Computational Models For Polydisperse

Particulate And Multiphase Systems Daniele L

Marchisio download

https://ebookbell.com/product/computational-models-for-
polydisperse-particulate-and-multiphase-systems-daniele-l-
marchisio-4327382

Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.

Computational Models For Neuroscience Human Cortical Information

Processing 2003th Edition Robert Hechtnielsen

https://ebookbell.com/product/computational-models-for-neuroscience-
human-cortical-information-processing-2003th-edition-robert-
hechtnielsen-55934754

Computational Models For The Human Body Special Volume Xii Handbook Of
Numerical Analysis 1st Edition Nicholas Ayache

https://ebookbell.com/product/computational-models-for-the-human-body-
special-volume-xii-handbook-of-numerical-analysis-1st-edition-
nicholas-ayache-2436464

Computational Models For Turbulent Reacting Flows Fox Ro

https://ebookbell.com/product/computational-models-for-turbulent-
reacting-flows-fox-ro-2600852

Computational Models For Co2 Geosequestration Compressed Air Energy

Storage Rafid Alkhoury Jochen Bundschuh

https://ebookbell.com/product/computational-models-for-
co2-geosequestration-compressed-air-energy-storage-rafid-alkhoury-
jochen-bundschuh-4732404
Computational Models For Cognitive Vision 1st Edition Hiranmay Ghosh

https://ebookbell.com/product/computational-models-for-cognitive-
vision-1st-edition-hiranmay-ghosh-11221104

Computational Models For Cognitive Vision Hiranmay Ghosh

https://ebookbell.com/product/computational-models-for-cognitive-
vision-hiranmay-ghosh-170932258

Computational Models For Turbulent Reacting Flows 1st Edition Rodney O

Fox

https://ebookbell.com/product/computational-models-for-turbulent-
reacting-flows-1st-edition-rodney-o-fox-1125236

Mathematical And Computational Models For Congestion Charging 1st

Edition Siriphong Lawphongpanich

https://ebookbell.com/product/mathematical-and-computational-models-
for-congestion-charging-1st-edition-siriphong-lawphongpanich-4795036

Transactions On Computational Systems Biology Xiv Special Issue On

Computational Models For Cell Processes 1st Edition Oana Andrei

https://ebookbell.com/product/transactions-on-computational-systems-
biology-xiv-special-issue-on-computational-models-for-cell-
processes-1st-edition-oana-andrei-4202966
more information – www.cambridge.org/9780521858489
Computational Models for Polydisperse Particulate and
Multiphase Systems

Providing a clear description of the theory of polydisperse multiphase flows,

with emphasis on the mesoscale modeling approach and its relationship
with microscale and macroscale models, this all-inclusive introduction is
ideal, whether you are working in industry or academia. Theory is linked to
practice through discussions of key real-world cases (particle/droplet/bubble
coalescence, breakup, nucleation, advection and diffusion, and physical- and
phase-space), providing valuable experience in simulating systems that can
be applied to your own applications. Practical cases of QMOM, DQMOM,
CQMOM, EQMOM, and ECQMOM are also discussed and compared, as
are realizable finite-volume methods. This provides the tools you need to use
quadrature-based moment methods, choose from the many available options,
and design high-order numerical methods that guarantee realizable moment
sets. In addition to the numerous practical examples, Matlab scripts for sev-
eral algorithms are also provided, so you can apply the methods described to
practical problems straight away.

Daniele L. Marchisio is an Associate Professor at the Politecnico di Torino,

Italy, where he received his Ph.D. in 2001. He has held visiting positions at the
Laboratoire des Sciences du Génie Chimique, CNRS–ENSIC (Nancy, France),
Iowa State University (USA), Eidgenössische Technische Hochschule Zürich
(Switzerland), and University College London (UK), and has been an invited
professor at Aalborg University (Denmark) and the University of Valladolid
(Spain). He acts as a referee for the key international journals in his field of
research. He has authored more than 60 scientific papers and 5 book chapters,
and co-edited the volume Multiphase Reacting Flows (Springer, 2007).
Rodney O. Fox is the Anson Marston Distinguished Professor of Engineering
at Iowa State University (USA), an Associate Scientist at the US-DOE Ames
Laboratory, and a Senior Research Fellow in the EM2C laboratory at the Ecole
Centrale Paris (France). His numerous professional awards include an NSF
Presidential Young Investigator Award in 1992 and Fellow of the American
Physical Society in 2007. The impact of Fox’s work touches every technologi-
cal area dealing with multiphase flow and chemical reactions. His monograph
Computational Models for Turbulent Reacting Flows (Cambridge University
Press, 2003) offers an authoritative treatment of the field.
Cambridge Series in Chemical Engineering

Series Editor
Arvind Varma, Purdue University

Editorial Board
Christopher Bowman, University of Colorado
Edward Cussler, University of Minnesota
Chaitan Khosla, Stanford University
Athanassios Z. Panagiotopoulos, Princeton University
Gregory Stephanopolous, Massachusetts Institute of Technology
Jackie Ying, Institute of Bioengineering and Nanotechnology, Singapore

Books in Series
Baldea and Daoutidis, Dynamics and Nonlinear Control of Integrated Process Systems
Chau, Process Control: A First Course with MATLAB
Cussler, Diffusion: Mass Transfer in Fluid Systems, Third Edition
Cussler and Moggridge, Chemical Product Design, Second Edition
Denn, Chemical Engineering: An Introduction
Denn, Polymer Melt Processing: Foundations in Fluid Mechanics and Heat Transfer
Duncan and Reimer, Chemical Engineering Design and Analysis: An Introduction
Fan and Zhu, Principles of Gas–Solid Flows
Fox, Computational Models for Turbulent Reacting Flows
Leal, Advanced Transport Phenomena: Fluid Mechanics and Convective Transport
Lim and Shin, Fed-Batch Cultures: Fundamentals, Modeling, Optimization, and Control of
Semi-Batch Bioreactors
Marchisio and Fox, Computational Models for Polydisperse Particulate and Multiphase Systems
Mewis and Wagner, Colloidal Suspension Rheology
Morbidelli, Gavriilidis, and Varma, Catalyst Design: Optimal Distribution of Catalyst in Pellets,
Reactors, and Membranes
Noble and Terry, Principles of Chemical Separations with Environmental Applications
Orbey and Sandler, Modeling Vapor–Liquid Equilibria: Cubic Equations of State and Their
Mixing Rules
Petyluk, Distillation Theory and its Applications to Optimal Design of Separation Units
Rao and Nott, An Introduction to Granular Flow
Russell, Robinson, and Wagner, Mass and Heat Transfer: Analysis of Mass Contactors and Heat
Exchangers
Schobert, Chemistry of Fossil Fuels and Biofuels
Slattery, Advanced Transport Phenomena
Varma, Morbidelli, and Wu, Parametric Sensitivity in Chemical Systems
Computational Models for
Polydisperse Particulate and
Multiphase Systems
DANIELE L. MARCHISIO
Politecnico di Torino

R O D N E Y O . F OX
Iowa State University
cambridge university press
Cambridge, New York, Melbourne, Madrid, Cape Town,
Singapore, São Paulo, Delhi, Mexico City
Cambridge University Press
The Edinburgh Building, Cambridge CB2 8RU, UK

Published in the United States of America by Cambridge University Press, New York

www.cambridge.org
Information on this title: www.cambridge.org/9780521858489


c Daniele L. Marchisio and Rodney O. Fox 2013

This publication is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.

First published 2013

Printed and bound in the United Kingdom by the MPG Books Group

A catalog record for this publication is available from the British Library

Library of Congress Cataloging in Publication Data

Marchisio, Daniele L.
Computational models for polydisperse particulate and multiphase systems / Daniele L.
Marchisio, Politecnico di Torino, Rodney O. Fox, Iowa State University.
pages cm. – (Cambridge series in chemical engineering)
ISBN 978-0-521-85848-9
1. Multiphase flow – Mathematical models. 2. Chemical reactions – Mathematical models.
3. Transport theory. 4. Dispersion – Mathematical models. I. Fox, Rodney O., 1959–
II. Title.
TA357.5.M84M37 2013
532 .56–dc23
2012044073

ISBN 978-0-521-85848-9 Hardback

Cambridge University Press has no responsibility for the persistence or

accuracy of URLs for external or third-party internet websites referred to
in this publication, and does not guarantee that any content on such
websites is, or will remain, accurate or appropriate.
a Giampaolo
à Roberte
Contents

Preface page xiii

Notation xvii

1 Introduction 1
1.1 Disperse multiphase flows 1
1.2 Two example systems 3
1.2.1 The population-balance equation for fine particles 3
1.2.2 The kinetic equation for gas–particle flow 8
1.3 The mesoscale modeling approach 14
1.3.1 Relation to microscale models 16
1.3.2 Number-density functions 18
1.3.3 The kinetic equation for the disperse phase 19
1.3.4 Closure at the mesoscale level 20
1.3.5 Relation to macroscale models 20
1.4 Closure methods for moment-transport equations 23
1.4.1 Hydrodynamic models 23
1.4.2 Moment methods 25
1.5 A road map to Chapters 2–8 27

2 Mesoscale description of polydisperse systems 30

2.1 Number-density functions (NDF) 30
2.1.1 Length-based NDF 32
2.1.2 Volume-based NDF 33
2.1.3 Mass-based NDF 33
2.1.4 Velocity-based NDF 34
2.2 The NDF transport equation 35
2.2.1 The population-balance equation (PBE) 35
2.2.2 The generalized population-balance equation (GPBE) 37
2.2.3 The closure problem 37
2.3 Moment-transport equations 38
2.3.1 Moment-transport equations for a PBE 38
2.3.2 Moment-transport equations for a GPBE 40
2.4 Flow regimes for the PBE 43
2.4.1 Laminar PBE 43
2.4.2 Turbulent PBE 44
2.5 The moment-closure problem 45

vii
viii Contents

3 Quadrature-based moment methods 47

3.1 Univariate distributions 47
3.1.1 Gaussian quadrature 49
3.1.2 The product–difference (PD) algorithm 51
3.1.3 The Wheeler algorithm 53
3.1.4 Consistency of a moment set 55
3.2 Multivariate distributions 62
3.2.1 Brute-force QMOM 63
3.2.2 Tensor-product QMOM 68
3.2.3 Conditional QMOM 74
3.3 The extended quadrature method of moments (EQMOM) 82
3.3.1 Relationship to orthogonal polynomials 83
3.3.2 Univariate EQMOM 84
3.3.3 Evaluation of integrals with the EQMOM 91
3.3.4 Multivariate EQMOM 93
3.4 The direct quadrature method of moments (DQMOM) 99

4 The generalized population-balance equation 102

4.1 Particle-based definition of the NDF 102
4.1.1 Definition of the NDF for granular systems 102
4.1.2 NDF estimation methods 105
4.1.3 Definition of the NDF for fluid–particle systems 107
4.2 From the multi-particle–fluid joint PDF to the GPBE 110
4.2.1 The transport equation for the multi-particle joint PDF 111
4.2.2 The transport equation for the single-particle joint PDF 112
4.2.3 The transport equation for the NDF 112
4.2.4 The closure problem 113
4.3 Moment-transport equations 114
4.3.1 A few words about phase-space integration 114
4.3.2 Disperse-phase number transport 116
4.3.3 Disperse-phase volume transport 116
4.3.4 Fluid-phase volume transport 117
4.3.5 Disperse-phase mass transport 118
4.3.6 Fluid-phase mass transport 121
4.3.7 Disperse-phase momentum transport 123
4.3.8 Fluid-phase momentum transport 124
4.3.9 Higher-order moment transport 127
4.4 Moment closures for the GPBE 130

5 Mesoscale models for physical and chemical processes 136

5.1 An overview of mesoscale modeling 136
5.1.1 Mesoscale models in the GPBE 137
5.1.2 Formulation of mesoscale models 141
5.1.3 Relation to macroscale models 145
5.2 Phase-space advection: mass and heat transfer 147
5.2.1 Mesoscale variables for particle size 149
5.2.2 Size change for crystalline and amorphous particles 152
5.2.3 Non-isothermal systems 155
5.2.4 Mass transfer to gas bubbles 156
Contents ix

5.2.5 Heat/mass transfer to liquid droplets 158

5.2.6 Momentum change due to mass transfer 160
5.3 Phase-space advection: momentum transfer 161
5.3.1 Buoyancy and drag forces 162
5.3.2 Virtual-mass and lift forces 171
5.3.3 Boussinesq–Basset, Brownian, and thermophoretic forces 173
5.3.4 Final expressions for the mesoscale acceleration models 175
5.4 Real-space advection 177
5.4.1 The pseudo-homogeneous or dusty-gas model 179
5.4.2 The equilibrium or algebraic Eulerian model 180
5.4.3 The Eulerian two-fluid model 181
5.4.4 Guidelines for real-space advection 182
5.5 Diffusion processes 183
5.5.1 Phase-space diffusion 184
5.5.2 Physical-space diffusion 187
5.5.3 Mixed phase- and physical-space diffusion 188
5.6 Zeroth-order point processes 189
5.6.1 Formation of the disperse phase 189
5.6.2 Nucleation of crystals from solution 191
5.6.3 Nucleation of vapor bubbles in a boiling liquid 191
5.7 First-order point processes 192
5.7.1 Particle filtration and deposition 193
5.7.2 Particle breakage 195
5.8 Second-order point processes 202
5.8.1 Derivation of the source term 203
5.8.2 Source terms for aggregation and coalescence 205
5.8.3 Aggregation kernels for fine particles 206
5.8.4 Coalescence kernels for droplets and bubbles 212

6 Hard-sphere collision models 214

6.1 Monodisperse hard-sphere collisions 215
6.1.1 The Boltzmann collision model 217
6.1.2 The collision term for arbitrary moments 218
6.1.3 Collision angles and the transformation matrix 221
6.1.4 Integrals over collision angles 223
6.1.5 The collision term for integer moments 230
6.2 Polydisperse hard-sphere collisions 236
6.2.1 Collision terms for arbitrary moments 237
6.2.2 The third integral over collision angles 242
6.2.3 Collision terms for integer moments 243
6.3 Kinetic models 246
6.3.1 Monodisperse particles 246
6.3.2 Polydisperse particles 248
6.4 Moment-transport equations 250
6.4.1 Monodisperse particles 251
6.4.2 Polydisperse particles 255
6.5 Application of quadrature to collision terms 261
6.5.1 Flux terms 261
6.5.2 Source terms 263
x Contents

7 Solution methods for homogeneous systems 266

7.1 Overview of methods 266
7.2 Class and sectional methods 269
7.2.1 Univariate PBE 269
7.2.2 Bivariate and multivariate PBE 279
7.2.3 Collisional KE 283
7.3 The method of moments 289
7.3.1 Univariate PBE 290
7.3.2 Bivariate and multivariate PBE 296
7.3.3 Collisional KE 297
7.4 Quadrature-based moment methods 300
7.4.1 Univariate PBE 301
7.4.2 Bivariate and multivariate PBE 307
7.4.3 Collisional KE 314
7.5 Monte Carlo methods 315
7.6 Example homogeneous PBE 319
7.6.1 A few words on the spatially homogeneous PBE 319
7.6.2 Comparison between the QMOM and the DQMOM 323
7.6.3 Comparison between the CQMOM and Monte Carlo 324

8 Moment methods for inhomogeneous systems 329

8.1 Overview of spatial modeling issues 329
8.1.1 Realizability 330
8.1.2 Particle trajectory crossing 332
8.1.3 Coupling between active and passive internal coordinates 335
8.1.4 The QMOM versus the DQMOM 337
8.2 Kinetics-based finite-volume methods 340
8.2.1 Application to PBE 341
8.2.2 Application to KE 345
8.2.3 Application to GPBE 347
8.3 Inhomogeneous PBE 349
8.3.1 Moment-transport equations 349
8.3.2 Standard finite-volume schemes for moments 350
8.3.3 Realizable finite-volume schemes for moments 353
8.3.4 Example results for an inhomogeneous PBE 358
8.4 Inhomogeneous KE 362
8.4.1 The moment-transport equation 363
8.4.2 Operator splitting for moment equations 363
8.4.3 A realizable finite-volume scheme for bivariate
velocity moments 364
8.4.4 Example results for an inhomogeneous KE 366
8.5 Inhomogeneous GPBE 373
8.5.1 Classes of GPBE 373
8.5.2 Spatial transport with known scalar-dependent velocity 376
8.5.3 Example results with known scalar-dependent velocity 377
8.5.4 Spatial transport with scalar-conditioned velocity 381
8.5.5 Example results with scalar-conditioned velocity 388
8.5.6 Spatial transport of the velocity-scalar NDF 396
8.6 Concluding remarks 401
Contents xi

Appendix A Moment-inversion algorithms 403

A.1 Univariate quadrature 403
A.1.1 The PD algorithm 403
A.1.2 The adaptive Wheeler algorithm 404
A.2 Moment-correction algorithms 405
A.2.1 The correction algorithm of McGraw 405
A.2.2 The correction algorithm of Wright 407
A.3 Multivariate quadrature 408
A.3.1 Brute-force QMOM 408
A.3.2 Tensor-product QMOM 410
A.3.3 The CQMOM 412
A.4 The EQMOM 413
A.4.1 Beta EQMOM 413
A.4.2 Gamma EQMOM 416
A.4.3 Gaussian EQMOM 418

Appendix B Kinetics-based finite-volume methods 421

B.1 Spatial dependence of GPBE 421
B.2 Realizable FVM 423
B.3 Advection 427
B.4 Free transport 429
B.5 Mixed advection 434
B.6 Diffusion 437

Appendix C Moment methods with hyperbolic equations 441

C.1 A model kinetic equation 441
C.2 Analytical solution for segregated initial conditions 442
C.2.1 Segregating solution 442
C.2.2 Mixing solution 443
C.3 Moments and the quadrature approximation 444
C.3.1 Moments of segregating solution 444
C.3.2 Moments of mixing solution 446
C.4 Application of QBMM 447
C.4.1 The moment-transport equation 447
C.4.2 Transport equations for weights and abscissas 448

Appendix D The direct quadrature method of moments fully conservative 450

D.1 Inhomogeneous PBE 450
D.2 Standard DQMOM 450
D.3 DQMOM-FC 453
D.4 Time integration 455

References 459

Index 488
Preface

This book is intended for graduate students in different branches of science and engineering
(i.e. chemical, mechanical, environmental, energetics, etc.) interested in the simulation of
polydisperse multiphase flows, as well as for scientists and engineers already working in
this field. The book provides, in fact, a systematic and consistent discussion of the basic
theory that governs polydisperse multiphase systems, which is suitable for a neophyte,
and presents a particular class of computational methods for their actual simulation, which
might interest the more experienced scholar.
As explained throughout the book, disperse multiphase systems are characterized by
multiple phases, with one phase continuous and the others dispersed (i.e. in the form of
distinct particles, droplets, or bubbles). The term polydisperse is used in this context to
specify that the relevant properties characterizing the elements of the disperse phases, such
as mass, momentum, or energy, change from element to element, generating what are com-
monly called distributions. Typical distributions, which are often used as characteristic
signatures of multiphase systems, are, for example, a crystal-size distribution (CSD), a
particle-size distribution (PSD), and a particle-velocity distribution.
The problem of describing the evolution (in space and time) of these distributions has
been treated in many ways by different scientific communities, focusing on aspects most
relevant to their community. For example, in the field of crystallization and precipitation,
the problem is described (often neglecting spatial inhomogeneities) in terms of crystal or
particle size, and the resulting governing equation is called a population-balance equation
(PBE). In the field of evaporating (and non-evaporating) sprays the problem is formu-
lated in terms of the particle surface area and the governing equation is referred to as the
Williams–Boltzmann equation. In this and other fields great emphasis has been placed on
the fact that the investigated systems are spatially inhomogeneous. Aerosols and ultra-fine
particles are often described in terms of particle mass, and the final governing equation is
called the particle-dynamics equation. Particulate systems involved in granular flows have
instead been investigated in terms of particle velocity only, and the governing equation is
the inelastic extension to multiphase systems of the well-known Boltzmann equation (BE)
used to describe molecular velocity distributions in gas dynamics.
Although these apparently different theoretical frameworks are referred to by different
names, the underlying theory (which has its foundation in classical statistical mechanics) is
exactly the same. This has also generated a plethora of numerical methods for the solution
of the governing equations, often sharing many common elements, but generally with a
specific focus on only part of the problem. For example, in a PBE the distribution repre-
senting the elements constituting the multiphase system is often discretized into classes or
sections, generating the so-called discretized population-balance equation (DPBE). Among
the many methods developed, one widely used among practitioners in computational fluid
dynamics (CFD) is the multiple-size-group (MUSIG) method. This approach resembles,

xiii
xiv Preface

in its basic ideas, the discretization carried out for the BE in the so-called discrete-velocity
method (DVM). Analogously, the method of moments (MOM) has been used for the solu-
tion of both PBE and BE, but the resulting closure problem is overcome by following
different strategies in the two cases. In the case of the BE the most popular moment clo-
sure is the one proposed by Grad, which is based on the solution of a subset of 13 or
26 moments, coupled with a presumed functional form for the velocity distribution. In
contrast, in the case of a PBE the closure strategy often involves interpolation among the
known moments (as in the method of moments with interpolative closure, MOMIC). Given
the plethora of approaches, for the novice it is often impossible to see the connections
between the methods employed by the different communities.
This book provides a consistent treatment of these issues that is based on a general
theoretical framework. This, in turn, stems from the generalized population-balance equa-
tion (GPBE), which includes as special cases all the other governing equations previously
mentioned (e.g. PBE and BE). After discussing how this equation originates, the different
computational models for its numerical solution are presented. The book is structured as
follows.

• Chapter 1 introduces key concepts, such as flow regimes and relevant dimensionless
numbers, by using two examples: the PBE for fine particles and the KE for gas–
particle flow. Subsequently the mesoscale modeling approach used throughout the
book is explained in detail, with particular focus on the relation to microscale and
macroscale models and the resulting closure problems.

• Chapter 2 provides a brief introduction to the mesoscale description of polydisperse

systems. In this chapter the many possible number-density functions (NDF), formu-
lated with different choices for the internal coordinates, are presented, followed by an
introduction to the PBE in their various forms. The chapter concludes with a short
discussion on the differences between the moment-transport equations associated
with the PBE, and those arising due to ensemble averaging in turbulence theory.

• Chapter 3 provides an introduction to Gaussian quadrature and the moment-inversion

algorithms used in quadrature-based moment methods (QBMM). In this chapter,
the product–difference (PD) and Wheeler algorithms employed for the classical
univariate quadrature method of moments (QMOM) are discussed, together with
the brute-force, tensor-product, and conditional QMOM developed for multivari-
ate problems. The chapter concludes with a discussion of the extended quadrature
method of moments (EQMOM) and the direct quadrature method of moments
(DQMOM).

• In Chapter 4 the GPBE is derived, highlighting the closures that must be introduced
for the passage from the microscale to the mesoscale model. This chapter also con-
tains an overview of the mathematical steps needed to derive the transport equations
for the moments of the NDF from the GPBE. The resulting moment-closure problem
is also throughly discussed.

• Chapter 5 focuses on selected mesoscale models from the literature for key phys-
ical and chemical processes. The chapter begins with a general discussion of the
mesoscale modeling philosophy and its mathematical framework. Since the number
of mesoscale models proposed in the literature is enormous, the goal of the chapter is
to introduce examples of models for advection and diffusion in real and phase space
Preface xv

and zeroth-, first-, and second-order point processes, such as nucleation, breakage,
and aggregation.

• Chapter 6 is devoted to the topic of hard-sphere collision models (and related simpler
kinetic models) in the context of QBMM. In particular, the exact source terms for
integer moments due to collisions are derived in the case of inelastic binary collisions
between two particles with different diameters/masses, and the use of QBMM to
overcome the closure problem is illustrated.

• Chapter 7 is devoted to solution methods of the spatially homogeneous GPBE,

including class and sectional methods, MOM and QBMM, and Monte Carlo meth-
ods. The chapter concludes with a few examples comparing solution methods for
selected homogeneous PBE.

• Chapter 8 focuses on the use of moment methods for solving a spatially inhomo-
geneous GPBE. Critical issues with spatially inhomogeneous systems are moment
realizability and corruption (due to numerical advection and diffusion operator) and
the presence of particle trajectory crossing (PTC). These are discussed after introduc-
ing kinetics-based finite-volume methods, by presenting numerical schemes capable
of preserving moment realizability and by demonstrating with practical examples
that QBMM are ideally suited for capturing PTC. The chapter concludes with a
number of spatially one-dimensional numerical examples.

• To complete the book, four appendices are included. Appendix A contains the
Matlab scripts for the most common moment-inversion algorithms presented in
Chapter 3. Appendix B discusses in more detail the kinetics-based finite-volume
methods introduced in Chapter 8. Finally, the key issues of PTC in phase space,
which occurs in systems far from collisional equilibrium, and moment conservation
with some QBMM are discussed in Appendix C and Appendix D, respectively.

The authors are greatly indebted to the many people who contributed in different ways
to the completion of this work. Central in this book is the pioneering research of Dr. Robert
L. McGraw, who was the first to develop QMOM and the Jacobian matrix transformation
(which is the basis for DQMOM) for the solution of the PBE, and brought to our atten-
tion the importance of moment corruption and realizability when using moment methods.
The authors are therefore especially grateful to Professor Daniel E. Rosner, who in 1999
directed their attention to the newly published work of Dr. McGraw on QMOM. They
would also like to thank Professor R. Dennis Vigil for recognizing the capability of QMOM
for solving aggregation and breakage problems, and Professor Prakash Vedula for provid-
ing the mathematical framework used to compute the moment source terms for hard-sphere
collisions reported in Chapter 6.
A central theme of the solution methods described in this book is the importance of
maintaining the realizability of moment sets in the numerical approximation. On this point,
the authors are especially indebted to Professor Marc Massot for enlightening them on
the subtleties of kinetics-based methods for hyperbolic systems and the general topic of
particle trajectory crossings. Thanks to the excellent numerical analysis skills of Professor
Olivier Desjardins and a key suggestion by Dr. Philippe Villedieu during the 2006 Summer
Program at the Center for Turbulence Research, Professor Massot’s remarks eventually
pointed us in the direction of the realizable finite-volume schemes described in Chapter 8.
In this regard, we also want to acknowledge the key contributions of Professor Z. J. Wang
xvi Preface

in the area of high-order finite-volume schemes and Dr. Varun Vikas for the development
and implementation of the realizable quasi-high-order schemes described in Appendix B.
The idea of publishing this book with Cambridge University Press is the result of
the interest shown in the topic by Professor Massimo Morbidelli. The contribution of
many other colleagues is also gratefully acknowledged, among them Antonello A. Barresi,
Marco Vanni, Giancarlo Baldi, Miroslav Soos, Jan Sefcik, Christophe Chalons, Frédérique
Laurent, Hailiang Liu, Alberto Passalacqua, Venkat Raman, Julien Reveillon, and Shankar
Subramaniam. All the graduate students and post-doctoral researchers supervised by the
authors in the last ten years who have contributed to the findings reported in this book are
gratefully acknowledged and their specific contributions are meticulously cited.
The research work behind this book has been funded by many institutions and among
them are worth mentioning the European Commission (DLM), the Italian Ministry of Edu-
cation, University, and Research (DLM), the ISI Foundation (DLM, ROF), the US National
Science Foundation (ROF), the US Department of Energy (ROF), the Ecole Centrale Paris
(ROF), and the Center for Turbulence Research at Stanford University (ROF). The constant
stimulus and financial support of numerous industrial collaborators (ENI, Italy; BASF,
Germany; BP Chemicals, USA; Conoco Phillips, USA; and Univation Technologies, USA)
are also deeply appreciated.
Notation

Upper-case Roman
A generic particle acceleration due to buoyancy, gravity, and drag
A coefficient matrix in DQMOM and brute-force QMOM for
determining the quadrature approximation
A coefficient matrix constituted by mixed moments for calculating
velocity parameters un in inhomogeneous systems

AD particle cross-section surface area

Aeq area of equivalent sphere
AH Hamaker parameter
Ap particle surface area

Af fluid acceleration due to body forces

Afp pure particle acceleration due to fluid–particle momentum exchange
Ap pure particle acceleration due to body forces
Ap,0 mean particle-acceleration term
Apf pure fluid acceleration due to fluid–particle momentum exchange

A(n) acceleration acting on nth particle due to body forces

(for particles in vacuum)
A(n)
f acceleration acting on the fluid in the neighborhood of the
nth particle due to pressure, body, and viscous forces
A(n)
fp acceleration acting on the nth particle
due to fluid–particle forces
A(n)
p acceleration acting on the nth particle due to body forces
(for particles suspended in a fluid)
A(n)
pf acceleration acting on the fluid in the neighborhood of the
nth particle due to fluid–particle forces

A∗fp global particle acceleration due to fluid–particle momentum

exchange (including diffusion terms)
A∗i coefficient matrix constituted by moments m∗j,k;i for calculating
velocity parameters un in inhomogeneous systems with FVM
Aiklj collision frequencies between particle-velocity classes in DVM

Af 1 multi-particle conditional expected fluid acceleration

due to body forces
Afp 1 multi-particle conditional expected particle acceleration
due to fluid–particle forces

xvii
xviii Notation
Ap 1 multi-particle conditional expected particle acceleration
due to field forces
Apf 1 multi-particle conditional expected fluid acceleration
due to fluid–particle forces
fp 1
A(n) single-particle conditional expected continuous particle acceleration
due to fluid–particle forces
p 1
A(n) single-particle conditional expected continuous particle acceleration
due to field forces

Af f total acceleration of the fluid seen by the particles

due to forces in the fluid phase
Afp N total acceleration acting on monodisperse particles
(of constant size and mass) due to drag, lift, and pressure forces
Afp p total acceleration acting on particles due to drag, lift,
and pressure forces
Ap N total acceleration acting on monodisperse particles
(of constant size and mass) due to body forces
Ap p total acceleration acting on particles due to body forces
Apf f total acceleration of fluid seen by the particles
due to momentum transfer between phases
Arp Archimedes number for disperse phase

B(x, y) beta function

B(g, x) hard-sphere collision kernel
Bi j rate of change of the particle-number density in intervals I1(i) and I2( j)
due to a generic point process (in CM for bivariate systems)
Bagg
i rate of change of the particle number in interval Ii due to
aggregation (in CM for univariate systems)
Bbreak
i rate of change of the particle number in interval Ii due to
breakage (in CM for univariate systems)

B coefficient matrix constituted by mixed moments for calculating

velocity parameters un (in inhomogeneous systems)
Bfv mixed phase-space diffusion tensor for fluid velocity
and particle velocity
Bfvf pure phase-space diffusion tensor for fluid velocity
Bfξ mixed phase-space diffusion tensor for fluid velocity
and particle internal coordinate
Bfξf mixed phase-space diffusion tensor for fluid velocity
and fluid internal coordinate

Bpv pure phase-space diffusion tensor for particle velocity

Bpvf mixed phase-space diffusion tensor for particle velocity
and fluid velocity
Bpξ mixed phase-space diffusion tensor for particle velocity
and particle internal coordinate
Bpξf mixed phase-space diffusion tensor for particle velocity
and fluid internal coordinate
Bo Bond number (equal to Eötvös number)

C constant appearing in the parabolic daughter distribution function

for particle breakage
C(ψ) collisional source term for monodisperse system
Notation xix
C1 coefficient appearing in definition of the fluid effective
(laminar plus turbulent) viscosity μf,eff
C1 –C7 constants appearing in breakage kernel

CD particle-drag coefficient
CD∗ particle-drag coefficient including the Cunningham correction
factor for rarefied continuous phase
Ci concentration of the potential-determining ions
CL lift-force coefficient
Cm momentum exchange coefficient appearing in
thermophoretic force
Cs thermal slip coefficient appearing in thermophoretic force
Ct thermal exchange coefficient appearing in
thermophoretic force

Cvm virtual-mass force coefficient identifying the fraction of

fluid volume moving with a particle
Cα coefficients appearing in the functional expansion of the NDF
Cαβ (ψ) collisional source term for polydisperse system
(particles of types α and β)
(m)
Cαβ (ψ) terms appearing in the expansion of the collisional source term
for polydisperse systems (particles of types α and β)
Ci∗γjk approximate collision source term for velocity moments
of global order γ (monodisperse systems)
Ci∗γjk,α approximate collision source term for velocity moments
of global order γ of particle type α (polydisperse systems)

C generic collisional source term

Cαβ collisional source term for particles of types α and β
CNp Np -particle collision operator
Cfv mixed phase-space diffusion tensor for fluid internal
coordinate and particle velocity
Cfvf mixed phase-space diffusion tensor for fluid internal
coordinate and fluid velocity
Cfξ mixed phase-space diffusion tensor for fluid internal
coordinate and particle internal coordinate

Cfξf pure phase-space diffusion tensor for fluid internal coordinate

Cpv mixed phase-space diffusion tensor for particle internal
coordinate and particle velocity
Cpvf mixed phase-space diffusion tensor for particle internal
coordinate and fluid velocity
Cpξ pure phase-space diffusion tensor for particle internal coordinate
Cpξf mixed phase-space diffusion tensor for particle internal
coordinate and fluid internal coordinate
Cipzc concentration of potential determining ions at point
of zero charge

CD(n) drag coefficient for the nth particle

C(n)
fξ rate of change of the internal coordinate vector for the fluid
surrounding the nth particle due to discontinuous events
C(n)
pU rate of change of velocity for the nth particle
due to collisions (particles suspended in fluid)
xx Notation
C(n)
pξ rate of change of particle internal coordinate vector for nth particle
due to collisions (particles suspended in fluid)

C(n)
U rate of change of velocity for the nth particle
due to discontinuous particle collisions (in vacuum)
C(n)
ξ rate of change of particle internal coordinate vector for nth particle
due to particle collisions (in vacuum)
C(n)
1 single-particle collision operator
Cl(m)
1 l2 l3
collision source terms for integer moments of orders l1 , l2 , and l3
with respect to the three velocity components

D solute molecular diffusion coefficient

Db diameter of a stable bridge between two aggregating particles
Df particle fractal dimension
DG average size of objects constituting a porous medium
D0 cut-off distance for calculating the Hamaker parameter
D symmetric N × N diffusion matrix
D̂ volume-average symmetric N × N diffusion matrix
D∗ dimensionless normalized diffusion matrix
Daa aggregation Damköhler number
Dab breakage Damköhler number

E bubble aspect ratio

Eαβ energy scaling factor in polydisperse systems (particle types α and β)
Ep total particle granular energy
Eo Eötvös number

F inter-particle force
F(ζ) dimensionless normalized NDF
Ft cumulative probability distribution for the quiescence time in
MC methods
γ
Fi,l 1 l2 l3
ith component of spatial flux for moment of global order γ
+
Fi,l 1 l2 l3
ith component of spatial flux for moment of global order γ
corresponding to positive velocity
−
Fi,l 1 l2 l3
ith component of spatial flux for moment of global order γ
corresponding to negative velocity

F(M) generic moment flux function

Ffp drag and buoyancy fluid–particle force
Fγl1 l2 l3 spatial flux for moment of global order γ
F+l1 l2 l3 spatial flux for moment of global order γ
corresponding to positive velocity
F−l1 l2 l3 spatial flux for moment of global order γ
corresponding to negative velocity

Fl flow number for particle aggregation

Frg Froude number for the continuous phase

G(nl , nr ) numerical flux function for inhomogeneous systems

discretized with FVM
Gf fluid shear rate
Notation xxi
Gi gain rate of particles with velocity ξi due to collisions in DVM
Gi jk,αβ Gaussian moments with mean velocity Uαβ and
covariance matrix σαβ
GL continuous rate of change of particle size (growth rate)
GL,k average particle growth rate for moment of order k

Gm (ψ) component m of the collisional-flux term

Gm,l1 l2 l3 collisional-flux term for integer moments of orders l1 , l2 , and l3
with respect to the three velocity components
Gp pure advection component for the rate of change of crystal size
Gof correction for fluid-dynamic interactions between particles
Gkn coefficients appearing in the calculation of the velocity parameters un
for inhomogeneous systems
G∗p global rate of change of crystal size
G∗p1 mass-transfer rate from fluid to particle

G vectorial numerical flux function for inhomogeneous systems

discretized with FVM
G(ψ) collisional-flux term for monodisperse systems
Gf pure mesoscale advection model for the fluid internal coordinate
Gl1 l2 l3 collisional-flux vector for velocity moment of order l1 l2 l3
Gp pure mesoscale advection model for particle internal coordinate
Gαβ (ψ) collisional-flux term for polydisperse systems constituted by
particle types α and β

G(n) continuous rate of change of the internal coordinate vector

for the nth particle (particles in vacuum)
G(n)
f continuous rate of change of the internal coordinate vector
for the fluid surrounding the nth particle
G(n)
p continuous rate of change of the internal coordinate vector
for the nth particle (particles suspended in fluid)
G(m)
αβ (ψ) terms appearing in the expansion for the collisional-flux term
for polydisperse systems with particle types α and β

Gf 1 multi-particle conditional expected continuous rate of change

of fluid internal coordinate vector
f 1
G(n) single-particle conditional expected continuous rate of change
of fluid internal coordinate vector
Gp 1 multi-particle conditional expected continuous rate of change
of particle internal coordinate vector
p 1
G(n) single-particle conditional expected continuous rate of change
of particle internal coordinate vector

Gf  rate of change of fluid-phase mass density due to

continuous processes
Gp  rate of change of disperse-phase mass density due to
continuous processes
Gf V rate of change of fluid-phase volume fraction due to
continuous processes
Gp V rate of change of disperse-phase volume fraction due
to continuous processes
Gf f global fluid momentum rate of change due to mass transfer
from fluid phase
xxii Notation
Gp p global particle momentum rate of change
due to mass transfer from liquid phase

H(ξ) Heaviside step function

H functional appearing in the general definition of the pair
distribution function (function of the moments)
Ho distance between two primary particles within an aggregate
Hm,l1 l2 l3 symmetric change of variable involving Il(m)
1 l2 l3

Ii finite ith internal coordinate interval used in CM

Ir relative turbulence intensity
I[x1 ,x2 ] (x) indicator function equal to unity if x ∈ [x1 , x2 ] and zero otherwise
I (i)
j finite ith interval for the jth internal coordinate
in CM when extended to multivariate problems
Il(m)
1 l2 l3
factor appearing in collision term for integer
velocity moments of orders l1 , l2 , and l3
Il(p,q)
1 l2 l3
factor appearing in the collision term for polydisperse systems
for integer velocity moments of orders l1 , l2 , and l3

J molar flux of solute molecules at particle surface

J(φf ) rate of particle formation
J(η̃, η) Jacobian of variable transformation relating phase-space
variables before and after collision
Jf rate of particle formation
Jˆf volume-average rate of particle formation
Jk kth moment of the rate of particle formation in univariate GPBE
Jk kth moment of the rate of particle formation in multivariate GPBE
J∗ dimensionless and normalized rate of particle formation

KB history-force kernel
Ki(m)
jk integral over collision angles
Ki(mn)
jk coefficients appearing in the third integral over collision angles
K exponent matrix used to build the quadrature approximation
K moment vector used in the definition of the moment set M
Kmi,α reconstructed K in the ith cell at time step m employed in FVM
K±α,l/r K evaluated with v+α,l or v−α,r

Kn Knudsen number for continuous phase (relative to particle diameter)

Kn∗ Knudsen number for continuous phase (relative to particle radius)
Knp Knudsen number for disperse phase

L characteristic length of the system under investigation

L particle length
Li loss rate of particles with velocity ξi due to collisions in DVM
Lv latent heat of evaporation
L10 number-average mean particle length
L32 area-average mean particle length or Sauter diameter
L43 volume-average mean particle length
L transformation matrix between laboratory and collision frames
Le Lewis number
Notation xxiii
M number of internal coordinates appearing in the NDF
M number of sections or classes used in CM and DVM

Mf number of fluid internal coordinates appearing in the NDF

MG ratio between particle and collector size
Mk number of intervals used for the kth internal
coordinate in multivariate CM
Mi mass of the particles in the interval Ii in CM
Mp number of particle internal coordinates appearing in the NDF
Mp particle mass
Mw molecular weight (relative molecular mass) of chemical species
Mw1 molecular weight of the evaporating component
Mw2 molecular weight of the stagnant component

Miagg rate of change of particle mass in interval Ii due to aggregation in CM

Mibreak rate of change of particle mass in interval Ii due to breakage in CM
Miγjk velocity distribution moment of global order γ = i + j + k
Miγ∗jk velocity equilibrium moment of global order γ = i + j + k
Miγjk,α velocity distribution moment of global order γ = i + j + k
for particles of type α in polydisperse systems
Mi,l+ 1 l2 l3 positive integer moment of the velocity distribution
in the ith direction
Mi,l− 1 l2 l3 negative integer moment of the velocity distribution
in the ith direction

M vector defining the tracked moment set

M+ positive half-moment set (integration for positive velocity)
M− negative half-moment set (integration for negative velocity)
Mmi volume-averaged moment set in the ith cell at time step m
defined in FVM for a 1D grid
Mmijk volume-averaged moment set at cell Ωi jk and time step m
defined in FVM for a structured 3D grid

(Mmijk ) p updated moment set at cell Ωi jk and time step m defined

in FVM for a structured 3D grid calculated with permutation p
M(h)
i jk moment set at cell Ωi jk updated after advection in the h direction
in FVM for a structured 3D grid
M(1)
i first-stage moment set in the ith cell at time step m
defined in FVM for a 1D grid
M+i jk positive half-moment set at cell Ωi jk in FVM (3D)
M−i jk negative half-moment set at cell Ωi jk in FVM (3D)
M∗i second-stage moment set in the ith cell at time step m
defined in FVM for a 1D grid

Ma Mach number for continuous phase

Map Mach number for disperse phase
Mo Morton number

N order of the quadrature approximation

N± number of quadrature nodes used in the calculation
of the positive and negative fluxes
xxiv Notation
N(t, x) total particle-number concentration (or density)
N(V|V  ) volume-based daughter distribution function
N(ξ |ξ) daughter-particle conditional NDF

Nd global number of degrees of freedom of the multiphase system

Ni number of particles belonging to the ith interval in CM and DVM
Ni j number of particles belonging to intervals I1(i) and I2( j)
for bivariate CM
Ni number density of particles with velocity equal to ξi in DVM
Np number of particles considered in NDF definition
Np number of primary particles forming a fractal object
Nt total particle-number density
N number of weights and abscissas for Laguerre-
polynomial recursion coefficients in EQMOM
Nu Nusselt number

P(ξ |ξ) daughter-particle conditional PDF

Pα (ξ) polynomial of order α orthogonal to the NDF
Pα,β elements of the matrix used in the PD algorithm
Pi j components of the second-order pressure tensor
Pi j probability for the encounter of particles i and j in MC simulations
P matrix used in the PD algorithm
P total particle stress tensor and second-order pressure tensor
Pα second-order pressure tensor in polydisperse systems
for particles of type α
Pe Péclet number
Pr Prandtl number

Q+ positive moment flux (integration for positive velocity)

Q− negative moment flux (integration for negative velocity)

R ideal-gas constant
Rg radius of gyration of the particle
Reg continuous (gas)-phase Reynolds number
Rep disperse (particulate)-phase Reynolds number
Recp critical particle Reynolds number
ReMp meta-critical particle Reynolds number
Re∗p modified disperse-phase Reynolds number

S supersaturation
S comprehensive source term in the GPBE including drift, diffusion,
and point processes
Sc particle collisional cross-sectional area
S k,i reconstructed slope in the ith cell for the solution of the
moment transport equation discretized with FVM (1D)
n
S αβ,i reconstructed slope in the ith cell at time instant n
from EQMOM discretized with FVM (1D)
S moment set source term

S generic source term due to discontinuous events for the GPBE

SfV source term due to discontinuous events for
the fluid-phase volume fraction
SM source term due to discontinuous events for
the disperse-phase mass density
Notation xxv
SN source term due to discontinuous events for the total
disperse-phase number density
SV source term due to discontinuous events for the disperse-phase
volume fraction
S1 generic source term due to discontinuous events for GPBE
Sf viscous and pressure stress tensor for fluid phase

Sk source term for the kth moment of the NDF

Sk source term for the moment of order k of the multivariate NDF
S i jk source term of the moments of orders i, j, and k with respect to the
particle-velocity components for the NDF
S l(m)
1 l2 l3
factor appearing in the collision term for velocity moments of the
NDF of orders l1 , l2 , and l3 with respect to the three components
S lγ1 l2 l3 source terms for moment of global order γ = l1 + l2 + l3

Sp rate of change for particle momentum due to discontinuous events

S+ collision cross section

Sc Schmidt number
Sh Sherwood number
St Stokes number
Stp particle Stokes number

Tf continuous (fluid) phase temperature

Tp disperse (particulate) phase temperature
T ref reference temperature for liquid boiling
Ts temperature on particle surface
T sat saturation temperature for the continuous phase

U characteristic continuous phase velocity

U(ξ) particle velocity conditioned on particle size ξ
U characteristic particle velocity
Ug continuous phase velocity
Up mean particle velocity
Ur impact, or relative, velocity for fragmenting particles
U∞ particle terminal velocity

Uf mean fluid velocity field

UfV volume-average fluid velocity
UM mass-average particle velocity
Umix mass-average mixture velocity
UN number-average particle velocity
Up mean particle velocity
Up,2 second-order particle velocity-moment tensor
Up,k conditional particle velocity for ξp = ξpk
UpM fluid-mass-average particle velocity
UV volume-average particle velocity
Uαβ mean velocities for polydisperse Gaussian distributions
Up† characteristic disperse-phase velocity

U(n) velocity of the nth particle (in vacuum)

U(n)
f fluid velocity in the neighborhood of the nth particle
xxvi Notation
U∗f mesoscale variable describing fluid velocity
U(n)
p velocity of the nth particle
U∗p mesoscale variable describing particle velocity

Uf  Reynolds-average fluid velocity field

Up |ξ = ζ particle conditional velocity for ξ = ζ
Up |ξp  expected mean particle velocity for internal coordinates equal to ξp
Uf p particle-mass-average fluid velocity

V particle volume
Vf fluid volume seen by the particle
VL length-based volume density function
Vp particle volume
VW sample volume used in the estimation of the NDF

∗
Vαβ;i particle velocity with EQMOM in the ith cell after the advection step
when using FVM (1D) with time splitting
†
Vαβ;i particle velocity with EQMOM in the ith cell after the advection step
when using FVM (1D) with time splitting
V dimensionless and normalized particle velocity
V(n)
f fluid-velocity space for fluid surrounding the nth particle
V(n)
p particle-velocity space for the nth particle

W(t) generic Wiener process

Wαβ product of wα and wαβ in EQMOM
∗
Wαβ product of w∗α and w∗αβ in EQMOM
and calculated after the advection step with time splitting
W(t) generic vectorial Wiener process
We Weber number

X dimensionless and normalized spatial coordinate

XT abscissa (or node) matrix
X(n) center of mass of the nth particle
X∗p mesoscale variable describing particle position

Yf1 molar fraction of evaporating component in the gas phase

Ys1 molar fraction of evaporating component on droplet surface
Y1 gas-phase molar fraction of evaporating component
Y2 gas-phase molar fraction of stagnant component

Yf fluid-phase species mass fractions

Yp particle species mass fractions
Yα = wα ξα weighted node (or abscissa) of the M-dimensional
quadrature approximation
YT weighted-abscissa (or weighted-node) matrix

Lower-case Roman
a aggregation kernel
a breakage kernel (Chapter 7)
a0 constant-breakage kernel (Chapter 7)
Notation xxvii
ai j affinity parameter for i– j aggregation (with i = A, B and j = A, B)
aα coefficients of recursive formula for orthogonal polynomials
aα source term for the evolution equation of weight wα in DQMOM
aα coefficients of recursive formula appearing in Wheeler algorithm
a volume-average aggregation kernel
amijk slope vector employed in second-order spatial reconstructions
for FVM (3D)

b frequency of first-order process (breakage kernel)

b(ξ |ξ) daughter distribution function (Chapter 7)
bi,α source term for the evolution equation of weighted node α
for the ith internal coordinate in DQMOM
bα coefficients of recursive formula for orthogonal polynomials
bα source term for the evolution equation of weighted node α
in univariate DQMOM
bα coefficients of recursive formula appearing in Wheeler algorithm
b∗ dimensionless and normalized kernel for first-order process

b̂ volume-average frequency of first-order process

b volume-average breakage kernel
k
bα moment transform of order k of the daughter distribution
function for ξα in univariate problems
k
bα moment transform of order k of the daughter distribution
function for ξα in multivariate problems

bfvf fluid velocity coefficient for fluid fluctuations

bk response vector for third-order differences to a unit increment
bpvf particle-velocity coefficient for fluid fluctuations

ceq equilibrium solute concentration

cpξp self-diffusion component of crystal size growth rate
cαβ constant appearing in the definition of the pair distribution function
c± coefficients appearing in upwind reconstruction schemes
cpf specific heat of fluid phase
cpp particle specific heat
cα lattice velocities used in LBM

d degree of accuracy of the quadrature approximation

d molecular diameter of a solute molecule
do diameter of primary particles in an aggregate
dp particle size

d10% size corresponding to 10% of the smaller particles

d90% size corresponding to 10% of the larger particles
dα size of the particles of type α taking part in a collision
dβ size of the particles of type β taking part in a collision
dαβ arithmetic average of colliding particle size (polydisperse systems)

dp∗ characteristic particle length

dki j net flux of particles for the kth internal coordinate
due to phase-space diffusion in bivariate CM
d vector containing the moments source terms in DQMOM
dn difference vector of order n of natural logarithm of NDF moments
xxviii Notation
e coefficient of restitution for particle–particle collisions
e elementary charge
ei particle specific energy for the ith velocity component
ep particle specific energy
eαβ restitution coefficient in polydisperse systems for
collisions between particles of types α and β

f (t, x, v) particle-velocity NDF for monodisperse system

f (ξ1 , ..., ξ M−1 ) marginal NDF used in CQMOM
f (ξ M |...) conditional NDF used in CQMOM
f (δ) shape function for colliding particles
ff one-point PDF for the fluid
fGf volume distribution of the fluid shear rate
fNp multi-particle joint PDF
ft probability density function for the quiescence time in MC methods
feq equilibrium velocity NDF

fα weights used in reconstructing the NDF in LBM

fα (t, x, v) particle velocity NDF for particle type α
fβ (t, x, v) particle velocity NDF for particle type β
fε volume distribution of the fluid turbulent dissipation rate

f (2) pair correlation velocity NDF

f ∗ (v) equilibrium distribution function in the BGK
kinetic model and in Grad’s moment closure
f1(n) single-particle joint PDF
fki j net flux of particles for the kth internal coordinate
due to phase-space drift in bivariate CM
fα∗ weights used in LBM corresponding to the equilibrium NDF

g gravity acceleration constant

g( ) Kuwabara function for particle deposition in porous media
g0 (αp /α∗p ) particle radial distribution function
gαβ pair correlation function for particles of types α and β
g0,αβ constant appearing in the pair correlation function
gn (ξ) velocity parameters used in conjunction with EQMOM
g mean velocity difference used to approximate |v1 − v2 |

h size of the regular discretization used in CM

h(ξ) function used to model the second-order tensor for mixed advection
hi (ξ) numerical NDF in the ith cell for FVM (1D)
hL discontinuous event term for length-based formulation
hL,k moment of order k of source term for discontinuous event
hW constant kernel function used as filter to estimate NDF
h1 particle collisional acceleration term
h+ rate of particle formation due to discontinuous events
h− rate of particle disappearance due to discontinuous events

i index vector identifying a discrete particle velocity in DVM

k order of moment for univariate NDF

kA particle surface shape factor
Notation xxix
kB Boltzmann constant
kc corrective growth crowding factor
kco coordination number for an aggregate
kd particle mass-transfer coefficient

kf fluid-phase turbulent kinetic energy

kf fluid-phase thermal conductivity
kg fractal scaling factor of order unity
kh particle heat-transfer coefficient
ki (ξ) polynomial used to represent the NDF in the Ii interval in CM
kp thermal conductivity of particle
kV particle volumetric shape factor

kA∗ equivalent particle surface shape factor

kV∗ equivalent particle volumetric shape factor
k exponent vector for the order of moment in multivariate NDF
kξ internal coordinate exponent vector

m particle mass (used in daughter NDF for breakage)

m(k) moment of order k of univariate NDF
m(k) moment of order k = (k1 , . . . , k M ) of multivariate NDF
mc mass of newly formed particle (nucleus mass)
ṁcp mass of liquid evaporating per unit volume and unit time
mk moment of order k of NDF
mi jk moment of order i, j, and k with respect to the three velocity
components of the particle velocity NDF
m j,k;i mixed moment of orders j and k for the ith cell in FVM (1D)

mL,k kth moment of length-based NDF

mM,k kth moment of mass-based NDF
mU,k kth moment of velocity-based NDF
mV,k kth moment of volume-based NDF
mξ,k moment of order k = (k1 , . . . , k M ) of the multivariate NDF

mα mass of particle of type α taking part in collision

mβ mass of particle of type β taking part in collision
m∗k moment definition used in EQMOM
mnk,i reconstructed moment for the ith cell at time instant n in FVM
m∗j,k;i mixed moment of order j and k for the ith cell in FVM
when using time splitting after convection
m†j,k;i mixed moment of orders j and k for the ith cell in FVM
when using time splitting after drag

n generic NDF appearing in GPBE

neq equilibrium NDF
nL length-based NDF
nM mass-based NDF
nU velocity-based NDF
nV volume-based NDF
nξ generic NDF
nξ  Reynolds-average NDF
n̂ volume-average NDF
xxx Notation
n∗ reconstructed NDF
n∗ equilibrium Maxwellian NDF
n[13] NDF reconstructed in Grad’s 13-moment closure
nmi (ξ) NDF in the ith cell at time step m reconstructed
in FVM (1D) from Mmi

n±α weights of the quadrature approximation calculated from the

positive and negative moments of the velocity distribution
n(h)
i jk NDF in Ωi jk updated after advection in the h direction for FVM
nmijk NDF in cell Ωi jk at time t = m Δt for FVM
n+i jk contribution for updating the NDF due to advection from
positive velocity in cell Ωi jk for FVM
n−i jk contribution for updating the NDF due to advection from
negative velocity in cell Ωi jk for FVM

pf pressure of the fluid phase

pf1 partial pressure of evaporating component in gas phase
pg pressure of the gas phase
pref reference pressure for boiling liquid
ps1 partial pressure of evaporating component on droplet surface
pα orthogonal polynomials of order α used
in functional expansion of NDF
pα granular pressure of particles of type α

q heat flux to surface of particle

qi skewness of the NDF with respect to the ith velocity component
q total particle energy flux

s specific surface area of the porous medium

s ratio of geometric grids employed in CM
s particle surface area

tαβ abscissa computed from Laguerre-polynomial recursion

coefficients used in EQMOM

u generic known disperse-phase velocity (1D)

u(ξ) known particle velocity conditioned on internal coordinate ξ
uk (with k = 0, 1, 2) flow-dependent velocity parameters
uα velocity node α of the quadrature approximation
u+i,l positive i-component of velocity evaluated at left face of
cell Ωi jk in FVM (1D)
u−i,r negative i-component of velocity evaluated at right face of
cell Ωi jk in FVM (1D)

up up N disperse-phase stress tensor

v disperse-phase velocity (1D)

vf fluid-phase velocity (1D)
v+i positive i-component of velocity evaluated at left face
of cell Ωi jk in FVM
v−i negative i-component of velocity evaluated at right face
cell of Ωi jk in FVM
Notation xxxi
v+α velocity node α calculated from positive half-moment set
v−α velocity node α calculated from negative half-moment set

v phase-space variable for particle velocity vector (in vacuum)

vf phase-space variable for fluid velocity
vp phase-space variable for particle velocity vector
vr relative velocity of two colliding particles
v12 velocity difference before a direct collision

v±α nodes of the quadrature approximation calculated from the

positive and negative moments of the velocity distribution
v12 velocity difference after a direct collision
v12 velocity difference before an inverse collision
vmijk,α velocity node α at cell Ωi jk and time m Δt in FVM (3D)
v+i jk,α velocity node α at cell Ωi jk calculated from M+i jk
in FVM for 3D structured grids
v−i jk,α velocity node α at cell Ωi jk calculated from M−i jk
in FVM for 3D structured grids

w weight function corresponding to a family of orthogonal

polynomials in functional expansion of NDF
wα generic weight of a quadrature approximation
w α1 weight in CQMOM
wα1 ;α2 conditional weight in CQMOM
wα1 ;α2 ;α3 conditional weight in CQMOM
wαβ weight computed from Laguerre polynomial
recursion coefficients in EQMOM

wmi,α reconstructed weight α in the ith cell at time instant m

employed in FVM (1D)
wmijk,α weight α at cell Ωi jk at time m Δt
employed in FVM for 3D structured grids
w+α weight calculated from the positive half-moment set
w−α weight calculated from the negative half-moment set
w+i jk,α weight α at cell Ωi jk calculated from M+i jk
in FVM for 3D structured grids
w−i jk,α weight α at Ωi jk calculated from M−i jk
in FVM for 3D structured grids
w weight vector

x cosine of collision angle α x

x physical point coordinate vector
x12 collision vector in laboratory frame
x†12 collision vector in collision frame
x(n) particle-position space for nth particle

y cosine of collision angle αy

z valency of potential-determining ions

Upper-case Greek
Γ disperse-phase diffusion coefficient
Γ(x) gamma function
xxxii Notation
Γp effective disperse-phase diffusion coefficient
ΓT turbulent disperse-phase diffusion coefficient
Γ second-order tensor for multidimensional diffusion

Δc mass-transfer driving force

Δt uniform time step employed in FVM
Δx uniform spatial grid spacing in x direction employed in FVM
Δy uniform spatial grid spacing in y direction employed in FVM
Δz uniform spatial grid spacing in z direction employed in FVM
Δξ size of velocity intervals in DVM

Θ generic granular temperature

Θp disperse-phase granular temperature
Θα granular temperature for particles of type α
Θβ granular temperature for particles of type β
Θαβ granular temperature for polydisperse systems of
particle types α and β

Πi, j,k moment of order i, j, and k of the collision integral with respect
to the three velocity components

Σ generic velocity covariance matrix

Σp disperse-phase velocity covariance matrix

Ψs particle surface potential

Ωx control volume in physical space

Ωi jk cell for a structured Cartesian grid centered at (i Δx, j Δy, k Δz)
|Ωi jk | volume of the cell Ωi jk
∂Ωx boundaries of physical-space control volume
Ωξ integration or support interval
Ωξ control volume in phase space
∂Ωξ boundaries of phase-space control volume

Lower-case Greek
α aggregation efficiency
αf volume fraction of fluid phase
αg volume fraction of gas phase
αikl rate of gain of particles with velocity ξi due to collision
between particles with velocities ξk and ξl in DVM
αp volume fraction of disperse phase
αV length-based volume-fraction density function
αx angle formed by ξ and x in the formulation of the
isotropic collision operator
αy angle formed by ξ∗ and x in the formulation of the
isotropic collision operator
α1 volume fraction of particles of type 1
α2 volume fraction of particles of type 2

αcf thermal conductivity of continuous phase

αcp thermal conductivity of disperse phase
α∗p packing-limit volume fraction for disperse phase
α vector with source terms for weights and nodes in DQMOM
Notation xxxiii
β evaporation frequency
β kernel for generic second-order process
β̂ volume-average kernel for generic second-order process
βx angle formed by ξ and x for the isotropic collision operator
βy angle formed by ξ∗ and x for the isotropic collision operator
β0 constant aggregation kernel
β∗ dimensionless normalized kernel for second-order process
βV collision rate in the computational box in MC simulations

γ = |k| global moment order

γ effective volume coefficient
γx ratio between time step and grid spacing in x direction
γy ratio between time step and grid spacing in y direction
γz ratio between time step and grid spacing in z direction
γii reassignment factor for CM to describe discontinuous events

δ size ratio of colliding particles

δ(x) Dirac delta function
δσ kernel density function in EQMOM
δ(x) multidimensional Dirac delta function

porosity of porous medium

ε turbulence-dissipation rate
ε̂ volume-average turbulence-dissipation rate
εf fluid-phase turbulence-dissipation rate
εr relative permittivity constant
ε0 absolute permittivity constant

ζ model constant for BGK collision model (inverse of Pr)

ζi pivot point for interval Ii used in CM
ζi adaptive pivot point in interval Ii in CM with cell averaging
ζα coefficients of the continued fraction in PD algorithm
ζk(i) pivot point for interval Ii for the kth internal
coordinate in multivariate CM
ζ dimensionless and normalized internal coordinate vector
ζ̇ dimensionless and normalized continuous rate of change
of the internal coordinate vector

η coalescence efficiency
ηD particle-deposition efficiency due to Brownian motion
ηi kurtosis of the NDF with respect to the ith velocity component
ηi j reassignment coefficient for I1(i) and I2( j) in bivariate CM
ηI particle-deposition efficiency due to interception

η phase-space vector for particles involved in second-order processes

η(n)
f internal-coordinate space for fluid surrounding nth particle
η(n)
p internal-coordinate space for nth particle

θ collision angle for x12

θ1 collision angle for v12

κ Debye–Hückel parameter
καβ parameter appearing in the radial distribution function
xxxiv Notation
καβ collision-rate parameters for polydisperse BGK-like models
κ symmetric second-order tensor constructed from Ii(0)
jk
appearing in approximate collision models

λ ratio between time step and grid spacing in FVM (1D)

λf mean free path of molecules in continuous phase
λk moment of order k of a log-normal distribution
λ1 diffusion coefficient appearing in the moment-transport
equation discretized with FVM
λ2 convective coefficient appearing in the moment-transport
equation discretized with FVM
λ± (ξ) function appearing in CM with the cell averaging
λ second-order velocity tensor in
the equilibrium NDF employed in BGK model
λ non-negative second-order tensor for mixed advection

μ mean internal coordinate for Gaussian and log-normal distributions

μk moments in the positive tail of a Gaussian distribution
μf dynamic viscosity for continuous phase
μf,eff effective dynamic viscosity (including turbulent viscosity)
for continuous phase
μp granular viscosity
μαβ reduced mass of particles in polydisperse system

ν(vp , ξp ) number of daughter particles formed by first-order process

νf kinematic viscosity of continuous fluid phase
νg kinematic viscosity of continuous (gas) phase
νk coefficients calculated from modified moments in Wheeler algorithm
νi j rate of loss of particles with velocity ξi due to collisions with
particles with velocity ξ j in DVM
ν̃ solute molecular volume

ξ internal coordinate
ξ̇ continuous rate of change of particle internal coordinate
ξα abscissa or node α of the quadrature approximation
ξβ,α component for the β internal coordinate of the
abscissa α of the M-dimensional quadrature approximation
ξαβ node of the quadrature approximation for EQMOM
ξαβ;i node of the quadrature approximation for EQMOM
in the ith cell with FVM (1D)
ξi,α
m
reconstructed node α in the ith cell at time m Δt in FVM
∗
ξαβ;i node of the quadrature approximation for EQMOM in the ith
cell with FVM and time splitting after the advection step

ξ internal coordinate vector

ξ test particle pre-collision velocity
ξ̇ continuous rate of change of particle internal coordinate
ξ̇ˆ volume-average continuous rate of change of internal coordinate
ξ test particle post-collision velocity

ξf phase-space variable for fluid characteristic properties

ξfc fluid internal coordinate seen by newly formed particles
ξi discrete particle velocity used in DVM
ξi discrete velocity with index i = (i1 , i2 , i3 ) used in DVM
Notation xxxv
ξp phase-space variable for particle internal coordinate
ξpc internal coordinate of newly formed particles
ξα node or abscissa of M-dimensional quadrature approximation
ξ∗ field particle pre-collision velocity

ξ∗ field particle post-collision velocity

ξ(n) internal coordinate vector (of size M) of the nth particle (in vacuum)

ξf(n) vector for characteristic properties of fluid surrounding nth particle

ξf∗ mesoscale variable describing fluid internal coordinate
ξi,α
m
reconstructed node α in the ith cell at time m Δt in FVM (1D)
ξimjk,α node α at cell Ωi jk at time m Δt in FVM (3D)
ξi+jk,α node α at cell Ωi jk calculated from M+i jk in FVM (3D)
ξi−jk,α node α at cell Ωi jk calculated from M−i jk in FVM (3D)
ξp(n) internal coordinate vector (of size M) of the nth particle
ξp∗ mesoscale variable describing particle internal coordinate

ξk  moment of order k of a univariate NDF

k1 k2
 1k ξ2 . . . ξ M 
ξ kM
moment of order k = (k1 , . . . , k M ) of a multivariate NDF
ξ 2
conditional moment of order k2
 2k 
ξ33 conditional moment of order k3

πα (ξ) set of basis functions employed in Wheeler algorithm

to calculate weights and abscissas

ρf material density of fluid phase

ρg material density of gas phase
ρp material density of particles
ρ1 material density of particles of type 1
ρ2 material density of particles of type 2

f fluid-phase mass density

fα fluid-phase mass density for component α
mix mixture mass density
p disperse-phase mass density
pα disperse-phase mass density for component α

σ parameter used in definition of kernel density function in EQMOM

σ interfacial tension between continuous and disperse phases
σαβ velocity covariance matrix for polydisperse system
σα,β matrix used to calculate Jacobi matrix in Wheeler algorithm

σ velocity-covariance tensor
σp velocity-covariance matrix
σα velocity-covariance matrix for particles of type α
in polydisperse system

ςα weighted node α of quadrature approximation in DQMOM

ςi,α weighted node α of quadrature approximation for
the ith internal coordinate in DQMOM

τa characteristic time scale for particle aggregation

τb characteristic time scale for particle breakage
τc characteristic time scale for particle collision
xxxvi Notation
τcol characteristic collisional time for BGK-like models
τd characteristic time scale for particle–fluid drag force
τD,i j characteristic time scale for diffusion in phase space
τf characteristic time scale for fluid phase
τJ characteristic time scale for particle formation
τM characteristic particle advection time scale
τn characteristic time scale for growing a connecting neck
between two colliding particles
τp particle relaxation time
τs aggregate strength
τW autocorrelation time related to spatial filter
used to estimate an NDF

τI characteristic time scale for first-order process

τII characteristic time scale for second-order process

τp traceless granular tensor

τα traceless granular tensor for particles of type α
in polydisperse system
τξ̇ characteristic time scale for continuous change
of internal coordinates

τ expected interval of quiescence in MC simulation

φ particle composition
φ collision angle for x12
φc size-correction factor
φmax energy barrier for particle aggregation
φp volume fraction of solid within an aggregate
φ1 spherical angle for v12
φ1 density ratio between disperse and continuous (gaseous) phases
φ2 mass ratio between disperse and continuous (gaseous) phases
φ3 characteristic velocity ratio of disperse and continuous phases
φf fluid composition vector

χ factor relating velocities for direct and inverse collisions

χαβ factor relating velocities for direct and inverse collisions
for polydisperse particles of types α and β

ψ(ξ) dimensionless self-similar PSD

ψ(v) generic property change during a collision

ω restitution parameter relating velocities for direct

and inverse collisions
ωαβ mass-average restitution parameter for polydisperse particles
of types α and β

Superscripts and angle brackets


fluctuating component

property of particle undergoing breakage

property of field particle undergoing aggregation
Notation xxxvii

property of particle after a direct collision

property of particle before an inverse collision
†
variable in collision frame of reference
˜ property of test particle undergoing aggregation
time average
 ensemble and volume average

Subscripts
f property of fluid phase
g property of gas phase
p property of disperse phase
i−1/2 interface between the grid cells i − 1 and i in FVM
i+1/2 interface between the grid cells i and i + 1 in FVM
l limit value approaching the cell interface from the left in FVM
r limit value approaching the cell interface from the right in FVM

Operators
∂/∂x gradient operator in physical space
∇x gradient operator in physical space
∂/∂ξ gradient operator in internal-coordinate phase space
∇ξ gradient operator in internal-coordinate phase space
∂/∂v gradient operator in velocity phase space
∇v gradient operator in velocity phase space
· scalar product
⊗ dyadic or tensorial product
: total tensor product

Abbreviations
BGK Bhatnagar–Gross–Krook
CFD computational fluid dynamics
CFL Courant–Friedrichs–Lewy
CM classes methods
CN Crank–Nicolson
CQMOM conditional quadrature method of moments
CSD crystal-size distribution
DEM discrete-element method
DNS direct numerical simulation
DPBE discrete population-balance equation
DQMOM direct quadrature method of moments
DQMOM-FC direct quadrature method of moments, fully conservative
DVM discrete-velocity method
ECQMOM extended conditional quadrature method of moments
EQMOM extended quadrature method of moments
ES-BGK ellipsoidal–statistical Bhatnagar–Gross–Krook
FEM finite-element method
FVM finite-volume method
xxxviii Notation
GPBE generalized population-balance equation
KBFVM kinetics-based finite-volume method
KE kinetic equation
LBM lattice Boltzmann method
LES large-eddy simulation
MCM Monte Carlo method
MOM method of moments
MOMIC method of moments with interpolative closure
MUSCL monotone upstream-centered scheme for conservation laws
MUSIG multiple-size group
MWR method of weighted residuals
NDF number-density function
NS Navier–Stokes equation
ODE ordinary differential equation
PBE population-balance equation
PD product difference
PDF probability density function
PSD particle-size distribution
PTC particle trajectory crossing
QBMM quadrature-based moment method
QMOM quadrature method of moments
QUICK quadratic upstream interpolation for convective kinematics
RANS Reynolds-average Navier–Stokes equation
RK2SSP second-order two-stage strong stability-preserving Runge–Kutta
1

Introduction

1.1 Disperse multiphase flows

The majority of the equipment used in the chemical process industry employs multiphase
flow. Bubble columns, fluidized beds, flame reactors, and equipment for liquid–liquid
extraction, for solid drying, and size enlargement or reduction are common examples. In
order to efficiently design, optimize, and scale up industrial systems, computational tools
for simulating multiphase flows are very important. Polydisperse multiphase flows are also
common in other areas, such as fuel sprays in auto and aircraft engines, brown-out condi-
tions in aerospace vehicles and particulate flows occurring in the environment. Although
at first glance the multifarious industrial and environmental multiphase flows appear to
be very different from each other, they have a very important common element: it is pos-
sible to identify a continuous phase and a disperse phase (usually with a distribution of
characteristic “particle sizes”).
Historically the development of the theoretical framework and of computational mod-
els for disperse multiphase flows has focused on two different aspects: (i) the evolution of
the disperse phase (e.g. breakage and coalescence of bubbles or droplets, particle–particle
collisions, etc.) and (ii) multiphase fluid dynamics. The first class of models is mainly
concerned with the description of the disperse phase, and is based on the solution of the
spatially homogeneous1 population balance equation (PBE). A PBE is a continuity state-
ment written in terms of a number density function (NDF), which will be described in detail
in Chapter 2. The NDF contains information about how the population constituted by the
discrete elements of the disperse phase is distributed over certain characteristic properties
that determine, for example, product quality. For example, in crystallization the final qual-
ity of the crystals very often depends on the crystal-size distribution (CSD). The CSD (or
NDF) defines how the population of crystals is distributed over crystal size, and the PBE is
a partial integro-differential equation that defines the evolution of the NDF, as described by
Ramkrishna (2000) and Randolph & Larson (1971). Generally these models consist of a
PBE coupled with spatially homogeneous mass and energy balances, and can be profitably
used in modeling, design, and scale-up of process equipment, and in the development of

1 In the more sophisticated treatments, spatial inhomogeneities are modeled by connected “zones” or regions of

space that are assumed to be homogeneous. The resulting population-balance equation (PBE) does not explicitly
account for local variations in the flow field.

1
2 Introduction

control strategies for particulate systems (Chiu & Christofides, 1999; Crowley et al., 2000;
Nagy, 2009).
However, in many industrial and environmental applications the fluid-dynamic
interactions between the disperse and continuous phases are also very important,
and much research effort has focused on these aspects of the problem (e.g. Delnoij et al.,
1997; Fox, 2012; Gavi et al., 2007; Gerber & Mousavi, 2007; Lain et al., 2002; Laurent &
Massot, 2001; Marchisio & Fox, 2007; Monahan & Fox, 2007; Petitti et al., 2010; Prat &
Ducoste, 2006; Rigopoulos, 2010; Sanyal et al., 2005; Venneker et al., 2002). For exam-
ple, in gas–solid systems the application of the kinetic theory of gases to granular flows and
the development of multiphase computational models have led to a deeper understanding
of these issues (Gidaspow, 1994; Goldhirsch, 2003; Jenkins & Mancini, 1989). Gener-
ally these transport-phenomena-based models consist of spatially inhomogeneous mass,
momentum, and energy (thermal and granular) balances between the disperse phase and the
continuous phase, and spatial dependences are handled by using a finite-volume approach
(Leveque, 2002) in the context of computational fluid dynamics (CFD). Important indus-
trial examples of such flows include fluidized beds and riser flows, slurry-flow reactors,
and bubble columns. In all of these examples, the coupling due to mass, momentum, and
energy exchange between the disperse and continuous phases results in flow dynamics that
are distinctly different than that observed in single-phase flows. Thus, the CFD models
used for describing the fluid dynamics of disperse multiphase flows usually involve multi-
ple continuity and momentum equations that are tightly coupled through phase-interaction
terms (Drew & Passman, 1999; van der Hoef et al., 2008; Ishii, 1975; Portela & Oliemans,
2006). For polydisperse multiphase flows, the situation becomes more complex because it
is necessary to describe the “particle-size” distribution of the disperse phase (De Chaise-
martin et al., 2009; Fan et al., 2004; Fox, 2007; Riber et al., 2009), as well as the coupling
with the continuous phase for particles of different sizes. One of the primary goals of the
present work is to present a systematic modeling framework for accomplishing this task.
In general, the first class of models mentioned above is able to describe the evolution
of the NDF that characterizes the disperse phase. For example, if this approach is applied
to crystallization problems, it is possible to describe the evolution of the CSD in a crystal-
lizer working under certain operating conditions, taking into account all possible physical
and chemical processes such as nucleation, molecular growth, aggregation, and breakage.
However, these models are unable to take into account spatial gradients of properties and
how they relate to the continuous or disperse phases, fluid-dynamic interactions between
crystals and the continuous phase, and interactions among the crystals themselves. As a
consequence, the first class of models is unable to predict the effect of local flow interac-
tions on the evolution of the CSD. On the other hand, the second class of models is capable
of describing detailed fluid-dynamic interactions, but for simplicity often assumes a con-
stant (or even monodisperse) CSD. The principal aim of this book is to present the common
underlying theory and, through the introduction of appropriate computational methods, to
create a bridge between the two approaches. The resulting CFD-based computational mod-
els can then be used to solve a large variety of polydisperse particulate and multiphase flow
problems.

A working definition of polydisperse multiphase flows

For the purposes of this book, it will be useful to have a clear definition of what is meant
by disperse and polydisperse in the context of multiphase flows. By disperse, we mean
that one or more of the phases is composed of clearly identifiable discrete entities such
as solid particles, drops, or bubbles. By polydisperse, we mean that the properties of the
1.2 Two example systems 3

disperse-phase entities can be different for each entity (e.g. particles with different mass,
composition, or temperature). For example, fuel sprays have a region near the nozzle where
the liquid jet is not disperse, followed by a region after breakup of the primary jet that is
composed of individual droplets. The latter region would be considered a polydisperse mul-
tiphase flow and could be modeled using the methods described in this book. In contrast,
bubble-column flows, for example, can be entirely monodisperse because all of the bubbles
have (approximately) the same properties (e.g. bubble diameter). In practice, monodisperse
multiphase flows are relatively rare and, hence, it will be important to have a modeling
framework that naturally accounts for polydispersity.
Another important manifestation of “polydispersity” is the presence of disperse-phase
entities (even those with identical physical properties) with different velocities. The reader
familiar with the kinetic theory of molecular gases will recognize this type of polydis-
persity as leading to the velocity-distribution function, which plays an important role in
the transport theory of polydisperse multiphase flows. As in the molecular kinetic theory,
the mesoscale description of a monodisperse multiphase flow can be formulated in terms
of a kinetic equation for the velocity-distribution function. However, unlike in molecu-
lar gases at standard temperatures and pressures, the disperse-phase entities often interact
infrequently due to collisions, so the standard hydrodynamic approximations valid in the
collision-dominated regime are no longer accurate. In analogy to molecular gases, such
multiphase flows behave as rarefied granular gases wherein processes besides collisions
(e.g. momentum exchange with the continuous phase) are dominant in determining the
flow regime. For this reason, it is often necessary to retain the mesoscale description as the
starting point for describing the disperse-phase flow dynamics.
In summary, a polydisperse multiphase flow consists of one (or more) disperse phases
with entities of possibly different physical properties and velocities. The mesoscale mod-
eling approach (described in more detail in Section 1.3) for describing such flows is the
primary focus of this book. For multiphase flows that are not composed of a clearly distinct
disperse phase, other modeling approaches must be followed (e.g. methods that resolve the
dynamics of the interface separating the phases, or volume-averaging approaches and the
concomitant ad hoc closures of the phase-interaction and transport terms). As will become
clearer to the reader in subsequent chapters, the mesoscale modeling approach allows rig-
orous derivation of the macroscale transport equations, thereby ensuring that the resulting
CFD models will be as accurate as possible when simulating real polydisperse multiphase
flows.

1.2 Two example systems

In this section, we provide two examples of polydisperse multiphase systems. In the first
example, the particles are assumed to be very small so that their inertia is small enough to
be negligible relative to that of the fluid phase. Thus, the particle-size distribution (PSD) is
governed by a PBE and undergoes changes due to aggregation and breakage. In the second
example, the particles are assumed to be large enough to have finite inertia relative to the
surrounding fluid, and we let all particles be identical and consider the distribution of the
particle velocity, which is described by a kinetic equation (KE).

1.2.1 The population-balance equation for fine particles

As a first example, we consider a population of solid particles suspended in a liquid. We
assume that the density and size of the particles are such that they move together with
4 Introduction

the fluid. Local velocity gradients in the fluid, which are generally quantified by the shear
rate, induce particle aggregation and breakage, changing the PSD, as dictated by the cor-
responding PBE. In this context we will assume (as is often the case for colloidal systems)
that aggregation and breakage are completely reversible, which means that, when particles
aggregate, the aggregates formed can later be fragmented by breakage, and that the frag-
ments formed by breakage can, in turn, aggregate together. In what follows we will analyze
the PBE and introduce a set of dimensionless numbers and characteristic time scales that
will turn out to be useful in the investigation of multiphase systems.

Relevant dimensionless numbers and time scales

As described in Chapter 4, the following PBE governs the evolution of the PSD,
representing the state of the solid aggregates:
∂n ∂vf n ∂2 n
+ = Γ 2 + C, (1.1)
∂t ∂x ∂x
where vf is the known fluid velocity with which the particles are advected and Γ is a diffu-
sion coefficient. For simplicity, in this example, only one spatial coordinate x is considered
and the diffusion coefficient Γ is constant. The PSD, denoted by n, is a number-density
function (NDF) representing the number concentration of particles with volumes between
V and V + dV. The PSD is a function of time t, of the spatial coordinate x, and of the
volume of the aggregates V. The aggregates are composed of spherical primary parti-
cles with fixed diameter dp . If just one primary particle is included in the aggregate then
V = πdp3 /6, whereas for an aggregate composed of two primary particles V = πdp3 /3, and
so on for larger aggregates. As described in Chapter 5, the source term due to aggregation
and breakage is
  ∞
1 V
C= a(V − V  , V  )n(V − V  )n(V  )dV  − a(V, V  )n(V)n(V  )dV 
2 0
 ∞ 0

+ b(V  )N(V|V  )n(V  )dV  − b(V)n(V), (1.2)

V

where a(V, V ) and b(V) are the aggregation and breakage kernels, N(V|V  ) is the daughter


distribution function, and the time and space dependences are omitted for clarity. As will
be described in more detail in later chapters, the aggregation and breakage kernels express
the tendency of the particles to aggregate and break. Generally speaking, these kernels
depend strongly on the spatial coordinate x, since real multiphase systems are characterized
by regions where aggregation and breakage occur at very different rates. For example,
particles suspended in a stirred tank might experience intense breakage near the impeller
and strong aggregation in other regions. The daughter distribution function gives instead
the size distribution of particles generated by the breakup of a single aggregate.
In order to highlight the different regimes exhibited by Eq. (1.1) and the different treat-
ments that are more suitable for its solution in each regime, the PBE has to be normalized
and made dimensionless. Different quantities can be used to this end, such as the moments
of order zero and one of the PSD:
 ∞
m0 (t, x) = n(t, x, V)dV,
0 ∞ (1.3)
m1 (t, x) = Vn(t, x, V)dV.
0
1.2 Two example systems 5

The zeroth-order moment identifies the total aggregate number density (i.e. the total num-
ber of aggregates per unit volume), whereas the first-order moment is the disperse-phase
volume fraction, more frequently denoted by αp . For this simple two-phase system, if αf
is used to indicate the fluid-phase volume fraction, by definition αp + αf = 1. By using
these moments, one can define a mean particle volume corresponding to the ratio between
the moments of order one and zero and, in addition, one can define the dimensionless PSD
ψ(ξ) = n (t, x, V) m1 (t, x)/m20 (t, x) with ξ = Vm0 /m1 . The dimensionless PSD has first- and
second-order moments equal to unity, and is useful when searching for the self-similar
solution that characterizes pure aggregation, pure breakage and simultaneous aggregation
and breakage problems (Ramkrishna, 2000). The moment of order two of the PSD,
 ∞
m2 (t, x) = V 2 n(t, x, V)dV, (1.4)
0

is instead useful in the definition of the standard deviation of the PSD, which is equal
to m2 − m21 /m0 . Here the PSD and Eq. (1.1) are made dimensionless by introducing the
characteristic length L, velocity U, aggregate volume ζ, and total aggregate number density
Nt . For example, for particles suspended in a stirred tank, L could be the tank diameter and
U the impeller-tip velocity, whereas ζ and Nt could be extracted by the volume-average
PSD in the vessel at steady state. The last two quantities can in turn be used to define the
characteristic disperse-phase volume fraction: φp = ζNt .
Knowledge of these characteristic quantities suffices for the calculation of the charac-
teristic time scales for aggregation τa = [a(ζ, ζ)Nt ]−1 and breakage τb = [b(ζ)]−1 . These
two time scales represent the average time interval between two subsequent aggregation or
breakage events. When these time scales are very short, aggregation and breakage events
are very frequent, resulting in rapid evolution of the PSD. In fact, as particles move and
diffuse in the domain, many aggregation and breakage events occur, resulting in a quick
adaptation of the PSD to its local “equilibrium” value neq (t, x, V), which is dictated by
the local values of the aggregation and breakage kernels. Here the term equilibrium refers
to the steady state reached when aggregation and breakage counterbalance each other.2
Because of the hypothesis of reversible aggregation and breakage, this equilibrium will
be different from point to point, since it depends on the local values of the aggregation
and breakage kernels. The equilibrium solution corresponds to the resulting PSD, in the
case of particles unable to move away from point x, but evolving in time according to
the local aggregation and breakage kernels at point x. In contrast, when the characteris-
tic time scales are very long, aggregation and breakage events are not very frequent, and
the evolution of the PSD becomes quite slow. Therefore as particles move and diffuse in the
domain, due to the low aggregation and breakage frequencies and the slow changes in the
PSD, only the averaged rates are perceived. The evolution of the PSD is then determined
by the volume-average kernels, rather than by their local values.
Closer observation of the aggregation time scale shows that it is inversely proportional
to the characteristic total aggregate number density Nt , which is related to the character-
istic disperse-phase volume fraction φp . When the system is dilute (i.e. φp 1), Nt is
2 Depending on the functional form of the aggregation and breakage kernels, different scenarios are possible.

When aggregation prevails, the system can undergo extensive aggregation. Since when particles aggregate their
number is reduced, the gelling limit, where the total number of particles tends to zero, can be reached. In contrast,
when breakage prevails, the total number of particles can go to infinity, giving the opposite limit known as
shattering. When aggregation and breakage counterbalance each other, a population of particles can evolve to a
steady state. This steady state can be calculated by setting the collision source term in Eq. (1.2) equal to zero. We
refer to this steady-state solution as the equilibrium PSD.
6 Introduction

small and as a consequence the aggregation time scale is large. Under these conditions
aggregation events are separated by long time intervals. In fact, for dilute systems with
very few particles, longer time intervals are necessary to observe a collision between two
particles. In contrast, as the system becomes more concentrated (i.e. as Nt and φp increase)
the time interval between two aggregation events becomes very short, since as more par-
ticles are suspended in the fluid more collisions (and aggregation events) will occur. The
characteristic disperse-phase volume fraction also affects the breakage time scales; in fact,
as aggregation events become more frequent the resulting characteristic aggregate volume
ζ increases. Since generally the breakage kernel increases with increasing particle size (i.e.
larger aggregates are more likely to break than smaller ones), the breakage time scale also
becomes shorter. In summary, dilute systems are characterized by large aggregation and
breakage time scales, whereas for dense systems these time scales are very short.
A more quantitative definition of dilute and dense systems is possible only by analyzing
the normalized and dimensionless PBE. By letting ξ = V/ζ, n∗ = nζ/Nt , x∗ = x/L, v∗f =
vf /U, and t∗ = tΓ/L2 the following dimensionless PBE is obtained:
  ξ
∂n∗ ∂v∗f n∗ ∂2 n∗ 1
+ Pe − ∗2 = Daa a∗ (ξ − ξ , ξ )n∗ (ξ − ξ )n∗ (ξ )dξ
∂t∗ ∂x∗ ∂x 2 0
 +∞ 
∗  ∗ ∗  
− a (ξ, ξ )n (ξ)n (ξ )dξ
0
 +∞ 
∗   ∗   ∗ ∗
+ Dab b (ξ )N(ξ|ξ )n (ξ )dξ − b (ξ)n (ξ) , (1.5)
ξ

where the normalized aggregation and breakage kernels are given by a∗ (ξ, ξ ) = a(V, V  )/
a(ζ, ζ) and b∗ (ξ) = b(V)/b(ζ). The Péclet number, defined by
UL L2 /Γ
Pe = = , (1.6)
Γ L/U
is the ratio between the particle-diffusion and particle-advection time scales. The aggrega-
tion Damköhler number, defined by
a(ζ, ζ)L2 Nt L2 /Γ
Daa = = , (1.7)
Γ 1/(a(ζ, ζ)Nt )
is the ratio between the particle-diffusion and particle-aggregation time scales. The
breakage Damköhler number, defined by
b(ζ)L2 L2 /Γ
Dab = = , (1.8)
Γ 1/b(ζ)
is the ratio between the particle-diffusion and breakage time scales. Depending on the
values of Pe, Daa , and Dab different regimes can be identified. For example, when Pe
becomes much smaller than unity, diffusive particle transport becomes more important than
advective transport. Under this condition, different scenarios are still possible, depending
on the values of Daa and Dab . When both numbers are much smaller than unity, diffusion
is much faster than aggregation and breakage, whereas when both are much greater than
unity aggregation and breakage are faster than particle diffusion. It is important to recall
here that particle diffusion tends to homogenize the PSD in the spatial domain, whereas,
since kernels generally contain some form of spatial dependence, aggregation and breakage
tend to create gradients in the PSD.
1.2 Two example systems 7

Major operating regimes

From the discussion above, three major operating regimes can be identified from Eq. (1.5).

(i) Under dilute conditions such that φp 1, aggregation and breakage are slower
than particle diffusion, or, in other words, Daa 1 and Dab 1. The evolution
of the PSD is not controlled by the local values of the aggregation and breakage
kernels, but rather by their volume-average values. Under dilute conditions, since
diffusion mixes particles faster than the kernels let them aggregate and break, the
two-phase system is generally considered as spatially homogeneous or well mixed
(i.e. the PSDs at different points of the domain are identical). The evolution of the
PSD for a well-mixed system n̂(t, V) is governed by a volume-average PBE:
  ∞
∂n̂ 1 V
= â(V − V  , V  )n̂(V − V  )n̂(V  )dV  − â(V, V  )n̂(V)n̂(V  )dV 
∂t 2
0 ∞ 0
   
+ b̂(V )N(V|V )n̂(V )dV − b̂(V)n̂(V), (1.9)
V

where â and b̂ are the volume-average aggregation and breakage kernels, which,
due to their non-linear dependence on the spatial coordinates, are generally quite
different from the local kernels a and b. A system is generally considered dilute (and
therefore well mixed) when Daa and Dab are smaller than 10−2 . Depending on the
type of problem and the functional form of the aggregation and breakage kernels,
these conditions result in different constraints for the characteristic disperse-phase
volume fraction. For most applications, it is reasonable to consider the operating
regime as dilute when φp < 10−3 .

(ii) Under moderately dense conditions, Daa and Dab are greater than 10−2 but smaller
than 102 . The characteristic time scales for aggregation and breakage are then com-
parable to that of particle diffusion. Under these conditions the system cannot be
considered to be well mixed, and the PSD is subjected to large variations across
the spatial domain. As a consequence, the movement of particles in physical space
and their aggregation and breakage cannot be decoupled, and the evolution of the
two-phase system is found by solving Eq. (1.5).

(iii) Under dense conditions Daa and Dab are both greater than 102 , and hence aggre-
gation and breakage are very fast (or almost instantaneous) compared with particle
diffusion. Thanks to the hypothesis of reversibility of aggregation and breakage, as
particles diffuse in the domain, they instantaneously adapt to the local equilibrium
solution, i.e. neq . In this case the evolution of the PSD is completely determined by
the local aggregation and breakage kernels and the resulting local steady state.

In summary, for fine particles undergoing aggregation and breakage different modeling
approaches should be used for each of the regimes described above. Under dilute con-
ditions, the system can be approximated as a well-mixed system, since spatial gradients
in the PSD are not relevant. The evolution of the system is governed by Eq. (1.9) using
the volume-average aggregation and breakage kernels. Under moderately dense condi-
tions, spatial gradients of the PSD are not negligible and therefore the evolution of the
system is governed by Eq. (1.5). Under dense conditions the system evolves according to
the equilibrium solution found by setting the right-hand side of Eq. (1.2) equal to zero.
8 Introduction

1.2.2 The kinetic equation for gas–particle flow

As a second example, we consider the kinetic equation (KE) for monodisperse, isothermal
solid particles suspended in a constant-density gas phase. For clarity, we assume that the
particle material density is significantly larger than that of the gas so that only the fluid
drag and buoyancy terms are needed to account for momentum exchange between the two
phases (Maxey & Riley, 1983). In this example, the particles are large enough to have finite
inertia and thus they evolve with a velocity that can be quite different than that of the gas
phase.

Relevant dimensionless numbers

As in all mathematical descriptions of transport phenomena, the theory of polydisperse
multiphase flows introduces a set of dimensionless numbers that are pertinent in describ-
ing the behavior of the flow. Depending on the complexity of the flow (e.g. variations in
physical properties due to chemical reactions, collisions, etc.), the set of dimensionless
numbers can be quite large. (Details on the physical models for momentum exchange are
given in Chapter 5.) As will be described in detail in Chapter 4, a kinetic equation can
be derived for the number-density function (NDF) of the velocity of the disperse phase
n(t, x, v). Also in this example, for clarity, we will assume that the problem has only one
particle velocity component v and is one-dimensional in physical space with coordinate x
at time t. Furthermore, we will assume that the NDF has been normalized (by multiplying
it by the volume of a particle) such that the first three velocity moments are
 +∞
αp ≡ n dv,
−∞
 +∞
αp Up ≡ vn dv, (1.10)
−∞
 +∞
αp Θp + αp Up2 ≡ v2 n dv,
−∞

where αp , Up , and Θp are the volume fraction, mean velocity, and granular temperature of
the disperse phase, respectively. The example kinetic equation is then given by

∂n ∂vn ∂An
+ + = C, (1.11)
∂t ∂x ∂v
where the acceleration (due to buoyancy, gravity, and drag) and collision terms are defined,
respectively, by

1 ∂pg 1
A≡− + g x + (Ug − v),
ρp ∂x τp
(1.12)
1
C ≡ (neq − n).
τc

In these expressions, g x is the x-component of the gravity force, ρp is the particle density,
and pg is the gas-phase pressure. τp and τc are characteristic time scales for drag and
collisions, respectively. Ug is the velocity of the gas phase, and neq is the equilibrium
1.2 Two example systems 9

distribution. For simplicity, we have used a linear collision model3 in which the first two
moments of neq are the same as those of n, and the third is
 +∞
1 + e2
αp Θp + αp Up2 = v2 neq dv, (1.13)
2 −∞

and 0 ≤ e ≤ 1 is the coefficient of restitution for particle–particle collisions. For elastic

collisions, e = 1, whereas e < 1 corresponds to inelastic collisions.
The disperse-phase kinetic equation is coupled to the continuity and momentum
equations for the continuous phase, which are given, respectively, by

∂αg ∂(αg Ug )
+ = 0,
∂t ∂x
(1.14)
∂(αg Ug ) ∂(αg Ug Ug ) ∂2 Ug αg ∂pg ρp αp
+ = αg νg − + αg g x + (Up − Ug ),
∂t ∂x ∂x2 ρg ∂x ρg τp

where αg , ρg , and νg are the gas-phase volume fraction, density, and kinematic viscosity,4
respectively. By definition, αp + αg = 1. In order to make Eqs. (1.14) dimensionless, we
introduce a characteristic gas-phase velocity U and characteristic length L. The dimen-
sionless time is then t∗ = tU/L. On noting that αg is dimensionless and letting x∗ = x/L,
Ug∗ = Ug /U, and Up∗ = Up /U, we then have

∂αg ∂αg Ug∗

+ = 0,
∂t∗ ∂x∗
∂(αg Ug∗ ) ∂(αg Ug∗ Ug∗ ) (1.15)
αg ∂2 Ug∗ ∂p∗g αg ∗ αg φ2 ∗
+ = − αg + g + (Up − Ug∗ ),
∂t∗ ∂x∗ Reg (∂x∗ )2 ∂x∗ Fr2g x St

where p∗g = p/(ρg U 2 ) and g∗x = g x /g. The gas-phase Reynolds and Froude numbers are
defined, respectively, by Reg = UL/νg and Frg = U/(Lg)1/2 , where g is the gravitational
acceleration constant. The Stokes number is defined by St = Uτp /L, and the phase-mass
ratio by φ2 = αp ρp /(αg ρg ). The latter is an important parameter insofar as it determines
whether the gas phase “sees” the disperse phase (i.e. momentum coupling is negligible
when φ2 1).
The kinetic equation given by Eq. (1.11) is made dimensionless by defining v∗ = v/U
and n∗ = Un:
∂n∗ ∂(v∗ n∗ ) ∂(A∗ n∗ )
+ + = C∗ , (1.16)
∂t∗ ∂x∗ ∂v∗
where n∗ (t∗ , x∗ , v∗ ) is the dimensionless velocity distribution function,5 and
∗
1 ∂pg g∗ 1
A∗ ≡ − ∗
+ x2 + (Ug∗ − v∗ ),
φ1 ∂x Frg St
(1.17)
φ3 ∗
C∗ ≡ (n − n).
Knp
3 See Chapter 6 for more details on collision models.
4 For a 1D flow the viscosity term would normally be zero. However, we include the viscosity term as a
placeholder for the fully 3D case in order to show the Reynolds-number dependence.

5 Note that n and n are made dimensionless in a manner such that α = n∗ dv∗ . Because the kinetic equation
eq p
is linear in n, no new dimensionless numbers are generated by this process.
10 Introduction

The phase-density ratio is defined by φ1 = ρp /ρg and, since φ1 1, the buoyancy term
is negligible. The two new dimensionless numbers generated in this process are the phase-
velocity ratio φ3 = Up† /U,6 and the disperse-phase Knudsen number Knp = Up† τc /L. In
addition, the dimensionless form of Eqs. (1.10),
 +∞
αp = n∗ dv∗ ,
−∞
 +∞
∗
αp Up = v∗ n∗ dv∗ , (1.18)
−∞
 +∞
φ23
αp + αp (Up∗ )2 = (v∗ )2 n∗ dv∗ ,
Ma2p −∞

introduces the disperse-phase Mach number Map = Up† /Θ1/2 p . By analogy to compressible
gas flows, Map is the ratio of the characteristic mean particle velocity Up† and the speed of
“sound” (Θ1/2
p ) in the disperse phase. Thus for Map 1 transport in the disperse phase
is predominantely due to velocity fluctuations, whereas for Map 1 it is due to mean
advection.
Following the convention used in gas dynamics, we will define the velocity ratio and
disperse-phase Knudsen number differently according to whether the disperse phase is
subsonic or supersonic:

Θ1/2
p Θ1/2
p τc π1/2 dp
for Map ≤ 1, φ3 = , Knp = = ;
U L 12αp g0 L
(1.19)
Up† Up† τc π1/2 dp
for Map ≥ 1, φ3 = , Knp = = Map .
U L 12αp g0 L

Here dp is the particle diameter,7 g0 (αp /α∗p ) is the radial distribution function, and α∗p ≈
0.63 is the maximum volume fraction at close packing. For αp α∗p , g0 ≈ 1. However, as
∗
αp approaches αp , g0 diverges to infinity.
Note that the ratio φ3 will depend strongly on the Stokes number,

Uτp ρp dp2 U
St = = , (1.20)
L 18ρg νg L

where this expression for τp is valid for small particle Reynolds numbers, defined by

dp |Up − Ug |
Rep ≡ . (1.21)
νg

When St 1 the kinetic equation will be uncoupled from the gas phase and the disperse
phase will behave as a granular gas. In the opposite limit where St 1, Θp ≈ 0 and
Up† ≈ U, so that the disperse-phase Mach number will be very large and φ3 ≈ 1. At
intermediate values of the Stokes number, a rich variety of flow phenomena depending on
all the values of the dimensionless parameters can be observed.
6 Because φ and Kn appear together in Eq. (1.17), they could be combined into one dimensionless number.
3 p
We keep them separate so that Knp and Map depend only on disperse-phase variables.
7 At fixed volume fraction α , reducing d increases the collision rate because the total number concentration
p p
of particles N increases significantly.
 Another Random Document on
Scribd Without Any Related Topics
tears, prevailed. After a brief struggle between the sentiment of love
on the one hand and those of ambition and of duty combined on the
other, Antony gave up the contest. Abandoning every thing else, he
surrendered himself wholly to Cleopatra’s control, and went with her
to Alexandria. He spent the winter there, giving himself up with her
to every species of sensual indulgence that the most remorseless
license could tolerate, and the most unbounded wealth procure.
There seemed, in fact, to be no bounds to the extravagance and
infatuation which Antony displayed during the winter in Alexandria.
Cleopatra devoted herself to him incessantly, day and night, filling up
every moment of time with some new form of pleasure, in order that
he might have no time to think of his absent wife, or to listen to the
reproaches of his conscience. Antony, on his part, surrendered
himself a willing victim to these wiles, and entered with all his heart
into the thousand plans of gayety and merry-making which
Cleopatra devised. They had each a separate establishment in the
city, which was maintained at an enormous cost, and they made a
regular arrangement by which each was the guest of the other on
alternate days. These visits were spent in games, sports, spectacles,
feasting, drinking, and in every species of riot, irregularity, and
excess.
A curious instance is afforded of the accidental manner in which
intelligence in respect to the scenes and incidents of private life in
those ancient days is sometimes obtained, in a circumstance which
occurred at this time at Antony’s court. It seems that there was a
young medical student at Alexandria that winter, named Philotas,
who happened, in some way or other, to have formed an
acquaintance with one of Antony’s domestics, a cook. Under the
guidance of this cook, Philotas went one day into the palace to see
what was to be seen. The cook took his friend into the kitchens,
where, to Philotas’s great surprise, he saw, among an infinite
number and variety of other preparations, eight wild boars roasting
before the fires, some being more and some less advanced in the
process. Philotas asked what great company was to dine there that
day. The cook smiled at this question, and replied that there was to
be no company at all, other than Antony’s ordinary party. “But,” said
the cook, in explanation, “we are obliged always to prepare several
suppers, and to have them ready in succession at different hours, for
no one can tell at what time they will order the entertainment to be
served. Sometimes, when the supper has been actually carried in,
Antony and Cleopatra will get engaged in some new turn of their
diversions, and conclude not to sit down just then to the table, and
so we have to take the supper away, and presently bring in another.”
Antony had a son with him at Alexandria at this time, the child of his
wife Fulvia. The name of the son, as well as that of the father, was
Antony. He was old enough to feel some sense of shame at his
father’s dereliction from duty, and to manifest some respectful
regard for the rights and the honor of his mother. Instead of this,
however, he imitated his father’s example, and, in his own way, was
as reckless and as extravagant as he. The same Philotas who is
above referred to was, after a time, appointed to some office or
other in the young Antony’s household, so that he was accustomed
to sit at his table and share in his convivial enjoyments. He relates
that once, while they were feasting together, there was a guest
present, a physician, who was a very vain and conceited man, and
so talkative that no one else had any opportunity to speak. All the
pleasure of conversation was spoiled by his excessive garrulity.
Philotas, however, at length puzzled him so completely with a
question of logic—of a kind similar to those often discussed with
great interest in ancient days—as to silence him for a time; and
young Antony was so much delighted with this feat, that he gave
Philotas all the gold and silver plate that there was upon the table,
and sent all the articles home to him, after the entertainment was
over, telling him to put his mark and stamp upon them, and lock
them up.
The question with which Philotas puzzled the self-conceited
physician was this. It must be premised, however, that in those days
it was considered that cold water in an intermittent fever was
extremely dangerous, except in some peculiar cases, and in those
the effect was good. Philotas then argued as follows: “In cases of a
certain kind it is best to give water to a patient in an ague. All cases
of ague are cases of a certain kind. Therefore it is best in all cases to
give the patient water.” Philotas having propounded his argument in
this way, challenged the physician to point out the fallacy of it; and
while the physician sat perplexed and puzzled in his attempts to
unravel the intricacy of it, the company enjoyed a temporary respite
from his excessive loquacity.
Philotas adds, in his account of this affair, that he sent the gold and
silver plate back to young Antony again, being afraid to keep them.
Antony said that perhaps it was as well that this should be done,
since many of the vessels were of great value on account of their
rare and antique workmanship, and his father might possibly miss
them and wish to know what had become of them.
As there were no limits, on the one hand, to the loftiness and
grandeur of the pleasures to which Antony and Cleopatra addicted
themselves, so there were none to the low and debasing tendencies
which characterized them on the other. Sometimes, at midnight,
after having been spending many hours in mirth and revelry in the
palace, Antony would disguise himself in the dress of a slave, and
sally forth into the streets, excited with wine, in search of
adventures. In many cases, Cleopatra herself, similarly disguised,
would go out with him. On these excursions Antony would take
pleasure in involving himself in all sorts of difficulties and dangers—
in street riots, drunken brawls, and desperate quarrels with the
populace—all for Cleopatra’s amusement and his own. Stories of
these adventures would circulate afterward among the people, some
of whom would admire the free and jovial character of their
eccentric visitor, and others would despise him as a prince degrading
himself to the level of a brute.
Some of the amusements and pleasures which Antony and Cleopatra
pursued were innocent in themselves, though wholly unworthy to be
made the serious business of life by personages on whom such
exalted duties rightfully devolved. They made various excursions
upon the Nile, and arranged parties of pleasure to go out on the
water in the harbor, and to various rural retreats in the environs of
the city. Once they went out on a fishing-party, in boats, in the port.
Antony was unsuccessful; and feeling chagrined that Cleopatra
should witness his ill luck, he made a secret arrangement with some
of the fishermen to dive down, where they could do so unobserved,
and fasten fishes to his hook under the water. By this plan he caught
very large and fine fish very fast. Cleopatra, however, was too wary
to be easily deceived by such a stratagem as this. She observed the
maneuver, but pretended not to observe it; she expressed, on the
other hand, the greatest surprise and delight at Antony’s good luck,
and the extraordinary skill which it indicated.
The next day she wished to go a fishing again, and a party was
accordingly made as on the day before. She had, however, secretly
instructed another fisherman to procure a dried and salted fish from
the market, and, watching his opportunity, to get down into the
water under the boats and attach it to the hook, before Antony’s
divers could get there. This plan succeeded, and Antony, in the midst
of a large and gay party that were looking on, pulled out an
excellent fish, cured and dried, such as was known to every one as
an imported article, bought in the market. It was a fish of a kind that
was brought originally from Asia Minor. The boats, and the water all
around them, resounded with the shouts of merriment and laughter
which this incident occasioned.

In the mean time, while Antony was thus spending his time in low
and ignoble pursuits and in guilty pleasures at Alexandria, his wife
Fulvia, after exhausting all other means of inducing her husband to
return to her, became desperate, and took measures for fomenting
an open war, which she thought would compel him to return. The
extraordinary energy, influence, and talent which Fulvia possessed,
enabled her to do this in an effectual manner. She organized an
army, formed a camp, placed herself at the head of the troops, and
sent such tidings to Antony of the dangers which threatened his
cause as greatly alarmed him. At the same time news came of great
disasters in Asia Minor, and of alarming insurrections among the
provinces which had been committed to his charge there. Antony
saw that he must arouse himself from the spell which had enchanted
him and break away from Cleopatra, or that he would be wholly and
irretrievably ruined. He made, accordingly, a desperate effort to get
free. He bade the queen farewell, embarked hastily in a fleet of
galleys, and sailed away to Tyre, leaving Cleopatra in her palace,
vexed, disappointed, and chagrined.
 Chapter XI.

The Battle of Actium.

C LEOPATRA, in parting with Antony as described in the last

chapter, lost him for two or three years. During this time Antony
himself was involved in a great variety of difficulties and dangers,
and passed through many eventful scenes, which, however, can not
here be described in detail. His life, during this period, was full of
vicissitude and excitement, and was spent probably in alternations of
remorse for the past and anxiety for the future. On landing at Tyre,
he was at first extremely perplexed whether to go to Asia Minor or to
Rome. His presence was imperiously demanded in both places. The
war which Fulvia had fomented was caused, in part, by the rivalry of
Octavius, and the collision of his interests with those of her husband.
Antony was very angry with her for having managed his affairs in
such a way as to bring about a war. After a time Antony and Fulvia
met at Athens. Fulvia had retreated to that city, and was very
seriously sick there, either from bodily disease, or from the influence
of long-continued anxiety, vexation, and distress. They had a stormy
meeting. Neither party was disposed to exercise any mercy toward
the other. Antony left his wife rudely and roughly, after loading her
with reproaches. A short time afterward, she sank down in sorrow to
the grave.
The death of Fulvia was an event which proved to be of advantage
to Antony. It opened the way to a reconciliation between him and
Octavius. Fulvia had been extremely active in opposing Octavius’s
designs, and in organizing plans for resisting him. He felt, therefore,
a special hostility against her, and, through her, against Antony. Now,
however, that she was dead, the way seemed to be in some sense
opened for a reconciliation.
Octavius had a sister, Octavia, who had been the wife of a Roman
general named Marcellus. She was a very beautiful and a very
accomplished woman, and of a spirit very different from that of
Fulvia. She was gentle, affectionate, and kind, a lover of peace and
harmony, and not at all disposed, like Fulvia, to assert and maintain
her influence over others by an overbearing and violent demeanor.
Octavia’s husband died about this time, and, in the course of the
movements and negotiations between Antony and Octavius, the plan
was proposed of a marriage between Antony and Octavia, which, it
was thought, would ratify and confirm the reconciliation. This
proposal was finally agreed upon. Antony was glad to find so easy a
mode of settling his difficulties. The people of Rome, too, and the
authorities there, knowing that the peace of the world depended
upon the terms on which these two men stood with regard to each
other, were extremely desirous that this arrangement should be
carried into effect. There was a law of the commonwealth forbidding
the marriage of a widow within a specified period after the death of
her husband. That period had not, in Octavia’s case, yet expired.
There was, however, so strong a desire that no obstacle should be
allowed to prevent this proposed union, or even to occasion delay,
that the law was altered expressly for this case, and Antony and
Octavia were married. The empire was divided between Octavius
and Antony, Octavius receiving the western portion as his share,
while the eastern was assigned to Antony.
It is not probable that Antony felt any very strong affection for his
new wife, beautiful and gentle as she was. A man, in fact, who had
led such a life as his had been, must have become by this time
incapable of any strong and pure attachment. He, however, was
pleased with the novelty of his acquisition, and seemed to forget for
a time the loss of Cleopatra. He remained with Octavia a year. After
that he went away on certain military enterprises which kept him
some time from her. He returned again, and again he went away. All
this time Octavia’s influence over him and over her brother was of
the most salutary and excellent character. She soothed their
animosities, quieted their suspicions and jealousies, and at one time,
when they were on the brink of open war, she effected a
reconciliation between them by the most courageous and energetic,
and at the same time, gentle and unassuming efforts. At the time of
this danger she was with her husband in Greece; but she persuaded
him to send her to her brother at Rome, saying that she was
confident that she could arrange a settlement of the difficulties
impending. Antony allowed her to go. She proceeded to Rome, and
procured an interview with her brother in the presence of his two
principal officers of state. Here she pleaded her husband’s cause
with tears in her eyes; she defended his conduct, explained what
seemed to be against him, and entreated her brother not to take
such a course as should cast her down from being the happiest of
women to being the most miserable. “Consider the circumstances of
my case,” said she. “The eyes of the world are upon me. Of the two
most powerful men in the world, I am the wife of one and the sister
of another. If you allow rash counsels to go on and war to ensue, I
am hopelessly ruined; for, whichever is conquered, my husband or
my brother, my own happiness will be for ever gone.”
Octavius sincerely loved his sister, and he was so far softened by her
entreaties that he consented to appoint an interview with Antony in
order to see if their difficulties could be settled. This interview was
accordingly held. The two generals came to a river, where, at the
opposite banks, each embarked in a boat, and, being rowed out
toward each other, they met in the middle of the stream. A
conference ensued, at which all the questions at issue were, for a
time at least, very happily arranged.
Antony, however, after a time, began to become tired of his wife,
and to sigh for Cleopatra once more. He left Octavia at Rome and
proceeded to the eastward, under pretense of attending to the
affairs of that portion of the empire; but, instead of doing this, he
went to Alexandria, and there renewed again his former intimacy
with the Egyptian queen.
Octavius was very indignant at this. His former hostility to Antony,
which had been in a measure appeased by the kind influence of
Octavia, now broke forth anew, and was heightened by the feeling
of resentment naturally awakened by his sister’s wrongs. Public
sentiment in Rome, too, was setting very strongly against Antony.
Lampoons were written against him to ridicule him and Cleopatra,
and the most decided censures were passed upon his conduct.
Octavia was universally beloved, and the sympathy which was every
where felt for her increased and heightened very much the popular
indignation which was felt against the man who could wrong so
deeply such sweetness, and gentleness, and affectionate fidelity as
hers.
After remaining for some time in Alexandria, and renewing his
connection and intimacy with Cleopatra, Antony went away again,
crossing the sea into Asia, with the intention of prosecuting certain
military undertakings there which imperiously demanded his
attention. His plan was to return as soon as possible to Egypt after
the object of his expedition should be accomplished. He found,
however, that he could not bear even a temporary absence from
Cleopatra. His mind dwelled so much upon her, and upon the
pleasures which he had enjoyed with her in Egypt, and he longed so
much to see her again, that he was wholly unfit for the discharge of
his duties in the camp. He became timid, inefficient, and remiss, and
almost every thing that he undertook ended disastrously. The army,
who understood perfectly well the reason of their commander’s
remissness and consequent ill fortune, were extremely indignant at
his conduct, and the camp was filled with suppressed murmurs and
complaints. Antony, however, like other persons in his situation, was
blind to all these indications of dissatisfaction; probably he would
have disregarded them if he had observed them. At length, finding
that he could bear his absence from his mistress no longer, he set
out to march across the country, in the depth of the winter, to the
sea-shore, to a point where he had sent for Cleopatra to come to
join him. The army endured incredible hardships and exposures in
this march. When Antony had once commenced the journey, he was
so impatient to get forward that he compelled his troops to advance
with a rapidity greater than their strength would bear. They were,
besides, not provided with proper tents or with proper supplies of
provision. They were often obliged, therefore, after a long and
fatiguing march during the day, to bivouac at night in the open air
among the mountains, with scanty means of appeasing their hunger,
and very little shelter from the cold rain, or from the storms of
driving snow. Eight thousand men died on this march, from cold,
fatigue, and exposure; a greater sacrifice, perhaps, than had ever
been made before to the mere ardor and impatience of a lover.
When Antony reached the shore, he advanced to a certain sea-port,
near Sidon, where Cleopatra was to land. At the time of his arrival
but a small part of his army was left, and the few men that survived
were in a miserably destitute condition. Antony’s eagerness to see
Cleopatra became more and more excited as the time drew nigh.
She did not come so soon as he had expected, and during the delay
he seemed to pine away under the influence of love and sorrow. He
was silent, absent-minded, and sad. He had no thoughts for any
thing but the coming of Cleopatra, and felt no interest in any other
plans. He watched for her incessantly, and would sometimes leave
his place at the table, in the midst of the supper, and go down alone
to the shore, where he would stand gazing out upon the sea, and
saying mournfully to himself, “Why does not she come?” The
animosity and the ridicule which these things awakened against him,
on the part of the army, were extreme; but he was so utterly
infatuated that he disregarded all the manifestations of public
sentiment around him, and continued to allow his mind to be wholly
engrossed with the single idea of Cleopatra’s coming.
She arrived at last. She brought a great supply of clothes and other
necessaries for the use of Antony’s army, so that her coming not
only gratified his love, but afforded him, also, a very essential relief,
in respect to the military difficulties in which he was involved.
After some time spent in the enjoyment of the pleasure which being
thus reunited to Cleopatra afforded him, Antony began again to
think of the affairs of his government, which every month more and
more imperiously demanded his attention. He began to receive
urgent calls from various quarters, urging him to action. In the mean
time, Octavia—who had been all this while waiting in distress and
anxiety at Rome, hearing continually the most gloomy accounts of
her husband’s affairs, and the most humiliating tidings in respect to
his infatuated devotion to Cleopatra—resolved to make one more
effort to save him. She interceded with her brother to allow her to
raise troops and to collect supplies, and then proceed to the
eastward to re-enforce him. Octavius consented to this. He, in fact,
assisted Octavia in making her preparations. It is said, however, that
he was influenced in this plan by his confident belief that this noble
attempt of his sister to reclaim her husband would fail, and that, by
the failure of it, Antony would be put in the wrong, in the estimation
of the Roman people, more absolutely and hopelessly than ever, and
that the way would thus be prepared for his complete and final
destruction.
Octavia was rejoiced to obtain her brother’s aid to her undertaking,
whatever the motive might be which induced him to afford it. She
accordingly levied a considerable body of troops, raised a large sum
of money, provided clothes, and tents, and military stores for the
army; and when all was ready, she left Italy and put to sea, having
previously dispatched a messenger to her husband to inform him
that she was coming.
Cleopatra began now to be afraid that she was to lose Antony again,
and she at once began to resort to the usual artifices employed in
such cases, in order to retain her power over him. She said nothing,
but assumed the appearance of one pining under the influence of
some secret suffering or sorrow. She contrived to be often surprised
in tears. In such cases she would hastily brush her tears away, and
assume a countenance of smiles and good humor, as if making every
effort to be happy, though really oppressed with a heavy burden of
anxiety and grief. When Antony was near her she would seem
overjoyed at his presence, and gaze upon him with an expression of
the most devoted fondness. When absent from him, she spent her
time alone, always silent and dejected, and often in tears; and she
took care that the secret sorrows and sufferings that she endured
should be duly made known to Antony, and that he should
understand that they were all occasioned by her love for him, and by
the danger which she apprehended that he was about to leave her.
The friends and secret agents of Cleopatra, who reported these
things to Antony, made, moreover, direct representations to him, for
the purpose of inclining his mind in her favor. They had, in fact, the
astonishing audacity to argue that Cleopatra’s claims upon Antony
for a continuance of his love were paramount to those of Octavia.
She, that is, Octavia, had been his wife, they said, only for a very
short time. Cleopatra had been most devotedly attached to him for
many years. Octavia was married to him, they alleged, not under the
impulse of love, but from political considerations alone, to please her
brother, and to ratify and confirm a political league made with him.
Cleopatra, on the other hand, had given herself up to him in the
most absolute and unconditional manner, under the influence solely
of a personal affection which she could not control. She had
surrendered and sacrificed every thing to him. For him she had lost
her good name, alienated the affections of her subjects, made
herself the object of reproach and censure to all mankind, and now
she had left her native land to come and join him in his adverse
fortunes. Considering how much she had done, and suffered, and
sacrificed for his sake, it would be extreme and unjustifiable cruelty
in him to forsake her now. She never would survive such an
abandonment. Her whole soul was so wrapped up in him, that she
would pine away and die if he were now to forsake her.
Antony was distressed and agitated beyond measure by the
entanglements in which he found that he was involved. His duty, his
inclination perhaps, certainly his ambition, and every dictate of
prudence and policy, required that he should break away from these
snares at once and go to meet Octavia. But the spell that bound him
was too mighty to be dissolved. He yielded to Cleopatra’s sorrows
and tears. He dispatched a messenger to Octavia, who had by this
time reached Athens, in Greece, directing her not to come any
farther. Octavia, who seemed incapable of resentment or anger
against her husband, sent back to ask what she should do with the
troops, and money, and the military stores which she was bringing.
Antony directed her to leave them in Greece. Octavia did so, and
mournfully returned to her home.
As soon as she arrived at Rome, Octavius, her brother, whose
indignation was now thoroughly aroused at the baseness of Antony,
sent to his sister to say that she must leave Antony’s house and
come to him. A proper self-respect, he said, forbade her remaining
any longer under the roof of such a man. Octavia replied that she
would not leave her husband’s house. That house was her post of
duty, whatever her husband might do, and there she would remain.
She accordingly retired within the precincts of her old home, and
devoted herself in patient and uncomplaining sorrow to the care of
the family and the children. Among these children was one young
son of Antony’s, born during his marriage with her predecessor
Fulvia. In the mean time, while Octavia was thus faithfully though
mournfully fulfilling her duties as wife and mother, in her husband’s
house at Rome, Antony himself had gone with Cleopatra to
Alexandria, and was abandoning himself once more to a life of guilty
pleasure there. The greatness of mind which this beautiful and
devoted wife thus displayed, attracted the admiration of all mankind.
It produced, however, one other effect, which Octavia must have
greatly deprecated. It aroused a strong and universal feeling of
indignation against the unworthy object toward whom this
extraordinary magnanimity was displayed.
In the mean time, Antony gave himself up wholly to Cleopatra’s
influence and control, and managed all the affairs of the Roman
empire in the East in the way best fitted to promote her
aggrandizement and honor. He made Alexandria his capital,
celebrated triumphs there, arranged ostentatious expeditions into
Asia and Syria with Cleopatra and her train, gave her whole
provinces as presents, and exalted her two sons, Alexander and
Ptolemy, children born during the period of his first acquaintance
with her, to positions of the highest rank and station, as his own
acknowledged sons. The consequences of these and similar
measures at Rome were fatal to Antony’s character and standing.
Octavius reported every thing to the Roman senate and people, and
made Antony’s misgovernment and his various misdemeanors the
ground of the heaviest accusations against him. Antony, hearing of
these things, sent his agents to Rome and made accusations against
Octavius; but these counter accusations were of no avail. Public
sentiment was very strong and decided against him at the capital,
and Octavius began to prepare for war.
Antony perceived that he must prepare to defend himself. Cleopatra
entered into the plans which he formed for this purpose with great
ardor. Antony began to levy troops, and collect and equip galleys
and ships of war, and to make requisitions of money and military
stores from all the eastern provinces and kingdoms. Cleopatra put all
the resources of Egypt at his disposal. She furnished him with
immense sums of money, and with an inexhaustible supply of corn,
which she procured for this purpose from her dominions in the valley
of the Nile. The various divisions of the immense armament which
was thus provided for were ordered to rendezvous at Ephesus,
where Antony and Cleopatra were awaiting to receive them, having
proceeded there when their arrangements in Egypt were completed,
and they were ready to commence the campaign.
When all was ready for the expedition to set sail from Ephesus, it
was Antony’s judgment that it would be best for Cleopatra to return
to Egypt, and leave him to go forth with the fleet to meet Octavius
alone. Cleopatra was, however, determined not to go away. She did
not dare to leave Antony at all to himself, for fear that in some way
or other a peace would be effected between himself and Octavius,
which would result in his returning to Octavia and abandoning her.
She accordingly contrived to persuade Antony to retain her with him,
by bribing his chief counselor to advise him to do so. His counselor’s
name was Canidius. Canidius, having received Cleopatra’s money,
while yet he pretended to be wholly disinterested in his advice,
represented to Antony that it would not be reasonable to send
Cleopatra away, and deprive her of all participation in the glory of
the war, when she was defraying so large a part of the expense of it.
Besides, a large portion of the army consisted of Egyptian troops,
who would feel discouraged and disheartened if Cleopatra were to
leave them, and would probably act far less efficiently in the conflict
than they would do if animated by the presence of their queen.
Then, moreover, such a woman as Cleopatra was not to be
considered, as many women would be, an embarrassment and a
source of care to a military expedition which she might join, but a
very efficient counselor and aid to it. She was, he said, a very
sagacious, energetic, and powerful queen, accustomed to the
command of armies and to the management of affairs of state, and
her aid in the conduct of the expedition might be expected to
conduce very materially to its success.
Antony was easily won by such persuasions as these, and it was at
length decided that Cleopatra should accompany him.
Antony then ordered the fleet to move forward to the island of
Samos.[9] Here it was brought to anchor and remained for some
time, waiting for the coming in of new re-enforcements, and for the
completion of the other arrangements. Antony, as if becoming more
and more infatuated as he approached the brink of his ruin, spent
his time while the expedition remained at Samos, not in maturing his
plans and perfecting his arrangements for the tremendous conflict
which was approaching, but in festivities, games, revelings, and
every species of riot and dissolute excess. This, however, is not
surprising. Men almost always, when in a situation analogous to his,
fly to similar means of protecting themselves, in some small degree,
from the pangs of remorse, and from the forebodings which stand
ready to terrify and torment them at every instant in which these
gloomy specters are not driven away by intoxication and revelry. At
least Antony found it so. Accordingly, an immense company of
players, tumblers, fools, jesters, and mountebanks were ordered to
assemble at Samos, and to devote themselves with all zeal to the
amusement of Antony’s court. The island was one universal scene of
riot and revelry. People were astonished at such celebrations and
displays, wholly unsuitable, as they considered them, to the
occasion. If such are the rejoicings, said they, which Antony
celebrates before going into the battle, what festivities will he
contrive on his return, joyous enough to express his pleasure if he
shall gain the victory?
After a time, Antony and Cleopatra, with a magnificent train of
attendants, left Samos, and, passing across the Ægean Sea, landed
in Greece, and advanced to Athens; while the fleet, proceeding
westward from Samos, passed around Tænarus, the southern
promontory of Greece, and then moved northward along the
western coast of the peninsula. Cleopatra wished to go to Athens for
a special reason. It was there that Octavia had stopped on her
journey toward her husband with re-enforcements and aid; and
while she was there, the people of Athens, pitying her sad condition,
and admiring the noble spirit of mind which she displayed in her
misfortunes, had paid her great attention, and during her stay
among them had bestowed upon her many honors. Cleopatra now
wished to go to the same place, and to triumph over her rival there,
by making so great a display of her wealth and magnificence, and of
her ascendency over the mind of Antony, as should entirely
transcend and outshine the more unassuming pretensions of
Octavia. She was not willing, it seems, to leave to the unhappy wife
whom she had so cruelly wronged even the possession of a place in
the hearts of the people of this foreign city, but must go and
enviously strive to efface the impression which injured innocence
had made, by an ostentatious exhibition of the triumphant prosperity
of her own shameless wickedness. She succeeded well in her plans.
The people of Athens were amazed and bewildered at the immense
magnificence that Cleopatra exhibited before them. She distributed
vast sums of money among the people. The city, in return, decreed
to her the most exalted honors. They sent a solemn embassy to her
to present her with these decrees. Antony himself, in the character
of a citizen of Athens, was one of the embassadors. Cleopatra
received the deputation at her palace. The reception was attended
with the most splendid and imposing ceremonies.
One would have supposed that Cleopatra’s cruel and unnatural
hostility to Octavia might now have been satisfied; but it was not.
Antony, while he was at Athens, and doubtless at Cleopatra’s
instigation, sent a messenger to Rome with a notice of divorcement
to Octavia, and with an order that she should leave his house.
Octavia obeyed. She went forth from her home, taking the children
with her, and bitterly lamenting her cruel destiny.
In the mean time, while all these events had been transpiring in the
East, Octavius had been making his preparations for the coming
crisis, and was now advancing with a powerful fleet across the sea.
He was armed with authority from the Roman senate and people, for
he had obtained from them a decree deposing Antony from his
power. The charges made against him all related to misdemeanors
and offenses arising out of his connection with Cleopatra. Octavius
contrived to get possession of a will which Antony had written before
leaving Rome, and which he had placed there in what he supposed a
very sacred place of deposit. The custodians who had it in charge
replied to Octavius, when he demanded it, that they would not give
it to him, but if he wished to take it they would not hinder him.
Octavius then took the will, and read it to the Roman senate. It
provided, among other things, that at his death, if his death should
happen at Rome, his body should be sent to Alexandria to be given
to Cleopatra; and it evinced in other ways a degree of subserviency
and devotedness to the Egyptian queen which was considered
wholly unworthy of a Roman chief magistrate. Antony was accused,
too, of having plundered cities and provinces to make presents to
Cleopatra; of having sent a library of two hundred thousand volumes
to her from Pergamus, to replace the one which Julius Cæsar had
accidentally burned; of having raised her sons, ignoble as their birth
was, to high places of trust and power in the Roman government,
and of having in many ways compromised the dignity of a Roman
officer by his unworthy conduct in reference to her. He used, for
example, when presiding at a judicial tribunal, to receive love-letters
sent him from Cleopatra, and then at once turn off his attention from
the proceedings going forward before him to read the letters.[10]
Sometimes he did this when sitting in the chair of state, giving
audience to embassadors and princes. Cleopatra probably sent these
letters in at such times under the influence of a wanton disposition
to show her power. At one time, as Octavius said in his arguments
before the Roman senate, Antony was hearing a cause of the
greatest importance, and during a time in the progress of the cause
when one of the principal orators of the city was addressing him,
Cleopatra came passing by, when Antony suddenly arose, and,
leaving the court without any ceremony, ran out to follow her. These
and a thousand similar tales exhibited Antony in so odious a light,
that his friends forsook his cause, and his enemies gained a
complete triumph. The decree was passed against him, and Octavius
was authorized to carry it into effect; and accordingly, while Antony,
with his fleet and army, was moving westward from Samos and the
Ægean Sea, Octavius was coming eastward and southward down the
Adriatic to meet him.
In process of time, after various maneuvers and delays, the two
armaments came into the vicinity of each other at a place called
Actium, which will be found upon the map on the western coast of
Epirus, north of Greece. Both of the commanders had powerful fleets
at sea, and both had great armies upon the land. Antony was
strongest in land troops, but his fleet was inferior to that of
Octavius, and he was himself inclined to remain on the land and
fight the principal battle there. But Cleopatra would not consent to
this. She urged him to give Octavius battle at sea. The motive which
induced her to do this has been supposed to be her wish to provide
a more sure way of escape in case of an unfavorable issue to the
conflict. She thought that in her galleys she could make sail at once
across the sea to Alexandria in case of defeat, whereas she knew
not what would become of her if beaten at the head of an army on
the land. The ablest counselors and chief officers in the army urged
Antony very strongly not to trust himself to the sea. To all their
arguments and remonstrances, however, Antony turned a deaf ear.
Cleopatra must be allowed to have her way.
On the morning of the battle, when the ships were drawn up in
array, Cleopatra held the command of a division of fifty or sixty
Egyptian vessels, which were all completely manned, and well
equipped with masts and sails. She took good care to have every
thing in perfect order for flight, in case flight should prove to be
necessary. With these ships she took a station in reserve, and for a
time remained there a quiet witness of the battle. The ships of
Octavius advanced to the attack of those of Antony, and the men
fought from deck to deck with spears, boarding-pikes, flaming darts,
and every other destructive missile which the military art had then
devised. Antony’s ships had to contend against great disadvantages.
They were not only outnumbered by those of Octavius, but were far
surpassed by them in the efficiency with which they were manned
and armed. Still, it was a very obstinate conflict. Cleopatra, however,
did not wait to see how it was to be finally decided. As Antony’s
forces did not immediately gain the victory, she soon began to yield
to her fears in respect to the result, and, finally, fell into a panic and
resolved to fly. She ordered the oars to be manned and the sails to
be hoisted, and then forcing her way through a portion of the fleet
that was engaged in the contest, and throwing the vessels into
confusion as she passed, she succeeded in getting to sea, and then
pressed on, under full sail, down the coast to the southward. Antony,
as soon as he perceived that she was going, abandoning every other
thought, and impelled by his insane devotedness to her, hastily
called up a galley of five banks of oars, and, leaping on board of it,
ordered the oarsmen to pull with all their force after Cleopatra’s
flying squadron.
Cleopatra, looking back from the deck of her vessel, saw this swift
galley pressing on toward her. She raised a signal at the stern of the
vessel which she was in, that Antony might know for which of the
fifty flying ships he was to steer. Guided by the signal, Antony came
up to the vessel, and the sailors hoisted him up the side and helped
him in. Cleopatra had, however, disappeared. Overcome with shame
and confusion, she did not dare, it seems, to meet the look of the
wretched victim of her arts whom she had now irretrievably ruined.
Antony did not seek her. He did not speak a word. He went forward
to the prow of the ship, and, throwing himself down there alone,
pressed his head between his hands, and seemed stunned and
stupefied, and utterly overwhelmed with horror and despair.
He was, however, soon aroused from his stupor by an alarm raised
on board his galley that they were pursued. He rose from his seat,
seized a spear, and, on ascending to the quarter-deck, saw that
there were a number of small light boats, full of men and of arms,
coming up behind them, and gaining rapidly upon his galley. Antony,
now free for a moment from his enchantress’s sway, and acting
under the impulse of his own indomitable boldness and decision,
instead of urging the oarsmen to press forward more rapidly in order
to make good their escape, ordered the helm to be put about, and
thus, turning the galley around, he faced his pursuers, and drove his
ship into the midst of them. A violent conflict ensued, the din and
confusion of which was increased by the shocks and collisions
between the boats and the galley. In the end, the boats were beaten
off, all excepting one: that one kept still hovering near, and the
commander of it, who stood upon the deck, poising his spear with
an aim at Antony, and seeking eagerly an opportunity to throw it,
seemed by his attitude and the expression of his countenance to be
animated by some peculiarly bitter feeling of hostility and hate.
Antony asked him who he was, that dared so fiercely to threaten
him. The man replied by giving his name, and saying that he came
to avenge the death of his father. It proved that he was the son of a
man whom Antony had at a previous time, on some account or
other, caused to be beheaded.
There followed an obstinate contest between Antony and this fierce
assailant, in the end of which the latter was beaten off. The boats
then, having succeeded in making some prizes from Antony’s fleet,
though they had failed in capturing Antony himself, gave up the
pursuit and returned. Antony then went back to his place, sat down
in the prow, buried his face in his hands, and sank into the same
condition of hopeless distress and anguish as before.
When husband and wife are overwhelmed with misfortune and
suffering, each instinctively seeks a refuge in the sympathy and
support of the other. It is, however, far otherwise with such
connections as that of Antony and Cleopatra. Conscience, which
remains calm and quiet in prosperity and sunshine, rises up with
sudden and unexpected violence as soon as the hour of calamity
comes; and thus, instead of mutual comfort and help, each finds in
the thoughts of the other only the means of adding the horrors of
remorse to the anguish of disappointment and despair. So extreme
was Antony’s distress, that for three days he and Cleopatra neither
saw nor spoke to each other. She was overwhelmed with confusion
and chagrin, and he was in such a condition of mental excitement
that she did not dare to approach him. In a word, reason seemed to
have wholly lost its sway—his mind, in the alternations of his
insanity, rising sometimes to fearful excitement, in paroxysms of
uncontrollable rage, and then sinking again for a time into the stupor
of despair.
In the mean time, the ships were passing down as rapidly as
possible on the western coast of Greece. When they reached
Tænarus, the southern promontory of the peninsula, it was
necessary to pause and consider what was to be done. Cleopatra’s
women went to Antony and attempted to quiet and calm him. They
brought him food. They persuaded him to see Cleopatra. A great
number of merchant ships from the ports along the coast gathered
around Antony’s little fleet and offered their services. His cause, they
said, was by no means desperate. The army on the land had not
been beaten. It was not even certain that his fleet had been
conquered. They endeavored thus to revive the ruined commander’s
sinking courage, and to urge him to make a new effort to retrieve his
fortunes. But all was in vain. Antony was sunk in a hopeless
despondency. Cleopatra was determined on going to Egypt, and he
must go too. He distributed what treasure remained at his disposal
among his immediate followers and friends, and gave them advice
about the means of concealing themselves until they could make
peace with Octavius. Then, giving up all as lost, he followed
Cleopatra across the sea to Alexandria.
 Chapter XII.

The End of Cleopatra.

T HE case of Mark Antony affords one of the most extraordinary

examples of the power of unlawful love to lead its deluded and
infatuated victim into the very jaws of open and recognized
destruction that history records. Cases similar in character occur by
thousands in common life; but Antony’s, though perhaps not more
striking in itself than a great multitude of others have been, is the
most conspicuous instance that has ever been held up to the
observation of mankind.
In early life, Antony was remarkable, as we have already seen, for a
certain savage ruggedness of character, and for a stern and
indomitable recklessness of will, so great that it seemed impossible
that any thing human should be able to tame him. He was under the
control, too, of an ambition so lofty and aspiring that it appeared to
know no bounds; and yet we find him taken possession of, in the
very midst of his career, and in the height of his prosperity and
success, by a woman, and so subdued by her arts and fascinations
as to yield himself wholly to her guidance, and allow himself to be
led about by her entirely at her will. She displaces whatever there
might have been that was noble and generous in his heart, and
substitutes therefor her own principles of malice and cruelty. She
extinguishes all the fires of his ambition, originally so magnificent in
its aims that the world seemed hardly large enough to afford it
scope, and instead of this lofty passion, fills his soul with a love of
the lowest, vilest, and most ignoble pleasures. She leads him to
betray every public trust, to alienate from himself all the affections
of his countrymen, to repel most cruelly the kindness and
devotedness of a beautiful and faithful wife, and, finally, to expel this
wife and all of his own legitimate family from his house; and now, at
last, she conducts him away in a most cowardly and ignoble flight
from the field of his duty as a soldier—he knowing, all the time, that
she is hurrying him to disgrace and destruction, and yet utterly
without power to break from the control of his invisible chains.

The indignation which Antony’s base abandonment of his fleet and

army at the battle of Actium excited, over all that part of the empire
which had been under his command, was extreme. There was not
the slightest possible excuse for such a flight. His army, in which his
greatest strength lay, remained unharmed, and even his fleet was
not defeated. The ships continued the combat until night,
notwithstanding the betrayal of their cause by their commander.
They were at length, however, subdued. The army, also, being
discouraged, and losing all motive for resistance, yielded too. In a
very short time the whole country went over to Octavius’s side.
In the mean time, Cleopatra and Antony, on their first return to
Egypt, were completely beside themselves with terror. Cleopatra
formed a plan for having all the treasures that she could save, and a
certain number of galleys sufficient for the transportation of these
treasures and a small company of friends, carried across the isthmus
of Suez and launched upon the Red Sea, in order that she might
escape in that direction, and find some remote hiding-place and safe
retreat on the shores of Arabia or India, beyond the reach of
Octavius’s dreaded power. She actually commenced this undertaking,
and sent one or two of her galleys across the isthmus; but the Arabs
seized them as soon as they reached their place of destination, and
killed or captured the men that had them in charge, so that this
desperate scheme was soon abandoned. She and Antony then finally
concluded to establish themselves at Alexandria, and made
preparation, as well as they could, for defending themselves against
Octavius there.
Antony, when the first effects of his panic subsided, began to grow
mad with vexation and resentment against all mankind. He
determined that he would have nothing to do with Cleopatra or with
any of her friends, but went off in a fit of sullen rage, and built a
hermitage in a lonely place on the island of Pharos, where he lived
for a time, cursing his folly and his wretched fate, and uttering the
bitterest invectives against all who had been concerned in it. Here
tidings came continually in, informing him of the defection of one
after another of his armies, of the fall of his provinces in Greece and
Asia Minor, and of the irresistible progress which Octavius was now
making toward universal dominion. The tidings of these disasters
coming incessantly upon him kept him in a continual fever of
resentment and rage.
At last he became tired of his misanthropic solitude, a sort of
reconciliation ensued between himself and Cleopatra, and he went
back again to the city. Here he joined himself once more to
Cleopatra, and, collecting together what remained of their joint
resources, they plunged again into a life of dissipation and vice, with
the vain attempt to drown in mirth and wine the bitter regrets and
the anxious forebodings which filled their souls. They joined with
them a company of revelers as abandoned as themselves, and
strove very hard to disguise and conceal their cares in their forced
and unnatural gayety. They could not, however, accomplish this
purpose. Octavius was gradually advancing in his progress, and they
knew very well that the time of his dreadful reckoning with them
must soon come; nor was there any place on earth in which they
could look with any hope of finding a refuge in it from his vindictive
hostility.
Cleopatra, warned by dreadful presentiments of what would
probably at last be her fate, amused herself in studying the nature of
poisons—not theoretically, but practically—making experiments with
them on wretched prisoners and captives whom she compelled to
take them, in order that she and Antony might see the effects which
they produced. She made a collection of all the poisons which she
could procure, and administered portions of them all, that she might
see which were sudden and which were slow in their effects, and
also learn which produced the greatest distress and suffering, and
which, on the other hand, only benumbed and stupefied the
faculties, and thus extinguished life with the least infliction of pain.
These experiments were not confined to such vegetable and mineral
poisons as could be mingled with the food or administered in a
potion. Cleopatra took an equal interest in the effects of the bite of
venomous serpents and reptiles. She procured specimens of all
these animals, and tried them upon her prisoners, causing the men
to be stung and bitten by them, and then watching the effects.
These investigations were made, not directly with a view to any
practical use which she was to make of the knowledge thus
acquired, but rather as an agreeable occupation, to divert her mind,
and to amuse Antony and her guests. The variety in the forms and
expressions which the agony of her poisoned victims assumed—their
writhings, their cries, their convulsions, and the distortions of their
features when struggling with death, furnished exactly the kind and
degree of excitement which she needed to occupy and amuse her
mind.
Antony was not entirely at ease, however, during the progress of
these terrible experiments. His foolish and childish fondness for
Cleopatra was mingled with jealousy, suspicion, and distrust; and he
was so afraid that Cleopatra might secretly poison him, that he
would never take any food or wine without requiring that she should
taste it before him. At length, one day, Cleopatra caused the petals
of some flowers to be poisoned, and then had the flowers woven
into the chaplet which Antony was to wear at supper. In the midst of
the feast, she pulled off the leaves of the flowers from her own
chaplet and put them playfully into her wine, and then proposed that
Antony should do the same with his chaplet, and that they should
then drink the wine, tinctured, as it would be, with the color and the
perfume of the flowers. Antony entered very readily into this
proposal, and when he was about to drink the wine, she arrested his
hand, and told him that it was poisoned. “You see now,” said she,
“how vain it is for you to watch against me. If it were possible for
me to live without you, how easy it would be for me to devise ways
and means to kill you.” Then, to prove that her words were true, she
ordered one of the servants to drink Antony’s wine. He did so, and
died before their sight in dreadful agony.
The experiments which Cleopatra thus made on the nature and
effects of poison were not, however, wholly without practical result.
Cleopatra learned from them, it is said, that the bite of the asp was
the easiest and least painful mode of death. The effect of the venom
of that animal appeared to her to be the lulling of the sensorium into
a lethargy or stupor, which soon ended in death, without the
intervention of pain. This knowledge she seems to have laid up in
her mind for future use.
The thoughts of Cleopatra appear, in fact, to have been much
disposed, at this time, to flow in gloomy channels, for she occupied
herself a great deal in building for herself a sepulchral monument in
a certain sacred portion of the city. This monument had, in fact,
been commenced many years ago, in accordance with a custom
prevailing among Egyptian sovereigns, of expending a portion of
their revenues during their life-time in building and decorating their
own tombs. Cleopatra now turned her mind with new interest to her
own mausoleum. She finished it, provided it with the strongest
possible bolts and bars, and, in a word, seemed to be preparing it in
all respects for occupation.
In the mean time, Octavius, having made himself master of all the
countries which had formerly been under Antony’s sway, now
advanced, meeting none to oppose him, from Asia Minor into Syria,
and from Syria toward Egypt. Antony and Cleopatra made one
attempt, while he was thus advancing toward Alexandria, to avert
the storm which was impending over them, by sending an
embassage to ask for some terms of peace. Antony proposed, in this
embassage, to give up every thing to his conqueror on condition that
he might be permitted to retire unmolested with Cleopatra to
Athens, and allowed to spend the remainder of their days there in
peace; and that the kingdom of Egypt might descend to their
children. Octavius replied that he could not make any terms with
Antony, though he was willing to consent to any thing that was
reasonable in behalf of Cleopatra. The messenger who came back
from Octavius with this reply spent some time in private interviews
with Cleopatra. This aroused Antony’s jealousy and anger. He
accordingly ordered the unfortunate messenger to be scourged and
then sent back to Octavius, all lacerated with wounds, with orders to
say to Octavius that if it displeased him to have one of his servants
thus punished, he might revenge himself by scourging a servant of
Antony’s, who was then, as it happened, in Octavius’s power.
The news at length suddenly arrived at Alexandria that Octavius had
appeared before Pelusium, and that the city had fallen into his
hands. The next thing Antony and Cleopatra well knew would be,
that they should see him at the gates of Alexandria. Neither Antony
nor Cleopatra had any means of resisting his progress, and there
was no place to which they could fly. Nothing was to be done but to
await, in consternation and terror, the sure and inevitable doom
which was now so near.
Cleopatra gathered together all her treasures and sent them to her
tomb. These treasures consisted of great and valuable stores of
gold, silver, precious stones, garments of the highest cost, and
weapons, and vessels of exquisite workmanship and great value, the
hereditary possessions of the Egyptian kings. She also sent to the
mausoleum an immense quantity of flax, tow, torches, and other
combustibles. These she stored in the lower apartments of the
monument, with the desperate determination of burning herself and
her treasures together rather than to fall into the hands of the
Romans.
In the mean time, the army of Octavius steadily continued its march
across the desert from Pelusium to Alexandria. On the way, Octavius
learned, through the agents in communication with him from within
the city, what were the arrangements which Cleopatra had made for
the destruction of her treasure whenever the danger should become
imminent of its falling into his hands. He was extremely unwilling
that this treasure should be lost. Besides its intrinsic value, it was an
object of immense importance to him to get possession of it for the
purpose of carrying it to Rome as a trophy of his triumph. He
accordingly sent secret messengers to Cleopatra, endeavoring to
separate her from Antony, and to amuse her mind with the
profession that he felt only friendship for her, and did not mean to
do her any injury, being in pursuit of Antony only. These negotiations
were continued from day to day while Octavius was advancing. At
last the Roman army reached Alexandria, and invested it on every
side.
As soon as Octavius was established in his camp under the walls of
the city, Antony planned a sally, and he executed it, in fact, with
considerable energy and success. He issued suddenly from the
gates, at the head of as strong a force as he could command, and
attacked a body of Octavius’s horsemen. He succeeded in driving
these horsemen away from their position, but he was soon driven
back in his turn, and compelled to retreat to the city, fighting as he
fled, to beat back his pursuers. He was extremely elated at the
success of this skirmish. He came to Cleopatra with a countenance
full of animation and pleasure, took her in his arms and kissed her,
all accoutered for battle as he was, and boasted greatly of the
exploit which he had performed. He praised, too, in the highest
terms, the valor of one of the officers who had gone out with him to
the fight, and whom he had now brought to the palace to present to
Cleopatra. Cleopatra rewarded the faithful captain’s prowess with a
magnificent suit of armor made of gold. Notwithstanding this
reward, however, the man deserted Antony that very night, and
went over to the enemy. Almost all of Antony’s adherents were in
the same state of mind. They would have gladly gone over to the
camp of Octavius, if they could have found an opportunity to do so.
In fact, when the final battle was fought, the fate of it was decided
by a grand defection in the fleet, which went over in a body to the
side of Octavius. Antony was planning the operations of the day, and
reconnoitering the movements of the enemy from an eminence
which he occupied at the head of a body of foot soldiers—all the
land forces that now remained to him—and looking off from the
eminence on which he stood toward the harbor, he observed a
movement among the galleys. They were going out to meet the
ships of Octavius, which were lying at anchor not very far from
them. Antony supposed that his vessels were going to attack those
of the enemy, and he looked to see what exploits they would
perform. They advanced toward Octavius’s ships, and when they
met them, Antony observed, to his utter amazement, that, instead of
the furious combat that he had expected to see, the ships only
exchanged friendly salutations, by the use of the customary naval
signals; and then his ships, passing quietly round, took their
positions in the lines of the other fleet. The two fleets had thus
become merged and mingled into one.
Antony immediately decided that this was Cleopatra’s treason. She
had made peace with Octavius, he thought, and surrendered the
fleet to him as one of the conditions of it. Antony ran through the
city, crying out that he was betrayed, and in a phrensy of rage
sought the palace. Cleopatra fled to her tomb. She took in with her
one or two attendants, and bolted and barred the doors, securing
the fastenings with the heavy catches and springs that she had
previously made ready. She then directed her women to call out
through the door that she had killed herself within the tomb.
The tidings of her death were borne to Antony. It changed his anger
to grief and despair. His mind, in fact, was now wholly lost to all
balance and control, and it passed from the dominion of one stormy
passion to another with the most capricious facility. He cried out with
the most bitter expressions of sorrow, mourning, he said, not so
much Cleopatra’s death, for he should soon follow and join her, as
the fact that she had proved herself so superior to him in courage at
last, in having thus anticipated him in the work of self-destruction.
He was at this time in one of the chambers of the palace, whither he
had fled in his despair, and was standing by a fire, for the morning
was cold. He had a favorite servant named Eros, whom he greatly
trusted, and whom he had made to take an oath long before, that
whenever it should become necessary for him to die, Eros should kill
him. This Eros he now called to him, and telling him that the time
was come, ordered him to take the sword and strike the blow.
Eros took the sword while Antony stood up before him. Eros turned
his head aside as if wishing that his eyes should not see the deed
which his hands were about to perform. Instead, however, of
piercing his master with it, he plunged it into his own breast, fell
down at Antony’s feet, and died.
Antony gazed a moment at the shocking spectacle, and then said, “I
thank thee for this, noble Eros. Thou hast set me an example. I
must do for myself what thou couldst not do for me.” So saying, he
took the sword from his servant’s hands, plunged it into his body,
and staggering to a little bed that was near, fell over upon it in a
swoon. He had received a mortal wound.
The pressure, however, which was produced by the position in which
he lay upon the bed, stanched the wound a little and stopped the
flow of blood. Antony came presently to himself again, and then
began to beg and implore those around him to take the sword and
put him out of his misery. But no one would do it. He lay for a time
suffering great pain, and moaning incessantly, until, at length, an
officer came into the apartment and told him that the story which he
had heard of Cleopatra’s death was not true; that she was still alive,
shut up in her monument, and that she desired to see him there.
This intelligence was the source of new excitement and agitation.
Antony implored the by-standers to carry him to Cleopatra, that he
might see her once more before he died. They shrank from the
attempt; but, after some hesitation and delay, they concluded to
undertake to remove him. So, taking him in their arms, they bore
him along, faint and dying, and marking their track with his blood,
toward the tomb.
Cleopatra would not open the gates to let the party in. The city was
all in uproar and confusion through the terror of the assault which
Octavius was making upon it, and she did not know what treachery
might be intended. She therefore went up to a window above, and
letting down ropes and chains, she directed those below to fasten
the dying body to them, that she and the two women with her might
draw it up. This was done. Those who witnessed it said that it was a
most piteous sight to behold—Cleopatra and her women above
exhausting their strength in drawing the wounded and bleeding
sufferer up the wall, while he, when he approached the window,
feebly raised his arms to them, that they might lift him in. The
women had hardly strength sufficient to draw the body up. At one
time it seemed that the attempt would have to be abandoned; but
Cleopatra reached down from the window as far as she could to get
hold of Antony’s arms, and thus, by dint of great effort, they
succeeded at last in taking him in. They bore him to a couch which
was in the upper room from which the window opened, and laid him
down, while Cleopatra wrung her hands, and tore her hair, and
uttered the most piercing lamentations and cries. She leaned over
the dying Antony, crying out incessantly with the most piteous
exclamations of grief. She bathed his face, which was covered with
blood, and vainly endeavored to stanch his wound.
Antony urged her to be calm, and not to mourn his fate. He asked
for some wine. They brought it to him, and he drank it. He then
entreated Cleopatra to save her life, if she possibly could do so, and
to make some terms or other with Octavius, so as to continue to
live. Very soon after this he expired.
 The raising of Antony to the upper Window of the Tomb

In the mean time, Octavius had heard of the mortal wound which
Antony had given himself; for one of the by-standers had seized the
sword the moment that the deed was done, and had hastened to
carry it to Octavius, and to announce to him the death of his enemy.
Octavius immediately desired to get Cleopatra into his power. He
sent a messenger, therefore, to the tomb, who attempted to open a
parley there with her. Cleopatra talked with the messenger through
the keyholes or crevices, but could not be induced to open the door.
The messenger reported these facts to Octavius. Octavius then sent
another man with the messenger, and while one was engaging the
attention of Cleopatra and her women at the door below, the other
obtained ladders, and succeeded in gaining admission into the
window above. Cleopatra was warned of the success of this
stratagem by the shriek of her woman, who saw the officer coming
down the stairs. She looked around, and observing at a glance that
she was betrayed, and that the officer was coming to seize her, she
drew a little dagger from her robe, and was about to plunge it into
her breast, when the officer grasped her arm just in time to prevent
the blow. He took the dagger from her, and then examined her
clothes to see that there were no other secret weapons concealed
there.
The capture of the queen being reported to Octavius, he appointed
an officer to take her into close custody. This officer was charged to
treat her with all possible courtesy, but to keep a close and constant
watch over her, and particularly to guard against allowing her any
possible means or opportunity for self-destruction.
In the mean time, Octavius took formal possession of the city,
marching in at the head of his troops with the most imposing pomp
and parade. A chair of state, magnificently decorated, was set up for
him on a high elevation in a public square; and here he sat, with
circles of guards around him, while the people of the city, assembled
before him in the dress of suppliants, and kneeling upon the
pavement, begged his forgiveness, and implored him to spare the
city. These petitions the great conqueror graciously condescended to
grant.
Many of the princes and generals who had served under Antony
came next to beg the body of their commander, that they might give
it an honorable burial. These requests, however, Octavius would not
accede to, saying that he could not take the body away from
Cleopatra. He, however, gave Cleopatra leave to make such
arrangements for the obsequies as she thought fit, and allowed her
to appropriate such sums of money from her treasures for this
purpose as she desired. Cleopatra accordingly made the necessary
arrangements, and superintended the execution of them; not,
however, with any degree of calmness and composure, but in a
state, on the contrary, of extreme agitation and distress. In fact, she
had been living now so long under the unlimited and unrestrained
dominion of caprice and passion, that reason was pretty effectually
dethroned, and all self-control was gone. She was now nearly forty
years of age, and, though traces of her inexpressible beauty
remained, her bloom was faded, and her countenance was wan with
the effects of weeping, anxiety, and despair. She was, in a word,
both in body and mind, only the wreck and ruin of what she once
had been.
When the burial ceremonies were performed, and she found that all
was over—that Antony was forever gone, and she herself hopelessly
and irremediably ruined—she gave herself up to a perfect phrensy of
grief. She beat her breast, and scratched and tore her flesh so
dreadfully, in the vain efforts which she made to kill herself, in the
paroxysms of her despair, that she was soon covered with
contusions and wounds, which, becoming inflamed and swelled,
made her a shocking spectacle to see, and threw her into a fever.
She then conceived the idea of pretending to be more sick than she
was, and so refusing food and starving herself to death. She
attempted to execute this design. She rejected every medical
remedy that was offered her, and would not eat, and lived thus
some days without food. Octavius, to whom every thing relating to
his captive was minutely reported by her attendants, suspected her
design. He was very unwilling that she should die, having set his
heart on exhibiting her to the Roman people, on his return to the
capital, in his triumphal procession. He accordingly sent her orders,
requiring that she should submit to the treatment prescribed by the
physician, and take her food, enforcing these his commands with a
certain threat which he imagined might have some influence over
her. And what threat does the reader imagine could possibly be
devised to reach a mind so sunk, so desperate, so wretched as hers?
Every thing seemed already lost but life, and life was only an
insupportable burden. What interests, then, had she still remaining
upon which a threat could take hold?
Octavius, in looking for some avenue by which he could reach her,
reflected that she was a mother. Cæsarion, the son of Julius Cæsar,
and Alexander, Cleopatra, and Ptolemy, Antony’s children, were still
alive. Octavius imagined that in the secret recesses of her wrecked
and ruined soul there might be some lingering principle of maternal
affection remaining which he could goad into life and action. He
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge

connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and

personal growth every day!

ebookbell.com