS, 9 avenue A. Savary, BP 47870, 21078
Dijon, France, e-mail: [email protected]
Simona Bennici Institut de Recherches sur la Catalyse et l’Environnement de
Lyon, UMR 5256 CNRS/Université Lyon1, 2 avenue Einstein, 69626 Villeurb-
anne, France, e-mail: [email protected]
Vera Bolis Dipartimento di Chimica, Università di Torino, Via Pietro Giuria 7,
10125 Turin, Italy, e-mail: [email protected]
Christian Brylinski Laboratoire des Multimatériaux et Interfaces, Université
Claude Bernard, Lyon 1, UMR CNRS 5615 Bât. Berthollet, 43 Bddu 11 Novembre
1918, 69622 Villeurbanne, France, e-mail: [email protected]
Paolo Carniti Dipartimento di Chimica, Università degli Studi di Milano, Via
Camillo Golgi, 19, 20133 Milan, Italy, e-mail: [email protected]
Rodica Chiriac Laboratoire des Multimatériaux et Interfaces, Université Claude
Bernard, Lyon 1, UMR CNRS 5615 Bât. Berthollet, 43 Bddu 11 Novembre 1918,
69622 Villeurbanne, France, e-mail: [email protected]
Jean-Yves Coxam Institut de Chimie de Clermont-Ferrand, Clermont Université,
Université Blaise Pascal, BP 10448, 63000 Clermont-Ferrand, France, e-mail: j-
[email protected]
Ljiljana Damjanovic Faculty of Physical Chemistry, University of Belgrade,
Studentski trg 12-16, P.O. Box 47, 11158 Belgrade 118, Serbia, e-mail: ljiljana@
ffh.bg.ac.rs

xv
xvi Contributors

Eva Diaz Faculty of Chemistry, Department of Chemical Engineering and

Environmental Technology, University of Oviedo, 33006 Oviedo, Spain, e-mail:
[email protected]
Antonella Gervasini Dipartimento di Chimica, Università degli Studi di Milano,
Via Camillo Golgi 19, 20133 Milan, Italy, e-mail: [email protected]
Pierre Leparlouer SETARAM Instrumentation, 7 Rue de l’Oratoire, 69300
Caluire-et-Cuire, France, e-mail: [email protected]
Salvador Ordonez Faculty of Chemistry, Department of Chemical Engineering
and Environmental Technology, University of Oviedo, 33006 Oviedo, Spain, e-
mail: [email protected]
Georgeta Postole Institut de Recherches sur la Catalyse et l’Environnement de
Lyon, UMR 5256 CNRS/Université Lyon1, 2 avenue Einstein, 69626 Villeurb-
anne, France, e-mail: [email protected]
Vesna Rakic Faculty of Agriculture, Department of Chemistry, University of
Belgrade, Nemanjina 6, 11080 Zemun, Serbia, e-mail: [email protected]
Dusan Stosic Institut de Recherches sur la Catalyse et l0 Environnement de Lyon,
UMR 5256 CNRS/Université Lyon1, 2 avenue Einstein, 69626 Villeurbanne,
France, e-mail: [email protected]
François Toche Laboratoire des Multimatériaux et Interfaces, Université Claude
Bernard, Lyon 1, UMR CNRS 5615 Bât. Berthollet, 43 Bddu 11 Novembre 1918,
69622 Villeurbanne, France, e-mail: [email protected]
Jerzy Zajac Institut Charles Gerhardt Montpellier, UMR-5253 CNRS-UM2-
ENSCM-UM1, C. C. 1502 Place Eugène Bataillon, 34095 Montpellier cedex 5,
France, e-mail: [email protected]
 Part I
Fundamentals and Techniques
Chapter 1
Fundamentals in Adsorption at the Solid-Gas
Interface. Concepts and Thermodynamics

Vera Bolis

Abstract Some fundamental concepts about the features of a solid material surface
and the adsorption at the gas-solid interface are illustrated. The basic tools dealing
with the thermodynamics aspects of adsorption processes are also discussed. The
stepwise adsorption microcalorimetry technique, which is a tool of greatest quanti-
tative merit in surface chemistry studies, is described in detail through a selection
of gas-solid interface systems, taken from different materials science fields. Criteria
for discriminating physical and chemical adsorption are given, based on the nature
of the forces involved in the process and the heat of adsorption values. The mole-
cular interpretation of the volumetric-calorimetric data, favored by the joint use of
adsorption microcalorimetry, spectroscopic and/or ab initio modeling techniques, is
also stressed by illustrating a number of examples dealing with either physical or
associative/dissociative chemical adsorption.

1.1 Introduction

Phenomena taking place at the solid-fluid interface are governed by specific and/or
aspecific interactions between the atoms at the solid surface and the molecules
approaching the surface from the gas (or the liquid) phase. In particular, heteroge-
neous catalysis is based on a sequence of steps which involve adsorption of reactants
at the surface of the solid material, surface reactions and desorption of final products
[1, 2]. On the other hand, the adsorption of (bio)molecules at the solid surface of bio-
materials in contact with physiological fluids is recognized to be the initial step of a
chain of molecular events leading to the favorable integration of the implanted mate-
rial [3, 4]. The adsorption features of (probe) molecules on solid surfaces have been

V. Bolis (B)
Dipartimento di Chimica and NIS Centre of Excellence, Università di Torino,
Via Pietro Giuria 7, 10125 Torino, Italy
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 3

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_1,
© Springer-Verlag Berlin Heidelberg 2013
4 V. Bolis

studied over the years by a variety of different techniques, which allowed to describe
the nature of the fluid-solid interactions and give insight into the properties of the
solid surface [1, 2, 5, 6]. On one hand, spectroscopic techniques (in particular, IR
and Raman, UV-vis, NMR, XPS, EXAFS-XANES) are suitable methods to describe
the microscopic features of the solid, i.e., the nature and structure of both surface
terminations and adsorbed species formed upon contact with molecules [7–9]. On
the other hand, adsorption microcalorimetry represents a tool of greatest quantitative
merit in that the heat evolved when a fluid contacts the solid surface is related to the
nature and energy of the adsorbed species/surface atoms interactions [10–16]. Fur-
ther, the knowledge of the energetics of chemical and physical events responsible for
the process as well as the assessment of the associated thermodynamic parameters
contributes to a thorough understanding of phenomena taking place at both catalytic
and biological interfaces [13–17]. Coupling the molar volumetric-calorimetric data
with the molecular information on the nature of the interaction arising from both
spectroscopic methods and ab initio computational results, has been proved very
fruitful in characterizing in the surface acidity of materials [16, 18–29].

1.2 The Solid Surface

When a molecule (or an atom) from the gas-phase approaches a solid, it is more or
less strongly attracted by the atoms exposed at the surface, according to the nature
of both the molecule and the solid material.
A crystalline solid is described through the periodic infinite repetition of an ele-
mental pattern (unit cell) [1]. A real solid is however necessarily finite: the periodic
repetition of the unit cell terminates generating a surface, the structure of which
depends on the cleavage of the crystal, the chemical (either ionic or covalent) nature
of the solid and the origin of the surface (either chemical or mechanical). The surface
atoms arrangement depends on the plane preferentially exposed during the forma-
tion of the surface, according to the preparation conditions of the real material (either
in the single crystal form or as nanosized powder) [30]. If no major reconstruction
processes are required in order to minimize the surface atoms energy, and if no
structural/compositional defects are present, an ideal perfect homogeneous surface
is obtained which can be properly represented by cutting a slab of the solid struc-
ture. Such an ideal perfect homogeneous surface is very rarely encountered, unless
especially prepared for surface science studies. Real solid surfaces (mostly in the
case of finely divided, nanometric sized solids) are made up of a combination of
flat regions (terraces), structural defects (steps, kinks, corners, edges), point defects
(vacancies of ions/atoms in the solid), as schematically illustrated in Fig. 1.1 Com-
positional defects may contribute to the “imperfections” of the solid surface. They
include a variety of oxidation states of the atoms constituting the solid and/or a vari-
ety of heteroatoms present either as impurities, or especially introduced in order to
modify the physico-chemical properties of the surface. This means that in a real solid
1 Fundamentals in Adsorption at the Solid-Gas Interface 5

Fig. 1.1 Cartoon of a piece of realistic MgO nanocrystal, which exhibits both structural (steps,
kinks edges and corners) and point (anionic and cationic vacancies) defects along the flat regions
(terraces). The presence of a compositional defect (substitutional cation) is also outlined. By courtesy
of Prof. Piero Ugliengo, University of Torino

a heterogeneous distribution of surface sites potentially active in catalytic reactions

and/or in interface processes is generally expected.
Another type of heterogeneous solid surface is represented by an active material
dispersed ad hoc over the surface of another solid (the support) [31].
In recent years, with the advent of high resolution electron microscopes it has
become possible to image atomic details in nanocrystals [9]. Some of the above men-
tioned structural defects can be imaged by the high resolution transmission electron
microscopy (HR-TEM), as illustrated in Fig. 1.2 for monoclinic ZrO2 nanocrystals,
which terminate with structural defects as steps, kinks edges and corners [19].
Owing to their intrinsic coordinative and/or valence unsaturation, species making
up such defects act as surface highly reactive sites capable of taking up molecules
from the environment.

1.2.1 Porous Materials

Finely divided solids possess not only a geometrical surface, as defined by the dif-
ferent planes exposed by the solid, but also an internal surface due to the primary
particles aggregation, which generates pores of different size according to both the
nature of the solid and origin of the surface. Pores are classified on the basis of their
width w, which represents either the diameter of a cylindrical pore, or the distance
between the sides of a slit-shaped pore [32]. The smallest pores, characterized by a
width w < 20 Å (2 nm) are defined micropores; the intermediate pores, characterized
by a width comprised in the 20 Å ≤ w ≤ 500 Å (2 and 50 nm) range are classified as
6 V. Bolis

Fig. 1.2 High resolution transmission electron microscopy (HR-TEM) of monoclinic ZrO2
nanocrystals. Adapted from Ref. [19], Fig.2a

mesopores, whereas the largest pores, characterized by a width w > 500 Å (50 nm)
as macropores [31, 32].
Some materials, like charcoal and silico-alumina, have irregular pores with widely
variable diameters in a normal shape. Conversely, other materials such as zeolites
and clay minerals are entirely micro- or meso-porous, respectively. In these cases,
the porosity does not arise from the primary particles aggregation but is an intrinsic
structural feature of the solid material [31, 33, 34].
Zeolites are either natural or synthetic crystalline alumino-silicates, the structure
of which is based upon a three dimensional polymeric framework, with nanosized
cages and channels [1, 2, 33–35]. The basic building block of such materials, of
 
n+ (AlO ) (SiO ) x− ·zH O, is the [TO ] unit with T = Si,Al.
general formula Mx/n 2 x 2 y 2 4
This unit is a tetrahedron centered on one T atom bound to four O atoms located at the
corners; each O atom is in turn shared between two T atoms. These tetrahedral units
join each another through T−O−T linkages in a variety of open-structure frameworks
characterized by (interconnected) channels and voids which are occupied by cations
and water molecules. The presence of charge-balancing (extra-framework) cations is
required in order to compensate the negative charge of the tetrahedral [AlO4 ]− units in
which Al is in isomorphous substitution of Si atoms. The density of charge-balancing
cations depends upon the Al Si
ratio, which span in the 1 to ∞ range (for AlSi
→∞
the so-called all-silica zeolites are obtained) [36–38]. The nature and distribution of
extra-framework cations, which are intrinsically mobile and can be exchanged by
other cations (including the acidic proton) give specific chemical properties to the
material. On the other hand, the presence of nanosized cages and channels within
the crystalline structure of zeolites gives to these materials unique molecular sieve
1 Fundamentals in Adsorption at the Solid-Gas Interface 7

Fig. 1.3 Schematic picture of adsorption process at the surface of a solid material

and shape selectivity properties, of greatest interest in catalysis and in gas separation
processes [1, 2, 33, 34, 38, 39].
In all cases, porous materials exhibit high surface areas, which maximize the
extension of the interface region.

1.3 Adsorption Processes

The surface atoms of a solid, which are coordinatively unsaturated with respect to
the bulk atoms, become saturated thanks to the interaction with molecules of the
environment. Adsorption is the process whereby molecules from the gas (or liquid)
phase are taken up by a solid surface; it is distinguished from absorption which refers
to molecules entering into the lattice (bulk) of the solid material. The adsorptive is
the material in the gas phase capable of being adsorbed, whereas the adsorbate is the
material actually adsorbed by the solid. The solid, which exposes the surface sites
responsible for the process is called the adsorbent. In Fig. 1.3 the adsorption process
at the surface of a solid material is schematically illustrated.
Adsorption is governed by either physical or chemical forces. In the former case
the adsorption is named physical adsorption (physisorption) whereas in the latter
case chemical adsorption (chemisorption). Details on the nature of these forces will
be dealt in Sect. 1.6, through the description of a selection of examples.
When a solid is exposed in a closed space to a gas at pressure p, the weight
of the solid typically increases and the pressure of the gas decreases: the gas is
adsorbed by the solid. After a time, the pressure p does not change any more and
correspondingly the weight ceases to increase any further: a dynamic equilibrium is
reached. The amount of gas adsorbed is experimentally determined: (i) by gravimetry
(the increase in weight of the solid is monitored by a spring balance); (ii) by volumetry
(the fall in the gas pressure is monitored by manometers/transducer gauges); (iii) by
8 V. Bolis

monitoring the change of any other physical parameter related to the adsorption of
matter, such as the evolved heat (if the heat of adsorption is known and constant) or
the integrated IR absorbance (if the specific molar absorbance of adsorbed species
is known) [20, 40–42].

1.3.1 Adsorption Isotherms

Adsorption is described through isotherms, i.e., through the functions connecting the
amount of adsorbate taken up by the adsorbent (or the change of any other physical
parameter related to the adsorption of matter) with the adsorptive equilibrium pres-
sure p, the temperature T and all other parameters being constant. Below the critical
temperature the pressure is properly normalized to the saturation vapor pressure p◦ ,
and the adsorbed amounts are so referred to the dimensionless relative pressure, pp◦ .
Adsorption isotherms are currently classified in five classes (I - V) according to
the Brunauer, Deming, Deming, Teller (BDDT) original classification, [43] which
is often referred to as the Brunauer, Emmet, and Teller (BET), [44] or simply to
as the Brunauer [45] classification. An extra type of isotherm (the stepped Type VI
isotherm, which is relatively rare) is also reported. Type IV and V isotherms typi-
cally exhibit a hysteresis loop, which is characteristic of porous systems, involving
capillary condensation [32].
The fractional coverage θ of the adsorbate, at a given equilibrium pressure p,
is defined as the ratio of NS surface sites occupied by the adsorbate over the total
available adsorption sites N, i.e. the total number of substrate surface sites which
are active towards the given adsorptive. The first layer of adsorbed phase is due to
either chemisorption or physisorption, or both, according to the nature of the forces
governing the adsorbate/adsorbent interactions (vide infra Sect. 1.6). Conversely, the
second layer is originated by physical forces, similar to the forces that lead to the non-
ideal behavior of gases and eventually to the condensation to the liquid. Subsequent
layers are expected to approach a liquid-like phase.
When the number of NS occupied sites matches the number of total available
sites N, the adsorbate monolayer is complete (θ = 1). In Fig. 1.4 the formation
of subsequent layers of adsorbate at the surface of a solid sample is schematically
illustrated.
The amount of gas taken up by a solid surface depends upon the solid and the gas
nature, the pressure p of the gas and the temperature T . The uptake being proportional
to the mass m and the surface area A of the sample, adsorbed amounts (often expressed
as mass or volume STP of the gas) are properly normalized either to the unit mass or
to the unit surface area. Here, in view of describing the process at molecular details,
the adsorbed amounts nads are properly expressed as adsorbate moles (or molecules)
per either unit mass or unit surface area of the adsorbent.
As an example, in Fig. 1.5 the equilibrium data for CO adsorbed at T = 303 K on
Na− and K−MFI zeolites, are reported as volumetric (Fig. 1.5a) and calorimetric
(Fig. 1.5b) isotherms. In volumetric isotherms the adsorbed amounts (nads ), in calori-
1 Fundamentals in Adsorption at the Solid-Gas Interface 9

Fig. 1.4 Schematic illustration of the first and second layer of adsorption; the multi-layers adsorp-
tion approaches a liquid-like phase

(a) 0.4 (b) 15

0.3
nads (mmol/g)

10
Qint (J/g)

0.2
5
0.1

0.0 0
0 30 60 90 0 30 60 90

pCO (Torr) pCO (Torr)

Fig. 1.5 Adsorption of CO at T = 303 K on Na–MFI (square) and K–MFI (circle) zeolites out-
gassed at T = 673 K. a Volumetric isotherms (adsorbed amounts vs. equilibrium pressure). b
Calorimetric isotherms (evolved heats vs. equilibrium pressure). Solid symbols first run, open sym-
bols second run of adsorption. Experimental points interpolated by the Langmuir equation (vide
infra)

metric isotherms the integral heat (Qint ) evolved during the process are plotted as a
function of the CO equilibrium pressure (pCO ), expressed in Torr (1 Torr = 133.3 Pa).
In this particular case, the first (1st) and second (2nd) run isotherms are virtually
coincident indicating that CO adsorption was entirely reversible upon evacuation of
the CO equilibrium pressure. For the experimental and samples details vide infra
Sect. 1.4. It is here only recalled that the 2nd run isotherms were performed after the
overnight outgassing of the reversible adsorbed phase. The isotherms experimental
points were interpolated by the Langmuir model equation (vide infra).
As an example of the influence of the adsorption temperature, the equilibrium data
for water (H2 O) adsorbed at T = 303, 353 and 423 K on a H-BEA zeolite specimen
(outgassed at T = 873 K) are illustrated in Fig. 1.6. In Fig. 1.6a the three volumetric
isotherms are reported (for experimental and samples details vide infra Sect. 1.4): as
far as the equilibrium pressure pH2O increases, the adsorbed amounts also increase
more or less steeply according to the adsorption temperature. In Fig. 1.6b the amounts
adsorbed at a constant equilibrium pressure (pH2O = 6 Torr) are plotted against the
adsorption temperature giving rise to an adsorption isobar (nads vs. Tads ). Note that
10 V. Bolis

(a)
nads (mmol/g) 6 (b) 6

nads (mmol/g)
4 4

2 2

0 0
0 2 4 6 300 350 400 450
pH2O (Torr) Tads (K)

Fig. 1.6 a 1st run adsorption volumetric isotherms of H2 O on H-BEA zeolite at T = 303 K
(square), T = 353 K (circle) and at T = 423 K (triangle). Experimental points interpolated by the
Freundlich equation (vide infra). b Adsorption isobar (nads vs. Tads ) at constant pH2O = 6 Torr;
best fitting by exponential decay

in this case the adsorption process was depressed by the increasing temperature,
according to the enhanced mobility of the adsorptive molecules. In all isotherms,
the experimental points were interpolated by using the Freundlich isotherm equation
(vide infra).
The isotherm experimental points can be interpolated by a variety of equations
according to the mechanism of the adsorption process, which in turn depends on
the nature of the gas/solid interaction. Among the different equations proposed to
describe quantitatively the isotherms, the only one based on a physical model is the
Langmuir equation [1, 2, 30, 32, 46].
The Langmuir model assumes a dynamic equilibrium at constant T between the
gas (at pressure p) and the adsorbed layer, and requires a number of well-defined
conditions: (i) the adsorption, which is limited to a monolayer, takes place at a surface
consisting of a distribution of energetically equivalent, non-interacting sites; (ii) the
ability of a molecule to bind at a site is independent of whether or not a nearby
site is occupied (absence of lateral interactions); (iii) once adsorbed, the molecules
are localized in that the activation barrier hindering migration to an adjacent site is
supposed to be much larger than kT ; (iv) the enthalpy of adsorption a H (per site)
is constant with θ.
The Langmuir adsorption isotherm is derived from a kinetic mechanism. Let us
assume for sake of simplicity that the molecule M is adsorbed molecularly (i.e.
without rupture/formation of chemical bonds) from the gas at a surface site S. The
fractional monolayer coverage of the sites occupied by adsorbate molecules is θ =
NS
N . The rate of adsorption is given by the Eq. 1.1:

adsorption rate = ka p (1 − θ) (1.1)

ka being the rate constant for the adsorption and (1 − θ) the fractional monolayer
coverage of sites not occupied yet by the adsorbate molecules.
1 Fundamentals in Adsorption at the Solid-Gas Interface 11

The rate of desorption, kd being the rate constant for desorption, is given by the
Eq. 1.2:
desorption rate = kd θ (1.2)

When the dynamic equilibrium is reached (adsorption rate = desorption rate) the
Eq. 1.3 is obtained:
ka p (1 − θ) = kd θ (1.3)

Equation 1.4 represents the Langmuir equation:

θ
= Kp (1.4)
1−θ

Note that the constant K is obtained by the

 ratio of the rate constant for adsorption
ka
over the rate constant for desorption kd .
The Langmuir equation is often written as reported by the Eq. 1.5

V Kp
θ= = (1.5)
Vmon (1 + Kp)

The term V represents the adsorbate volume and Vmon the monolayer volume, i.e.
the volume of adsorbate required to complete the monolayer.
At very low pressure the equation reduces to a linear dependence of the coverage
upon the equilibrium pressure (θ = hp). Conversely, at high pressure the equation
reduces to the case of coverage approaching the monolayer (θ ≈1).
The monolayer coverage (Vmon ) is hardly determined experimentally with accu-
racy. So, for practical purposes the Langmuir equation is suitably transformed in the
so-called reciprocal linear form, as illustrated by the Eq. 1.6:
   
1 1 1 1
= + (1.6)
V K Vmon p Vmon

For isotherms obeying the Langmuir model, the reciprocal volume V1 against recip-
 
rocal pressure 1p plot is linear (Langmuir-type isotherms). Conversely, if the exper-
imental data plot is not linear, Langmuir equation does not hold in describing the
given adsorption process.
The monolayer capacity is obtained from the intercept i = Vmon 1
of the straight
line. Once determined Vmon , the equilibrium constant K is obtained by the slope
s = K V1mon of the plot.
The monolayer volume and the equilibrium constant are typical of the adsor-
bent/adsorbate pairs at a given temperature. In particular, the value of K is bound to
the strength of the adsorbent-adsorbate interaction: high values of K indicate large
strength, low values little strength.
12 V. Bolis

(a) 0.4 (b) 0.5

0.4
0.3
θ = NS /N

θ/1− θ
0.3
0.2
0.2
0.1
0.1

0.0 0.0
0 30 60 90 0 30 60 90
pCO (Torr) pCO (Torr)

Fig. 1.7 Adsorption of CO at T = 303 K on dehydrated zeolites Na–MFI (square) and K–MFI
(circle). Solid symbols 1st run, open symbols 2nd run of adsorption. a θp versus pCO plot. b
θ
1−θ versus pCO plot

In the following, the very simple case of CO adsorbed at T = 303 K on dehy-

drated Na− and K−MFI will be discussed (vide supra in Fig. 1.5 the experimental
volumetric and calorimetric isotherms). The number of CO molecules adsorbed per
gram of zeolite at pCO represents the number of occupied sites (NS ), whereas the
number of charge-balancing cations exposed per gram of zeolite represents the total
available sites (N).
θ
In Fig. 1.7, the coverage θ = NNS (Fig. 1.7a) and the 1−θ quantity (Fig. 1.7b) are
θ
plotted against pCO . The slope of the 1−θ versus pCO plot is the Langmuir constant
K (see Eq. 1.4).
CO is a soft Lewis base which is easily polarized by the electrostatic field generated
by the extra-framework alkaline-metal cations located in the MFI zeolite nanocavi-
ties. As a consequence, it is reversibly taken up by the surface when put in contact
with the activated zeolite [23].
The equilibrium constant K for Na–MFI (4.88 ± 0.02 10−3 Torr −1 ) is larger than
for K–MFI (1.15 ± 0.02 10−3 Torr −1 ), in agreement with the different polarizing
power of the cations. In fact, the local electric field generated by the coordinatively
unsaturated (cus) cations depends on the charge/ionic radius ratio, which is larger
for Na+ than for K+ , the ionic radius of the former being 0.97 Å and that of the latter
1.33 Å [47]. Note also that, according to the charge/ionic radius ratio, the maximum
coverage attained at pCO = 90 Torr was larger for Na–MFI (θ ≈0.3) than for K–MFI
(θ ≈0.1).
The standard free energy a G◦ for CO adsorption at the two alkaline-metal sites
is obtained by the Langmuir equilibrium constant K by employing the Eq. 1.7:

a G◦ = −RT ln K (1.7)

In both cases the adsorption process in standard conditions is endoergonic, being

a G◦ = +13.4 kJ mol−1 for Na−MFI and +17.0 kJ mol−1 for K−MFI. The endo-
ergonic character of the process is witnessed by the non-spontaneity of the adsorption
1 Fundamentals in Adsorption at the Solid-Gas Interface 13

unless a CO pressure is applied. In fact, by evacuating the CO pressure, the electro-

static Na+ · · · CO and K+ · · · CO adspecies are completely destroyed, as confirmed
by the overlap of the 1st and 2nd run of adsorption (see Figs. 1.5 and 1.7).
From a G◦ the standard entropy of adsorption a S ◦ is obtained, if a H ◦ is
known. The CO adsorption enthalpy change was measured calorimetrically during
the same experiments in which the adsorbed amounts were measured (vide infra,
Sect. 1.4.2.3). We will come back to this point and to the evaluation of the a S ◦ in
the section devoted to the entropy of adsorption (vide infra, Sect. 1.5.3).
Deviations from the Langmuir model are often observed in real systems. The
Langmuir model assumptions listed above are indeed very limitative and severe: (i)
the solid surface is rarely uniform: there are always “imperfections” at the surface,
(ii) the mechanism of adsorption is not the same for the first molecules as for the
last to adsorb. When two or more kind of sites characterized by different adsorption
energies are present at the surface (as stated in point i), and when lateral interactions
among adsorbed species occur (as stated in point ii), the equivalence/independence
of adsorption sites assumption fails. The most energetic sites are expected to be
occupied first, and the adsorption enthalpy a H (per site) instead of keeping a
constant, coverage-independent value, exhibits a declining trend as far as the coverage
θ increases.
Further, the adsorbed molecules are not necessarily inert, and on the top of the
monolayer other molecules may adsorb and multi-layers build up: this is properly
described by the Brunauer, Emmet and Teller (BET) model [2, 30, 32].
Freundlich and Temkin isotherms, which refer to the case of the adsorption at sur-
faces characterized by a heterogeneous distribution of active sites, will be discussed
briefly [30].
Freundlich isotherm is mathematically expressed by the Eq. 1.8:

Vads = kp1/n (1.8)

This is a purely empirical formula, where the term Vads represents the adsorbed
amount, p the adsorptive pressure, whereas k and n are suitable empirical constants
for a given adsorbent-adsorbate pair at temperature T . The adsorbed amount are
normalized either to the mass of the adsorbent or to the exposed surface area. As
an example see. Fig. 1.6a, where the adsorption volumetric isotherms of H2 O on
H−BEA zeolite are reported: the experimental points were interpolated by the Fre-
undlich isotherm equation. The Freundlich isotherm assumes that the adsorption
enthalpy a H (per site) varies exponentially with increasing equilibrium pressure.
In fact, the experimental points in the correspondent heat of adsorption versus cover-
age plot were properly interpolated by an exponential fitting, as illustrated in Fig. 1.15
(vide infra Sect. 1.4.2.3).
Temkin isotherm is mathematically expressed by the Eq. 1.9:

Vads = k1 ln (k2 p) (1.9)

14 V. Bolis

Temkin equation too is a purely empirical formula, where Vads represents the
adsorbed amount and p the adsorptive pressure; k1 and k2 are suitable empirical
constants for a given adsorbent-adsorbate pair at temperature T . Also in this case,
the adsorbed amount are normalized either to the mass of the adsorbent or to the
exposed surface area.
The Temkin isotherm assumes that the adsorption enthalpy a H (per site)
decreases linearly upon increasing coverage. Examples of heats of adsorption
decreasing linearly with coverage are reported in the literature, as for instance in
the case of NH3 adsorbed on hydroxylated silica, either crystalline, [48] or amor-
phous, [49] as well as in the case of CH3 OH adsorption on silica-based materials
[26].
Further, it is worth noticing that at sufficiently low pressure all adsorption
isotherms are linear and may be regarded as obeying the Henry’s law, which is
reported in Eq. 1.10:
Vads = h p (1.10)

The Henry constant h is typical of the individual adsorbate-adsorbent pair, and is

obtained by the slope of the straight line representing the isotherm at low coverage.
The isotherms classification, which is of high merit in terms of generality, deals
with ideal cases which in practical work are rarely encountered. In fact, most often
the adsorption process over the whole interval of pressure is described by an exper-
imental isotherm which does not fit into the classification. Nonetheless, each of the
equations described above may be used over restricted ranges of equilibrium pres-
sure, so allowing to describe the experimental isotherm through the combination of
individual components to the process. In such a way the surface properties of the
solid, and the thermodynamics features of processes taking place at the interface can
be quantitatively described [30].
As an example, it is here mentioned that the adsorption NH3 on a highly dehy-
drated silica specimen was satisfactorily described by the combination of the Lang-
muir and Henry isotherms. The former accounted for H−bonding interactions on
isolated silanols (Si-OH), whereas the latter accounted for the aspecific adsorption
on dehydrated patches of the surface, dominated by dispersion forces interactions
[28].

1.4 Adsorption Microcalorimetry

The measurement of the heat of adsorption by a suitable calorimeter is the most reli-
able method for evaluating the strength of adsorption (either physical or chemical).
Tian-Calvet heat-flow microcalorimeters are an example of high sensitivity apparatus
which are suitably adapted to the study of gas-solid interactions when connected to
sensitive volumetric systems [10–14, 50–55]. Volumetric-calorimetric data reported
in the following were measured by means of either a C-80 or MS standard heat-flow
microcalorimeter (both by Setaram, F), connected to a high vacuum (residual pressure
1 Fundamentals in Adsorption at the Solid-Gas Interface 15

p ≤ 10−5 Torr) gas-volumetric glass apparatus. During the same experiment, both
integral heats evolved and adsorbed amounts were measured for small increments of
the adsorptive, from the gas or vapor phase. Two identical calorimetric vessels, one
containing the sample under investigation, the other (usually empty) serving as refer-
ence element were connected in opposition. Thanks to the differential construction of
the apparatus, all parasitic phenomena (i.e., all thermal effects other than the one due
to the interaction of the gas with the solid surface) were successfully compensated.
C-80 microcalorimeters allow the heats of adsorption to be measured at constant T
in the room temperature −573 K range, whereas MS standard microcalorimeters in
the room temperature −473 K range. The adsorptive pressure in the measurements
were monitored by either a Varian Ceramicell or a Baratron MKS transducer gauge
(0 − 100 Torr).
A well-established stepwise procedure was followed [16, 23, 25, 56]. Small suc-
cessive doses of the adsorptive were admitted and left in contact with the adsorbent
until the thermal equilibrium was attained. The 1st run of adsorption performed
on the activated sample (pretreated in high vacuum conditions and/or in controlled
atmosphere) will be hereafter referred to as ads. I. At any individual dose of gas
introduced in the system, the evolved heat Qint was measured within the calori-
metric cells, while the adsorbed amount nads was measured by volumetry. Ads. I
was followed by a desorption run (des. I), performed by simple evacuation of the
cell. In such a way the reversibly adsorbed phase was desorbed and either the pris-
tine surface was restored, in case of an entirely reversible adsorption, or the pristine
surface was not recovered, in case of a (partially) irreversible adsorption. Ads. II
was subsequently performed in order to assess which fraction (if any) of the pristine
surface sites was irreversibly occupied by the adsorbed phase (in the adopted con-
ditions). By subtracting the ads. II curve from the ads. I one, the adsorbed fraction
not removed by evacuation is evaluated. The ads. II component will be hereafter
referred to as the reversible adsorbed phase, whereas the (ads. I - ads. II) component
will be referred to as the irreversible phase (in the adopted conditions). Subsequent
runs of adsorption (ads. III, IV etc.) are performed in some cases, if the irreversible
modification of the surface is expected/suspected not to be extinguished during the
ads. I [21, 23, 26]. Adsorption measurements are usually performed at least twice on
a virgin portion of the same batch of the material, activated in the same conditions,
to check the experiments reproducibility. The routinely run protocol of adsorption-
desorption-adsorption cycles is schematically illustrated in Fig. 1.8.

1.4.1 Materials

Before illustrating an instructive selection of experimental data obtained by the

method described above, it is worth doing to report a brief description of the investi-
gated materials. Some data have been already published (as will be reported), other
are original.
16 V. Bolis

Fig. 1.8 Schematic illustration of the adsorption-desorption-adsorption cycle routinely run in order
to collect the equilibrium data (evolved heats and adsorbed amounts) which are measured within
the calorimetric cells at increasing equilibrium pressure

1.4.1.1 Solid Materials

H-BEA (H-BETA): a proton exchanged BEA zeolite specimen, characterized by a

three-dimensional network of pores consisting of three families of 12-ring intercon-
nected channels [57]. The specimen here illustrated was characterized by a silica-
SiO2
to-alumina ratio Al 2 O3
= 4.9, corresponding to a distribution of Al atoms per unit
cell uc = 5.9 (see Ref. [25]. The acidic strength of such material is related to the
Al

presence of both Brønsted and Lewis acidic sites. Such latter kind of sites gives
H−BEA zeolites unique catalytic properties [58–60].
BEA (BETA): an all-silica BEA zeolite specimen, characterized by the same
three-dimensional network as H−BEA, but virtually free of Al species in that the
SiO2
silica-to-alumina ratio was Al 2 O3
= 255, corresponding to a negligible distribution
of Al atoms per unit cell uc  0.1 (see Ref. [25, 61]).
Al

Prior to the adsorption experiments, H−BEA sample was outgassed for 2 h at T =

873 K, a temperature which ensured a maximum surface dehydration, still compatible
with the stability of the structure, and yielding the maximum density of Brønsted
and Lewis acidic sites. For the all-silica specimen, which was less hydrophilic than
the proton-exchanged counterpart, a 2h-outgassing at T = 673 K was sufficient to
get rid of all adsorbed water, so yielding the maximum density of Si−OH hydroxyl
nests.
1 Fundamentals in Adsorption at the Solid-Gas Interface 17

H−MFI (H−ZSM5): a proton exchanged MFI zeolite specimen, characterized

by a three-dimensional network of pores consisting of sinusoidal and intersecting
straight 10-ring channels [35]. The specimen here illustrated was characterized by a
SiO2
silica-to-alumina ratio Al 2 O3
= 7.5, corresponding to a distribution of Al atoms per
unit cell uc = 6.0, very close to that of the H-BEA specimen illustrated above [25].
Al

The acidic strength of H–MFI materials is related to the presence of Brønsted

acidic sites whereas, opposite to H-BEA zeolites, Lewis acidic sites represent only
a minor feature.
MFI–Silicalite: a Na- and Al-free defective all-silica MFI specimen, characterized
by the same three-dimensional network as H–MFI, but with an extremely large silica-
SiO2
to-alumina ratio Al 2 O3
→ ∞. See Ref. [25] for details.
The all-silica zeolites, both BEA and MFI, are in general characterized by a vari-
able amount of internal defects consisting of hydroxyl nests made up of H–bonding
interacting Si-OH species, located within the zeolite nanopores. A virtually perfect
(i.e. defect-free) MFI–Silicalite was also investigated for comparison purposes. See
Ref. [24] and references therein for details on both defective and perfect all-silica
specimens.
Prior to the adsorption experiments, both H–MFI and MFI samples were outgassed
for 2 h at T = 673 K in order to achieve the maximum dehydration of the surface
compatible with the stability of the structure and yielding the maximum density of
Brønsted acidic sites in H–MFI and of polar Si-OH hydroxyl nests in MFI–Silicalite.
Me(I)−MFI (Me(I)−ZSM5): cation-exchanged zeolites (MFI) with Me(I) =
Cu+ or Ag+ (both belonging to the group 11 of transition metals), or Me(I) = Na+
or K+ (both belonging to the group 1 of alkaline-metals) as extra-framework species.
The samples were prepared starting from the same NH4 –MFI precursor (character-
SiO2
ized by a silica-to-alumina ratio Al 2 O3
= 7) either by conventional wet exchange
(Ag(I)−, Na− and K−MFI), or by direct CuCl gas phase exchange (Cu(I)−MFI). In
all cases a nearly total exchange of the parent material extra-framework cations was
achieved, as confirmed by IR spectroscopy: one Me(I) cation for every framework
Al atom was present in all examined materials. For samples details see Ref. [21, 23,
62]
TiO2 : (a) a crystallographic pure anatase obtained by a sulphate preparation and
thoroughly freed from sulphate impurities, following the preparation route described
in Ref. [63]; (b) the same crystallographic pure anatase but still carrying sulphate
surface impurities (4–5 % SO4 by weight) [64]. Prior to the adsorption experiments
all samples were outgassed at T = 673 K for 2 h, in order to dehydrate the surface
and yield the maximum density of Lewis acidic sites (i.e. cus Ti4+ cations). After
outgassing, the samples were contacted at the same temperature with ≈100 Torr of
O2 for 10 min to ensure stoichiometry.
A Ca-modified silica (8% mol CaO) was obtained by adding dosed amounts of an
aqueous solution of Ca(NO3 )2 ·4H2 O to a dry amorphous nonporous silica (Aerosil
200 from Degussa, Frankfurt A.M., D) using the incipient-wetness impregnation
technique described in Refs. [26, 65] Prior to the adsorption experiments the sample
was outgassed at T = 423 K for 2 h, in order to get rid of physically adsorbed water but
18 V. Bolis

without inducing any appreciable surface dehydroxylation. The choice of a vacuum

activation temperature only slightly higher than room temperature was determined
by the need to study the surface properties of still highly hydrated samples (i.e. of
solids taken under conditions not too far from those experienced by biomaterials in
contact with the biological medium).

1.4.1.2 Molecular Probes

CO specpure from either Matheson or Praxair was used as a molecular probe in order
to assess the Lewis acidic properties of coordinatively unsaturated (cus) cations either
located in the dehydrated zeolite nanocavities as charge-balancing cations, or exposed
at the dehydrated surface of oxidic materials. CO is capable of interacting with the
cus cations leading to the formation of adducts of different stability according to
the chemical nature of the cation. Weak electrostatic adducts are formed on alkaline
metal cations, σ-coordinated species of intermediate stability on non d/d0 metal
cations, whereas high-stability carbonyl-like species originated by a σ-coordination
+π-back donation of d electrons are formed on d block metal cations.
H2 Ovap was used as a molecular probe to assess the hydrophilic and/or hydropho-
bic features of protonic (H−BEA) and all-silica (BEA) zeolites. Water (from Milli-
pore) was distilled several times in vacuo and rendered gas-free by several ‘freeze-
pump-thaw’ cycles. The vapor pressure of H2 O at T = 303 K is 31.8 Torr, and the
standard molar enthalpy of liquefaction (i.e. the latent heat of liquefaction, qL ) is
−L H ◦ = 44 kJ mol−1 .
NH3 gas (from Praxair) was used as molecular probe of moderate basic strength
(PA = 854 kJ mol−1 ), [66]. in order to characterize the acidic strength of Brønsted
(and Lewis, if any) acidic sites in protonic and all-silica zeolites.
CH3 OH vapor, obtained by distilling in vacuo liquid methanol (Sigma-Aldrich),
was rendered gas-free by several “freeze-pump-thaw” cycles. The vapor pressure of
CH3 OH at T = 303 K is 164 Torr, and the standard molar enthalpy of liquefaction
(i.e. the latent heat of liquefaction, qL ) is −L H ◦ = 38 kJ mol−1 .

1.4.2 Equilibrium Data

1.4.2.1 Volumetric and Calorimetric Isotherms

Adsorbed amounts and integral heat evolved will be suitably reported as a function
of the increasing equilibrium pressure, i.e.

as volumetric and calorimetric isotherms,
respectively. Adsorbed amounts nads = nads were obtained by adding the indi-
vidual doses amounts, nads , and will be reported either as mol per unit mass
(mol g−1 ) or per unit surface area (mol m−2 ), or as molecules per square nanometer.
In zeolites, in order to compare from a structural point of view the affinity of different
zeolites towards the given adsorptive, the adsorbed amounts will be more suitably
1 Fundamentals in Adsorption at the Solid-Gas Interface 19

(a)
nads (mmol/g) 6 (b) 450

4 300

Q int (J/g)
2 150

0 0
0 2 4 6 0 2 4 6
pH2O (Torr) pH2O (Torr)

Fig. 1.9 Adsorption of H2 Ovap adsorbed at T = 303 K on proton-exchanged (H-BEA, square)

and all-silica (BEA, up triangle) zeolites pre-outgassed at T = 873 and 673 K, respectively. a
Volumetric isotherms. b Calorimetric isotherms. Solid symbols ads. I; open symbols ads. II. Adapted
from Ref. [25] Fig.4

reported as molecules per unit cell (uc) or per Al atom. Integral heats Qint = Qint
were obtained by adding the individual doses evolved heats, Qint , and will be
reported per gram (J g−1 ) or per unit surface area (J m−2 ) of the adsorbent. Integral
heats plotted as a function of the adsorbed amounts will be referred to as the integral
heat curve: Qint versus nads .
A selection of adsorption isotherms obtained for a variety of materials and probe
molecules will be illustrated. Note that all the adsorption measurements reported in
the following were performed at T = 303 K.
H2 Ovap adsorbed on H–BEA and all-silica BEA zeolites. In Fig. 1.9, ads. I and
ads. II volumetric (section a) and calorimetric (section b) isotherms of H2 Ovap
adsorbed on proton-exchanged (H–BEA) and all-silica (BEA) zeolites are reported.
In H−BEA an irreversible adsorption component was revealed by the non-coincidence
of the ads. I and ads. II isotherms (both volumetric and calorimetric). Conversely, in
the all-silica case the process was entirely reversible upon evacuation of the vapor
phase, as witnessed by the coincidence of ads. I and ads. II isotherms. The isotherms
experimental points were interpolated by the Freundlich equation.
H−BEA exhibited, as expected, a much higher affinity towards H2 O than the
Al−free systems, owing to the presence of Si(OH)+ Al− species, characteristic of
proton-exchanged zeolites and acting as Brønsted acidic sites. Such species are
located within the zeolite nanocavities and are able to adsorb guest molecules by
strong H–bonding interactions, often leading to the formation of protonated species
[25, 67–73]. In addition, in H−BEA zeolites structural defects acting as Lewis acidic
sites (i.e. strong electron acceptors) are often present [58, 60, 74–76]. It is still
under debate whether such species consist of framework trigonal Al (III) atoms,
[73, 77, 78] or of extra-framework Al (III) species (EFAL) located within the pores
[75, 76]. Anyway, both Lewis and Brønsted acidic sites are responsible for the for-
mation of water complexes which are stable upon room temperature evacuation.
20 V. Bolis

(a) 2.0 (b) 150

1.5 120
nads (mmol/g)

Q (J/g)
90
1.0

int
60
0.5
30

0.0 0
0 30 60 90 0 30 60 90
pCO (Torr) pCO (Torr)

Fig. 1.10 CO adsorbed at T = 303 K on zeolites Cu(I)–MFI (diamond) and Na–MFI (square):
volumetric (a) and calorimetric (b) isotherms. Both samples were pre-outgassed at T = 673 K.
Solid symbol ads. I; open symbols ads. II. Volumetric isotherms: adapted from Ref. [23] Fig. 3a

The affinity towards water of the all-silica counterpart was lower than that of
the proton exchanged zeolite, as expected, but it was not negligible. The reported
isotherms indicated that hydrophilic sites, responsible for weak and reversible water
H–bonding adducts, are developed in zeolites even in the absence of framework Al
atoms. Structural defects generating polar species consisting of Si−OH nests (which
are characterized by a weak Brønsted acidic strength),[25, 61] are always present
in Al-free zeolites, unless especially prepared in order to obtain hydrophobic, inert
materials, as claimed by Flanigen et al.[36]. See also Ref. [24].
CO adsorbed on Me(I)-exchanged MFI zeolites. In Fig. 1.10 the ads. I and ads.
II volumetric (section a) and calorimetric (section b) isotherms of CO adsorbed on
Cu(I)− and Na−MFI are reported. In Fig. 1.11 ads. I and ads. II volumetric (section
a) and calorimetric isotherms (section b) of CO adsorbed on Ag(I)− and K−MFI
are reported. Note that the ordinate scale of the isotherms plots for Ag(I)− and
K−MFI is twice as large as the Cu(I)− and Na−MFI scale, owing to the much
lower adsorption capacity of the Ag(I)− and K−MFI zeolites with respect to the
Cu(I)− and Na−MFI ones.
Cu(I)− and Ag(I)−MFI ads. II isotherms (both volumetric and calorimetric) lie
below the ads. I correspondent isotherms, indicating the presence of irreversible
phenomena. The irreversible adsorption component was quantified by taking the
(ads. I – ads. II) difference in the volumetric isotherms at pCO = 90 Torr. It was
≈30 % of total uptake (ads. I) for copper- and ≈20 % for silver-exchanged zeolites.
Conversely, in both Na− and K–MFI cases the coincidence of ads. I and ads. II
isotherms confirmed the reversibility of CO adsorption. The much lower adsorption
capacity witnessed by the group 1 metal cations volumetric isotherms with respect
to the two group 11 ones is striking. The calorimetric isotherms too confirmed the
much lower affinity of the former with respect to the latter. The isotherms exper-
imental points of the group 1 metals exchanged zeolites were interpolated by the
Langmuir equation (vide supra, Sect. 1.3.1), whereas in the case of the group 11
metals isotherms the curves were drawn by employing a B-Spline function, just as
an aid to the eye. The two d-block metals isotherms dramatically deviate from the
1 Fundamentals in Adsorption at the Solid-Gas Interface 21

(a)
nads (mmol/g) 1.0 (b) 75

Q int (J/g)
50
0.5

25

0.0 0
0 30 60 90 0 30 60 90
pCO (Torr) pCO (Torr)

Fig. 1.11 CO adsorbed at T = 303 K on zeolites Ag(I)–MFI (triangle) and K–MFI (circle):
volumetric (a) and calorimetric (b) isotherms. K–MFI was pre-outgassed at T = 673 K, Ag(I)–
MFI at T = 400 K. Solid symbol ads. I; open symbols ads. II. Volumetric isotherms: adapted from
Ref. [23], Fig. 3b

Langmuir behavior, in that in this latter case the adsorption of CO was driven by
forces other than the simple electrostatic polarization, and concerns a heterogeneous
distribution of active sites. In fact, Cu(I)− and Ag(I)−cations hosted in the zeolite
nanocavities, besides the electrostatic polarization of CO molecule, interacted chem-
ically with the molecule. Stable carbonyl-like species were formed at the d-block
metal cations sites, through a σ-coordination of the C− end lone pair plus a partial
π-back-donation of d electrons. The stoichiometry of the heterogeneous di-carbonyl
 +  +
Cu (CO)2 and mono-carbonyl Ag (CO) complexes was obtained from the
quantitative data reported in Figs. 1.10 and 1.11. These results, confirmed also by
IR spectroscopic data (see Ref. [23] and references therein for details), are in good
agreement with the stoichiometry reported for the correspondent Cu(I) and Ag(I)
complexes formed in homogeneous conditions [79, 80].
The differences between the two d-block metal cations can be explained on one
hand from an electrostatic point of view, since the charge density of Cu(I) is much
larger than that of Ag(I) cations (rCu(I) = 0.96 Å and rAg(I) = 1.26 Å) [47]. On
the other hand, the overlap of the metal cations and CO orbitals in the carbonyl bond
is expected to be larger for Cu(I) than for Ag(I). The adsorption of CO on Na+ and
K+ cations hosted in the same zeolite framework allowed to roughly single out the
electrostatic contribution to the two d-block metal cations/CO interaction. Na+ and
K+ cations possess indeed a charge/radius ratio very close to that of Cu(I) and Ag(I),
respectively (0.97 Å for Na+ and 1.33 Å for K+ ) [47].
NH3 adsorbed on H−MFI and all-silica MFI zeolites. Adsorption of NH3 has
been widely used to assess the acidic strength of both Brønsted and Lewis sites at
solid surfaces [68, 70, 71]. In Fig. 1.12, the ads. I and ads. II volumetric (section a)
and calorimetric (section b) isotherms of NH3 adsorbed on: (i) the proton-exchanged
H−MFI zeolite, (ii) one defective MFI–Silicalite (MFI−def) and (iii) the perfect
(defect-free) all-silica MFI–Silicalite (MFI−perf), are reported. In H−MFI, the
adsorption was only partially reversible in agreement with the proton-transfer from
the Brønsted acidic site Si(OH)+ Al− to NH3 , as reported in Ref. [81]. Conversely,
22 V. Bolis

(a)
nads (mmol/g) 6 (b) 300

4 200

Q int (J/g)
2 100

0 0
0 30 60 90 0 30 60 90
pNH3 (Torr) pNH3 (Torr)

Fig. 1.12 Adsorption of NH3 at T = 303 K on proton-exchanged H–MFI zeolite (square), defec-
tive MFI–Silicalite (MFI−def, diamond), and perfect (defect-free) MFI–Silicalite (MFI−perf, up
triangle). a Volumetric isotherms. b Calorimetric isotherms. H−MFI zeolite was pre-outgassed at
T = 873 K, MFI–Silicalite samples at T = 673 K. Solid symbols ads. I, open symbols ads. II

in both defective and perfect all-silica zeolites the process was entirely reversible
upon evacuation of the gas phase. In this latter case NH3 interacted only via hydro-
gen bond with Si−OH nests [24]. In the defect-free MFI–Silicalite, which exposes
only unreactive siloxane bridges, the interaction was aspecific in that governed by
dispersion forces due to the nanopores walls (confinement effect) [82–84].

1.4.2.2 Integral Heat of Adsorption

Integral heats normalized to the adsorbed amounts are referred to asthe integral
 molar
Qint
heat of adsorption at the given equilibrium pressure p: (qmol )p = nads expressed
p
in kJ mol−1 . The (qmol )p quantity is an intrinsically average value, as it refers to
the thermal response of the surface as a whole, and is comprehensive of all thermal
contributions from the variety of interactions the gas molecules have experienced
from the beginning of the process up to the chosen equilibrium pressure p.
By plotting the integral heats evolved against the adsorbed amounts the so-called
integral heats curve is obtained (vide infra as an example the insets of Fig. 1.14). In
the Langmuir-like adsorption characterized by a uniform distribution of equivalent,
non-interacting sites the heat of adsorption is constant upon increasing coverage: the
integral heats curve is thus a straight line through the origin, the slope of which gives
the differential heat of adsorption (qdiff ). In cases other than this particular one, qdiff
is obtained by differentiating the non-linear Qint = f (nads ) function.
The interest in dealing with differential heats stems on the fact that the differential
quantities are more adequate than the average ones in describing the evolution with
the increasing coverage of the adsorbate/surface sites energy of interaction.
1 Fundamentals in Adsorption at the Solid-Gas Interface 23

1.4.2.3 Differential Heat of Adsorption

Differential heats of adsorption represent a reasonable measure of the energy of

interaction of a (probe) molecule with the individual sites, at any adsorbate cover-
age. The magnitude of the heat evolved during adsorption, which depends on the
nature of the adsorbate/surface sites bonding, varies upon increasing coverage as a
consequence of the presence of either a heterogeneous distribution of surface sites, or
lateral interactions among adsorbed species (vide infra Sect. 1.5.1). The shape of the
qdiff versus nads plots depends on, and actually describes, the surface heterogeneity.
δQint
Differential heats of adsorption are properly defined as qdiff = δn ads
, i.e., the deriv-
ative of the Qint = f (nads ) function which best fits the Qint versus nads equilibrium
data.
An alternative route for evaluating qdiff is however first discussed here. The method
Qint
is based on the use of the partial molar heats n ads
quantities,[10, 12, 85] i.e. the
ratio of the integral heat evolved over the correspondent amount adsorbed for the
Qint
individual incremental doses of the adsorptive. The n ads
quantity, expressed in
kJ mol−1 is still an integral molar heat (and thus average in nature) but it refers to the
thermal response of very small regions of the surface, provided that the individual
Qint
doses were prepared as small as possible. Note that the limit of n ads
quantity for
nads approaching zero is the true differential heat, as illustrated by the Eq. 1.11:

Qint
lim = qdiff (1.11)
nads →0 nads

Q int
By plotting, in the form of an histogram, n ads
values as a function of the adsorbed
amounts nads , the evolution of the heat of adsorption upon increasing coverage is
Qint
properly described. By taking the middle point of each partial molar heat n ads
block,
the mean heat values correspondent to small portions of the surface are obtained,
which represent a reasonable measure of the differential heat, as reported by several
authors [10, 14, 56, 85–88].
The qdiff experimental points versus the nads adsorbed amounts are interpolated
by functions which best fit the experimental points (vide infra in Fig. 1.13a the case
of H2 Ovap adsorbed on zeolites).
For most purposes, it is convenient to define the zero-coverage differential heat
of adsorption q0 , which corresponds to the energy of interaction of the molecular
probe with the most energetic sites, expected to be active in the earliest stages of the
adsorption process. The q0 value is estimated by extrapolating the qdiff versus nads
plot to vanishing coverage. The extrapolated quantities of experimental origin can
be properly and often fruitfully compared to the computed energy of interaction
of a probe molecule with an individual model site, as obtained through ab initio
calculations [24–26, 29, 73, 89].
24 V. Bolis

0 1 2 3 4
(a) 200 (b) 200
H 2O molecules/Al atom
160 160

qdiff (kJ/mol)
q diff (kJ/mol)

120 120

80 80

40 40
q L (H2O) qL (H2O)
0 0
0 6 12 18 24 0 2 4 6
H 2 O molecules/unit cell pH2O (Torr)

Fig. 1.13 a qdiff versus nads of H2 Ovap at T = 303 K on proton-exchanged (H−BEA, square) and
all-silica (BEA, up triangle) zeolites, pre-outgassed at either T = 873 K (H−BEA) or T = 673 K
(BEA). b qdiff versus pH2O . Solid symbols ads. I; open symbols ads. II. Adapted from Ref. [25],Fig. 5

H2 Ovap adsorbed on H − BEA and BEA zeolites. In Fig. 1.13 the differential heats
of adsorption of H2 Ovap on H−BEA and BEA zeolites are reported as a function of
water adsorbed amounts (Fig. 1.13a) or water equilibrium pressure (Fig. 1.13b).
Figure 1.13a curves are the exponential functions which best fitted the partial
molar heats experimental points. The qdiff values were plotted against the number of
H2 O molecules adsorbed per unit cell, so facilitating the interpretation of the results
in terms of structural features. As already outlined in describing the correspondent
volumetric isotherms (vide supra Fig 1.9), in H−BEA case water adsorption was only
partially reversible upon evacuation of pH2 O . The ads. I and ads. II qdiff versus nads
curves are well distinguished in the early stage of the process, namely up to the
adsorption of ≈1H2 O molecule per Al, i.e. per acidic (either Brønsted or Lewis) site.
Afterwards, the two curves merge and remain constant at a heat value larger than the
latent heat of liquefaction of water (qL = 44 kJ mol−1 ). The ads. I and ads. II curves
for the all-silica BEA zeolite are virtually coincident and rapidly approach qL .
In all cases, the qdiff versus nads curves are typical of heterogeneous surfaces. In
H−BEA several interactions, of different strength, take place on sites of different
nature located either within the zeolite nanocavities or at the external surface.
All interactions taking place simultaneously contributed to the calorimetrically
measured heat, and consequently it is hard to single out the energetics of the individual
contributions to the interaction. At the Brønsted Si(OH)+ Al− sites H2 O molecules
are either strongly H−bonded or protonated, whereas at the Lewis acidic sites, i.e. the
cus framework Al (III) cations, H2 O molecules are oxygen-down coordinated [25].
H2 O molecules interacted also via H−bonding with Si−OH nests located within the
zeolite nanocavities and, more weakly, with (isolated) Si−OH species exposed at
the external surface. Aspecific interactions generated by confinement effect,[82–84]
also contributed to the overall measured heat.
1 Fundamentals in Adsorption at the Solid-Gas Interface 25

(a) 150 (b) 150

Qint (J/g) Q int (J/g)
100 50
120 120
qdiff (kJ/mol)

50

qdiff (kJ/mol)
25
90 n ads (mmol/g) 90
0 n ads (mmol/g)
0 1 0
0.0 0.5
60 60

30 30

0 0
0.0 0.5 1.0 1.5 2.0 0.00 0.25 0.50 0.75 1.00
n ads (mmol/g) n ads (mmol/g)

Fig. 1.14 Differential heats of adsorption versus CO uptake (T = 303 K). a zeolites Cu(I)–MFI
(diamond) and Na–MFI (square). b zeolites Ag(I)–MFI (triangle) and K−MFI (circle). Insets:
interpolated integral heats of adsorption curves Qint versus nads . Zeolites Cu(I)−, Na− and K−MFI
were pre-outgassed at T = 673 K, Ag(I)–MFI at T = 400 K. Solid symbols: ads. I; open symbols
ads. II. Agreement between the experimental points (partial molar heats) and the derivative of the
integral heat curves is quite good (see the text for details). Adapted from Ref. [23], Fig.4

The heats of adsorption started from a quite high zero-coverage value (q0 ≈160 kJ
mol−1 ), which is compatible with a chemisorption process, either the protonation
of H2 O at the Brønsted acidic site or the strong oxygen-down coordination at the
Lewis acidic sites. According to the ab initio modeling results, indicating that the
H2 O/Lewis site energy of interaction is comprised in the 160-109 kJ mol−1 range
(depending on the local coordination of Al (III) atom),[25] the zero-coverage heat
of adsorption for H−BEA could be assigned to the H2 O/Lewis complex formation,
which dominated the early stage of the adsorption. At increasing coverage the heat
values decreased exponentially but remained well above the latent heat of liquefaction
of water even after the adsorption of ≈4 molecules per Al atom. For one H2 O
molecule adsorbed per Al atom, on average, the heat values were comprised in the
160 < qdiff < 80 kJ mol−1 range, whereas for the second-to-fourth H2 O adsorbed
molecules in the 80 < qdiff < 60 kJ mol−1 range.
In the all-silica BEA specimen the zero-coverage heats of adsorption were much
lower than for H–BEA (q0 ≈70 vs.160 kJ mol−1 , respectively) and both ads. I and
ads. II curves dropped very fast down to an almost constant value, only slightly
higher than qL = 44 kJ mol−1 . This result indicates that, despite the absence of spe-
cific Brønsted/Lewis acidic sites, the Si−OH nests manifest a H−bonding capacity
of medium strength giving to the all-silica zeolite a substantial non-hydrophobic
character.
Figure 1.14b plot indicates that for adsorption leading to the same residual pressure
in different runs and/or on different samples, qdiff values virtually coincide. Differ-
ences between the protonic H−BEA and the all-silica BEA zeolites (and between
ads. I and ads. II for H−BEA) are evident at a residual pressure close to zero, at
which the irreversible adsorption took place only on H−BEA.
Two examples of qdiff obtained by differentiating the Qint = f (nads ) functions
best fitting the Qint versus nads experimental points will be discussed in the following.
26 V. Bolis

CO adsorbed on Me(I)−exchanged MFI zeolites. In Fig. 1.14 the differential heat

of adsorption of CO on Cu(I)− and Na−MFI (Fig. 1.14a) and on Ag(I)− and K−MFI
(Fig. 1.14b) zeolites are reported as a function of the uptake. The differential heat
δQint
of adsorption curves are the analytical derivative qdiff = δn ads
of the integral heats
plots which are reported in the inset of the figures. In both Cu(I)− and Ag(I)−MFI
cases the integral heats curves were reasonably fitted by a polynomial of order five,
whereas in the case of alkaline metal cations by a linear equation (in agreement with
the Langmuirian behavior of such systems, vide supra Sect. 1.3.1). The experimental
points reported in the qdiff versus nads plot were obtained by taking the middle points
Qint
of the partial molar heats n ads
histogram. A reasonably good agreement exists
between the two methods over the whole examined adsorption range. From both
Fig. 1.14 plots it is clearly evident that at least a fraction of CO adspecies formed at
the Cu(I) and Ag(I) sites are stable complexes.
The ads. I q0 was estimated as high as ≈120 and ≈100 kJ mol−1 for Cu(I) and
Ag(I) sites, respectively. Such values are compatible with the bond energies typical
of a chemisorption process, i.e. the formation of stable carbonyl-like species. The
value for the early formation of the reversible carbonyl-like species (ads. II) was
lower (≈90 kJ mol−1 ) than for ads. I and very close for the two systems. As far as
the coverage increased, the heat decreased to values typical of labile species and
eventually fell down to values as low as ≈35 kJ mol−1 for Cu(I)−MFI and ≈25 kJ
mol−1 for Ag(I)−MFI. These latter values were ascribed to the aspecific interaction
of CO with the zeolite nanopores walls. It is worth noting that the heat of adsorption
associated to this interaction was much larger than the latent heat of liquefaction of
CO (q( L) = 6 kJ mol−1 ).
In agreement with the Langmuir-like behavior of the Na− and K–MFI isotherms
(vide supra Sect. 1.3.1), their Qint versus nads curves were reasonably fitted by linear
equations. A constant value for the differential heat was obtained: qdiff ≈35 kJ mol−1
for Na–MFI and ≈28 kJ mol−1 for K–MFI. The linear fit of the integral heat curves
seemed the most realistic, in spite of the fact that in both cases at very low and at
high coverage the middle points of the experimental histogram deviated from the
constant value. At both low and high coverage, however, the heat values cannot
be assigned to processes involving a specific Me+ · · · CO interaction and must be
disregarded in evaluating the heat of formation of the adducts. In fact, the low-
coverage heterogeneity was due to the presence of a few defective centers (1–2 %
of the total active sites) interacting with CO more strongly than the alkaline metal
cations. Conversely, the high-coverage low heat values correspond to the aspecific
interaction with the zeolite nanopores walls, similarly to what observed for Cu(I)−
and Ag(I)–MFI as explained above.
NH3 adsorbed on H−MFI and all-silica MFI zeolites. In Fig. 1.15 integral
(Fig. 1.15a) and differential (Fig. 1.15b) heats of the reversible adsorption of NH3
on a variety of defective MFI–Silicalite (Sil−A, Sil−B, Sil−C) and on a perfect
(i.e. defect-free) MFI–Silicalite (Sil−D) are illustrated as a function of NH3 uptake.
Note that Sil−A and Sil−D are the same specimens as the ones named in Fig. 1.12
MFI−def and MFI−perf, respectively. See Ref.[24] for the experimental details.
1 Fundamentals in Adsorption at the Solid-Gas Interface 27

(a) 120 (b)

80
1

qdiff (kJ/mol)
80 60
Qint (J/g)

3 1
40
3
40
2 2
20

4 4
0 0
0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0

nads (mmol/g) nads (mmol/g)

Fig. 1.15 Reversible adsorption (ads. II) of NH3 at T = 303 K on defective MFI–Silicalite:
Sil−A(diamond, 1), Sil-B(up triangle, 2), Sil−C(square, 3) and on perfect (i.e. defect-free) MFI–
Silicalite: Sil−D(circle, 4). a Integral heats of adsorption versus NH3 uptake. Curves interpolating
the experimental points (1–4) are polynomials of order 3. b Differential heats of adsorption versus
NH3 uptake. Differential heats were obtained by differentiating the section (a) polynomial func-
tions. All samples were outgassed at T = 673 K. Adapted from Ref. [24] Fig. 4. Note that Sil−A and
Sil−D were the same specimens as the ones named in Fig. 1.12 MFI-def and MFI-perf, respectively

The curves interpolating the integral heats plot experimental points (Fig. 1.15a)
were in all cases polynomials of order 3. Differential heats reported in Fig. 1.15b were
δQint
obtained by differentiating the integral heats polynomial functions. The qdiff = δn ads
curves were all typical of heterogeneous surfaces, in that the heat values decreased
upon increasing coverage, but for the perfect Silicalite (Sil−D). The defective MFI–
Silicalite heat curves (Sil−A, Sil−B and Sil−C) decreased from the initial values
q0 ≈80 kJ mol−1 down to q ≈20 kJ mol−1 , a value close to the latent heat of lique-
faction of NH3 , qL = 21 kJ mol−1 . The evolution of qdiff with increasing coverage
was found to vary according to the different population and/or geometrical arrange-
ment of the polar sites (Si−OH nests) in the defective MFI–Silicalite specimens [24,
90]. Conversely, the heat of adsorption on defects-free MFI–Silicalite (Sil−D) was
found to be virtually coverage-independent and assessed to a very low value (q ≈10
kJ mol−1 ), according to the lack of polar sites capable of specifically interacting with
NH3 . In this latter case, the adsorption was bound to the confinement effect due to
dispersion forces. The constant heat value, lower than latent heat of liquefaction of
NH3 , could be reasonably taken as a measure of the aspecific interaction of ammonia
with the siloxane surface, which takes place also at the flat surface of amorphous
non-porous silica, as documented in Ref. [28].
In Fig. 1.16 the differential heat of NH3 adsorption on one of the all-silica defec-
tive specimens discussed above (Sil−A) will be compared with the correspondent
heat of adsorption on the Brønsted acidic H−MFI zeolite. Note that Sil−A will be
hereafter named MFI−def, in agreement with the nomenclature of the correspondent
volumetric-calorimetric isotherms illustrated in Fig. 1.12.
28 V. Bolis

120

80

qdiff (kJ/mol)
40

qL (NH3)
0
0 1 2 3 4 5
nads (mmol/g)

Fig. 1.16 Differential heat of adsorption versus NH3 uptake at T = 303 K on proton-exchanged
H−MFI (square) and defective all-silica MFI−def (diamond) zeolites. The samples were pre-
outgassed at either T = 873 K (H−MFI) or T = 673 K (MFI−def). Solid symbols ads. I; open
symbols ads. II. Solid lines: polynomial best fitting of H−MFI ads. I and ads. II partial molar heats
experimental points .

The larger affinity towards water of the Brønsted acidic H−MFI zeolite with
respect to the all-silica MFI−def is well evident, as already pointed out by the
Fig. 1.12 volumetric-calorimetric isotherms. The zero-coverage heat of adsorption
on the proton-exchanged zeolite (q0 ≈120 kJ mol−1 ) was compatible with the pro-
ton transfer from the Si(OH)+ Al− Brønsted acidic site to the NH3 molecule, in
agreement with the standard enthalpy of NH3 adsorption a H ◦ = −128 ± 5 kJ
mol−1 measured as isosteric heat (vide infra Sect. 1.5.2) reported in Ref. [81]. After-
wards, as far as the coverage increased the H–MFI heat values decreased similarly to
what observed for the all-silica MFI−def specimen, indicating that a heterogeneous
distribution of surface sites was present at the surface of H−MFI as well. The con-
tribution of the aspecific interaction with the nanopores walls (confinement effect)
characterized by a very low heat of adsorption (q ≈10 kJ mol−1 , as measured for
Sil−D/MFI−perf, see Fig. 1.15b) was most likely one of the causes for the progres-
sive decrease of the heat of adsorption.
The H−MFI ads. II heat curve was initially virtually coincident with that
correspondent to the reversible adsorption of NH3 on the all-silica MFI−def:
q0 ≈80 kJ mol−1 . The heat values for H−MFI ads. II and for MFI−def ads. I =
ads. II followed lying in a common curve up to the coverage of ≈1.5 mmol g−1 .
Afterwards, the trend of the H−MFI ads. II curve changed and reached a coverage
much larger than that of MFI−def.
In conclusion, the most accurate method for evaluating the differential heat
depends on the specific features of the investigated systems, i.e. on the mechanism
of the surface sites filling. It is worth mentioning that if the volumetric/calorimetric
isotherms are properly fitted by suitable equations and if the mechanism of filling
the sites is known, the qdiff values are obtained by processing the isotherms, as it
was done in Ref. [28], dealing with the adsorption of NH3 on a highly dehydrated
1 Fundamentals in Adsorption at the Solid-Gas Interface 29

amorphous silica specimen. In that case the assumption of two independent sites was
suggested/confirmed by IR spectroscopy and ab initio modeling: both volumetric
and calorimetric isotherms were successfully simulated by the superposition of two
local isotherms, one of which was Langmuir-like.

1.5 Thermodynamics of Adsorption

The adsorption of a gas at a solid surface is exothermic. This is required by the

thermodynamic condition for a spontaneous process, illustrated by the Eq. 1.12:

a G = a H − T a S < 0 (1.12)

In fact, adsorption being necessarily accompanied by a decrease in entropy (a S <

0) in that the degrees of freedom of the molecules in the adsorbed state are lower
than in the gaseous state (vide infra Sect. 1.5.3 for details), it turns out that the a H
term, i.e. the enthalpy change accompanying adsorption, must be negative [1, 13,
30].
Heat is not a state function and the value of the heat of adsorption depends on
both the experimental conditions and the employed method of measurement. As a
consequence, any physical interpretation of the experimentally determined heats of
adsorption requires an accurate thermodynamic definition.
Only a few fundamental concepts will be summarized and discussed here, in that
a detailed description of the thermodynamics of adsorption is out of the scopes of this
Chapter. The interested reader is addressed to the exhaustive review on this subject
published in 1992 by Cardona-Martinez and Dumesic [13].
Thermodynamics requires a precise and accurate use of quantities; the convention
suggested by IUPAC for a proper thermodynamics nomenclature is here summarized.
For any extensive quantity X we define: (i) the mean molar quantity x = Xn , which
is the quantity normalized to the amount of matter expressed as moles n, and is
indicated by the lower case; ii) dX dn , the correspondent differential quantity, which
is defined as the derivative of the quantity X with respect to the amount of matter
expressed as moles n, and is indicated by the bar lower case: x.
The quantities of interest are: (i) n, moles of adsorbate; (ii) m, mass of adsorbent;
(iii) V , volume; (iv) p, pressure; (v) T , absolute temperature; (vi) R, molar ideal
gas constant; (vii) A, surface area of the adsorbent; (viii) Q heat; (ix) U, internal
energy; (x) H, enthalpy; (xi) S, entropy and (xii) G, Gibbs free energy. Superscripts
refer to: differential quantities (d); experimentally measured quantities (exp); integral
quantities (int); gas phase (g), adsorbed phase (s) and solid adsorbent (sol) quantities;
standard state quantities (◦ ). Subscript (a) refers to adsorption phenomena (e.g. a H)
[13, 91].
30 V. Bolis

Fig. 1.17 Schematic picture of a real solid surface. Adsorption on: a homogeneous distribution of
surface sites; b heterogeneous distribution of surface sites; c homogeneous distribution of surface
sites in presence of lateral interaction among adsorbed species, either through the space or through
the solid

1.5.1 Heat of Adsorption from Direct Calorimetric Methods

Heats of adsorption derived from direct calorimetric methods are based on the mea-
surement of the heat evolved when a known amount of gas is allowed to adsorb onto a
“clean” surface. A “clean” surface is a solid surface kept in high vacuum conditions
after having been activated either in vacuo in order to eliminate (either totally or
partially) the surface contaminants, or in controlled atmosphere/conditions.
The simplest way envisaged to perform this experiment is to measure the tempera-
ture rise in the solid. This very simple and intuitive way is of limited interest because
it does not give any information about the kinetics of the heat release [10]. Further, it
is worth recalling that not only the magnitude of the heat evolved during adsorption
but also its variation upon increasing coverage may reveal useful information con-
cerning the type of adsorbate/surface sites bonding, and its evolution according to the
surface heterogeneity. For these reasons the equipment described above (a heat-flow
Calvet-type microcalorimeter connected to a high-vacuum volumetric apparatus) is
suitable for characterizing in detail the surface chemical features of a solid material.
As pointed out in Sect. 1.2 (see Figs. 1.1 and 1.2), the surface of a real solid material
is in general characterized by a structural and/or a chemical heterogeneity of the sites,
owing to the presence of either structural defects and/or (hetero)atoms in different
oxidation state. Another kind of surface heterogeneity, originated by the presence of
lateral interactions among adsorbed species, is the so-called induced heterogeneity.
The adsorption on either a homogeneous or a heterogeneous distribution of surface
sites are schematically illustrated in Fig. 1.17a and b, respectively. Lateral interac-
tions among adsorbed species, either through space or through solid, are envisaged
in Fig. 1.17c.
In the following, the thermodynamic features of the direct calorimetric methods
for measuring the heat of adsorption will be discussed. The integral heat Qint is the
heat evolved when ns moles are adsorbed at constant T in a closed gas-solid system.
Because no volume work is done, according to the First Law, the change in internal
energy during the adsorption of ns moles:

n s u s − u g = n s a U (1.13)
1 Fundamentals in Adsorption at the Solid-Gas Interface 31

is given by the integral heat Qint :


Qint = ns us − ug = ns a U (1.14)

and the molar integral heat of adsorption is the molar change in internal energy:

Qint
qint = = us − ug = a U (1.15)
ns

Differential heat of adsorption qd is related to the integral heat according to the

Eq. 1.16:
 int 
δQ
q =
d
= a u (1.16)
δns T , A

In a gas-solid open system, as the one described above (i.e., a Tian-Calvet type
microcalorimeter connected to a high-vacuum gas-volumetric apparatus) the system
will exchange with the environment not only heat but also work and matter. Work is
due to the reversible, isothermal transfer of matter to both gas (V g dp) and adsorbed
(RT dns ) phase. So, the isothermal heat measured in the calorimeter at constant T is
defined by the Eq. 1.17:

Qexp = ns us − ug − RT ns − V g p (1.17)

and is related to the molar change in internal energy by the Eq. 1.18:

(Qexp + RT ns + V g p)
= us − ug = a u (1.18)
ns

It turns out that the two Qexp and Qint quantities differ by the amount of work
exchanged, as illustrated by the Eq. 1.19:

Qint = Qexp + RT ns + V g p (1.19)

The differential heat of adsorption qd , obtained by differentiating the experimental

integral heats Qint measured in isothermal conditions, is expressed by the Eq. 1.20:
     
δQint δQexp δp
qd = = + RT + V g (1.20)
δns T, A δns T ,A δns T ,A

The last term in Eq. 1.20 represents either the compression work during adsorption
or the expansion work during desorption, and is null when two calorimetric cells
(one with the sample, the other as reference element) are connected differentially, as
in the case of the equipment here described. Equation 1.20 so simplifies in Eq. 1.21:
32 V. Bolis
   
δQint δQexp
q =
d
= + RT (1.21)
δns T, A δns T ,A

which is the differential heat of adsorption at constant T and surface area A.

By combining Eqs. 1.21 and 1.18 it turns out that the change in internal energy is
related to the experimental heat measured in the way illustrated by the Eq. 1.22:
   
δQint δQexp
a U = = + RT (1.22)
δns T, A δns T ,A

For processes at constant pressure, we are in general more interested to the enthalpy
than to the internal energy change. By considering that the condensed phases depend
very little upon pressure p and that H s ≈U s , it turns out that the differential form of
the enthalpy change is defined by the Eq. 1.23:
 
δQexp
a h = a u − RT = (1.23)
δns T ,A

From Eq. 1.23 it turns out that the experimental heat measured in a gas-solid open sys-
tem, operating in a differential assembly of calorimetric cells, represents the enthalpy
change associated to the adsorption. This result applies to adsorption processes per-
formed in a gas-solid open system through the admission of the adsorptive on the
solid material kept isothermally within a heat-flow microcalorimeter consisting of
two cells in opposition.
The IUPAC nomenclature adopted in this section is not strictly followed by most
Authors, as witnessed by the quoted literature and will not be followed here in
describing the reported examples.

1.5.2 Heat of Adsorption from Indirect Non-Calorimetric Methods

Another classical method for determining the heat of adsorption is based on the
application of the Clausius-Clapeyron equation:
 
δ ln p
q = RT
st 2
= a H (1.24)
δT ns,A

to a series of adsorption isotherms obtained at (at least) two different temperatures. No

calorimetric measurements are needed: the heat of adsorption is evaluated by using
only the equilibrium data from the isotherms. The isostere of adsorption(p versus T )θ
is obtained by plotting the equilibrium pressure as a function of the adsorption tem-
perature, at constant coverage θ. The so called isosteric heat of adsorption qst is
determined by introducing in Eq. 1.25 (which is obtained by integrating Eq. 1.24) the
1 Fundamentals in Adsorption at the Solid-Gas Interface 33

p1 and p2 values at T1 and T2 , for a given constant coverage θ:

     
p1 qst 1 1
ln = − (1.25)
p2 R T2 T1

It is worth recalling that severe restrictions are imposed in order to obtain accurate
isosteric heats data: (i) the adsorption is reversible; (ii) the gas partial molar volume
is much greater than the adsorbate volume; (iii) the gas behaves ideally; (iv) the
surface state does not vary during measurements; (iv) the heat of adsorption does not
vary with T . For this last reason T1 and T2 should be as close as possible.
Dunne et al. [54] employed the Clapeyron equation for determining the isosteric
heat of a variety of molecules adsorbed on Silicalite. The obtained data were com-
pared by the same authors to the heat of adsorption measured calorimetrically by
means of a home-made volumetric-calorimetric apparatus. The agreement between
the two methods was fairly good.
The isotherms required to evaluate the isosteric heat may be obtained by IR
spectroscopy measurements. In Ref. [92] it was demonstrated that the variable-
temperature FTIR spectroscopy, with the simultaneous measurement of temperature
and equilibrium pressure, may be a suitable method for obtaining the isosteric heat
for an ideal Langmuir-like adsorption process. Optical isotherms of N2 adsorption on
H−ZSM5 were determined (in the 104 –183 K temperature range) from the change
upon adsorption of the integrated intensity of the νOH stretching band at 3616 cm−1 ,
typical of the Brønsted acidic site Si(OH)+ Al− . The adsorption enthalpy so evalu-
ated was a H ◦ = −19.7 ± 0.5 kJ mol−1 , in good agreement with the calorimetric
heats (q = 19 kJ mol−1 ) measured at T = 195 K for N2 on H−ZSM5 by Savitz
et al.[93]. By the same method the standard enthalpy of adsorption of NH3 on a
H−ZSM5 zeolite was evaluated as a H ◦ = −128 ± 5 kJ mol−1 , [81] which is in
reasonably good agreement with the zero-coverage calorimetric heat of adsorption
reported in Fig. 1.16 (q0 ≈120 kJ mol−1 ).

1.5.3 Entropy of Adsorption

In order to illustrate the role played by entropy S in adsorption processes, the entropy
change a S accompanying the adsorption in a very simple case was computed, by
using the standard statistical mechanics rigid rotor/harmonic oscillator formula[94].
In Fig. 1.18 the adsorption of an Ar atom at the surface of an apolar solid is schemat-
ically illustrated. The Ar atom approaches the solid surface from the gas phase: the
translation entropy of the solid, which is fixed in the space, is taken as zero, whereas
the free Ar atoms, before interacting with the solid surface, possess a translational
entropy St which amounts to 150 and 170 J mol−1 K−1 at T = 100 and 298 K,
respectively, at pAr = 100 Torr.
The adsorbed atom, which is not allowed to translate freely anymore, starts vibrat-
ing. Taking as reasonable vibrational frequency ≈100 cm−1 , the adsorbed atoms
34 V. Bolis

Fig. 1.18 Cartoon representing the adsorption of an Ar atom approaching a solid (apolar) surface
from the gas phase

vibrational entropy is Sv = 18 and 43 J mol−1 K−1 at pAr = 100 Torr and at T = 100
and 298 K, respectively. As expected, the adsorbed atoms entropy is much lower than
the free atoms entropy, and the entropy change a S = (Sv − St ) is in all cases neg-
ative: a S ◦ = −132 J mol−1 K−1 at T = 100 K and a S ◦ = −127 J mol−1 K−1 at
T = 298 K.
This means that in a spontaneous process, which requires a negative free energy
change (see Eq. 1.12), the enthalpy of adsorption must be negative in order to com-
pensate the loss of entropy. In other words, the process must be exothermic of an
amount of heat evolved at least as high as the decrease of the T a S term.
The adsorbate can be held at the surface more or less strongly and so the negative
entropy change (with respect to the free molecules in the gas phase) will be more or
less pronounced, according to the nature/structure of the adsorbate and to the nature
and strength of the surface atoms/molecules bonds.
From a molar point of view, the term entropy of adsorption covers a great number
of different functions and it is required to specify whether the function considered
is a derivative or an integral, and if it refers to an equilibrium state (p, T ) or to a
standard state (p◦ , T ) [95, 96].
The integral molar entropy of adsorption is the difference between the entropy of
an adsorbed mole and the entropy of the adsorptive in the ideal gas state, at given p
and T . It is a mean integral quantity taken over the whole amount adsorbed and it is
characteristic of a given state of equilibrium. This is distinguished by the standard
integral molar entropy of adsorption, which is the entropy of one adsorbed mole with
respect to the entropy of the adsorptive in the ideal gas state at the same T , but under
standard pressure.
It is worth recalling that the entropy of adsorption may be obtained from
calorimetric experiments only if the heat exchange is reversible. A formula for evalu-
ating the standard adsorption entropy a S ◦ from a reversible adsorption volumetric-
calorimetric data was proposed by Garrone et al. [91] and applied to a selection of
quasi-ideal systems, [97] consisting of CO adsorbed on non d/d0 metal oxides, at the
surface of which cus cations acting as Lewis acidic sites were exposed. An isothermal
microcalorimeter with a discontinuous (stepwise) introduction of the adsorptive, as
the one described here, was fruitfully employed.
The formation of surface Lewis complexes through a plain σ-coordination being
fully reversible upon evacuation of pCO , this process entirely fulfilled the require-
ments imposed by the method. Equation 1.26 was employed:
1 Fundamentals in Adsorption at the Solid-Gas Interface 35

q p1/2
a S = + R − RT ln ◦ (1.26)
T p

In the right-hand side of the formula, the enthalpic and the ergonic contributions to
the standard entropy of adsorption are recognized. The enthalpic term represented
by the Eq. 1.27:
q
a H = + R (1.27)
T
is obtained from the calorimetric data, whereas the ergonic term represented by the
Eq. 1.28:
p1/2
a G = −RT ln ◦ (1.28)
p

is obtained from the adsorption isotherms. The term p1/2 is the equilibrium pressure
at half-coverage.
The method is here applied to a very simple but instructive case: CO adsorbed via
electrostatic polarization,[23] on Na–MFI and K–MFI pre-outgassed at T = 673 K.
The adsorption represents in both cases an ideal process, which is characterized
(within the experimental error) by a Langmuir-like behavior, as evidenced by the
adsorption isotherms (vide supra Fig. 1.7 in Sect. 1.3.1). The calorimetric heat of
adsorption was ca. 35 and 28 kJ mol−1 for Na–MFI and K–MFI, respectively (vide
supra Sect. 1.4.2.3, Fig. 1.14a, b). The half-coverage equilibrium pressure (obtained
by the adsorption isotherms) were p1/2 = 200 Torr for Na−MFI and 850 Torr for
K−MFI. In both cases the reference state for the gas phase was p◦ = 1 Torr, as done
in previous work (see Ref. [18, 97]). The obtained standard adsorption entropy was
a S ◦ = −151 J mol−1 K−1 for Na–MFI and −140 J mol−1 K−1 for K−MFI.
Note that: (i) the loss of entropy for CO adsorbed at a polar surface through
electrostatic forces was slightly larger than for the aspecific interaction of Ar atoms
adsorbed at an apolar surface discussed above; (ii) the loss of entropy for CO adsorbed
on Na−MFI was larger than on K−MFI, in agreement with the larger strength of
adsorption of CO on Na+ than on K+ sites.
In the following, the evaluation of the adsorption entropy change for the slightly
more complex case of CO σ-coordinated (at T = 303 K) on a variety of TiO2 −
anatase specimens (pre-outgassed at T = 673 K) will be illustrated. At the TiO2
dehydrated surface, CO was adsorbed giving rise to two adspecies, as witnessed
by two distinct IR bands located at νCO = 2188 and 2206 cm−1 , as reported in
Ref. [18] As illustrated schematically in Fig. 1.19 the two adspecies were formed on
two different Lewis acidic sites made up of structurally different cus Ti4+ cations.
They were named species A and B, and their spectroscopic and energetic features
are summarized in the figure.
Species A (νCO = 2188 cm−1 ) were formed at the 5-coord Ti4+ cations
typically exposed at the flat planes of anatase nanocrystals, whereas species B
(νCO = 2206 cm−1 ) were formed at the 4-coord Ti4+ cations, which are exposed
at the steps, corners and kinks of the flat planes [98]. As extensively illustrated
in Ref. [18] species − A νCO moved from 2188 down to 2184 cm−1 as far as the
36 V. Bolis

Fig. 1.19 Spectroscopic and energetic features of species A and B formed upon adsorption of CO at
T = 303 K on TiO2 −anatase pre−outgassed at T = 673 K. Left side Species A formed on 5-coord
Ti4+ cations exposed at the flat planes of anatase nanocrystals. Right side Species B formed on
4-coord Ti4+ cations, exposed at steps, corners and kinks of the flat planes

coverage increased, suggesting the presence of surface heterogeneity among CO

adspecies. Conversely, species-B νCO remained constant at 2206 cm −1 at any cov-
erage examined, suggesting that such species were formed at structurally equivalent
and non-interacting sites. The heat of adsorption values were found to be coverage-
independent for both species A and B: qdiff ≈52 kJ mol−1 and ≈69 kJ mol−1 , respec-
tively.
According to the constancy of both νCO spectral position and heat of adsorption,
species−B adsorption was considered Langmuirian and the Eq. 1.26 could be applied
straightforward. The standard molar entropy of adsorption was so evaluated from
the calorimetric data (qdiff ≈69 kJ mol−1 ) and the Langmuir adsorption isotherm
(p1/2 ≈7 Torr), and turned out to be a S ◦ = −237 J mol−1 K−1 . Note that this value
is larger than both a S ◦ = −151 J mol−1 K−1 and a S ◦ = −140 J mol−1 K−1
obtained for the weak physical adsorption of CO on Na− and K−MFI (vide supra).
The larger loss of entropy in the TiO2 case with respect to the alkaline-metal
exchanged zeolites was due to the stronger Ti4+ ← CO bond, responsible for the
associative chemisorption.
Conversely, as far as species A are concerned, the adsorption was found to deviate
from the ideal Langmuirian behavior, as witnessed by the variation of νCO with
coverage. The apparent contradiction with the constancy of the heat of adsorption
was interpreted as due to the quite scarce population of sites A active towards CO at
T = 303 K, which prevented an accurate determination of the evolution of the heat
of adsorption with coverage [18].
In any case, Langmuir model did not hold in describing species−A thermody-
namic features and Eq. 1.26 employed for species B did not apply. So, the entropy of
CO adsorption at A sites was obtained by considering that, in the very early stage of
adsorption, the coverage θ increased linearly with the pCO [18]. By employing the
Eq. 1.29, based on the Henry-like isotherm:
q
a S = + R + R ln h (1.29)
T
1 Fundamentals in Adsorption at the Solid-Gas Interface 37

Fig. 1.20 Spectroscopic and

energetic features of species
formed on 5-coord cations
exposed at the flat planes of
dehydrated sulphated TiO2 −
anatase nanocrystals

the low-coverage standard molar entropy of adsorption was evaluated from the
calorimetric data (qdiff ≈52 kJ mol−1 ) and the Henry adsorption isotherm (h =
0.027 Torr −1 ). It turned out to be a S ◦ = −194 J mol−1 K−1 . This datum, which
is lower than a S ◦ = −237 J mol−1 K−1 obtained for species B, is in agreement
with the less stronger interaction of CO with sites A with respect to sites B, which
implied a lower loss of degree of freedom of the CO molecule.
The heterogeneity features of the 5-coord Ti4+ ← CO complex were investigated
in more detail on the dehydrated sulphated−TiO2 anatase, which exhibited a popu-
lation of 5-coord Ti4+ ← CO adducts larger than that of the unsulphated specimen
[18]. No B−type CO adspecies were formed, as witnessed by the presence of a single
νCO band, initially located at 2203 cm−1 , as reported in Ref. [18]. The spectroscopic
and energetic features of the single CO adspecies formed at the sulphated − TiO2
surface, which will be named species A , are summarized and schematically illus-
trated in Fig. 1.20. The presence at the surface of strong electron-withdrawing sul-
phate moieties caused the CO oscillators frequency to increase for sulphated anatase
with respect to the unsulphated specimen (2203 cm −1 vs. 2188 cm−1 , respectively).
The same occurred for the heat of adsorption which was found to increase from
qdiff ≈52 kJ mol−1 (unsulphated anatase) up to ≈59 kJ mol−1 (sulphated specimen).
Note also that both CO stretching frequency (νCO = 2203 − 2194 cm−1 ) and heat
of adsorption (59−53 kJ mol−1 ) varied upon increasing coverage, clearly indicating
the presence of induced heterogeneity among the 5-coord Ti4+ ← CO adspecies.
This datum confirmed the ability of TiO2 to transmit electronic effects [18, 22].
The adsorption of species A being non-Langmuirian, Eq. 1.26 could not be
applied and the low-coverage entropy change for species A was evaluated by employ-
ing the formula based on the Henry isotherm, with h = 0.073 Torr −1 . A value of
a S ◦ = −208 J mol−1 K−1 was obtained, which is larger than that obtained for
species A (a S ◦ = −194 J mol−1 K−1 ) and lower than that obtained for species
B (a S ◦ = −237 J mol−1 K−1 ). According to the fact that the CO entropy loss
is lower when the molecule is less tightly bound, the sequence of entropy values
for species A, A and B was in agreement with the sequence of the enthalpy values
(qdiff ≈52, 59 and 69 kJ mol−1 , respectively).
a G◦ values, obtained by combining the a S ◦ and a H ◦ terms, were positive
in all cases reported. The endoergonic character of the process indicated the non-
spontaneity of the adsorption unless a CO pressure is applied: by evacuating the
38 V. Bolis

equilibrium CO pressure, Ti4+ ← CO complexes were completely destroyed, as

confirmed by the coincidence of the 1st and 2nd run of adsorption (not reported
for the sake of brevity; for details see Ref. [18]). It is worth of noticing that a G◦
values were more positive for the weak electrostatic Na+ · · · CO and K+ · · · CO
adducts (a G◦ = +13 and +17 kJ mol−1 , respectively), than for the Ti4+ ←
CO σ−coordinated complexes (a G◦ = +9, +7 and +6 kJ mol−1 for A, A and B
species, respectively), in agreement with the increased strength of the coordinative
chemisorption with respect to the plain polarization of the CO molecule.
In the reported examples, an isothermal microcalorimeter with a discontinuous
(stepwise) introduction of the adsorptive was fruitfully employed. An alternative,
suitable way to collect the experimental data required for determining the enthalpy
and entropy changes accompanying the adsorption process is to follow the proce-
dure implying the slow-and-constant adsorptive introduction which was extensively
described in Ref. [99].
The adsorption entropy of N2 and of NH3 on a H−MFI zeolite was evalu-
ated starting from the correspondent optical Langmuir-like isotherms (vide supra
Sect. 1.5.2), as reported for N2 in Ref. [92] and for NH3 in Ref. [81]. The entropy
change was negative for both adsorptives, but much more negative for the lat-
ter (a S ◦ = −184 ± 10 J mol−1 K−1 ) than for the former adsorbate (a S ◦ =
−125 ± 5 J mol−1 K−1 ), as expected on the basis of the larger adsorption strength
of NH3 than of N2 .

1.6 Adsorption of a Single Component: Physisorption Versus

Chemisorption

Forces governing adsorption depend on the nature of both adsorbent and adsorptive.
They are distinguished in: (i) van der Waals/London-dispersion forces, similar to
the forces leading to the non-ideal behavior of gases and eventually to the forma-
tion of a liquid phase; (ii) chemical (covalent) forces leading to the formation of
a true chemical bond between the adsorbate and surface atoms. In the former case
the process is called physical adsorption (physisorption), in the latter case chemical
adsorption (chemisorption), [2, 32]; (iii) H–bonding interactions are variably clas-
sified as responsible for either physical or chemical adsorption. In fact, H–bonding
interactions originate surface complexes of variable stability according to the nature
of the involved species [100].
The total potential energy of the interacting moieties is quantitatively described
by the Lennard-Jones potential, [1, 2, 30, 32] which includes attractive and repulsive
energy terms.
Physisorption is intrinsically weak and is characterized by heats of adsorption
relatively small, close to the enthalpy of liquefaction of the adsorptive, typically
comprised in the 5−45 kJ mol−1 range; it is in general favored by temperatures close
1 Fundamentals in Adsorption at the Solid-Gas Interface 39

to the boiling point of the adsorptive. As a consequence of the non-covalent nature

of the interaction, adsorbate and adsorptive are chemically equivalent moieties.
Chemisorption, consisting of a chemical reaction confined to the solid surface,
does involve rearrangement of electrons of both adsorptive molecules and surface
atoms, yielding new surface terminations. Adsorptive and adsorbate being chemically
different species in dissociative chemisorption, spectroscopic and/or ab initio model-
ing methods are required to assess the nature of surface species formed upon contact
of the adsorptive with the reactive surface atoms [25, 26, 29]. Further, chemisorption
is structure-sensitive in that the features of the process depend on the solid crystal
structure (see for instance anatase vs. rutile, [56, 101] and amorphous silica vs. crys-
talline quartz, [15, 85, 102]) and on the crystal faces exposed by the solid material
[103].
Chemisorption heats are generally larger than physisorption heats, and are typi-
cally comprised in the 80−400 kJ mol−1 range. Dissociative chemisorption is dis-
tinguished from associative chemisorption, which involves the coordination of the
molecules as such [16, 23, 26, 29]. In this latter case, the enthalpy change is not in
general as large as for dissociation: it can be even lower than ≈80 kJ mol−1 , as for
instance for the case of CO adsorbed on cus non d/d 0 metal cations described above.
The enthalpy change for adsorption driven by H–bonding interactions may cover
a large range of values, in that the correspondent interaction energy depends in a
complex way on the chemical nature of the participating atoms, as well as on the
distances and angles between them. For instance, the zero-coverage enthalpy change
for H2 O adsorption on hydroxylated silica, which is characterized by two H−bonds
per adsorbed molecule on adjacent Si−OH terminations, is − (a H)0 ≈60 kJ mol−1
for amorphous silica and ≈80 kJ mol−1 for crystalline silica [15, 49]. This datum
does suggest that (at least at low coverage) H−bonding interactions exhibit a certain
degree of specificity, similarly to chemisorption. The energy of H−bonding inter-
actions at the water/all-silica zeolites interface is even larger than for plain silica as
indicated by a zero-coverage adsorption enthalpy of ≈100 kJ mol−1 [25].
Conversely, the enthalpy change for H2 O adsorption on dehydroxylated silica,
which is characterized by only one H−bond per adsorbed molecule on isolated
Si−OH terminations, is quite low: −a H < 44 kJ mol−1 , and virtually surface-
coverage independent [15, 56, 85].
Also CH3 OH adsorption on hydroxylated silica is structure-sensitive: on crys-
talline silica the H−bonding adsorption enthalpy at vanishing coverage is −(a H)0
≈90 kJ mol−1 ,[15] whereas on amorphous silica is only ≈70 kJ mol−1 [26].
NH3 is adsorbed on all-silica zeolites and on hydroxylated amorphous silica with
a zero-coverage adsorption enthalpy as high as ≈80 kJ mol−1 ,[24] whereas on dehy-
droxylated silica − (a H)0 is only ≈60 kJ mol−1 [28].
In all cases, at increasing equilibrium pressure the adsorption approaches the
adsorptive liquid phase, and the adsorption enthalpy approaches the adsorptive latent
enthalpy of liquefaction −L H, as typically occurs for physical adsorption at pp◦ → 1
[28, 56, 85, 104].
40 V. Bolis

(a) 100 160 (b) 90

nads (mmol/g)
120
60
75 80

Δ a H (kJ/mol)
q diff (kJ/mol)

40 30
0
50 0 20 40 60 80 0
pCO (Torr)
-30
25 r(P)
qL(CO) -60
r(C)
0
-90
0 40 80 120 160 1 2 3 4 5
n ads (mmol/g) r (Angstrom)

Fig. 1.21 a Differential heats of adsorption of CO, at T = 303 K, on proton-exchanged zeolite

H−BEA zeolite (circle) and on defective MFI–Silicalite (up triangle) as a function of the increasing
coverage. Inset: volumetric isotherms. H–BEA zeolite was pre-outgassed at T = 873 K, MFI–
Silicalite at T = 673 K. Solid symbols ads. I, open symbols ads. II. Dashed line Latent heat of
liquefaction of CO, qL = 6 kJ mol−1 . b One dimension potential energy diagram for physical
adsorption (dotted line) and associative chemical adsorption (solid line). See text for details

In the following, the adsorption of CO on a H−BEA zeolite specimen, charac-

terized by both Brønsted and Lewis acidic sites (consisting these latter of structural
defects in zeolite framework), will be discussed.
The room temperature adsorption of CO on protonic zeolites allowed Lewis acidic
sites to be calorimetrically discriminated from the Brønsted acidic ones: Al(III) ←
CO adducts of relative stability were formed at the cus Al(III) Lewis sites, whereas
only very weak H−bonding adducts were formed at the Brønsted Si(OH)+ Al− acidic
sites [73].
In Fig. 1.21a, the differential heats of adsorption of CO on H−BEA zeolite and
on MFI–Silicalite are reported as a function of the adsorbed amounts. Volumet-
ric isotherms are illustrated in the figure inset. In both cases the adsorption was
fully reversible upon evacuation of the CO pressure, as typical of both physical
and weak, associative chemical adsorption. For H−BEA a constant heat plateau
at q ≈60 kJ mol−1 was measured. This value is typical of a specific interaction of
CO with coordinative unsaturated Al(III) atoms, as it was confirmed by combining
adsorption microcalorimetry and molecular modeling [73, 74, 78, 89] Note that the
heat value was close to the heat of adsorption of CO at cus Al(III) sites on tran-
sition catalytic alumina, a typical Lewis acidic oxide [55, 73]. Once saturated the
Al(III) defects, the heat of adsorption started decreasing down to values typical of the
H−bonding interaction of CO with the Brønsted acidic sites (−a H ≈30 kJ mol−1 ,
as reported by Savitz et al. [93]) and with polar defects, either confined in the zeolite
nanopores or at the external surface.
Conversely, the heat of adsorption of CO at polar sites located in the all-silica
nanopores (Si−OH nests) was much lower (q ≈17 kJ mol−1 ) than on H−BEA, and
virtually constant over the whole coverage examined in agreement with the formation
of very weak H−bonding adducts, as described in Refs. [73, 105]. The adsorption
capacity of the all-silica zeolite was dramatically lower than that of the proton-
exchanged zeolite, as witnessed by the correspondent isotherms reported in the inset
1 Fundamentals in Adsorption at the Solid-Gas Interface 41

Fig. 1.22 Cartoon representing the CO adducts formed within a zeolite nanopore. a CO σ-
coordinated on trigonal Al(III) atom (Lewis acidic site). b CO H−bonded at polar sites Si−OH
(nests). Courtesy of Prof. Piero Ugliengo, University of Torino

of Fig. 1.21. In Fig. 1.21b a schematic one-dimension potential energy diagram is

illustrated, including on one hand the CO physical adsorption on MFI–Silicalite
and on the other hand the associative chemical adsorption on H−BEA. The classical
Lennard-Jones formula,[2, 30, 32, 106] was used. Note that both physical and chem-
ical adsorption of CO, which occur in different zeolites in the present example, are
reported in the same plot just for comparison purposes. The right-side curve repre-
sents physisorption. The Si−OH nests in MFI–Silicalite are the P sites at which CO
is adsorbed via weak H−bonding interaction and r(P) is the equilibrium separation
between the two interacting moieties (in this case set to ≈3 Å). The enthalpy change
−a H(P) is set at ≈17 kJ mol−1 , according to the heat of adsorption reported in
Fig. 1.21a. The left-side curve represents the enthalpy diagram for chemisorption.
C are the associative-chemisorption sites at which CO is σ−coordinated, i.e. the
cus Al(III) cations acting as Lewis acidic sites. r(C) is the equilibrium separation
between the two interacting moieties (in this case set to ≈2.5Å); −a H(C) is the
negative enthalpy change (≈60 kJ mol−1 ) according to the heat of adsorption plateau
reported in Fig. 1.21a.
In Fig. 1.22 a cartoon illustrating the formation in a zeolite nanopore of the
Al(III) ← CO complex (Fig. 1.22a) at the Lewis acidic site, and of the Si−OH· · ·CO
H–bonding adduct (Fig. 1.22b) is reported.
The same kind of one-dimension potential energy diagram (not reported for the
sake of brevity) can be applied to compare NH3 interacting via H−bonding with polar
sites (Si−OH nests) and NH3 interacting via dispersion forces with the nanopore
walls of an all-silica zeolite (confinement effect). In Fig. 1.23a the former interac-
tion is illustrated, in Fig. 1.23b the latter one. The H−bonding interaction of NH3
with Si−OH nests described in Fig. 1.23a cartoon is quite strong, as witnessed by
the relatively high zero-coverage heat values measured for defective MFI–Silicalite
(q0 ≈80 kJ mol−1 , see the heat plots in Figs. 1.15b and 1.16). The process could be
reasonably classified as associative chemisorption. Conversely, the aspecific inter-
action of NH3 with the nanopore walls (Fig. 1.15b) is much weaker, as inferred
by the adsorption heat measured for NH3 adsorbed on defects-free MFI–Silicalite
42 V. Bolis

Fig. 1.23 Cartoon repre-

senting the NH3 adducts
formed within an all-silica
zeolite nanopore. a NH3 H−
bonded at polar sites (Si−OH
nests). b NH3 interacting with
the nanopore walls through
dispersion forces. Courtesy
of Prof. Piero Ugliengo,
University of Torino

(q ≈10 kJ mol−1 , see Fig. 1.15b), which remained constant with increasing coverage,
in that insensitive to possible surface heterogeneity.
As an example of associative/dissociative chemisorption, the interaction of CH3 OH
with a Ca-modified amorphous silica specimen will be discussed. This latter material
was adopted as a model system to investigate the role of Ca species in the development
of bioactivity in sol-gel glasses (see Ref. [26] and references therein). The interac-
tion involved both physical and chemical adsorption. The former process consisted
of H−bonding interactions (the only ones occurring at the surface of the parent plain
silica specimen, data not reported for the sake of brevity), which originated rather
labile adducts on silica polar terminations and were eliminated by pumping off at
the adsorption temperature. Chemisorption involved both associative and dissociative
processes. The associative adsorption at the surface cus Ca 2+ cations (acting as Lewis
acidic sites), yielded coordinated CH3 OH species, a fraction of which was eliminated
by evacuation of the methanol pressure at the adsorption temperature (weak associa-
tive chemisorption). Another fraction was eliminated only by outgassing at temper-
atures as high as T = 423 K, according to the relatively high energy of interaction,
q0 ≈100 kJ mol−1 , measured in the early stage of the adsorption (strong associative
chemisorption). A further fraction was found to resist outgassing up to T > 573 K,
and was assigned to a dissociative chemical adsorption. Dissociative adsorption con-
sisted of a chemical reaction at Ca sites, yielding Si − OCH3 and Ca−OH as new
surface terminations. This process was slow, slightly less exothermic than associative
chemisorption and was found to depend on both CH3 OH pressure and contact time,
as illustrated in Fig. 1.24, in which the volumetric isotherms (Fig. 1.24a) and the
qdiff versus nads plots (Fig. 1.24b) are reported for a first (ads. IA and ads. IIA) and a
second set of measurements (ads. IB and ads. IIB). A large difference between ads. IA
and ads. IIA curves is well evident in both volumetric isotherms and heat of adsorp-
tion plots, indicating the presence of an irreversible component due to a chemical
adsorption. The standard ads-des-ads protocol was followed until a fourth adsorption
run without getting evidence that irreversible adsorption was extinguished. Data for
the third (ads. IIIA) and the fourth (ads. IVA) runs, see Ref. [26], are not reported
for the sake of brevity. The reported results indicated that the process originating
irreversibly adsorbed species, consisting of both strong associative and dissociative
chemisorption, as demonstrated by IR spectroscopy and ab initio modeling, [26] was
not accomplished within the second adsorption (ads. IIA), as typically observed, but
1 Fundamentals in Adsorption at the Solid-Gas Interface 43

(a) 12
n ads ( μmol/m 2 ) (b) 120

q diff (kJ/mol)
80

40
3 q L(CH3OH)

0 0
0 30 60 90 0 3 6 9 12
pCH3OH (Torr)
nads ( μ mol/m2 )

Fig. 1.24 Adsorption (at T = 303 K) of CH3 OH vapor on Ca–modified silica outgassed at T =
423 K. a Volumetric isotherms. b qdiff versus nads plots. Ads. IA: diamond, solid symbols; ads. IIA:
diamond, open symbols. Ads. IB (run after evacuating overnight the sample previously kept 3 days
in contact with ≈80 Torr of CH3 OH): triangle, solid symbols; ads. IIB: triangle, open symbols.
Adapted from Ref. [26] Fig. 12

continued in all subsequent adsorption steps. The irreversible adsorption was found
to represent a progressively decreasing component of the overall uptake, and for the
fourth adsorption run became almost negligible. After the fourth adsorption the sam-
ple (still in the calorimeter at T = 303 K) was kept in contact with CH3 OH pressure
(p ≈80 Torr) for 3 days, and then it was evacuated overnight at the same temperature.
A new set of ads-des-ads runs were then performed (ads. IB and ads. IIB). The pres-
ence of a new measurable irreversible component was detected, as witnessed by the
non-coincidence of ads. IB and ads. IIB isotherms reported in Fig. 1.24a. This datum
confirmed that the prolonged contact of the surface with a large CH3 OH pressure
allowed (some of the) coordinated species to react, yielding an increased amount
of Si−OCH3 and Ca−OH surface species. The qdiff versus nads plots reported in
Fig. 1.24b were typical of highly heterogeneous surfaces, in that heat values progres-
sively decreased with increasing coverage. The zero-coverage heat values for the first
set of data were q0 ≈100 kJ mol−1 for ads. IA and q0 ≈80 kJ mol−1 for ads. IIA. The
ads. IA curve crossed the heat of liquefaction of methanol line (qL = 38 kJ mol−1 )
at a CH3 OH coverage of ≈9.0 μmol m−2 , corresponding to ≈5.4 molecules nm−2 ,
whereas the ads. IIA curve at a CH3 OH coverage of ≈6.5 μmol m−2 , corresponding
to ≈3.9 molecules nm−2 . The ads. IB and IIB curves, even starting from a q0 value
close to the of ads. IIA, decreased together much more sharply than that of ads.
IA and ads. IIA and approached steeply values lower than the heat of liquefaction
of methanol, already at a CH3 OH coverage of ≈4.2 μmol m−2 , corresponding to
≈2.5 molecules m−2 . IR spectroscopy data, also combined with ab initio calcula-
tions, confirmed a precursor-mediated kinetics, in which coordinated CH3 OH was a
precursor for the methoxylation reaction [26].
In Fig. 1.25 the one-dimension potential enthalpy diagram for physical adsorption
(dotted line) and for associative (dot-dashed line) and dissociative (solid line) chem-
ical adsorption schematically illustrates the features of the interactions described
44 V. Bolis

Fig. 1.25 One-dimension 120

potential energy diagram for
physical adsorption (P, dotted

Δa H (kJ/mol)
60
line), and for associative (C,
dashed line) and dissocia-
tive (D, solid line) chemical 0
adsorption. See text for details
r(P)
-60

-120 r(D) r(C)

2 3 4 5
r (Angstrom)

above. The classical Lennard-Jones formula,[106] was used for physisorption and
coordinative chemisorption, whereas a Morse expression was used for dissociative
chemisorption [107].
The right-side curve represents physisorption. P indicates the physical adsorption
site exposed at the silica matrix surface (polar terminations Si−OH) at which CH3 OH
is weakly H−bonded; r (P) is the equilibrium separation between the two interacting
moieties (set to ≈3Å) and −a H (P) ≈35 kJ mol−1 is the enthalpy of adsorption.
The left-side curves represent the enthalpy diagram for chemisorption. Curve C rep-
resents the coordination of CH3 OH at Ca sites (associative chemisorption), whereas
curve D represents the methoxylation reaction of coordinated methanol (dissocia-
tive chemisorption). The equilibrium separation r(C) between the two interacting
moieties is set to ≈2.5Å, whereas the equilibrium separation r(D) in dissociative
chemisorption is set to ≈2.0Å, a lower value than r(C). The enthalpy of adsorption
−a H(C) at r(C) was set to ≈100 kJ mol−1 , whereas −a H(D) at r(D) was set to
≈95 kJ mol−1 , according to the obtained experimental calorimetric and computed
ab initio values [26]. By the inspection of the enthalpy profile illustrated in Fig. 1.25,
it is clearly evident that the presence of molecularly adsorbed precursor C facilitates
the rupture and formation of new chemical bonds at methoxylation site D, in that
the energy barrier hindering the dissociation of methanol is dramatically lowered
starting from the coordinated state instead of from the gas phase. It was calculated
that coordination of CH3 OH at Ca sites is thermodynamically favored with respect
to the methoxylation reaction, in agreement with the experimental results indicating
that, before the prolonged contact with methanol pressure, only a fraction of the total
uptake remained irreversibly bound to the surface.

1.7 Conclusions

In this preliminary Chapter it has been illustrated that adsorption microcalorimetry

is very fruitfully employed in describing quantitatively the processes occurring at
the gas-solid interface. The population of the surface sites active towards suitably
chosen probe molecules is evaluated through the volumetric adsorption isotherms; the
1 Fundamentals in Adsorption at the Solid-Gas Interface 45

surface sites heterogeneity (either structural, chemical and/or induced) is described

by monitoring the evolution of the heat of adsorption with increasing coverage.
By combining the adsorption microcalorimetry results with spectroscopic and/or
ab initio modeling methods, the knowledge of the chemical nature of the pristine
surface terminations and of the adsorbed species allows to interpret at molecular level
the intrinsically molar volumetric-calorimetric data. We will come back to develop
this point in more details in Chap. 15.
Further, it is worth noting that much of the concepts and methods developed for
characterizing the surface chemistry of solid materials of interest in heterogeneous
catalysis may be used for a better understanding of the structure and properties of
solid materials of interest in other fields of the material science as for instance the
materials for biomedical applications (bioactive glasses, biosensors, materials for
controlled drug release).

Acknowledgments Prof. Bice Fubini (University of Torino) teachings in the early stage of my work
in adsorption microcalorimetry are greatly acknowledged. Further, I would like to acknowledge the
contribution of Dr Claudia Busco (zeolites work, University Piemonte Orientale “A. Avogadro”)
and Dr Valentina Aina (Ca-modified silica work, University of Torino) in collecting the volumetric-
calorimetric data published in the quoted references and employed to describe the use of adsorption
microcalorimetry in surface chemistry studies. Without their enthusiasm and helpful contribution
much of this work would not have been carried out. Mr Raffaele Disa (Disa Raffaele e F.lli s.a.s -
Milano, I) is also greatly acknowledged for the endless, patient help in building up and maintaining
the high-vacuum volumetric line.

References

1. J.M. Thomas, W.J. Thomas, Principles and Practice of Heterogeneous Catalysis (VCH,
Weinheim (Germany), 1997)
2. R.A. Van Santen, P.W.N.M. van Leeuwen, J.A. Moulijn, B.A. Averill, Catalysis an Integrated
Approach. Studies in Surface Science and Catalysis, vol 123 (Elsevier, Amsterdam, 1999)
3. L.L. Hench, J. Wilson, Introduction to Bioceramics (World Scientific, Singapore, 1993)
4. D.F. Williams, J. Black, P.J. Doherty, Advances in Biomaterials: Biomaterial-Tissue Inter-
faces, vol. 10 (Elsevier Science, Amsterdam, 1992)
5. B. Kasemo, Biological surface science. Curr. Opin. Solid State Mater. Sci. 3(5), 451–459
(1998). doi:10.1016/S1359-0286(98)80006-5
6. B. Kasemo, Biological surface science. Surf. Sci. 500(1–3), 656–677 (2002). doi:10.1016/
S0039-6028(01)01809-X
7. R. Schlogl, in Handbook of Heterogeneous Catalysis, vol. 8, 2nd edn., ed. by G. Ertl,
H. Knozinger, F. Schuth, J. Weitkamp (Wiley-VCH Verlag, Weinheim, 2008)
8. G. Ertl, in Encyclopedia of Catalysis, vol. 1, ed. by J.T. Horvath (John Wiley & Sons, Hoboken,
NJ, 2003), pp. 329–352
9. A. Zecchina, D. Scarano, S. Bordiga, G. Spoto, C. Lamberti, Surface structures of oxides
and halides and their relationships to catalytic properties. Adv. Catal. 46, 265–397 (2001).
doi:10.1016/S0360-0564(02)46024-5
10. Gravelle PC, in Heat-Flow Microcalorimetry and Its Application to Heterogeneous Catalysis,
ed. by D.D. Eley HP, Paul BW. Advances in Catalysis, vol 22 (Academic Press, 1972).
pp. 191–263
46 V. Bolis

11. V. Bolis, G. Della Gatta, B. Fubini, E. Giamello, L. Stradella, G. Venturello, Identification

of surface sites on potentially catalytic solids by adsorption calorimetry. Gazz Chim Ital 112,
83–89 (1982)
12. B. Fubini, Adsorption Calorimetry in Surface-Chemistry. Thermochim. Acta 135, 19–29
(1988)
13. N. Cardona-Martinez, J.A. Dumesic, Applications Of adsorption microcalorimetry to the
study of heterogeneous catalysis. Adv. Catal. 38, 149–244 (1992). doi:10.1016/s0360-
0564(08)60007-3
14. A. Auroux, Physical Techniques for Solid Materials (Plenum Press, New York, 1994)
15. V. Bolis, A. Cavenago, B. Fubini, Surface heterogeneity on hydrophilic and hydrophobic
silicas: Water and alcohols as probes for H-bonding and dispersion forces. Langmuir 13(5),
895–902 (1997). doi:10.1021/la951006i
16. V. Bolis, S. Bordiga, C. Lamberti, A. Zecchina, A. Carati, F. Rivetti, G. Spano, G. Petrini,
Heterogeneity of framework Ti(IV) in Ti-silicalite as revealed by the adsorption of NH3 .
Combined calorimetric and spectroscopic study. Langmuir 15(18), 5753–5764 (1999). doi:10.
1021/la981420t
17. V.A. Basiuk in Encyclopedia of Surface and Colloid Science, ed. by S. Ponisseril. Adsorption
of Biomolecules at Silica, (Marcel Dekker, Inc., New York, 2002), p. 277
18. V. Bolis, B. Fubini, E. Garrone, C. Morterra, Thermodynamic and vibrational characterization
of CO adsorption on variously pretreated anatase. J. Chem. Soc. Faraday Trans. I 85, 1383–
1395 (1989). doi:10.1039/f19898501383
19. C. Morterra, V. Bolis, B. Fubini, L. Orio, T.B. Williams, A Ftir and Hrem study of some
morphological and adsorptive properties of Monoclinic ZrO2 microcrystals. Surf. Sci. 251,
540–545 (1991). doi:10.1016/0039-6028(91)91051-X
20. V. Bolis, G. Magnacca, C. Morterra, Surface properties of catalytic aluminas modified by
alkaline-earth metal cations: a microcalorimetric and IR-spectroscopic study. Res. Chem.
Intermed. 25(1), 25–56 (1999). doi:10.1163/156856799X00374
21. V. Bolis, S. Maggiorini, L. Meda, F. D’Acapito, G.T. Palomino, S. Bordiga, C. Lamberti,
X-ray photoelectron spectroscopy and x-ray absorption near edge structure study of copper
sites hosted at the internal surface of ZSM-5 zeolite: A comparison with quantitative and
energetic data on the CO and NH3 adsorption. J. Chem. Phys. 113(20), 9248–9261 (2000).
doi:10.1063/I.1319318
22. E. Garrone, V. Bolis, B. Fubini, C. Morterra, Thermodynamic and spectroscopic character-
ization of heterogeneity among adsorption Sites—CO on anatase at ambient-temperature.
Langmuir 5(4), 892–899 (1989). doi:10.1021/la00088a002
23. V. Bolis, A. Barbaglia, S. Bordiga, C. Lamberti, A. Zecchina, Heterogeneous nonclassical
carbonyls stabilized in Cu(I)- and Ag(I)-ZSM-5 zeolites: Thermodynamic and spectroscopic
features. J. Phys. Chem. B 108(28), 9970–9983 (2004). doi:10.1021/Jp049613e
24. V. Bolis, C. Busco, S. Bordiga, P. Ugliengo, C. Lamberti, A. Zecchina, Calorimetric and IR
spectroscopic study of the interaction of NH3 with variously prepared defective silicalites—
Comparison with ab initio computational data. Appl. Surf. Sci. 196(1–4), 56–70 (2002).
doi:10.1016/S0169-4332(02)00046-6
25. V. Bolis, C. Busco, P. Ugliengo, Thermodynamic study of water adsorption in high-silica
zeolites. J. Phys. Chem. B 110(30), 14849–14859 (2006). doi:10.1021/Jp061078q
26. V. Bolis, C. Busco, V. Aina, C. Morterra, P. Ugliengo, Surface properties of silica-based
biomaterials: Ca species at the surface of amorphous silica as model sites. J. Phys. Chem. C
112(43), 16879–16892 (2008). doi:10.1021/Jp805206z
27. M. Corno, A. Rimola, V. Bolis, P. Ugliengo, Hydroxyapatite as a key biomaterial: quantum-
mechanical simulation of its surfaces in interaction with biomolecules. Phys. Chem. Chem.
Phys. 12(24), 6309–6329 (2010). doi:10.1039/C002146f
28. M. Armandi, V. Bolis, B. Bonelli, C.O. Arean, P. Ugliengo, E. Garrone, Silanol-Related and
Unspecific Adsorption of Molecular Ammonia on Highly Dehydrated Silica. J. Phys. Chem.
C 115(47), 23344–23353 (2011). doi:10.1021/Jp206301c
1 Fundamentals in Adsorption at the Solid-Gas Interface 47

29. V. Bolis, C. Busco, G. Martra, L. Bertinetti, Y. Sakhno, P. Ugliengo, F. Chiatti, M. Corno,

N. Roveri, Coordination chemistry of Ca sites at the surface of nanosized hydroxyapatite:
interaction with H2 O and CO. Philos. Trans. R. Soc. A: Math. Phys. Eng. Sci. 370(1963),
1313–1336 (2012). doi:10.1098/rsta.2011.0273
30. G. Attard, C. Barnes (1998) Surfaces. Oxford Chemistry Primers N. 59 (Oxford Science
Publications, Oxford, 1998)
31. I.M. Campbell, Catalysis at Surfaces (Chapman and Hall, London and New York, 1988)
32. S.J. Gregg, K.S.W. Sing, Adsorption, Surface Area and Porosity (Academic Press, London,
1982)
33. J.A. Rabo, Zeolites Chemistry and Catalysis, vol. 171 (ACS Monograph American Chemical
Society, Washington, D.C., 1976)
34. D.W. Breck, Zeolites Molecular Sieves, vol. 4 (John Wiley, New York, 1974)
35. W.M. Meier, D.H. Olson Atlas of Zeolites Structure Types, 2nd edn. (Butterworths, London,
1993)
36. E.M. Flanigen, J.M. Bennett, R.W. Grose, J.P. Cohen, R.L. Patton, R.M. Kirchner, J.V. Smith,
Silicalite, a new hydrophobic crystalline silica molecular sieve. Nature 271(5645), 512–516
(1978)
37. W.O. Haag, R.M. Lago, P.B. Weisz, The active site of acidic aluminosilicate catalysts. Nature
309(5969), 589–591 (1984). doi:10.1038/309589a0
38. A. Corma, From microporous to mesoporous molecular sieve materials and their use in catal-
ysis. Chem Rev 97(6), 2373–2419 (1997). doi:10.1021/cr960406n
39. A. Corma, State of the art and future challenges of zeolites as catalysts. J. Catal. 216(1–2),
298–312 (2003). doi:10.1016/s0021-9517(02)00132-x
40. E. Garrone, B. Fubini, B. Bonelli, B. Onida, C.O. Arean, Thermodynamics of CO adsorption
on the zeolite Na-ZSM-5—a combined microcalorimetric and FTIR spectroscopic study.
Phys. Chem. Chem. Phys. 1(4), 513–518 (1999). doi:10.1039/a806973e
41. C. Morterra, V. Bolis, E. Fisicaro, The hydrated layer and the adsorption of CO at the surface of
TiO2 (Anatase). Colloids Surf. 41(1–2), 177–188 (1989). doi:10.1016/0166-6622(89)80051-
4
42. C. Morterra, G. Magnacca, V. Bolis, On the critical use of molar absorption coefficients for
adsorbed species: the methanol/silica system. Catal. Today 70(1–3), 43–58 (2001). doi:10.
1016/S0920-5861(01)00406-0
43. S. Brunauer, L.S. Deming, W.E. Deming, E. Teller, On a theory of the van der Waals adsorption
of gases. J. Am. Chem. Soc. 62(7), 1723–1732 (1940). doi:10.1021/ja01864a025
44. S. Brunauer, P.H. Emmett, E. Teller, Adsorption of gases in multimolecular layers. J. Am.
Chem. Soc. 60(2), 309–319 (1938). doi:10.1021/ja01269a023
45. S. Brunauer, The Adsorption of Gases and Vapours (Oxford University Press, Oxford, 1945)
46. I. Langmuir, The adsorption of gases on plane surfaces of glass, mica and platinum. J. Am.
Chem. Soc. 40(9), 1361–1403 (1918). doi:10.1021/ja02242a004
47. R. Shannon, Revised effective ionic radii and systematic studies of interatomic distances
in halides and chalcogenides. Acta Crystallogr. A 32(5), 751–767 (1976). doi:10.1107/
S0567739476001551
48. B. Fubini, V. Bolis, A. Cavenago, P. Ugliengo, Ammonia and water as probes for the surface
reactivity of covalent solids—cristobalite and silicon-carbide. J. Chem. Soc., Faraday Trans.
88(3), 277–290 (1992). doi:10.1039/ft9928800277
49. B. Fubini, V. Bolis, A. Cavenago, E. Garrone, P. Ugliengo, Structural and induced hetero-
geneity at the surface of some SiO2 polymorphs from the enthalpy of adsorption of various
molecules. Langmuir 9(10), 2712–2720 (1993). doi:10.1021/la00034a034
50. P.C. Gravelle, Application of adsorption calorimetry to the study of heterogeneous catalysis
reactions. Thermochim. Acta 96(2), 365–376 (1985). doi:10.1016/0040-6031(85)80075-7
51. G. Della Gatta, Direct determination of adsorption heats. Thermochim. Acta 96(2), 349–363
(1985). doi:10.1016/0040-6031(85)80074-5
52. A. Auroux, Acidity characterization by microcalorimetry and relationship with reactivity.
Top. Catal. 4(1–2), 71–89 (1997). doi:10.1023/A:1019127919907
48 V. Bolis

53. B. Fubini, Rev. Gen. Thermique 18, 297 (1979)

54. J.A. Dunne, M. Rao, S. Sircar, R.J. Gorte, A.L. Myers, Calorimetric heats of adsorption
and adsorption isotherms.2. O2 , N2 , Ar, CO2 , CH4 , C2 H6 , and SF6 on NaX, H-ZSM-5, and
Na-ZSM-5 zeolites. Langmuir 12(24), 5896–5904 (1996)
55. V. Bolis, G. Cerrato, G. Magnacca, C. Morterra, Surface acidity of metal oxides. Com-
bined microcalorimetric and IR-spectroscopic studies of variously dehydrated systems. Ther-
mochim. Acta 312(1–2), 63–77 (1998). doi:10.1016/S0040-6031(97)00440-1
56. Fubini B, Bolis V, Bailes M, Stone FS (1989) The reactivity of oxides with water vapor. Solid
State Ionics 32–33, (Part 1 (0))258–272. doi:10.1016/0167-2738(89)90230-0
57. J.M. Newsam, M.M.J. Treacy, W.T. Koetsier, C.B.D. Gruyter, Structural characterization of
zeolite beta. Proc. R. Soc. Lond. Math. Phys. Sci. 420(1859), 375–405 (1988). doi:10.1098/
rspa.1988.0131
58. M.M. Huang, A. Auroux, S. Kaliaguine, Crystallinity dependence of acid site distribution
in HA, HX and HY Zeolites. Microporous Mater. 5(1–2), 17–27 (1995). doi:10.1016/0927-
6513(95)00028-8
59. A.L. Blumenfeld, J.J. Fripiat, Acid sites topology in aluminas and zeolites from
high-resolution solid-state NMR. Top. Catal. 4(1–2), 119–129 (1997). doi:10.1023/A:
1019119718089
60. C. Otero Arean, G. Turnes Palomino, E. Escalona Platero, M. Penarroya Mentruit, Zeolite-
supported metal carbonyls: sensitive probes for infrared spectroscopic characterization of the
zeolite surface. J. Chem. Soc., Dalton Trans. 5, 873–880 (1997). doi:10.1039/A604775K
61. M. Bregolato, V. Bolis, C. Busco, P. Ugliengo, S. Bordiga, F. Cavani, N. Ballarini, L. Maselli,
S. Passeri, I. Rossetti, L. Forni, Methylation of phenol over high-silica beta zeolite: effect of
zeolite acidity and crystal size on catalyst behaviour. J. Catal. 245(2), 285–300 (2007). doi:10.
1016/j.jcat.2006.10.024
62. V. Bolis, S. Bordiga, G.T. Palomino, A. Zecchina, C. Lamberti, Calorimetric and spectroscopic
study of the coordinative unsaturation of copper(I) and silver(I) cations in ZSM-5 zeolite -
Room temperature adsorption of NH3. Thermochim. Acta 379(1–2), 131–145 (2001). doi:10.
1016/S0040-6031(01)00612-8
63. C. Morterra, A. Chiorino, G. Ghiotti, E. Fisicaro, Spectroscopic study of anatase properties.
Part 5—Surface modifications caused by K2 O addition. J. Chem. Soc., Faraday Trans. 1:
Phys. Chem. Condensed Phases 78(9), 2649–2659 (1982). doi:10.1039/F19827802649
64. C. Morterra, G. Ghiotti, E. Garrone, E. Fisicaro, Spectroscopic study of anatase properties.
Part 3—Surface acidity. J. Chem. Soc., Faraday Trans. 1: Phys. Chem. Condensed Phases 76,
2102–2113 (1980). doi:10.1039/F19807602102
65. M. Cerruti, G. Magnacca, V. Bolis, C. Morterra, Characterization of sol-gel bioglasses with
the use of simple model systems: a surface-chemistry approach. J. Mater. Chem. 13(6), 1279–
1286 (2003). doi:10.1039/B300961k
66. E.P.L. Hunter, S.G. Lias, Evaluated gas phase basicities and proton affinities of molecules:
an update. J. Phys. Chem. Ref. Data 27(3), 413–656 (1998)
67. A. Auroux, Microcalorimetry methods to study the acidity and reactivity of zeolites, pil-
lared clays and mesoporous materials. Top. Catal. 19(3–4), 205–213 (2002). doi:10.1023/A:
1015367708955
68. W.E. Farneth, R.J. Gorte, Methods for characterizing zeolite acidity. Chem. Rev. 95(3), 615–
635 (1995). doi:10.1021/cr00035a007
69. A. Zecchina, Otero Arean C (1996) Diatomic molecular probes for mid-IR studies of zeolites.
Chem. Soc. Rev. 25 (3):187–197. doi:10.1039/cs9962500187.
70. A. Zecchina, C. Lamberti, S. Bordiga, Surface acidity and basicity: general concepts. Catal.
Today 41(1–3), 169–177 (1998). doi:10.1016/S0920-5861(98)00047-9
71. R.J. Gorte, What we do know about the acidity of solid acids? Catal. Lett. 62, 1–13 (1999)
72. C. Otero Arean, E. Garrone, Trends in infrared spectroscopy of zeolites. Trends Inorg. Chem.
7, 119–133 (2001)
1 Fundamentals in Adsorption at the Solid-Gas Interface 49

73. V. Bolis, M. Broyer, A. Barbaglia, C. Busco, G.M. Foddanu, P. Ugliengo, Van der Waals
interactions on acidic centres localized in zeolites nanocavities: a calorimetric and com-
puter modeling study. J. Mol. Catal. A: Chem. 204, 561–569 (2003). doi:10.1016/S1381-
1169(03)00339-X
74. C. Busco, A. Barbaglia, M. Broyer, V. Bolis, G.M. Foddanu, P. Ugliengo, Characterisation
of Lewis and Bronsted acidic sites in H-MFI and H-BEA zeolites: a thermodynamic and ab
initio study. Thermochim. Acta 418(1–2), 3–9 (2004). doi:10.1016/j.tca.2003.11.050
75. A. Poppl, T. Rudolf, D. Michel, A pulsed electron nuclear double resonance study of the Lewis
acid site nitric oxide complex in zeolite H-ZSM-5. J. Am. Chem. Soc. 120(19), 4879–4880
(1998). doi:10.1021/ja9741685
76. R.D. Shannon, K.H. Gardner, R.H. Staley, G. Bergeret, P. Gallezot, A. Auroux, The nature
of the nonframework aluminum species formed during the dehydroxylation of H-Y. J. Phys.
Chem. 89(22), 4778–4788 (1985). doi:10.1021/j100268a025
77. P.A. Jacobs, H.K. Beyer, Evidence for the nature of true Lewis sites in faujasite-type zeolites.
J. Phys. Chem. 83(9), 1174–1177 (1979). doi:10.1021/j100472a013
78. C. Busco, V. Bolis, P. Ugliengo, Masked Lewis sites in proton-exchanged zeolites: a compu-
tational and microcalorimetric investigation. J. Phys. Chem. C 111(15), 5561–5567 (2007).
doi:10.1021/Jp0705471
79. H. Willner, F. Aubke, Homoleptic metal carbonyl cations of the electron-rich metals: their gen-
eration in superacid media together with their spectroscopic and structural characterization.
Angewandte Chemie-Int Ed. Engl. 36(22), 2403–2425 (1997). doi:10.1002/anie.199724021
80. A.J. Lupinetti, S.H. Strauss, G. Frenking, Nonclassical metal carbonyls. Prog. Inorg. Chem.
49, 1–112 (2001). doi:10.1002/9780470166512.ch1
81. M. Armandi, B. Bonelli, I. Bottero, C.O. Arean, E. Garrone, Thermodynamic features of the
reaction of ammonia with the acidic proton of H-ZSM-5 as studied by variable-temperature
IR Spectroscopy. J. Phys. Chem. C 114(14), 6658–6662 (2010). doi:10.1021/Jp100799k
82. E.G. Derouane, C.D. Chang, Confinement effects in the adsorption of simple bases by zeolites.
Microporous Mesoporous Mater. 35–6, 425–433 (2000)
83. L. Yang, K. Trafford, O. Kresnawahjuesa, J. Sepa, R.J. Gorte, D. White, An examination
of confinement effects in high-silica zeolites. Russ. J. Phys. Chem. B 105(10), 1935–1942
(2001). doi:10.1021/jp002964i
84. R.J. Gorte, D. White, Measuring sorption effects at zeolite acid sites: pursuing ideas from
W.O. Haag. Microporous Mesoporous Mater. 35–36, 447–455 (2000)
85. V. Bolis, B. Fubini, L. Marchese, G. Martra, D. Costa, Hydrophilic and hydrophobic sites on
dehydrated crystalline and amorphous silicas. J. Chem. Soc., Faraday Trans. 87(3), 497–505
(1991). doi:10.1039/ft9918700497
86. V. Bolis, C. Morterra, M. Volante, L. Orio, B. Fubini, Development and suppression of surface-
acidity on monoclinic zirconia—a spectroscopic and calorimetric investigation. Langmuir
6(3), 695–701 (1990). doi:10.1021/la00093a028
87. V. Bolis, B. Fubini, E. Garrone, E. Giamello, C. Morterra, in Studies in Surface Science and
Catalysis: Structure and Reactivity of Surfaces , vol. 48, ed. by C. Morterra, A. Zecchina, G.
Costa (Elsevier Sci. Publ. B.V., Amsterdam, 1989), pp. 159–166
88. V. Solinas, I. Ferino, Microcalorimetric characterisation of acid-basic catalysts. Catal. Today
41(1–3), 179–189 (1998). doi:10.1016/S0920-5861(98)00048-0
89. V. Bolis, A. Barbaglia, M. Broyer, C. Busco, B. Civalleri, P. Ugliengo, Entrapping molecules in
zeolites nanocavities: a thermodynamic and ab-initio study. Orig. Life Evol. Biosph. 34(1–2),
69–77 (2004). doi:10.1023/B:ORIG.0000009829.11244.d1
90. S. Bordiga, I. Roggero, P. Ugliengo, A. Zecchina, V. Bolis, G. Artioli, R. Buzzoni, G. Marra,
F. Rivetti, G. Spano, C. Lamberti, Characterisation of defective silicalites. J. Chem. Soc.,
Dalton Trans. 21, 3921–3929 (2000). doi:10.1039/B004794p
91. E. Garrone, F. Rouquerol, B. Fubini, G. Della Gatta, Entropy of adsorption and related ther-
modynamic functions in an open isothermal system. J Chim Phys 76, 528–532 (1979)
50 V. Bolis

92. C. Otero Arean, O.V. Manoilova, G.T. Palomino, M.R. Delgado, A.A. Tsyganenko, B. Bonelli,
E. Garrone, Variable-temperature infrared spectroscopy: an access to adsorption thermody-
namics of weakly interacting systems. Phys. Chem. Chem. Phys. 4(23), 5713–5715 (2002).
doi:10.1039/B209299a
93. S. Savitz, A.L. Myers, R.J. Gorte, Calorimetric investigation of CO and N2 for characterization
of acidity in zeolite H-MFI. J. Phys. Chem. B 103(18), 3687–3690 (1999). doi:10.1021/
jp990157h
94. D. McQuarrie, Statistical Mechanics (Harper and Row, New York, 1986)
95. D.H. Everett, Thermodynamics of adsorption. Part I—General considerations. Trans. Faraday
Soc. 46, 453–459 (1950)
96. T.L. Hill, Statistical mechanics of adsorption. V. Thermodynamics and heat of adsorption.
J. Chem. Phys. 17(6), 520–535 (1949). doi:10.1063/1.1747314
97. E. Garrone, G. Ghiotti, E. Giamello, B. Fubini, Entropy of adsorption by microcalorimetry.
Part 1—Quasi-ideal chemisorption of CO onto various oxidic systems. J. Chem. Soc., Faraday
Trans. 1: Phys. Chem. Condensed Phases 77(11), 2613–2620 (1981)
98. G. Busca, H. Saussey, O. Saur, J.C. Lavalley, V. Lorenzelli, Ft-Ir Characterization of the
surface-acidity of different titanium-dioxide anatase preparations. Appl. Catal. 14(1–3),
245–260 (1985). doi:10.1016/S0166-9834(00)84358-4
99. F. Rouquerol, J. Rouquerol, C. Letoquart, Use of isothermal microcalorimetry data for
the determination of integral molar entropies of adsorption at the gas–solid interface by a
quasi-equilibrium procedure. Thermochim. Acta 39(2), 151–158 (1980). doi:10.1016/0040-
6031(80)80008-6
100. H. Knozinger, The Hydrogen Bond (North Holland, Amsterdam, 1976)
101. U. Diebold, The surface science of titanium dioxide. Surf. Sci. Rep. 48(5–8), 53–229 (2003)
102. V. Bolis, B. Fubini, E. Giamello, Effect of form on the surface-chemistry of finely divided
solids. Mater. Chem. Phys. 29(1–4), 153–164 (1991). doi:10.1016/0254-0584(91)90012-J
103. L. Kieu, P. Boyd, H. Idriss, Trends within the adsorption energy of alcohols over rutile
TiO2 (110) and (011) clusters. J. Mol. Catal. A: Chem. 188(1–2), 153–161 (2002). doi:10.
1016/S1381-1169(02)00210-8
104. A.C. Zettlemoyer, F.T. Micale, K. Klier, Water in dispersed Systems, vol. 5 (Plenum, New
York, 1975)
105. S. Savitz, A.L. Myers, R.J. Gorte, A calorimetric investigation of CO, N2 , and O2 in
alkali-exchanged MFI. Microporous Mesoporous Mater. 37(1–2), 33–40 (2000). doi:10.1016/
S1387-1811(99)00190-0
106. J.E. Jones, On the determination of molecular fields. II. From the equation of state of a gas.
Proc. R. Soc. Lond. Ser. A 106(738), 463–477 (1924). doi:10.1098/rspa.1924.0082
107. P.M. Morse, Diatomic molecules according to the wave mechanics. II. Vibrational levels.
Physical Review 34(1), 57–64 (1929). doi:10.1098/rspa.1924.0082
Chapter 2
Thermal Analysis and Calorimetry
Techniques for Catalytic Investigations

Pierre Le Parlouër

Abstract The use of thermal analysis and calorimetry techniques is quite an old and
known field of applications for the catalytic investigations and many publications
have been published on the various topics including analysis of catalysts, investiga-
tion of the processes during the preparation of catalysts, desactivation of catalysts
and interaction of reactants or catalytic poisons with the catalysts. Differential ther-
mal analysis, calorimetry and thermogravimetry are also used to characterize the
catalysts, especially in the field of gas–solid and gas–liquid interactions. Since the
last years, many technical improvements have appeared in the design and the use
of thermal analyzers and calorimeters, particularly for the characterization of cat-
alysts. This chapter gives a detailed overview of the uptodate thermal techniques
covering various techniques including Differential Thermal Analysis (DTA), Differ-
ential Scanning Calorimetry (DSC), the calorimetric techniques (including Isother-
mal Calorimetry, Titration Calorimetry), Thermogravimetric Analysis (TGA), the
combined techniques (including TG-DTA and TG-DSC), the Evolved Gas Analysis
(including TG-MS, TG-FTIR). Some examples of applications are given to illustrate
the catalyst characterizations.

2.1 Introduction

The use of thermal analysis and calorimetry techniques is quite an old and known
field of applications for the catalytic investigations and many publications have been
published on the various topics. In 1977, Habersberger [19] has published a short
review of the possibilities of applications of thermal analysis to the investigation
of catalysts including analysis of catalysts, investigation of the processes during
the preparation of catalysts, desactivation of catalysts and interaction of reactants

P. Le Parlouër (B)
Setaram, 7 Rue de l’Oratoire, 69300 Caluire, France
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 51

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_2,
© Springer-Verlag Berlin Heidelberg 2013
52 P. L. Parlouër

or catalytic poisons with the catalysts. In 1994, Auroux [4, 8] has described the
different thermal techniques including differential thermal analysis, calorimetry and
thermogravimetry that are used to characterize the catalysts, especially in the field of
gas–solid and gas–liquid interactions. More recently in 2003, Pawelec and Fierro [45]
have prepared a chapter called “Applications of thermal analysis in the preparation of
catalysts and in catalysis” for the “Handbook of Thermal Analysis and Calorimetry”.
Since the last years, many technical improvements have appeared in the design and
the use of thermal analyzers and calorimeters, particularly for the characterization
of catalysts. Following the summer school of calorimetry named “Calorimetry and
thermal methods in catalysis” organized in Lyon (France) since 2007, this chapter
will give a detailed overview of the uptodate thermal techniques presented during
this meeting with some examples of applications for catalyst characterization.

2.2 Thermal Analysis and Calorimetry:

Techniques and Applications

Thermal Analysis (TA) means the analysis of a change in a sample property which is
related to an imposed temperature alteration and calorimetry means the measurement
of heat [24]. The corresponding property of the sample is measured according to time
and temperature with the main following techniques:
• Temperature difference >>> Differential Thermal Analysis (DTA).
• Heat flux variation >>> Differential Scanning Calorimeter (DSC).
• Heat variation >>> Calorimetry.
• Mass change >>> Thermogravimetry (TG).
• Length or volume change >>> Thermomechanical Analysis (TMA).
There is no one method of measurement in thermal analysis and calorimetry that
can be used for all materials or to cover all the possible thermal properties and
applications. A measurement method has to be selected depending on the following
criteria:
• the parameter to be measured (temperature, heat, mass or length),
• sample type (solid or liquid),
• sample size,
• sample reactivity (which affects choice of crucible, protective gas),
• range of temperature,
• atmosphere around the sample (inert, oxidative, reducing),
• operation under pressure,
• operation in a corrosive gas,
• operation in humid atmosphere,
• sensitivity of detection in changes of mass, heat and length,
• whether a combined technique is necessary, . . .
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 53

According to the different thermal techniques and the selection of the experimental
parameters, a large selection of applications is available for the characterization of
catalysts and related materials, and also the simulation of the catalytic processes.
The Table 2.1 provides a brief overview of the main applications that have been
developed.

2.3 The Differential Thermal Analysis Technique

2.3.1 Principle

ICTAC (International Confederation for Thermal Analysis and Calorimetry) gives

the following definition for the DTA technique: “A technique where the tempera-
ture difference between a sample and a reference material is measured while they
are subjected to the same temperature variation (heated or cooled) in a controlled
atmosphere” [25]
The Differential Thermal Analysis technique is based on a differential mounting
of thermocouples in the sample (S) crucible and reference (R) crucible (Fig. 2.1).
The device is located in a block heated (or cooled) at a controlled temperature.
The electric signal (emf) measured at the ends of the thermocouple is proportional
to the difference of temperature, T, between the sample and reference sides.
The temperature of the furnace (Tp ) is programmed with a linear heating rate
(Fig. 2.2). The reference temperature (Tr ), measured in an inert material (that means
a material that does not undergo a thermal transition for the given temperature range

Table 2.1 Techniques and applications in thermal analysis

Physical Technique Application
property
Temperature Differential thermal Phase change, dehydration, desorption, decomposition,
analysis (DTA) reduction, oxidation, . . .
Heat flow Differential scanning Phase change, adsorption, absorption, hydrogenation,
calorimetry (DSC) dehydration, desorption, decomposition,
dehydrogenation, reduction, oxidation, catalysis,
heat capacity, kinetics, . . .
Heat Calorimetry Reaction, adsorption, absorption, hydration, mixing,
formation, catalysis, thermodynamics, heat capacity,
kinetics, . . .
Mass Thermogravimetric Adsorption, hydrogenation, dehydration, desorption,
analysis (TGA) decomposition, dehydrogenation, oxidation,
kinetics, . . .
Length, Thermomechanical Phase change, expansion, sintering, . . .
volume analysis (TMA)
54 P. L. Parlouër

Fig. 2.1 DTA principle

Fig. 2.2 DTA principle—recording of the different temperatures

of investigation) follows this heating profile with a small delay due to the thermal
gradient in the crucible.
The sample temperature (Te ), measured in the sample material under investi-
gation, follows in the same way the heating profile until a thermal transition in the
sample is reached. In the case of melting, the temperature of the sample remains
constant during melting (this is the melting plateau). When melting is complete the
sample temperature changes to then follow the heating profile.
By measuring the difference between the sample temperature and the reference
temperature, the thermogram (Fig. 2.3) is obtained, describing the melting peak.
From this curve, a first information is obtained: the temperature of melting of
the material. In a case of a pure sample, this temperature corresponds to the onset
temperature of the melting peak.
Attention: Such a principle described on Fig. 2.1 does not experimentally exist.
In fact there is a risk of interaction between the thermocouple and the material that
will affect the measure. In the experimental DTA device, a crucible has to be used
as an interface between the thermocouple and the sample.
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 55

Fig. 2.3 Melting thermogram

Fig. 2.4 The DTA thermo-

gram

The DTA technique also gives another important information on the type of tran-
sition, either endothermic or exothermic effect. The endothermic tranformation pro-
duces a decrease of the temperature that means that T is negative. For an exothermic
transformation, a positive T is recorded. The DTA thermogram (Fig. 2.4) allows
making a clear identification of the different types of transformations according to
the side of the peaks.
Among the different types of endothermic and exothermic effects, the following
list can be given when catalysts are under investigation:
• Endothermic: melting, evaporation, sublimation, dehydration, dehydroxylation,
desorption, and pyrolysis.
• Exothermic: crystallization, adsorption, oxidation, combustion, hydrogenation,
and decomposition.
A variation of the base line may also be detected on the DTA curve (the first transition
on Fig. 2.4). This is the signature of a glass transition, expressing that the material
contains an amorphous phase.
56 P. L. Parlouër

2.3.2 Detectors

A standard DTA detector is shown in Fig. 2.5. Two different types of metals are used
to build the DTA arrangement. Different types of thermocouples are used to produce
DTA detectors according to the range of temperature to be investigated (Table 2.2).
The sensitivity of the detector varies with the nature of the thermocouple. Higher
sensitivity is obtained with types of thermocouples in the top part of Table 2.2.
A crucible is adjusted on each side of the DTA rod for the sample and the inert
material. The crucible volume varies from 20 to 100 μl.
The choice of the crucible is a very important step in the DTA experimentation. It is
needed that there is no interaction between the crucible and the sample during the
test. In the same time it is needed that the crucible is a good heat conductor in order

Fig. 2.5 Principle and cross section of a DTA detector

Table 2.2 Different types of thermocouples used for DTA detectors

Type Composition Temperature range
Platinel 55Pd 31Pt 14Au/65Au 35Pd −150 ◦ C → 1000 ◦ C
E Ni Cr/Cu Ni = Chromel/Constantan −50 ◦ C → 800 ◦ C (500 ◦ C under air)
S Pt Rh10 % / Pt 1600 ◦ C (1000 ◦ C under H2 )
B Pt Rh30 % / Pt Rh6 % 1750 ◦ C (1000 ◦ C under H2 )
W5 W Re26 % / W Re5 % 2400 ◦ C / inert atmosphere
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 57

to ensure a good heat transfer between the sample and the thermocouple. Different
materials are used to produce adequate crucibles:
• Aluminium (up to 500 ◦ C),
• Alumina (especially required for metallic samples),
• Platinum (especially required for inorganic samples),
• Graphite and tungsten (for temperature above 1750 ◦ C).
According to experimentation and corrosion problems, other types of materials are
available such as magnesia, zirconia, boron nitride, . . .

2.3.3 Operation

In most experimentations, the DTA crucible is used open. A cover is needed especially
in the two following situations:
• The vapour has to be retained above the sample, in the case of dehydration or
decomposition. For such an experiment a cover with a pin hole is used.
• There is a risk of overflowing of the sample when decomposing. In such a case
it is recommended to close the crucible with a lid to avoid the destruction of the
DTA rod.
As the crucible is open, the control of the gas above the sample is a very important
parameter. Different types of gas atmosphere are available according to the experi-
ment to be run:
• Inert atmosphere (nitrogen, argon, helium) to protect the sample from any
oxidation.
• Oxidative atmosphere (air, oxygen) to be used for oxidation and combustion inves-
tigations.
• Reactive gas (hydrogen, carbon monoxide, ammonia, . . .) for adsorption and
absorption reactions.
• Water vapour for hydration reactions.

2.3.4 Applications

The applications of the DTA technique are quite wide for the characterization of the
catalysts and related products, but they are mainly oriented on the determination of
their structural properties. Here below is a non exhaustive list of DTA applications:
• Melting and crystallization,
• Phase transitions (glass transition, order–disorder, . . .),
• Preparation of catalyst: dehydration and dehydroxylation,
• Rapid screening of potential catalysts,
• Evaluation of the effects of various pretreatments,
58 P. L. Parlouër

• Adsorption, desorption,
• Decomposition,
• Oxydation, reduction,
• Regeneration of catalyst, . . .
The DTA applications applied to catalysts and more generally to inorganic materials
have been extensively reviewed in various books [23, 33, 44, 45, 48, 61].

2.4 The Differential Scanning Calorimetry Technique

2.4.1 Principle

Differential Scanning Calorimetry is defined as followed by ICTAC:

“A technique in which the heat flux (thermal power) to (or from) a sample is measured
versus time or temperature while the temperature of the sample is programmed, in
a controlled atmosphere. The difference of heat flux between a crucible containing
the sample and a reference crucible (empty or not) is measured.” [25].
In a heat flux DSC, the specimen surroundings (generally the furnace and detector)
are at constant temperature (isothermal mode) or at a variable temperature (scanning
temperature mode). A defined exchange of heat takes place between the specimen and
its surroundings (Fig. 2.6). The amount of heat flux (heat flow rate) is determined

Fig. 2.6 Heat flux DSC

principle
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 59

on the basis of the temperature difference along a thermal resistance between the
specimen and its surroundings.
The heat flux for a given sample at a temperature Ts is equivalent to:

dqs /dt = −dh/dt + Cs dTs /dt (2.1)

where:
• dh/dt: Heat flux produced by the transformation of the sample or the reaction.
• Cs : Heat capacity of the sample, including the container.
The heat flux dqs /dt is exchanged with the thermostatic block at a temperature Tp
through a thermal resistance R according to the following relation;

dqs /dt = (Tp − Ts )/R. (2.2)

From the relation (2.1), it is seen that the thermal contribution due to the heat capacity
of the sample and container is significant and will provide a major disturbance at
the introduction of the container in the calorimeter. From the relation (2.2), it is also
evident that any temperature perturbation of the thermostatic block will affect the
calorimetric measurement.
In order to solve these different problems, a symmetrical design is used. An
identical crucible with an inert material (the reference container can also be empty)
is placed on the detector at the same Tp temperature.
The difference of heat flux is measured between the two sides:

dq/dt = dqs /dt − dqr /dt = −dh/dt + Cs dTs /dt − Cr dTr /dt (2.3)

where:
• Cr : Heat capacity of the reference, including the container.
• Tr : Temperature of the reference.
The relation (2.2) becomes:

dq/dt = (Tr − Ts )/R

or by derivation
Rd2 q/dt 2 = dTr /dt − dTs /dt (2.4)

By combination of relations (2.3) and (2.4), the characteristic equation for the calori-
metric measurement is obtained:

dh/dt = −dq/dt + (Cs − Cr )dTp /dt − RCs d2 q/dt 2 . (2.5)

If dh/dt corresponds to an absorbed thermal power due to an endothermic transfor-

mation or reaction, the dh/dt value is positive.
60 P. L. Parlouër

If dh/dt corresponds to a released thermal power due to an exothermic transformation

or reaction, the dh/dt value is negative.
dh/dt (heat flow rate) is related to the kinetics of the transformation. The shape of
the DSC peak gives a first indication on the rate of reaction.
When dh/dt is null (no transformation), the Eq. (2.5) allows to access the heat
capacity of the sample. DSC provides a direct and accurate measurement of the
specific heat of any type of material.
If the DSC test is run isothermally, the parameter dTp /dt is null. In the case of a
small perturbation of the temperature Tp of the thermostatic block, the corresponding
thermal effect will be minimized if the Cs and Cr heat capacities are similar.
The last term R Cs d2 q/dt 2 (called also thermal lag) that affects the measurement
mostly depends on the thermal resistance or also the time of response of the DSC
on one side, and the heat capacity of the sample and container on the other side. For
long period of time (t>>RCs ) it will be negligible.
As seen before with DTA (Fig. 2.4), the DSC curve will show endothermic and/or
exothermic effects according to the transformations and the reactions.
If DTA is more considered as a qualitative technique, the sophisticated DSC detectors
define a quantitative and accurate determination of the heat effects.
From the DSC curve different important information are obtained (Fig. 2.7):
• the temperature of the transformation (generally measured at the top of the peak),
• the heat related to the transformation, and
• the rate of the transformation.
In order to get accurate measurements of temperature of transition, the DSC has to
be calibrated using reference materials [20]. The experimental procedures for such
a correction are described in different standards, especially ISO standard [26].
As the DSC signal is equivalent to a thermal power (expressed in mW), the
integration of the DSC peak between the time ti (start of the peak) and tf (end of the

Ti
Tf

ti tf

Fig. 2.7 DSC curve

2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 61

peak) corresponds to the area under the peak, delimited by the line drawn between
the two points.
Before getting the corresponding heat value, a calibration of the DSC detector is also
needed. The raw DSC signal is an electrical signal provided by the thermocouples,
expressed in microvolt. To convert this signal in microwatt, the use of reference
materials with known heat of melting is required [20].
However another calibration technique is available when the Calvet type DSC is
used. The principle is to apply a known amount of power in a dedicated calibration
vessel. To reach this target, a resistance is embedded in the crucible. A known current
I is delivered and the corresponding tension U measured, providing the power P = UI
that is applied. The corresponding Joule effect provides a DSC exothermic deviation
in microvolt (Fig. 2.8). Such an electrical calibration is very interesting as it can
apply at any temperature, even at constant temperature. This will be more detailed
in paragraph 5 for the calorimetric techniques.
The other important information associated to the DSC peak is related to the
kinetics of the transformation. As the DSC signal corresponds to a heat flow rate, the
shape of the peak directly informs on the rate of the transformation. For example a
sharp peak indicates a high rate of reaction. At the top of the peak, dh/dt max (Fig. 2.7),
the rate of the reaction reaches the maximum value. This information will be used
in many models for the determination of the kinetic parameters of a transformation.

2.4.2 Detectors

Different types of DSC techniques are available and are described in Table 2.3.

Fig. 2.8 Principle of the DSC electrical calibration using the Joule effect
62 P. L. Parlouër

Table 2.3 Different types of DSC techniques

Property Technique Description
Heat flow rate Heat flux DSC A technique in which the difference in heat flow
rate to a test sample and to a reference sample
is analysed while they are subjected to the
same temperature variation (heated or
cooled). The detector is a plate-DSC detector.
Heat flow rate Power compensated DSC A technique in which the temperature difference
between a test specimen and a reference
specimen occurring through subjecting both
specimens to the same controlled temperature
program is compensated by appropriately
adjusting the difference of heating power to
the test and reference specimen. The
differential heating power is recorded against
temperature or time.
Heat flow rate Calvet DSC A technique in which the difference of heat flow
rate to the sample and to the reference sample
is analysed while they are subjected to a
temperature variation (heated or cooled). The
detector is a cylindrical DSC detector.

Fig. 2.9 DSC power compensation principle

For the heat flux plate DSC (Fig. 2.6) and the power compensation DSC (Fig. 2.9),
the common characteristic is that they both use a flat-shaped sensor.
The principle of the power compensated DSC [37] is different from the heat flux
DSC described previously. The heaters beneath the pans aim to minimise the dif-
ference in temperature between a specimen and an inert material. When a transition
occurs in the specimen, the reference heater will aim to compensate for this and keep
the reference pan at a similar temperature. Individual platinum resistance thermome-
ters measure the temperature of each pan, and the power required to maintain this
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 63

state is recorded. Endothermic and exothermic transitions are recorded depending if

power is added or subtracted.
Even if the measurement principles are quite different, the heat transfer from (or
to) the sample is about the same. The sample, contained in a metallic crucible, is
placed and centered on the plate acting as a flat-shaped sensor. A reference crucible
(empty or containing an inert material) is placed on the other plate.
The basics of the plate DSC is that the heat exchange between the sample and
the detector is done through the bottom of the crucible, corresponding to a two
dimensional detection. In fact only a part of this heat transfer is measured, as a
significant part is dissipated through the walls and the cover of the crucible (Fig. 2.6).
If the efficiency ratio is considered, that means the heat flux measured by the sensor
to the total heat flux produced by the thermal event, a simulation using a thermal
modeling software shows that only around half of the heat flux is dissipated through
the plate [13, 31]. The Fig. 2.10 clearly shows that the efficiency rapidly decreases
with the temperature and the thickness for the plate. The efficiency is also affected by
the amount of the sample under investigation. This is why it is recommended to work
with small amounts of material (about 5–10 mg) when using a plate DSC, in order to
minimize the heat losses. The thermal conductivities of the crucible and the gas used
in the experimental chambers also are very important parameters to be considered in
the efficiency of the heat exchange. For example a very heat conductive gas (helium)
will favour the heat transfer between the crucible and the detector, but in the same
time increase the heat losses. So the calibration of a plate-type DSC is very critical
and has to be run with the experimental conditions that are selected for the sample test.
The Calvet type DSC offers another technical option [38].
The detection concept is based on a three dimensional fluxmeter sensor. The fluxme-
ter element consists in a ring of several thermocouples in series (Fig. 2.11). The
corresponding thermopile of high thermal conductivity surrounds the experimental
space within the calorimetric block. The radial arrangement of the thermopiles guar-
antees an almost complete integration of the heat. This is verified by the calculation
of the efficiency ratio that indicates that an average value of 94 ± 1 % of heat is
transmitted through the sensor on the full range of temperature of the Calvet type
DSC (Fig. 2.12). Another very important point is that the sensitivity of the DSC is
no more affected by the type of crucible, the type of purge gas and flow rate.

Fig. 2.10 Efficiency ratio of a 50

flat-shaped DSC as a function
of the sensor plate thickness 40
% Flux

30

20 0,01 mm

10 0,05 mm
0,10 mm
0
0 150 300 450 600
Temperature ( °C)
64 P. L. Parlouër

Fig. 2.11 Schematic of the

Calvet principle

One of the main advantages of the Calvet-type DSC is that larger amounts of sample
can be investigated. The open tube detection allows the adaptation of different types
of experimental crucibles, especially with the possibility of introduction of various
types of gas under normal or high pressure. This design specificity is of a great
interest for the applications on catalysts, and more generally for all the adsorption
investigations.

2.4.3 Operation

The selection of the crucible for the sample is one of the most important task to get
a good experiment. The choice depends on a series of parameters that has to be very
clearly identified:
• The nature of the crucible: various materials are available (aluminium, platinum,
gold, stainless steel, . . .) according to the temperature range. It is important that
no interaction occurs between the sample and the crucible during the test.
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 65

Ratio of energy transmitted through the thermopile

95
94,8

Ratio of energy (%)

94,6
94,4
94,2
94
93,8
93,6
93,4
93,2
93
0 100 200 300 400 500
T (°C)

Fig. 2.12 Efficiency ratio of a Calvet type calorimeter versus temperature

• The way of using the crucible:

– closed with a lid: this is the most common use for a DSC crucible. It is very
convenient for any type of transition or transformation as soon as the vapour
pressure of the sample remains low during the test
– closed with a pierced lid: the hole is to favour the escape of vapour from the
crucible, for example during a dehydration
– open: this situation is needed when a gas interaction with the sample is investi-
gated, for example an adsorption
– tightly closed: dedicated crucibles are available to keep the vapour inside the
crucible during a reaction. As this will produce an increase of the vapour pres-
sure, a tight closing of the crucible is strictly needed.
Two other types of crucibles, that can be only used with a Calvet type DSC are also
described as they find specific applications for catalytic investigations.
The first type of crucible is a quartz reactor that is designed for the investigation
of gas adsorption on powders (Fig. 2.13).
The quartz tube is introduced in the Calvet type DSC (set in the vertical position).
A fritted glass substrate is located in the middle of the tube to receive the powdered
sample in order to be surrounded by the calorimetric detector. Tight connections are
adjusted at both ends of the tubes for the gas inlet and outlet.
By using a gas injection loop, pulses of known volume of reactive gas are introduced
on the sample [4, 9, 16, 46, 50]. The DSC signal gives the corresponding heat of
adsorption. In order to know if the total volume of gas is adsorbed on the sample, a GC
detector is set on-line with the DSC. The outlet gas is analyzed and the corresponding
amount of non adsorbed gas measured.
This DSC-GC coupling allows correlating the heat of adsorption to the amount of
gas adsorbed on the sample.
The use of a quartz tube makes also possible to use any type of gas (even corrosive
gas) for such an investigation.
66 P. L. Parlouër

Fig. 2.13 Quartz reactor for adsorption studies and coupling with the gas introduction and analysis
device

Fig. 2.14 Controlled high pressure crucible with the Calvet type DSC

The second interesting type of crucible is dedicated to work under a controlled

gas pressure. Today more and more works are done on the adsorption of hydrogen
to form metal hydrides used for the storage of hydrogen. Such investigations have to
be run under a controlled pressure of hydrogen and a given temperature, or range of
temperature.
After introducing the sample, the crucible (made of incoloy) is tightly closed with
a screwable stopper, then introduced in the tube of the Calvet type DSC (Fig. 2.14).
A pressure of gas is applied through an external high pressure panel.
Using such an experimental design, only the small volume containing the sample is
under pressure. The DSC detector itself remains under normal pressure.
This is very convenient for safety reasons especially when hydrogen is used, but also
for the DSC calibration as it is not affected by the gas pressure. A current pressure
of 400 bar is available on such a vessel up to 600 ◦ C.
Two different modes are used for the investigation of absorption and decomposition
of metal hydrides:
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 67

• Operation in temperature scanning mode with a constant gas pressure.

• Operation at constant temperature with a variable pressure.

2.4.4 Applications

The DSC applications for catalytic investigations are very similar to the ones
described previously for the DTA technique. However the DSC measurement gives
more possibilities as described in Sect. 4.3, especially for adsorption investigations
under normal [57] or high pressure [51]
The DSC technique also offers more flexibility for coupling with other techniques
(thermogravimetry, gas analysis).
DSC can also be used for the Thermal Programmed Desorption (TPD) determination
coupled with a gas analyzer [11, 34]
Among the applications that are specific to the DSC technique, the determination
of heat capacity is one of the most important. Different methods can apply for this
measurement and are described below.

2.4.4.1 Heat Capacity Determination

Heat capacity plays an important role in thermal processes in any type of industry.
Heat loads and processing times, and industrial equipment sizes are influenced by the
heat capacity of the material. Combined with thermal conductivity and thermal dif-
fusivity, heat capacity data are needed for the modelisation of the thermal processes.
Heat capacity varies with temperature, composition and also water content. As the
material can be under the solid or the liquid form, different ways of measuring heat
capacity using the calorimetric techniques have been developed.
Heat capacity is thermodynamically defined as the ratio of a small amount of heat
δQ added to the substance, to the corresponding small increase in its temperature dT:

C = δQ/dT

For processes at constant pressure, the heat capacity is expressed as:

Cp = (δH/δT)p

Though DSC is a very well adapted technique to measure heat capacity [52], only
one procedure has been essentially developed using a continuous heating mode for
solid samples. In this chapter, another procedure is described using a step heating
mode.
68 P. L. Parlouër

cp Determination in the Temperature Scanning Mode

If there is no transformation for the considered temperature range, the calorimetric

signal for a given mass of sample heated at a constant heating rate dT/dt is relative
to the following relation for the sample side:

(dq/dt)s = (ms cp(s) + mcs cp(cs) )dT/dt (2.6)

where:
• ms and mcs respectively sample mass and vessel mass (including the cover),
• cp(s) and cp(cs) respectively specific heat capacity of sample and its vessel.
On the reference side, an empty vessel is used giving the corresponding signal:

(dq/dt)r = mcr cp(cr) dT/dt (2.7)

where:
• mcr respectively reference vessel mass,
• cp(cr) respectively specific heat capacity of reference vessel (equal to cp(cs) as the
nature of the vessel is identical).
The differential calorimetric signal is given by the following relation:

(dq/dt)s = (ms cp(s) + mcs cp(cs) − mcr cp(cr) )dT/dt (2.8)

In order to get rid of the thermal effect of both vessels, the same test (called
blank test) is run with identical empty containers. From a practical point of view, the
reference vessel is not removed from the calorimeter. The corresponding relation is
obtained:
(dq/dt)b = (mcs cp(cs) − mcr cp(cr) )dT/dt (2.9)

By subtracting the two calorimetric traces, the specific heat capacity of the sample
is extracted (Fig. 2.15):

Cp(s) = (1/ms )[(dq/dt)s − (dq/dt)b ]/dT/dt. (2.10)

As seen before in the calibration paragraph, the calibration using the Joule effect
technique allows converting any calorimetric signal in mW without the need of
standard reference materials. That means that in relation (2.10) all the parameters
(sample mass, calorimetric signals, heating rate) are accurately known to determine
the specific heat capacity of the sample Cp(s) (expressed in J.g−1 .◦ C−1 ) at a given
temperature. The variation of Cp(s) versus temperature can be determined.
With the DSC technique, a third test is needed using a standard reference material
(sapphire) that has a known specific heat capacity.
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 69

Fig. 2.15 cp determination Heat flow (mW)

in the temperature scanning
mode

Ab
As

T
time

Fig. 2.16 cp determination in Heat flow (mW)

the step heating mode Qb

Qs

ΔT
T

time

cp Determination in the Temperature Step Mode

The previous described technique is very easy to use, but shows some drawback as
soon as the accuracy of the cp determination is concerned. With the temperature scan-
ning mode, the sample is continuously heated and is never at the thermal equilibrium.
However cp is a thermodynamical parameter, defined at the thermal equilibrium.
The temperature step mode has been developed to answer this limitation. A step of
temperature is applied to the sample and the thermal equilibrium (characterized by
the signal return to the baseline) is waited after each step (Fig. 2.16).
If the relation (2.8) is integrated between time to (beginning of the step) and time
tn (return to the baseline):

∫(dq/dt)s dt = ∫[(ms cp(s) + mcs cp(cs) − mcr cp(cr) )dT/dt]dt. (2.11)

the corresponding equation is obtained:

to = (ms c̄p(s) + mcs c̄p(cs) − mcr c̄p(cr) )T

[Q]tn (2.12)

where cp corresponds to the mean cp value between the two temperatures defining
the step of temperature.
70 P. L. Parlouër

Q is obtained by integrating the corresponding surface defined by the calorimetric

signal between to and tn .
If again the signal corresponding to the blank test is subtracted when an identical
step of temperature is applied, the final relation giving the mean cp of the sample is
obtained:
Cp(s) = (1/ms )[Qs − Qb ]/T.

In that case, the cp value is obtained by a difference of surface. That means that the
result is not depending on fluctuations of the baseline between the different tests that
can be seen with the previous technique.

2.5 The Calorimetric Techniques

The DSC technique, described in the previous paragraph, has a lot of advantages but
also some drawbacks as soon as it is needed to work on larger amounts of sample,
to investigate gas or liquid interactions, to simulate mixing or reactions between two
or more components, to work with higher sensitivity.
A calorimeter is mainly characterized by a measurement chamber surrounded by a
detector (thermocouples, resistance wires, thermistors, thermopiles) to integrate the
heat flux exchanged by the sample contained in an adapted vessel. The measurement
chamber is insulated in a surrounding heat sink, made of a high thermal conductivity
material.
The main improvement with calorimetry is that it becomes possible to increase
the size of the experimental vessel, and consequently the size of the sample without
affecting the accuracy of the calorimetric measurement. According to this funda-
mental property, calorimeters of different sizes have been produced to adapt various
types of applications. The calorimeter offers an experimental space in which differ-
ent types of vessels are designed to especially make possible the investigations of
interactions between solid and liquid materials.

2.5.1 Calorimetric Principles

Many different types of calorimeters are commercially available and most of the
calorimetric principles are summarized in Table 2.4:
The main calorimetric modes can be resumed using Fig. 2.17,
where:
• A is the sample contained in a vessel,
• B is the thermostated jacket (liquid or metallic),
• Ti is the temperature of the sample, and
• Tj is the temperature of the jacket.
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 71

Table 2.4 Different calorimetric principles

Property Technique Description
Heat Isothermal calorimetry In an isothermal calorimeter, a defined exchange of
heat occurs between the sample and the
surroundings at a given environment
temperature. The heat flow rate is determined on
the basis of the temperature difference along a
thermal resistance between the sample and its
surroundings.
Heat Isoperibolic calorimetry In an isoperibolic calorimeter, the temperature of the
surroundings remains constant, while the
temperature of the sample can differ from the
surrounding temperature.
Heat Adiabatic calorimetry An adiabatic calorimeter is used to study chemical
reactions and is one in which there is no net heat
gain or loss during the chemical reaction.
Heat Calvet Calorimetry A Calvet calorimeter is a heat exchanging
calorimeter with a cylindrical type detector
working in isothermal and scanning modes.
Heat Titration calorimetry In a titration calorimeter, a small volume of solution
A is injected into the calorimetric vessel
containing the solution B. The corresponding
heat of interaction is measured.
Heat Flow calorimetry In a flow calorimeter, the sample is a flowing liquid.
Heat Solution calorimetry In a solution calorimeter the mixing of the samples is
performed in a static (batch) mode.
Heat Reaction calorimetry A reaction calorimeter is designed for the
investigation of reactions between liquids or
solids. The calorimetric technique can be
isothermal, isoperibolic or adiabatic.
Heat Bomb calorimetry A bomb calorimeter is designed for measuring heat of
combustion.
The sample is contained in a heavy-walled metal
container containing oxygen under pressure,
followed by ignition to obtain the heat of combustion
of the sample.

In the isothermal mode, the sample temperature Ti has to remain constant. To reach
this target, the temperature of the jacket Tj is permanently adjusted. In this situation,
a heat flux is measured between A and B and relates to any transformation or reaction
that occurs in A.
In the adiabatic mode, the temperature of the sample Ti and the thermostated jacket
Tj are kept at the same temperature in order to prevent any heat transfer between A
and B. In that case, the temperature of the sample increases if the occurring reaction
in A is exothermic or decreases if the reaction is endothermic.
The isoperibolic principle is an intermediate mode where the temperature of
the thermostated jacket Tj is maintained at a constant temperature. The sample
72 P. L. Parlouër

Fig. 2.17 Sample and ther-

mostated jacket in calorimetric
TI
principles

TJ

temperature Ti varies according to the transformation in A. The difference of tem-

perature is measured and is related to the heat generated by the transformation.
According to the catalyst to be investigated or the catalytic process to be simulated,
the calorimetric mode, and accordingly the type of calorimeter, has to be carefully
selected.

2.5.2 Isothermal Calorimetry

The Calvet type calorimeter is one of the most commonly used for the catalytic
investigations. Different types of experimental vessels have been developed to fulfill
the different experimental requirements.

2.5.2.1 Calvet Principle

The Calvet principle has already been described in Sect. 2.4.2.

To understand the direct correlation between the electrical signal and the heat
flux, it is needed to consider that a power W is fully dissipated in a calibration
vessel surrounded by a fluxmeter composed of crowns of thermocouples (Fig. 2.11).
An elementary power wi is dissipated through each thermocouple producing an
elementary variation of temperature Ti between the internal and external weldings
(Fig. 2.18):
wi = δi Ti (2.13)

where δi is the conductance of the thermocouple.

The corresponding variation of temperature generates an elementary electromo-
tice force (emf) according to the Oersted law:
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 73

Fig. 2.18 Joule effect cali-

bration principle

ei = εi Ti (2.14)

where εi is the thermoelectric constant of the thermocouple.

By combining the relations (2.13) and (2.14), and considering that all the thermo-
couples are in series:
E = ei = (εi /δi )wi (2.15)

As all the thermocouples are identical and made from the same materials, the rela-
tion (2.15) can be expressed as:

E = (ε/δ)wi or E = (ε/δ)W (2.16)

The relation (2.16) shows that the power dissipated in the vessel is directly correlated
to the heat flux. The term ε/δ corresponds to the calibration factor of the calorimeter.
This relation provides the correlation between the electrical output of the calorimetric
detector and the corresponding heat flux exchanged with the sample.
The main advantages of this type of calibration:
• it is an absolute calibration;
• it is not needed to use metallic reference materials;
• the calibration can be performed at a constant temperature, in the heating mode
and in the cooling mode;
• it can apply to any experimental vessel volume; and
• it provides a very accurate calibration of the calorimetric detector.
According to the Calvet principle, many different calorimeters have been designed
with various temperature ranges, from small to large size volumes, with a large
variety of sensitivity. In the next paragraph is more precisely described one Calvet
74 P. L. Parlouër

calorimeter, produced by the SETARAM company, and that is used worldwide for
catalytic investigations.

2.5.2.2 The C80 Microcalorimeter

The C80 microcalorimeter (produced by the SETARAM company) is a versatile tool

in the field of catalytic investigations as it works according to different calorimetric
modes:
• Isothermal calorimetry.
• Scanning calorimetry.
• Gas adsorption calorimetry.
• Liquid adsorption calorimetry.
• Mixing calorimetry.
• Percolation calorimetry.
• Reaction calorimetry.
• Pressure calorimetry.
In a Calvet calorimeter, the most important part is the thermoelectric element that
provides the performances of the calorimetric detector.
The microcalorimeter is built around a high metallic conductive block with two
cavities containing the thermopiles. Each thermopile is made of crowns of thermo-
couples, mounted on a metallic tube, and defines the experimental zone (Fig. 2.19).
The two thermopiles (measure and reference) are inserted in a thermostated heating
block that fixes the temperature of the calorimeter (Fig. 2.20). The block itself is sur-
rounded by the heating element and arranged in an insulated chamber. The detectors
define the experimental zone, in which the vessels are very tightly introduced. The top
of the calorimeter is removable in order to make possible the fixation of dedicated ves-
sels (mixing) inside the calorimeter and also for the arrangements of temperature pre-
stabilisation features for fluid samples before introduction in the calorimetric zone.
The microcalorimeter is characterized by a large experimental volume (15 cm 3 )
that has allowed the design of a large variety of experimental vessels according to
investigations to be carried out.
The main characteristic of the calorimeter is to be fitted on a rotating mechanism.
A special mixing vessel has been designed to be used in such experimental conditions.
The microcalorimeter offers a large choice of experimental vessels according to
the applications to be run:
• Isothermal and scanning calorimetry.
– The batch standard vessel is designed for investigating any type of transforma-
tion at a constant temperature, when heating or cooling large volume of samples
in the solid or liquid form. It is also dedicated to the determination of heat
capacity.
• Pressure calorimetry
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 75

Fig. 2.19 Cross-section of

the C80 thermopile

Fig. 2.20 Cross section of the

calorimeter

– The batch high pressure vessel is designed for the simulation of reaction and
decomposition under pressure in a closed vessel or under controlled pressure
(max : 100 bar). It is used to define the safety conditions of reaction operations
but also investigations under supercrital gas conditions (e.g. adsorption of CO2
on zeolites).
76 P. L. Parlouër

A dedicated very high pressure vessel (max 350 bar) is connected to a high
sensitivity pressure sensor through a special capillary tube. This vessel allows
to record simultaneously the heat and pressure released by the sample during
the reaction as a function of time or temperature [32].
For studies of adsorption at very high pressures (e.g. formation of gas hydrates),
specific vessels are able to operate up to 1000 bar.
• Gas adsorption calorimetry.
– The gas circulation vessel is fitted with two coaxial tubes and is used to produce
a circulation of gas (inert or active) around the sample. It is especially conve-
nient for the investigation of adsorption/absorption on a catalyst under normal
pressure of reactive gas such as hydrogen, ammonia, CO, . . . [18, 22, 35] or
high pressure [15].
Such a vessel can also be fitted with a relative humidity generator in order to
introduce a humid gas with a known rate of humidity on a solid sample.
This type of vessel has especially been used by Auroux [5, 14] to adapt a volumetric
gas line to the calorimeter that is described on Fig. 2.21. Using such a calorimetric
and volumetric device, the determination of the surface acidity and basicity of various
types of zeolites and related materials was performed [6].

• Mixing and reaction calorimetry.

– The mixing vessel using the rotating mechanism is divided in two chambers and
separated by a metallic lid (Fig. 2.22). The two materials are placed, one in the

Fig. 2.21 Calorimetric et volumetric devices coupling (from Ref. [14])

2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 77

Fig. 2.22 Mixing vessel (by

reversing)

lower chamber (i.e. powder) and the other in the upper chamber (i.e. liquid).
The mixing of the two components is obtained by rotating the calorimeter, the
metallic lid acting as a stirrer. This mixing vessel is designed for investigating any
liquid–liquid mixing (dilution, neutralization, reaction, . . .), solid–liquid mixing
(dissolution, hydration, wetting, reaction, . . .)
– The membrane mixing vessel is especially dedicated to the mixing of viscous
samples and for applications when the rotation of the calorimeter cannot be used.
In such a vessel, the separation between both chambers is performed with a very
thin membrane (metal or PTFE) (Fig. 2.23). The vessel is fitted with a metallic rod
that is operated from outside the calorimeter. In this situation, the mixing of both
components is obtained by pushing the rod in order to break the membrane. The
rod is also used as a stirrer during the test. The applications are similar to the ones
previously described.

• Sorption calorimetry.
– The ampoule sorption vessel is designed for slow dissolution process and for
wetting operation [43, 56]. The sample is introduced inside a special glass
ampoule the upper part of which is connected to a vacuum pump and the bottom
part of which is very thin. After outgassing during 30 min, the ampoule is sealed
with a torch and the upper part of the ampoule is cut and withdrawn. The vacuum
operation allows desorbing the surface of the solid sample (especially powder)
and makes easier the interaction. The sealed ampoule is introduced in the vessel
78 P. L. Parlouër

Fig. 2.23 Mixing vessel

(metallic membrane)

containing the solution. By breaking the ampoule using a piston rod, the solid
and liquid samples are put into contact
• Percolation calorimetry.
– The percolation vessel is used for adsorption of liquid on a powder (especially
catalyst) [41]. However when a liquid is flowed through a powder, the first
interaction that occurs is wetting of the powder followed by adsorption. In
order to make distinct these two thermal interactions, a special calorimetric
design has been developed with the percolation vessel. The powder is introduced
in a metallic cylinder on a sintered metallic section. A first carrier liquid is
flowed through the powder to get the wetting (Fig. 2.24). Then the pump is
switched to the liquid solute for the adsorption phase. According to the reaction,
it will be possible to be back with the first carrier liquid in order to proceed
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 79

Fig. 2.24 Experimental set-up for percolation calorimetry

to the desorption phase. This calorimetric vessel is very convenient for the
investigations of surface reactivity, catalyst contamination, chemisorption under
normal and high pressure.
• Flow mixing calorimetry.
– The use of the chemical sorbents is today one of the most popular absorption
technique for the CO2 capture in postcombustion techniques. In such an indus-
trial process, the amine solution is introduced at the top of an absorption tower
while the exhausted fume containing carbon dioxide is introduced at the bottom.
As an intimate contact is reached in the absorption tower, the amine solution
chemically absorbs the carbon dioxide from the gaseous stream. Such a process
especially requires two types of thermodynamic parameters: gas solubility and
enthalpy of absorption. The enthalpy of absorption, according to the amount
of absorbed gas and the corresponding heat capacities of solutions, define the
temperatures of the fluids when they exit the absorption columns. Flow mixing
calorimetry is the ideal technique for measuring such enthalpies of absorption
[2, 28]. In order to work under pressure, a dedicated high pressure mixing ves-
sel is adapted to be used on the Setaram C80 calorimeter. The mixing vessel is
made of a stainless steel tube in a helicoidal shape into a cylindrical container
(Fig. 2.25). The length of the tube in closed thermal contact with the cylinder is
about 240 cm. The fluids (CO2 and amine solution) are introduced at the bottom
80 P. L. Parlouër

Fig. 2.25 The high pressure Outlet A+B

flow mixing vessel

Preheater

Fluid A

Fluid B

Mixing device

part of the vessel in two vertical and concentrical tubes. The mixing (dissolu-
tion, reaction) starts when the thinner part of the tube is reached. The heat that
is associated with the reaction, is exchanged between the rolled tube and the
calorimetric block through the wall of the vessel in an isothermal mode.
The flow mixing vessel operates from room temperature to 200 ◦ C and for a
range of fluid pressure from 0.1 to 20 MPa. The fluid flowrates vary from 50 to
1500 μ · min−1 , that allow to cover a wide range of mixture composition.

2.5.3 Isothermal Titration Calorimetry

The different mixing vessels that are described in the previous paragraph are designed
for the interaction between two components and are not well adapted for titration
studies. In the Isothermal Titration Calorimetry, aliquots of sample B is added in
a volume of sample A, located in the calorimetric vessel and maintained under a
constant stirring (Fig. 2.26).
Another sample C can also be added according to the reaction to be simulated.
To allow such experimentations, the C80 calorimetric block is modified to adapt
a magnetic stirring system at its bottom part. This leaves more space to adapt the
different liquid and gas introduction lines. A heating cover has also to be used for the
thermostatisation of the fluids at the temperature of the calorimeter. The liquids are
introduced through a motor-driven syringe pump that allow a continuous or a step
injection of defined volumes.
Such a calorimetric technique is especially designed to follow the liquid adsorp-
tion of organic compounds on catalysts or zeolites [12, 21, 49]. The adsorption of
n-butylamine (in decane) on a zeolite (in a solution of decane) illustrates the calori-
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 81

Fig. 2.26 Principle of the

Titrys calorimeter

metric titration mode. After activation of the zeolite at 400 ◦ C, the calorimetric test
is run at 40 ◦ C. Aliquots of 0.25 ml of adsorbant are added until the saturation of the
zeolite is reached (Fig. 2.27).
Such a vessel is adapted to follow an organic reaction when several compounds
have to be added during the process.
It is also used to simulate the injection of very small amounts of water to produce a
hydrolysis reaction (hydrolysis of ammonia borane for the production of hydrogen).

2.6 The Thermogravimetric Technique

2.6.1 Principle

Thermogravimetric Analysis (TGA) or Thermogravimetry (TG) is defined as fol-

lowed by ICTAC: “A technique in which the mass of the sample is recorded versus
time or temperature while the temperature of the sample is programmed, in a con-
trolled atmosphere. The instrument is called a thermogravimetric analyzer (TGA) or
a thermobalance.” [17].
Two different types of transformations can occur:
• Transformation with a mass loss: dehydration, dehydroxylation, evaporation,
decomposition, desorption, pyrolysis, . . .
• Transformation with a mass gain: adsorption, hydration, reaction, . . .
The Fig. 2.28 shows a classical mass loss with the determination of the derivative
curve (DTG).
82 P. L. Parlouër

Fig. 2.27 Adsorption of n-butylamine (in decane) on a zeolite at 40 ◦ C

Fig. 2.28 A TG mass loss

curve with the derivative
curve (DTG)

If mi is the initial mass and mf the final mass, the mass loss is equal to (mi − mf ) or
(mi − mf )/mi in percentage.
It is interesting to notice that the DTG peak is very similar to the DTA/DSC peak (see
also paragraph on couplings). The integration of the DTG peak also gives access to
the variation of mass.
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 83

The DTG peak is related to the kinetics of the transformation, as the DTG signal
corresponds to a mass variation rate. The shape of the peak directly informs on the
rate of the transformation.

2.6.2 Detectors

Different types of TGA techniques are available and are described in Table 2.5.
A thermobalance is built around a highly sensitive balance module, a furnace and
a controlled atmosphere cabinet. Different types of balance detectors are available
but the most commonly used is based on the principle of the null position balance.
An example is given in Fig. 2.29.
The balance is made of an articulated beam, suspended on a torsion wire. A rod
is fixed in the middle of a beam and is coming with a window in the upper part.
Four magnets are mounted in the middle part of the rod. Each of them is plunging
in a solenoid fixed on the mechanical frame. On the same frame, a device of optical
detection delivers a light that is fully crossing the slot of the window. The light signal
is measured by means of a photodiode.
When there is a mass variation inducing a movement of the balance beam, the light
is partially hidden. A compensation current is sent through the solenoids in order
to have the beam back to the null position. With such a detection principle, the
mass variation is proportional to the compensation current. It is positive or negative
depending if it is a mass loss or a mass gain.
Based on this principle, different types of thermobalances are available: top load-
ing balance, bottom loading balance, horizontal balance.

Table 2.5 Different types of TGA techniques

Property Technique Description
Mass Mono TGA A technique in which the change of sample mass is
analysed while the sample is subjected to a
temperature variation. A mono TGA has only
one furnace, for the sample.
Mass Symmetrical TGA A technique in which the change of sample mass is
analysed while the sample is subjected to a
temperature variation. A symmetrical TGA has
two furnaces, one for the sample and one for the
reference material.
Mass High pressure TGA A high pressure TGA is a TGA (mono or
symmetrical) working with the test sample under
pressure.
Mass Corrosive TGA A corrosive TGA is a TGA (mono or symmetrical)
working with the test sample in a corrosive
atmosphere.
84 P. L. Parlouër

Fig. 2.29 Principle of a null position balance (SETARAM balance)

The nature of the furnace heating element depends on the range of temperature
for the investigations. The most used materials are metallic (nickelchrome, kanthal,
platinum, tungsten) but also SiC and graphite.
As for the DTA experimentation, the choice of the crucible for the TGA test has
to be done very carefully. The main point is to select a material that will not react
with the sample, especially at high temperature;
• Aluminum (up to 500 ◦ C),
• Silica (especially used below 1000 ◦ C as it is very inert and easy to clean),
• Alumina (especially required for metallic samples),
• Platinum (especially required for inorganic samples),
• Graphite and tungsten (for temperature above 1750 ◦ C).
Other types of materials are available such as magnesia, zirconia, boron nitride,
according to the sample under investigation.
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 85

2.6.3 Operation

2.6.3.1 Selection of the Atmosphere

As the crucible is open for the TGA test, the atmosphere around the sample has to
be carefully controlled, especially in the case of catalysts. Different gas options are
available to work in a thermobalance:
• Operation in inert gas:
– This is the most common situation where it is needed to protect the sample from
oxidation during the heating process. A vacuum purge is recommended prior to
the introduction of the inert gas in order to empty the balance, the furnace and
all the gas lines.
If it is needed for desorbing the sample prior to the heating, a primary vacuum
or even a secondary vacuum pumping is to be used for this operation before the
introduction of the inert gas.
Most of the investigations on catalysts need a good initial desorption of their
surfaces before starting the adsorption investigations.
• Operation in reactive gas:
– During catalytic investigations, different types of reactive gases are used and it
is needed to know the risk of interactions with the different metallic and ceramic
parts of the thermobalance. Here is the situation for the most common reactive
gases:
Air, oxygen for oxidation studies up to 1750 ◦ C without specific problem.
Reactive gas such as butane for investigations of catalytic reactions [36].
Hydrogen: with such a gas two main concerns have to be considered: the
concentration and the corrosion.
A non-explosive mixture of less than 4 % H2 in an inert gas is recommended.
For higher concentrations, a safety device is required on the thermobalance. The
use of H2 requires that no trace of air is present in the thermobalance to avoid
an explosion.
Above 1000 ◦ C, H2 becomes a poison for platinum that is becoming brittle. The
platinum–rhodium temperature control thermocouple has to be replaced to work
above this temperature. Tungsten rhenium thermocouple is one option.
• Operation in corrosive gas:
– Adsorption or reaction under corrosive gas, such as CO, NH3 but also halogens
(chlorine, fluorine), can be investigated using the TGA technique but caution
has to be taken to avoid the contact between the corrosive gas and any metallic
part of the thermobalance (especially the thermocouple).
Dedicated experimental set-up need to be used to guarantee such a protection,
as described on Fig. 2.30. A silica tube, with a restriction in the lower part, is
introduced in the furnace. This tube shape allows having the thermocouple out
86 P. L. Parlouër

Fig. 2.30 TGA corrosive set-

up

of the corrosive gas stream. The protection of the balance has to be carried out
with an inert gas. The crucible and the suspension also are in silica. With such
an adaptation it is possible to work with TGA with any type of corrosive gas,
even at high concentration up to 1000 ◦ C.
• Operation in humid atmosphere:
– Adsorption of water vapour at a given pressure or at a relative humidity percent-
age is an important test to characterize the capabilibity of solid sorbents to fix
water. It is also known that water vapour will enhance the adsorption of CO2 on
solid sorbents.
To perform such a measurement a relative humidity generator has to be con-
nected to the TGA through a thermostated gas line (Fig. 2.31).
A dry gas is splitted in two lines, one remaining dry and the other one saturated
with water. The two gas streams are mixed in a chamber. According to the desired
relative humidity, the respective flowrates are adjusted for the gas mixture. The
humid gas resulting from this operation is then transferred to the TGA via a
heated transfer line.
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 87

Fig. 2.31 Principle of the Wetsys relative humidity generator (Setaram)

2.6.3.2 Buoyancy Effect

The buoyancy effect is most of the time a problem that is misunderstood in the TGA
measurement or often hidden by the numerical treatment of the TGA curves.
Buoyancy is the upward force exerted on an object when it is immersed, partially or
fully, in a fluid (liquid or gas). Its value is equal to the weight of the fluid displaced by
the object. In the case of the TGA experiment with a given crucible (the object), this
results in an apparent increase of the mass when the sample is heated. The effect is
observed on all conventional balances. In fact the term buoyancy includes different
parameters: the true buoyancy as described before, the convection currents, the gas
flow drag effects, the gas velocity effects, the thermomolecular forces, the thermal
effects on the balance mechanism.
The buoyancy, according to these differents factors, is especially significant at low
temperature and will decrease at high temperature. When a mass variation has to be
accurately measured at low temperature (for example water content), a correction
of the buoyancy has to be performed. The common way is to run a blank test with
an empty crucible with the same experimental conditions. However this numerical
correction remains dependent on the reproducibility of such a blank curve and the
correction is affected with a certain uncertainty.
Another way to solve the problem is to use a symmetrical thermobalance.
In such a configuration (Fig. 2.32), the crucibles containing respectively the sample
and the inert reference material are hung on each side of the balance. One or two
furnaces are used according to the symmetrical model.
If the gas flowrates are adjusted on both sides, the buoyancy effect coming from the
sample and the reference, as it is identical, is compensated.

2.6.4 Applications

As described in the previous paragraphs, the TGA technique provides a wide range
of applications for the investigation of catalysts and related material:
• Gas adsorption and desorption.
88 P. L. Parlouër

Fig. 2.32 A symmetrical dual

furnace thermobalance

• Decomposition.
• Dehydration and dehydroxylation.
• Oxidation.
• Preparation of catalysts.
• Regeneration of catalysts.
• Investigation under various reactive gas (H2 , CO, . . .).
• Investigation under relative humidity.
• Investigation under corrosive gas.
In a more general way, thermogravimetry, with simple and short experiments allow
preliminary screenings of catalysts where multiple variables are being considered. In
just one experiment, the capacity of an adsorbent can be evaluated over an entire tem-
perature range. It is also possible to collect qualitative information about the initial
adsorption rates. Thermogravimetry, which has been applied to the preparation and
characterization of adsorbents, has also proved to be a useful technique for prelimi-
nary adsorption capacity assessment. This is especially the case for the CO2 capture
performance of the sorbents and their thermal stability. For example, the evaluation
of aminated solid sorbents for the CO2 capture performance was evaluated [47].
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 89

Thermogravimetry is a very useful tool to evaluate the performances of catalytic

filters used for the removal of soot from diesel exhaust gas [42].
TGA is also very widely used to investigate the NOx -carbon reactions to control the
emission of NOx during the combustion process [60].
More recently, thermogravimetry has been applied to the evaluation of the
chemical-looping combustion (CLC) process. It is considered as an energy-efficient
method for the capture of carbon dioxide from combustion and provides an advantage
of no energy loss for CO2 separation without NOx formation. This process consists of
oxidation and reduction reactors where metal oxides particles are circulating through
these two reactors [29, 54, 58]. The TGA set up for such an application is described
on Fig. 2.33.

2.7 The Simultaneous Techniques

The TGA technique provides information on the mass variations that occurs within
the sample, but most of the time, if the composition of the sample is unknown, the
interpretation of the data are not easy. This is why the TGA technique is mostly used
combined with other techniques such as DTA or DSC [59].

Fig. 2.33 Schematic diagram of a thermal gravimetric analyzer (TGA) system for the reactivity
test; 1 water reservoir and bubbler, 2 gas pre-heater, 3 electric balance, 4 sample basket, 5 heater, 6
personal computer (from Ref. [29])
90 P. L. Parlouër

The TG-DTA or TG-DSC combination allows obtaining the simultaneous TGA and
DTA or DSC measurements on the same sample (Fig. 2.34). The heat effects are
correlated with the mass variations. The DTA or DSC signal gives information on
heat effects not correlated with mass variation (melting, phase transitions, . . .)
The main interest of such a simultaneous measurement is that only one sample is
used. This is especially very important for the characterization of catalysts as sample
size (mass, surface, porosity, and composition), the gas flowrate and composition are
critical parameters for an accurate determination.
Another important factor of the combination is that any endo or exo effect is related
to the mass fraction that has been adsorbing or reacting.
The derivative DTG curve is similar to the DTA or DSC curve for the transformation
with mass variation. Both curves are used for the kinetic interpretation of the sample
transformation.
As an example of the interest of such a coupling technique, the investigation of
the hydrogenation of a poisoned catalyst is given (Fig. 2.35).
A sample of dead catalyst is heated under pure hydrogen using a protected DTA rod.
The combination of both techniques allows to understand the different processes that
are occurring during the regeneration reaction:
• The first mass moss is associated with a small endothermic effect corresponding
to the degradation of the organic compounds adsorbed on the catalyst.
• The second mass loss gives rise to a significant exothermic effect, due to the
reduction of sulphur contained in the poisoned catalyst and producing H2 S.
The applications of the TG-DTA (or DSC) technique are very similar to the ones
developed in the previous paragraph. However the DTA (or DSC) technique allows
to give more information on the thermal properties of the catalyst, for example the
characterization of materials after synthesis [1, 3, 27].

Fig. 2.34 The TG-DTA (or

DSC) curve
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 91

Fig. 2.35 TG, DTG and DTA curves of the hydrogenation of a poisoned catalyst

In most TG-DTA or DSC instruments, the DTA or DSC probe is attached to the
balance. Such a combination results in a reduction of the performances of the TGA
and the DTA or DSC detectors compared to the characteristics when they are used
alone.
The Sensys TG-DSC is an exception in the field as the TGA and the DSC detectors
are not mechanically attached (Fig. 2.36).
The Sensys TG-DSC is based on the Calvet type DSC (Fig. 2.36) used in the verti-
cal position [30]. On top of the DSC, is adjusted a symmetrical balance corresponding
to the principle described on Fig. 2.29. The crucibles containing the sample and the
inert material are hung on each side of the balance and introduced in the calorimetric
zone of the DSC without touching the walls. In such a situation, the crucibles are
fully surrounded by the fluxmeters, providing an accurate DSC determination. In the
same time, the symmetrical balance allows a compensation of the buoyancy effect
resulting on a very high sensitive TG determination.
This technique is very powerful for the investigation of gas interaction on solid or
liquid sample, such as adsorption, absorption, desorption, catalytic reaction under
various types of gas (inert, oxidative, reducing, corrosive, humid) [55], the synthesis
and catalytic behaviour of catalysts [53].
An example of the use of such a TG-DSC combination is given on the adsorption
and desorption of CO2 on a catalyst (Fig. 2.37). After preparation of the sample under
helium, a mixture of helium (90 %) and CO2 (10 %) is introduced on the catalyst. The
exothermic effect is related to the amount of CO2 adsorbed on the catalyst. When
92 P. L. Parlouër

Fig. 2.36 Cross section of the

Sensys TG-DSC

switching back to helium, it is noticed that desorption of the catalyst is not complete
at this temperature.
The TG-DSC technique was also used for the investigation of amminoborane for the
production of hydrogen [7]

2.8 Evolved Gas Analysis

Evolved Gas Analysis (EGA) is defined as followed by ICTAC:

“A technique of determining the nature and amount of volatile product or products
formed during thermal analysis.” [40].
As indicated in the previous chapter, the TGA technique only informs on the
variation of sample mass, but hardly on the mechanisms linked to the decomposition
or the reaction. The idea behind the EGA coupling is to analyze the vapours emitted
by the sample during the test [39]. Three main gas coupling techniques are used for
such an investigation as described in Table 2.6:
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 93

Heat TG/m
Ex
2.

ΔH = -233 mJ

Δ H = 218.1

1. -

100% He 90% He + 10% CO2 100% He

1.
Δm = -236
Δm = +298 -

0 3 6 9 120 15 Time/mi

Fig. 2.37 Adsorption/desorption of CO2 on a catalyst at 40 ◦ C (Sensys TG-DSC)

Table 2.6 The main EGA techniques

Mass and MS spectrum TGA-MS In TGA-MS, the TG is connected online with a mass
spectroscopy detector to determine the nature and
amount of volatile product emitted by the sample
during a thermogravimetric experiment
Mass and FTIR spectrum TGA-FTIR TGA-FTIR is a thermogravimetric analyzer (TGA)
coupled online with a Fourier-Transform Infrared
(FTIR) spectrometer to determine the nature and
amount of volatile product emitted by the sample
during a thermogravimetric experiment
Mass and GC chromatogram TGA-GC In TGA-GC, the TGA is connected online with a gas
chromatograph(GC) detector to determine the
nature and amount of volatile product emitted by
the sample during a thermogravimetric
experiment

2.8.1 TG-MS Coupling

During a long period, the main difficulty in the TG-MS coupling has been to define
the right interface that will solve the following technical coupling problem:
94 P. L. Parlouër

• Pressure in the balance is equal to 1 bar, and about 10−6 bar in the mass spectrom-
eter. The pressure has to be reduced by 6 decades during the gas transfer.
• Gas sampling has to be representative of the gas emitted by the sample.
• Emitted gas has not to be degradated in the transfer line.
• Condensation of the emitted gas have to be prevented in the transfer line.
• Time of gas transfer has to be short (about 100 ms) in order to have a simultaneous
information between the TGA and the MS curves.
To solve these different technical requirements, two different interfaces are today
available, according to the mass analysis to be performed:
• the capillary interface, and
• the skimmer or supersonic interface.

2.8.1.1 The Capillary Interface

The transfer line is made of a very thin silica capillary contained in a heated
jacket to prevent the condensation, especially at the outlet of the furnace (Fig. 2.38).
A fraction of the gas emitted by the sample is picked as closed as possible from the

Fig. 2.38 The heated capil-

lary interface (Setaram)
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 95

TG/% HF./µV
Exo 15

0
10

-20 5

0
-40

-5

-60
-10

200 400 600 800 Température /°C

Fig. 2.39 TG, DTG, DSC and MS curves of the calcium oxalate decomposition. TG and FTIR
curves of the calcium oxalate decomposition

TGA crucible to avoid gas dilution. The gas is then transferred through the capillary
towards the MS detector within 100 ms.
The MS detector will provide a spectrum for each injection. According to the sample
under study in the TGA, two situations can occur:
• The nature of the sample is known and it is possible to predict the types of gas
that will be emitted by the sample during the TGA experiment. For such a test,
characteristic mass numbers (amu), corresponding to the emitted gas, are selected
and their variation is recorded versus time to be correlated with the TGA and DTG
curves.
• The nature of the sample is unknown. In that case, it is needed to investigate
the different MS spectra, especially at temperatures (times) corresponding to the
different mass variations. When the significant mass numbers (amu) are identified,
their variation can be drawn as described previously.
An example of the use of the heated capillary is given with the investigation of
calcium oxalate CaC2 O4 , H2 O (Fig. 2.39)
The different uma traces show very clearly and respectively the different transfor-
mations that are expected and seen on the TG and DSC curves:
• the decomposition of the hydrate (H2 0+ );
• the decomposition of calcium oxalate in calcium carbonate (CO+ ); and
• the decomposition of the calcium carbonate in calcium oxide (CO+
2 ).
However the MS curves also allow to detect some CO2 emission during the decom-
position of the oxalate, and some CO during the decarbonatation.
96 P. L. Parlouër

It indicates the presence of some traces of air in the furnace that reacts with CO. Such
a reaction is slightly seen on the DSC curve (small exothermic deviation at the end of
the second endothermic peak) but it is not detectable on the TGA curve. This example,
even based on a very common sample decomposition, shows the contribution of the
MS coupling to the TG-DSC technique.
When coupled to MS, the thermogravimetric equipment can be used as a Temper-
ature Programmed Oxidation (TPO) or/and a Temperature Programmed Desorption
(TPD) technique to follow catalytic oxidation processes [62].
The TG-MS technique is also used to run Temperature Programmed Reduction (TPR)
experiments in reducing atmosphere [10].

2.8.1.2 The Skimmer or Supersonic Interface

The capillary interface is easy to install but its use remains limited to a certain range
of mass numbers. For high mass species (above 150 amu), condensation occurs in
the transfer line. To avoid this problem, the skimmer, also called supersonic interface
has to be selected.
Using a mass spectrometer in molecular mode requires to work under very low
pressures, in order to avoid any saturation of the quadrupole and to make the pro-
duction of the electron beam easier. Using the supersonic system, it is necessary to
have an important difference of pressure between the thermobalance and the mass
spectrometer, in order to obtain a free molecular beam.
A gas at a stable temperature and at atmospheric pressure is in a disturbed mole-
cular state. The rates of the molecules have an isotropic distribution in the space,
varying as the temperature square root varies. If a chamber under atmospheric pres-
sure and a chamber under low pressure are connected together through a small hole,
it is possible to obtain a thermal effusion or a supersonic beam. The isentropic pres-
sure reduction enables the molecules to reach some supersonic speeds in a few micro
seconds. Therefore, intermolecular recombinations are not possible and a molecu-
lar beam in only one direction is obtained. Pressure reduction releases the impact
between the molecules.
On a technical point of view, the supersonic system is adjusted on the thermobal-
ance furnace, in order to have the MS detector straight to the sample container
(Fig. 2.40).
Gas emitted by the sample flows through a taking hole at the top of a first cham-
ber working under primary vacuum. A second chamber, working under secondary
vacuum, is located inside the first chamber. The secondary chamber receives a cone
with a hole at the top. The chamber is in line with the quadrupole detector of the gas
analysis system.
The Supersonic or skimmer system is more costly and more difficult to use but it
gives many advantages on the capillary transfer line: no condensation in the transfer
line, no collision between the molecules and with the transfer line wall, possibility
to detect high mass species (above 500 amu)
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 97

Fig. 2.40 The supersonic interface (Setaram)

2.8.2 TG-FTIR Coupling

The TG-FTIR coupling is easier to carry out as both analyzers are working at the
same normal pressure. A heated transfer line has to be installed between the outlet
of the thermobalance and the inlet of the FTIR spectrometer.
As in the TG-MS coupling, the time of transfer is very short. Caution has to be taken
to avoid any condensation in the transfer line.
The infra red cell that is mostly KBr made of, has to be adapted (use of ZnSe cell)
in order to avoid degradation with water vapour.
The same investigation of calcium oxalate, described previously with the TG-
MS coupling, is used to see the difference between the two coupling techniques
(Fig. 2.41).
The characteristic wavelengths of the expected emitted gas are recorded together
with the three mass losses corresponding to the different steps of decomposition of
the calcium oxalate:
• H2 O (3734 cm−1 ) for the decomposition of the hydrate.
• CO (2068 cm−1 ) for the decomposition of the calcium oxalate in calcium carbonate.
• CO2 (2359 cm−1 ) for the decomposition of the calcium carbonate in calcium oxide.
98 P. L. Parlouër

CO 2 (2359cm-1)

CO (2068cm-1)

CO 2 (2359cm-1)
H2O (3734cm-1)

Fig. 2.41 TG and FTIR curves of the calcium oxalate decomposition

As seen before with the TG-MS coupling, a fraction of CO2 is detected during the
second loss, and a fraction of CO during the third mass loss, indicating the presence
of some trace of air in the furnace during the experiment.
On such an example, very similar results are obtained with both EGA techniques.

2.9 Conclusion

The thermal analysis and calorimetric techniques provide a very large variety of
possibilities of experimentations, of combinations with other analytic techniques that
make unlimited the number of applications, especially in the field of characterization
of catalysts and evaluation of catalytic processes.
The different techniques, together with the corresponding detectors and the experi-
mental vessels, can be adapted and adjusted to the investigations of new materials or
new processes. This is the case for example with the research on hydrogen storage
on metal hydrides, all the works related to the capture and sequestration of CO2 with
the evaluation of solid and liquid sorbents, the development of the Chemical Looping
Combustion with improved catalysts. . .
With this efficient faculty of adaptation to new research fields, the thermal analysis
and calorimetric techniques are promised to a brilliant future
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 99

References

1. M.A. Aramendia, V. Borau, C. Jiménez, A. Marinas, J.M. Marinas, J.A. Navio, J.R. Ruiz, F.J.
Urbano, Synthesis and textural-structural characterization of magnesia, magnesia–titania and
magnesia–zirconia catalysts. Colloids Surf. A: Physicochem. Eng. Aspects 234, 17–25 (2004)
2. H. Arcis, L. Rodier, J.Y. Coxam, Enthalpy of solution of CO2 in aqueous solutions of 2-amino-
2-methyl-1-propanol. J. Chem. Thermodyn. 39, 878–887 (2007)
3. A. Arnold, M. Hunger, J. Weitkamp, Dry-gel synthesis of zeolites [Al]EU-1 and [Ga]EU-1.
Microporous Mesoporous Mater. 67, 205–213 (2004)
4. A. Auroux, Thermal methods: calorimetry, differential thermal analysis and calorimetry
(Chap. 22), in Catalyst Characterization: Physical Techniques for Solid Materials, ed. by
B. Imelik, J.C. Vedrine (Plenum Press, New York, 1994)
5. A. Auroux, Acidity characterization by microcalorimetry and relationship with reactivity. Top.
Catal. 4, 71–89 (1997)
6. A. Auroux, Acidity and Basicity: Determination by Adsorption Microcalorimetry Molecular
Sieves (Springer, Berlin, 2006)
7. J. Baumann, F. Baitalow, G. Wolf, Thermal decomposition of polymeric aminoborane
(H2 BNH2 )× under hydrogen release. Thermochim. Acta 430, 9–14 (2005)
8. S. Bennici, A. Auroux, Thermal analysis and calorimetry methods, in Metal Oxide Catalysis,
vol. 1, ed. by S.D. Jackson, J.S.J. Hargreaves (Wiley, Weinheim, 2009)
9. J.J.P. Biermann, P.P. Coelen, H. Den Daas, F.J.J.G. Janssen, The use of a heat-flow differential
scanning calorimeter as a plug-flow fixed bed reactor in heterogeneous catalysis. Thermochim.
Acta 144, 329–337 (1989)
10. V.Y. Bychkov, Y.P. Tyulenin, V.N. Korchak, E.L. Aptekar, Study of nickel catalyst in oscillating
regime of methane oxidation by means of gravimetry and mass-spectrometry. Appl. Catal. A:
Gen. 304, 21–29 (2006)
11. K. Chakarova, K. Hadjiivanov, G. Atanasova, K. Tenchev, Effect of preparation technique on
the properties of platinum in NaY zeolite: a study by FTIR spectroscopy of adsorbed CO. J.
Mol. Catal. A: Chem. 264, 270–279 (2007)
12. F. Dan, M.H. Hamedi, J.P.E. Grolier, New developments and applications in titration calorime-
try and reaction calorimetry. J. Thermal Anal. Calorim. 85, 531–540 (2006)
13. J.L. Daudon, Heat flux devices and methods for optimum specific heat measurements, in
Fourteenth European Conference on Thermophysical Properties, Lyon, France, 1996
14. L. Dussault, J.C. Dupin, E. Dumitriu, A. Auroux, C. Guimon, Microcalorimetry TPR and
XPS studies of acid–base properties of NiCuMgAl mixed oxides using LDHs as precursors.
Thermochim. Acta 434, 93–99 (2005)
15. A.F.P. Ferreira, M.C. Mittelmeijer-Hazeleger, A. Bliek, Adsorption and differential heats of
adsorption of normal and iso-butane on zeolite MFI. Microporous Mesoporous Mater. 91,
47–52 (2006)
16. S.P. Felix, C. Savill-Jowitt, D.R. Brown, Base adsorption calorimetry for characterising surface
acidity: a comparison between pulse flow and conventional “static” techniques. Thermochim.
Acta 433, 59–65 (2005)
17. P. Gallagher, Thermogravimetry and thermomagnetometry, in Handbook of Thermal Analysis
and Calorimetry, vol. 1, Principles and Practice, ed. by M.E. Brown (Elsevier, The Netherlands,
1998)
18. A. Gervasini, P. Carniti, J. Kerainen, L. Niinisto, A. Auroux, Surface characteristics and activity
in selective oxidation of o-xylene of supported V2 O5 catalysts prepared by standard impreg-
nation and atomic layer deposition. Catal. Today 96, 187–194 (2004)
19. K. Habersberger, Application of thermal analysis to the investigation of catalysts. J. Thermal
Anal. 12, 55–58 (1977)
20. P.J. Haines, M. Reading, F.W. Wilburn, Differential thermal analysis and differential scanning
calorimetry, in Handbook of Thermal Analysis and Calorimetry, vol. 1, Principles and Practice,
ed. by M.E. Brown (Elsevier, The Netherlands, 1998)
100 P. L. Parlouër

21. M. Hart, G. Fuller, D.R. Brown, J.A. Dale, S. Plant, Sulfonated poly(styrene-co-divinylbenzene)
ion-exchange resins: acidities and catalytic activities in aqueous reactions. J. Mol. Catal. A:
Chem. 182–183, 439–445 (2002)
22. M.P. Hart, D.R. Brown, Surface acidities and catalytic activities of acid-activated clays. J. Mol.
Catal. A: Chem. 212, 315–321 (2004)
23. T. Hatakeyama, Z. Liu, Handbook of Thermal Analysis (Wiley, New York, 1998)
24. W. Hemmiger, S.M. Sarge, Definitions, nomenclature, terms and litterature, in Handbook of
Thermal Analysis and Calorimetry, vol. 1, Principles and Practice, ed. by M.E. Brown (Elsevier,
The Netherlands, 1998)
25. International Confederation for Thermal Analysis, in For Better Thermal Analysis and
Calorimetry, 3rd edn., ed. by J.O. Hill (RSC Publishing, Cambridge, 1991). http://www.ictac.
org
26. ISO 11357-1, Plastics: Differential Scanning Calorimetry—Part1: General principles, (Uni-
versidade De Aveiro, Aveiro, 1997)
27. D.S. Kim, J.S. Chang, J.S. Hwang, S.E. Park, J.M. Kim, Synthesis of zeolite beta in fluoride
media under microwave irradiation. Microporous Mesoporous Mater. 68, 77–82 (2004)
28. D. Koschel, J.Y. Coxam, L. Rodier, V. Majer, Enthalpy and solubility data of CO2 in water
and NaCl(aq) at conditions of interest for geological sequestration. Fluid Phase Equilib. 247,
107–120 (2006)
29. A. Lambert, C. Delquié, I. Clémençon, E. Comte, V. Lefebvre, J. Rousseau, B. Durand, Synthe-
sis and characterization of bimetallic Fe/Mn oxides for chemical looping combustion. Energy
Procedia 1, 375–381 (2009)
30. P. Le Parlouër, Simultaneous TG-DSC: a new technique for thermal analysis. Thermochim.
Acta 121, 307–322 (1987)
31. P. Le Parlouër, C. Mathonat, Sensys: an innovative concept for the Calvet DSC111 and TG-
DSC111, in 33rd NATAS Proceedings, Universal City, USA (2005), p. 44
32. X.R. Li, H. Koseki, Hazard evaluation of self-decomposition materials by the combination of
pressure and heat flux measurements. Thermochim. Acta 423, 77–82 (2004)
33. R.C. Mackenzie, Differential Thermal Analysis, vols. 1, 2 (Academic, London, 1972)
34. E. Manova, T. Tsoncheva, C. Estournès, D. Paneva, K. Tenchev, I. Mitov, L. Petrov, Nanosized
iron and iron-cobalt spinel oxides as catalysts for methanol decomposition. Appl. Catal. A:
Gen. 300, 170–180 (2006)
35. A. Maroto-Valiente, I. Rodrıguez-Ramos, A. Guerrero-Ruiz, Surface study of rhodium nanopar-
ticles supported on alumina. Catal. Today 93–95, 567–574 (2004)
36. V. Martin, J.M.M. Millet, J.C. Volta, Thermogravimetric analysis of vanadium phosphorus
oxide catalysts doped with cobalt and iron. J. Thermal Anal. 53, 111–121 (1998)
37. J.L. McNaughton, C.T. Mortimer, Physical Chemistry Series, vol. 2 (Butterworths, London,
1975)
38. J. Mercier, J. Thermal Anal. 14, 161 (1978)
39. L. Meublat, P. Le Parlouër, Couplage thermogravimétrie: analyse de gaz. Spectra 2000 19(161),
59–62 (1991)
40. J. Mullens, EGA: evolved gas analysis, in Handbook of Thermal Analysis and Calorimetry,
vol. 1, Principles and Practice, ed. by M.E. Brown (Elsevier, The Netherlands, 1998)
41. F.T.T. Ng, A. Rahman, T. Ohasi, M. Jiang, A study of the adsorption of thiophenic sulphur
compounds using flow calorimetry. Appl. Catal. B: Environ. 56, 127–136 (2005)
42. Y. Nguyen Huu Nhon, H.M. Magan, C. Petit, Catalytic diesel particulate filter: evaluation of
parameters for laboratory studies. Appl. Catal. B: Environ. 49, 127–133 (2004)
43. F. Omota, A.C. Dimian, A. Bliek, Partially hydrophobized silica supported Pd catalyst for
hydrogenation reactions in aqueous media. Appl. Catal. A: Gen. 294, 121–130 (2005)
44. J. Paulik, F. Paulik, in Comprehensive Analytical Chemistry, vol. 12, ed. by G. Svehla (Elsevier,
Amsterdam, 1981)
45. B. Pawelec, J.L.G. Fierro, Applications of thermal analysis in the preparation of catalysts and in
catalysis, in Handbook of Thermal Analysis and Calorimetry, vol. 2, Applications to Inorganic
and Miscellaneous Materials, ed. by M.E. Brown, P.K. Gallagher (Elsevier, The Netherlands,
2003)
2 Thermal Analysis and Calorimetry Techniques for Catalytic Investigations 101

46. L. Pinard, J. Mijoin, P. Magnoux, M. Guisnet, Dichloromethane transformation over bifunc-

tional PtFAU catalysts. Influence of the acidobasicity of the zeolite. C. R. Chim. 8, 457–463
(2005)
47. M.G. Plaza, C. Pevida, B. Arias, J. Fermoso, A. Arenillas, F. Rubiera, J.J. Pis, Application of
thermogravimetric analysis to the evaluation of aminated solid sorbents for CO2 capture. J.
Thermal Anal. Calorim. 92(2), 601–606 (2008)
48. M.I. Pope, M.D. Judd, Differential Thermal Analysis: A Guide to the Technique and its Appli-
cations (Heyden, London, 1977)
49. V. Quaschning, A. Auroux, J. Deutsch, H. Lieske, E. Kemnitz, Microcalorimetric and catalytic
studies on sulfated zirconia catalysts of different preparations. J. Catal. 203, 426–433 (2001)
50. V. Rakic, V. Dondur, R. Hercigonja, Thermal effects of the interaction of carbon monoxide
with zeolites. Thermochim. Acta 379, 77–84 (2001)
51. M. Resan, M.D. Hampton, J.K. Lomness, D.K. Slattery, Effects of various catalysts on hydrogen
release and uptake characteristics of LiAlH4 . Int. J. Hydrogen Energy 30, 1413–1416 (2005)
52. M.J. Richardson, E.L. Charsley, Calibration and standardisation in DSC, in Handbook of Ther-
mal Analysis and Calorimetry, vol. 1, Principles and Practice, ed. by M.E. Brown (Elsevier,
The Netherlands, 1998)
53. R. Ruiz, C. Pesquera, F. González, C. Blanco, Synthesis and catalytic behaviour of heteroge-
nized rhodium catalysts on modified clays. Appl. Catal. A: Gen. 257, 165–175 (2004)
54. L. Shen, M. Zheng, J. Xiao, R. Xiao, A mechanistic investigation of a calcium-based oxygen
carrier for chemical looping combustion. Combust. Flame 154, 489–506 (2008)
55. S. Siffert, L. Gaillard, B.L. Su, Alkylation of benzene by propene on a series of beta zeolites:
toward a better understanding of the mechanisms. J. Mol. Catal. A: Chem. 153, 267–279 (2000)
56. J. Silvestre-Albero, C. Gomez de Salazar, Characterization of microporous solids by immersion
calorimetry. Colloids Surf. A: Physicochem. Eng. Aspects 187–188, 151–165 (2001)
57. P.F. Siril, A.D. Davison, J.K. Randhawa, D.R. Brown, Acid strengths and catalytic activities
of sulfonic acid on polymeric and silica supports. J. Mol. Catal. A: Chem. 267, 72–78 (2007)
58. S.R. Son, S.D. Kim, Semi-continuous operation of chemical looping combustion with metal
oxides supported on bentonite in an annular fluidized bed reactor, in Proceedings of the Asian
Conference on Fluidized Bed and Three Phase Reactors (2006)
59. J. van Humbeeck, Simultaneous thermal analysis, in Handbook of Thermal Analysis and
Calorimetry, vol. 1, Principles and Practice, ed. by M.E. Brown (Elsevier, The Netherlands,
1998)
60. S. Wang, V. Slovak, B.S. Haynes, Kinetic studies of graphon and coal-char reaction with NO
and O2 : direct non-linear regression from TG curves. Fuel Process. Technol. 86, 651–660
(2005)
61. W.W. Wendlandt, Thermal Analysis, vol. 19, 3rd edn. (Wiley, New York, 1986)
62. J. Zhao, Z. Liu, D. Sun, TPO-TPD study of an activated carbon-supported copper catalyst-
sorbent used for catalytic dry oxidation of phenol. J. Catal. 227, 297–303 (2004)
Chapter 3
Couplings

Dušan Stošić and Aline Auroux

Abstract Basic principles of calorimetry coupled with other techniques are intro-
duced. These methods are used in heterogeneous catalysis for characterization of
acidic, basic and red-ox properties of solid catalysts. Estimation of these features is
achieved by monitoring the interaction of various probe molecules with the surface
of such materials. Overview of gas phase, as well as liquid phase techniques is given.
Special attention is devoted to coupled calorimetry–volumetry method. Furthermore,
the influence of different experimental parameters on the results of these techniques
is discussed, since it is known that they can significantly influence the evaluation of
catalytic properties of investigated materials.

3.1 Introduction

Catalytic cycle in heterogeneous catalysis can be divided in five consecutive steps;

diffusion of reactants to surface, adsorption of reactants, surface reaction processes,
desorption of products and diffusion of reactants from the surface [1]. A fundamental
understanding of the great diversity of chemistry occurring at the surface requires
familiarity with molecule–surface interaction. The energetics of chemical events
occurring on the surface play an important role in determining the catalytic properties
of the surface, since different surfaces will cause a substance to react in different ways.
It is known from the literature that oxidation of methanol can lead to different reaction
products, and different selectivities depending on surface properties (structural and
chemical) of catalyst used in this reaction [2, 3]. Study of adsorption of probe and
reactive gas molecules onto surface of catalysts lead to the better understanding of the

D. Stošić (B) · A. Auroux

Institut de Recherches sur la Catalyse et l’Environnement de Lyon,
UMR 5256 CNRS/Université Lyon 1,
2 avenue Albert Einstein, 69626 Villeurbanne Cedex, France
e-mail: [email protected]; [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 103

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_3,
© Springer-Verlag Berlin Heidelberg 2013
104 D. Stošić and A. Auroux

nature of gas-solid interactions and give valuable information about the properties
of the adsorbent surface and thus is of primary importance in catalysis.
Adsorption is an exothermic process and when a reactive molecule interacts with
the surface of the solid heat is evolved. This heat is related to the energy of the
bonds formed between the adsorbed species and the adsorbent and hence to the
nature of the bonds and to the chemical reactivity of the surface. From a number of
techniques used to study this interaction only a few provide information about the
strength of chemisorption itself. The determination of the differential heats evolved,
by a suitable microcalorimeter, when known amounts of gas probe molecules are
adsorbed on catalytic surface is the most suitable and accurate method which allows
the determination of the number, strength and energy distribution of the adsorption
sites [4–12]. The effective use of this technique in heterogeneous catalysis depends
on choice of probe molecule, since the nature of the probe determines which surface
property is going to be investigated.
The definition of various heats of adsorption/desorption and their relationship with
thermodynamic quantities directly obtainable by calorimetry are given in Table 3.1.

3.2 Coupled Calorimetry–Volumetry

Adsorption microcalorimetry is a powerful technique for the energetic characteri-

zation of solid surfaces, as well as for the thermodynamic description of solid–gas
interfacial phenomena.
To access the energetic heterogeneity of surface, small doses of probe gas (typ-
ically <10 µmol g−1 ) have to be added successively on the solid to saturate active
sites progressively. Corresponding heats range from 100 to 1000 mJ, and require few
hours for being evolved. Since the development of commercial instrumentation able

Table 3.1 Definition of various heats of adsorption/desorption

Name and symbol Unit Definition Measurement method
Integral heat of J Quantity of heat evolved when Isotherm calorimetry
adsorption: qint n a moles are adsorbed at
constant temperature on an
adsorbent initially in vacuo,
without a change in the volume
of the cell: Q int = n a (Ua − Ug )
Molar integral heat J mol−1 Integral heat per mole adsorbed Isotherm calorimetry
Q int = qninta
Differential heat of J mol−1 Defined 
from integral
 heat by: Isotherm calorimetry
adsorption: qdiff Q diff = ∂q int
∂n a T,A
Isosteric heat: qst J mol−1 Heats of adsorption calculated Adsorption isotherm
from adsorption
 isotherms:

qst = −RT ∂ ∂T
2 ln P
a
n
3 Couplings 105

to measure quantitatively such low heats and adsorbed amounts, microcalorimetry

has gained importance as one of the most reliable methods for the study of gas solid
interactions [5, 7, 8].
A calorimeter that is suitable for measurement of heats of adsorption is required
to have the following characteristics:
(1) High sensitivity, with sufficient precision, to follow the changes of differen-
tial and integral heats with coverage, allowing the use of small doses of probe
molecules.
(2) High thermal stability, to have a well-stabilized base line, which is necessary to
study the slow process of adsorption.
(3) Large interval of utilization temperature. This is important for studying the inter-
actions of different reactants with surface.
(4) Good accessibility of the calorimeter proper allowing the connections with vol-
umetric apparatus. Relatively large dimensions as well as suitable geometry of
the calorimetric cells are also required for same purposes.
Most commonly used are heat-flow microcalorimeters of the Tian-Calvet type
[5, 8]. The detailed theory and operation of this calorimeter can be found elsewhere
[11]. The apparatus is composed of an experimental vessel, where the studied system
is located, which is placed into a calorimetric block (Fig. 3.1). The temperature of the
block, which functions as heat sink, is controlled very precisely. The heat generated
in the system flows to the heat sink and is accurately measured by means of detector.
This is made of a large numbers of identical thermocouples (a thermopile) that
surrounds the vessel and connected to the block (Fig. 3.2) in such a way that vessel
and block temperature are always close to each other. A signal is generated by the
detector that is proportional to the heat transfer per unit time. Undesired signals due
to the external temperature fluctuations in the calorimetric block are minimized by
connecting in opposition two heat flow detectors from two identical vessels, one
of which is used to perform the experiment, the other being used as a reference.
Heat related to the introduction of the probe and other parasitic phenomena are thus
compensated.
The heat generated by the process inside the vessel is transferred completely
through the thermocouples, provided that radiation, conduction along the wall of
the vessel, conduction and convection through the fluid phase are minimized by the
design of the apparatus. A voltage signal, E, is generated by the thermocouples and
recorded over the time of the thermal event, whose end is indicated by the returning
of the thermopile output signal to the base line. The total heat evolved during the
event is given by the area under the thermogram:

Q exp . = K Edt (3.1)
exp .

where K is the instrument constant [11].

This highly sensitive heat flow calorimeter linked to sensitive volumetric sys-
tem makes it possible to study gas–solid interactions and catalytic reactions. The
106 D. Stošić and A. Auroux

Fig. 3.1 Calvet

microcalorimeter

apparatus used for this kind of measurement is presented in Fig. 3.3. The volumetric
determination of the adsorbed amount of gas is performed in a constant volume vessel
linked to a vacuum pump. The apparatus consists of two parts: the measuring element
equipped with a capacitance manometer and a cell section, which includes the cells
placed in the calorimeter (a sample cell in which the adsorbent solid is placed, and
an empty reference cell). The volume of this vessel is determined by the expansion
of a known quantity of gas, contained in the measuring part of the assembly, into the
previously evacuated cell section. This calibration must be made with the same gas
and at the same temperature as the proposed study is going to be done.
In a typical experimental procedure the sample is first outgassed at the desired
pretreatment temperature and under high vacuum in the calorimetric cell. After cool-
ing to the adsorption temperature and when the thermal equilibrium of the system is
reached, successive doses of gaseous probe molecules are brought into contact with
catalyst sample. The heat flow signal and the corresponding pressure change are
3 Couplings 107

heatflux transducers

reference sample

thermostated calorimetric block

Fig. 3.2 Thermostated calorimetric block

monitored and recorded until equilibrium is reached after admission of each dose.
The volume of doses is minimized in order to detect any change in adsorption heats
evolved with increasing the coverage of surface active sites. The adsorption is con-
sidered complete when for a significant increase in pressure there is no detectable
heat evolution or gas adsorption. The adsorption temperature is maintained constant
during the experiment. The irreversibly chemisorbed amount (Virr ) can be evalu-
ated from the difference between the primary adsorption isotherm (adsorbed amount
of gas as function of equilibrium pressure over sample) and a secondary isotherm
obtained after desorption under vacuum and re-adsorption of the gaseous probe at the
same temperature. By subtraction of re-adsorption isotherm from the first adsorption
isotherm amount of strong sites can be roughly estimated. The data obtained directly
from adsorption calorimetry measurements are represented in Fig. 3.4, where both
108 D. Stošić and A. Auroux

Fig. 3.3 Schematic diagram of the volumetric–calorimetric line

Fig. 3.4 Recorded data,

pressure and heat flow signal,
as a function of time for each
dose of probe molecules

the calorimetric signal and the evolution of pressure are plotted as a function of time
for each dose added.
From accumulation of these raw data until the full coverage is attained we can
get information about the sample properties that can be summarized on various plots
(Fig. 3.5):
(a) The volumetric isotherms (n a = f (P)) for a cycle of adsorption (I), desorption
by pumping at the same temperature and then re-adsorption (II). The irreversibly
adsorbed volume (Virr ), which characterizes the strong sites of the catalyst, can
then be calculated as the difference between the adsorption volume and the
re-adsorption volume at a given equilibrium pressure.
(b) The corresponding calorimetric isotherms (Q int = f (P)).
3 Couplings 109

Fig. 3.5 Data obtained from volumetric–calorimetric adsorption experiments

(c) The integral heats (Q int ) as a function of the adsorbed quantities (n a ). This
representation leads to the detection of coverage ranges with constant heat of
adsorption, those for which the evolved
 heat is a linear function of coverage.
(d) The differential heat, qdiff = Q int ∂n a (molar heat of each dose of adsorbate) as
a function of n a . The ratio of the amount of heat evolved for each increment to
the number of moles adsorbed in the same period is equal to the average value
of the differential enthalpy of adsorption in the interval of the adsorbed quantity.
The curve of differential heat as a function of surface coverage is traditionally
represented as histograms. For simplification the steps of histogram are often
replaced by a continuous curve connecting the centres of steps. Differential
heats of chemisorption usually decrease with increasing the surface coverage.
The way in which the profile of differential heat changes with increasing surface
coverage depends on nature of adsorbate and adsorbent. Sharp heat falls at low
surface coverage are in general regarded to surface
 heterogeneity.
(e) The energy distribution spectra in which −dn a dqdiff is expressed as a function
of qdiff . The area under this curve represents the number of molecules that are
adsorbed with a given evolved heat. This kind of representation is not as accurate
as the previous one.
Although it is not possible to determine the nature of adsorbed species, or even
to distinguish between different kinds of adsorbed species from the calorimetric
110 D. Stošić and A. Auroux

Fig. 3.6 Regions in a typical curve of differential heats of adsorption versus adsorbed amount

data, the variation of the differential heats of adsorption with coverage quite clearly
shows energy distribution of surface active sites with respect to a given adsorbate and
their varying reactivity on given adsorbents. A general representation of differential
heats of adsorption as a function of uptake of probe molecule (Fig. 3.6) presents the
following features:
(a) An initial region of high heats of adsorption, representing adsorption on the
strongest sites, which are usually thought to be of Lewis type. After this region
the curve drops abruptly.
(b) One or more regions of intermediate strength sites. Regions of constant heat in
this region are attributed to sets of sites of constant strength. This is taken as an
indication of discrete inhomogeneity. These features, when adsorption of base
molecule on the surface of zeolite is in question, are attributed to Brönsted acid
sites.
(c) A region where heats decrease more or less steeply depending on the hetero-
geneity of the sites. In this area a bump on the curve can sometimes be observed,
which indicates lateral interactions between adsorbed molecules.
(d) At a high coverage heat of adsorption approaches a constant value characteristic
of hydrogen bonding between the molecule and the sample or physisorption
of the molecules. This constant value is dependent on the nature of the probe
molecule.
All regions (a, b, c, d) can be observed for zeolite samples presenting both Lewis
and Brönsted acid sites, as probed by ammonia adsorption. For oxides presenting
only Lewis acid sites, the regions a, c and d are observed.
This kind of representation of calorimetric data can be used to assess the
uniformity–non-uniformity of the surface of the adsorbent with respect to energy,
the energy of the lateral (adsorbate–adsorbate) interactions, and the structural or
substructural (textural) changes that the adsorbent often undergoes as a result of
interaction with the adsorbed substance [13].
3 Couplings 111

Fig. 3.7 Thermokinetic para-

meter as a function of pyridine
uptake on mordenite samples
(from [16])

Study of the thermokinetics of the heat evolution combined with volumetric and
calorimetric data can provide a better insight into the adsorption mechanism and into
the location of adsorbing sites on the catalyst surface [14–16]. It has been shown
that interpretation of the shapes of individual thermograms facilitates identification
of simultaneous processes with different kinetics as well as the variation of kinetics
with coverage.
The kinetics of heat release during adsorption can be monitored by the change
in thermokinetic parameter τ [14]. The calorimetric signal decreases exponentially
with the time of adsorption after the maximum of each adsorption peak. This can
be expressed in the form D = Dm e−t/τ , where D and Dm are the deviation at time
t and maximum deviation of calorimetric signal, respectively. The thermokinetic
parameter in this expression can be computed as the inverse of the slope of the
logarithm of the evolved heat curve during the return to equilibrium. Variations of
the thermokinetic parameter, determined in this way, can then be plotted versus the
amount of adsorbed probe.
When ammonia or pyridine were used as probe molecules to study the acidity
of zeolites at a given temperature [14, 16], the time needed to establish equilibrium
after the addition of doses at first increases with increasing coverage (Fig. 3.7). This
is due to the fast adsorption rate because the molecules are bonded irreversibly to the
112 D. Stošić and A. Auroux

strongest acid sites. The curve then passes through a maximum as the adsorption rate
decreases because a smaller number of strong sites is available and the molecules must
diffuse over greater distances on the surface. This coverage should correspond to the
filling of the acid sites of intermediate strength. Finally only the exchange between
probe molecules among the weaker sites occurs. This is a fast process resulting in
decrease of equilibrium time, which reaches a value close to the time constant of the
calorimeter. In this way differentiation between strong and weak adsorptions can be
done.
The time needed to establish the equilibrium depends on the quantity of adsorbed
probe, on the temperature, number and strength of surface active sites and on the
inertia of calorimeter. At lower temperatures, a slower adsorption is observed in
covering the strong adsorption centres than at higher temperatures. The long time to
establish the equilibrium is apparently related to redistribution of the adsorbed probe
on the centres that are energetically more favourable [17]. When the time to establish
thermal equilibrium is determined solely by inertia of calorimeter, one can be sure
that the adsorption temperature was well chosen.
Entropy of adsorption, determined from adsorption equilibrium constants obtained
from adsorption data (adsorption isotherms and differential heats of adsorption) can
be also estimated.
Adsorption is an exothermic process and takes place with decrease in both free
energy (G < 0) and entropy (S < 0). With respect to the adsorbate, the gas–
solid interaction results in a decrease in entropy of the system. The differential molar
entropies of the system can be plotted as a function of the coverage. The interest
of a continuous measurement of derivative entropy of adsorption stems from the
fact that any variation of its magnitude provides information on: (i) a variation of
the adsorbate-adsorbent interactions as a function of coverage (it allows the hetero-
geneities of the adsorbing surface to be detected), (ii) a collective modification of
the state of the interaction of the molecules already adsorbed, (iii) or a combination
of both [17, 18].

3.3 Coupled Calorimetry–Gravimetry

Gravimetry can be associated with adsorption calorimetry either in vacuum or in gas

flow system. When vacuum system is used, methodology of experiment is similar as
used in calorimetry–volumetry technique. Desired amount of the sample is placed
on microbalance which is surrounded by Tian-Calvet thermopile. Adsorbed amount
of gas is determined by measuring the change in the sample mass. The heat flow
signal and the corresponding mass change are monitored and recorded until equi-
librium is reached after admission of each dose. Results obtained by this technique
are comparable to that obtained when volumetric system is used to determine the
amount of adsorbed gas. Volumetric measurements are to be preferred to gravimetric
measurements for these determinations because it is very difficult to ensure a good
3 Couplings 113

Fig. 3.8 Gas flow cells

and reproducible thermal contact between a sample of adsorbent, hanging from a

balance beam, and the inner cell of a heat-flow calorimeter [11].
Flow adsorption calorimetry involves the use of carrier gas passing continuously
through the adsorption cell. The catalyst is placed on a glass frit in a gas circulation
cell in the calorimeter. Examples of gas flow cells are represented in Fig. 3.8. In
order to determine the amounts of adsorbed gas, flow calorimetry is used in com-
bination with thermogravimetry (TG). When this system is used in pulse mode it
allows the measurement of heats of adsorption of a gaseous reactant on a solid or
interaction heats between a gaseous reactant and pre-adsorbed species. When used
as a flow reactor, it allows the kinetic study of catalytic reactions as well as the study
of the activation or the aging of the catalyst. This is also a suitable system to perform
calorimetric temperature programmed reduction (TPR), temperature programmed
oxidation (TPO) or temperature programmed desorption (TPD) experiments. In addi-
tion to calorimetry, temperature programmed desorption (TPD) of adsorbed probe
molecules can in principle also be used to estimate heats of adsorption [19].

3.4 Temperature Programmed Desorption Technique

The temperature programmed desorption of probe molecules is a commonly used

method for measuring basic or acidic properties of catalyst surfaces. In this procedure
the sample is subjected to temperature programmed heating and pre-adsorbed gas is
progressively desorbed. The rate of desorption increases with temperature increase,
goes through a maximum, and finally goes back to zero as the surface is cleared from
adsorbate. TG-DSC-MS is experimental setup (Fig. 3.9) that can be used for this kind
of measurements. During this experiment endothermic heat signal, temperature of
desorption and concentration of desorbed molecules is measured. Desorption energy
distribution of probe molecules that is available from these data, can be related to the
114 D. Stošić and A. Auroux

Fig. 3.9 Schematic representation and a photograph of a TG-DSC-MS apparatus

acid strength distribution of surface active sites by hypothesis that desorption energy
is equal to the energy of acid re-protonation.
Despite the fact that TPD is widely used because of its simplicity the main draw-
backs of this technique have to be pointed out.
• TPD gives an average value of acid strength rather than a distribution.
• Results obtained by this technique are strongly affected by mass or heat transfer
and diffusion.
• TPD enables only to distinguish the strength of active sites, but not their types.
• Desorption on weak sites is hindered by adsorbates on strong sites.
• During desorption, a readsorption may occur.
• Peaks overlap, so TPD can be used only to roughly distinguish the various acid
site strengths.
TPD of reactive amines is a very useful technique for measuring Brönsted site
densities in zeolites [20–23]. Alkylammonium ions, formed by protonation of amines
at Brönsted sites react in a very narrow temperature range in TPD through a reaction
similar to Hoffman elimination reaction. Figure 3.10 shows a result for TPD-TGA-
MS of isopropylamine from an H-MFI sample having a bulk Si/Al ratio of 26 [24].
The TGA results are useful since one can determine whether there are residual
species on the sample after desorption and one can quantify the amounts which
desorb without having to calibrate the mass spectrometer.
From MS results the nature of acid sites can be determined. Upon heating in vac-
uum two stages of weight loss and two stages of product evolution can be observed.
The molecules associated with low-temperature desorption can probably be assigned
to a number of different types of species, including molecules associated with Lewis
3 Couplings 115

Fig. 3.10 TPD-TGA-MS results for isopropyl amine in H-MFI sample. The desorption of propylene
and ammonia above 550 K corresponds to decomposition of the amine at the Brönsted acid sites
(from [24])

acid sites, molecules associated with hydroxyl defects, and molecules which are
hydrogen bonded to protonated amines at the Brönsted sites. At high temperature
molecules desorb as propylene and ammonia. This desorption feature can be associ-
ated with protonated molecules at Brönsted sites. Fact that the measured site density
is independent of the choice of alkylamine gives further confirmation that TPD-TGA-
MS results provide a true evaluation measure of the Brönsted site density [25].

3.5 Temperature Programmed Reduction

From calorimetric TPR-TPO experiments heats of reduction and reoxidation of

cationic or metal species can be determined. Data obtained from these experiments
may provide kinetic data of theoretical significance as well as an insight into the
mechanism of the reduction processes.
TPR profiles of copper exchanged ETS-10, obtained by differential scanning
calorimetry (DSC) and carried out with a hydrogen stream in helium, are presented
in Fig. 3.11 [26]. All the reduction profiles show complex structure, in which a step-
wise reduction of CuO to Cu through Cu2 O intermediates occurred with overlapping
of the two peaks. By analysis of this profiles authors gained useful information con-
116 D. Stošić and A. Auroux

Fig. 3.11 DSC curves of reduction by hydrogen as a function of temperature for Cu-ETS samples
(from [26])

cerning the identification of supported species and the determination of intermediate

oxidation states of metals during reduction.
In another study in order to compare reducibility power of hydrogen, ethane, and
ethylene authors performed the reduction of fresh vanadium pentoxide by each of
them (diluted in helium) in a TG-DSC apparatus [27]. Figure 3.12 shows the heats of
reduction which are exothermic and the derivative of the thermogravimetric curves
which are negative if associated with a weight loss occurring during the reduction.
According to the weight loss, a similar reduction extent was recorded for all three
cases. Heat flow measurements have shown that reducibility levels of V2 O5 by ethane
and hydrogen are similar, whereas ethylene is much more active that ethane over that
oxide.

3.6 Calorimetry–Gas Chromatography/Mass Spectrometry

This coupling can be used for characterization of acidic or basic character of solid
catalysts, and for study of solid–gas reactions. Systems used for these measurements
are usually modified DSC instruments connected with GC/MS by a heated capillary
tube [28–32]. In a typical experiment small pulses of probe gas are injected at regular
intervals in to the carrier gas stream from a gas sampling valve. The gas flows are
regulated by mass flow controllers. The concentration of ammonia down-stream of
the sample is monitored with the GC/MS and heat evolution with the calorimeter.
3 Couplings 117

Fig. 3.12 Heats of reduction of V2 O5 and associated derivatives of the thermogravimetric curves
as a function of temperature, for various reducing agents (from [27])

When gas–solid reactions are under investigation this system allows simultaneous
determination of heats of reaction and identification of products. Scheme of the
apparatus is shown in Fig. 3.13.

3.7 Calorimetry-Syringe Pump-UV-Vis Fluorescence

Spectrometry

Isothermal titration calorimetry is a technique that can directly measure liquid–solid

or liquid–liquid interactions. This method has been used for determination of acid–
base properties of solid catalysts in an inert phase (n-decane), for determination of
heats of reaction in liquid phase (alkylation, anisole + benzoic anhydride on H-Beta
zeolite), and for determination of heats of adsorption of different pollutants on an
adsorbent [33]. It is also recognized as the only technique that can directly measure
the binding energies of biological processes such as protein–ligand binding and
protein–protein binding [34].
Instrument that is usually employed in this technique is differential heat flow
calorimeter of Tian-Calvet type equipped by a stirring system. A programmable
118 D. Stošić and A. Auroux

Fig. 3.13 Calorimetry-GC/MS experimental device

syringe pump is linked to the calorimeter by capillary tubes. Using this syringe pump,
successive pulses of known amounts of a solution of a probe molecule can be sent on
the sample which is maintained at constant temperature during the experiment. UV-
Vis spectrometer is connected to the calorimetric cell by an optic fiber, and it is used
to monitor the changes in concentration of the probe molecule. From equilibrium
concentration after each dose of solution, adsorption isotherms can be constructed.
Scheme of the instrument is presented in Fig. 3.14.
This method has been used for characterization of acidity of solid catalysts, or to
investigate adsorptive properties of solid materials by measuring the adsorption heats
evolved upon injection of dilute solutions of chosen molecules onto the solid itself in
slurry of appropriate solvent. The output of typical microcalorimetric experiment is
shown in Fig. 3.15. Each pulse injection of dosed amount of solution of probe mole-
cule gives as a result the exothermic signal that formed a specific peak related to the
heat of adsorption. This heat derives from two different contributions: the exothermic
enthalpy of adsorption and the endothermic enthalpy of displacement of adsorption;
while the enthalpy effects originating from dilution and mixing phenomena can be
neglected due to the differential design of the calorimeter and the preheating of the
probe solution. The amount of adsorbate in solution is measured simultaneously with
a UV-Vis spectrometer, leading to an adsorption isotherm which is measured over
the range of base addition used in calorimetric titration. Study of evolution of UV-
Vis signal with time (Fig. 3.16) during adsorption of each dose added gives a better
insight into the adsorption mechanism.
3 Couplings 119

Fig. 3.14 Schematic representation of a TITRYS calorimeter connected to UV-Vis spectrometer

Combination of data obtained from calorimetric measurements and adsorption

isotherms permits a determination of the number of active sites, the equilibrium
constants, and the enthalpy of a binding to each type of the site [35–39].
For microcalorimetric determination of the strength of solid acids in solution,
pyridine, n-butylamine and aniline are the usual probe molecules that are used. The
measurements are done in a non-interacting hydrocarbon solvent (e.g. cyclohexane)
whose molecular mass is close to that of donor in order to cancel out contribution from
a dispersion component to the measured enthalpy [36]. Throughout the experiment
the molar enthalpy of neutralization, Hneut◦ , is measured. If H◦
neut values are going
to be used as measurements of relative acid strength, some assumptions have to be
made. This is because the measured enthalpy of neutralization is made up from
contributions from several sources: proton loss from the acid and proton gain by the
base, changes in solvation as the acid and base are converted to the conjugate base
and acid, plus any enthalpy changes associated with formation of ion pairs that form
or break up during the neutralization process. In order for the differences in measured
◦
Hneut values to reflect the differences in the acid strength of acid solids requires
◦
that all contributions to Hneut other than the proton affinity of the acid are constant.
This can be taken as reasonable assumption only when closely related solid acids
are compared [36, 40–42]. Another point is that titration calorimetric measurements
of this kind are usually performed at temperatures close to room temperature so
the process of adsorption is more likely to be kinetically controlled. In this way
adsorption could occur simultaneously on different types of active sites and average
120 D. Stošić and A. Auroux

Fig. 3.15 The output of typical microcalorimetric experiment for successive doses

Fig. 3.16 UV-Vis signal evolution as a function of time for one injection

◦
value of Hneut is obtained for each dose of probe molecule. Because of this, real
strength distribution of acid sites is not always available from these measurements.
Useful feature of this method is possibility to change solvent and investigate the
influence of solvent on acid strength. Niobium oxide and niobium phosphate were
investigated by this technique, using aniline and 2-phenyl-ethylamine as probe mole-
cules, in order to understand the effective acidity of these solids measured in various
3 Couplings 121

solvents (decane, cyclohexane, toluene, methanol, and isopropanol) of very differ-

ent polarities and proticities. By using this technique it is possible to discriminate
the acid site strength distribution more accurately than with conventional gas-solid
phase titration with ammonia [43]. In another study acid strength of sulfonic acids
supported on polystyrene and silica solids were determined in water and cyclohexane.
It was shown how the relative acid strength of these very similar types of supported
acids depends markedly on the nature of solvent. The relative acidities of these solid
materials are reversed when going from water to cyclohexane. This leads to a con-
clusion that if these measuremenrs are to be used for predicting catalytic activity, it
is essential that they are made in solvents similar to those in which the catalysts are
ultimately to be used [44].

3.8 Limitations of Technique

Results obtained by adsorption microcalorimetry are not sufficient to determine the

nature of adsorbed species, or even to distinguish between different kinds of adsorbed
species. When examination of catalyst surface is done, that possesses both Lewis and
Brönsted acidity, by adsorption of base probe molecule such as ammonia, it is difficult
to discriminate strong Lewis from strong Brönsted acid sites solely by adsorption
microcalorimetry. This is due to the fact that differential heats of NH3 interaction with
strong Lewis and Brönsted acid sites are relatively close to each other. For this reason
complementary informations from suitable IR, MAS NMR and XPS investigations
are necessary to identify these sites [45]. However, because of the complex nature
of the active site strength distribution it is not possible to make a detailed correlation
between sites of different nature and their strength.
Initial values representing very small concentrations of the strongest active sites
can be easily missed in the measurement if the gas doses are not small enough.
If kinetics of adsorption is determined by diffusion limitations, determined heats of
adsorption may reflect an average heat of the various strong sites, rather than a specific
effect. In this case adsorption on active but less-accessible sites may occur only
after better-exposed but less-active sites have interacted. Differential heat profiles
determined under diffusion restrictions may reflect less surface heterogeneity than
actually exists on surface of the adsorbent [11]. The sample bed thickness has also
to be considered because the molecular mobility in the sample bed is a limitation to
rapid data collection [7].
Finding correlations of surface properties of catalyst (in the case of adsorption
calorimetry: acid, basic and redox properties) with its catalytic behaviour is some-
times difficult. The lack of correlation is due to the fact that by microcalorimetry
measurements the total amount of acid or basic sites is obtained and only a fraction
of these active sites may actually be involved during the catalytic reaction [46].
There could be surface sites that are either too weak to activate the reactants or are
too strong, leading to strongly held species which block and deactivate these sites or
causing excessive fragmentation of reactant or products. It has been recognized that
122 D. Stošić and A. Auroux

the catalytic properties of Brönsted acid sites (proton donors) can be different from
Lewis acid sites (electron pairs acceptors). For example, it has been suggested that
Brönsted acids are much more active than Lewis acids for skeletal transformations
of hydrocarbons [47], and as said before it is impossible to differentiate them solely
by adsorption microcalorimetry technique.

3.9 Influence of the Adsorption Temperature on the Acid/Base

Determination

In order that adsorption microcalorimetry can give an accurate representation of

active site strength distribution the adsorbed probe molecule must be equilibrated
among all sites on the catalyst surface within the time frame of the experiment. If
external or internal mass-transfer limitation exists the adsorption on surface sites
is controlled by kinetics instead of thermodynamics [5]. In this case the profile
of differential heats as function of coverage would not accurately represent energy
spectrum of the surface active sites and the adsorbent surface would appear to be more
homogeneous. For these reasons, it is important to study the effect of the adsorption
temperature and verify that molecules possess sufficient thermal energy to obtain the
thermodynamically stable site occupation.
When the process of adsorption was studied at various temperatures two different
kinds of temperature dependence were observed [48, 49]. One type was described as
a change of the profile of differential heat of adsorption with increasing coverage of
pyridine on HY zeolite and silica–alumina as temperature was raised. The adsorp-
tion of pyridine was performed at 423 and 473 K. For lower coverage, the profile of
integral heats obtained for the adsorption at 473 K, lies above that obtained at 423 K.
This difference is decreasing as the surface coverage is increasing, and at high cov-
erage the situation is opposite. The difference between the differential heat curves
with temperature suggests that there is difference in selectivity when adsorption is
performed at higher temperatures compared to adsorption at lower temperatures.
Adsorption occurred on stronger acid sites in preference to weaker sites at higher
temperature, whereas random adsorption occurred simultaneously on both weak and
strong sites at low temperatures. The authors [48, 49] proved this hypothesis by
monitoring the change in the intensity of the hydroxyl infrared adsorption bands on
a HY zeolite at different temperatures for the progressive titration of acid sites with
pyridine. This behaviour proves the importance of high surface mobility of probe
molecules on the catalysts surface for equilibration on active sites.
The second type of behaviour was a slight decrease in the differential heats of
ammonia adsorption on HY above 473 K and NaY above 313 K with an increase of
temperature. This decrease was attributed to the fact that the temperature dependence
of the heat of adsorption was thermodynamically defined by the difference in molar
heat capacity between adsorbed state and gaseous state.
3 Couplings 123

Unusual behaviour was observed for adsorption of ammonia on protonated ZSM

zeolite at 416 K [50]. Differential heat curve passed through a maximum at relatively
low coverage. This behaviour was explained by conjunction of three independent
phenomena: immobile adsorption, mass-transfer limitations and preferential location
of the most energetic acid sites in the internal pores of zeolite structure. The maximum
was eliminated by heating of samples between doses to increase the surface mobility
of the preadsorbed ammonia and allow it to migrate and adsorb on the most acidic
sites. In this way most accessible sites are made available for new doses of ammonia
at lower temperature. After this the heat curve showed a slightly higher value than
the initial heat for the conventional method of adsorption, and it did not show any
maximum.
From the discussed experimental results it is clear that results obtained by adsorp-
tion microcalorimetry technique are strongly dependent on the temperature at which
the experiments are performed. For samples which possess strong sites, increasing
the adsorption temperature yields a better description of the thermodynamic site
strength distribution. For samples which possess weaker sites adsorption at higher
temperatures produces a large decline in the adsorption capacity. In any case the
choice of adsorption temperature is constrained and involves compromise between
using high temperatures to reduce equilibrium times, and using lower temperatures
to achieve high coverage of the active sites at reasonable pressures.

3.10 Probe Molecules

The key for effective utilisation of microcalorimetry in heterogeneous catalysis is

the judicious choice of gas-phase molecules for study. Although total number of
surface active sites and potentially active centres can be estimated by this tech-
nique, the obtained values are strongly dependent on the nature and size of the probe
molecule. As a general rule, no matter which surface property is examined (acidic,
basic or redox); probe molecule should be stable with temperature and with time.
Furthermore, as discussed before, the adsorbed probe should be sufficiently mobile
to equilibrate on active sites at temperature of the experiment.

3.10.1 Probing Surface Acidic Properties

Though it is very reliable, the calorimetric measurement of surface acidity of solid

catalysts depends on the choice of the basic probe molecule used to neutralize the
acid sites. Ammonia (pKa = 9.24, proton affinity in gas phase = 857.7 kJmol−1 )
and pyridine (pK a = 5.19, proton affinity in gas phase = 922.2 kJmol−1 ) are the
favoured probe molecules to probe overall surface acidity of catalysts, because both
Lewis and Brönsted acid sites retain these molecules.
124 D. Stošić and A. Auroux

Ammonia is among the smallest strongly basic molecules and its diffusion is little
affected by the porous structure, if at all. Because of this, it is the most commonly used
probe molecule for testing surface acid properties. It is adsorbed as an ammonium
ion, and the corresponding heat of adsorption depends both on the proton mobility
and on the affinity of ammonia for the proton.
Heats of adsorption of strong base molecules, are related to intrinsic acidity of the
site and to the interaction energy between the deprotonated surface of the sample and
the protonated base [51]. From this it is a logical question how strongly do differ-
ent bases interact with the Brönsted and Lewis sites, and which base interacts more
strongly. A study of surface acidity by adsorption of various basic probe molecules
(ammonia, pyrrole, dimethyl ether and acetonitrile) on dealuminated Y type zeolites
proved that obtained acid site distribution strongly depends on the basicity of probe
molecule [52]. Ammonia proved to be a reliable probe molecule when Brönsted acid
sites were being investigated. It helped to unveil two site populations whose proper-
ties varied with Si/Al ratio. These two populations differ by local environment of the
associated AlO4 tetrahedra. Ammonia, however, failed to reveal inhomogeneity of
one particular acid population. Dimethyl ether appeared as a too weak base, whereas
pyrrole, showing no steric hindrance in this case, appeared as a rather amphoteric
probe which helped visualize the basicity difference between the parent and dealumi-
nated materials. Acetonitrile proved to be a reliable probe to monitor quantitatively
and qualitatively Lewis acidity.
In another study with various base probe molecules experiments have been per-
formed on ferrierite sample, covering a wide range of basic strength, in the order
acetonitrile > dimethyl ether > water > pyrrole > ammonia [53]. It has been
shown that the strength distribution depends on the basicity of the probe molecule.
Moreover the total number of neutralized sites appeared constant irrespective of the
nature of the probe, except for pyrrole, and nearly equal to the number of protons
(Fig. 3.17).
Using a series of amines as probe molecules, Parrilo et al. found a good correlation
between heats of adsorption and gas phase proton affinities, but not with the proton
transfer energies of those bases in aqueous phase [54–56]. These results indicate
that the proton transfer dominates the interaction between the adsorbate and the
acid sites. However, in a theoretical study published by Teraishi the fact that the
heat of ammonia adsorption depends both on the proton affinity and the ammonium
ion affinity was underlined [57]. With regard to the catalytic reaction in zeolites, the
activity depends not only on the proton affinity but also on the stability of the cationic
intermediate in the zeolite. The heat of ammonia adsorption, which includes the later
effect, is thus in disagreement with proton affinity and provides a different measure
of acidity, which is better suited to evaluate the acid strength of the zeolite in relation
with its catalytic activity.
3 Couplings 125

Fig. 3.17 Differential heats of adsorption of various probe molecules on a ferrierite sample versus
the probe uptake (from [53])

3.10.2 Probing Surface Basic Properties

The number of acidic probe molecules that are able to cover wide range of strength
is rather small. Moreover the difficulty in evaluating surface basic properties stems
from the fact that these molecules may interact simultaneously with cations (such as
Na+ ). The ideal probe molecule should be specific to basic sites, it should distinguish
interaction with oxide ion and hydroxyls and it should not give rise to chemical
reactions.
Carbon dioxide (pK a = 6.37), owing to its acidic character, is commonly used to
characterize the basicity of solids. A large number of species of this molecule on solid
surfaces can be formed; on basic hydroxyl groups there is a formation of hydrogen
carbonate species, on basic oxygen ions different kinds of carbonate species can be
formed. IR evidence concerning the formation of carbonate-like species of different
configurations has been reported for metal oxides, which accounts for the heterogene-
ity of the surface revealed by microcalorimetric measurements [58]. Localization of
the carbonate species formed either on surface (in the form of unidentate, bidentate
of bridged species) or in the bulk (polydentate species) is essential for the results
obtained by microcalorimetry. The heat evolved from the CO2 interaction can result
not only from its adsorption on basic surface sites, but also from its reaction with the
bulk. Moreover, if basic character of zeolites is investigated, CO2 can interact with
extra framework cation.
Sulphur dioxide (pKa = 1.89) is another molecule commonly used to investigate
basicity of solid catalysts. SO2 adsorption on the surface of metal oxides is complex.
Several types of species can be formed according to hydroxylation state and the acid-
ity or basicity of surface [58]. Interaction of SO2 with basic O2− leads to formation
126 D. Stošić and A. Auroux

of sulfite type species. The sulfite ion is pyramidal and it can coordinate to metal in
monodentate, bidentate or bridged form. Sulfur dioxide can also interact with sur-
face hydroxyl groups and form hydrogen sulfite species. It can also be coordinated
to Lewis acid sites where it forms weakly bonded species.
The number of CO2 adsorption sites is always much lower than that found for SO2
suggesting that adsorption of CO2 is more specific than that of SO2 . The latter, having
a very strong acidic character, therefore probes almost all the basic sites regardless
of their strength.

3.10.3 Probing Surface Redox Properties

The heats of reduction of oxide samples can be determined by studying the adsorption
of hydrogen, CO and various hydrocarbons on the fully oxidized catalysts [27, 59].
The extent of reduction of the catalyst surface can be evaluated in particular using
H2 . The measurement of hydrocarbon (e.g. propene, propane, acrolein, etc.) adsorp-
tion heats is complicated by the subsequent reaction of the adsorbed species or by
incomplete desorption of the products [60].
In the case of CO reduction, the catalyst–oxygen bond energy has to be calculated
by subtracting the heat of formation of CO2 .
However, it is known that, in the absence of processes other than plain surface
coordination, CO acts as a weak Lewis base and can interact with the strongest
surface Lewis acid sites. NO can also be employed either as a probe to identify
Lewis acid sites or as a reducing agent. However, NO may disproportionate into
N2 O and oxygen and it is also very likely to form nitrosyl complexes in the presence
of transition metal ions [60].
The heats of oxidation of the reduced oxides can be further measured using O2
adsorption. Large variations of the reoxidation heat can be sometimes observed when
any further oxidation is limited by the diffusion of oxygen into the reduced portion
of the particle [61].

References

1. L. Arnaut, S. Formosinho, H. Burrows, Chemical Kinetics: From Molecular Structure to Chem-

ical Reactivity (Elsevier, Amsterdam, 2007)
2. J.M. Tatibouet et al., Methanol oxidation as a catalytic surface probe. Appl. Catal. A Gen. 148,
213–252 (1997)
3. M. Baldani, I.E. Wachs et al., Methanol: a “smart” chemical probe molecule. Catal. Lett. 75,
137–149 (2001)
4. W.E. Farneth, J. Gorte et al., Methods for characterizing zeolite acidity. Chem. Rev. 95, 615–635
(1995)
5. N. Cardona-Martinez, J.A. Dumesic, Applications of adsorption microcalorlmetry to the study
of heterogeneous catalysis, in Advances in Catalysis, vol. 42, ed. by D.D. Eley, H. Pines, P.B.
Weisz (Academic, San Diego, 1992), pp. 149–244
3 Couplings 127

6. P.J. Andersen, H.H. Kung, Characterization of catalysts with microcalorimetry, in Catalysis,

vol. 11, ed. by J.J. Spivey, S.K. Agarwal (Royal Society of Chemistry, Cambridge, 1994), pp.
441–466
7. A. Auroux et al., Acidity characterization by microcalorimetry and relationship with reactivity.
Top. Catal. 4, 71–89 (1997)
8. V. Solinas, I. Ferino et al., Microcalorimetric characterisation of acid–basic catalysts. Catal.
Today 41, 179–189 (1998)
9. B.E. Spiewak, J.A. Dumesic et al., Microcalorimetric measurements of differential heats of
adsorption on reactive catalyst surfaces. Thermochim. Acta 290, 43–53 (1996)
10. A. Auroux, Thermal methods: calorimetry, differential thermal analysis, and thermogravimetry,
in Catalyst Characterization: Physical Techniques for Solid Materials, ed. by B. Imelik, J.C.
Védrine (Plenum Press, New York, 1994), pp. 611–635
11. P.C. Gravelle et al., Heat-flow microcalorimetry and its application to heterogeneous catalysis,
in Advances in Catalysis, vol. 22, ed. by D.D. Eley, H. Pines, P.B. Weisz (Academic, San Diego,
1972), pp. 191–263
12. A. Auroux et al., Acidity and basicity: determination by adsorption microcalorimetry, in Molec-
ular Sieves—Science and Technology: Acidity and Basicity, vol. 6, ed. by H. Karge, J. Weitkamp
(Springer, Heidelberg, 2008), pp. 45–152
13. Y.I. Tarasevich et al., Interaction of water and other polar substances with various sorbents
according to calorimetric and chromatographic data. Theor. Exp. Chem. 37, 197–214 (2001)
14. A. Auroux, M. Huang, S. Kaliaguine et al., Decrystallization process of HNaY zeolites upon
mechanical milling: a microcalorimetric and thermokinetic study. Langmuir 12, 4803–4807
(1996)
15. N. Cardona-Martinez, J.A. Dumesic et al., Acid strength of silica–alumina and silica studied
by microcalorimetric measurements of pyridine adsorption. J. Catal. 125, 427–444 (1990)
16. I. Ferino, R. Monaci, E. Rombi, V. Solinas et al., Microcalorimetric investigation of mordenite
and Y zeolites for 1-methylnaphthalene isomerisation. J. Chem. Soc. Faraday Trans. 95, 2647–
2652 (1998)
17. G.I. Kapustin, T.R. Brueva, A.L. Klyacho, A.M. Rubinshtein et al., Analysis of distribution
of adsorbate molecules within zeolite crystals by thermokinetics. Kinet. Catal. 22, 183–195
(1981)
18. A. Gervasini, A. Auroux et al., Thermodynamics of adsorbed molecules for a new acid–base
topochemistry of alumina. J. Phys. Chem. 97, 2628–2639 (1993)
19. Lj Damjanović, A. Auroux et al., Determination of acid/base properties by temperature pro-
grammed desorption (TPD) and adsorption calorimetry, in Zeolite Chemistry and Catalysis:
An Integrated Approach and Tutorial, ed. by A. Chester, E. Derouane (Springer, Berlin, 2009),
pp. 107–167
20. T.J. Gricus Kofke, R.J. Gorte, W.E. Farneth et al., Stoichiometric adsorption complexes in
H-ZSM-5. J. Catal. 114, 34–45 (1988)
21. T.J. Gricus Kofke, R.J. Gorte, G.T. Kokotailo, W.E. Farneth et al., Stoichiometric adsorption
complexes in H-ZSM-5, H-ZSM-12, and H-mordenite zeolites. J. Catal. 115, 265–272 (1989)
22. T.J. Gricus Kofke, G.T. Kokotailo, R.J. Gorte et al., Stoichiometric adsorption complexes in
[B]- and [Fe]-ZSM-5 zeolites. J. Catal. 116, 252–262 (1989)
23. J. Tittensor, R.J. Gorte, D. Chapman et al., Isopropylamine adsorption for the characterization
of acid sites in silica-alumina catalysts. J. Catal. 138, 714–720 (1992)
24. R.J. Gorte et al., What do we know about the acidity of solid acids? Catal. Lett. 62, 1–13 (1999)
25. C. Pereira, R.J. Gorte et al., Method for distinguishing Brönsted-acid sites in mixtures of
H-ZSM-5 H-Y and silica–alumina. Appl. Catal. A 90, 145–157 (1992)
26. A. Gervasini, C. Picciau, A. Auroux et al., Characterization of copper-exchanged ZSM-5 and
ETS-10 catalysts with low and high degrees of exchange. Micropor. Mesopor. Mater. 35–36,
457–469 (2000)
27. J. Le Bars, J.C. Védrine, A. Auroux, B. Pommier, G.M. Pajonk et al., Calorimetric study of
vanadium pentoxide catalysts used in the reaction of ethane oxidative dehydrogenation. J. Phys.
Chem. 96, 2217–2221 (1992)
128 D. Stošić and A. Auroux

28. P.F. Siril, A.D. Davison, J.K. Randhava, D.R. Brown et al., Calorimetric study of vanadium
pentoxide catalysts used in the reaction of ethane oxidative dehydrogenation. J. Mol. Catal. A:
Chem. 267, 72–78 (2007)
29. P.R. Westmoreland, T. Inguilizian, K. Rotem et al., Flammability kinetics from
TGA/DSC/GCMS, microcalorimetry and computational quantum chemistry. Thermochim.
Acta 367–368, 401–405 (2001)
30. A.J. Groszek et al., Flow adsorption microcalorimetry. Thermochim. Acta 312, 133–143 (1998)
31. P.F. Siril, D.R. Brown et al., Acid site accessibility in sulfonated polystyrene acid catalysts:
calorimetric study of NH3 adsorption from flowing gas stream. J. Mol. Catal. A: Chem. 252,
125–131 (2006)
32. S.P. Felix, C.S. Jowitt, D.R. Brown et al., Base adsorption calorimetry for characterising surface
acidity: a comparison between pulse flow and conventional “static” techniques. Thermochim.
Acta 433, 59–65 (2005)
33. V. Rakić, Lj Damjanović, V. Rac, D. Stošić, V. Dondur, A. Auroux et al., The adsorption of
nicotine from aqueous solutions on different zeolite structures. Water Res. 44, 2047–2057
(2010)
34. J.A. Thomson, J.E. Ladbury, Isothermal titration calorimetry, in Biocalorimetry 2 Applications
of Calorimetry in the Biological Sciences, ed. by J.E. Ladbury, M.L. Doyle (Wiley-VCH,
Weinheim, 2004), pp. 35–58
35. Y.Y. Lim, R.S. Drago, M.W. Babich, N. Wong, P.E. Doan et al., Thermodynamic studies of
donor binding to heterogeneous catalysts. J. Am. Chem. Soc. 109, 169–179 (1987)
36. C.W. Chronister, R.S. Drago et al., Determination of hydrogen-bonding acid sites on silica
using the Cal-Ad method. J. Am. Chem. Soc. 115, 4793–4798 (1993)
37. R.S. Drago, S.C. Dias, M. Torrealba, L. de Lima et al., Calorimetric and spectroscopic inves-
tigation of the acidity of HZSM-5. J. Am. Chem. Soc. 119, 4444–4452 (1997)
38. N. Kob, R.S. Drago, V. Young et al., Preparation, characterization, and acidity of a silica
gel/tungsten oxide solid acid. Inorg. Chem. 36, 5127–5131 (1997)
39. R.S. Drago, C.S. Dias, J.M. McGilvray, A.L.M.L. Mateus et al., Acidity and hydrophobicity
of TS-1. J. Phys. Chem. B 102, 1508–1514 (1998)
40. R.S. Drago, N. Kob et al., Acidity and reactivity of sulfated zirconia and metal-doped sulfated
zirconia. J. Phys. Chem. B 101, 3360–3364 (1997)
41. E.M. Arnet, R.A. Haaksma, B. Chawla, M.H. Healy et al., Thermochemical comparisons of
homogeneous and heterogeneous acids and bases. 1: sulfonic acid solutions and resins as
prototype Brönsted acids. J. Am. Chem. Soc. 108, 4888–4896 (1986)
42. E.M. Arnet, T. Absan, K. Amarnath et al., Thermochemical comparisons of solid and homoge-
neous acids and bases: pyridine and polyvinylpyridine as prototype bases. J. Am. Chem. Soc.
113, 6858–6861 (1991)
43. P. Carniti, A. Gervasini, S. Biella, A. Auroux et al., Intrinsic and effective acidity study of niobic
acid and niobium phosphate by a multitechnique approach. Chem. Mater. 17, 6128–6136 (2005)
44. S. Koujout, D.R. Brown et al., Calorimetric base adsorption and neutralisation studies of sup-
ported sulfonic acids. Thermochim. Acta 434, 158–164 (2005)
45. C. Guimon, H. Martinez, Recent Research Developments in Catalysis, vol. 2 (Research Sign-
post, Kerala, 2003)
46. H. Knözinger, Specific poisoning and characterization of catalytically active oxide surfaces, in
Advances in Catalysis, vol. 25, ed. by D.D. Eley, H. Pines, P.B. Weisz (Academic, San Diego,
1976), pp. 184–271
47. H.A. Benesi, B.H. Winquist, Surface acidity of solid catalysts, in Advances in Catalysis, vol.
27, ed. by D.D. Eley, H. Pines, P.B. Weisz (Academic, San Diego, 1997), pp. 97–182
48. Y. Mitani, K. Tsutsumi, H. Takahashi et al., Direct measurement of the interaction energy
between solids and gases X. acidic properties of hydroxyl sites of H-Y zeolite determined by
high-temperature calorimetry. Bull. Chem. Soc. Jpn. 56, 1917–1920 (1983)
49. K. Tsutsumi, Y. Mitani, H. Takahashi et al., Direct measurement of the interaction energy
between solids and gases IX. heats of adsorption of ammonia and pyridine on several solids at
high temperature. Bull. Chem. Soc. Jpn. 56, 1912–1916 (1983)
3 Couplings 129

50. J.C. Védrine, A. Auroux, G. Coudurier, Combined physical techniques in the characteriza-
tion of zeolite ZSM-5 and ZSM-11 acidity and basicity, in Catalytic Materials: Relationship
Between Structure and Reactivity, vol. 248, ASC Symposium Series, ed. by T.W. Whyte Jr, R.A.
Dalla Betta, E.G. Derouane, R.T.K. Baker (American Chemical Society, Washington, 1984),
pp. 253–273
51. Lj Damjanović, A. Auroux, Heterogeneous catalysis on solids, in Handbook of Thermal Analy-
sis and Calorimetry, vol. 5, ed. by M.E. Brown, P.K. Gallagher (Elsevier, Amsterdam, 2008),
pp. 387–438
52. A. Auroux, Z.C. Chi, N. Echoufi, Y. Ben Taarit, Calorimetric investigation of the acidity of
dealuminated Y-type zeolites using various basic probes, in Zeolites as Catalysts, Sorbents and
Detergent Builders, vol. 46, ed. by H.G. Karge, J. Weitkamp (Elsevier, Amsterdam, 1989),
pp. 377–387
53. A. Auroux, New probes for an accurate calorimetric determination of the acidity of zeolites,
in Innovation in Zeolite Materials Science, vol. 37, Studies in Surface Science and Catalysis,
ed. by P.J. Grobet, W.J. Mortier, G. Schulz-Ekloff (Elsevier, Amsterdam, 1988), pp. 385–391
54. D.J. Parrillo, R.J. Gorte, W.E. Farneth et al., A calorimetric study of simple bases in H-ZSM-5: a
comparison with gas-phase and solution-phase acidities. J. Am. Chem. Soc. 115, 12441–12445
(1993)
55. D.J. Parrillo, A. Biaglow, R.J. Gorte, D. White, Quantification of acidity in H-ZSM-5, in Zeolites
and Related Microporous Materials: State of the Art, vol. 34, Studies in Surface Science and
Catalysis, ed. by J. Weitkamp, H.G. Karge, H. Pfeifer, W. Hölderich (Elsevier, Amsterdam,
1994), pp. 701–708
56. D.J. Parrillo, R.J. Gorte et al., Characterization of stoichiometric adsorption complexes in
H-ZSM-5 using microcalorimetry. Catal. Lett. 16, 17–25 (1992)
57. K. Teraishi et al., Effect of Si to A1 substitution at next-nearest neighbor sites on the acid
strength: ab initio calculation of the proton affinity and the heat of ammonia adsorption. Micro-
por. Mater. 5, 233–244 (1995)
58. J.C. Lavalley et al., Infrared spectrometric studies of the surface basicity of metal oxides and
zeolites using adsorbed probe molecules. Catal. Today 27, 377–401 (1996)
59. M. Cabrejas Manchado, J.M. Guil, A. Perez Masia, A. Ruiz Paniego, J.M. Trejo Menayo et al.,
Adsorption of H2 , O2 , CO, and CO2 on a y-alumina: volumetric and calorimetric studies.
Langmuir 10, 685–691 (1994)
60. F. Witzel, H.G. Karge, A. Gutze et al., in Proceedings from the 9th International Zeolite
Conference, Montreal, Canada, 5–10 July 1992, vol. 2, ed. by R. von Ballamos, J.B. Higgins,
M.M.J. Treacy (Butterworth-Heineman, Boston, 1993), p. 283
61. L. Yang, O. Kresnawahjuesa, R.J. Gorte, A calorimetric study of oxygen-storage in Pd/ceria
and Pd/ceria–zirconia catalysts. Catal. Lett. 72, 33–37 (2001)
Chapter 4
Temperature-Programmed Desorption (TPD)
Methods

Vesna Rakić and Ljiljana Damjanović

Abstract This chapter presents the fundamentals, the experimental setups and the
applications of temperature-programmed desorption (TPD), method used to inves-
tigate the events that take place at the surface of solid material while its tempera-
ture is changed in a controlled manner. At the beginning, fundamental principles of
adsorption and desorption phenomena, as well as the data concerning first exper-
imental setups are given. Further, important information related to the construc-
tion of nowadays used equipment and the organization of common experiments are
underlined. The significance of data directly obtained from temperature-programmed
experiment—TPD profile, which are the area under it and the position of peak
maximum, are highlighted. Particular attention is given to the results that can be
derived from these data—characterization of active sites that can be found on the
surface of solid material and determination of kinetic and thermodynamic parame-
ters of desorption process. In this regard, the influence of important experimental
parameters on derived values is explained. Besides, the distinctions between TPD
experiments performed in ultra-high vacuum and in the flow systems (differences
in experimental setups and in the derivation of kinetic and thermodynamic parame-
ters) are explained. Also, the modification of temperature-programmed techniques,
known as temperature-programmed oxidation and temperature-programmed reduc-
tion are shortly explained and compared with temperature-programmed desorption
method. In the end, a brief comparison of the TPD and adsorption calorimetry, two
most widely used techniques for the study of acid/base properties of catalysts, is
given.

V. Rakić (B)
University of Belgrade, Nemanjina 6, 11080 Zemun, Serbia
e-mail: [email protected]
L. Damjanović
University of Belgrade, Studentski trg 12-16, 11000 Belgrade, Serbia
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 131

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_4,
© Springer-Verlag Berlin Heidelberg 2013
132 V. Rakić and L. Damjanović

4.1 Introduction

Temperature-programmed desorption (TPD) belongs to the group of techniques

in which an event that takes place at the surface of solid substance is moni-
tored, while the temperature of investigated sample is changed with a temperature
program β(t) = dT/dt. When the technique is applied to a system in which the
adsorption process is (at least in part) irreversible and surface reaction takes place,
this technique is often known as temperature-programmed reaction spectroscopy
(TPRS). Although most often investigated surface process is desorption, reactions
such as reduction, oxidation or sulfidation can be monitored using the techniques
named temperature-programmed reduction (TPR), temperature-programmed oxi-
dation (TPO) or temperature-programmed sulfidation (TPS), respectively. Usually,
temperature T is a linear function of the time t; in that case the heating rate β is
constant value.
In general, temperature-programmed methods are applicable for the investiga-
tion of both porous materials (such as real catalysts) and well-defined surfaces of
single-crystalline samples. In addition, their application is experimentally simple and
inexpensive, what explains their wide application in several scientific domains.
Temperature-programmed desorption originates from so-called flash desorption,
which was originally developed in early fifties of twentieth century to quantitatively
investigate the kinetics of molecular desorption from well-defined single crystal
surfaces in high vacuum [1]. Flash desorption involved the adsorption of a known gas
on the sample (in the form of a ribbon or wire, rigorously cleaned previously) while
kept in vacuum. Subsequently, desorption was provoked by heating while the pressure
in the system was recorded: as the temperature increased, certain previously absorbed
species had enough energy to desorb from the surface and would be detected as a
rise in pressure. The resulting pressure-time curve was referred to as a “desorption
spectrum” [1, 2]. In flash desorption, temperature was raised very quickly (from
1 to 1000 Ks−1 ). Two heating schedules were applied: a linear variation of sample
temperature with time: T = T0 + βt and its reciprocal variation 1/T = 1/T0 − αt (T0
is starting temperature, α and β are constants) [1]. From the beginning, the possibility
to extract the heat of adsorption from the obtained results was demonstrated.
Flash desorption was mainly applied for the investigations of low-surface area
substances such as metals. The method was later adapted for the investigations of
high-surface area materials, under carrier gas and ambient pressure. In the domain
of catalysis, the pioneering work in this direction was done by Amnenomiya and
Cvetanović [3], who studied the catalysts’ active sites, and needed the methodology
that enabled conditions similar to those ordinarily used in catalytic reaction. Their
apparatus allowed the pre-treatment of solid material by heating and evacuation, its
exposure to molecular species of interest (at low temperature) and subsequent pro-
grammed desorption performed by heating in a controlled manner with the possibility
to detect the desorbed gas in the carrier. This equipment consisted of two important
parts: a temperature programming controller and a thermistor type thermal conduc-
tivity cell for measuring the rate of desorption of pre-adsorbed molecules from the
4 Temperature-Programmed Desorption (TPD) Methods 133

surface. In this way, it was enabled to increase the temperature in a controlled manner,
slowly (about 1 to tenth degrees K per minute), in a manner similar to those applied in
real catalytic reactions.1 An additional important feature of their modified technique
was that it permitted simultaneous study of a chemisorption process and the surface
reaction which accompanied it. Applying the “temperature-programmed desorption”
(the name they used for this modification of flash desorption), Amnenomiya and Cve-
tanović discovered the existence of two different active sites for the adsorption of
ethylene on alumina [4]. Importantly, they employed the method to calculate the
values of energies for ethylene desorption.
Since that time, the method is widely developed; experimental setups are improved
and adjusted to many different purposes (e.g. for the investigations of oxidation and
reduction reactions). Today, two main types of equipment are available: those oper-
ating under ultrahigh vacuum and so-called “flow” systems. Well-defined surfaces
of single-crystalline samples are investigated in a continuously pumped ultra-high
vacuum (UHV) chamber (this technique is often referred to as thermal desorption
spectroscopy—TDS [5]). The equipment that is constructed to allow adsorption–
desorption in the gas flow are most often used for the investigation of porous materi-
als (catalysts, for example). Vacuum setups are customarily used for surface science
studies, but they can be also useful for the characterization of porous materials.
Generally, TPD can be described as the measurement of the rate of desorption
of adsorbed molecules, as a function of temperature. Therefore, this method can be
useful in the extraction of very important information. It can be used in the identifi-
cation and characterization of sites active in adsorption and catalytic reactions, in the
study of adsorption states, binding energies, surface concentration and desorption
kinetics. To summarize, this method is nowadays very important and often applied
for the characterization of materials used as catalysts. In this domain, two main areas
of applications are: the characterisation of acid/base properties of solid materials,
what is essential for understanding their reactivity2 ; and the determination of kinetic
and thermodynamic parameters of desorption processes or decomposition reactions.
In the following sections, the basic concepts of temperature-programmed methods
(primarily temperature-programmed desorption, but also temperature-programmed
reactions) are outlined. At the beginning, fundamental principles of adsorption and
desorption—their thermodynamic and kinetic aspects, are presented. Furthermore,
the descriptions of experimental setups, the data that can be obtained from the exper-
iments and their interpretation are given. The possibilities to extract the adsorption
energies and kinetic parameters from experimental results are discussed. Finally, the
examples of possible applications and the comparison of results obtained by TPD
with those obtained from adsorption calorimetry, are presented.

1 Significant difference in heating rates makes main distinction between “flash desorption”, where
the heating rate is very high (the desired temperature is reached in seconds) and temperature-
programmed desorption (where the sample is heated in minutes or even hours).
2 TPD is perhaps the most often used for estimation of acid/base properties of solid catalysts.
134 V. Rakić and L. Damjanović

4.2 Adsorption–Desorption; Fundamental Principles

Adsorption is defined as the enrichment of gas or liquid (adsorbate) at the surface

of a solid material (adsorbent); or as the increase in the density of the fluid in the
vicinity of an interface. Adsorption takes place at the active sites—specific points
on the solids’ surface that possess affinity toward the particles coming from the gas
or liquid phase; it happens when an attractive interaction between a particle from
adsorbate phase and a surface is strong enough to overcome the disordering effect
of thermal motion [6, 7].
In discussing the fundamentals of adsorption it is usual to distinguish between
physisorption, which takes place when the attractive interactions are essentially the
result of weak intermolecular forces; and chemisorption, which involves the over-
lapping between the molecular orbitals of the adsorbed particle and the surface
atoms (i.e. electron transfer which leads to the formation of chemical—covalent
bond between adsorbate molecule and the active site at the surface). Although this
distinction is very useful, it has to be pointed out that there are many intermediate
cases, and that it is not always possible to categorize a particular event at the surface
as physi- or chemisorption. However, there are some general characteristics which
distinguish these two possible types of adsorption. Chemisorption is highly specific,
while quite contrary, physisorption is non-specific; chemisorption may involve dis-
sociation, it is possible over a wide range of temperature and a monolayer is formed;
while physisorption takes place only at relatively low temperatures, multilayer can
be formed, dissociation of adsorbed species does not happen. However, the most
important distinction is the amount of heat that is associated to either one of these
two general types of adsorption: physisorption is characterized by low heat of adsorp-
tion (below approximately 50 kJ mol−1 ); while chemisorption is characterized by
high heat of adsorption, typically exceeding 50 kJ mol−1 . Hence, as a result of mole-
cular chemisorption, the weakening of intramolecular bonds inside the adsorbed
molecule can happen, which may lead to its dissociation [6–8]. The differences in
potential energies curves that present processes of physisorption and chemisorption
(dissociative and non-dissociative) are presented in Fig. 4.1.3
In addition, chemisorption is often an activated process, which means that the for-
mation of a chemisorptive bond requires overcoming the activation barrier (Fig. 4.1b);
it may be slow and irreversible. By contrast, physisorption is rapid, non-activated
and reversible process.

3 Dissociative chemisorption of a diatomic molecule can also happen through the dissociation in a
gas phase and a creation of two gas phase atoms; these two atomic species can be then adsorbed on
the surface (this way is almost always non-activated). If the curves describing molecular and atomic
adsorption intersect at or below the zero potential energy line, then the precursor physisorbed mole-
cule can experience non-activated dissociation, followed by chemisorption (Fig. 4.1a). In contrast,
if the energetic for these two pathways are such that the intersection occurs above the zero energy
plane, then chemisorption will be activated with activation energy, Ead , as indicated in Fig. 4.1b.
4 Temperature-Programmed Desorption (TPD) Methods 135

Fig. 4.1 Potential energy curves for (1) physical and (2) chemical adsorption: (a) non-activated
(b) activated. Epot - potential energy, Qc - heats of chemisorption, Qp - heats of physisorption, Ead -
energy of activation for desorption, Ediss - dissociation energy for the diatomic molecule. The sum:
Edes = Ead + Qc is the the heat of hemisorption, in the activated processes [8]

4.2.1 Thermodynamic View

Both physisorption and chemisorption are exothermic processes, what can be con-
cluded from a simple thermodynamic consideration. The adsorbed molecule has
at most two degrees of translational freedom on the surface; in addition, its rota-
tional freedom must always be less than that of the gas phase molecule. In total the
adsorbed molecule possesses less degree of freedom than the same molecule in the
gas phase. Consequently, the entropy change of adsorption Sads = Sads − Sgas is
obligatory negative. However, the adsorption is spontaneous process, which means
that the free energy change (G = H −TS) must be negative. Negative value
of entropy change Sads means that the second term in previous relation is positive
(−TSads ), what requires the value of H necessarily negative. Hence, adsorption
is always exothermic process [6].
The heats of adsorption provide a direct measure of the strength of the bonding
between adsorbate and the active site at the surface of solid substance. Therefore,
it is of importance to estimate these values, particularly in the domain of catalysis
where the strength of active sites determines the mechanism and the yield of certain
process. One possible way to determine the heat of adsorption is to apply calorimetry,
experimental technique which provides the heats of adsorption as a function of the
adsorbed amount (−H = f(na ), where n a is the adsorbed amount and −H is, in
that case, so-called differential heat) [9]. The heats of adsorption can be derived also
from the variation of adsorption with temperature. In that case, Clausius-Clapeyron
equation and the data from isosteric measurements are used (in that way, so-called
isosteric enthalpy of adsorption can be obtained).
136 V. Rakić and L. Damjanović

The extent of surface coverage (or simply surface coverage), reached as a result
of adsorption, is usually denoted as θ . It is a ratio between adsorbed particles number
(Nads ) and the number of adsorption sites available at a surface (usually denoted as
active sites −Nsur f ): θ = Nads / Nsur f . The chemical equilibrium between adsorbed
species and gas phase particles is reached when chemical potentials of adsorbate
particles in both phases are equal (the rates of adsorption and desorption are equal);
and it is characterized by constant value of surface coverage θ . The temperature
dependence of the gas pressure p required for equilibrium between the adsorption and
desorption can be calculated from the Clausius-Clapeyron equation [6]. Neglecting
the volume of the condensed surface phase, this relation becomes:
 
∂ln p qisost
= (4.1)
∂T θ RT 2

where qisost denotes so-called isosteric heats of adsorption. Evidently, these values
can be calculated from the temperature dependence of the adsorption isotherms, i.e.
from the isosteres, for each average temperature:
   
∂H ∂(ln P)
qisost = = −RT 2 (4.2)
∂θ T ∂T nr

where T is the absolute temperature, R the gas constant and nr the number of
reversibly adsorbed molecules.4
Third possibility to obtain the adsorption heats is to extract them from the data
acquired from temperature-programmed desorption experiments. This possibility
will be exposed in detail later, in the Sect. 4.5.2. However, for that purpose, it is
obligatory to know some basic postulates about the kinetics of adsorption and des-
orption; what is given in the following section.

4.2.2 Kinetics of Adsorption and Desorption

In the case of gas-phase adsorbate, the surface coverage θ is dependent on the gas
pressure. Adsorption isotherms relay the surface coverage and the gas pressure (at
constant temperature); the most known equation of this type is Langmuir adsorption
isotherm. It is based on the following assumptions [6]:

– one adsorbed particle interacts with one active site at the surface (once adsorbed,
it is immobile on the surface);

4 Instead to define equilibrium by constant surface coverage, it is possible to keep constant pressure
at the surface; in that case the equilibrium heat of adsorption qeq is incorporated in Clausius-
Clapeyron equation.
4 Temperature-Programmed Desorption (TPD) Methods 137

– the surface of adsorbent is saturated when all adsorption sites are covered with
the particles of adsorbate, i.e. the adsorption goes on until monolayer of adsorbate
particles is spread over the solid (then, θ equals 1);
– there are no interactions between the adsorbed particles.

The rates of adsorption (rads ) and desorption (rdes ) are proportional to the numbers
of empty or occupied active sites, θ or (1 – θ ), respectively:

rads (θ ) = An p (1 − θ )x
(4.3)
rdes (θ ) = Bn θ x

where rads and rdes are the rates of adsorption and desorption, respectively, An and
Bn are constants, while x is the kinetic order of surface event (x can be 1 or 2,
although adsorption and desorption are usually considered as the first order events).
In the equilibrium, the rates of adsorption and desorption are equal: |rads | = |rdes |,
therefore, surface coverage θ can be expressed as:

(bn p)1/x
θn = (4.4)
1 + (bn p)1/x

(where bn (T ) = An /Bn = const, x = 1, 2).

If adsorption is a reversible process (i.e. backward process—desorption, passes
through exactly the same states), the rates of both processes can be described using
the same equation:
dθ
r =− = k · θx (4.5)
dt
However, in contrast to adsorption which may or may not be activated process, des-
orption is always activated, with a minimum activation energy denoted as activation
a ). The rate constant for desorption can be expressed by
energy for desorption (E des
Arrhenius equation:  a 
E des
kdes = ν (θ ) · exp − (4.6)
RT

where ν (θ ) is pre-exponential factor, which is in general dependent on surface

coverage θ . Now, the rate of desorption becomes:
 a 
dθ E des
rdes =− = ν(θ ) · exp − · θx (4.7)
dt RT

The relation (4.7) is customarily used to describe the rate of desorption, and is
known as Polanyi-Wigner equation.
Desorption is often explained assuming the existence of transition state:

(A)ads → (A)ads
#
→ A gas + ()ads (4.8)
138 V. Rakić and L. Damjanović

where (A)ads is molecule of specific adsorbate adsorbed on the active site at the
#
surface ()ads , Agas is the gas-phase molecule and (A)ads is the transition complex.
The equilibrium between the adsorbate and transition state is defined as:

(A)ads
K# = (4.9)
(A)ads
#

Equilibrium constant for overall desorption process (A)ads → Agas + ()ads ,

(K) can be considered as equal to K # . The reaction rate constant is related with the
equilibrium constant K# (hence, with equilibrium constant K), as:

kT # kT
kdes = K = K (4.10)
h h
From the other side, thermodynamic functions are related with the equilibrium
constant through the known relation:

G = H − T S = −RT ln K (4.11)

The rate of desorption is customarily extracted from the experiments of

temperature-programmed desorption. Since the rate of desorption is related to the
equilibrium constant through the relations (4.5) and (4.10); it is evident that the rate
constant for desorption can be expressed as:
     
kT kT G des kT Sdes Hdes
kdes = K = exp − = · exp · − (4.12)
h h RT h R RT

Obviously, thermodynamic quantities (Hdes , Sdes and G des ) for the acti-
vated process such as desorption can be extracted from the data obtained from des-
orption experiments. Importantly, since the adsorption is spontaneous process (it
does not need the energy of activation), the heat of adsorption equals, in general,
the activation energy for desorption. Hence, desorption experiments provide also
thermodynamic parameters of adsorption.

4.3 Experimental Setups

The experimental setups for temperature-programmed desorption have evolved with

time. As it has been already stated, there are numerous experimental designs that
allow the application of this method under the conditions that are the same or very
similar to those applied in real catalytic reaction (or any other surface event). Nev-
ertheless, the various equipment used for these experiments, although different, is
all constructed to allow two main steps that are common for all thermal desorption
methods:
4 Temperature-Programmed Desorption (TPD) Methods 139

Fig. 4.2 (a) Experimental setup for temperature-programmed desorption performed in vacuum.
The sample (monocrystal or a thin layer of powder) is exposed at the sample holder, which is
connected to the system for temperature control. The pumping speed has to be high enough to allow
the monitoring of desorbed species by mass spectrometer. Inset presents the output of one TPD
experiment—the TPD profile [5]. (b) In the flow systems, sample is placed in the sample holder,
inside the furnace [10]

1. The admission of desired gas to the sample.

2. The heating of the sample in a programmed way.
In TPD experiment, gaseous molecules (atoms) of interest are adsorbed at the
surface, at constant temperature. The adsorption is very often performed at ambient
temperature, but can be sometimes done at sub-ambient or at elevated temperature.
In the modifications of technique such as TPO or TPR, gaseous species are con-
sumed while temperature is increased in a programmed manner. In the case of TPD
procedure, desorption of adsorbate is monitored while increasing the solid sample
temperature in a controlled fashion; while in the case of TPR/TPO, the consumption
of active gas is monitored during temperature increase, as explained later in more
details (see Sect. 4.4).
From previously stated, it follows that temperature-programmed experiments can
be performed under ultrahigh vacuum or in the flow of gas. Still, whatever is the
experimental design, three main parts of equipment are always necessary to perform
this kind of investigations:

• The system for the controlled admission of (different) gases. The adsorption
is commonly performed as isothermal process. Nowadays, it is possible to con-
struct the equipment which enables the adsorption of desired and precisely known
amount of adsorbate. In the past, small polar molecules (NH3 , CO, CO2 , SO2 ,
H2 O) have been usual adsorbates in the TPD studies. More recently, larger mole-
cules (such as hydrocarbons) and non-polar molecules (such as Ar2 or N2 ) have
been applied as adsorbates. Usually, adsorbates are denoted as “probes” or “probe
species”. Most often used are the probes customarily applied to titrate acidic or
basic surface sites (NH3 , pyridine, CO, CO2 , and SO2 ).
140 V. Rakić and L. Damjanović

• “Reactor” or sample holder, placed in a heated area (or furnace) where tem-
perature can be controlled. In the case of experimental setup constructed to allow
TPD in ultra-high vacuum (schematic presentation shown in Fig. 4.2a) sample is
deposited on a sample holder as a single crystal or monolayer, and connected with
a system for temperature control. In the systems that are designed to work in the
flow of gas, sample is placed in a reactor, which is usually a quartz tube placed in
furnace (Fig. 4.2b). By far, the most common approach is to increase the sample
temperature linearly with time at constant rates (β = dT /dt = const) that have
values between 0.5 K s−1 and 25 K s−1 .
• A system for detection of evolved gases. Heating of the sample provokes the
evolution of species from the surface back into the gas phase, which has to be
monitored. The detectors used for detection and possible quantification of evolved
gases are: thermal conductivity detector (catharometer), flame ionization detector,
conductometric titration and mass spectrometer.

– Catharometer serves for measuring the difference in thermal conductivity between

reference gas and the gas that flows through the sample (it is used only in flow
systems). This kind of detectors is often used in the variations of temperature-
programmed (TP) methods, known as TPO or TPR.
– Flame ionization detector is in specific use for the detection of organic effluents:
effluent enters in a flame obtained by combustion of hydrogen and air, then, the
ions that are formed are trapped by two electrodes (with a potential difference
between them). As a result, electric current appears and can be detected.
– Conductometric titration is applied if it is possible to entrap the evolved gas in an
aqueous solution; then, the change in conductivity can be detected.
– In modern implementations of temperature-programmed techniques the detector
of choice is a small, quadrupole mass spectrometer (QMS). The application of
mass spectrometer enables simultaneous acquisition of single or multiple masses
desorbed during heating. In fact, the application of this kind of detector enabled the
distinction between different species desorbed in the same time from the surface.

Nowadays, two main techniques that are most often used for detection of effluents
are mass spectrometry and thermal conductivity; the whole process is most often
controlled by computer.

4.4 The Design of Temperature-Programmed Experiment;

Obtained Data

The equipment used for TPD experiments have to be designed in a way which
allows the performance of certain steps that may be necessary in particular experi-
ment. Firstly, sample is placed in the sample holder (reactor) and pre-treated in the
appropriate way (in vacuum or in the flow of desired gas, at desired temperature); the
pre-treatment procedure depends on the characteristics of investigated material and
4 Temperature-Programmed Desorption (TPD) Methods 141

Fig. 4.3 Experimental setups for temperature-programmed desorption, reduction and oxidation.
(a) The reactor is placed inside the furnace which is connected with temperature programmer.
Detection of evolved gas(es) is performed by monitoring the variations in thermal conductivity of
gas mixture. (b) The TPD apparatus equipped with mass spectrometer as a detector [5]

Fig. 4.4 (a) Time dependence of adsorbate concentration upon exposing the solid sample. (b) An
example of TPD profile, drown as a signal of detector versus temperature. Common TPD profile
is a complex shaped curve. Figure presents the interaction (adsorption and desorption) of CO with
CoY zeolite [11]

the purpose of TPD experiment. Afterwards, the sample is exposed to the adsorbate.
Usually, the adsorption is performed isothermally, at appropriate temperature (fre-
quently at 300 K, but also at temperatures higher then this one, or even at sub-ambient
temperatures). Subsequently, physisorbed part of adsorbed gas is removed from the
surface, either by the evacuation, or by inert gas flow. The residual chemisorbed
adsorbate is desorbed by heating the sample in a controlled manner, preferably in
a way to give a linear temperature ramp; the analysis of the evolved gas (gases) is
performed to establish its identity and the amount, using the appropriate detection
system. The whole procedure is performed in situ.
Schematic presentation of one experimental setup that enables realisation of all
mentioned steps in vacuum is shown in Fig. 4.2, while Fig. 4.3 shows two typical
constructions designed for the experiments in the flow of appropriate gases.
The data obtained from one temperature-programmed experiment are presented
as the variation of a detector signal intensity (presented at y axis) as a function of time
(or temperature, presented at x axis). Consequently, as detector signal is proportional
to the concentration of the species desorbed from the surface, y axis values are
proportional to the rate of desorption (rdes ).
142 V. Rakić and L. Damjanović

It is important to point out that in the case of flow systems, temperature-

programmed techniques can be used for quantitative measurements. For that purpose,
the gas flow has to be controlled and constant in time. If this condition is achieved, the
detector signal can be properly calibrated by using dilute gas streams of known con-
centrations: the signal of such a known mixture is passed through the empty reactor
(or bypass reactor) and the signal intensity is monitored. Furthermore, the intensity
of signal obtained as a result of the same gaseous species evolved from the sample
can be considered as proportional to the value obtained for the known mixture. Once
calibrated in that way, the intensity of detector signal can be given in concentration
units (Fig. 4.4), and the measurement of evolved gas concentration becomes possi-
ble. The calibration of a detector signal enables the precise determination of both
adsorbed and desorbed amounts.
The precise amount of adsorbed gas can be obtained by passing it through “bypass”
reactor, and subsequently, through the sample (Fig. 4.4a). If the detector signal has
been previously properly calibrated, the monitoring of two signals (bypass and
through the sample) provides the true adsorption amount of respective gas at a given
temperature, which can be derived from the surface in between these two signals.
Precise amount of desorbed gas can be obtained from the surface under the TPD
profile (Fig. 4.4b).
It is very important to keep in mind that many different chemical species can
evaporate from the sample in the same temperature region (particularly in the case of
real catalytic systems). In the old versions of TPD setups, the overall desorbed amount
would be recorded as a rise in the pressure. It is especially important to point out
that the incorporation of mass spectrometer as a detector enabled the discrimination
of different products, desorbed in the same time (in the same temperature region).
Figure 4.5 shows one example: evidently, the evolution of ammonia and hydrocar-
bons happens in the same temperature region (from ∼370 ◦ C up to ∼440 ◦ C), during
the thermal decomposition of 1,2-diaminopropane entrapped inside a zincophosphate
structure.
The insight in the typical TPD profiles presented in Figs. 4.4 and 4.5 reveals what
information can be obtained from one TPD experiment. There are two main classes
of data that we can “read” from the desorption profile:
1. The area under the TPD profile, which is proportional to the amount of adsorbate
originally adsorbed, in other words, to the surface coverage θ . Under particular
circumstances, id est, if the limitations such are diffusion and/or readsorption can
be neglected (see later, in the Sect. 4.5.2.2), temperature-programmed technique
can be employed as an excellent tool for determination of surface coverage.
2. The position of the peak maximum (along the temperature scale, Tmax ), which
is related to the activation energy for desorption. Generally, the higher the value
of temperature of peak maximum (Tmax ) is, desorption is more difficult, which
is an indication of stronger interaction between the adsorbate species and the
active site on the surface. As it has been explained previously, the adsorption
is spontaneous process, therefore, the heat of adsorption equals, in general,
the activation energy for desorption. Hence, the value of Tmax is related to the
4 Temperature-Programmed Desorption (TPD) Methods 143

Fig. 4.5 Top TPD profiles of ammonia (m/z = 17) and hydrocarbons (Cx H y , m/z = 26) obtained as
a result of thermal degradation of 1,2-diaminopropane entrapped inside a zincophosphate structure-
ZnPO-HDAP. Bottom Thermogravimetric and differential thermogravimetric signal of this decom-
position [12]

heat (enthalpy) of adsorption; in other words, to the strength of adsorbate binding

to the surface. The methods for deriving the activation energy for desorption from
the position of Tmax will be discussed latter. However, it is very important to note
here that a simple TPD profile that possesses only one peak (and one Tmax , like a
desorption profile of mass 17, presented in Fig. 4.5) is not a common case that can
be found for the majority of investigated systems. In fact, very often there is more
than one binding state for the adsorbate molecules on a surface, which express
144 V. Rakić and L. Damjanović

significantly different adsorption enthalpies (consequently, significantly different

activation energies for desorption); therefore, this will give rise to multiple peaks
in the TPD spectrum (one example is desorption profile presented in Fig. 4.4b).
In that case, the determination of Tmax positions demands particular attention.

4.4.1 The Design of TPR/TPO Experiments; Obtained Data

In the modification of temperature-programmed methods known as temperature-

programmed reduction (TPR), a reductive gas (usually H2 mixed with an inert gas) is
consumed by the sample, while the temperature is increased with a constant heating
rate β. In the case of temperature-programmed oxidation, the gas or gas mixture
which can perform oxidation (O2 mixed with an inert gas, air, N2 O, etc.) is passed
through the sample. The reductive or oxidant gas consumption is monitored either
by mass spectrometer or by catharometer. If reduction is done by hydrogen, the
analysis is usually performed using a catharometer as a detector; the same stands
for TPO performed with oxygen as oxidative gas.5 The experiments of temperature-
programmed reactions (reduction, oxidation, or any other reaction, such is sulfidation,
for example) have to be performed in the systems that allow the calibrations of
detector signals [14]. Therefore, these experiments can be performed only using the
equipment that allows flow of different gases (as those setups presented in Fig. 4.3).
TPR and TPO profiles give information concerning the reduction or oxidation
(red-ox) state of the solid which is analyzed. These features are very important data
for commercial catalysts; what explains the vast application of TPR and/or TPO in the
characterization of solid materials for industrial applications. For example, reduction
is an inevitable step in the preparation of metallic catalysts [5, 15]. In addition, it is
often a critical step—if it is not performed correctly the catalyst may sinter or may
not reach its optimum state of reduction.
Similarly to the case of TPD, the data obtained from TPR or TPO experiments
are presented as the variation of detector signal intensity as a function of time (or
temperature). Generally, the data that can be obtained from temperature-programmed
reduction or oxidation are:
1. The difference in reduction (oxidation) temperature between different materials;
these differences are recognized from different positions of Tmax ;
2. The profile of TPR (TPO) pattern, which indicates the presence of one (or more)
species that can be reduced (oxidized);
3. The area under a TPR or TPO curve represents the total hydrogen (oxygen)
consumption, and is commonly expressed in moles of reductive (oxidative) gas
consumed per mole of metal atoms. Hence, the calculation of exact amounts of
those species which were reduced (oxidized) at the surface is possible.

5 In those cases, the consumption of either reductive or oxidative gas by the catalyst is derived
from the change in thermal conductivity of the gas mixture.
4 Temperature-Programmed Desorption (TPD) Methods 145

Table 4.1 The organization of temperature-programmed techniques

Temperature-programmed desorption Temperature-programmed reduction or
oxidation
Sample is pre-treated if necessary (by Sample is pre-treated if necessary (by
heating, or by flushing isothermally, in the heating, or by flushing isothermally, in the
desired atmosphere). desired atmosphere)
The sample is exposed to the adsorbate, The sample is purged by inactive gas,
isothermally. isothermally
Desorption of physisorbed part of a gas by The exposure of a sample to the reductive or
inert gas flow (isothermally, at the oxidative gas, in the linear heating regime.
appropriate temperature). The detector monitors the consumption of
this active gas. Subsequently, the sample is
cooled and purged in the inert atmosphere,
in order to remove the traces of active gas
Heating of a sample in the appropriate If necessary, the sample is exposed to
atmosphere, with the analysis of the oxidative (if the sample was previously
evolved gases. reduced) or reductive gas (if the sample
was previously oxidized). The detector
monitors the consumption of this active gas

4. The activation energy of the reduction can be estimated from the temperature
Tmax at which the reduction rate is maximal by using appropriate equations.
The organization of one TPR or TPO experiment is somehow different in com-
parison with that one applied for temperature-programmed desorption. For example,
the sample has to be purged with inactive gas, before exposure to active (reductive
or oxidative) gas.6 These differences can be seen in Table 4.1, which presents the
organisation of both TPD and TPR/TPO experiments.

4.5 The Interpretation of Results Obtained from

Temperature-Programmed Desorption Experiments

Temperature-programmed desorption technique offers very useful and important

methodology which can be applied for the characterization of materials used as
catalysts. There are two main fields of applications:
1. The characterisation of active sites of solid materials. It is of outmost impor-
tance to determine and understand the acid/base character of solid catalysts,
because these features are essential for their reactivity [16–20]. There are sev-
eral groups of techniques developed and particularly adapted for the investiga-
tion of acidity/basicity of solid catalysts; most of these methods are based on

6 If catharometer is used as detector, it is very important to remove traces of water or any other
impurities from the gas flows, because they would affect the thermal conductivity measurements.
146 V. Rakić and L. Damjanović

the adsorption of gas-phase probe molecules, which are chosen on the basis of
their reactivity, molecular shape and size [21–26, 13]. Among the other tech-
niques, TPD is of particular importance because its experimental conditions can
be organized in the same (or very similar) way as the conditions of real catalytic
reaction. The investigation of acid/base character of solids is perhaps the most
common application of TPD. For that purpose, many different chemical species
can be used as adsorbates (probes). In addition, the strength and the population of
specific active sites can be estimated, using the appropriate probes and applying
appropriate experimental conditions.
2. The determination of kinetic and thermodynamic parameters of desorption
processes, decomposition or other reactions. The interpretation of experimen-
tally obtained data and derivation of kinetic and thermodynamic parameters from
TPD results depends on the type of TP experiment: specific experimental con-
ditions have influence on the overall TP profile and on the position of Tmax
obtained either in ultra-high vacuum or in the flow system.
The details that explain more closely how the data obtained from TPD experiment
can be used to get the information concerning the characterization of active sites,
kinetic and thermodynamic parameters, are given in the following sections.

4.5.1 The Application of Temperature-Programmed

Desorption in Active Sites Characterisation

The characterization of active sites of solid catalysts includes the determination

of active sites nature, the estimation of their density (or population, i.e. the num-
ber of active sites per unit of mass or per unit of surface area), their strength and
strength distribution. Active sites can be acidic, basic and, in certain cases, ampho-
teric. All mentioned characteristics are very important for catalysts functionality;
therefore, many experimental techniques are invented and adapted for their investi-
gation. Among others, mainly spectroscopic methods (like NMR, IR, XPS, XRF…),
temperature-programmed desorption is particularly important because it can be use-
ful in the characterization of all mentioned features.
The strategy which is employed in order to get the above mentioned features of
catalysts’ active sites is the adsorption of appropriate gas phase probe, under the spe-
cific experimental conditions (that are chosen in a way to be similar to those applied
in the particular catalytic reaction), followed by subsequent desorption, monitored
with appropriate detector. One experiment, in which the characterisation of acid/base
properties of solid material is performed, is designed as follows:
– The sample is pre-treated in situ, in desired atmosphere (or in vacuum), at appro-
priate temperature, and during the appropriate time. Usually, the purpose of pre-
treatment is to remove water (eventually, some impurities) and/or to perform
degasification;
4 Temperature-Programmed Desorption (TPD) Methods 147

– The probe gas is admitted and adsorbed on the solid surface up to some specific
surface coverage or up to the saturation;
– Desorption process is performed.
According to the Lowry-Brönsted theory, a Brönsted acid is a proton donor, while
a Brönsted base is a proton acceptor. In Lewis’ concept, acid acts as electron-pair
acceptor, while base is electron donor (such as molecules possessing electron lone
pairs). Hence, a Lewis base is in practice equivalent to a Brønsted base. However,
the concepts of acidity are markedly different [27].
In the case of solid catalysts, any atomic (ionic) group at the surface that can
donate a proton is a Brönsted acid; while any place where one empty electron orbital
exists is Lewis acid. For example, in the case of zeolites, Brönsted acid site is a part
of microporous aluminosilicate framework—a bridging [≡ Si · · ·(OH) · ·· Al ≡]
configuration which is able to donate a proton to an acceptor; while Lewis acid site
is either tri-coordinated Al atom or charge-balancing cation Mez+ which are able to
accept the electron pair. Accordingly to the same theories, any place at the solid sur-
face which can accept proton is a Brönsted base; while any place which can donate
electron(s) is a Lewis basic site. For example, in the case of MeOx (metallic oxides),
the oxygen ions (Oz− ) behave as Brønsted bases (because they are proton accep-
tors); while cations at the surface possess Lewis acidity (they are electron acceptors)
[27, 28].
The probe molecules that are used to investigate surface acidity should be chosen
accordingly to their ability to accept proton from the surface active site, or to donate
electron pair to the solid surface. The molecules that fulfil these demands are, for
example, ammonia, pyridine, or hydrocarbons. Similarly, the probe molecules that
can be used to “trace” the basic site of solid catalysts must be able either to donate a
proton or to accept electron(s). Importantly, many species (that even do not contain
hydrogen in their formula, which is a demand according to Lowry-Brönsted theory)
can function as Lewis acid, accepting electron pair. Hence, the molecules that could be
chosen to investigate surface basicity are, for example, dioxides of carbon or sulphur.
However, acidity or basicity of a gas-phase adsorbate is not a sole criterion for its
choice as a probe molecule. Firstly, the strength of an acidic or basic probe should
be distinguished accordingly to its acid- or base-dissociation constant (K a or K b ). In
addition, very important feature of probe molecule is its radius. If there is a need to
locate all active sites in the structure of microporous solid material, the radius of probe
molecule has to be smaller then the diameter of pore(s) opening(s). In other words,
probe molecules have to be of appropriate size, so the entrance in the micropores
of the solid and the access of adsorbate to each active site become possible. For
example, ammonia, which is frequently used to reveal the acidic property of solids,
is selected as a probe due to its basicity and due to the size of the molecule. Its
molecule is smaller than the diameter of the pores in the zeolites’ structures, and also
in many other solids. The other probe often used for investigation of solids’ acidity
is pyridine; however, the application of other chemical species is also possible.
As it has been already stated, the value of Tmax is the indication of the strength of
the interaction adsorbate—active site. The stronger the active site is, the stronger the
148 V. Rakić and L. Damjanović

Fig. 4.6 TPD profiles of ammonia obtained from H-mordenite (SiO2 /Al2 O3 = 15.0) and
H-ZSM-5 (SiO2 /Al2 O3 = 23.8) [54]

interaction with probe molecule is, which causes more difficult desorption: higher
Tmax indicates that desorption is more difficult. The energy that have to be consumed
for desorption is related with the bond energy between surface active site and adsor-
bate; hence, the position of peak maximum provides information on the strength of
this bond. Solid materials possess active sites of different strength, i.e. they express
energetic heterogeneity. The origin of active sites strengths heterogeneity is usually
the consequence of the solids’ structure, or it can be result of different topologies
and chemical environments of active sites.
In the case of energetically heterogeneous surface, TPD curves are generally
complex-shaped profiles. Figure 4.6 presents two typical cases. Sometimes, desorp-
tion profile is composed of well resolved peaks, like upper TPD curve in the Fig. 4.6,
where two desorption peaks are denoted as low (l) and high (h), accordingly to the
temperature region of appearance. More often, desorption of probe molecules takes
place simultaneously from different sites, what gives more or less pronounced over-
lapping of the peaks (bottom TPD profile in Fig. 4.6, TPD profiles already presented
in the Figs. 4.4 and 4.5).
It is important to notice the influence of adsorption temperature (Tads ) on the
shape of TPD profile. Desorption takes part consequently to adsorption as a result of
thermal motion which kinetic energy is high enough to break the bond between the
adsorbate and weak active sites. If adsorption is performed at high temperature, TPD
profiles are either single-peak shaped, or overlapping of peaks is less pronounced. By
contrast, low temperature of adsorption allows the bonding of adsorbate with all active
sites present in the investigated structure. In that case, complex-shaped TPD profile
is obviously obtained, in the case of heterogeneous solid surface. The adsorption
temperature is apparently very important experimental condition: its influence on
the shape of TPD profile (hence, on the conclusions that can be derived from TPD
experiment) is illustrated by the example shown in Fig. 4.7. Evidently, the lower the
Tads value is, the more complex TPD profile is obtained, and vice versa. Of course,
4 Temperature-Programmed Desorption (TPD) Methods 149

Fig. 4.7 TPD signals obtained after pyridine was adsorbed at Y-type zeolite. The applied Tads :
(a) 150 ◦ C (b) 200 ◦ C (c) 250◦ C (d) 300 ◦ C [29]
150 V. Rakić and L. Damjanović

it has to be kept in mind that the terms “low” or “high temperature” are relative—
the choice of adsorption temperature depends on the particular investigated system
(on the pair adsorbate—solid surface). In addition, it is important to notice that, the
higher the Tads value is, the overlapping of peaks is less visible, while desorption
profile is shifted more to high-temperature region.
Usually, sites denoted as “weak”, “medium”, “strong” or “very strong” are recog-
nized from temperature-programmed desorption experiments, accordingly to the
temperature region of appearance.
Apparently, the overlapping of desorption peaks that has its origin in the surface
heterogeneity imposes the necessity to resolve the complex TPD curve and to assign
particular desorption profiles to the sites of definite strength. Generally, there are
two possible methods that can be applied in order to get information concerning the
presence and population of particular active sites on the surface of solid material.
Firstly, if a complex desorption profile has been obtained as result of probe adsorp-
tion, the mathematical procedure of deconvolution can be applied. Usually, decon-
volution is based on the assumption that the desorptions from different sites are
parallel and independent events of first order (in surface coverage) [31, 33]. Then,
desorption from the sites of same type and same strength would give symmetric
desorption profile, with well-defined single temperature of maximum. In that way,
certain number of symmetric desorption traces can be obtained, their sum should
give the overall TPD profile obtained from experiment. This procedure enables to
get information about the presence of different active sites in the investigated system
(from the number of single-peak symmetrical curves), their population (from the
percent with which each of these curves contribute in the area of overall desorption
profile) and their strength (from the positions of Tmax ). The procedure of deconvo-
lution can be performed relatively simply: the Tmax positions could be recognized
from the experimental TPD profile, the whole numerical procedure should be per-
formed as to choose the set of parameters in such a way to enable minimization
of standard deviations (in comparison between the linear sum of single desorption
profiles and the overall complex experimental curve). However, even though this
numerical procedure can be performed to give a unique deconvolution of the experi-
mental curves, it is recommended to compare the results obtained by deconvolution
with the information provided by other experimental techniques (for example, the
adsorption-desorption studied by FTIR spectroscopy).
Another possibility to investigate desorption from a heterogeneous surface and to
recognize the presence of some particular active sites is to perform step-wise filling
of the surface with the probe. When the active (probe) gas is admitted to the solids’
surface, the first interactions would be those between the strongest active sites and
the probe molecules. Therefore, in this approach, the main idea is to admit small
quantities of probe gas to enable the adsorption on the most active (the strongest)
sites separately, and to continue with the filling of surface, step by step. This task
can be fulfilled in two ways:
– The usage of experimental setup which enables the admission of controlled
amounts of a gas-phase probe;
4 Temperature-Programmed Desorption (TPD) Methods 151

Fig. 4.8 TPD signals,

obtained after pyridine was
adsorbed on AlPO4 − Al2 O3 .
The step-wise filling of the
surface is achieved by apply-
ing different Tads : (a) 50 ◦ C
(b) 100 ◦ C (c) 150 ◦ C (d)
200 ◦ C (e) 300 ◦ C [30]

– The variation of adsorption temperature in order to start the adsorption at the

highest possible temperature.
It is evident from the example presented in Fig. 4.7 that when the lower the temper-
ature of adsorption is applied more complex TPD profiles are obtained. By contrast,
the higher the adsorption temperature is, less complicated TPD profile is obtained.
Hence, if high enough temperature is applied, the interaction with one single type
of energetically homogeneous centres should be expected; what should give a sym-
metric, single-peak desorption curve. Figure 4.8 presents more obvious example.
Evidently, step-wise filling of AlPO4 − Al2 O3 surface with pyridine has enabled
more information about the population and strength distribution of active sites of this
solid catalyst. For example, the adsorption of pyridine performed at high temperature
(300 ◦ C) enabled to reveal the existence of some very strong active sites, which
population is low.
This presentation of active sites characterisation by TPD has been started with
the statement that the applied strategy in investigation of active sites’ characteristics
is the adsorption of appropriate gas phase probe, under the specific experimental
conditions. Evidently, the first criterion that has to be applied in order to choose
the gas-phase probe is its acidity/basicity. Through this text, the importance of gas-
phase nature, the size of its molecules and the temperature of adsorption, has been
considered. At this place, the importance of the nature of gas-phase probe will be
underlined. Figure 4.9 presents the example in which the investigation of same solids
has been performed using two different probes: HY and dealuminated HY zeolites
have been investigated using ammonia and pyridine, respectively. It is possible to
see from Fig. 4.9a that the adsorption of ammonia revealed the existence of three or
four different types of acid centres (in the case of HY and dealuminated HY zeolite,
respectively). However, the same solids possess only two different types of sites
152 V. Rakić and L. Damjanović

Fig. 4.9 (a) TPD profiles of ammonia obtained from HY and dealuminated HY zeolites. (b) TPD
profiles of pyridine obtained from HY and dealuminated HY zeolites. Overall TPD profile is pre-
sented by spotted line; dashed lines, obtained by deconvolution, present desorption from the acid
sites of different strength [31]

for pyridine adsorption, Fig. 4.9b. Apparently, pyridine molecule could not reach
all active sites in microporous zeolitic structure. Evidently, not only the basicity of
gas-phase molecule is important for its application as a probe, the diameter of its
molecule seems to be decisive in the example presented in Fig. 4.9.
It should be emphasized that the choice of a probe molecule should be done by
taking into account all relevant parameters, and having in mind the features of solid
material at which surface this probe should be adsorbed. In fact, the solid surface and
the gas which is chosen as a probe for the characterization of its active sites should
be considered as a pair. Very often, the separate adsorption-desorption experiments
of more then one gas-phase probe is necessary in order to obtain reliable information
concerning all active sites for particular solid material. The adsorption-desorption
of more than one probe molecule should complete the picture about the catalysts’
active sites, particularly in the case of complex systems, where different types of
active sites and energetic heterogeneity could be expected.
It is worth noting that the improvement of equipment available for temperature-
programmed desorption, made from the first experiments in the field until nowadays,
has enabled the application of many gases as probes. In fact, the possibility to perform
adsorption at low temperatures (even sub-ambient) and the improvements in the
detection systems allowed to introduce the gases which molecules are poorly polar
(such as CO), or even non-polar (such are inert gases and saturated hydrocarbons).
4 Temperature-Programmed Desorption (TPD) Methods 153

Fig. 4.10 TPD profiles of

argon, obtained from different
solid acids. Tads = 113 K [32]

The interactions of these probes with the active sites are often realised through
dispersion forces. Applied as probes, these kinds of molecules enable the recognition
and “titration” of very weak active sites. Being on the level of dispersion forces, the
interactions of inert gases with the active sites can be considered as specific, in some
cases. In Fig. 4.10, the application of argon as probe gas is shown.
The application of same probe for the characterization of different solids revealed
important differences in the strength and strength distributions of acid sites. It can
be seen from the results presented in Fig. 4.10 that Tmax positions and the shapes of
desorption profiles differ for all investigated systems; while the strongest sites and
the most pronounced heterogeneity are found for two specific solids (mordenite-type
zeolite and Cs-salt of heteropolyacid). It is important to note that if some less specific
probe would be used (such are ammonia or pyridine) all investigated solids should
express significant acidity and heterogeneity. In that case, the differences among
these catalysts would not be noticed.
The adsorption of non-polar gases offers the possibility to find an appropriate
probe which would allow desorption in the desired range of temperature. Figure 4.11
presents one example of hydrocarbons’ application as probe gases. Single crystals
possess energetically homogeneous surface for adsorption-desorption of these mole-
cules. Figure 4.11 shows that, in the case of adsorption on Rh(111), progressively
higher peak desorption temperature is noticed with increasing molecular weight of
adsorbate (saturated hydrocarbon).
At the end of this section, it can be concluded that in the domain of active sites
characterisation, there is a large body of methods and techniques that are developed
154 V. Rakić and L. Damjanović

Fig. 4.11 TPD profiles of dif-

ferent saturated hydrocarbons
obtained from single-crystal
Rh (111) surface [34]

for the determination of nature, population, strength and strength distribution of

active sites. Among them, temperature-programmed desorption, a well-established
and simple technique, continues to be a very useful and often applied in active sites
characterisation. It is important to keep in mind that, although TPD of appropriate
probe can be successfully used for distinction between the sites of different strength,
it does not enable to distinguish their nature. From the interaction with the probe gas
it can be concluded whether some active sites are acidic or basic; but, it can not be
concluded if they belong to Brönsted or Lewis type. In order to get precise data on
the active sites’ nature, the application of other techniques is needed.
Despite the evident richness of data derived from TPD experiment, there are
several significant limitations of this technique. A short summary of data that can
be obtained from TPD experiments and the limitations of technique are given in
Table 4.2.

4.5.2 The Application of TPD in the Determination of Kinetic

and Thermodynamic Parameters of Desorption Processes

Temperature-programmed desorption is by far the most often used technique for

determination of kinetic parameters on both model and real systems. From these
experiments, kinetic and thermodynamic information can be extracted under the
conditions of variable temperature. In the following section, the procedures of eval-
uation of these important parameters will be presented.
4 Temperature-Programmed Desorption (TPD) Methods 155

Table 4.2 Experimental data that can be obtained from TPD experiments; data that can be derived
and the limitations of technique
The obtained data Evaluated data Comments/limitations
Surface under the TPD profile, Surface coverage (θ) The possibility to determine the
which is proportional to the surface coverage is one of the
amount of desorbed (i.e. major advantages of TPD.
adsorbed) gas However, for that purpose, the
quantification of detector signal
is necessary
The values presented on y axis are The activation energy The strength and population of
proportional to the rate of for desorption, hence, sites active for adsorption
desorption (rdes ). The position the enthalpy of (surface reactions) can be
of the peak maximum (Tmax ), adsorption. evaluated. Often, complex
which is related to the activation Pre-exponential factor desorption profiles are obtained;
energy for desorption. Tmax is a for desorption additional procedures of
temperature where rdes is deconvolution or adsorption
maximal under different experimental
conditions is necessary. It is not
possible to distinguish the type
of active sites

Most often, the estimation of kinetic parameters is based on the assumption of

independent desorption that takes part from different active sites as the first order
event. In the interpretation of data obtained from TPD experiments, it is also assumed
that desorption is the sole surface event. However, in reality the readsorption and dif-
fusion of probe molecules take part, as events consecutive to desorption. These effects
are particularly prominent in the case of microporous solids, and in the experiments
performed in the flow systems. In those cases, the results obtained in TPD experiment
can be often misinterpreted. The readsorption and diffusion can be avoided by adjust-
ing some important experimental parameters. First of all, high heating rate would not
favour these processes. However, the choice of very high heating rate is not recom-
mended, particularly in modern implementations of TPD; where the investigation of
solid materials used in real catalytic systems is the most common application.
If readsorption is important part of surface event, its influence on temperature-
programmed desorption is that TPD profile broadens towards higher desorption tem-
peratures; the same stands for diffusion limitations. Therefore, the task to neglect or
minimize the effects of these processes is imposed; and it is relatively easy if TPD
experiments are performed in UHV setups. The pumping speed should be sufficiently
high to prevent readsorption of the desorbed species back onto the surface. However,
in the case of flow systems there are many experimental conditions that have to be
adjusted; even though, the interpretation of data obtained from TPD experiment is
not simple—in the estimation of kinetic parameters, those experimental conditions
have to be taken into consideration. Therefore, the derivation of kinetic and thermo-
dynamic parameters from the results obtained in UHV and in the flow system will
be discussed separately.
156 V. Rakić and L. Damjanović

4.5.2.1 The Interpretation of Data Obtained in Ultra-High Vacuum Systems

If the pumping speed is high enough, readsorption may be ignored and the rate of
desorption, defined as the change in adsorbate coverage per unit of time, is given
by Eq. (4.7). In a TPD experiment temperature (T) is usually increased linearly with
time from some initial temperature T0 , with the heating rate β:

T = T0 + β · t , dT = β · dt (4.13)

where all symbols have the same meaning as previously stated (see Sect. 4.2.2).
The intensity of the desorption signal, I(T ), which is proportional to the rate at
which the surface concentration of adsorbed species is decreasing, i.e. to the rate of
desorption, can be expressed by combining Eqs. (4.7) and (4.13):
 a 
dθ ν (θ ) θ x −E des
I (T ) ∝ − = exp (4.14)
dT β RT

Molecular adsorption and desorption are often the first order events (x = 1). The
maximum desorption signal will occur when the first derivative of signal intensity
with temperature equals 0 (dI/dT = 0):
  a 
d ν (θ ) θ −E des
exp =0 (4.15)
dT β RT

Since surface coverage changes with temperature, i.e. θ = f(T), this derivative is:
 a   a 
a
ν (θ ) θ E des −E des ν (θ ) −E des dθ
· exp + exp · =0 (4.16)
β RT 2 RT β RT dT

The derivative of surface coverage with temperature (dθ/dT ) can be substituted from
Eq. (4.14). In that way, Eq. (4.16) is transformed to:
  a   a 
ν (θ ) θ a
E des ν (θ ) −E des −E des
− exp exp =0 (4.17)
β RT 2 β RT RT

The solution of Eq. (4.17) can be obtained by setting the expression in square brackets
to be equal to zero; from where the relation between the temperature at which the
desorption maximum (Tmax ) appears and E des a is obtained as:

 a   2 ν
a
E des ν (θ ) −E des RTmax
= exp and: E des = RTM ln
a
(4.18)
2
RTmax β RTmax a β
E des

These equations give the relations between the temperature Tmax at which the
desorption maximum appears and the activation energy for desorption. Hence, a
a should be to analyze the TPD curve in order
simple approach to obtain the value E des
4 Temperature-Programmed Desorption (TPD) Methods 157

to get easily accessible parameter such as the temperature Tmax . Unfortunately, the
differential equation in (4.7) and (4.14) can not be solved analytically, so the value
a can not be obtained simply by substituting T
E des max value in Eq. (4.18). Therefore,
the derivation of kinetic parameters can be rather complicated task, in particular
because the kinetic parameters usually depend on surface coverage. However, we can
note that several facts can be stated from each temperature-programmed experiment:
– as the activation energy for desorption increases peak temperature Tmax increases;
– the peak temperature is not dependent upon, and consequently, does not change
with the initial coverage, θt=0 ;
– after the desorption maximum, the shape of the desorption peak tend to be asym-
metric, with the signal which decreases rapidly.
Consequently, the values of Tmax are evident from the experimental result. The
procedure that can be applied to derive kinetic parameters is to solve Eq. (4.18) itera-
tively, applying a suitable choice for pre-exponential factor ν(θ ) (for chemisorption
this value is typically 1013 s−1 ). The procedure is to read T max from the measurement,
a in the right-hand side of Eq. (4.18) and to calcu-
to insert an estimated value for E des
a
late the resulting E des value. The obtained value has to be fed back into Eq. (4.18) to
yield an improved value. The iterations should be done until the difference between
two subsequent iterations becomes negligible [5].
In the case of second-order desorption, a similar, although more complicated
expression exists for second-order desorption kinetics. In this case, the maximum
desorption signal will occur when the second derivative of surface coverage is equal
to zero: 
a β a 
−E des
d 2θ E des dθ dθ
2
= 2
− ν exp 2 θ =0 (4.19)
dt RTmax dt RTmax dt

From Eq. (4.19) the relation analogue to (4.18) can be derived:

 2 ν 
RTmax
a
E des = RTmax ln a β2θ (4.20)
E des

Again, iterations are necessary in order to estimate the desorption energy. The
insight in the Eq. (4.20) reveals evidence that θ has to be known (or estimated) at the
point where Tmax is reached.
Apart from this approach which implies the evidence of Tmax , there is another
which includes the value of peak width in the analysis. Also, many authors rely on
the application of other, even more simplified methods that enable the calculation of
kinetic parameters. Particularly popular among surface scientists are the Redhead’s
and Kissinger’s methods.
From all previously stated, it can be inferred that the starting point for extraction
of kinetic parameters from thermodesorpion profiles is desorption rate equation pro-
posed by Polanyi and Wigner (Eqs. (4.7) and (4.14)) [6]. However, it has to be kept in
mind that the term θ n is just one particular case of one general function f(α), where
α denotes the reacted (desorbed) fraction (the degree of surface event) and f(α) is the
158 V. Rakić and L. Damjanović

reaction kinetic model. Therefore, generalized form of Eq. (4.14) can be written as:
 a 
dθ ν (θ ) E des
= f (α) exp (4.21)
dT β RT

The methods that are derived for the calculation of kinetic parameters from TPD
profiles can bee divided in two big groups, shortly presented by following text.
(i) Integral methods are based on the temperature of desorption rate maximum
(Tmax ) and/or peak half-widths. These methods assume that pre-exponential
factor, reaction order and activation energy are coverage independent values.
The most known is Redhead method [1], where Eq. (4.7) is solved in order to
find the temperature at which desorption rate expresses its maximum. For the
first-order desorption (x = 1), the relation between the temperature of peak
maximum (Tmax ), activation energy, heating rate and pre-exponential factor is:
 a 
E ν1 E des
= exp − (4.22)
2
RTmax β RTmax

where ν1 is pre-exponential factor for the first-order desorption. The relation

between activation energy and Tmax is almost linear; therefore for ν/β values
which are between 108 and 1013 ◦ , Eq. (4.22) can be written as:

E ν1 Tmax
= ln − 3.46 (4.23)
2
RTmax β

The activation energy can be determined by varying heating rate β and plotting
ln(Tmax ) values against lnβ, without assuming the value of rate constant.
For the second-order desorption (x = 2), the relation analogue to (4.23) is:
a  a 
E des 2θmax ν2 E des
= exp − (4.24)
2
RTmax β RTmax

where ν2 is pre-exponential factor for second-order desorption, θmax is the adsor-

bate coverage at T = Tmax , and it is assumed that θmax = θ0 /2, with θ0 being
the initial surface coverage.
Equation is approximately correct for the first-order desorption and for values
of ν/β between 108 and 1013 ◦ . It is very often applied to determine E des from
a single TPD spectrum.
(ii) Differential methods are based on the assumption that at the temperature of
maximum, the second derivative of desorption rate (Eqs. (4.7), (4.14) or (4.21))
is equal to zero. The most known is Kissinger method [40]. If reaction rate is
expressed by (4.21), the second derivative is:
4 Temperature-Programmed Desorption (TPD) Methods 159
 a β  a   
d 2α E des  E des dα
= + ν f (αmax ) exp − =0 (4.25)
dt 2 2
RTmax RTmax dt max

where αmax and (dα/dt)max are reacted fraction and reaction rate at the maxi-
mum; while heating rate β should be constant. From (4.25) it follows:
 a 
Eβ  E des
2
= −ν f (αmax ) exp − (4.26)
RTmax RTmax

Equation (4.26) can be rearranged after taking logarithms:

   
β νR Ea
ln 2
= − a f  (αmax ) − des (4.27)
Tmax E des RTmax

Evidently, for the first order reaction, f’(α) = −1, and (4.27) becomes:
  a
β νR E des
ln 2
= ln a − (4.28)
Tmax E des RTmax

The procedure of extracting the activation energy for desorption is to analyse a set
of TPD profiles measured with different constant heating rates β, and to plot graphs
of left hand side of (4.28) versus 1/Tmax , what should lead to a straight line whose
slope gives the activation energy, independently of the value of reacted fraction,
αmax , at this point.
Evidently, for application of either integral or differential methods, the values of
Tmax have to be detectable. In case of poorly resolved TP profiles, their application
would not provide reliable kinetic parameters. In those cases, either deconvolution
of complex desorption profiles should be performed or desorption would be done
under different experimental conditions, so the resolving of simple desorption profiles
becomes possible.

4.5.2.2 The Interpretation of Data Obtained in the Flow Systems

Previously presented procedures for evaluation of kinetic parameters would give reli-
able values only if desorption is a lone surface process that takes part as a result of
temperature increase. In the case of UHV systems, where samples are usually spread
in a thin layer, diffusion takes part in a very limited extent; while readsorption can
be avoided using sufficiently high pumping speed. However, if experiment is orga-
nized in a flow of a gas, in one usual physical situation, desorption and readsorption
are occurring simultaneously with diffusion. In these systems, the construction of
sample holders (Fig. 4.2b) does not allow neglecting mass transfer and readsorption
limitations. These effects are particularly significant in the case of porous samples.
Therefore, the interpretation of results obtained in the systems designed to allow
temperature-programmed experiments in the flow of gas requires consideration of
160 V. Rakić and L. Damjanović

all parameters that may induce mass transfer and readsorption limitations. The para-
meters that have to be considered are related to the gas (carrier or adsorbate), the
geometry of furnace and sample holder, and the features of the sample.
The nature of both carrier and probe gas should be important, as well as their
purities and flow rate. High gas flow can provoke desorption of weakly bound species.
High amount of desorbed species that arrive in carrier gas can change its purity,
so the sensitivity of experiment can be reduced. The consequences of low flow
are: diffusion and readsorption effects become more probable, the time between
desorption and detection is longer. In addition, appropriate gas flow has to be chosen
to avoid concentration gradients within the catalyst particles and along the length of
the bed. Hence, a compromise between low and high flow must be found.
The characteristics of furnace and sample holder that may influence the desorption
profile are bed length, diameter and porosity, while the characteristics of the sample
that could be important are its weight, particle size radius, sample density, particle
porosity and number of active sites. In order to avoid temperature gradients, the
reactor can not be of big size; hence, mass of the sample is limited by its size and
geometry. Diameters of sample particles are another important factor - small particles
decrease the possibility of intra-granular diffusion.
Some additional parameters such are temperature range, heating rate, the tempera-
ture detection and monitoring, and distance between sample holder and detector may
influence the shape of temperature-programmed profile. Distance between reactor
and detector has to be the smallest possible, so the answer of detector is instanta-
neous. Thermocouple has to be precise enough to enable the time of furnace response
appropriate, so the temperature rise is absolutely linear. In addition, it has to be kept
in mind that temperature of desorbed gas (which has to be analysed) can change dur-
ing the experiment, what can cause a so-called “apparent” concentration. This effect
can be minimized if high flow of gas, low mass of the sample and the equipment
situated in the constant temperature area, are employed.
Evidently, in order to calculate reliable kinetic parameters, the temperature of peak
maximum Tmax has to be exactly the temperature at which the rate of desorption
is maximal. However, additional surface events (diffusion and readsorption) can
influence the TPD profile. The coupling of readsorption and mass transfer effects
with desorption can shift the peak of TPD curve significantly.
Therefore, it is necessary to select suitable operating conditions that enable to
avoid effects that could have influence on temperature-programmed profile. The ways
how to find experimental conditions required to obtain reliable activation energies
have been discussed in the literature. The recommendations that help to find appro-
priate sets of experimental parameters for experiments of temperature-programmed
desorption [41–43] or other tempetrature-programmed techniques can be found
[14, 44]. Once limitations that arrive from diffusion and readsorption are minimized,
simplified procedures can be applied to evaluate kinetic parameters.
In the case of desorption which takes part in the flow of gas, material balance
can be obtained from the assumption of equilibrium which is reached in a time t,
between the adsorbed and the gas-phase molecules. In the absence of diffusion, this
equilibrium can be expressed by following relations:
4 Temperature-Programmed Desorption (TPD) Methods 161

dθ
− Csm = Csm kdes θ − k ads C g (1 − θ ) (4.29)
dt

where Csm is concentration of adsorbed molecules (in mol kg−1 , for θ = 1) and C g
is concentration of adsorbate molecules in the gas phase (in mol dm−3 ); while kdes
and kads are rate constants of desorption and adsorption. If temperature increase is
linear, with constant heating rate (β), previous relation is transformed:

dθ
− Csm β = Csm kdes θ − k ads C g (1 − θ ) (4.30)
dT

If the gas flow (F, dm3 s−1 ) is constant in time and the mass of adsorbent is known
(W, kg), the same equilibrium can be expressed as:

FC g = Csm W kdes θ − W kads C g (1 − θ ) (4.31)

From Eq. (4.31), the concentration of adsorbate molecules in the gas phase can be
obtained from:
C sm W kdes θ
Cg = (4.32)
F + W kads (1 − θ )

The same value can be related with the rate of desorption (dθ/dT ) through the
equation obtained from (4.30) and (4.31):

Csm Wβ dθ
Cg = − (4.33)
F dT

Readsorption is negligible if the gas flow is high enough (F >> Wk ads (1−θ )). In
that case the concentration of adsorbate molecules in the gas phase can be obtained
from:
C sm W kdes θ
Cg = (4.34)
F
At the temperature of peak maximum (Tmax ), the concentration of adsorbate
molecules in gas phase reaches its maximum, so dC g /dT = 0. Since, the rate of
desorption is expressed by Eq. (4.7), for the first order desorption which takes part
without readsorption and diffusion limitations, the value of rate constant reached at
Tmax , (kdes )m , is related with Tmax value as:
a
E des
(kd )m = β 2
(4.35)
RTmax

The relation between Tmax and activation energy for desorption is given by equa-
tion identical to (4.27):
162 V. Rakić and L. Damjanović
  a
2
Tmax E des Ea
ln = ln + des (4.36)
β νR RTmax

It can be seen that for constant heating rate β, the value Tmax is independent on
surface coverage θ. The
 activation energy for desorption can be obtained from the
T2
slope of a plot: ln βm = f (1/Tmax ).
If a flow of carrier gas is not high enough (F << W k ads (1 − θ )) Eq. (4.32) is
transformed to:
C sm kdes θ θ
Cg = = C sm K (4.37)
kads (1 − θ ) 1−θ
   
where K = exp SRdes · − H des
RT . Again, having in mind that at the temperature
of peak maximum dC g /dT is equal to zero, the relation between Tmax and the heat
of adsorption can be obtained:
⎛ ⎞
 
2
Tmax (1 − θ ) 2
W H H
= ln ⎝  ⎠+
max
ln  (4.38)
β R F exp Sdes RTmax
R

Evidently, the heat of adsorption can be obtained from the slope of the same plot as
previously. Similarly, it can be shown that for desorption which takes part as a second-
order surface event, the activation energy for desorption (in the case of negligible
readsorption) or the heat of adsorption (in the case of significant readsorption) could
be obtained from the slope of the same plot. Hopefully, even in the cases when
desorption is significantly affected by side effects such are readsorption or diffusion,
kinetic parameters can be obtained using relatively simple procedures.
After everything that has been said about temperature-programmed methods, few
examples will be considered further in the text. The two experimental techniques most
commonly used for the study of acid/base properties of porous solid materials are
temperature-programmed desorption (TPD) and adsorption calorimetry. Application
of these techniques for characterisation of several different classes of materials will
be presented as well as comparison of data obtained by both techniques.

4.6 The Examples of TPD Application; the Comparison

with Data Obtained by Adsorption Calorimetry

As discussed in details previously, in TPD experiments, temperature increases lin-

early and the concentration of desorbed gas is recorded as a function of temperature,
whereas calorimetry involves the adsorption of gases onto the sample’s surface while
it is kept at a constant temperature and a heat-flow detector emits a signal propor-
tional to the amount of heat transferred per unit time. The peak maxima temperatures
in the TPD spectra are influenced by the active site strength, the number of active
4 Temperature-Programmed Desorption (TPD) Methods 163

sites, the structure of investigated material and the heating rate [45, 46]. In particular,
adsorption microcalorimetry gives access to the number, strength, and strength dis-
tribution of the acid sites in a single experiment [21]. This information is of outmost
importance for design of catalysts with high activity and selectivity.
Every micro or mesoporous material can be investigated by these two techniques.
Among many, few examples will be presented for the most often investigated cata-
lysts; such are zeolite, oxides and metals.

4.6.1 Zeolites

Zeolites are known to be important catalysts for a number of industrially important

reactions. A question of basic interest, which provides opportunity for development
of catalyst with suitable and tailored characteristics, is to determine the correlation
between number, strength and strength distribution of active sites and the promotion
of catalytic activity. Therefore, the investigation of acid sites, both Lewis and Brön-
sted type, is very important subject. Properties of zeolites as catalysts will depend
on many factors: the adsorption or desorption temperature of the probe, pretreatment
of the sample, proton exchange level, influence of coking as well as Si/Al ratio and
dealumination and influence of exchanged cations [47].

4.6.1.1 Influence of the Si/Al Ratio and Dealumination

The Si/Al ratio plays a significant role, since the aluminum atom is directly related
to the acidic site. Dealumination processes can promote porous structure modifica-
tions, which may improve some interesting properties of zeolites, like thermal and
hydrothermal stability, acidity, catalytic activity, resistance to aging and low coking
rate, and matter transfer. However, a severe dealumination may also cause a loss of
crystallinity [47].
Different dealumination processes have been proposed, namely steaming and acid
treatments, as well as reactions with SiCl4 or SiF2− 6 [47]. The removal of aluminum
from zeolite crystals leads to products with high framework Si/Al ratios. Some of
the aluminum atoms are released from the framework and form non-framework
aluminum-containing species. The non-framework aluminum species can be elimi-
nated by treatment with diluted hydrochloric acid. Dealumination generally brings a
decrease in the acid site concentration. However, the extent of the indicated decrease
varies with the kind of base probe, and a significant change was observed by Mitani
et al. [48] in the ratio of acid site concentrations when titrated with pyridine instead of
ammonia. An important increase of the initial heat values and of the site strength het-
erogeneity was observed for samples presenting many extra-framework aluminum
species. Samples subjected to a moderate dealumination and nearly total removal of
the extra-framework aluminum displayed a homogeneous acid strength [49].
164 V. Rakić and L. Damjanović

Fig. 4.12 Influence of dealumination on NH3 -TPD profiles of HY zeolites [14]

The effect of dealumination on NH3 -TPD profiles of HY zeolites is shown in

Fig. 4.12 [14]. Evidently l-peak became smaller upon dealumination, while the h-
peak increased up to a maximum for Si/Al = 15 and then decreased upon further
dealumination.
Figure 4.13 shows the effect of varying the Si/Al ratio of a MFI sample on its
TPD profile. As shown at the figure, both the l- and h-peaks became smaller upon
dealumination. A curve-fitting analysis led to the determination of average adsorption
heats that were almost constant for all investigated MFI samples, ca. 130 kJ mol−1
[35].
The acidity of H-ZSM-5 zeolites synthesized with different Al contents has been
characterized by microcalorimetric measurements of the differential heats of adsorp-
tion of ammonia [50]. The strength of the strongest acid sites increased with the Al
content to a maximum for Si/Al = 17.5 and then decreased notably. The total acidity
increased regularly with Al content. The importance of selecting appropriate Si/Al
ratios for specific catalytic applications is therefore obvious.
Figure 4.14 represents the differential heats of adsorption on dealuminated H-
ZSM-5 zeolite samples at 393 K versus the adsorbed amount of ammonia [36]. It
appears that the adsorption heat decreases gradually with the amount of adsorbed
ammonia and exhibits various steps, attributed to populations of sites of different
strengths. A heterogeneous distribution of sites is clearly evidenced by the shape
of the curves and can be attributed to the presence of extra-framework aluminium
spices. The samples adsorb ammonia with initial differential heats that vary between
160 and 177 kJ mol−1 . Karge et al. [51] have reported that, in first approximation,
the sites evolving Qdi f f > 150 kJ mol−1 can be assigned to strong Lewis sites, while
the sites with Qdi f f. = 150 − 100 kJ mol−1 typically correspond to strong Brönsted
4 Temperature-Programmed Desorption (TPD) Methods 165

Fig. 4.13 NH3-TPD profiles observed on H-MFI with Si/Al ratios: (a) 11.9 (b) 12.5 (c) 19 (d) 38
(e) 50 [35]

Fig. 4.14 Differential heats of ammonia adsorption at 393 K on H-ZSM-5 zeolites with Si/Al
ratios: (•) 14 (◦) 25 () 37.5 () 50 () 75 [36]

acid sites. This assumption is supported by a comparison of the numbers of acidic

sites obtained from infrared and microcalorimetric measurements. The initial heat
of adsorption for the zeolite with Si/ Al = 75 is lower than that of the other zeolites
except for Si/Al = 14, and the decrease in heat with coverage is steeper. The heat
evolved falls abruptly from an initial value of 150 kJ mol−1 to 70 kJ mol−1 at around
50 % coverage.
166 V. Rakić and L. Damjanović

4.6.1.2 Influence of Substitution by Other Cations

The nature of the exchanged cation is one of the key points that determine acidity
in zeolites. It is very important to use an acidic probe able to distinguish the alkali
cations from the basic sites.
Differential heats of ammonia adsorption and NH3 -TPD profiles of HZSM-5
zeolite as well as FeZSM-5, Cu-ZSM-5 and MnZSM-5 zeolites are presented in
Figs. 4.15 and 4.16. As can be seen in Fig. 4.15, the overall acidity of investigated sam-
ples was not significantly modified by ion-exchange procedure. However, changes
in the Qdi f f versus NH3 uptake profiles, particularly in their middle parts (140 –
65 kJ mol−1 ), indicate that the distribution of strength of acid sites was affected by
ion exchange. Ion exchange with Cu and Mn resulted in enhanced heterogeneity of

Fig. 4.15 Differential heat of adsorption versus NH3 uptake on HZSM-5 and ion exchanged ZSM-5
zeolites [37]
Fig. 4.16 NH3 − TPD pro-
files for HZSM-5 and ion
exchanged ZSM-5 zeolites
[37]
4 Temperature-Programmed Desorption (TPD) Methods 167

the acid site strength, as confirmed by the NH3 -TPD profiles, which present poorly
separated desorption peaks and even a single very broad peak for sample MnZSM-5
(Fig. 4.16) [37].
This example clearly shows that TPD and adsorption microcalorimetry are com-
plementary techniques, and combination of these two techniques provides very good
characterisation of active sites number, strength and distribution on particular solid
material.

4.6.2 Metal Oxides

Metal oxides, either bulk, doped, supported or mixed, are widely used as catalysts
in chemical industry. Catalytic behavior of these materials, in terms of activity and
selectivity, is related to their acid/base properties.
Metal oxide surfaces react with gases or solutions; they can be used as active
phases or as supports for catalysts. The behavior of metal oxide surfaces is con-
trolled by: (i) coordination—sites of low coordination are in general more reac-
tive than sites of high coordination; (ii) acid/base properties—clean and anhydrous
metal oxide surfaces present two different types of active sites, cations and anions
(acid/base pairs) which determine reactivity towards gas-phase adsorbates; (iii) the
redox mechanism—when the oxide deviates from the stoichiometry due to the pres-
ence of defects such as vacancies or adatoms, the oxidation state of surface atoms
varies [52].
Molecular and dissociative adsorption can be understood as acid/base processes.
Molecules adsorbing without dissociation always bind to one or several metal cations.
Ammonia and pyridine are the most commonly used probes for determining the acid
site strength of oxides.
In the case of supported metal oxide catalysts, the role of the support is to disperse
the active phase and to create new active surface species by host (active phase)
– guest (support) interaction. The dispersion of the active phase plays a fundamental
role, and very often a maximum of strength of the active sites is observed when the
monolayer coverage is reached. The pure oxides, such as Al2 O3 , ZrO2 , TiO2 (the
most frequently used catalyst supports) carry both basic and acidic Lewis sites on their
surface; depending on the probe molecules used (CO2 or NH3 ), they can exhibit either
acidic or basic character. Excess negative or positive charges can be induced, and
therefore acidity (Brönsted or Lewis) or basicity can be generated by mixing oxides.
Modifying the surface with a minor anionic, cationic or metallic component enhances
or decreases the acidic or basic strength of the sites. For example, the incorporation
of chloride, fluoride or sulfate ions increases the acidity of carrier oxides (Al2 O3 ,
ZrO2 , TiO2 ), while alkali cations enhance the basic strength of alumina or silica [53].
Silica-aluminas (amorphous aluminosilicates) are widely used as catalyst supports
due to their high acidity and surface area. The behavior of silica-alumina surfaces
is similar to that of zeolites concerning the initial differential heats of ammonia and
168 V. Rakić and L. Damjanović

Fig. 4.17 TPD curve for γ -alumina, using pyridine as probe molecule [38]

pyridine, but the total number of acidic sites varies with the preparation method and
the Si/Al ratio.
Auroux et al. [38] have studied acid properties of silica, alumina and multilayers
of silica on alumina (SA) and alumina on silica (AS), obtained by grafting. The
surface acidity of the pure oxides and samples obtained by grafting, SA and AS, of
both Lewis and Brönsted type, has been investigated by TPD and microcalorimetry,
using pyridine as probe molecule.
Figure 4.17 shows the TPD curve of pure alumina; a semi-quantitative estimation
of the number of acid site can be derived from the TPD plots such as this one. Three
classes of acid sites have been considered, depending on the desorption temperature:
weak, (T D < 523 K), medium (523 < T D < 673 K) and strong, (T D > 673 K) acid
sites.
Calorimetric results are presented in Fig. 4.18 where the differential heat of adsorp-
tion (Q diff ) of pyridine is plotted versus coverage. From the calorimetric data, the
number of sites and their distribution, according to the adsorption energies, can
be determined. Analyzing calorimetric profile of differential heats vs. surface cov-
erage, following sites active for adsorption can be distinguished: weak acid sites,
(90 ≤ Q diff <120k J mol −1 ), medium acid sites, m (120 ≤ Q di f f <150k J mol −1 )
and strong acid sites, s (Q diff ≥ 150k J mol −1 ).
From Fig. 4.18 it can be seen that calorimetric curve for pure alumina shows three
type of sites with different strength, in accordance with TPD result. Also, it can be
seen that the grafted mixed oxides SA and AS have acidic properties different from
those of the pure alumina and silica supports used as starting materials, i.e. properties
of catalysts depend on preparation procedure.
It can be noticed that some differences exist between results obtained from calori-
metric measurements and those revealed from TPD results, with regard to the sites
4 Temperature-Programmed Desorption (TPD) Methods 169

Fig. 4.18 Differential heats of pyridine adsorption as function of the coverage degree [38]

strength distribution: the percent of medium acid sites, derived from TPD results is
always remarkably higher than one found using microcalomitery. This is due to the
fact that strong acid sites having Qdi f f > 180 kJ mol−1 might not completely release
the organic base at 693 K and therefore are not disclosed by the TPD technique. This
implies importance of knowing limitations of techniques used for characterizations
as well as operative conditions adopted during experiments [38].

4.6.3 Metals

TPD and microcalorimetric methods provide an effective means for measuring the
strengths of adsorbate-surface interactions, not only on clean metal surfaces, but also
on metal surfaces that have been exposed to reaction conditions. Many recent studies
use H2 , CO, O2 and hydrocarbons as probe molecules since they are involved in
numerous commercial catalytic processes.
Bimetallic catalysts have been the subject of great interest for a long time because
of their exceptional properties compared to the monometallic catalysts, yet the reason
behind their improved activity is still a question of debate and they are subject of
many recent studies. Tanskale et al. [39] have studied the promoting effect of Pt and
Pd in bimetallic Ni–Pt and Ni–Pd catalysts supported on alumina nanofibre (Alnf)
for the liquid phase reforming of sorbitol to produce hydrogen. Fig. 4.19 shows
TPD profiles for CO desorption for several monometallic and bimetallic catalysts
dispersed on alumina nanofibre.
Results obtained by temperature-programmed desorption suggested that in the
case of bimetallic catalysts there was a reduction in the number of strong CO-
adsorption sites. This finding allows conclusion that the alloying effect of these sys-
tems leads to the lowering of the CO heat of adsorption. This finding was confirmed by
direct measurement of differential heat of CO chemisorption in the microcalorimetry
experiment (see Fig. 4.20).
170 V. Rakić and L. Damjanović

Fig. 4.19 Results from CO-TPD of monometallic and bimetallic catalysts over alumina nano-
fibre support; rate of CO desorption. Pt/Alnf (), Ni/Alnf (•), Pd/Alnf (), Ni/Pd/Alnf (), Ni-
Pd/Alnf(♦) [39]

Fig. 4.20 Differential heats of CO adsorption as a function of coverage, adsorption temperature

3 ◦ C. Pt/Alnf (), Ni/Alnf (•), Pd/Alnf (), Ni/Pd/Alnf (), Ni-Pd/Alnf (♦) [39]

The differential heat of adsorption for Ni–Pt/Alnf was reduced to 111.28 kJ mol−1 ,
which was 11.45 and 5.51 kJ mol−1 lower than Pt/Alnf and Ni/Alnf, respectively.
A similar result was obtained for the Ni–Pd bimetallic catalyst. This was interesting
because the atomic ratio of Pt:Ni and Pd:Ni was only 1:33 and 1:18, hence even
small amount of surface concentration of solute in the alloy catalysts is sufficient to
significantly alter the properties of constituent monometallic catalysts.
The differential heat profiles are characterised by a plateau of nearly constant heat
of adsorption at low coverage (15–20 μmol g m−1 ) followed by an abrupt decrease
as the surface saturation limit is reached. In the low coverage plateau region the
4 Temperature-Programmed Desorption (TPD) Methods 171

adsorption of CO can be considered to be strongly bonded and at high coverage

the differential heat of CO adsorption represents an average heat from the various
adsorption sites on the surface of a given catalyst particle. Therefore, one can compare
only the initial heat of adsorption when the CO chemisorption may be considered
to be equilibrated. The initial heat of CO adsorption on Pd/Alnf was found to be
the highest (130.09 kJ mol−1 ) followed by Pt/Alnf (122.73 kJ mol−1 ) and Ni/ALnf
(116.79 kJ mol−1 ). The initial heat of CO adsorption on Ni–Pt/Alnf was measured
as 111.28 kJ mol−1 , which is lower than both Pt/Alnf and Ni/Alnf. Similarly, the
initial differential heat of Ni–Pd/Alnf (109.60 kJ mol−1 ) was lower than both the
corresponding monometallic catalysts. This indicates that the addition of noble met-
als, even in small fractions, to the Ni catalyst has a significant but weaker promoting
effect on the adsorption of CO. Finding that the CO differential heat of adsorption
is lowered in the bimetallic catalysts is substantial because with reduction of the
CO binding strength poisoning of the active metal sites can be avoided which will
increase its catalytic activity [39].

4.7 Conclusion

This chapter has presented the fundamentals, the experimental setups and the appli-
cations of temperature-programmed desorption (TPD). TPD is widely utilized for
characterization of active sites present on the surface of solid materials and deter-
mination of kinetic and thermodynamic parameters of desorption processes. It is a
powerful technique, even if it does not provide direct information about molecular
nature of adsorbed species. Given its relative simplicity and low cost, this technique
will continue to find more applications in the future.
Thermal method which is complementary to TPD, adsorption calorimetry, pro-
vides tools necessary for measuring the energy of an adsorption system as a function
of coverage, allowing precise determination of number of surface active sites as well
as their relative populations and strengths. In general, adsorption calorimetry pro-
vides a much better description of the surface active site strength distribution than
TPD. Combination of these two methods allows detailed characterization of various
materials, in particular catalysts, and will surely find more application in different
research areas such as hydrogen production and storage or environmental chemistry.

References

1. P.A. Redhead, Thermal desorption of gases. Vacuum 12, 203–211 (1962). doi:10.1016/0042-
207X(62)90978-8
2. G. Ehrlich, Modern Methods in surface kinetics: flash desorption, field emission microscopy,
and ultrahigh vacuum techniques. Adv. Catal. 14, 255–427 (1963). doi:10.1016/S0360-
0564(08)60341-7
172 V. Rakić and L. Damjanović

3. R.J. Cvetanović, Y. Amenomiya, Application of temperature-programmed dessorption tech-

nique to catalyst studies. Adv. Catal. 17, 103–149 (1967). doi:10.1016/S0360-0564(08)60686-
0
4. R.J. Cvetanović, Y. Amenomiya, A temperature programmed desorption technique for inves-
tigation of practical catalysts. Catal. Rev. 6, 21–48 (1972). doi:10.1080/01614947208078690
5. I. Chorkendorff, J.W. Niemantsverdriet, Concepts of Modern Catalysis and Kinetics (Wiley-
VCH Verlag GmbH & Co. KGaA, Weinheim, 2003)
6. F. Rouquerol, J. Rouquerol, K. Sing, Adsorption by Powders and Porous Solids, Principles
Methodology and Applications (Academic Press, San Diego, 1999)
7. D.M. Ruthven, Principles of Adsorption and Adsorption Processes (Wiley, New York, 1984)
8. M.A. Vannice, Kinetics of Catalytic Reactions (Springer Science Business Media, Inc., New
York, 2005)
9. Lj. Damjanović, A. Auroux, Determination of acid/base properties by temperature-
programmed desorption (TPD) and adsorption calorimetry, in Zeolite Chemistry and Catalysis:
An Integrated Approach and Tutorial, ed. by A.W. Chester, E.G. Derouane (Springer, Heidel-
berg, 2009)
10. J.M. Kanervo, T.J. Keskitalo, R.I. Slioor, A.O.I. Krause, Temperature-programmed desorption
as a tool to extract quantitative kinetic or energetic information for porous catalysts. J. Catal.
238, 382–393 (2006). doi:10.1016/j.jcat.2005.12.026
11. V. Rakić, V. Dondur, U. Mioč, D. Jovanović, Microcalorimetry in the identification and charac-
terization of the most reactive active sites of heterogeneous catalysts. Top Catal. 19, 241–247
(2002). doi:10.1023/A:1015328526702
12. Dj. Stojaković, N. Rajić, V. Rakić, N. Zabukovec Logar, V. Kaučić, Structure and thermal behav-
ior of the layered zincophosphate [NH3 −CH2 −CH(NH3 )−CH3 ](ZnPO4 )2 . Inorg. Chim. Acta.
362, 1991–1995 (2009). doi:10.1016/j.ica.2008.09.020
13. B. Brunner, Solid state NMR–a powerful tool for the investigation of surface hydroxyl groups
in zeolites and their interactions with adsorbed probe molecules. J. Mol. Struct. 355, 61–85
(1995). doi:10.1016/0022-2860(95)08867-U
14. D. Delahay, Méthodes en température programmée, in Les matériaux micro et mésoporeux,
Charactérisation. ed. by F. Thibault-Starzyk, Groupe français des zéolithes, (EDP sciences,
Les Ulis, France, 2004)
15. S. Bennici, A. Auroux, Thermal analysis and calorimetric methods, in Metal Oxide Catalysis,
eds. by S.D. Jackson, J.S.J. Hargreaves (Wiley, New York, 2009)
16. G.A. Somorjai, Modern surface science and surface technologies: an introduction. Chem. Rev.
96, 1223–1236 (1996). doi:10.1021/cr950234e
17. W.E. Farneth, R.J. Gorte, Methods for analyzing zeolite acidity. Chem. Rev. 95, 615–635
(1995). doi:10.1021/cr00035a007
18. A. Corma, Inorganic solid acids and their use in acid-catalyzed hydrocarbon reactions. Chem.
Rev. 95, 559–614 (1995). doi:10.1021/cr00035a006
19. J. Weitkamp, U. Weis, E. Ernst, New aspects and trends in zeolite catalysis, in Catalysis by
Microporous Materials, Studies in Surface Science and Catalysis, vol. 94, eds. by H.K. Beyer,
H.G. Karge, I. Kiricsi, J.B. Nagy (Elsevier, Amsterdam), p. 363
20. V. Solinas, I. Ferino, Microcalorimetric characterisation of acid-base catalysts. Catal. Today.
41, 179–189 (1998). doi:10.1016/S0920-5861(98)00048-0
21. A. Auroux, Microcalorimetry methods to study the acidity and reactivity of zeolites, pil-
lared clays and mesoporous materials. Top Catal. 19, 205–213 (2002). doi:10.1023/A:
1015367708955
22. A. Auroux, Innovation in zeolite materials science, in Proceedings of the International Sumpo-
sium on Studies in Surface Science and Catalysis, Nieuwpoort, vol. 37 (Elsevier, Amsterdam,
1988), 13–17 September 1987, eds. by P.J. Grobet, W.J. Mortier, E.F. Vansant, G.G Schulz-
Eklo, p. 385
23. A. Zecchina, S. Bordiga, G. Spoto, L. Marchese, G. Pterini, G. Leofanti, M. Padovan, Silicalite
characterization. 2. IR spectroscopy of the interaction of carbon monoxide with internal and
external hydroxyl groups. J. Phys. Chem. 96, 4991–4997 (1992). doi:10.1021/j100191a048
4 Temperature-Programmed Desorption (TPD) Methods 173

24. M.A. Makarova, K.M. Al-Gefaili, J. Dwyer, Brönsted acid strength in US-Y: FTIR study of CO
adsorption. J. Chem. Soc. Faraday Trans. 90, 383–386 (1994). doi:10.1039/FT9949000383
25. Y. Kuroda, T. Mori, Y. Yoshikawa, S. Kittaba, R. Kumashiro, M. Nagao, What are the important
factors determining the state of copper ion on various supports? Analysis using spectroscopic
methods and adsorption calorimetry. Phys. Chem. Chem. Phys. 1, 3807–3816 (1999). doi:10.
1039/A904754I
26. E. Garrone, B. Fubini, B. Bonelli, B. Onida, C.O. Arean, Thermodynamics of CO adsorption
on the zeolite Na-ZSM-5 A combined microcalorimetric and FTIR spectroscopic study. Phys.
Chem. Chem. Phys. 1, 513–518 (1999). doi:10.1039/A806973E
27. J. Sauer, Acidic sites in heterogeneous catalysis: structure, properties and activity. J. Mol. Catal.
54, 312–323 (1989). doi:10.1016/0304-5102(89)80149-X
28. P.A. Jacobs, Carboniogenic Activity of Zeolites (Elsevier Scientific Publishing Company, Ams-
terdam, 1977)
29. E. Selli, L. Forni, Comparison between the surface acidity of solid catalysts determined by
TPD and FTIR analysis of pre-adsorbed pyridine. Microporous Mesoporous Mater. 31, 129–
140 (1999). doi:10.1016/S1387-1811(99)00063-3
30. J.M. Campelo, A. Garcia, D. Luna, J.M. Marinas, A.A. Romero, Characterization of acidity in
AlPO4 -Al2 O3 (5–15 wt% Al2 O3 ) catalysts using pyridine temperature-programmed desorp-
tion. Thermochim. Acta. 265, 103–110 (1995). doi:10.1016/0040-6031(95)02379-G
31. H.G. Karge, V. Dondur, J. Weitkamp, Investigation of the distribution of acidity strength in
zeolites by temperature-programmed desorption of probe molecules. 2. Dealuminated Y-type
zeolites. J. Phys. Chem. 95, 283–288 (1991). doi:10.1021/j100154a053
32. H. Matsuhashi, K. Arata, Temperature-programmed desorption of argon for evaluation of sur-
face acidity of solid superacids. Chem. Commun. 387–388, (2000). doi:10.1039/A909844E
33. H.G. Karge, V. Dondur, Investigation of the distribution of acidity in zeolites by temperature-
programmed desorption of probe molecules. I. Dealuminate mordenites. J. Phys. Chem. 94,
765–772 (1990). doi:10.1021/j100365a047
34. J. Wilson, H. Guo, R. Morales, E. Podgornov, I. Lee, F. Zaera, Kinetic measurements of
hydrocarbon conversion reactions on model metal surfaces. Phys. Chem. Chem. Phys. 9, 3830–
3852 (2007). doi:10.1039/B702652H
35. N. Katada, H. Igi, J.-H. Kim, M. Niwa, Determination of the acidic properties of zeolite by
theoretical analysis of temperature-programmed desorption of ammonia based on adsorption
equilibrium. J. Phys. Chem. B. 101, 5969–5977 (1997). doi:10.1021/jp9639152
36. S. Narayanan, A. Sultana, Q.T. Le, A. Auroux, A comparative and multitechnical approach to
the acid character of templated and non-templated ZSM-5 zeolites. Appl. Catal A-Gen. 168,
373–384 (1998). doi:10.1016/S0926-860X(97)00368-2
37. V. Rac, V. Rakić, S. Gajinov, V. Dondur, A. Auroux, Room-temperature interaction of n-hexane
with ZSM-5 zeolites. Microcalorimetric and temperature-programmed desorption studies. J.
Therm. Anal. Cal. 84, 239–245 (2006). doi:10.1007/s10973-005-7164-z
38. A. Auroux, R. Monaci, E. Rombi, V. Solinas, A. Sorrentino, E. Santacesaria, Acid sites inves-
tigation of simple and mixed oxides by TPD and microcalorimetric techniques. Thermochim.
Acta. 379, 227–231 (2001). doi:10.1016/S0040-6031(01)00620-7
39. A. Tanksale, J.N. Beltramini, J.A. Dumesic, G.Q. Lu, Effect of Pt and Pd promoter on Ni
supported catalysts–A TPR/TPO/TPD and microcalorimetry study. J. Catal. 258, 366–377
(2008). doi:10.1016/j.jcat.2008.06.024
40. K. Kissinger, Reaction kinetics in differential thermal analysis. Anal. Chem. 29, 1702–1706
(1957). doi:10.1021/ac60131a045
41. R.J. Gorte, Design parameters for temperature programmed desorption from porous catalysts.
J. Catal. 75, 164–174 (1982). doi:10.1016/0021-9517(82)90131-2
42. R.A. Demmin, R.J. Gorte, Design parameters for temperature-programmed desorption from a
packed bed. J. Catal. 90, 32–39 (1984). doi:10.1016/0021-9517(84)90081-2
43. R.J. Gorte, Temperature-programmed desorption for the characterization of oxide catalysts.
Catal. Today 28, 405–414 (1996). doi:10.1016/S0920-5861(96)00249-0
174 V. Rakić and L. Damjanović

44. D.A.M. Monti, A. Baiker, Temperature-programmed reduction. Parametric sensitivity and esti-
mation of kinetic parameters. J. Catal. 83, 323–335 (1983). doi:10.1016/0021-9517(83)90058-
1
45. J. Shen, A. Auroux, The determination of acidity in fluid cracking catalysts (FCCs) from
adsorption microcalorimetry of probe molecules, in Proceedings of International Symposium
on Fluid Catalytic Cracking VI, Preparation and Characterization of Catalysts, New York,
7–11 September 2003, eds. by M. Occelli, p. 35. Studies in Surface Science and Catalysis, vol.
149, pp. 35–70 (2004). doi:10.1016/S0167-2991(04)80756-0
46. A. Corma, From microporous to mesoporous molecular sieve materials and their use in catal-
ysis. Chem. Rev. 97, 2373–2420 (1997). doi:10.1021/cr960406n
47. A. Auroux, Acidity and basicity, in Molecular Sieves- Science and Technology, vol. 6. (
Springer, Berlin, 2008), p. 45
48. Y. Mitani, K. Tsutsumi, H. Takahashi, Direct measurement of the interaction energy between
solids and gases. XI. Calorimetric measurements of acidities of aluminium deficient H-Y
zeolites. Bull. Chem. Soc. Japan. 56, 1921–1923 (1983). doi:10.1246/bcsj.56.1921
49. A. Auroux, Y. Ben Taarit, Calorimetric investigation of the effect of dealumination on the
acidity of zeolites. Thermochim. Acta. 122, 63–70 (1987). doi:10.1016/0040-6031(87)80105-
3
50. A. Auroux, P.C. Gravelle, J.C. Védrine, M. Rekas, in Proceedings of the 5th International
Conference on Zeolite, Naples, 2–6 June 1980, ed. by L.V.C. Rees, L.V. Heyden, p. 433
51. H.G. Karge, L.C. Jozefowicz, A comparative study of the acidity of various zeolites using
the differential heats of ammonia adsorption as measured by high-vacuum microcalorimetry,
in Proceedings of the 10th International Conference on Zeolites and Related Microporous
Materials: State of the Art 1994, Garmisch-Partenkirchen, Elsevier, Amsterdam, 17–22 July
1994, eds. by J. Weitkamp, H.G. Karge, H. Pfeifer, Hö.W. lderich, p. 685. Studies in Surface
Science and Catalysis, vol. 84, pp. 685–692 (1994). doi:10.1016/S0167-2991(08)64174-9
52. M. Calatayud, A. Markovits, C. Minot, Electron-count control on adsorption upon reducible
and irreducible clean metal-oxide surfaces. Catal. Today 89, 269–278 (2004). doi:10.1016/j.
cattod.2003.12.015
53. Lj. Damjanović, A. Auroux, Heterogeneous catalysis on solids, in The Handbook of Thermal
Analysis and Calorimetry. Further advances, techniques and applications, vol. 5, ed. by M.
Brown, P. Gallagher (Elsevier, Amsterdam, 2008)
54. M. Niwa, N. Katada, Measurements of acidic property of zeolites by temperature programmed
desorption of ammonia. Catal. Surv. Japan. 1, 215–226 (1997). doi:10.1023/A:1019033115091
Chapter 5
Temperature Programmed
Reduction/Oxidation (TPR/TPO)
Methods

Antonella Gervasini

Abstract The redox properties of the metal oxides impart them peculiar catalytic
activity which is exploited in reactions of oxidation and reduction of high applicative
importance. It is possible to measure the extent of oxidation/reduction of given metal
oxide by thermal methods which are become very popular: TPR and TPO analyses.
By successive experiments of reduction and oxidation (TPR-TPO cycles) it is possi-
ble to control the reversible redox ability of a given oxide in view of its use as catalyst.
The two methods are here presented with explanation on some possibility of exploita-
tion of kinetic study to derive quantitative information on the reduction/oxidation of
the oxide. Examples of selected metal oxides with well-established redox properties
which have been used in catalytic processes are shown.

5.1 Redox Properties of Metal Oxides and Catalytic

Implications

The important properties of metal oxides that are of certain interest in different fields
of applied science and technology are connected with magnetic, electrical, dielectric,
optical, acid-base, and redox behavior. In particular, transition metal oxides with
redox properties are of interest for applications in catalysis [1]. Metal oxides and in
particular transition metal oxides can possess more than one stable oxidation state
making them possible to catalyze reactions which necessitate electron exchanges
between reactant(s) and surface active sites (e.g., oxidation, dehydrogenation, etc.).
Conversely, other kinds of metal oxides are known to be almost complete irreducible.
Oxides of copper, nickel, cobalt, molybdenum, are only few examples of reducible

A. Gervasini (B)
Dipartimento di Chimica,
Università degli Studi di Milano,
Via Camillo Golgi, Milan, Italy
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 175

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_5,
© Springer-Verlag Berlin Heidelberg 2013
176 A. Gervasini

oxides and oxides of gallium or niobium are examples of quasi-irreducible oxides

(for Ga2 O3 , any reduction is not observed up to 1000 ◦ C and for Nb2 O5 , less than
50 % of Nb(V) can be reduced to Nb(IV) at temperature higher than 900 ◦ C).
The reducing properties of metal oxides can be quantitatively established evalu-
ating the reaction between the metal oxide (MO) and hydrogen to form metal (M)
and water vapor:
MO(s) + H2(g) → M(s) + H2 O(g) (5.1)

The standard free energy change for the reduction (G ◦ ) is negative for a number
of oxides, thus indicating that for these oxides the reactions are thermodynamically
feasible (Fig. 5.1). However, it may still possible for the reduction to proceed even
when G ◦ is moderately positive. The following equation:

G = G◦ + RT ln (PH2O /PH2 ) (5.2)

where PH2O and PH2 denote partial pressures of water and hydrogen, respectively,
has to be considered. If the experimental method of performing the reductions is that
the water vapor is constantly swept from the reaction zone as it is formed, then PH2O
is lowered and it is possible that the term RT ln (PH2O /PH2 ) could be sufficiently
negative to nullify the positive G◦ value. This is likely to occur at high temperatures
where it is possible to observe reduction phenomena for the oxides of vanadium, tin,
chromium, which have positive G◦ values in the range 40–100 kJ/mol (Fig. 5.1).
During some reactions catalyzed by metal oxides, they undergo reduction and re-
oxidation simultaneously by loss and gain of surface lattice oxygen to and from the
gas phase. This phenomenon is called redox catalysis. The redox property as well as
the acidic and basic feature is the most important properties of metal oxide catalysts.
Some simple metal oxides, like V2 O5 , CoO2 , NiO, MnO2 , CeO2 , MgO and some
mixed metal oxides have redox properties and they found application as they are or
supported on ceramic carriers as oxidation catalysts.
The use of oxides and mixed metal oxides as heterogeneous catalysts in selective
oxidation reactions, in particular, is of great industrial importance. In selective oxi-
dation reactions, organic feeds are converted to useful products by oxygen insertion
without any change of the number of carbons. The major processes include allylic
oxidation (i.e., selective oxidation of olefins at the allylic position) to give aldehy-
des, nitriles, and acids; aromatic oxidation to acids and anhydrides, epoxidation to
olefins, methanol oxidation to formaldehyde, etc. [3]. The desired selective oxida-
tion reactions are, of course, thermodynamically favorable and the oxide catalysts are
necessary to overcome kinetic limitations. Moreover, they have to select the desired
oxidized product limiting the formation of the more thermodynamically favored deep
oxidation products, like CO2 , H2 O, and HCN (Table 5.1 [3]).
The research and study of selective oxidation oxide catalysts with developed
redox properties began more than five decades ago with the concept that the lattice
oxygen of a reducible metal oxide could serve as a more versatile and useful oxidizing
agent for hydrocarbons than would gaseous molecular oxygen. The role of molecular
oxygen is only to replenish lattice oxygen vacancies and therefore an oxidation-
5 Temperature Programmed Reduction/Oxidation (TPR/TPO) Methods 177

Fig. 5.1 Standard free energy change as a function of temperature (from Ref. [2])

Table 5.1 Thermodynamics of several selective and total oxidation reactions

Reactions G ◦427◦ C (kcal/mol)
(A) C3 H6 + O2 → CH2 = CHCHO + H2 O − 80.82
(B) C3 H6 + 23 O2 → CH2 = CHCOOH + H2 O −131.42
(C) C3 H6 + 3O2 → 3CO + 3H2 O −304.95
(D) C3 H6 + 29 O2 → 3CO2 + 3H2 O −463.86
(E) C3 H6 + NH3 + 23 O2 → CH2 = CHCN + 3H2 O −136.09
(F) C3 H6 + 23 NH3 + 23 O2 → 23 CH3 CN + 3H2 O −142.31
(G) C3 H6 + 3NH3 + 3O2 → 3HCN + 6H2 O −273.48
178 A. Gervasini

Fig. 5.2 Scheme of a catalytic selective oxidation-reduction cycle (from Ref. [3])

reductant cycle is recognized to hold in the selective catalytic oxidations (Fig. 5.2).
Confirmation of the involvement of lattice oxygen in selective oxidation reaction (i.e.,
propylene oxidation) comes from experiments with 18 O2 as a vapor phase oxidant.
From a mechanistic standpoint, reoxidation of the catalyst involves two processes:
adsorption and activation of dioxygen and its incorporation into the solid vacancies
(i.e., transformation of O2 to O2− ).
This step requires electron transfer from the lattice to activate and dissociate the
dioxygen before incorporation into the vacancies. When the oxygen vacancies are
located near the surface, the reoxidation is very fast and activation energy for this
surface vacancy reoxidation of the order of 5–10 kJ/mol. Reoxidation of subsurface
vacancies is much slower and it is limited by the ability of the catalyst to transport
the oxygen from the surface O2 -chemisorption sites to the subsurface vacancies. For
example, γ -bismuth molybdate with its layer structure results in low-energy pathway
for diffusion of the oxygen anions and reoxidation proceeds with an activation energy
of only 35–38 kJ/mol. In more closed packed structures, like α-bismuth molybdate,
rapid diffusion cannot occur and the reoxidation rate is diffusion limited with activa-
tion energy as high as 105–110 kJ/mol. Moreover, the presence of a redox couple in
a catalyst can create a lower energy pathway for diffusion by promoting electron and
oxygen transfer between the surface and the bulk. Incorporation of iron (Fe3+ /Fe2+
redox couple) into the α-bismuth molybdate structure, single Bi3 FeMo2 O12 phase is
formed, decreases the activation energy for the reoxidation process to 29–35 kJ/mol.
The real breakthrough in the appreciation of the idea of lattice oxygen depletion
and subsequent replenishment governing the selective oxidation reactions took place
after the appearance of a series of papers by Mars and Van Krevelen (MvK), in 1954
[4] describing some applications of their idea in the kinetics (Fig. 5.3). However,
5 Temperature Programmed Reduction/Oxidation (TPR/TPO) Methods 179

Fig. 5.3 Conversion of propylene into acroleine on bismuth molybdate following the Mars and
Van Krevelen mechanism (from Ref. [4])

all oxidations are redox reactions, even when the prevailing mechanism can not be
called a MVK mechanism
A property of key importance for an efficient selective oxidation catalyst is its
tendency to donate the structural oxygens. If reduction of the catalyst is too facile,
it may be well active but it ceases to be selective. Although the reoxidation rate is
usually not rate determining in the overall redox cycle, the catalyst ability to replenish
its reservoir of lattice oxygen is of fundamental importance in order to sustain the
catalyzed reaction. While reoxidation at higher temperatures is generally fast for
all the selective oxidation catalysts, it becomes more difficult at low temperatures,
where selectivities to the usual products are highest.
Examples of reactions of industrial and environmental importance that run accord-
ing MvK mechanism are alcohol selective oxidations, VOC total oxidations, selec-
tive catalytic reductions of NOx by ammonia or hydrocarbons, and more recently
photocatalytic oxidation of hydrocarbons [5] etc. Revisitation of MvK mechanism
for various reactions appeared in Ref. [6] For each one of these reactions, a metal
oxide-based catalyst with well-developed redox properties has been optimized.
The redox property of any metal oxide catalyst can be characterized by using the
techniques of temperature-programmed reduction/temperature programmed oxida-
tion (TPR/TPO). TPR is a relatively new technique used for the chemical characteri-
zation of solid materials. Since the first publications in this area (1975–1985), many
papers have appeared in the literature showing TPR applicability in a variety of scien-
tific fields, most particularly in characterization of catalysts. In 1982 the first review
on the subject of TPR appeared in the literature [2]. TPR technique is based on the
reducibility of species in/on solids by a gas, in general by hydrogen, at the same time
that the temperature of the system is changed in a predetermined way (in general, lin-
ear temperature increase). The decrease in reducing gas concentration in the effluent
gas with respect to its initial concentration allows monitoring the reduction progress.
The obtained results enable obtaining information not only of a purely analytical
nature but also, and more importantly, about the condition of species present in/on
solids. The technique is relatively simple in concept and application and as a result,
TPR and related techniques are becoming commonly used among catalysis scientists.
180 A. Gervasini

More recently, combination of TPR technique with the temperature programmed

oxidation technique (TPO) is imposing in the catalyst characterization studies [7]. In
TPO, the catalyst is in the reduced form and is submitted to a programmed tempera-
ture increase, while an oxidizing mixture of gas, conventionally containing oxygen,
is flowed over the sample. TPO is used in applications such as the study of the kinet-
ics of coking, evaluation of carbon burn off, determination of the different forms of
carbonaceous deposits present on a catalyst surface after a given reaction involving
organic compounds, etc. Coupled TPR/TPO technique offers many advantages. The
two techniques can provide useful information in the study of the reactivity and
redox behavior of catalysts. The combination of the two reactions of reduction and
oxidation represents a real titration of any reducing/oxidant phase by quantification
of hydrogen/oxygen consumption.

5.2 Temperature-Programmed Reduction/Oxidation

Technique

The developed technique foresees an increase of the sample temperature at a uniform

rate (Ti = T0 + β t, where β is the rate of temperature increase) measuring and
recording the rate at which the sample (or the active component in the sample)
reduces/oxidizes under a reducing/oxidizing atmosphere of known concentration.
The following reactions hold:
TPR: Mx Oy(s) + H2(g) → M(s) + H2 O
TPO: xM(s) + y/2O2 → Mx Oy(s)
A typical TPR/TPO apparatus can be assembled following some general princi-
ples. The apparatus essentially consists of three parts (i) the gas line for pre-treatments
and analysis; (ii) the reactor electrically controlled; (iii) the detector for quantitative
evaluation of gas-consumption. At present, all the experimental apparatus used oper-
ate at ambient (in flowing gas) or sub-ambient (under vacuum) pressure.
Typically, TCD detectors are mounted (even if many other different types are
reported in the literature: detection by pressure changes of the reacting gas, detection
by weight change of the solid, etc.), they produced on output signal that is proportional
to the concentration of hydrogen or oxygen in the carrier gas. A TC-detector works
by measuring the difference in thermal conductivity between the pure carrier gas and
the mixture of carrier and reactant gas. For this reason, it is most sensitive when the
thermal conductivities of the two gases in mixture are very different. This is why
argon or nitrogen are preferentially used as carrier gases in TPR and helium in TPO
experiments (thermal conductivities of Ar, N2 , and He of 45.46×10−6 , 65.71×10−6 ,
and 376.07×10−6 cal/(cm · s · K), respectively and thermal conductivities of H2
and O2 of 471.11×10−6 and 68.19×10−6 cal/(cms K), respectively). Moreover, the
change in thermal conductivity is proportional to the mole fraction of reactant gas in
the mixture only at low concentrations of reactant gas. To maintain a high detector
5 Temperature Programmed Reduction/Oxidation (TPR/TPO) Methods 181

sensitivity the concentration of reactant gas in the mixture has to be limited in the
1–10 % range.
When setting a TPR/TPO apparatus (Fig. 5.4), some practical considerations have
to be followed. Particular care has to be taken concerning the determination of
time/temperature of the reactive events. If the period between the reduction/oxidation
of the sample and the detection of hydrogen/oxygen concentration is significantly
long, then the measured temperatures (Tmeas ) of the reactive events require correction
according to:
Tcorr = Tmeas − (β ·V/F) (5.3)

where Tcorr is the corrected temperature, V is the volume between the reactor and
detector, and F is the used flow rate. If such corrections are large, the precision of the
temperature measurement is reduced and it is advised maintaining high flow rates
during experiment. The flow rate must also be compatible with the geometry of the
reactor such that the gas flow maintains uniform velocity distribution conditions.
Moreover, the flow rate has also to be in relation to the amount of reducible/oxidized
sample, as detailed here below.
During the TPR/TPO experiments, several products as H2 O, CO, or CO2 can be
formed. It is important to remove all the undesired gas molecules that can interfere
in the signal output. Besides a correct pre-treatment procedure of the sample, use of
suitable traps to stop secondary products are necessary and they have to be present
in the apparatus configuration.

5.2.1 General Operative Procedure

The procedure to collect TPR/TPO data is also comprehensive of the sample pre-
treatment. Several types of procedures can be chosen in relation with the sample
nature and type of information required. Generally, before starting a TPR analysis,
the sample should be in its oxidized form. In this case, the pretreatment consists in
oxidizing the sample in flowing air or pure oxygen and then purging the product
formed (like water and carbon residues) by flowing an inert gas. The two procedures
have to be carried out at given temperature to assure the effectiveness of the actions.
In the case of TPO analysis, the sample has to be previously reduced; in this case
too, the reducing procedure has to be carried out at given temperature.
At the beginning of a TPR or TPO experiment, reducing/oxidizing gas is made to
flow over a fixed amount of solid at a temperature low enough to prevent reaction.
The starting temperature of the experiment has then to be below the expected reduc-
tion/oxidation temperature of the phase under study. In this condition, a perfect flat
baseline is observed because any hydrogen/oxygen is consumed from the sample (the
two branches of TCD are at same concentration of active gas component). Next, the
temperature of the sample is increased at a constant rate (β, ◦ C/min) and, as reduc-
tion/oxidation begins, hydrogen/oxygen is consumed from the gas mixture (con-
ventionally, H2 /Ar and O2 /He mixtures), which is detected by the detector (TCD).
182 A. Gervasini

Fig. 5.4 Scheme of a typical TPR/TPO apparatus. Items: P pressure controller; V2 shutoff valve;
F purification stage; M mass flow controller; V metering valve; V3, V5, V7 3-way, 5-way, and 7-way
switching valves, respectively; V4 crossover valve; F temperature controlled furnace; R reactor; PFL
pretreatment flow line; RFLK and MFL reference and measure flow line of TCD

The TCD signal rises proportionally to the rate of hydrogen/oxygen consumption.

When reduction/oxidation ceases, no more hydrogen/oxygen is consumed and the
TCD signal returns to its baseline. A peak is obtained (Fig. 5.5) which maximum
position represents the maximum rate of hydrogen/oxygen consumed (i.e., temper-
ature of maximum rate of reduction/oxidation, Tmax ). Several peaks (Tmax,i ) may
be detected over the course of the temperature ramp if different reduction/oxidation
steps may occur and each one requires given thermal energy value. The amplitude
of each peak is proportional to the reaction rate. On the basis of suitable calibration
tests, the amount of consumed hydrogen/oxygen can be obtained from the graphical
integration of each peak. Calibration tests are readily achieved by injecting into the
5 Temperature Programmed Reduction/Oxidation (TPR/TPO) Methods 183

Fig. 5.5 Example of typical

TPR and TPO profiles

TPR apparatus known amounts of hydrogen, for example, usually several different
amounts. The linear calibration curve obtained (detector signal vs. moles of H2 ) is
used for the quantitative calculation of the extent of reduction (or oxidation) of the
actual sample under study.

5.2.2 Analytical Parameters

The choice of the analytical parameters for a TPR/TPO experiment, in particular

sample volume/mass, temperature increasing rate, and flow concentration and flow
rate, is fundamental to obtain significant reaction profiles.
Theory predicts that the Tmax values should be independent of the mass of the
sample [8, 9]. Minimal changes in Tmax were experimentally observed from many
authors. The volume (and mass) of the sample is best kept to a minimum as this
diminishes problems due to back pressure, temperature differences between sensor
and sample, and no uniform gas concentration within the solid. The smaller amount
of sample used, it is easier to handle the whole analysis, in particular, less water
is formed thus avoiding problems in the water removal system. The main effect
of changing the sample mass was that resolution of two separate chemical events
obtained with, e.g., 50-mg sample was lost when the sample mass was increased
to, e.g., 400 mg with higher value of Tmax for the composite resultant peak. This
behavior is due to the temperature difference between the temperature sensor and
the sample that becomes marked for large mass of sample. Therefore, care must be
taken in comparing results when sample sizes are markedly different. For example,
the two well known rate processes for the reduction of Cu2+ to Cu0 through Cu+
intermediate, are well distinguished if TPR is performed with low sample mass and
low hydrogen concentration [9, 10].
Sample particle size has to be controlled. Particles that are too small can result
in the creation of back pressure and consequent problems in maintaining constant
flow along the sample, while too large particles can result in intraparticle diffu-
sion problems, causing distorted peaks. Mass transport processes, both inter- and
intraparticles, are characterized by low-activation energies and they can thus alter
184 A. Gervasini

Fig. 5.6 Sensitivity of the

temperature of the maximum
reduction rate for the reduc-
tion process of NiO with β
variation

Fig. 5.7 Sensitivity of the

temperature of the maximum
reduction rate for the reduction
process of NiO with H2
concentration

the shape of reduction profiles. It is common practice to experimentally check for

mass transport limitations by looking for changes in TPR/TPO profiles when particle
size is varied. Heat transfer limitations can be also considered to be present during
reduction/oxidation processes. To avoid or highly limited them, sample dilution with
quartz or inert material is commonly performed, this procedure does not alter the
TPR/TPO profile.
Concerning the rate of temperature increase, there are advantages in having the
heating rate as fast as all the other factors allow. This results in the obtainment of better
defined-peaks. Moreover, variation of heating rate (β) leads to variation of Tmax , this
behavior is most commonly used to obtain activation energies for the reactive process.
When the change in Tmax is due to kinetic parameters, other parameters remaining
constant, the sensitivity of Tmax to changes in heating rate is established. For example,
systematic study of hydrogen reduction of NiO (simple one-step reduction, from Ni2+
to Ni(0), first-order in hydrogen and NiO), showed that an increase in heating rate
from 5 to 20 ◦ C/min produced a shift in Tmax of 30 ◦ C (Fig. 5.6). Exploiting the
observed Tmax with β, activation energy of about 134 kJ/mol was calculated for the
NiO reduction process [8].
5 Temperature Programmed Reduction/Oxidation (TPR/TPO) Methods 185

Concentration of hydrogen/oxygen in TPR/TPO experiment affects the sensitivity

of the detector and the Tmax values vary. For example, an increase on H2 concen-
tration from 3 % (1.23 μmol/cm 3 ) to 15 % (6.15 μmol/cm 3 ) produced a decrease
in Tmax from 330 to 290 ◦ C for the reduction process of NiO [8] (Fig. 5.7). This
indicates that the sensitivity is inversely proportional to the H2 concentration in the
reactant mixture.
The variation of the total flow rate influences the results obtained in a TPR/TPO
experiment, too. In the case of hydrogen reduction of Cu-exchanged zeolite, an
increase in total flow rate from 10 to 20 cm3 /min (with 4 % H2 concentration)
lowered the value of Tmax by 15−20 ◦ C [9]. Recalling the flow reactor theory, it is
expected that an increase in total flow rate for a reactant consumed by a first-order
process results in a lowering of the degree of conversion and then an increase of
reactant concentration in the reactor. The observed increase of H2 concentration in
the reactor and decrease of Tmax value of the reduction event with total flow rate are
in agreement with theory.

5.2.3 Selection of Operating Parameters

To obtain meaningful TPR/TPO profiles, certain restrictions on the choice of com-

binations of operating parameters are self-evident. Of particular importance in the
case of a TPR experiment, is that the hydrogen feed rate should be equal to the max-
imum reduction rate, otherwise distortion of the reduction profile will occur. Certain
combinations of total flow rate, reactant concentration, sample volume (mass), and
heating rate are allowed while other ones are not.
The possibility to carry out a TPR experiment (but the same holds for TPO exper-
iment) with quantitative evaluation of results is based on the simplifying assumption
that the mean hydrogen concentration between the reactor inlet and outlet is the
driving force for the reduction; that is, the mean hydrogen concentration into the
reactor. This holds only if low fractional conversion of hydrogen occurs all along the
experiment. On the other hand, the difference of hydrogen concentration between
reactor inlet and outlet has to be well detectable, that is, the change in hydrogen con-
centration has to be large compared to the statistical noise of the detector baseline.
Two criteria have to be fulfilled to meet the two requirements above exposed: (i) the
amount of hydrogen consumed at the peak maximum should not exceed 2/3 of the
hydrogen fed to the reactor and (ii) the minimum conversion at the peak maximum
should be 10 %. The two criteria can be met with sets of operating variables. A
characteristic number, K (usually expressed in s), has been defined [8] to facilitate
the selection of appropriate operating variables:

K = So /(V∗ Co ) (5.4)

where V* is the total flow rate of the reducing gas (cm3 /s); So , the initial amount of
reducible species in the sample (μmol); Co is the initial concentration of the reducing
186 A. Gervasini

gas (μmol/cm3 ). The values of K have to be comprised between 55 s and 140 s, with
possibility to match So , V*, and Co parameters to obtain the appropriate K value.
For values of K below 55 s, the sensitivity of the analysis becomes too low, while
for values exceeding 140 s, the amount of reducing gas consumed is too large, so
violating the assumption of a linear concentration profile. Concerning the physical
meaning of the K parameter, it represents the sensitivity parameter. Once optimized
K, a second characteristic number can be determined, P (usually expressed in ◦ C),
that is defined as:
P =K·β (5.5)

P represents the shape and resolution parameter and it has to be comprised between
20 and 50 ◦ C. With the K and P numbers in the correct range, accurate and reliable
results from the TPR/TPO analysis and above all comparable data are obtained. The
sensitivity of the analysis becomes lower for small heating rates with an inverse
proportionality to the hydrogen concentration in the feed. A typical example is in the
TPR analysis of CuO: changing the rate of heating and of total flow, two different
reaction profiles result. It is possible or not to distinguish the two steps of the reduction
process identified by two reducing peaks, Cu(II) → Cu(I) → Cu(0), or to obtain
one only peak comprehensive of the total hydrogen consumption involved in the two
steps. In the second case, the advantage is to calculate more easily the total amount of
the consumed reactive gas. In general, when the sample contains only one component
it is useful to perform the analysis with a low temperature rate (b) to observe the
mechanism of the reaction process.

5.3 Kinetics and Reduction Mechanisms

The aim of the analysis of the TPR/TPO data is to derive kinetic parameters relating to
the reduction/oxidation process. It is common to observe the reduced/oxidized frac-
tion (α) as a function of time for various temperatures and pressures/concentrations
of reducing/oxidant gas. The reduction/oxidation of a sample with redox property
or of part of it (in the case of supported phase system) is a bulk phenomenon and its
degree of reduction/oxidation is interpreted in terms of mechanism by which the reac-
tion occurs. Nucleation model or contracting sphere model are the most successful
utilized kinetic models [2].
In synthesis, at constant temperature, the rate of the reaction: MOx + H2 →
reduced products, may be expressed as:

rate = −d[MOx ]/dt or − d[H2 ]/dt = k[H2 ]P [MOx ]Q (5.6)

with P and Q, the reaction orders for the MOx and H2 species, respectively.
If the rate constant k is expressed with its dependence on temperature by the
classical Arrhenius equation (with A, the pre-exponential factor and Ea , the activation
energy) and the linear dependence of temperature on time is introduced by the linear
5 Temperature Programmed Reduction/Oxidation (TPR/TPO) Methods 187

heating rate (β = dT/dt), we can write the following expression:

− β d[MOx ]/dT = A exp(−Ea /RT)[H2 ]P [MOx ]Q (5.7)

If the H2 consumed is little (in accord with the criteria above cited in Sect. 5.2.3, the
reaction rate can be considered independent of H2 concentration and Eq. (5.7) can
be rewritten as:

− β d[MOx ]/dT = A exp(−Ea /RT)[MOx ]Q (5.8)

we can now introduce the reduced fraction at given time t (αt ), instead of the residual
concentration of sample to be reduced, and being: β d(1 − α)/dT = β d(α)/dT, we
can write:
β d(α)/dT = A exp(−Ea /RT)(1 − α)Q (5.9)

integration of Eq. (5.9) gives:

α T
d(α)/(1 − α) = A/ β
Q
exp(−Ea /RT)dT (5.10)
0 0

or
α t
d(α)/(1 − α) = A
Q
exp(−Ea /R(To + β t)dt (5.11)
0 0

with To representing the starting temperature of the linear heating rate.

The problem is now to define the correct expression for the variation of α versus
t to solve Eq. (5.11) that is to describe the reduction process by a suitable model.
For the derivation of the kinetic parameters of reaction, it is more generally use-
ful to utilize equations able to directly interpret the raw TPR data. Most experi-
mental TPR systems monitor the rate of change of reducing gas concentration as a
function of temperature, producing reduction profiles with peaks corresponding to
maxima for the reaction rates. Methods originally developed for other temperature-
programmed techniques are employed, like thermogravimetry analysis [11–16]. In
these approaches, multiple reaction analyses have to be carried out with several
different constant heating rates (by varying β), from which the shift of peak maxi-
mum position (Tmax ) with β is exploited. The proposed equations converge on linear
expressions like that here below recalled, from which both Ea and A can be obtained
from the slope and intercept of the relevant line:

ln(Tmax2 [H2 ]/ β) = Ea /R(1/Tmax ) + ln(Ea /RA) (5.12)

normally, first order kinetics with respect to the reducing sample and hydrogen is
assumed and mean hydrogen concentration at the temperature of the maximum reac-
tion rate is considered.
188 A. Gervasini

5.3.1 Nucleation Model

When a reducing sample (e.g., oxide) and hydrogen come into contact, the reaction
starts and after some time, t1 (induction time), the first nuclei of the reduced product
form. Reduction removes oxygen ions from the lattice and when the concentration
of vacancies reached a certain critical value they are annihilated by rearrangement of
the lattice with formation of metal nuclei. Oxygen ions may be removed by inward
diffusion of hydrogen to the metal/metal-oxide interface or outwards diffusion of
oxygen ions from the metal oxide to the metal gas interface. The reaction interface
increases more and more rapidly thanks to two processes: (i) the growth of the nuclei
already formed and (ii) the appearance of new ones. At a given time of reduction, the
metal nuclei have grown at the surface of the oxide grains to such an extent that they
begin to make contact with each other. Starting from this moment, a decrease of the
reaction interface is observed because of the overlapping of the metal nuclei and the
steady consumption of the oxide grains. These processes continue until the complete
reduction of the sample; during the final stage in which the reaction interface is
shrinking as the reduced layer grows, the process is equivalent to the contraction
sphere model. The nucleation mechanism results in the S-shaped α against t plot and
in the maximum in the dα/dt against α, as illustrated in Fig. 5.8 (left). The same plots
are observed for autocatalytic reductions; for example, hydrogen can be dissociated
and activated by the reducing oxide; this phenomenon is observed on oxide promoted
by metal dopants (doped-CuO and NiO samples).
The mathematical treatment that describes the nucleation mechanism of reduction
can be found in Ref. [2]. The final integral equation rate, Eq. 5.13), represents the
rising part of the α-vs-t curve before the inflection point (Fig. 5.8 left) and shows that
the rate of nucleation is proportional to (time)q :

1C1 C2
α= p+q+1
(t − t1 )q+p+1 (5.13)
Vfinal

with C1 and C2 , proportionality constants; p, the dimension of nuclei; V, the final

volume of nuclei; and q, an integer number representing the order of nucleation rate.

5.3.2 Contracting Sphere Model

In many cases, the reaction interface between the oxidized and freshly reduced sample
decreases continuously from the beginning of the reaction when it has its maximum
value. This can be interpreted in terms of a very rapid nucleation resulting in the total
coverage of the reducing grain with a thin layer of the reduced product in the first
instant of the reaction. The reaction interface then decreases as the substrate grain is
continuously consumed in the course of the reaction. Figure 5.8 (right) shows the con-
tracting sphere model with its decreasing reaction interface resulting in continuously
decreasing rate (dα/dt vs. t).
5 Temperature Programmed Reduction/Oxidation (TPR/TPO) Methods 189

Fig. 5.8 Reduction process by the nucleation mechanism (left) and the contracting sphere mecha-
nism (right)

The distinction between the nucleation and contracting sphere model is somewhat
artificial; in fact the contracting sphere model starts with very rapid nucleation and
the nucleation mechanism finishes by an essentially contracting sphere model. The
real distinction is between a reaction interface, and reaction rate, that is increasing
in the early stages of the reactive process (nucleation model) and a reaction interface
that is contracting throughout the reaction (contracting sphere model).
The mathematical treatment that describes the contracting sphere mechanism of
reduction can be found in Ref. [2]. The rate of the reaction is controlled by both
diffusion from the gas stream to the sample particle and through the reduced layer
freshly formed and by chemical reaction at the interface. The final integral equation
rate, Eq. (5.14) gives the degree of reduction α as a function of time. The rate constant,
kp , may be higher or lower than the constant rate of hydrogen diffusion, kd , with the
chemical or hydrogen diffusion stage controlling the reaction kinetics, respectively.

kp r0 kp 1 α 2
(c0 − Ceq )t = [1 − (1 − α)1/3 ] + [ − − (1 − α) 3 ] (5.14)
r0 d0 kd 2 3
190 A. Gervasini

with Co and Ceq , the hydrogen concentration when reaction rate is zero, ro , and when
it is at equilibrium, respectively; and do , the sample density.

5.4 Examples

Typical TPR and TPO analyses of catalysts (bulk, supported, doped, reducing/oxidi-
zing phases) are aimed at evaluating the position of Tmax peaks and the amount
of consumed reducing/oxidizing gas which provides information on (i) easiness to
reduction/oxidation; (ii) degree of reduction/oxidation attained; (iii) average oxida-
tion state of given (metal) phases after a given reducing/oxidizing treatment; (iv)
dispersion and aggregation state of supported (reducing/oxidizing) phases; v) ability
to complete a redox cycle.
In the preparation of supported metal catalysts, salt and molecular complex pre-
cursors of the metal active components are used; procedures of air-calcination or
H2 -reduction are then needed for activating the catalyst for the reaction. Platinum
is one among the most important metals widely used as catalytic component for
a host of chemical processes. For the preparation of Pt-containing catalysts, the
highest Pt-dispersion over suitable supports is highly desirable to avoid the con-
sumption of the costly metal. Therefore, a wide range of platinum compounds have
been studied to determine their reduction conditions to obtain the Pt-phase, e.g.,
PtCl4 , H2 PtCl6 , Pt(NH3 )2 (NO2 )2 , Pt(NH3 )4 (OH)2 , [Pt(NH3 )4 ]Cl2 , H2 [Pt(OH)6 ],
ecc. Some Pt-compounds require a four-electron process for reduction (PtCl4 , H2
PtCl6 , and H2 [Pt(OH)6 ]) while the reduction of some others require a two-electron
process. The TPR profiles of some dried Pt-samples are shown in Fig. 5.9. It is inter-
esting to note that the Cl-containing salts/complexes are significantly more difficult
to reduce than the OH- containing compounds (e.g., in Fig. 5.9 compares the TPR
profiles A with B and C; and the D with the E profiles).
In the catalyst preparation, not only the choice of the active phase precursor is cru-
cial, the method of catalyst preparation is decisive, too, for obtaining good dispersion
of the active phase. Active phase can be deposited on supports by impregnation, ion-
exchange, adsorption, etc. Once selected the nature of support and active phase, the
observed differences in dispersion should only be due to the method of preparation.
Dispersed iron oxide catalysts (FeOx ) have received much attention because their
potentiality for many applications in environmental catalysis (N2 O decomposition
and reduction) and in fine chemical industry (Friedel-Crafts, isomerisations, etc.).
For most applications, high dispersion of the metal centres is desirable to enhance
the activity-selectivity pattern of the catalysts.
Iron oxide is a well reducible phase, the comparison of reduction profiles of sup-
ported Fe-catalysts may give information on the dispersion state of iron. In Fig. 5.10,
two TPR spectra of two dispersed FeOx catalysts on zirconia prepared by con-
ventional impregnation and adsorption methods are shown [17]. Reduction of pure
hematite (Fe2 O3 ) by hydrogen is a complex event that can proceed in different steps
5 Temperature Programmed Reduction/Oxidation (TPR/TPO) Methods 191

Fig. 5.9 TPR profiles for

some Pt salts used in catalyst
preparation (from Ref. [2])

via intermediate oxides (i.e., magnetite, Fe3 O4 , and wüstite, FeOx ). The possible
reduction reactions, involving hematite, are written here below.
(i) Fe2 O3 + 3H2 → 2Fe + 3H2 O
(ii) 3Fe2 O3 + H2 → 2Fe3 O4 + H2 O
(iii) 3Fe2 O3 + 3H2 → 6FeO + 3H2 O
(iv) Fe3 O4 + 4H2 → 3Fe + 4H2 O
(v) Fe3 O4 + H2 → 3FeO + H2 O
(vi) (1 − x)Fe3 O4 + (1 − 4x)H2 → 3Fe( 1 − x)O + (1 − 4x)H2 O
(vii) FeO + H2 → Fe + H2 O
(viii) Fe( 1 − x)O + H2 → (1 − x)Fe + H2 O
Generally, the first step in the hematite reduction is the formation of Fe3 O4 . Accord-
ingly, observation of a TPR peak at low temperature (the exact position diverges to
a large extent dependent on impurities) corresponds to the Fe3 O4 formation. The
Fe3 O4 reduction profile gave one symmetrical peak with Tmax at around 550 ◦ C,
while the reduction of FeO was observed at very higher temperature (> 750 ◦ C).
A very different and complex situation occurs when supported iron phases are
concerned. The TPR profile of the FeOx /ZrO2 catalysts prepared by impregnation
presents a convoluted curve (Fig. 5.10, left). The complexity of the TPR profile reveals
192 A. Gervasini

1000
1000

Temperature / °C
Temperature / °C
800
800
signal / a.u.

signal / a.u.
600 600
400 400
200 200
0 0
0 2000 4000 6000 8000 10000 0 1500 3000 4500 6000 7500 9000
time / s time / s

Fig. 5.10 TPR profiles of FeOx /ZrO2 (with 5 wt % of Fe) prepared by the impregnation (left) and
by adsorption equilibrium method (right)

the high heterogeneity of the iron surface in terms of nature of the oxide species,
dimension of aggregates, and oxide-support interaction. A completely different sit-
uation emerges when FeOx /ZrO2 was prepared by adsorption method (Fig. 5.10,
right). The spectrum is dominated by two well defined and intense peaks positioned
at low (Tmax,1 at ca.430 ◦ C) and high (Tmax,2 at ca.850 ◦ C) temperature. The two
well defined and sharp reduction peaks suggest that iron oxide aggregates are of
smaller dimension than those of the impregnated catalyst. The nature of the support
can deeply modify the reducing path of the supported Fe2 O3 phase. Examples can
be found in Ref. [18], in which a series of dispersed Fe-catalysts with increasing Fe
loading (from 3 to 17 wt.%) has been prepared over a mesoporous silica as support.
From a qualitative point of view, all the reducing profiles present the same feature
independent of the Fe-content (Fig. 5.11): a main reduction peak with well defined
maximum at ca. 400 ◦ C dominates the spectra, other weakly resolved maxima with
lower intensity are between 700 and 800 ◦ C. The samples at progressively higher
Fe loading consume increasing amount of H2 without any important modification of
the shape and position of the reducing peaks. Concerning the quantitative point of
view, the experimental H2 -consumptions are lower than those that can be calculated
assuming total reduction of the F2 O3 , assumed as the starting species. It has been
determined that the silica matrix had a strong inhibiting effect on the complete Fe(3+)
reduction to Fe(0) due to fayalitic phase formation (Fe2 SiO4 ). Only the reducible iron
oxide aggregates took part in the reaction [18], the high dispersion of the Fe-oxide
species imparts acidic- properties of Lewis type to the surface.
The reducibility of any metal oxide can be improved or got worse by doping the
oxide. Detailed studies were made on the reducibility of copper oxide. Hurst et al.
[2] report an exhaustive study on the CuO phase doped with 2 mol % of a large series
of metal ions (Cr, Mn, Fe, Co, Ni, Ru, Rh, Pd, Ag, and Ir, Pt, and Au). The doping
with the first-row transition metal ions gives TPR profiles with the same shape as that
obtained from CuO with a shift to lower temperatures (293 ◦ C < Tmax < 313 ◦ C)
compared with Tmax of 333 ◦ C for CuO under the same experimental conditions.
The improving effect of these ions on the reducing property of CuO likely arises
from nonspecific modification of the CuO lattice, resulting in the increase of copper
nucleation sites. The TPR profiles of Pd- and Ru-doped CuO are distinguished from
5 Temperature Programmed Reduction/Oxidation (TPR/TPO) Methods 193

H2 consumed / mmol g -1
2.0
1000
Fe17/SIM 1.6 Fe17/SIM
Signal / a.u.

800 Fe12/SIM
Fe12/SIM

T / °C
Fe6/SIM 1.2 Fe6/SIM
600
Fe3/SIM Fe3/SIM
400 0.8

200 0.4

0 0.0
0 2000 4000 6000 8000 10000 1000 2000 3000 4000 5000
time / s time / s

Fig. 5.11 TPR profiles of FeOx catalysts dispersed on a mesoporous silica (SIM) with different
amount of Fe2 O3 , from 3 to 17 wt.% (left) and H2 consumed in the reduction (right) (from Ref. [18])

all the others in that reduction is complete at very low temperatures (150−230 ◦ C)
and the TPR profiles consist of two peaks which indicate a low-temperature process
followed by a more intense high-temperature reduction. Preferential reduction of Pd
and Ru ions to form Pd and Ru nucleation sites occurred followed by an improvement
of the CuO reducibility. The ion-doping effect of Ag, Au, Rh, Ir, and Pt on CuO is
more complex; CuO reduction is completed at about the same temperature as that
observed for CuO with a more complex shape of the TPR profile.
The knowledge of the average oxidation state (AOS, nox,av ) attained for a given
oxide under defined reducing conditions (nature of the reducing agent and tempera-
ture, in particular) is of importance for several applications. TPR offers a simple way
to compute AOS for any given reducible metal species as an alternative to expensive
analytical analyses (i.e., XPS). The simple starting idea is to compare the theoretical
amount of electrons needed for the reduction of one mole of species (Ne◦theor ) with the
amount of effective electrons exchanged during reduction (Ne◦exptl ), obtained from
the TPR experiment (derived from the experimental knowledge of H2 consumed: one
mole of H2 gives 2 electrons during the reduction process). For a mole of reducing
species, nox,av can be calculated as:

nox,av = (Ne◦theor − Ne◦exptl )/SF (5.15)

where SF indicates the stoichiometric factor, number of reducible atoms in the

species; for example, SF is 2 for V2 O5 .
In alternative, AOS can be calculated from the balanced redox reaction; here
below written for a trivalent metal species, M3+ , which can be reduced to bivalent
or univalent or zerovalent metal species:

M3+ + (y − 2)/2H2 = M(5−y)+ + (y − 2)H+ (5.16)

where y indicates the experimental amount of hydrogen consumed, obtained from

TPR analysis. Once obtained y, the final average oxidation state (nox,av ) of M can be
calculated on the basis of Eq. (5.16).
194 A. Gervasini

Many examples can be found in the literature on this point. The reducibility of
vanadia catalysts has catalytic implications for selective oxidation reactions where
they found real use. The support nature and the preparation method affect the
reducibility of the vanadia phase. In Ref. [19]; pure titania or bilayered titania/silica
supports were chosen and concerning the vanadia deposition method, impregnation
and atomic layer deposition procedures were performed. The reducibility of vana-
dia improved with increasing titania loading as shown by the calculated AOS. The
lowest AOS were associated to vanadia on pure titania supports (nox,av = 3.5) while
vanadia on titania-silica supports achieved at maximum nox,av of 3.7–3.8. AOS of
vanadium after reduction was independent of the preparation method.
When supported metal oxide phases are concerned, the kind of surface species
that can be present at the surface depends on the support nature, being active a
strong or weak metal-support interaction. CuO is an important catalytic phase, easy
to disperse on acid supports that can interact with it with strong metal-support bond;
this interaction can influence the redox properties of CuO. Modified silicas with
amount of alumina (SA), titania (ST), and zirconia (SZ) in 12–14 wt.% concentration
were used to support CuO (8–9 wt.%) [20]; a commercial silica-alumina support was
comparatively studied (SAG). On the different supports, the CuO redox properties
were controlled by combining TPR and successive TPO experiments.
As depicted in Fig. 5.12, the profiles for the two silico–aluminas (SA and SAG)
were well different between each other, in terms of Tmax values and line widths.
For Cu/SAG, two well distinguished peaks of similar intensity and for Cu/SA, four
peaks with very different line widths were otained. A set of three peaks described

Fig. 5.12 Profiles of reduction (TPR) and oxidation (TPO) at programmed temperature for Cu/SAG
(a), Cu/SA (b), Cu/ST (c), and Cu/SZ (d). Experimental and decomposed lines are indicated (from
Ref. [20])
5 Temperature Programmed Reduction/Oxidation (TPR/TPO) Methods 195

the TPR profile of Cu/SZ, while quite an unique peak was obtained for Cu/ST. The
reduction of the copper phase on the supports of high acidity (SA and SZ) showed low
defined and very broadened TPR peaks at very high temperatures, (380−425 ◦ C).
The temperature and feature of these peaks suggest that they could be assigned to
copper species in interaction with the support. These compounds were more stable
to reduction than the dispersed CuO species.
The reoxidation of the reduced copper phase was accomplished by tempera-
ture programmed oxydation (TPO). All the studied catalysts could be almost com-
pletely reoxidized to CuO, as indicated by the values of oxidation-percent calculated
(>60 %). Quite the same TPO profiles were observed on all the catalysts with two oxi-
dation steps: at .first, oxidation of surface copper species and at higher temperature,
bulk oxidation. The effective redox properties of CuO allow it to find application
in several reaction processes demanding cyclic reduction and reoxidation process
with the reactants (i.e., selective catalytic reduction of NOx with hydrocarbons as
reducing agent, [20–22]).

References

1. J. Hu, L. Chen, R. Richards, in Metal Oxide Catalysis, vol. 2, ed. by S.D. Jackson, J.S.J.
Hargreaves (WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, 2009)
2. N.W. Hurst, S.J. Gentry, A. Jones, B.D. McNicol, Catal.-Rev.-Sci. Eng. 24, 233 (1982)
3. R.K. Grasselli, J.D. Burrington, Adv. Catal. 30, 133 (1981)
4. P. Mars, D.W. van Krevelen, Chem. Eng. Sci. Spec. Suppl. 3, 41 (1954)
5. A.R. Almeida, J.A. Moulijn, G. Mul, J. Phys. Chem. C. 115, 1330 (2011)
6. C. Doornkamp, V. Ponec, J. Mol. Catal. A 162, 19 (2000)
7. B. Jouguet, A. Gervasini, A. Auroux, Chem. Eng. Technol. 18, 243 (1995)
8. D.A.M. Monti, A. Baiker, J. Catal. 83, 323 (1983)
9. S.J. Gentry, N.W. Hurst, A. Jones, J. Chem. Soc. Faraday Trans. I, 75, 1688 (1979)
10. P.A. Jacobs, J.P. Linart, H. Nijs, J.B. Uytterhoeven, J. Chem. Soc. Faraday Trans. I, 73, 1745
(1977)
11. R.J. Cvetanovic, Y. Amenomiya, Catal. Rev. 16, 21 (1972)
12. V. Rakic, V. Dondur, D. Misljenovic, J. Therm. Alnal. 38, 879 (1992)
13. J.H. Chan, S.T. Balke, Polym. Degrad. Stabil. 57, 135 (1997)
14. P. Carniti, A. Gervasini, Thermochim. Acta 379, 51 (2001)
15. H. Tanaka, Thermochim. Acta 267, 29 (1995)
16. F. Eigenmann, M. Maciejewski, A. Baiker, Thermochim. Acta 359, 131 (2000)
17. C. Messi, P. Carniti, A. Gervasini, J. Thermal Anal. 91, 93 (2008)
18. A. Gervasini, C. Messi, P. Carniti, A. Ponti, N. Ravasio, F. Zaccheria, J. Catal. 262, 224 (2009)
19. J. Keranen, P. Carniti, A. Gervasini, E. Iiskola, A. Auroux, L. Niinistö, Catal. Today 91–92, 67
(2004)
20. S. Bennici, P. Carniti, A. Gervasini, Cat. Lett. 98, 187 (2004)
21. A. Gervasini, M. Manzoli, G. Martra, A. Ponti, N. Ravasio, L. Sordelli, F. Zaccheria, J. Phys.
Chem. B 110, 7851 (2006)
22. S. Bennici, A. Gervasini, Appl. Catal. B 62, 336 (2006)
Chapter 6
Calorimetry at the Solid–Liquid Interface

Jerzy Jozef Zajac

Abstract Broad principles of Solid-Liquid calorimetry together with some

illustrative examples of its use in the field of catalysis are presented here. The first
use is related to the determination of surface properties of catalysts, adsorbents and
solid materials in contact with liquids. In particular, it is shown how to evaluate the
capacity of a given solid to establish different types of interaction with its liquid
environment or to calculate its specific surface area accessible to liquids. The second
use includes the measurement of the heat effects accompanying catalytic reactions
and the related interfacial phenomena at Solid-Liquid and Liquid-Liquid interfaces.
Examples of competitive ion adsorption from dilute aqueous solutions, as well as
the formation of surfactant aggregates either in aqueous solution or at the Solid-
Liquid interface are considered in view of potential applications in Environmental
Remediation and Micellar Catalysis.

6.1 Introduction

Microcalorimetry, also nanocalorimetry to follow the recent trends in thermal instru-

mentation and analysis, is a measuring technique that can be used to study interfacial
phenomena occurring at the Solid-Liquid interface. Immersion of a solid in a pure
liquid or a solution, wetting of a solid initially in contact with a gas or vapour by
a liquid, adhesion between two condensed phases upon their “molecular” contact
are examples of exothermic phenomena which are accompanied by significant heat
evolvement. Competitive adsorption from solution is an important exception to the
exothermicity of interfacial phenomena. This is because certain components of the

J. J. Zajac (B)
Institut Charles Gerhardt Montpellier, UMR-5253 CNRS-UM2 Equipe Agrégats,
Interface, et Matériaux pour l’Energie (AIME), Université Montpellier 2,
C.C. 1502 Place Eugène Bataillon, 34095 Montpellier Cedex 5, France
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 197

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_6,
© Springer-Verlag Berlin Heidelberg 2013
198 J. J. Zajac

solution may compete against each other to adsorb at the interface. Adsorption from
solution is thus considered as an exchange process and formally split into several
adsorption and desorption steps. If one of the components is to be preferentially
accumulated at the interface, the transfer of its molecules to the interface must be
accompanied by the transfer of an equivalent amount of molecules of another com-
ponent in the reverse direction, i.e., from the interface to the interior of the solution.
The “displacement” is a frequent term for this process. The overall effect of such a
displacement may be endothermic in numerous systems, thereby giving rise to an
entropy-driven phenomenon. The complexity of displacement process is the main
reason why the van’t Hoff procedure for heat determination based on the measure-
ment of the temperature dependence of adsorption isotherms frequently leads to
unreliable values. In consequence, direct measurement of the thermal effect in these
systems by calorimetry is strongly recommended.
Frequently the heat values for the displacement are relatively small, and thereby
difficult to be detected. With the recent progress in ultra-sensitive heat flow mea-
surements and the use of a wide variety of accessories to control the experimen-
tal conditions, new commercially available calorimeters offer maximum sensitivity,
flexibility, and productivity. Nowadays it becomes possible to study the competi-
tive adsorption phenomena with increased sensitivity and lower detection limits than
previously possible.
In the field of catalysis, calorimetry may be used in two manners. The first use is
related to the determination of surface properties of catalysts, adsorbents and solid
materials in contact with liquids. In particular, it is possible to evaluate the capacity of
a given solid to establish different types of interaction with its liquid environment or to
calculate its specific surface area accessible to liquids. The second use of calorimetry
includes the measurement of the heat effects accompanying catalytic reactions and
the related interfacial phenomena at Solid-Liquid and Liquid-Liquid interfaces. In the
present chapter, this group of calorimetry applications will be illustrated by following
the examples of competitive ion adsorption from dilute aqueous solutions and the
formation of surfactant aggregates either in aqueous solution or at the Solid-Liquid
interface.
The present chapter does not pretend to be an exhaustive record of Solid-Liquid
calorimetry applications in Surface Science and Technology. It should be rather
regarded as an introductory course with some illustrative examples. It is important
to realise that the individual author’s experience in the field has been the principal
criterion for selection of specific instruments and their uses, without any intention
of neglecting other contributions. The presentation of calorimetry methods will be
restricted only to interfacial systems composed of a pure liquid or a dilute binary,
at the most, solution in contact with a solid which does not dissolve in the liquid
phase. This formalism may be still employed in the case of solutions which are
not strictly binary but may be viewed as such (e.g., solutions containing ionizable
solutes, background electrolytes or other additives that may be lumped together as
constituting a mean solvent or a mean solute).
6 Calorimetry at the Solid–Liquid Interface 199

6.2 Thermodynamic Treatment of the Solid–Liquid

Interface and the Related Interfacial Phenomena

This opening paragraph reviews some basic ideas and methods relating to interfacial
phenomena at the Solid-Liquid interface. Usually, the subject is characterised by two
main approaches: the presentation of these phenomena in terms of thermodynamics
and their molecular interpretation. The detailed treatment of such general concepts
and relationships can be found in numerous standard texts [1–5] and the interested
reader should consult these texts. Here only a very brief review is provided on which
to base the entire text, in particular the language that will be widely used in further
discussion.

6.2.1 Surface Excess Functions and Surface Phase Model

Contrary to the bulk liquid phase which is homogeneous in three directions in space,
has a characteristic composition, and is also autonomous (i.e., its extensive properties
depend only on the intensive variables characterising this phase such as the temper-
ature T, the pressure P, and the chemical potentials of the solvent μ1 and the solute
μ2 ), the formal thermodynamic description of a Solid-Liquid interface presents a
serious difficulty. In the interfacial region, the density ω of any extensive quantity 
changes continuously throughout the thickness (Fig. 6.1a).
For real solids, even the two-dimensional homogeneity of the interface is very
difficult to attain, because the solid boundary is heterogeneous both in a physical
and in a chemical sense (surface heterogeneity and roughness) [6, 7]. In order to
overcome this difficulty, the only possible way is to introduce the so-called excess
thermodynamic functions [1, 6, 8].
The widely used definitions of excess functions are based on the Gibbs model of
the system, in which a flat interface is regarded as a mathematical dividing plane (the
Gibbs dividing surface—GDS) [1, 6]. The two phases α and β in contact are assumed
to remain homogeneous up to the GDS (Fig. 6.1b depicts such a model). Since the
Gibbs model provides a complete description of the heterogeneous system in physico-
chemical equilibrium, the formal thermodynamic study of interfacial phenomena is
commonly based on this approach. Compared to bulk phases, the thermodynamic
expressions for the interface contain additional terms relating to interfacial tension
and adsorption of chemical species. This model has the advantage of leading rapidly
to the principal thermodynamic relationships between the interfacial quantities (e.g.,
the so-called Gibbs adsorption equation) [1, 6].
Nevertheless, the Gibbs formalism has some serious drawbacks. Firstly, the inter-
facial properties (i.e., the Gibbs excess functions) adopt different values depending
on the position of the GDS and, consequently, have no direct experimental signifi-
cance. For dilute solutions, this imperfection may be removed through introducing
the so-called relative interfacial quantities. Secondly, the Gibbs surface has zero
200 J. J. Zajac

Fig. 6.1 A hypothetical profile of the density  of some extensive property  (e.g., number of
moles, internal energy, free energy, enthalpy, Gibbs energy, entropy) in the heterogeneous system
as a function of the distance x perpendicular to the planar interface. (a) Real system: values of  α
and  β are determined at such a distance from the interfacial region that the two phases have their
bulk properties, (b) Gibbs model of the interface: value of the interfacial excess σ is given by a
sum of areas I and II [1, 6]; (c) Surface phase model of the interface: interfacial property s is
defined such that areas I, II and III compensate for one another [6, 8]

thickness and volume and this is at variance with the obvious physical picture of
an interface. Furthermore, the physical meaning of the Gibbs excess functions is
difficult to translate into molecular terms.
6 Calorimetry at the Solid–Liquid Interface 201

The alternative surface phase model popularized by Guggenheim is conceptually

simpler since the interfacial region is approximated by a thin, homogeneous layer
having an arbitrary thickness [6, 8]. This is illustrated in Fig. 6.1c where the two
phases α and β remain homogeneous up to the imaginary planes which constitute
the boundaries of the surface phase. This representation may be directly used for the
purpose of constructing molecular models of the interface. The related interfacial
properties have a real physical significance and are compatible with the experimen-
tally measured quantities. For systems containing a thermodynamically inert solid in
contact with a pure liquid or a dilute solution, one of the separating planes is chosen
to coincide with the surface of the solid. Therefore, the area A of the interface (i.e.,
its cross-sectional area) is identified with the surface area of the solid phase acces-
sible to the liquid. In practice, the main challenge is always to evaluate correctly the
thickness τ s of the interfacial region.
In the Guggenheim convention [8], the value of any extensive property of the
surface phase, per unit area of the interface, may be expressed as

1   
s = τ s ·  s =  − αV α + βV β (6.1)
A

where  is the total extensive property of the whole system;  α ,  β and  s are
the densities of , respectively, in the two bulk phases and in the surface phase of
thickness τ s and surface area A; Vα and V β represent the volume of the bulk phases
α and β. The interfacial enthalpy H s , interfacial Gibbs energy G s , and interfacial
entropy S s are defined in such a manner.
It is satisfactory to define the interfacial tension γ as the work required to cre-
ate isothermally and reversibly a unit area of an interface [1, 8]: γ SG (Solid-Gas
interface), γ S L (Solid-Liquid interface), and γ LG (Liquid-Gas interface). Since γ is
referred to as an energy per unit area in this formulation, the privileged SI unit is
J m−2 . Nevertheless, interfacial tensions reported in J m−2 and N m−1 have the same
numerical value. Usually more convenient is the submultiple mJ m−2 or mN m−1
(numerically equivalent to the previously used c.g.s. units). Conceptually, the inter-
facial tension can be also seen as a new excess quantity that is attributed to the surface
phase (or to the GDS) for the Guggenheim (or Gibbs) model to be thermodynamically
equivalent to the real system [1, 8]. In heterogeneous systems where adsorption does
not occur, like those containing the interface between one-component liquid and
gas phases, interfacial tension is numerically equal to interfacial Gibbs energy G s
(per unit area of the interface). Otherwise, adsorption takes place with a change in
interfacial tension.
From a mechanical standpoint, the interface between a pure liquid and its own
equilibrium vapour (or air, as adsorption is to be neglected here) behaves as a mem-
brane of infinitesimal thickness stretched uniformly and isotropically by a force
exerted tangential to it. Rapid relaxation towards equilibrium is the hallmark of liq-
uid surfaces: when the viscosity of the liquid is not too high, the freshly formed
area has enough time to relax completely and the equilibrium interfacial tension
will attain the same value in all surface parts. It is important to realise that, owing
202 J. J. Zajac

to non-equilibrated cohesive forces operating in the interfacial region, the liquid

squeezes itself together until it has the locally lowest surface area possible. There-
fore, γ LG is regarded as a force (per unit length of surface edge) which opposes any
attempt to increase the surface area. The terms surface and interfacial tensions are
used interchangeably for γ LG . The surface tension of most liquids (against equilib-
rium vapour or air) near room temperature ranges between 10 and 80 mJ m−2 , and
decreases in a nearly linear fashion as the temperature rises. The description of the
common experimental methods with comments on their suitability may be found in
Refs. [4, 9].
In pure water, the collective action of intermolecular hydrogen bonds together
with classical Van der Waals forces make water molecules stay close to one another
(one water molecule is capable of forming four hydrogen bonds since it can accept
two and donate two hydrogen atoms). When pure water is in contact with air, the
great surface tension tends to minimize the area of hydrophobic-hydrophilic contact:
the experimental value of γ LG at 298.15 K is equal to 71.99 ± 0.05 mJ m−2 [10].
In the temperature range 273.15–323.15 K, the effect of temperature on the surface
tension of water against air is given by [11, 12]:

γ LG = 75.716 − 0.1416 · (T − 273.15) +

− 0.25054 × 10−3 · (T − 273.15)2 (6.2)

Figure 6.2 illustrates the temperature dependence of the surface tension and the
s of water at
numerical procedure leading to the estimate of the surface enthalpy HLG
room temperature.

Fig. 6.2 Temperature dependence of the surface tension γ LG of water in contact with air [13] and
s at 298.15 K. The surface entropy S s at
the numerical determination of the surface enthalpy HLG LG
−2
298.15 K is equal to 0.157 mJ m K −1
6 Calorimetry at the Solid–Liquid Interface 203

The value of γ LG , which is still equal to 58.9 mN m−1 at T = 373.15 K, approaches

zero in the vicinity of the critical temperature T = 647.4 K, where there is no longer
an interface between the liquid and the vapour.
In the case of solids, the term surface tension is used only to designate the interfa-
cial tension γ S0 operating at the boundary between the solid phase and the surround-
ing vacuum. The determination of γ S0 for solid surfaces by stretching the surface area
against the surface stress is not possible, since the latter is not equal to γ S0 . Solids
do not deform reversibly and are capable of retaining their non-equilibrium shapes
for a long time [4]. The determination of γ S0 is sometimes possible from the calcu-
lation of reversible work of cleaving a crystal, i.e., by creating fresh surface having
the same properties as the original. Several known examples are given in Table 6.1.
These results are, however, subject to considerable uncertainty, because the cleavage
technique is not entirely reversible [14]. Furthermore, the different crystallographic
faces have somewhat different surface tensions owing to the differences in packing
density of the atoms.

6.2.2 Adhesion and Cohesion

Adhesion between a pure liquid and a solid may be described in terms of the interfacial
and surface tensions [18]. Consider the reversible process of splitting a unit area of
the Solid-Liquid interface in such a way as to create a unit area of the Solid-Vacuum
interface and a unit area of the Liquid-Gas interface (as shown in Fig. 6.3a).
Adhesion between the two phases is defined as the reversed process and the Gibbs
energy of adhesion is given, at fixed P and T, by the Dupré equation [4, 19–21]

adh G S L = γ S L − γ S0 − γ LG , per unit area of the interface (6.3)

where γ S L is the interfacial tension between both phases; γ S0 and γ LG are the individ-
ual surface tensions of the solid against vacuum and the liquid against its equilibrium
vapour (or air). For a single solid or liquid phase an analogous procedure (as shown
in Fig. 6.3b) yields the Gibbs energy of cohesion (per unit area of the interface)

coh G S = −2γ S0 or
coh G L = −2γ LG (6.4)

Table 6.1 Surface tensions

Solid Cleavage plane γ S0 (mJ m−2 )
of several solid crystals
against vacuum, as obtained Mica − 4500
from the work of cleavage, MgO (100) 1200
wcliv = 2 · γ S0 [14–17] CaF2 (111) 450
LiF (100) 340
CaCO3 (001) 230
NaCl (100) 110
204 J. J. Zajac

Fig. 6.3 Schematic illustration of the reversible process of (a) adhesion and (b) cohesion

Combination of Eqs. (6.3) and (6.4) yields

1
γ S L =
adh G S L − (
coh G S +
coh G L ) (6.5)
2
In a sense, the right-hand side of Eq. (6.5) may be considered as a generalised Gibbs
energy of mixing. If the two phases mix spontaneously in all proportions, the Gibbs
energy will decrease during such a process, thereby rendering γ S L negative. In ther-
modynamic terms this means that there is no stable interface. When the phases are
immiscible, separation is spontaneous and the interfacial tension becomes positive.
In that case, the interface is stable.
In broad outline, the cohesion of molecules (atoms, ions) to form the bulk phase of
matter is due to long-range physical interactions (mainly of the van der Waals type)
and short-range chemical forces (giving rise to covalent, ionic, metal, or hydro-
gen bonds). Amongst these various types of interactions encountered more or less
frequently in interfacial phenomena involving liquid and solid phases, the London
(dispersion) forces and the Lewis acid-base ones are really crucial in the construction
of a thermodynamic treatment of interfaces [19, 21–23]. Solid and liquid substances
may be classified according to their capacity of forming Lewis acid-base bonding.
Materials that can be both Lewis acids (electron acceptors) and Lewis bases (electron
donors) are termed bipolar. The monopolar acidic or monopolar basic substance can
act exclusively as a Lewis acid or a Lewis base (the other property is negligible).
6 Calorimetry at the Solid–Liquid Interface 205

Inert materials, capable of neither acid nor base interactions, are called apolar. It
should be noted here that this nomenclature [22, 23] has nothing in common with
polarity of molecules per se, as measured by their respective dipole moments.
Dispersion forces are universal because they attract all molecules together, regard-
less of their specific chemical nature. The potential energy of dispersion attraction
between two isolated molecules decays with the sixth power of the separation dis-
tance. Based on the so-called Hamaker theory (i.e., the method of pair-wise summa-
tion of intermolecular forces) or the more modern Lifshitz macroscopic treatment of
strictly additive London forces, it is possible to develop the so-called Lifshitz-Van
der Waals expression for the macroscopic interactions between macroscopic-in-size
objects (i.e., macrobodies) [19, 21]. Such an expression strongly depends on the
shapes of the interacting macrobodies as well as on the separation distance (non-
retarded or retarded interaction). For two portions of the same phase of infinite
extent bounded by parallel flat surfaces, at a distance h apart, the potential energy of
macroscopic attraction is:
A11
U (h) = − (6.6)
12π · h 2
where A11 is the so-called Hamaker constant which depends on the chemical nature
of the molecules (atoms, ions) constituting the phase under consideration and the
number of molecules per unit volume in two interacting bodies. The more gradual
fall-off of the potential energy (6.6) with distance compared to the molecule-molecule
interaction indicates that macroscopic attractions are of a more long-range type and
they are expected to make a significant contribution to the total energy of attraction
even at longer distances.
The values of A11 for the various substances interacting across vacuum or across
a medium can be found in Refs. [24, 25]. It should be noted that Hamaker constants
for interaction across a medium are usually much lower in comparison with the
related values under vacuum (e.g., in the case of two macroscopic bodies of quartz
at short distances apart, A11 equals 41.3 × 10−20 J for interaction across vacuum and
1.3 × 10−20 J for interaction across water [25]).
The macroscopic effect of cohesion due to dispersion forces is usually calculated
from the Lifshitz-Van der Waals expression (6.6), providing that the separation dis-
tance h is known. Israelachvili [21, 22] has proposed a universal value of 0.165 nm
to describe the effective spacing h between molecular planes in all liquids with mole-
cules interacting solely through dispersion forces. In this case, the Gibbs energy of
cohesion may be evaluated as

coh G L ≈ 9.74 × 1017 A11 in J m −2 , for apolar liquids (6.7)

where A11 is the Hamaker constant for the liquid substance. In general, consider-
able theoretical and experimental evidence is consistent with the postulate that the
Keesom-Debye contribution to the Gibbs energy of cohesion is very small and has
no significant importance between macroscopic bodies in the condensed systems
[21, 26]. As a result, the Lifshitz-Van der Waals (LW) component
coh G L W of the
206 J. J. Zajac

Gibbs energy of cohesion for any liquid or solid material is commonly identified
with the dispersion contribution.
According to Eq. 6.4, the apolar component of the surface tension of a solid against
vacuum or a liquid against its equilibrium vapour (or air) becomes:

1 1
γ S0
LW
= −
coh G SL W or γ LG
LW
= −
coh G LL W (6.8)
2 2
Lewis acid-base interaction between molecules (atoms, ions) differs from a clas-
sical covalent bond in that only one of the partners supplies the pair of electrons
[27]. Electron pair donors (EPD) are molecules which donate the lone pair of non-
bonding electrons (n-EPD), the electron pair of a σ -bond ( σ -EPD), or the pair of
π -electrons ( π -EPD). Electron pair acceptor (EPA) molecules may use a vacant
valence orbital (n-EPA), a nonbonding σ -orbital ( σ -EPA), or a π -bond system
with electron-withdrawing substituents (π -EPA). The combinations between all the
above donor and acceptor types result in nine types of EPD-EPA complexes, with the
bond strength ranging from high values for n-EPD/n-EPA associations to very week
π -EPD/π -EPA interactions between neutral molecules. Formally, Lewis acid-base
interaction includes a hydrogen bond which is usually situated at the lower end of
the chemical bond range.
It has long been an operational premise that the Gibbs energy of cohesion for any
non-metallic liquid or solid phase can be split into two contributions: an apolar one

coh G L W originating chiefly from dispersion forces and a polar one

coh G AB aris-
ing from the Lewis acid-base interaction between the constituent molecules (atoms,
ions). Several semi-empirical methods have been proposed in the literature, e.g.,
[18, 26, 28–30], to determine both contributions. The results can be collated for
different liquids and solids and subsequently used to predict the behaviour of new
interfaces. Nevertheless, it is important to understand the approximate nature of this
approach and to consider the resulting conclusions with caution.
In consequence, the total surface tension of a given non-metallic material against
vacuum or its own equilibrium vapour is expressed by the sum of the Lifshitz-van
der Waals (LW) and Lewis acid-base (AB) contributions [19, 22]

1 
γ S0 = γ S0
LW
+ γ S0
AB
=−
coh G SL W +
coh G SAB (6.9a)
2
1 
γ LG = γ LG
LW
+ γ LG
AB
= −
coh G LL W +
coh G LAB (6.9b)
2
The change of Gibbs energy during adhesion between two phases is the macroscopic
outcome of interactions between the microscopic constituents of the different phases.
According to the empirical Berthelot principle [21], the London energy of attraction
between two dissimilar macrobodies is a geometric mean of the mutual interactions
between similar objects, so that

A12 ≈ A11 A12 (6.10)
6 Calorimetry at the Solid–Liquid Interface 207

where A12 is the Hamaker constant referring to macroscopic interaction between

two different phases.
To obtain the apolar (LW) contribution to the Gibbs energy of adhesion
adh G SL LW
between a solid and a liquid, it is assumed that expressions analogous to that given
by Eq. 6.7 are still valid. The Berthelot principle Eq. 6.10 may be therefore used to
evaluate
adh G SL LW . The combining rule for this component of
adh G S L is given by
the Good-Girifalco-Fowkes relation [18, 31]:


adh G SL LW =
coh G SL W ·
coh G LL W = −2 γ S0
LW · γ LW
LG (6.11)

If one of the condensed phases is apolar, dispersion forces are the only important
type of interaction operating across the interface and
adh G S L =
adh G SL LW . The
apolar (LW) surface tension component of any solid can be thus determined by
the measurement of the Gibbs energy of adhesion between this material and an
apolar probe substance. The latter may be a liquid alkane, methylene iodide, or α-
bromonaphthalene, for which the surface tension γ LG = γ LG L W has already been

measured.
The two-condensed-phase analog of Eq. 6.9a, 6.9b is:

adh G S L =
adh G SL LW +
adh G SAB
L (6.12)

The polar (AB) contributions to the surface tension and Gibbs energy of adhe-
sion are sometimes expressed in terms of Van Oss-Chaudhury-Good parameters
[18, 19, 22, 32]

γ AB = 2 γ A · γ B (6.13a)


adh G SAB
L = −2 γ A ·γB −2 γA ·γB
S0 LG S0 LG (6.13b)

where γ A and γ B are the Lewis acid and Lewis base parameters of surface tension
γ , respectively. This approximation leads to the following interesting conclusions:
the polar (AB) surface tension component of a pure substance is equal to zero, if
the substance is monopolar or apolar. There is no acid-base interaction across the
interface, i.e.,
adh G SABL = 0, if one of the components is apolar or if the two
components are monopolar in the same sense, i.e., both being monofunctional acids
or both monofunctional bases.
The Lewis acid and base contributions to the surface tension of solids can be
derived from measurements of the Gibbs energy of adhesion between the material
and a polar probe liquid. At least two different polar liquids (e.g., water and appro-
priate monofunctional liquid) must be used as probes, provided that their surface
components γ LG L W , γ A , γ B are known.
LG LG
208 J. J. Zajac

6.2.3 Wetting in Solid–Liquid Systems

Wetting includes the spreading of a pure liquid over the surface of a solid, displacing
the gas (or vapour) initially in contact with that surface [18, 33]. Hence the phe-
nomenon involves three interfaces, namely Solid-Gas, Solid-Liquid, and Liquid-Gas
ones. The spreading coefficient W S is defined as [4, 33]

W S = γ SG − (γ S L + γ LG ) (6.14)

where γ SG , γ S L and γ LG are the appropriate interfacial tensions for the three inter-
faces at equilibrium. When a portion of the liquid is placed on a uniform, perfectly
flat, and non-deformable solid surface and the two phases are allowed to come to
equilibrium with the surrounding gas phase, one of the two events may happen:
1. When W S is positive or zero, the liquid wets the solid material, i.e., spreads out
spontaneously over its surface, providing there is enough liquid to eliminate a
unit area of the Solid-Gas interface while exposing a corresponding amount of
the Solid-Liquid and Liquid-Gas interfaces.
2. When W S is negative, the liquid remains as a drop having, at equilibrium, a definite
angle of contact with the solid surface (the liquid does not wet the solid). This
case is illustrated in Fig. 6.4.
The equilibrium contact angle  between the liquid and the solid phases is deter-
mined by the following balance of interfacial tensions [4, 33, 34]:

γ SG − γ S L
cos  = (6.15)
γ LG

known as Young’s equation. This relation applies for contact angles  less than,
equal to, or greater than 90◦ . In the limiting case where  = 0, the liquid wets out
the solid.
The classical form of Young’s equation, which describes the equilibrium balance
of forces meeting at the three-phase contact line in the plane of the solid surface (see
Fig. 6.4), is one of the most controversial expressions in Surface Science and there is

Fig. 6.4 A drop of a non-spreading liquid on a flat solid surface together with the traditional
representation of the vectorial equilibrium between respective interfacial tensions viewed as forces
acting along the perimeter of the drop
6 Calorimetry at the Solid–Liquid Interface 209

a long list of objections to it (e.g., see Refs. [4, 33, 34] for details). In spite of that, it is
still very widely encountered in the literature. To derive it, one must assume an ideal
solid: chemically homogeneous, thermodynamically inert (e.g., it cannot swell under
the action of the liquid neither can dissolve in the liquid), and flat at an atomic scale.
In practice, appreciable hysteresis of the contact angle is observed in real systems
(chemical heterogeneity and roughness of solid surfaces), depending on whether
the liquid is advancing or receding across the solid surface [33]. Advancing contact
angles are larger than receding angles, and the difference may be sometimes as much
as 20◦ –30◦ . A very detailed critical discussion of the various methods for measuring
contact angles can be found in Ref. [34]. The contact angles of powdered solids (e.g.,
clay minerals) are technically important but are difficult to measure. In the case of
numerous fine-grained minerals, which do not occur as large, perfect single crystals
with well-developed faces, the contact angles are determined indirectly by column
and thin layer wicking [35, 36].
If the liquid is volatile, the gas phase will contain its vapour. In consequence, even
though both the fluid phases are nominally pure components, there is in general finite
adsorption at the Solid-Gas interface. The equilibrium value of γ SG for this interface
will be, therefore, lower than its pure-component value γ S0 by an experimentally
determinable quantity, which is called the two-dimensional or surface pressure π SG
[4, 5]. For the one-component gas phase under the conditions of sufficiently low
pressures in contact with an inert solid adsorbent, the value of π SG can be evaluated
using the adsorption isotherm for the vapour of the liquid on the solid surface:

p∗
 ∗

π SG = γ S0 − γ SG p = RT s d ln p T, P = const (6.16)
0

where R is the molar gas constant; s denotes the number of moles of gas adsorbed
per unit area of the solid adsorbent related to the equilibrium bulk pressure p in
the bulk gas phase, at constant temperature T and pressure P; p ∗ is the equilibrium
pressure at which the actual adsorbed film has been formed on the solid surface: the
integration of the adsorption isotherm s = s ( p) is carried out over a p-range from
0 to p ∗ . On applying Eq. 6.16 to the solid surface saturated with the vapour, i.e., when
the latter forms an equilibrium, physically adsorbed film on the available adsorbent
surface at p equal to the saturation vapour pressure psat , the surface concentration
s and the surface pressure π SG are found to reach their (positive) maximum values,
ms and π m , respectively.
SG
In the case of solids having relatively small values of surface tension γ S0 against
vacuum (usually less than 100 mN m−1 [2]), the effect of gas adsorption is thought
to be of little importance. Polymers and many other solid organic compounds are
usually given as examples of this category of substrates. Solid materials for which
π SG
m is small for any adsorbate are named the low-energy solids. For high-energy

solids (e.g., mineral oxides, metal sulphides, inorganic salts), the decrease in surface
tension due to adsorption is significant [30, 37, 38].
210 J. J. Zajac

Equations 6.3, 6.15, and 6.16 may be combined to give the Gibbs energy of
adhesion between the solid and liquid phases

adh G S L = −γ LG (1 + cos ) − π SG
m
(6.17)

This is a very useful relation, in which γ LG , cos , and π SG

m can be measured and

calculated quite easily and accurately. Therefore, the operating procedures for the
adhesion experiment are usually based on Eq. 6.17.

6.2.4 Hydrophobic and Hydrophilic Substances

Commonly the distinction between hydrophobic and hydrophilic substances is based

on the analysis of interactions between their molecules and water as a solvent. A more
precise classification of liquid and solid substances as hydrophobic and hydrophilic
may be constructed basing on the apolar (LW) and polar (AB) components of their
surface tensions. This three-parameter approach is of great importance for the under-
standing of surface behaviour [22, 32, 39, 40]. A non-metallic substance is hydropho-
bic if it interacts with water by exhibiting only LW character. It has very little (or none
at all) Lewis acid or Lewis base character. Typical substances at the hydrophobic end
have low γ L W surface parameters and their γ A and γ B components are equal to
zero. Hydrophilic substances have non-zero γ L W components of the surface tension
and at least one of their γ A and γ B parameters is significant.
The values of surface tension components that have been derived from the appro-
priate measurements of the Gibbs energy of adhesion for numerous liquids and solids
A and γ B values for probe
are listed in Table 6.2. The determination of a set of γ LG LG
liquids is based on the choice of the first reference liquid. For this purpose, Van
Oss et al. [39] assumed that γ LG A = γ B for water. In consequence, all of acid-base
LG
parameters in Table 6.2 are relative to those of water.

6.3 Calorimetry Applied to Evaluate Surface

Properties of Solids

The determination of the apolar and polar components of the surface tension of
liquids and solids may be a powerful tool for classification of various substances
with respect to their hydrophobic-hydrophilic character. In the case of solids, this
macroscopic approach provides important information about the potential “global”
behaviour of their surfaces against a given environment, however, giving no direct
indication in regard with the heterogeneity of the solid surface. In consequence,
it cannot replace methods based on the adsorption of probe molecules from the gas
phase, but it does complement them by providing a different level of surface scanning.
6 Calorimetry at the Solid–Liquid Interface 211

Table 6.2 Surface tension parameters (in mJ m−2 ) of some liquids and polymers [23, 32, 39]
Substance γ LW γA γB γ S0 /γ LG
Water 21.8 25.5 25.5 72.8
Liquids
n-Heptane 20.1 0 0 20.1
n-Decane 23.8 0 0 23.8
Chloroform 27.15 3.8 0 27.15
n-Hexadecane 27.5 0 0 27.5
α-Bromonaphthalene 44.4 ≈0 ≈0 44.4
Methylene iodide 50.8 ≈0 ≈0 50.8
Ethylene glycol 29.0 1.92 47.0 48.0
Formamide 39.0 2.28 39.6 58.0
Glycerol 34.0 3.92 57.4 64.0
Polymers
Poly(methylmethacrylate),
cast film 39–43 ≈0 9.5–22.4 39–43
Poly(vinylchloride) 43 0.04 3.5 43.72
Poly(oxyethylene):
PEG 6000 45 ≈0 66 45
Cellulose acetate 35 0.3 22.7 40.2
Cellulose nitrate 45 0 16 45
Poly(styrene) 42 0 1.1 42

In practice, the experimental procedure is quite long and fastidious referring to the
successive determinations of the Gibbs energy of adhesion
adh G S L between a given
solid and an apolar or polar liquid, which requires, in accordance with Eq. 6.17,
measurement of the surface tension γ LG , contact angle , and vapour adsorption
m ) for each solid-liquid couple. To make matters worse,
isotherm (to calculate π SG
the very precise measurement of  is possible only for atomically smooth surfaces.
Finally, the additive approximation expressed by Eqs. 6.9a, 6.9b and 6.12 is better
suited to calculation of the enthalpy term than to that of the free energy, since the
interactions may have both mechanical and entropic contributions [41].
s and H s from direct
The determination of the related surface enthalpy terms HS0 LG
calorimetry measurements provides an alternative way to evaluate the hydrophobic-
hydrophilic character of a solid surface [38, 42–44], since the following deconvo-
lution procedures may be proposed for the surface enthalpy in analogy with those
holding for the Gibbs energy (Eqs. 6.8–6.13a, 6.13b):
• surface enthalpy of a non-metallic solid against vacuum:

s
HS0 = HS0
LW
+ HS0
AB
= HS0
LW
+ 2 HS0
A · HB
S0 (6.18a)

• surface enthalpy of a non-metallic liquid against its equilibrium vapour (or air):
212 J. J. Zajac

s
HLG = HLG
LW
+ HLG
AB
= HLG
LW
+ 2 HLG
A · HB
LG (6.18b)

• Solid-Liquid interfacial enthalpy (Berthelot principle):

HSs L = HSLLW + HSAB

L = HS0 + HLG +
s s
 
−2 HS0 · HLG + HS0 · HLG + HS0 · HLG
L W L W A B B A (6.18c)

where HS0 L W , H L W and H AB , H AB are, respectively, the apolar (LW) and polar (AB)
LG S0 LG
contributions to the appropriate surface enthalpy; HS0 A , H A denote the Lewis acid
LG
and HS0 , HLG the Lewis base components of the surface enthalpy ; the HSLLW , HSAB
A A
L
components refer to the Solid-Liquid interface.
It should be always remembered that the global treatment of surface hydrophobi-
city-hydrophilicity based on surface and interfacial enthalpies does not include the
entropy effects.
For liquids and solids, specific orientation and conformation of unsymmetrical
molecules (ions) in the interfacial regions result not only in the maximization of
their interaction energy, but also yield entropy effects that cannot be neglected. For
example, molecular dynamics calculations of the intermolecular potential function
points to a predominant orientation of the water dipoles at the Liquid-Gas interface
[45]. Other examples are an icelike structuring of water molecules in the vicinity of
crystalline solid surfaces [46] and a specific orientation of the alcohol molecules in
the interface between a liquid n-alkanol and water [47].
Immersional and wetting calorimetry is an important method for studying inter-
actions at the Solid-Liquid interface [48–50], especially in the case of finely divided
and porous solids where the direct measurements of contact angle are hardly possible.

6.3.1 Enthalpy Changes in the Thermodynamic Cycle

of Immersion-Adsorption–Wetting

Consider a simple experiment in which a clean solid surface (free of adsorbed liquid
and vapour impurities) is immersed in an excess of pure liquid (Path 1 in Fig. 6.5).
If thermal effects arising from absorption, solubility, and swelling of a solid may be
eliminated, the whole enthalpy change on immersion is ascribed only to the interface.
Sometimes the immersion of a solid in a liquid is accompanied by the formation of an
electrical double layer. For mineral oxide-water systems [51, 52], the double-layer
effects (i.e., generation of surface charge by protonation or deprotonation of some
surface hydroxyl groups, and adsorption of counterions in the Stern or/and diffuse
layers) are clearly secondary in comparison with the basic wetting (this contribution
is 10–15 % of the total heat effect, at the most).
The total enthalpy change, at constant P and T, is then written as
6 Calorimetry at the Solid–Liquid Interface 213

Fig. 6.5 Schematic representation of the difference between immersion (clean solid surface) and
immersional wetting (solid surface pre-covered with vapour); the excess of liquid is high enough
for the enthalpy change in the bulk liquid phase during adsorption or immersion to be neglected

 

imm H = A S L ·
imm H ∗ = A S L HSs L − HS0
s
(6.19)

where
imm H is called the enthalpy of immersion; A S L is the area of the Solid-
Liquid interface (often identified with the surface area of a non-microporous solid);
HSs L and HS0s are the interfacial enthalpies per unit area for the Solid-Liquid and

Solid-Vacuum interfaces, respectively.

The experiment of immersion is sometimes performed under completely different
conditions. The solid may be first put in equilibrium with the vapour of the immer-
sional liquid at a given equilibrium pressure p (Paths 2 or 3 in Fig. 6.5). The adsorbed
gas may be at submonolayer, monolayer or multilayer coverage, depending chiefly
on the value of p, but also on the nature of liquid and solid. When the solid is subse-
quently immersed in the liquid, the measured enthalpy change, called the enthalpy
of immersional wetting,
W H , will be different from
imm H . The various stages
of the immersion-adsorption-wetting cycle are shown in Fig. 6.5.
In the case of ideal wetting (this means that the contact angle among the solid,
immersional liquid, and vapour of this immersional liquid  = 0), the final state is
214 J. J. Zajac

always the same, irrespective of the path followed. Therefore, one can write for Path
2 in Fig. 6.5:
 

W H = A S L ·
W H ∗ = A S L HSs L − HSG
s
( p)
=
imm H −
ads H ( p) (6.20)

where HSG s
( p) is the interfacial enthalpy for the Solid-Gas interface at equilibrium
pressure p. The area of the Solid-Gas interface is taken to be identical with A S L (the
accessibility of the solid surface does not change when passing from the vapour to
the liquid phase).
The enthalpy change
 s σ


ads H ( p) = A S L HSG ( p) − HS0 (6.21)

refers to the formation of an adsorbed film onto solid in equilibrium with the current
gas phase and is therefore named the enthalpy of adsorption from vapour.
Since adsorption at the solid-gas interface is in general exothermic,
W H
increases (becomes less negative) monotonously with precoverage from the initial
value
imm H at p = 0 to a limiting steady value
W Hm at p = psat . Examples of
such curves are presented in Fig. 6.6.
As far as the enthalpy is concerned, the interface between the saturated (multilayer)
film on the solid surface and the equilibrium vapour phase at p = psat may be identified
with the interface between the liquid and its own vapour [49, 54, 55]. Therefore,
 

W Hm = A S L HSs L − HSG
s
( psat ) = −A S L HLG
s
(6.22)

s is the surface enthalpy of the immersional liquid.

where HLG

6.3.2 Immersional and Wetting Calorimetry Experiments

Measurements of the enthalpy changes accompanying the immersion and wetting

phenomena may be performed with a Tian-Calvet type differential calorimeter
[49, 54]. The experimental procedure includes several intermediate stages which
may take much time, especially when the wetting enthalpy is measured as a function
of the surface coverage by the vapour of the immersional liquid.
The first stage corresponds to the sample preparation during which a solid sample
of a given mass is placed in a bulb made of high temperature glass and closed by
a brittle tail (Fig. 6.7a). The subsequent sample evacuation and pre-coverage steps
are performed with the solid enclosed in this bulb, as represented schematically in
Fig. 6.7. The bulb fixed to the end of a glass tube is placed in the outgassing rig where
the sample is evacuated at a high temperature to remove all adsorbed impurities from
its surface. Then the tube is sealed off a few cm from the bulb end, fixed to the end
6 Calorimetry at the Solid–Liquid Interface 215

Fig. 6.6 Enthalpy of immersional wetting per unit surface area

W H ∗ (taken with the opposite
sign) as a function of the surface pre-coverage for two powdered solid samples in two immersional
liquids: water (H2 0) and n-decane (n-C10 ) [53]

of glass rod and transferred to the calorimeter. Prior to wetting experiment, the solid
sample is outgassed in a glass bulb and then brought into contact with the vapour
of the immersional liquid at a given pressure and at constant temperature TM . The
equilibrium pressure of the pre-coverage step is controlled by the temperature Tb of
the thermostated bath (Tb must be lower than ambient temperature and TM ).
The Tian-Calvet type calorimeter system contains a massive calorimetric bloc
which acts as a heat sink (its temperature is constant) and a removable calorimetric
cell made of stainless steel and designed to fit in a cylindrical hole inside the calori-
metric bloc. The instrumental signal is obtained by measuring the heat flux between
the cell and the bloc. The temperature signal is derived from a sensor in the bloc:
two symmetrical thermal flux meters, each constructed by a series of thermocouples
surrounding a cylindrical hole for the measurement cell. The electric signal deliv-
ered by the difference in output voltage of the two flux meters is proportional to the
temperature difference θ between the bloc and the cell. The electric signal is directly
fed into a computer; the digitized signal is recorded on the computer hard disk and
216 J. J. Zajac

Fig. 6.7 Schematic representation of the sample preparation stage in immersional and wetting
calorimetry experiments: (a) solid sample enclosed in the glass bulb, (b) sample evacuation and
pre-coverage, (c) sealing the end of the tube, (d) assembling

then processed using special software. To control the heat evolvement (or absorption)
during each calorimetric run, a pen recorder may be used in analog recording of the
signal; here, the vertical pen deflection
l perpendicular to the direction of feed of
the recording chart is proportional to the temperature difference θ . The scheme of the
calorimetric cell is shown in Fig. 6.8, together with a trace showing a representative
thermal profile for immersion.
According to the theoretical equation of Tian [56–59] for a conduction calorimeter
working under ideal conditions (e.g., the thermal delay between the temperature
change θ in the calorimetric cell and the response of the sensor is to be neglected),
the total heat effect occurring in the calorimetric cell during the time of experiment
tex p is given by the following expression:

texp
texp

λ 
2
Q exp = P(t)dt =
dt + d
(6.23)
g g
1
0 0

where P(t) is the heat generation (absorption) rate at time t, λ is the thermal con-
ductivity of conducting surface separating the calorimetric cell and the calorimetric
bloc,  is the heat capacity constant of the calorimetric cell with its content,
is the
voltage signal produced by the sensor at time t, and g is the proportional constant in
the relation
= g · θ ;
1 and
2 denote the initial (t = 0) and final (t = tex p ) sensor
indications, respectively. The first term on the right hand side of Eq. 6.23 represents
the heat exchange between the bloc and the cell monitored by the sensor during the
6 Calorimetry at the Solid–Liquid Interface 217

Fig. 6.8 Scheme of the calorimetric cell in a Tian-Calvet type differential calorimeter and a trace
showing a representative thermal profile for immersion. Q ex p is the overall thermal effect recorded

experiment, whereas the second term corresponds to the temperature rise (decrease)
in the calorimetric cell. When the duration of the experiment tex p is chosen such that

1 =
2 (i.e., the sensor signal returns to the baseline), this second term is always
equal to zero and the total heat effect Q ex p is determined by integrating the sensor
record
=
(t) from
1 to
2 :

texp
texp
λ
Q exp =
dt = K
dt (6.24)
g
0 0

The calorimeter constant K = λg is evaluated during the calibration run. Calibration

of the area under each thermal peak in the thermogram
=
(t) is carried out by
dissipating a known amount of energy in the cell (heating through Joule effect). For
this purpose, a special calibration resistor is placed in a glass bulb and introduced into
the calorimetric cell under the conditions of real experiment. The operator decides
both the power dissipated in the resistor I 2 · R(I—current flowing through the resistor
and R—its resistance) and the duration of the calibration step tcal . The integration
of the resulting thermal peak allows the calibration constant K to be calculated as
follows:
Acal
K = 2 (6.25)
I · R · tcal

where Acal is the area under the calibration peak.

The standard operating procedure for heat measurement is as follows [49, 54]. The
glass bulb containing the solid sample after the sample evacuation (and pre-coverage)
stage is introduced into the calorimetric cell, which has been previously filled with
218 J. J. Zajac

the immersional liquid (usually about 15cc). A large part of the glass rod remains
outside when the cell is closed. Complete gas-tightnest of the cell is ensured by a
special toric seal placed around the rod. At a given temperature, there is still some
vapour of the immersional liquid occupying the dead volume inside the calorimetric
cell. Then the calorimeter is left overnight to come to thermal equilibrium, giving a
steady baseline on the recorder.
After attaining thermal equilibrium, the glass rod is pushed gently down and the
glass tail of the bulb is broken against the bottom side of the cell. The immersional
liquid penetrates into the bulb and comes into contact with the solid sample. Thermal
effects accompanying the related exothermic and endothermic phenomena induce
changes in the temperature inside the calorimetric cell and the concomitant heat flux
between the bloc and the cell. The global thermal effect is recorded as a thermal peak
Aex p which can be integrated and transformed to the heat quantity Qex p (Fig. 6.8)
making use of the calibration constant K:

Q exp = K · Aexp (6.26)

To extract the net enthalpy of immersion

imm H or net enthalpy of wetting
W H ,
several correction terms have to be subtracted from Qex p . These correction terms are
related to (i) breaking the tail of the glass bulb inside the calorimetric cell, (ii) changes
in the dead volume of the calorimetric cell and the bulb during the experiment, (iii)
evaporation of the immersional liquid and condensation of its vapour, (iv) decrease
in the adsorbent mass during the evacuation step. They may be determined in a
blank test recording. Nevertheless, only two of them were found to give a noticeable
contribution to the total heat effect [43, 54].
When a given volume of the immersional liquid enters the bulb, the liquid level
in the cell is lowered and there is some evaporation of the liquid to equilibrate the
vapour pressure in the dead volume. This phenomenon yields an endothermic effect
which is proportional to the dead volume V0 of the bulb:

Q cor = −V0 ·
vap h (6.27)

where
vap h is the evaporation enthalpy per unit volume measured in a blank run
with an empty bulb (without a solid sample). The values of
vap h obtained for some
immersional liquids are given in Table 6.3.
The second correction term of great importance concerns the loss of the adsorbent
mass during the evacuation step [43]. Prior to immersion and wetting measurements,
the solid sample is subjected to thermal treatment in the glass bulb under vacuum. The
total mass of the sample certainly decreases, depending on the amount of liquid and
vapour impurities pre-adsorbed on its surface. The present experimental procedure
does not allow the actual mass to be determined precisely because the bulb containing
the sample cannot be weighed at the end of the outgassing process. In consequence,
the measured enthalpy values may be underestimated, especially if the weight loss
is significant (e.g., hydrophilic solids with a high specific surface area may contain
much water vapour adsorbed on the surface). The most reliable way of quantifying
6 Calorimetry at the Solid–Liquid Interface 219

Table 6.3 Enthalpies of

Immersional liquid
vap h(mJ cm−3 )
evaporation per unit volume

vap h for selected n-heptane 126

immersional liquids, as Isooctane 59
measured in a blank Chloroforme 254
calorimety run with an empty Benzene 249
bulb [60] Water 78
Formamide 24
l-butanol 160

the mass decrease is to carry out the outgassing procedure separately under the same
conditions as those used in the calorimetry experiment (the outgassed sample may
be weighed and the mass difference calculated).
Finally, the enthalpy change upon immersion of a solid in a given liquid per unit
surface area of the solid is calculated as follows:

imm H 1  

imm H ∗ = =− Q exp − V0 ·
vap h , in mJ m−2 (6.28)
AS L mS · S

where m S and S are the mass and the specific surface area of the “dry” solid sample,
respectively.

6.3.3 Hydrophilic-Hydrophobic Series and Harkins-Jura

Method

Using approximation Eq. 6.18c to express the interfacial enthalpy HSs L in terms of
the apolar (LW), Lewis acid (A) and Lewis base (B) components of the surface
enthalpy for the solid and immersional liquid, the final explicit form for the enthalpy
of immersion is as follows [42–44]:
 

imm H = HLG
s
−2· LW · H LW +
HS0 LG
A · HB +
HS0 LG
B · HA
HS0 LG (6.29)

LW , H A , H B
Provided that the surface enthalpy components for the liquids, i.e., HLG LG LG
are known, the apolar (HS0 ), Lewis acid (HS0 ), and Lewis base (HS0 ) contributions
L W A B

to the surface enthalpy of solids can be derived from measurements of the enthalpy
of immersion. At least one apolar liquid and two polar liquids must be used. The
values of HLGL W , H A , H B for selected immersional liquids, i.e., apolar n-heptane,
LG LG
bifunctional water, and monofunctional basic formamide, are given in Table 6.4.
The examples of the use of apolar n-heptane, bifunctional water, and monofunc-
tional basic formamide to study the surface hydrophobic-hydrophilic character of
several solids are shown in Figs. 6.9, 6.10, and 6.11 [38, 43, 44, 60]. n-heptane
220 J. J. Zajac

Table 6.4 Surface enthalpy components for n-heptane, water and formamide [43]
Liquid LW
HLG A
HLG B
HLG s
HLG
mJ m−2 mJ m−2 mJ m−2 mJ m−2
n-heptane 54.5 0 0 54.5
Water 35.0 41.5 41.5 118.0
Formamide 55.6 3.25 56.4 82.6

Fig. 6.9 Enthalpy of immersion per unit surface area of the solid
imm H ∗ for several solid materials
in n-heptane

and formamide (Merck HPLC grade materials with purity exceeding 99 %), were
additionally dried with 3A zeolite molecular sieves. This purification procedure to
remove even traces of water from organic solvents is of great importance and strongly
recommended for all immersion experiments. Water was deionised and purified with
a Millipore Super Q System.
The immersion experiment was repeated three times for each solid sample and
the average value taken. The amount of immersional liquid penetrating into the glass
bulb upon immersion was determined by weighing the bulb after every run (to eval-
uate the correction term 6.27). Reproducibility of the calorimetry measurement in
regard to the enthalpy of immersion expressed by weight of the solid was within
3 %. To obtain the enthalpy of immersion per unit surface area of the solid
imm H ∗ ,
6 Calorimetry at the Solid–Liquid Interface 221

Fig. 6.10 Apolar (LW) and polar (AB) contributions to the surface enthalpy for a series of solids,
as derived from measurements of the enthalpy of immersion
imm H ∗ of each solid in n-heptane,
water and formamide

the experimental enthalpy values were divided by the corresponding BET specific
surface areas (the adsorption model of Brunauer, Emmet, and Teller applied to the
experimental results of gaseous nitrogen adsorption at 77 K taking a cross sectional
area of 0.162 nm2 per N2 molecule), i.e., S = S B E T . Finally, each solid was sepa-
rately outgassed under the same conditions as those used in the immersion experiment
with the purpose of quantifying the mass of dried sample. Then the immersion data
were corrected for the mass loss during outgassing (following Eq. 6.28). In conse-
quence of all these additional steps, the uncertainty in the enthalpy determination
increased, but was, on average, better than 5 %.
Figure 6.9 shows the experimental values of
imm H ∗ for one series of solids
immersed in n-heptane. Here the interfacial phenomenon involves only van der Waals
interactions between the immersional liquid and the solid, irrespective of the actual
surface hydrophobic-hydrophilic balance (SHB) of the latter. Based on the enthalpy
results obtained, it is possible to classify these solids with respect to hydrophobic
character of their surface.
According to the criteria of hydrophobicity given in Sect. 6.2.4, the most hydropho-
bic substance has the lowest HS0 L W (more precisely γ L W ) surface parameter. There-
S0
fore, its enthalpy of immersion in n-heptane per unit area of the solid should
have the smallest value. Among the solids presented in Fig. 6.9, only sulphur and
222 J. J. Zajac

Fig. 6.11 PLewis acid (A) and Lewis base (B) contributions to the surface enthalpy for a series
of solids, as derived from measurements of the enthalpy of immersion
imm H ∗ of each solid in
n-heptane, water and formamide

polytetrafluoroethylene (PTFE) appear clearly at the hydrophobic end. Such min-

eral oxides as clay minerals with a lamellar structure (kaolinite, illite, talc, chlorite),
precipitated amorphous SiO2 (silica), crystalline SiO2 (quartz), amorphous Al2 O3
(alumina), or crystalline aluminosilicates (zeolite) are known to be more or less
hydrophilic (commonly, the presence of numerous functional groups with a polar
character in the surface of these materials is advanced as a typical argument). The
comparison of the
imm H ∗ values reported in Fig. 6.9 indicates that the overall inten-
sity of Lifshitz-Van der Waals interactions between apolar n-heptane and a unit area
of the solid surface is not the same for various solids. This means that all microscop-
ically unsaturated structures which may be encountered in solid surfaces differ also
in “apolar character” and, consequently, the apolar (LW) component to the surface
enthalpy HS0 L W is a solid-dependent parameter. Surprisingly, non-porous graphitized

carbon black (Graphon), regarded as a weakly hydrophilic solid, yield

imm H ∗
greater than those of kaolinite, illite, or quartz which certainly possess more polar
groups per unit surface area. Of course, it may be argued that graphitic basal planes
in the surface of Graphon act as electron pair donors, thereby participating in π -EPD/
π -EPA or π -complexation interactions with the foreign molecules. Nevertheless, it
is more reasonable to evaluate both the apolar (LW) and the polar (AB) components
to the surface enthalpy HS0 s in order to compare the differences in the SHB among

various solids.
6 Calorimetry at the Solid–Liquid Interface 223

It is worth noting that any arrangement of solid surfaces by hydrophobic character,

as well as any ordering or ranking of solids in regard to their SHB may be used only
for comparative purposes and do not give the “absolute ranking position” of a given
solid within the class of hydrophobic and hydrophilic substances.
Figure 6.10 shows examples of several hydrophobic and hydrophilic solids,
together with their HS0 L W and H AB parameters obtained on the basis of Eq. 6.29.
S0
SiC14, SiAl32C14, and SiAl8C14 are mesoporous silica-based materials of the
MCM-41 type prepared by a surfactant-assisted synthesis and doping with aluminium
(e.g., abbreviation “SiAl32” refers to a molar Si-to-Al ratio of 32). Only PTFE is
a clearly hydrophobic substance since it has low HS0 L W and zero H AB component.
S0
Amorphous fumed SiO2 (Cab-O-Sil) and mesoporous MCM-41 silica or aluminosli-
cates can be reckoned among low-energy solids possessing surfaces with a weakly
hydrophilic character. For high-energy solids with surface enthalpy HS0 s greater than
−2 AB L W
500 mJ m , the ratio between HS0 and HS0 varies from 7 (illite) to 0.2 (SiC);
alumina is characterized by an intermediate SHB value of 0.8.
Figure 6.11 illustrates the Lewis acid-base character of the solid surfaces presented
in Fig. 6.10. The Lewis acid HS0 A and Lewis base H B components are related to the
S0
polar (AB) contribution to the surface enthalpy of each solid: HS0 AB = 2 H A · H B .
S0 S0
Crystalline SiO2 (quartz) and, to a smaller extent, amorphous precipitated SiO2
(silica) provide strongly acidic surfaces, whereas the surfaces of MCM-41 meso-
porous silicas and Cab-O-Sil have a predominantly basic character.
The Harkins-Jura method for estimating the specific surface area S H J of a solid
available to a given immersional liquid is based on Eq. 6.22 [20, 54, 55]. In practice,
water is by far the most frequently used liquid because of the small size of its
molecules (the van der Waals diameter of a water molecule is about 0.28 nm). The
specific surface area of a hydrophilic solid is thus calculated as follows:

W Hm
W h m
W h m
SH J = − =− s =− (6.30)
m S · HLG
s HLG 0.118

where
W h m , expressed in J g−1 , denotes the minimum value of the specific enthalpy
of immersional wetting, i.e., wetting enthalpy per unit mass of the solid sample

W h m =
Wm SHm . This enthalpy value is determined from the plot of
W h against
p
the relative pressure psat at which surface pre-coverage with water vapour has been
carried out (Fig. 6.12).
The application of the above procedure is limited only to non-microporous mate-
rials, for which the solid surface area available to water is equal to the area of the
interface between the adsorbed water film on the surface and the equilibrium water
vapour phase. It is always necessary to check the condition of the ideal wetting.
Water does not spread on hydrophobic surfaces, like the surface of the Graphon, and
the value of
W H decreases (becomes more negative) as the surface adsorbs more
water vapour [20]. This proves that the empty surface has less affinity for water than
has the surface of the adsorbed water. Equations 6.22 and 6.30 cannot be applied to
such a system.
224 J. J. Zajac

Fig. 6.12 Specific enthalpy of immersional wetting

W h (taken with the opposite sign) in water for
silica SIFRACO C-600 pre-covered with water vapour as a function of the relative vapour pressure
[60]. The resulting HJ specific surface area is also given (S H J = 5.5 m2 g−1 )

6.4 Enthalpy Changes Accompanying Competitive

Adsorption from Dilute Solution

The competitive aspect of the phenomenon, nowadays commonly accepted, shows up

most clearly in adsorption from concentrated solutions (e.g., mixtures of completely
miscible liquids) [6]. In dilute solution, only a very detailed analysis of the adsorp-
tion system may reveal that the phenomenon is indeed competitive. In this context,
calorimetry has proven very useful in studying the competitivity of adsorption from
dilute solution onto solids.
The driving force of competitive adsorption from binary solution onto solids is a
macroscopic outcome of intermolecular forces belonging to the following categories:
(i) interactions of the solute and the solvent with the solid surface, (ii) interactions
among the solution components in the interfacial region, and (iii) solute—solvent,
solvent—solvent, and solute—solute interactions in the bulk solution. In numerous
adsorption systems, a considerable enhancement of the amount adsorbed is often the
result of co-operative effects involving various interaction types. The most spectacu-
lar examples are related to the adsorption of amphiphilic substances (e.g., surfactants
and polymers) that exhibit reduced water solubility [61, 62]. The general behaviour
of amphiphilic molecules in aqueous solution reflects the opposing tendencies of
the hydrophobic portion of the molecule to escape from the aqueous environment
while the hydrophilic moiety tends to remain immersed in the water. Depending on
the surface hydrophilic-hydrophobic character of the solid particules immersed in
the solution, such molecules can be compelled to accumulate in an oriented fashion
at the Solid-Solution interface or to associate between themselves within the solu-
tion (i.e., micellization). Aggregation of surfactant units both at the Solid-Solution
6 Calorimetry at the Solid–Liquid Interface 225

interface and in the bulk solution is mainly governed by hydrophobic bonding, which
refers to the apparent attraction between hydrophobic moieties being much stronger
in water than it is in non-aqueous media [26, 61, 63]. Hydrophobic bonding is a typ-
ical example of the so-called thermodynamic interactions having both mechanical
and entropic contributions.
Adsorption experiment may be carried out in two different ways: as (1) immersion
of a dry adsorbent in a binary solution of a given composition or (2) displacement
of the solvent from the solid-solvent interface by the adsorbing solute supplied in a
stock solution [6, 64, 65]. Additionally, the effects of displacement are determined
either in a batch displacement or in a flow displacement experiment. Depending
on the experimental procedure applied, the accompanying changes of macroscopic
properties differ to a great extent. The amount adsorbed and enthalpy change on
adsorption can be precisely defined in operational terms by considering models of
the immersion and displacement experiments.

6.4.1 Thermal Properties of Dilute Solutions

A dilute solution typically means a solution containing no more than about 10−2
mol l−1 of solute. For the solvent, a convenient reference state is its own pure liquid
state at 1 bar. However, a more useful reference state for the solute, including aqueous
solutions of electrolytes and non-electrolytes, is that of infinite dilution in the solvent
[66, 67].
The general form of the chemical potential of a real solute in a binary solution of
non-electrolytes is therefore [66, 67]
m 
μ2 = μ∗2 + RT ln
2
· f 2 , f 2 → 1 as m 2 → 0 (6.31a)
 c1 
∗ 2
μ2 = μ2 + RT ln · f 2 , f 2 → 1 as c2 → 0 (6.31b)
1

where μ∗2 is the standard chemical potential of the solute; m 2 is the molality of this
component and c2 the corresponding molar concentration. The unity on the right-
hand side is written to remind the reader that the molality (the molar concentration)
is expressed in mol kg−1 (mol L−1 ), whereas the activity is dimensionless. The
employment of concentration units is often discouraged because the concentration
of a given solution varies with temperature, while the molality is independent of
temperature. The latter also allows the composition determination to be achieved
with greater precision, since any solution may be prepared by weighing the solute
and solvent, or the stock solution and solvent (dilution of the stock solution).
The interactions between ions contained in an ionic solution are so strong that
the solution approaches ideality in the sense of obeying Henry’s law only at very
low values of total ion concentration, usually less than 10−3 mol kg−1 [68, 69]. In
precise considerations, ion activities must be used. In the case of strong electrolytes
226 J. J. Zajac

in aqueous solution, the total Gibbs energy (and any other extensive property) of the
solute is the sum of the partial molar Gibbs energies for the individual ions produced
by the solute. Suppose that a strong electrolyte M p Xq dissociates into ions according
to the following reaction

M p Xq = p · Mz+ + q · Xz− , p · z+ = q · z− (6.32)

where p and q are the number of ions of M and X type, respectively; z + and z −
represent the valencies of the respective ions (positive number for cations and negative
number for anions). The chemical potential of the solute j in a real solution may be
written as [68, 69]

μ j = μ∗j + ( p + q)RT ln m ± + ( p + q)RT ln f ± (6.33a)

where the mean ionic molality m ± and the mean ionic activity coefficient f ± are
defined as follows:
  1   1
m ± = (m + ) p (m − )q p+q and f ± = ( f + ) p ( f − )q p+q (6.33b)

Now both types of ion share equal responsibility for the non-ideality.
Strong and long-range Coulombic forces acting between ions are primarily
responsible for the departures from ideality (the activity coefficients are lowered)
and dominate all other contributions. The effect has been evaluated in the Debye-
Hückel theory and there exist several equations, which are useful in estimating the
mean activity coefficient [68, 69]. The latter is related to the ionic strength of the
solution:
1 2
I = zj · mj (6.34)
2
j

where the sum extends over all ions present in the solution; m j is the molality of
the jth ion. For example, the value of f ± can be calculated from the Debye-Hückel
limiting law 
I
log f ± = −A |z + · z − | · (6.35a)
1

where A is the dimensionless constant characteristic of the solvent at the specified

temperature and pressure (A = 0.5085 for an aqueous solution at 298.15 K [68, 69]).
The approximation (6.35a) is in good agreement with the experiment at very
low molalities: less than 0.01–0.001 mol kg−1 , depending on charge type (the range
of validity becomes narrower when divalent or multivalent ions are present in the
solution). Nevertheless, when the departures from the experimental results are large,
the activity coefficient may be estimated from the extended Debye-Hückel law
6 Calorimetry at the Solid–Liquid Interface 227

A |z + · z − | · 1I
log f ± = − (6.35b)
1 + B · 1I

where B is an adjustable empirical parameter (which involves numerical factors, the

dielectric constant, the temperature and the mean distance of nearest approach of the
ions [68, 69]).
In thermodynamics of solutions, the partial molal quantity and the apparent molal
quantity for a component are usually defined [66]. In the properties of enthalpy,
one deals with quantities which cannot be measured in an absolute sense. It is thus
necessary to use a reference state from which to make the evaluations. Since the state
of infinite dilution of the solute in the solvent is taken as reference, the specification
of the composition of the solution by means of the molality m 2 of the solute is
consequently most convenient for the calculation of enthalpy.
If the enthalpy H of a two-component solution is expressed as a function of the
composition, at constant pressure and temperature, the partial molal enthalpy h 2
for the solute is defined as the rate of change of the enthalpy H with change in the
number of moles of this component, with the number of moles of the solvent being
held constant; that is  
∂H
h2 = (6.36)
∂n 2 T,P,n 1

The basic partial molal equation for the total enthalpy of the binary solution is

2
H= n j · h j , T, P = const (6.37)
j=1

where n j is the number of moles of the jth component in the solution.

The partial molal enthalpy of the solute in a solution may not always be conve-
niently measured experimentally with the required precision. Usually, the apparent
molal property for the solute is most used in connection with dilute solutions. For a
solution of two components, the apparent molal enthalpy of the solute  H is

H
H = , T, P = const (6.38)
n2

where
H is the enthalpy change produced when the solution is formed by adding
n 2 moles of solute to n 1 moles of pure solvent.
The total enthalpy of the solution may be conveniently expressed in terms of the
apparent molal enthalpy of the solute as

M1
H= m 2  H + M1 h ∗1 (6.39)
103
228 J. J. Zajac

where M1 is the total mass of the solvent in the solution and h ∗1 is the specific enthalpy
(per unit mass) of the pure solvent.
In the infinitely dilute solution, the value of the molal enthalpy for the solute is
identical with the corresponding value of the partial molal enthalpy, namely

 H (∞) = h 2 (∞) at infinite dilution (6.40)

For any other solution, the partial molal enthalpy of the solute referred to the state
of infinite dilution may be calculated if the apparent molal enthalpy is known as a
function of the composition. In such a case,

d [ H −  H (∞)]
h 2 − h 2 (∞) =  H −  H (∞) + m 2 (6.41)
dm 2

It frequently happens that the apparent molal enthalpy is expressible as an appropriate

function of the composition in order to obtain a curve which does not depart greatly
from a simple function and which has not too great a curvature over the given range.
For aqueous solutions of strong electrolytes, the value of  H is sometimes linear
with the square root of the molality, over a given range of molality [68]. Converting
√
Eq. 6.41 to the slope of  H against m 2 , one obtains

1 √ d [ H −  H (∞)]
h 2 − h 2 (∞) =  H −  H (∞) + m2 √  (6.42)
2 d m2

The difference  H −  H (∞), called the relative apparent molal enthalpy, is the
quantity most readily evaluated from calorimetric measurements of enthalpies of
dilution. For the following reaction of dilution:
 
M10 0  
m 2 × solute + M1
0
× solvent (liq) + ∞ − M 0
1 × solvent(liq)
103
 
M10 0
→ m × solute + ∞ × solvent (liq) (6.43)
103 2

the change in the enthalpy is equal to

  M0

dil H m 02 → ∞ = − 13 m 02 [ H −  H (∞)] (6.44)
10

where M10 denotes the initial mass of the solvent and m 02 the initial molality of the
solute. Values of the relative apparent molal enthalpy may be obtained by measuring
enthalpy changes
dil H for the dilution of a given stock solution of molality m 02 to
a series of different molalities m 2 ,
6 Calorimetry at the Solid–Liquid Interface 229

  M0    

dil H m 02 → m 2 = − 13 m 02  H m 02 −  H (m 2 ) (6.45)
10
and suitably extrapolating the curve
dil H =
dil H (m 2 ) to infinite dilution.

6.4.2 Macroscopic Description of Competitive Adsorption

Compared to interfacial phenomena occurring at the boundary between a solid and a

pure liquid, Coulombic forces constitute an additional type of interaction that should
be taken into account when analysing the adsorption phenomenon at an electrified
interface from solution containing free charges (ions and electrons) and associated
charges (dipolar molecules and polarised atoms). In such systems, there is a varia-
tion in the charge density across the interfacial region and an electric double layer
forms [70–72] (the Gouy-Chapman-Stern-Grahame model of the EDL is shown in
Fig. 6.13).
Transfer of the individual ionic species between the interface and the solution
leads to a thermodynamic equilibrium at which the interface is electrically neutral
as a whole (and so is the equilibrium bulk solution).
Based on isothermal reversible work done in transferring one mole of ions of the
solute 2 from infinity (in vacuum) to a given part in the interfacial region which has
a non-zero average charge and where the electrostatic potential is, it is possible to
define the so-called electrochemical potential of component 2 [70]

μ2 = μ2 + z 2 · F ·  , T, P = const (6.46)

where μ2 is the chemical potential of the solute, z 2 is the valency of an ion of the
solute, and F is the Faraday constant (F is a product of the Avogadro number L and
the elementary charge e). If the transferred species carries no net charge, i.e., z 2 = 0,
the electrochemical potential μ2 becomes simply the chemical potential μ2 . The new
potential μ2 is now the quantity which must have the same value everywhere in the
system for thermodynamic equilibrium to be established at constant temperature and
pressure [70].
According to the surface phase model (cf., Sect. 6.2.1), when the concentrations
of the components of the binary solution β in the solid phase α are to be neglected,
i.e., cαj = 0 (j =1,2), the interfacial concentration sj of component j is given by
[6, 8]
n sj 1  β

sj ≡ = τ s · csj = n j − c j V β , (j = 1, 2) (6.47)
A A
β
where c j and csj are the concentrations (in moles per unit volume of solution) of the
jth component in the bulk solution β and in the surface phase of thickness τ s and
surface area A; V β represents the volume of the bulk phase β in the surface phase
model. The subscripts 1 and 2 denote the solvent and the solute, respectively; this
notation will be used throughout the present chapter.
230 J. J. Zajac

Fig. 6.13 Schematic representation of the structure of the electric double layer according to the
Gouy-Chapman-Stern-Grahame model and variations of the interfacial electric potential  across
the interfacial region. By convention, the potential  of any part of the electrified interface is defined
with respect to the solution at infinite distance where the average charge is zero. The distance x is
measured in reference to the position of a plane containing solid surface charge σ0 . The solution
part of the EDL with an excess of counter-ions is composed of two distinct regions: a thin Stern
layer (next to the surface charge plane) including specifically adsorbed (Coulombic and other non-
electrostatic forces) counter-ions which may be partially dehydrated and an extended diffuse layer
where the hydrated counter-ions are distributed non-uniformly according to the combined action of
electrostatic forces and random thermal motion. Electroneutrality of the charged interface requires
that σ0 + σβ + σd = 0; σβ and σd represent the charge densities of the Stern and diffuse layers,
respectively [70]

The case of sj > 0 defines the preferential adsorption of component j at the

Solid-Liquid interface. In a given adsorption system, the interfacial concentration
β β
sj is a function of both the actual composition of the bulk solution (i.e., c1 and c2 )
and the relative affinities of the components for the interface (chiefly, for the solid
surface). Contrary to the case of adsorption at the Solid-Gas interface, the interfacial
region is always completely filled with molecules of solvent and solute. Further
6 Calorimetry at the Solid–Liquid Interface 231

consideration will be based on the preferential adsorption of solute against solvent,

taking the interface between the solid and the pure solvent as a starting point for
thermodynamic treatment of competitive adsorption. In such a case, the adsorption
equilibrium can be represented schematically as follows:

β β
r · n2 + n1s  r · n2s + n1 (6.48)

where r is the molar ratio of displacement introduced to account for unequal mole-
cular sizes of both the components. It is worth noting that the same formalism may
be applied in the treatment of adsorption at the interface between a binary liquid
solution and a gas phase.
When the pure liquid phase is replaced by a binary liquid solution β, the interfacial
tensions γ S L and γ LG will be lowered owing to the preferential adsorption of the
solute at the Solid-Liquid and Liquid-Gas interface. The relationship between the
extent of adsorption and the resulting interfacial (surface) tension change (T, P =
const) is given by the following equation developed by Guggenheim [6, 8]
  β
   β

c2 n2
dγ = − 2s − 1s β
dμ2 = − 2s − 1s β
dμ2
c1 n1
   
M1s β
= − 2s − · m j dμ2 (6.49)
103 A

where μ2 is the chemical potential of solute in the bulk solution β: at equilibrium,

μ2 is the same throughout the system; when expressing the solution composition
β
in terms of the molality m 2 (in moles per kilogram of solvent), M1s is the mass of
solvent in the surface phase. It is worth noting that Eq. 6.49 is formally analogous to
the Gibbs adsorption isotherm for a binary solution [6, 8].

6.4.3 Competitive Adsorption Measurements

The amount of solute molecules preferentially adsorbed from dilute solution onto a
given solid can be measured in a separate adsorption experiment, independently of
the calorimetry measurement. In the case of the titration calorimetry procedure, this
is even the only possibility to determine the amount adsorbed after each injection step
and subsequently calculate the differential molar enthalpy of adsorption. The main
difficulty here, contributing to a significant uncertainty of the experimental result,
is related to the necessity of reproducing strictly the same experimental conditions
in both types of experiment (i.e., the same solid surface-to-solution volume ratio,
evolution of the pH and ionic strength in the equilibrium bulk solution, charging
behaviour of the solid surface, etc.).
The quantity of adsorption is usually measured by means of the solution depletion
method [6] in glass stoppered tubes or flasks (Fig. 6.14). A known mass of the solid
232 J. J. Zajac

Fig. 6.14 Schematic representation of the solution depletion technique. The solid particles are
separated from the supernatant solution by centrifugation or filtration. The way of calculating the
quantity of adsorption depends on whether the solution composition is expressed in terms of molality
β β
m 2 or the molar concentration C2

sample and a certain amount of dilute binary solution of given composition are put
into each tube or flask, and the mixture is shaken at a constant temperature (e.g., in a
thermostated box) for a period of time necessary to attain the adsorption equilibrium.
Then the solid particles are separated from the supernatant by centrifugation or
filtration (special attention should be paid during fitration to avoid the retention of the
solute on the filtration membrane). After centrifugation when the solid particles have
been precipitated at the bottom of the tube or flask, a sample of the supernatant may be
collected by using a special syringe with a long needle. The equilibrium composition
of the supernatant is determined by referring to the appropriate analytical technique
(UV spectroscopy, refractometry, total carbon analysis, etc).
The amount n s2 of solute adsorbed at the Solid-Liquid interface is calculated by
means of the following formulas:

n s2 V0  0 β
 n s2 M0  0 β

= C2 − C2 or = 3 1 m2 − m2 (6.50)
mS mS mS 10 · m S

where m S is the mass of the solid sample in a given tube or flask; C20 and m 02 are
respectively the molarity and the molality of the initial solution (before adsorption);
β β
C2 and m 2 are respectively the molarity and the molality of the supernatant solution
β (after the attainment of adsorption equilibrium); V 0 denotes the volume of the
solution put initially into the tube or flask and M10 is the total mass of the solvent
in the heterogeneous system. When the specific surface area of the solid S B E T or
S H J is known, it is even possible to calculate the interfacial concentration 2s of the
solute. It should be realised that expressions 6.50 hold only for dilute binary solutions
containing a solute that is preferentially adsorbed at the Solid-Liquid interface.
For a given adsorption system, the amount of solute adsorbed at equilibrium
depends on the temperature T, the pressure P, and the composition of the equilibrium
solution phase β. The experimental results of adsorption measurements are usually
ns
reported in the form of individual adsorption isotherms showing the quantity m2S or
6 Calorimetry at the Solid–Liquid Interface 233

2s as a function of the solution composition:

n s2 ns  β  n s2 ns  β 
= 2 C2 or = 2 m 2 , T, P = const (6.51a)
mS mS mS m
   S 
β β
2 = 2 C 2
s s
or 2 = 2 m 2 , T, P = const
s s
(6.51b)

Figure 6.15 shows three selected shapes of adsorption isotherms at the Solid-Liquid
interface for various charged and uncharged solutes exhibiting a limited solubility
in the solvent employed. In general, typical adsorption curves present the amount
adsorbed as a smooth, monotonically increasing function of the solute content in the
equilibrium bulk solution. Many adsorption isotherms end up in an upper composition
range at a constant amount adsorbed (a plateau of the isotherm).
The appearance of an adsorption plateau region at equilibrium concentrations
(molalities) approaching the solubility limit indicates that the phenomenon involves
only single solute species that are individually dissolved in the solvent. When the
adsorption plateau is observed at lower concentrations (molalities), it is usually
argued that surface sites of a given type have been saturated by the adsorbing solute
species.
The experimental adsorption isotherms may be plotted on different scales, thereby
allowing a more detailed analysis of the subsequent adsorption stages to be made.
When the adsorption of homologous substances onto the same solid is due only to
their different solubilities in the solvent, the use of a reduced concentration (molality)

Fig. 6.15 Adsorption isotherms of (a) zwitterionic (C12N1C), and cationic (BDDA+ ) surfactants
from aqueous solutions onto non-porous negatively charged silica at 298 K [73, 74], (b) buckmin-
sterfullerene C60 from its tollune solutions onto mesoporous activated carbon at 293 K [75]; C12N1C
and BDDA+ denote (dodecyldimethylammonio)ethanoate and benzyldimethyldodecylammonium
cation, respectively
234 J. J. Zajac

scale to represent the corresponding isotherms results in a single adsorption curve.

Any changes in the adsorption mechanism with increasing bulk concentration (molal-
ity) show up most clearly in double logarithmic (log-log) plots. The transitions
between such adsorption regions are obviously more gradual in reality than they
seem to be on a log-log scale. The plateau adsorption region markedly manifests
itself only on a double linear (lin-lin) scale. In the linear-logarithmic (lin-log) plot,
it is difficult to observe the effects of surface heterogeneity and co-operative adsorp-
tion at low bulk concentrations (molalities) but the top parts of the isotherm become
clearer.
By analogy to Eq. 6.16, the surface pressure per unit surface area of the Solid–
Liquid interface may be determined as follows:
∗

C2
    RT
π S L C2∗ = γ S L1 − γ S L C2∗ = n s2 d ln C2 (6.52)
mS · S
C2 =0

where γ S L1 is the interfacial tension of the solid in equilibrium with the pure solvent;
S is the specific surface area of the solid (i.e., S = S B E T or S = S H J ); C2∗ denotes
the equilibrium bulk concentration taken into consideration.

6.4.4 Immersion in Dilute Solutions

The solution depletion procedure can be generalised to construct a simplified model

of immersion experiment [64, 65]. This model is very useful to define the measurable
enthalpy quantities, which does not necessarily mean that they can be easily obtained
experimentally. The immersion process is schematically represented in Fig. 6.16.
Before immersion the system separately contains an outgassed solid sample of
a given mass m S and a given surface area S in vacuum and a binary solution of
molality m 02 consisting M10 grams of the solvent, all at constant temperature T and
pressure P. Then the solid is immersed in the solution under conditions of constant T,
P. The phenomenon of adsorption induces an uneven partition of solvent and solute
molecules between the solid-solution interface and the bulk solution. After attaining
the adsorption equilibrium, the molality of the solute in the bulk of the solution (suf-
β
ficiently far from the solid-liquid interface) becomes equal to m 2 . Since absorption,
dissolution and swelling effects are not taken into consideration, the adsorbent is
considered as thermodynamically inert, i.e., its mass, specific surface area and bulk
phase properties do not change during adsorption. Furthermore, the adsorption of the
solute at the Solution-Gas interface and the relative surface enthalpy for this interface
are omitted for convenience: changes of both effects during immersion in an excess
of dilute solution are negligible small.
The balance of enthalpy in the initial (init) and final (fin) states of the model
system can be written as follows (cf. Sect. 6.4.1):
6 Calorimetry at the Solid–Liquid Interface 235

Fig. 6.16 Schematic representation of the immersion experiment (explanation of the symbols in
the text). The adsorption phenomenon is assumed not to affect the enthalpy of the bulk of solid,
m S · h S ; changes in the mass of the solvent are to be neglected when immersion is carried out in a
sufficient excess of dilute solution containing a preferentially adsorbed solute (i.e., M1 = M10 )

M10 0  
H init = m S · S · HS0
s
+ m S · h S + M10 · h ∗1 + m ·  H m 02 (6.53a)
103 2
M0 β  
β
H f in = m S · S · HSs L + m S · h S + 13 m 2 ·  H m 2 + M10 · h ∗1 (6.53b)
10

where HS0 s and H s are the interfacial enthalpies for solid in vacuum and for Solid-
SL
β
Liquid interface in equilibrium with a bulk solution of molality m 2 , respectively; h S
is the specific (per unit mass) enthalpy of the bulk of the solid phase;  H (m 2 ) is the
value of the apparent molal enthalpy of the solute in a binary solution corresponding
to the molality m 2 ; h ∗1 is the specific enthalpy (per unit mass) of the pure liquid
solvent. Equation (6.53b) has been developed based on the hypothesis that changes
in the mass of the solvent can be neglected during immersion in excess solution.
The total thermal effect of the immersion experiment is:
 

Hexp = H f in − H init = m S · S · HSs L − HS0
s

M0  β  
β
 
+ 13 m 2 ·  H m 2 − m 02 ·  H m 02 (6.54a)
10
Taking into account the reference state for the solute defined in Sect. 6.4.1 and using
Eqs. (6.40), (6.47), and (6.50), this expression may be transformed as follows:
236 J. J. Zajac
 

Hexp = m S · S · HSs L − HS0
s
− m S · S · 2s · h 2 (∞) +
    M0     
β β
− m S · S · 2s  H m 2 −  H (∞) − 13 m 02  H m 02 −  H m 2
10
(6.54b)

Since the adsorption of the solvent is here considered negligible (i.e.,M1 = M10 ),
the first two terms on the right hand side of Eq. (6.54b) represents the enthalpy of
immersion in a binary solution (1+2) with respect to the bulk solution at infinite
dilution taken as the reference state [64]; the third term corresponds to the enthalpy
β
of dilution of m S · S · 2s moles of the solute from molality m 2 to infinite dilution
in the solvent (cf., Eq. 6.44) and it is often incorporated into the previous enthalpy
β
contribution to give the enthalpy of immersion
imm H12 m 2 of the solid in excess
β
solution of molality m 2 [48, 76, 77]. The last term represents the dilution effects in
M10
the bulk solution: 103
m 02 moles of the solute are diluted from molality m 01 to molality
β
m 2 . Finally,    
β β

Hexp =
imm H12 m 2 +
dil H m 02 → m 2 (6.55)

When the dilution term is evaluated independently

  in appropriate dilution experi-
β
ment, the enthalpy of immersion
imm H12 m 2 can be determined experimentally
by means of the same calorimetry equipment as that used to measure the enthalpy of
immersion in a pure liquid (Sect. 6.3.2). However, great difficulties may be encoun-
tered when evaluating the usual correction terms in case of solutes which are volatile
or surface-active (the composition of the vapour occupying the dead volume of the
bulb is unknown). This method is also tedious. For systems containting electrified
interfaces, the effects of EDL formation additionally contribute to the complexity of
the adsorption phenomenon. Here the displacement experiment yields the enthalpy
data easier to interpret.

6.4.5 Model of Flow Calorimetry Experiment

Liquid-flow microcalorimetry is a reliable method to measure simultaneously the

enthalpy changes and amounts of adsorption under dynamic conditions. Calorime-
try experiments may be carried out in two different ways by following a pulse or
saturation operating mode [64, 78–83]. In the pulse mode, small aliquots of a stock
solution at a known concentration are injected into the carrier liquid (pure solvent)
flowing through the adsorbent bed placed inside the calorimetric cell. In this case,
the calorimetric system contains an additional loop injection facility (a manual injec-
tion valve with appropriate injection loops). The interpretation of the enthalpy data
obtained is straightforward only when the whole amount of the solute injected is
irreversibly adsorbed on the solid surface.
6 Calorimetry at the Solid–Liquid Interface 237

The saturation mode (continuous-flow method) is more frequently used. Here

changes in enthalpy and amount adsorbed of the solute correspond to the formation
of a Solid-Liquid interface being in thermal and material equilibrium with the per-
colating stock solution of a given composition. Repeated adsorption and desorption
cycles with the liquid phase in contact with the solid surface for a time required to
reach equilibrium can be used to assess reversibility of the phenomenon, and quan-
tify the reversible and irreversible adsorption components [79, 80]. In addition, the
same equipment allows probing for some active sites in the solid surface.
The physical meaning of thermodynamic quantities measured in the flow calorime-
try experiment, may be discussed based on a simplified model of the system
(Fig. 6.17). The model system is composed of three parts: (i) a reservoir R con-
taining a given volume of the stock solution of molality m 02 , (ii) a cell C with the
solid sample of mass m S , in contact with the solvent or the stock solution, (iii) a trap
T for the effluent [65, 78].
Initially, the outgassed solid sample is immersed in M1C 0 grams of pure solvent

and the total quantity of stock solution in the reservoir is given by the mass of solvent
M1R0 ; the trap is empty. Then the flow of stock solution from the reservoir is directed

to the cell under constant liquid-flow conditions. The temperature and pressure are
assumed to be uniform throughout the system and there is neither loss of energy
nor loss of matter between the reservoir, the cell and the trap. As the stock solution

Fig. 6.17 Schematic representation of the flow calorimetry experiment (explanation of the symbols
in the text). The model system is composed of three parts: R Reservoir, C Calorimetric cell, T Trap
238 J. J. Zajac

flows through the cell with the adsorbent, the boundary between the solution and
the solvent remains sharp, and the effects of adsorption and desorption change its
position according to the ideal chromatographic behaviour. In practice, this means
that the change of liquid composition should be sufficiently small for heat of mixing
contribution to be neglected. The final state is achieved just before the front of the
stock solution reaches the trap. At this point, the trap is filled with M1T grams of
pure solvent and the equilibrium molality of the solute within the cell (far from the
solid surface) is equal to m 02 . The composition of the stock solution in the reservoir
does not change but its amount decreases; the final mass of the solvent is M1R .
For the initial state, the amount of the solute n init
2 in the model system and the
total enthalpy of the system H init are given by (cf. Sect. 6.4.1):

M1R0
n init = m0 (6.56a)
2
103 2
H init = m S · S · HSs L1 + m S · h S + M1C
0
· h ∗1
M0 0  
+ 1R m ·  H m 02 + M1R0
· h ∗1 (6.56b)
103 2

where HSs L1 is the interfacial enthalpy for the solid-solvent interface; other quantities
have been introduced previously (cf., Eqs. 6.53a, 6.53b).
Applying the law of mass conservation to the solvent in the system, the corre-
sponding quantities in the final state may be written as follows:

0 + M0 − M
f in M1R 0 M1R 1C 1R − M1T 0
n2 = m S · S · 2s + 3
m2 + m2 (6.57a)
10 103
M1R 0  
H f in = m S · S · HSs L + m S · h S + m 2 ·  H m 02 + M1R · h ∗1
103
M 0 + M1C 0 −M  
1R − M1T 0
+ 1R 3
m 2 ·  H m 02
 10 
+ M1R
0
+ M1C
0
− M1R − M1T · h ∗1 (6.57b)

where HSs L refers to the interfacial enthalpy for the solid-solution interface in equi-
librium with a bulk solution of molality m 02 . The comparison of expressions (6.56b)
and (6.57b) indicates that the adsorbent is considered as thermodynamically inert,
i.e., its parameters m S , S, and h S do not change during adsorption.
The law of mass conservation applied to the solute gives rise to the following
relation:  
M1T − M1C 0
2 =
s
m0 (6.58a)
103 · m S · S 2

This relation shows that the quantity of adsorption for the solute may be directly
measured in the flow calorimetry experiment if the values of M1C 0 and M
1T are
known. When the solid bed in the cell contains a non-adsorbing solid, the difference
6 Calorimetry at the Solid–Liquid Interface 239

M1T − M1C 0 is equal to zero; the value of M 0 can thus be measured in the appropriate
1C
“blank” run. Otherwise, this difference depends on how much longer the solute is
retained by the adsorbent. In the model flow experiment, the concentration curves
obtained with a non-adsorbing and adsorbing solid represent obviously square pro-
files (Fig. 6.18a).
If the feed solution flow rate d pump is constant and the solvent migrates at the
same velocity throughout the whole system, the mass of solvent entering the trap is
directly proportional to the retention time τ A :

103 · d pump
M1T = · τA (6.58b)
103 + m 02 · Msolute

where Msolute is the molar mass of the solute. The amount of solute adsorbed onto
solid sample can be thus calculated from the corrected retention time, i.e., τ A − τ N A
(Fig. 6.18a). In a real calorimetry run, the shape of the m 2 vs. time plot depends on
the underlying equilibrium isotherm of solute adsorption, as well as diffusion and
mass-transfer kinetics. This results in a breakthrough curve, as shown in Fig. 6.18b.
The quantity of solute adsorption is calculated from the difference between the areas
over the breakthrough curves obtained with a non-adsorbing and adsorbing solid
[79, 80].
When using a pen recorder, the retention measurements are made in terms of chart
distances and the area difference is determined by graphical integration. In modern
systems, this theoretically correct analysis is also easy to perform when the digitized
signal is recorded on the computer hard disk. In practice, the calculation of the areas
always includes signal noise and it is very sensitive to the integration limits (the mass
transfer is often slow and the plateau concentration m 02 is reached slowly, thereby
resulting in significant systematic errors). Therefore, it is easier and better to handle

Fig. 6.18 Square profiles of the solute concentration obtained in the model flow experiments (a) and
breakthrough curves of the solute registered during real flow calorimetry runs (b). The dashed lines
refer to the profiles obtained with a non-adsorbing solid; τ A and τ N A are the retention times of the
retained solute and the completely unretained solute, respectively. The dashed areas represent the
amount of solute adsorbed (the calibration factor is needed)
240 J. J. Zajac

the retention times (chart distances) defined by characteristic points method (e.g.,
the retention parameters derived from the inflection point or the half-height [84]).
The thermal balance in the flow calorimetry experiment can now be expressed in
the following way:
   

Hexp = H f in − H init = m S · S · HSs L − HS0
s
− 2s ·  H m 02 +
 
− m S · S · HSs L1 − HS0
s
(6.59a)

Taking into account the discussion in Sect. 6.4.4,the first term on the right hand side
represents the enthalpy of immersion
imm H12 m 02 of the solid in excess solution
of molality m 02 and one obtains:
 

Hexp =
imm H12 m 02 −
imm H1 =
d pl H (6.59b)

where
d pl H denotes the integral enthalpy of displacement [64], which is the main
enthalpy effect measured in the liquid-flow calorimetry experiment.
The molar integral enthalpy of displacement is calculated as follows:

Hexp

d pl h = (6.60)
m S · S · 2s

This quantity provides information about the excess of component-adsorbent interac-

tions averaged over all surface domains from which the solvent has been displaced by
the adsorbing solute species. In consequence, it is not easy to monitor subtle changes
in the adsorption mechanism based on usually small variations of the
d pl h values
with increasing quantity of adsorption. Compared to the differential molar enthalpy
of displacement, the enthalpy
d pl h is less sensitive to the energetic heterogeneity
of the solid surface.

6.4.6 Model of Batch Calorimetry Experiment

Liquid titration calorimetry is a microcalorimetric batch technique most often used

to study the mechanism of solute adsorption onto solids from binary dilute solution
[83, 85, 86]. It differs from the flow variant in that a stock solution is injected into the
calorimetric cell where the solid sample is kept in homogeneous suspension in a liq-
uid (solvent or solution). The possibility of measuring the pseudo-differential molar
enthalpy effects is the most important advantage of this technique. The differential
enthalpy is very sensitive to various partial processes of adsorption occurring at the
Solid-Liquid interface, thus allowing changes in the interfacial properties with sur-
face coverage to be continuously monitored. From this standpoint, calorimetric data
are very useful for theoretical consideration and modelling. Additionally, titration
6 Calorimetry at the Solid–Liquid Interface 241

calorimetry may be an important analytical tool for determining the enthalpy effects
accompanying the dilution of solutions with given compositions.
A typical operational procedure involves injection of a stock solution by small
steps into the calorimetric cell containing either dilute solution (dilution experiment)
or suspension of powder solid sample in a dilute solution (adsorption experiment).
Serious drawbacks to the use of this calorimetry technique in studying the adsorp-
tion phenomena are due to the difficulty of direct evaluation of the related amounts
adsorbed and to insufficient control of the environment of the liquid phase (e.g.,
the pH and ionic strength cannot be maintained constant through the whole run)
[73, 87]. The progress in the adsorption quantity during successive injections of the
adsorbate to the cell is quantified from the adsorption isotherm measured separately
under exactly the same experimental conditions. For the adsorption of ionic solutes
at electrified interfaces, it is impossible to well reproduce the charging behaviour
of the adsorbent at a constant pH since this parameter cannot be re-adjusted during
injections.
The model of dilution experiment is depicted in Fig. 6.19. Prior to injection
sequence, there is M10 grams of pure solvent in the calorimetric cell. Then, small
aliquots of the stock solution of molality m 02 are injected by the syringe pump oper-
ating at a constant flow rate d pump . When the time of injection tin j is maintained
in j
constant, the amount of the solute n 2 introduced into the calorimetric cell is always
the same. As a result, the equilibrium solution in the cell becomes more and more
concentrated, i.e., m i2 → m 02 . The dilution data may be further processed in two
different ways, by calculating either differential molar or cumulative molar enthalpy
changes.

Fig. 6.19 Schematic representation of an injection step in the batch dilution experiment. The flow
rate d pump of the pump and the time tin j of injection are kept constant during the experiment. The
equation shows how to calculate the number of moles of the solute injected; m 02 and Msolute are the
molality of the stock solution in the syringe and the molar mass of the solute, respectively;  H (m 2 )
is the value of the apparent molal enthalpy of the solute in a binary solution corresponding to the
molality m 2
242 J. J. Zajac

The total enthalpy change during the ith injection per mole of the solute is
expressed as follows [73, 74, 88]:
  i  

in j Hi    H m 2 −  H m i−1
2
in j
in j
=  H m i2 + (i − 1) · n 2 · in j
+
n2 n2
 
−  H m 02 (6.61a)

in j
where n 2 depends on d pump , tin j , and m 02 , according to the formula given in
Fig. 6.19.
The equilibrium molality of the solution m i2 in the cell after this injection is
calculated by means of the following expression:

in j
103 · i · n 2
m i2 = in j
(6.61b)
103 ·i·n 2
M10 + m 02

in j Hi
The determination of all in j terms in function of m i−1
2 and m i2 yields the results
n2
in the form of a histogram, as shown in Fig. 6.20. Referring to the general definition
of the partial molal enthalpy h 2 and apparent molal enthalpy  H for any particular

Fig. 6.20 Titration calorimetry of successive dilutions of a 0.028 mol kg−1 aqueous solution of
benzyldimethyldodecylammonium bromide (BDDAB) in the presence of 0.01 mol kg−1 NaBr at
298 K [89]: (a) histogram of the molar enthalpy change during the ith injection as a function of
the equilibrium solution molality, (b) plot of the partial molal enthalpy of the solute against its
bulk molality. The solution is regarded as a binary system: the solute ions BDDAB+ and Br − form
together the mean solute, whereas NaBr is included in the mean solvent. The smooth curve of h 2
against the equilibrium
 solution
 molality m 2 is achieved by choosing the value of m 2 to lie in the
middle of interval m i−1
2 , m i
2
6 Calorimetry at the Solid–Liquid Interface 243

solute in a solution of two components (Eqs. 6.36 and 6.39), it is to be noted that,
in j
for the sufficiently small quantities n 2 of the solute injected during each injection,
Eq. (6.61a) can be transformed to give:
   

in j Hi H m i2 − H m i−1
2
in j
≈ lim in j
− const = h 2 (m 2 ) − const,
in j
n2 n 2 →0 n2
m i−1
2 < m 2 < m i2 (6.62)

where H denotes the enthalpy of a binary solution in the calorimetric cell.

Thus, values of h 2 may be obtained experimentally by measuring both enthalpy
in j
changes
in j Hi and very small quantities n 2 of the solute injected during successive
dilution of the stock solution. It is clear from Eq. (6.62) that the above relation does
not hold for the first injection: the first points in the calorimetric curves of dilution
should not be taken into consideration.
The thermal effects of successive injections can be also summed up to obtain the
molar cumulative enthalpy of dilution
dil h cum . After k injections, one obtains:

k

in j Hi     

dil h cum = in j
= k ·  H m k2 −  H m 02 ,
i=1 n2
with m 2 = m k2 (6.63)

The experimentally measured values of

dil h cum are further plotted against the
equilibrium solute molality m 2 and only such a representation may have clear phys-
ical meaning from a thermodynamic standpoint. Nevertheless, it is sometimes more
useful to present the enthalpy of dilution curve in terms of
dil h cum as a function of
the injection number k, especially when the dependence of
dil h cum vs m 2 is linear
or contains several linear portions (cf., Fig. 6.21).
in j
In the adsorption calorimetry experiment, a small amount n 2 of the stock solution
injected during a given injection is diluted in the supernatant liquid inside the cell and
some of the resulting species subsequently adsorb onto solid particles. They displace
a certain amount of solvent molecules and can exchange with some pre-adsorbed
molecules or ions, because of the limited extent of the adsorption space. The effects
of desolvation and re-solvation of various compounds taking part in the displacement
process contribute to the competitive character of adsorption at the solid-solution
interface. The flow chart of the batch displacement experiment is shown in Fig. 6.22.
Since the enthalpy effects accompanying dilution of the stock solution inside the cell
should be known, both the dilution and adsorption experiments are carried out under
the same conditions (cf., Fig. 6.19).
When the adsorbent is initially immersed in M10 grams of pure solvent and the final
state corresponds to the formation of a Solid-Liquid interface in equilibrium with a
bulk solution of molality m k2 , the mass balance inside the cell leads to the following
expression for the amount of solute adsorbed at the Solid-Liquid interface:
244 J. J. Zajac

Fig. 6.21 Cumulative enthalpy of dilution of a 0.009 mol kg−1 aqueous solution of hexade-
cyltrimethylammonium bromide (HTAB) at 303 K as a function of (a) the solute molality and
(b) the injection number [88]. The solution is regarded as a binary system: the solute ions HTA+
and Br− form together the mean solute. Plot a: the intersection of the two linear portions provides
estimate of the critical micelle concentration for HTAB; Plot b: the enthalpy of HTAB micellisation
is determined directly from the difference between the slopes of the two linear regression segments

Fig. 6.22 Schematic representation of an injection step in the batch adsorption experiment. The
in j
values of d pump , tin j , n 2 , and m 02 are the same as those in Fig. 6.19. The equation shows how to
calculate the mass of solvent injected with the stock solution. HSs L (m 2 ) refers to the Solid-Liquid
interface in equilibrium with a bulk solution of molality m 2
6 Calorimetry at the Solid–Liquid Interface 245
 
  k
mk
k
in j in j
n s2,k = n s2 0 → m k2 = n 2 − 23 M10 + M1 (6.64)
10
i=1 i=1

in j
where M1 is the mass of solvent introduced into the calorimetric cell during one
injection of the stock solution (see the appropriate equation in Fig. 6.22).
Equation 6.64 represents the so-called calorimetric line, since the amount adsorbed
n s2,k varies linearly with m k2 between two characteristic points: (P1) the whole amount
of the solute injected is retained by the adsorbent surface and m k2 = 0; (P2) the whole
amount of the solute injected remains in the bulk solution and n s2,k = 0. In practice,
it is impossible to measure directly the values of n s2 for all injections (it would not
be reasonable to interrupt the calorimetry run after each injection and remove the
calorimetric cell in order to analyse its content). Nevertheless, the partition of the
solute between the adsorbed and bulk phases is strictly determined by the adsorption
equilibrium at a given temperature and does not depend on the path by which the
adsorption system passes from its initial state to the equilibrium. Hence, the values
of m k2 and n s2,k are evaluated from the intersection between the calorimetric line
 
β
(6.64) and the experimental isotherm 2s = 2s m 2 , obtained under exactly the
same experimental conditions (Fig. 6.23).
The total change in enthalpy during the ith injection in the adsorption experiment
may be written as follows:

Fig. 6.23 Graphical determination of the amount adsorbed n s2,k after a given injection step in
the titration calorimetry run and the related molality of the equilibrium bulk solution m k2 from
the intersection between the calorimetric line 6.64 (passing through two characteristic points P1
and P2) and the experimental adsorption isotherm measured under the same conditions
246 J. J. Zajac
      

in j Hi = m S · S · HSs L m i2 − HSs L m i−1
2 − n s2,i ·  H m i2 +
 
+ n s2,i−1 ·  H m i−1
2 +
⎧      ⎫
⎨    H m i2 −  H m i−1 2  ⎬
in j in j
+ n 2 ·  H m i2 + (i − 1) · n 2 · −  H m 02
⎩ n2
in j ⎭
(6.65a)

where S is the specific surface area of the adsorbent; HSs L (m 2 ) is the interfacial
enthalpy for a Solid-Liquid interface in equilibrium with a bulk solution of molal-
ity m 2 . The above equation has been developed by neglecting any changes in the
enthalpy of the bulk of the adsorbent m S · h S . The comparison with Eqs. (6.54a,
6.54b) and (6.59a, 6.59b) in Sects. 6.4.4 and 6.4.5 indicates that the first three terms
on the right hand side of Eq.(6.65a)  represent
 the difference in the enthalpy of dis-
 i i−1
placement
d pl H m 2 −
d pl H m 2 when passing from m i−1 i
2 to m 2 . The forth
term has the same form as the overall enthalpy change recorded in the dilution exper-
in j
iment (cf., Eq. 6.61a). For small n 2 values accompanied by small increments in the
amount adsorbed n s2 , the batch adsorption experiment allows measuring the differ-
ential enthalpy of displacement per mole of the solute adsorbed at the Solid-Liquid
interface, since:
   

d pl H m i2 −
d pl H m i−1 2

d pl h di f f = lim

n 2 →0
s n 2,i − n 2,i−1
s s

 i  

d pl H m 2 −
d pl H m i−1 2
≈ =
n s2,i − n s2,i−1
   
in j β

in j Hi − n 2 · h 2 m 2 − const
= (6.65b)
n s2,i − n s2,i−1
 
β
where the equilibrium molality m 2 is located in the middle of interval m i−1 2 , m2 .
i

The experimental procedure and data processing leading to a smooth curve of

d pl h di f f as a function of the amount adsorbed n s2 include the following stages: the

in j
quantity of the solute injected n 2 and the successive enthalpy changes
in j Hi are
measured in the adsorption calorimetry experiment; the limit molality values m i−1 2
and m i2 for each injection class and the related increments in the amount adsorbed
n s2,i−1 and n s2,i are evaluated with the aid of the adsorption isotherm; based on the
“differential” enthalpy  curve
 obtained in the dilution experiment, the correction terms
in j β β
for dilution, n 2 · h 2 m 2 − const , are determined for the values of m 2 taken
 
as the middle of appropriate intervals m i−1 2 , m 2 ; with small increments
n 2 =
i s
6 Calorimetry at the Solid–Liquid Interface 247

n s2,i − n s2,i−1 , the differential displacement terms

d pl h di f f are finally calculated
β
and ascribed to the n s2 values in relation with the equilibrium molalities m 2 . It should
be noted again that the enthalpy value corresponding to the first injection cannot be
considered as the differential molar enthalpy of displacement and has to be removed
from the experimental curve.

6.5 Calorimetry Applied to Study Competitive Adsorption

from Dilute Solution

The integral and differential enthalpies of displacement can be measured directly

using calorimeters of the isothermal type. Such instruments either are produced
and marketed by some manufacturers of high-performance calorimetric systems for
different applications (e.g., SETARAM Instrumentation, LKB-ThermoMetric, TA
Instruments, Microscal Ltd), or are home-mode prototypes based on original work
carried out in university laboratories and therefore documented in considerable detail
in scientific publications. This section focuses on the properties of two calorimetry
systems used by the author to study the competitive adsorption from dilute solution.

6.5.1 Flow Calorimetry System

A liquid-flow calorimetry system (commercialised by Microscal Ltd) is represented

schematically in Fig. 6.24 [79, 80]. Here the calorimetric cell is simply a cylindrical
cavity inside the calorimetric bloc and its volume is limited by two removable tubes.
The outlet tube is also used as a holder for a powdered solid sample. Prior to each
calorimetry run, the inlet and outlet tubes are taken away from the calorimeter and
the calorimetric cell is cleaned.
The measuring termistors detect the temperature of the cell content via a PTFE
cell wall membrane which protects them from chemical attack. A separate temper-
ature sensor is installed in the calorimetric block to permit the digital display of
block temperature. The thermistor sensors are part of a Wheatstone bridge, which
has the advantage that only temperature differences between the calorimetric cell
and the block will put the bridge out of balance. When the solution containing a
preferentially adsorbing solute or the pure solvent reaches the adsorbent bed there is
an exothermic or endothermic displacement with resultant evolution or absorption
of heat. This causes a small resistance change to the measuring thermistors and the
effect is registered in the form of a thermal peak, the area of which is proportional
to the heat measured.
It is necessary to have an evenly packed adsorbent bed in the calorimetric cell and
a smooth, steady flow of liquid to pass through the bed so as to obtain reproducible
results. Since the volume of the cell is limited (about 0.17 mL), a suitable volume
248 J. J. Zajac

Fig. 6.24 Schematic representation of a liquid-flow microcalorimetry system operating in

continuous-flow mode, together with traces showing the thermal and mass exchange profiles for
adsorption of solute from its solution in the solvent, followed by the desorption of the solute by
flow of pure solvent: 1 adsorbent bed, 2 inlet tube, 3 outlet tube, 4 toric seals, 5 aluminium block,
6 measuring thermistors, 7 syringe pomp, 8 downstream detector

of powder or granular solid is weighed out in a special measuring tube provided by

the manufacturer. Then the outlet tube is fitted in the calorimeter and the powder
is poured into the sample holding space by means of an extended stainless steel
funnel. Contrary to appearances, this operation is crucial for further measurements.
Sometimes, the sample should be gently crushed to decrease the dead volume of
the adsorbent bed. The use of monodispersed particles can be beneficial. When the
particles are agglomerates of much finer material, they should not disintegrate in the
liquid flow through the adsorbent. For solids having significant differences in their
dry and wet packing densities (e.g., solids swelling in carrier liquids), the sample
volume placed in the cell must take account of the volume change occurring on sample
evacuation and immersion with the liquid. It is still possible to check whether the
quantity of solid sample introduced and its packing in the cell are correct by fitting
the inlet tube carefully. Any adsorbed species can be subsequently removed from
the sample surface by evacuation, making use of a vacuum pump connected to the
6 Calorimetry at the Solid–Liquid Interface 249

outlet tube. The temperature may be raised on evacuation by heating the calorimetric
block with the aid of a stabilised D.C. power supply. The temperature limit depends
on the thermal resistance of the materials of construction used. The efficiency of
the evacuation step can be enhanced by flushing the adsorbent bed with the carrier
liquid.
Further operating procedure for adsorption studies includes establishing a steady
flow of the solvent used as a carrier liquid, awaiting thermal equilibrium, and setting
the sensitivity controls. Different flow rates should be usually tested to optimise the
operation for peak height, shape, sensitivity and duration. It should be remembered
that the pump flow rate is limited by the syringe volume (normally 10 or 20 mL),
compressibility of the solvent utilised, size of the solid particles, and their packing
density in the sample holding space: the most commonly applicable flow rates are
included in the range 0.05–0.1 mL per minute. Moreover, all occluded air should be
carefully removed from the syringe and the whole line purged. As the equilibration
proceeds, the thermal signal in function of time approaches a straight line asymptot-
ically and this state is the criterion by which to judge the equilibration and to set the
sensitivity of the calorimeter.
After attaining thermal equilibrium in the system, the flow of carrier liquid sol-
vent is replaced with the identical flow rate of the solution of adsorbate at defined
concentration by means of a changeover valve. The solution is fed by a syringe
pump other than the one directing the solvent via the valve to the calorimeter. The
progress of solute adsorption is monitored by the evolution or absorption of heat
measured by thermistors sensing temperature changes in the calorimetric cell and
simultaneously adsorbate transfer from the bulk phase to the Solid-Liquid interface
is monitored by measuring composition changes in the effluent leaving the adsorbent
and passing through the downstream detector. The principle of detection (UV, refer-
actometry, thermal conductivity) and thus the choice of the detector depend on the
chemical nature of both the solute and the solvent, the flow rates to be used, as well
as the technical specification of the detector (e.g., operating parameters, sensitivity
of detection, linearity of the signal with concentration, baseline stability, ease of
operation). The adsorbent bed and the solution are allowed to remain in contact until
heat evolution or absorption ceases (the signal returns to the thermal baseline) and
no further change in the effluent composition is detected (the recorder trace shows a
straight line), as illustrated in Fig. 6.24.
Since the calorimetric system is particularly well adapted to the study of the
thermodynamic reversibility of the adsorption phenomenon, it is always worth car-
rying out the first desorption stage under exactly the same experimental conditions,
just by returning to the flow of the carrier liquid through the adsorbent bed. The
reader should be reminded that the heat effects of adsorption and desorption are
opposite in sign. When the thermal and detector signals return to those of pure
solvent, the adsorption-desorption cycle can be repeated to test the attainment of
adsorption reversibility (e.g., the heat effects of adsorption and desorption are equal
and repeatable). For some porous materials, the desorption process may last longer
than the corresponding adsorption due to the slower mass transfer kinetics and the
250 J. J. Zajac

peak tailings effects are observed, thereby reducing the measurement accuracy and
repeatability.
Calibration of the instrument is required at some stage for the particular operating
conditions employed. Calibration of the areas under the thermal peaks is carried
out by dissipating a known amount of energy in the adsorbed bed with the aid of a
calibration probe incorporated into the outlet tube and encapsulated in PTFE. The
related “exothermic” peak may be integrated making use of appropriate software
facilities to process the digitized signal recorded on the computer hard disk. The data
processing is the same as that described in Sect. 6.3.2 for immersional and wetting
calorimetry. The downstream detector provides a plot of the effluent composition,
the profile of this curve being influenced by the amount of solute molecules retained
on the adsorbent surface during adsorption or released from the surface during des-
orption. To determine the related amount of solute adsorbed or desorbed, this profile
is to be compared with composition changes obtained in a “blank” experiment with
the use of a “non-adsorbing” solid through which pure solvent and the solution are
passed at the same flow rate and temperature as those for the adsorbing sample.
Glass or PTFE balls of low specific surface areas are usually used as non-adsorbing
adsorbents, but the main difficulty here is to reproduce the same flow conditions of
liquids through the adsorbent bed in the blank run (e.g., packing density of the solid,
hydrophobic-hydrophilic character of its surface, pressure drop over the adsorbent
bed). The two types of composition profile can be matched and presented in the
form of net mass-transfer for adsorption and desorption. The areas of the segments,
resulting from the subtraction of the peaks obtained on adsorbing and non-adsorbing
solids, provide a direct measure of the quantity of the solute which either fails to
emerge in the effluent solution due to its adsorption by the adsorbent bed in the
calorimetric cell or is released to the carrier stream by the flow of pure solvent. The
detector calibration factor is determined from the injection of a given volume of the
solution into the stream of the solvent percolating through the detector (calibrated
injection loop facility).

6.5.2 Measurements of Integral Enthalpy of Displacement

Despite the limited physical meaning of the integral enthalpy of displacement, mea-
suring
d pl H or
d pl h may be very useful in several cases.
The first case certainly corresponds to the study of the thermodynamic reversibility
of adsorption onto solids from binary solutions. Liquid-flow calorimetry measure-
ments usually provide clear, unambiguous arguments for irreversible character of
the phenomenon in numerous systems. An example of such systems is illustrated
in Fig. 6.25. With non-porous Graphon possessing a very small number of surface
polar sites, the adsorption of C60 fullerene from toluene is completely reversible.
In the case of porous active carbons, the phenomenon is only partially reversible,
the degree of reversibility being evaluated from the difference between the values of

d pl h measured for the adsorption and desorption stage.

6 Calorimetry at the Solid–Liquid Interface 251

Fig. 6.25 Integral molar enthalpies (in kJ mol−1 ) of C60 fullerene adsorption (black bars) and
desorption (grey bars) obtained in one adsorption-desorption cycle from a 0.5 g L−1 toluene solution
onto graphitised carbon black (Graphon) and three active carbons (S-51 MB, S-51 WTX, Darco
G-60) at 293 K [75]

The irreversible enthalpy component may be calculated from the following

formula:

ads n ·
ads h +
des n ·
des h

d pl h I R = (6.66)

ads n −
des n

where
ads n and
des n are positive changes in the number of moles of solute
measured during the adsorption and desorption run, respectively;
ads h and
des h
are the corresponding molar enthalpies of displacement observed in both stages.
Low sensitivity of the
d pl H or
d pl h values to the surface heterogeneity effects
makes the integral enthalpy of displacement useful for probing specific sites on the
surface of solid materials. The principle of this method lies in measuring the enthalpy
of displacement per unit area of the adsorbent
d pl H ∗ during adsorption of specific
probe species (solute) capable of displacing non-specific solvent molecules from the
targeted surface sites. For example, the polar contribution to the interfacial enthalpy
HSs L may be thus approximated by determining the integral enthalpy of displacement
of an apolar solvent by a polar solute. Prior to calorimetry measurements, the solution
composition should be carefully optimised to ensure monomolecular adsorption of
the solute on the solid surface [90, 91].
Figure 6.26 shows the effect of heteroatom incorporation into the framework
of ordered mesoporous silica, as inferred from the adsorption of 1-butanol from
n-heptane.
When diluted in an apolar solvent, the molecules of 1-butanol (BuOH) may be
regarded as monomer species, potentially acting as both hydrogen-bond donors and
252 J. J. Zajac

Fig. 6.26 Integral enthalpy of displacement per unit surface area

dpl H ∗ (taken with the opposite
sign) measured in the liquid-flow calorimetry experiment for adsorption of 1-butanol from a 2 g
L−1 solution in n-heptane at 298 K onto ordered mesoporous silica of the SBA-15 type and three
mesoporous silica-based materials doped with various heteroatoms [92]

acceptors. Since the doping procedure aims at isomorphic substitutions of silicon

by such heteroatoms as Al, Ti or Zr in the tetrahedral structures, the hydrophilic
surface of the four mineral oxides should be dominated by surface hydroxyl groups.
Butanol molecules can form hydrogen bonds with these silanols (Si-OH), alumi-
nols (Al-OH), titanols (Ti-OH), or zirconols (Zr-OH). The monolayer adsorption of
BuOH is accompanied by simultaneous desorption of apolar heptane molecules. In
consequence, the integral enthalpy of displacement per unit area of the adsorbent
surface
d pl H ∗ is a function of the surface density of hydroxyl groups. In the case
of materials doped with Ti(IV) and Zr(IV), there is no reason for a marked change
in the surface density of hydroxyl groups.
This hypothesis is well illustrated by quite similar (within the experimental error)

d pl H ∗ values obtained for SBA-15, Zr-SBA-15, and Ti-SBA-15, irrespective of

differences in the surface area and porous structure among the samples. The partial
replacement of Si(IV) by Al(III) results in additional surface hydroxyls related to the
‘bridging’ Si(OH)Al hydroxyl structures, thereby enhancing the value of
d pl H ∗
(Fig. 6.26).
In the liquid-flow calorimetry experiment, the purified adsorbent bed remains in
contact with a stock solution of constant composition. It is clear that the environment
of the liquid phase does not change during the measurement. This is an important
advantage of the flow calorimetry, especially in the case of solid-solution systems
containing electrified interfaces. The study of ions adsorption from aqueous solutions
6 Calorimetry at the Solid–Liquid Interface 253

Fig. 6.27 Adsorption of Cd2+ cations from Cd(NO3 )2 aqueous solutions of varying concentration
at pH 7 onto Spherosil (S B E T = 25 m2 g−1 ) at 298 K [94]: a record of successive saturation and
desorption runs showing heat absorption and evolution: calibration (cal), adsorption (ads), and
desorption (des). For each concentration of the stock solution m 02 , the areas under the adsorption
and desorption peaks are equal

onto mineral oxides bearing a pH-dependent surface charge always requires constant
pH and ionic strength [71, 93].
Figure 6.27 presents the thermogram resulting from adsorption of a heavy metal
cation from aqueous solution on the negatively charged surface of Spherosil regis-
tered during liquid-flow calorimetry measurements.
For a given concentration of the stock solution m 02 , continuous percolation of the
solution through the calorimetric cell containing the solid sample leads to saturation
of the adsorbent with the solute giving rise to a negative heat effect (i.e., adsorption
is endothermic) in the form of a peak in which the beginning and the end depend
on the solution concentration, flow rate of solution throughout the adsorbent bed,
and the kinetics of adsorption. Then the solute is removed from the adsorbent by
exchanging the flow of the solution for that of pure solvent. The solute desorption is
exothermic and complete in each cycle since the adsorption and desorption enthalpy
effects have the same absolute value. Therefore, the successive saturation-desorption
cycles can be performed without changing the solid sample in the cell. Although the
thermogram in Fig. 6.27 shows that the resulting values of
d pl H increases with
increasing concentration, the molar enthalpy
d pl h is proven to be a monotonously
decreasing function of m 02 [94].
The comparison among thermal displacement effects accompanying the individ-
ual adsorption of an alkaline earth metal from aqueous solution is given in Fig. 6.28.
The pH and ionic strength of the aqueous phase were identical in the four systems
studied.
According to the appropriate speciation diagrams, each metal forms divalent
cationic species in aqueous solution under the experimental conditions applied. In
spite of the same electric charge of the four cations, the positive enthalpy values are
very different, indicating that electrostatic attraction is not the only driving force of
adsorption. Since there are Na+ ions in the heterogeneous system, the total displace-
ment effect should also include ion exchange between sodium and a given divalent
cation. Modelling attempts to reproduce the positive displacement effects lead to the
conclusion that metal cations may form multidentate complexes with oxygen atoms
of ionised silanol groups and changes in the hydration layers of the adsorbing and
254 J. J. Zajac

Fig. 6.28 Integral molar enthalpy of displacement related to the adsorption of a metal cation from
10−2 M nitrate solution onto Spherosil in the presence of 10−1 M NaNO3 at pH 7 and 298 K for
various divalent cations [95]. The number of water molecules in the hydration layer and the total
enthalpy of hydration are given for each metal cation in aqueous solution

desorbing cations are the main reason for the endothermic character of the overall
process [96].

6.5.3 Titration Calorimetry System

Liquid titration calorimeters contain a stirring device, which ensures the homogene-
ity of the liquid solution or solid suspension in the calorimetric cell, and an injection
system permitting the controlled introduction of the reagents from outside the calori-
metric cell. Contrary to the flow system, nothing flows out of the calorimeter to the
surroundings since the reagents from a stock solution fed to the injection device are
collected within the calorimetric cell. The heat detection is usually based on the prin-
ciples of the isothermal batch or flow microcalorimeters. An example of home-made
microcalorimeter [86] designed for study of the enthalpies of mixing of liquid and
adsorption from dilute solution onto divided solids is represented schematically in
Fig. 6.29.
The variations of temperature in the calorimetric cell induced by dilution or
adsorption phenomena, as well as by the electrical calibration are recorded by ther-
mistors arranged as a Wheatstone bridge. The two measuring thermistors, calibration
coil and the inlet end of the injection tube are immersed in the solution or suspen-
sion. A precision syringe pump injects a stock solution into the cell at a constant rate
ranging between 0.01 and 0.2 g min−1 without introducing any significant thermal
perturbation. The injected solution flows through a heat exchanger tube where it is
heated up to attain the temperature of the calorimetric block. The measurement of the
6 Calorimetry at the Solid–Liquid Interface 255

Fig. 6.29 Schematic representation of the calorimetric cell, stirring device, liquid injection and
heat effect recording systems of the Montcal titration microcalorimeter [86], together with a thermal
record of successive injections of a stock solution by small steps of 0.1 g mn−1 into the calorimetric
cell: 1,2 measuring thermistors, 3,4 reference thermistors, 5 calibration coil, 6 stainless steel calori-
metric cel (12–30 mL), 7 inert cover for calorimetric cell, 8 stirrer, 9 aluminium calorimetric block,
10 injection tube, 11 heat exchanger tube, 12 pre-heater, 13 magnet attached for stirrer, 14 magnet
attached to electric motor, 15 electric motor, 16 aluminium cylinder supporting the exchanger tube,
17 syringe pump, 18 bearing; cal.1, cal.2—calibration peaks, inj.1, …, inj.5—injection peaks

enthalpy changes may be carried out also at higher temperatures (this microcalorime-
ter has been proven to give satisfactory results at temperatures ranging from 20 and
50◦ C [86]). For this purpose, an additional power unit provides a stabilised D.C.
power supply for heating the calorimetric block to the desired temperature. To reduce
the heat loss or temperature variations during injection, it is sometimes necessary to
thermostate both the stock solution in the syringe and the injection tube outside the
calorimetric block. A horizontal agitator with a variable speed of rotation is driven
by a stepper motor fitted through a magnetic transmission.
The dissipation of a known amount of electrical energy inside the calorimetric
cell by means of a calibration coil (i.e., the Joule effect) is used to relate the area
of the thermal peaks recorded to the enthalpy effects which this represents. The
difficulty with this type of calibration in the titration calorimetry systems is related
to the fact that the mass of solution in the calorimetric cell is constantly increased by
successive injections, thereby changing the calorific capacity of the cell. Therefore,
thermal calibration should be regularly repeated after each series of injections in the
same calorimetric run.
256 J. J. Zajac

The calorimetric cell, together with the measuring termistors, stirring device and
injection system, can be removed from the calorimetric block to facilitate cleaning
and refilling the cell. After the instrument set up and attainment of thermal equi-
librium, all steps of the run (injections of the stock solution containing the adsor-
bate, calibration processing, and recording of the heat effects) are carried out by the
appropriate computer system. It is a best practice to weigh the calorimetric cell at
the end of each run in order to calculate the real (mean) pump rate. One of the orig-
inal features of the construction of this calorimeter is also the possibility of easily
changing the thickness of the insulating barrier between the calorimetric cell and the
metal block. This operation permits the modification of the calorimeter sensitivity
and the time of the return of the thermal signal to the baseline.
The titration calorimetry technique presents some limitations due to the necessity
of a strict correlation between the dilution and adsorption measurements. A special
care must be taken to avoid experimental artefacts and erroneous interpretations in
the study of solute adsorption onto solid supports which dissolve to a great extent in
solutions or when the quantity of foreign substances released from the solid surface
to the bulk solution is significant. In such cases, the composition of the equilibrium
supernatant does not correspond to the pure solvent and it may even change con-
stantly with increasing adsorption of the preferentially adsorbed solute [87, 93, 97].
Firstly, this evolution of the supernatant liquid should be monitored thoroughly dur-
ing adsorption. Then, the evaluation of the correction term for dilution may pose
serious problems since the stock solution has to be prepared by dissolving solute in
the actual supernatant liquid and not in the pure solvent.
An example of the adsorption system investigated by means of titration calorime-
try technique is given in Fig. 6.30 [74]. The experimental adsorption isotherm
(Fig. 6.30b) has been determined separately based on the solution depletion method.
The solute (dodecyldimethylammonio) butanoate (C12N3C) is a zwitterionic surfac-
tant possessing a dipolar head-group and a linear aliphatic tail. In aqueous solution,
the solute molecules self-assemble into aggregates called micelles. The concentra-
tion of the surrounding aqueous phase at which the surfactant monomers begin to
form micelles is known as the critical micelle concentration (CMC).
When a micellar stock solution (i.e., its concentration is 10 times the CMC)
is injected into a more dilute solution in the calorimetric cell, the constant value of
partial molal enthalpy h 2 in the premicellar region is due to destruction of micelles and
dilution of unmicellized species (Fig. 6.30a); the constant h 2 value in the postmicellar
region is ascribed to dilution of micelles.
The molar change in partial molal enthalpy of the surfactant when monomers
associate into a micelle at the cmc represents the standard enthalpy of micellisation

mic h o per mole of surfactant monomers [98]. In accordance with the variations of
β
h 2 as a function of m 2 in Fig. 6.30a, the enthalpy of micellisation for C12N3C at 298
K is positive, indicating that the micellisation process is endothermic.
The experimental curves describing the adsorption of C12N3C onto Spherosil
XOB015 from aqueous solution at 298 K (i.e., Figs. 6.30b and 6.30c) suggest that
the phenomenon generally occurs in two stages. At very small quantities of adsorp-
6 Calorimetry at the Solid–Liquid Interface 257

Fig. 6.30 Dilution of aqueous solution of (dodecyldimethylammonio) butanoate (C12N3C) and

its adsorption onto Spherosil XOB015 (S B E T = 25m2 g−1 ) at 298 K: (a) enthalpy of dilution,
(b) adsorption isotherm, (c) differential molar enthalpy of displacement. In both types of titration
calorimetry experiment, a 0.3 mol kg−1 C12N3C solution in pure H2 O was used

tion n s2 , the values of

d pl h di f f are negative so that the phenomenon is exothermic.
The negative enthalpies of displacement are usually attributed to individual adsorp-
tion of surfactant molecules on an empty surface, where there may be only a few
adsorbed molecules, and therefore lateral adsorbate-adsorbate interactions can be
neglected [74]. The subsequent adsorption stage is dominated by adsorbate-adsorbate
interactions. At moderate and great adsorption amounts, the driving force of adsorp-
tion derives from the hydrophobic effect, i.e., lateral chain-chain attractions and the
tendency of hydrophobic tails to escape from an aqueous environment. This mode of
adsorption is characterised by a constant, positive enthalpy of displacement, showing
much similarity to micelle formation in the bulk solution [74]. This surface aggrega-
tion is likely controlled by a pseudo nucleation step, i.e., individual adsorption: the
first adsorbed monomers act as nucleation centres for future surface-bound surfactant
aggregates formed through chain-chain association.
258 J. J. Zajac

6.5.4 Scanning of Surfactant Aggregation

by Titration Calorimetry

Surface-active molecules or ions with an amphiphilic structure are known to have

low solubility in water and self-assemble into large aggregates [61]. According to the
enthalpy curves presented in Fig. 6.30, surfactant aggregation may occur not only in
aqueous solution but also at the Solid-Liquid interface. Since the topic of surfactant
aggregation is a well-developed field of research [62, 71, 73, 93, 99, 100], only a
brief review of the broad principles is proposed in the present paragraph to better
illustrate the contribution of titration calorimetry.
In aqueous solution, the formation of surfactant aggregates is driven by the ten-
dency of the surfactant units to densely pack their tails. The hydrophobic tails remain
inside the liquid-like micellar core due to unfavourable interactions with water mole-
cules, whereas the polar head-groups, due to favourable interactions with the solvent,
form a hydrophilic outer layer protecting the hydrophobic core. In the case of ionic
surfactants, the Coulombic repulsion among the ionised head-groups is moderated
by the specific adsorption of some counter-ions close to them within the Stern part
of a curved ionic double layer surrounding the micelle. The micelle morphology
and size depend primarily on the nature and relative sizes of the hydrophilic and
hydrophobic moieties, as well as on the composition and environment of the aque-
ous phase [61, 62]. Titration calorimetry may be very useful when studying surfactant
micellisation under different experimental conditions. The fundamental thermody-
namic parameters, namely the CMC and the standard molar enthalpy of micellisation

mic h ◦ , can be easily inferred from calorimetric measurements of successive dilu-

tions of a micellar stock solution injected by small steps into aqueous solution having
a given composition [73, 74]. Several examples of cationic and zwitterionic surfac-
tants are given in Table 6.5.
It is important to note that the formation of micelles in pure water may be an
exothermic, endothermic or even athermic phenomenon, depending on the detailed
molecular structure of the surfactant. The CMC value is related to the standard
Gibbs energy of micellization,
mic G ◦ , which always takes negative values. This
negative energy results rather from a large increase in entropy, which is ascribed
either to structural changes in the solvent, associated with loss of hydration of the
hydrophobic tail when the surfactant enters the micelle [61] or to increased freedom
of the hydrophobic chain in the interior of the micelle compared to the bulk aqueous
medium [101]. Besides the volume and length of the hydrophobic tail, the area per
head-group at a curved interface between the micelle core and the aqueous solution
σmic is a critical packing factor having an impact on the ultimate micelle structure
[102]. For conventional ionic and zwitterionic surfactants with a single hydrocarbon
chain, globular micelles are formed above but near the CMC. A significant increase
in the overall surfactant content in aqueous solution may induce a change from spher-
ical micelles to cylindrical (prolate) or disc-like (oblate) aggregates. This change is
parralled by a decrease in σmic which allows a larger number of monomers to be
inserted into each aggregate. The maximum cohesion is attained in large lamellar
6 Calorimetry at the Solid–Liquid Interface 259

Table 6.5 Critical micelle concentrations, CMC, and standard enthalpies of micellisation,
mic h ◦ ,
per mole of surfactant monomers for selected quaternary ammonium surfactants in pure water at
298 K [73, 74, 88, 97, 103, 104]
Surfactant acronym and formula CMC
mic h ◦
mmol kg−1 kJ mol−1
Zwitterionic surfactants
C12N1C: C12 H25 (CH3 )2 N+ (CH2 )CO− 2 1.9 4.6
C12N3C: C12 H25 (CH3 )2 N+ (CH2 )3 CO− 2 4.6 8.8
C12N3S: C12 H25 (CH3 )2 N+ (CH2 )3 SO− 3 3.0 3.6
Classical cationic surfactants
BDDAB: (C6 H5 )(CH2 )N+ (CH3 )2 (C12 H25 )Br − 5.6 −5.3
TTAB: C14 H29 N+ (CH3 )3 Br − 4.0 −4.7
DTAB: C12 H25 N+ (CH3 )3 Br − 14.8 −1.6
Gemini cationic surfactants with a hydrophobic spacer
C12S2C12:
C12 H25 (CH3 )2 N+ (CH2 )2 N+ (CH3 )2 C12 H25 Br − 0.84 −22
C12S6C12:
C12 H25 (CH3 )2 N+ (CH2 )6 N+ (CH3 )2 C12 H25 Br − 1.03 −8.5
C12S12C12:
C12 H25 (CH3 )2 N+ (CH2 )12 N+ (CH3 )2 C12 H25 Br − 0.37 −12.2
Gemini cationic surfactants with a hydrophilic spacer
C12EO3C12:
C12 H25 (CH3 )2 N+ (C2 H4 O)3 N+ (CH3 )2 C12 H25 Br − 1.02 −6.9
C12EO7C12:
C12 H25 (CH3 )2 N+ (C2 H4 O)7 N+ (CH3 )2 C12 H25 Br − 1.58 0.0
C12EO12C12:
C12 H25 (CH3 )2 N+ (C2 H4 O)12 N+ (CH3 )2 C12 H25 Br − 1.93 6.3

sheets (flat bilayers) two molecules thick, though surfactant tails never attain such a
close-packed arrangement in dilute solutions.
Dimeric or Gemini surfactants (composed of two surfactant units connected by a
hydrophobic or hydrophilic chain—spacer) usually show a much stronger tendency
for micellar growth and self-assemble into larger aggregates with a lower degree of
curvature (e.g., linear thread-like and tree-like micelles, or spheroids) [105, 106].
The addition of solid particles (porous or non porous) into aqueous solution
induces a decrease in the chemical (or electrochemical) potential of the surfactant
solute and, in consequence, the adsorbing surfactant units form, at the solid surface,
some periodic adsorbate self-assemblies closely related to the micellar structures
encountered in the bulk solution at higher monomer concentrations. Such surface-
bound aggregates are spoken of as interfacial aggregates or solloids [107]. In the
general case, the shape and size of solloids are considered to be a compromise
between the free curvature, as defined by the energetic, geometrical and packing fac-
tors arising from the molecular structure of the surfactant in a given environment, and
some influences and constraints imposed, on the one hand, by direct solute-surface
260 J. J. Zajac

interactions and, on the other hand, by the porosity of the adsorbent. For example,
the images of extended aggregate structures showing a closer registry with the under-
lying surface have been obtained by atomic force microscopy (AFM) on atomically
smooth crystalline surfaces [99]. Such solloids have cylindrical and hemi-cylindrical
morphologies, depending on the hydrophilic-hydrophobic character of the solid sur-
face. In the case of powders or porous solids, where such microscopy techniques as
AFM or ellipsometry have very limited applicability, the titration calorimetry mea-
surements of the differential molar enthalpy of displacement as a function of the
surfactant adsorption may provide important information about the self-assembled
surfactant structures when compared with the thermal effects of surfactant micel-
lisation in aqueous solution under the same experimental conditions. Nevertheless,
calorimetry alone cannot be used to scan for the detailed solloid morphology and
appropriate modelling of the adsorption system is necessary.
As far as the adsorption of ionic surfactants on the oppositely charged (and macro-
scopically flat) surfaces of mineral oxides is concerned, the following three types of
solloid are frequently used in the empirical explanation or modelling of the experi-
mental data:
(1) monolayered hemimicelles, composed of surfactant units oriented ‘head-on’
towards the surface, with the surfactant tails forming a hydrophobic film in
contact with the equilibrium aqueous solution [108],
(2) bilayered admicelles, containing two adsorbed layers of the surfactant monomers
directed ‘head-on’ and ‘head-out’ with respect to the surface [109],
(3) small surface micelles, i.e., spherical isolated aggregates anchored to certain
surface sites with aggregation numbers markedly smaller than in bulk micelles
[110].
Each of these solloid morphologies is claimed to have a micelle-like character,
although the contact area between the water molecules and the hydrophobic
surfactant moieties is not always reduced to the minimum.
Typical plots of the differential molar enthalpy of displacement against the amount
of the surfactant adsorbed reveal significant variations in the enthalpy value when
the adsorption progresses. An example of the enthalpy curve is given in Fig. 6.31.
Such trends in
d pl h di f f with increasing n s2 suggest almost continuous evolution
of the solloid morphology and size: the aggregates self-assembled from the adsorb-
ing surfactant monomers at equilibrium concentrations in the bulk phase lower than
the CMC may grow in the direction parallel and perpendicular to the solid sur-
face. The most successful theoretical attempts to mimic complex shapes of both the
experimental adsorption isotherms and enthalpy of displacement curves have been
based on the assumption that the adsorbed phase at a given surface coverage can be
seen as a mixture of mutually interacting surface-bound monomers, monolayered
hemimicelles, and bilayered admicelles varying in size and number [65, 111]. The
proportion between the various types of adsorbate species is shown to undergo signifi-
cant changes with increasing surfactant adsorption, first monomers and monolayered
aggregates and then bilayered admicelles dominating on the surface.
6 Calorimetry at the Solid–Liquid Interface 261

Fig. 6.31 Variations of the differential molar enthalpy of displacement as a function of the adsorp-
tion of benzyldimethylammonium bromide (BDDAB) onto silica powder S91-16 (Rhône-Poulenc,
France) from aqueous solutions at 298 K at the initial pH 8 [89]. The arrows indicate the critical
micelle concentration (CMC) and the isoelectric point (IEP) at which the effective charge of the
silica particles together with the specifically adsorbed surfactant cations becomes equal to zero.
The region of particle flocculation (where the silica particles covered with the adsorbed species are
predominantly hydrophobic) is also shown

When surfactants are adsorbed onto fine-pore solids, the growth of solloids in the
direction perpendicular to the pore walls should be limited by the pore volume [93].
For mesoporous ordered mineral oxides, it may be even that the surfactant monomers
adsorb only head-on with respect to the hydrophilic surface and the hydrophobic tails
of the surfactant units adsorbed on the opposite walls interpenetrate themselves in
such a way as to produce “internal” aggregates, which fill the pore space. To check
whether the head-out adsorption of surfactant monomers is to be excluded in such
systems, one can refer to the micellar solubilisation of water-insoluble, hydrophobic
materials as well as polar substances, which dissolve in water only to a limited extent
[62, 69, 112, 113]. For low contents in aqueous solution, some small molecules
may behave as molecular probes, occupying specific sites in surfactant aggregates
without greatly disturbing their morphology and size. The exact locus of a given probe
molecule in a micelle reflects the type of forces operating between the aggregate and
the solubilised material. This justifies the use of titration calorimetry to study micellar
solubilisation. One of the possible methodologies is to lump water and the additive
together into the mean solvent and investigate the micellisation of the surfactant in this
new medium. The molar enthalpies of micellisation for several cationic surfactants
in the absence and the presence of phenol at various additive contents are compared
schematically in Fig. 6.32.
On the addition of phenol to the aqueous phase, the enthalpy of micellisation per
mole of the surfactant becomes more negative (and the CMC value is decreased),
thereby indicating more favourable phenol-surfactant interactions after the transfer
262 J. J. Zajac

Fig. 6.32 Molar enthalpies of micellisation,

mic h, for selected quaternary ammonium surfactants
in pure water and in the presence of phenol molecules in the aqueous phase at 298 K [88, 114, 115].
The overall phenol (PhOH) content (in mmol kg−1 ) is reported on the X-axis. The surfactant
acronyms are explained in Table 6.5. For hexadecyltrimethylammonium bromide (HTAB), the
calorimetry measurements were carried out at 303 K safely above the Krafft point

of the additive to the micellar phase. Therefore, phenol molecules are preferentially
located in the outer portions of cationic micelles close to the surfactant head-groups,
without involving much rearrangement of the micelle structure. When the phenol
concentration increases (e.g., PhOH-HTAB systems), the existence of an endother-
mic contribution to
mic h may be deduced from the evolution of the enthalpy value.
To better understand this positive enthalpy component, one may refer to the detailed
analysis of 1 H NMR spectra recorded with the various PhOH-HTAB solutions: addi-
tional phenol units penetrate deeper into the micelle core producing unfavorable inter-
actions with cationic micelles of HTAB [88]. Consequently, if phenol is to be used as
a molecular probe for detecting the presence of the head-out adsorbed surfactants at
the Solid-Liquid interface, the overall additive content in the system should remain
low.
Based on the assumption that surfactant aggregation on the solid surface is a pre-
requisite for the uptake of phenol and the aromatic molecules can be located only
close to the “free” head-groups of the surfactant units within the interfacial aggre-
gates, it may be helpful to compare the curves presenting the differential enthalpy of
displacement as a function of the surface coverage by the surfactant adsorbate in the
absence and in the presence of the additive. Figure 6.33 illustrates such a comparison
for a cationic surfactant adsorbed onto ordered mesoporous aluminosilicate of the
MCM-41 type.
The difference between the two curves is clearly pronounced at higher surface cov-
erage ratios where surface aggregation is considered to be the predominant sorption
mode. With a small amount of phenol added to the aquous phase, the displacement
phenomenon is more exothermic in this region: the related portion of the enthalpy
6 Calorimetry at the Solid–Liquid Interface 263

Fig. 6.33 Effect of phenol addition on the differential molar enthalpy of displacement upon adsorp-
tion of cationic Gemini C12C12C12 onto ordered mesoporous aluminosilicate of the MCM-41 type
(S B E T = 860 m2 g−1 , mean pore diameter = 5 nm, Si:Al = 32) from aqueous solution at 298 K and
the initial pH 8 [114]. The
dpl hdi f f enthalpy is plotted against the adsorption coverage of the solid
surface by the surfactant cations

curve can be viewed as shifted towards more negative values by a constant value
in comparison with the curve corresponding to the system without phenol. Similar
enthalpy behaviour of this surfactant has been observed during its micellisation in
aqueous solution (cf., Fig. 6.32). This analysis provides strong indication for phe-
nol incorporation in the interfacial aggregates having their head-groups oriented
outwards. Therefore, the image of all surfactant units interacting directly with the
negatively charged surface and their hydrophobic tails filling the whole pore space
available is rather to be excluded.
The coexistence of hydrophilic and hydrophobic nano-domains separated in space,
with a local order and fluidity typical of liquids, confer to supramolecular sur-
factant structures remarkable properties, which are advantageous in applications
involving molecular confinement within nanoscopic regions and reactivity in micro-
heterogeneous media. Micelle-mediated reactions constitute the basis of the so-called
micellar catalysis [62, 116], admicellar catalysis [117] or admicellar polymerisation
[118] in which reaction mechanisms may be controlled at a molecular level to save
energy and raw materials, as well as to avoid lengthy post-reaction purification and
analytical steps.

6.6 Concluding Remarks

The intention of the present chapter was to present mostly the prospective advantages,
but also some limitations, of the use of isothermal calorimetry at the Solid-Liquid
interface as a powerful tool in the study of interactions between solid surfaces and
264 J. J. Zajac

the surrounding liquid phase. The operating principles and examples of applications
were described for three calorimetry techniques frequently used to date in Surface
Science and Technology: (i) wetting and immersional calorimetry, (ii) liquid flow
calorimetry, (iii) titration batch calorimetry. The interested reader is encouraged to
search for other outstanding examples of commercial or home-made instruments and
their specific uses which have not been included here.
Wetting and immersional calorimetry may be employed to determine surface
properties of catalysts, adsorbents and other solid materials in contact with liquids.
Based on the calorimtery measurements of the various contributions to the total sur-
face enthalpy of a solid, it is possible to evaluate the hydrophobic-hydrophilic charac-
ter of its surface. The Harkins-Jura method for the evaluation of the specific surface
area of a solid based on enthalpy changes in the so-called immersion-adsorption-
wetting cycle gives the surface area of contact between this solid and a pure liquid
or a solution, thereby sheding light on the availability of the solid surface under real
experimental conditions.
Liquid flow or batch titration calorimetry techniques offer an opportunity for
studying the macroscopic outcome of the various interactions involved in interfacial
phenomena occurring at the Solid-Liquid Interface. The enthalpy changes appear
very sensitive to the partial mechanisms through which a given phenomenon can
occur. In particular, the effects of heterogeneity of a solid surface (i.e., surface sites
with different adsorption energies, “confinement effects” due to adsorbent poros-
ity) show up more clearly in heat quantities than in adsorption isotherms, thereby
allowing easier interpretation of the phenomenon studied. The assessment of the ther-
modynamic reversibility of competitive adsorption from multicomponent solutions
is one of the different possibilities of isothermal calorimetry. Self-assembled surfac-
tant structures, defined by the regular assembly of small molecular entities into larger
supra-molecular structures either in aquous solution or at the Solid-Liquid interface,
may be thermodynamically described based on the results of titration calorime-
try measurements. Here the next step would be to use high sensitivity isothermal
calorimeters to determine the thermal effects of micelle-mediated reactions in micel-
lar or admicellar catalysis.
In spite of many advantages, Solid-Liquid calorimetry alone is not capable of
solving satisfactorily many detailed problems concerning the resulting interfacial
mechanisms. It certainly cannot provide much information on entropy changes.
In consequence, calorimetric measurements have to be always supplemented by
other experimental studies reported on the system so as to obtain a more complete
description of the phenomenon

References

1. R. Defay, I. Prigogine, A. Bellemans, D.H. Everett, Surface Tension and Adsorption (Long-
mans, London, 1966)
2. P.C. Hiemenz, Principles of Colloid and Surface Chemistry, 2nd edn. (Marcel Dekker, New
York, 1986)
6 Calorimetry at the Solid–Liquid Interface 265

3. R.J. Hunter, Foundations of Colloid Science,, vol. 1 and 2, (Oxford University Press, Oxford,
1989)
4. A.W. Adamson, Physical Chemistry of Surfaces, 5th edn. (Wiley-Interscience, New York,
1990)
5. J. Lyklema, Fundamentals of Interface and Colloid Science, vol. 1–3, (Academic Press, Lon-
don, 1991–2000)
6. D.H. Everett, Reporting data on adsorption from solution at the solid/solution inter-
face (Recommendations 1986). Pure Appl. Chem. 58(7), 967–984 (1986). doi:10.1351/
pac198658070967
7. W. Rudzinski, D.H. Everett, Adsorption of Gases on Heterogeneous Surfaces (Academic
Press, London, 1992)
8. E.A. Guggenheim, Thermodynamics, 5th edn. (North Holland Publishing Co., Amsterdam,
1967)
9. J.F. Padday, in Surface Tension. II. The Measurement of Surface Tension, vol 1, ed. by E.
Matijevic, F. Eirich. Surface and Colloid Science, vol 1 (Wiley-Interscience, New York, 1969),
pp. 101–149
10. N.R. Pallas, Y. Harrison, An automated drop shape apparatus and the surface tension of pure
water. Colloids Surf. 43(2), 169–194 (1990). doi:10.1016/0166-6622(90)80287-E
11. R. Cini, G. Loglio, A. Ficalbi, Temperature dependence of the surface tension of water by the
equilibrium ring method. J. Colloid Interface Sci. 41(2), 287–297 (1972). doi:10.1016/0021-
9797(72)90113-0
12. G. Loglio, A. Ficalbi, R. Cini, A new evaluation of the surface tension temperature coefficients
for water. J. Colloid Interface Sci. 64(1), 198–198 (1978). doi:10.1016/0021-9797(78)90352-
1
13. R.C. Weast (ed.), Handbook of Chemistry and Physics, 45th edn. (CRC, Cleveland, 1964)
14. A.R.C. Westwood, T.T. Hitch, Surface energy of 100 potassium chloride. J. Appl. Phys.
34(10), 3085–3089 (1963)
15. S. Boffi, M. Ricci, On the cleavage energy of magnesium oxide. Mater. Chem. 1(4), 289–296
(1976). doi:10.1016/0390-6035(76)90030-4
16. J.J. Gilman, Direct measurements of the surface energies of crystals. J. Appl. Phys. 31(12),
2208–2218 (1960)
17. E. Orowan, Die Zugfestigkeit von Glimmer und das Problem der technischen Festigkeit. Z.
für Phys. A Hadrons Nuclei 82(3), 235–266 (1933). doi:10.1007/bf01341490
18. R.J. Good, Contact angle, Wetting, and Adhesion: a critical review, in Contact Angle, ed. by
K.L. Mittal, Wettability and Adhesion (VSP, Utrecht, 1993), pp 3–36
19. C.J. Van Oss, M.K. Chaudhury, R.J. Good, Interfacial Lifshitz-van der Waals and polar interac-
tions in macroscopic systems. Chem. Rev. 88(6), 927–941 (1988). doi:10.1021/cr00088a006
20. A.C. Zettlemoyer, Hydrophobic surfaces. J. Colloid Interface Sci. 28(3–4), 343–369 (1968).
doi:10.1016/0021-9797(68)90066-0
21. J.N. Israelachvili, Intermolecular and Surface Forces, 2nd edn. (Academic Press, London,
1991)
22. J.C. Berg, The Role of Acid-Base Interactions in Wetting and Related Phenomena, in Wetta-
bility, ed. by J.C. Berg (Marcel Dekker, New York, 1993), pp. 75–148
23. C.J. Van Oss, M.K. Chaudhury, R.J. Good, Monopolar surfaces. Adv. Colloid Interface Sci.
28, 35–64 (1987). doi:10.1016/0001-8686(87)80008-8
24. D.B. Hough, L.R. White, The calculation of Hamaker constants from Liftshitz theory with
applications to wetting phenomena. Adv. Colloid Interface Sci. 14(1), 3–41 (1980). doi:10.
1016/0001-8686(80)80006-6
25. J. Visser, On Hamaker constants: a comparison between Hamaker constants and Lifshitz-van
der Waals constants. Adv. Colloid Interface Sci. 3(4), 331–363 (1972). doi:10.1016/0001-
8686(72)85001-2
26. C.J. Van Oss, R.J. Good, M.K. Chaudhury, The role of van der Waals forces and hydrogen
bonds in "hydrophobic interactions" between biopolymers and low energy surfaces. J. Colloid
Interface Sci. 111(2), 378–390 (1986). doi:10.1016/0021-9797(86)90041-X
266 J. J. Zajac

27. R.S. Drago, Quantitative evolution and prediction of donor-acceptor interactions. Struct.
Bond. (Berlin) 15, 73–139 (1973)
28. F.M. Fowkes, Quantitative characterization of the acid-base properties of solvents, polymers,
and inorganic surfaces. in Acid-Base Interactions—Relevance to Adhesion Science and Tech-
nology ed. by K.L. Mittal, H.R.J. Anderson (VSP, Utrecht, 1991), pp. 93–115
29. M.K. Chaudhury, Interfacial interaction between low-energy surfaces. Mater. Sci. Eng. R Rep.
16(3), 97–159 (1996). doi:10.1016/0927-796X(95)00185-9
30. J.M. Douillard, T. Zoungrana, S. Partyka, Surface Gibbs free energy of minerals: some values.
J. Pet. Sci. Eng. 14(1–2), 51–57 (1995). doi:10.1016/0920-4105(95)00018-6
31. L.A. Girifalco, R.J. Good, A theory for the estimation of surface and interfacial energies.
I. Derivation and application to interfacial tension. J. Phys. Chem. 61(7), 904–909 (1957).
doi:10.1021/j150553a013
32. C.J. Van Oss, Acid-base interfacial interactions in aqueous media. Colloids Surf. A Physico-
chemical Eng. Aspects 78, 1–49 (1993). doi:10.1016/0927-7757(93)80308-2
33. P.G. De Gennes, Wetting: statics and dynamics. Rev. Mod. Phys. 57(3), 827 (1985)
34. A.W. Neumann, R.J. Good, Technique of measuring contact angle, in Surface and Colloid
Science, vol. 11, ed. by R.J. Good, R.R. Stromberg (Plenum Press, New York, 1979), pp.
31–91
35. C.J. Van Oss, R.F. Giese, Z. Li, K. Murphy, J. Norris, M.K. Chaudhury, R.J. Good, Determi-
nation of contact angles and pore sizes of porous media by column and thin layer wicking. J.
Adhes. Sci. Technol. 6, 413–428 (1992). doi:10.1163/156856192X00755
36. H.G. Bruil, J.J. van Aartsen, The determination of contact angles of aqueous surfactant solu-
tions on powders. Colloid Polym. Sci. 252(1), 32–38 (1974). doi:10.1007/bf01381692
37. J.M. Douillard, V. Médout-Marère, A new interpretation of contact angle variations in view of
a recent analysis of immersion calorimetry. J. Colloid Interface Sci. 223(2), 255–260 (2000).
doi:10.1006/jcis.1999.6679
38. V. Médout-Marère, S. Partyka, G. Chauveteau, J.M. Douillard, R. Dutartre, Surface hetero-
geneity of passively oxidized silicon carbide particles: vapor adsorption isotherms. J. Colloid
Interface Sci. 262(2), 309–320 (2003). doi:10.1016/S0021-9797(03)00198-X
39. C.J. Van Oss, The Apolar and Polar Properties of Liquid Water and Other Condensed-Phase
Materials. in Interface Science and Technology, vol 16 (Elsevier, Amsterdam, 2008), pp.
13–30. doi:10.1016/S1573-4285(08)00202-0
40. V. Médout-Marère, A. El Ghzaoui, C. Charnay, J.M. Douillard, G. Chauveteau, S. Partyka, Sur-
face Heterogeneity of Passively Oxidized Silicon Carbide particles: Hydrophobic-Hydrophilic
partition. J. Colloid Interface Sci. 223(2), 205–214 (2000). doi:10.1006/jcis.1999.6625
41. J.M. Douillard, Concerning the thermodynamic consistency of the "Surface Tension Com-
ponents" equations. J. Colloid Interface Sci. 188(2), 511–515 (1997). doi:10.1006/jcis.1997.
4768
42. J.M. Douillard, J. Zajac, H. Malandrini, F. Clauss, Contact angle and film pressure: study of
a talc surface. J. Colloid Interface Sci. 255(2), 341–351 (2002). doi:10.1006/jcis.2002.8611
43. M.J. Meziani, J. Zajac, J.-M. Douillard, D.J. Jones, S. Partyka, J. Rozière, Evaluation of surface
enthalpy of porous aluminosilicates of the MCM-41 type using immersional calorimetry:
effect of the pore size and framework Si:Al ratio. J. Colloid Interface Sci. 233(2), 219–226
(2001). doi:10.1006/jcis.2002.8611
44. V. Médout-Marère, H. Belarbi, P. Thomas, F. Morato, J.C. Giuntini, J.M. Douillard, Ther-
modynamic analysis of the immersion of a swelling clay. J. Colloid Interface Sci. 202(1),
139–148 (1998). doi:10.1006/jcis.1998.5400
45. M.A. Wilson, A. Pohorille, L.R. Pratt, Molecular dynamics of the water liquid-vapor interface.
J. Phys. Chem. 91(19), 4873–4878 (1987). doi:10.1021/j100303a002
46. W. Drost-Hansen, Structure of water near solid interfaces. Ind. Eng. Chem. 61(11), 10–47
(1969). doi:10.1021/ie50719a005
47. W.D. Harkins, The Physical Chemistry of Surface Films (Reinhold, New York, 1952)
48. Proceedings of BP Symposium on the Significance of the Heats of Adsorption at the Solid-
Liquid Interface. in A.J. Groszek, Sunbury-on-Thames, BP Research Centre (1971)
6 Calorimetry at the Solid–Liquid Interface 267

49. S. Partyka, J.M. Douillard, Nature of interactions between organic pure liquids and model
rocks: a calorimetric investigation. J. Pet. Sci. Eng. 13(2), 95–102 (1995). doi:10.1016/0920-
4105(94)00065-C
50. J.M. Douillard, What can really be deduced from enthalpy of immersional wetting experi-
ments? J. Colloid Interface Sci. 182(1), 308–311 (1996). doi:10.1006/jcis.1996.0468
51. T.W. Healy, D.W. Fuerstenau, The oxide-water interface-Interrelation of the zero point of
charge and the heat of immersion. J. Colloid Sci. 20(4), 376–386 (1965). doi:10.1016/0095-
8522(65)90083-8
52. D.A. Griffiths, D.W. Fuerstenau, The effect of pH and temperature on the heat of immersion of
alumina. J. Colloid Interface Sci. 80(1), 271–283 (1981). doi:10.1016/0021-9797(81)90181-
8
53. M. El Wafir, Approche thermodynamique des interactions entre les liquides et les solides
modeles issus des roches reservoirs de petrole. Ph.D. Thesis, (University of Montpellier 2,
Montpellier 1991)
54. S. Partyka, F. Rouquerol, J. Rouquerol, Calorimetric determination of surface areas: Possibil-
ities of a modified Harkins and Jura procedure. J. Colloid Interface Sci. 68(1), 21–31 (1979).
doi:10.1016/0021-9797(79)90255-8
55. J. Fripiat, J. Cases, M. Francois, M. Letellier, Thermodynamic and microdynamic behavior
of water in clay suspensions and gels. J. Colloid Interface Sci. 89(2), 378–400 (1982). doi:10.
1016/0021-9797(82)90191-6
56. X.-C. Zeng, Y. Chen, X.-N. Chen, J.-Q. Xie, F.-B. Jiang, Thermo-kinetic research method
for faster reactions: modifier method of distorted thermoanalytical curve. Thermochim. Acta
332(1), 97–102 (1999). doi:10.1016/S0040-6031(99)00092-1
57. W. Hemminger, G. Höhne, Calorimetry—Fundamentals and Practice (Verlag Chemie, Wein-
heim/Basel, 1984)
58. W. Zielenkiewicz, E. Margas, Theory of Calorimetry (Kluwer Academic Publishers, Dor-
drecht, 2002)
59. E. Calvet, H. Prat, Microcalorimetrie, Applications Physico-Chimiques et Biologiques (Mas-
son, Paris, 1956)
60. H. Malandrini, Une etude thermodynamique de l’energie superficielle des solides divises :
Determination de la tension superficielle de poudres talco-chloriteuses. Ph.D. Thesis, (Uni-
versity of Montpellier 2, Montpellier, 1995)
61. C. Tanford, The Hydrophobic Effect. Formation of Micelles and Biological Membranes, 2nd
edn. (Wiley, New York, 1980)
62. M.J. Rosen, Surfactants and Interfacial Phenomena, 2nd edn. (Wiley, New York, 1989)
63. J. Lyklema, Adsorption at solid-liquid interfaces with special reference to emulsion sys-
tems. Colloids Surf. A Physicochemical Eng. Aspects 91, 25–38 (1994). doi:10.1016/0927-
7757(94)02718-8
64. R. Denoyel, F. Rouquerol, J. Rouquerol, Thermodynamics of adsorption from solution: Exper-
imental and formal assessment of the enthalpies of displacement. Journal of Colloid and
Interface Science 136(2), 375–384 (1990). doi:10.1016/0021-9797(90)90384-Z
65. W. Rudzinski, J. Narkiewicz-Michalek, R. Charmas MD, Piasecki W, Zajac J, Thermody-
namics of adsorption at heterogeneous solid—liquid interfaces, in Interfacial Dynamics, ed.
by N. Kallay, Surfactant Science Series (Marcel Dekker, New York , 1999), pp. 83–162
66. F.D. Rossini, Chemical Thermodynamics, 3rd edn. (Wiley, New York, 1961)
67. I. Prigogine, A. Bellemans, V. Mathot, The Molecular Theory of Solutions (North-Holland
Publishing Company, Amsterdam, 1957)
68. H.S. Harned, B.B. Owen, The Physical Chemistry of Electrolytic Solutions (Reinhold, New
York, 1958)
69. R.M. Garrels, C.L. Christ, Solutions, Minerals and Equilibria (Freeman, Cooper & Co., San
Francisco, 1965)
70. M.J. Sparnaay, The Electrical Double Layer, vol 4, 1st edn. Properties of Interfaces, (Pergamon
Press, Glasgow, 1972)
268 J. J. Zajac

71. L.K. Koopal, Adsorption of ions and surfactants, in Coagulation and Flocculation: Theory
and Applications, ed. by B. Dobias, vol. 47, Surfactant Science Series, (Marcel Dekker, New
York, 1993), pp. 101–208
72. H.-H. Kohler, Surface charge and surface potential, in Coagulation and Flocculation: Theory
and Applications, ed. by B. Dobias, vol. 47, Surfactant Science Series, (Marcel Dekker, New
York, 1993), pp. 37–56
73. J. Zajac, Adsorption microcalorimetry used to study interfacial aggregation of quaternary
ammonium surfactants (zwitterionic and cationic) on powdered silica supports in dilute aque-
ous solutions. Colloids Surf. A Physicochemical Eng. Aspects 167(1–2), 3–19 (2000). doi:10.
1016/S0927-7757(99)00479-3
74. J. Zajac, C. Chorro, M. Lindheimer, S. Partyka, Thermodynamics of Micellization and Adsorp-
tion of Zwitterionic Surfactants in Aqueous Media. Langmuir 13(6), 1486–1495 (1997).
doi:10.1021/la960926d
75. J. Zajac, A.J. Groszek, Adsorption of C60 fullerene from its toluene solutions on active
carbons: Application of flow microcalorimetry. Carbon 35(8), 1053–1060 (1997). doi:10.
1016/S0008-6223(97)00058-4
76. W. Rudzinski, J. Zajac, I. Dekany, F. Szanto, Heats of immersion in monolayer adsorption
from binary liquid mixtures on heterogeneous solid surfaces: equations for excess isotherms
and heats of immersion corresponding to condensation approximation and Rudzinski-
Jagiello approach. J. Colloid Interface Sci. 112(2), 473–483 (1986). doi:10.1016/0021-
9797(86)90115-3
77. D.H. Everett, Enthalpy and entropy effects in adsorption from solution. J. Phys. Chem. 85(22),
3263–3265 (1981). doi:10.1021/j150622a012
78. G.W. Woodbury Jr, L.A. Noll, Heat of adsorption of liquid mixtures on solid surfaces:
comparison of theory and experiment. Colloids Surf. 8(1), 1–15 (1983). doi:10.1016/0166-
6622(83)80068-7
79. A.J. Groszek, Flow adsorption microcalorimetry. Thermochim. Acta 312(1–2), 133–143
(1998). doi:10.1016/S0040-6031(97)00447-4
80. A.J. Groszek, M.J. Templer, Innovative flow-adsorption microcalorimetry. ChemTech 29(11),
19–26 (1999)
81. R. Denoyel, F. Rouquerol, J. Rouquerol, Interest and requirements of liquid-flow
microcalorimetry in the study of adsorption from solution in the scope of tertiary oil recovery,
in Adsorption from Solution, ed. by C. Rochester (Academic Press, London, 1982), pp. 1–10
82. G.W. Woodbury Jr, L.A. Noll, Heats of adsorption from flow calorimetry: relationships
between heats measured by different methods. Colloids Surf. 28, 233–245 (1987). doi:10.
1016/0166-6622(87)80187-7
83. R. Denoyel, F. Rouquerol, J. Rouquerol, Adsorption of anionic surfactants on alumina: com-
plementarity of the information provided by batch and liquid flow microcalorimetry. Colloids
Surf. 37, 295–307 (1989). doi:10.1016/0166-6622(89)80126-X
84. J.M. Miller, Chromatography: Concepts and Contrasts, 2nd edn. (Wiley, New York, 2005)
85. Z. Kiraly, R.H.K. Borner, G.H. Findenegg, Adsorption and Aggregation of C8E4 and C8G1
Nonionic Surfactants on Hydrophilic Silica Studied by Calorimetry. Langmuir 13(13), 3308–
3315 (1997). doi:10.1021/la9620768
86. S. Partyka, E. Keh, M. Lindheimer, A. Groszek, A new microcalorimeter for the study of
solutions, adsorption and suspensions. Colloids Surf. 37, 309–318 (1989). doi:10.1016/0166-
6622(89)80127-1
87. M. Chorro, C. Chorro, O. Dolladille, S. Partyka, R. Zana, Adsorption mechanism of conven-
tional and dimeric cationic surfactants on silica surface: effect of the state of the surface. J.
Colloid Interface Sci. 210(1), 134–143 (1999). doi:10.1006/jcis.1998.5936
88. R. Chaghi, L.-C. de Ménorval, C. Charnay, G. Derrien, J. Zajac, Interactions of phenol with
cationic micelles of hexadecyltrimethylammonium bromide studied by titration calorimetry,
conductimetry, and 1H NMR in the range of low additive and surfactant concentrations. J.
Colloid Interface Sci. 326(1), 227–234 (2008). doi:10.1016/j.jcis.2008.07.035
6 Calorimetry at the Solid–Liquid Interface 269

89. J.L. Trompette, Contribution de la calorimetrie a l’etude de l’interaction tensioactif cationique

- solide divise. Ph.D. Thesis (University of Montpellier 2, Montpellier, 1995)
90. A.J. Groszek, Graphitic and polar surface sites in carbonaceous solids. Carbon 25(6), 717–722
(1987). doi:10.1016/0008-6223(87)90140-0
91. A.J. Groszek, S. Partyka, Measurements of hydrophobic and hydrophilic surface sites by flow
microcalorimetry. Langmuir 9(10), 2721–2725 (1993)
92. K. Szczodrowski, B. Prélot, S. Lantenois, J.-M. Douillard, J. Zajac, Effect of heteroatom
doping on surface acidity and hydrophilicity of Al, Ti, Zr-doped mesoporous SBA-15. Micro-
porous Mesoporous Mater. 124(1–3), 84–93 (2009). doi:10.1016/j.micromeso.2009.04.035
93. J. Zajac, Mechanism of ionic and zwitterionic surfactant adsorption from dilute solutions
onto charged non-porous and porous mineral oxides inferred from thermodynamic studies,
in Recent Research Developments in Surface and Colloids, ed. by S.G. Pandalai (Research
Signpost, Kerala, 2004), pp. 265–300
94. S. Lantenois, B. Prélot, J.-M. Douillard, K. Szczodrowski, M.-C. Charbonnel, Flow
microcalorimetry: experimental development and application to adsorption of heavy metal
cations on silica. Appl. Surface Sci. 253(13), 5807–5813 (2007). doi:10.1016/j.apsusc.2006.
12.064
95. B. Prelot, S. Lantenois, M.-C. Charbonnel, F. Marchandeau, J.M. Douillard, J. Zajac, What
are the main contributions to the total enthalpy of displacement accompanying the adsorption
of some multivalent metals at the silica–electrolyte interface? J. Colloid Interface Sci. 396,
205–209 (2013) . doi:10.1016/j.jcis.2012.12.049
96. B. Prelot, S. Lantenois, C. Chorro, M.-C. Charbonnel, J. Zajac, J.M. Douillard, Effect of
nanoscale pore space confinement on cadmium adsorption from aqueous solution onto ordered
mesoporous silica: a combined adsorption and flow calorimetry study. J. Phys. Chem. C
115(40), 19686–19695 (2011). doi:10.1021/jp2015885
97. J.L. Trompette, J. Zajac, E. Keh, S. Partyka, Scanning of the cationic surfactant adsorption
on a hydrophilic silica surface at low surface coverages. Langmuir 10(3), 812–818 (1994)
98. R. De Lisi, C. Ostiguy, G. Perron, J.E. Desnoyers, Complete thermodynamic properties of
nonyl- and decyltrimethylammonium bromides in water. J. Colloid Interface Sci. 71(1), 147–
166 (1979). doi:10.1016/0021-9797(79)90229-7
99. R. Atkin, V.S.J. Craig, E.J. Wanless, S. Biggs, Mechanism of cationic surfactant adsorption
at the solid-aqueous interface. Adv. Colloid Interface Sci. 103(3), 219–304 (2003). doi:10.
1016/S0001-8686(03)00002-2
100. J.H. Clint, Surfactant Aggregation (Blackie, Glasgow/London, 1992)
101. R.H. Aranow, L. Witten, The environmental influence on the behavior of long chain molecules.
J. Phys. Chem. 64(11), 1643–1648 (1960). doi:10.1021/j100840a010
102. H. Hoffmann, Fascinating phenomena in surfactant chemistry. Adv. Mater. 6(2), 116–129
(1994). doi:10.1002/adma.19940060204
103. A. Bendjeriou, G. Derrien, P. Hartmann, C. Charnay, S. Partyka, Microcalorimetric studies
of cationic gemini surfactant with a hydrophilic spacer group. Thermochim. Acta 434(1–2),
165–170 (2005). doi:10.1016/j.tca.2005.01.034
104. L. Grosmaire, M. Chorro, C. Chorro, S. Partyka, R. Zana, Alkanediyl-alpha, omega-
bis(dimethylalkylammonium bromide) surfactants - 9. Effect of the spacer carbon number
and temperature on the enthalpy of micellization. J. Colloid Interface Sci. 246(1), 175–181
(2002). doi:10.1006/jcis.2001.8001
105. M. Pisarcik, M.J. Rosen, M. Polakovicova, F. Devinsky, I. Lacko, Area per surfactant molecule
values of gemini surfactants at the liquid-hydrophobic solid interface. J. Colloid Interface Sci.
289(2), 560–565 (2005). doi:10.1016/j.jcis.2005.03.092
106. R. Zana, Dimeric (Gemini) surfactants: effect of the spacer group on the association behavior
in aqueous solution. J. Colloid Interface Sci. 248(2), 203–220 (2002). doi:10.1006/jcis.2001.
8104
107. P. Somasundaran, J.T. Kunjappu, In-situ investigation of adsorbed surfactants and polymers
on solids in solution. Colloids Surf. 37, 245–268 (1989). doi:10.1016/0166-6622(89)80123-
4
270 J. J. Zajac

108. P. Somasundaran, D.W. Fuerstenau, Mechanisms of Alkyl Sulfonate Adsorption at the

Alumina-Water Interface. J. Phys. Chem. 70(1), 90–96 (1966). doi:10.1021/j100873a014
109. M.A. Yeskie, J.H. Harwell, On the structure of aggregates of adsorbed surfactants: the surface
charge density at the hemimicelle/admicelle transition. J. Phys. Chem. 92(8), 2346–2352
(1988). doi:10.1021/j100319a048
110. H. Rupprecht, T. Gu, Structure of adsorption layers of ionic surfactants at the solid/liquid
interface. Colloid Polym. Sci. 269(5), 506–522 (1991). doi:10.1007/bf00655889
111. B. Li, E. Ruckenstein, Adsorption of Ionic Surfactants on charged solid surfaces from Aqueous
solutions. Langmuir 12(21), 5052–5063 (1996). doi:10.1021/la951559t
112. S.D. Christian, J.F. Scamehorn (eds.), Solubilisation in Surfactant Aggregates, Surfactant
Science Series, vol. 55 (Marcel Dekker, New York, 1995)
113. R. Zana, Aqueous surfactant-alcohol systems: a review. Adv. Colloid Interface Sci. 57, 1–64
(1995). doi:10.1016/0001-8686(95)00235-I
114. M.J. Meziani, H. Benalla, J. Zajac, S. Partyka, D.J. Jones, Adsorption of a cationic gemini
surfactant from aqueous solution onto aluminosilicate powders of the MCM-41 type: effect
of pore size and co-adsorption of phenol. J. Colloid Interface Sci. 262(2), 362–371 (2003).
doi:10.1016/S0021-9797(03)00204-2
115. H. Benalla, J. Zajac, S. Partyka, J. Rozière, Calorimetric study of phenol adsolubilisation by
cationic surfactants adsorbed on a flat silica surface or confined within small mesopores of
powdered MCM-41 aluminosilicates. Colloids Surf. A Physicochem. Eng. Aspects 203(1–3),
259–271 (2002). doi:10.1016/S0927-7757(01)01109-8
116. M.N. Khan, Micellar Catalysis. Surfactant Science Series, vol. 133 (Taylor and Francis Group,
Boca Raton, 2006)
117. C.-C. Yu, L. Lobban Lance, Admicellar catalysis. in Surfactant Adsorption and Surface Solu-
bilization, vol. 615, ACS Symposium Series. American Chemical Society, (1996), pp 67–76.
doi:doi:10.1021/bk-1995-0615.ch005.
118. A.D.W. Carswell, E.A. O’Rea, B.P. Grady, Adsorbed surfactants as templates for the synthesis
of morphologically controlled polyaniline and polypyrrole nanostructures on flat surfaces:
from spheres to wires to flat films. J. Am. Chem. Soc. 125(48), 14793–14800 (2003). doi:10.
1021/ja0365983
 Part II
Applications and Case Studies
Chapter 7
Study of Selective Adsorption of Gases
by Calorimetry

Jean-Pierre Bellat

Abstract This chapter is devoted to the study of coadsorption of gases in nanoporous

solids by using the differential calorimetry. In the first part, the thermodynamic
principles of adsorption of gases are recalled. Some of them have already presented
in chapter one. However a special attention has been paid here to the determination of
the adsorption enthalpies and entropies and we focused on the selective adsorption of
binary mixtures. Then the specific experimental technique based on the combination
of differential calorimetry with manometry and gas phase chromatography or mass
spectrometry is shown in details. In the last part, the thermodynamic concepts on
coadsorption are illustrated with experimental results taken from studies on gas
separation by selective adsorption in microporous solids.

7.1 Introduction

Adsorption is a phenomenon, which systematically occurs when a solid is in contact

with a liquid or a gas. The molecules of the adsorptive fluid are accumulated at the pore
surface of the solid to form another phase, called the adsorbate, which has different
physical and chemical properties than the fluid bulk. When several components are
present in the adsorptive fluid, some components can be adsorbed in greater amount
than the other ones. It results that the composition of the adsorbed phase is different
that the composition of the adsorptive fluid. In this case, the adsorption process is
called selective adsorption or competitive adsorption.
The adsorption process is all the more important as the specific surface area of
the solid is large. So, this phenomenon will be predominant when the solid is highly
divided as in nanometric powdered material or highly porous as in microporous or
mesoporous materials.

J.-P. Bellat (B)

Laboratoire Interdisciplinaire Carnot de Bourgogne, UMR 6303 CNRS-Université de
Bourgogne,9 avenue A. Savary, BP 47870, 21078 Dijon, France
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 273

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_7,
© Springer-Verlag Berlin Heidelberg 2013
274 J.-P. Bellat

Adsorption is the first step in the mechanism of reactivity of solids: corrosion,

oxidation of metals, catalytic reactions begin always by an adsorption process. There-
fore, the understanding of the adsorption step is of a relevant interest from a fun-
damental point of view in the study of reactivity of solids. Moreover, adsorption
finds many applications in various fields of investigations as in characterization of
porous solids, gas separation or purification, heterogeneous catalysis, energy storage,
toxic or green-house gas sequestration, gas detection, storage of active molecules…
This concerns several industrial sectors as for example the petroleum chemistry, the
environment, the renewable energy, the electronics, the nuclear energy, the pharma-
cology, the agriculture and food chemistry… This is certainly the reason why we
observe today a real renewed interest for adsorption.
The aim of this chapter is to present, in a simple and didactic way, the bases
of the selective gas adsorption theory and how we can experimentally study the
coadsorption of gases by using the calorimetry technique. It focuses on the energetic
aspect of adsorption and on the selective physisorption of gases in microporous solids
in order to use the selective adsorption for gas separation or gas purification. Thus, the
thermodynamics concepts presented in this chapter are illustrated at the end by two
experimental case studies on the selective adsorption in faujasite zeolites. The former
is devoted to the separation of xylenes isomers, the latter to the desulphurization of
natural gas.

7.2 Definition of Selective Adsorption

7.2.1 Adsorption of Single Component

a. Definition
The adsorption of a single gaseous component on a solid is the accumulation process
of the gas at the surface of this solid as illustrated in Fig. 7.1a [1, 2]. However this
picture is restrictive. Indeed, if we can imagine that the molecules adsorbed on a
homogeneous plane surface form a film with a given thickness, in the case of micro-
porous solids with pores having a size close to that one of adsorptive molecules, the
notion of adsorbed layer has no more physical meaning (Fig. 7.1b). Adsorption is dif-
ferent than absorption. In adsorption, the molecules are localized on the surface where
they can form a film while in absorption the molecules can diffuse inside the solid
structure. Adsorption occurs on a surface, absorption in a volume. The accumulation
of the gas on the surface is due to the fact that the atoms have not the same surround-
ing at the surface than within the solids. The surface atoms can interact with the
molecules coming from the outside by forming physical or chemical bonds in order
to minimize their energy. When the molecules interact with the surface by specific or
non-specific physical interactions (van der Waals interactions and hydrogen bonds),
the adsorption process is called physisorption. When the molecules are bound to
the surface by chemical bonds with electronic transfer and modification of chemical
7 Study of Selective Adsorption of Gases by Calorimetry 275

Channel

Gas phase = adsorptive

Cage

Adsorbed phase = adsorbate

Solid = adsorbent

(a) Plane surface (b) Micropores

Fig. 7.1 Schematic representation of gas adsorption on a plane surface and inside micropores.
a Plane surface. b Micropores

Table 7.1 Main differences between physisorption and chemisorption

Physisorption Chemisorption
Adsorbate-adsorbent interaction Electrostatic Electronic
Interaction energy E ∼ Eliquefaction E >> Eliquefaction
Reversibility of the adsorption process Reversible Non reversible
Kinetics Fast Slow

natures of gas and solid surface, the adsorption process is called chemisorption or
reactive adsorption. The main differences between physisorption and chemisorp-
tion are reported in Table 7.1. The term sorption is often used to describe in a general
manner adsorption, absorption, physisorption and chemisorption without distinction.
b. Adsorbed amount and surface excess amount
As the adsorption process is an accumulation of gas at the solid surface, the gas
concentration decreases progressively with increased distance z from the surface to
the bulk phase (Fig. 7.2).
The adsorbed amount, na , is defined as the amount of matter present in the layer
of thickness t adsorbed at the surface [3]. This amount is given by the relation:

t
n =A
a
cdz (7.1)
0

where A is the specific surface area of the solid. This value corresponds to the white
area (a) below the curve c = f(z) on Fig. 7.2a.
The surface excess amount, nσ , is the amount of matter localized at the surface,
in addition to the amount of gas present in the layer of thickness t. This amount is
then defined by the relation:
276 J.-P. Bellat

(a) Layer model (b) Gibbs representation

c c

(σ)
(a)

cg cg
(g)
0 0
0 t z 0 z
adsorbed amount counted in the
amount n a amount remaining
surface excess
amount n σ gas phase
in the gas phase

Fig. 7.2 Concentration of the gas c as a function of the distance z from the solid surface. Graphic
representation of adsorbed amount and Gibbs surface excess amount. a Layer model b Gibbs
representation

t
σ
n =A (c − cg )dz (7.2)
0

This value is represented by the area (σ) below the curve c = f(z) on Fig. 7.2b.
Adsorbed amount and surface excess amount are two different quantities, which
are often used without clear distinction. We will see later in which conditions these
two values can be considered as equal.
c. Variance and representation of adsorption equilibria
The adsorption equilibrium of a single component i on a solid can be written:

i(g) ↔ i(ads)

The adsorption equilibrium constant is [4, 5]:

γθ
K(T) = (7.3)
p/p◦

with:
• γ the activity coefficient of the adsorbed phase,
• p the vapour pressure at equilibrium,
• p◦ the standard pressure (101 325 Pa),
7 Study of Selective Adsorption of Gases by Calorimetry 277

Fig. 7.3 Pressure range of

pressure
adsorption at constant temper-
ature in the liquid-gas phase
diagram of the adsorptive. The
pressure cannot exceed the liquid
saturated vapour pressure ps
of the fluid at the adsorption gas
temperature Tads
ps
pressure range of
adsorption

Tads temperature

a
• θ = nn a the filling coefficient or surface coverage coefficient, ratio of the amount
s
adsorbed under the pressure p on the amount adsorbed under the saturated vapour
pressure ps of the adsorptive at the adsorption temperature.
Thus K(T) is dimensionless. The adsorption process of a gas on a solid can occur
only in a restricted pressure range limited by the saturated vapour pressure of the
component i at the same temperature as the adsorption temperature. If the vapour
pressure is above ps then adsorption is replaced by liquefaction (Fig. 7.3). The satu-
ration of the adsorbent is established when p is equal to ps . We often use the notion
of relative pressure defined by the ratio p/ps . Its value lies between 0 and 1. When
the relative pressure is equal to one, the filling coefficient is also equal to one.
The variance of the adsorption equilibrium can be regarding as the number of
intensive variables that must be set to know completely the thermodynamic state of
the system. It can be calculated by using the well-known phase rule of Gibbs [6]:

v=N−r+P−ϕ (7.4)

where N is the number of components in the system, r the number of independent

equilibria between phases, P the number of intensive variables and ϕ the number of
phases.
In the case of adsorption of a single component, N, r and ϕ are easily determined:

N = 3 (adsorbent, gas and adsorbate),

r = 1 (adsorption equilibrium),
ϕ = 3 (adsorbent, gas and adsorbate).

The determination of the number of intensive variables P is less evident. Indeed,

if the pressure and the temperature must obviously be taken into account, a new
intensive variable must be introduced, the spreading pressure π. As a gas enclosed
in a bulb of volume V creates a pressure due to the collision of gas molecules against
the wall, the molecules adsorbed at the surface. A exert a bi-dimensional pressure
(Fig. 7.4). As for the gas the elementary mechanical work is given by −pdV, the
278 J.-P. Bellat

Adsorbed phase Gas phase phase

π : spreading pressure p: pressure
(bi-dimensional space) (tri-dimensional space)
Mechanical work of surface strength: Mechanical work of volume strength:
δW = -πdA δW = -pdV

Fig. 7.4 Illustration of the spreading pressure exerted by molecules adsorbed on a surface compared
to the pressure exerted by a gas in a volume

elementary mechanical work of surface strength is given by the term − πdA. The
physical meaning of spreading pressure is analogous to that of surface tension of a
monomolecular film at the gas–liquid interface.
The spreading pressure defines the lowering of the surface tension at the gas–solid
interface upon adsorption π = σ◦ − σ, where σ◦ and σ are the surface tensions of
the clean and monolayer covered surfaces, respectively. We will show below that the
spreading pressure is related to the adsorbed amount. Thus, the number of intensive
variables is P = 3 (T, p and π) and the variance of the adsorption equilibrium is
equal to 2. The system is “bivariant” and has two degrees of freedom. This means
that the spreading pressure or the adsorbed amount depends on the temperature and
the pressure. Thus, as illustrated in Fig. 7.5, there are three manners to represent an
adsorption equilibrium by plotting:
• an adsorption isotherm na = f(p)T
• an adsorption isobar na = f(T)p
 
• an adsorption isoster p = f(T)na or rather Ln(p) = f T1 na .
The adsorption isotherm is the representation, which is the more often used.

7.2.2 Adsorption of Gas Mixtures

a. Definition
Selective adsorption so-called competitive adsorption occurs when a gas contain-
ing more than one component is in interaction with a solid in such a way that the
7 Study of Selective Adsorption of Gases by Calorimetry 279

ISOTHERM ISOBAR ISOSTER

T1

Adsorbed amount n a
Adsorbed amount n a

Ln(Pressure)
T2
n 1a

T3 p1 n a2
p2
p3 n a3
Pressure p Temperature T 1/Temperature

Fig. 7.5 The three representations of adsorption equilibrium of gas on solid at constant temperature
(isotherm), constant pressure (isobar) and constant adsorbed amount (isoster)

yA and yB : mole fractions in the gas phase

x A and xB : mole fractions in the adsorbed phase

Solid = adsorbent

Fig. 7.6 Schematic representation of the adsorption of a binary mixture A (red) + B (blue) on a
solid. The adsorbate is rich in component A, the gas phase in component B. The adsorption process
is in favour of component A

composition of the gas is different than the composition of the adsorbate. This implies
that one component is more adsorbed by the solid than the other ones. For example,
the Fig. 7.6 is an illustration of a selective adsorption process of two components
A and B where the component A is favourably adsorbed. The adsorption process is
called to be “selective for A”.
b. Variance and representation of coadsorption equilibria
The adsorption equilibrium of binary mixture A + B on a solid can be written:

A(g) + B(ads) ↔ A(ads) + B(g)

The equilibrium constant is defined by:

γA θA pB KA
K(T) = = (7.5)
γB θB pA KB

with:
• γi the activity coefficient of component i in the adsorbed phase,
• θi the partial filling coefficient of component i
280 J.-P. Bellat

• pi the partial pressure of component i in the gas phase.

• Ki the adsorption equilibrium constant of single component i.
There are seven thermodynamics parameters characterizing this equilibrium (p, T,
π, yA , y B , xA , and xB ). In that case, the application of the phase rule of Gibbs leads
to a variance equal to 3. The representation of the binary adsorption equilibrium is
obviously more complicated than for the adsorption of a single component. Coad-
sorption equilibrium could be represented in a three dimensional space by plotting
for example the amounts adsorbed as a function of the pressure in one direction and
the composition of the adsorbate in another one. However, this kind of plot is not
common because it requires a lot of data measured at different composition and pres-
sure. We prefer to represent the coadsorption equilibria in a two-dimension space by
maintaining constant certain variables. The representations that are the more often
used and easily understandable are:
• the total and partial isotherms at constant initial composition of the gas phase:
na = f(p)T,yio and nia = f(p)T,yio
• the x-y diagram at constant temperature and pressure: xi = f(yi )T,p
• the adsorption selectivity as a function of the adsorbed amount at constant initial
composition of gas mixture: α = f(na )T,yio
• the coadsorption heat and the molar entropy of the adsorbate as a function of the
adsorbed amount at constant initial composition of gas mixture: r Ha = f(na )T,yio
and Sam,i = f(na )T,yio , respectively.
These different representations are illustrated in Fig. 7.7.
c. Adsorption selectivity
The adsorption selectivity is the more appropriate parameter to describe the prefer-
ential adsorption of one component with respect to another one. For the adsorption
equilibrium of a binary mixture,

A(g) + B(ads) ↔ A(ads) + B(g)

the adsorption selectivity is defined by the ratio:

xA /yA
α = (7.6)
A/B xB /yB

with xi and yi the mole fractions of component i in the adsorbed phase and the
gas phase, respectively. This parameter is analogous to a separation factor of two
components between two phases. It may be pointed out that the selectivity is the
equilibrium constant of the binary adsorption equilibrium in the case of an ideal
adsorbed solution. We have only to replace in Eq. 7.5 the partial filling coefficients
and pressures by their relation with the mole fractions of each component in the
adsorbed and gas phases (θi = xi θ and pi = yi p) and to take the activity coefficients
equal to unity.
7 Study of Selective Adsorption of Gases by Calorimetry 281

COADSORPTION
ENTHALPY ENTROPY
ISOTHERM
y oi = cte y oi = cte y oi = cte
adsorbed amount n a

A+B

enthalpy Δ r H a

entropy S am,i
A

pressure p adsorbed amount na adsorbed amount n a

x-y DIAGRAMM SELECTIVITY

y oi = cte
Adsorption No preferential
in favor of B p = cte adsorption

selectivity αA/B
yA (gas phase)

selective for A

Adsorption
in favor of A selective for B

x A (adsorbed phase) adsorbed amount na

Fig. 7.7 Different representations of coadsorption equilibrium which are often used for a binary
mixture

When the solid adsorbs two components A and B in the same proportions, the
composition of the mixture in the gas and adsorbed phases are the same and the
selectivity is equal to unity. The adsorption process is not selective. If the solid adsorbs
preferentially the component A, the mole fraction of this component is higher in the
adsorbed phase than in the gas phase and the selectivity is higher than unity. On the
opposite, if the adsorption process is in favour of component B, the selectivity is
lower than unity. Seeing that the selectivity is the equilibrium constant in the case
of an ideal adsorbed solution, it must be constant whatever the composition. The
fact that the selectivity changes with the composition indicates that the adsorption
process is not ideal.
d. Different types of selective adsorption
The selective adsorption of several components on a solid is a very complex phenom-
enon. The reason why, an adsorption process is in favour of one component rather
than another one is far to be elucidated. In practical applications, we consider there
are three types of separation by selective adsorption (Fig. 7.8):
• the steric exclusion, which occurs when the size of one component in the mixture
is higher than the pore opening of the adsorbent. The small molecules can be
adsorbed in the microporosity while the biggest remain in the gas phase. This is a
real molecular sieving with a selectivity that tends to infinity.
282 J.-P. Bellat

Steric exclusion: αA → ∞
(molecular sieving) B

KA DA
Kinetic separation: αA =
B KB DB

K A xA y B
Energetic separation: αA = =
B K B y A xB

Fig. 7.8 The three types of separation by selective adsorption [Ki is the adsorption equilibrium
constant and Di is the diffusion coefficient of component i]

• the kinetic separation based on the difference of diffusivity of components in the

microporosity. All the molecules of the mixture can pass through the adsorbent
but some of them diffuse faster than the others. In that case the expression of the
selectivity is different than in Eq. 7.6. It takes into account the diffusion coefficient
of each component in the microprosity.
• the energetic separation which results of difference in adsorption affinity of the
adsorbent towards the components of the mixture. All the molecules can penetrate
the microporosity but are adsorbed with different energies. At equilibrium, the
component adsorbed in favour is that one having the strong adsorption energy
with the solid.

7.3 Adsorption Enthalpies and Entropies

Adsorption is a spontaneous and exothermic process. The adsorption Gibbs energy

and the adsorption enthalpy are then negative. As the molecules are adsorbed in
micropores and bond to the solid by physical or chemical interactions, they lose
degrees of freedom. The adsorbate forms a phase, which is more ordered compared
to the gas phase. Therefore the adsorption entropy is negative too. Thus the sign
of the Gibbs energy depends on the enthalpic term which is always negative and
the entropic term which is positive: r Ga = r Ha − Tr Sa . In most cases, the
adsorption process is spontaneous owing to the enthalpic term, which is generally
more important than the entropic one. However, the weight of the entropic term
must not be neglected because it can compensate the enthalpic term, especially when
adsorption occurs in very confined spaces like nanopores. One of particularities of
the adsorption process is that the adsorption enthalpy and entropy depend on the
adsorbed amount and, sometimes, the composition of gas mixture. The knowledge
7 Study of Selective Adsorption of Gases by Calorimetry 283

of these two thermodynamics values, as a function of the adsorbed amount, is then

of a relevant interest for the understanding of adsorption mechanisms.

7.3.1 Adsorption Enthalpy

Adsorption enthalpy versus adsorbed amount gives interesting information about

molecular interactions between the solid surface and the adsorbed molecules
(adsorbate-adsorbent interactions) and between the adsorbed molecules themselves
(adsorbate-adsorbate interactions). The adsorbate-adsorbent interactions depend on
the energetic properties of the surface. If all the adsorption sites have the same energy,
the adsorbate-adsorbent interactions are almost constant whatever the surface cov-
erage. Each molecule interacts in the same way with each site. On the opposite, if
the adsorption sites are very different in energy, the adsorbate-adsorbent interactions
decrease with the surface coverage. First, the adsorption occurs on the more ener-
getic sites and then goes on the less energetic ones. As for the adsorbate-adsorbate
interactions, they are always increasing with the surface coverage. If they can be
neglected at the beginning of the adsorption process, they are maximal at complete
surface coverage and can be very important especially when adsorption occurs in
nanopores in which the molecules are very confined. The adsorption enthalpy is the
sum of these two kinds of interactions and the shape of the plot adsorption enthalpy
versus adsorbed amount will be different according to the adsorption process. As
illustrated on Fig. 7.9, adsorption on a heterogeneous surface is characterized by a
sharp decrease of the adsorption enthalpy at very low filling while this value is almost
constant when the surface is homogeneous. An increase of the adsorption enthalpy
at high filling indicates the presence of adsorbate-adsorbate interactions. In all cases,
at saturation, when the adsorption process is complete, the adsorption enthalpy tends
to the liquefaction enthalpy. It is not possible to distinguish adsorbate-adsorbent
and adsorbate-adsorbate interactions from the adsorption enthalpy. However, if we
know that the surface is homogeneous, the adsorbate-adsorbate interactions can be
estimated by subtracting to the adsorption enthalpy its value extrapolated at zero
filling. Moreover, in the case of coadsorption, the adsorption enthalpy curve can
show additional changes resulting from endothermic or exothermic effects consec-
utive to displacement of a species by another one or specific interactions between
components in the adsorbed phase.

7.3.2 Adsorption Entropy and Molar Adsorbate Entropy

Adsorption entropy and molar adsorbate entropy are related by the relation:

g◦
Sam = r Sa + Sm (7.7)

g◦
where Sm is the standard molar entropy of the gaseous adsorptive.
284 J.-P. Bellat

HETEROGENEOUS SURFACE HOMOGENEOUS SURFACE

- Adsorption Enthalpy: - ΔH a
- Adsorption enthalpy: - ΔH a

I total
Itotal

I aA IaA

ΔH liquefaction ΔH liquefaction

I aa I aa

Adsorbed amount n a Adsorbed amount na

Fig. 7.9 Change of the adsorption enthalpy as a function of the adsorbed amount for an adsorption
process energetically homogeneous or heterogeneous. Iaa is the adsorbate-adsorbate interaction, IAa
is the adsorbate-adsorbent interaction and Itotal = Iaa + IaA corresponds to the adsorption enthalpy

The knowledge of molar adsorbate entropy compared to molar entropies of solid,

liquid or gaseous adsorptive is particularly interesting when represented as a function
of the adsorbed amount. These values give interesting information about the physical
state of the adsorbate and allow us to know if the molecules adsorbed on the solid
surface flutter or are rather frozen around their adsorption sites. It is currently admitted
that the adsorbed phase is in a physical state similar to a liquid particularly at complete
filling. Thus, to calculate the adsorbed volume we used the density of the liquid fluid.
However the determination of the molar adsorbate entropy shows that this hypothesis
is not always true. Indeed, if in the case of a mobile adsorption, the molar adsorbate
entropy is often close to the molar entropy of the liquid adsorptive, for a localized
adsorption, it can be near to that one of the solid adsorptive, the molecules having
lost numerous modes of vibration. This result is particularly true when the molecules
are adsorbed in narrow nanopores. Figure 7.10 shows two examples of molar entropy
curve corresponding to these two cases. However, the entropy curves have often more
complicated shapes that make their interpretation very difficult.

7.4 Calculation of Adsorption Enthalpy and Entropy

from Single Adsorption Isotherms

The adsorption enthalpie and entropie of single component can be extracted from
adsorption data obtained under isothermal conditions by using two different methods.
The first one is based on the van’t Hoff equation and the second one on the Clausius-
Clapeyron equation. These methods only need a set of adsorption isotherms measured
at various temperatures.
7 Study of Selective Adsorption of Gases by Calorimetry 285

LOCALIZED ADSORPTION MOBILE ADSORPTION

S°gas S°gas

Molar entropy: S am
Molar entropy: S am

S°liquid S °liquid

S°solid S °solid

Adsorbed amount na Adsorbed amount na

Fig. 7.10 Change of the molar adsorbate entropy as a function of the adsorbed amount for localized
and mobile adsorption process. S◦ is the standard molar entropy of the adsorptive in the gas, liquid
or solid state

7.4.1 Van’t Hoff Method

This method consists to find the thermodynamic model (Henry, Langmuir, Fowler,
Freundlich…) which gives the best fit with adsorption isotherms measured at various
temperatures [7]. The parameter of the model Kmodel is temperature dependent. It
can be related to the dimensionless equilibrium constant of adsorption K(T) by the
relation:
K(T)
Kmodel = ◦ (7.8)
p

where p◦ is the standard pressure (p◦ = 101325 ∼ 105 Pa).

At equilibrium we have:

r Ga = r Ha − Tr Sa = −RTLnK(T ) (7.9)

and
r Ha 1 r Sa
LnK(T) = − + (7.10)
R T R
After differentiation of this equation we retrieve the well-known van’t Hoff equation:
 
dLnK(T)
r H = RT
a 2
(7.11)
dT

Thus, by plotting the curve Ln K (T) = f(1/T) it is easy to determine the adsorption
enthalpy and entropy from the slope and the intercept of this straight-line (Fig. 7.11a).
In this method we have to keep in mind that we consider that the adsorption enthalpy
and entropy are constant i.e. independent on the temperature and the adsorbed
286 J.-P. Bellat

amount. In fact this is not true. However, this is a simple way that is often used
to roughly estimate the adsorption enthalpy and entropy.

7.4.2 Isosteric Method So-called Clausius-Clapeyron Method

Let us consider the adsorption equilibrium:

gas ↔ adsorbate

If equilibrium exists between gas and adsorbed phases we have, for a given adsorbed
amount, equality in chemical potentials:

μg (T, p) = μa (T, p, na ) (7.12)

The chemical potential of the gas assumed as ideal is defined by:

◦ p
μg (T, p) = μg (T) + RT Ln (7.13)
p◦
◦
where μg is the standard chemical potential of the gas at the temperature T.
The chemical potentials can be expressed as a function of molar enthalpies and
entropies as follows:
◦ g◦ g◦
μg (T) = Hm (T) − TS m (T) (7.14)

μa (T, p, n a ) = Ham (T, p, n a ) − TS am (T, p, n a ) (7.15)

By combining Eqs. 7.13, 7.14 and 7.15 we get:

  g◦ g◦
p Ha (T,p,na ) − Hm (T) 1 Sa (T, p, na ) − Sm (T)
Ln = m − m (7.16)
p◦ na R T R

If we define the isosteric enthalpy and entropy by:

g◦
Hiso (na ) = Ham (T,p,na ) − Hm (T) (7.17)

g◦
Siso (na ) = Sam (T,p,na ) − Sm (T) (7.18)

we obtain:  
p Hiso (na ) 1 Siso (na )
Ln = − (7.19)
p◦ na R T R
7 Study of Selective Adsorption of Gases by Calorimetry 287

Adsorption isotherms
(a) van’t Hoff method
na T1
Thermodynamic model of adsorption:
T2
Henry: n a = K Hp
T3 K Lp
Langmuir: n a = n as
1 + K Lp
pi p

(b) Isosteric method

⎛p⎞
Ln ⎜⎜ ⎟⎟ 1 LnK (T) 1
⎝ p° ⎠ n a T T
gas ⇔
liquid

ΔS a
intercept = − − ΔH a
R gas slope =
⇔ ΔS a R
intercept =
ad
sor
ba
na2 R
te
ΔH a
slope =
R n1a

Fig. 7.11 Graphic determination of adsorption enthalpy and entropy from isothermal adsorption
data measured at various temperatures. a Van’t Hoff method applied to the adsorption equilibrium
constant deduced from the parameter of a thermodynamic model like Henry or Langmuir fitting
well the adsorption isotherms. b Clausius-Clapeyron or isosteric method: the isosters lie below a
limit line, which corresponds to the liquefaction equilibrium of the adsorptive

 
Thus the adsorption enthalpy and entropy can be determined by plotting Ln pp◦ a =
  n
f T1 at constant adsorbed amount. The slope and the intercept of this plot called
“adsorption isoster” give Hiso (na ) and Siso (na ) as illustrated in Fig. 7.11b. We
need to apply this method a series of adsorption isotherms in a large range of pressure
and temperature in order to explore the complete domain of adsorbed amounts.
This second method has the advantage to take into account the dependence of the
adsorbed amount on the adsorption enthalpy and entropy and allows a more accurate
determination of these values than the van’t Hoff method.
However we consider that the enthalpy and entropy are not function of the tem-
perature. It may be noticed that by differentiating Eq. 7.19 we obtain:

dLnp Hiso (na )

=− (7.20)
dT RT2
This relation is analogous to the Clausius-Clapeyron equation characterizing the
equilibrium between a liquid and a gas.
288 J.-P. Bellat

7.5 Determination of Coadsorption Enthalpy

and Entropy by Calorimetry

If the van’t Hoff and isosteric methods are simple ways for estimating the adsorption
enthalpy of single component from isothermal adsorption data, they have the dis-
advantage to not take into account the temperature dependence on the enthalpy and
entropy and to be not enough accurate. Moreover they are not adapted to the adsorp-
tion of gas mixtures. The best mean to determine the adsorption and coadsorption
enthalpy is to measure them by using a differential calorimetry technique coupled
with others techniques allowing the measure of adsorbed amount and composition
as for example the manometry and the chromatography.

7.5.1 Experimental Calorimetric Technique

a. Description of the experimental device

The experimental technique, which is particularly well adapted for determining the
molar adsorption enthalpy of a single component or a mixture as a function on the
adsorbed amount at constant temperature, is the differential calorimetry coupled
with a manometry and a gas phase chromatography or a mass spectrometry [8].
A schematic representation of the experimental device is shown on Fig. 7.12. The
calorimeter is a Setaram C80 differential calorimeter type Calvet. Composed of
a sample cell containing the adsorbent and a reference cell surrounded by a vast
number of thermocouples, it allows measuring the heat flow being given off during
the adsorption of a small dose of gas put in contact with the adsorbent. The amount
of gas introduced in the calorimeter is controlled with the manometric device by
measuring the pressure of gas mixture in glass bulbs with calibrated volumes VS
and VF . The composition of the gas is analyzed with the gas chromatograph or the
mass spectrometer. A bypass system connected to the vacuum line (turbomolecular
pump and dry primary pump) and composed of a six-way gas injection valve (GIV)
with sample loops allows the collection of a small dose of gas under low pressure for
analysis. The pressure is measured with several MKS Baratron pressure sensors in
the range 10−4 –103 hPa. This apparatus can be used to study the adsorption of single
component as well as mixtures. It simultaneously allows the determination of the total
and partial adsorption isotherms at constant initial gas composition, the coadsorption
enthalpy and the adsorption selectivity as a function of the total adsorbed amount.
b. Operating procedure
After activation by heating under dynamic vacuum, the sample is brought to the
adsorption temperature. The gas mixture is prepared in the flask of volume Vs by
introducing each component of the mixture, one after the other one, while controlling
the partial pressures. The flask is then slightly heated in order to homogenize the
mixture by convection. The composition of the mixture is controlled by GPC or MS.
Coadsorption isotherms are obtained by successive introduction of small doses of
7 Study of Selective Adsorption of Gases by Calorimetry 289

p Sample loop
p Vacuum
GPC

GAS ANALYZIS
MANOMETRY

VS VF GIV

He Vacuum MS

A B

Reference cell Sample cell

Data
recorder

CALORIMETRY

sample loop

Load position Injection position

vacuum He

sample in column

Gas Injection Valve

Fig. 7.12 Schematic representation of the calorimeter coupled with a manometric device and a gas
analysis technique used to study the coadsorption of gas. Details of the connection of the six-way
valve used for GPC or MS sampling

gas mixture stored in bulb V S first into the flask VF , previously evacuated and then
in the adsorption cells. The pressure is recorded and the gas composition is analyzed
before and after each step.

7.5.2 Measurement of Adsorbed Amounts

In the manometry technique coupled with CPG or MS, the determination of the
amount of each component i adsorbed (nai ) consists to perform a mass balance on
the gas phase from the total pressure and the mole fraction of component i in the gas
phase, measured before and after adsorption [9]. As the system is closed, we consider
that the amount of matter leaving the gas phase is equal to that one adsorbing on the
solid.
First, let us consider the adsorption of a single component in a closed system
(Fig. 7.13). The volume of the system containing the adsorbent previously evacuated
290 J.-P. Bellat

Before adsorption At equilibrium

gas gas
nog
ng
closed system
po, T, V p, T, Vg
° at constant V and T

adsorbate
solid solid

Fig. 7.13 Equivalent thermodynamic system for measuring the adsorbed amount by manometry.
Vo : dead volume of reactor (known from calibration with helium); po and p: pressure of gas before
and after adsorption (measured); V g = volume of gas at equilibrium (unknown)

is Vo . This volume can be precisely calibrated with helium by using a pycnometric

technique. We introduced in the system a small dose of gas under the initial pressure
po . At equilibrium, the pressure of the gas is p. The volume of the gas phase is
Vg < V◦ because the adsorbed phase occupies a finite volume in the system. Let us
assume the gas as ideal, the amounts of gas and adsorbed phase before adsorption
and at equilibrium are:
A(g) → A(ads)
g Vo
Initial condition no = p 0
RT o
Vg
Equilibrium ng = p na
RT
The adsorbed amount is given by:

Vo Vg
na = po − p (7.21)
RT RT
p
As for an ideal gas the concentration is defined by c = , Eq. 7.21 becomes:
RT
g
na = co Vo − cVg (7.22)

c◦ , c and V◦ are experimentally measured but Vg is unknown. As the volume is

constant,
Vo = Va + Vg (7.23)

we can write:
g
(co − cg )Vo = na − cg Va (7.24)

By introducing Eq. 7.1 in the right hand side and as the volume of the adsorbed phase
is
7 Study of Selective Adsorption of Gases by Calorimetry 291

t
V =Aa
dz (7.25)
o

Eq. 7.24 becomes:

(co − cg )Vo = nσ
g
(7.26)

In the manometric technique, the measure of the concentration (pressure) of the gas
before and after adsorption does not lead to the adsorbed amount but to the surface
excess. However, if the volume of the adsorbed phase is small compared to that one
of the system (Va << Vg ) and if the concentration of the gas is not too much high,
Eq. 7.24 can be reduced to:
g
na = Vo (co − cg ) (7.27)

Thus, the surface excess can be assimilated to the adsorbed amount if the volume of
the reactor is enough large and the pressure enough low.
In the case of the adsorption of a gas mixture, we have just to introduce in Eq. 7.27
the composition of the gas before adsorption and after adsorption. Let us define by
yio and yi the mole fraction of the component i in the gas phase before adsorption
and at equilibrium, respectively, the adsorbed amount of each component i is given
by:
g
nai = Vo (yio co − yi cg ) (7.28)

with obviously:
nai = natotal (7.29)
i

Practically, the system composed of a volume V◦ containing the adsorbent under

the initial pressure po , as illustrated in Fig. 7.13, is not conceivable because the
gas is adsorbed as soon as it is in contact with the solid. It is the reason why the
manometric device shown in Fig. 7.12 is composed of two volumes: the volume Vc
of the calorimetric cells, which contains the adsorbent and the volume VF , which is
used to prepare the dose of gas. By applying the Eq. 7.28 to this system, we find that
the amount of each component of gas mixture adsorbed on the solid at the adsorption
step j is given by the relation:

VF   Vc  
nai,j = yio,j po,j − yi,j pj − yi,j pj − yi,j−1 pj-1 (7.30)
RTF RTc

TF and TC being the temperature of volume VF and VC , respectively.

The total amount of component i adsorbed since the first step and expressed per
unity of mass of solid is:
1 a
j
Nai,j = ni,j (7.31)
m
j=1
292 J.-P. Bellat

and the total amount of mixture adsorbed after j adsorption steps is:

i
Natotal, j = Nai,j (7.32)
i=1

where m is the mass of the adsorbent.

Then, the mole fraction of component i in the adsorbed phase can be calculated
by mean of the relation:
Nai,j
xi,j = a (7.33)
Ntotal,j

Finally, the adsorption selectivity at the adsorption step j can be calculated with
Eq. 7.6.

7.5.3 Measurement of Differential Adsorption Enthalpy

The calorimetry technique coupled with manometry shown in Fig. 7.12 allows the
measure of the heat emanated by the adsorption of a given amount of gas on the
solid. In thermodynamics it is essential to know at what function of state corresponds
this adsorption heat. This depends on the calorimetric technique and experimental
conditions used.
When an amount of gas is introduced in the calorimetric cells, two energetic
exchanges must be taken into account: thermal energy due to the adsorption and
mechanical energy due to the expansion of the gas in the cells. Thermodynamically,
the operation, which consists to introduce a given amount of gas into a vessel contain-
ing or not the adsorbent, is equivalent to the compression of gas enclosed in a cylinder
by moving the piston [10]. This equivalent thermodynamic system is represented in
Fig. 7.14.
Let us consider the adsorption of a single component in the sample cell. The
amount of matter present in the system and their molar internal energies are defined
as below:

gas ↔ adsorbate
g
before adsorption no , ug , po 0

after adsorption ng , ug , p na , ua

with ug and ua the molar internal energy of gas and adsorbate, respectively.
The change in internal energy is:
g
Us = na ua + ng ug − no ug (7.34)
7 Study of Selective Adsorption of Gases by Calorimetry 293

piston

piston
gas gas
ngo
ng
na
adsorbent adsorbent

Before adsorption After adsorption

Fig. 7.14 Thermodynamic system composed of a cylinder closed by a frictionless piston equivalent
to the sample cell of the manometric system. For the reference cell, the equivalent system is the
same but without adsorbent

As the system is closed:

g
no = na + ng (7.35)

we have:
Us = na (ua − ug ) (7.36)

Then, by derivation of Eq. 7.36 we obtain the differential energy:

dUs = (ua − ug )dna + na dua = δQs + δWs (7.37)

The differential mechanical work exchanged with the outside world, assuming the
transformation as reversible, is expressed by:

δWs = −pdVs (7.38)

Let us consider the gas as ideal (pVs = ng RT) and as the system is close (dng =
−dna ), this equation becomes:

δWs = Vs dp − dng RT = Vs dp + dna RT (7.39)

Thus the differential heat in the sample is:

δQs = dUs − δWs (7.40)

a
δQs = (ua − ug )dn + na dua − RTdna − Vs dp (7.41)
294 J.-P. Bellat

Let us consider now the reference cell in which we introduced the same amount of
gas as in the sample cell. As the internal energy of an ideal gas depends only on
the temperature, at constant temperature the differential internal energy of gas in the
reference cell is null:
dUr = δQr + δWr = 0 (7.42)

The system being closed and isothermal, we have:

pVr = ng RT = constant (7.43)

pdVr + Vr dp = 0 (7.44)

Thus the differential heat in the reference cell is:

δQr = −Vr dp (7.45)

The adsorption heat measured by differential calorimetry is equal to the difference

between the heat emanated in the sample cell and that one emanated in the reference
cell.
δQa = δQs − δQr (7.46)

By introducing Eqs. 7.41 and 7.45 in this relation we obtain:

δQa = (ua − ug )dna + na dua − RTdna + (Vr − Vs )dp (7.47)

Let us define now the molar internal energy of the adsorbate at the adsorbed amount
na by:
dua
ua (na ) = ua + na a (7.48)
dn
We obtain:
δQa = ua (na )dna − ug dna − RTdna + (Vr − Vs )dp (7.49)

If the volume of sample and reference cells are identical and larger than that one of
the adsorbent we can assume:

(Vr − Vs )dp = 0 (7.50)

Therefore the molar heat of adsorption measured by differential calorimetry is equal

to:
δQa
Qam = a = ua (na ) − ug − RT = r Uam − RT (7.51)
dn

Qam is called the differential calorimetric heat of adsorption and r Uam the differential
molar internal energy of adsorption. They depend on the adsorbed amount na .
7 Study of Selective Adsorption of Gases by Calorimetry 295

The differential molar internal energy is related to the differential molar enthalpy
by:
r Ham = r Uam + pr V (7.52)

For the adsorption equilibrium gas ↔ adsorbate, the change in molar volume is equal
to:
g
r V = Vam − Vm (7.53)

Let us consider the molar volume of the adsorbed phase as negligible compared to
that one of the ideal gas we have:

g RT
r V ≈ −Vm = − (7.54)
p

Finally, by combining Eqs. 7.51, 7.52 and 7.54 we obtain:

Qam = r Ham (7.55)

Thus, the molar adsorption heat measured by differential calorimetry at constant

temperature corresponds to the differential molar adsorption enthalpy. This value is
equivalent to the isosteric enthalpy of adsorption [11].
Practically, when an adsorption step j is performed at constant temperature
between the initial state j-1 defined by the couple of parameters pj-1 and Naj−1 and the
final state j defined by the couple pj and Naj as illustrated on Fig. 7.15, the calorimetric
Naj−1 +Naj
molar adsorption heat at the mean total adsorbed amount 2 is calculated by
the relation:

Adsorption isotherm Calorimetric signal

Adsorbed amount N a

Heat flux J

a Q aj
Nj

a
N j-1

p j-1 pj
Time t
Pressure p

Fig. 7.15 Schematic representation of an isothermal adsorption step from initial equilibrium state
j-1 to final equilibrium state j and corresponding thermal effect measured by differential calorimetry
296 J.-P. Bellat

Qaj Naj ua (Naj ) − Naj−1 ua (Naj−1 )

Qam,j = = − ug − RT (7.56)
Naj − Naj−1 Naj − Naj−1

where Qaj is the heat emanated by the adsorption of the Naj − Naj−1 moles of gas per
unity of mass of solid. Qaj is measured by integration of the calorimetric signal giving
the heat flux J as a function of time (Fig. 7.15):

t
Qaj = Jj dt (7.57)
0

Experimentally, the difference Naj −Naj−1 must be as small as possible in order that the
calorimetric adsorption heat can be considered as a differential value. This implies
to have accurate pressure sensors particularly for exploring the low adsorbed amount
range where the pressure variations between the initial and final states can be very
small.
It may be noted that some authors [3] use a different procedure to determine
the molar differential enthalpy of adsorption by mean of a differential calorimeter
coupled with a manometric device. They prefer to work with only one calorimetric
vessel connected to the manometric device, the reference cell being maintained under
vacuum. In these conditions we obtain from Eq. 7.48 that the molar heat of adsorption
is equal to:
δ Qa dp
Qam = a = r Ham − Vs a (7.58)
dn dn
This procedure needs to know precisely the volume Vs of the sample cell, which
is located within the calorimetric detector and to reduce the dead volume, which is
outside the detector. Vs can be estimated by liquid weighting or from geometrical
measurements and by taking into account the volume occupied by the sample.
Let us consider now the coadsorption of two gases or more, the definition of the
calorimetric heat is exactly the same as for the adsorption of a single component. In
this case, it corresponds obviously to the differential molar enthalpy of coadsorption.
It is not possible to measure directly by calorimetry the differential enthalpy of
adsorption of each component present in the mixture. Thus, for the coadsorption of
two components A and B, the molar calorimetric coadsorption heat is equal to the
molar differential enthalpy of coadsorption:

Qam,A+B = r Ham,A+B (7.59)

However, if the adsorbed phase can be considered as an ideal solution, in which

the molecular interactions are the same as for the adsorption of single components,
it is possible to calculate the molar differential coadsorption enthalpy by mean of
the relation:
∗ ∗
r Ham, A+B = xA r Ham, A + xB r Ham, B (7.60)
7 Study of Selective Adsorption of Gases by Calorimetry 297

∗
where r Ham,i is the molar differential enthalpy of adsorption of single component i
at the same adsorbed amount as in the mixture.

7.5.4 Adsorption Gibbs Energy

a. Gibbs adsorption isotherm

For a three-dimensional fluid phase like, for example, a gas enclosed in a volume,
the free enthalpy so-called Gibbs energy depends on the temperature, the pressure
and the amount of gas: Gg = f(T, p, ng ). The differential Gibbs energy is:
     
∂Gg ∂Gg ∂Gg
dGg = dT + dp + dng (7.61)
∂T p,ng ∂p T,ng ∂ng T,p

and we obtain by expressing the partial derivatives according to the Maxwell relations
[12]:
dGg = −Sg dT + Vg dp + μg dng (7.62)

For the adsorbed phase, which is a two-dimensional phase, we have to take into
account the specific surface area A of the solid on which the molecules are adsorbed.
Thus, the Gibbs energy of the adsorbed phase is a function: Ga = f(T, p, A, na ). The
differential Gibbs energy is given by:
       
∂Ga ∂Ga ∂Ga ∂Ga
dG =a
dT + dp + dA + dna
∂T p,A,na ∂p T,A,na ∂A T,p,na ∂na T,p,A
(7.63)
and the Maxwell relations give:

dGa = −Sa dT + Va dp − π dA + μa dna (7.64)

We retrieve here the spreading pressure p, which corresponds to the change of Gibbs
energy due to a change of the interface area between solid and adsorbate. At constant
temperature and pressure Eq. 7.64 is reduced to:

dGa = − π dA + μa dna (7.65)

The application of the Euler theorem gives:

Ga = − π A+na μa (7.66)

By derivation of this equation we have:

dGa = − π dA-Ad π +na dμa + μa dna (7.67)

298 J.-P. Bellat

The subtraction of Eqs. 7.65 and 7.67 gives:

Ad π = na dμa (7.68)

The equilibrium condition between the adsorbed phase and the gas phase is:

μg (T, p) = μa (T, p, na ) (7.69)

dμg = dμa (7.70)

with the chemical potential of the gas assumed as ideal:

◦ p
μg (T, p) = μg (T) + RTLn (7.71)
p◦
p
dμg = RTdLn (7.72)
p◦

Thus we obtain with Eqs. 7.68, 7.70 and 7.72:

na p
dπ = RTLn ◦ (7.73)
A p

This is the Gibbs adsorption isotherm. The spreading pressure cannot be directly
measured but can be calculated owing to this equation. The integration of Eq. 7.73
from 0 to p gives:
p
RT p
π= na dLn ◦ (7.74)
A p
0

Thus the spreading pressure can

 be calculated
 from the integration of the adsorp-
tion isotherm in the form na = f Ln pp◦ (Fig. 7.16).
T
It may be noted that a problem of integration appears because Ln pp◦ tends to
infinity when p tends to zero. To overcome this problem we can use the Henry law,
which states that, in the low pressure region, the adsorbed amount is proportional to
the pressure:
na = KH p (7.75)

KH , the Henry constant is given by the slope of the adsorption isotherm in the very
low pressure region (Fig. 7.16).
7 Study of Selective Adsorption of Gases by Calorimetry 299

Adsorption isotherm
p
⎛ p⎞
∫ N dLn⎜⎜⎝ p° ⎟⎟⎠
RT a
A
Adsorbed amount N a

Adsorbed amount N a
pH

RT
K p
A H H
Henry’s law region

pH Ln(pH /p°) Ln(p/p°)

Pressure p

Fig. 7.16 Determination of the spreading pressure by integration of the adsorption isotherm

Thus the spreading pressure is calculated by mean of the relation:

⎛ ⎞
p
RT ⎜ p⎟
π= ⎝KH pH + na dLn ⎠ (7.76)
A p◦
pH

where pH is the pressure up to which Henry’s law is valid.

In the case of multi-components adsorption the total spreading pressure is:
⎛ p ⎞

RT ⎝ y p
π= πi = nai dLn i◦ ⎠ (7.77)
A p
i i 0

It can be calculated from partial adsorption isotherms in the same way as for a single
component.
In the case of microporous solids like zeolites for example, where the curvature of
the surface can be very important, the notion of specific surface area is questionable.
It is the reason why some authors recommend replacing the term πA by  that they
call the surface potential.
b. Differential adsorption Gibbs energy
Let us consider the coadsorption of two components A and B in a closed system at
constant temperature. As for the adsorption of a single component, we can define
an equivalent thermodynamic system composed of a closed volume containing the
gas mixture in contact with the sample (Fig. 7.17). The pressure of gas mixture is
po = pAo + pBo in the initial condition and p= pA + pB at the equilibrium. The molar
Gibbs energy of each component is Gmi .
300 J.-P. Bellat

Before adsorption At equilibrium

gas gas
g g g g
n Ao + n Bo nA + nB
closed system
pAo − pBo pA − p B
at constant V and T
T, Vg
T, V°
adsorbate
solid
solid

Fig. 7.17 Equivalent thermodynamic system for measuring the adsorbed amount by manometry
in the case of the adsorption of a binary mixture A + B

The coadsorption equilibrium corresponding to the adsorption of a dose of gas

mixture A+B at a given composition is:

A(g) + B(g) ↔ A(ads) + B(ads)

g g g g
before adsorption nAo , GmAo , pAo nBo , GmBo , pBo 0 0
g g g g
after adsorption nA , GmA , pA nB , GmB , pB naA , GamA naB , GamB

The total Gibbs energy of the system in the initial conditions is:
g g g g g g
Go = nio Gmio = nAo GmAo + nBo GmBo (7.78)
i

For each component i in the gas phase we can write:

g g g◦ pio
Gmio = μio (T,p) = μi (T) + RTLn (7.79)
p◦

Thus we have before adsorption:

g g◦ pio
Go = nio (μi (T) + RTLn ) (7.80)
p◦
i
   
g g◦ p g g◦ pBo
Go = nAo μA (T) + RTLn Ao + n μ (T) + RTLn (7.81)
p◦ Bo B
p◦

The total Gibbs energy of the system after adsorption is:

g g
G= nai Gami + ni Gmi (7.82)
i i
7 Study of Selective Adsorption of Gases by Calorimetry 301

with:
g g
ni = nio − nai (7.83)

and pi yp
g g g◦ g◦
Gmi = μi (T,p) = μi (T) + RTLn ◦
= μi (T) + RTLn i◦ (7.84)
p p

For each component i adsorbed, we can write from Eq. 7.66:

− πai A
Gami = + μai (T, π) (7.85)
nai

with μai (T, π) the chemical potential of the adsorbed phase defined by:
∗
μai (T, π) = μai (T, π) + RTLnγi xi (7.86)
∗
where μai (T, π) is the chemical potential of the single component i adsorbed at the
same spreading pressure as the mixture.
Thus the expression of the total Gibbs energy after adsorption becomes:
g  g◦ pi 
G=− πi A + nai μai (T, π) + (nio − nai ) μi (T ) + RT Ln ◦ (7.87)
p
i i i

As at equilibrium we have equality of the chemical potentials of each component

present in the gas phase and in the adsorbed phase:
g
μai (T, π) = μi (T,p) (7.88)

the total Gibbs energy after adsorption is:

g◦ g◦ pi
G=− πi A + ni (μi (T) + RTLn ) (7.89)
p◦
i i

The change of coadsorption Gibbs energy between the final and initial states is then
obtained by subtracting Eq. 7.89 with Eq. 7.80:
pi
G = G − Go = − πi A + nio RTLn (7.90)
pio
i i

For the adsorption of the mixture A+B we obtain:

pA p
G = − πA A − πB A + nAo RT Ln + nBo RTLn B (7.91)
pAo pBo
302 J.-P. Bellat

G is an integral value.
 This is the change of Gibbs energy of the system for the
g g
coadsorption of the i (nio − ni )moles of gas mixture in a given proportion. G
can be easily calculated by measuring the pressure of gas mixture in the system
before and after adsorption and by determining the partial spreading pressures from
integration of the partial adsorption isotherm of each component i:

pi
pi
πA = πi A = RT RTdLn (7.92)
p◦
i i 0

The differential molar coadsorption Gibbs energy can be calculated at given adsorbed
amount by derivation of the integral coadsorption Gibbs energy:
 
∂G
r Gam,A+B = (7.93)
∂na (T, A)

Thus we can obtain the dependence of the adsorbed amount on the differential coad-
sorption Gibbs energy in a similar way as for the differential coadsorption enthalpy
measured by calorimetry. Practically, r Gam,A+B can be calculated by numerical
derivation of the plot G = f(na ). In the case of the adsorption of a single compo-
nent, the equations for the calculation of the Gibbs adsorption energy are exactly the
same, we have just to take i equal to one.

7.5.5 Differential Adsorption Entropy and Molar

Entropy of Adsorbate

The differential molar coadsorption entropy can be easily calculated knowing the
differential molar coadsorption enthalpy measured by calorimetry and the differential
molar coadsorption Gibbs energy calculated as explained above by mean of the
relation:
r Ham,A+B − r Gam,A+B
r Sam,A+B = (7.94)
T

The molar entropy of the adsorbate at the total adsorbed amount naA+B and for a
given composition of the gas mixture is given by:

g◦
Sam,A+B = r Sam,A+B + Sm,A+B (7.95)

g◦
where Sm,A+B is the molar entropy of the gas mixture. Its value can be calculated
assuming the mixture as ideal with the relation:

g◦ g◦ g◦
Sm,A+B = yAo Sm,A + yBo Sm,B (7.96)
7 Study of Selective Adsorption of Gases by Calorimetry 303

g◦
yio being the initial mole fraction of component i in the gas phase and Sm,i the molar
entropy of gaseous single component i.

7.5.6 Partial Adsorption Enthalpy and Entropy

In the case of multi-components adsorption, the partial molar differential adsorption

enthalpies and entropies of each component i present in the gas mixture cannot be
directly measured by experiments. However, it is possible to estimate them by mean
of the “tangent method” based on the well-known Gibbs-Duhem relation.
Let us consider the adsorption of a binary gas mixture A+B and Z a function
of state, which denotes either the enthalpy or the entropy. The differential molar
coadsorption value r Zam,A+B is equal to:

r Zam,A+B = xA r Zam,A + xB r Zam,B (7.97)

where r Zam,i is the partial molar differential adsorption value of component i and
xi is the mole fraction of component i in the adsorbed phase for a given adsorbed
amount of mixture. As xA + xB =1, we obtain:

r Zam,A+B = r Zam,B + (r Zam,A − r Zam,B )xA (7.98)


The application of the Gibbs-Duhem relation xi dZam,i = 0 leads to:
i

xA dr Zam,A + xB dr Zam,B = 0 (7.99)

The derivation of Eq. 7.96 with respect to xA gives:

∂r Zam,A+B
= r Zam,A + xA dr Zam,A − r Zam,B + (1 − xA )dr Zm,B (7.100)
∂xA

By combining with Eq. 7.99 and by reporting in Eq. 7.98 we have:

∂r Zam,A+B
r Zm , Ba = r Zam,A+B − xA (7.101)
∂xA

In the same way, by derivation of Eq. 7.97 with respect to x B we obtain:

∂r Zam,A+B
r Zam,A = r Zam,A+B − xB (7.102)
∂xB

As xA + xB = 1 and dx A = −dx B , this equation can be written as:

304 J.-P. Bellat

Δr Zam, A + B

xA xB

0 1 xA

Δr Zam*, A
Δr Zam*, B
Δr Zam, B
Δr Zam, A + B
Δ r Zam , A

B A

Fig. 7.18 Graphical method for the determination of the partial molar differential adsorption
enthalpies or entropies in the case of a binary mixture A + B

∂r Zam,A+B
r Zam,A = r Zam,A+B + (1 − xA ) (7.103)
∂xA

Thus, the partial molar differential values r Zam,A and r Zam,B can be graphically
determined from the curve r Zam,A+B = f(xa ) plotted at a constant total adsorbed
amount. As illustrated on Fig. 7.18, they are given by the intercepts of the straight-
line tangent to the curve at xA = 1 and xA = 0. We just need for that to have a lot
of accurate measurements of r Zam,A+B in a large range of composition and total
adsorbed amount of mixtures.
It may be noticed that in the case of an ideal adsorbed solution the partial molar
differential values r Zam,i is constant and equal to the partial molar value of single
∗
component r Zam,i . If Z represents the enthalpy, we retrieve from the Eq. 7.97 the
Eq. 7.60, which allows the prediction of the coadsorption enthalpy from the adsorp-
tion enthalpies of single components.
7 Study of Selective Adsorption of Gases by Calorimetry 305

O 2-

Si 4+

Na + or Ba 2+ II
II’

I’
I

p-xylene
III m-xylene

α-cage
β -cage
Øα = 0.74 nm Øβ = 0.22 nm

Fig. 7.19 Structure of the Y type zeolite. For clarity reason the aluminium, silicon and oxygen
atoms are not represented. Only the compensation cations (in red) located on sites II inside the
supercages accessible to xylenes isomers are shown. Black points represent the cationic sites

7.6 Case Studies

7.6.1 Separation of Xylenes Isomers by Selective

Adsorption on FAU Type Zeolite

This first example of application concerns the separation of xylenes isomers (o-, m-
and p-dimethylbenzene) by selective adsorption on FAU type zeolites exchanged with
various compensation cations. As these isomers have very close boiling points, they
cannot be separated by distillation. So, they are separated by competitive adsorption
using the liquid chromatography technique and the simulated moving bed technol-
ogy [13, 14]. The aim of this work was to perform a complete thermodynamic study
of the adsorption of xylenes in the gas phase in order to obtain a detailed understand-
ing of the selective adsorption process. We focused here only on the separation of
p-xylene (pX) and m-xylene (mX), which is particularly difficult. This basic study
was essential for the development of a more efficient separation process [15] and, in
particular, for the optimization of adsorbent formulation and operating conditions of
the process.
The FAU type zeolite is a microporous aluminosilicate [16]. Its chemical for-
mula can be written as: Mn+ x/n (AlO2 )x (SiO2 )y , mH2 O where M is a compensation
cation (alkaline or alkaline-earth metals) at the oxidation state +n. The simplified
structure of FAU type zeolite is shown on Fig. 7.19. It is composed of SiO4 and
AlO4 tetrahedrons arranged in such a way that they form sodalite cages (so-called β-
cages) and supercages (so-called α-cages) having a free aperture of 0.22 and 0.74 nm
306 J.-P. Bellat

NaY BaY
4 4
mX pX
N / molec/cage

N / molec/cage
3 3 mX
pX
2 2

1 1
a

a
0 0
0 2 4 6 8 0 2 4 6 8
p / hPa p / hPa

Fig. 7.20 Adsorption isotherms of single p-xylene and m-xylene on NaY and BaY zeolites
at 423 K

respectively. The internal diameter of the supercage is around 1.3 nm. As the kinetic
diameters of p-xylene and m-xylene are about 0.67 and 0.71 nm respectively, these
isomers can only enter the supercages. FAU zeolite is usually called X zeolite if the
Silicon/Aluminium ratio is lower than 1.5 and Y zeolite if this ratio lies between
1.5 and 3. According to the Si/Al ratio, the compensation cation can be located on
different sites inside supercages (sites II and III), sodalite cages (sites I’ and II’) or
hexagonal prisms linking two sodalite cages (sites I) [17, 18]. In this study, the Si/Al
ratio was equal to 2.4 and the compensation cations were sodium or barium. They
were essentially located at the center of the hexagonal window of each sodalite cage
but some of them are also present inside sodalite cages and hexagonal prims.
First, the adsorption of single xylenes has been studied on NaY and BaY zeolites
at 423 K by calorimetry coupled with manometry [8, 19]. The adsorption isotherms
are given in Fig. 7.20. No significant difference is observed between p-xylene and m-
xylene. Adsorption isotherms are perfectly superimposed on BaY while the amount
of m-xylene is slightly higher for m-xylene on NaY. This indicates that whatever the
compensation cation, the adsorption affinities for p-xylene and m-xylene are very
close and the supercage can accommodate around 3 aromatic molecules [20]. The
adsorption enthalpies as a function of supercage filling are shown on Fig. 7.21. The
values measured by calorimetry are higher than those estimated by using the van’t
Hoff method as shown for the NaY zeolite in Table 7.2 [21]. Relevant information
can be extracted from these calorimetric curves. Firstly, on both zeolites, the adsorp-
tion enthalpy at zero filling is higher with m-xylene than with p-xylene. However,
the difference is rather low (5 kJ.mol−1 ). This means that the aromatic molecule-
zeolite interactions are almost the same for the two isomers and consequently this
suggests that the adsorption sites for p-xylene and m-xylene are probably very sim-
ilar. This result has been confirmed by a detailed structural study on the location of
xylenes adsorbed in the supercages by neutron diffraction [22, 23] and molecular
simulation [24, 25]. Figure 7.22 shows the location of xylene molecules inside the
supercage. Both isomers have a very similar position. Their aromatic rings are cen-
tred on compensation cation in site II and their methyl groups are in interaction with
7 Study of Selective Adsorption of Gases by Calorimetry 307

NaY BaY
150 150
-1

-1
mX
/ kJ.mol

pX

/ kJ.mol
100 100
mX
pX
m

m
50 50
a

a
-Δ H

-Δ H
ΔH ΔH
r

r
liquefaction liquefaction
0 0
0 1 2 3 4 0 1 2 3 4
a a
N / molec/cage N / molec/cage

Fig. 7.21 Differential molar adsorption enthalpies of single p-xylene and m-xylene as a function
of filling in NaY and BaY zeolites at 423 K

Table 7.2 Comparison of adsorption enthalpies of p-xylene and m-xylene on NaY measured by
calorimetry and estimated from adsorption isotherms using the Van’t Hoff method
r Ha /kJ.mol−1 p-xylene m-xylene
Experimental Calorimetry 75 80
Henry 50 58
Van’t Hoff Langmuir 64 67

p-xylene m-xylene

Fig. 7.22 Schematic representation of the location of p-xylene and m-xylene molecules adsorbed
in the supercage of FAU zeolite [blue:oxygen atoms; yellow: compensation cation; red: carbon
atoms; white: hydrogen atoms]

oxygen atoms of the double 6-ring unities joining two sodalite cages. The fact that
the adsorption enthalpies are higher with m-xylene could be explained by a better
interaction of methyl groups with the oxygen atoms of hexagonal prisms, which are
also in the meta position to the hexagonal window.
Secondly, the adsorption enthalpy at zero filling is largely higher on BaY than on
NaY (around 75 against 120 kJ.mol−1 ). This indicates that the adsorbate–adsorbent
interactions are stronger with barium than with sodium. This important differ-
ence in adsorption enthalpies can be explained from structural considerations. The
308 J.-P. Bellat

xylene

Sodalite cage Ba 2+

A
Oxygen Na+

Fig. 7.23 Schematic representation of the location of sodium and barium cation in site II in the
supercage and in interaction with a xylene molecule

adsorption enthalpy is a function of the local electric field in the supercages and the
adsorbate-adsorbent interactions (mainly aromatic ring—compensation cation and
methyl group—framework oxygen interactions). These interactions depend obvi-
ously of the electric charge of the compensation cation (barium is divalent while
sodium is monovalent) but are also strongly related to the size and the location of the
compensation cation in the supercage. As schematically shown in Fig. 7.23, sodium
cation is located at site II in the plane of the hexagonal window of the sodalite cage.
The xylene molecule is adsorbed on this site, involving aromatic-Na+ and methyl–
oxygen interactions. As for the barium cation, it lies out of this plane, inside the
supercage, because of its bigger size (rNa+ = 0.095nm and rBa2+ = 0.143 nm). It is
more accessible to xylene molecule than the sodium cation, which is rather hidden
by the oxygen atoms. With barium, the aromatic—ring cation distance is shorter than
with sodium, leading thus to stronger aromatic-ring cation interaction and weaker
methyl-oxygen interaction because the methyl groups, in this case, are farther from
the framework oxygen atoms. But this weakening is however in a less extent and it
results that the adsorption enthalpy at zero filling is higher with barium than with
sodium. Thirdly, as the filling increases, above 2 molec/cage, the adsorption enthalpy
increases slightly for NaY while it decreases for BaY. This means that the xylene
molecules do not interact in the same way with both zeolites. For NaY, the increase
of the adsorption enthalpy is due to the increase of adsorbate-adsorbate interactions,
which is not compensated by the decrease of the adsorbate-adsorbent interactions. As
7 Study of Selective Adsorption of Gases by Calorimetry 309

in a supercage, each xylene molecules interact with one sodium cation, the adsorbate-
adsorbent interactions are almost constant during the adsorption. But when the last
molecule is adsorbed, because of steric hindrance, the molecular interactions become
important. For BaY, if the adsorbate-adsorbate interactions increase also with the
filling, the adsorbate-adsorbent interactions are not constant. As there are only two
cations in the supercage, contrary to the two first molecules, the last xylene molecule
cannot interact with one barium cation and is adsorbed on a non cationic site with
a lower energy. Finally, once the supercages are completely filled, the adsorption
enthalpy decreases sharply down to the liquefaction enthalpy of xylenes. This last
part of the enthalpy curve corresponds to the adsorption of some xylene molecules
on the external surface of zeolite particles.
Then, the coadsorption of p-xylene and m-xylene has been studied on NaY and
BaY zeolites. From the study on adsorption of single components indicating no
important difference in adsorption capacities and enthalpies between p-xylene and m-
xylene, we could expect that the coadsorption of these two isomers on these materials
will not be selective. Nevertheless, results show that the FAU zeolites can selectively
adsorb the xylene isomers. However these systems are very particular because the
selectivity appears only at high filling of supercages [26, 27]. Adsorption isotherms
of equimolar mixture of p-xylene and m-xylene on NaY and BaY zeolites at 423 K
are displayed on Fig. 7.24. For both zeolites, in the low pressure range the amounts of
p-xylene and m-xylene adsorbed are quite the same, but at the plateau of the partial
isotherms, a small difference in adsorption capacities appears. NaY adsorbs more
m-xylene while BaY adsorbs more p-xylene. The adsorption selectivities of p-xylene
with respect to m-xylene versus filling are given in Fig. 7.25. Below 2 molec/cage
and whatever the zeolite, the selectivity is equal to one. Both isomers are adsorbed
in the same proportions. The adsorption process is not selective. Above this filling,
during the adsorption of the third and last molecule in the supercages, the selectivity
decreases slightly for NaY (m-xylene is preferentially adsorbed) while it increases
sharply for BaY (p-xylene is preferentially adsorbed).
Figure 7.26 shows the coadsorption enthalpies as a function of the filling. For
NaY, the calorimetric curve measured with the equimolar mixture is very similar to

49% pX + 51% mX / NaY 49% pX + 51% mX / BaY

4 4
pX + mX
N / molec./cage

N / molec/cage

3 pX + mX
mX 3

mX pX
2 2

pX
a
a

1 1 mX

0 0
0 1 2 3 0 1 2 3
p / hPa p / hPa

Fig. 7.24 Coadsorption isotherms of p-xylene and m-xylene on NaY and BaY zeolites at 423 K
310 J.-P. Bellat

49% pX + 51 % mX / NaY 49% pX + 51% mX / BaY

3 3
pX/mX

pX/mX
Selectivity α

Selectivity α
2 2

1 1

0 0
0 1 2 3 4 0 1 2 3 4
a a
N / molec/cage N / molec/cage

Fig. 7.25 Dependence of the total filling on the adsorption selectivity of p-xylene with respect to
m-xylene on NaY and BaY zeolites at 423 K

49% pX + 51% mX / NaY 49% pX + 51% mX / BaY

150 150
pX+mX
-1

-1
kJ.mol
kJ.mol

mX
pX+mX
100 mX 100 pX
/
/

pX
m

m
a
a

-Δ H
-Δ H

50 liquefaction 50 liquefaction
r

0 0
0 1 2 3 4 0 1 2 3 4
a a
N / molec/cage N / molec/cage

Fig. 7.26 Coadsorption enthalpies of p-xylene and m-xylene on NaY and BaY zeolites at 423 K
as a function of the total filling [Doted lines adsorption enthalpies of single components]

those obtained with single xylenes. The coadsorption enthalpy is even very close
to the adsorption enthalpy of single m-xylene, the isomer, which is adsorbed in
favour at high filling. No particular change in the adsorption enthalpy during the
filling is observed. The fact that the NaY adsorbs preferentially the m-xylene is
certainly due to the adsorbate-adsorbent interactions which are stronger with m-
xylene owing to a better orientation of methyl groups with respect to oxygen atoms
of hexagonal prisms as explained above. On the contrary, for BaY the coadsorption
enthalpy is very different than the adsorption enthalpy of single xylenes. If up to
1 molec/cage the coadsorption enthalpy is equivalent to the adsorption enthalpy of
single m-xylene, it decreases during the adsorption of the second molecule. Then it
increases slightly above 2 molec/cage, as the adsorption process becomes selective
for p-xylene, before to go down to the liquefaction enthalpy when the saturation of
supercages is approaching. This result shows that during the coadsorption process,
the xylene-zeolite interactions as well as the xylene-xylene interactions are changed
compared to the adsorption of single components. The fact that the BaY adsorbs
selectively the p-xylene at high filling, while not the NaY, is attributed to steric
hindrances caused by the barium cation. Although they are two times less numerous
7 Study of Selective Adsorption of Gases by Calorimetry 311

than sodium, the free space available to xylene isomers in the supercages is reduced
because of their bigger size and their location. Thus, at high filling of supercages, the
last xylene molecule, which is adsorbed is extremely confined and cannot directly
interact with a barium cation. Moreover, this last molecule cannot be accommodated
in the supercage without molecular rearrangement of the others. These phenomena
are accompanied by exo- or endothermic effects, which are at the origin of the changes
observed in coadsorption calorimetric curve. In these conditions, this is the isomer
having the smallest size and the most symmetric geometry, i.e. the p-xylene, which
is preferentially adsorbed. The compensation cation appears as the key parameter.
The selectivity depends strongly of its size, number, electric charge and location.
In such systems, the selective adsorption process does not result of differences in
adsorbate-adsorbent interactions between the isomers. These ones are too much low
to induce a competitive adsorption. The selectivity is essentially governed by the
entropic effects, consecutive to the steric hindrance caused by the compensation
cations and leading to important molecular rearrangement of xylenes confined in the
zeolitic cavities.

7.6.2 Desulphurization of Natural Gas by Selective

Adsorption on FAU Type Zeolite

This second example of application is taken from a complete thermodynamic study

devoted to the adsorption of light mercaptans, paraffins and aromatics on FAU type
zeolite [28–30]. The aim of this work was to find a good adsorbent able to perform
a deep desulphurization of natural gas by selective adsorption. Results presented
below concern the adsorption and coadsorption of ethylmercaptan (ESH), n-heptane
(HEP) and toluene (TOL) on the NaX zeolite a 298 K. NaX zeolite has exactly the
same structure as NaY excepted that it contains more compensation cations located
in sites II and also in sites III around the dodecagonal window (Fig. 7.27).
Adsorption isotherms of single components are shown in Fig. 7.28. In order to
compare the three adsorbates, they are represented by using the loading rate θ and
the relative pressure p/ps . In the low relative pressure range, isotherms show that the
adsorption affinity increases according to the sequence: ESH > TOL > HEP, which
is the same as for the dipolar moment (Table 7.3). Moreover, the adsorption affinity of
the NaX zeolite for the sulphur compound is strong enough to consider this adsorbent
as a good candidate for a deep desulphurisation of natural gas. Regarding the adsorp-
tion capacities, they are very similar. The volume of each molecule adsorbed is very
close to the crystallographic volume of supercages (Table 7.3). Figure. 7.29 shows
the adsorption enthalpy of ethylmercaptan measured by differential calorimetry and
calculated with the isosteric method from a set of isotherms measured at different tem-
peratures (Fig. 7.30). A very good agreement is obtained between the two methods.
The filling range investigated with the isosteric method is however limited to mod-
erate fillings, the determination of equilibrium pressure at fixed adsorbed amount on
312 J.-P. Bellat

O 2-
HEP
Al 3+ Si 4+

Na+

ESH

TOL

Fig. 7.27 Structure of the NaX zeolite. The compensation cations inside the supercage are located
on sites II and III

1
TOL
0.9 ESH
loading rate θ

0.8

0.7 HEP

0.6

0.5

0.4
-6 -5
10 10 0.0001 0.001 0.01 0.1
relative pressure p/p
s

Fig. 7.28 Adsorption isotherms of ethylmercaptan, n-heptane and toluene on NaX zeolite at 298 K

Table 7.3 Dipolar moment and volume adsorbed in NaX at 298 K of ethylmercaptan, n-heptane
and toluene
ESH TOL HEP
μ/D 1.58 0.36 0
Va /cm 3 .g−1 0.294 0.265 0.270
Crystallographic volume of supercages: Vα = 0.294 cm3 .g−1
7 Study of Selective Adsorption of Gases by Calorimetry 313

ESH / NaX
100

-1
80 calorimetric

- Δ r H m / kJ.mol
60
isosteric
a

40

20 liquefaction

0
0 5 10 15 20 25 30
a
m / wt%

Fig. 7.29 Comparison of adsorption enthalpies of ethylmercaptan on NaX zeolite at 298 K mea-
sured by calorimetry (red points) and calculated with the isosteric method (green points)

ESH / NaX
25

20
m a/ wt%

15
298 K
10 323 K
348 K
5
373 K

0
0 0.5 1 1.5 2
p / hPa

Fig. 7.30 Adsorption isotherms of ethylmercaptan on NaX zeolite measured at different tempera-
tures

the adsorption isotherms being difficult in low and high pressure ranges. Adsorption
enthalpies of three single adsorptives are compared in Fig. 7.31. At zero filling the
adsorption enthalpies of ESH and TOL are equivalent and slightly higher than with
HEP. This is attributed to the fact that ESH and TOL have a dipolar moment contrary
to ESH. Their adsorption involves specific interactions. As the filling increases, the
adsorption enthalpies of toluene and ethylmercaptan are rather constant. The adsorp-
tion process of these two molecules is relatively homogeneous from the energetic
point of view. The molecules are in interaction with the compensation cation in the
supercages without important interactions between them. The adsorption enthalpy is
higher with toluene. This could be explained by a stronger interaction with the com-
pensation cation owing to the aromatic ring. It may be pointed out with this molecule a
314 J.-P. Bellat

120

100 TOL

-1
Δ H m / kJ.mol
80 HEP

60
ESH
a

40
r

20

0
0 0.2 0.4 0.6 0.8 1
Loading rate θ

Fig. 7.31 Adsorption enthalpies of ESH, HEP and TOL on NaX zeolite determined by differential
calorimetry at 298 K as a function of the loading rate

sharp increase of the adsorption enthalpy at low filling. This effect is still unexplained.
It could be due to the presence of compensation cation in sites III at the entrance of
the supercage. For n-heptane, which is a non-polar molecule, the adsorption involves
only non-specific interactions. The adsorption enthalpy increases continuously with
the filling because of the adsorbate-adsorbate interactions, which appear stronger
than with ethylmercaptan and toluene. This is probably due to the length of this lin-
ear paraffin (1.16 nm), which is close to the supercage diameter (1.3 nm). The molar
entropies of ESH, HEP and TOL adsorbed in NaX, calculated by combining isosteric
method and calorimetric measurements, are given as a function of the filling coeffi-
cient in Fig. 7.32. Their values obtained at a loading rate equal to 0.5 are compared
with those of standard molar entropies of solid, liquid and gas phases at 298 K in
Table 7.4. The molar entropies of adsorbates vary following the sequence: TOL <
ESH < HEP. The molar entropy of toluene adsorbed in NaX is very low. Its value
is lower than that one of the solid state. This means that the toluene is extremely
confined in the supercage. The molecule is “frozen” on its adsorption site. The molar
entropies of ethylmercaptan and n-heptane are higher and lie between the standard
molar entropies of the liquid and solid states. These molecules have lost less degrees
of freedom and are more flexible and mobile on their adsorption sites than toluene.
The coadsorption of ESH with HEP or TOL has then been studied in order to
check if the presence of long paraffins or aromatics in the natural gas can perturb
the capture of sulphur compound. Although the study on the adsorption of single
components shows that the strongest adsorption affinity of NaX is for ESH, in the case
of coadsorption, the behaviour of NaX can be different. The adsorption selectivity
of ESH with respect to HEP or TOL as a function of the total filling, for an initial
gas mixture containing 25 % of sulphur compound, is given in Fig. 7.33. Similar
selectivity curves are obtained for other compositions of gas mixtures. When ESH is
in the presence of HEP in the gas phase, the adsorption selectivity is largely higher
than one. Whatever the filling, NaX adsorbs preferentially the mercaptan.
7 Study of Selective Adsorption of Gases by Calorimetry 315

300

-1
HEP

Sm / J.K .mol
200
-1
ESH
a

100

TOL
0
0 0.2 0.4 0.6 0.8 1
Loading rate θ

Fig. 7.32 Molar entropies of ESH, HEP and TOL adsorbed in NaX zeolite at 298 K as a function
of the loading rate

Table 7.4 Comparison of molar entropies of ESH, HEP and TOL adsorbed in NaX at 298 K at a
loading rate equal to 0.5 with the standard molar entropies of solid, liquid and gas states at 298 K
S/J.K−1 .mol−1 ESH HEP TOL
Sam at θ = 0.5 128 182 71
S◦solid 70 127 81
S◦liquid 207 305 192
S◦gas 296 428 320

25% ESH + 75% HEP 25% ESH + 75% TOL

60 2
ESH/TOL
ESH/HEP

50
Selectivity α

40
Selectivity α

30 1
20
10
0 0
0 1 2 3 4 5 0 1 2 3 4 5
a a
N / molec/cage N / molec/cage

Fig. 7.33 Adsorption selectivity of ESH with respect to HEP or TOL as a function of the total
filling of NaX zeolite at 298 K and for a sulphur content in the initial gas mixture of 25 %

However, the selectivity curve exhibits some peaks of selectivity at certain fill-
ings which are less or more intense according to the initial gas composition. When
ESH is in the presence of TOL, the selectivity is in favour of the mercaptan in a
large range of filling but decreases sharply as the complete filling of supercages is
approaching. Then, NaX adsorbs preferentially the toluene and we observe, in the
partial adsorbed amounts, a displacement of ESH by TOL. It may be noticed that
316 J.-P. Bellat

NaX is much less selective for ethylmercaptan in the presence of toluene than in the
presence of n-heptane. In order to better understand these unusual changes in the
adsorption selectivity, the coadsorption heats have been determined by using the dif-
ferential calorimetry technique coupled with manometry, chromatography and mass
spectrometry. By this way, we expected to observe on the calorimetric curve particu-
lar thermal effects corresponding to these modifications of selectivity. Coadsorption
enthalpies of ethylmercaptan with n-heptane or toluene are given in Fig. 7.34. Sur-
prisingly, no special thermal effect appears on the calorimetric curves. They are
rather very similar to those obtained with single n-heptane and toluene which are
the major components in the gas phase. The partial adsorption enthalpies of each
component in the mixture have been determined by using the “tangent method” with
Eqs. 7.101 and 7.102. Results are given in Fig. 7.35. They show that whatever the
composition, the partial adsorption enthalpies are very close to those of single compo-
nent excepted for the toluene for which, the partial molar enthalpy increases slightly
when the mixture becomes very reach in mercaptan. As the partial enthalpies are
constant and equal to the adsorption enthalpies of single components in a large range

25%ESH + 75% HEP 25% ESH + 75% TOL

100 100
-1

-1

80 80
-Δ H m / kJ.mol

-ΔrH m / kJ.mol

60 60

40 40
a

liquefaction
liquefaction
20 20

0 0
0 1 2 3 4 5 0 1 2 3 4 5
a a
N / molec/cagne N / molec/cage

Fig. 7.34 Coadsorption enthalpies of ESH + HEP and ESH + TOL as a function of the total filling
of NaX zeolite at 298 K and for a sulphur content in the initial gas mixture of 25 %. [Red points
experimental; green solid line calculated from adsorption enthalpies of single components with
Eq. 7.60]

ESH + HEP / NaX ESH+TOL / NaX

110 110
-Δ Hma / kJ.mol-1
-1

a N atotal = 1 molec/cage TOL

-Δ Hm / kJ.mol

N = 1molec/cage 100
100 total

90 90
HEP ESH+TOL
80 80
a

ESH+HEP
r
r

70 70
ESH
ESH
60 60
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x in adsorbate x in adsorbate
ESH ESH

Fig. 7.35 Partial adsorption enthalpies of ESH, HEP and TOL coadsorbed in NaX zeolite at 298
K and for a total amount of mixture adsorbed equal to 1 molec/cage
7 Study of Selective Adsorption of Gases by Calorimetry 317

of composition, the coadsorption enthalpies of ESH + HEP and ESH + TOL can be
calculated by the adsorption enthalpies of single components using the Eq. 7.60. As
shown on Fig. 7.34, a very good agreement is obtained between the calculation and
the calorimetric measurements. Therefore the variations of the adsorption selectivity
observed as filling increases are not related to particular thermal effects. At least
these effects are not important enough to be detected by our calorimetry technique.
These results suggest that the coadsorption affinity is more dependent on the entropic
effects than the enthalpic ones. As in the case of the adsorption of xylenes, in the
supercages of NaX, the molecules of ESH, HEP and TOL are probably confronted to
steric hindrance leading to molecular rearrangements. In these systems the selectivity
is also essentially governed by entropic effects.

7.7 Conclusions

The development of new gas separation or purification processes by selective adsorp-

tion requires the elaboration of more and more efficient adsorbents. New solids having
large adsorption capacity, strong adsorption affinity, good hydrothermal and mechan-
ical stability and high selectivity must be discovered. This imposes to have a detailed
knowledge of the adsorption process at macroscopic and molecular levels, especially
when the adsorption involves several components. Since the last two decades, the
numerical molecular simulation methods (molecular dynamics and Monte Carlo sim-
ulations) have shown fantastic advances in the field of gas adsorption in nanoporous
solids. It is possible now to predict with acceptable precision the adsorption capaci-
ties, energies and selectivities in shorter time than by experiments. Moreover these
methods give interesting information at the microscopic level as the location of the
adsorbed species, the energy of molecular interactions, the diffusion coefficients that
are very difficult or impossible to measure by experiments. Simulations are becoming
more numerous than experiments. Nevertheless, adsorption experiments are essential
for checking the validity of numerical simulations codes and to find the real physico-
chemical phenomena involved in the adsorption process. It is the reason why the
experimental approach is essential in any thermodynamic study of gas adsorption. In
this field, the differential calorimetry coupled with manometry and gas chromatog-
raphy or mass spectrometry is a powerful technique. It allows performing a complete
and detailed thermodynamic study of gas adsorption on solids. The knowledge of
the adsorption enthalpies and entropies are now very useful to design the porous
structure and the surface chemical properties of efficient selective adsorbents. This
is a complementary technique, which can give rich information even at the molecular
level. The use of this technique remains however delicate and is time consuming.
It needs a lot of calibrations and adsorption experiments must be repeated several
times. The calorimetry requires the greatest care and patience but this technique is
really indispensable to have a detailed knowledge of adsorption mechanisms that
take place at the gas-solid interface.
318 J.-P. Bellat

References

1. S.J. Gregg, K.S.W. Sing, Adsorption, Surface Area and Porosity (Academic Press, London,
1982)
2. K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T.
Siemienewska, Pure Appl. Chem. 57(4), 603 (1985)
3. F. Rouquerol, J. Rouquerol, K. Sing, Adsorption by Powders and Porous Solids (Academic
Press, London, 1999)
4. R.M. Barrer, Zeolites and clay minerals as sorbents and molecular sieves (Academic Press,
New York, 1978)
5. F. Thibault-Starzyk, M.H. Simonot-Grange, Les matériaux micro et mésoporeux—
Caractérisation (EDP Sciences, Les Ulis, 2004)
6. P.W. Atkins, Physical Chemistry (Oxford University Press, Oxford, 1990)
7. D.D. Do, Adsorption Analysis: Equilibria and Kinetics (Imperial College Press, London, 1998)
8. M.H. Simonot-Grange, O. Bertrand, E. Pilverdier, J.P. Bellat, C. Paulin, J. Thermal Anal. 48,
741 (1997)
9. V. Cottier, J.P. Bellat, M.H. Simonot-Grange, A. Méthivier, J. Phys. Chem. B 101(24), 4798
(1997)
10. S. Sircar, J. Chem. Soc. Faraday Trans. 1 81, 1527 (1985)
11. S. Ross, J.P. Olivier, On Physical Adsorption (Interscience Publishers John Wiley & Sons, New
York USA, 1964)
12. A. Boutin, G. Dosseh, A. Fuchs, Eléments de Thermodynamique (Masson, Paris, 1997)
13. D.B. Broughton, R.W. Neuzil, J.M. Pharis, C.S. Breatly, Chem. Eng. Prog. 66(9), 70 (1970)
14. R.T. Yang, Gas separation by adsorption processes (Imperial College Press, London, 1987)
15. G. Hotier, B. Balannec, Revue de l’Institut Français du Pétrole 46, 803 (1991)
16. D.W. Breck, Zeolite Molecular Sieves (Wiley, New York, 1974)
17. C. Beauvais, A. Boutin, A.H. Fuchs, Chem. Phys. Chem. 5, 1791 (2004)
18. C. Beauvais, A. Boutin, A.H. Fuchs, C.R. Chimie 8, 485 (2005)
19. V. Cottier, E. Pilverdier, M.H. Simonot-Grange, J.P. Bellat, J. Thermal Anal. 58, 121 (1999)
20. J.P. Bellat, M.H. Simonot-Grange, S. Jullian, Zeolite 15, 124 (1995)
21. J.P. Bellat, M.H. Simonot-Grange, Zeolite 15, 219 (1995)
22. M. Czjzek, H. Fuess, T. Vogt, J. Phys. Chem. 95, 5255 (1991)
23. C. Mellot, M.H. Simonot-Grange, E. Pilverdier, J.P. Bellat, D. Espinat, Langmuir 11(5), 1727
(1995)
24. V. Lachet, A. Boutin, B. Tavitian, A. Fuchs, J. Phys. Chem. 102, 9224 (1998)
25. V. Lachet, S. Buttefey, A. Boutin, A. Fuchs, Phys. Chem. Chem. Phys. 3, 80 (2001)
26. V. Cottier, J.P. Bellat, M.H. Simonot-Grange, J. Phys. Chem. B 101(24), 4798 (1997)
27. J.C. Moïse, J.P. Bellat, J. Phys. Chem. B 109, 17239 (2005)
28. G. Weber, J.P. Bellat, F. Benoit, C. Paulin, S. Limborg-Noetinger, M. Thomas, Adsorption 11,
183 (2005)
29. G. Weber, F. Benoit, J.P. Bellat, C. Paulin, P. Mougin, M. Thomas, Microporous Mesoporous
Mater. 109, 184 (2008)
30. F. Benoit, G. Weber, J.P. Bellat, C. Paulin, P. Mougin, M. Thomas, Adsorption 14, 501 (2008)
Chapter 8
Characterization of Acid–Base Sites in Oxides

Antonella Gervasini

Abstract The acidity and basicity of the oxides and metal oxide surfaces is one
among the most peculiar properties of this group of inorganic products that have
been exploited in catalysis from decades. The different types of oxides (single oxides,
doped and modified oxides, supported oxides, mixed and complex oxides) are here
described with emphasis on their acid-base functionalities in relation with the cat-
alytic activity. The prediction of the acidity insurgence from the oxide composition is
discussed. The possibility to have a deep knowledge of the amount of acid sites and
of the acid strength and strength distribution at the oxide surface is here presented.
Adsorption by suitable probes with basic/acidic characteristics can be realized in a
microcalorimetry of adsorption giving information on the heterogeneity of the sur-
face. Alternatively, probe desorption can be realized by a temperature increasing
program to obtain the acid-site strength distribution. In any case, quantitative infor-
mation on the acid oxide surfaces can be obtained with the relevant thermodynamic
of kinetic parameters which determine the amount and strength of the acid sites.

8.1 Introduction

Oxides are relevant products of the inorganic chemistry that are extensively used in
many applied fields such as adsorption technology, pigment technology, ceramics,
and heterogeneous catalysis. In any applied field, the oxide functionality is governed
by surface interaction with the environment. In particular, the interactions of the
oxide surfaces with gases and liquids, active in the fields of adsorption and catalysis,
are mainly governed by the acid-base properties of the oxide. These properties are
tightly connected to the catalytic functionality of several divided, wide surface area
oxide surfaces. Acid oxide catalysts have great importance in petroleum chemistry

A. Gervasini (B)
Dipartimento di Chimica, Università degli Studi di Milano, Milan, Italy
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 319

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_8,
© Springer-Verlag Berlin Heidelberg 2013
320 A. Gervasini

(e.g., amorphous silica-alumina cracking catalyst) and in any reaction involving

hydrocarbons and they have been extensively investigated, while basic oxide catalysts
have been studied to a lesser extent. The nature of the acid and base sites on oxide
surfaces is an intriguing and moot subject. Surface exposed unsaturated cations and
hydroxyl groups give rise to Lewis and Brønsted acid sites, respectively, while basic
sites could be associated with surface oxygen atoms able to interact attractively with
a proton. Severe pretreatment conditions of the oxides are needed for generating
active basic sites by removal of carbon dioxide, water, and in some cases, oxygen
from the surface.
Among the various types of heterogeneous acid and basic catalysts that found
real application in various reactions and catalytic processes (Table 8.1, [1–4]), var-
ious types of oxide structures are found. The surface properties, nature and struc-
ture of acid and base sites of oxides have been investigated for many decades by
various approaches, comprising spectroscopic, thermal, and theoretical studies, and
newly developed measurement methods using modern techniques are nowadays used
for this purpose. Other studies are aimed at developing novel acid and base oxide
formulations with improved functionality. In the following, the acidity-basicity of
oxide materials of catalytic importance by adsorption calorimetry by conventional
and novel methodologies of study is presented. Beside the conventional gas-solid
adsorption calorimetry studies, liquid-solid adsorption calorimetry has been devel-
oped to disclose sound relations between acid-base properties and activity of catalysts
working in liquid-solid heterogeneous processes.

8.2 The Surface Acido–Basicity of Metal Oxides

To introduce the concept of acidity and basicity of surface sites of any solid, it is
necessary to start with the very general definition of acids and bases. Two different
definitions are used today; they are the most relevant and accepted ones among all
the others. Brønsted and Lowry in 1923 defined that an acid (HA) is any hydrogen-
containing species able to release a proton and a base (B) is any species capable of
combining with a proton. In this view, the acid-base interaction consists in an equi-
librium exchange of a proton and the final product of the exchange is the conjugated
base (A− ) and acid (BH+ ), respectively. Alternatively, Lewis proposed in the same
year (1923) a different view to define the acidity-basicity concept. An acid (A) is any
species possessing an incomplete electronic grouping that can accept an electron pair
−
to form a dative or coordination bond (Aδ ). A base (B) is any species possessing a
non-bonding electron pair that can be donated to form a dative or coordination bond
(Bδ+ ). This second definition is more general than the first one and can be applied
to any species, even if hydrogen is not concerned in the exchange. The Brønsted
and Lewis acid and basic strength can be quantitatively determined. According to
the Brønsted theory, pK a measures the position of the equilibrium in water solution
or in gas-phase so allowing the quantitative determination of the relative strength
8 Characterization of Acid–Base Sites in Oxides 321

Table 8.1 Types of heterogeneous acid and basic catalysts [1–4]

Acid catalyst Basic catalyst
(1) Alumina (1) Single component metal oxides
chloridrated alumina alkaline earth oxide
sulfated alumina alkali metal oxide
rare earth oxide
(2) Silica (2) Zeolites
silicalite-1 alkali ion-exchanged zeolite
solid phosphoric acid (kiselghur) alkali ion-added zeolite
(3) Silica-alumina (3) Supported alkali metal ions
…aluminated silica alkali metal ions on alumina
alkali metal ions on silica
alkali metal ions on alkaline earth
oxide
alkali metals and alkali metal
hydroxides on alumina
(4) H-zeolites (4) Clay minerals
Hydrotalcite
chrysotile
sepiolite
(5) SAPOs (silicoaluminophosphates) (5) Non-oxide
KF supported on alumina
Lanthanide imide and nitride
zeolites
(6) Pure/mixed/supported metal oxides
(titania, zirconia, tungsta,
molibdena, etc.)
(7) Zirconia
sulfated zirconia
tungstated zirconia
(8) Resins
Sulfonic acid resins
(9) Clays
acid-treated montmorillonite
(10) Niobic acid and niobium phosphate
(11) Heteropolyacids

of the acids and bases. The pK a of the various acids/bases is usually reported with
respect to water, while in the gas-phase, where solvation effects are absent, the acid-
ity strength scale of given acids can be greatly different. If HA bond is fully ionic,
the equilibrium of the acid-base couple is modified and now the Hammett acidity
function (Ho ) is introduced. Ho holds in concentrated solutions where the pH con-
cept is no longer applicable. The Hammett function is determined by the extent to
which an indicator (weak or very weak organic base) is protonated, by measuring
the [BH+ ]/[B] ratio by UV-vis spectroscopy. The acid strength concept of the Lewis
322 A. Gervasini

acids is more difficult to be quantified. A simple scale of acid strength for Lewis
acids is not feasible because in this case the acidity depends upon the nature of the
base involved in the acid-base equilibrium. Only few quantitative measurements of
acid strength are available compared to Brønsted acids.
Other definitions of the acidity concept have been proposed in the literature; the
concept of “Hard and Soft Acids and Bases (HSAB)” initially proposed by R. Pearson
in 1963 [5] and extended by G. Klopman merits to be mentioned. Pearson classified
Lewis acids and Lewis bases as hard, borderline, or soft and G. Klopman attempted
to quantify Pearson’s HSAB principle using frontier molecular orbital (FMO) theory
[6]. Hard acids are defined as small-sized, highly positively charged, and not easily
polarizable electron acceptors. Hard bases are species holding their electrons tightly
as a consequence of large electronegativities, low polarizabilities, and difficulty of
oxidation of their donor atoms. Hard acids prefer to associate with hard bases, while
soft acids prefer to associate with soft bases. The formed complexes are thermody-
namically more stable than mixed complexes and also form faster. According with
this theory, protons and hydroxides and oxide ions are all hard species, as are most
O- and N-terminated anions. C- and S- terminated anions, sulphides, phosphines,
and aromatic hydrocarbons are soft bases.
It may be taken into consideration that the proton affinity, PA, of a anion or of
a neutral atom or molecule is a measure of its gas-phase basicity. It is the energy
released in the following reactions: A− + H+ → HA or B + H+ → BH+ . For
any species, the higher the proton affinity, the stronger the base and the weaker
the conjugate acid in the gas phase. Recently, the relationships between the proton
affinity and atomic charge and with the transfer of charge have been reexamined [7].
When getting on to the acidity-basicity of solids, we face with a different problem.
The quantitative concepts that hold in dilute and concentrated solutions (pK a and Ho )
cannot be extended to solid surfaces. In general terms, a solid acid may be viewed
as a surface on which a base is chemically adsorbed and, if the base acts as an
indicator, its color changes upon adsorption. Following both the Brønsted and Lewis
definitions, a solid acid shows a tendency to donate protons or to accept electron pairs,
respectively, while a solid base should tend to accept protons or to donate electron
pairs. For a complete evaluation of the solid acidity/basicity, the knowledge of the
nature, amount and strength of the acid/base sites is of crucial importance, besides
the knowledge of the distribution and density of the sites in nature and strength. This
last point is of great importance when the solid acts as catalyst since the catalyst
activity mainly depends on the amount of active sites but the catalyst selectivity may
be different as a function of the site nature and their distribution and density on the
surface. The amount of acid/base on a solid is usually expressed as the number of
moles of acid/base sites per unit mass or per unit surface area of the solid, and it is
obtained by measuring the amount of a base which is adsorbed on it. The extent of
the adsorption reaction can be studied with different analytical techniques, among
these UV-vis spectroscopy, volumetric and gravimetric apparatus. Others techniques
of study have to be selected to investigate on the nature, strength, and distribution
of the acid/base sites. The need for integrated approach of study clearly emerges to
successful face the solid acidity/basicity question.
8 Characterization of Acid–Base Sites in Oxides 323

Confining our interest on metal oxides, knowing that they are solid compounds of
oxygen and metal or semimetal elements, their surfaces are constituted by positive
and negative arrays of coordinated ions which are bound through ionic or covalent
bonds, depending on the nature of the cations and structure of the oxide. The pos-
itive and negative centers can behave as Lewis acids and Lewis bases that con be
described as a bidimensional organization of AB acid-base pairs (AB)n . Above that,
hydroxyl groups variably cover any real oxide surfaces, as these groups are potential
proton donors or acceptors, it is evident that Brønsted acidity and basicity should
be considered to appear on such surfaces. Of course Lewis and Brønsted species
are often simultaneously present on the same surface where they form densely popu-
lated patches. Consequently due to the complexity of any real metal oxide surface, the
quantitative study of their acid-base properties is not a trivial matter and it deserves
considerable work.

8.3 Acid, Basic, and Amphoteric Oxides

Before providing an overview of the acid-base properties of metal oxide surfaces,

a brief systematic description of their bulk and structural properties and surface
chemistry is needed. According to basic inorganic chemistry concepts, the oxides of
non-metals as well as the oxides of the metals in very high oxidation states are defined
as acidic oxides and anhydrides, respectively, the oxides of metals as basic oxides,
and the oxides having both acidic and basic characters are denoted as amphoteric.
Basic oxides possess low-valency metals (oxidation number lower than or equal
to +4) and they are typically ionic compounds obtained in crystalline forms. In such
oxides the coordination of the cations (from 4 to 8) is generally higher than the
valency (from 1 to 4) and the same occurs for the coordination of the oxide ions
(from 3 to 6). The basic nature of the ionic oxides is associated with the strong
polarization of the metal-oxygen bond and with the formation of basic products
(metal hydroxides) formed from the reaction with water. Examples of relevance in
catalysis are: Cu2 O, CuO, Ag2 O, MgO, BaO, MnO, FeO, La2 O3 , etc. They found
application in various base-catalyzed reactions (alcohol dehydrogenation, olefines,
aromatic carboxylic acids hydrogenation, amination of primary and secondary
amines, Meerwein-Ponndorf-Verley reduction, side-chain alkylation of aromatics,
aldol addition and condensation, Tischenko reaction to produce esters by aldehydes
dimerization, Michael addition, Knoevenagel condensation, etc.) [2].
Acidic oxides are oxides of non-metal elements with high electronegativity char-
acterized by a nearly covalent bond between the non-metal and oxygen. In nor-
mal conditions, some of these oxide structures are gaseous or liquids while others
are solids (e.g., Se4 O12 , P4 O10 ). Decreasing the electronegativity of the element,
covalent network structures are formed (e.g., SiO2 , B2 O3 ). In this case, the cova-
lent nature of the non-metal oxides is evidenced by the coordination number of the
non-metal element which corresponds to its valency or even lower and coordina-
tion of oxygen is two or one. They can form crystalline phases but they usually
324 A. Gervasini

exist in very stable amorphous states that are utilized for wide applications (e.g.,
silica) and convert into stable crystalline phases after treatment at high tempera-
tures. The acid nature of these oxides is associated with the acidity of the products
of their reaction with water: oxo-acids. Different types of acid oxides, denoted as
anhydrides, are constituted of metal elements in very high oxidation states (e.g.,
Mn2 O7 , CrO3 , MoO3 , WO3 , Sb2 O5 , Nb2 O5 , etc.). The very high electronegativity
of the metal imparts them nearly the same properties of non metal oxides. Concern-
ing the structure of these oxides, ii depends on the position of the metal in the period
(from group 4 to group 7); network structure, as TiO2 , layered structure as V2 O5 ,
linear polymeric structure, asCrO3 , and molecular structure, as Mn2 O7 . These metal
oxides react with water giving rise to acid and polyoxoacid products.
Some oxides exhibit both acidic and basic properties, that is, they have amphoteric
properties. In a strongly acidic environment, they act as bases while in a strongly
basic environment they act as acids. In general the electropositive character of the
oxide central atom determines whether the oxide will be acidic or basic. The more
electropositive the metal is the more basic the oxide will be and the more electroneg-
ative the central atom is, the more acidic oxide will be. Electropositive character
increases from right to left across the period and increases down the column for each
period column. A resultant borderline between basic and acidic oxides occurs along
the diagonal of the period including oxides of Zn, Al, Ga, Sb, etc. Alumina represents
the most well-known example of amphoteric oxide able to react with acids and d
bases:
−
Al2 O3(s) + 6 HCl(aq) = 2 Al3+
(aq) + 6Cl(aq) + 3 H2 O(aq) (8.1)
Al2 O3(s) + 2 NaOH(aq) + 3 H2 O(aq) = 2 Na+
(aq) +2 Al(OH)4 −
(aq) (8.2)

Taking into account a given series of metal oxides of different acid-base properties,
including Al2 O3 , BeO, CaO, Cr 2 O3 , Ga2 O3 , La2 O3 , MgO, MoO3 , Nb2 O5 , Nd2 O3 ,
Pr 6 O11 , SiO2 , Ta2 O5 , ThO2 , TiO2 , V2 O5 , WO3 , ZnO, ZrO2 [8, 9], those able to
adsorb at their surface both carbon dioxide (an acid molecular probe) and ammonia
(a basic molecular probe) with evolution of an appreciable amount of heat, typi-
cally in the range of chemisorbed reactions (from 50 to 200 kJ·mol−1 ) are classed
as amphoteric oxides, while if only ammonia or carbon dioxide are retained on a
given surface, the complete acidic or basic character of the oxide, respectively, can
be evinced. Figure 8.1 shows a quantitative description of the acid base characteris-
tics of the above cited series of oxides by using a three-dimensional picture in which
each experimental point of each curve refers to three axes: the amount of adsorbed
ammonia and the amount of adsorbed carbon dioxide (expressed per unit surface
area of oxide sample), and the relevant evolved heat of adsorption (expressed by
the differential heat, kJ·mol−1 ). The pure acidic oxides are settled completely in the
right hand side plane of the diagrams, the basic ones which reacted noticeably with
carbon dioxide only stayed completely in the left hand side of the diagrams, and the
amphoteric oxides can be found in the middle of the diagram planes with their exper-
imental points in the forehead space. For instance, ZnO showed high acidic tendency
8 Characterization of Acid–Base Sites in Oxides 325

Fig. 8.1 Heat evolved for given amounts of CO2 and NH3 adsorbed over a series of metal oxides
(from Ref. [9]); top oxides of Al, Be, La, Mo, Nd, Pr, Si, Ta, Th, and Zr; bottom oxides of Cr, Ca,
Ga, Mg, Nb, Mg, Ti, V, and W

for its strong sites and much more basic tendency for its weak sites (starting from
100 kJ·mol−1 ).

8.4 Heterogeneous Character of Oxides

The heterogeneous character of oxides is due to (i) irregularity in the surface crystal-
lographic structure (different types of crystal planes, growth steps, crystal edges and
corners, Fig. 8.2); (ii) presence of various functional groups (M–O− , M–O–M, M=O,
326 A. Gervasini

Fig. 8.2 Schematic representation of a real oxide particle with irregularity in crystal planes, growth
steps, edges, and corners which causes the presence of surface species with different coordinative
unsaturation (from Ref. [12])

Scheme 8.1 Hydroxyl

groups of metal oxide
surfaces [11]: Type 1,
mono bridging hydroxyl
group; Type II, doubly bridg-
ing hydroxyl group; Type III,
triply bridging hydroxyl group

(M)x –OH, M(–OH)2 , M–[ ], etc. [8, 10]); and (iii) presence of impurities strongly
bounded to the surface. Concerning the acid and basic functionalites of any oxide
surface, different species can be distinguished: (i) exposed oxide species (mono-oxo
or di-oxo surface species), potentially acting as basic oxides; (ii) exposed hydroxyl
groups arising from water adsorption, potentially acting as acid Brønsted sites or
basic sites; (iii) exposed cationic centers, potentially acting as Lewis acid sites; and
(iv) cationic or anionic vacancies. The acid or basic character of the hydroxyl species
is dependent on the nucleophilic character of the hydroxyl group that is governed
by the M–OH bond strength; the proton acidity increases with the increase of the
coordination number of the OH group (Scheme 8.1 [10, 11]).
Other than the intrinsic heterogeneous character of any surface, it is possible
to induce surface heterogeneity by several actions as (i) the mode of formation
of the solid (e.g., by precipitation from liquid phase, flocculation of colloidal sus-
pension, sintering of powder, partial/total re-crystallization, sputtering/condensation
from vapor phase, surface chemical reaction, etc.); (ii) the thermal treatment of the
solid; in fact, heating causes roughness of initially smooth surface by atom dis-
placement and migration to sites of higher energy on top of the surface; and (iii) the
introduction of additional phase (active phase or promoter) that imparts functionality
to the surface.
The complex picture of oxide surfaces can be studied from both qualitative
and quantitative point of view by spectroscopic techniques, conventionally infrared
(FT-IR) but also Raman and NMR spectroscopies [11, 13–15], spectrophotometric
8 Characterization of Acid–Base Sites in Oxides 327

techniques, in particular XPS [16, 17], by directly observing the surface and by the
help of suitable probe molecules adsorption [2, 4, 18–21] which give the possibil-
ity to study the molecular adduct formed between the acid/base site and the probe
[2, 10, 11]. Reasons for the choice of one or another probe are essentially linked
with their size (i.e., kinetic diameter) and basicity [10, 22].
The most used molecular probes are among different chemical families: cyclic
amines (piperidine), alkyl amine (n-butylamine), heterocyclic amines (pyridine,
2-phenylethylamine), phosphines (trimethyl-phosphine), ammonia, carbon monox-
ide. The probe basicity can be expressed as pK a of the conjugated acid and as proton
affinity (PA). The former parameter, pK a , is relevant with respect to acid-base inter-
actions occurring in water solution, also implying solvation with water of both the
base and its conjugated acid. The latter parameter, PA, is more appropriate for gas-
phase interactions. The two basic scales are not coincident, for example ammonia
is associated with a higher pK a and a lower PA than pyridine: This suggests that
ammonia is more basic than pyridine in water but it is less basic than pyridine in
gas phase. The inversion of the relative acidity in water with respect to the gas phase
observed for the ammonia and pyridine probes is due to the greater solvation of
ammonium ion in water compared with the pyridinium ion. At last, the choice of the
probe is also dependent on the analytical technique and the method used to study the
adsorbed molecular acid-base complex formed. For example, a high molecular mass
probe is preferred when the acid-sites of a surface are determined by gravimetric
analysis (by adsorption or thermal desorption), in this case 2-phenylethylamine is a
better choice than pyridine [23, 24].
The quantitative evaluation of the amount of acid/base sites and relative acid/base
strength of a surface can be more easily determined by gravimetric, volumetric,
microcalorimetric, and thermal techniques (Temperature Programmed Desorption,
TPD, among others) [8, 24–35]. Moreover, calorimetric and thermal techniques are
able to provide quantitative information on the strength of interaction between the
acid/base site and the probe, that is, the energy of the bond formed, through the direct
determination of the heat of adsorption or desorption, depending on the experimen-
tal method adopted for the measurements. Such information let us to classify the
acid/base sites of any solid following a scale of acid/base strength with important
relationship with the solid reactivity [8, 36, 37]. When a heterogeneous surface is
concerned, the acid/base site strength distribution can be determined by mathematical
analysis of the experimental data obtained from adsorption and desorption thermal
techniques [30–32, 34, 38]. In this context, the studies of adsorption calorimetry can
give an estimation of the energy distribution of acid or base sites of heterogeneous
surfaces by the direct inspection of the collected calorimetric profiles which show a
more or less pronounced decreasing trend of the differential heat values as a func-
tion of the degree of coverage of the used probe (qdiff versus adsorbed moles). The
decreasing trend reflects heterogeneity of the surface sites but also induced hetero-
geneity, due to lateral interactions between the adsorbed molecules [38, 39], leads to
the observed decreasing trends, so making complicated the quantitative evaluation
of the energy distribution of any surface. Following a common and simplified view-
ing, the different types of site are assumed to be covered by the probe successively
328 A. Gervasini

starting from the strongest ones. This assumption can hold only if large differences
among the adsorption constants are involved [40, 41].
Carniti and Gervasini [42] presented the possibility to study the heterogeneity of
oxide surfaces by performing adsorption calorimetry experiments, by using a suit-
able base probe, at different adsorption temperatures. If on a same sample, adsorp-
tion calorimetry measurements are performed at different temperatures, different
distributions of sites could be observed [42–44]. If no modification of the surface
occurs in the temperature range considered, the different observed distributions can
be due to either kinetic or thermodynamic effects. Regarding the kinetic effects,
at low adsorption temperature, the strongest sites may not be activated or, even if
sufficiently activated, may be covered by the probe at an adsorption rate too low
to achieve equilibrium during the experiment. Moreover, surface diffusion could
not be accomplished if enough time does not elapse in the equilibration. When the
temperature and the time employed in the measurements ensure that the adsorp-
tion equilibrium has been obtained on all sites, the differences among the observed
distributions have to be ascribed to the influence of temperature on the thermody-
namic parameters of adsorption, particularly on the adsorption constants. Accurate
adsorption calorimetric measurements carried out at different temperatures can give
the true site distribution; besides, the thermodynamic parameters of adsorption can
be obtained as functions of temperature. The true distribution can be different from
the observed one obtained by using the above-mentioned oversimplification from a
single adsorption calorimetric curve. Concerning the temperature range employed,
temperatures should be sufficiently high that the adsorption process is activated
on all types of site, but low enough that the adsorption equilibrium constants are
favorable.
Once obtained the data (sets of adsorption and calorimetric isotherms collected
at different temperatures, n and Q, adsorbed moles and evolved heat, respectively),
a computation fitting procedure has to be applied [42]. In the model, the adsorption
isotherms are interpreted as summations of single Langmuir isotherms, each of them
relevant to the sites of a definite type (with nmax,i representing the maximum number
of moles adsorbed on sites of ith type and having a defined adsorption enthalpy with
the probe, a Hi ). The temperature dependence of the Langmuir adsorption constant
of each site-type (bi ) was obtained through the integration of the vant’Hoff differential
equation, considering the adsorption enthalpy either independent or dependent on
temperature by introducing the heat capacity of adsorption (a Cp,i ). On the basis
of the experimental values of n and Q obtained at the investigated temperatures as a
function of equilibrium pressure and surface coverage, respectively, the characteristic
parameters of adsorption (nmax,i , a Hi , number of acid sites and adsorption enthalpy
for each type of ith site, and any other parameters related to them) can be evaluated by
a computational program including an optimization subroutine. When the vant’Hoff
differential equation is considered,

bi = b0,i · exp(−a Hi /RT) or bi = b0,i · exp(−a Hi /RT + (a Cp,i /R) · ln T)

(8.3)
8 Characterization of Acid–Base Sites in Oxides 329

the b0,i pre-exponential parameter has to be optimized too, in order to obtain the
energy distribution of the surface, that is nmax,i versus a Hi . The validity of the
b0,I parameter is difficult to evaluate because it does not have a simple physical
meaning and the range of its numerical values is far from being clearly defined.
A new parameter [45] related to b0,i and defined as half-coverage temperature at unit
0
pressure, T1/2,i , can be optimized instead of b0,i . The new parameter corresponds to
the temperature at which the adsorption constant, bi , becomes unity. In the case of a
Langmuir isotherm, this condition corresponds to half-coverage; therefore the sites
of ith type are half-covered at T = T1/2
0 . Then b
0,I can be easily calculated from the
0
optimized T1/2,i , values.

bi = b0,i · exp(−a Hi /R · T1/2,i

0
) = 1 or bi = b0,i · exp(−a Hi /R · T1/2,i
0

+ (a Cp,i /R) · ln · T1/2,i

0
)=1 (8.4)
b0,i · = exp(a Hi /R · T1/2,i
0
) or b0,i = exp(a Hi /R · T1/2,i
0

− (a Cp,i /R) · ln · T1/2,i

0
) (8.5)

Following the approach above described, the energy distribution of several acidic
simple metal oxides has been determined by using ammonia as probe [42]. The
obtained total number of acid sites (nmax = i nmax,i ) allowed to disclose the fol-
lowing acidity scale: WO3 (5.55 μmol/m2 ) > ZrO2 (5.50 μmol/m2 ) > Nb2 O5
(4.76 μmol/m2 ) < BeO(4.56 μmol/m2 ) > TiO2 (3.20 μmol/m2 ) > Al2 O3
(1.98 μmol/m2 ) (Fig. 8.3). No more than three or four different types of site were
found for each oxide. All the oxide surfaces showed a significant quantity of sites

Fig. 8.3 Acid site energy distribution for several oxides of catalytic relevance (from Ref. [42])
330 A. Gervasini

(70–80 %) associated with a Hi of −40 kJ/mol, corresponding to hydrogen-bonded

ammonia (the weakest mode of interaction). The other 20–30 % of sites was dis-
tributed in a a Hi range of −160 to −280 kJ/mol, corresponding to more energetic
interactions of ammonia with the sites.

8.5 Single Oxides, Doped and Modified Oxides, Supported

Oxides, Mixed Oxides, and Complex Oxides

In catalysis, oxides with well defined acidic and basic properties are used in dif-
ferent forms that have found application in numerous catalytic applications in the
gas-solid and liquid-solid heterogeneous catalysis [3, 46, 47]. Among the most used
oxide materials in catalysis, we find: (i) bulk oxides (one component metal oxides);
(ii) doped and modified oxides; (iii) supported metal oxides (dispersed active oxide
component onto a support oxide component); (iv) bulk and supported binary metal
oxides to quaternary metal oxides (mixed oxide compositions); (v) complex oxides
(e.g., spinels, perovskites, hexa-aluminates, bulk and supported hydrotalcites, pil-
lared clays, bulk and supported heteropolyacids, layered silicas, etc.).
One component metal oxides can crystallize with different morphologies (isotropic,
anisoptropic, or remain amorphous) and local coordination. All one component metal
oxide phases crystallize at high temperatures, providing detectable crystalline XRD
spectra, and many phases can remain amorphous at moderate calcinations tempera-
tures (SiO2 , Al2 O3 , Nb2 O3 , MoO3 , etc.). Their surfaces terminate with various func-
tionalities; for example, amorphous silica terminates with isolated Si–OH, hydrox-
ylpairs of (Si–OH)2 , and Si(–OH)2 , bridging Si–O–Si bonds [48]; niobia terminates
with Nb–OH, Nb–O–Nb, and Nb=O bonds [49, 50]. In general, surfaces terminating
with M=O functionality typically possess M+7 , M+6 , and M+5 cations, while sur-
faces showing M+4 , M+3 , M+2 , and M+1 cations do not possess enough electrons
to form terminal bonds and typically terminate with M–OH or M–O–M bonds. The
literature reports several investigations on acid-base properties of amorphous metal
oxides by adsorption calorimetry in research and review articles [8, 9, 41].
The presence of the different functionalities on the various oxide surfaces leads to
different reactivity properties. When molecules, metal salts, or metal complexes are
adsorbed on the SiO2 surface, for example, the anchoring sites are the more reactive
isolated Si–OH groups rather than the less reactive bridged Si–O–Si groups. Acid
and base sites can often be present simultaneously on oxide surfaces and they can
work independently or in a concerted way. Alumina is the best known example of
amphoteric oxide widespread used in catalysis (Scheme 8.2) [51].
In this case, multiple surface hydroxyl groups varying by the number of Al sites
and Al coordination are present as well as oxygen vacancies of defects, such deter-
mining the characteristic surface chemistry of alumina [48]. As for SiO2 , the isolated
Al–OH sites are the most reactive towards molecules, metal salts, and metal com-
plexes that have to be adsorbed on the alumina surface, these sites are basic, for
8 Characterization of Acid–Base Sites in Oxides 331

Scheme 8.2 Picture of acid and base sites creation on the alumina surface (from Ref. [51])

the most part, while the bridging Al–OH–Al sites are the most acidic ones, and the
Al–[ ] defects possess Lewis acidic character.
There are some other metal oxides crystallize with an anisotropic morphology,
such as platelets [52]. Both MoO3 and V2 O5 crystallize with platelet morphology. In
these structures, the terminal M=O and bridging M–O–M groups are present on the
basal planes and the terminal M–OH predominate on the edge plains. Concerning
surface reactivity of such surfaces, it was demonstrated that the MoO3 basal planes
do not chemisorb methanol while the flat portions of surface containing Mo–OH
sites on the edges react with methanol to form Mo–OCH3 and water [53]. The same
behavior was observed for V2 O5 and ZnO when exposed to methanol [54–56]. On
such materials, the number of catalytically active sites, i.e., acid sites, are only a small
fraction of the total exposed metal oxide sites and we assist to structure sensitivity
of active sites to different crystalline planes.
The recent discovery of mesoporous silica or in general mesoporous molecular
sieves (MMS) has attracted a great deal of attention [57, 58]. The adjustable porosity
of silica-based MMS allows large reactant molecules to penetrate into the internal
void space to be processed at the active-acid sites and then diffuse out freely as
products. Because of the low acidity of the silanol groups of such materials non-
silica mesoporous transition-metal oxide materials have been recently prepared for
catalytic purposes.
Many early transition metal oxides possess high acidity, especially in sulfated and
phosphated forms; titanium, niobium, and tantalum oxides demonstrated very high
activity for large series of acid-catalytic reactions such as benzylation, alkilation,
and isomerisation [57]. For example, sulfated mesoporous Nb oxide has excellent
activity (ca. 200 times greater than that of sulfated bulk Nb oxide) in the benzylation
of anisole or toluene with benzyl alcohol in liquid phase. The mesoporous structure
is maintained even after acid treatment (both with sulfuric or phosphoric acids),
Brønsted and Lewis acid sites coexist in a roughly 50:50 mixture on the surface of
mesoporous Nb oxide while a strong dominance of Brønsted acid sites is present
on the sulfated and phosphated materials [59]. The very high surface area of these
mesoporous materials results in a great increase in the exposed acid sites per mass,
so justifying the observed excellent catalytic activity.
Doped and modified oxides are a wide family of samples which are synthesized
with the aim of modifying some surface property of a given oxide. In particular,
when an acid or base solid has to be used as catalyst, it can happen that the average
acid/base strength of its surface active sites is not useful for the studied reaction.
332 A. Gervasini

The acidic or basic sites can be either too strong causing some irreversible adsorption
of the reactant species or too weakly acidic or basic to activate the reactants. Then, the
oxide surface can be easily modified by incorporation of a second oxide component or
by adding doping species that can regulate their acid-base strength by modification of
the electronic and geometric characteristics of the acid or base sites. The effect of the
addition of small amounts of various ionic species (Ca2+ , Li+ , Nd3+ , Ni2+ , Zr 4+ ,
and SO2− γ
4 ,) on the acid-base properties of -alumina, silica, and magnesia surfaces
was studied by ammonia and sulfur dioxide probe molecules, respectively, in an
adsorption calorimeter [60].
The added metal ion concentration was in the range from 0.1 to 0.3 μmol·m2
corresponding to a surface coverage of about 0.5–1 % of moles of metal ion per
mole of support. It was found that the modification of γ -Al2 O3 surface properties by
the ion dopants only moderately changed its amphoteric properties, the surfaces of
the doped alumina samples remained amphoteric with acid/base pairs independent
of the additives. More substantial changes were observed on magnesia concerning
its basic properties, formation of even more basic sites in the domain of medium
weak sites, but not in the strong sites domain, were revealed (Fig. 8.4 upper). On the
weakly acid silica sample, the number of acid-base centers was strongly affected by
the introduction of the doping ions, as shown in Fig. 8.4 down.
Concerning the relationships between the surface acid-base properties and more
general intrinsic properties of the ions, such as electronegativity, ionicity character
of the cation-oxygen bond, etc., it was found that the acidity of the samples corre-
lated with the charge/radius ratio of the ionic species and with the electronegativity
of the doping ions. The basicity correlated well with the partial oxygen charge asso-
ciated with the cationic dopant [60]. Cardona-Martíinez and Dumesic [61] arrived
at the same conclusion studying the acid properties of a series of doped silicas
with small amounts (0.2–0.3 μmol·m2 ) of Mg2+ , Sc3+ , Fe2+ , Fe3+ , Al3+ , Zn2+ ,
and Ga3+ ions by adsorption microcalorimetry using pyridine as probe. The new
acidity created by the ion introduction could be correlated with the Sanderson elec-
tronegativity of the corresponding oxide formed on the surface (Fig. 8.5). In particu-
lar, the Ga-, Al-, and Sc-silica samples were found to have both Brønsted and Lewis
acidity while all the other samples showed only Lewis acidity.
Besides the doped oxides, modified oxides are gained a prominent role in the cat-
alytic scenario. Many different oxide materials were synthesized chemically mod-
ifying in particular, but not exclusively, silica structures with alumina, titanium,
niobium, tantalum, and zirconium giving rise to many successful materials used in
catalysis as active phases or support phase [23, 24, 34, 62–67].
The possibility of modifying the acid properties of the silica surface by chemical
modification, for example, by covalent anchoring of acid groups or by introducing
elements of other valence [68–71] which create a defect of charge causing the for-
mation of a Brønsted site to balance it, is well known from long time and it has been
well exploited in catalysis to synthesize solid acids with modulated acidity strength.
The integral and differential heats evolved from ammonia adsorption on a series of
oxides comprising silica, alumina, and three modified silicas with amounts of Al,
Ti, and Zr, are shown in Fig. 8.6. The curves for alumina (A), silica-alumina (SA),
8 Characterization of Acid–Base Sites in Oxides 333

Fig. 8.4 Heat of adsorption for NH3 and SO2 probes on doped magnesia samples (upper) and on
doped silica samples (bottom) (from Ref. [60])

and silica-zirconia (SZ) run together while that for silica-titania (ST) lays well below
those of the other samples and the curves for all the samples are very higher than
that of silica. This suggests that by chemical modification of the silica structure it
is possible to enhance the surface acidity of the resultant samples and in particu-
lar that alumina, silica-alumina, and silica-zirconia samples have surfaces with the
highest acidity in terms of number and strength of sites with high heterogeneity of
the surfaces as indicated by the continuous decreasing heat values as a function of
coverage [34].
Among these kinds of oxides, the most important example is silica-alumina that
gained success as substitute for acidic zeolites in many catalytic processes of petrol
chemistry and refining and in general of hydrocarbon chemistry [72]. Synthetic
334 A. Gervasini

Fig. 8.5 Integral heat of

pyridine adsorption as a
function of the Sanderson
electronegativity of the cations
introduced on silica surface
(from Ref. [61])

silica-aluminas are amorphous materials with structure consisting of a random array

of silica and alumina tetrahedral interconnected over three dimensions. In order to
maximize the number of acid sites in a silica-alumina it is important to prepare sam-
ples with the maximum amount of tetrahedrally coordinated Al; in this way, one
acidic proton would be generated by each Al atom, while formation of Al–O–Al
should be avoided. Figure 8.7 reports a comparative view of the acidity of some zeo-
lite and silica-alumina samples [33]. In the case of a zeolite, a typical differential
heat plot shows three regions; the sharp decrease of adsorption heats at low coverage
indicates the presence of small concentration of very strong Lewis type-acid sites.
The plateau of constant heats of adsorption reveals the presence of Brønsted type-
acid sites. The differential heat versus coverage curves for silica-aluminas reveal a
more heterogeneity of the acid sites with a more balanced presence of Brønsted and
Lewis acid sites than zeolite materials that have a predominant Brønsted acidity.
Not only the number and strength of the acid sites but also the accessibility to the
reactant molecules is a target in the case of the solid acids when reaction forecasts
large reactants to transform. A new family of mesoporous aluminosilicates with a
regular array of uniform pores of 20–100 Å diameter has been discovered [73]. These
MCM-41 materials can be synthesized over a large range of framework Si/Al ratios
developing acidity properties [74]. The presence of these very large uniform pores
combined with acidic properties opened new possibilities for cracking heavier feeds
and other hydrocarbon transformation.
Supported metal oxides obtained by spreading an active metal oxide component
over an oxide support, in prevalence of ceramic type (Al2 OI3 , SiO2 , ZrO2 , TiO2 ,
etc.), represent a very large family of oxides widely used in heterogeneous catalysis.
The driving force for this surface wetting of metal oxides is the lower surface free
energy of the final supported metal oxide system. The hydroxylated oxide support
surface possesses a much higher surface free energy than the terminal M=O bonds
8 Characterization of Acid–Base Sites in Oxides 335

Fig. 8.6 Integral (upper) and differential (bottom) heats of ammonia adsorption versus coverage
for a series of modified silicas with alumina, SA, titania, ST, and zirconia, SZ; for comparison the
curves for silica, S, and alumina, A, are reported (from Ref. [34])

formed in the surface metal oxide monolayer. The electronic and molecular struc-
tures of the surface metal oxide species dispersed on oxide supports have received
enormous attention over the past decades because (i) the industrial significance as
catalysts for numerous applications and (ii) the ability to serve as model mixed metal
oxides systems due to the essentially completed dispersed state.
Depending on the mutual characteristics and nature of the supported oxide and
oxide components, supported systems with different properties and stability can
be formed [10]. As a general trend, it is possible to support ionic oxides on ionic
336 A. Gervasini

Fig. 8.7 Differential heat of

ammonia adsorption versus
coverage for a series of zeo-
lites (H-ZSM-5, H-BETA, and
H-MCM-22) and two silica-
aluminas (SAH and SAG)
(from Ref. [33])

oxides even if the resultant compositions are quite unstable due to the ability of
the two oxides to react with each other giving rise to solid solutions or ternary
phases. This occurs, for example, for CuO on alumina which forms CuAl2 O4 spinel.
Only when the supported oxide is formed by big cations (e.g., K2 O, CaO, La2 O3 ,
etc.) unable to enter the close packing of the oxide ions forming the support (e.g.,
Al2 O3 , TiO2 , etc.), fairly stable oxide-on-oxide structures can be formed. Systems
based on covalent oxides deposited on covalent oxides are frequently used in catalysis
(e.g., MoO3 /SiO2 , WO3 /SiO2 , etc.). They are generally viewed as constituted of
small particles of the supported phase weakly interacting with the support surface.
Also covalent oxides supported on ionic oxides are frequently used in catalysis.
Oxides such as vanadia, tungsta, niobia, rhenia can be usefully supported on oxides as
alumina, titania, and zirconia. This possibility gives rise to monolayer-like supported
phases which are quite stable because of the slow reactivity of the two components.
If, however, the ionic oxide support has a too pronounced basic feature the reactivity
between the two components is high and a salt is easily produced (e.g., vanadia
on magnesia gives rise to Mg vanadate). Ionic oxides supported on covalent oxides
are sometimes used in catalysis. The covalent support is almost amorphous silica,
on which many ionic oxides can be supported. The stability of such compositions
appears to be quite limited, due to salt formation.
The acidic metal oxides (e.g., CrOx , MoOx , WOx , VOx , NbOx , TaOx , etc.) usu-
ally anchor to the oxide substrate by preferentially titrating the basic surface hydrox-
yls of the support surfaces [47]. The active basic metal oxides (e.g., NiOx , CoOx ,
InOx , CuO, etc.) usually anchor to the oxide substrate by preferentially titrating the
surface Lewis acid sites of the oxide supports [47, 75]. For many supported metal
oxide systems, the active component can be present as a 100 % dispersion (typically
when Al2 O3 , TiO2 , ZrO2 are used as oxide supports) when its loading on the support
is not high and the support coverage is below the monolayer. Less than 100 % disper-
sion is usually obtained for metal oxides at high loading and when the support surface
8 Characterization of Acid–Base Sites in Oxides 337

Fig. 8.8 Examples of differ-

ent supported oxide systems
concerning the two oxide
component interaction (from
Ref. [77]): a weak interac-
tion; b medium interaction; c
strong interaction, with forma-
tion of α-surface compound,
β-bulk compound, and γ-solid
solution; d composition with
gradient-concentration; e lay-
ered support with adsorbed
particles, f diamond layered
structure

presents lower reactivity of the surface hydroxyls. In addition, it is known that some
active basic metal oxides do not interact strongly with the different oxide functional-
ities of the oxide supports and, consequently, do not disperse to form well-dispersed
phases (e.g., MnOx , CeOx , [47, 76]).
Another important distinction has to be made based on the strength of the inter-
action between the support and the supported phase, designated as active phase, in
view of its catalytic role [77]. By varying the support and the active oxide nature,
different situations can be met which cover a continuous range of interaction strength
between the two phases (Fig. 8.8). When weak interaction strength occurs (Fig. 8.8a),
the supported phase gives rise to isolated crystallites deposited on, but not necessarily
covering, the support phase. For medium interaction strength (Fig. 8.8b), the active
phase tends to spread over the support surface if its loading is low, while, for increas-
ing concentrations, multiple layers, and even distinct crystallites, of active oxide are
built. For strong interaction strength (Fig. 8.8c), a spreading of the supported phase id
favored, but a strong interaction can also lead to a new surface compound (Fig. 8.8c, α)
or to bulk compound (Fig. 8.8c, β) or to solid solution (Fig. 8.8c, γ). Figure 8.8d–f
illustrate other situations with formation of a gradient composition compound, the
layered support with adsorbed particles, and a diamond-layered model, respectively.
The electronic and molecular structures of surface metal oxide species present
on oxide supports have received enormous attention over the past three decades
338 A. Gervasini

because of their industrial significance as catalysts for numerous applications and

their ability to serve as model mixed metal oxide catalytic systems due to the quite
completely dispersed state. Here below, some examples of supported oxide catalysts
are illustrated with emphasis on the possibility to judiciously develop surfaces with
defined acid-base properties.
The modification of the acid-base properties of oxides following the deposition
of variable amounts (from 1 to 50 % of the support surface coverage) of Li+ , Ni2+ ,
and SO2−4 species on supports like alumina, magnesia, and silica have been studied
[78]. Once again the adsorption microcalorimetric experiments using a basic and an
acidic probes, NH3 and SO2 , respectively, was used to study the acid properties of
the surfaces. It has been found that any linear changes in the amount and in strength
of acid-base sites with the increasing addition of lithium, nickel, or sulfate ions to
alumina, silica, or magnesia were found. The addition of small amounts of additives
to alumina has very slight impact on the heat of adsorption and density of acid/base
sites until the attainment of a sudden change when half of its surface was covered.
The possibility to adjust the acid-base properties of bulk oxides by a second oxide
phase deposited on them was more effective on silica owing to its very weak intrinsic
acid character.
Niobia is a fascinating oxide phase used in catalysis because it is a typical strong-
metal-support-interaction (SMSI) oxide; for this property it is used as a support phase
for metal and metal oxides [50, 79]. The strong interaction between niobia surface
and the supported metal species enhances both the redox of the reducible metal
oxides species and the acid properties of the supported phases. Such modifications
could be favorable for given catalytic processes; for example, some group III oxides
(B2 O3 , Al2 O3 , Ga2 O3 , and In2 O3 ) are frequently studied as semiconductor materials
used as sensor devices for the detection of NOx , O2 , H2 O, CO [80–82], and more
recently as de-NOx catalysts [83, 84], or as catalysts for the dehydrogenation or
aromatization of light alkanes [85] in the form of supported phases. When niobia is
concerned as support phase for Al2 O3 (Al/Nb), Ga2 O3 (Ga/Nb), and In2 O3 (In/Nb)
oxides, amphoteric surfaces are obtained giving interesting catalytic properties in
the reaction of dehydrogenation of propene [86]. The Ga/Nb and In/Nb samples
presented differential heat curves of ammonia adsorption clearly different from that
of niobia. Up to a certain adsorbed amount (ca. 25 μmol·g−1 ) the samples have
adsorption heat values higher than those of niobia, attributed to newly created centers
of guest oxide; then the adsorption heats become lower than those of the bare support
(Fig. 8.9, left). This means that gallium and indium oxides are preferentially deposited
on the acid sites of niobia. Concerning the basicity of the surfaces, as determined by
SO2 adsorption, the differential heat curves obtained (Fig. 8.9, right) do not reflect the
basicity order of the group III bulk oxides. Despite the well assessed basic properties
of In2 O3 , the poor In-dispersion on niobia (In/Nb sample) was responsible for the very
lower adsorbed amount of SO2 than on Ga/Nb and, in particular on Al/Nb (possessing
the highest surface area). At this subject, the influence of the support nature in
promoting the dispersion of the supported phase is of fundamental importance. Still
considering the dispersed In2 O3 active phase, it was shown [84] that among a series
of oxide supports (Al2 O3 , Nb2 O5 , SiO2 , and TiO2 ), Al2 O3 and, to a lower extent,
8 Characterization of Acid–Base Sites in Oxides 339

Fig. 8.9 Sulfur dioxide adsorption (left) and ammonia adsorption (right) on some supported group
III oxides (Al2 O3 , Al/Nb, Ga2 O3 , Ga/Nb, and In2 O3 , In/Nb) on niobia. Bulk niobia does not show
any basicity [84]

Fig. 8.10 Representation of

the evolution of supported
WOx domains from isolated
mono tungstates to two-
dimensional polytungstates
and three-dimensional WO3
clusters. The increasing W–O–
W connectivity with increas-
ing domain size is related to
the extent of oligomerization
(from Ref. [87])

TiO2 were found to be the best ones for obtaining active de-NOx catalysts, since the
good In-dispersion and high active surface area.
A peculiar characteristic of the highly dispersed oxides in nanometer size is that
their local structure and electronic properties vary with domain size [87]. For exam-
ple, in the catalysis applications, turnover rates vary as oxide domains evolve from
isolated monomers to two-dimensional oligomers, and into three-dimensional clus-
ters with bulk-like properties. Many examples on dispersed metal oxide domains
in relation to acid catalyzed reactions can be recalled. For example, WOx domains
dispersed on ZrO2 constitute an excellent acid catalyst for xylene isomerization at
low temperatures [87]. The maximum rates of reaction, expressed per W-atom, were
observed at WO3 surface densities of ca. 10 atomW ·nm−2 , corresponding to the
two-dimensional polytungstate monolayer formation. Raman evidences suggested
that W–O–W connectivity (vibrational mode at 807 cm−1 ) which predominate over
the terminal W=O groups (vibrational mode at 1019 cm−1 ) in two-dimensional and
three-dimensional extended WOx oligomers constitute the active acid sites, they
could stabilize the cationic transition state involved in the xylene isomerization.
Figure 8.10 reports the evolution of WOx species on zirconia surface from the small-
est to most wide aggregates.
340 A. Gervasini

Bulk and supported mixed oxide compositions, from binary metal oxides to qua-
ternary metal oxides, consist in general of large crystalline phases possessing low
surface area values (typically from 1 to 10 m2 ·g−1 ). Examples of oxides of this type
of catalytic relevance are V–Nb–O, Mo–Nb–O, Co–Ti–O, Ni–Ti–O, etc. The acid-
base properties of mixed metal oxides have been found to change with the nature
of the constituents and their relative concentrations, preparation (co-precipitation
and sol-gel synthesis among are the most popular methods used), and pre-treatments
procedures. Appropriately choosing the mentioned variables, mixed oxides can be
used to prepare catalysts with the desired-acid-base characteristics.
The example of binary mixed oxides constituted of silica and a second compo-
nent like alumina, titania, and zirconia has been above reported. The so constituted
surfaces have different acidity properties in terms of number of sites and site strength
distribution, depending on the nature of the second component (Al, Ti, or Zr), the
Si/Al, Si/Ti, and Si/Zr ratios, and the final sample structure [23, 34]. In catalysis
such oxides can be used as support phases as well as acid catalysts. When they are
used as support phases, the acid sites serve as anchoring points for the supported
phase stabilizing it in the dispersed state, even when the samples operate under high
temperature conditions. It was demonstrated by XPS spectrophotometry [23] that for
silica-zirconia compositions, Si/Zr, it was possible to regularly increase the amount
of surface acid sites by tuning the amount of Zr during preparation (Six Zr 1−x O2 with
0.715 < x < 1, corresponding to ZrO2 from 5 to 45 wt% in the silica) (Fig. 8.11). On
the different synthesized Si/Zr acidic surfaces, the iron oxide dispersed phase was
supported and new acidic surfaces were created, Fe/Six Zr 1−x O2 , differing from the
relevant bare supports for the nature, strength, and amount of the acid sites.
The average acid strength of the Fe-samples was higher than that of Six Zr 1−x O2
supports and tuned towards a prominent Lewis acidity. Concerning the amount of
acid sites of the Fe-catalysts, they were higher or lesser compared with that of the
relevant supports depending on the acidity of the bare support. When a portion of
highly acidic surface of support (Si/Zr samples at high Zr content) was covered by
iron oxide having lower surface area than its support, a decrease of the number of
surface acid sites was observed (Fig. 8.12, see the SZ-15, SZ-30, and SZ-45 supports
and relevant Fe-catalysts) while when low acidic surfaces, as pure silica or low
containing zirconium samples were covered by iron oxide an increase of the amount
of acid sites was observed (Fig. 8.12, see the S and SZ-5 supports and relevant Fe-
catalysts). It is then possible to tune not only the amount of the acid sites but also the
Brønsted and Lewis nature and the acid strength of a surface by a judicious presence
of different components at the surface.
Silica-alumina has been used to support the CuO coupled with Ga2 O3 , and SnO2
dispersed phases to enhance the catalytic properties of CuO-based catalysts in reac-
tions of environmental importance (hydrocarbon combustion, NO and N2 O decom-
position and reduction [88]). The acidic properties of such oxide systems were studied
from a qualitative (nature of the acid sites) and quantitative (number, acid strength,
and strength distribution of acid sites) points of view through the adsorption and
desorption of two basic probes (ammonia and pyridine) by coupled volumetric-
calorimetric technique and XPS and FT-IR spectroscopy.
8 Characterization of Acid–Base Sites in Oxides 341

Fig. 8.11 Relation between

the amount of acid sites of
Six Zr 1−x O2 oxides deter-
mined by 2-phenylethylamine
probe and the amount of
surface Zr concentration
determined by XPS (from
Ref. [23])

Fig. 8.12 Amount

of acid sites of the
Six Zr 1−x O2 supports in com-
parison with the Fe oxide
covered surfaces, determined
by 2-phenylethylamine probe
(S, pure silica, and SZ-X,
silica-zirconia oxides with 5
< x < 45 zirconia mass con-
centration) (from Ref. [23])

The Brønsted acidity of the silica-alumina support was converted into predom-
inantly Lewis acidity upon metal oxide deposition, in particular that of the CuO
supported phase. The highest amount of Lewis acid centers were formed on the
CuO sample, the partial coverage of the Cu phase by Ga-phase caused a marked
decrease of the amount of Lewis acid centers of the Cu/Ga surfaces. The Lewis
to Brønsted acid site ratio of the Cu/Sn samples was higher compared to that
of Cu/Ga samples, accounting for a higher surface Sn-dispersion. Concerning the
acid strength of the surfaces, moderate acidity was associated with the Cu sites
(100 kJ·mol−1 < qdiff < 150 kJ·mol−1 ) whereas the most acidic fraction of the sites
(qdiff > 150 kJ·mol−1 ) increased with the presence of Ga and Sn (Fig. 8.13). The
results have been discussed in the light of the intrinsic acidity of the Cu, Ga, and Sn
metal ions derived from electronic parameters (in particular electronegativity).
Complex oxides comprise a large variety of structures which gained importance in
the catalysis field due to their acidity or basicity properties (e.g., spinels, perovskites,
hexa-aluminates, bulk and supported hydrotalcites, pillared clays, bulk and supported
heteropolyacids, etc.).
342 A. Gervasini

Fig. 8.13 Acid site strength percent distributions of the Cu/Ga (a) and Cu/Sn (b) series samples sup-
ported on silica-alumina (SA) as obtained from ammonia adsorption by calorimetric measurements
(the Ga/Cu and Sn/Cu atomic ratios, 25, 50, or 75, are indicated) (from Ref. [88])

Hydrotalcites are gained an important position, in particular, in basic catalysis.

They are layered double hydroxides containing a divalent ion (e.g., Mg2+ , Ca2+ ,
Zn2+ , or Ni2+ ) a trivalent ion (e.g., Al3+ , Fe3+ , Cr 3+ ) and a charge compensating
anion (e.g., OH− , Cl− , NO− 2− 2−
3 , CO3 , or SO4 ). Thermal treatments cause dehydra-
tion and dehydroxylation and loss of anions giving rise to mixed oxide with MgO-type
structures. The acid-base properties of Mg–Al oxides are governed by the Mg to Al
molar ratio, besides the calcination temperature and preparation procedure. Increas-
ing the Mg:Al ratio of the hydrotalcites, increase of the total number of basic sites
is observed. The thermal decomposition of hydrotalcites gives rise to mixed oxides
with basic properties and they can act as precursors for the preparation of basic oxides
with catalytic activity [89]. Several reviews have been reported recently concerning
the application of Mg–Al or calcined hydrotalcites as basic materials in particular in
the field of fine organic chemistry [90–92].
Among the mixed oxides, those with perovskite-like structure are among the
most widely studied. They have general formula ABO3 where A is usually a large-
radius metallic cation (e.g., lanthanide, alkali-earth metal ion) and B is a lower
size transition metal ions. The principal advantages of this oxide family are the
high versatility of their composition that can give compounds with widely different
properties and their exceptional stability of the structures. The crystal field force of the
perovskite structure can force some ions to assume unusual oxidation states. A typical
example is Cu in Ba-substituted lanthanum cuprate, in which Cu is present in both
2+ and 3+ oxidation states [93]. Another interesting property of perovskites is their
nonstoichimetry. As the whole structure has to be neutral, so cation charge have to
be distributed to give an overall +6 sum: A1+ B5+ O3 , or A2+ B4+ O3 , or A3+ B3+ O3 .
The presence of defects, like anionic or cationic vacancies, is very common in real
structures. Anionic vacancies are the most frequent and they can extend up to a whole
oxygen layer, leading to a different family of compounds, the brownmillerites [94],
8 Characterization of Acid–Base Sites in Oxides 343

like Ca2 Fe2 O5 and La2 Ni2 O5 in which one sixth of the oxygen atoms are vacant.
In other cases, an excess oxygen can be found; this is the case of LaMnO3+δ .which
shows very good catalytic activity for total oxidation reactions. This defect or excess
of charge is connected with acid or basic properties of these structures which found
wide application in the hydrocarbon catalytic combustion processes rather than in the
acid-base catalytic reactions, due to their excellent thermal and chemical stability.
Another interesting family of oxide compounds are the heteropolycompounds
possessing Keggin-type structure. They consist of heteropolyanions and counter-
cations such as NH4+ , Cs+ , H+ . When the counter-cations are protons, the com-
pounds are called heteropolyacids, they can have very strong Brønsted acid sites,
like H3 PW12 O40 . They can have different acid strength depending on the nature of
the compounds but always presents a plateau of sites of the same strength. For the cat-
alytic application, they are often supported on high-surface area oxides or activated
carbons to increase the surface contact with the reactants of the fluid phase. They
are used in several acid-catalyzed reactions, like the hydration of alkanes, synthesis
of methacrolein, isobuturric acid, etc. Various studies are reported in the literature
[95–97] concerning the acid strength of heteropolyacids, influence of the support
on acidity, influence of the acid site distribution on the substitution of protons with
cationic species, change in acidity upon heat treatment, etc.

8.6 Acidity Prediction from Composition

From practical and theoretical points of view concerning binary metal oxides, it is
interesting to find oxide combinations having well defined and tunable acid or basic
properties. On a catalytic oxide surface, the acid or basic sites can be either too
strong causing some irreversible adsorption of the substrate species or the sites can
be too weak to activate the substrate species. Therefore, the possibility to regulate
the acid-base strength, besides the acid site amount of the oxide surfaces, appears a
necessary tool for catalytic purposes.
The acid/base strength of an oxide surface may be enhanced or decreased by the
addition of a secondary component which modifies the electronic and/or geomet-
ric characteristics of the parent acid/base sites or creates new acid/base centers. The
oxides belonging to the silica-alumina system, wherein the Al introduction generates
new Brønsted or Lewis acid sites, are representative examples of acid binary oxides.
Several authors have found generation of acidity also for other mixed oxide sys-
tems (silica-zirconia, silica-niobia, silica-titania, [23, 51, 98]). Then, several models
appeared in the literature (Thomas [99], Tanaka [100], Tanabe [101], Seiyama [102],
Dumesic [103], and Gervasini [60]) which attempted to generalize the experimental
observations picturing acidity generation mechanisms.
The Tanabe model for acidity prediction in mixed oxide compositions is based on
the interaction of the oxides at molecular level, the interaction generates an excess
of negative or positive charge in the mixed composition localized around the guest
element. According to this model, the substitution/introduction of a metal ion into
344 A. Gervasini

the structure of a host oxide, whether the charge is an excess or not, and whether it is
positive or negative, may generate acid sites. They are determined by the coordination
numbers, C, and valences, V, of the positive and negative elements in the model oxide
structure pictured according to two postulates: (i) the coordination numbers of the
positive elements of the metal oxides are maintained even when mixed; (ii) the
coordination number of oxygen of the major component oxide is retained for all the
oxygen atoms in the mixed composition. It is then possible to explain the mechanisms
of the acidity generation of mixed oxides and to predict whether the generated acid
sites will be of the Brønsted or the Lewis type.

 = (VA /NCA − VO /NCO ) · NCA (8.6)

where  is the excess/ of charge, V/NC is the ionic valence to coordination number
ratio, and the A and O subscripts concern the positive element of the added ion and
the negative element (oxygen) of the major component oxide, respectively. In any
case, Lewis acidity is assumed to appear upon the presence of an excess of positive
charge and Brønsted acidity for an excess of negative charge.
Some model oxide structures pictured according to Tanabe’s model are here below
reported.
Silica-titania, with silica the major component oxide:
 = (+4/6 − 2/2) · 2 = −2

O OO
O Si O Ti O
O OO

Titania-silica, with titania the major component oxide:

 = (+4/4 − 2/3) · 4 = +4/3

OO O
O Ti O Si O
OO O

Silica-niobia with Nb tetrahedral coordination (NbO4 ):

8 Characterization of Acid–Base Sites in Oxides 345

 = (+5/4 − 2/2) · 4 = +1

O O
O Si O Nb O
O O

Silica-niobia with Nb octahedral coordination (NbO6 ):

 = (+5/6 − 2/2) · 6 = −1

OO O
O Si O Nb O
O OO

Silica-zirconia wirh 8-fold coordination of Zr (fluorite-like structure):

 = (+4/8 − 2/2) · 8 = −4

O OO
O
O Si O Zr O
O
O OO

Seiyama has presented a different model in which the oxygen bridging between
the two different metal ions develops a positive or negative charge as a consequence
of the different coordination of the two cations. In this case too, it is possible to
calculate the effective charge of oxygen as the sum of the boundary charges of the
two oxides:
 = (VA /NCA + VS /NCS ) · −2 (8.7)

where the S subscript concerns the positive element of the major component oxide
(support oxide).
There is not very often accordance among the different models for the acidity
prediction of a given oxide structure. Moreover, it is also hard to justify the predicted
acid properties with the catalytic activity of the oxide composition; this is the case of
titania-silica system [67, 104]. TiO2 −SiO2 mixed oxide is a very important industrial
material and catalyst in both the amorphous and crystalline phases which found
several industrial applications (e.g., isomerization of olefins, epoxidation of olefins
346 A. Gervasini

with hydroperoxides, selective oxidation of a number of organic compounds, etc.).

The TiO2 −SiO2 activity in the 1-butene isomerization, phenol ammination, and
ethane hydration was attributed to the formation of Brønsted acidity [105]. While
for the oxidation reactions there is a general consensus that the active sites are
tetrahedrally coordinated Ti4+ isolated in SiO2 matrix, less clear is the source of
the activity in the other reactions which demand acid sites. Because the catalytic
activity could not be explained by the acid properties of the two pure oxides: in
fact TiO2 has only Lewis acid sites and the silanol groups of SiO2 are too weakly
acidic for any reaction requiring an acid catalysis. The performances of TiO2 −SiO2
as acid catalyst were at first explained by Tanabe on the basis of his theory of acidity
generation in mixed compositions. From his first hypothesis, many authors studied
the TiO2 −SiO2 system and support the Tanabe hypothesis. Notari et al. [67] did not
agree with the literature evidences on the acidity of TiO2 −SiO2 and confuted them
by selected experiments. In particular, he proved that highly divided TiO2 operated
the same acid reactions that TiO2 −SiO2 did with a radical or anion mechanism.
In some other different examples, acidity prediction of an oxide system correlates
well with its activity. This is the case of TiO2 −SnO2 which generated strong new
acid sites. The number of acid sites become maximum at the composition TiO2 50 %.
At this composition, the oxide shows the maximum catalytic activity in the butane
isomerization reaction [106].

8.7 Intrinsic and Effective Acidity of Oxide Surfaces

The distinction between the intrinsic acidity of a solid and the effective acidity dis-
played when the surface acidic groups are screened by interaction with solvent mole-
cules becomes a topic of prominent importance when the solid has to work in contact
with liquids for its practical uses. This is the case of liquid-solid heterogeneous cat-
alysts in which the activity of the catalyst can be modified by the presence of solvent
which may establish physical or chemical interactions with the acid sites of the sur-
face. For reactions carried out in liquid phase, the knowledge of the effective acidity
(in terms of number and strength of the sites) of the catalyst in given liquids allows
determining sound relationships between the catalytic activity and surface acidity.
Adsorption calorimetry in liquid is a complex matter of study because many vari-
ables are simultaneously present: solids with different acidity in terms of distributions
of acid sites (nature and acid strength); liquids of different polarity, proticity, and
solvating ability; and probes with different basicity (pKa scale). Therefore, results
permit comparing for a given solid and probe couple, the influence of the solvent;
for a given solid and liquid couple, the influence of the probe; and for a given liquid
and probe couple, the properties of different solids.
One of the main goals of the calorimetric experiments of acid-base titrations in
liquid is to evaluate the effective acid strengths of the surfaces when suspended in
liquids of different nature. As for the gas-solid acid-base titrations, solutions of base
probe in liquids of various natures are used for titrating the acid sites of the solid which
8 Characterization of Acid–Base Sites in Oxides 347

is placed in a reaction calorimeter equipped with a stirring system. Successive pulse

injections of dosed amounts of the probe solution are sent onto the sample maintained
at constant temperature up to surface saturation with the probe. In this case, the
evolved heat from the adsorption reaction derives from two different contributions:
the exothermic enthalpy of adsorption, ad Hprobe , and the endothermic enthalpy
of displacement of the liquid, dpl Hliquid , the enthalpy effects describing dilution
and mixing phenomena can be neglected depending on the differential design and
preheating of the probe solution.
Various papers have recently been published concerning the liquid-phase titration
of acid solids, such as acidic polymeric resins [107–109]. Example is here pre-
sented on the study of the intrinsic and effective acidities of two catalysts based on
niobium: niobium oxide (NBO) and niobium phosphate (NBP) [110] which found
application in reaction of acid transformation of monosaccharides (fructose, glu-
cose, in particular) to useful chemicals, like 5-hydroxymethyl-2-furaldehyde (HMF)
[111, 112].
When a strongly basic probe, like 2-phenylethylamine (PEA), in the decane sol-
vent is used for titrating the NBO and NBP surfaces (Fig. 8.14), the information
obtained on the two acid surfaces support the conclusions obtained from the more
classical determinations of acidity using gas-solid calorimetric adsorption and ther-
mal desorption approaches [110]. The calorimetric curve of acid-base titration in
decane for NBP always lies above that of NBO, and the difference between the two
curves increases as the titration progresses further. This seems confirming that the
main difference between the NBO and NBP surfaces has to be ascribed to the medium
and weak strength acid sites.

5
Evolved heat for each dose (J·g -1)

2
NBO

1 NBP

0
0 10 20 30 40 50 60 70
PEA introduced (µmol)

Fig. 8.14 Evolved heat for each dose of PEA solution injected as a function of the total amount
of PEA injected in the liquid-solid adsorption experiments performed at 40 ◦ C in decane (from
Ref. [110])
348 A. Gervasini

(a)
decane
4
cyclohexane
Evolved heat for each dose (J·g -1)

iso-propanol
toluene
3

0
0 100 200 300 400 500 600
Aniline introduced ( µmol·g-1 )

(b)
decane
4
cyclohexane
Evolved heat for each dose (J·g -1)

iso-propanol
toluene
3

0
0 200 400 600 800 1000

Aniline introduced ( µmol·g )

-1

Fig. 8.15 Evolved heat for each dose of aniline solution injected as a function of the total amount
of aniline injected during the liquid-solid adsorption experiments performed at 40 ◦ C in various
liquids: NBO (a) and NBP (b) (from Ref. [110])

The different effective acid strengths of the surfaces of NBO and NBP can be evi-
denced by the comparative experimental results summarized in Fig. 8.15a,b, respec-
tively, obtained by using aniline as base probe. In these figures, the areas of the heat
flow peaks obtained during aniline titration in cyclohexane, decane, toluene, and
isopropanol are plotted as functions of the amount of aniline introduced. The higher
8 Characterization of Acid–Base Sites in Oxides 349

acidity of NBP compared to that of NBO, in terms of both the number and strength
of the sites, is clearly evident. The total heat evolved following the introduction of
a definite amount of aniline, and the amount of aniline necessary to complete the
adsorption reaction, are in all cases higher for NBP than for NBO, independently
of the nature of the liquid. Since the evolved heat is a sum of two contributions
with opposite signs (ad Hprobe , adsorption enthalpy, exothermic, and dpl Hliquid ,
enthalpy of displacement of the liquid, endothermic), the higher the endothermic
contribution, the lower the experimental heat flow measured. For a given solid and
probe, ad Hprobe depends only on the solid acidity, while dpl Hliquid can be differ-
ent, depending on solid-solvent interactions. It may then be argued that the surface
acid sites of NBO (Fig. 8.15a) present a stronger interaction with cyclohexane and
decane than with toluene; the dpl Hliquid contribution is then higher in cyclohexane
and decane, which results in lower measured heats than in toluene.
A completely different picture emerges when aniline titration is carried out in
isopropanol, a protic solvent with high polarity. On both NBO and NBP, very low
heats evolved have been measured; the initial value of evolved heat obtained for NBP
is much higher than that for NBO (Fig. 8.15a,b). This confirms the ability of NBP to
maintain a highly acidic character in solvents of high polarity, like alcohols or water.
The acid performances of the NBO and NBP catalysts were compared in the dehy-
dration of fructose to 5-hydroxymethyl-2-furaldehyde (HMF). The reaction runs with
high conversion and good selectivity to HMF in organic solvents but in agreement
with a green chemistry process development, friendly solvents have to be used, like
water or alcohol mixtures. In water, NBP shows higher fructose conversion and
HMF selectivity than NBO, so justifying its higher effective acidity in polar and
protic liquids [110].
The calorimetric titration of the acid sites in liquid phase makes it possible to dis-
criminate the acid site strength more accurately than the more conventional gas-solid
phase titration in order to understand the effective acidity of the solids measured in
various liquids of very different polarities and proticities. The obtained calorimetric
results are of not easy interpretation since solute-solvent and solid-solvent interac-
tions have to be taken into account, besides the solute-site interaction.

References

1. G. Busca, Chem. Rev. 107, 5366 (2007)

2. H. Hattori, Chem. Rev. 95, 537 (1995)
3. K. Tanabe, W.F. Hölderich, Appl. Catal. A 181, 399 (1999)
4. G. Busca, Chem. Rev. 110, 2217 (2010)
5. R.G. Pearson, J. Am. Chem. Soc. 85, 3533 (1963)
6. G. Klopman, J. Am. Chem. Soc. 90, 223 (1968)
7. J.L. Reed, J. Phys. Chem. 98, 10477 (1994)
8. A. Auroux, A. Gervasini, J. Phys. Chem. 94, 6371 (1990)
9. A. Gervasini, A. Auroux, J. Therm. Anal. 37, 1737 (1991)
10. G. Busca, Phys. Chem. Chem. Phys. 1, 723 (1999)
11. J.C. Lavalley, Catal. Today 27, 377 (1996)
350 A. Gervasini

12. S. Coluccia, A.J. Tench, Proceedings of the 7th International Congress on Catalysis, Tokyo,
Japan, Kodansha, 1980
13. I.E. Wachs, Catal. Today 27, 437 (1996)
14. G. Busca, Catal. Today 41, 191 (1998)
15. F. Haw, I.S. Chuang, B.L. Hawkins, G.E. Maciel, J. Am. Chem. Soc. 105, 7206 (1983)
16. M. Johansson, K. Klier, Top. Catal. 4, 99 (1997)
17. R. Borade, A. Sayari, A. Adnot, S. Kaliaguine, J. Phys. Chem. 94, 5989 (1990)
18. D. Barthomeuf, G. Coudurier, J.C. Vedrine, Mater. Chem. Phys. 18, 553 (1988)
19. R.J. Gorte, D. White, Top. Catal. 4, 57 (1997)
20. A. Boréave, A. Auroux, C. Guimon, Microporous Mater. 11, 275 (1997)
21. J. Kotrla, L. Kubelková, C.-C. Lee, R.J. Gorte, J. Phys. Chem. B 102, 1437 (1998)
22. G. Busca, in Metal Oxide Catalysis, ed. by S.D. Jackson, J.S.J. Hargreaves. The Use of Infrared
Spectroscopy Methods in the Field of Heterogeneous Catalysis by Metal Oxide, vol 1 (Wiley-
VCH, Weinheim, 2009), p. 95
23. A. Gervasini, C. Messi, D. Flahaut, C. Guimon, Appl. Catal. A 367, 113 (2009)
24. P. Carniti, A. Gervasini, S. Bennici, J. Phys. Chem. B 109, 1528 (2005)
25. V. Solinas, I. Ferino, Catal. Today 41, 179 (1998)
26. A. Auroux, Mol. Sieves 6, 45 (2008)
27. A. Auroux, Top. Catal. 19, 205 (2002)
28. A. Auroux, in Catalyst characterization: physical øøøtechniques for solid materials, ed. by B.
Imelik, J.C. Vedriine (Plenum Press, New York, 1994), p. 611
29. M. Fadoni, L. Lucarelli, in Adsorption and its Applications in Industry and Environmental
Protection, ed. by A. Dabrowski. Applications in Industry, vol 1 (Stud. Surf. Sci. Catal. 120A)
(1999)
30. H.G. Karge, V. Dondur, J. Phys. Chem. 94, 765 (1990)
31. H.G. Karge, V. Dondur, J. Weitkamp, J. Phys. Chem. 95, 283 (1991)
32. F. Arena, R. Dario, A. Parmaliana, Appl. Catal. A 170, 127 (1998)
33. B. Dragoi, A. Gervasini, E. Dumitriu, A. Auroux, Thermochim. Acta 420, 127 (2004)
34. A. Gervasini, P. Carniti, A. Auroux, Thermochim. Acta 434, 42 (2005)
35. S. Bennici, A. Auroux, in Metal Oxide Catalysis, ed. by S.D. Jackson, J.S.J. Hargreaves.
Thermal Analysis and Calorimetric Methods, vol 1 (Wiley-VCH, Weinheim, 2009), p. 391
36. A. Auroux, Top. Catal. 4, 71 (1997)
37. A. Auroux, A. Gervasini, L. Nemeth, G. Gati, I.S. Pap, G. Mink, Surf. Interface Anal. 19, 529
(1992)
38. W. Rudzinski, T. Borowiecki, T. Panczyk, A. Dominko, Langmuir 16, 8037 (2000)
39. M.V. Gargiulo, J.L. Sales, M. Ciacera, G. Zgrablich, Surf. Sci. 501, 282 (2002)
40. W. Rudzinski, D.H. Everett, Adsorption of Gases on Heterogeneous Surfaces (Academic Press,
London, 1992)
41. N. Cardona-Martinez, J.A. Dumesic, Adv. Catal. 38, 149 (1992)
42. P. Carniti, A. Gervasini, A. Auroux, J. Catal. 150, 274 (1994)
43. K. Tsutsumi, Y. Mitani, H. Takahashi, Bull. Chem. Soc. Jpn. 56, 1912 (1983)
44. N. Cardona-Martinez, J.A. Dumesic, J. Catal. 125, 427 (1990)
45. P. Carniti, A. Gervasini, React. Kinet. Catal. Lett. 52, 285 (1994)
46. S.D. Jackson, J.S.J. Hargreaves (eds.), Wiley-VCH, Weinheim, vol I and II, 2009 (ISBN 978-
3-527-31815-5).
47. I.E. Wachs, Catal. Today 100, 79 (2005)
48. H.-P. Boehnm, H. Knozinger, in Nature and Estimation of Functional Groups on Solid Surfaces,
ed. by J.R. Anderson, M. Boudart. Catalysis, Science and Technology, vol 4 (Springer, Berlin,
1984)
49. I. Nowak, M. Ziolek, Chem. Rev. 99, 3603 (1999)
50. M. Ziolek, Catal. Today 78, 47 (2003)
51. K. Tanabe, in Solid Acid and Base Catalysts, ed. by J.R. Anderson, M. Boudart. Catalysis,
Science and Technology, vol 5 (Springer, Berlin, 1987), p. 232
8 Characterization of Acid–Base Sites in Oxides 351

52. J. Haber, in Crystallography of Catalyst Types, ed. by M. Boudart, J.R. Anderson. Catalysis,
vol 2 (Springer, Berlin, 1984)
53. W.E. Farneth, F. Ohuchi, R.H. Staley, U. Chowdhry, A.W. Sleight, J. Phys. Chem. 89, 2493
(1985)
54. M. Badlani, I.E. Wachs, Catal. Lett. 75, 137 (2001)
55. I.E. Wachs, Y. Chen, J.-M. Jehng, L.E. Briand, T. Tanaka, Catal. Today 78, 13 (2003)
56. L.E. Briand, J.-M. Jehng, L. Cornaglia, A.M. Hirt, I.E. Wachs, Catal. Today 78, 257 (2003)
57. Y. Rao, D.M. Antonelli, J. Mater. Chem. 19, 1937 (2009)
58. Y. Rao, J. Kang, M. Trudeau, D.M. Antonelli, J. Catal. 266, 1 (2009)
59. Y. Rao, M.L. Trudeau, D.M. Antonelli, J. Am. Chem. Soc. 128, 13996 (2006)
60. A. Gervasini, G. Bellussi, J. Fenyvesi, A. Auroux, J. Phys. Chem. 99, 5117 (1995)
61. N. Cardona-Martínez, J.A. Dumesic, J. Catal. 127, 706 (1991)
62. E. Cano-Serrano, G. Blanco-Brieva, J.M. Campos-Martin, J.L.G. Fierro, Langmuir 19, 7621
(2003)
63. A. Gervasini, C. Messi, A. Ponti, S. Cenedese, N. Ravasio, J. Phys. Chem. C 112, 4635 (2008)
64. C. Flego, L. Carluccio, C. Rizzo, C. Perego, Catal. Commun. 2, 43 (2001)
65. T. Klimova, M.-L. Rojas, P. Castello, R. Cuevas, J. Ramírez, Microporous Mesoporous Mater.
20, 293 (1998)
66. G. Guiu, P. Grange, J. Catal. 156, 132 (1995)
67. B. Notari, R.J. Willey, M. Panizza, G. Busca, Catal. Today 116, 99 (2006)
68. A. Desmartin-Chomel, J.L. Flores, A. Bourane, J.M. Clacens, F. Figueras, G. Delahay, A.
Giroir-Fendler, C. Lehaut-Burnouf, J. Phys. Chem. 110, 858 (2006)
69. S.Y. Kim, J.G. Goodwin, S. Hammache, A. Auroux, D. Galloway, J. Catal. 201, 1 (2001)
70. M. Occelli, D.A. Schiraldi, A. Auroux, K.A. Keogh, B.H. Davis, Appl. Catal. A 209, 165
(2001)
71. D. Deutsch, V. Quaschning, E. Kemnitz, A. Auroux, H. Ehwald, H. Lieske, Top. Catal. 13, 281
(2000)
72. A. Corma, Chem. Rev. 95, 559 (1995)
73. J.S. Beck, C.W. Chu, I.D. Johnson, C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, JWO
91/11390 (1991)
74. A. Corma, V. Fornés, M.T. Navarro, J. Pérez-Pariente, J. Catal. 148, 569 (1994)
75. M.A. Vuurman, D.J. Stufkens, A. Oskam, G. Deo, I.E. Wachs, J. Chem. Soc. Faraday Trans.
92, 3259 (1996)
76. F. Kapteijn, A.D. Vanlangeveld, J.A. Moulijn, A. Andreini, M.A. Vuurman, A.M. Turek, J.-M.
Jehng, I.E. Wachs, J. Catal. 150, 94 (1994)
77. A. Cimino, D. Gazzoli, M. Valigi, J. Electron Spectrosc, Relat. Phenom. 104, 1 (1999)
78. A. Gervasini, J. Fenyvesi, A. Auroux, Langmuir 12, 5356 (1996)
79. T. Uchijima, Catal. Today 28, 105 (1996)
80. C. Rizzo, A. Carati, M. Tagliabue, C. Perego, Stud. Surf. Sci. Catal. 128, 137 (2000)
81. R. Pohle, M. Fleischer, H. Meixner, Sens. Actuators B 68, 151 (2000)
82. M. Kudo, T. Kosaka, Y. Takahashi, H. Kokusen, N. Sotani, S. Hasegawa, Sens. Actuators B
69, 10 (2000)
83. J.A. Perdigón-Melón, A. Gervasini, A. Auroux, J. Catal. 234, 421 (2005)
84. A. Gervasini, J.A. Perdigón-Melón, C. Guimon, A. Auroux, J. Phys. Chem. B 110, 240 (2006)
85. K. Nakagawa, C. Kajita, Y. Ide, M. Okamura, S. Kato, H. Kasuya, N. Ikenaga, T. Kobayashi,
T. Suzuki, Catal. Lett. 64, 215 (2000)
86. A.L. Petre, J.A. Perdigón-Melón, A. Gervasini, A. Auroux, Catal. Today 78, 377 (2003)
87. J. Macht, E. Iglesia, Phys. Chem. Chem. Phys. 10, 5331 (2008)
88. S. Bennici, A. Auroux, C. Guimon, A. Gervasini, Chem. Mater. 18, 3641 (2006)
89. D. Tichit, M.H. Lhouty, A. Guida, B.H. Chiche, F. Figueras, A. Auroux, D. Bartolani, E.
Garrone, J. Catal. 151, 50 (1995)
90. A. Vaccari, Appl. Clay Sci. 14, 161 (1999)
91. D. Tichit, B. Coq, CATTECH 7, 206 (2003)
92. A. Corma, S. Iborra, Adv. Catal. 49, 239 (2006)
352 A. Gervasini

93. M.A. Peña, J.L.G. Fierro, Chem. Rev. 101, 1981 (2001)
94. M.J. Sayagues, M. Vallet-Regi, A. Caneiro, J.M. Gonzales Calbet, J. Solid State Chem. 110,
295 (1994)
95. F.X. Liu-Cai, B. Sahut, E. Faydi, A. Auroux, G. Hervé Appl, Catal. A 185, 75 (1999)
96. L. Damjanovic, V. Rakic, U.B. Mioc, A. Auroux, Thermochim. Acta 434, 81 (2005)
97. T. Nakato, M. Kimura, S. Nakata, T. Okuhara, Langmuir 14, 319 (1998)
98. P. Carniti, A. Gervasini, M. Marzo, J. Phys. Chem. C 112, 14064 (2008)
99. C.L. Thomas, Ind. Eng. Chem. 41, 2564 (1949)
100. K.I. Tanaka, A. Ozaki, J. Catal. 8, 1 (1967)
101. K. Tanabe, T. Sumiyoshi, K. Shibata, T. Kiyoura, J. Kitagawa, Bull. Chem. Soc. Japan 47,
1064 (1974)
102. T. Seiyama, Metal Oxides and Their Catalytic Actions (Kodansha, Tokyo, 1978)
103. G. Connell, J.A. Dumesic, J. Catal. 102, 216 (1986)
104. B. Notari, Adv. Catal. 41, 253 (1996)
105. M. Itoh, H. Hattori, K. Tanabe, J. Catal. 35, 225 (1974)
106. M. Itoh, H. Hattori, K. Tanabe, J. Catal. 43, 192 (1976)
107. M. Hart, G. Fuller, D.R. Brown, J.A. Dale, S. Plant, J. Mol. Catal. A: Chem. 182–183, 439
(2002)
108. S. Koujout, B.M. Kiernan, D.R. Brown, H.G.M. Edwards, J.A. Dale, S. Plant, Catal. Lett. 85,
33 (2003)
109. S. Koujout, D.R. Brown, Catal. Lett. 98, 195 (2004)
110. P. Carniti, A. Gervasini, S. Biella, A. Auroux, Chem. Mater. 17, 6128 (2005)
111. P. Carniti, A. Gervasini, S. Biella, A. Auroux, Catal. Today 118, 373 (2006)
112. P. Carniti, A. Gervasini, M. Marzo, Catal. Today 152, 42 (2010)
Chapter 9
Characterization of Acid–Base Sites in Zeolites

Dušan Stošić and Aline Auroux

Abstract The review of the use of adsorption microcalorimetry for the charac-
terization of acid-base properties of various types of zeolites is provided. Factors
influencing these properties are introduced and explained. Furthermore, the relation-
ship between the data obtained by this technique and catalytic activity of investigated
materials is discussed.

9.1 Introduction

Zeolites are natural or synthetic microporous materials with an ordered structure:

zeolite frameworks consist of 4-fold connected TO4 tetrahedra (T = Si, Al) that
form three-dimensional networks. In these networks, each oxygen atom is shared
between two neighbouring tetrahedra; in other words, the linkage in between tetra-
hedra is accomplished through T-O-T bridges [1–3]. It should be mentioned that alu-
minate tetrahedra cannot be neighbours in the frameworks of zeolites. In other words
Al-O-Al linkages are forbidden; this requirement being known as the Loewenstein
rule [1, 2]. In zeolitic network, the binding capability in between tetrahedra reaches
its maximum; thus, ideal zeolite crystals should have terminal silanol groups only
on their external surface [4].
The linkage of tetrahedra within zeolites leads to open network structures: the
tetrahedra (primary building units) form rings of various sizes which are linked to
form complex units (secondary building units).These secondary building units may
be connected in many different ways to give a large number of different zeolite struc-
ture types. Thus formed network of interconnected tetrahedra results in the zeolite

D. Stošić (B) · A. Auroux

Institut de Recherches sur la Catalyse et l’Environnement de Lyon,
UMR 5256 CNRS/Université Lyon 1, 2 avenue Albert Einstein, 69626 Villeurbanne
Cedex, France
e-mail: [email protected]; [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 353

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_9,
© Springer-Verlag Berlin Heidelberg 2013
354 D. Stošić and A. Auroux

framework. From the other point of view, it can be told that zeolite crystals contain
pore or channel systems of molecular dimensions with fixed geometry and size [3].
According to the pore size, zeolites are classified into small- (pore size up to
5 Å), medium-(pore size 5–6 Å), and large-pore types (pore size 7–8 Å). Typical
representatives of the different types are zeolite A, ZSM-5, and the faujasites (X and
Y), respectively [5]. According to the International Zeolite Association Website,
about 190 different framework types of zeolites, zeolitic silicates and phosphates
with precisely estimated structures are known to date (2010) [6].
In zeolite network, any tetrahedral T atom (Si4+ or Al3+ ) is surrounded by four
2−
O ions. Network constructed of SiO2 units should be therefore, neutral, because
each O2− ion is shared by two tetrahedra. However, the isomorphous substitution
of Si4+ by Al3+ causes a negative excess charge of the framework. This framework
anionic charge is compensated by charge-balancing cations located in the channels
[7–9]. In contrast to the semipolar character of bonds within the zeolite framework,
the interaction between cations and the framework is ionic. Oxygen atoms with radii
of ca. 1.36 Å surround the smaller central atoms of the tetrahedra nearly completely
[3]. Consequently, the interior surface of zeolites is nearly entirely composed of
oxygen atoms. Nonframework charge-balancing cations and molecules within the
pore system coordinate framework oxygen atoms. Water, if present, can coordinate
these cations and interact with other sorbed molecules [1]. Finally, the overall com-
position of zeolitic unit cell is given by the formula:

1/m Mem+ [(SiO2 )n AlO−

2 ](H2 O)z

where n denotes the Si/Al ratio and z is the number of water molecules adsorbed per
framework Al atom.
Previously mentioned isomorphous substitution of Si4+ ions with Al3+ that pro-
vokes a negative excess charge of the framework is at the origin of their acidity
[1, 2], one of the most important characteristics of these materials which explains their
numerous catalytic applications. Unlike framework T atoms, the charge-balancing
cations can be exchanged by other cations from aqueous solutions [10, 11]. If cations
are exchanged by protons, the zeolite acquires considerable Brönsted acidic proper-
ties and for this reason they can be viewed as solid acids. Importantly, both Lewis and
Brönsted acid sites can be found in the structure of zeolites [12]. The way to generate
Brönsted acid sites (proton donor sites) is based either on the thermal decomposi-
tion of ammonium form of as-synthesized zeolites or on the direct ion exchange
of Na+ (usually) by H+ (with mineral or organic acids). Brönsted acid sites can
be converted into Lewis acid sites by dehydroxylation at elevated temperatures [5].
These processes are presented in Fig. 9.1. However, it is important to notice that
charge-balancing cations can also play a role of Lewis acidity.
Altogether, the structural diversity of zeolites discussed above is responsible for
a wide range of interesting zeolite properties such as ion-exchange capacity, specific
adsorption behaviour, and catalytic activity due to acidity, shape selectivity caused by
size and polarity of molecules, high thermal stability and resistance against solvents,
and wide flexibility for adjustments by post synthesis modification.
9 Characterization of Acid–Base Sites in Zeolites 355

Fig. 9.1 Shematic presentation of generation of Brönsted and Lewis acidity in zeolites. a the
presentation of charge-balancing cation; b cation-exchange Na + to NH+ +
4 ; c from NH4 cation,
isolated proton (proton-donor or Brönsted acid site) arises by calcination; d Lewis acid site (Al
atom with the empty electron orbital, electron acceptor) is formed by dehydroxylation

Because their thermal and mechanical stabilities are not enough to be used for
the industrial purposes, and the zeolite synthesis requires too much cost and time,
so many kinds of zeolite species are not available industrially. In fact, only about
ten kinds of zeolites have been applied in industrial catalytic processes. In practice,
Y-zeolite, ZSM-5, mordenite, MCM-22, and β zeolites are most typical zeolite cata-
lysts [13]. Accordingly to the needs of specific catalytic reaction, they can be mod-
ified appropriately. For example, steaming, dealumination by HCl, cation-exchange
or metal loading, are processes that can be applied to tune zeolites’ features for
their applications as industrial catalysts. In addition, since the remarkable develop-
ments of the mesoporous materials such as SBA-15, MCM-41 and FSM-16, many
kinds of mesoporous materials are also synthesized and studied. Evidently, many
356 D. Stošić and A. Auroux

opportunities to develop a new catalyst based on zeolites and zeolite like materials
are opened.
To summarize, industrial processes in which zeolites are used are mostly reactions
catalysed by acid sites [14–26]. In other words, acidity with a controlled distribution
of acid sites strengths is probably the most important property observed in the zeolite
catalyst. Therefore, the estimation of features related to acidity is evidently crucial
for understanding the reactivity of zeolites. The understanding of zeolites’ acidity
comprises knowledge of related concepts:

1. Nature of the acid sites.

2. Number of the acid sites.
3. Acid strength of the sites.
4. Distribution of the acid strength of the sites.

Experimentally, the physicochemical properties of the solid per se, obtained by

solid-state NMR spectroscopy [27–31], or IR [32, 33] spectroscopy, give insight
into the surface properties related to the acidity. However, more detailed informa-
tion is accessible if probe molecules are brought into contact with the surface sites
and the mode of interaction is studied. Among techniques funded on adsorption of
specific probe molecules, several methods such as: temperature-programmed desorp-
tion (TPD) of amines [34–37]; UV-Vis [38–40], IR [41–43] or XPS spectroscopies
[44, 45] of adsorbed probe molecules; and adsorption microcalorimetry [46–52] are
applicable for the characterization of the solid materials’ acidities. It is important
to point out that none of mentioned methodologies can reveal all previously listed
acidity concepts. In any case it has to be specified whether the concentration, strength
or nature of acid sites is estimated.
Adsorption microcalorimetry allows an accurate determination of amount, strength
and strength distribution of surface sites, based on the heats of adsorption of appro-
priate probe molecules and differential heats as a function of surface coverage. Fol-
lowing text discuss the applicability of adsorption microcalorimetry for the determi-
nation of zeolites’ acidity, and the estimation of different factors that can influence
this important property.

9.2 Factors Influencing the Acid Properties of Zeolites

The strategy for estimation of zeolites’ acidity comprises (includes) the adsorption
and subsequent desorption of chosen probe molecules; both events can be studied
by different physical methods. From the obtained results, the facts about acidic sites
(such as: the strength, strength distribution…) can be derived. As it has been already
discussed (Chap. 3), the characteristics of probe molecule (its nature and size), as
well as the temperature dependence of the heats of adsorption have to be considered
during the estimation of zeolite’ acidity. However, additional parameters that are
related to zeolite of interest might influence its acidity, and have to be taken into
consideration. Here, the influence of:
9 Characterization of Acid–Base Sites in Zeolites 357

(1) zeolite topology;

(2) pre-treatment parameters;
(3) proton (cation) exchange level;
(4) Si/Al ratio and dealumination;
(5) the nature of framework T atom;
on acidity of zeolites and their determinations by the method of adsorption
microcalorimetry will be discussed.

9.2.1 Influence of the Zeolite Topology

Topology of zeolite structures plays an important role in the number of phenomena

related to this important group of solid catalysts, such as selective adsorption and
diffusion. It is well known that in the cases of adsorptions performed on the molec-
ular sieves that possess same chemical composition but different pore systems, the
topology of microporous solid (the structure of the pore system and the diameters of
pores openings) is responsible for the phenomenon of selective adsorption [53–60].
Although other factors have additional impact on selective adsorption (the intensity
of overall electrostatic field; outer surface acidity which can diminish adsorption
and even cause pore blocking by coke formation), topology of zeolite structure is
the most important factor responsible for the fact that in acid-catalazyd reactions,
zeolites often show shape selectivity [61–67].
The topology of the zeolite host is also crucial for the diffusion of the guest
through the interior; it controls the maximum uptake achievable. Zeolites with tridi-
rectional structures are more accessible and allow diffusion readily, whereas diffusion
in monodirectional zeolites is seriously affected [68].
Besides its significant role on the catalytic behavior of zeolites through the so
called ‘shape selectivity effects’, the zeolite geometry plays an important role in
determining the acidity of zeolites. From the overall knowledge concerning zeolitic
structures, it is known that the distances and bond angles in the Al-OH-Si group
can affect the acid strength of the hydroxyl group; besides, the zeolite structure may
also cause preferential location of acid sites [69, 70]. From the molecular orbital
calculations for protonated Si-O-Al bridge it was suggested that the deprotonation
energy decreases with the increase of the T-O-T bond angle, and consequently, the
corresponding acidity increases [71–75]. However, in the attempts to estimate the
influence of zeolite structure on acidity, the difficulties in experimental work arise
from the fact that it is not easy to obtain samples with the same chemical composition
but with different homogenous distribution of Al atoms through the zeolitic particles.
Nevertheless, there are results of investigation of zeolites’ topology influence on the
acidity reported in the literature. Here, the examples of such investigations done by
adsorption microcalorimetry are summarized.
Influence of structure on the number and strength of available Brönsted acid sites
was investigated on a series of dealuminated zeolites (HY, mordenite, ZSM-5 and
358 D. Stošić and A. Auroux

Fig. 9.2 Differential heat of

ammonia adsorption versus
the adsorbed amount for
various zeolites [46]

mazzite) with similar framework composition (Si/ Al ratios ≈ 15). It is known that in
the case of dealuminated zeolites, microporosity does not influence importantly their
acidity; while the bond angles play an important role. Differential heat curves plotted
as a function of coverage, represented in Fig. 9.2, shows that the number of sites is
higher on HY, but the strength of the sites is greater for the mazzite sample [46]. It
seems therefore that these systems contain domains of more extensive dealumination
than the other samples, including less accessible zones, thus leading to strength
distributions of acid sites dependent on the samples.
Contrary to this behavior, various structures with high Si/Al ratio (H-ZSM-5,
H-ZSM-12, HY and H-mordenite) were found to exhibit almost the same adsorption
energies measured at 480 K, suggesting that the isolated Brönsted acid sites in these
materials are identical [76]. Authors observed that the strength of Brönsted acid sites
associated with framework Al atoms does not vary in a given zeolite, and they suggest
that Brönsted acid sites in unsteamed zeolites are independent of the sample (what
means, independent of Si/ Al ratio) and equal in concentration to the framework
aluminum content [77].
A study performed by adsorption microcalorimetry of NH3 and SO2 on a series
of milled Na zeolites [78, 79] with different degrees of crystal collapse, revealed
that losses in zeolite crystallinity cause a decrease in the number and strength of
both Lewis acid sites and Lewis basic sites. It was observed that Lewis acid strength
is strongly cristallinity dependent while Lewis basic strength is less so. Combined
XPS and calorimetric results have shown that the extraframework Lewis acid sites
(Na cations) are remarkably influenced by the long range stabilization effect of the
zeolite lattice which gives a higher strength of Lewis acidity.
A similar study performed on crystals of HA, HX and HY zeolites [80] revealed
that milling caused the destruction of the long-range symmetry of the crystal
while the primary building units, namely Si(Al)-O4 tetrahedra, remained intact.
Microcalorimetry showed that the collapse in zeolite structure also caused decrease
in the number and strength of acid sites, and that the Brönsted acid strength is strongly
crystallinity-dependent. The destruction of long range crystal symmetry induces an
9 Characterization of Acid–Base Sites in Zeolites 359

Fig. 9.3 Differential heat NH3 of adsorption on initial and ball-milled zeolite samples (each milled
sample is designated by milling time noted behind the sample name) [80]

increase of the population of weak Brönsted acid sites and a decrease of the popula-
tion of strong Brönsted acid sites (Fig. 9.3).
It can be concluded that the heat curves should be similar only when the
solids contain exlusively isolated sites where the next-nearest neighbors (NNN) are
identical. It appears that the NNN determines Brönsted acidity, in fact, it can be said
that zeolite geometry has an influence on acidity due to long-range effects as well as
on the local structure of the acid sites.

9.2.2 Influence of the Si/Al Ratio

The Si/Al ratio has an important influence on the properties of zeolites. The Al content
determines the number of cations in the framework and the properties such as the
thermal and chemical stability or the polarity of internal species. Typically, zeolites
with high Al content are thermally and chemically less stable, so that dehydratation at
high temperature may cause partial dealumination that happens together with water
desorption, what results in a certain decrease of cristallinity[81]. The hydrophilicity
/ hydrophobicity of zeolites are related to the polarity of the pores. Zeolites without
framework Al are the most hydrophobic [82, 83].
360 D. Stošić and A. Auroux

The Si/ Al ratio plays also a significant role on acidic properties of zeolites.
Aluminum atom is directly related to the acidic site, since a Brönsted site can result
when the cation which balances the negative charge associated with each framework
Al is a “proton”. Because the “protons” are bonded to the bridging oxygen with
considerable covalent character, the site should be viewed as a hydroxyl group, but
with properties significantly different from that associated with pure silica [84].
Additionally, Al also accounts for the formation of carbenium and/ or carbonium
ions or possibly cation radicals inside the zeolite.
An important aspect of zeolites and other acidic molecular sieves is that each
of these materials contains a well defined, discrete number of acid sites. For high
silica zeolites in which all bridging hydroxyls are accessible to adsorbates, such as
H-ZSM-5, the site concentration is approximately equal to framework Al content
[85, 86]. In some other materials, such as HY, many of the sites are inaccessible to
adsorbate molecule, so the measured site concentration is significantly lower than
the Al content [87].
Tsutsumi et al. have studied heats of ammonia adsorption on mordenites with
different Si / Al ratio [88]. The number of acid sites determined from the heat curves
plotted against the theoretical number of protons (aluminum atoms) per unit cell
of mordenites, shows a slope that deviates little from unity (Fig. 9.4); the deviation
being smaller for higher Si / Al ratios. The authors suggested that the framework alu-
minium atoms participate in the formation of acid sites and that some of the generated
Brönsted acid sites are converted to Lewis acid sites. Displacement of aluminium
atoms from the zeolite framework may also be responsible for the observed deviation.
To investigate how Si/Al ratio influences the acidity for a given zeolite structure,
several H-ZSM-5 samples have been prepared, with Si/ Al values varying between
50 and 14 [89]. Auroux et. al. found that the Al atoms progressively incorporated
in larger proportion not only create new acid sites, but also modify the strength of
pre-existing acid sites as well. The initial heat of ammonia adsorption goes through a
maximum. When the Al content increases, the initial heat of adsorption, and thus the

Fig. 9.4 Relationship

between the number of acid
sites calculated from calori-
metric heat curves and the
calculated number of protons
on mordenite samples with
different Si/Al ratio [88]
9 Characterization of Acid–Base Sites in Zeolites 361

strength of the strongest acid sites increases, then remains constant (for Si/ Al between
35 and 18), and finally strongly decreases for low Si/ Al ratios. However, the integral
heats of irreversible adsorption, representative of the overall acidity, increase sharply
when the Al content increases from 0 to 2 atoms per unit cell, and then moderately for
higher Al loadings. These results show that when relating the acidity of material to
its catalytic properties, it is important to discriminate the strength of individual acid
sites and the total acidity. It has been discovered that appropriate Si/Al ratio has to be
discovered for a given reaction; so the importance of a careful selection of adequate
zeolite sample is emphasized. For example, to prevent an excessive polymerization,
the presence of very strong acid sites must be avoided, while the total acidity must
be sufficient; for that purpose, ZSM-5 zeolites with a Si/Al ratio between 10 and 40
have been found to be best suited for this purpose.
The chemical composition of synthetic zeolites can be controlled either dur-
ing synthesis (by varying the composition of mother gels) or by postsynthetic
modifications. Dealumination process can promote modifications of porous struc-
ture, which may improve some important properties of zeolites, like: thermal and
hydrothermal stability, acidity, catalytic activity, resistance to aging and low coking
rate, and material transfer. Different dealumination processes have been proposed:
steaming and acid treatments, as well as reactions with SiCl4 or SiF2− 6 . From many
theoretical and experimental investigations on the acidity of faujasites with differ-
ent framework aluminium content, it was concluded that the number of bridging
hydroxyl groups increases with the number of Al atoms in the lattice [90]. In contrast,
the dependence of the acid strength on aluminium content is more complicated. From
theoretical considerations, the acid strength is expected to increase with decreasing
number of aluminium atoms, whereas from the studies of Al topology in the frame-
work, a curve often presenting a maximum was derived from the number of strong
sites [90].
Figure 9.5 shows that on aluminium-deficient HY zeolites, dealumination gen-
erally causes a decrease in the acid sites concentration followed by an increase in
strength [91].
Mitani at al., in their work related to the investigation on the acidity of faujasites
with different framework aluminium content, reported that the extent of the decrease
in acid sites concentration varies with the kind of basic probe used in microcalorime-
try experiments (see Fig. 9.6) [92]. The experimental results were explained by dif-
ferences in molecular diameters of ammonia and pyridine, and the fact that pore
size of zeolite becomes widely distributed as the dealumination proceeds; so OH
groups in supercages have become available for adsorption of such large molecule
as pyridine is.
In another study performed on steamed deluminated faujasites presenting both
framework and non framework Al, authors found no evidence for the presence of
a small concentration of very strong acid sites [93]. For low coverage, differential
heats of adsorption did not show any dependence on the Si/Al ratio of the samples or
on the sample preparation. Since the samples with different degrees of dealumination
showed considerable different catalytic activities, it was suggested that factors other
than acid strength are responsible for this behaviour. It was concluded that when
362 D. Stošić and A. Auroux

Fig. 9.5 Differential heats of ammonia adsorption over HY zeolites with various Al content as a
function of coverage [91]

heterogeneity in the strength of sites is revealed by microcalorimetry (i.e. a fall of

differential heat with coverage), this is the result of molecular interactions of mole-
cules adsorbed at neighbouring sites, rather than a true indication of the differences
between the sites. Strong repulsive interactions between adsorbate molecules arising
from ionic repulsion of two positively charged adsorbates, from steric interactions
between neighbouring adsorbates, or from changes in a protonic site due to adsorp-
tion on adjacent sites, could lead to a large decrease in measured adsorption heats
with coverage. The slope of this decrease depends on the adsorbate.
Generally, it can be seen that, when differential heats of adsorption of ammonia or
pyridine are plotted as a function of surface coverage for the samples with different
aluminium content, at least three domains of acidity are revealed, whose relative
importance depends on the aluminum content, and whether aluminium is exclusively
located in framework positions (Fig. 9.6). First domain, with a sharp initial decrease
in qdi f f , is generally associated with non-framework Al and a very strong Lewis
sites (heats of pyridine adsorption above ∼150 kJ/mol). A region of intermediate
9 Characterization of Acid–Base Sites in Zeolites 363

Fig. 9.6 Differential heats of adsorption of ammonia (left) and pyridine (right) at 473 K [92]

heats, represented by a plateau, follows. This plateau corresponds to strong Brönsted

acid sites associated with framework Al around 140–150 kJ/mol depending on the
zeolite. At the end there is again a decrease of heat, corresponding to weak Lewis
acid sites associated with alkali cations and weak Brönsted sites associated with
non-framework Al or neighboring species.
The heats of adsorption of ammonia and the catalytic activity in isooctane cracking
are reported by Mishin et al. [94] for high-silica Y zeolites with Si/Al ratios varying
from 1.25 to 100. The aluminium-deficient Y zeolites have been found to possess
stronger acid sites than the parent zeolites. It has been revealed using adsorption
microcalorimetry measurements that even a minor increase in the Si/Al framework
ratio results in an increase of the bond strength between NH+ 4 ions and the lattice.
In agreement with the changes in acidity (number of sites with qdiff ≈ 130 kJ/mol),
progressive dealumination of Y zeolites results firstly in an increase followed by a
decrease in catalytic activity, with a maximum for 25–30 Al atoms per unit cell. The
increase in catalytic activity is associated with the appearance of strong acid sites
(qNH3 > 120 kJ/mol). The effect of acidity on activity of dealuminated zeolites is
explained in terms of isolated acid-site generation. It is postulated that nonframework
Al atoms do not contribute to catalytic activity.
The effect of steaming on the number and strength of the acid sites is obvious
from a comparison of differential heat curves for samples with different degrees of
dealumination. The microcalorimetric curves show also that the strength of the sites
corresponding to the intermediate plateau region first increases and then progressively
decreases with steaming severity. The dependence of the acid strength distribution of
dealuminated mordenites and dealuminated faujasites on the Si/Al ratio is depicted in
Fig. 9.7. The number of strong acid sites presents a distinct maximum. The abscissa
of the maximum corresponds to an Al content of 4.6 per unit cell (u.c.) (and Si/ Al
ratio equal to 9.5) for dealuminated mordenites and an Al content of about 29 / u.c.
(and Si / Al ratio around 5.5) for dealuminated Y zeolites. These values are in good
agreement with the values predicted theoretically [95], where the limit values, mlim ,
364 D. Stošić and A. Auroux

Fig. 9.7 a Acid strength distribution dependence on Al(IV) content per unit cell of dealuminated
H-mordenites. Na number of acid sites in molecules g−1 with Q >80 kJ mol−1 (◦), with Q >100 kJ
mol−1 (), with Q > 120 kJ mol−1 (, *) from [96]. Differential molar heats of ammonia adsorption
were measured at 423 K. b Acid strength distribution dependence on the framework aluminum
content per unit cell in dealuminated Y zeolites. Curve 1 total number of acid sites with Q >80 kJ
mol−1 (*,◦ ,♦). Curve 2: number of acid sites with Q > 100 kJ mol−1 (,). Curve 3 number of
acid sites with Q >120 kJ mol−1 (+,×,) from [90]. Differential molar heats of ammonia adsorption
were measured at 423 K

for mordenite and H-Y zeolites were calculated. It was found that they correspond
to mlim = 0.096 (or Si/ Al = 9.4) for mordenite and mlim = 0.150 (Si/ Al = 5.8) for
H-Y zeolites. According to Barthomeuf [95], below this maximum no Al atom has
another Al atom as a next-nearest neighbour, and therefore all acid sites show a high
acid strength. Above these maxima, there are Al atoms in the second coordination
sphere, with the consequence that some Brönsted sites show weak acidity (although
the total number of acid sites is still increasing), and their concentration decreases
with increasing of framework Al content. Measurements from Stach et al. [90, 96]
and Macedo et al. [97] were found to confirm this model and to be in very good
9 Characterization of Acid–Base Sites in Zeolites 365

agreement, though the dealumination and the measurements were not performed
under the same conditions.

9.2.3 Influence of the Pre-treatment Parameters

It is known that increasing of the pre-treatment temperature modifies the surface

acidity of the solids. High-temperature calcination is a well-known method of reduc-
ing total acidity of zeolites via dehydroxylation and dealumination. Although for the
estimation of the real nature of an active site other techniques (spectroscopic inves-
tigation of adsorption/desorption of probe molecules) are needed, it can be told that,
in principle, by using adsorption calorimetry technique it should be possible to deter-
mine the number of “strong” Brönsted sites, assuming that activation at 673K before
NH3 adsorption gives rise to a maximum in H+ , i.e. that no dehydroxylation occurs
at that temperature. This is a crude approximation but can be considered as valid for
comparison of different samples [48]. After calcination at increasing temperatures,
dehydroxylation of the zeolite is observed: above 675K, the number of Brönsted
acid sites decreases, while that of strong Lewis acid sites increases. However, a
limited dealumination occurs and the constraining character of the intracrystalline
voids increases. Microcalorimetric studies of ammonia adsorption confirm the very
strong acidic character of the acid sites and show their dependence in strength and
heterogeneity upon calcination temperature [89, 98, 99].
In a ZSM-5 zeolite (Si/Al = 19), a thermal treatment at high temperatures leads
to condensation of two OH groups (representative of two protons) to form a Lewis
site. Changes in pre-treatment temperature modify the differential heat curve and
thus allow identification of the nature (Lewis or Brönsted) of acid sites. Figure 9.8
presents the acidity spectra (the values of dna /dq ratios) obtained for the adsorption

Fig. 9.8 Acidity spectra

(−dna /dq versus qdi f f ) of a
ZSM-5 sample for various
pretreatment temperatures
[46]
366 D. Stošić and A. Auroux

of ammonia at 416 K on these ZSM-5 zeolites after the samples were pre-treated
at various temperatures. The initial heat of adsorption increases as a function of
outgassing temperature. From the results presented by Fig. 9.8, it is evident that the
main band of the spectrum is progressively shifted towards higher adsorption heats,
while its area (and thus the total number of sites) simultaneously decreases, when
the sample is dehydrated at increasing temperatures (743, 923 or 1073 K). It follows
that a high activation temperature (1073 K) decreases the total number of acid sites,
while the acid strength of some of them is increased. The hypothesis of formation of
Lewis sites stronger than Brönsted ones is thus confirmed [46].
However, the formation of Lewis acid sites cannot be revealed only from this
shifting of the band in the acidity spectrum, since as already mentioned, calorimetry
alone is not able to provide a simple procedure for identification of the nature of
sites, and needs to be associated with another technique like infrared spectroscopy.
It is important to notice that the dehydroxylation at high temperatures produces an
irreversible transformation; a sample outgased at 1073 K, rehydrated at room temper-
ature and outgased again at 673 K, does not exhibit the initial acidity spectrum [46].
A similar microcalorimetric study has been performed by Karge et al. [100, 101]
on a well-crystallized H-ZSM-5 activated at 673 K which contains very few Lewis
sites, i.e. with aluminium located almost exclusively in framework positions. The
microcalorimetric curve of ammonia adsorption at 423 K showed a nearly horizontal
plateau around 150 kJ/mol of homogeneous acidic strength due to the interaction
of ammonia with Brönsted acid sites. After high temperature dehydroxylation in
vacuum at 1073 K, true Lewis sites were created, inducing a large increase in the
number of sites evolving differential heat over 150 kJ/mol. These were considered
to be heats which have a contribution both from the Lewis sites induced by the high
activation temperature and from the remaining Brönsted sites. The overall number
of titrated strong and medium acidic sites had simultaneously decreased from 2.7 to
1.6 NH3 molecules per u.c. while the width and the homogeneity of the plateau were
considerably reduced. After dehydroxylation at high temperature, no Brönsted sites
of homogeneous energy distribution remained on dealuminated ZSM-5 samples.
Only the acidic sites of broad energy distribution remained unaffected by the high
temperature treatment, i.e. the sites of strength between 140 and 80 kJ/mol.
Another example of pre-treatment influence on acidity is the study concerning
high temperature calcination of H-mordenite (Si/Al = 13) in air at 1008 K, that was
done by Chen et al. [102] using adsorption microcalorimetry of pyridine at 473 K.
This treatment caused a significant reduction in the total number and in the strength
of the acid sites. The plateau characteristic of a large number of sites of uniform
strength (near 200 kJ mol−1 ), observed on the sample pre-treated at 673 K, also
disappeared. High-temperature calcination is known to induce dealumination and
dehydroxylation, both of which are expected to reduce the number of acid sites [102].
Dehydroxylation of an H-mordenite sample at 923 K caused the appearance of centres
with heats of NH3 adsorption between 170 and 175 kJ mol−1 which were not present
in the same sample pre-treated at 703 K. Increasing the dehydroxylation temperature
to 1023 K provoked the increasing of the concentration of centers characterized by
a heat of ammonia adsorption of 175–170 kJ mol−1 to 0.2 mmol g−1 , and sharply
9 Characterization of Acid–Base Sites in Zeolites 367

Fig. 9.9 Calorimetrically determined molar heats of adsorption of ammonia at 473 K on HM-20
evacuated at various temperatures. Filled symbols represent heats of re-adsorption on samples which
were evacuated at 473 K after the first run of the adsorption measurement [88]

decreased the concentration of centres generating heats of 160–130 kJ mol−1 [103].

It was shown that the latter values are typical of the dissociation of ammonia over
Lewis acid centers.
The acidic properties of a mordenite zeolite with a Si/Al ratio of 10 were analyzed
by Tsutsumi et al. [88] through calorimetric measurements of the heats of ammonia
adsorption done at 473 K, on samples previously evacuated at temperatures varying
from 703 to 1073 K, as can be seen from Fig. 9.9. Drastic energy changes were
observed between low and high coverage, with the exception of the sample evacuated
at 1073 K. Both the number of the more energetic sites and their energy increased
with an increase in the evacuation temperature, and reached a maximum at 773 K
(around 170–175 kJ mol−1 ); then a decline was observed. The heat curve of the
sample evacuated at 1073 K did not exhibit a plateau, and its initial value was much
less than those of the other samples (the value of 160 kJ mol−1 was obtained, instead
of 175 kJ mol−1 ). This may be attributed to the breakdown of part of the mordenite
structure and/or to the local formation of amorphous species [88].
A sample of mordenite (98 % degree of ammonium-ion exchange) deammoniated
at various temperatures from 693 to 923 K was studied at 573 K by NH3 adsorption
microcalorimetry by Bankos et al. [104]. On increasing the pre-treatment temper-
ature, the number of acid sites passed through a maximum at 753 K as a result of
simultaneous decationation and dehydroxylation. The heat of adsorption of NH3 on
Brönsted acid sites formed by decationation was 110–160 kJ mol−1 . During dehy-
droxylation, two types of Lewis sites were formed, characterized by heats of NH3
adsorption of 170–185 kJ mol−1 and 95–100 kJ mol−1 respectively, on which disso-
ciative chemisorption of ammonia was evidenced by IR [103].
368 D. Stošić and A. Auroux

9.2.4 The Effect of Proton (Cation) Exchange Level

Ion exchange is a characteristic property manifested by most molecular sieves. This

property is used routinely for post synthesis modification during the preparation of
molecular sieves for major industrial applications [10, 11].
The origin of ion exchange properties resides in the fact that the net negative
charge of zeolite network, equal to the number of the constituent aluminium atoms,
is balanced by exchangeable cations, Mn+ , located in channels which normally also
contain water. According to the synthetic procedures, these cations which neutral-
ize the electrical charge of the aluminosilicate framework, are usually Na+ , K+ or
template cations. This charge-compensating entities can further be exchanged for
other cations. Conventionally, ion exchange is carried out by suspending the zeolite
powder in aqueous solutions of salts which contain the desired in-going cation, or
in mineral acids to introduce protons into acid resistant zeolites [10, 11]. The same
can be done by solid state ion exchange, where dry powders of zeolites and salts or
oxides containing the exchangeable cations are reacting [105].
Evidently, ion exchange capacities can be considered as functions of the quantity
and distribution of aluminium atoms within the structure of zeolites. In addition, it is
worth to notice that zeolites often possess high ion exchange selectivity for certain
cations, and this can be used for their concentration and isolation. Molecular sieving
and acid properties of zeolites are modified in this way, what gives possibility to
tailor their catalytic activity. The framework anions and exchangeable cations form
electrostatic fields which strongly interact with electronic structure of adsorbate
molecules [106]. The reactivity of adsorbed species depends strongly on the number
and kind of cations of the zeolite under investigation.
The exchange degree plays an important role in the heterogeneity of the acid
sites of zeolites. From the study performed on NaY parent sample and NH4 NaY
zeolites exchanged to 29, 56, 80, 90, and 94 % respectively, it was observed that
acidity appears slowly at the early stages of replacement, and that up to about 60 %
exchange, the number of acid sites present in the zeolites is four times lower than
the amount of sodium replaced [107, 108]. At higher extents of decationation, the
density of acid sites increases rapidly and approaches a theoretical value [108].
A question is how the framework charge balance can be achieved at low exchange
degree, if the number of cations that are removed is greater than that of the acid sites
formed. Two alternative interpretations were offered: either the hydroxyls generated
at low exchange levels are unstable and can be eliminated during pretreatment, or
an incomplete decomposition of NH+ 4 ions occurs upon thermal activation, the latter
hypothesis being supported by TPD spectra. To rationalize the observed acidity-
exchange level profile a model was developed, extending the shielding power of a
single Na+ ion to several neighboring AlO− 4 tetrahedra. The catalytic activity of
these decationated faujasites in the cracking of isooctane and the disproportionation
of ethylbenzene was also found to vary in the same manner as the number of acid sites,
increasing rapidly beyond 70 % exchange degree. However, the observed discrepancy
between the two increases (the catalytic activity increased much faster) prevented
9 Characterization of Acid–Base Sites in Zeolites 369

Fig. 9.10 Acid site strength distribution of Na,H-ZSM-5 zeolites as a function of the exchange
level. Ammonia adsorption at 393 K, pretreatment at 673 K (111)

direct interpretation of the catalytic activity in terms of acid sites population. It is

reasonable to assume that both the number and the strength of the acid sites are
affected by the exchange level.
Similar study was performed on decationated mordenites [107], which have nar-
rower pores and a higher Si/Al ratio than Y zeolites. From microcalorimetric mea-
surements it was found that the number of acid sites increases nearly linearly with
an increase of decationation level. Growth of the exchange level increased the acid
strength, and removal of the last NH+ 4 ions resulted in the formation of strong sites
(qdiff > 130 kJ mol−1 ). This increased concentration of strong acid sites explained
that mordenites are more active in the cracking of n-octane than HY zeolites.
However with bulkier reactants such as isooctane and ethylbenzene, the difference
in catalytic activity decreases due to shape selectivity effects.
In a study by Muscas et al., a variety of NaHZSM-5 zeolites with varying extent
of Na exchange was subjected to adsorption of various probe molecules in order
to determine the selectivity of adsorption of these molecules on their acid centres
of variable nature, such as Na+ in Na forms and acidic OHs in the corresponding
H forms [109]. Using ammonia as a basic probe it has been shown that the num-
ber of acid sites also increased quasi-linearly with the exchange level, while the
acid strength increased monotonously and moderately up to about 40 % of exchange
level, remained almost constant at intermediate exchange level, and then dramati-
cally increased when 80 % of the original Na+ ions were removed. The removal of
residual Na+ ions not only generated much stronger sites but resulted also in a gen-
eral increase of the acid site population already present. Figure 9.10 clearly shows
that at low exchange levels, most of the acid sites were rather weak sites. While this
population of weak sites remained almost constant as the exchange level increased,
the population of stronger sites increased progressively up to the point where, for
370 D. Stošić and A. Auroux

extensively exchanged samples, the strongest sites became predominant. The popu-
lation of sites, the heat of adsorption which was above 150 kJ mol−1 , illustrated the
effect of removal of the very last sodium ions on the acid strength, not only on that
of newly created sites but also on that of pre-existing ones. This phenomenon, fairly
common to those encountered with faujasites [96, 108, 109] and mordenites [107],
was interpreted in terms of remote perturbation of the structure of the acid centers
by the very few residual Na+ ions modifying significantly the T-O-T bond angles.
In addition, possible general cation redistribution induced by NH3 adsorption may
also affect the evolved heats of adsorption.
Figure 9.11 presents the number of sites evolving heats above 110 kJ mol−1 upon
NH3 adsorption, as a function of the exchange level; for ZSM-5 [109], faujasite [107]
and mordenite [108] samples. This number of strong sites increases in all cases, but
in different proportions; almost linearly for mordenite and ZSM-5, and first slowly
(for low exchange levels) and then sharply (at high exchange level) for faujasite.
Concomitantly, the catalytic efficiency increases as does the acid strength. But it is
not clear yet how the acid strength increases with increased exchange level.
The nature of the exchanged cation is one of the key points that determine acidity
in zeolites. The acid-base properties of alkali-exchanged X and Y zeolites have been
studied using microcalorimetry of ammonia, pyrrole or SO2 probes. Generally, the
alkali-exchanged X zeolites are more basic than the corresponding Y zeolites. It was
shown that the basic strength increases quasi-linearly with the negative charge of
oxygen atom, calculated from the Sanderson electronegativity equivalence method.
Pyrrole chemisorbed on the basic site creates a bonding between the framework oxy-
gen (Lewis basic site) and the H atom of the NH group of the pyrrole molecule, so the
stronger base (Cs) will produce a larger heat of adsorption. Briefly, a thermodynamic
scale of the Lewis basic strength distribution of alkali-exchanged X and Y zeolites
was obtained, supporting the conclusion that the Lewis basicity in alkali-exchanged
zeolites is a local property, strongly influenced by the adjacent alkali cation [110].
The acidic properties of a series of X faujasites exchanged with Li, Na, K, Rb
and Cs have been studied by adsorption microcalorimetry, using ammonia as acidic
probe. The heats of NH3 adsorption were found to decrease in the sequence from Li
to Cs (Fig. 9.12).
Li and Na zeolites presented much higher heats of NH3 adsorption and greater
coverage at the same pressure in comparison with the other zeolites. The acid-base
properties of alkali-metal ion exchanged X and Y zeolites have also been investigated
by ammonia and sulphur dioxide adsorption microcalorimetry, in parallel with the
study of a catalytic reaction, viz. 4-methylpentan-2-ol conversion [111].

9.2.5 The Influence of the Framework T Atom

Isomorphous substitution is an important way to modify zeolite properties for prac-

tical applications and has achieved considerable interest in the field of zeolite chem-
istry. The thermodynamics of this process have been considered by Barrer [112]. The
9 Characterization of Acid–Base Sites in Zeolites 371

Fig. 9.11 Number of acid sites evolving heats above 110 kJ mol−1 upon NH3 adsorption as a
function of exchange level for HNa-ZSM-5 (), HNaY (+) and HNaM (*) zeolites

Fig. 9.12 Differential heats of NH3 adsorption on alkaline X zeolites versus coverage [90]

phenomenon of isomorphous substitution is well-known in the field of mineralogy

[1, 113]. By isomorphous substitution, framework atoms of crystalline compounds
are replaced by atoms of other elements without changing the type of the crystal
structure. Most elements can undergo substitution, at least to a very low degree. Even
chromium, which prefers octahedral coordination, may substitute silicon in tetrahe-
dral framework positions, to a certain extent [114, 115]. Extremely low degrees of
substitution (0.2 atom %) can give rise to a remarkable change of properties.
The main factors governing isomorphous substitution are summarized in the fol-
lowing way [116]. (a) The tendency of substitution depends on the ratio of radii
372 D. Stošić and A. Auroux

of the atoms involved. With decreasing the difference of the radii r of atoms A
and B, the substitution becomes energetically more favoured. (b) If the replacement
leads to a decrease of the coordination number of atom A, larger atom A will replace
more easily smaller atom B. (c) Electronegativity, ratio and the ionization poten-
tials of exchanging atoms have an influence on substitution. (d) During isomorphous
replacement of A by B with preservation of local structure (coordination number),
a minimum free energy is being achieved with r/r = 0.025–0.03, r denoting the
radius of the atoms to be replaced in the framework by atoms of another element.
(e) Isomorphous substitution is facilitated as long as replacement of A by B does
not change long range electrostatic interactions. (f) Substitution may occur when the
charges of exchanged atoms differ by 1, 2 or 3 units. (g) Exchanging atoms should
not react with each other.
The main criteria for the occurrence of isomorphous substitution, which are pri-
marily derived from crystal chemistry and geometrical considerations, have been
formulated by Pauling [117]. The basic idea is that framework of crystals, i.e., sil-
icates including zeolites, consist of packages of negatively charged oxygen anions
(O2− ). Therefore, tetrahedral and octahedral vacancies are formed. The size of
these vacancies depends on the size of anions. According to Pauling, cations pre-
fer tetrahedral coordination if r Me /r O2− = 0.214 − 0.4, and octahedral sites if
r Me /r O2− = 0.4 − 0.6.
It should be said that all rules cited above are not strict laws. The ability of
the framework to change its fine structure (bond angles and distances) by rotation
of tetrahedra, by tilting, or by inversion is important in order to relax the strain
resulting from the substituting atom [118]. The aim of substitution is to integrate
the new element in the structure while preserving the structure type. Often a certain
percentage of the modifier remains in the extra-framework positions.
The motivation to replace aluminium in aluminosilicate zeolite structures by other
elements arose from the need to adjust their properties to intended applications. Since
the nature and strength of the bridging hydroxyl groups (Si – OH – T, T = Al, Fe,
Ga, B, ...) depend on T atom, and thus the proton – T distance and resulting acid
strength of the modified material. At comparable bond angles, the proton – T distance
decreases in order Fe, Ga, Al. This means that electrostatic repulsion between the
proton and T increases in the same way, and that the acidity is expected to increase in
the same order [48]. These changing in the acidity are easily accessible by adsorption
calorimetry technique.
The use of organic templates has rendered possible the substitution of many ele-
ments, including other trivalent (Cr 3+ ), bivalent (Be2+ ), and tetravalent ions (Ge4+ ,
Ti4+ ). Most of these metallosilicate compositions have been synthesized with ZSM-5
crystal structure. Adsorption microcalorimetry enabled the studying of acidic prop-
erties of thus obtained materials, as in the case of parent zeolites.
A co-incorporation of aluminium and boron in the zeolite lattice has revealed
weak acidity for boron-associated sites [119] in boron-substituted ZSM-5 and ZSM-
11 zeolites. Ammonia adsorption microcalorimetry gave initial heats of adsorption
of about 65 kJ mol−1 for H-B-ZSM-11 and showed that B-substituted pentasils have
only very weak acidity [120]. Calcination at 1073 K increased the heat of NH3 adsorp-
9 Characterization of Acid–Base Sites in Zeolites 373

tion to about 170 kJ mol−1 by creating the strong Lewis acid sites. The lack of strong
Brönsted acid sites in H-BZSM-11 was additionally confirmed by poor catalytic
activity in methanol conversion and in toluene alkylation with methanol.
Gallium has been successfully introduced into numerous zeolite frameworks
(β, MFI, offretite, faujasite, ...). The Ga3+ ions in zeolites can occupy both tetra-
hedral framework sites (T) and non-framework cationic positions. The isomorphous
substitution of gallium into aluminosilicate zeolites results in modified acidity and
subsequently modified catalytic activity such as enhanced selectivity towards aro-
matic hydrocarbons.
Microcalorimetric experiments with ammonia and pyridine as probe molecules
have been used to investigate the effects of framework Ga on the acidic properties
of several zeolites [121–127]. Experiments of NH3 adsorption microcalorimetry,
together with FTIR results from pyridine thermodesorption, have shown that the iso-
morphous substitution of Al by Ga in various zeolite frameworks (offretite, faujasite,
β) leads to reduced acid site strength, density, and distribution [122–125]. To a lower
extent, a similar behaviour has also been observed in the case of a MFI framework
[126, 127]. A drastic reduction in the acid site density of H,Ga-offretites has been
reported, while the initial acid site strength remained high [122, 124].
For Ga-β zeolite it was found that, when the Si/Ga ratio increased from 10 to 40,
the number of strong sites decreased drastically for Si/Ga between 10 and 25 and
then reached a plateau above Si/Ga = 25 [121]. The strength and density of acid sites
in H-(Ga,La)-Y have also been found to be lower than those in H-Y crystals of the
type used in FCC preparation (LZY-82) [123].
Parrillo et al. [128] have used microcalorimetric measurements of ammonia and
pyridine adsorption to compare the acid sites in H-[Fe]ZSM-5, H-[Ga]ZSM-5, and
H-[Al]ZSM-5. On each of the molecular sieves, the differential heats of adsorption
for both ammonia and pyridine were constant up to coverage of one molecule per
Brönsted site. The differential heats at a coverage below 1 : 1 were identical on
each of the materials, with values for ammonia of 145 ± 5, 150 ± 5, and 145 ±
5 kJ mol−1 on H-[Fe]ZSM-5, H-[Ga]ZSM-5, and H-[Al]ZSM-5, respectively; and
for pyridine of 195 ± 5, 200 ± 5, and 200 ± 5 kJ mol−1 on H-[Fe]ZSM-5, H-
[Ga]ZSM-5, and H-[Al]ZSM-5, respectively. The authors [128] concluded that the
microcalorimetric heats of adsorption for ammonia and pyridine at Brönsted acid sites
formed by framework Fe(III) and Ga(III) were very similar to heats of adsorption at
Al(OH)Si sites, and that the three samples were effectively equivalent proton donors.
By contrast, they found very different reactivity measurements for n-hexane cracking
and propene oligomerization on the same materials. The authors claimed that heats
of adsorption for strong bases do not reflect differences in inherent acid strength and
may not be related to catalytic activity in a simple manner.
Iron silicates of MFI structure with various Si/Fe ratios have also been studied
by NH3 adsorption microcalorimetry [127, 129], and compared with the Al and
Ga analogues. The intermediate strength sites (predominantly Brönsted sites) were
found to correspond to a plateau around 145 kJ mol−1 for H-[Al]-ZSM-5 (Si/Al
= 19), 140 kJ mol−1 for H-[Ga]-ZSM-5 (Si/Ga = 22), and 135 (Si/Fe = 41), 125
(Si/Fe = 26), or 120 (Si/Fe = 12) kJ mol−1 for H-[Fe]-ZSM-5, respectively.
374 D. Stošić and A. Auroux

Fig. 9.13 Differential molar heats of ammonia chemisorption at 423 K on MFI zeolites as a function
of the adsorbed amount:  Al-Sil,  Fe-Sil, ♦ In-Sil,  silicalite. Pretreatment temperature 673 K
([132])

Besides, the adsorption of acetonitrile as a probe molecule has been studied using
microcalorimetry at 400 K, in the case of H-[Fe] ZSM-5 and H-[Al] ZSM-5 [130].
The heats of formation of the complexes were found to differ slightly, ca. 95 kJ mol−1
on H-[Fe] ZSM-5 compared to ca. 110 kJ mol−1 on H-[Al] ZSM-5; suggesting that
the hydrogen bonds at FeOHSi sites may be slightly weaker.
Ga- and Fe-substituted MFI zeolites have been investigated using adsorption
microcalorimetry of different alkanes at 353 K by Auroux et al. [131]. The acid
strength of the zeolite protons decreased following the sequence H-[Al]MFI > H-
[Ga]MFI >H-[Fe]MFI. The heats of adsorption decreased with the basicity of the
alkane in the order n-butane > isobutane > propane.
The active sites of isomorphously substituted MFI structures activated at 673 K
have been characterized by Jänchen et al. [132] using microcalorimetric measure-
ments carried out at 423 K with ammonia as a probe. In accordance with decreasing
heats of NH3 adsorption, the Brönsted acid site strength of the modified MFI was
reported to decrease in the sequence Al > Fe > In > silicalite. In addition to these
strong sites, weaker Lewis centres due to the non-framework material were found
(Fig. 9.13).
Crystalline MFI-type indosilicates containing indium ions in framework posi-
tions were hydrothermally synthesized by Vorbeck et al. [133] and characterized
by adsorption microcalorimetry of NH3 at 423 K (after activation of the sample at
670 K). They exhibited rather weak acidity attributed to Brönsted sites. However,
the number of acid sites with a sorption heat between 80 and 110 kJ mol−1 was
significantly higher than for silicalite [133].
Titanium-substituted silicalite can be prepared with a homogeneous distribution
of Ti ions in the crystal. The Ti4+ ions seem to be all surrounded by four NNN Si
9 Characterization of Acid–Base Sites in Zeolites 375

ions and, thus, the catalytic site is an isolated Ti4+ ion. For TS-1 (titanium silicate
molecular sieve, zeolite of pentasil family), relatively low heats of adsorption due to
coordinatively bonded ammonia were detected by Jänchen et al. [132]. Moreover,
the amounts of adsorption with heats higher than found for silicalite correlated with
the amount of Ti in the sample [132].
Similarly, the acidity of titanium-silicalites with different titanium contents was
characterized by adsorption calorimetry at 353K of various probe molecules by
Muscas et al. [134]. These molecular sieves had a molar composition xTiO2
(1 − x)SiO2 , where x ranged from 0 to 0.02. Subjected to ammonia adsorp-
tion, these solids showed an acidic character compared to a pure silicalite-1 sam-
ple. A small amount of titanium induced a high increase in the strong acid sites
(Q init = 160–170 kJ mol−1 instead of 75 for silicalite-1). The integral heat and the
total amount of acid sites increased with increasing titanium loading and then reached
a plateau for x ≥ 0.014. All the curves showed a sharp decrease in Q diff at very low
coverage. The next region corresponded to a plateau with heats evolving around
70 kJ mol−1 , instead of 40 kJ mol−1 found for silicalite-1. Other basic probes such
as pyridine and substituted pyridines (DMP) were also used in an attempt to identify,
by selective adsorption, the different sites of these catalysts [134]. The heats and
amounts adsorbed were in the order: pyridine > 3.5 lutidine > 2.6 lutidine.
The adsorption properties of titanium silicalites-1 synthesized via two different
routes (in the presence or the absence of sodium in the precursor gel) have been com-
pared by Auroux et al. [135]. Adsorption calorimetric measurements of a basic probe
(NH3 ) and an acidic probe (SO2 ) showed that these solids were very acidic compared
to a silicalite-1 sample. The presence of Na in the different samples decreased the
number and the strength of the acid sites. The modification strongly depended on the
synthesis procedure [135].
The treatment of Ti-silicalite-1 (TS-1) and silicalite by aqueous solutions of
ammonium acetate has been shown to suppress the most energetic sites on these
two catalysts, as evidenced by the heats of adsorption which were much lower for
the treated samples than for the untreated ones [136], while the number of NH3
molecules absorbed per Ti atom was unaffected by the treatment in the case of TS-1.
In another study, the heterogeneity of framework Ti (IV) atoms in Ti-silicalite
(TS-1) was studied by NH3 adsorption calorimetry [137] and compared to a Ti-free
silicalite taken as reference material. The evolution of the heat of adsorption with
coverage was found to be typical of heterogeneous surfaces, not only due to the
presence of sites active towards ammonia on the silicalite matrix, but also due to the
heterogeneous distribution of Ti (IV) sites. This suggests that a considerable number
of framework sites (among the 12 available in the MFI framework) is occupied in a
nearly equally distributed manner.
376 D. Stošić and A. Auroux

9.3 Correlation Between Adsorption Heat and Catalytic Activity

Finding correlations in between catalysts surface properties and catalytic behaviour

is the main reason to study adsorption of probe molecules on catalysts surfaces. If
the probe molecules are chosen to resemble possible reaction intermediates of the
catalytic cycle, measurements of heat of adsorption of such probe molecules can
provide essential information about reaction mechanisms.
Finding correlations between the heats of adsorption and the catalytic behaviour
(activity and/ or selectivity) of the solid catalysts is not an easy task. Surface sites
may exist in different configurations what typically brings to the distribution of
adsorption sites’ strengths. Among the active sites present on the surface of catalyst
certain number may possess adequate strength to activate the adsorbed molecule and
to form a reactive intermediate [52]. In other words, some of the sites may be too
weak to activate the reactants or they can be too strong, leading to strongly held
species which block and deactivate these sites or cause excessive fragmentation of
reactants or products.
Therefore, the determination of the strength and strength distribution of surface
active sites is of fundamental importance, because in this way one can distinguish the
undesired sites from the desired ones. This further gives possibility of creating novel
catalysts, which properties can be properly tuned for the target reaction. In revealing
these very important features, adsorption microcalorimetry plays one irreplaceable
role.
In a study performed by Auroux et. al. [111] microcalorimetry experiments
of ammonia and sulfur dioxide were performed in order to analyze the possible
correlations between the acidity and basicity of the alkali-metal ion-exchanged X
and Y zeolite structures and their catalytic properties. The catalytic results for the
4-methylpentan-2-ol conversion show that activity and selectivity are both affected
to some extent by the acid-base character of the catalysts. The activity was found
to increase in order Cs > Rb > K > Na > Li for both X and Y zeolites. The
dehydrogenation reaction occurs only on CsX + Cs2 O, which presents very strong
basicity. The product selectivity of the reaction depends on both Lewis acidity and
basicity; Lewis basic or acidic sites of zeolites can be considered as acid-base pairs,
in which both framework basic oxygens and neighboring cations are important. The
selectivity ratio between the 1-alkene and (2-alkene+isomers) increases linearly with
the ratio between basic and acid sites number, n SO2 /n NH3 , for both X and Y zeolites,
as shown in Fig. 9.14.
In another study [138], authors presented evidence that heats of adsorption of
ammonia on three different kinds of zeolite structure (Y, mordenite, ZSM-5) can
be used to obtain the correlation plots that describe relationships between acidic
and catalytic properties of these catalysts. Catalytic properties were tested in simple
carbonium ion reactions. An approach to search “acidity-catalytic activity” corre-
lations was based on the straightforward concept which implies that the heat of
adsorption of the base is directly related to the energy needed for the protonation
of hydrocarbon molecules leading to the carbocation formation. For example it was
9 Characterization of Acid–Base Sites in Zeolites 377

Fig. 9.14 1-ene/(2-ene+isomers) selectivity ratio versus the basic/acidic site number for X samples
left and for Y samples right

Fig. 9.15 Variation in the number of strong acid sites with Q NH3 = 122–136 kJ/mol (a) and activity
in cracking of isooctane as a function of NAl for Y zeolites (b) prepared by treating the stabilized
Y zeolites with HCl (1), prepared by thermal dealumination at 600–700◦ C (2) and produced by
treatment of sodium faujasites with SiCl4 (3)

Fig. 9.16 Plots of number of strong Brønsted acid sites (♦) and catalytic activity for n-hexane
cracking () as a function of percentage XRD crystallinity of ZSM-5-based samples[139]
378 D. Stošić and A. Auroux

found that the variation of strong acidity and the change of activity with progressing
dealumination of Y zeolites show similar trend (Fig. 9.15).
In a study on ZSM-5 based materials [139], Nicolaides et. al. have shown that
the number of Brönsted acid sites, determined by adsorption microcalorimetry of
ammonia, increases with increasing XRD cristallinity. A strong correlation was
observed between the catalytic activity of these materials, in n-hexane cracking reac-
tion, and the number of strong acid sites, as presented in Fig. 9.16.
Microcalorimetry study of NH3 adsorption has been used to characterize the acid
sites of H-USY zeolite and another USY sample in which the strong Lewis sites
were poisoned with ammonia. Poisoning of Lewis acid sites did not affect the rate
of deactivation, the cracking activity, or the distribution of cracked products during
2-mehylpentene cracking. From these findings it was concluded that strong Lewis
sites does not play any important role in cracking reactions [140].
The effect of the Si/Al ratio of HZSM-5 zeolite-based catalysts on surface acid-
ity and on selectivity in the transformation of methanol into hydrocarbons has been
studied using adsorption calorimetry of ammonia and tert-butylamine. The observed
increase in light olefin selectivity and decrease in methanol conversion with increas-
ing Si/Al ratio can be explained by a decrease in total acidity [141].
However, it is worth noting that experiments aimed at directly comparing the
cracking activity and the enthalpy of adsorption of basic probe molecules have
sometimes failed to make correlative conclusions. For example, Gorte et al. [128]
detected no difference in the enthalpies of ammonia or pyridine adsorption on H-
[Al] ZSM5, H-[Ga] ZSM5 and H-[Fe] ZSM5 samples whose cracking activities were
quite different. Likewise, in a study on dealuminated Y zeolites, no significant differ-
ences in the highest enthalpies of adsorption could be detected by microcalorimetry
for samples of different cracking activities [93]. Kuehne et al. [142] confirm these
conclusions by the results of another study where a 57 kJ mol−1 difference in the
enthalpy of ammonia adsorption was detected between HY and (H, NH4)- USY
sample, although the latter was 33 times more active than HY. It means that there is
only a certain category of sites with a given strength that are active in the reaction.

References

1. D.W. Breck, Zeolite Molecular Sieves: Structure Chemistry and Use (Wiley, New York, 1974)
2. P.A. Jacobs, Carboniogenic Activity of Zeolites (Elsevier Science Ltd., Amsterdam, 1977)
3. J.V. Smith et al., Topochemistry of zeolites and related materials. 1. Topology geometry.
Chem. Rev. 88, 149–182 (1988)
4. W.M. Meier, Molecular Sieves (Society of Chemical industry, London, 1968)
5. R. Fricke, H. Kosslick, G. Lische, M. Richter et al., Incorporation of gallium into zeolites:
syntheses, properties and catalytic application. Chem. Rev. 100, 2303–2405 (2000)
6. Database of zeolite structures (2000) The International Zeolite Association, http://www.iza-
structure.org/databases/, Accessed 24 Sept. 2011
7. A. Dyer, An Introduction to Zeolite Molecular Sieves (Wiley, New York, 1988)
8. W.M. Meier, Zeolites and zeolite like materials. Stud. Surf. Sci. Catal. 28, 13–22 (1986)
9 Characterization of Acid–Base Sites in Zeolites 379

9. J.B. Uytterhoeven, L.C. Christner, W.K. Hall et al., Studies of the hydrogen held by solids.
VIII. The decationated zeolites. J. Phys. Chem. 69, 2117–2126 (1965)
10. R.P. Townsend, R. Harjula, in Ion Exchange in Molecular Sieves by Conventional Techniques,
vol. 3, ed. by H. Karge, J. Weitkamp. Molecular Sieves–Science and Technology, (Springer,
Berlin, 2002), pp. 2–42
11. R.P. Townsend, Ion exchange in zeolites. Stud. Surf. Sci. Catal. 58, 359–390 (1991)
12. W.E. Farneth, R.J. Gorte et al., Methods for characterizing zeolites acidity. Chem. Rev. 95,
615–635 (1995)
13. M. Niwa, N. Katada, K. Okumura, Characterization and Design of Zeolite Catalysts (Springer,
Berlin, 2010)
14. H. Ichihashi, M. Ishida, A. Shiga, M. Kitamura, T. Suzuki, K. Suenobu, K. Sugita et al., The
catalysis of vapor-phase Beckmann rearrangement for the production of ε-caprolactam. Catal.
Surv. Asia. 7, 261–270 (2003)
15. S. Shimizu, N. Abe, A. Iguchi, H. Sato et al., Synthesis of pyridine bases: general methods
and recent advances in gas phase synthesis over ZSM-5 zeolite. Catal. Surv. Jpn. 2, 71–76
(1998)
16. H. Ishida et al., Liquid-phase hydration process of cyclohexene with zeolites. Catal. Surv.
Jpn. 1, 241–246 (1997)
17. H. Tsuneki, M. Kirishiki, T. Oku et al., Development of 2,2’-iminodiethanol selective pro-
duction process using shape-selective pentasil-type zeolite catalyst. Bull. Chem. Soc. Jpn. 80,
1075–1090 (2007)
18. M. Iwamoto, H. Yahiro, K. Tanda, N. Mizuno, Y. Mine, S. Kagawa et al., Removal of nitrogen
monoxide through a novel catalytic process. 1. Decomposition on excessively copper-ion-
exchanged ZSM-5 zeolites. J. Phys. Chem. 95, 3727–3730 (1991)
19. X.B. Feng, W.K. Hall et al., FeZSM-5: a durable SCR catalyst for NOx removal from com-
bustion streams. J. Catal. 166, 368–376 (1997)
20. Y.J. Li, J.N. Armor et al., Catalytic reduction of nitrogen oxides with methane in the presence
of excess oxygen. Appl. Catal. B Environ. 1, L31–L40 (1992)
21. D.J. Wang, J.H. Lunsford, M.P. Rosynek et al., Characterization of a Mo/ZSM-5 catalyst for
the conversion of methane to benzene. J. Catal. 169, 347–357 (1997)
22. Y. Ashina, T. Fujita, M. Fukatsu, K. Niwa, J. Yagi et al., Manufacture of dimethylamine using
zeolite catalyst. Stud. Surf. Sci. Catal. 28, 779–786 (1986)
23. A. Corma, V. Martinez-Soria, E. Schnoeveld et al., Alkylation of benzene with short-chain
olefins over MCM-22 zeolite: catalytic behaviour and kinetic mechanism. J. Catal. 192, 163–
173 (2000)
24. U. Freese, F. Heinrich, F. Roessner et al., Acylation of aromatic compounds on H-Beta zeolites.
Catal. Today. 49, 237–244 (1999)
25. T.R. Hughes, W.C. Buss, P.W. Tamm, R.L. Jacobson et al., Aromatization of hydrocarbons
over platinum alkaline earth zeolites. Stud. Surf. Sci. Catal. 28, 725–732 (1986)
26. R.E. Jentoft, M. Tsapatsis, M.E. Davis, B.C. Gates et al., Platinum clusters supported in zeolite
LTL: influence of catalyst morphology on performance in n-hexane reforming. J. Catal. 179,
565–580 (1998)
27. H. Pfiefer, D. Freude, J. Kärger, Catalysis and Adsorption by Zeolites (Elsevier, Amsterdam,
1991)
28. M. Hunger, D. Freude, D. Fenzke, H. Pfeifer et al., 1 H solid-state NMR studies of the geometry
of Brönsted acid sites in zeolites H-ZSM-5. Chem. Phys. Lett. 191, 391–395 (1992)
29. T.J. Gluszak, D.T. Chen, S.B. Sharma, J.A. Dumesic, T.W. Root et al., Observation of Brönsted
acid sites of D-Y zeolites with deuterium NMR. Chem. Phys. Lett. 190, 36–41 (1992)
30. P. Batamack, C. Doremieux-Morin, J. Fraissard et al., Broad-line 1 H NMR: a new application
for studying the Brönsted acid strength of solids. J. Phys. Chem. 97, 9779–9783 (1993)
31. A.L. Blumenfeld, J.J. Fripiat et al., Characterization of Brönsted and Lewis acidity in zeolites
by solid-state NMR and the recent progress in the REDOR technique. Magn. Reson. Chem.
37, S118–S125 (1999)
380 D. Stošić and A. Auroux

32. A.G. Pelmenshchikov, E.A. Paukshtis, V.G. Stepanov, V.I. Pavlov, E.N. Yurchenko, K.G.
Ione, S. Beran et al., Scattering vector dependence of mutual diffusion coefficients for rodlike
mlcelles in aqueous sodlum halide solutions. J. Phys. Chem. 93, 6720–6725 (1989)
33. K.P. Schröder, J. Sauer, M. Leslie, C.R.A. Catlow, J.M. Thomas et al., Bridging hydroxyl
groups in zeolitic catalysts: a computer simulation of their structure, vibrational properties
and acidity in protonated faujasites (H-Y zeolites). Chem. Phys. Lett. 188, 320–325 (1992)
34. K. Suzuki, T. Noda, N. Katada, M. Niwa et al., IRMS-TPD of ammonia: Direct and individual
measurement of Brönsted acidity in zeolites and its relationship with the catalytic cracking
activity. J. Catal. 250, 151–160 (2007)
35. N. Katada, H. Igi, J.H. Kim, M. Niwa et al., Determination of the acidic properties of zeolite by
theoretical analysis of temperature-programmed desorption of ammonia based on adsorption
equilibrium. J. Phys. Chem. B 101, 5969–5977 (1997)
36. J.R. Anderson, K. Foger, T. Mole, R.A. Rajadhyaksha, J.V. Sanders et al., Reactions on ZSM-
5-type zeolite catalysts. J. Catal. 58, 114–130 (1979)
37. V.S. Nayak, V.R. Choudhary et al., Acid strength distribution and catalytic properties of
H-ZSM-5: effect of deammoniation conditions of NH4 -ZSM-5. J. Catal. 81, 26–45 (1983)
38. D. Atkinson, G. Curthoys et al., The acidity of solid surfaces and its determination by amine
titration and adsorption of coloured indicators. Chem. Soc. Rev. 8, 475–479 (1979)
39. B.S. Umansky, W.K. Hall et al., A spectrophotometric study of the acidity of some solid acids.
J. Catal. 124, 97–108 (1990)
40. D. Farcasiu, A. Ghenciu et al., Acidity functions from 13 C-NMR. J. Am. Chem. Soc. 115,
10901–10908 (1993)
41. E.P. Parry, An infrared study of pyridine adsorbed on acidic solids. Characterization of surface
acidity. J. Catal. 2, 371–379 (1963)
42. C.A. Emeis, Determination of integrated molar extinction coefficients for infrared absorption
bands of pyridine adsorbed on solid acid catalysts. J. Catal. 141, 347–354 (1993)
43. O. Cairon, T. Chevreau et al., Quantitative FTIR studies of hexagonal and cubic faujasites by
pyridine and CO adsorption. J. Chem. Soc. Faraday Trans. 94, 323–330 (1998)
44. R. Borade, A. Sayari, A. Adnot, S. Kaliaguine et al., Characterization of acidity in ZSM-5
zeolites: an X-ray photoelectron and IR spectroscopy study. J. Phys. Chem. 94, 5989–5994
(1990)
45. C. Guimon, A. Boreave, G. Pfister-Guillouzo et al., Study of the bulk and surface acidity of
protonated Y zeolites by TPD and XPS. Surf. Interface Anal. 22, 407–411 (1994)
46. A. Auroux, Acidity characterization by microcalorimetry and relationship with reactivity.
Top. Catal. 4, 71–89 (1997)
47. A. Auroux, Microcalorimetry methods to study the acidity and reactivity of zeolites, pillared
clays and mesoporous materials. Top. Catal. 19, 205–213 (2002)
48. A. Auroux, in Acidity and Basicity: Determination by Adsorption Microcalorimetry, vol. 6.
Molecular Sieves–Science and Technology: Acidity and Basicity, (Springer, Berlin, 2000),
pp. 46–154
49. N. Cardona-Martinez, J.A. Dumesic et al., Applications of adsorption microcalorimetry to
the study of heterogeneous catalysis. Adv. Catal. 38, 149–244 (1992)
50. V. Solinas, I. Ferino et al., Microcalorimetric characterization of acid-basic catalysts. Catal.
Today 41, 179–189 (1998)
51. A. Auroux, Thermal methods: calorimetry, differential thermal analysis and thermogravime-
try, in Thermal Methods in Catalysts Characterization, ed. by B. Imelik, J.C. Vedrine (Plenum
Press, New York, 1994), pp. 611–650
52. Lj. Damjanović, A. Auroux, Heterogeneous catalysis on solids, in Handbook of Thermal
Analysis and Calotimetry, vol. 5, ed. by M.E. Brown, P.K. Gallagher (Elsevier, Amsterdam,
2008), pp. 387–438
53. A. Corma, Inorganic solid acids and their use in acid-catalyzed hydrocarbon reactions. Chem.
Rev. 95, 559–614 (1995)
54. P.B. Weisz, Zeolites-new horizons in catalysis. Chemtech 3, 498–505 (1973)
9 Characterization of Acid–Base Sites in Zeolites 381

55. P.B. Weisz, V.J. Frilette et al., Intercrystalline and molecular-shape-selective catalysis by
zeolite salts. J. Phys. Chem. 64, 382 (1960)
56. W.O. Haag, R.M. Lago, P.B. Weisz et al., Transport and reactivity of hydrocarbon molecules
in a shape-selective zeolite. Faraday Discuss. Chem. Soc. 72, 317–331 (1982)
57. V.J. Frilette, W.O. Haag, R.M. Lago et al., Catalysis by crystalline aluminosilicates: charac-
terization of intermediate pore-size zeolites by the “Constraint Index”. J. Catal. 67, 218–222
(1981)
58. S.M. Csicsery, In: J.A. Rabo (eds), Zeolite chemistry and catalysis. ACS Symposium series
171, 680–701 (1976)
59. S.M. Csicsery, Shape-selective catalysis in zeolites. Zeolites 4, 202–213 (1984)
60. S.M. Csicsery, The reactions of 1-methyl-2-ethylbenzene I. Exploring the structure of
intracrystalline void space and the catalytic properties of molecular sieves and other cata-
lysts. J. Catal. 108, 433–443 (1987)
61. E.G. Derouane, Shape selectivity in catalysis by zeolites: the nest effect. J. Catal. 100, 541–544
(1986)
62. E. G. Derouane, New aspects of molecular shape-selectivity: catalysis by zeolite ZSM - 5.
Stud. Surf. Sci. Catal. 4, 5–18 (1980)
63. E.G. Derouane, Molecular shape-selective catalysis in Zeolites - Selected topics. Stud. Surf.
Sci. Catal. 19, 1–17 (1984)
64. J. Dwyer, Uses of natural zeolites. Chem. Ind. 7, 241–245 (1984)
65. P.A. Jacobs, J.A. Martens et al., Exploration of the void size and structure of zeolites and
molecular sieves using chemical reactions. Stud. Surf. Sci. Catal. 28, 23–32 (1986)
66. U. Hammon, G.T. Kokotailo, L. Riekert, J.Q. Zhou et al., Deactivation of dealuminated zeolite
Y in the catalytic cracking of n-hexane. Zeolites 8, 338–339 (1988)
67. E.G. Derouane, Z. Gabelica et al., A novel effect of shape selectivity: molecular traffic control
in zeolite ZSM-5. J. Catal. 65, 486–489 (1980)
68. J. Kärger, D.M. Ruthven, Diffusion in Zeolites and Other Microporous Solids (Wiley, New
York, 1992)
69. J. Datka, M. Boczar, P. Rymarowicz et al., Heterogeneity of OH groups in NaH-ZSM-5 zeolite
studied by infrared spectroscopy. J. Catal. 114, 368–376 (1988)
70. A. Chatterjee, T. Iwasaki, T. Ebina, H. Tsuruya, T. Kanougi, Y. Oumi, M. Kubo, A. Miyamoto
et al., Effects of structural characteristics of zeolites on the properties of their bridging and
terminal hydroxyl groups. Appl. Surf. Sci. 130–132, 555–560 (1998)
71. J. Sauer, Molecular models in ab initio studies of solids and surfaces: from ionic crystals and
semiconductors to catalysts. Chem. Rev. 89, 199–255 (1989)
72. J. Sauer, C.M. Kölmel, J.R. Hill, R. Ahlrichs et al., Brönsted sites in zeolitic catalysts. An
ab initio study of local geometries and of the barrier for proton jumps between neighbouring
sites. Chem. Phys. Lett. 164, 193–198 (1989)
73. J.B. Nicholas, R.E. Winans, R.J. Harrison, L.E. Iton, L.A. Curtiss, A.J. Hopfinger et al.,
Ab initio molecular orbital study of the effects of basis set size on the calculated structure and
acidity of hydroxyl groups in framework molecular sieves. J. Phys. Chem. 96, 10247–10257
(1992)
74. J. Limatrakul, S. Hannongbua et al., Catalytic properties of a free hydroxyl on a silica, a
zeolite, and modified zeolites: quantum-chemical calculations. J. Mol. Struc. (Theochem)
280, 139–145 (1993)
75. R. Carson, E.M. Cooke, J. Dwyer, A. Hinchliffe, P.J. O’Malley, Molecular orbital calculations
of structural and acidic characteristics of aluminophosphates (AlPO), silicoaluminophos-
phates (SAPO) and metaluminophosphate (MeAPO) based molecular sieves. Stud. Surf. Sci.
Catal. 46, 39–48 (1989)
76. D.J. Parrillo, R.J. Gorte et al., Characterization of acidity in H-ZSM-5, H-ZSM-12,
H-mordenite, and H-Y using microcalorimetry. J. Phys. Chem. 97, 8786–8792 (1993)
77. D.J. Parrillo, C. Lee, R.J. Gorte et al., Heats of adsorption for ammonia and pyridine in
H-ZSM-5: evidence for identical Brönsted-acid sites. Appl. Catal. A 110, 67–74 (1994)
382 D. Stošić and A. Auroux

78. M. Huang, S. Kaliaguine, A. Auroux et al., Crystallinity dependence of Lewis acidity and
Lewis basicity in Na zeolites. J. Phys. Chem. 99, 9952–9959 (1995)
79. M. Huang, S. Kaliaguine, A. Auroux et al., Lewis basic and Lewis acidic sites in zeolites.
Stud. Surf. Sci. Catal. 97, 311–318 (1995)
80. M. Huang, A. Auroux, S. Kaliaguine et al., Crystallinity dependence of acid site distribution
in HA, HX and HY zeolites. Micropor. Mater. 5, 17–27 (1995)
81. C.V. McDaniel, P.K. Maher et al., In: J.A. Rabo (eds). Zeolite chemistry and catalysis. ACS
Symposium series 171, 171–180 (1976)
82. M.A. Camblor, A. Corma, S. Valencia et al., Spontaneous nucleation and growth of pure silica
zeolite-P free of connectivity defects, Chem. Commun. 2365–2366 (1996)
83. M.A. Camblor, A. Corma, A. Mifsud, J. Perez-Pariente, S. Valencia et al., Synthesis of
nanocrystalline zeolite beta in the absence of alkali metal cations. Stud. Surf. Sc. Catal.
105, 341–348 (1997)
84. P.A. Jacobs, R. von Ballmoos et al., Framework hydroxyl groups of H-ZSM-5 zeolites. J.
Phys. Chem. 86, 3050–3052 (1982)
85. T.J. Gricus Kofke, R.J. Gorte, W.E. Farneth et al., Stoichiometric adsorption complexes in
H-ZSM-5. J. Catal. 114, 34–45 (1988)
86. T.J. Gricus Kofke, R.J. Gorte, G.T. Kokotailo, W.E. Farneth et al., Stoichiometric adsorption
complexes in H-ZSM-5, H-ZSM-12, and H-mordenite zeolites. J. Catal. 115, 265–272 (1989)
87. A.I. Biaglow, D.J. Parrillo, R.J. Gorte et al., Characterization of H, Na-Y using amine des-
orption. J. Catal. 144, 193–201 (1993)
88. K. Tsutsumi, K. Nishimiya et al., Differential molar heats of adsorption of ammonia on
silicious mordenites at high temperature. Thermochim. Acta. 143, 299–309 (1989)
89. A. Auroux, P.C. Gravelle, J.C. Védrine, M. Rekas, in Proc 5th Int Zeolite Conf, Naples, Italy,
June 2–6, 1980, ed. by L.V.C. Rees (LV, Heyden, London, 1980), pp. 433–438
90. H. Stach, J. Jänchen, U. Lohse et al., Relationship between acid strength and framework
aluminium content in dealuminated faujasites. Catal. Lett. 13, 389–393 (1992)
91. A. Auroux, Y. Ben Taarit et al., Calorimetric investigation of the effect of dealumination on
the acidity of zeolites. Thermochim. Acta. 122, 63–70 (1987)
92. Y. Mitani, K. Tsutsumi, H. Takahashi et al., Direct measurement of the interaction energy
between solids and gases. XI. Calorimetric measurements of acidities of aluminum deficient
H-Y zeolites. Bull. Chem. Soc. Jpn. 56, 1921–1923 (1983)
93. A.I. Biaglow, D.J. Parrillo, G.T. Kokotailo, R.J. Gorte et al., A study of dealuminated faujasites.
J. Catal. 148, 213–223 (1994)
94. I.V. Mishin, A.L. Klyachko, T.R. Brueva, O.P. Tkachenko, H.K. Beyer et al., Composition,
acidity and catalytic activity of high silica fujasites. Kinet. Catal. 34, 502–508 (1993)
95. D. Barthomeuf, Aluminium topological density and correlations with acidic and catalytic
properties of zeolites. Stud. Surf. Sci. Catal. 38, 177–186 (1987)
96. H. Stach, J. Jänchen, H.G. Jerschkewitz, U. Lohse, B. Parlitz, M. Hunger et al., Mordenite
acidity: dependence on the silicon/aluminum ratio and the framework aluminum topology. 2
Acidity investigations. J. Phys. Chem. 96, 8480–8485 (1992)
97. A. Macedo, A. Auroux, F. Raatz, E. Jacquinot, R. Boulet et al., Strong acid sites of dealu-
minated Y zeolites prepared by conventional treatments and isomorphous substitution. ACS
Symp. Ser. 368, 98–104 (1988)
98. J.C. Védrine, A. Auroux, G. Coudurier, Catalytic materials, relationship between structure
and reactivity. ACS Symp. Ser. 248(13), 254 (1984)
99. J.C. Védrine, A. Auroux, V. Bolis, P. Dejaifve, C. Naccache, P. Wierzchowski, E.G. Derouane,
J.C.H. van Hoff et al., Infrared, microcalorimetric, and electron spin resonance investigations
of the acidic properties of the H-ZSM-5 zeolite. J. Catal. 59, 248–262 (1979)
100. H.G. Karge, L.C. Jozefowicz et al., A comparative study of the acidity of various zeolites using
the differential heats of ammonia adsorption as measured by high-vacuum microcalorimetry.
Stud. Surf. Sci. Catal. 84, 685–692 (1994)
101. L.C. Jozefowicz, H.G. Karge, E.N. Coker et al., Microcalorimetric investigation of H-ZSM-5
zeolites using an ultrahigh-vacuum system for gas adsorption. J. Phys. Chem. 98, 8053–8060
(1994)
9 Characterization of Acid–Base Sites in Zeolites 383

102. D.T. Chen, S.B. Sharma, I. Filimonov, J.A. Dumesic et al., Microcalorimetric studies of zeolite
acidity. Catal. Lett. 12, 201–211 (1992)
103. G.I. Kapustin, L.M. Kustov, G.O. Glonti, T.R. Brueva, V.Y. Borovkov, A.L. Klyachko, A.M.
Rubinshtein, V.B. Kazanskii et al., A microcalorimetric study of NH3 adsorption on H, Na-mor
catalysts. Kinet. Katal. 25, 959–964 (1984)
104. I. Bankós, J. Valyon, G.I. Kapustin, D. Kallo, A.L. Klyachko, T.R. Brueva et al., Acidic and
catalytic properties of hydrogen sodium mordenite. Zeolites 8, 189–195 (1988)
105. H.G. Karge, H.K. Beyer, Solid-state ion exchange in microporous and mesoporous materi-
als, in Molecular Sieves–Science and Technology, vol. 3, ed. by H.G. Karge, J. Weitkamp
(Springer, New York, 2002)
106. H. Garcia, H.D. Roth et al., Generation and reactions of organic radical cations in zeolites.
Chem. Rev. 102, 3947–4008 (2002)
107. I.V. Mishin, A.L. Klyachko, G.I. Kapustin, H.G. Karge et al., The effect of exchange degrees
on the heterogeneity of acid sites in decationated mordenites, Kinet. Katal . 34, 828–834
(1993)
108. I.V. Mishin, A.L. Klyachko, T.R. Brueva, V.D. Nissenbaum, H.G. Karge et al., Effect of the
exchange degree on the heterogeneity of acid sites of decationed zeolites. Kinet. Catal. 34,
835–840 (1993)
109. M. Muscas, J.F. Dutel, V. Solinas, A. Auroux, Y. Ben Taarit et al., A dynamic assessment of
MFI acidity using microcalorimetric techniques. J. Mol. Catal. A 106, 169–175 (1996)
110. M. Huang, S. Kaliaguine, M. Muscas, A. Auroux et al., Microcalorimetric characterization
of the basicity in alkali-exchanged X zeolites. J. Catal. 157, 266–269 (1995)
111. A. Auroux, P. Artizzu, I. Ferino, R. Monaci, E. Rombi, V. Solinas et al., Conversion of
4-methylpentan-2-ol over alkali-metal ion-exchanged X and Y zeolites: a microcalorimetric
and catalytic investigation. Micropor. Mater. 11, 117–126 (1997)
112. R.M. Barrer, Hydrothermal Chemistry of Zeolites (Academic Press, London, 1982)
113. R. Szostak, Molecular Sieves: Principle of Synthesis and Identification (Van Nostrand Rein-
hold, New York, 1989)
114. T. Chapus, A. Tuel, Y. Ben Taarit, C. Naccache et al., Synthesis and characterization of a
chromium silicalite-1. Zeolites 14, 349–355 (1994)
115. K.G. Ione, L.A. Vostrikova, V.M. Mastikin et al., Synthesis of crystalline metal silicates having
zeolite structure and study of their catalytic properties. J. Mol. Catal. 31, 355–370 (1985)
116. K.G. Ione, L.A. Vostrikova et al., The isomorphism and catalytic properties of silicates with
the zeolite structure. Uspechi. Chimii. 56, 231–252 (1987)
117. L. Pauling, The Nature of the Chemical Bond, 3rd ed. (Cornell University Press, Ithaka,
New York, 1967)
118. F. Liebau, Structural Chemistry of Silicates (Springer, Berlin, 1985)
119. M. Sayed, A. Auroux, J.C. Védrine et al., The effect of boron on ZSM-5 zeolite shape selec-
tivity and activity: II. Coincorporation of aluminium and boron in the zeolite lattice. J. Catal.
116, 1–10 (1989)
120. G. Coudurier, A. Auroux, J.C. Védrine, R.D. Farlee, L. Abrams, R.D. Shannon et al., Properties
of boron-substituted ZSM-5 and ZSM-11 zeolites. J. Catal. 108, 1–14 (1987)
121. M.L. Occelli, H. Eckert, A. Wölker, A. Auroux et al., Crystalline galliosilicate molecular
sieves with the beta structure. Micropor. Mesopor. Mat. 30, 219–232 (1999)
122. M.L. Occelli, H. Eckert, C. Hudalla, A. Auroux, P. Ritz, P.S. Iyer et al., Acidic properties
of galliosilicate molecular sieves with the offretite structure. Stud. Surf. Sci. Catal. 105,
1981–1987 (1997)
123. M.L. Occelli, G. Schwering, C. Fild, H. Eckert, A. Auroux, P.S. Iyer et al., Galliosilicate
molecular sieves with the faujasite structure. Micropor. Mesopor. Mat. 34, 15–22 (2000)
124. H. Eckert, C. Hudalla, A. Wölker, A. Auroux, M.L. Occelli et al., Synthesis and structural
characterization of mixed aluminum-gallium-offretites. Solid State NMR 9, 143–153 (1997)
125. M.L. Occelli, A.E. Schweizer, C. Fild, G. Schwering, H. Eckert, A. Auroux et al., Gallioalu-
minosilicate molecular sieves with the faujasite structure. J. Catal. 192, 119–127 (2000)
384 D. Stošić and A. Auroux

126. B. Ducourty, M.L. Occelli, A. Auroux et al., Processes of ammonia adsorption in gallium
zeolites as studied by microcalorimetry. Thermochim. Acta 312, 27–32 (1998)
127. E. Dumitriu, V. Hulea, I. Fechete, C. Catrinescu, A. Auroux, J.F. Lacaze, C. Guimon et al.,
Prins condensation of isobutylene and formaldehyde over Fe-silicates of MFI structure. Appl.
Catal. A Gen. 181, 15–28 (1999)
128. D.J. Parrillo, C. Lee, R.J. Gorte, D. White, W.E. Farneth et al., Comparison of the acidic prop-
erties of H-[Al]ZSM-5, H-[Fe]ZSM-5, and H-[Ga]ZSM-5 using microcalorimetry, hexane
cracking, and propene oligomerization. J. Phys. Chem. 99, 8745–8749 (1995)
129. E. Dumitriu, V. Hulea, I. Fechete, A. Auroux, J.F. Lacaze, C. Guimon et al., The aldol con-
densation of lower aldehydes over MFI zeolites with different acidic properties. Micropor.
Mesopor. Mat. 43, 341–359 (2001)
130. J. Kotrla, L. Kubelkova, C.C. Lee, R.J. Gorte, Calorimetric and FTIR studies of acetonitrile
on H-[Fe]ZSM-5 and H-[Al]ZSM-5. J. Phys. Chem. B 102, 1437–1443 (1998)
131. A. Auroux, A. Tuel, J. Bandiera, J.M. Guil et al., Calorimetric and catalytic investigation of
alkanes reactivity over a variety of MFI structures. Appl. Catal. 93, 181–190 (1993)
132. J. Jänchen, G. Vorbeck, H. Stach, B. Parlitz, J.H.C. van Hooff et al., Adsorption calorimetric
and spectroscopic studies on isomorphous substituted (Al, Fe, In, Ti) MFI zeolites. Stud. Surf.
Sci. Catal. 94, 108–115 (1995)
133. G. Vorbeck, J. Jänchen, B. Parlitz, M. Schneider, R. Fricke et al., Synthesis and characterization
of crystalline indosilicates with the MFI structure. J. Chem. Soc. Chem. Com. 123–124 (1994)
134. M. Muscas, V. Solinas, S. Gontier, A. Tuel, A. Auroux et al., Microcalorimetry studies of the
acidic properties of titanium-silicalites-1. Stud. Surf. Sci. Catal. 94, 101–107 (1995)
135. A. Auroux, A. Gervasini, E. Jorda, A. Tuel et al., Acidic properties of titanium-silicalites-1.
Stud. Surf. Sci. Catal. 84, 653–659 (1994)
136. V. Bolis, S. Bordiga, C. Lamberti, A. Zecchina, A. Carati, F. Rivetti, G. Spano, G. Petrini et
al., A calorimetric, IR, XANES and EXAFS study of the adsorption of NH3 on Ti-silicalite
as a function of the sample pre-treatment. Micropor. Mater. 30, 67–76 (1999)
137. V. Bolis, S. Bordiga, C. Lamberti, A. Zecchina, A. Carati, F. Rivetti, G. Spano, G. Petrini
et al., Heterogeneity of framework Ti(IV) in Ti-silicalite as revealed by the adsorption of NH3 .
Combined calorimetric and spectroscopic study. Langmuir 15, 5753–5764 (1999)
138. I.V. Mishin, T.R. Brueva, G.I. Kapustin et al., Heats of adsorption of ammonia and correlation
of activity and acidity in heterogeneous catalysis. Adsorption 11, 415–424 (2005)
139. C.P. Nicolaides, H.H. Kung, N.P. Makgoba, N.P. Sincadu, M.S. Scurrell et al., Characterization
by ammonia adsorption microcalorimetry of substantially amorphous or partially crystalline
ZSM-5 materials and correlation with catalytic activity. Appl. Catal. A Gen. 223, 29–33 (2002)
140. S.M. Babitz, M.A. Kuehne, H.H. Kung, J.T. Miller et al., Role of Lewis acidity in the deactiva-
tion of USY zeolites during 2-methylpentane cracking. Ind. Eng. Chem. Res. 36, 3027–3031
(1997)
141. A.G. Gayubo, P.L. Benito, A.T. Aguayo, M. Olazar, J. Bilbao et al., Effect of Si/Al ratio and
of acidity of H-ZSM5 zeolites on the primary products of methanol to gasoline conversion.
J. Chem. Tech. Biotechnol. 65, 183–191 (1996)
142. M.A. Kuehne, H.H. Kung, J.T. Miller et al., Effect of steam dealumination on H-Y acidity
and 2-methylpentane cracking activity. J. Catal. 171, 293–304 (1997)
Chapter 10
Adsorption/Desorption of Simple Pollutants

Vesna Rakić

Abstract This text reviews the possibilities to apply microcalorimetry and thermo-
analytical methods of analysis in the fields related to the environment pollution.
At the beginning, short overview of chemical species that can be found as com-
mon pollutants in the atmosphere, waters and soils is given. Further, it is shown
how the mentioned techniques can be applied for direct investigation of some event
that includes specific pollutants. The possibilities to use calorimetry and thermo-
analytical methods for the characterization of substances used either as adsorbents
or catalysts in the processes of pollutants’ abatement are presented. Besides, it is
shown how all mentioned methods can provide data useful in the removal of cer-
tain pollutant. The importance of microcalorimetry and thermo-analytical methods
in environment protection is underlined.

10.1 Introduction

Nowadays, environment is in the middle of our concerns. The pollution of the envi-
ronment is very complex problem: numerous chemicals that have been recognized
as poisonous or dangerous in any other sense come in the environment; mainly as a
result of anthropogenic actions. Atmosphere, waters and soils may contain a number
of different organic and inorganic residues present in a wide range of concentrations.
Therefore, their de-pollution has become one of the most important tasks worldwide,
imposed also by legislative procedures for environment protection.
The drastic increase of fossil fuel consumption linked to the industrialization of
the world is at the origin of the emission of the major part of atmospheric pollu-
tion. The energy stored in fossil fuels is freed mostly in air-combustion processes
by flame. The operative temperatures thermodynamically favour the formation of

V. Rakić (B)
Faculty of Agriculture, University of Belgrade, Nemanjina 6, 11080 Zemun, Serbia
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 385

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_10,
© Springer-Verlag Berlin Heidelberg 2013
386 V. Rakić

nitrogen oxides: N2 O, and in particular NO and NO2 , jointly known as NOx . Besides,
flame combustion of fossil fuels is a common source of CO2 and SOx . The oxides
of nitrogen (NOx ) and sulphur (SOx ) are responsible for occurrence of the acid-
rain phenomenon. The increasing N2 O and CO2 emissions have been identified as
main contributors to global warming what is commonly known as the "greenhouse
effect" [1–7]. In addition, there are the other pollutants which make the problem of
atmospheric pollution even more complicated. For example, the emissions of volatile
organic compounds (VOCs) into the atmosphere have the potential to cause adverse
effects on human health and the environment [8]. Coal-burning power plants, iron
and steel plants, paper mills, internal-combustion engines are the major sources of
these pollutants. Depending on the chemical nature and sources of contaminants, the
concentrations of VOCs in the resulting gas stream may vary from tens of ppb to
a few percent by volume. Therefore, not only the elimination of pollutants present
in trace amounts is important; even their detection in the environment can be a real
challenge.
During the past decades, widespread chemical contamination of lakes, streams
and groundwater, and consequently, soils; have been documented worldwide [9].
A large variety of compounds such as polyaromatic and heterocyclic compounds, pes-
ticides, pharmaceuticals (for example, antibiotics) have been detected. More recently,
other pharmaceuticals (nonsteroidal and steroidal anti-inflammatory drugs, antihy-
pertensive drugs, etc.) and personal care products (PPCPs) have been discovered in
various surface and ground waters, some of which have been linked to ecological
impacts even at trace concentrations [10, 11].
Besides, there are reports that certain synthetic and natural compounds could
mimic natural hormones in the endocrine systems of animals. These substances are
now collectively known as endocrine-disrupting compounds (EDCs), and have been
linked to a variety of adverse effects in both humans and wildlife. EDCs are released
into aquatic environments as a result of industrial, agricultural and sewage runoff.
Phenolic compounds, such as alkylphenols and bisphenol-A, many pesticides and
steroidal hormones (17α-estradiol, synthetic agent in contraceptives with estrogenic
activity of about tenfold that of female hormone 17β-estradiol, which is found in
wastewaters coming from pharmaceutical industries) are among the other compounds
recognized as hormone disruptors [12].
Particular problem is a frequent occurrence of phenol and phenolic compounds
(phenols) in rivers, what occasionally disturbs the production of drinkable water [13].
Phenol often originates from coal power plants or as a result of pesticides degradation.
In addition, large quantities of wastewater are generated during tobacco processing,
and they can be toxic due to the presence of alkaloids such as nicotine, the strongest
addictive agent from tobacco [14]. Nicotine and the majority of EDCs and PPCPs
are polar molecules - they possess acidic or basic functional groups. Consequently,
they are usually soluble in water. These properties, coupled with occurrence at trace
levels, create unique challenge for both analytical detection and removal processes.
All previously mentioned indicate that the production of drinkable water could be a
very complex problem.
10 Adsorption/Desorption of Simple Pollutants 387

Table 10.1 The systematization of possible environment pollutants

Eco system Pollutant(s) Comment
Atmosphere NOx Cause acidic rains. N2 O is
responsible for greenhouse
effect (global warming).
SOx Cause acidic rains.
Cox CO2 is responsible for
greenhouse effect.
VOCs.
Waters Polyaromatic compounds
(PAHs, among them - phenolic compounds …);
Heterocyclic compounds
(alcaloides, nicotine …);
Pesticides;
Heavy metals;
Pharmaceuticals and personal
care products (PPCPs);
Endocrine disrupting
compounds (EDCs).
Soils May contain all chemical
species that arrive in surface
and ground waters.

Table 10.1 summarizes previously stated facts and gives a systematization of

atmospheric, waters’ and soils’ pollutants.
Environment pollution control is presently one of major thrust areas of both sci-
entific research and technological development. Over the last few decades, concerns
about the potential ecological and health impacts of exposure to substances recog-
nized as atmospheric, water or soil pollutants have led to the establishment of new,
multi-stakeholder research and testing initiatives, expert groups, committees, data-
bases, etc. throughout the world. The combined efforts of scientists, governments and
communities as a whole gave noticeable results in this domain. However, emission
of pollutants coming from industrial and anthropogenic sources is a persistent prob-
lem. It has to be pointed out additionally that new pollutants in the environment can
be expected, due to the continuous development of new technologies. Evidently, the
pollution of environment presents a complex problem which has to be permanently
controlled.

10.1.1 Possible Abatement Procedures

Several main strategies have been established for the abatement of environment
pollution. In general, there are four classes of possibilities to decrease the presence
of pollutants in the environment:
388 V. Rakić

• reduction of pollutant emissions,

• reuse of pollutants (if possible),
• their recycling,
• treatment of waste(s) in order to obtain benign or less dangerous substances.
For the sake of environment remediation, the procedures that decrease pollutant
emissions or those that enable the elimination of pollutants already discharged into
the environment are still of the highest interest. For that purposes, physical (filtration,
adsorption), chemical (oxidation, catalytic conversions, ozonization in the presence
of a catalyst or without one) or biological methods are available. The choice of
particular method depends on the nature and the concentration of pollutant, and
on its physical state. All mentioned processes are permanently objects of further
development.
In the domain of atmospheric de-pollution, still the most important is to remove
several main classes of pollutants (CO2 , NOx , SOx and VOCs) from the gas-phase
effluents. The scientific efforts have been, and still are, directed towards several
classes of procedures that can be generalized in a following way: the usage of vari-
ous solid materials as adsorbents, chemical treatment of effluent gases either in het-
erogeneous catalytic processes or with ozone, the treatment of pollutant gases with
non-thermal plasmas, and plasma-assisted adsorption and/or catalytic processes.
For all processes that involve the adsorption step, such as physical processes of
separation or catalytic transformations, the usage of solid materials with optimised
activity as adsorbents and catalysts is necessary. Various solids, such as porous mate-
rials (zeolites—molecular sieves with hierarchical porosities and natural clays), acti-
vated carbons, mesoporous silica-based materials, pillared clays and metal oxides,
have shown the ability to act as adsorbents or as catalysts for the conversions of
previously mentioned atmospheric pollutants. Solid materials are also used for the
removal of pollutants that can be found in wastewaters. The possibilities to remove
polyaromatic hydrocarbons (PAHs) and heavy metal particles using the adsorptive
characteristics of activated carbon and porous materials from wastewaters have been
proven [15–17]. The same classes of solids are used for the elimination of organic
pollutants form wastewaters by heterogeneous catalytic oxidation processes; one of
the most important tasks is to eliminate phenolic compounds [13].
Evidently, the synthesis, characterizations and the investigations of adsorption or
catalytic capabilities of particular materials are necessary for enabling the complete
processes of pollutant eliminations. Besides, there is a necessity to find new or to
improve the existing methods for the de-pollution of all mentioned eco-systems.
It can be inferred that the permanent investigation of de-pollution procedures as
a whole is imposed to the researchers. Numerous scientific methods are available
for investigations of materials and for development of procedures applied in the
previously mentioned domains: from the methods that enable the investigation of
materials’ structure (based primarily on the diffraction of X-rays) or texture (such
as the low-temperature adsorption of different gases, or microscopies) up to the
application of different spectroscopies that can give information on both de-pollution
procedures and on the characteristics of applied adsorbents or catalysts.
10 Adsorption/Desorption of Simple Pollutants 389

This text will discuss the possibilities to use calorimetry, temperature-programmed

techniques, particularly, temperature-programmed desorption (TPD), and thermo-
analytical methods (such as thermogravimetry or differential scanning calorimetry)
in the fields of environment protection and remediation.

10.2 The Application of Calorimetry in Environment Protection

Calorimetry is the science of measuring the amount of heat which is involved in

some chemical or physical processes. Heat may be exchanged between some system
and the environment (generated in the case of exothermic or consumed in the case of
endothermic process), or dissipated by the sample. Since calorimetry’s advent in the
late eighteenth century, a large number of calorimetric techniques have been devel-
oped. Initially they were based on simple measurements of temperature, but more
recently, it became possible to collect data and maintain samples under conditions
that enable a wide range of applications [18].
Any process that results in heat being exchanged with the environment is a can-
didate for a calorimetric study; therefore, calorimetry has a very broad range of
applicability. The classification of calorimeters may be based on three different cri-
teria: the measuring principle (heat conduction, heat accumulation or heat exchange
calorimeters have been constructed), the mode of operation (isothermal, isoperibol or
adiabatic calorimeters exist) and the principle of construction (single or differential
measuring system is possible). Additionally, calorimeters may be classified as those
that enable measurements at constant pressure, or those working at constant volume
[18]. Besides, another classification of these apparatuses can be done according to
the amounts of samples used and the heat effects monitored. Microcalorimetry is
the calorimetry of small samples, specifically microgram samples, designed to work
in the microwatt range. Recently, a new generation of miniaturized calorimeters
has been developed with highly reduced dimensions compared to the state-of-the-
art microcalorimeters. Accordingly to the amounts of sample that can be used, this
improved version of calorimetry is denoted as nanocalorimetry [19].
The wide range of different constructions allow measurements in many different
fields: examples can be ranged from drug design in the pharmaceutical industry, to
quality control of processes in the chemical industry and the study of metabolic rates
in biological (people included) systems. Therefore, this methodology can be also very
useful in the domain of environment protection - pollutants detection and abatement.
There are two main ways in which calorimetry can be useful in this domain.
390 V. Rakić

10.2.1 Calorimeters Can be Applied for Direct Investigation of

Some Event that Includes Specific Pollutant(s)

In those cases, the heat of pollutants’ degradation or their chemical transformations

can be measured; the obtained results can be useful either in the process of monitoring
of dangerous substances, or for discovering the abatement possibilities. As it has
been already stated, the elimination of pollutants can be performed using physical
processes of adsorption, or catalytic procedures can be applied for their chemical
transformations. These processes may be monitored and characterized through the
measurement of the heats that they exchange with the environment.
It may be inferred generally that calorimeters are used to register the heat
evolved/consumed as a result of a contact between the adsorbent (reactant) and
adsorbate (the other reactant) which can be admitted from either gas or liquid phase.
The enthalpy of adsorption provide an insight into the mechanism of adsorption and
hence, into the mechanism of chemical transformation if it happens after the adsorp-
tion step. These values can be determined from the isosteric measurements, by the
application of Clausius-Clapeyron equation—isosteric heats of adsorption are deter-
mined from the variation of adsorption isotherms with temperature. Another way to
determine the enthalpy of adsorption is to perform direct calorimetric measurements
in order to obtain the desired values as –H = f (n a )T (n a —the adsorbed amount,
T—temperature). In the case of gaseous adsorbate, its adsorbed amount can be deter-
mined in several different ways: from the variations of gas pressure in a known
volume (volumetric measurements), from the variations of solid sample’s weight in
a static or continuous-flow lines (gravimetric measurements) or from the variations
of a mass spectrometer signal [20]. The insight into the available literature data gives
evidence that calorimetry linked to the volumetric lines is the most commonly used
method.
Being the methodology that reveals the heat exchanged as a result of any physical
or chemical event, calorimetry can help to obtain the information about the effects that
versatile pollutants provoke in the environment, in a number of different ways. It can
be stated that there is a growing application of calorimetric techniques that enables
the monitoring of pollutant’s presence as well as the investigation of effects that they
produce in the environment. Several examples of the application of calorimetry in
obtaining the informations concerning soils’, waters’ and atmosphere’ pollution will
be presented here.

10.2.1.1 Calorimetry Can be Applied to Assess and Compare the Toxic Effect
of Different Pollutants on the Chosen Ecosystem

As and example, the case of soil microbial community’ pollution will be presented.
Soil microorganisms form an essential component of most terrestrial ecosystems and
play a critical role in the environment due to their role in cycling mineral compounds
and in the decomposition of organic material. Environmental stress caused by the
10 Adsorption/Desorption of Simple Pollutants 391

presence of versatile chemicals generally decreases the diversity and activity of

soil microbial populations. Their biological functionality is mostly affected by the
presence of heavy metals or agrochemicals (such as pesticides) [21–25]. Heavy metal
contamination of soil is widespread due to metal processing industries, combustion
of wood, coal and mineral oil. Different pesticides are used on a vast and expanding
scale in modern agriculture, with the aim of eliminating undesirable weeds, insects,
and diseases. Although some of these compounds are directly applied to soil, the
great majority can reach plants and animals.
Soil microorganisms release heat when decomposing the substrate. It is well
known that the heat released closely correlates with the CO2 evolution. Therefore,
CO2 is considered as one of the most important indicator of soil microbial activity
[26]. Because it is independent on the organisms and reaction pathways [27], it
is enough only to know the initial and final energetic states of investigated soil,
which means that application of calorimetry can provide specific information on
soil microbial activity. Currently, microcalorimetry is sensitive enough to detect
very low values of heat and it is applied in studies of soil microbial activity
[28–30]. In calorimetric measurements, the microbial activity which is affected by the
pollutants’ presence produces typical power vs. time curves. The variation in thermal
effects that occurred after programmed applications of particular pollutant to the soil
containing microorganisms can be monitored as important information. Apart from
these values, microcalorimetry permits to reach several parameters simultaneously,
such as microbial growth rate constant, the heat yield of the microbial growth and
the metabolic enthalpy.
In order to study the effects of contamination by agrochemicals, selected pesticide
can be applied to soil containing a degradable source of carbon. Then, the variation
in thermal effects that occurred after programmed applications of this pesticide to
an appropriate, nutrient-containing system can be monitored as important and valu-
able information. For example, the pesticide denoted as picloram is toxic for the
microorganisms residing in the soil and causes the inhibition of their metabolism.
Consequently, the progressive addition of pesticide picloram affects the metabolism
of the microflora in the soil, what can be recognized through the changes in the pro-
files of power–time curves, as presented in Fig. 10.1. The profile of each recorded
curve reflects the thermal effect generated by the activity of the microorganisms
present in the system. The effects caused by distinct doses of picloram on the total
thermal effect (Q) evolved by microorganisms are evident.
Similar results are obtained in the cases of soil contamination by other agrochem-
icals. Figure 10.2 presents the variation of enthalpy with time, obtained for samples
of Red Latosol soil that contained certain ammount of glucose as degradable source
of carbon and variable amounts of herbicide diquat. These results are presented as
the integral heats vs. time. The curves show that the increase of the applied agro-
chemicals’ doses produces the decreases in activity generated by microbes.
The toxic effects of heavy metals on soil microorganisms arrive mainly from
the interaction of metals with proteins (enzymes) what provoke the inhibition of
metabolic processes. These effects of heavy metals on natural soil microbial com-
munities are complex and difficult to characterize [21]. However, similarly to the
392 V. Rakić

Fig. 10.1 Power–time curves of the microbial metabolism of 6.0 mg of glucose plus 6.0 mg of
ammonium sulfate at 34.8 % humidity in a sample of 1.50 g red Latosol soil, with different doses
of picloram: (A) 0.00, (B) 1.00, (C) 2.50, (D) 5.00, (E) 10.00 mg (g soil) [24]

Fig. 10.2 Variation of enthalpy with time for samples with 1.50 g of Red Latosol soil, 6.0 mg of
glucose, 6.0 mg of ammonium sulfate with 35 % of moisture content, control (A), and variable
amounts of diquat: 1.30 (B); 2.70 (C), 5.30 (D), 6.70 (E), and 8.00 (F) mg [25]
10 Adsorption/Desorption of Simple Pollutants 393

Fig. 10.3 Power–time curves for soil amended with glucose and ammonium sulphate and different
selected heavy metals (1.0 mg g−1 As, Cu, Pb, Zn, Cd, Co and 0.01 mg g−1 Cr) [21]

Fig. 10.4 The thermogenic curves of the growth of pure culture of Bacillus subtilis for the different
concentrations of Cd2+ at 28–30◦ C [31]

pollution by different agrochemicals, the inhibition of metabolic processes that arrive

from the presence of heavy metals can be revealed using microcalorimetric mea-
surements. Figure 10.3 presents power vs. time curves, obtained for one “clean” and
several polluted soils. Different thermogenic curves obtained for different metals
clearly indicate the influence of heavy metal character on microbial activity. How-
ever, the influence of metal content is also obvious, as revealed by microcalorimetric
investigation of toxic effects of cadmium on pure microbes (see Fig. 10.4) [31]. From
the results presented in Figs. 10.1– 10.4, and being aware of all other results reported
in the literature, it can be inferred that microcalorimetry is a powerful tool to provide
qualitative and quantitative data on microbial activities in soils.
394 V. Rakić

10.2.1.2 Calorimetry Can Be Applied in the Investigation of Degradation and

Biogedradation of Different Pollutants

Calorimetry can be applied in the investigation of degradation and biogedradation

of different pollutants such as polycyclic aromatic hydrocarbons (PAH’s) or other
persistent organic pollutants. Due to their high hydrophobicity, PAHs tend to interact
with non-aqueous phases and natural organic matter and, as a consequence, are poorly
bioavailable for microbial degradation. Common methods for the determination of
biodegradation parameters of the polycyclic aromatic hydrocarbon (PAH) either rely
on continuous assessment of oxygen consumption and/or CO2 -production or on the
chemical analysis of the PAH-removal rates. Calorimetry appears to be a convenient
and powerful alternative of sufficient sensitivity that overcomes some of the prob-
lems, since the rates of PAH biodegradation by appropriate microorganisms can be
directly followed as absolute values of the reaction heat production in the bioreactor.
For example, in the case of biodegradation of anthracene, it has been shown that the
achievable thermal sensitivity between few nanowatts and 1 μW corresponds to a
degradation rate of few μgL −1 h−1 and is sufficient to follow the biodegradation of
ultra traces in the magnitude of few nanograms [29].
As it has been already mentioned, “isothermal microcalorimeters” are those
calorimeters designed to work in the microwatt range under essentially isothermal
conditions. Isothermal titration microcalorimetry (ITμC) is designed to connect
extremely sensitive thermal measurement equipment (approx. 20–100 nanowatts)
with an automatic syringe able to add reactants in successive injections to the solu-
tion with a precision of few nanoliters [32, 33]. Each injection produces specific
heat effect, as shown in Fig. 10.5. The determination of heats evolved as a result of
interaction between molecules is a main application of this variation of calorime-
try. Consequently, isothermal titration calorimetry is a suitable method for studying
degradations and biodegradations of versatile pollutants.

10.2.1.3 Calorimetry Can Be Applied in the Development of Different

Bio-Sensors

The presence of versatile pollutants in different mediums (soils, water, and

atmosphere) can be noticed by the application of microcalorimetry. For example,
lichen tissue can be used to detect the presence of benzene, because it can metabolise
benzene derivatives in high quantities. Therefore, lichen tissue can be used to detect
traces of benzene in aqueous solutions. It has been shown that by measuring the
enthalpy changes the interaction that occurs between lichen and 2-Chlorophenol can
be recognized and quantified [34].
10 Adsorption/Desorption of Simple Pollutants 395

Fig. 10.5 Typical data obtained from an isothermal titration microcalorimeter [33]

10.2.1.4 Calorimetry Can Be Applied to Obtain the Data Concerning the

Adsorption of Pollutants, From Gas or Liquid Phase

As mentioned previously, calorimetry linked to the volumetric technique is the most

commonly used. The adsorbate (admitted either from gas or liquid phase) can be
added in desired increments (doses) and the heat can be measured after the equili-
bration is reached, for each dose. In such a way, differential heats, defined as Qdiff
= ∂Qint/∂na (the ratio of the amount of heat evolved for each dose to the number
of moles adsorbed, or molar adsorption heat for each dose of adsorbate), become
available. The results are most commonly presented as differential heat variations in
relation to the adsorbed amount; in the form of histograms, or by a continuous curve
connecting the centers of the steps (increments or doses). In this variation of calori-
metric technique, the calorimeter is connected to a gas handling and a volumetric
system equipped with manometers for precision pressure measurement. In the case
of liquid adsorbates, the calorimeter is connected with a stirring system and a system
for the introduction of liquids (a programmable syringe pump). Successive small
doses of gas (or liquid) are introduced until appropriate final equilibrium pressure
(or saturation of adsorbent with liquid adsorbate) is achieved [20, 35, 36].
The obtained differential heats of adsorption allow the estimation of adsorption
affinity toward the gas-phase pollutant; the amounts adsorbed are also available. This
possibility is illustrated by example presented in Fig. 10.6: the values of differential
heats and the amounts adsorbed reveal that alumina-supported tin oxide expresses
396 V. Rakić

Fig. 10.6 Differential heats vs. adsorbate coverage obtained for alumina-supported tin oxide [37]

CO adsorption
140 CuZ
120 FeZ
MnZ
Qdif, kJ/mol

100
CoZ
80
60
40
20
0
0 50 100 150 200
V, µmol/g

Fig. 10.7 Differential heats vs. adsorbate coverage obtained for CO adsorption on different cation
exchanged ZSM-5 zeolites [38]

significant affinity toward ammonia adsorption, moderate affinity toward NO2 and
SO2 adsorption, and does not express the affinity toward NO adsorption [37].
The values of differential heats of adsorption and the profile that they form plotted
versus the adsorbed amounts, allow to chose an efficient adsorbent (among several
investigated) for the adsorption of particular gas. Figure 10.7 illustrates the abilities
of different cation-exchanged ZSM-5 zeolites toward CO adsorption. Evidently, the
values of differential heats clearly indicated that Cu(II)-containing ZSM-5 zeolite
expresses the highest possibility to adsorb carbon monoxide, among investigated
ones.
The values of differential heats of adsorption enable the knowledge about the
pollutant which will be preferably adsorbed on a given solid material. The example
presented in Fig. 10.8 show that ZSM-5 zeolite which contain both Cu(II) and Fe(III)
as charge-balancing cations is better adsorbent for CO then for N2 O (both the amount
adsorbed up to equilibrium and the values of differential heats are higher for CO then
for N2 O adsorption).
10 Adsorption/Desorption of Simple Pollutants 397

150

N2O
adsorption

Qdif, kJ/mol
100
CO
adsorption

50

0
0 50 100 150 200
V, µmol/g

Fig. 10.8 Differential heats vs. adsorbate coverage obtained for CO and N2 O adsorption on Cu,
FeZSM-5 zeolite [39]

In the case of water pollution, the estimation of adsorption affinity of poten-

tial solid adsorbent toward the specific pollutant can be done using the so-called
liquid microcalorimetry. The instruments used for this purpose are differential heat
flow microcalorimeters modified to allow continuous stirring of liquid samples. The
adsorbate is added to both sample and reference cells simultaneously using a pro-
grammable twin syringe pump, linked to the calorimeter. The heat evolved as a
result of adsorption can be obtained by integration of the area under the calorimeter
signal, for each particular injection (dose). The output of typical microcalorimetric
experiment of this type is shown in Fig. 10.9.

Fig. 10.9 The output of typical liquid-phase microcalorimetric experiment. Each pulse injection
of dosed amount of particular solution gives as a result the exothermic signal that formed a specific
peak related to the heat of adsorption [14]
398 V. Rakić

Fig. 10.10 Differential heats vs. adsorbate coverage obtained for nicotine adsorption from the
aqueous solution on activated carbon and β-zeolites [14]

Figure 10.10 shows the profiles of differential heats revealed as a function of the
amounts of aqueous solution of nicotine adsorbed on different solids. Based on the
values and profile of differential heats, it can be concluded that adsorption capabilities
of β-zeolites are comparable with that of activated carbon, solid known to be effective
adsorbent for remediation of wastewaters. There is evidence in the literature on
the studies of other water pollutants adsorption, such as phenol, aldehydes and the
ketones, done by the application of microcalorimetry [40].
Evidently, calorimetry can bring noticeable and valuable data concerning soil,
water and atmosphere pollution. Having in mind numerous modifications of this
technique that have been done in the past decades; it can be expected that new
versions of apparatuses and related equipments will give further improvement in
the possibilities to detect small amount of heats that are related to the processes
significant in both environment pollution and protection.

10.2.2 Calorimeters Can Be Applied for the Characterization

of Solid Materials

If pollutants’ abatement is performed using solid adsorbent or catalysts, the charac-

terization of these substances is necessary. For that purpose, numerous methods have
to be used: those enabling information about structure; the others that provide data
concerning texture and those that reveal the existence of active sites for adsorption
(conversion) of particular pollutants. Evidently, multi-technique approach has to be
applied in order to perform full characterization of materials applied for pollutants
abatement. Among the other techniques, calorimetry may be applied for the charac-
10 Adsorption/Desorption of Simple Pollutants 399

terization of materials that can be used for capturing of pollutants (as adsorbents) or
as catalysts (for conversion of pollutants) [20, 35, 36].
It is important to point out that acidic/basic character may be decisive for the
reactivity of a solid material in the reactions important for pollutants abatement.
The data concerning acidity/basicity can be obtained using several methods, such
as IR, XPS and Raman spectroscopies; temperature-programmed desorption, and
calorimetry. Among them, calorimetry gives multi-indicative data: the adsorption
of successive injections (doses) of so-called “probe” molecules happens onto the
sample’s surface while it is kept at a constant temperature and a heat-flow detector
notice the amount of heat transferred per time.
An experiment of adsorption from the gas-phase, performed in microcalorimeter
coupled with volumetric line can give: a profile of Qdi f versus the amount adsorbed,
integral heats of adsorption, adsorption isotherms (adsorbed amounts vs. equilibrium
pressure) and irreversibly absorbed amount of a chemisorbed gas; the same stands
for the adsorption from the liquid-phase, where the adsorbate (titrant) is added to
both sample and reference cells simultaneously. The profile of differential heats
versus the uptake of “probe” gives the data concerning the amount, strength and
distribution of the active sites. Besides, the values of initial heats of adsorption char-
acterize the strongest sites active in adsorption process. For the sake of acidic/basic
characterization of solids’ surface, the most commonly used gas-phase probes are
ammonia, pyridine or some amines for the interaction with acidic sites. SO2 and CO2
are the probes used to notice and characterize the basic sites. In microporous solids,
the accessibility of active sites is not the same for the molecules of different sizes.
Therefore, many different probes can be applied to study acidity or basicity of same
solid materials; this approach brings additional information. For example, acidity of
zeolites can be characterized by adsorption of ammonia, but also by adsorption of
pyridine (from the gas phase) and aniline (from the liquid phase) [20–22]. Liquid
microcalorimetry can be also used for the determination of acidic character of solid
adsorbent; the common liquid-phase probe is aniline dissolved in n-decane [40].
The obtained information about the acidic/basic character can be correlated with
the adsorption possibilities. The possibility of some solid material to adsorb (or
modify) the atmospheric or water pollutant depends very often on these features.
Therefore, there are literature sources that report the correlations between acid-
ity/basicity of different materials and their abilities in pollutants’ abatement. For
example, balanced acidic/basic properties are indispensable for an efficient conver-
sion of NOx, that is, high activity and selectivity, over a broad temperature range.
In case of alumina-supported indium oxide catalysts, microcalorimetry of ammonia
and SO2 adsorption helped to find effective formulations [41].
Finally, it can be summarized that, in the domain of environment protection
microcalorimetry gives possibilities to:
1. correlate the acidic/basic character with the adsorption capacities and thus, to
find the appropriate formulations of adsorbents/catalysts;
2. determine the adsorption capacities toward different gas-phase or liquid phase
pollutants, and thus:
400 V. Rakić

3. choose an efficient adsorbent, among several investigated, for the adsorption of:
– particular gas-phase pollutant,
– particular water pollutant;
4. recognize the pollutant which will be preferably adsorbed on a given solid mate-
rial.

10.3 The Application of Temperature-Programmed Techniques

in Environment Protection

Temperature-programmed desorption (TPD) can be described as the measurement

of the rate of desorption of preadsorbed molecules as a function of temperature. It
involves heating of sample while contained in a sample holder—as the temperature
rises, certain absorbed species will have enough energy to desorb and will be detected
simultaneously by means of specific detector (for example, mass spectrometer). In
TPD experiments, temperature increases linearly and the concentration of desorbed
gas is recorded as a function of temperature. Therefore, TPD profile is traced as
desorption rate versus time (or temperature). If detector signal is properly calibrated,
concentration of desorbed gas can be plotted as a function of temperature.
TPD provides the estimation (or precise determination in the case of quantified
detector response) of the amount of adsorbed species; the enthalpy of adsorption (i.e.
the strength of binding to the surface) and the kinetics of desorption. The tempera-
tures of peak maxima in the TPD profile are influenced by the strength of interaction
adsorbate molecule—active site, while the area under this profile is affected by
the population of active sites. Therefore, this method is an excellent tool for deter-
mination of surface coverage and for studying the adsorption states of adsorbate
molecules, their binding energies and surface concentration, and desorption kinet-
ics [20]. Two other temperature-programmed techniques: temperature-programmed
oxidation (TPO) and temperature-programmed reduction (TPR) provide the red-ox
features of solid materials: the amount of reducible/oxidable species, their types and
kinetic parameters concerning reduction-oxidation [42].
One typical TPD experiment is designed in a way to enable the pre-treatment of the
sample, in situ; the admission of specific adsorbate (probe) up to some specific surface
coverage or up to the saturation; and subsequently, desorption which is performed
in a temperature-controlled regime. Many different chemical species can be used as
probes: if a chosen probe can titrate acid or basic sites at the surface, TPD can be used
for the characterization of acidity/basicity of some adsorbent. In fact, temperature-
programmed desorption and adsorption calorimetry are most commonly used for
the study of acid/base properties of solid materials [20, 35, 36]. The same probe
molecules that are used for adsorption calorimetry experiments are applicable in the
case of TPD; while the investigation of acidic/basic character of solids is perhaps
10 Adsorption/Desorption of Simple Pollutants 401

the most common application of this technique. TPD method has been used to study
the acidity/basicity of wide variety of materials [20].
In the domain of environment protection, temperature-programmed techniques
provide: (i) the characterization of important features of materials that should be
employed for pollutants’ abatement (as adsorbents or catalysts), (ii) the detection
of environment contamination, (iii) the estimation of the possibility to remove the
pollutant.
Often, there is a correlation between acidic/basic or red-ox properties of some
solid material and its ability to adsorb or catalytically convert certain pollutant. For
example, acidity of different zeolites and mesoporous materials is important for
their ability to adsorb aldehydes and ketones (from the gas phase) or phenol and
nicotine (from the aqueous phase) [40, 43]. Red-ox properties are often correlated
with catalysts’ efficiency; for example, red-ox features of ceria-based mixed oxides
are of importance for their ability to catalyse direct conversion of methane to synthesis
gas, and they can be affected by incorporation of ZrO2 [44]. The oxidability of
mixed oxides is an important feature, which determines the possibility of their use
as catalysts for combustion reactions; and it can be also determined using TPR-TPO
techniques [45].
Apart from investigation of adsorption-desorption of probe molecules used to
estimate acidic/basic character, TPD can be applied to study all relevant parameters
concerning the adsorption of any other adsorbate. Among other possibilities, TPD
can be applied to get the information about adsorption/desorption of many different
pollutants of soils, waters and atmosphere. In that way, the contamination of environ-
ment by different pollutants can be detected and quantified using TPD procedures;
besides, the possibility to remove it can be estimated. Figure 10.11a presents the des-
orption profiles of sulphur dioxide from activated carbon [46]; while TPD profiles
presented in Fig. 10.11b prove the pollution of investigated soil by agrochemical
phenantrene [47].

Fig. 10.11 (a) The profiles of temperature-programmed desorption of SO2 (m/e = 64) as
atmospheric pollutant from activated carbon [46]; and (b) the profiles of phenantrene (soil pol-
lutant) released during temperature programmed heating [47]
402 V. Rakić

Fig. 10.12 Temperature programmed techniques allow to monitor the adsorption and chemical
transformation of pollutants: (a) TPD spectra of adsorbed tert-butyl alcohol on TiO2 in He stream;
(b) TPD spectra after 300 s of photocatalytic oxidation of tert-butyl alcohol on TiO2 (right) [48]

It is important to point out that the products of catalytic conversion can be moni-
tored in temperature-programmed regime, if appropriate detector is chosen. Hence,
temperature-programmed decomposition of some pollutant can be studied using the
same methodology. Figure 10.12 presents TPD profiles of tert-butyl alcohol desorp-
tion and catalytic conversion on TiO2 [48].
Evidently, the possibility to make a choice of appropriate detection technique
makes possible to adapt temperature programmed techniques for many different
applications. From all presented examples, it becomes evident that these methods
can be of big help in the field of investigations related to environment protection.

10.4 The Application of Thermo-analytical Methods in

Environment Protection

The possibility to use any other technique which belongs to the group of thermo-
analytical methods for detection of pollutants or for the development of abatement
procedures can be also discussed. Thermal analysis is a group of techniques in which
one (or more) property of the sample is studied while it is subjected to a controlled
temperature programme. Although it can take many different forms, most often, the
sample is subjected to a constant heating (or cooling) rate (β). Here, the possibility to
use thermogravimetry (TG), differential scanning calorimetry (DSC) and differential
thermal analysis (DTA) in the field of environment protection will be exposed.
Thermogravimetry (TG) is defined by ICTAC (International Confederation for
Thermal Analysis and Calorimetry) as a technique in which the mass change of a
substance is measured as a function of temperature whilst the substance is subjected
to a controlled temperature programme. In TG experiment, the sample is placed in
a crucible which is positioned in a furnace; balance detects the mass loss, while
the results are presented as a plot of mass against temperature (T) or time (t). The
alternative presentation is the derivative of the original experimental curve: dm/dt
(or dm/dT) plotted against temperature T or time t.
10 Adsorption/Desorption of Simple Pollutants 403

Mass loss is only noticed if a volatile component is lost in the investigated process.
Therefore, the processes which happen without mass loss can not be investigated
by TG. For these kinds of processes, other thermal techniques such as DTA or
DSC are available. The term “differential” emphasises an important feature of these
techniques—two identical measuring sensors are used: one for the sample and the
other for the reference; the signal from the instrument depends on the difference
between the responses of the two sensors.
The practical distinction between DTA and DSC is in the nature of the signal
obtained from the equipment. In the case of DTA it is proportional to the temperature
difference established between the sample and an inert reference when both are
subjected to the same temperature program: T = TS − T R (subscripts S and
R indicate the sample and reference respectively). DSC signal can be regarded as
proportional to the difference in thermal power between the sample and reference:
dq/dt. The results from DTA and DSC experiments are displayed as a thermal
analysis curve in which the instrument signal is plotted against temperature (usually
the sample temperature) or time. Combining different heating or cooling rates with
isothermal periods is also possible and often employed.
Often, TG is coupled with DTA or DSC technique. In case of all these techniques,
the investigated processes are usually performed in a controlled atmosphere. In cases
where gaseous products are evolved, their identification can be performed if a system
is coupled with a detector (e.g. mass spectrometer - MS, infrared spectrophotometer
with Fourier transformation—FTIR or gas chromatograph—GC). So, the combined
techniques such as: TG-FTIR, TG/DSC-FTIR, DSC-FTIR, DSC-MS, DSC-GC or
TG/DSC-MS exist, and are employed in versatile investigations.
Similarly to previously exposed possibilities of microcalorimetry and temperature-
programmed techniques, TG, DSC, DTA and related coupled techniques can be also
helpful in the domain of environment protection. They enable to: (i) obtain data
useful in the preparation of materials applied in the abatement of pollutants; (ii)
investigate the processes of pollutant removal or degradation; (iii) identify the nature
of the decomposition products; (iv) monitor the processes that indicate the presence
of pollutants and to investigate their effects. Here, several typical applications will
be shown.
In the processes of preparation of solid materials that would be used as adsorbents
or catalysts, TG is usually applied to determine the temperature of calcinations of
crude samples. Similarly TG coupled with DSC or DTA is often applied to reveal the
stability of solid catalysts to heat. In addition, TG can be used to get the adsorption
capacities of some adsorbents toward specific gas or liquid phase pollutants.
By observing the difference in heat flow between the sample and reference, dif-
ferential scanning calorimeter is able to measure the amount of heat adsorbed or
released during some event. The enthalpy changes found as a result of certain process
are always significant information which elucidates its mechanism. In the field of
environment protection, the enthalpy changes that appear as a result of pollutant
adsorption, degradation or interaction with some living system, can be monitored.
Therefore, DSC can play an important role in elucidation of abatement mechanisms
for particular pollutants.
404 V. Rakić

Fig. 10.13 Thermal analysis of the dried sewage sludge. (a) Thermogravimetric analysis of dried
seawage sludge; (b) total ion current (TIC) curve of the evolved gas phase in the TG-MS analysis
of the dried sewage sludge; (c) gas chromatographic elutions of the gas phase evolved during the
pyrolysis of the sewage sludge [49]

Besides, TG or TG/DSC coupled with appropriate detection systems such are

MS, GC or FTIR are often used to investigate the processes of pollutant removal or
degradation. As an example, Fig. 10.13 presents the results obtained using coupled
TG-MS and TG-GC-MS systems for studying the products of an urban plant sewage
sludge pyrolisis carried out under helium atmosphere. Gas chromatographic analyses
of the gas released allow the identification of various chemical species.
TG and DTG curves recorded in the TG-MS measurement (Fig. 10.13a) show
the profile of mass loss along temperature increase. The profile of total ion cur-
rent (TIC) curve of the evolved gas phase in the TG-MS analysis of dried sewage
sludge (Fig. 10.13b) indicate the temperature regions where the majority of prod-
ucts appear; while isothermally performed mass spectrometric analysis gives precise
data about mass fragmentation that appear as a result of pyrolisis (Fig. 10.13 b, in the
insets mass spectra recorded corresponding to selected pyrolysis temperatures are
presented). Finally, a detailed identification of gaseous species, evolved during the
thermal treatment up to 1000 ◦ C, has been enabled by gas chromatographic analysis
(Fig. 10.13c) [49].
10 Adsorption/Desorption of Simple Pollutants 405

DSC or DSC coupled with appropriate detector can be useful in investigation of

effects that some particular pollutant can provoke on humans, animals or plants. Espe-
cially interesting is the case of persistent pollutants and effects that they can cause.
For example, perfluorooctanesulfonic acid (which belongs to the group of perfluori-
nated surfactants - PFOS) is a persistent pollutant of the environment, which extreme
inertness towards chemical and biological degradation, as well as its hydrophobic
character, contributes to their worldwide distribution in the environment and their
accumulation in the food chain. Recent bio-monitoring studies have shown a wide-
spread prevalence of PFOS in humans, thus giving raise to human health concerns.
One example of DSC application in this field is the investigation of interaction of
perfluorooctane-1-sulfonic acid (PFOS) with biological systems. It has been shown
that DSC can be useful in studying the changes in the structure of lipid bi-layers
caused by pollutant presence [50]. Also, DSC can be employed to notice the enthalpy
changes that arise as a result of altered protein denaturation in polluted seawater. For
that purpose, the denaturation of proteins in mussels Mytilus galloprovincialis col-
lected from polluted and non-polluted sites can be compared. In that way, DSC helps
finding a reliable biomarker of seawater pollution [51].

10.5 Conclusion

It can be inferred that all discussed techniques: calorimetry, temperature-programmed

techniques, as well as TG, DSC, DTA and coupled techniques, have important place
in the domain of environment science: in characterisation of materials used for the
abatement of pollutants; in investigation of processes of pollutant removal or degra-
dation; and generally in monitoring the processes that include numerous environment
pollutants. Having in mind the versatility of possible atmospheric, water or soil pol-
lutants, we can estimate that these techniques will be further developed and adjusted
for the applications in domain of environment protection.

References

1. P.G. Robertson, Abatement of nitrous oxide, methane, and the other non-CO2 greenhouse gases:
the need for a systems approach, in Global Carbon Cycle, Scientific Committee on Problem
of the Environment - SCOPE 62, ed. by C.B. Field, M.R. Raupach (Island Press, Washington,
DC, 2004)
2. R.G. Prinn, Non-CO2 greenhouse gases, in Global Carbon Cycle, Scientific Committee on
Problem of the Environment - SCOPE 62, ed. by C.B. Field, M.R. Raupach (Island Press,
Washington, DC, 2004)
3. M. Dalla Valle, E. Codato, A. Marcomini, Climate change influence on POPs distribution and
fate: a case study. Chemosphere 67, 1287–1295 (2007). doi:10.1016/j.chemosphere.2006.12.
028
406 V. Rakić

4. P.F. Verdes, Global warming is driven by anthropogenic emissions: a time series analy-
sis approach. Phys. Rev. Lett. 99, 048501-1–048501-4 (2007). doi:10.1103/PhysRevLett.99.
048501
5. J.C. Kramlich, W.P. Linak, Nitrous oxide behavior in the atmosphere, and in combus-
tion and industrial systems. Prog. Energy. Combust 20, 149–202 (1994). doi:10.1016/0360-
1285(94)90009-4
6. S. Seitzinger, C. Kroeze, R. Styles, Global distribution of N2 O emissions from aquatic systems:
natural emissions and anthropogenic effects. Chemosphere Glob. Change Sci. 2, 267–279
(2000). doi:10.1016/S1465-9972(00)00015-5
7. Y. Kamata, A. Matsunami, K. Kitagawa, N. Arai, FFT analysis of atmospheric trace concentra-
tion of N2 O continuously monitored by gas chromatography and cross-correlation to climate
parameters. Microchem. J. 71, 83–93 (2002). doi:10.1016/S0026-265X(01)00140-0
8. D.A. Sarigiannis, S.P. Karakitsios, A. Gotti, I.L. Liakos, A. Katsoyiannis, Exposure to major
volatile organic compounds and carbonyls in European indoor environments and associated
health risk. Environ. Int. 37, 743–765 (2011). doi:10.1016/j.envint.2011.01.005
9. D.J. Lapworth, N. Baran, M.E. Stuart, R.S. Ward, Emerging organic contaminants in ground-
water: a review of sources, fate and occurrence. Environ. Pollut. 163, 287–303 (2012). doi:10.
1016/j.envpol.2011.12.034
10. S.T. Glassmeyer, D.W. Kolpin, E.T. Furlong, M.J. Focazio, Environmental presence and persis-
tence of pharmaceuticals: an overview, in Fate of Pharmaceuticals in the Environment and in
Water Treatment Systems, ed. by D.S. Aga (CRC Press, Taylor & Francis Group, Boca Raton,
2008)
11. S.K. Khetan, T.J. Collins, Human pharmaceuticals in the aquatic environment: a challenge to
green chemistry. Chem. Rev. 107, 2319–2364 (2007). doi:10.1021/cr020441w
12. J.D. Meeker, Exposure to environmental endocrine disrupting compounds and men’s health.
Maturitas 66, 236–241 (2010). doi:10.1016/j.maturitas.2010.03.001
13. G. Busca, S. Berardinelli, C. Resini, L. Arrighi, Technologies for the removal of phenol from
fluid streams: a short review of recent developments. J. Hazard. Mater. 160, 265–288 (2008).
doi:10.1016/j.jhazmat.2008.03.045
14. V. Rakic, L.J. Damjanovic, The adsorption of nicotine from aqueous solutions on different
zeolite structures. Water. Res. 44, 2047–2057 (2010). doi:10.1016/j.watres.2009.12.019
15. C.Y. Yin, M.K. Aroua, W.M.A.W. Daud, Review of modifications of activated carbon for
enhancing contaminant uptakes from aqueous solutions. Sep. Purif. Technol. 52, 403–415
(2007). doi:10.1016/j.seppur.2006.06.009
16. K.G. Bhattacharyya, S.S. Gupta, Adsorption of a few heavy metals on natural and modified
kaolinite and montmorillonite: a review. Adv. Colloid Interfac 140, 114–131 (2008). doi:10.
1016/j.cis.2007.12.008
17. N. Rajic, D.J. Stojakovic, Removal of aqueous manganese using the natural zeolitic tuff from
the Vranjska Banja deposit in Serbia. J. Hazard. Mater. 172, 1450–1457 (2009). doi:10.1016/
j.jhazmat.2009.08.011
18. R.J. Willson, Calorimetry, in Principles of thermal analysis and calorimetry, ed. by P.J. Haines
(The Royal Society of Chemistry, Cambridge, 2002)
19. G.A. Urban, T. Weiss, Hydrogels for biosensors, in Hydrogel Sensors and Actuators, Springer
Series on Chemical Sensors and Biosensors Volume 6, ed. by G. Gerlach, K.F. Arndt (Springer-
Verlag, Berlin, 2010)
20. Lj Damjanovic, A. Auroux, Determination of acid/base properties by temperature programmed
desorption (TPD) and adsorption calorimetry, in Zeolite Chemistry and Catalysis: An Integrated
Approach and Tutorial, ed. by A.W. Chester, E.G. Derouane (Springer, Berlin, 2009)
21. F. Wang, J. Yao, Y. Si, H. Chen, M. Russel, K. Chen, Y. Qian, G. Zaray, E. Bramanti, Short-
time effect of heavy metals upon microbial community activity. J. Hazard. Mater. 173, 510–516
(2010). doi:10.1016/j.jhazmat.2009.08.114
22. A.G.S. Prado, C. Airoldi, The effect of the herbicide diuron on soil microbial activity. Pest.
Manag. Sci. 57, 640–644 (2001). doi:10.1002/ps.321
10 Adsorption/Desorption of Simple Pollutants 407

23. K.E. Giller, E. Witter, S.P. McGrath, Toxicity of heavy metals to microorganisms and microbial
processes in agricultural soils: a review. Soil Biol. Biochem. 30, 1389–1414 (1998). doi:10.
1016/S0038-0717(97)00270-8
24. A.G.S. Prado, C. Airoldi, Toxic effect caused on microflora of soil by pesticide picloram
application. J. Environ. Monit. 3, 394–397 (2001). doi:10.1039/b103872a
25. S.A.M. Critter, C. Airoldi, Application of calorimetry to microbial biodegradation studies of
agrochemicals in oxisols. J. Environ. Qual. 30, 954–959 (2001)
26. A. Tancho, R. Merckx, F. Schoovaerts, K. Vlassak, Relation between substrate induced respi-
ration and heat loss from soil samples amended with various contaminants. Thermochim. Acta
251, 21–28 (1995). doi:10.1016/0040-6031(94)02044-O
27. A.G.C. Prado, C. Airoldi, Effect of the pesticide 2, 4-D on microbial activity of the soil
monitored by microcalorimetry. Thermochim. Acta 349, 17–22 (2000). doi:10.1016/S0040-
6031(99)00492-X
28. N. Barros, S. Feijóo, S. Fernández, Microcalorimetric determination of the cell specific heat
rate in soils: relationship with the soil microbial population and biophysic significance. Ther-
mochim. Acta 406, 161–170 (2003). doi:10.1016/S0040-6031(03)00255-7
29. N. Barros, S. Feijóo, S. Fernández, J.A. Simoni, C. Airoldi, Application of the metabolic
enthalpy change in studies of soil microbial activity. Thermochim. Acta 356, 1–7 (2000).
doi:10.1016/S0040-6031(00)00495-0
30. N. Barros, M. Gallego, S. Feijóo, Sensitivity of calorimetric indicators of soil microbial activity.
Thermochim. Acta 458, 18–22 (2007). doi:10.1016/j.tca.2006.12.020
31. H. Chen, J. Yao, Y. Zhou, H. Chen, F. Wang, N. Gai, R. Zhuang, B. Ceccanti, Th Maskow, G.
Zaray, The toxic effect of cadmium on pure microbes using a microcalorimetric method and a
biosensor technique. J. Environ. Sci. Health Part A 43, 1639–1649 (2008)
32. F. Buchholz, L.Y. Wick, H. Harms, Th Maskow, The kinetics of polycyclic aromatic hydrocar-
bon (PAH) biodegradation assessed by isothermal titration calorimetry (ITC). Thermochim.
Acta. 458, 47–53 (2007). doi:10.1016/j.tca.2007.01.028
33. A.C. Dos Santos Pires, N. de Fátima Ferreira Soares , L.H.M. da Silva, N.J. de Andrade, Silveira
MFA, A.F. de Carvalho, Polydiacetylene as a biosensor: fundamentals and applications in the
food industry. Food Bioprocess Tech. 3, 172–181 (2010). doi:10.1007/s11947-008-0171-x
34. M.L. Antonelli, L. Campanella, P. Ercole, Lichen-based biosensor for the determination of ben-
zene and 2-chlorophenol: microcalorimetric and amperometric investigations. Anal. Bioanal.
Chem. 381, 1041–1048 (2005). doi:10.1007/s00216-004-3014-2
35. V. Solinas, I. Ferino, Microcalorimetric characterisation of acid-basic catalysts. Catal. Today
41, 179–189 (1998). doi:10.1016/S0920-5861(98)00048-0
36. A. Auroux, Microcalorimetry methods to study the acidity and reactivity of zeolites, pil-
lared clays and mesoporous materials. Top. Catal. 19, 205–213 (2002). doi:10.1023/A:
1015367708955
37. B. Gergely, A. Auroux, Calorimetric study of the adsorption of air pollutants on alumina-
supported tin and gallium oxides. Res. Chem. Intermed. 25, 13–24 (1999)
38. V. Rakic, V. Rac, A. Auroux, unpublished results
39. V. Rakic, V. Rac, V. Dondur, A. Auroux, Competitive adsorption of N2 O and CO on CuZSM-
5, FeZSM-5, CoZSM-5 and bimetallic forms of ZSM-5 zeolite. Catal. Today 110, 272–280
(2005). doi:10.1016/j.cattod.2005.09.027
40. B. Dragoi, V. Rakic, E. Dumitriu, A. Auroux, Adsorption of organic pollutants over microporous
solids investigated by microcalorimetry techniques. J. Therm. Anal. Calorim. 99, 733–740
(2010). doi:10.1007/s10973-009-0353-4
41. J.A. Perdigon-Melon, A. Gervasini, A. Auroux, Study of the influence of the In2 O3 loading on
γ -alumina for the development of de-NOx catalysts. J. Catal. 234, 421–430 (2005). doi:10.
1016/j.jcat.2005.07.001
42. S. Bennici, A. Auroux, Thermal analysis and calorimetric methods, in Metal Oxide Catalysis,
ed. by S.D. Jackson, J.S.J. Hargreaves (Wiley, New York, 2009)
43. B. Dragoi, E. Dumitriu, C. Guimon, A. Auroux, Acidic and adsorptive properties of SBA-
15 modified by aluminum incorporation. Microporous Mesoporous Mater. 121, 7–17 (2009).
doi:10.1016/j.micromeso.2008.12.023
408 V. Rakić

44. K. Otsuka, Y. Wang, M. Nakamura, Direct conversion of methane to synthesis gas through
gas-solid reaction using CeO2 -ZrO2 solid solution at moderate temperature. Appl. Catal. A
183, 317–324 (1999). doi:10.1016/S0926-860X(99)00070-8
45. T. Yuzhakova, V. Rakić, C. Guimon, A. Auroux, Preparation and characterization of Me2 O3 -
CeO2 (Me = B, Al, Ga, In) mixed-oxide catalysts. Chem. Mater. 19, 2970–2981 (2007). doi:10.
1021/cm062912r
46. C. Martin, A. Perrard, J.P. Joly, F. Gaillard, V. Delecroix, Dynamic adsorption on activated
carbons of SO2 traces in air I. Adsorption capacities. Carbon 40, 2235–2246 (2002)
47. A. Abu, S. Smith, Mechanistic characterization of adsorption and slow desorption of Phenan-
threne aged in soils. Environ. Sci. Technol. 40, 5409–5414 (2006). doi:10.1021/es060489h
48. S. Preis, J.L. Falconer, R. del Prado Asensio, N.C. Santiago, A. Kachina, J. Kallas, Photo-
catalytic oxidation of gas-phase methyl tert-butyl ether and tert-butyl alcohol. Appl. Catal.
B-Environ. 64, 79–87 (2006). doi:10.1016/j.apcatb.2005.11.002
49. M. Ischia, C. Perazzolli, R. Dal Maschio, R. Campostrini, Pyrolysis study of sewage sludge
by TG-MS and TG-GC-MS coupled analyses. J. Therm. Anal. Calorim. 87, 567–574 (2007).
doi:10.1007/s10973-006-7690-3
50. H.J. Lehmler, W. Xie , G.D. Bothun, P.M. Bummer, B.L. Knutson (2006) Mixing of perflu-
orooctanesulfonic acid (PFOS) potassium salt with dipalmitoyl phosphatidylcholine (DPPC).
Colloid. Surf. B 51: 25–29. doi:10.1016/j.colsurfb.2006.05.013
51. S. Gorinstein, S. Moncheva, F. Toledo, P. Arancibia-Avila, S. Trakhtenberg, A. Gorinstein, I.
Goshev, J. Namiesnik, Relationship between seawater pollution and qualitative changes in the
extracted proteins from mussels mytilus galloprovincialis. Sci. Total. Environ. 364, 251–259
(2006). doi:10.1016/j.scitotenv.2005.06.013
Chapter 11
Hydrogen and Calorimetry: Case Studies

Simona Bennici and Aline Auroux

Abstract In this chapter the use of calorimetric data is shown to be of primary

interest for the determination of the reaction/diffusion and reaction/absorption mech-
anisms involved in hydrogen storage or hydrogen production reactions. Examples
of calorimetric measurements (alone or coupled to volumetric devices working at
atmospheric or high pressure) and thermal analysis experiments applied to differ-
ent hydrogen storage systems are reported. In particular, applications of calorimetric
tools to the study of irreversible and reversible hydrogen storage systems are detailed,
such as borohydrides hydrolysis and hydrogen desorption from magnesium hydrides.

11.1 Introduction

Hydrogen as energy vector is one of the most promising approaches to the challenges
posed by global warming, and hydrogen-based fuel cell technologies are expected
to become one of the most prevalent energy sources in the near future [1, 2].
Now that technologies to use hydrogen as a clean fuel are readily available, like
the Proton Exchange Membrane Fuel Cell (PEMFC), and can be developed at an
industrial scale, research mainly focuses on the barrier of development which is
hydrogen storage for delayed use. In fact, if nowadays the H2 production methods
are well known and controlled, the storage and transportation of the fuel remain
major obstacles to its use [3].
During the last decade a lot of research effort has been put into the develop-
ment of suitable and safe technologies for hydrogen storage, such as materials for
high-pressure cylinders, liquefaction processes, hydrogen adsorption materials, and
metal hydrides [4, 5], adaptable to a wide range of applications from stationary and
automotive to portable devices. Although H2 adsorption capacities have recently been

S. Bennici (B) · A. Auroux

Institut de recherches sur la catalyse et l’environnement de Lyon, UMR5256,
CNRS-Université Lyon1, 2 avenue Einstein, 69626 Villeurbanne, France
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 409

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_11,
© Springer-Verlag Berlin Heidelberg 2013
410 S. Bennici and A. Auroux

brought up to values near 6–8 wt%, this storage method requires a high pressure and
low temperature.
Portable applications in particular require the implementation of H2 technologies
in the short term, due to a rapidly growing market and ever increasing energy demand.
In this case, conception of a H2 storage/generation system operating under ambient
conditions is a primary requirement. The ultra-pure hydrogen produced by hydrolysis
of metal borohydrides, possessing the right humidity level, can be directly used in
PEM fuel cells, which represents a very interesting alternative to lithium-ion battery.
At parity of mass, PEM based systems are 5–6 times more performing than lithium
batteries [6]. Moreover, chemical hydrides have an excellent potential for high energy
density storage at room temperature and ambient pressure [4, 5, 7].
To enable the commercialisation of fuel cells and hydrogen technologies, govern-
mental institutions worldwide are working to identify the most promising hydrogen
storage systems in order to promote the future R&D activities [7–9]. U.S Department
of Energy in particular has excluded the use of NaBH4 for on-board vehicle hydro-
gen storage due to its drawbacks in terms of exothermic reaction and formation of
very stable dehydrogenated products. The high energy consumption required for the
regeneration of these stable products back into hydrogen-containing fuel contributes
to an inefficient overall energy balance [10]. Despite this, NaBH4 remains a promis-
ing candidate for mobile, portable and stationary isolated site applications in which
the targets are less demanding than for automotive applications and the final product
price can be relatively high [11].
An appropriate catalyst is necessary to carry out the hydrolysis reaction at a
high rate. Noble metal based catalysts were initially developed and studied for this
purpose [5, 12–15], but the high cost of these materials associated with the low
supply available shifted the focus of research towards cheaper catalytic materials. In
fact, non-noble metals that form boride compounds such as Ni-B or Co-B alloys are
efficient and low-cost catalysts for this reaction [6, 8, 11, 12, 16–19].
Besides borohydrides, solid state complex hydrides [20] such as magnesium,
sodium, and lithium aluminium hydrides [21, 22], lithium amides [23], and their
combination with MgH2 [24, 25] have also proven to be potential candidates, in this
case for reversible hydrogen storage. The goal of the recent research is to overcome
the kinetic limitation and/or thermodynamic stability of Mg-based hydrogen storage
materials adding additives and catalysts to these mixtures [26, 27].
A third class of materials potentially useful in hydrogen storage is nanostruc-
tured materials (i.e. carbon nanofibers, carbon nanotubes, zeolites, clathrate hydrates,
metal-organic frameworks (MOFs)) that by virtue of their large surface-to-volume
ratios can adsorb considerable amounts of hydrogen in the molecular state via weak
molecular-surfacic interactions (physisorption) [4, 28–31]. For these materials the
main challenge is to find ways by which they can be engineered to store reversibly
high quantity of hydrogen close to room temperature.
For all the cited hydrogen storage systems, a precise determination of the heat
of reaction is needed for an industrially applicable system design and evaluation
of feasibility. The measurement of the heat evolved during a catalytic reaction or a
11 Hydrogen and Calorimetry: Case Studies 411

hydrogen desorption process is important from both the practical and fundamental
points of view.
Firstly it is an essential tool for the assessment of thermal risks related to the
performance of the reacting system at industrial scale (i.e. the capability of a system
to enter into a runaway reaction). This type of safety data is particularly important for
reactions like hydrogen generation by hydrolysis of borohydrides, where the rapid
increase in temperature may result in a sharp pressure increase. Secondly the thermo-
dynamic data are of primary interest for the determination of the reaction/diffusion
and reaction/absorption mechanisms.
Therefore, calorimetry and thermal analysis techniques alone or preferentially
combined to volumetric systems play a fundamental role in the study of hydrogen
storage materials and related phenomena.

11.2 First Case Study: Irreversible H2 Storage (Borohydrides)

The potential applications of sodium borohydride have been widely studied and most
of the research efforts and industrial devices developed up to now are based on the
catalyzed hydrolysis of sodium borohydride solutions (stabilized with NaOH) [32],
Eq. 11.1.

NaBH4 (aq) + 4H2 O(l) → NaB(OH)4 (aq) + 4H2 (g) (11.1)

◦ −1
r H = −250.5 kJ.molNaBH4

Sodium borohydride hydrolysis in aqueous solution can be represented in terms

of the overall stoichiometric equation (Eq. 11.1) where NaBH4 reacts with 4 mole-
cules of water to produce 4 molecules of H2 [33]. Although the reaction of NaBH4
hydrolysis has been studied since the discovery of sodium borohydride by Stock
in 1933 [34], the theoretical, calculated energy of the reaction is often cited in an
incompatible manner to the application, and real experimental data are very scarce
[35–38]. The thermodynamic features of the catalyzed hydrolysis in fact are not
yet well understood, as the evolved energy depends on the physical state and the
hydration degrees of borohydride and metaborate and on-side reactions.
Calorimetric techniques, and liquid phase calorimetry in particular, are promising
methods to study catalytic reactions [39]. Notably, the use of a differential reaction
calorimeter (DRC) makes it possible to determine the most important thermodynamic
data such as the heat of reaction and heat capacity of the system [40–42].

11.2.1 Hydrolysis of NaBH4 Stabilized Solutions

The study of Garron et al. [43] is one of the first investigations by liquid-phase
calorimetry of the mechanism of hydrogen generation by hydrolysis of sodium
412 S. Bennici and A. Auroux

borohydride catalysed by Co2 B nanoparticles generated in situ. The differential

reaction calorimeter (Fig. 11.1) was coupled with a hydrogen volumetric measure-
ment, allowing simultaneous thermodynamic and kinetic studies of the reaction.
In a typical experiment the “sample” vessel was filled with the reactants and the
“reference” cell with the solvent (NaOH solution). The oil circulating in the jack-
ets maintains the surroundings of the reactor at a constant temperature. The used
calorimeter principle is based on the continuous measurement of the temperature
difference between the two vessels during the experiment. In order to determine the
heat capacity (Cp) of the system and correlate the temperature difference (T) to
the heat flow, a Joule-effect calibration is performed before and after each reaction
step (see Fig. 11.2), and the corresponding energy is obtained by the Q = Cp x T
equation. The measured heat corresponds to the global energy evolved during the
NaBH4 hydrolysis reaction after each addition of fresh reactants (NaBH4 solution
stabilized by NaOH) (Fig. 11.2). The endothermic effects observed at the beginning
of each peak correspond to the addition of the reactive solution at a temperature
slightly lower than that of the thermostated vessel. In their paper [43] the authors
show that the total enthalpy of the catalytic process is strongly influenced by the
evolution of the catalyst during the hydrolysis reaction and by some water evapora-
tion which is related to the NaBH4 concentration. For the first time the calorimetric

Fig. 11.1 Picture of the DRC system coupled with a volumetric gas-meter. a Julabo Thermostat. b
IKA parallel stirrers. c DRC CS 32 control unit. d liquid N2 trap. e Gas volume-meter Ritter TG01.
f measurement cell. g reference cell
11 Hydrogen and Calorimetry: Case Studies 413

2,0 30

1,8 29
Differential temperature (°C)

1,6 NaBH 4 addition 28

1,4 27
1,2 Joule effect 26
1,0
25
0,8
24
0,6
23
0,4
0,2 22

0,0 21
0 3000 6000 9000 12000 15000 18000 21000 24000
-0,2 20
Time (s)

Fig. 11.2 Typical profile of a DRC analysis with 4 injections (10 mL) of NaOH-stabilized solution
of NaBH4 (19 wt.%) on 1 mmol Co-nanoparticles suspended in 20 mL water. Before and after each
injection peak a Joule effect is performed for determining the relative heat of reaction

250

200
10°C
Volume / mL

150

100

50

0
0 50 100 150 200
Time / min

Fig. 11.3 Evolution versus time of the hydrogen generation at different temperatures for the NaBH4
hydrolysis reaction (Injection of 0.2 mL of water on a solid mixture of 200 mg NaBH4 + 20 mg of
Co(H2 O)6 Cl2

technique was used to study the catalytic reaction in aqueous phase as complement
of “ex situ” characterization of the solid catalyst and the reacting solution, thus pro-
viding “operando” information about the reaction thermodynamics and kinetics and
consequently the mechanism.
The setup shown in Fig. 11.1 was thermostated so that experiments reported in
Fig. 11.3 could be performed at constant temperature between −5 ◦ C and 70 ◦ C, with
414 S. Bennici and A. Auroux

Fig. 11.4 Evolution versus time of the hydrogen generation (a) and of the differential temperature
(b) for the four successive additions of 10 mL of a NaOH-stabilized solution of sodium borohydride
(2 wt% NaBH4 stabilized by 5.7 mmol NaOH ) at 30 ◦ C on the in situ generated catalyst (1 mmol
Co(H2 O)6 Cl2 + 20 mL of H2 O + 5 mmol NaBH4 dissolved in 5 mL of H2 O)

the purpose of determining the activation energy by studying the dependence of the
rate of hydrogen production on the temperature. The activation energy determined
in this manner was −42.7 kJ.mol−1 [43].
Figure 11.4 shows the evolution of the hydrogen volume and differential tem-
perature versus time upon four successive additions of NaBH4 solution (Fig. 11.4a,
b). The curves of the amounts of evolved hydrogen (Fig. 11.4a) for each addition
show similar profiles, with a kinetic response that slightly decreases with each
successive addition as the concentration of metaborate and the total volume of
the reactant solution increase. In fact, the presence of residual sodium metaborate
11 Hydrogen and Calorimetry: Case Studies 415

during the reaction lowers the ability of borohydride to reach the catalyst surface
[43, 44]. The four differential temperature peaks obtained after successive additions
of solution (Fig. 11.4b) present very similar shapes, but the increasing of the width
at half-height (thermokinetic parameter), confirmed that thermal transfer within
the solution becomes increasingly difficult as its viscosity increases (Fig. 11.4b).
Nonetheless, the total area of the peak, and hence the total energy evolved during the
reaction, remain constant.
The activity towards NaBH4 hydrolysis of metals which can be readily oxidized
and form stable boride species was also studied by liquid reaction calorimetry [45].
Co, Ni, and Fe are stable in alkaline aqueous media and they are not affected by the
concentration of residual metaborate in the solution. In Ref. [45] the use of liquid-
phase calorimetric methods provided deep insight on the catalytic behaviour of the
samples by studying the energy production and its evolution as a function of time.
The volume of hydrogen generated at a given time is reported in Fig. 11.5 for Co, Ni
and Fe salts acting as catalysts in the NaBH4 hydrolysis reaction.

Co Ni Fe
600

500
Volume H2 / mL

400

300

200

100

0
1 3 8 15 25
Time / min

Fig. 11.5 Hydrogen volumes generated at 1, 3, 8, 15, and 25 min for Co, Ni and Fe salts acting as
catalysts. Catalyst generated in situ by addition of 5 mL of aqueous solution containing 200 mg of
NaBH4 on the chloride salt (1 mmol equivalent of metal in 20 mL water), pH = 14 by addition of
5 mL NaOH 6 M. Test: additions of 10 mL of NaBH4 (2 wt.% solution) stabilized by 10 mmol of
NaOH (4 wt.%) [45]
416 S. Bennici and A. Auroux

11.2.2 Hydrolysis of NaBH4 and KBH4 Powders

The hydrogen storage capacity of sodium borohydride depends on the quantity

of water involved in the whole storage system. For example, for a standard com-
mercial solution containing 20 wt% of NaBH4 the storage capacity corresponds to
4.2 wt% H2 . Higher hydrogen storage capacities can be achieved by the reaction of
stoichiometric quantity of water with solid sodium borohydride. If the water content
in the Na,KBH4 (s)/catalyst(s)/H2 O(l) system is high enough to dissolve the metabo-
rate product at room temperature (25 ◦ C), the advantages of this configuration are the
possibility to store the reagents separately and to remove the products of the reaction
from the system in liquid form, avoiding the formation of a borate solid crust that
will decrease the diffusion of water in the solid reactants [46].
To mimic and study this storage configuration the instrument used was a Cal-
vet type differential heat flow microcalorimeter which allows continuous stirring of
liquid-solid mixtures (Fig. 11.6). After stabilization of the calorimeter baseline, the
water (or the hydrolyzing solution) was added to the sample cell which contains the
potassium or sodium borohydride powders (eventually activated with a catalyst) and
to the empty reference cell in the same way using a programmable twin syringe pump
linked to the calorimeter by capillary tubes. The reaction cell had a home-made air-
tight cap with attached inox outlet tubing. The inox tube was linked to a gas volume
meter allowing continuous measurement of generated hydrogen.
Because hydrolysis of NaBH4 and KBH4 is a very slow process and the rate of
hydrogen generation from this reaction depends on the addition of a catalyst and on

Fig. 11.6 Scheme of the Titrys calorimeter coupled with a volumetric gas-meter and a program-
mable twin syringe pump
11 Hydrogen and Calorimetry: Case Studies 417

its preparation method, the study of non-catalyzed hydrolysis can be very difficult as
well as the experimental determination of the reaction heat. In fact, long time scale
experiments are challenging in terms of accuracy of the total evolved heat (integral
method) determination.
Self-hydrolysis occurs spontaneously but slowly when NaBH4 and KBH4 are
in contact with water, and liquid calorimetry coupled with a gas-meter (Fig. 11.6)
permits to precisely determine the heats of the hydrolysis reaction for NaBH4 and
KBH4 that are respectively of −236 and −220 kJmolNa,KBH4 −1 [46]. These values
were obtained by integration of the calorimetric peaks shown in Fig. 11.7 that display
the heat flow evolved during the hydrolysis reactions of NaBH4 and KBH4 without
catalyst. The endothermic signal visible on the heat flow signal versus time curves
for both investigated borohydrides is attributed to the water injection. The hydrolysis
reaction is an exothermic process and, after an initial increase, the heat flow signal
reaches an almost constant value until the end of the hydrolysis reaction, when it
exhibits an abrupt decrease. The value of the measured heat includes the enthalpy of
dissolution of the borohydride, the enthalpy of the hydrolysis reaction, the enthalpy
of dissolution of the by-product (NaBO2 or KBO2 ) and the evaporation of some
water during the exothermic reaction.
Therefore, to increase the decomposition rate of NaBH4 by hydrolysis a catalyst
is needed. As already pointed out, cobalt-based catalysts have shown promising
activity, and acids are also suitable catalytic accelerators [47, 48].
In Fig. 11.8 the heat flow peaks for the NaBH4 hydrolysis reaction in pres-
ence of malic acid (a), Co-nanoparticules (b) and pure water (c) are reported, and
the heats of reaction obtained by integration of these peaks were respectively of
−298, −244, and − 236 kJ · molNaBH4 −1 . By analyzing the shape of the calorimetric
peaks a perception of the hydrogen release kinetics can be obtained; the most rapid

Fig. 11.7 Heat flow signals versus time for NaBH4 and KBH4 hydrolysis without catalyst (1 mL
water on 30 mg borohydride)
418 S. Bennici and A. Auroux

400
a

300

Heat Flow / mW
200

b
100

0
c
3.5 4.0 4.5 5.0
Time / h

Fig. 11.8 Heat flow generated during NaBH4 hydrolysis by addition of a 1 mL of 1 M maleic acid,
b 1 mmol Co-nanoparticules dispersed in 1 mL of water, and c pure water

Fig. 11.9 Heat flow signals versus time for KBH4 hydrolysis reaction in the presence of different
Co-based solids (30 mg of KBH4 and an amount of catalyst equivalent to 10 wt% of pure Co)
11 Hydrogen and Calorimetry: Case Studies 419

hydrogen production corresponds to the use of malic acid and it can be noticed that
the faster reaction corresponds to the higher heat generated.
The experimental calorimetric curves obtained for the hydrolysis reaction of
NaBH4 in the presence of different catalysts (Co nanoparticles, CoO and Co3 O4 ) are
shown in Fig. 11.9, as well as the curve obtained for the NaBH4 hydrolysis reaction
without catalyst reported as reference. The measured heats of the NaBH4 hydrol-
ysis reaction were −243, −235, and − 236 kJ · mol−1 NaBH4 in the presence of Co
nanoparticles, CoO, and Co3 O4 , respectively [46]. The reaction times for the com-
plete hydrogen generation were: 33 min for Co nanoparticles, 120 min for Co3 O4 and
240 min for CoO. In all cases, 100 % of the stoichiometric amount of hydrogen was
generated during the stated reaction time [46]. Much higher hydrogen production
rates (6000 mL min−1 gCo , corresponding to a reaction time of 7 min) were measured
in the same way on Co based catalysts with the cobalt phase dispersed on acidic
heteropolyanions supports [49].
By means of liquid calorimetry coupled with a volumetric gas-meter Damjanovic
et al. [50] have shown that addition of solid NaOH to the reacting mixture (NaBH4 +
catalyst) increases the rate of NaBH4 hydrolysis reaction in the presence of Co-based
catalysts. Comparing the results available in the literature it is clear that the effect
of NaOH concentration on the hydrolysis generation rate greatly depends on the
type of catalysts [51]. Consequently, it is important to find an optimum range for
NaOH concentration to improve the hydrogen generation rate in the presence of Co
based catalysts. Figure 11.10 shows the experimental calorimetric curves obtained
for hydrolysis of 30 mg of NaBH4 in the presence of 4.1 mg of Co3 O4 (corre-
sponding to 10 wt% of pure Co with respect to sodium borohydride), and different
amounts of solid NaOH. The measured heats after injection of 0.5 ml of water were

Fig. 11.10 Heat flow signals versus time for injection of 0.5 mL of water on 30 mg of NaBH4
hydrolysis in the presence of 4.1 mg of Co3 O4 , and different amounts of solid NaOH (as reported
on the figure)
420 S. Bennici and A. Auroux

−225, −420, −390 and − 250 kJmol−1 for 5.5, 30, 66 and 125 mg of solid NaOH,
respectively [52]. As it can be seen on Fig. 11.5, a clear shift from 20 to about 10 min
of the peak maximum occurs when increasing the amount of solid NaOH added from
5.5 to 30 mg. Adding a balanced amount of solid NaOH (around 15–20 mg) to the
“NaBH4 + catalyst” mixture improves the kinetics of the reaction and creates a real
benefit in terms of reaction kinetics. In parallel to the calorimetry measurements,
the volume of generated H2 versus time during the NaBH4 hydrolysis reaction in
the presence of Co3 O4 , and different amounts of solid NaOH was determined. Full
conversion was achieved after 35, 12, 15 and 40 min for 5.5, 30, 66 and 125 mg of
solid NaOH, respectively. The maximum rates reached over Co3 O4 were 8, 16, 16
and 10 ml min−1 for 5.5, 30, 66 and 125 mg of solid NaOH, respectively, as presented
in Fig. 11.11 [50].
Another experimental solution is to inject very small amounts of water (that will
act as limiting reactant) and to avoid any separation step involving reactants and
reacted products. In this simpler configuration the issues to solve are the improvement
of the water diffusion in the solid (once the first injection performed and consequently
further NaBO2 ·xH2 O formed) and the optimisation of water consumption (by reduc-
ing the NaBO2 hydration ratio).
To give an answer to these issues, interesting studies about the thermal properties
of NaBH4 solid mixtures were obtained with other thermal techniques, as reported
in Ref. [16], where an infrared camera has been used to evaluate the temperature

Fig. 11.11 Evolution versus time of the hydrogen generation for injection of 0.5 mL of water on
30 mg of NaBH4 in the presence of 4.1 mg of Co3 O4 , and different amounts of solid NaOH (as
reported on the figure)
11 Hydrogen and Calorimetry: Case Studies 421

37.0 °C

35

30

25

21.2
Temperature (°C)

0 5 10 15
Distance (mm)

Fig. 11.12 Thermal imaging and temperature profile during the hydrolysis of NaBH4 + 10 wt%
nCo at 25 ◦ C in presence of a less-than-stoichiometric amount of water

profile in a NaBH4 + Co-nanoparticles solid mixture (with or without addition of

solid NaOH) when a drop of water is added (Fig. 11.12).
Knowing that the major limitation to reach the theoretical 10.9 wt% H2 storage
density by hydrolysis of NaBH4 powder with water is the hydration level of the
metaborate product, to attain a higher H2 storage density, a higher temperature of
the system is required for dehydrating the metaborate products, and an easy way to
reach this objective is to utilize the thermal effects of hydrolysis.
These thermal effects associated to the hydrolysis reaction have been stud-
ied on a fully dehydrated NaBH4 powder by means of an IR imaging camera
and a differential titration calorimeter. Various amounts of solid sodium hydrox-
ide were added to the system (NaBH4 + metallic nanoCobalt catalyst) allowing an
increase of the maximum reaction temperature (up to 140 ◦ C). The reaction maxi-
mum temperature and the hydrogen yield were considerably modified by varying
the amount of NaOH and the amount of catalyst (Fig. 11.13). At a temperature
of more than 140 ◦ C, it is reasonable to expect the formation of low hydration
borate phases. In fact, at temperatures above 105 ◦ C water is expected to partic-
ipate preferentially in the hydrolysis reaction rather than in the hydration of the
422 S. Bennici and A. Auroux

150 3000
140
130
Two times more 2500
120
Catalyst (20 wt.% Co)

Peak width at half-height (s)

Maximum temperature (°C)

110
100 2000
90
80
1500
70
60
50 1000
40
30
500
20
10
0 0
0 1 2 3 4 5 6 7 8 9 10
NaOH quantity (wt.%)

Fig. 11.13 Maximum temperature (left side) and peak width at half-height (right side) as functions
of the sodium hydroxide loading, and comparison with the results obtained with twice the amount
of catalyst (the two curves are obtained in presence of 10 wt% nCo catalyst)

formed metaborate [52]. Moreover the absence of highly hydrated borate favors the
diffusion of water by avoiding the formation of the hydrophobic NaBO2 · 2H2 O
layer.

11.3 Second Case Study: Reversible H2 Storage

(Mg-Based Materials)

As mentioned in the introductory section of this chapter, for automotive applications

the requirements for on-board hydrogen storage are very severe, as mainly stated
by the United States Department of Energy [7] and the Japanese Government. The
same kind of restrictions are also associated to the storage of intermittent energies
where the amount of energy involved can be extremely high and consequently a
stable storage system is required. To reach the demanded targets for the application
in the automotive field research efforts have been made to develop interstitial, binary
or even more complex hydrides capable to store and release hydrogen at tempera-
ture and pressure compatible with the different applications. For example, Mg (non
toxic and inexpensive) shows a hydrogen storage capacity of 7.7 wt%, but the major
impediment is its high H2 desorption temperature (>300 ◦ C). In order to decrease
the desorption temperature of MgH2 many research groups have tried to add other
hydrides to MgH2 in order to form complex hydrides with lower hydrogen desorption
11 Hydrogen and Calorimetry: Case Studies 423

Fig. 11.14 DSC profiles for (Mg(NH2 )2 + 2LiH) and Mg(NH2 )2 at various scan rates [30]

temperature. Temperature programmed desorption techniques (TPD, DSC, TG) are

the right tools to follow the hydrogen desorption process, determining the desorption
temperature, evaluating the amount of desorbed H2 and in DSC measuring the heat
associated to the hydrogen release.
As example, the Li-Mg-N-H system was studied by differential scanning calorime-
try by Gross’s group [53]. The major issue for metal-N-H storage systems is the
formation of NH3 , that takes place in parallel with H2 release, and that acts as a

Fig. 11.15 HP-DSC of the desorption reaction of MgH2 –2LiBH4 composites after milling (1),
after initial hydrogenation (2) and after milling with additive (3). Heating rate 5◦ C · min−1 ; 20
mLH2 · min−1 ; 3 bar H2 [57]
424 S. Bennici and A. Auroux

Fig. 11.16 a TPD comparison of LiBNH + nMgH2 without additive and with 2 mol% Ni, Cu,
Mn, Co and Fe at a constant ramping rate of 1◦ C/min and b Ramping kinetics measurements of
LiBNH + nMgH2 without and with 2 mol% nano Mn, Fe, Co, Cu, Ni and Fe + Ni [24]
11 Hydrogen and Calorimetry: Case Studies 425

Fig. 11.17 Scheme of the calorimetric device: (1) calorimetric chambers; (2) calorimetric cells
(reaction and blank); (3–8) needle valves; (9) vacuum gauge; (10) pressure gauge; (11) hydrogen
source

poison for the PEM fuel cell catalysts [54]. As reported in Fig. 11.14, the lower tem-
perature peak obtained by DSC method and centered at 200 ◦ C is related to hydrogen
release by the Mg(NH2 )2 + 2LiH → Li2 Mg(NH2 )2 + H2 reaction, while the higher
temperature peak corresponds to the self-decomposition of Mg amide (irreversible
reaction) related to NH3 production. The identification of these phenomena and of
their relative temperatures helps to identify the maximum working temperature nec-
essary to avoid the Mg(NH2 )2 decomposition and improve the cyclic stability of the
material.
Addition of LiBH4 to Mg allows to decrease the hydrogen desorption temperature
by approximately 30 ◦ C by formation of MgB2 intermediary species [55, 56]. New
techniques as high-pressure differential scanning calorimetry (HP-DSC) have been
used to verify the mechanism of decomposition of MgH2 –LiBH4 composites as well
as the influence of the addition of different additives such as titanium isopropoxide
and VCl3, under 3 bar hydrogen pressure [57]. As shown in Fig. 11.15, the absorption-
desorption cycling and the addition of additives deeply influence the kinetics of
hydrogen release. As consequence, the peaks C and D shift at lower temperature
for the MgH2 –LiBH4 materials activated by the absorption reaction (curve 2) or by
addition of titanium isopropoxide (curve 3).
Novel complex hydrides (LiBH4 /LiNH2 /MgH2 ) are also under study and the
addition of various nanosized additives as nickel, cobalt, iron, copper, manganese
was followed by thermoprogrammed desorption technique (TPD) [24].
The TPD tool was used in Ref. [24] to identify the optimal hydrogen release
temperature and to get an indication of the H2 desorption rate by measuring the peak
width. In Fig. 11.16, the addition of the different additives has a big effect on reducing
the temperature of the second H2 releasing peak.
Unusual are set-ups combining Tian-Calvet calorimetry and high pressure volu-
metric lines. One of the few examples is reported in the work of Anikina and Verbetsky
426 S. Bennici and A. Auroux

in Ref. [58]: the schematic diagram of their set-up is reported on Fig. 11.17. This tech-
nique was used to study TiZrMnV compounds used as H2 storage system. Desorption
molar enthalpies were determined and used in support to the researcher assumption
that the material was composed of two different hydride phases, Ti0.9 Zr 0.1 Mn1.1 H1.0
and Ti0.9 Zr 0.1 Mn1.1 H2.0 , respectively at low and high hydrogen content.

11.4 Conclusions

Different kinds of hydrogen storage materials have been studied and reported in
the recent literature, but only few candidates are able to answer to the require-
ments for portable and on-board applications. The main problems still to be solved
are the continuous control of the reaction for safety reasons, the increasing of the
kinetics of hydrogen release, and the reversibility of the reaction to perform adsorp-
tion/desorption cycling.
For an extensive use of these materials it is important to deeply understand the ther-
mal phenomena associated to hydrogen release reactions and adsorption/desorption
cycles.
The thermodynamic features of these systems (enthalpies, entropies) as well as the
thermal properties of the involved compounds (thermal conductivity, heat capacity)
are crucial data to predict the thermal behavior of large quantity of material, as those
used in real applications.
Calorimetry and thermal analysis techniques (alone or coupled to other instru-
ments) have been developed and applied to the study of hydrogen storage systems.
As example, information about the heat evolved during the hydrogen release as a
function of time has given the possibility to contemporarily evaluate the kinetics of
reaction and to point out the presence of diffusion problems, connected to borohy-
dride hydrolysis. In other showcases the identification of specific thermal phenomena
helped in understanding the reaction mechanism and the material structure.
It is certain that calorimetry and thermal analysis are indispensable tools in
the study of hydrogen storage and further developments and applications of these
techniques have to be expected in the near future.

References

1. G. Marban, T. Valdes-Solis, Int. J. Hydrogen Energy 32, 1625 (2007)

2. D.A.J. Rand, R.M. Dell, Hydrogen Energy (The Royal Society of Chemistry, Cambridge, 2008)
3. F. Jackow, J. Loughead, European Technology Platform for Hydrogen and Fuel Cells
(HFP), Implementation plan-status 2006 http://www.fch-ju.eu/sites/default/files/documents/
hfp-ip06-final-apr2007.pdf/
4. L. Schlappbach, A. Züttel, Nature 414, 353 (2001)
5. B. Sakintuna, F. Lamari-Darkrimb, M. Hirscher, Int. J. Hydrogen Energy 32, 1121 (2007)
6. S.F.J. Flipsen, J. Power Sour. 162, 927 (2006)
11 Hydrogen and Calorimetry: Case Studies 427

7. U.S. Department of Energy Hydrogen Program, http://www.hydrogen.energy.gov/

8. U.B. Demirci, O. Akdim, P. Miele, Int. J. Hydrogen Energy 34, 2638 (2009)
9. S. Bakker, Int. J. Hydrogen Energy 35, 6784 (2010)
10. C. Cakanyildirim, M. Guru, Int. J. Hydrogen Energy 33, 4634 (2008)
11. S.U. Jeong, R.K. Kim, E.A. Chob, H.-J. Kim, S.-W. Nam, I.-H. Oh, S.-A. Hong, S.H. Kim, J.
Power Sour. 144, 129 (2005)
12. B.H. Liu, Z.P. Li, J. Power Sour. 187, 527 (2009)
13. S.C. Amendola, S.L. Sharp-Goldman, M.S. Janjua, N.C. Spencer, M.T. Kelly, P.J. Petillo, Int.
J. Hydrogen Energy 25, 969 (2000)
14. Y. Kojima, K.I. Suzuki, K. Fukumoto, M. Sasaki, T. Yamamoto, Y. Hawai, Int. J. Hydrogen
Energy 27, 1029 (2002)
15. J.S. Zhang, W.N. Delgass, T.S. Fisher, J.P. Gore, J. Power Sour. 164, 772 (2007)
16. S. Bennici, A. Garron, A. Auroux, Int. J. Hydrogen Energy 35, 8621 (2010)
17. B.H. Liu, Z.P. Li, S. Suda, J. Alloys Compd. 415, 288 (2006)
18. P. Krishnan, K.L. Hsueh, S.D. Yim, Appl. Catal. B 77, 206 (2007)
19. R. Chen, X. Wang, L. Xu, L. Chen, S. Li, C. Chen, Mater. Chem. Phys. 124, 83 (2010)
20. A. Züttel, P. Sudan, P. Mauron, P. Wenger, Mater. Sci. Eng. B 108, 9 (2004)
21. N. Eigen, F. Gosch, M. Dornheim, T. Klassen, R. Bormann, J. Alloys Compds. 465, 310 (2008)
22. M. Ismail, Y. Zhao, X.B. Yu, S.X. Dou, Int. J. Hydrogen Energy 35, 2361 (2010)
23. K.R. Graham, T. Kemmitt, M.E. Bowden, Energy Environ. Sci. 2, 706 (2009)
24. S.S. Srinivasan, M.U. Niemann, J.R. Hattrick-Simpers, K. McGrath, P.C. Sharma, D.Y.
Goswami, E.K. Stefanakos, Int. J. Hydrogen Energy 35, 9646 (2010)
25. P. Chen, Z. Xion, J. Luo, J. Lin, K.L. Tan, Nature 420, 302 (2002)
26. W.N. Yang, C.X. Shang, Z.X. Guo, Int. J. Hydrogen Energy 35, 4534 (2010)
27. C.X. Shang, Z.X. Guo, Int. J. Hydrogen Energy 32, 2920 (2007)
28. D. Dybtsev, C. Serre, B. Schmitz, B. Panella, M. Hirscher, M. Latroche, P.L. Llewellyn, S.
Cordier, Y. Molard, M. Haouas, F. Taulelle, G. Férey, Langmuir 26, 11283 (2010)
29. M. Latroche, S. Suble, C. Serre, C. Mellot-Draznieks, P.L. Llewellyn, J.-H. Lee, J.-S. Chang,
S.H. Jhung, G. Ferey, Angew. Chemie Int. Ed. 45, 8227 (2006)
30. H.W. Langmi, D. Book, A. Walton, S.R. Johnson, M.M. Al-Mamouri, J.D. Speight, P.P.
Edwards, I.R. Harris, P.A. Anderson, J. Alloys Compds. 637, 404 (2005)
31. G.E. Ioannatos, X.E. Verykios, Int. J. Hydrogen Energy 35, 622 (2010)
32. V.G. Minkina, S.I. Shabunya, V.I. Kalinin, V.V. Martynenko, A.L. Smirnova, Int J. Hydrogen
Energy 33, 5629 (2008)
33. D.R. Lide, CRC Handbook of Chemistry and Physics, 82nd edn. (CRC Press, Boca Raton,
2001)
34. A. Stock, Hydrides of Boron and Silicon (Cornell University Press, Ithaca, 1933)
35. J.C. Ingersoll, N. Mani, J.C. Thenmozhiyal, A. Muthaiah, J. Power Sour. 173, 450 (2007)
36. J. Zhang, T.S. Fisher, J.P. Gore, D. Hazra, P.V. Ramachandran, Int. J. Hydrogen Energy 31,
2292 (2006)
37. N.O. Gonzales, M.E. Levin, L.W. Zimmerman, J. Hazard. Mater. 142, 369 (2007)
38. J. Li, B. Li, S. Gao, Phys. Chem. Miner. 27, 342 (2000)
39. S. Bennici, A. Auroux, Thermal Analysis and Calorimetric Methods, in Metal Oxide Catalysis,
vol. 1, ed. by S.D. Jackson, S.J. Hargreaves (Wiley-VCH Verlag GmbH & Co, Weinheim, 2009),
p. 391
40. H. Nogent, X. Le Tacon, L. Vincent, N. Sbirrazzuoli, J. Loss Prev. Process Ind. 18, 43 (2005)
41. R. Andre, L. Bou-Diab, P. Lerena, F. Stoessel, M. Giordano, C. Mathonat, Organ. Process Res.
Dev. 6, 915 (2002)
42. J. Andrieux, D. Swierczynski, L. Laversenne, A. Garron, S. Bennici, C. Goutaudier, P. Miele,
A. Auroux, B. Bonnetot, Int. J. Hydrogen Energy 34, 938 (2009)
43. A. Garron, D. Swierczynski, S. Bennici, A. Auroux, Int. J. Hydrogen Energy 34, 1185 (2009)
44. S.J. Kim, J. Lee, K.Y. Kong, C.R. Jung, I.G. Min, S.-Y. Lee, H.-J. Kima, S.W. Nam, T.-H. Lim,
J. Power Sour. 170, 412 (2007)
45. A. Garron, S. Bennici, A. Auroux, Appl. Catal. A 378, 90 (2010)
428 S. Bennici and A. Auroux

46. L. Damjanovic, S. Bennici, A. Auroux, J. Power Sour. 195, 3284 (2010)

47. O. Akdim, U.B. Demirci, P. Miele, Int. J. Hydrogen Energy 34, 4780 (2009)
48. H.I. Schlesinger, H.C. Brown, A.E. Finholt, J.R. Gilbreath, H.R. Hoekstra, E.K. Hyde, J. Am.
Chem. Soc. 75, 215 (1953)
49. S. Bennici, H. Yu, E. Obeid, A. Auroux, Int. J. Hydrogen Energy 36, 7431 (2011)
50. L. Damjanovic, S. Bennici, A. Auroux, Int. J. Hydrogen Energy 36, 1991 (2011)
51. H. Tian, Q. Guo, D. Xu, J. Power Sour. 195, 2136 (2010)
52. E.Y. Marrero-Alfonso, J.R. Gray, T.A. Davis, M.A. Matthews, Int. J. Hydrogen Energy 32,
4723 (2007)
53. W. Luo, J. Wang, K. Steward, M. Clift, K. Gross, J. Alloys Compds. 336, 446 (2007)
54. G. Postole, A. Auroux, Int. J. Hydrogen Energy 36, 6817 (2011)
55. S. Orimo, Y. Nakamori, G. Kitahara, K. Miwa, A. Ninomiya, H. Li, N. Ohba, S. Towata, A.
Zuttel, J. Alloys Compd. 427, 404 (2005)
56. J.J. Vajo, S.L. Skeith, F. Mertens, J. Phys. Chem. B Lett. 109, 3719 (2005)
57. U. Bosenberg, S. Doppiu, L. Mosegaard, G. Barkhordarian, N. Eigen, A. Borgschulte, T.R.
Jensen, Y. Cerenius, O. Gutfleisch, T. Klassen, M. Dornheim, R. Bormann, Acta Materialia 55,
3951 (2007)
58. E.Yu. Anikina, V.N. Verbetsky, J. Alloys Compd. 45, 330 (2002)
Chapter 12
Adsorption Microcalorimetry as a Tool
to Study the CO–Pt Interaction for PEMFC
Applications: A Case Study

Georgeta Postole and Aline Auroux

Abstract To date, microcalorimetry of CO adsorption onto supported metal catalysts

was mainly used to study the effects induced by the nature and the particle size of
supported metallic clusters, the conditions of pretreatment and the support materi-
als on the surface properties of the supported metallic particles. The present chapter
focuses on the employ of adsorption microcalorimetry for studying the interaction of
carbon monoxide with platinum-based catalyst aimed to be used in proton exchange
membrane fuel cells (PEMFCs ) applications.

12.1 Evolution and Types of Fuel Cell

Fuel cells are one of the oldest electrical energy conversion technologies known to
man for more than 160 years. A fuel cell is a galvanic cell, in which the free energy
of a chemical reaction is converted into electrical energy. Their development lacked
a drive at the beginning, as primary energy sources were abundant, unrestricted,
and inexpensive. After the second world war, this technology became the subject
of intense research; one of the major factors that have influenced their development
have been the increasing concern about the environmental consequences of fossil
fuel use in production of electricity, and for the propulsion of vehicles. The historical
development of fuel cells has been described by Carrete et al. [1], Litster et al. [2] and
Boudghene Stambouli and Traversa [3]. In spite of the attractive system efficiencies
and environmental benefits associated with fuel-cell technology, it is still a difficult
task to transform the early scientific experiments into commercially viable industrial
products. These problems have been often associated with the lack of appropriate

G. Postole (B) · A. Auroux

Université Lyon 1, CNRS, UMR 5256, IRCELYON, Institut de recherches
sur la catalyse et l’environnement de Lyon, F-69626 Villeurbanne Cedex, France
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 429

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_12,
© Springer-Verlag Berlin Heidelberg 2013
430 G. Postole and A. Auroux

Fig. 12.1 Summary of fuel-cell types [5]

materials or manufacturing routes that would enable the cost of electricity per kWh
to compete with the existing technology [4].
However, fuel cells may help to reduce our dependence on fossil fuels and diminish
poisonous emissions into the atmosphere and that is the reason for their continuous
development. For example, fuel cells using pure hydrogen as fuel produce only
water, thus eliminating locally all noxious byproducts otherwise caused by electricity
production.
Different types of fuel cells under active development are usually classified by
the electrolyte employed as the ionic conductor in the cell, or by their operating
temperature. Figure 12.1 summarizes the types of fuel cells in order of increasing
operating temperature [5], while an overview of their characteristics is given in
Table 12.1 [1, 3].
As it can be seen, low-temperature and high-temperature fuel cells can be distin-
guished. Low-temperature fuel cells are the Alkaline Fuel Cell (AFC), the Polymer
Electrolyte Fuel Cell (PEMFC), and the Phosphoric Acid Fuel Cell (PAFC). The
high-temperature fuel cells operate in the temperatures region from 500 to 1000 ◦ C;
two different types have been developed: the Molten Carbonate Fuel Cell (MCFC)
and the Solid Oxide Fuel Cell (SOFC). They have the ability of using methane as
fuel and thus present high inherent generation efficiency (45–60 % for common fuels
such as natural gas, 90 % with heat recovery [3]).
Table 12.1 The characteristics of different fuel cells that have been realized and are currently in use and development
Fuel cells AFC (Alkaline) PEMFC (Polymer DMFC (Direct PAFC (Phosphoric MCFC (Molten SOFC (Solid oxide)
electrolyte methanol) acid) carbonate)
membrane)
Operating <100 60–120 60–120 160–200 600–800 800–1000 low
temperature temperature
(◦ C) (500–600) possible
Anode reaction H2 + 2OH− → H2 → 2H+ + 2e− CH3 OH + H2 O → H2 → 2H+ + H2 + CO2− 3 → H2 + O2− → H2 O
2H2 O + 2e− CO2 + 6H+ + 2e− H2 O + CO2 + + 2e−
6e− 2e−
Cathode reaction 1/2O2 + H2 O + 1/2O2 + 2H+ + 3/2O2 + 6H+ + 6e− 1/2O2 + 2H+ + 1/2O2 + CO2 + 1/2O2 + 2e− → O2−
2e− → 2OH− 2e− → H2 O → 3H2 O 2e− → H2 O 2e− → CO3 2−
Electrolyte Potassium Polymer, proton Polymer Phosphoric acid Molten salt such as Ceramic as stabilized
hydroxide exchange membrane nitrate, sulphate, zirconia and doped
(KOH) carbonates… perovskite
Charge carrier in OH− H+ H+ H+ CO3 2− O2−
the electrolyte
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction

(continued)
431
 432

Table 12.1 (continued)

Fuel cells AFC (Alkaline) PEMFC (Polymer DMFC (Direct PAFC (Phosphoric MCFC (Molten SOFC (Solid
electrolyte methanol) acid) carbonate) oxide)
membrane)
Fuel Hydrogen or Hydrogen Liquid Hydrogen Hydrogen, carbon Hydrogen,
hydrazine methanol monoxide, natural hydrocarbons,
gas, propane natural gas,
renewable fuels
Applications Transportation, space, military, energy storage Combined heat and Combined heat and power for decentralized
system power for stationary power systems and for
decentralized stationary transportation (train, boats, …)
power systems
Realized power Small plants Small plants Small plants Small-medium plants Small power plants Small power plants
5–150 kW 5–250 kW 5 kW 50 kW–11 MW modular 100 kW–2 MW 100–200 kW
modular modular modular
Efficiency (%) 50–55 40–50 40–55 40–50 50–60 45–60
G. Postole and A. Auroux
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction 433

Not mentioned in this classification is the Direct Methanol Fuel Cell (DMFC)
working at 60–120 ◦ C and using methanol as a fuel. There are also other types of
fuel cells (e.g. air-depolarised cells, sodium amalgam cells, alkali metal-halogen
cells, etc.) which are less employed, but that can possibly find a specific application
in the future.
The basic structure of all fuel cells is similar: the cell consists of two electrodes
which are separated by the electrolyte and which are connected in an external circuit.
The electrodes are exposed to gas or liquid flows to supply the electrodes with fuel or
oxidant (e.g. hydrogen or oxygen). As it can be seen in Table 12.1, the anode reaction
in fuel cells is either the direct oxidation of hydrogen (low temperature fuel cells) or
the oxidation of methanol (DMFC). An indirect oxidation via a reforming step can
also occur in the case of high temperature operation fuel cells. The cathode reaction
is oxygen reduction, in most cases from air.
Both the low-temperature and the high-temperature fuel cells have their advan-
tages and disadvantages depending on the application. For example, among the fuel
cells used for transportation, the high-temperature fuel cells (such as solid oxide or
molten carbonate cells) present advantages for the use in ships or locomotives in
which frequent on/off cycling is not required. They allow more flexibility in fuel
selection and may be used without a reformer. The high-grade waste heat is more
easily used in a thermally integrated system. Conversely, the low-temperature fuel
cells, such as polymer electrolyte or alkaline, may be a better choice for passenger
cars to which rapid start-up and wide power range are important [6].
By far the greatest research interest throughout the world has focussed on proton
exchange membrane (PEM) and solid oxide (SO) cell stacks. PEMFCs are well
advanced type of fuel cells that has found widespread area of use, especially in
transportation applications, distributed generation (DG) units and portable electronic
equipments [7]. Some of the key advantages of PEMFC systems over the other
competitive types of fuel cells for its potential market competitiveness arise from
[8]:

• PEMFCs can operate at relatively low temperatures.

• PEMFCs are tolerant to CO2 ; so they can use the atmospheric air.
• PEMFCs have high voltage, current and power density.
• PEMFCs can work at low pressure (1 or 2 bars), which enhances security.
• PEMFCs have a good tolerance to the difference of pressure of the reactants.
• PEMFCs are compact and robust and have a simple mechanical design.
• PEMFCs use stable building materials.

12.2 Proton Exchange Membrane Fuel Cells

PEM fuel cells use a proton exchange membrane as an electrolyte and operates at low-
temperatures, between 60 and 120 ◦ C. From various types of fuel cells, PEMFCs are
suitable choice for both stationary and portable applications due to their fast start up,
434 G. Postole and A. Auroux

high power density, suitability for discontinuous operation as well as low operating
temperature [8–10]. Their components and related functions are well described in
a recent review [9]. The first application of a PEM fuel cell was in the 1960s as
an auxiliary power source in the Gemini space flights. It was also used to provide
the astronauts with clean drinking water. The membrane used was a polystyrene
sulfonate (PSS) polymer, which has been proven do not be enough stable. Advances
in the PEMFCs technology were stagnating until the late 1980s when the fundamental
design underwent significant reconfiguration. A major breakthrough in this field
came with the use of Nafion or Dow membranes, possessing a higher acidity
and conductivity and being more stable than the polystyrene sulfonate membranes
[1, 2].
Figure 12.2 presents the principle of a single proton exchange membrane fuel cell
fed with hydrogen which is oxidized at the anode, and oxygen that is reduced at the
cathode [1]. The protons released during the oxidation of hydrogen are conducted
through the proton exchange membrane (the electrolyte) to the cathode. Since the
membrane is not electrically conductive, the electrons released from the hydrogen
travel along the electrical detour provided and an electrical current is generated. The
reaction product is water, which is formed at the cathode [2].
In order to be efficient enough, the electrochemical reactions that take place in
fuel cell must be catalyzed. To date, platinum has proven to be the best catalyst for
both the hydrogen oxidation (anode) and the oxygen reduction (cathode) reactions.
For PEM fuel cell applications, platinum is usually implemented in the form of
Pt/C catalysts because of its significantly higher surface area compared with that of
platinum black catalysts resulting in its cost reduction [5, 11].
The oxidation of hydrogen occurs readily on Pt-based catalysts involving the
adsorption of the gas onto the catalyst surface followed by a dissociation of the
molecule and electrochemical reaction to two hydrogen ions as follows:

2Pt (s) + H2 → Pt−Hads + Pt−Hads (12.1)

Pt−Hads → H+ + e− + Pt(s) (12.2)

where Pt(s) is a free surface site and Pt–Hads is an adsorbed H-atom on the Pt active
site.
The overall reaction of hydrogen oxidation is:

H2 → 2H+ + 2e− (12.3)

The rate of the hydrogen oxidation process at the Pt-based catalyst at 80 ◦ C is very
high, as long as the catalyst surface is not contaminated by adsorbed impurities.
The highest performing PEMFC systems employ pure hydrogen as the fuel, but for
many applications, especially mobile, pure hydrogen is not yet a viable option due to
the technical difficulty of on-board storage and refuelling. Currently, the most viable
technology for on-site H2 generation is reforming technology (e.g. steam reform-
ing, partial oxidation or autothermal reforming) of hydrocarbons such as methanol,
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction 435

Fig. 12.2 Schematic presentation of a single typical proton exchange membrane fuel cell [1]. On
the anode, hydrogen gas is catalytically disassociated according to the reaction H2 → 2e− + 2H+ .
The hydrogen ions pass through the polymer electrolyte to the cathode where oxygen, in most cases
from air, is reduced (O2 + 4e− → O2− ). The overall reaction is: H2 + 1/2O2 → H2 O

gasoline or natural gas and alcohols that are readily available through existing distri-
bution channels [12]. Accordingly, the use of hydrocarbon or alcohol fuels requires
an external fuel processor (reformer) to be incorporated into the system. This item
not only increases the complexity and cost of the fuel cell system, but also decreases
the overall efficiency. The use of reformate fuel presents special challenges for a
PEM fuel cell in terms of system efficiency, reliability, and durability. Many of these
issues can be attributed to the detrimental effects of impurities such as CO, CO2 ,
NH3 and H2 S obtained as the by-products of the reforming process [13]. The hydro-
gen impurities due to manufacturing process are brought along with the fuel feed
stream into the anode of a PEMFC stack, causing performance degradation, and
sometimes permanent damage of the membrane electrode assemblies [14–16]. The
effect of different contaminants including: fuel impurities (CO,CO2 , H2 S, and NH3 ),
air pollutants (NOx , SOx , CO, and CO2 ), and cations resulting from the corrosion
of fuel cell stack system components (such are Fe3+ and Cu2+ ) on the efficiency of
PEM fuel cells was recently reviewed by Cheng et al. [17]. It was found that even
trace amounts of impurities present in either fuel or air streams or fuel cell system
components could severely poison the anode, membrane, and cathode, particularly
at low-temperature operation, what results in dramatic performance drop. Thus, elu-
cidation of the degradation mechanism of anode or cathode reactions by impurity
materials in polymer electrolyte fuel cells (PEMFCs) is a crucial topic, to attain its
longevity.
436 G. Postole and A. Auroux

Fig. 12.3 The performance losses caused by trace amounts (from 5 to 100 ppm) of carbon monoxide
in the fuel stream [6]

Poisoning of the anode catalyst is caused primarily by carbon monoxide, either

brought into the cell with the fuel feed stream or generated in situ by the reduction
of CO2 [9, 17–19]. The hydrogen produced by steam reforming, contains more than
1 % of CO [6, 13, 20]. Since a PEMFC cannot tolerate such high CO levels, the
carbon monoxide content is reduced through a series of high- and low-temperature
water-gas-shift (WGS) and preferential oxidation (PROX) reactions to bring the
CO level to less than 10 ppm [21]. However, due to the typical low temperature of
operation (ca. 80 ◦ C) and the choice of Pt as the electrocatalyst, the CO-poisoning
effect could significantly affect the long-term performance of the PEMFC stack,
even at this low CO level as it can be seen in Fig. 12.3 [6, 22, 23]. CO poisoning
on Pt electrocatalysts becomes more severe with increases in CO concentration and
exposure time. For example, Fig. 12.3 illustrates typical fuel cell stack polarization
curves obtained at 80 ◦ C in both the absence and presence of various concentrations
of CO [6]. The figure indicates that the CO impurities from fuel streams, even at a
level of a few ppm, can cause a substantial degradation in cell performance, especially
at high current densities.
This problem is therefore particularly severe with fuel feed streams derived from
the steam reforming of hydrocarbons, methanol or other liquid fuels, and is of lesser
concern when the PEMFC operates on neat hydrogen. Nevertheless, the level of CO
which brings about significant poisoning at the hydrogen anode in a PEMFC is so
small (several parts per million) that even in the case of relatively pure hydrogen feeds,
e.g., bottled hydrogen of nominal 99.99 % purity, some long-term platinum anode
performance loss has been observed [9]. Overcoming the CO poisoning problem is
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction 437

thus of paramount interest and needs to be addressed in order to make reformate gas
a viable fuel for PEM fuel cells. Extensive research has focused on gaining a better
understanding of poisoning effect that CO has on PEMFC performance. Baschuk
and Li [24] reviewed the CO poisoning of platinum electrocatalyts used in PEM
fuel cells in terms of characteristics, mechanism, mitigation, and theoretical models.
Many electrochemical, spectroscopic and theoretical studies [23, 25–31] have been
employed to investigate the CO adsorption process on platinum and Pt-based elec-
trocatalysts, including in situ attenuated total reflectance-Fourier transform infrared
(ATR-FTIR) study [32, 33], CO isotope exchange experiments [34], cyclic and strip-
ping voltammetry [35], etc. Although these studies provide valuable information for
the design of new catalysts, extrapolations of the conclusions with respect to the
CO tolerance to the real systems are not straightforward, mainly because of reaction
conditions that are different from the actual PEM fuel cell operating environment.
However, it has been shown without doubts that the CO poisoning effect was strongly
related to the concentration of CO, the exposure time to CO, the cell operation tem-
perature, and anode catalyst types.

12.3 CO Adsorption Microcalorimetry on Pt-Based

Materials: Literature Survey

Gas-solid interactions are fundamental for the understanding of adsorption, which

is the first step in a variety of processes in surface science. The characterization of
many gas-solid surface processes and the study of the gas interaction with heteroge-
neous surfaces (surfaces composed of different kinds of sites) are therefore important
aspects. One of the challenges in the field of gas-solid interactions is to envisage
methods for the determination of the energetic topography of heterogeneous sub-
strates from adsorption experiments. In this context, the adsorption microcalorime-
try technique coupled to volumetry is a powerful tool able to supply information
about the strength of gas-surface interactions [36–39]. Moreover, the adsorption
microcalorimetry also provides a direct measurement of heat of adsorption and their
evolutions with the coverage and can contribute in the study of all phenomena which
can be involved in one catalyzed process, e.g. activation/deactivation of the catalyst,
coke production, pore blocking, sintering, and adsorption of poisoning in the feed
gases [40, 41].
Chemisorption of carbon monoxide on transition metal surfaces (either single
crystals or supported clusters) is a tool of general use to study the active sites present
over this type of solid surfaces [42]. CO adsorption on noble metals has been the
subject of a large number of papers. For instance, the adsorption behaviour of CO on
surfaces of Pt single crystals, polycrystalline Pt films, and supported Pt catalysts has
been discussed in terms of adsorbed species or adsorption structures formed during
interaction of CO with the metal surface.
438 G. Postole and A. Auroux

The adsorption of CO on a series of platinum based catalysts was also carried out
by microcalorimetry technique, supplying information about the number, the strength
distribution and the heat associated to the carbon monoxide adsorption on available
Pt surface sites. Table 12.2 lists some literature reports on the average heats of CO
adsorption over different platinum supported catalysts. In this table, the experimental
data on powdered catalysts were collected from the vicinity of room temperature up
to 130 ◦ C. Prior to CO adsorption, the samples were reduced at 200 or 500 ◦ C under
hydrogen flow.
In these literature reports [43–58], the data obtained by adsorption microcalorime-
try are considered together with those obtained from complementary techniques
(i.e. infrared spectroscopy, temperature programmed desorption, X-ray photoelec-
tron spectroscopy) in order to elucidate the influence of loading and dispersion of
Pt, type of support material and the reduction temperature, on energetics and mech-
anism of CO adsorption on supported Pt catalysts for a better understanding of their
catalytic performances.
For example, by examining the initial and differential heats of adsorption mea-
sured on Pt/Al2 O3 powders calcined at different temperatures, Uner and Uner [50]
concluded that CO adsorption processes is not structure-sensitive. CO heats of
adsorption values obtained by the authors are plotted against carbon monoxide cov-
erage in Fig. 12.4. The heat of adsorption data for all catalysts fell on the same curve,

Table 12.2 Literature data of CO adsorption over supported Pt surfaces

Catalyst Pt (wt%) Temperature of Average heat of Reference
adsorption (◦ C) adsorption (kJ/mol)
Pt/C (Norit RX-3) 0.22 27 115 43
Pt/C (Vulcan) 5.00 27 110 44
Pt/graphite 2.00 57 115 45
Pt/C 1.00 25 80 46
PtSn/TiO2 2.00 25 105 47
Pt/TiO2 2.00 27 90 48
Pt/A2 O3 1.88 25 120 49
Pt/Al2 O3 2.00 30 135 50
Pt/Al2 O3 nano-fibre 2.91 30 120 51
Pt/η-Al2 O3 2.10 47 100 48
Pt/Al2 O3 3.00 57 130 42,52
Pt/Al2 O3 5.00 50 130 53
Pt/SiO2 10.00 27 135 54
Pt/SiO2 2.10 47 113 48
Pt/SiO2 1.20 130 130 55,56
Pt/SiO2 –Al2 O3 1.50 27 115 48
Pt/NaX zeolite 1.00 27 130 57
Pt/K-L zeolite 1.00 130 120 58
Pt/K-ZSM5 zeolite 1.00 130 110 58
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction 439

Fig. 12.4 Differential heat of carbon monoxide adsorption at 30 ◦ C over 2 % Pt/γ -Al2 O3 reduced
under hydrogen at 270 ◦ C. The temperatures indicate the calcination temperature of the catalysts
[50]

but the adsorbed gas amounts decreased with increasing the calcinations temperature
in agreement with the decrease in metal dispersions.
Serrano-Ruiz et al. [49] used the adsorption microcalorimetry of CO at room
temperature, X-ray photoelectron spectroscopy (XPS), and 119 Sn Mössbauer spec-
troscopy to study the effect of adding Sn to Pt/CeO2 –Al2 O3 and Pt/Al2 O3 catalysts.
Microcalorimetric analysis indicated that adding cerium caused the appearance of a
more heterogeneous distribution of active sites, whereas adding tin led to a higher
homogeneity of these sites. The influence of reduction conditions on the Pt–CO
adsorption strength have also been investigated in detail for Pt/Al2 O3 . The authors
[49] observed that the catalyst reduction at higher temperature caused a decrease of
the initial CO adsorption heat (from 140 to 120 kJ/mol) and of CO saturation coverage
from 55 to 45 µmolCO /gcat . Thus a higher homogeneity of the surface metal atoms
for CO adsorption was reported when catalyst was reduced at 500 ◦ C. The higher
initial heat of adsorption obtained in the sample reduced at low temperature (200 ◦ C)
was attributed to the interaction of CO with highly unsaturated metal atoms at cor-
ners and edges. The decrease of the total amount of CO adsorbed with increasing
the pretreatment temperature was explained to be due to sintering of the Pt particles
after high-temperature reduction.
The CO adsorption microcalorimetry was also used to explain the promoting effect
of Pt in bimetallic Ni–Pt catalysts supported on alumina nano-fibre (Alnf) tested for
the liquid phase reforming of sorbitol to produce hydrogen [51]. The differential
heat of adsorption for Ni–Pt/Alnf reduced to around 111 kJ/mol, which was 12 and
440 G. Postole and A. Auroux

6 kJ/mol lower than Pt/Alnf and Ni/Alnf respectively. This is substantial because
reducing the CO binding strength can avoid the poisoning of the active metal sites.
These results suggest that in the case of bimetallic catalysts there was a reduction in
the number of strong CO-adsorption sites.
The influence of support on CO–Pt interaction was studied by Vannice et al. [48]
using adsorption microcalorimetry. The supports used were SiO2 , η-Al2 O3 , SiO2 –
Al2 O3 , and TiO2 . The studied catalysts possessed a range of differential heats of
CO adsorption at 27 ◦ C between 88 and 135 kJ/mol (21 and 32 kcal/mol) with the
Pt/TiO2 sample having the lowest values. This variation in the values of differential
heats of adsorption appears to be a strong function of crystallite size, with weaker
Pt–CO bonding occurring as Pt dispersion increases.
Although one of the most commonly used catalysts today is Pt/Al2 O3 , the use of
carbonaceous materials as catalyst supports is continuously increasing. Their porous
texture can be easily tailored, yielding high surface-area supports where the active
phase can be well dispersed, and with the required pore-size distribution to facilitate
the diffusion of reactants and products to, and from, the active phase [59]. In the case
of noble-metal-based catalysts, the metal dispersion in the final catalyst depends on a
number of factors: porous texture of the support, nature of the metal precursor, types
and amount of surface complexes on the support, etc. [60]. Guerrero-Ruiz et al. [45]
applied CO adsorption microcalorimetry to study the factors affecting the Pt disper-
sions over high surface area graphite. Using different carbon supports (e.g. with, and
without, oxygen surface groups) and different platinum precursors (H2 PtCl6 and
Pt(NH3 )4 (OH)2 ) for the catalysts preparation, the authors showed that adsorption
microcalorimetry is one useful tool to provide some evidence concerning the spe-
cific interaction that takes place between graphite carbon and metal particles. Thus,
microcalorimetry of CO adsorption evidences that the presence of oxygen surface
groups diminishes the metal-support interaction having as results lower differential
heats of CO adsorption.
Heat of adsorption of a gas on a solid is, in general, composed of several contri-
butions, including energy of the formed surface bond, energy associated with per-
turbation (or even dissociation) of the adsorbate, energy of interactions between the
adparticles and energy associated with the surface relaxation or rearrangement [61].
Thus, in the basic research of adsorption, it is frequently desirable to separate the
individual effects. In order to facilitate such a separation, and to minimize the compli-
cating effects of the polycrystallinity and contamination of the surface, the adsorption
microcalorimetry is often carried out for the surfaces of well defined crystal structure.
Such surfaces can be obtained in the forms of metal filaments, vacuum-evaporated
thin films, and single crystals of metals. The comparison of adsorption heats obtained
by microcalorimetry over supported metallic clusters with those determined over
metallic single crystals can help to understand the actual surface structure of the
metal aggregates, and then their catalytic properties. Moreover, the adsorptive prop-
erties of CO on platinum single crystals can provide a better understanding of CO–Pt
interaction, what can bring important information for designing of new catalysts and
for the purpose of understanding the activity of Pt metal nanoparticles employed
in fuel cells in the size range of few nanometers. Information concerning the CO
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction 441

adsorption on single platinum crystals has been mainly gained from kinetic studies
and from spectroscopic techniques, whereas the direct determination of adsorption
heats has been less applied. Karmazyn and coworkers [62] studied, for example, the
chemisorption of CO on the stepped Pt{2 1 1} surface using first-principles density
functional theory (DFT) and microcalorimetry experiments. The heat of adsorption
of CO on Pt{2 1 1} were measured at room temperature as a function of coverage and
an initial heat of adsorption of 185 kJ/mol was found. Yeo et al. [63, 64] reported the
values of 180 and 215 kJ/mol for the initial CO heats of adsorption on unsupported
Pt{1 1 1} and Pt{1 0 0}, respectively. It can be observed that the heat of adsorption
data obtained for Pt single crystal surfaces with two different orientations is different,
beyond experimental errors.
The insight into the data presented in Table 12.2 gives evidence that it is difficult to
elucidate the structure dependency of CO chemisorptions from the heats of adsorp-
tion found over supported catalysts. Even though, based on literature reports it can
be concluded that the adsorption microcalorimetry, although not widely used, is a
powerful technique for surface characterization of supported metal clusters, because
it enables obtaining data concerning the strength of interaction and population of
active sites. Concerning CO–Pt interaction, it is evident that CO is chemisorbed
on all Pt-based catalysts (average heats of adsorption is higher than 100 kJ/mol,
see Table 12.2). The differential heat profiles are usually characterised by a plateau
of nearly constant heat of adsorption at low CO coverage, followed by an abrupt
decrease as the surface saturation limit is reached.

12.4 The CO Poisoning Effects on Pt/C Studied

by Adsorption Microcalorimetry: A Case Study

However, although the determination of the heat of adsorption of carbon monoxide

on platinum have been reported in the literature (Table 12.2), the possible use of
adsorption microcalorimetry as an available technique for giving valuable informa-
tion in PEM fuel cells studies has not been reported. In what follows, recent results
from our work in the field of microcalorimetry of CO adsorption on Pt/C catalysts
are presented [65–67]. The catalysts used in these studies were different commer-
cial carbon-supported platinum, with high Pt loading, aimed to be used in PEMFCs
applications. Particular emphasis on the sample history (preparation method, the
support material, the metal loading) and the pre-treatment of the catalysts on the CO
poisoning effect on supported platinum is paid.
The results of CO adsorption microcalorimetry described in this overview were
collected with a differential and isothermal microcalorimeter (Tian–Calvet Micro-
calorimeter) linked to a static volumetric system. The equipment permits the intro-
duction of successive small doses of CO onto the catalyst. Both the calorimetric and
the volumetric data were stored and analyzed by microcomputer processing. The
obtained data are presented as differential heats versus the amount of CO adsorbed
442 G. Postole and A. Auroux

while the amount of gas adsorbed at constant temperature are plotted as a function of
the equilibrium pressure, thus giving adsorption isotherms. By using microcalorime-
try, the structure sensitivity of hydrogen and carbon monoxide adsorption was inves-
tigated by using the Pt/C commercial catalysts with different Pt loading on carbon
support. The studied catalysts were thus Pt/C powders provided by E-Tek (lot#E
1280702, 16.8 wt% Pt), Tanaka (lot 103-1341R, 24.5 wt% Pt), and Johnson Mathhey
(lot 128372001, 16.6 wt% Pt) companies. As in a PEM fuel cell CO comes at the
anode surface as an impurity in the hydrogen flow, H2 adsorption microcalorimetry
was also applied for a better understanding of CO effect on Pt based catalysts.
Due to higher CO poisoning effect at lower temperatures, it is important to know
its effect at room temperature at which the start-up of the fuel cell system takes
place. It is why the adsorption studies were carried out at two different temperatures:
near room temperature (30 ◦ C) and at 80 ◦ C (the PEMFCs operation temperature).
Prior to the adsorption experiments, the catalysts were reduced under static hydrogen
(27 kPa) at 25, 100 or 200 ◦ C.
The obtained microcalorimetric results showed that both H2 and CO can be
chemisorbed on all Pt/C catalysts. Indifferent of their provenience (different carbon
support and different method of preparation), Pt/C exhibited significantly higher dif-
ferential heats of carbon monoxide adsorption in comparison with hydrogen adsorp-
tion as can be seen in Fig. 12.5.
It is thus evident that in case of co-adsorption of these two gases (similar situation to
that in fuel cell), CO will be primarily adsorbed. In addition, the adsorbed amount of
CO is much higher than the amount of adsorbed H2 , which means that the Pt based
catalysts present a larger number of sites active for carbon monoxide adsorption, in
comparison with those active for hydrogen adsorption. By comparison with results

200 CO adsorption

H2 adsorption

150
(kJ/mol)

100
diff
Q

50

0
0 50 100 150 200 250
probe molecule uptake (µmol/gcat )

Fig. 12.5 Differential heats of H2 and CO adsorption at 30 ◦ C as a function of surface coverage

for Pt/C sample (E-Tek, lot#E 1280702) pre-treated at 25 ◦ C in static hydrogen (27 kPa)
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction 443

obtained for different Pt/C catalysts [65, 67], no structure dependency was observed
for hydrogen initial heats of adsorption. Similar values were obtained for hydrogen
adsorption for all Pt/C catalysts, in good agreement with previously published values
for supported and unsupported platinum catalysts [56, 58, 68–70]. Higher values of
CO adsorption heats (average values ranged from 135 to 155 kJ/mol for the three
tested catalysts) where observed when compared with those reported in Table 12.2.
The difference observed in the adsorption heats values could be attributed to dif-
ferent experimental conditions used, different loading of Pt and possibly to some
remaining contaminants species (H2 , O2 , H2 O…) on the platinum, due to a poor
cleaning of its surface when the sample is pre-treated at room temperature (25 ◦ C).
As reported in the literature, platinum shows a preference for a linear mode of car-
bon monoxide adsorption [71]. Anyway, for low surface coverage, two IR bands
are observed from spectroscopic data of CO adsorbed on Pt supported powders: at
around 1870 and 2060 cm−1 [72–74]. These bands have been attributed either to CO
adsorbed on various sites of the metal [71] or to linear (2060 cm−1 ) and bridged
(1870 cm−1 ) forms of adsorbed CO [18, 75]. The microcalorimetric results, with
rather similar adsorption heat values in a large domain, can suggest that the main
adsorbed species, over all Pt/C catalysts, are linear, as may also be inferred from
spectroscopic data [71, 73, 74]. However, the presence of surface oxygen functional
groups (SOFGs) on the support could modify the strength of the Pt–CO bonds. Thus,
the differences observed in between Pt/C catalysts for CO adsorption were attributed
to the contribution of the support to the surface properties, knowing that electron-
donating supports produce an enrichment of the electron density of metal atoms
interacting with them [76], inducing changes in the chemisorption properties of the
metal particles. Nørskov et al. [77] have also demonstrated that the heats of adsorp-
tion of a species is directly related to the local structure of the catalysts, the step
sites are more active unless poisoned, and they bind the adsorbates more strongly.
The heats of adsorption are closely related to the adsorbate-substrate bond strength.
Furthermore, the differential heat of adsorption can be dependent on the surface cov-
erage of the adsorbate due to the lateral adsorbate-adsorbate interactions or due to
the surface heterogeneity [50].
In contrast to other techniques for studying adsorption, heat-flow calorimetry
yields both kinetic and thermodynamic information. The kinetics of heat release
during adsorption can be monitored by changes in the thermokinetic parameter,
τ . The calorimetric signal decreases exponentially with the adsorption time after
the maximum of each adsorption peak. This can be expressed in the form D(t) =
Dm exp(−t/τ ), where D(t) and Dm are the deviation of the calorimetric signal at
time t and the maximum deviation, respectively. The thermokinetic parameter τ in
this expression can thus be computed as the inverse of the slope of the logarithm
of the evolved heat curve during the return to equilibrium, and depends mainly on
the accessibility of the adsorption sites in the samples. Thus, it was found [65] that
the pore architecture of Pt/C catalysts (e.g. pore volume and pore size distribution)
influenced the kinetics of heat release during CO adsorption.
The accessibility of CO molecules to the adsorption sites increased with the meso-
porosity of the catalyst. When the Pt/C catalyst is predominantly mesoporous, the
444 G. Postole and A. Auroux

diffusion is not limited by movement of CO through micropores. More mesoporosity

facilitates the creation of larger paths for diffusion, leading to an increased accessi-
bility of CO molecules to the adsorption sites in the samples.
Different behaviour was also observed for Pt/C powders when their CO adsorption
properties were studied [66, 67] after the catalysts were reduced at various tempera-
tures. When the pre-treatment temperature was increased from 25 to 100 and further
to 200 ◦ C, the adsorptive properties of high loading Pt/C catalysts were influenced
as shown for example in Fig. 12.6 [66]. Increasing the pre-treatment temperature
provoked a decrease in the amount of chemisorbed carbon monoxide as well as in
the values of differential heats of CO adsorption for the sample provided by E-Tek.
These results were related with the changes in Pt particle sizes and dispersion as well
as to a better cleaning of the catalyst surface when a higher pre-treatment tempera-
ture was used. This demonstrates once more that adsorption microcalorimetry is a
powerful tool allowing the detection of different changes in the catalyst surface.
It has to be pointed out that not all the tested Pt/C samples presented the same
behaviour for CO adsorption with increasing temperature of pre-treatment. Obser-
vation of the calorimetric profiles (Fig. 12.7) reveals a plateau at around 135 kJ/mol
for Pt/C Tanaka, which did not change with increasing of reduction temperature.
The constant value of the plateau extends over a wide range of surface coverage,
thus indicating a high homogeneity of the surface metal atoms of this sample for
the CO adsorption. The drop in the differential heats at higher coverage is indicative
of saturation of the accessible platinum metallic surface sites. The changing of the
reduction conditions influenced only the total capacity (the length of plateau) of CO
adsorption, which decreased with increasing pre-treatment temperature, due to the
changes in Pt particle sizes and dispersion. Moreover, no differences in the adsorptive

200 Pretreatment at 25°C

Pretreatment at 100°C
Pretreatment at 200°C
150
Qdiff (kJ/mol)

100

50

CO adsorption
0
0 50 100 150 200 250
CO uptake (µmol/gcat )

Fig. 12.6 Differential heats of CO adsorption at 30 ◦ C as a function of surface coverage for Pt/C
catalyst (E-Tek, lot#E 1280702) pre-treated at 25, 100 and 200 ◦ C in static hydrogen (27 kPa) [66]
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction 445

200

150
Q diff (kJ/mol)

100

Red-25°C, Ads-30°C
50
Red-100°C, Ads-30°C
Red-200°C, Ads-30°C
Red-100°C, Ads-80°C
0
0 50 100 150 200 250 300 350

CO uptake (µmol/gcat )

Fig. 12.7 Differential heats of CO adsorption at 30 and 80 ◦ C as a function of surface coverage for
Pt/C catalyst (Tanaka, lot 103-1341R) pre-treated at 25, 100 and 200 ◦ C in static hydrogen (27 kPa)

properties of Pt/C catalyst for CO were observed when the adsorption temperature
was increased from 30 to 80 ◦ C and the same pre-treatment procedure was employed,
which was the case for all studied powders.
When the thermokinetic parameter was measured, a more detailed characteri-
zation of the available sites for CO adsorption on the Pt/C surface was provided.
For example, it was possible to differentiate reversible and irreversible adsorption
processes. Because chemisorption may be a slow, irreversible process, involving acti-
vation of the adsorbate, a longer time and, therefore, a broader thermogram would
distinguish such a process from a faster, reversible physisorption process. This feature
was thus exploited to monitor the change in adsorption with coverage. As presented
in Fig. 12.8, the adsorption process was initially slow and became slower, reaching a
minimum of rate, before a significant acceleration of the process which was observed
on approaching the physisorbed state at high coverage. The minimum rate appears
as a maximum in a plot of the thermokinetic parameter as a function of the surface
coverage, being indication of a change from irreversible to reversible adsorption.
As it can be seen in Fig. 12.8, the kinetic of CO adsorption on Pt based catalysts
did not change when the adsorption and the pre-treatment were performed at higher
temperature. As expected, the thermokinetic parameter firstly increases to reach a
maximum and then slowly decreases showing that CO is almost completely irre-
versibly adsorbed on the surface of Pt-based catalyst. The amount held by the strong
chemisorption sites at a certain adsorption temperature gives valuable information
about the catalysts behaviour towards poisoning. Indeed, as deduced from volumetric
data, 93 and 85 % of the total amount of CO was irreversibly adsorbed on Pt/C Tanaka
catalyst for the powder pre-treated at 25 ◦ C and CO adsorbed at 30 ◦ C, and for the
same sample pre-treated at 100 ◦ C and CO adsorbed at 80 ◦ C, respectively. Similar
446 G. Postole and A. Auroux

5
Red25°C, Ads 30°C
(min)
Red100°C, Ads 80C
Thermokinetic parameter,

2
0 50 100 150 200 250 300 350

CO uptake (µmol/gcat )

Fig. 12.8 Variation of the thermokinetic parameter, τ , versus CO uptake for Pt/C sample (Tanaka,
lot 103-1341R)

results were also obtained for the catalysts provided by E-Tek and Johnson Matthey
companies, with more than 90 % of CO irreversibly adsorbed for both powders.
A different and unexpected behavior was observed for the Pt/C catalyst provided
by Johnson Matthey showing that the results obtained by direct adsorption calori-
metric measurements provide accurate values of the heats of CO adsorption, and
that they are directly related with the catalyst history (carbon used as support, the
method of preparation). The calorimetric results obtained for this sample are given
in Fig. 12.9 and Table 12.3.
The differential heat profiles of CO adsorption presented in Fig. 12.9 are very
similar at low CO coverage (up to around 100 µmolCO /gcat ) in spite of different
reduction temperatures used. After a CO coverage of 100 µmolCO /gcat , when the
catalyst was pretreated at 25 and 100 ◦ C, an increase in the adsorption heats could be
observed with increasing CO uptake, before a drastic drop to the values corresponding
to weakly adsorbed CO (40 kJ/mol). The total metal stoichiometries at saturation are
not influenced by increasing the pre-treatment temperature.
As can be seen in Table 12.3, the irreversibly chemisorbed amount of CO (nirrev )
can also be measured from the calorimetric studies. Obviously, this volume corre-
sponds to the total amount held by the strong sites at the adsorption temperature
over the catalysts. In order to accurately determine the chemisorbed amount from
the overall adsorption isotherm, the catalyst was outgassed after the first adsorption
run at the same temperature to remove the physically adsorbed amount, after which
a new adsorption procedure was carried out to obtain a second isotherm. The differ-
ence between the first and second isotherm gives the extent of irreversible adsorption
(nirrev ) at a given temperature. However, in the first approximation, the magnitude of
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction 447

200

150
Q diff (kJ/mol)

100

50 Pretreatment at 25°C
Pretreatment at 100°C
Pretreatment at 200°C
0
0 50 100 150 200 250 300 350
CO uptake (µmol/gcat )

Fig. 12.9 Differential heats of CO adsorption at 30 ◦ C as a function of surface coverage for Pt/C
catalyst (Johnson Matthey, lot 128372001) pre-treated at 25, 100 and 200 ◦ C in static hydrogen
(27 kPa) [67]

Table 12.3 Data obtained from microcalorimetric measurements for Pt/C Johnson Matthey sample
(lot 128372001)
Temperature H2 adsorption at 30 ◦ C CO adsorption at 30 ◦ C
of reduction (◦ C)
D (%) Qinit a H2 uptake D (%) Qinit a CO uptake
(kJ/mol) (µmol/gPt ) (kJ/mol) (µmol/gPt )
ntotal b nirrev c ntotal b nirrev c
25 26.12 89 551 72 31.95 169 1413 1253
100 30.98 89 669 207 35.75 152 1744 1587
200 31.56 94 653 291 30.31 151 1490 1322
a Initial differential heats of H2 and CO adsorption; b Amount of H2 and CO adsorbed under
an equilibrium pressure of 27 Pa; c Amount of irreversibly chemisorbed H2 and CO under an
equilibrium pressure of 27 Pa

the heat of adsorption can be considered as a simple criterion to distinguish between

physical and chemical adsorption. As can be deduced from Table 12.3 which sum-
marizes nirrev , ntotal as well as the dispersion obtained from the total amounts of H2
and CO uptake at the monolayer, CO is almost completely irreversibly adsorbed on
the surface of the Pt/C catalyst (i.e. 87, 91 and 89 % for the sample pre-treated at 25,
100 and 200 ◦ C, respectively).
It was found that the available hydrogen adsorption sites have a broader site
energy distribution, while the intermediate and low adsorption sites are almost not
observed for carbon monoxide. The broader site energy distribution monitored by
H2 adsorption microcalorimetry can be attributed to the higher surface mobility of
hydrogen.
448 G. Postole and A. Auroux

From the data presented in Table 12.3, it is clear that at 30 ◦ C an irreversible form of
adsorbed hydrogen is detected on platinum, together with a reversible one for higher
coverage. It means that carbon monoxide was adsorbed on the catalyst previously
contacted with hydrogen. Even if the sample is evacuated before CO adsorption,
CO was adsorbed on a surface partly covered with strongly bonded hydrogen which
was not completely removed during outgassing under vacuum. It is evident from the
presented data that the strong adsorption of CO at the Pt surface can directly block
the surface active sites used for H2 adsorption. As reported also in the literature [17],
the representative mechanism of hydrogen adsorption containing small amounts of
carbon monoxide can be expressed as follows:

H2 + 2Pt → 2Pt−Hads (12.4)

CO + Pt → Pt−COads (12.5)

2CO + 2Pt−Hads → 2Pt−COads + H2 (12.6)

The adsorption of CO occurs not only at bare Pt sites through reaction (12.5) but
also at Pt hydride sites via reaction (12.6). This is not surprising, since the heats of
adsorption of CO on Pt are much higher than that of irreversibly adsorbed hydrogen.
As observed in Fig. 12.9, up to coverage of about 100 µmolCO /gcat , the CO mole-
cules are adsorbed on unoccupied active sites on the catalyst surface. At higher CO
coverage, the noticed increase of differential heats can be attributed to the replace-
ment of hydrogen irreversibly adsorbed by CO (reaction 12.6) which explains the
different platinum dispersion found from H2 and CO adsorption.
When the catalyst was pre-treated at a temperature of 200 ◦ C, no increase of
the heats was observed and similar values for the hydrogen and CO dispersion were
found. This means that with increasing activation temperature, CO adsorption occurs
on a reduced and clean surface. Therefore, for this catalyst, the temperature of pre-
treatment does not seem to influence the adsorption capacity, but plays a role in the
type of available sites on the surface. It has to be noted that the bell shaped profile
was not observed for the other investigated catalyst which means that the adsorption
properties are strongly dependent on the catalyst surface structure. Thus, the history
of the investigated materials is of importance: platinum is supported on different
carbons, in different quantities and possibly different preparation procedures have
been used; which influenced its dispersion and orientation. All these features can be
responsible as a whole for the different adsorptive properties of the studied materials.
But, in spite of different differential heats profiles observed, there is no doubt that
CO adsorbs quickly and essentially irreversibly on Pt even in the presence of pre-
adsorbed hydrogen on the surface of catalyst. In the absence of CO, hydrogen adsorbs
onto active platinum sites. Nevertheless, when only traces of CO are presented in the
anode gas mixture, it will gradually accumulate on the platinum surface through a
replacement reaction or a free site adsorption. Thus, the CO-poisoning arises from
the adsorption of CO molecules on the platinum catalyst sites, which avoids the
hydrogen to reach the platinum particles.
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction 449

200

150
Q diff (kJ/mol)

100

CO adsorption I
50 CO adsorption II
CO adsorption III
CO adsorption IV
0
0 50 100 150 200 250 300 350
CO uptake (µmol/g)

Fig. 12.10 Differential heats of CO adsorption at 80 ◦ C as a function of surface coverage for Pt/C
powder (Johnson Matthey, lot 128372001) pre-treated at 100 ◦ C in static hydrogen (27 kPa) [67]

Finally, the response to several cycles poisoning-recovering has also been assessed
by using the adsorption microcalorimetry. The poisoning degree of Pt/C and the cat-
alysts regeneration was followed during four successive CO adsorption/desorption/
readsorption cycles on the same bath of sample [65, 67]. For this, in between two
successive CO adsorption runs, the catalysts were kept under air overnight and again
reduced under static hydrogen before a new adsorption cycle. The results obtained
for Pt/C Johnson Matthey, are given as example in Fig. 12.10. Here, the CO adsorp-
tion was performed at 80 ◦ C after the catalyst was reduced at 100 ◦ C. As it can be
seen, this sample seems to be tolerant to CO poisoning in spite of the fact that CO
binds strongly on Pt and cannot be desorbed at the operation temperature of PEM
fuel cells (lower than 100 ◦ C). The same result was obtained when a similar study
was performed at room temperature [65]. This behaviour suggests that the platinum
morphology, structure and adsorbed CO species are restored or remain unchanged,
independently of the repeated exposure to air/H2 /CO. It was not the case for the
other studied samples as shown in the same reference [65]. Differences appeared
in the total amount of CO adsorbed after each cycle, which decreases considerably
at the end of the experiments. The degree of catalyst poisoning by CO upon suc-
cessive air/H2 /CO cycles varied between 2 and 30 % for different studied samples.
The decrease in CO adsorption capacity, which means a decrease in the number of
adsorption sites, can be attributed to an irreversible poisoning of the surface. No rela-
tion in between the loading of Pt and the tolerance to CO poisoning could be found.
These results confirm that the surface chemistry of the catalyst affects the surface site
energy distribution and consequently the adsorptive properties towards H2 and CO.
Thus it can be concluded that if the catalyst characteristics are well controlled, the
reformate gas can be used as fuel for operation in PEMFC but an exhaustive control
of CO must be anyway taken into account.
450 G. Postole and A. Auroux

12.5 Summary

The CO poisoning effect on PEM fuel cells efficiency is still under active study.
The detailed mechanism is still not well understood and there are contradictory
observations that need to be clarified. The goal of our work was to verify if the
adsorption microcalorimetry is an available technique for further understanding of
the effects of impurities on hydrogen activation and hydrogen surface coverage of
Pt/C catalysts used as anode in PEM fuel cells. It is evident that a correlation between
poisoning effects and the adsorptive properties exists, the rate of the former being
catalyst structure sensitive. The CO-poisoning can be a reversible process through
air bleed at the anode and the injection of clean hydrogen. Anyway it was shown that
the degree of recovery performances is directly related to the specific structure of the
surface metallic atoms of the catalyst. It was proven that microcalorimetry technique
is quite well developed and very useful in providing information on the strength and
distribution of active sites for CO adsorption on the catalyst surface.

References

1. L. Carrette, K.A. Friedrich, U. Stimming, Fuel cells-fundamentals and applications. Fuel Cells
1, 5–39 (2001)
2. S. Litster, G. McLean, PEM fuel cell electrodes. J. Power Sources 130, 61–76 (2004)
3. A. Boudghene Stambouli, E. Traversa, Solid oxide fuel cells (SOFCs): a review of an envi-
ronmentally clean and efficient source of energy. Renew. Sustain. Energy Rev. 6, 433–455
(2002)
4. B.C.H. Steele, Material science and engineering: the enabling technology for the commercial-
isation of fuel cell systems. J. Mater. Sci. 36, 1053–1068 (2001)
5. B.C.H. Steele, A. Heinzel, Materials for fuel-cell technologies. Nature 414, 345–352 (2001)
6. R.A. Lemons, Fuel cells for transportation. J. Power Sources 29, 251–264 (1990)
7. T. Yalcinoz, M.S. Alam, Improved dynamic performance of hybrid PEM fuel cells and ultra-
capacitors for portable applications. Int. J. Hydrogen Energy 33, 1932–1940 (2008)
8. O. Erdinc, M. Uzunoglu, Recent trends in PEM fuel cell-powered hybrid systems: investigation
of application areas, design architectures and energy management approaches. Renew. Sustain.
Energy Rev. 14, 2874–2884 (2010)
9. A. Bıyıkoğlu, Review of proton exchange membrane fuel cell models. Int. J. Hydrogen Energy
30, 1181–1212 (2005)
10. M.L. Perry, R.M. Darling, S. Kandoi, T.W. Patterson, C. Reiser, in Polymer Electrolyte Fuel
Cell Durability, ed. by F.N. Buchi, M. Inaba, T.J. Schmids (LLC, Springer Science and Business
Media, New York , 2009), p. 405
11. H. Liu, F.D. Coms, J. Zhang, H.A. Gasteiger, A.B. LaConti, in Polymer Electrolyte Fuel Cell
Durability, ed. by F.N. Buchi, M. Inaba, T.J. Schmids (LLC, Springer Science and Business
Media, New York , 2009), p. 80
12. A. Midilli, M. Ay, I. Dincer, M.A. Rosen, On hydrogen and hydrogen energy strategies I:
current status and needs. Renew. Sustain. Energy Rev. 9, 255–271 (2005)
13. R. Jiang, H.R. Kunz, J.M. Fenton, Influence of temperature and relative humidity on perfor-
mance and CO tolerance of PEM fuel cells with Nafion® -Teflon® -Zr(HPO4 )2 higher temper-
ature composite membranes. Electrochim. Acta 51, 5596–5605 (2006)
14. S. Gottesfeld, J. Pafford, A new approach to the problem of carbon monoxide poisoning in fuel
cells operating at low temperatures. J. Electrochem. Soc. 135, 2651–2652 (1988)
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction 451

15. J. Wu, X.Z. Yuan, J.J. Martin, H. Wang, J. Zhang, J. Shen, S. Wu, W. Merida, A review of PEM
fuel cell durability: degradation mechanisms and mitigation strategies. J. Power Sources 184,
104–119 (2008)
16. H.-F. Oetjen, V.M. Schmidt, U. Stimming, F. Trila, Performance data of a proton exchange
membrane fuel cell using H2 /CO as fuel gas. J. Electrochem. Soc. 143, 3838–3842 (1996)
17. X. Cheng, Z. Shi, N. Glass, L. Zhang, J. Zhang, D. Song, Z.-S. Liu, H. Wang, J. Shen, A
review of PEM hydrogen fuel cell contamination: impacts, mechanisms, and mitigation. J.
Power Sources 165, 739–756 (2007)
18. J.Z. Zhang, Z. Liu, The effect of low concentrations of CO on H2 adsorption and activation on
Pt/C. Part 1: in the absence of humidity. J. Power Sources 195, 3060–3068 (2010)
19. F.A. De Bruijn, D.C. Papageorgopoulos, E.F. Sitters, G.J.M. Janssen, The influence of carbon
dioxide on PEM fuel cell anodes. J. Power Sources 110, 117–124 (2002)
20. A. Pitois, A. Pilenga, G. Tsotridis, CO desorption kinetics at concentrations and temperatures
relevant to PEM fuel cells operating with reformate gas and PtRu/C anodes. Appl. Catal. A
374, 95–102 (2010)
21. B. Du, R. Pollard, J.F. Elter, M. Ramani, in Polymer Electrolyte Fuel Cell Durability, ed. by
F.N. Buchi, M. Inaba, T.J. Schmids (LLC, Springer Science and Business Media, New York ,
2009), pp. 341–366
22. L. Li, G. Wu, B.Q. Xu, Electro-catalytic oxidation of CO on Pt catalyst supported on carbon
nanotubes pretreated with oxidative acids. Carbon 44, 2973–2983 (2006)
23. H. Igarashi, T. Fujino, M. Watanabe, Hydrogen electro-oxidation on platinum catalysts in the
presence of trace carbon monoxide. J. Electroanal. Chem. 391, 119–123 (1995)
24. J.J. Baschuk, X.G. Li, Carbon monoxide poisoning of proton exchange membrane fuel cells.
Int. J. Energy Res. 25, 695–713 (2001)
25. K. Wang, H.A. Gasteiger, N.M. Markovic, On the reaction pathway for methanol and carbon
monoxide electrooxidation on Pt-Sn alloy versus Pt-Ru alloy surfaces. Electrochim. Acta 41,
2587–2593 (1996)
26. G. Garcıá, J.A. Silva-Chong, O. Guillén-Villafuerte, J.L. Rodríguez, E.R. González, E. Pastor,
CO tolerant catalysts for PEM fuel cells. Spectroelectrochemical studies. Catal. Today 116,
415–421 (2006)
27. J. Sobkowski, A. Czerwinski, Voltammetric study of CO and CO2 adsorption on smooth and
platinized platinum electrodes. J. Phys. Chem. 89, 365–369 (1985)
28. B. Beden, C. Lamy, N.R. de Tacconi, J. Arvia, The electrooxidation of CO: a test reaction in
electrocatalysis. Electrochim. Acta 35, 691–704 (1990)
29. M. Hachkar, T. Napporn, J.-M. LeÂger, B. Beden, C. Lamy, An electrochemical quartz crystal
microbalance investigation of the adsorption and oxidation of CO on a platinum electrode.
Electrochim. Acta 41, 2721–2730 (1996)
30. S.J. Lee, S. Mukerjee, E.A. Ticianelli, J. McBreen, Electrocatalysis of CO tolerance in hydrogen
oxidation reaction in PEM fuel cells. Electrochim. Acta 44, 3283–3293 (1999)
31. H. Steininger, S. Lehwald, H. Ibach, On the adsorption of CO on Pt(111). Surf. Sci. 123,
264–282 (1982)
32. H. Hanawa, K. Kunimatsu, H. Uchida, M. Watanabe, In situ ATR-FTIR study of bulk CO
oxidation on a polycrystalline Pt electrode. Electrochim. Acta 54, 6276–6285 (2009)
33. T. Sato, K. Kunimatsu, H. Uchida, M. Watanabe, Adsorption/oxidation of CO on highly dis-
persed Pt catalyst studied by combined electrochemical and ATR-FTIRAS methods. Part 1:
ATR-FTIRAS spectra of CO adsorbed on highly dispersed Pt catalyst on carbon black and
carbon un-supported Pt black. Electrochim. Acta 53, 1265–1278 (2007)
34. A. Pitois, J.C. Davies, A. Pilenga, A. Pfrang, G. Tsotridis, Kinetic study of CO desorption
from PtRu/C PEM fuel cell anodes: temperature dependence and associated microstructural
transformations. J. Catal. 265, 199–208 (2009)
35. A.-K. Meland, S. Kjelstrup, Three steps in the anode reaction of the polymer electrolyte mem-
brane fuel cell. Effect of CO. J. Electroanal. Chem. 610, 171–178 (2007)
36. A. Auroux, in Catalyst Characterization: Physical Techniques for Solid Materials, ed. by B.
Imelik, J.C. Vedrine (Plenum Press, New York, 1994), p. 611
452 G. Postole and A. Auroux

37. P.C. Gravelle, Application of adsorption calorimetry to the study of heterogeneous catalysis
reactions. Thermochim. Acta 96, 365–376 (1985)
38. B.E. Spiewak, J.A. Dumesic, Microcalorimetric measurements of differential heats of adsorp-
tion on reactive catalyst surfaces. Thermochim. Acta 290, 43–53 (1996)
39. V.E. Ostrovskii, Molar heats of chemisorption of gases at metals: review of experimental results
and technical problems. Thermochim. Acta 489, 5–21 (2009)
40. P.L. Llewellyn, G. Maurin, Gas adsorption microcalorimetry and modelling to characterise
zeolites and related materials. C. R. Chim. 8, 283–302 (2005)
41. G. Postole, A. Auroux, in Advances, in Fluid Catalytic Cracking: Testing, Characterization,
and Environmental Regulations, ed. by M.L. Occelli (CRC Press Taylor & Francis Group, Boca
Raton, 2010), pp. 199–256
42. A. Guerrero-Ruiz, A. Maroto-Valiente, M. Cerro-Alarcon, B. Bachiller-Baeza, I. Rodriguez-
Ramos, Surface properties of supported metallic clusters as determined by microcalorimetry
of CO chemisorption. Top. Catal. 19, 303–311 (2002)
43. J. Silvestre-Albero, J.C. Serrano-Ruiz, A. Sepulveda-Escribano, F. Rodriguez-Reinoso, Mod-
ification of the catalytic behaviour of platinum by zinc in crotonaldehyde hydrogenation and
iso-butane dehydrogenation. Appl. Catal. A 292, 244–251 (2005)
44. J.C. Serrano-Ruiz, A. Lopez-Cudero, J. Solla-Gullon, A. Sepulveda-Escribano, A. Aldaz, F.
Rodriguez-Reinoso, Hydrogenation of α, β unsaturated aldehydes over polycrystalline, (111)
and (100) preferentially oriented Pt nanoparticles supported on carbon. J. Catal. 253, 159–166
(2008)
45. A. Guerrero-Ruiz, P. Badenes, I. Rodriguez-Ramos, Study of some factors affecting the Ru
and Pt dispersions over high surface area graphite-supported catalysts. Appl. Catal. A 173,
313–321 (1998)
46. J.C. Serrano-Ruiz, A. Sepulveda-Escribano, F. Rodriguez-Reinoso, Bimetallic PtSn/C catalysts
promoted by ceria: application in the nonoxidative dehydrogenation of isobutene. J. Catal. 246,
158–165 (2007)
47. J. Ruiz-Martinez, A. Sepulveda-Escribano, J.A. Anderson, F. Rodriguez-Reinoso, Influence
of the preparation method on the catalytic behavior of PtSn/TiO2 catalysts. Catal. Today 123,
235–244 (2007)
48. M.A. Vannice, L.C. Hasselbring, B. Sen, Direct measurements of heats of adsorption on plat-
inum catalysts. J. Catal. 97, 66–74 (1986)
49. J.C. Serrano-Ruiz, G.W. Huber, M.A. Sanchez-Castillo, J.A. Dumesic, F. Rodriguez-Reinoso,
A. Sepulveda-Escribano, Effect of Sn addition to Pt/CeO2 -Al2 O3 and Pt/Al2 O3 catalysts: an
XPS, 119 Sn Mössbauer and microcalorimetry study. J. Catal. 241, 378–388 (2006)
50. D. Uner, M. Uner, Adsorption calorimetry in supported catalyst characterization: adsorption
structure sensitivity on Pt/γ -Al2 O3 . Thermochim. Acta 434, 107–112 (2005)
51. A. Tanksale, J.N. Beltramini, J.A. Dumesic, G.Q. Lu, Effect of Pt and Pd promoter on Ni sup-
ported catalysts-a TPR/TPO/TPD and microcalorimetry study. J. Catal. 258, 366–377 (2008)
52. A. Maroto-Valiente, I. Rodriguez-Ramos, A. Guerrero-Ruiz, Determination of the surface
states of metallic clusters supported on alumina using microcalorimetry of CO adsorption.
Thermochim. Acta 379, 195–199 (2001)
53. H. Kivrak, A. Mastalir, Z. Kiraly, D. Uner, Determination of the dispersion of supported Pt
particles by gas-phase and liquid-phase measurements. Catal. Commun. 10, 1002–1005 (2009)
54. R. Alcala, J.W. Shabaker, G.W. Huber, M.A. Sanchez-Castillo, J.A. Dumesic, Experimental
and DFT studies of the conversion of ethanol and acetic acid on PtSn-based catalysts. J. Phys.
Chem. B 109, 2074–2085 (2005)
55. R.D. Cortright, J.A. Dumesic, Microcalorimetric, spectroscopic, and kinetic studies of silica
supported Pt and Pt/Sn catalysts for isobutene dehydrogenation. J. Catal. 148, 771–778 (1994)
56. R.D. Cortright, J.A. Dumesic, Effect of potassium on silica-supported Pt and Pt/Sn catalysts
for isobutene dehydrogenation. J. Catal. 157, 576–583 (1995)
57. N.D. Gangal, N.M. Gupta, R.M. Iyer, Microcalorimetric study of the adsorption and reaction
of CO, O2 , CO+O2 , and CO2 on NaX zeolite, Pt/NaX, and platinum metal: effect of oxidizing
and reducing pre-treatment. J. Catal. 140, 443–452 (1993)
12 Adsorption Microcalorimetry as a Tool to Study the CO–Pt Interaction 453

58. S.B. Sharma, J.T. Miller, J.A. Dimesic, Microcalorimetric study of silica- and zeolite-supported
platinum catalysts. J. Catal. 148, 198–204 (1994)
59. F. Rodríguez-Reinoso, I. Rodríguez-Ramos, A. Guerrero-Ruiz, J.D. López-González, Platinum
catalysts supported on activated carbons. Part I: Preparation and characterization. J. Catal. 99,
171–183 (1986)
60. A. Sepúlveda-Escribano, F. Coloma, F. Rodríguez-Reinoso, Platinum catalysts supported on
carbon blacks with different surface chemical properties. Appl. Catal. A 173, 247–257 (1998)
61. S. Ćerný, Adsorption calorimetry on filaments, vacuum-evaporated films and single crystals
of metals. Thermochim. Acta 312, 3–16 (1998)
62. A.D. Karmazyn, V. Fiorin, S.J. Jenkins, D.A. King, First-principles theory and microcalorime-
try of CO adsorption on the 211 surfaces of Pt and Ni. Surf. Sci. 538, 171–183 (2003)
63. Y.Y. Yeo, L. Vattunoe, D.A. King, Calorimetric heats for CO and oxygen adsorption and for
the catalytic CO oxidation reaction on Pt111. J. Chem. Phys. 106, 392–403 (1997)
64. Y.Y. Yeo, L. Vattunoe, D.A. King, Energetics and kinetics of CO and NO adsorption on Pt100:
restructuring and lateral interactions. J. Chem. Phys. 104, 3810–3821 (1996)
65. G. Postole, S. Bennici, A. Auroux, Calorimetric study of the reversibility of CO pollutant
adsorption on high loaded Pt/carbon catalysts used in PEM fuel cells. Appl. Catal. B 92,
307–317 (2009)
66. G. Postole, S. Bennici, A. Auroux, Poisoning by CO of Pt/C commercial catalysts used in
PEMFCs. A microcalorimetric study, in Proceedings of International Symposium on Funda-
mentals and Developments of Fuel Cells (FDFC) 2011 Conference, ISBN: 978-2-7466-2970-7
(2011).
67. G. Postole, A. Auroux, The poisoning level of Pt/C catalysts used in PEM fuel cells by the
hydrogen feed gas impurities: the bonding strength. Int. J. Hydrogen Energy 36, 6817–6825
(2011)
68. S. Cerny, M. Smutek, F. Buzek, The calorimetric heats of adsorption of hydrogen on platinum
films. J. Catal. 38, 245–256 (1975)
69. J.B. Lantz, R.D. Gonzalez, Development of a new isothermal calorimeter; heats of hydrogen
adsorption on supported platinum versus crystallite size. J. Catal. 41, 293–302 (1976)
70. J.M. Herrmann, M. Gravelle-Rumeau-Maillot, P.C. Gravelle, A microcalorimetric study of
metal-support interaction in the Pt/TiO2 system. J. Catal. 104, 136–146 (1987)
71. P. Hollins, The influence of surface defects on the infrared spectra of adsorbed species. Surf.
Sci. Rep. 16, 51–94 (1992)
72. M. Primet, J.M. Basset, M.V. Mathieu, M. Prettre, Infrared study of CO adsorbed on Pt/Al2 O3 .
A method for determining metal-adsorbate interactions. J. Catal. 29, 213–223 (1973)
73. S.D. Jackson, B.M. Glanville, J. Willis, G.D. McLellan, G. Webb, R.B. Moyes, S. Simpson,
P.B. Wells, R. Whyman, Supported metal catalysts: preparation, characterization, and function.
Part II: carbon monoxide and dioxygen adsorption on platinum catalysts. J. Catal. 139, 207–220
(1993)
74. D. Liu, G.H. Que, Z.X. Wang, Z.F. Yan, In situ FT-IR study of CO and H2 adsorption on a
Pt/Al2 O3 catalyst. Catal. Today 68, 155–160 (2001)
75. M. Primet, J.M. Basset, M.V. Mathieu, M. Prettre, Infrared investigation of hydrogen adsorption
on alumina-supported platinum. J. Catal. 28, 368–375 (1973)
76. P. Gallezot, D. Richard, Selective hydrogenation of α, ß-unsaturated aldehydes. Catal. Rev.
Sci. Eng. 40, 81–126 (1998)
77. J.K. Nørskov, T. Bligaard, A. Logadottir, S. Bahn, L.B. Hansen, M. Bollinger, H. Bengaard,
B. Hammer, Z. Slivancanin, M. Mavrikakis, Y. Xu, S. Dahl, C.J.H. Jakobsen, Universality in
heterogeneous catalysis. J. Catal. 209, 275–278 (2002)
Chapter 13
Biodiesel: Characterization by DSC and P-DSC

Rodica Chiriac, François Toche and Christian Brylinski

The use of vegetable oils as engine fuel may seem negligible

today. Nevertheless, such oils may become, in the passing years,
as important as oil and coal tar presently.
Rudolf Diesel, 1912

Abstract Thermal analytical methods such as differential scanning calorimetry

(DSC) have been successfully applied to neat petrodiesel and engine oils in the last
25 years. This chapter shows how DSC and P-DSC (pressurized DSC) techniques
can be used to compare, characterize, and predict some properties of alternative
non-petroleum fuels, such as cold flow behavior and oxidative stability. These two
properties are extremely important with respect to the operability, transport, and
long-term storage of biodiesel fuel. It is shown that the quantity of unsaturated fatty
acids in the fuel composition has an important impact on both properties. In addition,
it is shown that the impact of fuel additives on the oxidative stability or the cold flow
behavior of biodiesel can be studied by means of DSC and P-DSC techniques. Ther-
momicroscopy can also be used to study the cold flow behavior of biodiesel, giving
information on the size and the morphology of crystals formed at low temperature.

13.1 Introduction

Nowadays, the world’s primary energy consumption is based on fossil fuels. How-
ever, the rising cost of petroleum oil, together with global warming, has led
researchers to develop alternative non-petroleum fuels such as biofuels. These are
divided in two main classes: bioethanol and biodiesel. The former is obtained from

R. Chiriac (B) · F. Toche · C. Brylinski

Laboratoire des Multimatériaux et Interfaces, Université Lyon 1, CNRS, UMR 5615,
43 boulevard du 11 Novembre 1918, 69622 Villeurbanne, France
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 455

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_13,
© Springer-Verlag Berlin Heidelberg 2013
456 R. Chiriac et al.

Table 13.1 Fatty acids profiles of the most common VOs and animal fats
Fatty acid Structure Common acronym Methyl Ester
Palmitic acid CH3 − (CH2 )14 − COOH C16 : 0 Methyl palmitate
Stearic acid CH3 − (CH2 )16 − COOH C18 : 0 Methyl stearate
Oleic acid CH3 − (CH2 )7 − CH = C18 : 1 Methyl oleate
CH − (CH2 )7 − COOH
Linoleic acid CH3 − (CH2 )7 − CH = C18 : 2 Methyl linoleate
CH − CH2 − CH =
CH − (CH2 )4 − COOH
Linolenic acid CH3 − (CH2 )7 − (CH = C18 : 3 Methyl linolenate
CH − CH2 )3 − COOH

sugar and starch crops and is used only in gasoline motors in its pure form (ethanol)
or as a gasoline additive. Bioethanol is mostly produced in the USA and Brazil. The
second class is obtained from vegetable oils, animal fats or recycled greases and is
used only in diesel engines mostly in Europe whether in its pure form, transformed
by transesterification, or as a diesel additive. More recently, it has been found that
micro-algae can also be a source of biofuel. Algae can produce bioethanol (from
biomass via hydrolysis and fermentation) and biodiesel (from algal oil). Algae fuel,
also called oilgae is a promising field because it does not require fertile land or food
crops [13, 14, 34, 40].
This chapter will however treat only biodiesel from vegetable oil (VO), the use of
which in internal combustion engines is not a recent innovation. Indeed, in 1900, at
the World Fair in Paris, the new diesel engine created by Rudolph Diesel (1858–1913)
was tested for the first time with peanut oil.
There are four main ways to make biodiesel: direct use and blending, microemul-
sions, thermal cracking (pyrolysis) and transesterification. Raw VO or recycled
greases (also called waste cooking oils) are not registered by the US Environmental
Protection Agency (EPA) as biodiesel for legal use in vehicles. Biodiesel is in fact
defined as a mixture of mono-alkyl esters of fatty acids derived from VOs or animal
fats. Fatty acids are represented by two numbers: the first denotes the total number
of carbon atoms in the fatty acid chain and the second is the number of double bonds
present in the chain. For example, C18:1 designates oleic acid, which has 18 carbon
atoms and one double bond. Table 13.1 lists the fatty acid profiles of a number of
common VOs and animal fats.
Pure VO has both a higher viscosity and auto-ignition temperature than those of
diesel fuel. Thus, one of the most common methods used to reduce oil viscosity in
the biodiesel industry is transesterification. Biodiesel is the product obtained when
a VO or animal fat is chemically reacted with an alcohol to produce fatty acid alkyl
esters. These reactions are often catalyzed by a base or acid [6, 7, 18, 37]. Bases
can catalyze the reaction by removing a proton from the alcohol thus making it
more reactive, while acids catalyze the reaction by donating a proton to the carbonyl
group, also making it more reactive. Glycerol is produced as a co-product and typical
“contaminants” are water, free glycerin, bonded glycerin, free fatty acids, catalysts,
soaps and oxidation products [31].
13 Biodiesel: Characterization by DSC and P-DSC 457

VO
(vegetable oils)

Transesterification
with methanol or ethanol
(catalyst NaOH)

B100 MEVO* EEVO* B100

+ Diesel +

Biodiesel
B10, B20, B30 ….. blends

*MEVO = Methyl Ester of Vegetable Oil

*EEVO = Ethyl Ester of Vegetable Oil

Fig. 13.1 Simplified schematic of how to obtain biodiesel and biodiesel blends

Figure 13.1 summarizes the most common way to obtain biodiesel and biodiesel
blends. Blends are denoted by acronyms such as B10, meaning 10 % of alkyl ester
of VO and 90 % petrodiesel. The B designates biodiesel. B100 means 100 % alkyl
ester of VO. Biodiesel blends are usually found at low percentages of alkyl ester of
VO rather than 100 %.
The suitability of biodiesel as a fuel comes from similarities to petrodiesel or at
least to some of its properties. The main properties are: the cetane number, flash
point, heat of combustion, and lubricity, etc. For example, the cetane number (CN)
is one of the most important properties of a diesel fuel. This indicator measures
how quickly the fuel ignites under diesel engine conditions. The CN test (ASTM D
613) compares the test fuel with a blend of two reference fuels, n-cetane (hexadecane,
with CN = 100, that ignites very easily under compression) and hepta-methyl-nonane
(iso-cetane, CN = 15), on their tendency to auto-ignite. Fuels with a high CN value
will have short ignition delays. Generally, diesel engines run well with fuel with a CN
value from 40 to 55. Most biodiesel fuels have a CN higher than that of petroleum-
based diesel fuels. Biodiesel from VO sources have been recorded as having a range
of 46 to 52, and animal-fat based biodiesels (more saturated esters) a CN range from
56 to 65.
The advantages of biodiesel over petrodiesel are summarized as follows [31]:
• derived from a renewable domestic source → reducing dependence on and pre-
serving petroleum;
• biodegradability;
• reduction of most exhaust emissions, except NOx ;
458 R. Chiriac et al.

• higher flash point, leading to safer handling and storage;

• excellent lubricity: B1 or B2 restores the lubricity of low-sulfur petrodiesel fuels,
which have greatly reduced lubricity.
Problems associated with biodiesel refer to:
• its higher price, which can be offset by reduced excise taxes or the use of less
expensive feedstock;
• stability when exposed to air (oxidative stability);
• cold flow properties.
In conclusion, biodiesel is overall a suitable substitute for petrodiesel fuel.
Many publications present research that has been done on biodiesel from pro-
duction to properties, such as oxidative stability, cold flow properties, and thermal
decomposition, etc. [2, 11, 32, 39, 46]. Differential Scanning Calorimetry (DSC),
Modulated Temperature-DSC and Pressure-DSC are among the techniques used to
characterize biodiesel. Some of these techniques will be discussed in the following
chapter.

13.2 Study of the Cold Flow Behavior of Biodiesel

by DSC and Thermomicroscopy

13.2.1 Introduction

The cold flow properties of biodiesel and conventional petrodiesel are extremely
important. Unlike gasoline, petrodiesel and biodiesel can both start to gel as the tem-
perature gets colder. If this happens, it can clog filters or even become too thick to
be pumped from the fuel tank to the engine. The high molecular weight paraffins (or
n-alkanes with C18–C30) of a diesel fuel crystallize with decreasing temperature.
Firstly, paraffin crystals appear and the liquid becomes cloudy, then with the decreas-
ing temperature crystals grow, agglomerate and stop flowing. The same occurs for a
liquid fatty material which becomes cloudy due to formation of crystals and solidifi-
cation of the saturated fatty acid methyl esters (chains with no double bonds) present
in biodiesel. This can be explained by the fact that these saturates have significantly
higher melting points than the unsaturated fatty compounds (chains with double
bonds), also present in biodiesel, and in a mixture they crystallize at a higher tem-
perature than unsaturates. Thus the cold flow behavior of biodiesel greatly depends
on the quantity and chemical structure of saturates.
The tendency of a (petrodiesel) fuel to gel or solidify at low temperature can
be quantified by several experimental parameters such as Cloud Point (CP), Cold
Filter Plugging Point (CFPP), and Pour Point (PP) [23, 31]. They are described in
more detail below. The American and corresponding French and European standard
methods are also given.
13 Biodiesel: Characterization by DSC and P-DSC 459

Cloud point is the temperature at which haziness is observed (small wax crystals
start to form and the fuel appears cloudy) when the sample is cooled at a specified
rate and examined at 1 ◦ C intervals. CP is an index of the lowest temperature of the
fuel’s utility under certain applications. Operating at temperatures below the CP for a
diesel fuel can result in fuel filter clogging due to the wax crystals. The repeatability
of the CP test is <0.5 ◦ C and the reproducibility is <2.6 ◦ C (using the standards:
ASTM D-2500 (or AFNOR T60-105)).
Cold filter plugging point is the lowest temperature at which a 20-mL sam-
ple passes through a 45-μm wire mesh filter under cooling at a specified rate and
0.0194 atm vacuum within 60 s (using the standard: ASTM D6371 (EN 116)).
Pour point represents the lowest temperature at which a fuel sample will still flow,
when the temperature is decreased at a specified rate below the CP and the crystals
are large enough (and agglomerate) to plug fuel filter systems. Examination of the
sample is made at 3 ◦ C intervals. The PP provides an index of the lowest temperature
of the fuel’s utility for certain applications. In addition, the PP has implications for
the handling of fuels at cold temperature. The repeatability of the PP test is <3 ◦ C
and the reproducibility is <6 ◦ C (ASTM D-97 (AFNOR T60-105)). Biodiesel fuels
derived from fats or oils which have significant amounts of saturated fatty compounds
will display high CP and PP values.
The definition of a maximum value for the CP, PP, and CFPP properties depends
on the fuel purpose, the region/country and season. In Europe, diesel fuels can be
classified in 10 classes as a function of their cold flow characteristics: six classes for
“temperate climates” with CFPP limits comprised between 5 and −20 ◦ C; and four
classes for “arctic regions” with the CFPP between −20 and −44 ◦ C. Each country
chooses one or more classes as a function of its climatic conditions. Thus, in France,
there are 3 classes for summer, winter and “extremely cold” periods with CFPP limits
of 0, −15 and −20 ◦ C, respectively [23]. Only a few approaches are possible in order
to achieve a given value of these properties for a fuel. One of them [40] combines
the high CP or PP fuel with fuels of lower CP. The second approach to improve the
cold flow behavior of a diesel fuel is either to decrease the distillation final point in
order to reduce the heavy fractions containing n-paraffins in fuel, or to choose more
naphthenic and aromatic fractions. However, these options are restrictive and costly
[23]. The third option [17] is to decrease the total saturated ester concentration in
biodiesel via winterization, but the yields are sometimes poor. Another approach uses
cold flow improvers or additives such as CFPP and PP depressants. These additives
are known to improve the paraffin crystals dispersion in order to avoid agglomerates
forming and so the plugging of filter pores. Generally, copolymers of ethylene and
vinyl acetate are used as CFPP and PP depressants at concentrations between 200
and 600 ppm; this diesel fuel treatment cost is relatively low at only a few cents per
liter of fuel [23].
There are also additives having an impact only on the CP. These are called cloud
point depressants and are obtained by a radical polymerization between an olefin and
acrylic or vinyl compound. These macromolecules prevent the formation of paraffin
crystals in solution. The use of all these additives necessitates new fundamental
460 R. Chiriac et al.

knowledge to be acquired relating to nucleation and crystal growth in biodiesel and

biodiesel blends [17].
Different standard methods are used to determine or predict the low temperature
behavior of a fuel in terms of CP, CFPP and PP. The repeatability and reproducibility
tests are not always very good (see above: <6 ◦ C) and other “non-standardized”
methods can be used such as DSC and thermomicroscopy. These thermal analytical
methods were successfully applied to neat petrodiesel and crude oils [9, 19, 26, 33]
but also to engine oils [12, 16, 17, 27, 35].
As the crystallization and melting of paraffins, or saturated fatty compounds in the
case of biodiesel, are accompanied by a low release of heat and adsorption of heat,
respectively, calorimetry can therefore be used to accurately detect and study the
process of crystallization and melting. Furthermore, DSC has the advantage of the
rapid determination of cold flow parameters from heating and cooling scans. Among
these parameters are glass transition temperatures and percentage of crystallizable
fraction from the heating scan, crystallization onset temperature, PP and percentages
of crystals at different temperatures ( i.e. −10 ◦ C, −20 ◦ C) from the cooling scan.
Claudy [9] and Heino [26] have shown that CP, PP and CFPP of diesel fuels can be
obtained by DSC with more accuracy, better reproducibility and more rapidly than
with standard tests. It has been found that the CP determined with standard tests
corresponds to the onset temperature of the crystallization peak obtained by DSC,
the latter being a few degrees lower than the CP.
Thermomicroscopy is useful for correlating morphological or structural changes
with the thermal effects observed by DSC as has been shown in other studies of
diesel and alternative fuels [12, 27, 33].

13.2.2 Experimental Procedures

13.2.2.1 Materials

Prior to DSC and thermomicroscopic measurements, gas chromatography analyses

were carried out for both palm oil and rapeseed oil methyl esters.
Blends of palm oil methyl ester and rapeseed oil methyl ester with diesel fuel
(DF) have been realized on a gravimetric basis. Blend ratios were 0, 10, 20, 30 and
100 %.
Palm and rapeseed biodiesels were mixed with cold flow improvers (additives)
after adding them to DF. The additives were mixtures of ethylene vinyl acetate (EVA)
copolymers and naphthenic distillates.

13.2.2.2 DSC Analyses

DSC measurements were carried out using a DSC 820 Mettler-Toledo apparatus.
Calibrations for temperature and enthalpy were performed using the melting point
13 Biodiesel: Characterization by DSC and P-DSC 461

and enthalpy of high purity metals and volatile organic compounds. The precision
of these temperature and enthalpy measurements was within ±0.2 ◦ C and 0.4 J/g,
respectively. The instrument was flushed with nitrogen.
For each scan, approximately 20 mg of sample was hermetically sealed in a 40 μL
aluminum pan and tested against an empty pan. Two scans (heating and cooling)
have been run for each sample. For the heating run, the samples were cooled and
held isothermally at −150 ◦ C for 2 min before heating to 50 ◦ C at 5 ◦ C/min, while for
the cooling run the samples were cooled from 25 to −30 ◦ C at a rate of 0.5 ◦ C/min.

13.2.2.3 Thermomicroscopic Analyses

The equipment used for the thermomicroscopic analyses was composed of: a Zeiss
Axioplan microscope using transmitted light and equipped with polarized light and
phase contrast devices; a LTS350 Heating/Freezing Stage unit which could operate
in the temperature range +350 to −50 ◦ C, controlled by a LINKAM TMS 94 control
unit; and an AxioCam MRc5 Zeiss colour camera and an AxioVision 4.1 software.
A Dewar flask filled with liquid nitrogen was used for cooling.
Samples were placed in an 8 μL glass crucible and analyzed in the temperature
range +20 to −15 ◦ C at a cooling rate of 0.5 ◦ C/min, the same as for DSC analyses.

13.2.3 DSC and Thermomicroscopy for the Study

of Biodiesel and Biodiesel blends

This section shows how DSC and thermomicroscopy can be used to study the cold
flow behavior of two biodiesels (palm oil methyl ester (ME1) and rapeseed oil methyl
ester (ME2)) and their blends with a conventional diesel fuel (DF). The impact of a
cold flow improver on the quantity and size of crystals is also presented.

13.2.3.1 Pure Compounds

Figures 13.2 and 13.3 show the heating and cooling curves obtained by DSC for pure
DF, ME1 and ME2.
The heating scan of DF (Fig. 13.2a) firstly shows an increase in heat capacity,
Cp, of 0.4 J g−1 K−1 , corresponding to the glass transition of the hydrocarbon
matrix at about −120 ◦ C (midpoint value), followed by a small exothermal effect due
to the crystallization of species that cannot crystallize on cooling. The same effect
is observed in Fig. 13.2b for ME1 and ME2 at −38 ◦ C and −34 ◦ C, respectively,
and on the DF scan, a broad endothermic effect is also seen, corresponding to the
dissolution of wax (n-alkanes + iso-alkanes) in the liquid matrix. This effect on
462 R. Chiriac et al.

the heating curves for both biodiesels (Fig. 13.2b) corresponds to the dissolution of
FAMEs (fatty acid methyl esters) in the composition of the ME1 and ME2 biodiesels.
Determination of the amount of paraffin, i.e. the crystallizable fraction (CF), on
the heating of DF, requires the use of the equation Hdiss = f(T ◦ C). The way
this equation is established is explained elsewhere [5, 8]. This relation between the
paraffin dissolution enthalpy and temperature should of course be reviewed when
paraffins and FAME precipitate together. For the latter, a new equation needs to be
assessed.
The cooling scan of the pure compounds (Fig. 13.3) shows a broad exother-
mal peak corresponding to the crystallization of paraffins (for DF) or FAME (for
biodiesels). The first cold flow parameters obtained by this analysis are the onset
crystallization temperature (CT) and the slope of crystallization. The CT is deter-
mined by the intersection of the tangents to the base line and the exothermal peak. The
cooling rate should be as low as possible to get close to the equilibrium temperature:
solid–liquid.
The analysis of each of the most abundant FAMEs by DSC at 5 ◦ C/min has
permitted us to identify the peaks obtained on the cooling and heating scans of
rapeseed and palm biodiesels. Table 13.2 lists the FAME composition of rapeseed
(ME2) and palm (ME1) biodiesels as well as the melting and crystallization onset
and enthalpy. Methyl oleate and elaidate are the predominant FAMEs in ME1 and
ME2 at 53.2 % and 57.1 %, respectively. For palm biodiesel, methyl palmitate was
the next most abundant FAME (27.4 %), followed by methyl stearate (9.4 %) and
methyl linoleate (6.4 %). Concerning the rapeseed biodiesel, methyl linoleate was
the next most abundant FAME (24.5 %) after methyl oleate, followed by methyl
linolenate (8.6 %) and methyl palmitate (4.6 %). Rapeseed biodiesel (ME2) has a
higher amount of unsaturated fatty acids (UFA, ≈92 %) than palm biodiesel (UFA,
≈61 %). The total saturated fatty acid (SFA) content for palm ester was ≈38 %, and
≈7 % for rapeseed ester.
As shown in Table 13.2, the unsaturates present lower melting and crystallization
points (temperatures and enthalpies) than saturates. Therefore, it can be assumed
that the cold flow parameters will be better (i.e. low CPs and PPs) for the rapeseed
oil methyl ester as it contains more unsaturated compounds compared to the palm
oil methyl ester. This affirmation is also supported by Knothe [29] who stated that

(a) (b)
exo exo
ME1

0.2
Diesel Fuel
Wg^-1
1
Wg^-1

ME2

-140-130-120-110-100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 °C
-140 -130-120-110-100 -90 -80 -70 -60 -50 -40 -30 -20 -10 -0 10 20 30 40°C
22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 min
22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 min

Fig. 13.2 Heating scans for pure compounds: a DF; b ME1 & ME2
13 Biodiesel: Characterization by DSC and P-DSC 463

(a) (b)
exo exo
0.2
Wg^-1

ME1 ME1
5
Wg^-1

0.05 ME2
Wg^-1
ME2

DF
18 16 14 12 10 8 6 4 2 0 -2 -4 -6 -8 -10 -12 -14-16-18 -20-22-24 -26 °C 20 10 -0 -10 -20 -30 -40 -50 -60 -70 -80 -90 -100 -110 -120 °C
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 min 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15min

Fig. 13.3 Cooling scans for pure compounds. a DF, ME1, ME2 cooled at 0.5 ◦ C/min until −30 ◦ C.
b ME1 & ME2 cooled at 5 ◦ C/min until −130 ◦ C

biodiesel fuels derived from fats or oils with significant amounts of saturated fatty
compounds will display higher CPs and PPs.
The cooling scan of pure compounds (ME1 and ME2) at 5 ◦ C/min (Fig. 13.3b)
shows two broad exothermic peaks. As the melting and crystallization properties of
FAME have been obtained and presented in Table 13.2, the identification of saturated
and unsaturated compounds on these DSC cooling curves can therefore be made. The
saturated ones start crystallizing at 6 and −7 ◦ C for ME1 and ME2, respectively,
while the unsaturates crystallize at very low temperatures (about −40/−50 ◦ C).
Palm oil methyl ester presents an onset of crystallization above 0 ◦ C (i.e. 6 ◦ C)
and a steep crystallization slope (Fig. 13.3a). Thus, its cold flow properties will be
poorer than those of the rapeseed methyl ester. The crystallized fractions (CFs)
determined at −15 ◦ C for ME1, ME2 and DF are: 2.3, 1.5 and 25.9 %, respectively.
These percentages were calculated using the enthalpy of crystallization of DF (210
J/g) obtained elsewhere [9]. The same linear relation obtained for diesel fuels has
been used for biodiesel and biodiesel blends:

Hcryst. = 1.5167T + 211.7 Jg−1 (with T in◦ C)

Even if the enthalpy of 210 J/g is not the corresponding value for the crystallization
enthalpy of ME1 and ME2 it has been chosen in order to compare the variation of
the crystallized fraction for all studied blends. The ME2 and DF have similar low
percentages of crystals at −15 ◦ C. However, ME1 shows a high percentage of crystals
at −15 ◦ C. The cold flow parameters of pure compounds are listed in Table 13.3. The
CFPP and PP were determined by DSC at 0.45 and 1 % of crystals, respectively,
using the same Hcryst as mentioned above. Claudy et al. [9] realized a study on 40
different diesel fuels and established correlations between CFPP and PP determined
by ASTM standards and CFPP and PP obtained by DSC measurements. The best
correlations were obtained at 0.45 and 1 % of crystals, respectively. In our study, the
same values have been retained in order to compare the samples.
Palm oil methyl ester has a high % of crystals at −15 ◦ C and high CT, PP and CFPP
values compared with DF and ME2. This is due to the presence, in high quantity, of
saturated fatty esters (no double bonds) which have high melting and crystallization
points as shown in Table 13.2. Thus, from these results, we can conclude that ME1
464 R. Chiriac et al.

Table 13.2 FAME compositions and DSC parameters for palm and rapeseed methyl esters
Compounds MW ME1a ME2a Melting Crystallization
(g/mol) (wt)% (wt)%
Onset ◦ C H J/g Onset ◦ C H J/g
Methyl Palmitate 270.45 27.4 4.6 28.6 214.9 24.3 −214.1
(C16 : 0)
Methyl Palmitoleate 268.43 0.0 0.2 n.a. n.a. n.a. n.a.
(C16 : 1n9c)
Methyl 284.48 0.1 0.0 n.a. n.a. n.a. n.a.
Heptadecanoate
(C17 : 0)
Methyl Stearate 298.5 9.4 1.8 38.3 229.9 34.8 −219.1
(C18 : 0)
Methyl Oleate (cis) 296.49 53.2 57.1 −20.4 160.2 −38.3 −127.7
& Elaidate 8.1 189.8 2.8 −192.0
(trans) (C18 :
1n9c)
Methyl Linoleate 294.47 6.4 24.5 −42.7 107.2 −77.5 −92.3
(C18 : 2n6c)
Methyl Linolenate 292.46 0.5 8.6 −52.9 80.4 −78.8 −70.8
(C18 : 3n3c)
Methyl Arachidate 326.56 0.5 0.5 47.1 242.8 42.0 −228.2
(C20 : 0)
Methyl cis-11- 324.54 0.4 1.4 −15.6 138.7 −21.1 −143.4
Eicosenoate
(C20 : 1)
Methyl Behenate 354.61 0.2 0.4 n.a. n.a. n.a. n.a.
(C22 : 0)
Methyl Erucate (C22 352.59 0.1 0.2 n.a. n.a. n.a. n.a.
: 1n9c)
Methyl Lignocerate 382.66 0.4 0.1 n.a. n.a. n.a. n.a.
(C24 : 0)
Total non identified − 1.5 0.5 − − − −
compounds
Total saturated − 38 7.4 − − − −
compounds
Total unsaturated − 60.6 92 − − − −
compounds
a determined by GC; n.a. non-analyzed

cannot be used as a conventional diesel fuel in its pure form in spite of its high cetane
number (54).
To visualize the crystallization evolution, the pure compounds were analyzed by
thermomicroscopy. During cooling, three photos were taken for each compound at
different temperatures (Fig. 13.4). ME2 and DF show continuous and homogenous
crystallization of about 20 and 80 μm, respectively, but with no crystal agglomeration.
In contrast, ME1 behaves differently with crystals agglomerating and freezing over
only a 0.4 ◦ C period (from 5.5 to 5.1 ◦ C). These observations show that ME1 cannot
be used as a diesel fuel as confirmed by the DSC results.
13 Biodiesel: Characterization by DSC and P-DSC 465

DF
(photos at
-8,-10, -15°C)

ME2
(photos at
-10.3,-12.7, -15°C)

ME1
(photos at
5.5, 5.3, 5.1°C)

Fig. 13.4 Evolution of crystallization of pure compounds with decreasing temperature by ther-
momicroscopy

Table 13.3 Cold flow parameters of pure compounds (Hcryst. = 210 J/g)
CT (◦ C) CFPP (◦ C) PP (◦ C) % of crystals
at 0.45 % crystals at 1 % crystals at −15 ◦ C
DF −6.9 −8.1 −8.5 2.3
ME2 −7.0 −10.4 −13.5 1.5
ME1 6.0 5.9 5.8 25.9

13.2.3.2 Diesel Fuel and Biodiesel Blends

The cold flow properties of diesel fuel and biodiesel blends are given in Table 13.4.
As can be seen from these results, there is almost no influence when ME2 is blended
with DF. This situation changes when ME1 is blended with DF, where the quantity
of precipitated crystals and the cold flow parameters are higher than those of neat
DF and increase when ME1 is increased from 10 to 30 %. As was expected after
the analyses of the pure compounds, blends of ME1 with DF have lower cold flow
parameters compared with blends of ME2 with DF.
The blends of ME1 and ME2 with DF have also been tested by thermomicroscopy.
The results show that there is no impact on the morphology and the size of crystals
when ME2 is added to DF (Fig. 13.5).
Concerning the blends of ME1 with DF, the crystals size increases and there is
agglomeration as the quantity of ester increases from 10 to 30 % (see Fig. 13.6).
In addition, the morphology of crystals is not uniform. The DF/ME1 70/30 shows
a large increase in crystal size of about 120–150 μm, knowing that in DF they are
466 R. Chiriac et al.

Table 13.4 Cold flow parameters of DF and biodiesel blends by DSC (Hcryst. = 210 J/g)
CT (◦ C) CFPP ( ◦ C) PP ( ◦ C) % of crystals at
at 0.45 % crystals at 1 % crystals −15 ◦ C
Diesel fuel −6.9 −8.1 −9.5 2.3
DF/ME1 90/10 −5.3 −6.3 −8.7 2.7
DF/ME1 80/20 −5.9 −6.4 −7.9 3.7
DF/ME1 70/30 −5.0 −5.3 −5.9 5.8
DF/ME2 90/10 −6.3 −7.1 −9.6 2.0
DF/ME2 80/20 −5.6 −6.8 −10.0 1.9
DF/ME2 70/30 −6.3 −7.1 −10.4 1.7

DF (T=-15°C) ME2 (T=-15.3°C)

DF/ME2 90/10 (T=-15°C) DF/ME2 80/20 (T=-15°C) DF/ME2 70/30 (T=-15°C)

Fig. 13.5 Visualization of crystallization of pure compounds (DF and ME2) and corresponding
biodiesel blends at −15 ◦ C by thermomicroscopy

about 80 μm. Thus, this blend cannot be used in diesel engines unless a cold flow
additive is used to reduce the crystal size.

13.2.3.3 DF and Biodiesel Blends: Impact of Cold Flow Improver

For any diesel fuel, the crystal size at low temperature is a critical parameter. Indeed,
the standard tests for the determination of cold flow parameters are done on a 45 μm
diameter filter. In reality, the injection pump filter pores can have different sizes
depending on the vehicle make, varying from approximately 1–45 μm. Thus, the
size of paraffin or FAME crystals needs to be drastically reduced.
The same analytical techniques as those presented before are used to study the
impact of 200 ppm of a cold flow improver (additive “A”) on the morphology of
13 Biodiesel: Characterization by DSC and P-DSC 467

DF (T=-15°C) ME1 (T=+5.1°C)

DF/ME1 90/10 (T=-15°C) DF/ME1 80/20 (T=-15°C) DF/ME1 70/30 (T=-15°C)

Fig. 13.6 Visualization of crystallization of pure compounds (DF and ME1) and corresponding
biodiesel blends at −15 ◦ C by thermomicroscopy

DF/ME2 90/10 DF/ME2 80/20 DF/ME2 70/30

↓ ↓ ↓

DF/ME2 90/10 + A DF/ME2 80/20 + A DF/ME2 70/30 + A

Fig. 13.7 Comparison of crystal morphology at −15 ◦ C between biodiesel blends (DF/ME2) with
and without cold flow improver (additive A)

precipitated crystals. The DSC results permitted the following conclusions to be

drawn:
• no significant impact on the quantity of crystals was obtained in the presence of A;
• a slight decrease of the crystallization slope was observed for all studied blends,
except for the blend DF/ME1 70/30, knowing that a small crystallization slope
signifies the formation of crystals with small size. The latter result is confirmed by
thermomicroscopy as shown in Figs. 13.7 and 13.8, where the photos were taken
at −15 ◦ C.
468 R. Chiriac et al.

DF/ME1 90/10 DF/ME1 80/20 DF/ME1 70/30

↓ ↓ ↓

DF/ME1 90/10 + A DF/ME1 80/20 + A DF/ME1 70/30 + A

Fig. 13.8 Comparison of crystal morphologies at −15 ◦ C between biodiesel blends (DF/ME1) with
and without cold flow improver (additive A)

The crystal size in the blend DF/ME2 greatly decreased from 80 (without A) to
5 μm (with 200 ppm of A) for B10, but it slowly increased from 5 to 20 μm when
the percentage of ME2 was increased from 10 to 30 %.
It has been shown that the addition of additive to the DF/ME1 blends led to a
decrease in crystal size in the three studied blends, but less significantly for those
with 20 and 30 % of ester. The crystals were also more heterogeneous. In the presence
of additive A, the size of the precipitated crystals decreased from 60 to 5 μm for B10,
from 60 to 40 μm for B20 and from 100 to 30 μm for B30. Additive addition clearly
gives the best results for the blend with 10 % of ester (B10).

13.2.4 Conclusion

The cold flow properties of biodiesel and conventional diesel fuel are extremely
important. If with decreasing temperature the fuel begins to gel, it can clog filters or
even become too thick to be pumped from the fuel tank to the engine. We therefore
characterized the cold flow behavior of biodiesel and biodiesel blends by two com-
plementary techniques: DSC and thermomicroscopy. The former has permitted us
to quantitatively determine the precipitated crystals and other cold flow parameters,
whereas, the size and morphology of crystals were obtained by thermomicroscopy.
We observed that the rapeseed methyl ester (ME2) with a high percentage of unsatu-
rated compounds, presents very good cold flow behavior, because these compounds
crystallize at lower temperatures than saturates. The latter are responsible for poor
cold flow properties, thus their proportion should be controlled.
13 Biodiesel: Characterization by DSC and P-DSC 469

The palm oil methyl ester (ME1) and its blends with DF have shown poor cold
flow behavior especially for B30 and this can be explained by the fact that this ester
naturally contains more saturated fatty acids (38 %) than, for example, ME2 (7 %) .
These saturated compounds have higher melting and crystallization points.
By adding 200 ppm of a cold flow improver, the size of crystals was greatly
reduced for all the biodiesels and biodiesel blends studied. Therefore, the cold flow
behavior was improved, but no impact on the quantity of crystals was obtained. This
means that the additive used only had an impact on the nucleation and growth of
particles.

13.3 Oxidative Stability of Biodiesel by P-DSC

13.3.1 Introduction

Although biodiesel is an excellent source of fuel, it has some disadvantages which

include relatively poor cold flow properties and difficulty in maintaining fuel quality
during long-term storage. The latter is a matter of oxidation stability and will be
discussed in this part.
Degradation of the quality of this alternative fuel leads to an increase of deposits
on injectors and pump parts and therefore an increase in pressure across filters [31].
Biodiesel is easily subject to oxidation under ambient conditions. This is mainly due
to the presence of double bonds in the chains of fatty compounds. However, these
unsaturated chains with low melting points lead to a good cold flow behavior (see
Sect. 13.1), which is essential for a biodiesel.
The reaction rate of oxidation depends on the number and position of double
bonds [24, 31, 44]. The greater the level of unsaturation in a fatty oil or ester, the
more susceptible it will be to oxidation. The CH2 , allylic or bis-allylic to the double
bonds, is the site of the first attack. After hydrogen is removed from such a carbon, a
carbon-based free radical is formed. Then, oxygen quickly attacks to form a peroxy
radical. This is sufficiently reactive to remove another hydrogen from a carbon chain
to form another radical and a hydroperoxide. This is a chain reaction including a
set of reactions such as initiation, propagation, and termination that can proceed
rapidly after the initial induction period has occurred. Once the hydroperoxides have
formed (primary oxidation), they decompose and react forming numerous secondary
oxidation products such as aldehydes, alcohols, shorter chain carboxylic acids (which
increase biodiesel acidity) and polymers (which increase biodiesel viscosity).
Other factors can have an impact on the oxidative stability of fatty oils and/or
esters, such as, water, metals and light, etc. The presence of water can cause hydrol-
ysis reactions and damage the fuel quality. Moreover, water can be dispersed in the
biodiesel by mono-, di-glycerides and glycerol (co-products of the esterification).
These compounds can play the role of emulsifier [3]. Furthermore, the presence of
metals, even trace levels, can catalyze the oxidation reaction. Copper [30] has the
470 R. Chiriac et al.

strongest catalytic effect, but iron, nickel and tin also reduce the oxidative stability.
Besides this, the presence of a commonly-used additive, fuel acid corrosion inhibitor,
can greatly increase the formation of secondary oxidation products. When only 20
ppm of additive in a B50 blend was added, it has been found that polymeric gums
were formed [45]. In addition, some polymers are soluble in biodiesel, due to its polar
nature, and become insoluble when mixed with diesel fuel [4]. Some polymers are
also produced in the biodiesel when the temperature reaches 250−300 ◦ C (thermal
polymerization). The mechanism is a Diels-Alder reaction [48]: a conjugated diene
group from one fatty acid chain reacts with a single olefin group from another fatty
acid chain to form a cyclohexene ring. Exposure to light is another factor accelerat-
ing oxidation of fatty oils and esters. This process, called photo-oxidation involves
a different mechanism whereby oxygen directly attacks the olefinic carbons [44].
In order to prevent oxidation reactions, chemicals called antioxidants (AOs) are
added to the fuel. There are two types depending on their action during the reaction
process: some decompose hydroperoxides (e.g. zinc dithiophosphates, organophos-
phates, and organothio derivatives), and others break the reaction chain (phenolic
and amine compounds). Most publications treat the latter type and they will also
be discussed in this study. Such compounds contain highly labile hydrogen that is
more easily removed by a peroxy-radical than the hydrogen of a fatty oil or ester.
The resulting antioxidant free radical is either stable or reacts further to form a stable
molecule that does not contribute to the chain oxidation process, as shown here:

ROO• + AH → ROOH + A•

A• → stable products

Some AOs can be naturally present in vegetable oils (VOs), such as tocopherols, or
chemically synthesized. Tocopherol is a phenolic compound that is found in four
isomeric forms (α, β, γ, and δ). They all occur naturally in VOs but not in animal-
derived fats (except at trace levels). Several studies have shown that the naturally-
occurring levels of tocopherols are optimized with respect to AO ability. In general,
further addition of tocopherols has either no further benefit or may even decrease the
oxidative stability [42]. The commonly synthesized AOs are:
• 3-t-butyl-4-hydroxyanisol (BHA)
• 2, 6-di-t-butyl-4-methylphenol (BHT)
• 2-t-butylhydroquinon (TBHQ)
• 1, 2, 3-trihydroxybenzene (Pyrogallol or PY)
AO effectiveness is generally measured by stressing a fatty oil or ester both with and
without AO and comparing the results. The methods used will be discussed in the
following paragraphs.
A literature study has shown that AOs have different effects depending on the
origin of the biodiesel [43]. The volatility of the AOs is also a critical parameter
because part of the additive may be lost depending on the procedures used. The
volatility of 3 AOs can be ranked as follows: BHT > BHA > TBHQ [44]. One study
13 Biodiesel: Characterization by DSC and P-DSC 471

Fig. 13.9 Mechanisms of synergism between two antioxidants as proposed by Guzman [25]

reports that the addition of AO drastically improves the oxidative stability of B100
biodiesel up to a concentration of 1000 ppm [15]. Further additions do not reveal
any significant benefit. Moreover, solids appeared in a B10 blend when 3000 ppm of
BHT were added.
Cooperative effects (positive synergy) between two phenolic antioxidants in
biodiesel have been brought to light. One of the mechanisms proposed by
Guzman et al. [25], points the regeneration of PY by TBHQ (Fig. 13.9a). PY being
more reactive, it easily donates the hydrogen from its hydroxyl group to fatty acid
free radicals creating an antioxidant radical in the process. TBHQ then transfers
hydrogen to the antioxidant radical to regenerate PY. As TBHQ was converted
to a radical, it can then form stable products with other free radicals. The other
mechanism (Fig. 13.9b) suggests the formation of a heterodimer. The as-made phe-
nolic compound may have a better efficiency than the parent antioxidants.
In the same way, a study of lubricating oils [1] reported a synergy between amine
and phenolic AOs. The same mechanisms could be applied, replacing PY by the
amine AO. The proportions for optimum synergy should be a low amount of amine
AO compared to that of the phenolic one, e.g. 1:4 ratio for binary blends. The same
study also showed a positive influence of some detergent on the oxidative stability
of the oil.
The most common method used to determine the oxidative stability of fatty com-
pounds is Rancimat (EN14214) also called OSI (Oxidation Stability Index). This
test method consists in bubbling air through a sample maintained at 110 ◦ C. This
air is then passed through deionized water while the conductivity is measured. The
sudden increase in water conductivity is due to soluble secondary oxidation products
(volatile organic acids). The time at which the conductivity drastically increases is
the induction time.
The oxidative stability of diesel fuel is determined by the standard method ASTM
D2274 or NF M 07-047, which consists in bubbling air through the sample main-
tained at 95 ◦ C for 16 h. The solids formed are then dried and weighed. A maximum
weight is set by the specification. These tests are however time-consuming. Thus,
it has become necessary to develop new oxidation stability tests. In preliminary
investigations, numerous pressurized DSC (P-DSC) methods have been developed
to determine the oxidation stability of engine oil lubricants [47]. As lubricants are
subject to harsh conditions, i.e. high temperature, pressure, atmosphere, and con-
472 R. Chiriac et al.

tact with metal, etc., P-DSC was found to be a good way to simulate these condi-
tions. In addition, P-DSC can be used to assess the resistance to oxidation during
storage. Accelerating the oxidation process by a pressurized oxygen atmosphere is
one way to determine the stability of a fuel during storage. The conditions are not
strictly the same, but this method could help to discriminate different types of fuels
and observe the impact of additives, such as antioxidants or detergents. However,
the P-DSC methods (crucible material—aluminum or platinum, sample mass, type
of gas—oxygen or air, and pressure applied) can vary from one study to another
[10, 21, 38, 39]. Nevertheless, standard methods for hydrocarbons allow us to deter-
mine two critical parameters by P-DSC: the oxidation onset temperature (OOT using
the ASTM E 1858-08 standard) and the oxidation induction time (OIT with the
ASTM E 2009-08 standard). These methods recommend using aluminum crucibles
and sample masses comprises between 3 and 3.30 mg. Therefore, the use of P-DSC
is a suitable technique to determine oxidation stability of biodiesel.

13.3.2 Experimental Procedure

Samples of conventional diesel fuel (DF), 3 rapeseed oil methyl esters and 3 antiox-
idants, BHT, TBHQ and EAP (mixture of N, N’-di-sec-butyl-p-phenylenediamine
(40–60 %) 2, 6-di-t-butylphenol (<40 %) and tri-t-butylphenol (<10 %)) were used
for this study.
P-DSC experiments were carried out using a High Pressure DSC827 from Mettler-
Toledo. Approximately 3.5 ± 0.3 mg of sample was placed in a sealed 40 μL alu-
minum pan with a pinhole. The analysis cell was pressurized (20 bars) under static
oxygen atmosphere.
To determine the OOT, a dynamic mode was used with a temperature range from
25 to 230 ◦ C and heating rate of 5 ◦ C min−1 . On the curve, the intersection of the
tangents of the baseline and the exothermic peak corresponds to the OOT.
To determine the OIT, an isothermal mode was used where the isothermal tem-
perature (140 ◦ C) was reached at a heating rate of ∼= 120 ◦ C min−1 . On the curve,
the time between the beginning of the isothermal step and the onset of the oxidation
peak corresponds to the OIT.
The P-DSC module was calibrated in terms of heat flow and temperature using
naphthalene (T f usion = 80.1◦ C, H f usion = 147.64 Jg−1 ), indium (T f usion =
156.6◦ C, H f usion = 28.45 Jg−1 ), and tin (T f usion = 231.9◦ C, H f usion =
60.10 Jg−1 ) standards.
Repeatability was assessed by testing the same sample X, 3 times under the
conditions described previously for both methods. Using Eq. (13.1), the average (X )
and standard deviation (σ ) were calculated while the Eq. (13.2) shows the variation
coefficient (VC). The results are given in Table 13.5 for both the OOT and OIT.
13 Biodiesel: Characterization by DSC and P-DSC 473

Table 13.5 Repeatability results for both P-DSC methods

Dynamic method OOT (◦ C) Isothermal method OIT (min)
Test 1 193.1 86.0
Test 2 192.4 85.9
Test 3 192.6 86.3
Average 192.7 86.1
Standard deviation 0.4 0.2
VC (%) 0.2 0.3


 n
1  1
n
σ = (X i − X )2 with X = X i and n = 3 (13.1)
n n
i=1 i=1

σ × 100
VC = (13.2)
X
From Table 13.5, one can conclude that both methods are repeatable when they
are carried out under identical conditions and by the same operator and with a freshly
prepared sample (i.e. not more than a few days between tests).

13.3.3 Results and Discussions

P-DSC was used to compare different methyl esters of vegetable oil with a conven-
tional diesel fuel, and to evaluate the efficiency of AO. In this study both methods,
dynamic and isothermal, were used for every sample in order to determine OOT and
OIT values.
First, P-DSC experiments were carried out on two rapeseed oil methyl esters
(RME1 and RME2) and a conventional diesel fuel (DF). OOT and OIT values are
given in Table 13.6. As expected, due to its chemical composition, DF is much more
stable than both RME1 and RME2. Furthermore, RME1 is more stable to oxidation
than RME2 (Fig. 13.10).
The two RMEs were blended with DF to make B7 blends. As shown in Table 13.6,
the OOT of DF decreases in the presence of methyl esters (also illustrated in
Fig. 13.11). The OIT of B7_1 is double that of B7_2. Therefore, B7_1 has a better
stability to oxidation than B7_2, and both B7s are less stable than neat DF. Further-
more, the exothermal peak of DF shows a smaller oxidation slope compared with
those of the two B7s. This is due to the kinetic of reaction which is slower for the
oxidation of DF than in the case of B7s. This is expected because of their different
chemical composition.
AOs were added to the ester and then blended with DF to obtain the B7 mixture
with a concentration of 1000 ppm of AO. These mixtures were analyzed by P-DSC in
terms of their OOT and OIT. Two AOs were tested: BHT and TBHQ. The presence
474 R. Chiriac et al.

RME2 RME1 DF
RME1 exo exo

RME2
0,2 10
Wg^-1 Wg^-1

100 °C 40 60 80 100 120 140 160 180 200 220 200 150 100 50°C
50 100 150 200 250 300 350 400 450 500 550 600 650 700 min 0 5 10 15 20 25 30 35 40 45 50 55 min

Fig. 13.10 P-DSC curves of DF and pure esters under dynamic and isothermal conditions

Table 13.6 OOT and OIT values of neat and blended compounds
OOT ◦ C OIT min (at 140 ◦ C)
DF 195.0 −
RME1 168.4 20.2
RME2 143.4 n.a.
DF+7%RME1 186.6 41.3
DF+7%RME2 181.9 20.6
Note n.a. = non analyzed because of its OOT value of 143.4 ◦ C, quite similar with that of
measurement.

exo exo DF+RME1

DF+RME2 DF

DF+RME1
DF+RME2
20
0,1
Wg^-1
Wg^-1 DF

140 °C 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 °C
20 40 60 80 100 120 140 160 180 200 220 min 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 min

Fig. 13.11 P-DSC curves of diesel fuel and B7 blends under dynamic and isothermal conditions

of AO (with BHT better than TBHQ) greatly improves the oxidative stability of both
biodiesel blends, as the OOT and OIT values have increased (Table 13.7).
An interesting point in this study is the relatively poor performance of TBHQ
compared to BHT. It has been well-established that TBHQ is highly effective for
fatty oils, esters and animal fats while BHT is more effective for diesel fuel and
lubricants [15, 22, 36, 44]. Two reasons for this are that, first, the greatly different
structures of fatty oils compared to non-polar hydrocarbons (fuels, mineral oil or
lubricants) may interact with the highly hindered polar phenol group of BHT to
13 Biodiesel: Characterization by DSC and P-DSC 475

Table 13.7 OOT and OIT values of B7 with and without AOs
OOT ◦ C OIT min (at 140 ◦ C)
RME1 RME2 RME1 RME2
B7 186.6 181.9 41.3 20.6
B7+TBHQ 1000 ppm 191.5 186.1 57.3 31.2
B7+BHT 1000 ppm 196.5 192.6 79.9 74.4

Table 13.8 OOT and OIT values of a B7 with and without AOs
OOT◦ C OIT min (at 140 ◦ C)
B7 186.5 22.0
B7+TBHQ 1000 ppm 193.2 35.8
B7+BHT 1000 ppm 196.5 103.5
B7+EAP 1000 ppm 211.1 319.9

exo BT+EAP exo

B7+BHT
0,2
Wg^-1
B7 B7+TBHQ
50
B7+TBHQ
B7+EAP Wg^-1 B7
B7+BHT

140 °C 120 130 140 150 160 170 180 190 200 210 220 230 210 190 170 150 130 110 °C
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 min 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 min

Fig. 13.12 P-DSC curves of a B7 with different AOs under dynamic and isothermal conditions

reduce its AO capacity. Second, BHT is relatively volatile, and under the conditions
used for many of the procedures much of the additive may be lost during the early
parts of the tests [22, 28, 36]. The blends used in this study being only 7 % ester in
diesel fuel, it is understandable that BHT works better as AO than TBHQ.
Table 13.8 gives the results comparing 3 AOs (BHT, TBHQ and EAP) tested on
a B7 (also illustrated in Fig. 13.12). B7 was different from the two other blends
previously presented, but was still a rapeseed oil methyl ester. Addition of BHT and
TBHQ led to improvements in OOT and OIT values. It is clear that EAP significantly
improves oxidative stability as the OOT increased by more than 24 ◦ C and the OIT
by almost 300 min. This could be explained by its composition: one amine and two
phenolic AOs, the whole mixture giving a positive synergy [1]. Thus, the three AOs
tested can be classified as a function of their performance as follows:

EAP >> BHT > TBHQ

476 R. Chiriac et al.

13.3.4 Conclusion

During long-term storage biodiesel can be easily subject to oxidation under ambient
conditions due to the presence of double bonds in the chains of fatty compounds. The
reaction of oxidation increases with the increasing of unsaturation level (polyunsat-
urated compounds are many times more reactive that monounsaturated ones). Some
parameters such as acid value, peroxide value or kinematic viscosity increase when
autoxidation occurs, thus the fuel’s quality is affected. Degradation of the quality of
this alternative fuel leads to an increase of deposits on injectors and pump parts and
therefore an increase in pressure across filters.
The work herein has shown the suitability of P-DSC as an accelerated oxidation
stability test method for biodiesel. Different biodiesel blends (B7) were compared
in terms of their oxidative stability by P-DSC using dynamic (linear heating up to
oxidation) and isothermal modes (maintaining constant temperature at a fixed value
up to oxidation). The OOT and OIT parameters were determined for each studied
blend and the positive influence of “chain-breaking” type additives (such as amine
and/or phenolic compounds) has been shown. BHT showed better results (high OOT
and OIT values) than TBHQ, but EAP, which is a mixture of 1 amine and 2 phenolic
compounds, was by far the best AO. Synergetic effects could explain these very good
results.
Nowadays, pressurized DSC has become a very adapted technique to discriminate
different samples in terms of oxidative stability or to study the impact of additives
on this property. The results are obtained rapidly and with better repeatability and
accuracy than standard methods. Furthermore, P-DSC can also be used to determine
the oxidative stability of lubricants, paintings, rubbers, polymers etc.

13.4 Overall Conclusion

In this chapter, it has been shown that DSC and P-DSC are both rapid and accurate
methods to compare, characterize and predict some properties of alternative biodiesel
fuels, such as cold flow behavior and oxidative stability which are closely related.
These two properties are extremely important with respect to the operability, trans-
port and long-term storage of biodiesels and diesel fuels. Therefore, much attention
should be paid to them. Indeed, the quantity of unsaturated fatty acids (compounds
with double bonds and very low melting points) has an important impact on both
properties. A high proportion of these compounds improves the cold flow behavior,
but also increases the reactivity and thus reduces the stability to oxidation. A good
compromise must therefore be found, and therefore the selection of the appropriate
cold flow improvers and antioxidants must be done with great care. Other tech-
niques such as thermomicroscopy can be employed to study the cold flow behavior
of biodiesel (i.e. the study of the evolution of paraffin and FAME crystallization with
information about the size and the morphology of crystals), such analytical technique
being complementary with DSC.
13 Biodiesel: Characterization by DSC and P-DSC 477

The chemical composition of a fuel is very complex and generally a lot of additives
are added in order to improve some properties such as viscosity, oxidative stability,
cold flow behavior, cetane number, color or smell etc. The impact on fuel properties
of a complex mixture of miscellaneous additives is not always very well known. It has
been already shown that it can bring a positive or negative synergy. Therefore, DSC
and P-DSC can be very useful in evaluating with high accuracy and in a reasonable
time the impact of a package of additives on the oxidative stability or the cold flow
behavior of an alternative or conventional diesel fuel.

References

1. J.Z. Adamczewska, C. Love, Oxidative stability of lubricants measured by PDSC CEC L-85-
T-99 test procedure. J. Therm. Anal. Calorim. 80, 753–759 (2005)
2. Y. Ali et al., Fuel properties of tallow and soybean oil esters. J. Am. Oil Chem. Soc. 72 (1995)
3. P. Bondioli, L. Folegatti, Evaluating the oxidation stability of biodiesel. An experimental con-
tribution. Riv. Ital. Sostanze Grasse 73, 349–353 (1996)
4. P. Bondioli, A. Gasparoli et al., Evaluation of biodiesel storage stability using reference meth-
ods. Eur. J. Lipid Sci. Technol. 104, 777–784 (2002)
5. F. Bosselet, Application de l’analyse calorimétrique différentielle à l’étude des gazoles moteurs
et de leur amélioration par dopage. Thesis, Science/Chimie-Saint-Étienne, (1984), pp. 1–85
6. L. Bournay et al., New heterogeneous process for biodiesel production: A way to improve the
quality and the value of the crude glycerine produced by biodiesel plants. Catal. Today 106,
190–192 (2005)
7. D.G. Cantrell, L.J. Gillie et al., Structure-reactivity correlations in MgAl hydrotalcite catalysts
for biodiesel synthesis. Appl. Catal. A General 287, 183–190 (2005)
8. P. Claudy, J.M. Létoffé et al., Crude oils and their distillates: characterization by differential
scanning calorimetry. Fuel 67, 58–61 (1988)
9. P. Claudy, J.M. Létoffé et al., Diesel fuels: determination of onset crystallization temperature,
pour point and filter plugging point by differential scanning calorimetry. Correlation with
standard test methods. Fuel 65, 861–864 (1986)
10. M.M. Conceiçao, M.B. Dantas et al., Evolution of the oxidative induction time of the ethilic
castor biodiesel. J. Therm. Anal. Calorim. 97, 643–646 (2009)
11. J. Cvengros et al., Production and treatment of rapeseed oil methyl esters as alternative fuels
for diesel engines. Bioresour. Technol. 55 (1996)
12. M. De Torres, G. Jiménez-Osés et al., Fatty acid derivatives and their use as CFPP additives in
biodiesel. Bioresour. Technol. 102, 2590–2594 (2011)
13. A. Demirbas, Biodiesel from oilgae, biofixation of carbon dioxide by microalgae: A solution
to pollution problems. Appl. Energy 88, 3541–3547 (2011)
14. M.F. Demirbas, Biofuels from algae for sustainable development. Appl. Energy 88, 3473–3480
(2011)
15. R.O. Dunn, Effect of antioxidants on the oxidative stability of methyl soyate (biodiesel). Fuel
Process. Technol. 86, 1071–1085 (2005)
16. R.O. Dunn, M.O. Bagby, Low-temperature filterability properties of alternative diesel fuels
from vegetable oils, in Proceedings of the third liquid fuel conference: Liquid fuel and industrial
products from renewable resources, American Society of Agricultural Engineers, St. Joseph,
MI, (1996), ed. by J. S. Cundiff et al., pp. 95–103
17. R.O. Dunn, Thermal analysis of alternative diesel fuels from vegetable oils. J. Am. Oil Chem.
Soc. 76, 109–115 (1999)
18. A.K. Endalew et al., Heterogeneous catalysis for biodiesel production from Jatropha curcas oil
(JCO). Energy 36, 2693–2700 (2011)
478 R. Chiriac et al.

19. H.R. Faust, The thermal analysis of waxes and petrolatums. Thermochim. Acta 26, 383–398
(1978)
20. C.S. Foon, Y.C. Liang et al., Crystallisation and melting behavior of methyl esters of palm oil.
Am. J. Appl. Sci. 3, 1859–1863 (2006)
21. L.M.S. Freire, T.C. Bicudo et al., Thermal investigation of oil and biodiesel from Jatropha
curcas L. J. Therm. Anal. Calorim. 96, 1029–1033 (2009)
22. M.E. Gordon, E. Mursi, Comparison of oil stability based on the metrohm rancimat with storage
at 20 C. J. Am. Oil Chem. Soc. 71, 649–651 (1994)
23. J.C. Guibet, Chapitre 4.5. In: Carburants et moteurs (Tome 1). (Editions OPHRYS, Paris, 1997)
24. F.D. Gunstone, T.P. Hilditch, The union of gaseous oxygen with methyl oleate, linoleate and
linolenate. J Chem Soc, 836–841 (1945)
25. R. Guzman, H. Tang et al., Synergistic effects of antioxidants on the oxidative stability of
soybean oil- and poultry fat-based biodiesel. J. Am. Oil Chem. Soc. 86, 459–467 (2009)
26. E.L. Heino, Determination of cloud point for petroleum middle distillates by differential scan-
ning calorimetry. Thermochim. Acta 114, 125–130 (1987)
27. J.C. Hipeaux, M. Born et al., Physico-chemical characterization of base stocks and thermal
analysis by differential scanning calorimetry and thermomicroscopy at low temperature. Ther-
mochim. Acta 348, 147–159 (2000)
28. A.W. Kirleis, C.M. Stine, Retention of synthetic phenolic antioxidants in model freeze-dried
food systems. J. Food Sci. 43(5), 1457–1460 (1978)
29. G. Knothe, Dependence of biodiesel fuel properties on the structure of fatty acid alkyl esters.
Fuel Process. Technol. 86, 1059–1070 (2005)
30. G. Knothe, R.O. Dunn, Dependence of oil stability index of fatty compounds on their structure
and concentration and presence of metals. J. Am. Oil Chem. Soc. 80, 1021–1026 (2003)
31. G. Knothe, J. Van Gerpen, J. Krahl, The Biodiesel Handbook (AOCS Press, Urbana, 2005)
32. I. Lee et al., Reducing the crystallization temperature of biodiesel by winterizing methyl soyate.
J. Am. Oil Chem. Soc. 73(5), 631–636 (1996)
33. J.M. Létoffé, P. Claudy et al., Antagonism between cloud point and cold filter plugging point
depressants in a diesel fuel. Fuel 74, 1830–1833 (1995)
34. R. Maceiras et al., Macroalgae: Raw material for biodiesel production. Appl. Energy 88, 3318–
3323 (2011)
35. F. Noel, Thermal analysis of lubricating oils. Thermochim. Acta 4, 377–392 (1972)
36. M. Peled, T. Gutfinger, A. Letan, Effect of water and BHT on stability of cottonseed during
frying. J. Sci. Food Agric. 26, 1655–1666 (1975)
37. G.R. Peterson et al., Rapeseed oil transesterification by heterogeneous catalysis. J. Am. Oil
Chem. Soc. 61(10), 1593–1597 (1984)
38. F.M.G. Rodrigues, A.G. Souza, I.M.G. Santos et al., Antioxidative properties of hydrogenated
cardanol for cotton biodiesel by PDSC and UV/Vis. J. Therm. Anal. Calorim. 97(2), 605–609
(2009)
39. N.A. Santos, J.R.J. Santos et al., Thermo-oxidative stability and cold flow properties of Babassu
biodiesel by PDSC and TMDSC technique. J. Therm. Anal. Calorim. 97(2), 611–614 (2009)
40. P.M. Schenk, S.R. Thomas-Hall et al., Second generation biofuels: High-efficiency microalgae
for biodiesel production. Bioenergy Res. 1, 20–43 (2008)
41. Y. Serpemen, F.W. Wenzel et al., Blending technology key to making new gasolines. Oil Gas
J. 89(11) (1991)
42. E.R. Sherwin, Oxidation and antioxidants in fat and oil processing. J. Am. Oil Chem. Soc. 55,
809–814 (1978)
43. H. Tang, A. Wang et al., The effect of natural and synthetic antioxidants on the oxidative
stability of biodiesel. J. Am. Oil Chem. Soc. 85, 373–382 (2008)
44. J.A. Waynick, Characterisation of biodiesel oxidation and oxidation products. The Coordinated
Research Council, Southwest Research Institute Project N◦ 08–10721, San Antonio, Texas
(2005)
45. J.A. Waynick, Evaluation of the stability, lubricity, and cold flow properties of biodiesel fuel,
in Proceedings of the 6th International Conference on Stability and Handling of Liquid Fuels,
Vancouver, B.C., Canada, pp. 805–829 (1997)
13 Biodiesel: Characterization by DSC and P-DSC 479

46. G. Wenzel et al., Boiling properties and thermal decomposition of vegetable oil methyl esters
with regard to their fuel suitability. J. Agric. Food Chem. 45 (1997)
47. A. Zeman, H.P. Binder, Determination of spontaneous ignition temperatures (SITs) of aviation
lubricants using pressure differential calorimetry (PDSC). Thermochim. Acta 98, 159–165
(1986)
48. O. Zovi, L. Lecamp et al., Stand reaction of linseed oil. Eur. J. Lipid Sci. Technol. 113, 616–626
(2011)
Chapter 14
CO2 Capture in Industrial Effluents.
Calorimetric Studies

Jean-Yves Coxam and Karine Ballerat-Busserolles

Abstract In order to reduce environmental impact of CO2 emissions, one possible

option is the decarbonation of the effluents coming from fixed sources. A description
of the different techniques proposed for a separation of CO2 from gaseous effluents
is explained with a focus on post-combustion processes. The design of specific sep-
aration units will require studies of gas dissolution in various selective absorbent
solutions. The thermodynamic approach for CO2 dissolution in aqueous solutions
of amine is depicted, showing the physicochemical background and the main prop-
erties required in this domain. An overview of the main experimental developments
for determining the enthalpy of solution of carbon dioxide in absorbent solutions is
presented together with some representative results.

14.1 Introduction

Carbon dioxide is considered as a main greenhouse gas. Its contribution to global

warming can be pointed up by correlations between the evolution of CO2 in
atmosphere and the average earth temperature [1]. Since the eighteenth century the
concentration of CO2 has approximately risen from 280–380 ppm when simultane-
ously temperature has increased by 0.6 K. Anthropogenic sources of carbon dioxide
represents currently about 25 Gt per year. If we consider constant industrial devel-
opment and population growth in the world, the emissions of carbon dioxide will

J.-Y. Coxam (B) · K. Ballerat-Busserolles

Institut de Chimie de Clermont-Ferrand,
Clermont Université, Université Blaise Pascal,
BP 10448, 63000 Ferrand, France
e-mail: [email protected]
K. Ballerat-Busserolles
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 481

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_14,
© Springer-Verlag Berlin Heidelberg 2013
482 J.-Y. Coxam and K. Ballerat-Busserolles

not decrease without an international mobilization. In order to avoid irreversible

environmental impact, CO2 emissions have to be drastically reduced [2].
The anthropogenic carbon dioxide emissions can roughly be classified in 3 main
categories: industrial, transportation and residential. The key solutions to reduce
these emissions will be found in an ecologic production and rational utilization of
energy. One of them will be to reduce CO2 emissions from industrial site called“fixed
sources”. It mainly concerns industries using fossil energy such as cement factory,
metallurgy, glass industry or power plant. For this purpose pre-combustion and post-
combustion processes have been proposed. The first ones consist in using specific
comburant and technology for the combustion of coal, fuel or natural gas in order to
release small quantity of CO2 in the fumes. The second ones refer to techniques of
separation of carbon dioxide from industrial effluent. In this last case carbon dioxide
will be valorized or stored in specific secured sites.
Capture processes were developed in the past to remove acid gases such as carbon
dioxide and hydrogen sulfide (H2 S) from natural gas. They are mainly based on
chemical and physical dissolution of the acid gas in aqueous solutions of amines.
The technique is considered as mature enough to be adapted in next future to the
treatment of post combustion effluents. The new processes should take into account
difference from natural gas treatment such as temperature, pressure or composition
of the effluent. However one major barrier for the integration into industrial sites in
the few coming years is the economical cost of the so called ton of C O2 avoided.
Specific researches are then carried out on both the technology and the choice of the
absorbent solutions.
Future absorbent solutions have to combine high carbon dioxide loading charges
(moles of dissolved carbon dioxide per mole of amine) with low energies of regener-
ation. Characterization of new absorbent solution can be performed by calorimetric
studies of gas dissolution. The experimental data collected are essential to develop
thermodynamic models representative of the {CO2 -absorbent solution} systems that
will be used to design the future capture units. The dissolution properties required
are the mainly the gas solubility and the enthalpy of solution. However some other
properties also have to be studied, such as heat capacity, vapor pressure, chemical and
thermal degradations. Then specific calorimetric techniques were set up to provide
the essential experimental data.

14.2 Presentation of Techniques for CO2 Separation from

Gaseous Effluents

Adsorption on micro-porous solids is a common industrial technique used for remov-

ing impurities (e.g. CO2 from gaseous streams such as hydrogen-rich gases resulting
from gasification, steam reforming or shift of fossil hydrocarbons. The molecules
to be adsorbed interact with the internal surface of microporous solids. Polar and/or
condensable species like CO2 will be strongly retained whereas “light components”
such as N2 , O2 , CO will be less or not retained on the adsorbent. The regenera-
14 CO2 Capture in Industrial Effluents. Calorimetric Studies 483

tion, or gas desorption, is performed by running thermal cycles (Temperature Swing

Adsorption, TSA) and/or pressure cycles (Pressure-Swing-Adsorption, PSA or
Vacuum-Swing-Adsorption, VSA). The energy required is thus analogous to that
of absorption. The most common adsorbents are activated carbons and zeolite mole-
cular sieves, mostly used in combination [3, 4].
For separation processes, solid adsorbents are packed as fixed beds and submitted
to successive adsorption and desorption steps. The operation is therefore transient
instead of steady-state. To achieve a continuous production, multiple packed-beds are
needed in parallel with shifted time-cycles. The technique necessitates a drastic pre-
treatment of gas effluent, especially for particle removal. A new class of nanoporous
crystalline solids, called metal-organic framework (MOF) materials, is intensively
studied as potential adsorbent materials for CO2 capture. The interest of such material
is their high internal surface area, large pore volume, and the possibility to adapt the
pores with a variety of functionalities.
The membrane processes proposed for CO2 separation consist in selective per-
meation of molecules through a thin layer of polymeric or mineral material. The two
intrinsic criteria determining their performance are the selectivity and the permeabil-
ity. Industrial membrane processes have been particularly developed for recovery of
relatively concentrated CO2 (>20 %) at high pressures from natural gas. [5, 6]. Selec-
tivity and permeability of membranes depend on their chemical nature and their
structure (microporosity, active layer thickness). Modern membranes are layered
composites, with a thin dense selective layer supported by thicker, more permeable
and less-selective or inert supports. There are basically two families of membrane
materials: inorganic ceramic type membranes, and organic polymeric membranes.
The former can stand high temperatures and pressures, and are corrosion resistant.
They can be made selective for H2 /CO2 separations, with hydrogen permeating pref-
erentially. However the CO2 /N2 separation selectivity remains relatively low in case
of CO2 capture from post combustion effluent.
The cryogenic process is used on a large scale for air separation but cryogenic
distillation does not seem to be currently considered as an alternative for CO2 cap-
ture, probably because of anticipated high costs when diluted streams are used [7].
However in the case oxy-fuel combustion technique where cryogenic air separation
unit is implemented, it may be conveniently integrated with c ondensation units for
separating water from CO2 .
Absorption in alkaline solution is a common principle used in acid gases capture
processes operating for decades in natural gas treatment. The reference absorbent
is aqueous solution of monoethanolamine (MEA). The mechanism of capture is a
combination of chemical and physical dissolution. The chemical dissolution is based
on an acido-basic reaction. The reaction must be reversible in order to regenerate the
absorbent solution and recover carbon dioxide for storage. The physical dissolution,
usually observed fot high partial pressure of CO2 , can be improved by addition of
specific physical solvent.
Some other more emerging techniques are proposed. It includes gas hydrate crys-
tallization [8], enzymatic capture [9] or dissolution in ionic liquids [10]. However
the separation process by dissolution in amine solutions (amine washing process) is
484 J.-Y. Coxam and K. Ballerat-Busserolles

presently the most advanced technology proposed to remove carbon dioxide from
industrial effluents. However the dilemma of the economic costs still constitutes a
barrier for integration such processes into industrial sites.

14.3 Industrial Processes Proposed for CO2 Capture in Post

Combustion Effluents

The CO2 capture processes will concern the industries that emit large volumes of
carbon dioxide i.e. those using fossil energy. For near future integration, the most
mature technologies are those based on gas separation by selective dissolution in
appropriate absorbent solutions. The industrial process is schematically represented
in Fig. 14.1. It consists in cycles of gas absorption and desorption that implicate a
reversible reaction of gas dissolution [11]. In the first step the gaseous effluent flows
counter-currently to the absorbent solution in an absorber unit. In the second step
the “rich” absorbent solution containing dissolved CO2 is pumped to the top of a
stripper (or regeneration vessel). The regeneration of the chemical solvent is carried
out by heating the solution using water vapor. The “lean” solvent is pumped back to
the absorber.
Industrial pilots based on this principle are now under study. The design of the
future installations must be tailored to the composition, flow rate, temperature and
pressure of the fumes. These effluent parameters largely depend on the type of indus-
try considered. The incoming gas flow rates will vary from 15,000 to 3,00,000 Nm3 /h

CO2
Off gas
Condensate
tank

Lean solvent
Absorber

Stripper
Rich solvent

Gas Effluent
Heat
exchanger

Rich solvent Lean solvent

Pumps
Reboiler

Fig. 14.1 Schematic representation of capture process

14 CO2 Capture in Industrial Effluents. Calorimetric Studies 485

as in the same way, CO2 compositions in the effluents will vary from 3 to 30 moles
percent.
The regeneration of the solvent solution is planned to be performed at temperatures
between 100 and 140 ◦ C and at pressures not very much higher than atmospheric
pressure. The costs of this step represent a major barrier to the integration of capture
process into industrial sites. Then active researches are conducted to elaborate better
absorbent solutions that will require less energy for regeneration.
The characterization of the new absorbent solutions will be achieved by a determi-
nation of its physical and chemical properties. We will focus here on some properties
that can be investigated by calorimetric techniques.
For example, the energy of regeneration, related to the opposite of the energy of gas
dissolution, will be studied through the determination of the enthalpy of solution of
CO2 in the absorbent solution. This enthalpy represents the energy needed to dissolve
one mole of gas and will contribute in a great extend to the calculi of economical
costs of carbon dioxide capture. Heat capacities will also be key properties for heat
exchange calculations as well as the gas solubility, required to estimate solvent flux
for given gas stream. In addition, the solvent characterization should be completed by
studies on vapor pressure, transport properties, chemical and thermal degradations,
kinetic of dissolution, corrosion...

14.4 Thermodynamic Approach of CO2 Dissolution in Aqueous

Solutions of Amine

14.4.1 Mechanism of CO2 Dissolution

The dissolution of CO2 in aqueous solution of amine is combination of chemical

reactions and physical dissolution. The chemical reactions involved depend on the
type of the amine.
In case of primary or secondary amine, the presence of hydrogen(s) branched on
nitrogen atom allows formation of carbamate. The carbon dioxide reacts first with
one molecule of amine to form a zwitterion (Eq. 14.1). Then the zwitterion reacts
with a second amine molecule to form a carbamate (Eq. 14.2). The stoechiometry for
chemical dissolution in primary or secondary amines is then limited to 2 molecules
of amine per molecule of carbon dioxide (Eq. 14.3).
−
CO2 + RR NH = RR N + HCOO (14.1)
 + −   −  +
RR N HCOO + RR NH = RR NCOO + RR N H2 (14.2)
  −  +
CO2 + 2RR NH = RR NCOO + RR N H2 (14.3)

In case of sterically hindered amine the carbamate is instable and primary or sec-
ondary amines will react as the tertiary amines. The instability of carbamate also
486 J.-Y. Coxam and K. Ballerat-Busserolles

increases with temperature. With tertiary amines, carbon dioxide cannot form car-
bamate because of no available free proton. The reaction mechanism then results
in hydration of carbon dioxide by alkaline catalysis to form hydrogenocarbonate
(Eq. 14.4).
+
RR R N + H2 O + CO2 = RR R NH + HCO− 3 (14.4)

The increase of partial pressure of carbon dioxide forces the physical dissolution.
This mechanism is purely mechanical and results in the apparition of molecular CO2
in the solution. This physical dissolution is particularly considerate in the treatment of
natural gas in which partial pressure of carbon dioxide can reach hundreds of bars.
In the case of CO2 capture in industrial effluent, partial pressure of CO2 remains
below atmospheric pressure and carbon dioxide is mainly chemically absorbed with
formation of carbamate and hydrogen carbonate.
The solution concentration and the choice of the amine will be adapted to the
conditions of temperature, pressure and composition of the gas stream to be treated.
It has been shown that amines can also be combined to improve kinetic of absorption
and quantity of dissolved gas [12]. In some cases, physical solvents can also be added
to the solution in order to increase physical dissolution.

14.4.2 Selection of Amines for CO2 Capture Processes

The design of specific amines for CO2 capture in industrial effluent is a difficult
task as no model really exists to predict the dissolution properties from structural
properties. Then current studies consist in a screening of commercial or synthe-
sized molecules, implicating numerous experimental investigations. The absorbent
solution must be characterized by properties such as density, viscosity, heat capac-
ity, thermal and chemical stability, or corrosion aspects. The gas dissolution can be
described by the determination of dissolution properties such as the limits of gas
solubility, enthalpies of solution or kinetic of gas dissolution. The experimental data
make it possible the development of theoretical models representative of the {CO2 -
water-amine} systems. These models can then be used for engineering calculations to
estimate the pre-cited properties of dissolution of the absorbent solution as function
of temperature, pressure or amine concentration.
First studies were carried out on alcanolamine such as monoethanolamine (MEA),
diethanolamine (DEA) or methyl diethanolamine (MDEA). Investigations are actu-
ally extended to “new” amines chosen within families such as aliphatic, cyclic, poly-
functional … Recently demixing amines solutions were proposed for new capture
process. Aqueous solutions containing such amines present a liquid-liquid phase
separation with a temperature change depending on the composition of the solution.
The corresponding process DMXTM has been proposed and patented by IFPEN [13].
A demixing unit was added at the output of the absorber unit in which the tempera-
ture of the gas charged solution is increased until reaching demixing of the ternary
solution CO2 -amine-water in two phases. Only the aqueous phase that contains
14 CO2 Capture in Industrial Effluents. Calorimetric Studies 487

dissolved carbon dioxide is sent to the stripper while the amine phase is directly
returned to the absorber. In these conditions the energy required for the regenera-
tion of the absorbent solution is significantly reduced. The characterization of these
demixing amine solutions is comparable to the classic one excepted an additional
determination of phase diagrams of binary {water-amine} and ternary {CO2 -water-
amine} systems.

14.4.3 Calorimetric Experimental Data Required

The properties of dissolution as gas solubility and enthalpy of solution can be derived
from vapor liquid equilibrium models representative of {CO2 -H2 O-amine} systems.
The developments of such models are based on a system of equations related to
phase equilibria and chemical reactions: electro-neutrality and mass balance. The
non ideality of the system can be taken into account in liquid phase by the expressions
of activity coefficients and by fugacity coefficients in vapor phase. Non ideality is
represented in activity and fugacity coefficient models through empirical interaction
parameters that have to be fitted to experimental data. Development of efficient
models will then depend on the quality and diversity of the experimental data.
Regarding literature data reported in Tables 14.1 and 14.2 collected for classi-
cal amines (MEA-DEA-AMP), it appears that solubility data are quite available
while enthalpy data remains very scarce. Solubility data in Table 14.1 are usually
obtained using pVT measurements and provide gas solubility expressed as loading
charge α (moles of gas per mole of amine) as function of temperature and pressure
(CO2 partial pressure or total pressure). The enthalpy data in Table 14.2 are obtained
from measurements of heat of mixing gas with absorbent solution using calorimetric
techniques. Different techniques and associated experimental protocols have been
proposed and will be presented in next paragraph. Mathonat et al. [14, 15], Arcis
et al. [16, 17] use a flow calorimetric technique where the heat of mixing is detected
by thermopiles surrounded a mixing cell. Kim and Svendsen [18], Carson et al. [19]
and Kierzkowska-Pawlak [20] use a reaction calorimeter based on power compen-
sation of the heat of mixing. Oscarson et al. [21], Merkley et al. [22] and Helton
et al. [23] use a flow technique where the heat of mixing is determined from power
compensation. In another hand enthalpies of solution can be derived from solubility
data or from thermodynamic model fitted to solubility data. However as shown by
Arcis et al. [24] and Kim et al. [25] such calculated values can show large divergence
with direct experimental.
The representation of chemical reactions in solution in the thermodynamic models
[21, 26, 27] necessitates the knowledge of the equilibrium constants of CO2 dissoci-
ations, water dissociation, amine protonation and carbamate formation. For original
amines the protonation or carbamate formation equilibrium constants are usually
not available and must be measured. In order to derive enthalpy properties using
Van’t Hoff equations, these equilibrium constants must be determined as function of
temperature. Such data can be obtained from a protonation constant determined at a
488 J.-Y. Coxam and K. Ballerat-Busserolles

Table 14.1 Literature solubility data of CO2 in aqueous solution

Source m mol.kg−1 TK pC O2 kPa ASD %
(a). Monethanolamine (MEA)
Mason and Dodge [28] 0.5–12.5(b) 273–348 1.32–100 2
Jones et al. [29] 15.3 313–413 0.0027–930 0.5
Lee et al. [30] 2.5–5.0(b) 313–373 1.15–6621 3
Lee et al. [31] 1.0–5.0(b) 298–393 0.1–10000 4
Lawson and Garst [32] 15.2 313–413 1.32–2750 9
Isaacs et al. [33] 2.5(a) 353–373 0.0066–1.75 15
Austgen and Rochelle [34] 2.5(a) 313–353 0.0934–229 NC
Shen and Li [35] 15.3–30.0 313–373 1.1–2550 12
Dawodu and Meisen [36] 4.2(a) 373 455–3863 13.5
Jou et al. [37] 30.0 273–423 0.0012–19954 3
Song et al. [38] 15.3 313 3.1–2359 12
Jane and Li [39] 2.5(a) 353 3.57–121.8∗ 5
Mathonat et al. [14] 30.0 313–393 5000–20000 7
Ma’mun et al. [40] 30.0 393 7.354–191.9 2
Arcis et al. [41] 323; 373 500–5000* 7
(b). Methyldiethanolamine (MDEA)
Jou et al. [42] 2.00–4.28(a) 298–393 0.001–6630 NC
Merkley et al. [22] 20.0–23.5 298–393 1121 5
Chakma and Meisen. [43] 1.69–4.28(a) 373–473 103–4930 14
Austgen and Rochelle [34] 2.00–4.28(a) 313 0.0056–93.6 NC
Shen and Li [35] 30 313–373 1.1–1979 12
Jou et al. [44] 35 313–373 0.963–236 3
Dawodu and Meisen [36] 4.28(a) 373–393 162–3832 14
Oscarson et al. [21] 20.0–60.0 289–422 22-6164 5
Kuranov et al. [27] 1.95–2.00(b) 313–413 73.5–5036.7 3
Xu et al. [45] 1.72–6.85(a) 328–363 137.5–808.5 NC
Mathonat et al. [15] 30 313–393 2000 − 10000∗ 7
Rho et al. [46] 5–75 323–373 0.775–268.3 5
Silkenbäumer et al. [47] 2.632(b) 313 12 − 4080∗ 2
Baek and Yoon [48] 30 313 1.02–1916 3
Rogers et al. [49] 23–50 313 0.00007–1.0018 6
Xu et al. [45] 3.04–4.28(a) 313–373 0.876–1013 NC
Lemoine et al. [50] 23.63 298 0.02–1.636 5
Pacheco et al. [51] 35–50 298–373 73–738 NC
Kamps et al. [52] 3.95–7.99(b) 313–353 176.5–7565 4
Park and Sandall [53] 50 298–373 0.78–140.40 10
Kierzkowska-Pawlak [20] 10-30 293–333 100–300 3
Bishnoi and Rochelle [54] 4.28(a) 313 0.108–0.730 NC
Sidi-Boumedine et al. [55] 25.73–46.88 298–348 2
Ali and Aroua [56] 2(a) 313–353 0.06–95.61 2
Kundu et al. [57] 23.8–30.0 303–323 1 3
Benamor and Aroua [58] 2 − 4(a) 303–323 0.1–98.2 5
Ma’mun et al. [40] 50 328–358 65.75–813.4 2
(continued)
14 CO2 Capture in Industrial Effluents. Calorimetric Studies 489

Table 14.1 (continued)

Source m mol.kg−1 TK pC O2 kPa ASD %
Jenab et al. [59] 2.0 − 2.5(a) 298–343 101–2320 4
Huttenhuis et al. [60] 35–50 283–298 0.054–986.80 NC
(c). Diethanolamine (DEA)
Lee et al. [30] 0.5–8.0(a) 273–413 0.69–6900 5
Oyevaar et al. [61] 0.984–2.463(a) 298 1.844–14.8 5
Kennard and Meisen [62] 10.0–30.0(c) 373-478 73.1–3746.7 10
Dawodu and Meisen [36] 4.2(a) 373 455–3863 13.5
Lal et al. [63] 2.0(a) 313–373 0.0026–3.336 15
Lee et al. [64] 3.0–4.0(a) 298–403 0.001- 5000 NC
Lawson and Garst [32] 25.0(c) 311–394 1.974–4315.79 9
Mason and Dodge [28] 0.5–8.3(a) 273–348 1.32–100 2
Sidi-Boumedine et al. [55] 41.78(c) 298- 348 2.46–4662.7 2
Seo and Hong [65] 30.0(c) 313–353 4.85–357.3 3
Benamor and Aroua [58] 2.0–4.0(a) 303–323 0.1–104.7 2
Haji-Sulaiman et al. [66] 2.0–4.0(a) 303–323 0.1–104.7 2
Haji-Sulaiman and Aroua [67] 2.0(a) 301–353 5–100.3 12
Arcis et al. [68] 15; 30(c) 323; 373 500 -5000 7
(a) molarity (mol.L−1 );(b) massic fraction; (c) wt%; ∗ ptot in kPa

Table 14.2 Literature references for experimental enthalpy of solution

Source wt % TK pC O2 kPa ASD %
MEA
Mathonat et al. [14] 30.0 313–393 2000–10000 7
Kim and Svendsen [18] 30.0 313–393 0.001 3
Carson et al. [19] 10.0–30.0 298 265 2
Arcis et al. [41] 15.0; 30.0 323; 373 500–5000* 5
DEA
Helton et al. [23] 20–50 300–400 87–1121 5
Carson et al. [19] 10.0–30.0 298 265 2
Oscarson et al. [21] 20.6–49.8 300–450 90–1121 5
Arcis et al. [68] 15.0; 30.0 323; 373 500–5000* 5
MDEA
Merkley et al. [22] 20.0–23.5 298–393 1121 5
Oscarson et al. [21] 20.0–60.0 289–422 156–1466 5
Mathonat et al. [15] 30.0 313–393 2000–10000 7
Kierzkowska-Pawlak [20] 10.0–40.0 293–333 100–300 2
Carson et al. [19] 10.0–30.0 298 265 2
Arcis et al. [17] 15.0–30.0 323 500–2000 5
Arcis et al. [16] 15.0–30.0 373 500–2000 5
∗p in kPa
tot
490 J.-Y. Coxam and K. Ballerat-Busserolles

reference temperature and the enthalpy of amine protonation. Experimental data can
be obtained by combining electrochemical and calorimetric techniques [21]. For pri-
mary and secondary amine it is also essential to determine the equilibrium constants
for carbamate formation. These constants can be obtained by the determination of
carbamate concentration when adding CO2 to the amine solution using available
speciation techniques. As for amine protonation the study can be completed by mea-
surements of enthalpies of carbamate formation using mixing calorimetry. Then it
appears that to develop thermodynamic models for CO2 dissolution in aqueous solu-
tion of amine, there is a real need of accurate experimental data for enthalpies of
solution, enthalpy of amine protonation as well as for enthalpy of carbamate forma-
tion.

14.5 Calorimetric Studies of CO2 Dissolution in Amine

Solutions

14.5.1 Calorimetric Techniques for Measuring Heat of Mixing

The calorimetric techniques for measuring heats of mixing two fluids can be classified
into their mode of measurement and their principle of heat detection. The isothermal
displacement calorimetry will refer to a “static” mode and flow calorimetry, to a
“dynamic mode”. The principles of heat detection in the following examples will be
power compensation or heat flux determination.

Isothermal Displacement Calorimeter

The technique used by Carson et al. [19] at University of Canterbur (New Zealand)
is a modification of the liquid–liquid isothermal displacement calorimeter originally
developed by Stokes et al. [69]. This static technique has then been primarily used
to study binary liquid mixtures. The technique consists in filling a calorimetric cell
with a known amount of one of the liquids (solvent) and injecting the solute from a
burette. The injection is performed after reaching thermal equilibrium in the calori-
metric cell. As reaction occurs, the temperature change is compensated in order to
maintain the calorimeter to its initial temperature. Heat of mixing can be determined
at temperatures range from 298 to 313 K and pressures comprised between 0.1 and
0.3 MPa.
In order to make it possible the determination of enthalpies of solution of gases in
organic liquid solvents, Battino and Marsh [70] set up a modified burette arrangement.
This modified technique proved to be effective for the particular systems where small
gas solubilities are observed. However the technique was impractical in the case of
CO2 dissolution in aqueous solution of amine. This was mainly due to high solubility
of carbon dioxide and consequently, large volumes of injected gas. Then Carson
14 CO2 Capture in Industrial Effluents. Calorimetric Studies 491

Fig. 14.2 Schematic representation of a classical displacement calorimeters [69]. (a annular heater;
b Peltier cooling unit; c inlet valve; d injection tube; e stirrer; f thermistor; g stirrer gland; h solution
outlet; i inlet valve control; j outlet to pipette; k by-pass outlet.) and of b modified calorimeter [19]
(a pressure transducer; b gas injection device; c calorimeter; d pipette; e carbon dioxide reservoir
tank)

et al. [19] carried out the last modifications of this technique. A special gas injection
device constructed from stainless steel was used in place of the former burette. The
mole number of gas introduced into the cell is calculated from the pressure changes
after gas injection at known temperature. The exothermic effect when mixing CO2
and amine solution is measured after successive injections of CO2 . The schematic
representation of the experiment is shown Fig. a,b.
Kim and Svendsen [18] at Norwegian University of Science and Technology
choose a similar technique for measuring heat of mixing of CO2 in amine solutions.
The main difference with Carson et al. modified calorimeter [19] results in a larger
range of experimental temperatures (253–473 K) and pressures (up to 10 MPa) that
can be investigated. Their experimental procedure (Fig. 14.3) consists as previously
in multiple successive injections of CO2 up to reach a CO2 pressure in the gas
reservoir tank close to pressure in the calorimeter vessel.

Isothermal Compensation Flow Calorimeter

This dynamic technique used by Oscarson and co-workers [21] at Brigham Young
University (USA) to study CO2 capture studies was developed in same University by
Christensen and coworkers [71]. The principle of isothermal compensation calorime-
ter is based on the measurement of an external heat power used to maintain a reaction
vessel at constant temperature during mixing. In flow mode the fluids flow to the
492 J.-Y. Coxam and K. Ballerat-Busserolles

Fig. 14.3 Experimental setup of Kim and Svendsen [18] calorimeter. 1 reaction calorimeter
(2000cm3 ); 2a, 2b CO2 reservoir tanks; 3 CO2 mass flow controller; 4 amine solution feed bottle;
5 vacuum pump

reaction vessel where they mix. Different types of reaction vessel were proposed
as for example that one described by Ott et al. [72] and represented in Fig. 14.4. In
this example heaters and mixing tube are wound around and vacuum silver braised
to a nickel plate cylinder. The mixing tube is made of hastelloy. Two concentric
tubes drive the fluids to the mixing point located at the bottom of the cylinder. The
energy generated by the reaction is balanced by adjusting the heater or the Peltier
cooler device. Hence the change in heater or cooler power is proportional to the
enthalpy change in the calorimeter. The two solutes are injected at defined flow rates
by used of two syringe pumps. The pressure is maintained constant using a back
pressure regulator. This kind of calorimetric technique can be be adapted to be used
at temperature from ambient up to 473 K and at pressures up to 20 MPa.

Heat Conduction Differential Flow Calorimeter

This dynamic technique was initially developed by Mathonat and coworkers at

University of Clermont-Ferrand (France). It is a flow technique where the heat of mix-
ing is measured in a mixing cell located inside a Calvet type calorimeter (Fig. 14.5a).
The Calvet sensors are a thermopiles constituted of thermocouples surrounding the
mixing cell and measuring the heat power exchanged with a thermostated calori-
metric block. The mixing cell represented in Fig. 14.5b consists of an hastelloy tube
14 CO2 Capture in Industrial Effluents. Calorimetric Studies 493

Fig. 14.4 Schematic repre-

sentation of a reaction vessel
[72] of isothermal compensa-
tion flow calorimeter. a inlets;
b outlet; c electrical leads; d
Peltier cooler; e controlled
heater; f isothermal cylinder;
g calibration heater; h inside
tube

tightly coiled inside a metallic confinement cylinder which fits into the well sur-
rounded by the thermopile detector. For measuring the heat of dissolution of carbon
dioxide in an absorbent solution, the fluids are injected at constant flow rates by
used of high pressure syringe pumps. The temperature of the fluids to be mixed
is controlled by pre-heaters positioned on the incoming tubes before the mixing
point.
The fluids encounter following two different mixing technologies schematically
represented in Fig. 14.5-c. The first one developed by Mathonat et al. [15] consists
in flowing the fluids to be mixed into two concentric tubes; the fluids meet where
the thinner tube ends. The second one presented by Arcis et al. [16, 17, 73] consists
in two parallel hastelloy tubes silver welded on a T type brass piece represented in
Fig. 14.5-c. Experiments are carried out at constant pressure range from 0.1–40 MPa.
The temperature of experiment will depend on the calorimeter. Using a commercial
Calvet type calorimeters from Setaram, this technique permits measurements from
200 to 573 K.
494 J.-Y. Coxam and K. Ballerat-Busserolles

Inner tube

Outer tube

(a) (b) (c)

Fig. 14.5 Schematic view of the thermopiles a, mixing cell b, and mixing points c in the Setaram
C80 or BT2.15 calorimeters a inlets; b outlet; c syringe pumps; d mixing cell; e thermostated
calorimetric bloc; f thermopiles; M mixing point

14.5.2 Calorimetric Investigations

The experimental protocol for measuring heat of mixing of gas in liquid absorbent
will depend on the chosen technique.
The static technique will consist in additions of known amount of gas in a cell
containing the absorbent solution. The heat of mixing is obtained by integration
of a peak corresponding to the heat power signal recorded as a function of time
during gas dissolution. The end of dissolution is considered as reached when the heat
power signal returns to baseline state. In the approximation of small gas addition,
the heats of mixing expressed per mole of CO2 can be assumed to be equal to
differential enthalpies of solution Hdi f f . Carson et al [19] measured the enthalpy of
solution of CO2 in aqueous solutions of monoethanolamine (MEA) and methyl di-
ethanolamine (MDEA) at 298.15 K. The experiments were carried out at atmospheric
pressure for small CO2 loading charges. In these conditions the gas is assumed to be
totally dissolved in amine solution and no vapor phase was considerate. The method
is particularly appropriate for the determination of enthalpy of solution at infinite
dilution at atmospheric pressure. The uncertainty is estimated by the authors to
±2 kJ · mol−1 . The results represented in Fig. 14.6 show that the enthalpy of solution
is almost constant down to α = 0.05 and then decreases.
Kim and Svendsen determined the enthalpy of solution of MEA for higher CO2
loading charges α (Fig. 14.7). The maximum loading charge reported in this work
is close to 0.8. The uncertainty on enthalpy of solution is estimated to 2.2 %. It is
observed that for loading charge below 0.5 the enthalpy of solution is more or less
constant. It is usually admitted that the enthalpy derived from the plateau (Fig. 14.7)
corresponds to an enthalpy of solution at infinite dilution. However this behavior
must be confirmed by measurements in the domain of small loading charges to con-
14 CO2 Capture in Industrial Effluents. Calorimetric Studies 495

85
54 84
52 83

- Δ H (kJ.mol-1 )
82
Δ H (kJ.mol -1 )

50
81
48 80
46 79

s
78
44
s

77
42 76
40 75
0.00 0.05 0.10 0.15 0.20 0.25 0.00 0.05 0.10 0.15
α(mol CO2 /mol MDEA) α (mol CO 2 /mol MEA)

Fig. 14.6 Enthalpy of solution of CO2 in aqueous solutions of monoethanolamine (MEA) and
methyl di-ethanolamine (MDEA) at low loading charges using an isothermal displacement calorime-
ter [19]. Mass fractions of amine solutions ♦ 0.1,  0.2, ◦ 0.3

Fig. 14.7 Enthalpy of 95

solution at 313.15 K
and 0.3 MPa of CO2
in aqueous solutions of
−Δ H (kJ.mol -1 )

monoethanolamine (MEA) 85
at mass fraction 0.3, using
static technique [18] on a wide
range of loading charges α
75
s

65
0.00 0.50 1.00
α (mol CO2 / mol MEA)

firm the enthalpy decrease observed by Carson et al. [19]. This domain is particularly
difficult to investigate due to problems of detection of small heat effects. The curve
representing the enthalpy of solution in Fig. 14.7 shows a split at loading charge
α = 0.5. As the monoethanolamine (MEA) is a primary amine that forms carba-
mate with CO2 , the stoechiometric limit of gas absorption should be 0.5. However,
when adding more gas, it was observed that the reaction of carbamate formation is
displaced to release the amine. Then dissolution of CO2 can continue with forma-
tion of carbonate as it is shown by the speciation curve in Fig. 14.9, representing
the solution composition as function of loading charge. The change in the chemical
mechanism of CO2 absorption will modify the enthalpy of solution after loading
charge 0.5 (Fig. 14.7).
Using flow techniques (or dynamic technique), the gas and the absorbent solution
flow separately in narrow tubes to a mixing point. Then the mixture flows in a common
mixing tube located in the sensible part of the calorimeter. The heat power due to
gas dissolution is then detected along the mixing tube, by a thermopile in Calvet
496 J.-Y. Coxam and K. Ballerat-Busserolles

70
90
60

−ΔsH (kJ.mol )
−ΔsH (kJ.mol -1)

-1
50 80

40
70
30

20 60
0.00 0.50 1.00 0.00 0.50 1.00
α(mol CO 2 / mol MDEA) α(mol CO2 / mol MEA)

Fig. 14.8 Enthalpy of solution of CO2 in aqueous solution of MDEA and MEA using dynamic
techniques. Results for amine solutions at mass fraction of 0.30, at temperature 322.5 K and pressure
5.17 MPa [41, 74]

type calorimeter or by a compensation heat power in Picker type calorimeter. The

enthalpy of solution per mole of CO2 is obtained by dividing the heat power by the
gas molar flow rate. This enthalpy of solution determined here is an integral enthalpy
of solution s H, related to the differential enthalpy following Eq. 14.5
 α
1
s H = Hdi f f dα. (14.5)
α 0

Special care is required with dynamic technique to adjust the fluid flow rates to the
kinetic of dissolution. The residence time of mixture in the sensible zone of the
calorimeter must be long enough in order to reach total dissolution. The volumetric
fluid flow rates are usually comprised between 0.05 and 1mL · min−1 . The heats of
mixing are measured at variable loading charges by changing fluid (gas/absorbent)
flow rates ratios. The experiments are carried out increasing loading charge α up to
the saturation of the absorbent solution. Investigation can be carried out at elevated
pressure where physical dissolution can be observed.
The graphs reported in Fig. 14.8 represent enthalpy of solution as function of load-
ing charge determined using fluxmetric detection; the graphs obtained with compen-
sation detection will be similar. As initially mentioned, the enthalpies remain nearly
constant in the domain of the low loading charges before decreasing when increasing
loading charges. With monoethanolamine (MEA), the decrease of the enthalpy of
solution starts at loading charge 0.5 corresponding to the change of chemical mech-
anism of absorption (Fig. 14.9). The speciation in the liquid can be estimated from
thermodynamic model (Fig. 14.9). It indicates that significant physical dissolution
starts for loading charge about 0.75. This is revealed by the presence of molecular
CO2 in the solution. Then enthalpy decrease is also physical dissolution of CO2 ; the
energy of dissolution associated to physical mechanisms is lower than the energy of
chemical dissolution. On the graph representing the enthalpy of solution of CO2 in
14 CO2 Capture in Industrial Effluents. Calorimetric Studies 497

3.00

2.50

m / mol .kg -1 2.00

1.50

1.00

0.50

CO32-
0.00
0 0.5 1 1.5
α

Fig. 14.9 Speciation in aqueous solution of MEA (3.00 mol · kg−1 ) as function of CO2 loading
charge. Value calculated at 313.15 K using thermodynamic model [75]

MDEA solution, the enthalpy decrease starts at loading charge about 0.9. The MDEA
is a tertiary amine and then does not form carbonate. The chemical mechanism of
CO2 absorption corresponds only to formation of carbonate. However the enthalpy of
solution decreases at loading charge about 0.9. This phenomena is observed because
of one part of CO2 is physically dissolved.
The thermodynamic models developed to represent rigorously CO2 dissolution
in aqueous solution of amine consider all possible chemical reactions. Then the sys-
tem is complex and numerous interaction parameters, molecule-molecule, ion-ion
and molecule-ion, have to be defined. These models correlate vapor liquid equilib-
ria data and are used to predict carbon dioxide solubility as a function of its partial
pressure in the gas phase, temperature and amine concentration of the absorbent solu-
tion. In addition the models can give a detailed composition of solution as function
of loading charge α as presented in Fig. 14.9. The model then predicts if gas dis-
solution leads to formation of hydrogenocarbonates or carbamates but experimental
data of the real speciation are very scarce to validate those results. However as the
energy of carbamate or carbonate reaction are different, the enthalpy data can provide
some indications on the nature of the compounds produced by reaction of CO2 with
the amine. For this purpose thermodynamic models are developed to decompose
the total enthalpy of solution into contribution terms associated to the formation of
all the different species in solution, such as protonated amine, carbonate, bicarbon-
ate or carbamate. Example of calculation of enthalpy contribution terms is given
in Fig. 14.10 for the dissolution at 313.15 K of CO2 in aqueous solution of MEA
of composition 3.00 mol · kg−1 . The comparison between the enthalpy of solution
obtained by summing the contribution terms and experimental data can then be used
to validate the model.
498 J.-Y. Coxam and K. Ballerat-Busserolles

120
Total enthalpy of solution
100

Enthalpy (kJ.mol-1)
80
Enthalpy of amine protonation
60

40 Enthalpy of carbamate formation

20
Enthalpy of physical dissolution
0
-20 Enthalpies of carbon dioxyde dissociation
0 0.2 0.4 0.6 0.8 1 1.2 1.4
α

Fig. 14.10 calculated enthalpy contribution at 313.15 K for CO2 dissolution in aqueous solution
of MEA 3.00 mol · kg−1 as function of loading charge

Other interest of flow techniques is the short time required to obtained enthalpy of
solution at given loading charge α, usually about 30 min. It is then not so much time
consuming to investigate a large domain of loading charge, covering the domain of
total and partial dissolutions of the CO2 gas flux. In this way the flow techniques allow
a simultaneous determination of the gas solubility at given temperature pressure and
absorbent composition. For this purpose the enthalpies are expressed as energy per
mole of amine and represented as function of loading charge. The graphic repre-
sentation then makes it possible to identify the domains of total and partial CO2
dissolutions. In the domain where the solution is unsaturated (all injected CO2 is dis-
solved), the enthalpies expressed per mol of amine (Figs. 14.11 and 14.12) increase
with gas loading charge. Over passing the loading charge corresponding to the limit
of solubility, the additional CO2 injected remains in vapor phase and the enthalpy
per mole amine remains constant. It corresponds to the apparition of a plateau as
showed in Fig. 14.11. The loading charge at the limit of gas solubility corresponds
to the point where the plateau is reached. If the mechanism of dissolution is chem-
ical reaction, the increase of enthalpy is almost linear and the intersection between
the unsaturated and saturated domain is easily determined (Fig. 14.11-I). This deter-
mination is somewhere more difficult when physical mechanism is involved as the
enthalpy slope decreases slowly before reaching the plateau (Fig. 14.11-II). In the
case of primary amine as MEA (Fig. 14.12) a change in enthalpy slope is observed
at loading charge α = 0.5, due to the formation of a carbamate and only the points
above α = 0.5 must be considered to determine the linear part of the unsaturated
domain that will cross the plateau. The uncertainties on the value of gas solubility are
obviously more important than those obtained from direct pVT techniques. However
it represents a simple and easy method for acquisition of solubility data at elevated
pressures and temperatures. In addition the comparison of solubility obtained by mix-
ing calorimetry with reference values obtained from pVT techniques will confirm that
good mixing and total gas dissolution is achieved in the mixing cell.
14 CO2 Capture in Industrial Effluents. Calorimetric Studies 499

(I) (II)
70 70

60 60
−ΔsH (kJ.mol -1 )

50 50

−ΔsH (kJ.mol-1)
40 40

30 30

20 20

10 10
0 0
0.00 0.50 1.00 1.50 0.00 0.50 1.00 1.50 2.00
α(mol CO2 / mol MDEA) α(mol CO2 / mol MDEA)

Fig. 14.11 Enthalpy of solution of CO2 expressed as energy per mole of amine as function
of loading charge. Case of gas dissolution at 322 K in aqueous solution of (I) MDEA 30 wt %,
p = 0.5 MPa and (II) MDEA 30 wt %, p = 5.17 MPa

70

60

50
− Δ sH (kJ.mol -1 )

40

30

20

10

0
0.00 0.50 1.00 1.50

α (mol CO 2 / mol MDEA)

Fig. 14.12 Enthalpy of solution of CO2 expressed as energy per mole of amine as function
of loading charge. Case of gas dissolution in aqueous solution of MEA 30 wt % at 322 K and
p = 5 MPa

14.6 Conclusion

The environmental impact of CO2 must be considered and it is urgent to reduce its
anthropogenic emission. Solutions have to be proposed for near future. One option
is a decarbonation of post combustion effluents. Design and development of future
industrial process for CO2 separation from fumes will require studies of gas disso-
lution in various selective absorbent solutions.
500 J.-Y. Coxam and K. Ballerat-Busserolles

Aqueous solutions of amine are considered as promising solvent but the selection
of the best adapted molecule is still open. Then important experimental work will be
necessary to characterize and test the dissolution of carbon dioxide. In this domain,
determination of calorimetric properties such as enthalpy of solution of carbon diox-
ide in absorbent solutions will be essential. The enthalpies of solution are particularly
important for the estimation of the energy required for the solvent regeneration step
of an industrial process. For theoretical point of view, this thermodynamic property
participates to the development of thermodynamic models representative of gas-
absorbent systems. This chapter has focused on mixing calorimetric techniques used
to investigate gas dissolution in liquid absorbents. The presented techniques can be
adapted, in studies of CO2 dissolution in aqueous solution of amines, to investigate
for example each chemical reaction involved such as amine protonation or carbamate
formation.

Symbols

wt %: mass percent composition

α: gas loading charge (mole of gas / moles of absorbent)
ASD: average standard deviation
s H : integral enthalpy of solution
Hdi f f : differential enthalpy of solution
MEA: mono-ethanolamine
DEA: di-ethanolamine
AMP: 2-Amino-2-methyl-1-propanol
MDEA: methyl di-ethanolamine

References

1. M.E. Mann, R.S. Bradley, M.K. Hughes, Global-scale temperature patterns and climate forcing
over the past six centuries. Nature 392(6678), 779–787 (1998)
2. N. Thybaud, D. Lebain (2010) Panorama des voies de valorisation du CO2 . ADEME.
3. D.M. Ruthven, S. Farooq, K.S. Knaebel, Pressure Swing Adsorption (VCH, New York, 1994)
4. E.S. Kikkinides, R.T. Yang, S.H. Cho, Concentration and recovery of carbon dioxide from flue
gas by pressure swing adsorption. Ind. Eng. Chem. Res. 32(11), 2714–2720 (1993)
5. U. Kragl, Book review: basic principles of membrane technology. (Second edition) By M.
Mulder. Angew. Chem. Int. Ed. Engl. 36(19), 2129–2130 (1997)
6. R. Klaassen, P.H.M. Feron, R. van der Vaart, A.E. Jansen, B. Dibakar, D.A. Butterfield (2003)
Chapter 6 Industrial applications and opportunities for membrane contactors. Membr. Sci.
Technol. 8, 125–145 (Elsevier)
7. D. Tondeur, F. Teng, M.L. Trevor, Carbon Capture and Storage for Greenhouse Effect Mitiga-
tion Future Energy (Elsevier, Oxford, 2008)
8. A. Fezoua, A. Bouchemoua-Benaissa, Y. Ouabbas, F. Chauvy, A. Cameirão, J-M Herri, CO2
capture by gas hydrate crystallization. in S4FE2009 (International Conference on Sustainable
Fossil Fuels for Future Energy), Rome, Italy, 2009
14 CO2 Capture in Industrial Effluents. Calorimetric Studies 501

9. N. Favre, M.L. Christ, A.C. Pierre, Biocatalytic capture of CO2 with carbonic anhydrase and
its transformation to solid carbonate. J. Mol. Catal. B Enzym. 60(3–4), 163–170 (2009)
10. D. Wappel, G. Gronald, R. Kalb, J. Draxler, Ionic liquids for post-combustion CO2 absorption.
Int. J. Greenhouse Gas Control 4(3), 486–494
11. B. Gwinner, D. Roizard, F. Lapicque, E. Favre, R. Cadours, P. Boucot, P.-L. Carrette, CO2
capture in flue gas: semiempirical approach to select a potential physical solvent. Ind. Eng.
Chem. Res. 45(14), 5044–5049 (2006)
12. G.F. Versteeg, L.A.J. Van Dijck, W.P.M. Van Swaaij, On the kinetics between CO2 and alka-
nolamines both in aqueous and non-aqueous solutions. An overview. Chem. Eng. Commun.
144(1), 113–158 (1996)
13. P.-A. Bouillon, M. Jacquin, L. Raynal (2012) Method for deacidifying a gas by an absorbing
solution with demixing control. FRANCE Patent EP 2 168 659 B1
14. C. Mathonat, V. Majer, A.E. Mather, J.-P.E. Grolier, Use of flow calorimetry for determining
enthalpies of absorption and the solubility of CO2 in aqueous monoethanolamine solutions.
Ind. Eng. Chem. Res. 37(10), 4136–4141 (1998)
15. C. Mathonat, V. Majer, A.E. Mather, J.-P.E. Grolier, Enthalpies of absorption and solubility of
CO2 in aqueous solutions of methyldiethanolamine. Fluid Phase Equilib. 140(1–2), 171–182
(1997)
16. H. Arcis, L. Rodier, K. Ballerat-Busserolles, J.Y. Coxam, Enthalpy of solution of CO2 in
aqueous solutions of methyldiethanolamine at T = 372.9 K and pressures up to 5 MPa. J.
Chem. Thermodyn. 41(7), 836–841 (2009)
17. H. Arcis, L. Rodier, K. Ballerat-Busserolles, J.Y. Coxam, Enthalpy of solution of CO2 in
aqueous solutions of methyldiethanolamine at T = 322.5 K and pressure up to 5 MPa. J. Chem.
Thermodyn. 40(6), 1022–1029 (2008)
18. I. Kim, H.F. Svendsen, Heat of absorption of carbon dioxide (CO2 ) in Monoethanolamine
(MEA) and 2-(Aminoethyl)ethanolamine (AEEA) solutions. Ind. Eng. Chem. Res. 46(17),
5803–5809 (2007)
19. J.K. Carson, K.N. Marsh, A.E. Mather, Enthalpy of solution of carbon dioxide in
(water + monoethanolamine, or diethanolamine, orN-methyldiethanolamine) and (water +
monoethanolamine + N-methyldiethanolamine) at T = 298.15 K. J. Chem. Thermodyn. 32(9),
1285–1296 (2000)
20. H. Kierzkowska-Pawlak, R. Zarzycki, Calorimetric measurements of CO2 absorption into
aqueous N -methyldiethanolamine solutions. Chem. Pap. 56, (2002)
21. J.L. Oscarson, X. Chen, R.M. Izatt (1995) A thermodynamically consistent model for the
prediction of solubilities and enthalpies of solution of acid gases in aqueous alkanolamine
solutions. Gas Processors Association, Research, Report, vol 130
22. K.E. Merkley, J.J. Christensen, R.M. Izatt (1986) Enthalpies of solution of CO2 in aqueous
methyldiethanolamine solutions. Gas Processors Association, Research, Report, vol 102
23. R. Helton, J.J. Christensen, R.M. Izatt (1987) Enthalpies of solution of CO2 in aqueous
diethanolamine solutions. Gas Processors Association, Research, Report, vol 108
24. H. Arcis, L. Rodier, K. Ballerat-Busserolles, J.Y. Coxam, Modeling of (vapor+ liquid) equi-
librium and enthalpy of solution of carbon dioxide (CO2 ) in aqueous methyldiethanolamine
(MDEA) solutions. J. Chem. Thermodyn. 41(6), 783–789 (2009)
25. I. Kim, K.A. Hoff, E.T. Hessen, T. Haug-Warberg, H.F. Svendsen, Enthalpy of absorption of
CO2 with alkanolamine solutions predicted from reaction equilibrium constants. Chem. Eng.
Sci. 64(9), 2027–2038 (2009)
26. J.I. Lee, F.D. Otto, A.E. Mather, Measurement and prediction of solubility of mixtures of
carbon-dioxide and hydrogen-sulfide in a 2.5 N monoethanolamine solution. Can. J. Chem.
Eng. 54(3), 214–219 (1976)
27. G. Kuranov, B. Rumpf, N.A. Smirnova, G. Maurer, Solubility of single gases carbon dioxide
and hydrogen sulfide in aqueous solutions of N-methyldiethanolamine in the temperature range
313 – 413 K at pressures up to 5 MPa. Ind. Eng. Chem. Res. 35(6), 1959–1966 (1996)
28. J.W. Mason, B.F. Dodge, Equilibrium absorption of carbon dioxide by solutions of the
ethanolamines. Trans AIChE 32, 27–48 (1936)
502 J.-Y. Coxam and K. Ballerat-Busserolles

29. J.H. Jones, H.R. Froning, E.E. Claytor, Solubility of acidic gases in aqueous monoethanolamine.
J. Chem. Eng. Data 4(1), 85–92 (1959)
30. J.I. Lee, F.D. Otto, A.E. Mather, Solubility of carbon dioxide in aqueous diethanolamine solu-
tions at high pressures. J. Chem. Eng. Data 17(4), 465–468 (1972)
31. J.I. Lee, F.D. Otto, A.E. Mather, Equilibrium between carbon dioxide and aqueous
monoethanolamine solutions. J. Appl. Chem. Biotechnol. 26(1), 541–549 (1976)
32. J.D. Lawson, A.W. Garst, Gas sweetening data: equilibrium solubility of hydrogen sulfide and
carbon dioxide in aqueous monoethanolamine and aqueous diethanolamine solutions. J. Chem.
Eng. Data 21(1), 20–30 (1976)
33. E.E. Isaacs, F.D. Otto, A.E. Mather, Solubility of mixtures of hydrogen sulfide and carbon
dioxide in a monoethanolamine solution at low partial pressures. J. Chem. Eng. Data 25(2),
118–120 (1980)
34. D.M. Austgen, G.T. Rochelle, C.C. Chen, Model of vapor-liquid equilibria for aqueous acid
gas-alkanolamine systems. 2. Representation of hydrogen sulfide and carbon dioxide solubility
in aqueous MDEA and carbon dioxide solubility in aqueous mixtures of MDEA with MEA or
DEA. Ind. Eng. Chem. Res. 30(3), 543–555 (1991)
35. K.P. Shen, M.H. Li, Solubility of carbon dioxide in aqueous mixtures of monoethanolamine
with methyldiethanolamine. J. Chem. Eng. Data 37(1), 96–100 (1992)
36. O.F. Dawodu, A. Meisen, Solubility of carbon dioxide in aqueous mixtures of alkanolamines.
J. Chem. Eng. Data 39(3), 548–552 (1994)
37. F.Y. Jou, A.E. Mather, F.D. Otto, The solubility of CO2 in A 30-mass-percent
monoethanolamine solution. Can. J. Chem. Eng. 73(1), 140–147 (1995)
38. J.H. Song, J.H. Yoon, H. Lee, K.H. Lee, Solubility of carbon dioxide in monoethanolamine
+ ethylene glycol + water and monoethanolamine + poly(ethylene glycol) + water. J. Chem.
Eng. Data 41(3), 497–499 (1996)
39. I.S. Jane, M.H. Li, Solubilities of mixtures of carbon dioxide and hydrogen sulfide in water +
diethanolamine + 2-amino-2-methyl-1-propanol. J. Chem. Eng. Data 42(1), 98–105 (1997)
40. S. Ma’mun, R. Nilsen, H.F. Svendsen, O. Juliussen, Solubility of carbon dioxide in 30 mass %
monoethanolamine and 50 mass % methyldiethanolamine solutions. J. Chem. Eng. Data 50(2),
630–634 (2005)
41. H. Arcis, K. Ballerat-Busserolles, L. Rodier, J.Y. Coxam, Enthalpy of solution of carbon diox-
ide in aqueous solutions of monoethanolamine at temperatures of 322.5 K and 372.9 K and
pressures up to 5 MPa. J. Chem. Eng. Data 56(8), 3351–3362 (2011)
42. F.Y. Jou, A.E. Mather, F.D. Otto, Solubility of hydrogen sulfide and carbon dioxide in aqueous
methyldiethanolamine solutions. Ind. Eng. Chem. Process Des. Dev. 21(4), 539–544 (1982)
43. A. Chakma, A. Meisen, Solubility of carbon dioxide in aqueous methyldiethanolamine and N,
N-bis(hydroxyethyl)piperazine solutions. Ind. Eng. Chem. Res. 26(12), 2461–2466 (1987)
44. F.Y. Jou, J.J. Carroll, A.E. Mather, F.D. Otto, The solubility of carbon-dioxide and hydrogen-
sulfide in a 35 wt-percent aqueous-solution of methyldiethanolamine. Can. J. Chem. Eng. 71(2),
264–268 (1993)
45. G.W. Xu, C.F. Zhang, S.J. Qin, W.H. Gao, H.B. Liu, Gas-liquid equilibrium in a CO2 -MDEA-
H2 O system and the effect of piperazine on it. Ind. Eng. Chem. Res. 37(4), 1473–1477 (1998)
46. S.W. Rho, K.P. Yoo, J.S. Lee, S.C. Nam, J.E. Son, B.M. Min, Solubility of CO2 in aqueous
methyldiethanolamine solutions. J. Chem. Eng. Data 42(6), 1161–1164 (1997)
47. D. Silkenbaumer, B. Rumpf, R.N. Lichtenthaler, Solubility of carbon dioxide in aqueous solu-
tions of 2-amino-2-methyl-1-propanol and N-methyldiethanolamine and their mixtures in the
temperature range from 313 to 353 K and pressures up to 2.7 MPa. Ind. Eng. Chem. Res. 37(8),
3133–3141 (1998)
48. J.I. Baek, J.H. Yoon, Solubility of carbon dioxide in aqueous solutions of 2-amino-2-methyl-
1,3-propanediol. J. Chem. Eng. Data 43(4), 635–637 (1998)
49. W.J. Rogers, J.A. Bullin, R.R. Davison, FTIR measurements of acid-gas-methyldiethanolamine
systems. AIChE J. 44(11), 2423–2430 (1998)
50. B. Lemoine, Y.G. Li, R. Cadours, C. Bouallou, D. Richon, Partial vapor pressure of CO2
and H2S over aqueous methyldiethanolamine solutions. Fluid Phase Equilib. 172(2), 261–277
(2000)
14 CO2 Capture in Industrial Effluents. Calorimetric Studies 503

51. M.A. Pacheco, S. Kaganoi, G.T. Rochelle, CO2 absorption into aqueous mixtures of diglyco-
lamine and methyldiethanolamine. Chem. Eng. Sci. 55(21), 5125–5140 (2000)
52. A.P.-S. Kamps, A. Balaban, M. Jodecke, G. Kuranov, N.A. Smirnova, G. Maurer, Sol-
ubility of single gases carbon dioxide and hydrogen sulfide in aqueous solutions of N-
methyldiethanolamine at temperatures from 313 to 393 K and pressures up to 7.6 MPa: new
experimental data and model extension. Ind. Eng. Chem. Res. 40(2), 696–706 (2000)
53. M.K. Park, O.C. Sandall, Solubility of carbon dioxide and nitrous oxide in 50 mass
methyldiethanolamine. J. Chem. Eng. Data 46(1), 166–168 (2000)
54. S. Bishnoi, G.T. Rochelle, Thermodynamics of piperazine/methyldiethanolamine/water/carbon
dioxide. Ind. Eng. Chem. Res. 41(3), 604–612 (2002)
55. R. Sidi-Boumedine, S. Horstmann, K. Fischer, E. Provost, W. Fürst, J. Gmehling, Experimental
determination of carbon dioxide solubility data in aqueous alkanolamine solutions. Fluid Phase
Equilib. 218(1), 85–94 (2004)
56. B.S. Ali, M.K. Aroua, Effect of piperazine on CO2 loading in aqueous solutions of MDEA at
low pressure. Int. J. Thermophys. 25(6), 1863–1870 (2004)
57. M. Kundu, S.S. Bandyopadhyay, Modelling vapour—liquid equilibrium of CO2 in aqueous
N-methyldiethanolamine through the simulate annealing algorithm. Can. J. Chem. Eng. 83(2),
344–353 (2005)
58. A. Benamor, M.K. Aroua, Modeling of CO2 solubility and carbamate concentration in DEA,
MDEA and their mixtures using the Deshmukh-Mather model. Fluid Phase Equilib. 231(2),
150–162 (2005)
59. M.H. Jenab,M. bedinzadegan Abdi, S.H. Najibi, M. Vahidi, N.S. Matin, Solubility of carbon
dioxide in aqueous mixtures of N-methyldiethanolamine + piperazine + sulfolane. J. Chem.
Eng. Data 50(2), 583–586 (2004)
60. P.J.G. Huttenhuis, N.J. Agrawal, J.A. Hogendoorn, G.F. Versteeg, Gas solubility of H2S and
CO2 in aqueous solutions of N-methyldiethanolamine. J. Pet. Sci. Eng. 55(1–2), 122–134
(2007)
61. M.H. Oyevaar, H.J. Fontein, K.R. Westerterp, Equilibria of carbon dioxide in solutions of
diethanolamine in aqueous ethylene glycol at 298 K. J. Chem. Eng. Data 34(4), 405–408
(1989)
62. M.L. Kennard, A. Meisen, Solubility of carbon dioxide in aqueous diethanolamine solutions
at elevated temperatures and pressures. J. Chem. Eng. Data 29(3), 309–312 (1984)
63. D. Lal, F.D. Otto, A.E. Mather, The solubility of H2S and CO2 in A diethanolamine solution
at low partial pressures. Can. J. Chem. Eng. 63(4), 681–685 (1985)
64. J.I. Lee, F.D. Otto, A.E. Mather, Solubility of mixtures of carbon-dioxide and hydrogen-sulfide
in aqueous diethanolamine solutions. Can. J. Chem. Eng. 52(1), 125–127 (1974)
65. D.J. Seo, W.H. Hong, Solubilities of carbon dioxide in aqueous mixtures of diethanolamine
and 2-amino-2-methyl-1-propanol. J. Chem. Eng. Data 41(2), 258–260 (1996)
66. M.Z. Haji-Sulaiman, M.K. Aroua, A. Benamor, Analysis of equilibrium data of CO2 in aqueous
solutions of diethanolamine (DEA), methyldiethanolamine (MDEA) and their mixtures using
the modified Kent Eisenberg model. Chem. Eng. Res. Des. 76(8), 961–968 (1998)
67. M.Z. Haji-Sulaiman, M.K. Aroua, Equilibrium of CO2 in aqueous diethanolamine (DEA) and
amino methyl propanol (AMP) solutions. Chem. Eng. Commun. 140(1), 157–171 (1994)
68. H. Arcis, K. Ballerat-Busserolles, L. Rodier, J.-Y. Coxam, Measurement and modeling of
enthalpy of solution of carbon dioxide in aqueous solutions of diethanolamine at temperatures
of (322.5 and 372.9) K and pressures up to 3 MPa. J. Chem. Eng. Data 57(3), 840–855 (2012)
69. R.H. Stokes, K.N. Marsh, R.P. Tomlins, An isothermal displacement calorimeter for endother-
mic enthalpies of mixing. J. Chem. Thermodyn. 1(2), 211–221 (1969)
70. R. Battino, K. Marsh, An isothermal displacement calorimeter for the measurement of the
enthalpy of solution of gases. Aust. J. Chem. 33(9), 1997–2003 (1980)
71. J.J. Christensen, L.D. Hansen, D.J. Eatough, R.M. Izatt, R.M. Hart, Isothermal high pressure
flow calorimeter. Rev. Sci. Instrum. 47(6), 730–734 (1976)
72. J.B. Ott, C.E. Stouffer, G.V. Cornett, B.F. Woodfield, R.C. Wirthlin, J.J. Christensen, U.K.
Deiters, Excess enthalpies for (ethanol + water) at 298.15 K and pressures of 0.4, 5, 10, and
15 MPa. J. Chem. Thermodyn. 18(1), 1–12 (1986)
504 J.-Y. Coxam and K. Ballerat-Busserolles

73. H. Arcis, L. Rodier, J.-Y. Coxam, Enthalpy of solution of CO2 in aqueous solutions of 2-amino-
2-methyl-1-propanol. J. Chem. Thermodyn. 39(6), 878–887 (2007)
74. H. Arcis, L. Rodier, K. Ballerat-Busserolles, J.Y. Coxam, Enthalpy of solution of CO(2) in
aqueous solutions of methyldiethanolamine at T=322.5 K and pressure up to 5 MPa. J. Chem.
Thermodyn. 40(6), 1022–1029 (2008)
75. H. Arcis, Etude thermodynamique de la dissolution du dioxyde de carbone dans des solutions
aqueuses d’alcanolamines (Université Blaise Pascal, Clermont-ferrand, 2008)
Chapter 15
Adsorption Microcalorimetry,
IR Spectroscopy and Molecular
Modelling in Surface Studies

Vera Bolis

Abstract The advantage in coupling adsorption microcalorimetry with IR

spectroscopy and/or ab initio modelling in surface studies is illustrated by a selection
of examples dealing with metal oxides and silica-based (either non porous or micro-
porous) materials. Correlations between thermodynamic and vibrational parameters
are illustrated for the adsorption of CO and of NH3 , employed as probe molecules
for studying the Lewis/Brønsted acidity of the investigated materials. Surface recon-
struction processes, responsible for endothermic effects, are invoked to interpret the
unexpectedly low heat measured in the calorimetric cell for the high-coverage adsorp-
tion of CO on transition aluminas, the adsorption of NH3 on a defective all-silica
zeolite and the adsorption of H2 O on an amorphous alumino-silicate.

15.1 Introduction

Heat evolved when (reactive) molecules contact the surface of a solid material is
related to the molecule/surface site bonding energy. By processing the overall set of
volumetric-calorimetric data at increasing equilibrium pressure, both magnitude of
the enthalpy of adsorption and its evolution with increasing coverage are obtained.
In such a way the possible heterogeneity of the surface (structural, chemical and/or
induced) is quantitatively described.
The volumetric-calorimetric data obtained by a Tian-Calvet heat-flow micro-
calorimeter connected to a gas-volumetric apparatus (as the one described in Chap. 1)
being intrinsically molar quantities, their molecular interpretation often requires a
multi-techniques approach.

V. Bolis (B)
Dipartimento di Chimica and NIS Centre of Excellence, Università di Torino,
Via Pietro Giuria 7, 10125 Torino, Italy
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 505

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_15,
© Springer-Verlag Berlin Heidelberg 2013
506 V. Bolis

Adsorption microcalorimetry is suitably and fruitfully coupled with complemen-

tary techniques, in particular with the ones which are intrinsically apt to detect the
nature of both the pristine surface sites and the new terminations created by the chem-
ical interaction of the (probe) molecules with the surface sites [1–3]. IR spectroscopy
is an ideal technique for this purpose [4, 5]. On the other hand, the increasing amount
of computational ab initio results on adspecies formed at simulated model sites is
ideally suited to be compared with experimental data from microcalorimetry [6–10].
IR spectroscopy experiments and/or ab initio calculations being strategically
planned and performed in parallel with the microcalorimetric measurements allow
to describe the surface features of a solid material in an almost exhaustive way
(nature/structure, population and strength of the surface sites) [11–13].
The combined use of adsorption microcalorimetry, IR spectroscopy and/or ab
initio calculations will be illustrated for a selection of non-dissociative chemical
adsorption cases. In the reported examples, no new surface terminations were orig-
inated during the adsorption: only a perturbation of the vibrational features of the
adsorptive when adsorbed at the surface was so detected by IR spectroscopy. The
experimental details of the microcalorimetric measurements, including the descrip-
tion of the investigated materials, were reported in Chap. 1.

15.2 CO Adsorbed on Coordinatively Unsaturated

Metal Cations

The adsorption of carbon monoxide has been widely used over the years in studies
aimed at characterizing the surface acidity of coordinatively unsaturated (cus) Men+
cations. The cus metal cations, either exposed at the dehydrated surface of oxidic
materials or located within the dehydrated zeolite nanocavities as charge-balancing
cations, are known to act as Lewis acidic sites of variable strength [14].
On non d/d0 cus metal cations, for which any d-electrons π -back-donation is
prevented, CO is either simply polarized by the local electric field generated by the
positive charge of the cus metal cation, or is σ -coordinated to the cus metal cation
through the C – end lone pair [1–15]. In the former case weak electrostatic adducts,
in the latter case reversible chemisorbed species of variable stability are formed. In
both electrostatic and σ -coordinated adducts, the C–O stretching frequency of the
adsorbed molecule is upwards shifted with respect to the stretching frequency of
the free molecule in the gas phase (νC Oads > νC Ogas = 2143 cm−1 ). It has been
observed that the extent of the C–O stretching frequency shift, defined as νC O =
νC Oads − νC Ogas , is a measure of the extent of the lone pair donation to the cus
cations and so of their electron-accepting properties [16–18]. In this respect, νC O
can be taken as a measure of the acidic strength of the cus cations, and is correlated to
the enthalpy of adsorption, which directly measures the strength of the Men+ ← CO
bond [4, 15, 19–21].
15 Adsorption Microcalorimetry, IR Spectroscopy 507

In the case of d-block cus metal cations, the strength of the σ -coordinative bond
in the carbonyl-like species is reinforced by the π -back donation of d electrons.
The actual spectral position of the νC O band turns out to be a compromise between
the upward shift due to the polarization/σ -coordination and the downwards shift
due to the π -back donation [1, 22]. As a consequence, the stretching frequency of
adsorbed CO can arrive to suffer a downwards shift with respect to the free molecule
(νC Oads < νC Ogas ).
In Fig. 15.1 the possible cus metal cation/CO interactions are schematically
illustrated. In Fig. 15.1a the plain electrostatic polarization/σ -coordination of the
molecule on non d/d0 cus metal cations, implying an upwards shift of the C–O
stretching frequency with respect to the free molecule, is reported. In Fig. 15.1b
the σ -coordination + π -back donation on d-block cus metal cations, for which the
C−O stretching frequency can be either slightly upward shifted or downward shifted
with respect to the free molecule, according to the weight of the π -back donation
contribution with respect to the electrostatic/σ -coordination contribution.
In Fig. 15.2 the shift of the CO stretching frequency (νco ) is reported as a func-
tion of the negative enthalpy of adsorption (−a H ) for CO adsorbed on cus Men+
cations (either non d, d0 or d in nature). The reported plot summarizes a large number
of experimental data. νco and −a H values were obtained through parallel exper-
iments carried out in different laboratories (see ref. [1] for details on the samples and
on the experimental conditions). The reported data will be discussed for CO adsorbed
at vanishing coverage on: (i) non-d/d0 metal cations (circle), (ii) copper and silver
metal carbonyls (square and diamonds, respectively), and for CO adsorbed at high
coverage on: (iii) transition catalytic alumina (star). Note that series (i) includes also
a d10 metal cation (cus Zn2+ ) which was found to behave similarly to the non-d
metal cations, both when hosted in Y-zeolite cages or when exposed at the surface
of dehydrated ZnO [23].
As far as non-d/d0 metal cations are concerned, a linear correlation between the
spectroscopic and thermodynamic quantities does exist (left-top side of the figure),
resulting in the Eq. 15.1:

Fig. 15.1 The possible cus metal cation/CO interactions are schematically illustrated. a electrostatic
polarization/σ -coordination (non d/d0 metal cations); b σ -coordination + π -back donation of d
electrons (d-block metal cations)
508 V. Bolis

Fig. 15.2 νco ver sus − a H for CO adsorbed at vanishing coverage on cus metal cations
exposed at the surface of either microporous or non-porous dehydrated systems (see Table 3 in
ref. [1] for details). Circle non-d/d0 metal cations; square Cu-carbonyls; diamond Ag-carbonyls;
star high CO coverage on transition catalytic alumina. Adapted from ref. [1] Fig. 6
 
νC O = (1.03 ± 0.05) cm−1 mol (kJ)−1 |a H | + (−4 ± 3) cm−1 (15.1)

with R = 0.981 [1]. The observed linearity for non-d/d0 metal cations in Fig. 15.2
arises from a statistically significant number of pairs of experimental data, so that
Eq. 15.1 can be considered as an empirical rule of general validity for all non-d/d0
CO complexes.
In the case of d-block metal cations, a lack of correlation with the Eq. 15.1 empir-
ical rule is expected as a consequence of the π -back donation of d-electrons. It was
already mentioned that the presence of a π -back-donation contribution does reinforce
the strength of the carbonyl bond and does cause a downwards shift of the stretching
frequency [22, 24, 25]. This was confirmed by plotting the pairs of IR spectroscopic
and microcalorimetric data for Cu- and Ag-carbonyls formed either within the zeo-
lite nanopores or at the surface of metal oxides. The reinforcement of the carbonyl
bond was witnessed by: (i) the large heat of adsorption (q0 > 100 kJ mol−1 ) mea-
sured for both for Cu- and Ag- carbonyls (see Chap. 1, Fig. 1.14) and (ii) the partial
irreversibility of the adsorption upon evacuation of pressure (see Chap. 1, Figs. 1.10
and 1.11).
A deviation from the non-d/d0 metal cations linear plot was clearly observed for
both copper and silver species, with remarkable differences between the two group 11
metals. The increase of −a H values for both Cu- and Ag- carbonyls with respect
to the non-d/d0 metal cations adspecies was not accompanied by a correspondent
increase of the upwards shift of the C − O stretching frequency.
The formation of well-defined [Cu (CO)]+ complexes within the zeolite nanopores
was characterized by (−a H )0 ≈120 kJ mol−1 and (νco ) comprised in the
16–14 cm−1 range, according to different authors [1, 26]. These pairs of values
well agreed with those obtained for CO adsorbed on Cu/SiO2 ·Al2 O3 (−a H =
115 − 112 kJ mol−1 and νco = 16 − 15 cm−1 ), as reported in ref. [27].
The low-enthalpy values for Cu-carbonyls reported in Fig. 15.2 refer to νco
and −a H pairs obtained for partially reduced Cu- species. In one case CO was
15 Adsorption Microcalorimetry, IR Spectroscopy 509

adsorbed on a Cu(I)-MFI specimen, in which Cu(I) sites were already engaged with
NH3 ligands, so yielding a mixed amino-carbonyl [Cu (NH3 ) CO]+ complex as illus-
trated in ref. [1]. The zero-coverage enthalpy change of ≈80 kJ mol−1 indicated
 +
a weakening of the CO/ Cu (NH3 )n interaction with respect to the CO/Cu(I)
one. The CO stretching frequency for the amino-carbonyl complex was negative
(νco = −43 cm−1 ), confirming the major change in the amino complex Cu/CO
interaction with respect to that of the bare Cu(I) cations. In fact, the actual charge
 +
density of copper cations in Cu (NH3 )n was much lower than that of the pristine
cus Cu(I) cations, owing to the presence of the charge-releasing NH3 ligands. This
was clearly demonstrated by XANES spectroscopy, as reported in ref. [1, 28]. In
the other reported cases, CO was adsorbed on Cu-species grafted on a non-porous
ZnO matrix. Cu- species were reduced by thermal treatments in H2 , either partially
in the Cu(δ+)/ZnO sample or totally in the Cu(0)/ZnO one, as reported in ref. [29].
For both partially and totally reduced Cu- species, the stretching frequency shift was
negative (νco = −45 and −74 cm−1 , respectively) and the enthalpy change was
significantly lower (−a H ≈66 and 43 kJ mol−1 , respectively) than that for cop-
per sites located within the zeolite nanocavities, and characterized by a well-defined
oxidation state of the metal.
By plotting νco against −a H values for the adsorption of CO on the various
Cu-sites, as illustrated in Fig. 15.2, it is rather evident that the spectroscopic and
thermodynamic parameters are once more linearly correlated, as witnessed by the
Eq. 15.2:
 
νC O = (1.21 ± 0.07) cm−1 mol (kJ)−1 |a H | + (−127 ± 6) cm−1 (15.2)

with R = 0.993 [1]. The linear decrease of νco with decreasing −a H by
 +  +
moving from Cu (CO)2 to Cu (NH3 )n (CO) complexes, and down to both
Cu(δ+)/ZnO and Cu(0)/ZnO species was interpreted as due to the decrease of the
actual charge of copper cations which brought about a progressive extinguishment of
the carbonyl bond electrostatic component. This result strongly suggests the major
role played by the electrostatic component in the carbonyl-like bonds.
By the inspection of Ag/CO data reported in Fig. 15.2, it is rather evident
that the deviation from the Eq. 15.1 empirical line for Ag-carbonyls was not as
large as for Cu- carbonyls. This datum indicates that the π -back-donation con-
tribution was lower for the former than for the latter. In fact, the zero-coverage
enthalpy of formation for Ag(I)-carbonyl (≈100 kJ mol−1 ) was lower than for
Cu(I)-carbonyl (≈120 kJ mol−1 ), as reported in ref. [1] and pointed out in Chap. 1
Fig. 1.14. As for the stretching frequency shift, the upwards shift measured for Ag(I)-
carbonyls was νco = 50 cm−1 , much larger than that measured for Cu(I)-carbonyls
(16 − 14 cm−1 ), confirming the π -back donation minor contribution in the Ag/CO
bond. Data for CO adsorbed on Ag(I) sites dispersed at the SiO2 ·Al2 O3 surface
were in fairly good agreement with those for Ag(I)-MFI, as it was reported in ref.
[27]. Conversely, a lower enthalpy change was measured for Ag sites characterized
by silver species in a not-well defined oxidation state, either hosted within a zeo-
510 V. Bolis

lite nanopores (≈80 kJ mol−1 ) or dispersed at the SiO2 surface (≈70 kJ mol−1 ).
The correspondent stretching frequency shift (νC O = 34 cm−1 for the former and
41 cm−1 for the latter) was lower than for Ag(I)-MFI (50 cm−1 ) [27].
In the silver case, owing to the limited set of partially reduced Ag-carbonyls data,
it was not possible to draw any reasonable correlation. A clear trend was however
observed: once more as far as the electrostatic component contribution decreases, as
a consequence of the reduced actual charge of the metal cation, the strength of the
interaction decreases.
By a further inspection of Fig. 15.2, it turns out that the pair of −a H and
νco values for CO adsorbed on transition catalytic aluminas is definitely out of
the Eq. 15.1 correlation line, if the high coverage values are considered.
In fact, the pairs of values for CO adsorbed at low coverage on both γ –
and δ, θ −Al2 O3 were found to fit quite well with the Eq. 15.1 correlation plot:
νC O = 60 − 70 cm−1 and −a H ≈60−70 kJ mol−1 . Conversely, at high CO cov-
erage the measured heat was surprisingly low (−a H ≈10 kJ mol−1 ) with respect
to the vibrational parameter (νC O = 55 cm−1 ).
It is here recalled that in cation-deficient spinel transition aluminas, the Al(III)
cations are located either in tetrahedral or octahedral cavities of the cubic close
packed array of oxide ions [15, 30]. In the γ – and δ, θ −Al2 O3 here discussed, two
CO adspecies were formed at the surface when outgassed at T = 773 K. The two
adspecies were defined as (CO) A and (CO) B in refs. [15, 31], being their presence
witnessed by the appearance of νC O bands at ≈2205 and 2215 cm−1 , respectively.
A third CO adspecies defined as (CO)C , witnessed by the appearance of a νC O
band at ≈ 2230 cm−1 , was observed only when the surface was outgassed at T =
1023 K [15]. The formation of more than one CO adspecies was ascribed to the
presence of the tetrahedral cus Al(III) cations located in different crystallographic
positions. The tetrahedral cus Al(III) cations are in fact known to be the only sites
able to adsorb at room temperature, being the octahedral cus ones too weakly acidic
to do so [4, 30, 31]. The vanishing coverage enthalpy of formation of the three species
(−a H ≈ 58, 73 and 83 kJ mol−1 for (CO) A , (CO) B and (CO)C , respectively) was
found correlated with the correspondent shift: 62, 72 and 87 cm−1 for (CO) A , (CO) B
and (CO)C , respectively) as reported in refs. [1, 15].
In Fig. 15.3 the heats of adsorption for γ - and δ, θ −Al2 O3 , both activated at T =
773 K, are reported as a function of CO uptake in comparison with the correspondent
data for TiO2 -anatase (activated at T = 673 K). The zero-coverage heat of CO
adsorption on γ - and δ, θ −Al2 O3 (≈60 kJ mol−1 ) was close to the value measured
for CO adsorbed on cus Ti(IV) cations on TiO2 [20] and was compatible with a
plain electrostatic polarization/σ -coordination of CO at the Lewis acidic AI(III)
sites [4, 32]. However, remarkable differences were observed between the titania
and alumina plots as far as the surface coverage increased. The heat of adsorption
on TiO2 decreased only slightly upon increasing coverage, whereas on Al2 O3 (both
γ − and δ, θ -phases) the heat values dropped abruptly down to very low values
(q ≈10 kJ mol−1 , at the highest coverage reached at pCO ≈80 Torr).
The assignment of this surprisingly low heat value to the adsorption of CO on
Lewis acidic sites other than the tetrahedral Al(III) cations was discarded by the IR
15 Adsorption Microcalorimetry, IR Spectroscopy 511

Fig. 15.3 Differential heat of

adsorption (at T = 303 K) as
a function of the increasing
CO uptake on: TiO2 -anatase
pre-outgassed at T = 673 K
(diamond); γ −Al2 O3 (tri-
angle) and δ, θ−Al2 O3
(square), both pre-outgassed
at T = 773 K. Adapted from
ref. [4] Fig. 3b

spectra which confirmed the presence of the sole σ -coordination of CO, through the
C-end, on the Lewis acidic tetrahedral Al(III) sites [4, 15, 30, 31].
A process other than a plain σ -coordination taking place at the Al2 O3 /CO inter-
face was invoked to justify the lack of correlation with the linear plot described
by Eq. 15.1 for the high-coverage −a H value. It was demonstrated that the heat
measured at high coverage comprised two different processes: (i) the exothermic
σ -coordination of CO on tetrahedral cus Al(III) cations, and (ii) an endothermic sur-
face reconstruction accompanying the adsorption process. In this respect, it is here
recalled that the overall heat measured within the calorimetric cell is irrespective
of how many and what kind of processes are actually occurring at the gas–solid
interface.
In Fig. 15.4a , the presence of surface Al(III) cations engaged with strained Al–O
bonds (and so not available as such for the interaction with molecules) is schemati-
cally illustrated [4, 30]. The adsorption of on this kind of Al(III) cations, which causes
the O–Al groups interaction to be broken, is schematically illustrated in Fig. 15.4b.
This effect was in fact evidenced at high coverage of CO by IR spectroscopy in the
region of the Al–O modes (1100–1000 cm−1 ), and was found to be entirely reversible
[4, 30, 31, 33].
The modification of the surface structure produced by the rupture of the
Al· · · O–Al moieties interaction is intrinsically endothermic, and caused the mea-
sured heat for the Al2 O3 /CO interaction to be lower than what expected for a plain
σ -coordination. This result does justify the dramatically low heat of interaction mea-
sured within the calorimetric cell, and does explain the lack of correlation between
the energetic and vibrational parameters [1, 20, 34, 35].
In conclusion, in transition aluminas only the zero-coverage heat of adsorption,
which is correlated to the shift of the stretching frequency, can be taken as a measure
of the electron accepting properties of cus Al(III) cations and of the strength of the
σ -dative bonds formed.
The lack of correlation between thermodynamic and spectroscopic data for CO
adsorbed on transition aluminas revealed that the interaction at the gas-solid interface
512 V. Bolis

Fig. 15.4 a Surface Al(III) cations engaged with strained Al–O bonds in transition aluminas; b the
adsorption of CO on the engaged Al(III) cations causes the Al(III)/O–Al interaction to be broken

is more complicated than expected. Also in this respect, the joint use of adsorption
microcalorimetry and IR spectroscopy was proved very useful in arriving at a detailed
molecular interpretation of the process.

15.3 NH3 Adsorbed on All-Silica MFI Zeolites (Silicalite)

Defective Silicalite is an all-silica MFI zeolite, which is non-hydrophobic and weakly

acidic as a consequence of the abundant polar defects (Si–OH nests) generated by
the structure to compensate the Si atoms vacancies [7]. IR spectra in the ν O H stretch-
ing frequency region give a clear evidence of the presence of Si–OH nests, which
are characterized by a different geometrical arrangement according to the synthesis
procedure and/or post-synthesis treatments, as illustrated in ref. [36].
The shift of the stretching frequency of the Si–OH terminations (ν O H ) when
the latter are engaged in H-bonding interaction with molecules is generally taken as
a measure of the strength of the H-bonding interaction [6, 13, 37–39]. In the case of
NH3 adsorbed on a variety of silica–based materials (defective MFI-Silicalite and
amorphous non-porous silica) ν O H was found to be linearly correlated to the zero-
coverage adsorption enthalpy (−a H )0 , as illustrated in Fig. 15.5. See ref. [36] for
details on the samples, and see also Chap. 1 Fig. 1.15b for the differential heat versus
coverage plots of the MFI-Silicalite investigated specimens.
The pair of ν O H and (−a H )0 values measured for one of the investigated MFI-
Silicalite samples (Sil-B, see Chap. 1 Fig. 1.15b) was found to deviate from the linear
correlation illustrated in the plot. The zero-coverage adsorption enthalpy for Sil-B

Fig. 15.5 Shift of the

νOH stretching frequency
(ν O H ) versus the zero-
coverage adsorption enthalpy
(−a H )0 measured for NH3
on a variety of silica–based
materials (defective MFI-
Silicalite and amorphous
non-porous silica). Adapted
from ref. [36] Fig. 6
15 Adsorption Microcalorimetry, IR Spectroscopy 513

Fig. 15.6 Cluster model of a secondary building unit of MFI-Silicalite simulating a Si–OH nest
arranged in a closed chain (ring), either free (C3) or in H-bonding interaction with one NH3 molecule
(C3/NH3 ). Geometries were fully optimized without constraints at B3-LYP/6-31+G(d,p) level.
Atomic symbols for cluster atoms are reported in C3 model; atomic symbol for ammonia N atom
on C3/NH3 model. Adapted from ref. [36] Fig. 8

(≈73 kJ mol−1 ) was lower than that expected on the basis of the νco value
(≈85 kJ mol−1 ). It is here recalled that the heat released for the adsorption of NH3
on Sil-B was lower than for the other investigated samples in the whole coverage
examined (see curve 3 in Chap. 1 Fig. 1.15).
Also in this case, the apparently anomalous behavior of one sample was interpreted
as an indication of the presence of phenomena others than a plain adsorption. In fact,
the adsorption of NH3 on Sil-B was demonstrated to imply an endothermic step
which lowered the measured heat value.
The Sil-B sample was characterized by a structured spectroscopic response in the
ν O H region. The peculiar structure of the ν O H band of this specimen was interpreted
as due to the presence of small highly structured rings of mutually interacting silanol
groups located within the nanopores of the zeolite. This was also suggested by neutron
diffraction data [40, 41] and was confirmed by an ab initio modeling study. A nest
arranged in a closed chain (ring) was simulated by a cluster model of a secondary
building unit of MFI-Silicalite. The model is reported in Fig. 15.6, either free (C3)
or in H-bonding interaction with one molecule (C3/NH3 ). Geometries were fully
optimized without constraints at B3-LYP/6-31+G(d,p) level. The ab initio modeling
study demonstrated that rings must be first “opened” in order to become available
for the interaction with NH3 . The insertion of one molecule in the ring required the
breaking of the Si–OH· · · OH–Si mutual H-bonding interaction, the energetic cost
of which was computed to be as high as ≈12 kJ mol−1 . This value turned out to
be in fairly good agreement with the experimental difference between the expected
(≈85 kJ mol−1 ) and the measured (≈73 kJ mol−1 ) enthalpy of adsorption [36].
Also in this case, the accurate interpretation at molecular detail of the volumetric-
calorimetric results was achieved thanks to the joint use of microcalorimetry, IR
spectroscopy and ab initio calculations.
514 V. Bolis

Fig. 15.7 a volumetric isotherms, b differential heats of adsorption versus coverage plots of H2 Ovap
adsorbed at T = 303 K on H-BEA (square) and on SiO2 ·Al2 O3 (diamond). H-BEA was pre-
outgassed at T = 873 K, SiO2 ·Al2 O3 at T = 673 K. Solid symbols first run; open symbols second
run of adsorption. Adapted from ref. [7] Fig. 8

15.4 H2 O Vapor Adsorbed on Crystalline and on Amorphous

Alumino-Silicates

The adsorption of water on a crystalline zeolite (H-BEA) and on an amorphous

alumino-silicate specimen (SiO2 ·Al2 O3 ) was studied in order to investigate the
nature/structure of the acidic species originated by the presence of Al moieties the
silica matrix [7]. It is generally accepted that proton-exchanged zeolites and alumino-
silicate materials of similar composition exhibit different acidic features [42]. In
particular, it was demonstrated that the Brønsted acidic strength, which is strongly
dependent on the structural features of the material, is highest for H-zeolites [43].
In Fig. 15.7, volumetric isotherms (Sect. a) and differential heats (Sect. b) of water
vapor (H2 Ovap ) adsorption for the two investigated specimens are reported. The two
samples,1 were pre-outgassed for 2 h at either T = 873 K (H-BEA) or T = 673 K
(SiO2 ·Al2 O3 ), at a residual pressure p ≤ 10−5 Torr in order to ensure a maximum
density of Lewis and Brønsted acidic sites.
The H – BEA specific adsorption capacity was much larger than that of SiO2 ·Al2 O3 ,
as witnessed by the volumetric isotherms. For instance at p = 6 Torr, uptake was
5.3 mmol g−1 for against only 1.7 mmol g−1 for SiO2 ·Al2 O3 . In both cases the
adsorption was found to be partially irreversible (2nd run isotherms lie below 1st
run isotherms) but the percent of the irreversible component was much larger for
SiO2 ·Al2 O3 than for H-BEA: ≈40 % versus ≈20 % of the total uptake, respectively.
This datum suggested that H2 Ovap adsorption on the amorphous alumino-silicate
caused an irreversible modification of the surface larger than on the crystalline
specimen.

SiO2
1 H-BEA ( Al 2 O3
= 4.9) features are reported in Chap. 1 (Sect. 1.4.1). The amorphous alumino-
SiO2
silicate ( Al 2 O3
= 5.8) was purchased by Strem Chemicals, Inc.
15 Adsorption Microcalorimetry, IR Spectroscopy 515

SiO2 ·Al2 O3 being more reactive towards water than H-BEA, a larger energy of
interaction with H2 Ovap was expected for the former than for the latter. Surprisingly,
the heat of adsorption curves reported in Fig. 15.7b indicate that the heat of adsorption
was dramatically larger for the crystalline than for the amorphous alumino-silicate,
in the whole range of coverage examined. The zero-coverage heat of adsorption was
≈160 kJ mol−1 for H-BEA and ≈110 kJ mol−1 for SiO2 ·Al2 O3 . At high coverage,
the former sample heat values lie well above the latent heat of liquefaction of water,
q L = 44 kJ mol−1 , while the latter sample heat values approach q L . The same trend
was observed for the 2nd run reversible adsorption.
Also in this case, an endothermic step during the adsorption was invoked as a
possible explanation for the large SiO2 ·Al2 O3 surface reactivity surprisingly not
accompanied by a high energy of interaction with water. The endothermic step of the
overall process was due to the deformation/reconstruction of the surface, consequent
to the interaction with water molecules. This process is expected to be facilitated by
the flexible structure of the amorphous alumino-silicate and inhibited by the rigidity
of the crystalline zeolite framework. In fact, the Lewis acidic cus Al(III) atoms,
when making part of a rigid zeolite framework, acquire a close similarity with the
Lewis acidic sites exposed at the ionic surface of transition aluminas [44]. This
was confirmed by the closeness of the zero-coverage heats of adsorption of H2 Ovap
on δ−Al2 O3 and H-BEA (≈180 and ≈160 kJ mol−1 , respectively) [7]. Conversely,
the covalent SiO2 ·Al2 O3 cus Al(III) species being exposed at a pliable amorphous
Si − O − Al surface were much less available for the interaction with molecules than
those making part of the above mentioned rigid structures. As a consequence of the
surface reconstruction allowing Al(III) species to interact with water molecules, the
measured heat for SiO2 ·Al2 O3 was lower than what expected for a plain adsorption.
This interpretation was supported by the ab initio modeling results described in
ref. [7]. Owing to the large uncertainty from the experiments of the local structure
around the Al atom, two topological different clusters (LS and LC structures) were
designed to simulate the Lewis acidic site, as illustrated in Fig. 15.8, top side row.
LS model was adopted to mimic highly strained moieties, typical of defects located
in nanocavities. [32] The cluster LC was conversely adopted to simulate the surface
species partially saturated by the coordination with an additional nearby framework
O atom. This structure can be reasonably assumed as a model for SiO2 ·Al2 O3 ,[7] in
which the Al atom is allowed to expand its coordination thanks to the flexibility of the
Si-O-Al amorphous structure. The structure of all clusters, either free or interacting
with water (vide infra), were fully optimized at ab initio level using the B3-LYP/6-
31+G(d,p) model chemistry [32, 45].
In the bottom side row of Fig. 15.8, the B3-LYP/6-31+G(d,p) optimized struc-
tures of LS and LC clusters interacting with one molecule (LSW and LCW, respec-
tively) are reported. The computed water binding energies (BE) were corrected
for the basis set superposition error, using the standard Boys-Bernardi counter-
poise method [46]. It turned out that H2 O interacts much less strongly with LC
(BE−LCW ≈109 kJ mol−1 ) than with LS (BE−LSW ≈160 kJ mol−1 ), in agree-
ment with the lower local coordinative unsaturation of the LC-Al(III) atom with
respect to that of the LS model. When H2 O is adsorbed at the LC site, a fraction of
516 V. Bolis

Fig. 15.8 Top side row B3-LYP/6-31+G(d,p) optimized clusters mimicking Lewis acidic Al(III)
sites. LS model mimics highly strained Al(III)atoms, typical of defects present in H-BEA zeolites;
LC cluster simulates the Al(III) atoms coordinating an additional nearby framework O atom. Bottom
side row B3-LYP/6-31+G(d,p) optimized structures of the LS and LC clusters interacting with one
H2 O molecule (LSW and LCW, respectively). Atomic symbols are reported in the model cluster
LS. Adapted from ref. [7] Figs. 2 and 3

the adsorption energy is lost to pull out the Al atom from the amorphous framework.
Conversely, in the crystalline material the acidic sites are already in place, as imposed
by the rigidity of the structure. The “extraction” of the Al-containing site requires
an energy cost which was computed to be ≈ 25 % of the total binding energy for the
LCW case [7].
Once more the ab initio modeling results allowed to properly interpret the appar-
ently anomalous results obtained by adsorption microcalorimetry, confirming the
hypothesis formulated to justify the experimental data.

15.5 Conclusions

A few examples of adsorption processes accompanied by an endothermic step due

to the deformation/reconstruction of the surface in interaction with molecules were
illustrated. In the reported cases, the heat measured within the calorimetric cell was
the combination of an exothermic (adsorption) and an endothermic (surface recon-
struction) effect, which caused the calorimetrically measured heat to be lower than
what expected on the basis of a plain adsorption. An extra-care in interpreting (at
molecular level) the experimental calorimetric results should be addressed in several
cases, and in this respect it is quite fruitful to complement the molar volumetric-
calorimetric data with results from other approaches, typically the various spectro-
scopic methods and/or the ab initio molecular modeling.
15 Adsorption Microcalorimetry, IR Spectroscopy 517

References
1. V. Bolis, A. Barbaglia, S. Bordiga, C. Lamberti, A. Zecchina, Heterogeneous nonclassical
carbonyls stabilized in Cu(I)- and Ag(I)-ZSM-5 zeolites: thermodynamic and spectroscopic
features. J. Phys. Chem. B 108(28), 9970–9983 (2004). doi:10.1021/Jp049613e
2. V. Bolis, S. Maggiorini, L. Meda, F. D’Acapito, G.T. Palomino, S. Bordiga, C. Lamberti, X-
ray photoelectron spectroscopy and x-ray absorption near edge structure study of copper sites
hosted at the internal surface of ZSM-5 zeolite: A comparison with quantitative and energetic
data on the CO and NH3 adsorption. J. Chem. Phys. 113(20), 9248–9261 (2000). doi:10.1063/
1.1319318
3. V. Bolis, S. Bordiga, C. Lamberti, A. Zecchina, A. Carati, F. Rivetti, G. Spano, G. Petrini,
Heterogeneity of framework Ti(IV) in Ti-silicalite as revealed by the adsorption of NH3 Com-
bined calorimetric and spectroscopic study. Langmuir 15(18), 5753–5764 (1999). doi:10.1021/
la981420t
4. V. Bolis, G. Cerrato, G. Magnacca, C. Morterra, Surface acidity of metal oxides. Combined
microcalorimetric and IR-spectroscopic studies of variously dehydrated systems. Thermochim
Acta 312, 63–77 (1998)
5. V. Aina, F. Bonino, C. Morterra, M. Miola, C.L. Bianchi, G. Malavasi, M. Marchetti, V.
Bolis, Influence of the chemical composition on nature and activity of the surface layer of Zn-
substituted Sol-Gel (bioactive) glasses. J. Phys. Chem. C 115(5), 2196–2210 (2011). doi:10.
1021/Jp1101708
6. J. Sauer, P. Ugliengo, E. Garrone, V.R. Saunders, Theoretical-study of Van-Der-Waals com-
plexes at surface sites in comparison with the experiment. Chem. Rev. 94(7), 2095–2160 (1994).
doi:10.1021/cr00031a014
7. V. Bolis, C. Busco, P. Ugliengo, Thermodynamic study of water adsorption in high-silica
zeolites. J. Phys. Chem. B 110(30), 14849–14859 (2006). doi:10.1021/Jp061078q
8. C. Busco, V. Bolis, P. Ugliengo, Masked Lewis sites in proton-exchanged zeolites: a compu-
tational and microcalorimetric investigation. J. Phys. Chem. C 111(15), 5561–5567 (2007).
doi:10.1021/Jp0705471
9. M. Corno, C. Busco, V. Bolis, S. Tosoni, P. Ugliengo, Water adsorption on the stoichiometric
(001) and (010) surfaces of hydroxyapatite: a periodic B3LYP study. Langmuir 25(4), 2188–
2198 (2009). doi:10.1021/La803253k
10. M. Corno, A. Rimola, V. Bolis, P. Ugliengo, Hydroxyapatite as a key biomaterial: quantum-
mechanical simulation of its surfaces in interaction with biomolecules. Phys. Chem. Chem.
Phys. 12(24), 6309–6329 (2010). doi:10.1039/C002146f
11. V. Bolis, C. Busco, V. Aina, C. Morterra, P. Ugliengo, Surface properties of silica-based bioma-
terials: Ca species at the surface of amorphous silica as model sites. J. Phys. Chem. C 112(43),
16879–16892 (2008). doi:10.1021/Jp805206z
12. V. Bolis, C. Busco, G. Martra, L. Bertinetti, Y. Sakhno, P. Ugliengo, F. Chiatti, M. Corno,
N. Roveri, Coordination chemistry of Ca sites at the surface of nanosized hydroxyapatite:
interaction with H2 O and CO. Phil. Trans. R. Soc. a-Math. Phys. Eng. Sci. 370(1963), 1313–
1336 (2012). doi:10.1098/rsta.2011.0273
13. M. Armandi, V. Bolis, B. Bonelli, C.O. Arean, P. Ugliengo, E. Garrone, Silanol-related and
unspecific adsorption of molecular ammonia on highly dehydrated silica. J. Phys. Chem. C
115(47), 23344–23353 (2011). doi:10.1021/Jp206301c
14. K.I. Hadjiivanov, G.N. Vayssilov, Characterization of oxide surfaces and zeolites by car-
bon monoxide as IR probe molecule. Adv. Catal. 47, 307–511 (2002). doi:10.1016/S0360-
0564(02)47008-3
15. V. Bolis, G. Magnacca, C. Morterra, Surface properties of catalytic aluminas modified by
alkaline-earth metal cations: a microcalorimetric and IR-spectroscopic study. Res. Chem.
Intermed. 25(1), 25–56 (1999). doi:10.1163/156856799X00374
16. A. Zecchina, C. Lamberti, S. Bordiga, Surface acidity and basicity: general concepts. Catal.
Today 41(1–3), 169–177 (1998). doi:10.1016/S0920-5861(98)00047-9
518 V. Bolis

17. A.M. Ferrari, P. Ugliengo, E. Garrone, Ab initio study of the adducts of carbon monoxide with
alkaline cations. J. Chem. Phys. 105(10), 4129–4139 (1996). doi:10.1063/1.472283
18. C. Lamberti, S. Bordiga, F. Geobaldo, A. Zecchina, C.O. Arean, Stretching frequencies of
cation CO adducts in alkali-metal exchanged zeolites—an elementary electrostatic approach.
J. Chem. Phys. 103(8), 3158–3165 (1995). doi:10.1063/1.470249
19. C. Morterra, E. Garrone, V. Bolis, B. Fubini, An infrared spectroscopic characterization of the
coordinative adsorption of carbon-monoxide on TiO2 . Spectrochim. Acta Part A-Mol. Biomol.
Spectrosc. 43(12), 1577–1581 (1987). doi:10.1016/S0584-8539(87)80051-X
20. V. Bolis, B. Fubini, E. Garrone, C. Morterra (1989) Thermodynamic and vibrational charac-
terization of CO adsorption on variously pretreated anatase. J. Chem. Soc. Faraday Trans. I 85,
(1383–1395)
21. E. Garrone, V. Bolis, B. Fubini, C. Morterra, Thermodynamic and spectroscopic characteriza-
tion of heterogeneity among adsorption sites—CO on anatase at ambient-temperature. Lang-
muir 5(4), 892–899 (1989). doi:10.1021/la00088a002
22. A.J. Lupinetti, S.H. Strauss, G. Frenking, Nonclassical metal carbonyls. Prog. Inorg. Chem.
49, 1–112 (2001). doi:10.1002/9780470166512.ch1
23. V. Bolis, B. Fubini, E. Garrone, E. Giamello, C. Morterra, in Studies in Surface Science and
Catalysis: Structure and Reactivity of Surfaces, ed. by C. Morterra, A. Zecchina, G. Costa, vol
48 (Elsevier Sci. Publ. B.V., 1989) pp. 159–166
24. S.H. Strauss, Copper(I) and silver(I) carbonyls. To be or not to be nonclassical. J. Chem. Soc.
Dalton Trans. 1, 1–6 (2000). doi:10.1039/a908459b
25. Q. Xu, Metal carbonyl cations: generation, characterization and catalytic application. Coord.
Chem. Rev. 231(1–2), 83–108 (2002). doi:10.1016/S0010-8545(02)00115-7
26. Y. Kuroda, Y. Yoshikawa, R. Kumashiro, M. Nagao, Analysis of active sites on copper ion-
exchanged ZSM-5 for CO adsorption through IR and adsorption-heat measurements. J. Phys.
Chem. B 101(33), 6497–6503 (1997). doi:10.1021/jp9710796
27. Y. Kuroda, H. Onishi, T. Mori, Y. Yoshikawa, R. Kumashiro, M. Nagao, H. Kobayashi, Char-
acteristics of silver ions exchanged in ZSM-5-type zeolite, aluminosilicate, and SiO2 samples:
In comparison with the properties of copper ions exchanged in these materials. J. Phys. Chem.
B 106(35), 8976–8987 (2002). doi:10.1021/Jp020507r
28. V. Bolis, S. Bordiga, G.T. Palomino, A. Zecchina, C. Lamberti, Calorimetric and spectroscopic
study of the coordinative unsaturation of copper(I) and silver(I) cations in ZSM-5 zeolite -
Room temperature adsorption of NH3 . Thermochim. Acta 379(1–2), 131–145 (2001). doi:10.
1016/S0040-6031(01)00612-8
29. E. Giamello, B. Fubini, V. Bolis, Microcalorimetric investigation of the interaction of carbon-
monoxide with coprecipitated cupric oxide zinc-oxide catalysts in well-defined oxidation-
states. Appl. Catal. 36(1–2), 287–298 (1988). doi:10.1016/S0166-9834(00)80122-0
30. C. Morterra, G. Magnacca, A case study: surface chemistry and surface structure of catalytic
aluminas, as studied by vibrational spectroscopy of adsorbed species. Catal. Today 27(3–4),
497–532 (1996). doi:10.1016/0920-5861(95)00163-8
31. C. Morterra, V. Bolis, G. Magnacca, IR spectroscopic and microcalorimetric characterization of
Lewis-acid sites on (transition phase) Al2 O3 using adsorbed CO. Langmuir 10(6), 1812–1824
(1994). doi:10.1021/la00018a033
32. V. Bolis, M. Broyer, A. Barbaglia, C. Busco, G.M. Foddanu, P. Ugliengo, Van der Waals interac-
tions on acidic centres localized in zeolites nanocavities: a calorimetric and computer modeling
study. J. Mol. Catal. a Chem. 204, 561–569 (2003). doi:10.1016/S1381-1169(03)00339-X
33. L. Marchese, S. Bordiga, S. Coluccia, G. Martra, A. Zecchina, Structure of the surface sites
of delta-Al2 O3 as determined by high-resolution transmission electron-microscopy, computer
modeling and infrared-spectroscopy of adsorbed CO. J. Chem. Soc. Faraday Trans. 1 Phys.
Chem. Condens. Phase 89(18), 3483–3489 (1993). doi:10.1039/ft9938903483
34. V. Bolis, B. Fubini, E. Garrone, C. Morterra, P. Ugliengo, Induced heterogeneity at the surface
of group-4 dioxides as revealed by CO adsorption at room-temperature. J. Chem. Soc. Faraday
Trans. 88(3), 391–398 (1992). doi:10.1039/ft9928800391
15 Adsorption Microcalorimetry, IR Spectroscopy 519

35. V. Bolis, C. Morterra, B. Fubini, P. Ugliengo, E. Garrone, Temkin-type model for the description
of induced heterogeneity—CO adsorption on group-4 transition-metal dioxides. Langmuir
9(6), 1521–1528 (1993). doi:10.1021/la00030a017
36. V. Bolis, C. Busco, S. Bordiga, P. Ugliengo, C. Lamberti, A. Zecchina, Calorimetric and IR
spectroscopic study of the interaction of NH3 with variously prepared defective silicalites—
comparison with ab initio computational data. Appl. Surf. Sci. 196(1–4), 56–70 (2002). doi:10.
1016/S0169-4332(02)00046-6
37. A. Zecchina, S. Bordiga, G. Spoto, D. Scarano, G. Petrini, G. Leofanti, M. Padovan, C.O.
Arean, Low-temperature fourier-transform infrared investigation of the interaction of CO with
nanosized ZSM5 and silicalite. J. Chem. Soc. Faraday Trans. 88(19), 2959–2969 (1992). doi:10.
1039/ft9928802959
38. H. Knozinger, Handbook of Heterogeneous Catalysis, vol. 2 (Wiley/VCH, Weinheim, 1997)
39. M. Armandi, B. Bonelli, I. Bottero, C.O. Arean, E. Garrone, Thermodynamic features of the
reaction of ammonia with the acidic proton of H-ZSM-5 as studied by variable-temperature IR
spectroscopy. J. Phys. Chem. C 114(14), 6658–6662 (2010). doi:10.1021/Jp100799k
40. G. Artioli, C. Lamberti, G.L. Marra, Neutron powder diffraction study of orthorhombic and
monoclinic defective silicalite. Acta Crystallogr. Sect. B Struct. Sci. 56, 2–10 (2000)
41. C. Lamberti, S. Bordiga, A. Zecchina, G. Artioli, G. Marra, G. Spano, Ti location in the MFI
framework of Ti-silicalite-1: a neutron powder diffraction study. J. Am. Chem. Soc. 123(10),
2204–2212 (2001). doi:10.1021/Ja003657t
42. M. Trombetta, G. Busca, S. Rossini, V. Piccoli, U. Cornaro, A. Guercio, R. Catani, R.J. Willey,
FT-IR studies on light olefin skeletal isomerization catalysis III. Surface acidity and activity
of amorphous and crystalline catalysts belonging to the SiO2 -Al2 O3 system. J. Catal. 179(2),
581–596 (1998). doi:10.1006/jcat.1998.2251
43. M.M. Huang, A. Auroux, S. Kaliaguine, Crystallinity dependence of acid site distribution
in HA HX and HY zeolites. Microporous Mater. 5(1–2), 17–27 (1995). doi:10.1016/0927-
6513(95)00028-8
44. G. Della Gatta, B. Fubini, L. Stradella, Energies of different surface rehydration processes on
"eta", "theta" and "alpha" aluminas. Journal of the Chemical Society. Faraday Trans 2 Mol.
Chem. Phys. 73 (7):1040–1049 (1977) doi:10.1039/F29777301040
45. V. Bolis, A. Barbaglia, M. Broyer, C. Busco, B. Civalleri, P. Ugliengo, Entrapping molecules in
zeolites nanocavities: a thermodynamic and ab-initio study. Orig. Life Evol. Biosph. 34(1–2),
69–77 (2004). doi:10.1023/B:ORIG.0000009829.11244.d1
46. S.F. Boys, F. Bernardi, The calculation of small molecular interactions by the differences
of separate total energies. Some procedures with reduced errors. Mol. Phys. 19(4), 553–566
(1970). doi:10.1080/00268977000101561
Chapter 16
Characterisation of Catalysts and Adsorbents
by Inverse Gas Chromatography

Eva Díaz and Salvador Ordóñez

Abstract Inverse Gas Chromatography (IGC), in contrast to analytical chromatog-

raphy, consists on adsorption of a known solute on an adsorbent whose properties
are to be determined. The shape and positions of the peaks supply information about
the nature and reactivity of the solid surface. If different probe molecules are used
(i.e. polar and apolar molecules, molecules with acid/base properties), it is possible
to study the specificity of these interactions. Therefore, IGC can be used both as a
tool for both characterizing the adsorption of a given compound on a given solid or
for studying the nature (in terms of acid-base properties, polar or apolar interactions,
etc.) of the active sites of a certain catalyst.

16.1 Introduction

Sorption measurements are a useful method in the characterization of solid materials.

From these data, it is possible to obtain information about the capacity of adsorption,
but also thermodynamic properties—enthalpies of adsorption, surface energy—as
well as kinetic information, such as diffusion rates. Sorption measurements can
be obtained either by static or dynamic methods. Static methods carried out the
adsorption measurements under vacuum, after a pre-treatment at high temperature
in order to clean the material surface. Dynamic methods use a flowing gas device.
Inverse gas chromatography (IGC) is a dynamic method. In comparison to static
adsorption systems, dynamic sorption techniques show shorter measurement time,
and a wider range of experimental possibilities.
In contrast to analytical chromatography, the stationary phase is the sample under
investigation, and the mobile phase acts as probe molecule. Thus, the roles of the

E. Díaz (B) · S. Ordóñez

Department of Chemical Engineering and Environmental Technology, Faculty of Chemistry,
University of Oviedo, C/ Julián Clavería s/n, 33006 Oviedo, Spain
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 521

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_16,
© Springer-Verlag Berlin Heidelberg 2013
522 E. Díaz and S. Ordóñez

phases are inverted and hence, the name of “inverse” gas chromatography. This
technique involves injecting a series of volatile probes and measuring their reten-
tion volumes. Retention volume is related to the interaction parameters between the
probes and the solid and can be converted into a number of surface thermodynamic
properties. Through an adequate choice of the probe molecules to be adsorbed, it is
possible to obtain information about the surface structure and/or surface functional-
ity of adsorbents. After the introduction of the technique and subsequent theoretical
developments, the application of IGC in the materials sciences has developed fast [1].
IGC has been used for the characterization of polymers, copolymers, polymer bends,
biopolymers, industrial fibers, wood and pulp fibers, composites, coatings, pigments,
catalysts, glass beads, coal, chemicals, and steel tubing [2].
Inverse chromatography can be used in the gas phase as well as in the liquid phase.
Although there is some interest in the research of the solid in liquid phase, a vast
literature has done o gas phase (IGC), thus this chapter will be centred in IGC.

16.2 Experimental

IGC measurements can be carried out using a pulse or continuous technique. The
pulse of probe molecule is introduced into the carrier gas stream. This pulse is
transported by the carrier gas through the system to the column with the solid sample.
On the stationary phase, adsorption and desorption occur and the result is a peak in
the chromatogram. The ratio of adsorption/desorption is governed by the partition
coefficient. At fixed conditions of temperature and flow rate, the time of retention of
a compound is characteristic of the system. An alternative is the frontal technique.
This is carried out by injection into the carrier gas stream of a continuous stream of
the probe molecule. When the sample enters into the column, there is a distribution
between phases, and the concentration profiles takes the shape of a plateau, preceded
by a breakthrough curve. The shape of this curve is characteristic of each system [3].
The benefit of the frontal technique is that equilibrium can be always established
due to its continuous nature while pulse chromatography requires the assumption of
a fast equilibration of the probe molecule adsorption on the surface. Between both
techniques, the main part of publications describes pulse experiences, since they are
faster, easier to control and more accurate, especially if interactions between probe
molecules and the adsorbent are weak.
The experimental set-up for the pulse chromatographic experiments consists of
a column inside an oven, with an inlet of the carrier gas with the probe molecule,
and the detector at the exit of the column. The pure carrier gas is introduced into the
column (packed with the material under study). The injection of the sample takes
place prior the oven of the chromatograph and it can be done by different methods [4]:
• by a syringe via the manual injector port of the chromatographic device, consisting
of vapor or liquid;
16 Characterisation of Catalysts and Adsorbents by Inverse Gas Chromatography 523

• by a vapour headspace system, in this case a carrier gas is passed through a reservoir
containing the prove molecule in its liquid form. The gas is saturated with the
prove molecule and then flowing through the injection loop. Concentration can be
controlled by the temperature in the reservoir and the loop volume. This saturated
carrier gas is injected into another pure carrier gas stream.
Concerning the carrier gas, helium or nitrogen, the most important requirement
is to ensure its purity and dryness, since some adsorption processes are highly sen-
sitive to traces of impurities, in particular moisture. Add to the retention time of the
probe molecule, it is necessary also to know the dead-time of the system—time that
the probe molecule would require travelling along the column without any interac-
tion. Obviously, this dead-time can not be measured using the probe molecule since
interactions will be always present. Thus, another molecule, knows as tracer, with
negligible interactions with both the adsorbents and the column walls is employed.
Usual tracers molecules are methane, hydrogen, nitrogen, or even air. The retention
time, as well as the area under the peak, is measured by flame ionisation (FID) or
thermal conductivity (TCD) detector. The FID has the benefit of the sensitive, but
it is limited to organic samples, while TCD is more versatile. Occasionally mass
spectrometric detectors are also used. This is particularly interesting for experiments
where two o more probe molecules are injected.
As far as columns are concerned, they are constructed from glass or metal tubes.
Furthermore, in the case of metals, they are usually stainless steel column, with
passivated inner walls to avoid interactions. In the literature, there is a wide variety
about column lengths and diameters for different applications. The main criteria for
selecting column dimensions are the following [4, 5]:
• Small column diameter, in order to keep gas-phase diffusion effects to minimum.
• Ratio between the column diameter and the particle diameter:

Din /d p > 10 (16.1)

This ratio between both diameters ensures minimization of the effect of channelling
at the wall.
• Ratio between the column length and the particle diameter:

L/d p > 50 (16.2)

This ratio minimizes the axial dispersion. The column length is not so crucial as the
bed length of the packed stationary phase. Usually packed beds are supported on a
porous filter or hold in place with glass wool plugs. For this reason, the column can
be longer than the packing. To avoid additional peak broadening is recommended
to pack the free place with inert material of the same particle size (glass), Fig. 16.1.
The length of the packing depends on the uptake capacity of the sample and the
amount of probe injected. It must be sure that the retention is strong enough (good
separation between probe and tracer peak) to obtain reproducible and accurate
results. This can be checked by repeating the column with different masses. If the
524 E. Díaz and S. Ordóñez

Glass Glass
wool wool

adsorbent

Fig. 16.1 Scheme of column packing and inert material disposition

measured parameters are independent of mass of the adsorbent, enough amount

of packing is used.
• Particle size should be selected to minimize the effects of intraparticular diffusion.
Concerning the flow rate, it is supposed that the lower the flow rate the more likely
the equilibrium of a system is reached. However, low flow rates mean longer exper-
imental times and broader peaks, thus accuracy in the retention time determination
is decreased. A chromatographic column can be considered as a sum of discrete but
contiguous narrow layers, or plates. At each plate, equilibration of the solute between
the mobile and stationary phase was assumed to take place. Movement of the solute
down the column was then treated as a stepwise transfer of equilibrated mobile phase
from one plate to the next. So, a chromatographic column is constituted by a number
of steps, N , with a length, L. Efficiency studies of a chromatographic column have
generally been carried out by determining H as a function of mobile-phase velocity
u, according to the van Deemter equation (Eq. 16.3) [6]:

B B
H=A+ + Cu = A + + (CS + CM )u (16.3)
u u
where H is the plate height in centimeters; u, the linear velocity of the mobile phase
in centimeters per second; and the quantities A, B, and C are coefficients related
to the phenomena of multiple flow paths, longitudinal diffusion, and mass transfer
between phases, respectively. The C coefficient can be divided into two coefficients,
one related to the stationary phase (Cs ) and one related to the mobile phase (C M ). The
van Deemter equation contains terms linearly and inversely proportional to, as well as
independent of, the mobile phase velocity. Taking into account these considerations,
it is recommended to repeat the experiment at different flow rates and determine the
optimum (minimum H ) via van Deemter equation, Fig. 16.2a.
16 Characterisation of Catalysts and Adsorbents by Inverse Gas Chromatography 525

Molecule concentration
Contribution to H (cm)
(a) (b)

(L - σ) (L + σ )

Linear velocity, u (cm/s) Column length (cm)

Fig. 16.2 a Plot of the plate height versus the mobile phase velocity; b Calculation of plate height

The plate height, H , is given by, Eq. (16.4):

σ2
H= (16.4)
L
Thus, the plate height can be obtained from the length of column that contains a
fraction of the probe molecule that lies between L − σ and L (Fig. 16.2b). Because
the area under a normal error curve bounded by σ is about 68 % of the total area, the
plate height, as defined, contains approximately 34 % of the probe molecule.
Once the plate height is known, the number of plates of the column is obtained
directly from the length of the chromatographic column (N = L/H ).
Taking into consideration the aforementioned points related to both the column
dimensions and packing and the gas flow rate, and working at very low concentra-
tions of adsorbate, symmetric peaks are obtained for many materials. Under these
conditions, the hypothesis of infinite dilution can be considered. Due to several both
experimental conditions and instrumental problems, broadening of the peaks can be
observed—avoiding the use of the Gaussian peak, explained later. Some of these
causes are:
• Inadequate column dimensions, amount of stationary phase or gas flow rate, thus
axial dispersion could be the responsible of the broadening.
• Large volume of adsorbate injections.
• Dead volumes in detector or injector.
• Imperfect column packing.
However, sometimes, even working at very low concentrations of the probe mole-
cule and following the aforementioned recommendations, peaks with a large broad-
ening are obtained. In these cases, it can be assumed that this asymmetry is not due
to instrumental problems or experimental conditions. This effect is characteristic of
the so-called “slow kinetic process” [7], Fig. 16.3, associated with markedly ener-
getically heterogeneous surfaces containing preferential sites were desorption takes
place in a slower way. Slow kinetic process depends on the concentration of the
526 E. Díaz and S. Ordóñez

Fig. 16.3 Effect of slow

kinetic process

Counts
Time (min)

solute, and is attenuated when the amount of adsorptive is decreased. Moreover, they
correspond to non equilibrium situations, thus the gas flow rate have a significant
effect on it.

16.3 Adsorption Isotherms

Adsorption isotherms of gases or vapors are the basis upon which the surface char-
acteristics of adsorbents are defined. From this magnitude, specific surface area,
porosity and other properties of the solid can be obtained. Even more, using adsor-
bates with diverse physical and chemical characteristics, it is possible to define the
type of adsorbate-adsorbent interactions involved and the nature of the adsorption
in the system tested. From the chromatographic peaks, the adsorption isotherms can
be directly obtained. Figure 16.4 presents a general relation between the chromato-
graphic peak and the adsorption isotherm shape. In the case of infinite dilution, a
symmetrical (gaussian) peak is observed representing a linear Henry isotherm. At
high concentration (finite dilution) tailing or leading will occur. In the case of a type
I, II, or IV isotherm there is a tailing because adsorbent/adsorbate interactions are
much stronger than adsorbate/adsorbate interactions.
Glückauf [8, 9] develop a method for obtaining the adsorption isotherms from
chromatographic peaks in which a continuous stream of adsorbate is injected into the
column until saturation, and the adsorbed material is then eluted by a pure carrier gas
stream. The adsorption isotherm is calculated from the shape of the desorption curve.
Gregg and Stock [10, 11] demonstrated that it was possible to obtained the adsorption
isotherms from chromatographic data for all types of Brunauer isotherm. For this
purpose, high concentrations of adsorbate are applied, but the effect of gradient
pressure was neglected in most of the experimental work published before 1968 [12].
Taking into account these considerations, adsorption isotherms can be obtained either
from the ideal GC, at conditions of infinite dilution, or by non-ideal and non-linear
chromatography, at conditions of finite dilution. Ideal GC is described here, whereas
finite chromatography isotherms determination can be found in the literature [3, 12].
16 Characterisation of Catalysts and Adsorbents by Inverse Gas Chromatography 527

Fig. 16.4 Correlation of (a) (b)

peak form a and adsorption
isotherm b for finite and
infinite dilution

time Partial pressure

Detector response
Amount adsorbed
time Partial pressure

time Partial pressure

At conditions in a chromatographic columns which approach the requirements

of ideal IGC (minimum of van Deemter curve), the following hypothesis can be
supposed:
• Flow through the column under isothermic and isobaric conditions
• Fast fluid-solid transfer
• Negligible intraparticular diffusion
• Axial transport due to convection
• No concentration or velocity gradients in radial direction
• Reversible adsorption and instantaneous equilibrium
In this case, diffusional and kinetical broadening of the chromatographic column
is reduced, and any distortion of the chromatogram is due to the deviation of the
adsorption isotherm from the Henry’s law.
Therefore, it can be described the change in adsorbate concentration across an
increment of the chromatographic column of length dx as:
     
∂c ∂c ∂ca
− u0 v =v + va (16.5)
∂x t ∂t x ∂t x

where u 0 is the lineal gas velocity; v, the gas phase volume in the column; va , the
volume of adsorbate retained on the adsorbent; c, the adsorbate concentration in gas
phase; ca , the adsorbate concentration on the adsorbent; t, the time since the injection,
and x, the length from the beginning of the column and the dx. The first part of the
528 E. Díaz and S. Ordóñez

Fig. 16.5 Graphic integration h

of chromatogram and deter- h
mination of the adsorption
isotherm by the ECP method h3

h2

h1

t0 time

mass balance corresponds to the mass of adsorbate accumulated, whereas the second
part represents the rate of change of the amount of adsorbate in the layer dx.
By successive calculations [12], it is possible to determine the magnitude of
adsorption, a, for an equilibrium concentration of adsorbate, c, in the mobile phase:

c
1
a= VR dc (16.6)
m
0

where m is the mass of adsorbent in the column, and V R , the retention volume.
The most common method of obtaining adsorption isotherms is the Elution of a
Characteristic Point (ECP) [13], which consists of obtaining the isotherm from just
one single injection. By introducing into Eq. (16.6) the magnitudes obtained from
the chromatogram, the detector constant, k, and giving the detector deflections, h,
the value of adsorption, a, is obtained:

m a Sads
a= (16.7)
m Speak

where m a is the mass of injected adsorbate; Sads , is the area bounded by the height
h between the tracer peak and the extender profile of the chromatogram (Fig. 16.5),
and Speak , the peak area.
The equilibrium concentration of adsorbate in the mobile phase can be expressed
as:
ma h
c= (16.8)
F Speak

where F is the flow rate of the carrier gas. The equilibrium pressure is determined
from the equation p = cRT. After substituting in Eq. (16.8), it is obtained the expres-
sion to calculate the equilibrium pressure:

m a h RT
p= (16.9)
F Speak
16 Characterisation of Catalysts and Adsorbents by Inverse Gas Chromatography 529

From a theoretical point of view, the method is accurate and a single elution profile
allows the determination of a complete isotherm [14]. The applicability is restricted
to very efficient columns allowing fast mass transfer, that it, columns possessing a
high plate number. Low concentration measurements allow also to determine Henry
constants, in this range the uptake is independent of the surface coverage. This regime
is ideal also for the measurement of thermodynamic parameters since they can be
obtained with the highest sensitivity. The span of the infinite dilution range depends
on the probe molecules and the heterogeneity of the material. Especially for polar
probe molecules adsorbing on very heterogeneous surface, non-symmetrical peaks
are often observed even with the smallest injection size/concentration. This suggests
that the values obtained under these conditions are not truly representing Henry
conditions; however, they are still use for practical considerations.

16.4 Thermodynamic Parameters

16.4.1 Retention Volume

Application of IGC to study the properties of a solid is based on the assumption that
the adsorbate equilibrium conditions are achieved between the mobile and stationary
phases. Thus, chromatogram should be symmetric and the maximum of the peak
must not depend on the amount of the injected adsorbate. Moreover, as the amount
of adsorbate is very small, the concentration of the adsorbate in the gas phase is
minimal and the adsorption process is conditioned by the real adsorbate-adsorbent
interactions. Under these conditions, the retention volume—key parameter in IGC-
of the solute depends on its partition between the stationary and mobile phase, and
is an indication of the interaction strength between the solute molecule and the
metal/adsorbent surface. The specific retention volume, Vg , in cm3 /g, is given as:
  
(t R − td ) p0 − pw T
Vg = F j (16.10)
m p0 Tmeter

where F is the uncorrected flow rate detected by a bubble flow meter; t R is the
retention time in min; td , the hold-up time or time of a tracer compound in pass
through the column; p0 , the outlet column pressure; pw , the vapor pressure of water
at the flowmeter temperature; T , the column temperature; Tmeter , the ambient tem-
perature, and j, the James-Martin compressibility factor. This parameter represents
the volume of dry gas to elute the adsorbate, corrected at 273 K and per gram of
stationary phase. Add to Vg , it is also very employed the net retention volume, VN ,
defined as the volume of dry gas to elute the adsorbate, corrected at 273 K.
In both cases, the James-Martin factor for the correction of gas compressibil-
ity under pressure difference between column inlet, pi , and column outlet, p0 , is
introduced:
530 E. Díaz and S. Ordóñez

Fig. 16.6 Chromatogram

presenting the Conder and
Young method to obtain the
skewness ratio (a/b) and
retention time, t R

a b
t tracer tL tT
Skewness ratio = a/b

 
3 ( pi / p0 )2 − 1
j= (16.11)
2 ( pi / p0 )3 − 1

In the case of perfect symmetric peaks, the retention time can be determined
directly from the peak maximum method, which is the simplest and most common.
The peak maximum method is useful for determination of retention time if the skew-
ness ratio is 0.7–1.3 [15]. The “skewness ratio” is defined as the ratio of tangent
slope to the peak leading part and tangent slope to the peak tailing part whereas both
tangents are drawn in the inflexion points. In such cases the skewness ratio is our of
this interval, t R is obtained from the first-order moment method or the Conder and
Young method. Between these two methods, Conder and Young is recommended
[16]:
t R = (t L + tT ) /2 (16.12)

where t L and tT are the times at which the tangents drawn to the peak leading and
tailing parts in their inflexion points intersect the zero line, Fig. 16.6.
The retention volume is related to the surface area and surface energy; that is,
the higher the surface area and energy, the higher the retention time, and therefore,
retention volume.
Moreover, the VN and the slope of adsorption isotherm are related by Eq. (16.13)
for small adsorbate injections, where conditions of “infinite dilution” are achieved:
q
VN = K S · A = ·A (16.13)
c
where K s is the inclination of the isotherm at infinite dilution, that is, the Henry’s con-
stant; A, the specific surface area of the adsorbent; q, concentration of the adsorbate
in the stationary phase, and c, concentration of the adsorbate in the gas phase.
16 Characterisation of Catalysts and Adsorbents by Inverse Gas Chromatography 531

16.4.2 Free Energy of Adsorption

Thermodynamics information of the adsorption process at infinite dilution can be

obtained from the retention volume. At infinite dilution, the standard free energy
to transfer 1 mol of adsorbate from the gas phase to the surface at standard state,
defined as the variation in the standard free energy of adsorption, G 0ads (J/mol), can
be expressed as:  
P Vg
G ads = −RT ln
0
(16.14)
π0 A

or its equivalent form:

G 0ads = −RT ln Vg + C (16.15)

where P is the gas phase pressure, A is the specific surface area of the adsorbent, and
π0 is the spreading pressure of the adsorbed gas. Two different standard states can be
considered: in the De Boer state, the spreading pressure has a value of 338 μN/m [17]
at p0 = 1.01 kN/m2 , whereas in the Kemball and Rideal state, at p0 = 1.01 kN/m2 ,
the spreading pressure is 0.0608 μN/m [18]. The parameter, C, is a constant related
to the standard reference states:
 
Aπ
C = −RT ln (16.16)
P

16.4.3 Enthalpy and Entropy of Adsorption

When zero coverage (infinite dilution) conditions are fulfilled the standard differential
heat of adsorption, q0 , is numerically equal to the opposite of the enthalpy of the
process. This value can be obtained from the variation of G 0ads with temperature.
For an equilibrium process this variation is given by Gibbs-Helmholtz equation:

  
∂ −G 0ads /T ∂ (ln VN )
= R = q0 (16.17)
∂(1/T ) ∂ (1/T ) P
P

In the Fig. 16.7 is illustrated the dependence of −G 0ads /T as a function of 1/T
for conventional carbon fibers and carbon fibers oxidized by electrochemical proce-
dure. In the figure is illustrated the dependence of n-heptane. This behaviour, typical
of hydrocarbons, implies that Hads 0 is constant within the temperature range of

characterisation. At this point, it is important to remark that comparison of differ-

ential heats of adsorption with the heats of liquefaction is recommended, in order
to ensure the nature of the interaction. In the cases where the adsorbate-adsorbent
interactions are stronger than adsorbate-adsorbate interactions, Hads0 is higher than

the liquefaction heat.

532 E. Díaz and S. Ordóñez

120

100

-1
-(ΔG/T)/JK mol
80

-1
60

40
CF
20 CFO

0
0.0028 0.0029 0.003 0.0031 0.0032 0.0033 0.0034
(1/T)/ K-1

Fig. 16.7 Variation of (−G 0ads /T ) with (1/T ) for the adsorption of n-heptane on carbon fibers
(adapted from Ref. [19])

From the adsorption standard free energies and standard enthalpies, adsorption
entropies can be calculated from:
0

q + G 0ads 0 − G 0
Hads
Sads
0
=− = ads
(16.18)
T T

In agreement with the linearity in the −G 0ads /T versus 1/T , the adsorption
entropies are independent of the temperature.
Figure 16.8 shows the existence of so-called “thermodynamic compensation
effect”, i.e. a linear dependence of Hads 0 on S 0 . Thermodynamic compensa-
ads
tion effect between the adsorption enthalpy and entropy was observed in different
studies for n-alkanes. It indicates that the stronger adsorption of longer n-alkanes is
accompanied by a higher loss of mobility of the molecules (it means higher inter-
action between the molecule and the surface). This type of plot is used currently
to highlight the differences in adsorbate-adsorbent interactions. A good fit of the
compensation effect data to a straight line indicates the non-specific nature of the
adsorbate-adsorbent interactions.
In the case of Fig. 16.8, three straight lines are depicted, one corresponding to the
data of HSAG-100 and HSAG-300, and two more fitting the CNTs and CNFs points.
Although the line of CNFs is clearly shifted with respect to graphites, the slope of
both lines is virtually the same and different from that of CNTs. Since the adsorbate
type is the same for both sets of data (n-alkanes), the difference in slopes can be
attributed to the existence of two different non-specific surfaces, one of the CNFs
and graphites and the other represented by the CNTs [19]. Likewise, the shift between
graphites and CNFs could be understood since HSAGs contain a large amount of
structural defects, so the interactions could be modified. It is also evident the higher
values of entropy for the nanotubes in comparison with the other materials, due to
high entropy of adsorbate located inside the tubes [21].
16 Characterisation of Catalysts and Adsorbents by Inverse Gas Chromatography 533

Fig. 16.8 Thermodynamic 60

compensation effect of n- CNT
alkanes on carbon nanotubes CNF
(CNT), carbon nanofibers 50

(-ΔΗ) / kJmol -1
HSAG-300
(CNF) and high-surface- HSAG-100
area-graphites (HSAG)
40
(adapted from Ref. [20])

30

20
0 20 40 60
-1 -1
(-ΔS)/JK mol

16.4.4 Work of Adhesion: Dispersive and Specific Contribution

In the absence of chemisorption and interdiffusion, the work of adhesion is the sum
of the different intermolecular forces involved and can be related to the surface free
energies, where a is the compound and the superscripts D and S denote dispersive
and specific interactions:
Wa = WaD + WaS (16.19)

Generally, the work of adhesion is coupled to G 0ads according to:

Gads = −NaWa (16.20)

where N is the Avogadro’s number and a is the surface area of a single probe
molecule.
For two materials interacting only via London dispersive forces across their inter-
face, Fowkes [22–24] suggested that the work of adhesion, Wa , could be described
as the geometric mean approach, where γ LD and γ SD are the dispersive component of
the surface energy of the liquid (the probe) and the solid, respectively:

Wa = WaD = 2 γLD γSD (16.21)

For probes interacting with the solid of interest via dispersive forces, a combina-
tion of Eqs. (16.15, 16.20, 16.21) will lead to:

RT ln VN = 2N · a · γLD γSD + K (16.22)

Thus, according to this approach, developed by Schultz et al. [25], by measuring

the net retention volume for various n-alkane probes and plotting RTlnV N ver-
sus a(γ LD )0.5 , the dispersive component of the surface free energy can be determined
534 E. Díaz and S. Ordóñez

from the slope of a linear fit. The surface free energy is the energy required to form
(or increase the surface by) a unit surface under reversible conditions and is the ana-
logue to the surface tension of a liquid. From the practical point of view, the higher
the surface energy, the more reactive the surface.
In the Schultz expression, it is necessary the molecular area. It can be determined
from the liquid density, ρ, assuming a spherical molecular shape in a hexagonal
close-packing configuration:
 2/3
M
a = 1.09 · 1014 (16.23)
ρN

where M is the molecular weight of the probe molecule.

However, some authors have stressed the difficulties associated with the determi-
nation of the molecular area, especially for non-spherical molecules such as straight
alkanes. To avoid the problems derived from the molecular area, Dorris and Gray [26]
considered the adsorption characteristics of single methylene groups in the n-alkane
probes. By defining the increment in free energy of adsorption per – CH2 - unit:

VN(n)
GCH2 = −RT ln (16.24)
VN(n+1)

where VN and VN +1 are the retention volumes of n-alkanes with (n) and (n + 1)
carbon atoms, respectively. In this way, γ SD can be determined from:

1 GCH2
2
γSD = (16.25)
4 γCH2 N2 aCH
2
2

The benefit of this approach is the fact that despite the use of various n-alkanes
probes, only methylene area, aCH 2 , and surface tension, γCH 2 , have to be known.
The CH2 area is takes as 0.06 nm2 , based on C–C length of 0.127 nm and an average
distance of 0.47 nm for two CH2 [26]. Jacob and Berg [27] have found extraordinary
an excellent agreement between n-alkane molecular areas as determined from the
fitting of experimental adsorption isotherms to the BET model and as obtained by
simply assuming the area of 0.06 nm2 for each methylene group. The parameter
γCH 2 is estimated from the surface tension of a linear polyethylene melt as function
of temperature:

γCH2 (mJ/m2 ) = 35.6 + 0.058 20 − T(◦ C) (16.26)

The validity of this approach has established on the basis that IGC and wettability
measurements lead to approximately the same γ SD value for poly(ethylene tereph-
thalate) [28]. Furthermore, Dorris and Gray stated that the molecular area could be
an adjustable parameter.
Once the dispersive interactions of a surface have been investigated, specific
interactions can be studied by injecting polar probes. For these probes, WaS is usually
16 Characterisation of Catalysts and Adsorbents by Inverse Gas Chromatography 535

Fig. 16.9 Determination 30

of the specific interaction decahydronaphtalene
25
parameter, I sp , for polar tetrahydronaphtalene

-1
probes on Al2 O3 at 250 ◦ C

-ΔG / kJmol
20
(adapted from Ref. [32]) cycloheptane
15 xylene

10 benzene ΔΔ G

0
0.2 0.3 0.4 0.5 0.6
a / nm2

larger than zero, which leads to increased net retention volumes as compared to n-
alkanes. The adsorption of these molecules on the stationary phase is influenced,
not only be dispersive interactions, but also by additional specific contributions.
These specific contributions include dipole-dipole and acid-base interactions, the
latter involving much higher energies than the former ones [35]. In fact, it is usually
assumed that the specific contribution of the adsorption of polar probes are actually
acid-base interaction only. In this way, Fowkes, by analogy with the dispersive work
of adhesion, described the “extended Fowkes equation”:

WaS = 2 γLS γSS (16.27)

However, this expression can not predict accurately the magnitude of the non
dispersive interactions, since it is wrong the assumption that the contribution to the
work of adhesion of two polar compounds could be represented by the geometric
mean value of their polar properties.
In order to quantify this specific contribution, several attempts were made. One is
relate the specific component of the surface free energy to the parameter of specific
interaction of polar solutes (I sp ). This parameter involves the surface properties
in terms of potential and acid-base interactions and may be determined from the
difference of free energy of adsorption, (G), between a polar solute and the real
or hypothetical n-alkane with the same surface area, a, (Fig. 16.9):

 (G) Gads
S
Isp = = (16.28)
Na Na
Add to the surface area as parameter of comparison, the boiling point or the
vapor pressure could also be used [29–31]. This treatment is essentially empirical,
but it allows to compare the specific interaction between the surface and the solute
molecules, based on a unified scale.
Donnet and coworkers [33, 34] obtained useless results with the above method
when analyzing material with relative high dispersive component of the surface
energy (γ SD > 100 mJ/m2 ), such as carbon nanofibers or graphite powders. For
536 E. Díaz and S. Ordóñez

Fig. 16.10 Graphical descrip- 50

tion of the method followed
to obtain the specific contri- 40

-Δ G ads (kJ/mol)
bution of the adsorption free
30 −ΔG ads
SP
energy measured for different
polar probes (adapted from n-alkanes
Ref. [20]) 20

10

0
5 7 9 11 13
1/2 49 3/2 2 -1/2
(hνL) α 0L10 (C m V )

these materials lower (−G 0ads ) values were found for the polar probes compared to
reference n-alkanes. This problem can be encountered by plotting the free energy of
adsorption as function of the molecular polarizability of the different polar adsorbates
(Fig. 16.10):
 
−Gads
0
= −G0D + −G0S

= kc (h υ)1/2 α0,s (h υ)1/2 α0 + −G0S (16.29)

where kc , is the constant of the chromatographic process; h, the Planck constant

(Js); υ, characteristic vibration frequency of the electron (s−1 ); α0 , polarizability
deformation (cm2 V−1 ).
S ) are determined from
Since the free energy of ad Specific interactions (−G ads
differences between (−G ads ) values of the polar probes and the reference line com-
0

posed with data obtained from the elution of n-alkanes. In this way, (−Hads S ) can

be calculated from the variation of (−G ads ) versus (1/T ), as stated in Eq. (16.17).
S

Therefore, the standard enthalpy of adsorption of polar probes is divided into two
contributions, dispersive and specific:

Hads
0
= Hads
D
+ Hads
S
(16.30)

The ability of the polar molecules to donate or accept electrons has been parameter-
ized by means of the donor (DN) and acceptor (AN) number [35]. These parameters
describe the basic and acidic nature, respectively. The DN values (kcal/mol) repre-
sent the enthalpy of formation for the adduct produced when the base in question
reacts with the reference Lewis acid SbCl5 in the 1,2-dichloromethane, as solvent.
However, for the characterization of acids, no similar reference system was found.
AN value (dimensionless) measuring the induced shift in the 31 P NMR spectra of
the base Et 3 PO4 when this compound was dissolved in the acid under investigation.
Riddle and Fowkes [36] corrected the AN scale to the enthalpy of reaction of Et 3 PO4
with SbCl5 . This new parameter, AN*, presents the same units as DN.
16 Characterisation of Catalysts and Adsorbents by Inverse Gas Chromatography 537

Thus, the measured (−HadsS ) can be correlated to the acid and base indices AN*

and DN of the solute probes [35], obtaining information about the surface acidity-
basicity:
− HadsS
= K a · D N + K b · AN ∗ (16.31)

where K a and K b are indices reflecting the acidity (electron acceptor) and basicity
(electron donor) of the solid surface. According to Eq. (16.31), a plot of versus
(DN/AN*) should yield a straight line from which K a can be obtained from the slope
and K b from the intercept. This equation is of empirical nature; other relationships
have been also proposed in the literature [37].

16.4.5 Surface Heterogeneity

There is an important point that must be taken into account in the IGC determined
parameters: for heterogeneous high energy surfaces, molecules will preferentially
adsorb on the highest energy sites [38]. The distribution of energetic sites is usually
called as surface heterogeneity. There exist two types of heterogeneity: structural
and energetic. A typical example of a structural heterogeneity is a wide pore size
distribution, where the geometrical effects determine the adsorption of the probe
molecules. Energetic heterogeneity occurs with a wide distribution of various surface
sites of different energetic levels. The energy heterogeneity can be described either
by the adsorption energy distribution or the adsorption potential distribution.
For the determination of the adsorption energy distribution, F(ε), from Eq. (16.32),
and assumption on the shape of the local isotherms has to be made and usually com-
plex numerical analysis is required [39, 40]:
εmax

θt ( p, T ) = θ1t (ε, p, T ) · F(ε) · dε (16.32)

ε min

where εmin and εmax indicate the range in potential energy of adsorption, and θt refers
to the local adsorption isotherm.
Concerning the adsorption potential distribution, it can be obtained easily form the
adsorption isotherm. Furthermore, is less affected by experimental noise and produce
reliable results. Once the adsorption isotherm is derived from IGC, the adsorption
potential, AP, is calculated according to:
 
ps
AP = RT ln (16.33)
p

where p is the partial pressure, and ps , the saturation pressure.

The distribution parameter, φ, represents the first derivation of the adsorbed
amount of molecules, n, with the adsorption potential:
538 E. Díaz and S. Ordóñez

Fig. 16.11 Heterogeneity 0.008

dn/dA (μmol·mol/g·J)
profiles of CNF (—) and
CNF-oxi (--) at 250 ◦ C for 0.006
benzene (adapted from Ref.
[41]) 0.004

0.002

0
0 20000 40000 60000
A( J/mol)

dn
φ=− (16.34)
dAP
Figure 16.11 shows, as example, the adsorption potential distribution of benzene over
two carbon nanofibers, a parent and an oxidized one. It can be seen that there is a good
coincidence of the profiles; therefore, the adsorbates interact with the same energy
sites. It means that the treatment of the nanofiber does not create new adsorption
sites, just varying slightly the adsorption capacities (area under the curve).

16.5 Applications and Comparison to Other Techniques

Adsorption techniques have been widely applied both to know the adsorption capac-
ity of the adsorbents and to obtain thermodynamic parameter for increasing the
knowledge of the surface of the material. Static methods are, probably, the most used
for these purposes, since they are considered the most accurate. Thus, comparison
of the adsorption parameters obtained by other techniques with static methods is a
usual way to ensure the reliability of a technique. Thielmann and Baumgarten [42]
investigated the adsorption properties of four aluminas with different microporosities
by both IGC and a static method. Sorption measurements obtained form IGC gave
similar results whereas the static experiments showed differences until 37 %. This
difference can be explained by a different micropore structure of the aluminas. This
is so because one of the hypotheses assumed for obtaining adsorption data from the
IGC eluted peak is to consider instantaneous adsorption equilibrium. In the case of
microporous materials, transport of the solutes through the porous structure could
delay the equilibrium, becoming more difficult to satisfy this hypothesis. Thus, the
application of the IGC to microporous materials has been discouraged in the litera-
ture for microporous materials [4]. With the same aim of comparison between IGC
and static techniques, different ion-exchanged zeolites (microporous materials) have
been studied by gas calorimetry coupled to a volumetric line and IGC [43]. n-Hexane
was used as adsorbate and isotherms of adsorption as well as the isosteric heats of
adsorption and enthalpies of adsorption obtained, respectively, by the earlier men-
tioned methods were determined. The comparison between the volumetric adsorption
16 Characterisation of Catalysts and Adsorbents by Inverse Gas Chromatography 539

isotherms of n-hexane at 250 ◦ C obtained both by IGC and calorimetry reveals that,
at low pressures (when both techniques are applicable), the results obtained by both
techniques are very similar. Moreover, the strength of adsorption is also quantify by
the heat of adsorption (IGC) and the isosteric heat (calorimetry), obtaining deviations
among them in the range of 4–20 %, depending on the material. Thus, at very low
partial pressures, IGC could be a reliable technique even for microporous materials.
The wettability of powders is a valuable parameter in different fields such as the
pharmacy [44], the ceramic [45], polymers [46] and nanomaterials [47]. Contact
angle measurements are the most used in order to obtain information about the sur-
face, however, IGC, by the dispersive component of the surface free energy could
also offer a sensitive approach to surface characterization. In this way, comparison
of the surface components using n-alkanes over theophylline and caffeine showed
a good agreement [44]. It is important to mention that the method of calculating
the γ SD value is applied to solid surfaces, flat at the molecular scales, presenting no
chemical, structural or energetic heterogeneities. Thus, application of the method to
other materials could give larger discordances. In this way, in a study of activated
carbons characterization by IGC, Herry et al. [48] have obtained values of γ SD 10
times higher than the ones obtained previously by capillary wetting. This difference
was attributed to the increase in the interaction potential in micropores. The discrep-
ancy in the values obtained by the two techniques could be explained by assuming
that the treatment of the column, prior to the IGC experiments, leads to carbon sur-
faces free from adsorbed species and other possible contaminants. Besides that, the
contribution of high-energy sites will be significantly outweighed in IGC measure-
ments under infinite dilution conditions. In fact, contact angle measurements are
known to provide an average of the surface energy [49]. It is important to remark
that in these situations, the G CH 2 is more reliable, since it keeps a precise physical
meaning [50].
Inverse gas chromatography parameters can also be applied in the field of cataly-
sis. In this way, as example, parent NaX and CaA zeolites, as well as transition metal
(Co2+ , Mn2+ , Fe3+ )-exchanged zeolites, were evaluated for the catalytic oxidation
of n-hexane. It was observed [51, 52], that although there was linear correlation
between the acidity and the adsorption enthalpy of the n-hexane, there was no rela-
tionship between the acidity and the activity for n-hexane oxidation. However, if
a reactivity parameter (such as T50 , temperature at which 50 % of conversion is
attained) is plotted versus the adsorption heat, a so-called Volcano plot is obtained
(Fig. 16.12), an optimum value of (−Hads 0 ) being observed, higher and lower values

yielding to worst catalytic performance.

IGC was also use to elucidate the nature of the interaction of both reactives and
products with the surface of heterogeneous catalysts in order to obtain information
for understanding the mechanism of the reaction. Xie et al. [31] proposed a mech-
anism for the partial oxidation of propylene to acrylic acid based on the adsorption
parameters of reactive and products on styrene divinylbenzene copolymer (SDB)
and Pd supported on SDB. Likewise, Díaz et al. [53] studied the performance of Fe-
ZSM-5 catalysts for benzylation of benzene with benzyl chloride in terms of their
chemical and adsorption properties.
540 E. Díaz and S. Ordóñez

Fig. 16.12 Influence of 500

the adsorption enthalpy of Fe-NaX NaX
n-hexane of metal transition- Co-NaX
CaA
exchanged zeolites on the 400

T 50 / C
T50 parameter (adapted from
Fe-CaA
Ref. [52]) Mn-NaX Mn-CaA
300

Co-CaA
200
20 30 40 50 60
0 -1
−ΔH ads / kJmol

Finally, a mention to the reverse-flow (RF) GC technique, which is a variation of

IGC where the probe molecule is injected at a middle point of the column containing
the adsorbent as stationary phase, and the direction of carrier gas flow is reversed
from time to time. This creates extra chromatographic peaks on the continuous signal.
This technique allows to measured different physicochemical quantities, including
rate constant of surface and gaseous reactions, and experimental isotherms [54, 55].

References

1. A.V. Kiselev, Adsorbents in gas chromatography, in Advances in Chromatography, ed. by J.C.

Giddings, R.A. Keller (Marcel Dekker, New York, 1967)
2. A. Voelkel, Inverse gas-chromatography—characterization of polymers, fibers, modi-
fied silicas, and surfactants. Crit. Rev. Anal. Chem. 22, 411–439 (1991). doi:10.1080/
10408349108051641
3. V.R. Choudhary, L.K. Doraiswamy, Applications of gas chromatography in catalysis. Ind. Eng.
Chem. Prod. Res. Develop. 10, 218–237 (1971). doi:10.1021/i360039a002
4. F. Thielmann, Introduction into the characterisation of porous materials by inverse gas chro-
matography. J. Chromatogr. A 1037, 115–123 (2004). doi:10.1016/j.chroma.2004.03.060
5. R.E. Hayes, S.T. Kolaczkowski, Introduction to Catalytic Combustion (Gordon and Breach
Science Publisher, Amsterdam, 1997)
6. J. van Deemter, F.J. Zuiderweg, A. Klinkerberg, Longitudinal diffusion and resistance to mass
transfer as causes of nonideality in chromatography. Chem. Eng. Sci. 5, 271–289 (1956). doi:10.
1016/0009-2509(56)80003-1
7. M. Montes-Morán, J.I. Paredes, A. Martínez-Alonso, J.M.D. Tascón, Adsorption of n-alkanes
on plasma-oxidized high-strength carbon fibers. J. Colloid Interface Sci. 247, 290–302 (2002).
doi:10.1006/jcis.2001.8134
8. E. Glückauf, Adsorption isotherms from chromatographic measurements. Nature 156, 748
(1945). doi:10.1038/156748c0
9. E. Glückauf, Theory of chromatography. Part II. Chromatograms of a single solute. J. Chem.
Soc. 1302–1308 (1947). doi:10.1039/JR9470001302
10. S.J. Gregg, R. Stock, Gas Chromatography (Desty D.H, London, 1958)
11. S.J. Gregg, The Surface Chemistry of Solids, 2nd edn. (Chapman and Hall, London, 1961)
12. T. Paryjczak, Gas Chromatography in Adsorption and Catalysis (J. Wiley & Sons, New York,
1987)
16 Characterisation of Catalysts and Adsorbents by Inverse Gas Chromatography 541

13. E. Cremer, H. Huber, in Gas Chromatogr: Instr Soc Amer Symp, vol. 3, ed. by N. Brenner, et
al. (Academic Press, New York, 1962), p. 169
14. A. Seidel-Morgenstern, Experimental determination of single solute and competitive adsorption
isotherms. J. Chromatogr. A 1037, 255–272 (2004). doi:10.1016/j.chroma.2003.11.108
15. J.R. Conder, S. McHale, M.A. Jones, Evaluation of methods of measuring gas-solid chro-
matographic retention on skewed peaks. Anal. Chem. 58, 2663–2668 (1986). doi:10.1021/
ac00126a019
16. B. Charmas, R. Leboda, Effect of surface heterogeneity on adsorption on solid surfaces: appli-
cation of inverse gas chromatography in the studies of energetic heterogeneity of adsorbents.
J. Chromatogr. A 886, 133–152 (2000). doi:10.1016/S0021-9673(00)00432-5
17. J.H. De Boer, The Dynamical Character of Adsorption (Clarendon Press, Oxford, 1953)
18. C. Kemball, E.K. Rideal, The adsorption of vapours on mercury. I. Non-polar substances. Proc.
R. Soc. A 187, 53–73 (1946). doi:10.1098/rspa.1946.0065
19. M.A. Montes-Morán, A. Martínez-Alonso, J.M.D. Tascón, Effect of sizing on the surface
properties of carbon fibres. J. Mater. Chem. 12, 3843–3850 (2002). doi:10.1039/B202902B
20. E. Díaz, S. Ordóñez, A. Vega, Adsorption of volatile organic compounds onto carbon nanotubes,
carbon nanofibers, and high-surface-area graphites. J. Colloid Interface Sci. 305, 7–16 (2007).
doi:10.1016/j.jcis.2006.09.036
21. S.Y. Bhide, S. Yashonath, Structure and dynamics of benzene in one-dimensional channels. J.
Phys. Chem. B 104, 11977–11986 (2000). doi:10.1021/jp002626h
22. F.M. Fowkes, Attractive forces at interface. Ind. Eng. Chem. 56, 40–52 (1964). doi:10.1021/
ie50660a008
23. F.M. Fowkes, Donor-acceptor interactions at interfaces. J. Adhesion 4, 155–159 (1972). doi:10.
1080/00218467208072219
24. F.M. Fowkes, M.A. Mostafa, Acid-base interactions in polymer adsorption. Ind. Eng. Chem.
Prod. Res. Dev. 17, 3–7 (1978). doi:10.1021/i360065a002
25. J. Schultz, L. Lavielle, C. Martin, The role of interface in carbon fiber-epoxy composites. J.
Adhesion 23, 45–60 (1987). doi:10.1080/00218468708080469
26. G.M. Dorris, D.G. Gray, Adsorption of n-alkanes at zero surface coverage on cellulose paper and
wood fibers. J. Colloid Interface Sci. 77, 353–362 (1980). doi:10.1016/0021-9797(80)90304-
5
27. P.N. Jacob, J.C. Berg, Acid-base surface energy characterization of microcrystalline cellulose
and two wood pulp fiber types using inverse gas chromatography. Langmuir 10, 3086–3093
(1994). doi:10.1021/la00021a036
28. A. Pizzi, K.L. Mittal, Handbook of Adhesive Technology (Marcel Dekker, New York, 2003)
29. U. Panzer, H.P. Schreiber, On the evaluation of surface interactions by inverse gas chromatog-
raphy. Macromolecules 25, 3633–3637 (1992). doi:10.1021/ma00040a005
30. A. van Asten, N. van Veenendaal, S. Koster, Surface characterization of industrial fibers with
inverse gas chromatography. J. Chromatogr. A 888, 175–196 (2000). doi:10.1016/S0021-
9673(00)00487-8
31. J. Xie, Q. Zhang, K.T. Chiang, An IGC study of Pd/SDB catalysts for partial oxidation of
propylene to acrylic acid. J. Catal. 191, 86–92 (2000). doi:10.1006/jcat.1999.2796
32. E. Díaz, S. Ordóñez, A. Vega, J. Coca, Adsorption properties of a Pd/γ -Al2 O3 catalyst using
inverse gas chromatography. Micropor. Mesopor. Mater. 70, 109–118 (2004). doi:10.1016/j.
micromeso.2004.03.005
33. S. Dong, M. Breadle, J.B. Donnet, Study of solid-surface polarity by inverse gas-
chromatography at infinite dilution. Chromatographia 28, 469–472 (1989). doi:10.1007/
BF02261062
34. J.B. Donnet, S.J. Park, H. Balard, Evaluation of specific interactions of solid-surfaces by inverse
gas-chromatography—a new approach based on polarizability of the probes. Chromatographia
31, 434–440 (1991). doi:10.1007/BF02262385
35. V. Gutmann, The Donor-Acceptor Approach to Molecular Interactions (Plenum Press, New
Cork, 1979)
542 E. Díaz and S. Ordóñez

36. F.L. Riddle, F.M. Fowkes, Spectral shifts in acid-base chemistry. 1. van der Waals contributions
to acceptor numbers. J. Am. Chem. Soc. 112, 3258–3264 (1990). doi:10.1021/ja00165a001
37. T. Hamieh, M. Nardin, M. Rageui-Lescourt, H. Haidara, J. Schultz, Study of acid-base inter-
actions between some metallic oxides and model organic molecules. Colloids Surf. A 125,
155–161 (1997). doi:10.1016/S0927-7757(96)03855-1
38. H. Ishida, Characterization of Composite Materials (Butterworth-Heinemann, London, 1994)
39. M. Pyda, G. Guiochon, Surface properties of silica-based adsorbents measured by inverse gas-
solid chromatography at finite concentration. Langmuir 13, 1020–1025 (1997). doi:10.1021/
la950541f
40. H. Balard, A. Saada, E. Papirer, B. Siffert, Energetic surface heterogeneity of illites and kaolin-
ites. Langmuir 13, 1256–1259 (1997). doi:10.1021/la9515276
41. M.R. Cuervo, E. Asedegbega-Nieto, E. Díaz, A. Vega, S. Ordóñez, E. Castillejos-López, I.
Rodríguez-Ramos, Effect of carbon nanofiber functionalization on the adsorption properties of
volatile organic compounds. J. Chromatogr. A 1188, 264–273 (2008). doi:10.1016/j.chroma.
2008.02.061
42. F. Thielmann, E. Baumgarten, Characterization of microporous aluminas by inverse gas chro-
matography. J. Colloid Interface Sci. 229, 418–422 (2000). doi:10.1006/jcis.2000.6958
43. E. Díaz, S. Ordóñez, A. Auroux, Comparative study on the gas-phase adsorption of hexane
over zeolites by calorimetry and inverse gas chromatography. J. Chromatogr. A 1095, 131–137
(2005). doi:10.1016/j.chroma.2005.07.117
44. J.W. Dove, G. Buckton, C. Doherty, A comparison of two contact angle measurement methods
and inverse gas chromatography to assess the surface energies of theophylline and caffeine.
Int. J. Pharm. 138, 199–206 (1996). doi:10.1016/0378-5173(96)04535-8
45. C.-W. Won, B. Siffert, Preparation by sol-gel method of SiO2 and mullite (3Al2 O3 , 2SiO2 )
powders and study of their surface characteristics by inverse gas chromatography and zetametry.
Colloids Surf. A 131, 161–172 (1998). doi:10.1016/S0927-7757(97)00149-0
46. R.A. Bailey, K.C. Persaud, Application of inverse gas chromatography to characterisa-
tion of a polypyrrole surface. Anal. Chim. Acta 363, 147–156 (1998). doi:10.1016/S0003-
2670(98)00084-1
47. K. Batko, A. Voelkel, Inverse gas chromatography as a tool for investigation of nanomaterials.
J. Colloid Interface Sci. 315, 768–771 (2007). doi:10.1016/j.jcis.2007.07.028
48. C. Herry, M. Baudu, D. Raveau, Estimation of the influence of structural elements of activated
carbons on the energetic components of adsorption. Carbon 39, 1879–1889 (2001). doi:10.
1016/S0008-6223(00)00310-9
49. M.A. Montes-Morán, J.I. Paredes, A. Martínez-Alonso, J.M.D. Tascón, Surface characteriza-
tion of PPTA fibers using inverse gas chromatography. Macromolecules 35, 5085–5096 (2002).
doi:10.1021/ma020069m
50. E. Papirer, E. Brendle, F. Ozil, H. Balard, Comparison of the surface properties of graphite,
carbon black and fullerene samples, measured by inverse gas chromatography. Carbon 37,
1265–1274 (1999). doi:10.1016/S0008-6223(98)00323-6
51. E. Díaz, S. Ordóñez, A. Vega, J. Coca, Characterization of Co, Fe and Mn-exchanged zeolites by
inverse gas chromatography. J. Chromatogr. A 1049, 161–169 (2004). doi:10.1016/j.chroma.
2004.07.065
52. E. Díaz, S. Ordóñez, A. Vega, J. Coca, Catalytic combustion of hexane over transition metal
modified zeolites NaX and CaA. Appl. Catal. B 56, 313–322 (2005). doi:10.1016/j.apcatb.
2004.09.016
53. E. Díaz, S. Ordóñez, A. Vega, A. Auroux, J. Coca, Benzylation of benzene over Fe-modified
ZSM-5 zeolites: correlation between activity and adsorption properties. Appl. Catal. A 295,
106–115 (2005). doi:10.1016/j.apcata.2005.07.059
54. N.A. Katsanos, N. Rakintzis, F. Roubani-Kalantzopoulou, E. Arvanitopoulou, A. Kalantzopou-
los, Measurement of adsorption energies on heterogeneous surfaces by inverse gas chromatog-
raphy. J. Chromatogr. A 845, 103–111 (1999). doi:10.1016/S0021-9673(99)00262-9
55. N.A. Katsanos, Physicochemical measurements by the reversed-flow version of inverse gas
chromatography. J. Chromatogr. A 969, 3–8 (2002). doi:10.1016/S0021-9673(02)00992-5
Chapter 17
Liquid-Solid Adsorption Properties:
Measurement of the Effective Surface Acidity
of Solid Catalysts

Paolo Carniti and Antonella Gervasini

Abstract Attention should be devoted to the measurements of the adsorption prop-

erties of catalytic surfaces when they have to work in liquid-solid heterogeneous
conditions. The mutual characteristics of the surface and the liquid affect the reagent
interactions with the surface sites which could be engaged with the liquid interaction
and then not-available for the reagent coordination. This leads to observe effective
adsorption properties that could be different from the intrinsic properties of the
surface. The possibility to quantitatively determine the effective acid properties of
catalytic surfaces by base adsorption is here showed. The adsorption can proceed
in any type of liquid of various characteristics (apolar, polar, aprotic, protic) with
dynamic (pulse liquid chromatographic method) or equilibrium (liquid recirculation
chromatographic method) methods. The measurements of effective acidity allows
finding more sound relations with the catalytic activity for a better comprehension
of the catalyst work and for a more correct determination of the turnover numbers in
liquid-solid catalysis.

17.1 Introduction

Heterogeneous catalysts have different ability in the chemical transformation of

reagents to form the desired reaction products due to various functionalities present at
their surface. When reactions run in liquid by a liquid-solid heterogeneous catalysis,
not only the characteristics of the surface (hydrophobic or hydrophilic surface) but the
liquid properties (polarity, proticity, solvating ability) govern the process reactivity.

P. Carniti (B) · A. Gervasini

Dipartimento di Chimica, Università degli Studi di Milano,
Via Camillo Golgi, 19, 20133 Milano, Italy
e-mail: [email protected]
A. Gervasini
e-mail: [email protected]

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 543

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5_17,
© Springer-Verlag Berlin Heidelberg 2013
544 P. Carniti and A. Gervasini

The liquid-solid interactions of physical (essentially van der Waals) and chemical
(essentially acid-base) type can quench or put out some surface sites, so modifying
their reactivity towards reagents. Therefore, in order to understand and predict the
catalytic ability of given surfaces in different liquids, it is very important to measure
the effective surface properties, besides the knowledge of the intrinsic properties.
Acid catalysts are among the most numerous and best studied solids finding use in
oil refining processes, base industrial chemistry, and, at present, biomass conversion,
and biorefinery processes [1, 2]. Acid surfaces possess either Brönsted or Lewis acid
sites, or both Brönsted and Lewis sites simultaneously, with the possibility to operate
different chemical transformations of given reagents. Depending on the properties of
the liquid solvent, some acid sites can be silent; for example the Lewis acid sites of
most surfaces are not lively any longer in water due to the solvating ability and weak
basic character of water. The acidity measurement of the solid surface performed
in the gas-solid phase (i.e., by using gaseous/vapor probes) could not be significant
when the catalyst is then employed in liquid-solid phase. Analogous considerations
apply if acidity characterization is performed in the liquid-solid phase with a solvent
having a proticity and polarity very different from those of the reaction solvent.
In this case, the influence of the solvent on the surface acidity cannot be correctly
taken into account. From this viewpoint, the acidity of sulphonic resins in water
[3, 4] and the influence of different solvents on the acidity of supported sulphonic
acid catalysts [5] as well as of niobium containing catalysts [6] have recently been
studied by acid-base titration calorimetry.
On a given acid catalytic surface, the amount and strength of the acid sites with
lively characteristics in defined liquids (effectivity acidity) is of primary importance
for its activity and they merit being determined. A technique widely used in the study
of acidity of solids in aqueous media is the measurement of the isoelectric point (IEP)
which expresses the overall ability of an acidic surface to attract protons [7], without
providing any quantitative information on the surface sites.
Recently, some techniques have been presented in the literature making use of
volumetric titrations of surface sites in liquids, of different polar and protic charac-
teristics, to determine the amount of acid sites and relevant acid strength (effective
acidity). Two different methods will be here discussed: the first one is based on a
pulse liquid chromatographic method (a dynamic method) [8] and the second one
on a liquid recirculation chromatographic method (a quasi-static method) [9]. When
surface acidity studies are concerned, the measurements may be performed in apolar,
aprotic liquid (like cyclohexane), for the determination of the intrinsic acidity, or in
several other liquids with polar/protic characteristics, for the determination of the
effective acidity. Basic probes of different basicity (following the pKa scale) may
be used (e.g., 2-phenylethylamine, PEA, aniline, AN, pyridine, PY, etc.). Titration
temperatures may be varied from room temperature (r.t.) up to the normal boiling
point of the liquid used in order to calculate, from the collected isotherms of adsorp-
tion, the isosteric heats of adsorption which can be related to the acid strength of the
surface sites [10, 11].
17 Liquid-Solid Adsorption Properties: Measurement of the Effective Surface 545

17.2 Pulse Liquid Chromatographic Method

This method enables the characterization of the acidity, or more general of surface
properties, of solids in liquids of different properties making use of an HPLC appa-
ratus. The pulse gas chromatographic technique has been widely employed in the
characterization of solid catalysts, particularly for the estimation of the dispersion
of metal deposited on supports [12, 13] while little has been reported on the use of
pulse liquid chromatographic technique for surface property determinations.
Experiments can be performed in a liquid chromatograph (HPLC), complete with
pump, injector, and detector. The catalyst sample, in a sample holder surrounded by
a thermostatted jacket, is placed in place of the chromatographic column and the
chosen liquid is continuously eluted on the solid. In the adsorption test, pulses of a
given amine solution of known concentration are sent at fixed time intervals onto the
catalyst sample at a constant liquid flow rate. The non-adsorbed amounts of amine
after each pulse are detected and measured up to sample saturation. The amount of
amine adsorbed (mequiv/g) on the sample after the i-th pulse is calculated by the
following equation:

(amine adsorbed)i = (VC/w)(Ā − Ai )/Ā (17.1)

where V (ml) corresponds to the volume of a single pulse, C (M) is the concentration
of the amine solution, w (g) is the mass of the catalyst sample, Ā is the mean
chromatographic area at saturation and Ai is the area corresponding to the i-th pulse.
When the adsorption stoichiometry is known, the total number of acidic sites is
related to the total amount of amine adsorbed on the sample. A typical example of a
chromatogram for an amine adsorption onto a catalyst sample is depicted in Fig. 17.1.

Fig. 17.1 Example of pulse chromatogram for the evaluation of the surface acidity of the IR-200
Amberlite in water employing PEA (from Ref. [8])
546 P. Carniti and A. Gervasini

Fig. 17.2 Evaluation of the surface acidity for some Amberlites measured by the pulse chromato-
graphic technique in water (from Ref. [8]): A120 (squares), A200 (triangles) and A50 (circles);
probes: PEA (open symbols) and AN (filled symbols)

The amount of amine adsorbed up to a given pulse was obtained by the summation
of the single adsorbed quantities via the equation above reported. Plot of the total
amount of amine adsorbed versus the total amine injected provides clear evidence
of the progressive coverage of the sample surface. At saturation, the curve attained
a plateau corresponding to the total amount of amine adsorbed.
Figure 17.2 reports, as an example, the titration in water of some Amberlite sam-
ples by two different basic probes, PEA (strong base, pKa = 9.84) and AN (weak
base, pK a = 4.58). The Amberlite IR-120 (A120) and IR-200 (A200) resins are
strongly acidic gel resins (styrene-divinylbenzene copolymers) with SO− 3 functional
groups, while Amberlite IRC-50 (A50) is a weakly acidic porous resin (metacrylic
acid-divinylbenzene copolymer) with COO− functional groups.
The A120 acidity measured by PEA is around 2.3–2.4 mequiv/g, and that of the
A200 and A50 resins are of the same order (2.0–2.1 mequiv/g). By AN as base probe,
the measured acidity of A120 and A200 was ca. three- or four-times less than that
measured with PEA. For the weak acidic resin A50, an acidity 13-times less than
that measured with PEA was observed by the AN base probe. These results are due
to partially reversible AN adsorption during the pulse flow titration; the formation
of a less stable adsorption complex between the acid sites and AN (in comparison to
the situation with PEA) could occur.
The acidity of acid oxides has been measured by this technique, in this case
the possibility to distinguish the intrinsic from the effective acidity is presented, by
performing the titrations in cyclohexane (aprotic and very weak polar) and water
(highly protic and polar), respectively [8]. A mixed oxide, silica-zirconia (SiZr)
and the same base probes (PEA and AN) used for the titration of Amberlites have
been chosen for this example. The acidity of SiZr determined by pulses of PEA in
cyclohexane is 0.49 mequiv/g, i.e. ca. four-times higher than that in water (Fig. 17.3).
This result can be compared with that obtained by titration with Hammett indicators
[14] in the same solvent employing n-butylamine (pKa = 10.77) as base probe.
17 Liquid-Solid Adsorption Properties: Measurement of the Effective Surface 547

Fig. 17.3 Evaluation of the surface acidity in cyclohexane (diamonds) and in water (circles) for
SiZr measured by the pulse chromatographic technique; probes: PEA (open symbols) and AN (filled
symbols) (from Ref. [8])

In particular, the amount of acid sites titrated with 4-phenylazo-1-naphthylamine

indicator was 0.48 mequiv/g, comparable to the value of 0.49 mequiv/g obtained in
the chromatographic titration measurement. The acid sites titrated with bromothymol
blue amounted to 0.15 mequiv/g. This value is very similar to that obtained in water
with PEA (0.13 mequiv/g by pulse). PEA in water should only be able of titrating the
strong sites of SiZr, due to the water presence which interacts with part of the surface
acid sites. In contrast with the behavior observed with PEA, very little evidence for
adsorption was observed when pulses of AN in cyclohexane were sent onto SiZr.
This unexpected behavior with AN suggests that when the acidity of a solid in a
liquid is measured, attention should be taken not only of the solvent–surface acid
site interactions but also of the solvent–base interactions. Likely, the low basicity
of the AN probe and the high AN affinity for the solvent are the cause of the weak
adsorption on AN on the acid sites and of its desorption during interval between the
successive pulses.

17.3 Liquid Recirculation Chromatographic Method

This method, which also makes use of an HPLC apparatus, enables the character-
ization of the acidity, or more general of surface properties, of solids in liquids of
different properties by collecting solid-liquid adsorption isotherms [9, 15].
Concerning the acid measurements of solid samples, the experiments of acid-base
titrations by base probes can be performed in various liquids, as above reported. In
this case, the HPLC line, comprising the pump and detector coupled to a personal
computer for the collection, storage, and processing of the data, works according to
a recirculating method (Fig. 17.4).
548 P. Carniti and A. Gervasini

Fig. 17.4 Scheme of the apparatus used for the acidity determination in liquid by recirculation
method (from Ref [9])

The sample holder, containing the sample and inserted in a glass jacket connected
with a thermostat to maintain a constant temperature of adsorption, is mounted in
place of the chromatographic column, as in the pulse method. During the titration
experiment, successive dosed amounts of base probe in the chosen liquid solution
are injected into a reservoir inserted into the line in which the solution continuously
circulates. At the beginning of the experiment, the reservoir contains pure solvent.
For each injection, the base solution recirculates onto the sample until adsorption
equilibrium is achieved, revealed by the attainment of the detector signal stability.
Figure 17.5 shows a typical example of raw data obtained from a titration experiment,
consisting on a series of increasing steps, each one representing the attainment of the
adsorption equilibrium. When the stoichiometry of the base adsorption on the acid
site is known, the amount of acid sites per sample mass (mequiv · g−1 ) or sample
area (mequiv · m−2 ) titrated at each equilibrium base concentration can be evaluated,
obtaining the relevant adsorption isotherm on the sample.
The intrinsic (measured in cyclohexane, apolar and aprotic solvent) and effective
(measured in water or other liquids) acidity of series of mixed silica-niobia oxides
have been determined by this method with the principal aim of disclosing relation-
ships between the surface acid properties and catalytic activity [9]. Dispersed niobia
17 Liquid-Solid Adsorption Properties: Measurement of the Effective Surface 549

Fig. 17.5 Example of raw data obtained from the acid–base titration in liquid by recirculation
method. The points representing the base probe injection are indicated by a cross (Ref. [9])

centers in silica (mixed SiO2 − Nb2 O5 oxides) constitute a good acid catalyst system
used in various reactions [15–18]. In particular, It has been proved to be suitable
for the dehydration of monosaccharides (like fructose) to obtain important chem-
ical intermediates (like 5-hydroxymethyl-2-furaldehyde, HMF). The development
of such reactive processes in agreement with a green chemistry approach imposes
the use of friendly solvents, as water. Like Nb2 O5 [19], SiO2 − Nb2 O5 catalysts are
able to maintain good activity in polar and protic liquids, and then to catalyze the
conversion of fructose to HMF in water. They also show better stability than Nb2 O5
in the reaction due to the dilution of the niobia centers into the silica matrix [9]. The
results of the acidity measurements of SiO2 − Nb2 O5 at low (Si-Nb/15org) and high
(Si-Nb/45org) niobia content in water by the recirculation chromatographic method
are shown in Fig. 17.6. As expected, both the Nb-samples have higher effective acid-
ity than the silica support and as higher the amount of niobia as more acid is the
sample. Comparing series of SiO2 − Nb2 O5 samples with different niobia concen-
tration, good relationship between the surface effective acidity of the samples and
their activity in the fructose dehydration reaction has been found [9]. This result sug-
gest some practical importance of the determination of effective surface properties
of solids to understand and predict their activity in given condition, that is, in the
presence of liquids which may interact with part of the surface sites.
Finally, by the liquid recirculation chromatographic method it is possible to deter-
mine the solid-liquid adsorption isotherms at different temperatures. Information
about the energetic of the adsorption process can be obtained through the calculation
of the isosteric heats of adsorption, once converting the isotherms into the corre-
sponding isosteres for given surface coverages of the adsorbed species. Example of
this procedure is given in Fig. 17.7 which shows the adsorption isotherms and rel-
evant isosteres of PEA adsorption in cyclohexane on a typical cracking catalysis, a
SiO2 −Al2 O3 sample (14.7 wt. % of Al2 O3 ) with highly acidity property. The amount
of acid sites titrated decreases with increasing temperature (from 17 to 57 ◦ C) and
550 P. Carniti and A. Gervasini

Fig. 17.6 Example of isotherms of adsorption of PEA on two silica-niobia samples at low
(Si-Nb/5org) and high (Si-Nb/45org) niobia content collected in water at 17 ◦ C by the recirculation
method (from Ref. [9])

0.003 0.0031 0.0032 0.0033 0.0034 0.0035

adsorbed PEA /mmol g -1

-9
1.2 17°C
-1
n / mmol g 0.35
27°C -9.5 0.4
0.8 0.45
42°C -10
ln c

0.5
57°C -10.5
0.4

-11
0.0
0 0.0002 0.0004 0.0006 0.0008 -11.5
[PEA] eq / M T-1 / K-1

Fig. 17.7 Example of isotherms of adsorption of PEA on a SiO2 -Al2 O3 sample collected in
cyclohexane at different adsorption temperatures (left) with corresponding isosteres of adsorption
at given surface PEA coverage (in mmol P E A · g−1 ) for the computation of the isosteric enthalpy of
adsorption (right)

the isosteric adsorption enthalpy values are around 20 kJ mol−1 for PEA coverage
from 0.35 to 0.50 meq g−1 .

17.4 Concluding Remarks

The adsorption properties of solids with their various surface functionalities are
governed in liquids by both the hydrophobic or hydrophilic properties of the surfaces
and the properties of liquids (polarity, proticity, solvating ability). Experiments based
on pulse liquid or recirculation chromatographic methods, here illustrated, allow
17 Liquid-Solid Adsorption Properties: Measurement of the Effective Surface 551

easily determining the adsorption properties of solids in various liquids which are
similar to or the same as those in which the solids operate. The knowledge of the
effective surface properties is useful to predict the solid activity in practical conditions
of work.

References

1. T. Okuhara, Chem. Rev. 102, 3641 (2002)

2. A. Corma, H. Garcia, Chem. Rev. 103, 4307 (2003)
3. M. Hart, G. Fuller, D.R. Brown, J.A. Dale, S. Plant, J. Mol. Catal. A 182–183, 439 (2002)
4. S. Koujout, B.M. Kierman, D.R. Brown, H.G.M. Edwards, J.A. Dale, S. Plant, Catal. Lett. 85,
33 (2003)
5. S. Koujout, D.R. Brown, Catal. Lett. 98, 195 (2004)
6. P. Carniti, A. Gervasini, S. Biella, A. Auroux, Chem. Mater. 17, 6128 (2005)
7. C. Sun, J.C. Berg, Adv. Colloid Interface Sci. 105, 151 (2003)
8. P. Carniti, A. Gervasini, S. Biella, Ads. Sci. Technol. 23, 739 (2005)
9. P. Carniti, A. Gervasini, M. Marzo, Catal. Today 152, 42 (2010)
10. J.A. Dunne, R. Mariwala, M. Rao, S. Sincar, R.J. Gorte, A.L. Myers, Langmuir 12, 5888 (1996)
11. B. Dragoi, A. Gervasini, E. Dumitriu, A. Auroux, Thermochim. Acta 420, 127 (2004)
12. T. Paryjczak, Gas Chromatography in Adsorption and Catalysis. (Ellis Horwood, Chichester,
1986), p 175
13. A. Gervasini, C. Flego, Appl. Catal. 72, 153 (1991)
14. H.A. Benesi, B.H.C. Winquist, Adv. Catal. 27, 97 (1978)
15. P. Carniti, A. Gervasini, M. Marzo, J. Phys. Chem. C 112, 14064 (2008)
16. I.E. Wachs, J.M. Jehng, G. Deo, H. Hu, N. Arora, Catal. Today 28, 19 (1996)
17. H. Yoshida, T. Tanaka, T. Yoshida, T. Funabiki, S. Yoshida, Catal. Today 28, 79 (1996)
18. T. Tanaka, H. Nojima, H. Yoshida, H. Nakagawa, T. Funabiki, S. Yoshida, Catal. Today 16,
297 (1993)
19. P. Carniti, A. Gervasini, S. Biella, A. Auroux, Catal. Today 118, 373 (2006)
Index

A Adsorption, 3, 7, 103, 119, 132, 273, 277, 281,

Ab initio calculations, 506 282, 284, 289, 291, 294, 305, 311, 317,
Ab initio computational results, 4 356, 385, 521, 549
Ab initio modeling study, 513 Adsorption at low temperatures, 152
Abatement, 387 Adsorption Calorimetry, 162, 327
Absorption, 274, 275 Adsorption capacity, 20
Absorption in alkaline solution, 483 Adsorption energy distribution, 537
Absorption reaction, 425 Adsorption enthalpies, 282, 303, 316
Acceptor, 536 Adsorption enthalpy change was, 34, 36
Acetonitrile, 374 Adsorption heats, 118, 292, 294–296, 362, 443
Acid strength, 119 Adsorption internal energy, 292, 294
Acid/base character, 145 Adsorption isotherms, 107, 136, 526
Acid/base processes, 167 Adsorption microcalorimetry, 4, 104,
Acid/base properties, 117, 133, 146, 370 121–123, 356, 357, 374, 441, 449, 450
Acid-base titrations, 547 Adsorption of carbon monoxide, 506
Acidic oxides, 323 Adsorption of pollutants, 395
Acidic probe (SO2), 375 Adsorption potential distribution, 537
Acidic properties, 123, 373 Adsorption site, 283, 284, 306, 314
Acidic sites, 399 Adsorption temperature, 148
Acidic strength, 18, 19 Adsorption-desorption, 150, 152
Acidity, 319, 320, 356 Adsorption/desorption cycling, 426
Acidity prediction, 343 Adsorption-desorption-adsorption cycles, 16
Activated carbons, 388 Adsorptive liquid phase, 39
Activation energy, 184 Agrochemicals, 391
Activation energy for desorption, 142, 155 Al atoms, 360
Active sites, 104, 119, 137, 145, 150, 171 Alcanolamine, 486
Additives, 459, 472 Alkali-exchanged, 370
Adiabatic mode, 71 Alkaloids, 386
Ads. I and Ads. II, 20 Alkanes, 374
Adsorbate, 273, 277, 279, 280, 282–284, 286, Alkylphenols, 386
292, 294, 295, 297, 302, 307–309, 311, All-silica nanopores, 40
314 All-silica zeolite, 25
Adsorbate-adsorbent interactions, 531 Alumina, 324
Adsorbed amounts, 19, 275, 276, 278, 280, Amberlite, 546
282–284, 286–291, 294–298, 302–304, Amino-carbonyl complex, 509
311, 315 Ammonia, 116, 123, 148, 167, 360, 362, 366,
Adsorbed hydrogen, 448 369, 370, 373, 376
Adsorbed impurities, 434 Ammonia, pyrrole or SO2, 370
Adsorbents, 277, 281–283, 288, 289, 291, 292, Amorphous alumino-silicate, 514
294, 305, 307, 311, 317, 388 Amount, 322

A. Auroux (ed.), Calorimetry and Thermal Methods in Catalysis, 553

Springer Series in Materials Science 154, DOI: 10.1007/978-3-642-11954-5,
Ó Springer-Verlag Berlin Heidelberg 2013
554 Index

Amphoteric properties, 324 Calorimetric Principles, 72

Anatase specimens, 35 Calorimetric signal, 108
Anisotropic morphology, 331 Calorimetry, 52, 70, 273, 274, 288, 292,
Anode, 433 294–296, 302, 306, 311, 316, 317, 389,
Antioxidants (AO), 470, 471, 473, 476 411
amine, 471 Calorimetry-Gas Chromatography/Mass Spec-
phenolic, 471 trometry, 116
Apolar and aprotic solvent, 548 Calorimetry-Gravimetry, 112
Applications, 53, 87 Calorimetry-volumetry, 104
Aspecific interaction, 26, 41 Calvet calorimeter, 73, 74
Associative chemisorption, 36, 41, 42, 44 Calvet principle, 72
Atmosphere, 85 Calvet type DSC, 63, 65
Atmospheric pollution, 385 Ca-modified silica, 17
Average oxidation state, 193 Capacitance manometer, 106
Capillary interface, 94
Carbon dioxide, 125, 126
B Carbon monoxide adsorption, 438
p-back-donation, 21 Carbonaceous materials, 440
ofd-electrons, 508 Carbonyl bond electrostatic component, 509
contribution, 509 Catalyst, 57, 87, 338, 410
Basic or acidic, 113 Catalytic activity, 543
Basic oxides, 323 Catharometer, 140
Basic probe (NH3), 375, 544 Cation-deficient spinel transition aluminas,
Basic sites, 370, 399 510
Basicity, 319, 320 Cation-exchanged zeolites, 17
Bimetallic catalysts, 169 Cetane number, 457
Binary mixed oxides, 340 Change in internal energy, 30
Binding energy, 516 Characterization, 57
Bioactivity in sol-gel glasses, 42 Charge-balancing cations, 354
Biodiesel, 455, 456, 458, 461 Chemical (covalent) forces, 38
advantages of, 457 Chemical contamination, 386
Biodiesel blends, 457 Chemical-looping combustion (CLC), 89
Bio-sensors, 394 Chemisorption, 7, 104, 134, 157
Bisphenol-A, 386 Chemisorption process, 26
Blank curve, 87 Chromatographic experiments, 522
Bond energies, 26 Chromatographic method, 547
Brönsted acid sites, 16, 40, 110, 123, 167, 334, Chromatography, 273, 288, 305, 316, 317
354, 358, 363, 365, 366, 373, 378 Clausius-Clapeyron equation, 32, 135
Brønsted acidic strength, 40, 514 Clean surface, 30
Brönsted acidity, 121, 167 Cloud point, 458
Brønsted and Lewis sites at solid surfaces, 21 CO, 396
Brönsted site, 114, 115, 360 CO2, 386
Brunauer, 13 CO2 capture, 481, 483, 485, 487, 489, 491,
Brunauer isotherm, 526 493, 495, 497, 499, 501, 503
Buoyancy effect, 87 CO2 loading charges, 494
CO adsorption, 443, 444, 446, 448
CO adsorption microcalorimetry, 437
C CO adsorption on the Pt/C surface, 445
C80 microcalorimeter, 74 CO adsorption/desorption/readsorption cycles,
Calcination, 365, 366 449
Calibration, 6, 73 CO heats of adsorption, 438
Calibration vessel, 72 CO irreversibly adsorbed, 446
Calorimeter, 70, 106 Cold flow
Calorimetric isotherms, 9, 108 behavior, 461, 476
Index 555

improver, 469 Decomposition rate, 417

parameters, 465 Deconvolution, 150, 159
properties, 463 Defective MFI, 27
Co nanoparticles, 419 Defective Silicalite, 512
CO poisoning, 436, 441, 442, 450 Deformation/reconstruction of the surface, 515
Coadsorption, 273, 274, 279, 280, 283, 288, Degree of reduction, 189
296, 299- 301, 304, 309, 311, 314, 316, Dehydroxylated silica, 39
317 Dehydroxylation, 365–367
Coadsorption enthalpy, 288, 296, 302, 304, 310 Demixing amines, 486
Coadsorption isotherms, 288 De-pollution, 388
Cold Filter Plugging Point, 458 Desorption kinetics, 133
Column diameter, 523 Desorption, 103, 132, 356, 385
Column length, 523 Desulphurisation, 311
Combination of the Langmuir and Henry iso- Diesel fuel, 456, 460, 461, 465
therms, 14 Different geometrical arrangement, 512
Combining adsorption microcalorimetry and Differential enthalpy, 296, 297
molecular modeling, 40 Differential heats, 104, 109, 110, 363, 366, 395
Complex oxides, 330 Differential heat of adsorption, 22, 164, 168,
Computed energy of interaction, 23 170
Conductometric titration, 140 Differential reaction calorimeter, 411
Confinement effect, 27 Differential scanning calorimetry (DSC), 58,
Contact angle measurements, 539 389, 402, 403, 423, 458, 460, 461
Continuous heating mode, 67 Differential thermal analysis (DTA), 402
Contracting sphere model, 188 Diffusion, 103, 114, 121, 155, 159, 357
Coordination, 44 Diffusion of water, 422
r-coordination, 21 Diffusional, 527
Coordinative chemisorption, 38 Discontinuous (stepwise) introduction of the
Copper and silver metal carbonyls, 507 adsorptive, 38
CO-Pt interaction, 440, 441 Dispersion forces, 22, 27
Correlation with catalysis, 376 Dispersive and specific interactions, 533
Corrosive gas, 85 Dispersive component of the surface free
Counder and Young method, 530 energy, 533
Coupled volumetric-calorimetric technique, Dissociative chemisorption, 42, 44
340 Distribution parameter, 537
coupling, 103 Donor, 536
Covalent oxides, 336 Doped, 331
Cracking, 378 Dorris and Gray, 534
Crucible, 56, 64, 84 DSC curve, 60
Cryogenic process, 483 DSC techniques, 62
Crystal DTA applications, 57
morphology, 465 DTA detector, 56
size, 466 DTG, 82
Crystallization Dynamic equilibrium, 11
enthalpy, 463
points, 462, 469
Crystalline zeolite, 514 E
Crystallizable fraction, 462 ECP method, 528
Crystallographic pure anatase, 17 Effective acidity, 346, 543, 544
Cycles, 449 Effective adsorption properties, 543
Electron-accepting properties, 506
Electronegativity, 332
D Electron-withdrawing sulphate moieties, 37
Dealuminated zeolites, 357, 366 Electrostatic field, 12
Dealumination, 163, 361, 363, 365–368 Electrostatic polarization, 21
556 Index

Elution of a characteristic point, 528 Flexibility of the Si-O-Al amorphous structure,

Emitted gas, 95 515
Emmet and Teller (BET) model, 13 Flexible structure of the amorphous alumino-
Endocrine-disrupting compounds (EDCs), 386 silicate, 515
Endothermic effect, 55 Flow mixing calorimetry, 79
Endothermic step, 513 Flow mixing vessel, 80
Endothermic surface reconstruction, 511 Flow systems, 159
Energy, 274, 275, 282, 292, 302, 317 Fluxmeter, 63
Energy cost, 516 Fossil fuels, 430
Energy distribution, 109, 329 Fractional coverage h, 8
Enthalpic term, 35 Framework aluminium, 361
Enthalpy, 119, 282–285, 287, 288, 292, Free energy of adsorption, 531
295–297, 302, 304–310, 313, 394 Freundlich isotherm, 13
Enthalpy and entropy of adsorption, 531 Fuel cells, 430
Enthalpy change, 32 Fully optimized at ab initio level, 515
Enthalpy of adsorption, 10, 505, 506 Functionality of adsorbents, 522
Enthalpy of dissolution, 417
Enthalpy of neutralization, 119
Entropy, 112, 135, 280, 284, 285, 287, 302, G
314, 382 Gas adsorption calorimetry, 76
Entropy change, 33 Gas calorimetry, 538
Environment, 385, 399 Gas coupling techniques, 92
Environment pollution, 387 Gas injection loop, 65
Environment remediation, 388 Gas interaction, 91
Equilibrium, 111 Gas-solid interactions, 15, 105
Equilibrium concentration, 118 Gas-solid open system, 31
Equilibrium constant, 11, 119 Gaussian peak, 525
Equilibrium pressure, 22, 107 Gibbs adsorption isotherm, 297, 298
Ergonic term, 35 Gibbs energy, 282, 297, 299, 300, 302
Evolution of pressure, 108 Gibbs-Helmholtz equation, 531
Evolution of the heat of adsorption with
increasing coverage, 45
Evolution with increasing coverage, 505 H
Evolved gas analysis (EGA), 92 H2adsorption microcalorimetry, 442
Exchange level, 370 H2and CO uptake, 447
Exothermic effect, 29, 55, 112 Hammett acidity function (Ho), 321
Exothermic r-coordination, 511 Hammett indicators, 546
Experimental Setups, 138 Hard and Soft Acids and Bases (HSAB), 322
Extended Fowkes equation, 535 H-bonding adducts, 20
Extent of reduction (or oxidation), 182 H-bonding interactions, 38
Extraction of the Al-containing site, 516 Heat, 53, 71
Heat capacity, 59, 67, 68
Heat flow, 116, 403
F Heat flux, 58, 59, 73
Fatty acid Heat flux DSC, 62
alkyl esters, 456 Heat of adsorption, 135, 162, 361, 370, 441,
methyl esters, 458 447, 539
profile, 456 Heat of liquefaction of methanol, 43
saturated, 458 Heat of reaction, 410
unsaturated, 458, 462 Heated capillary, 95
Faujasite, 274 Heat-flow microcalorimeters, 105
Flame ionization detector, 140 Heating rate b, 132, 158
Flash desorption, 132 Heats of adsorption, 117, 124
Index 557

Heats of reaction, 117 Intrinsic acidity, 346

Heavy metal particles, 388 Inverse Gas Chromatography, 521
Heavy metals, 391 Ion exchange, 354, 368
Hematite reduction, 191 Ion-doping, 192
Henry conditions, 529 Ionic oxides, 335
Henry’s law, 14 Ionicity, 332
Henry-like isotherm, 36 IR spectroscopy, 506
Heptane, 311, 314, 316 Iron oxide, 190
Heterogeneity of the surface, 505 Irreversible adsorption, 25, 361
Heterogeneous catalysis, 176, 319 Irreversible modification of the surface, 514
Heterogeneous solid surface, 5 Irreversible phenomena, 20
Heterogeneous surface, 24, 327 Isomorphous substitution, 354, 370
Heteropolyacids, 343 Isoperibolic principle, 71
Heteropolyanions, 419 Isoster, 278, 286, 287, 295, 311, 314
Heteropolycompounds, 343 Isostere of adsorption, 32
High pressure crucible, 66 Isosteric heat, 32, 539
High pressure volumetric lines, 425 Isosteric heats of adsorption, 390, 549
High resolution transmission elec-tron Isosteric measurements, 135
microscopy (HR-TEM), 5 Isothermal mode, 71
Histogram, 23 Isothermal titration microcalorimetry (ITlC),
Hydrogen, 90, 409, 433, 434 80, 117, 394
Hydrogen as fuel, 430 Isotherms, 8, 20, 30, 33, 278, 280, 284, 285,
Hydrogen bond, 22 287, 288, 294, 297, 298, 302, 306, 309,
Hydrogen diffusion, 189 311
Hydrogen oxidation, 434 Isp (specific interaction), 535
Hydrogen release, 426 IUPAC, 29
Hydrogen storage capacities, 416
Hydrogen/oxygen consumed, 182
Hydrolysis, 81 J
Hydrophilic and/or hydro-phobic features, 18 James-Martin factor, 529
Hydrophilicity / hydrophobicity, 359 Joule effect, 68
Hydrophobic-hydrophilic balance of solid
surfaces, 202, 210, 219, 221, 250, 264
Hydrotalcites, 342 K
Hydroxyl groups, 372 Kinetic parameters of reaction, 187
Hydroxyl nests, 16 Kinetic parameters, 146, 155
Hydroxylated silica, 39 Kinetical broadening, 527
5-hydroxymethyl-2-furaldehyde, 349 Kinetics, 61, 136
Kissinger method, 158

I
Induced heterogeneity, 30, 37 L
Inert gas, 85 Langmuir adsorption isotherm, 136
Infinite dilution, 529 Langmuir model equation, 9
Infrared camera, 420 Latent heat of liquefaction of water, 24
Integral heats, 109 Latent heat of liquefaction, 27
Integral molar heat of adsorption, 22 Lattice oxygen, 178
Interactions between solid surfaces and Length, volume, 53
molecular species, 206, 210, 263 Lennard-Jones formula, 41
International Confederation for Thermal Lewis, 110, 121, 123
Analysis and Calorimetry (ICTAC), 53, Lewis acidic sites, 16, 35, 40, 114, 115, 167,
81, 92 320, 334, 354, 358, 510
Internal surface, 5 Lewis basic sites, 358
Intraparticular diffusion, 524 Lewis basicity, 370
558 Index

Lewis sites, 362, 365, 366, 378 Molecular sieves, 360

Linear correlation between the spectroscopic Molecule/surface site bonding energy, 505
and thermodynamic quantities, 507 Monolayer coverage, 11
Linear Henry isotherm, 526 Monolayer, 10
Liquid adsorption, 80 MS curves, 95
Liquid chromatographic technique, 545 Multi-techniques approach, 505
Liquid flow calorimetry, 236, 237, 240, 247,
248, 250, 252, 253, 264
Liquid microcalorimetry, 397 N
Liquid phase calorimetry, 411 NaBH4, 410
Liquid reaction calorimetry, 415 NaBO2hydration ratio, 420
Liquid-like phase, 8 Nanocalorimetry, 389
Liquid-solid heterogeneous catalysis, 543 Nanopore walls, 41
Loewenstein rule, 353 Natural clays, 388
Low concentrations of adsorbate, 525 Nature of acid sites, 114, 322
NaX, 311, 314, 315, 317
Negative charge, 360
M Negative free energy, 34
Manometry, 273, 288, 289, 292, 306, 316, 317 Nicotine, 398
Mars & Van Krevelen, 178 Niobia, 338
Mass, 53, 81, 83 NO2, 396
Mass-transfer, 122 Nomenclature, 29
Maximum density of Lewis and Brønsted Non framework Al, 361
acidic sites, 514 Non-dissociative chemical adsorption, 506
Mechanical work, 277, 278, 293 NOx, 386
Mechanism, 413 Nucleation model, 188
Mechanism of the reaction, 539 Nuclei, 188
Mercaptan, 311, 314, 316
Mesoporous, 388
Mesoporous silica, 331 O
Metaborate product, 421 One-dimension potential energy diagram, 41
Metal oxides, 167, 175, 388 Oxidation, 469
Metal oxide surfaces, 319 Oxidation induction time (OIT), 472, 473, 476
Metallosilicate, 372 Oxidation onset temperature (OOT), 473, 476
Metals, 169 oxidative stability, 469
Methanol pressure, 42 Oxide supports, 337
Methoxylation reaction, 43, 44 Oxide surface, 326
MgH2, 410 Oxides, 331
Microcalorimeter, 74, 104
Microcalorimetry, 362, 385, 438, 440, 441
Microcalorimetry of adsorption, 319 P
Microcalorimetry of CO adsorption, 429, 437 Paraffins, 458
Micropore structure, 538 Parameter of specific interaction, 535
Microporous, 353 Partial adsorption enthalpy and entropy, 303
Mixing and reaction calorimetry, 76 Partial molar heats, 23
Mixing vessel, 74, 79 Partial molar value, 304
Model site, 23 P-DSC, 471, 473, 476
Modification of the surface structure, 511 Peak maximum (Tmax), 155
Modified oxides, 332 PEM fuel cells, 429, 433, 437, 450
Molar entropy, 280, 283, 302, 303, 314 PEMFC stack, 436
Molar quantities, 505 Percolation calorimetry, 78
Molecular interpretation, 505 Perovskite, 342
Molecular polarizability, 536 Personal care products (PPCPs), 386
Molecular probe, 18, 327 Pesticides, 386, 391
Index 559

Pharmaceuticals, 386 Quaternary metal oxides, 340

Phenol, 386
Phenolic compounds, 386
Physical dissolution, 485 R
Physisorption, 7, 110, 134 Rate, 137
Pillared clays, 388 Rate of adsorption, 10, 137
Plain electrostatic polarization/(r-coordina- Rate of desorption (rdes), 11, 138, 141, 155
tion), 510 Rate of reaction, 61
Plate DSC, 63 Reaction calorimeter, 347
Plain electrostatic polarization/-coordinatio, Reaction mechanisms, 376
510 Readsorption, 155, 159, 161
Platinum, 448 Reciprocal linear form, 11
Platinum-based catalyst, 429 Redhead method, 158
Polanyi-Wigner equation, 137 Redox properties, 126, 175, 401
Polar and protic liquids, 549 Reducing properties, 176
Polar defects, 512 Reduction, 115, 116
Polar sites, 27 Reduction mechanisms, 186
Polar solutes (Isp), 535 Reference materials, 60
Pollutants, 385, 387, 389, 391, 393, 395, 397, Reinforcement of the carbonyl bond, 508
399, 401, 403, 405, 407 Relative humidity, 86
Pollution, 385 Relative humidity generator, 87
Polyaromatic hydrocarbons (PAHs), 388, 394 Relative pressure, 277, 311
Population of the surface sites, 44 Reoxidation, 115
Pore size distribution, 537 Repeatability, 472
Pores, 5 Resolution parameter, 186
Porous materials, 133, 388 Retention time, 530
Pour point, 458 Retention volume, 529
Power compensation DSC, 62 Reverse-flow, 540
Precursor-mediated kinetics, 43 Reversibility, 20
Pressure calorimetry, 74, 75 Reversible adsorption, 34
Probe molecule, 106, 123, 147 Reversible H2Storage, 422
Processes at constant pressure, 32 Rigidity of the crystalline zeolite framework,
Protected DTA rod, 90 515
Proton, 360
Proton affinities, 124, 327
Protonation, 376 S
Pt based materials, 437 Seiyama, 345
Pt electrocatalysts, 436 Selective, 273, 274, 279, 281, 282, 305, 310,
Pt surface sites, 438 311, 317
Pt/C, 441, 444, 449 Selective adsorption, 375
Pt/C catalyst, 433, 442, 443, 445, 447, 450 Selective oxidation, 176
Pt-based catalyst, 434, 445 Selectivity, 280, 281, 288, 292, 309, 311, 314,
Pulse, 522 316, 317
Pulse liquid or recirculation chromatographic Sensitive volumetric systems, 15
methods, 550 Sensitivity parameter, 186
Purification, 274, 317 Separation, 273, 274, 280, 282, 305
Pyridine, 122, 123, 148, 167, 362, 366, 373 Shape selectivity, 354, 357
Pyrrole, 370 Si/Al ratio, 163, 359, 360, 363
Si4+or Al3+, 354
Si–O–Al amorphous structure, 515
Q Silanol groups, 353
Quadrupole mass spectrometer (QMS), 140 Silica-alumina, 333
Quantitative measurements, 142 Silica-based materials, 388
Quartz reactor, 65 Silicalite, 27
560 Index

Silica-niobia oxides, 548 Surface coverage (h), 155

Silica-zirconia, 546 Surface dehydration, 16
Siloxane surface, 27 Surface excess amount, 275, 276
Simultaneous TGA and DTA, 90 Surface heterogeneity, 537
Single crystals, 153 Surface lattice oxygen, 176
Site, 283, 308, 309, 314, 510 Surface Lewis complexes, 34
Skewness ratio, 530 Surface properties, 376
Skimmer system, 96 Surface science studies, 133
Slow kinetic process, 525 Surface sites heterogeneity, 45
Slow-and-constant adsorptive introduction, 38 Symmetrical balance, 91
Small highly structured rings of mutually Symmetrical dual furnace thermobalance, 88
interacting silanol group, 513
Sodium borohydride hydrolysis, 411
Soft Lewis base, 12 T
Soil pollutants, 387 Tanabe’smodel, 344
Solid NaOH, 420 Temkin equation, 14
Solid surface, 4 Temperature programmed oxidation (TPO), 96
Solid-fluid interface, 3 Temperature programmed reduction, 115
Sorption calorimetry, 77 Temperature step mode, 69
Sorption, 521 Temperature-programmed desorption (TPD),
SOx, 386 67, 96, 113, 114, 131, 132, 145, 146,
Speciation, 497 154, 162, 171, 389
Specific interactions, 534 Temperature-programmed experiment, 141,
Specific retention volume, 529 144, 157
Spectroscopic and energetic features, 35, 37 Temperatureprogrammed reaction spectros-
Spectroscopic methods, 4 copy (TPRS), 132
Spontaneous process, 34 Temperature-programmed reactions, 144
Spreading pressure, 277, 278, 297, 298, 301, Temperatureprogrammed reduction (TPR),
302 144
Stability Oxidative, 469, 476 Temperature-programmed techniques, 389,
Stable carbonyl-like species, 26 400
Stable complexes, 26 TGA techniques, 83
Standard entropy of adsorption, 13 TG-DSC, 90, 91
Standard free energy, 12 TG-DSC-MS, 113
Standard reference material, 68 TG-DTA, 90
Static techniques, 538 TG-FTIR coupling, 97
Steroidal hormones, 386 TG-MS, 93
Storage, 409 Thermal analysis, 52
Stretching frequency, 507 Thermal desorption spectroscopy (TDS), 133
Strong-metal-support-interaction, 338 Thermal equilibrium, 106, 112
Structural and/or a chemical heterogeneity of Thermal phenomena, 426
the sites, 30 Thermal resistance, 60
Structural defects, 5 Thermal risks, 411
Structure, 522 Thermal techniques, 327
Structurally equivalent and non-interacting Thermal transition, 53
sites, 36 Thermoanalytical methods, 385, 402
Subsequent runs of adsorption, 15 Thermobalances, 83
Substituted pyridines (DMP), 375 Thermocouples, 53, 56, 105
Sulphur dioxide, 125, 370, 396 Thermodynamic compensation effect, 532
Supercage, 305, 306, 308, 309, 311, 313, 317 Thermodynamic definition, 29
Supersonic system, 96 Thermodynamic models, 487, 497
Supported metal oxide, 194, 334 Thermodynamic parameter, 146, 154, 538
Supported oxides, mixed oxides, and complex Thermodynamics, 413
oxides, 330 Thermogram, 54, 105
Index 561

Thermogravimetric analysis (TGA), 81, 187 V

Thermogravimetry (TG), 113, 389, 402 Van deemter equation, 524
Thermokinetic parameter, 111, 415, 443, 445 Van der Waals/London-dispersion forces, 38
Thermokinetics, 111 Van’t hoff, 284, 285, 287, 306
Thermomicroscopy, 461 Vanadia, 194
Thermopiles, 63, 74 Variable-temperature FTIR spectroscopy, 33
Thermoprogrammed desorption, 425 Volatile organic compounds (VOCs), 386
Three-dimensional network of pores, 16 Volcano plot, 539
Tian-Calvet, 105, 117 Volumetric gas line, 76
Tian-Calvet heat-flow microcalorimeters, 14 Volumetric isotherms, 9, 108
Titania, 194 Volumetric titrations, 544
Titration batch calorimetry, 264
Tmax, 142, 150, 153
Toluene, 316 W
TPD (Temperature Programmed Desorption), Wastewaters, 388
67, 96, 113, 114, 131–133, 135, 137, Water diffusion, 420
139, 141, 143, 145, 147, 149, 151, 153, Weak associative chemisorption, 42
155, 157, 159, 161–163, 165, 167, 169, Weak physical adsorption, 36
171, 173 Wetting and immersional calorimetry,
TPD Profile, 142 212–214, 223, 250, 264
TPO (Temperature Programmed Oxidation), Work exchanged, 31
96 Work of adhesion, 533
TPR/TPO apparatus, 180
TPR/TPO experiments, 181
TPR/TPO profile, 179, 184 X
Transition catalytic alumina, 507 X-Y diagram, 280
Transition metal oxides, 331 Xylenes, 274, 305, 306, 309, 311, 317
Transportation, 409
Turnover rates, 339
Two independent sites, 29 Z
Two local isotherms, 29 Zeolite, 6, 80, 81, 163, 274, 299, 305, 306,
308–311, 334, 353, 354, 356, 388
Zeolite nanocavities, 24
U Zeolite topology, 357
Ultrahigh vacuum, 139, 156 Zero-coverage differential heat of adsorption,
Unsaturated fatty compounds, 458 23