Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.

fw001

Chemistry, Process Design,

and Safety for the
Nitration Industry

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.fw001

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 ACS SYMPOSIUM SERIES 1155

Chemistry, Process Design,

and Safety for the
Nitration Industry
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.fw001

Thomas L. Guggenheim, Editor

SABIC Innovative Plastics
Mt. Vernon, Indiana

Sponsored by the
ACS Division of Industrial and Engineering Chemistry

American Chemical Society, Washington, DC

Distributed in print by Oxford University Press

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Library of Congress Cataloging-in-Publication Data

Chemistry, process design, and safety for the nitration industry / Thomas L. Guggenheim,
editor, SABIC Innovative Plastics, Mt. Vernon, Indiana ; sponsored by the ACS Division of
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.fw001

Industrial and Engineering Chemistry.

pages cm -- (ACS symposium series, ISSN 0097-6156 ; 1155)
Includes bibliographical references and index.
ISBN 978-0-8412-2886-3 (alk. paper)
1. Nitrates--Safety measures. 2. Nitration--Congresses. 3. Chemical processes--
Safety measures--Congresses. 4. Chemical plants--Design and construction--Congresses.
5. Chemical process control--Congresses. I. Guggenheim, Thomas L. II. American
Chemical Society. Division of Industrial and Engineering Chemistry.
TP156.N5C44 2013
549′.732--dc23

The paper used in this publication meets the minimum requirements of American National
Standard for Information Sciences—Permanence of Paper for Printed Library Materials,
ANSI Z39.48n1984.

Copyright © 2013 American Chemical Society

Distributed in print by Oxford University Press

All Rights Reserved. Reprographic copying beyond that permitted by Sections 107 or 108
of the U.S. Copyright Act is allowed for internal use only, provided that a per-chapter fee of
$40.25 plus $0.75 per page is paid to the Copyright Clearance Center, Inc., 222 Rosewood
Drive, Danvers, MA 01923, USA. Republication or reproduction for sale of pages in this
book is permitted only under license from ACS. Direct these and other permission requests
to ACS Copyright Office, Publications Division, 1155 16th Street, N.W., Washington, DC
20036.

The citation of trade names and/or names of manufacturers in this publication is not to be
construed as an endorsement or as approval by ACS of the commercial products or services
referenced herein; nor should the mere reference herein to any drawing, specification,
chemical process, or other data be regarded as a license or as a conveyance of any right
or permission to the holder, reader, or any other person or corporation, to manufacture,
reproduce, use, or sell any patented invention or copyrighted work that may in any way be
related thereto. Registered names, trademarks, etc., used in this publication, even without
specific indication thereof, are not to be considered unprotected by law.

PRINTED IN THE UNITED STATES OF AMERICA

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Foreword
The ACS Symposium Series was first published in 1974 to provide a
mechanism for publishing symposia quickly in book form. The purpose of
the series is to publish timely, comprehensive books developed from the ACS
sponsored symposia based on current scientific research. Occasionally, books are
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.fw001

developed from symposia sponsored by other organizations when the topic is of

keen interest to the chemistry audience.

Before agreeing to publish a book, the proposed table of contents is reviewed

for appropriate and comprehensive coverage and for interest to the audience. Some
papers may be excluded to better focus the book; others may be added to provide
comprehensiveness. When appropriate, overview or introductory chapters are
added. Drafts of chapters are peer-reviewed prior to final acceptance or rejection,
and manuscripts are prepared in camera-ready format.

As a rule, only original research papers and original review papers are
included in the volumes. Verbatim reproductions of previous published papers
are not accepted.

ACS Books Department

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 James Dodgen 1921–2010
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.pr001

Jim Dodgen, born in Anniston, Alabama, graduated from Georgia Institute

of Technology with a B.S. in Chemical Engineering in 1943, whereupon he
entered the U.S. Navy. Lieutenant Dodgen was assigned to the Air Force, Pacific
Fleet, managing ordnance from 1943 until March 1945 in the Marshall Islands.
Jim married Charlene Ward in 1945 and they have two sons, James Jr. and
Charles. From 1945 until 1946, he had assignments pertaining to bomb- and
torpedo-handling equipment for the Bureau of Ordnance, and then distributing
aviation armaments with the Bureau of Aeronautics.
From 1946 to 1951, Jim worked as a senior engineer for Pennsalt, where he
designed chemical plants, while staying in the military as a reservist. In 1951,
he was called back to military service. From 1955 to 1958, Jim was head of
the propellants, explosives, chemicals, and pyrotechnic section of the Bureau
of Ordnance. He then worked at the Naval Propellant plant at Indian Head in
Maryland, serving as director from 1959 to 1962. During this time he worked on
propellant units for multiple systems including Talos, Sidewinder, Sparrow, and
Hawk. In 1962 he served as the representative of the Bureau of Naval Weapons at
Hercules in Utah, where he was responsible for engineering and inspection of the
second stage of Polaris. In 1965 he was transferred to the Naval Torpedo Station
in Washington, working on Mark torpedoes. He retired from the Navy with the
rank of Commander in 1968 and worked briefly at Lockheed, Olin, Aerojet Solid
Propulsion Co., and Cordova Chemical Co.
Charlene died in 1969. He remarried Virginia Britten in 1972 and became a
wonderful father to her three daughters. In 1974, Jim started Dodgen Engineering
Company, a one-man operation. He then consulted with many companies involved
in the manufacture of propellants, explosives, and chemicals up until his death in
2010.

ix
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 In 2003, the editor, working at General Electric at the time, got to work
with Jim when we started up a large-scale mixed acid nitration plant. Jim was
one of several consultants hired to oversee the engineering and safety aspects
of the process. He possessed the essential elements required when designing
and operating a plant that handles energetic material — namely, deep practical
experience and technical training. The plant started up and ran without incident;
and his insight and ability to teach others lent confidence to those who ran the
operation.
When a condenser failed in another nitration plant (one can read about
this in one of the chapters of this book), Jim was consulted. He had data in his
files on trinitromethane (the suspected culprit in the failure) that was not in the
public domain. This data proved very useful, resulting in the safe redesign of the
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.pr001

failed unit. Commander James Dodgen was a model technologist and wonderful
coworker.

x
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chester Grelecki 1927–2007
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.pr002

Chester (Chet) was born to Polish immigrants in Newton Township,

Pennsylvania in 1927, learning English when he went to school. In 1945, he left
high school and joined the Navy. He was discharged in 1946 and his older sister
pushed him to finish high school, after which he obtained a B.S. in chemistry
from Kings College (1950), an M.S. in biochemistry from Duquesne University,
and his Ph.D. in physical chemistry from F.O. Rice at the Catholic University of
America in Washington, D.C. (1956), whereupon he started working for Thiokol
Chem. Corp. in the Reaction Motors division.
In 1959, he became a manager directing work on propellant technology,
specifically mixed hydrazine fuel systems. This phase of Chet’s career concluded
with the successful landing of Surveyor 1 on the moon in 1966, which employed
the hydrazine fuel. Since the Surveyor briefly bounced on the surface during
the landing, Chet liked to claim that the fuel was also responsible for the first
successful launch of a vehicle from the moon’s surface.
While at Thiokol, Chet began testing propellants, commercial explosives,
and industrial chemicals to determine their thermal stability, detonation velocity,
critical diameter, ignition mechanisms, and shock sensitivity. In 1963, he founded
the Fire and Explosion Hazards Evaluation Service, a service to the chemical
process industries directed to the reduction of processing accidents. In 1968, Chet
was appointed Manager of Research Operation at Reaction Motors, directing work
in propellant and explosives research, combustion engineering, and pilot plant
process studies.
In 1970, Chet, with William Cruice, co-founded Hazards Research
Corporation (HRC) to continue safety studies for the chemical industry. From
that date until his death, Chet directed several thousand studies to access the
safety of chemicals and chemical processes in a multitude of industries. Work

xi
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 was performed for the Army, Navy, Air Force, Atomic Energy Commission,
Department of Transportation, the EPA, OSHA, and the chemical industry at
large. HRC determined the root cause of countless failures at chemical facilities,
leading to safe redesign efforts. In several cases, opposing parties hired Chet to
evaluate the circumstances of the failure in question, and based on his findings
settled the dispute, speaking to the high regard others placed in Chet. Chet
married the chemical nature of materials with the engineering used to handle
them. When interacting with him for the first time, it was not possible to discern
whether he was a chemical engineer or a chemist, or a physicist for that matter.
In the early 1970s Chet developed a course in Fire and Explosion Hazards
Evaluation for the American Institute of Chemical Engineers. This proved to be
an effective course, and was given hundreds of times at professional meetings and
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.pr002

companies around the world. Chet was a masterful educator, and special person
and tutor to authors Odle and Guggenheim. One can only ponder how many
industrial incidents and personnel injuries were averted because of the efforts of
Chet and all his associates at HRC. It is expertise and experience like Chet’s that
is required when designing and operating complex chemical operations.
Chet was a warm individual. He was once contracted to investigate a pump
explosion and he interviewed the people in the plant at the time of the event. He
asked how their ears were feeling. The question was part compassion and part
science: knowing the distance and orientation of the witness from the explosion,
whether the ear drum was intact or not, the metallurgy, and whether the pump
impellor housing failed in a brittle or ductile manner, quickly gave Chet an estimate
of the amount of material that had led to the explosion and if the event was a
detonation or a deflagration.
To see Chet’s photograph in color in the printed book, please see the color
insert.

xii
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Preface
This is the third ACS Symposium Series book dealing with nitration, the first
two having been published in 1976 and 1996. The nature of this 2013 publication
reflects the changes worldwide in process safety management, and geographies of
research and manufacturing. The contributions to this book were first presented
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.pr003

at the 243rd ACS National Meeting in San Diego, California in March of 2012, in
the Industrial and Chemical Engineering Division.
Several of the chapters deal with the burgeoning capacity increases in the
polyurethane industry, requiring improved methods to nitrate benzene and toluene,
to ultimately produce MDI and TDI. Methods to manage waste streams from these
nitrations plants are also discussed. There are several chapters on process safety
that discuss accident investigation, process redesign, and sensitivity testing of
energetic material. Hazards of laboratory and pilot plant nitration studies are
addressed. Several of the papers describe considerations which must be taken into
account when analyzing nitration reaction samples.
These chapters represent practical application of known principles and
concepts. Some of the chapters read more like a tutorial than a scientific paper.
Those new to nitration will benefit the most from reading this book, but it will
serve to remind the experienced of factors to consider when operating a nitration
facility. By no means are all hazards of nitration covered in this monograph.
Two Festschrifts are included in this publication, one for James Dodgen
and one for Chet Grelecki. Both these individuals were highly trained, deeply
experienced technologists who studied the processing and nature of energetic
materials. They remind us of the need to include minds such as theirs when
designing and operating nitration facilities.
The Editor wishes to thank those who made this book possible. Mary Moore
at Eastman Chemical Company assisted in organizing the nitration symposium
at the 243rd ACS National Meeting. The expert staff at ACS Books streamlined
the publishing process. Thanks to all the authors and reviewers who labored to
produce each chapter of the book. Finally, thanks to Jacob Oberholtzer and Roy
Odle, both working for SABIC, for encouragement and technical advice, and
SABIC for financial support.

Thomas L. Guggenheim
SABIC Innovative Plastics
Mt. Vernon, Indiana 47620

xiii
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 1

The Adiabatic Mononitrobenzene Process from

the Bench Scale in 1974
to a Total World Capacity Approaching
10 Million MTPY in 2012
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch001

Alfred Guenkel*

NORAM Engineering and Constructors Ltd., 200 Granville Street,

Suite 1800, Vancouver, BC, Canada V6C 1S4
*E-mail: [email protected]

The age of adiabatic mononitrobenzene (MNB) production

began with a meeting held in July 1974 at the Canadian
Industries Ltd. (CIL) Explosives Research Laboratory in
McMasterville, Quebec, Canada. Two senior scientists of
the American Cyanamid Company disclosed the adiabatic
MNB concept, and invited CIL to contribute its sulfuric acid
concentration technology, and lead the piloting of the adiabatic
process. Three simple questions had to be answered at that
time: What is the rate of by-products formation? Can the spent
acid be recycled indefinitely? What scale-up rules should be
applied to size industrial-scale stirred tank nitrators? The first
adiabatic MNB plant was brought on line in 1979, in Louisiana,
USA. At that time, the world’s MNB production was less than
1 million metric tonnes per year (MTPY), all coming from
plants based on the incumbent isothermal technology. The
world capacity in 2012 for MNB is now approaching 10 million
MTPY, predominantly from adiabatic plants. This paper is a
review of challenges which had to be overcome to bring the
now dominant adiabatic MNB process to its current state of
high reliability, high yield and energy efficiency, and excellent
safety record. MNB capacity estimates quoted in this paper
should be viewed as “best guesses” only. Producers keep
production records confidential.

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Background
The development of the adiabatic MNB process started with a meeting held
in July 1974 at CIL’s Explosives Research Laboratory. Two senior scientists of the
American Cyanamid Company disclosed, in a half-page document, the adiabatic
MNB process concept and invited CIL to participate in a joint development
project, where CIL would contribute its sulfuric acid concentration technology.
The process concept was simple: Nitrate benzene in a large excess of mixed acid,
sufficient to absorb all of the heat of nitration, decant the crude MNB phase,
then flash the spent acid under vacuum to boil off water generated chemically,
charging the reconcentrated sulfuric acid with additional nitric acid, and finally
sending the resulting mixed acid to the nitration stage. The potential benefits were
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch001

immediately evident; the heat of nitration was about the same as the heat required
to boil off the water from the spent acid. This would result in substantial energy
savings relative to the isothermal process, where the heat of nitration is dissipated
by cooling the nitrators and is thus wasted. Capital savings would come from the
elimination of almost all the heat-transfer surface areas in the isothermal nitrators
and in the associated sulfuric acid concentrator. During the meeting it was agreed
that three questions would have to be answered through a pilot program. What is
the rate of by-products formation? Can the sulfuric acid be recycled indefinitely?
What scale-up rules should be applied to size the proposed stirred-tank nitrators?
These nitrators had to achieve essentially complete conversion of nitric acid to
MNB for process economics and environmental reasons.

The MNB Market

American Cyanamid disclosed in the 1974 meeting that they were

contemplating building an MNB plant with a capacity of 350,000 MTPY, which
seemed surprising at the time. The total US MNB production in 1974 was only
230,000 MTPY (1). What wasn’t recognized by the CIL party, which included the
author of this paper, was that a new class of polymers, namely polyurethanes, had
become a commercial reality in the 1960’s, and was on a rapid growth trajectory,
which has continued to this day.
The first step in the synthesis of MDI-based urethanes requires MNB. In 1978
the total world production of MDI was 400,000 MTPY, which steadily increased to
2 million MTPY in 1998 (2). This corresponded to an average growth rate of 8.4%
per annum. To support this growth it has been necessary to build a world-scale
MNB plant almost every year, not counting capacity required by the replacement
of inefficient isothermal plants. Some large isothermal plants were, in fact, built
in the 1970’s, only to be scrapped in the 1980’s.
Extrapolation of production capacity through 2012 and beyond, accounting
for projects known to be in planning stages, suggests that current MDI capacity
may be over 6 million MTPY, and could exceed 10 million MTPY by 2015.
Much of the growth will take place in China, where currently at least seven major
projects are in various stages of execution. In 2010 the share of MNB production
worldwide was about 25% for the USA, 30% for Europe, 25% for China and 20%

2
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 for the rest of the world (3). China’s share will likely exceed 50% within the next
three years.
MNB has become a commodity chemical, and it is safe to assume that at least
one world-scale plant will be built every year for the foreseeable future. The other
trend has been to build plants with large capacities, of up to half a million MTPY.

First-Generation Adiabatic MNB Technology

Based on the outcome of the CIL/American Cyanamid pilot development
work, two adiabatic MNB plants came on stream in 1979, with individual
capacities of 172,000 and 50,000 MTPY (4). These two plants immediately
accounted for about a third of US capacity (1) and caused the subsequent
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch001

shut-down of a number of isothermal plants.

The pilot work did show that the rate of formation of nitrophenol by-products
was somewhat higher than that in the isothermal plants, but that it was not overly
problematic at the first two plant sites given the infrastructure available. On the
small pilot scale it was difficult to recycle the sulfuric acid more than 100 times,
but no negative effects could be observed. The problem in the acid recycle tests
was that the acid inventory had to be small in order to maximize the number of
times the sulfuric acid was recycled, and at the same time have enough material
to accommodate the sampling and analysis of the process. In a full-scale MNB
plant, the acid cycle time is typically about 5 minutes, so that 100 pilot plant cycles
corresponded to only a single plant operating shift.
Surprisingly, no meaningful data on the kinetics of heterogeneous benzene
nitration could be found in the literature at the time. In the isothermal MNB
process, kinetics had never been of much interest since the nitrator was always “big
enough”, provided that the required heat-transfer area could be accommodated
within the nitrator volume, provided that the acid concentration was such that
nitronium ions were present, and provided that the benzene was well dispersed
in the acid.
In the pilot plant the kinetics were studied through adiabatic batch nitration
experiments, where the extent of nitric acid conversion was established by
recording the rise in nitrator temperature as a function of time. Time zero was
when the total batch change of benzene was injected into the nitrator at a selected
initial nitrator temperature, mixed acid composition, and agitation intensity. All
three variables could be examined separately through this technique.
The outcome of the kinetics study – mostly on a 100 ml beaker scale, but also
in a nitrator vessel of 25 cm diameter, and even a single run in a 200 liter reactor
– was to scale up the nitrators on the basis of maintaining a constant power input
per unit volume. To the surprise of the research engineers and chemists, and of the
first commercial users of this technology, the nitration rates in the full-scale plants
were about two times faster than anticipated. None of the common stirred-tank
scaling rules could explain the high nitration rates. Evidently something else was
having a significant effect on the nitration rate.
Sometime after the completion of the adiabatic pilot plant work, a few studies
on the kinetics of nitration reactions were presented by a number of researchers
at the 169th Meeting of the American Chemical Society in 1975 (5). The findings
3
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 of these studies were useful for comparing relative nitration rates of different
aromatic compounds, but they could not be used to size the adiabatic nitrators.
Following the successful commissioning of the first two adiabatic plants, and
benefiting from the experience obtained during the start-ups, five more adiabatic
plants were built using the stirred nitrators as described in reference (4).
US Patents were granted to the American Cyanamid Company for the
adiabatic process in 1977 (6) and 1978 (7). In the course of a review of prior
art technology it was found that the DuPont Company had actually patented an
adiabatic batch process in 1941 (8).
The DuPont Company also recognized the potential of MDI growth in the
1970’s, and developed an alternative adiabatic process, called the azeotropic
process, where the heat of nitration was removed through the boiling of benzene
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch001

and water. Patents were granted for this process in 1975 (9) and 1976 (10).
With a number of adiabatic MNB plants having come on stream, and in view
of the rapid capacity growth, the Stanford Research Institute published a report
on the economics of the isothermal and adiabatic technologies in 1986 (11). The
isothermal MNB technology was still widely used at that time.
In 1990 a chapter on “Nitrobenzene and Nitrotoluene” was published in John
McKetta’s Encyclopedia of Chemical Processing and Design (12), where the
technical merits of the isothermal and adiabatic processes were compared.

Second-Generation Adiabatic MNB Technology

In the late 1980’s some of the engineers who played lead roles in the
development of the first-generation adiabatic MNB plants came together again
to see if anything could be done to improve the process. A motivating factor
was to answer the nagging question of why, in the first-generation plants, the
plant-scale nitrators gave twice the nitration rates compared to those obtained
in the pilot plant. It was then postulated that the benzene was not optimally
dispersed in the pilot nitrator, as had tacitly been assumed. Could it be that
the data generated in the kinetic studies were nothing more than the results of
a transient dispersion phenomenon? To test this hypothesis, some beaker-scale
experiments were carried out where benzene was pre-dispersed in sulfuric acid
for some time, and where the nitration reaction was initiated by “dumping in”
the stoichiometric amount of nitric acid. This addition sequence differed from
that used in the earlier kinetic studies, where benzene was injected into mixed
acid. As speculated, different nitration rates were observed, depending on the
length of time of benzene pre-dispersion. It was also noted that, when stopping
the agitator part-way through a run, the reaction would stop within seconds and
that the ensuing phase separation was very rapid, even though the benzene and
nitrobenzene droplets were very small.
It became apparent that the initial dispersion of benzene and the degree of
benzene-sulfuric acid coalescence do have significant effects on the nitration rate.
These effects are very difficult to quantify when scaling stirred-tank nitrators.
From these observations a concept for a new type of nitration reactor was
developed. Benzene would be uniformly added through a special inlet manifold,
which would disperse the benzene over the cross section of the mixed acid
4
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 inlet pipe. The dispersion would then pass through jet-impingement plates that
were spaced and sized to meet specific mixing requirements as the reacting
mixture passed through the plug-flow nitration train. It was soon recognized
that the nitration rate-controlling mechanisms changed as the reacting mixture
passed through the nitration train. A timely opportunity then offered itself to
demonstrate the jet impingement nitrator in a commercial first-generation plant,
and to subsequently retrofit and expand this plant.
In the early 1990’s the first grass-roots second-generation plant was
commissioned, and since that time 15 plants have been built, or are in the process
of being built, using this technology, with a combined capacity of about 5 million
MTPY. Details of the technology were publicized by the author of this paper in
a presentation given at the 209th National Meeting of the American Chemical
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch001

Society in 1995 (13).

The second-generation adiabatic process, which forms the basis of the modern
MNB plant, incorporates a number of important features which distinguish it from
the first-generation process. It uses lower nitric acid concentrations in the mixed
acid, which lowers the temperature rise through the nitration train, allowing
the MNB/Acid decanter to operate under a nitrogen blanket at atmospheric
pressure. This enhances the safety of the process in that the chance of developing
secondary exotherms, which are known to occur through a reaction between
MNB and sulfuric acid at temperatures above 180 °C in a pressurized nitrator,
is greatly reduced. In the first-generation process, the MNB/Acid decanter had
to be pressurized to prevent benzene flashing, and an emergency quench tank
had to be provided. The second generation process also uses a type of plug-flow
nitrator rather than the first-generation back-mixed nitrators in series (i.e., series
of CSTRs). This feature, together with the lower operating temperatures, results
in a 50% reduction of the nitrophenol generation rate. Several specific aspects of
the new process were patented (14, 15).

Operational Issues
In addition to energy efficiency and capital savings in the nitration train and
acid concentrators, there are many other important aspects which play a role in
MNB production economics, including plant reliability, safety, MNB purity, waste
treatment and disposal, and the impact on the environment.

Reliability

With reference to plant reliability, it has to be recognized that the MNB

plant represents just one step in the multi-step synthesis of MDI. The total
investment for a world-scale MDI complex is of the order of several hundred
million dollars, of which the MNB plant accounts for less than a quarter. Since
it is not desirable to hold large inventories of intermediates, the MNB plant
reliability and the on-stream factor are of great economic importance. Current
second-generation adiabatic MNB plants achieve on-stream factors of over 99%.
5
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The oldest first-generation MNB plant has now been in operation for 34 years,
and has similarly achieved very high on-stream factors.

Safety

A number of key safety aspects always have to be kept in mind during design
of an MNB plant, but also in the course of its long-term operation. These aspects
can be classified under the following headings:

Exotherms in the Nitration Train

Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch001

It has been found that significant exotherms occur in the nitration train if the
sulfuric acid/MNB mix reaches temperatures of about 180 °C in a pressurized
nitrator (16). The reactions between acid and MNB result in the formation of tar
and unknown gaseous by-products, which can cause overpressure in the nitration
train.

Exotherms in MNB Distillation

Several incidents have been reported where explosions occurred in the sump
of MNB distillation columns. The culprits have been leaky steam valves in
the reboiler during shut-downs leading to the slow concentration of unstable
impurities in the sump of the columns, or the accumulation of unstable sodium
salts of nitrophenols in the heat transfer area of the reboiler (17).

Nitric Acid/MNB

An explosion caused by a reaction between nitric acid and MNB leveled an

MNB plant in 1960 with a number of fatalities (18). It has been shown that nitric
acid/MNB mixtures can detonate (19).

Ammonium Nitrite

In some plants ammonia is used in MNB washing to remove nitrophenols

from the crude MNB. More commonly, a caustic solution is used for this purpose.
Introducing ammonia to a plant where NOX is produced as a by-product always
has to be viewed with concern. Unstable solid ammonium nitrite can form through
gas-phase reactions and settle in unexpected places (for example in “dead” vent
pockets), or can deposit in the casing of benzene pumps handling the benzene
recycle stream. Violent decompositions are known by the author to have occurred
in a number of installations.

6
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Benzene Handling

The major imported feedstock in an MNB plant is benzene, which is classified

as a carcinogen. Nitric acid is often produced on site. While methods for bulk
shipment and handling of benzene from refinery sources are well developed,
the disposal of the small benzene waste stream, which can be contaminated
with trace amounts of aliphatic compounds that were initially present in the
benzene feedstock, is of concern. This purge stream is commonly sent off-site for
disposal. Due consideration has to be given to handling this waste stream in an
environmentally safe manner. An aliphatics purge process which greatly reduces
benzene purge losses has been recently developed by NORAM.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch001

Crude MNB Purification

Crude MNB from a nitration train contains nitrophenols (NPh’s), dissolved
or entrained sulfuric acid, dinitrobenzene (DNB), dissolved nitric oxide (NO), and
excess benzene. The NPh’s and acid are commonly neutralized in an aqueous
washing system to form water-soluble salts. The adiabatic process operates with
a stoichiometric excess of benzene to ensure that essentially complete nitric acid
conversion is obtained in the nitration train. Excess benzene is removed by steam
stripping or vacuum distillation, and is then recycled back to the process. Aliphatic
impurities coming in with the benzene partially oxidize to carboxylic acids, but
also accumulate in the benzene recycle stream. NOX is stripped in the benzene
recovery process and in the sulfuric acid concentrator, and is further treated in the
NOX abatement area of the MNB plant. Some gas phase aniline processes require
very low concentrations of DNB (dinitrobenzene, <10ppm) in the feed MNB to
minimize catalyst poisoning. Therefore, MNB purification via distillation may be
required in some cases.

Waste Treatment
A number of liquid waste streams are generated in an MNB plant, including
wash water containing nitrophenolic compounds, an aliphatics-containing benzene
purge stream, a possible sulfuric acid purge, and a dinitrobenzene containing purge
stream. The following is a brief review of the status of current waste treatment
technologies.

Treatment of Nitrophenols

Nitrophenolic waste treatment was simple in the first two adiabatic plants;
one plant was permitted to use an existing “deep well” injection site, and the
other had available very large site-wide activated carbon beds. Nitrophenols
are toxic to the micro-organisms in biological treatment plants, even at low
concentrations. Treating nitrophenols in biological water treatment plants would
require massive dilution water volumes for a world-scale MNB plant, which is
usually not practical. Even then, there is doubt that some of the nitrophenolic
7
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 isomers are actually degraded. A hydrothermal process (20) has been developed
whereby nitrophenol in waste water is thermally degraded at high temperature
and pressure under slightly subcritical conditions. The effluent from this thermal
degradation process can be handled in biological treatment plants. NORAM has
built a dedicated biological treatment plant for nitrogen and BOD removal in the
effluent from an adiabatic MNB plant, using the thermal degradation process for
the pre-treatment of the nitrophenol-containing wash-water.
An alternative approach to dealing with nitrophenol wash-water is
incineration.

Aliphatics Purge
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch001

Various aliphatic compounds occur in trace quantities in nitration grade

benzene and are partially degraded through oxidation in the nitrators, but some
of the aliphatic species can also build up in the benzene recycle loop to a point
where they may have to be purged. As the purge contains both aliphatics and
a stoichiometric excess of benzene, as well as some MNB, this purge results
in benzene losses. If good quality benzene is available, a purge is usually not
required. However, benzene from certain supplies can contain stable species that
can build up in recycle streams, even if they are present at very low levels.

Sulfuric Acid Purge

In a typical world-scale plant, the sulphuric acid inventory in the nitration

train corresponds approximately to the hourly intake of nitric acid. Contaminants
present in the nitric acid will build up in the sulfuric acid loop, typically by a factor
of several hundred, until the contaminants reach a steady state concentration in the
process. Typically, this steady state is reached through purge from acid entrained in
the crude MNB, and acid spray entrainment in the vapor stream leaving the sulfuric
acid concentrator. Additional sulfuric acid purge for process reasons is normally
not required unless the feed nitric acid contains unusually high concentrations
of non-volatiles such as iron, calcium and lead. Sulfuric acid-containing wash-
water could possibly be recycled to the nitration train in order to reduce sulfate
concentrations in the aqueous plant effluent.

NOX Recovery

A patent (21) has been issued for a process operating at elevated pressure to
capture NOX generated in the nitration train for recycle as nitric acid. The benefit
is a slight improvement in the nitric acid yield, but more importantly, this process
substantially reduces the concentration of nitrites and nitrates in the effluent water,
and thus reduces water treatment costs.

8
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Dinitrobenzene Purge

In plants where MNB is distilled to remove heavy fractions and DNB, the
residue has to be purged from the still bottoms. This purge is typically incinerated
off-site.

Environmental
The total residual NOX and benzene vent rates from an MNB plant can be
kept below 1 kg/h, even in a world-scale plant, through conventional scrubbing
systems. This is, however, no longer sufficient. In new MNB plants the vent from
the plant is normally sent to a plant-wide thermal oxidizer.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch001

Patents and Technology Advancement

It is inevitable that a chemical having a long, steady and rapid growth profile,
such as MDI and its precursor MNB, will be of increasing commercial significance
to producers, and will, therefore, become the focus of dedicated R&D efforts.
In the case of MNB this is reflected in numerous patent applications (22, 23)
which have been filed over the past 40 years. Nowadays MNB technology is
advanced in small steps, through the know-how accumulated over the past 35 years
by plant designers and through the day-by-day experience of the plant operators.
Something new is learned from every project, and each new plant incorporates
incremental improvements.

Summary

• The world’s MNB plant capacity has grown almost 10-fold between 1974
and 2012, from less than one million MTPY to a capacity approaching
10 million MTPY, representing a growth rate of about 8% per year.
• Virtually all new MNB capacity has come from two generations of
adiabatic MNB processes. The first process, having been developed
in 1974, uses stirred nitrators in series under pressure, while the
second-generation technology, developed in 1988, uses plug flow
nitrators operating against an atmospheric back-pressure. Most of the
old isothermal plants have been shut down and scrapped.
• The driver for MNB growth has been the growth in MDI-based urethanes,
which were first commercialized in the 1960’s.
• Within the next few years China will account for about 50% of world
MNB production.
• The enormous size of world-scale MNB plants, some with capacities
in excess of 500 thousand MTPY (1600 MTPD), has necessitated
the refinement and optimization of MNB purification technologies,
and development of technologies to deal with by-products in an
environmentally sound manner.
• Benzene yields in the adiabatic process exceed 99.9% and nitric acid
yields exceed 99.7%.
9
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 • Economic technologies exist to degrade biotoxic nitrophenols such that
the aqueous effluents from an MNB plant can be treated in biological
treatment plants.
• Nitrite and sulfate concentrations in the effluent can be controlled to meet
site-specific regulations.
• In most plants there is typically only one aqueous effluent stream to be
dealt with, and a single plant vent, which normally is routed to a site-wide
thermal oxidizer.

References
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch001

1. Dickson, S. E.; Wahlen, J.; Kitai, A. Aniline and Nitrobenzene. In Chemical

Economics Handbook; Stanford Research Institute (SRI International):
Menlo Park, CA, 1981.
2. TDI/MDI; PERP 98/99-S8; Nexant ChemSystems: San Francisco, October
1999. This report is not in the public domain but can be purchased at
www.chemsystems.com/reports.
3. Lynch, M. K; Ryan, L. P. Nitrobenzene, Aniline, Methylenedianiline
Diisocyanate; PERP 2011-4; Nexant ChemSystems: San Francisco, May
2012. This report is not in the public domain but can be purchased at
www.chemsystems.com/reports.
4. Guenkel, A.; Prime, H.; Rae, J. Nitrobenzene via an adiabatic reaction.
Chem. Eng. 1981, 88 (16), 50.
5. Albright, L. F.; Hanson, C. Industrial and Laboratory Nitrations; Gould,
R. F., Ed.; ACS Symposium Series 22; American Chemical Society:
Washington, DC, 1976.
6. Alexanderson, V.; Trecek, J. B.; Vanderwaart, C. M. Adiabatic Process for
Nitration of Nitratable Aromatic Compounds. U.S. Patent 4,021,498, 1977.
7. Alexanderson, V.; Trecek, J. B.; Vanderwaart, C. M. Continuous Adiabatic
Process for the Mononitration of Benzene. U.S. Patent 4,091,042, 1978.
8. Castner, J. B. Nitration of Organic Compounds. U.S. Patent 2,256,999, 1941.
9. Dassel, M. W. Azeotropic Nitration of Benzene. U.S. Patent 3,928,475, 1974.
10. McCall, R. Azeotropic Nitration of Benzene. U.S. Patent 3,981,935, 1976.
11. Yen, Y. C.; Huang, F. H. Aromatic Amines; Process Economics Program
Report 76B; Stanford Research Institute: Menlo Park, CA, 1986.
12. McKetta, J.; Cunningham, W. A. Nitrobenzene and Nitrotoluene. In
Encyclopedia of Chemical Processing and Design; Marcel Dekker, Inc.:
New York, 1990; Vol. 31, p 165.
13. Guenkel, A. A.; Maloney, T. W. Recent Advances in Technology of
Mononitrobenzene Manufacture. In Nitration: Recent Laboratory and
Industrial Developments; Albright, L. F., Carr, R. V. C., Schmitt, R. J., Eds.;
ACS Symposium Series 623; American Chemical Society: Washington,
DC, 1996.
14. Guenkel, A. A.; Rae, J. M.; Hauptmann, E. G. Nitration Process. U.S. Patent
5,313,009, 1994.

10
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 15. Rae, J. M.; Hauptmann, E. G. Jet Impingement Reactor. U.S. Patent
4,994,242, 1991.
16. Silverstein, J. L.; Wood, B. H.; Leshaw, S. A. L. Case Study in reactor design
for hazards prevention. Loss Prev. 1981, 14, 78.
17. Badeen, C.; Turcotte, R.; Hobenshield, E.; Berretta, S. Thermal hazard
assessment of nitrobenzene/dinitrobenzene mixtures. J. Hazard. Mater.
2011, 188, 52–57.
18. Lodal, P. N. Distant replay: What can reinvestigation of a 40-year-old
incident tell you? A look at Eastman Chemical’s 1960 aniline plant
explosion. Process Saf. Prog. 2004, 23, 221–228.
19. Mason, C. M.; Van Dolah, R. W.; Ribovich, J. Detonability of the System
Nitrobenzene, Nitric Acid, and Water. J. Chem. Eng. Data 1965, 10,
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch001

173–175.
20. Larbig, W. Process for Working up Effluents Containing Nitro-Hydroxy-
Aromatic Compounds. U.S. Patent 4,230,567, 1980.
21. Brereton, C. M. H.; Guenkel, A. A. Nitration Process. U.S. Patent 5,963,878,
1999.
22. Hermann, H.; Gebauer, J. Process for the Nitration of Aromatic Compounds.
U.S. Patent 5,763,697, 1998.
23. Gillis, P. A.; Braun, H.; Schmidt, J.; Verwijs, J. W.; Velten, H.; Platkowski,
K. Process for Ring Nitrating Aromatic Compounds in a Tubular Reactor
Having Static Mixing Elements Separated by Coalescing Zones. U.S. Patent
6,506,949, 2003.

11
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 2

Effect of Reaction Conditions on the Formation

of Byproducts in the Adiabatic Mononitration
of Benzene into Mononitrobenzene (MNB)
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

Sergio Berretta*,1 and Brian Louie2

1NORAM Engineering and Constructors, Ltd., 200 Granville Street,
Suite 1800, Vancouver, B.C., V6C 1S4 Canada
2BC Research, Inc., 200 Granville Street,

Suite 1800, Vancouver, B.C., V6C 1S4 Canada

*E-mail: [email protected]

Two main impurities are made in the industrial production of

MNB. These impurities are nitrophenols and dinitrobenzene
(DNB). The formation rates of these impurities are significantly
affected by the initial reaction conditions. Understanding
these effects is an important first step in the continuous
on-going research aimed towards reducing the formation
of these impurities. However, very limited work has been
published on this subject. This paper presents the findings of
a study done by the authors, conducted in a laboratory setting,
examining the effect of relevant industrial operating conditions
on the formation rates of nitrophenols and DNB. The selected
operating conditions, which can usually be manipulated in most
industrial production MNB facilities, are: initial sulfuric acid
concentration, average reaction temperature, and nitric acid
concentration in the mixed acid feed.

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Introduction
Mononitrobenzene is produced industrially using a number of adiabatic
nitration technologies. The concept of adiabatic nitration was first introduced by
Castner (1) in the 1940s. However, it was not until the 1970’s when Alexanderson
(2, 3) proposed a new set of process conditions for the adiabatic technology that
Castner’s adiabatic nitration ideas finally led to a new commercial process.
Today, a few small industrial mononitrobenzene production facilities are still
built using the conditions proposed by Alexanderson. However, the majority
of new industrial mononitrobenzene adiabatic plants are now built based on the
process conditions proposed by Guenkel (4, 5) in the 1990’s. The main advantage
of Guenkel’s nitration process is a substantial reduction in the formation of
oxidation by-products, specifically nitrophenols, in the case of the nitration
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

of benzene. Since this last significant development, researchers at NORAM

Engineering have been working at further reducing the formation of these
by-products. As part of this challenge, and within the umbrella of a very large
research program, the authors ran a short test program with the aim of further
understanding how the process conditions described by the Guenkel process (4)
affect by-product formation. The findings from that work are the focus of this
paper.

Process Overview

Current “adiabatic” commercial processes for the manufacture of MNB

typically consist of a continuous addition of benzene to a mixture of sulfuric
acid and nitric acid, commonly called “mixed acid”. The sulfuric acid acts as a
catalyst disassociating the nitric acid into the reacting nitronium ion. It also acts
as a heat sink for the significant heat released in the formation of nitrobenzene. In
addition, it absorbs the water produced in the reaction. Following separation of
the organic and acid phases, the heat of reaction, which is mainly contained in the
large volume of sulfuric acid, is used to aid in the re-concentration of the sulfuric
acid in a flash evaporator.
Commercially, the nitration reaction of benzene follows Alexanderson’s or
Guenkel’s proposed operating conditions. Alexanderson proposed that MNB
should be commercially made most efficiently when the nitric acid concentration
in the mixed acid is 3 to 7.5 wt%, sulfuric acid concentration is 58.5 to 66.5
wt% with the balance as water. He also specifies that the temperature of the
initial mixed acid must be in the range of 80 °C to 120 °C (3). Compared to the
“isothermal” technologies of the day (i.e., prior art at the time), Alexanderson’s
conditions led to a significant reduction in the formation of the by-product DNB,
to within less than 500 ppm. However, these conditions still lead to significant
nitrophenol by-product formation. On the other hand, Guenkel proposed a
set of operating conditions using a nitric acid / sulfuric acid / water tertiary
diagram with the following limits: 82 wt% sulfuric acid and 18 wt% nitric acid,
55 wt% sulfuric acid and 45 wt% water, and 100 wt% sulfuric acid, and with
the additional constraint that the initial mixed acid temperature must be in the
range of 97 °C to 120 °C (4). Guenkel’s invention, which led to a significant
14
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 reduction in nitrophenol formation, is believed to be characterized by a mixed acid
composition in which nitric acid is more fully dissociated to nitronium ion leading
to an increase in the reaction rate. As shown by Guenkel (4), the formation of
by-product nitrophenol, for the conditions proposed, is in the range of 1700 ppm.

Description of the Test Program

Based on the referenced works (3, 4) and the objectives of the broader research
program, it was decided to focus the test program on the effect of the following
three process variables in the formation of nitrophenols and DNB by-products in
the production of MNB:
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

• Sulfuric Acid Concentration

• Nitric Acid to Sulfuric Acid Ratio (i.e., concentration of nitric acid in
mixed acid)
• Reaction Average Temperature

The aim of this work was to manipulate these process variables through a set
of experiments and measuring their effects on the formation of by-products. To
minimize the number of experiments, the study was done based on a factorial style
analysis.
A typical factorial designed experiment includes experimental runs for all
combinations of settings, both high and low, of the variables to be studied. The
minimum number of experiments to complete a study is then defined as 2N where N
is the number of variables to be studied. Since there were three variables of interest
included in the test program, then the minimum number of required experiments
was 8.
Based on objectives of the broader research program, it was decided that
changes on the process variables of interest would be limited to the following
ranges:

• Sulfuric Acid Concentration: 62 to 72 wt%

• Nitric Acid to Sulfuric Acid Ratio (mass basis): 0.022 to 0.033
• Reaction Average Temperature: 70 to 100 °C

The selected experimental conditions are graphically shown in Figure 1.

Overall, eight experiments were required to cover the “factorial cube”.
However, it was arbitrarily decided to add an additional two experiments to better
define trends, taking the total number of experiments to ten. The experiments
were repeated twice to check the reproducibility of the results, for a total of
twenty experiments.
All reactions were performed using an 8% molar excess of benzene, relative
to the nitric acid.
Table I summarizes the targeted operating conditions for each experiment.

15
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

Figure 1. Experimental Conditions. (Courtesy of Sergio Berretta).

Table I. Proposed Experimental Conditions for Each Run

Experiment Nitric Acid to Sulfuric Sulfuric Acid Reaction Average
No. Acid Ratio (mass basis) Concentration (wt%) Temperature (°C)
1 0.022 62 70
2 0.022 62 70
3 0.022 62 100
4 0.022 62 100
5 0.022 72 70
6 0.022 72 70
7 0.022 72 100
8 0.022 72 100
9 0.022 67 85
10 0.022 67 85
11 0.033 62 70
12 0.033 62 70
13 0.033 62 100
14 0.033 62 100
15 0.033 72 70
16 0.033 72 70
Continued on next page.

16
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table I. (Continued). Proposed Experimental Conditions for Each Run
Experiment Nitric Acid to Sulfuric Sulfuric Acid Reaction Average
No. Acid Ratio (mass basis) Concentration (wt%) Temperature (°C)
17 0.033 72 100
18 0.033 72 100
19 0.033 67 85
20 0.033 67 85

Experimental Set Up
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

All experiments were performed in the pressurized glass reactor (PGR) shown
in Figure 2.

Figure 2. Experimental Apparatus. (Courtesy of Sergio Berretta).

17
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The PGR consists of a 3” I.D., 450 ml, hemispherical bottomed glass reactor
clamped underneath a ¼” thick stainless steel plate. A gasket ensures a tight seal
between the reactor and the metal plate. Reactor contents are mixed using a 4-
bladed Teflon® impeller that is connected with a shaft seal to an overhead variable
speed mixer. Tantalum baffles attached to the top plate improve mixing while a
Teflon® covered Type J thermocouple was used to measure the temperature of
the solution. Nitrogen gas was used to pressurize the reactor and prevent benzene
boiling and also to inject liquids into the reactor from the overhead reservoirs. The
initial heat input was via a manually- controlled heating band.

Experimental Procedure
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

In a typical nitration experiment the sulfuric acid and the nitric acid were
charged to the PGR., The reactor was then attached to the top plate. An overhead
feed reservoir was then filled with benzene. The acid mixture was pressurized to
40 psig, then mixed and heated to the required initial temperature. The mixer was
then started. Benzene was injected at a predetermined rate from the feed reservoir
through a ¼” Teflon® dip tube into the high intensity mixing zone of the reactor.
Once the reaction was deemed to be complete, as assessed by the temperature
rise of the mixture, the reactor was depressurized and the contents poured into a
Pyrex® bottle. A portion of the organic layer, which quickly forms a layer on top
of the acid, was removed with a pipette into a separate vial for analysis. Then a
sample of the acid was also removed with a pipette into a separate vial and also sent
out for analysis. The samples were maintained in a refrigerator at 4 °C until their
time of analysis. To avoid possible contamination, the experimental apparatus was
rinsed thoroughly and dried before the next experiment.
Each of the two samples per experiment (i.e., MNB and acid samples) was
analyzed for:

• Picric acid concentration

• 2,4 dinitrophenol concentration
• 2,6 dinitrophenol concentration
• 2-mononitrophenol concentration
• 4-mononitrophenol concentration
• dinitrobenzene (all isomers)

Each sample was analyzed twice to check the reproducibility of the results.

Results and Discussion

The actual observed reaction conditions for each experiment varied somewhat
from the “targeted” conditions (Table I). Specifically, the “reaction average
temperature” variable proved hard to control. Table II presents the actual observed
experimental reaction conditions for each test.

18
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table II. Actual Reaction Conditions for Each Run
Sulfuric Acid
Experiment Nitric Acid to Sulfuric Reaction Average
Concentration
No. Acid Ratio (mass basis) Temperature (°C)
(wt%)
1 0.022 62 70
2 0.022 62 68
3 0.022 62 102
4 0.022 62 98
5 0.022 72 72
6 0.022 72 72
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

7 0.022 72 102
8 0.022 72 102
9 0.022 67 86
10 0.022 67 86
11 0.033 62 61
12 0.033 62 61
13 0.033 62 95
14 0.033 62 95
15 0.033 72 70
16 0.033 72 70
17 0.033 72 99
18 0.033 72 100
19 0.033 67 83
20 0.033 67 83

The test results are summarized on Figures 3 and 4. The figures show the
average analytical result for each combination of operating conditions.
Let us first analyze the experimental results with respect to nitrophenol
formation. The data in Figure 3 was introduced into Microsoft Excel and its
multivariable regression analysis tool was used to develop a correlation between
nitrophenol formation and the process variables: reaction average temperature,
sulfuric acid concentration, and nitric acid to sulfuric acid ratio. Equation 1 is the
output from the analysis. For convenience in the following equations 1 and 2, the
units for nitric acid to sulfuric acid ratio were changed to concentration of nitric
acid in the sulfuric acid in wt%.

19
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

Figure 3. Effect of Reaction Average Temperature, Sulfuric Acid Concentration,

and Nitric Acid to Sulfuric Acid Ratio on Nitrophenol Formation. (Courtesy
of Sergio Berretta). (see color insert)

where Y: concentration of nitrophenol in produced MNB, ppm

T: reaction average temperature, °C
S: sulfuric acid concentration, wt%
R: nitric acid concentration in mixed acid, wt%

The validity of equation 1 is bounded within the variable limits presented

under the heading “Description of Test Program”. The statistical errors on equation
1 are presented in Table III. The produced MNB under Y is defined as the total
MNB produced in the reaction which is the addition of both MNB in the organic
phase and dissolved MNB in the acid phase.
20
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

Figure 4. Effect of Reaction Average Temperature, Sulfuric Acid Concentration,

and Nitric Acid to Sulfuric Acid Ratio on DNB Formation. (Courtesy of Sergio
Berretta). (see color insert)

Table III. Statistical Errors of Equation 1

Coefficient Standard Error
Intercept -3900 485
T 31.72 2.42
S 32.67 8.13
R 39.07 50.69

21
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Equation 1 indicates that for every 1 °C change on reaction average
temperature, nitrophenol formation changes proportionally by 32 ppm. The
standard error of 2.42 on the coefficient of 31.72 suggests that the correlation
matches the experimental data very closely, as graphically shown in Figure 5.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

Figure 5. Effect of Reaction Average Temperature on Nitrophenol Formation.

(Courtesy of Sergio Berretta). (see color insert)

Similarly, Equation 1 indicates that for every 1 wt% change in sulfuric acid
concentration, nitrophenol formation changes proportionally by 33 ppm. The
standard error of 8.13 on the coefficient of 32.67 suggests that the correlation
matches the experimental data very closely, as graphically shown in Figure 6.
For the effect of nitric acid / sulfuric acid ratio on nitrophenol formation,
equation 1 provides a poor fit relative to the experimental data (i.e., standard error
of 50.69 on the coefficient of 39.07). In fact, the magnitude of the standard error
can lead to either a positive or negative coefficient as the multiplier of this variable,
meaning that it cannot be concluded whether the nitric acid / sulfuric acid ratio
has a proportional, inversely proportional, or any effect at all, on nitrophenol
formation.
Let us now analyze the experimental results in regards to effects on DNB
formation. The data in Figure 4 was introduced into Microsoft Excel and its
multivariable regression analysis tool was used to develop a correlation between
DNB formation and the process variables: reaction average temperature, sulfuric
acid concentration and nitric acid to sulfuric acid ratio. Equation 2 is the output
from the analysis.
22
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

Figure 6. Effect of Sulfuric Acid Concentration on Nitrophenol Formation.

(Courtesy of Sergio Berretta). (see color insert)

where Z: concentration of DNB in produced MNB, ppm

T: reaction average temperature, °C
S: sulfuric acid concentration, wt%
R: nitric acid concentration in mixed acid, wt%

The validity of equation 2 is bounded within the variable limits presented

under the heading “Description of Test Program”. The statistical errors on equation
2 are presented in Table IV.

Table IV. Statistical Errors of Equation 2

Coefficient Standard Error
Intercept -9878 5193
T 35.98 22.52
S 123.73 71.49
R -303.71 580.59

23
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Equation 2 indicates that for every 1 °C change on reaction average
temperature, DNB formation changes directly proportional by 36 ppm. However,
the standard error is significant (i.e., 22.52), indicating that the effect on DNB
formation may in reality be either mild or significant, but directly proportional
nevertheless. The poor fit is graphically shown in Figure 7.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

Figure 7. Effect of Reaction Average Temperature on DNB Formation. (Courtesy

of Sergio Berretta). (see color insert)

Similarly, Equation 2 indicates that for every 1 wt% change on sulfuric acid
concentration, DNB formation changes proportionally by 124 ppm. However,
the standard error is significant at 71.49. Regardless, the effect of sulfuric acid
concentration on DNB formation is significant overall, and directly proportional.
Figure 8 shows the fit of the correlation relative to the experimental data.
For the effect of nitric acid / sulfuric acid ratio on DNB formation, equation 2
provides a poor fit relative to the experimental data (i.e., standard error of 580.59
on the coefficient of -303.71). In fact, the magnitude of the standard error can lead
to either a positive or negative coefficient as the multiplier of this variable, meaning
that it cannot be concluded whether nitric acid concentration in the sulfuric acid
has a proportional, inversely proportional, or any effect at all, on DNB formation.
24
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

Figure 8. Effect of Sulfuric Acid Concentration on DNB Formation. (Courtesy

of Sergio Berretta). (see color insert)

Conclusions

The results from the experimental work show that both average reaction
temperature and sulfuric acid concentration affect the reaction rates of both
nitrophenol and DNB formations. However, these two process variables also
affect the reaction rate of MNB production. It is known that increasing the
reaction temperature or sulfuric acid concentration increases the reaction rate of
MNB formation (6).
At a high level, the findings from this work suggest that in the industrial
production of MNB a substantial reduction in by-product formation, relative to
current levels, is possible, by reducing the reaction average temperature and/or
sulfuric acid concentration. However, this benefit must be weighed against the
drop in the reaction rate of MNB production.
The expected change in nitrophenol and DNB formations due to manipulation
of temperature and sulfuric acid concentration can be approximately predicted
through equations 1 and 2 respectively.

25
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 References
1. Castner, J. B. U.S. Patent 2,256,999, 1941.
2. Alexanderson, V.; Trecek, J. B.; Vanderwaart, C. M. U.S. Patent 4,021,498
1977.
3. Alexanderson, V.; Trecek, J. B., Vanderwaart, C. M. U.S. Patent 4,091,042
1978.
4. Guenkel, A. A.; Hauptmann, E. G.; Rae, J. M. U.S. Patent 5,313,009 1994.
5. Guenkel, A. A.; Maloney, T. W. Recent Advances in the Technology of
Mononitrobenzene Manufacture. In Nitration: Recent Laboratory and
Industrial Developments; Albright, L. F., Carr, R. V. C., Schmitt, R. J., Eds.;
ACS Symposium Series 623; American Chemical Society, Washington, DC,
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch002

1996; Chapter 20, pp 223−233.

6. Marziano, N. C.; Tomasin, A.; Tortato, C.; Zaldivar, J. M. Thermodynamic
nitration rates of aromatic compounds. Part 4. Temperature dependence in
sulfuric acid of HNO3-NO2+ equilibrium, nitration rates and acidic properties
of the solvent. J. Chem. Soc. 1998, 2, 1973–1982.

26
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 3

Adiabatic Nitration for Mononitrotoluene

(MNT) Production
M. Gattrell*,1 and B. Louie2
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

1NORAM Engineering and Constructors, Ltd., Vancouver, B.C., Canada

2B.C. Research Institute, Burnaby, B.C., Canada
*E-mail: [email protected]

Adiabatic nitration has revolutionized mononitrobenzene

(MNB) production, but has not similarly impacted the
production of other nitro-aromatics. The issues related to
changing to adiabatic nitration are discussed by comparing the
nitration of toluene versus benzene using literature data and
adiabatic stirred reactor nitration tests. The topics discussed
include nitration rates, isomer distribution, and by-product
formation. The homogeneous chemical reaction rate for the
nitration of toluene is faster than benzene, but the overall
rate for interphase mass transport and reaction is found to be
fairly similar. The MNT isomer distribution is found to be a
function of sulfuric acid strength, temperature, and nitric acid
strength. The easier oxidation of toluene versus benzene results
in a greatly increased number of oxidation by-products. This
presents analytical difficulties in quantifying total by-products.
Deeply colored, oxidized by-products can also accumulate
in the recycled spent acid producing so called “black acid”.
However, once understood, the issues related to implementing
adiabatic MNT production appear manageable.

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Introduction
Interest in continuous adiabatic nitration versus continuous isothermal
nitration is due to the large energy savings possible by using the exothermic heat
of the nitration reaction for the re-concentration of the spent sulfuric acid. The
differences between the two processes can be seen by comparing the process flow
diagrams for the production of MNT from toluene in Figure 1.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

Figure 1. Simplified flow diagrams for the nitration of toluene to produce MNT
via the commercial isothermal process and a potential adiabatic process.
In the isothermal process (1–4), re-concentrated sulfuric acid is combined
with feed nitric acid to generate “mixed acid” or “nitrating acid”. This nitrating
acid is mixed with toluene in a series of cooled, stirred reactors to create a two-
phase liquid-liquid dispersion. Good mixing is required in the reactors both to
create a large interfacial area between the two phases for the reaction to occur
and to provide good heat transfer between the reaction mass and the reactor’s
28
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 cooling coils. Ample, reliable cooling is critical to remove the heat of reaction
to avoid a potentially catastrophic thermal runaway (5). After the reaction has
completed, the two-phase mixture is fed to a decanter where the phases are allowed
to separate into a crude organic product stream and a “spent acid” stream. The
spent acid phase consists of the starting sulfuric acid diluted by water generated
in the nitration reaction and water that enters with the nitric acid feed. This spent
acid is heated and the water is evaporated under vacuum to re-concentrate the
acid. The high temperature involved in this re-concentration step (~130-180 °C)
also helps to decompose or strip out organic contaminants that may build-up in the
acid. One advantage of this approach is that the nitration temperature and the acid
re-concentration temperature can be independently optimized. Two disadvantages
are the danger of thermal runaway if the reactor cooling malfunctions and the large
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

energy input required to re-concentrate the spent sulfuric acid.

For an adiabatic process (6), the toluene would also be mixed with the nitrating
acid, but the heat of reaction would not be allowed to dissipate. The reactor(s)
can either be a series of stirred tank reactors or a plug flow reactor with static
mixing elements. The temperature of the reaction mixture will then rise due to the
heat of reaction, with the temperature rise controlled to a safe limit by controlling
the amount of nitric acid and organic reactant added to the recirculating sulfuric
acid. The hot two-phase mixture exits the reactor(s) and continues to the decanter
where it is allowed to separate into a crude organic product stream and a “spent
acid” stream. For the adiabatic process this resulting hot spent acid is then flashed
under vacuum to re-concentrate the sulfuric acid. Most of the energy for this flash
re-concentration comes from the heat contained in the hot spent acid, with very
little external energy required compared to the isothermal process. In this way
adiabatic nitration uses the heat of the nitration reaction to provide the majority of
the energy for the sulfuric acid re-concentration.
A drawback of the adiabatic approach is that it links the nitrator starting
temperature to the acid re-concentration temperature which results can result in a
higher nitration reactor temperature and a lower acid re-concentration temperature
than in an isothermal process. A higher nitration temperature can cause increased
by-products. A lower acid re-concentration temperature decreases the destruction
and stripping of contaminants from the spent acid. Together these effects raise
concerns about the build-up of contaminants in the acid recycling loop over time.
When adiabatic nitration of benzene to MNB was implemented, these
problems were not found to be significant. While the adiabatic benzene nitration
reaction was carried out at higher temperatures which should increase by-products,
the higher temperature also led to lower reaction times. Further, limitations on the
safe temperature rise (i.e. the safe maximum acid loop temperature) combined
with a much lower cost for acid recycling drove a shift to reaction conditions
using less total reaction per pass (i.e. with lower concentrations of nitric acid and
benzene in the nitrator feed). These new conditions (7–9) allowed for adiabatic
operation with little increase in by-products. The shorter reaction times and
the lack of need for cooling coils with adiabatic operation also allowed a shift
to plug flow type reactors with static mixing elements, resulting in improved
volumetric efficiency. Overall, these factors have resulted in the adiabatic route
now dominating commercial MNB production (10).

29
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 However, while it should be possible to adiabatically nitrate many
compounds, so far the technology is not widely used beyond MNB. To understand
some of the issues, this paper will investigate the example of toluene nitration to
MNT.

Experimental
The adiabatic toluene and benzene nitration tests were carried out in a batch
manner using the 500 mL pressurized, insulated, stirred glass reactor shown in
Figure 2. During operation the reactor was kept pressurized with nitrogen gas at
~3 barg to prevent boiling of the toluene or benzene. In the tests about 400 g of a
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

mixture of sulfuric and nitric acids in water was first put in the reactor and a heating
rate chosen to bring it to the target starting temperature. To begin the experiment,
room temperature toluene or benzene was then injected into the stirred reactor
using nitrogen pressure (over ~1-2 s). The reaction was followed by monitoring
the reactor temperature with time, with an example curve shown in Figure 3. In
Figure 3, the acid temperature started at 90 °C. Room temperature toluene was then
injected at time 0 causing cooling, but this was followed by a rapid temperature
increase due to the exothermic nitration reaction. At 3.5 minutes, the temperature
rise leveled off indicating that the reaction was essentially over, though the reactor
was run until 5-6 minutes to ensure complete reaction. Note, that the measured
temperature rise is less than the theoretical value due to the thermal mass of the
glass reactor and some heat losses. After the reaction was finished the reactor
contents were emptied into a graduated cylinder and left to separate into organic
and acid phases at room temperature.
The organic phase was analyzed for residual reactant and products by
gas chromatography using a flame ionization detector (GC-FID). The organic
phase was also extracted with 0.1 M aqueous NaOH to recover acidic oxidation
by-products. This extract was analyzed using high performance liquid
chromatography (HPLC) using a C-18 column at 35 °C. The mobile phase was
pumped at 0.4 mL/min and gradient elution was used starting from a 1 wt%
acetic acid-sodium acetate buffer (pH ~4.5) and changing to acetonitrile over 12
minutes. Detection used a UV absorption detector using a range of wavelengths
(270, 320, 360 and 450 nm) to give the best signal to noise for the individual
compounds (example chromatograms are presented later in the paper).

30
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

Figure 2. Batch type, stirred, insulated, pressurized glass reactor used for the
adiabatic nitration screening tests presented in this paper.

Figure 3. An example temperature-time curve obtained with the reactor shown

in Figure 2. Example is for: 64.6 wt% H2SO4, 3.0 wt% HNO3, 1.10 mole
toluene/mole nitric acid, and stirring at 1200 rpm (~16 watts/liter, W/L). The
starting acid temperature was 90 °C and the injected toluene was at room
temperature.
31
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Results and Discussion
Reaction Rate

The toluene nitration rates observed from the temperature rise data were
not much different from those for benzene nitration. This might be considered
to be surprising since the reported chemical reaction rate constants for toluene
nitration at room temperature in 63 to 73 wt% sulfuric acid are from 11 to 18
times higher than those for benzene (11). However, under normal industrial
conditions the nitration of toluene can be described as a fast 2 phase reaction
(12–14) with the overall reaction rate controlled both by mass transport and
chemical reaction kinetics. Figure 4 shows a diagram of the expected shape
of the toluene concentration profile in the acid phase adjacent to a drop of the
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

organic phase. Toluene is assumed to dissolve in the acid phase at the interface at
a concentration roughly determined by its pure component solubility in the acid
phase and its mole fraction in the organic phase. Under fast kinetic conditions
this toluene then reacts while it diffuses into the acid phase, becoming essentially
completely reacted before reaching the bulk acid phase (i.e. within the mass
transfer boundary layer). If it is assumed that the small amount of dissolved
toluene does not significantly perturb the nitric acid concentration profile, then
the overall reaction rate can be written as given in equation 1.

Figure 4. The expected concentration profile shape for the fast 2 phase reaction
of toluene as it diffuses into the mixed acid phase (13, 14).

32
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Where:
a is the organic phase–acid phase interfacial area, which is influenced by
factors such as: mixing intensity, interfacial tension, coalescence rate and organic
phase volume fraction.
CTol* is the solubility of pure toluene in the mixed acid. This is influenced by
temperature and the mixed acid composition.
xTol is the mole fraction of toluene in the organic phase, which will change as
the reaction proceeds.
kC is the apparent nitration chemical reaction rate constant (i.e. in terms
of nitric acid), which is influenced by temperature and sulfuric acid (catalyst)
strength.
CHNO3 is the concentration of nitric acid in the mixed acid.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

DTol is the diffusion coefficient of toluene in the mixed acid, which is a

function of temperature and the mixed acid composition.

It can be seen from equation 1 that an increase in the chemical reaction rate
constant (kC) by 11 to 18 times will only increase the observed overall rate in
the fast two-phase reaction by the square root of these values and so a factor of
3.3 to 4.2. In addition, the room temperature solubility of toluene in 70 to 76
wt% sulfuric acid (CTol*) is about 3.1 to 4.1 times lower than for benzene (15)
which, based on equation 1, will essentially cancel the reaction rate increase. The
diffusion coefficient for toluene is also slightly lower than for benzene (14). The
exact balance of these different factors will vary with temperature and sulfuric
acid strength, but the resulting effect is that the overall toluene nitration rate is
not greatly different than that for benzene nitration. Thus, for adiabatic MNT
production, a nitrator with a similar size and mixing intensity (pressure drop) to
those used for MNB production could be used.

Isomer Distribution
When nitrating toluene, three isomers are produced as shown in Figure 5. If
MNT is the desired end nitration product, the different isomers can be separated
and sold, with different isomers having different market values. However, the
bulk of MNT production is for further nitration to dinitrotoluene (DNT), which
is used for the production of toluene diisocyanate (TDI) based polyurethanes (1).
For polyurethane production, vicinal DNT isomers (2,3 and 3,4 DNT) must first
be removed from the DNT mixture. Thus for DNT production, forming 3-MNT
during toluene nitration not only represents a waste of starting chemicals, but also
adds costs for by-product removal and disposal. The 3-MNT is also undesirable
for the production of trinitrotoluene (TNT) (16). Unfortunately, the yield of 3-
MNT increases with temperature, and adiabatic nitration would operate at higher
temperatures than isothermal nitration.
Figure 6 shows a compilation of mononitrotoluene isomer distribution data
under different conditions from our adiabatic nitration work reported here, as
well as literature adiabatic (17) and isothermal nitrations (18, 19). (Note that for
the adiabatic tests, average temperatures were plotted.) In spite of differences
in the reaction conditions, the yield of 3-MNT can be seen to clearly increase
33
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 with temperature showing that it is a dominate factor. As well as temperature,
the isomer distribution is also influenced by the make-up of the toluene hydration
sphere (i.e. the arrangement of the molecules around the toluene when it dissolves
in the acid phase) and so the acid phase composition will have some effect.
Multi-variable regression was therefore used to correlate all the data in Figure
6 with temperature, sulfuric acid concentration and nitric acid concentration.
Example plots of the multi-variable fits for 4-MNT are shown in Figure 7, and the
ranges over which the variables were correlated can also be seen. The results of
the correlations are given in equations 2 to 5.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

Figure 5. Different isomers produced by toluene nitration to MNT and the main
paths for further nitration of MNT to DNT. (Percentages are typical values for
isothermal nitration (1)).

Figure 6. Data for the isomer distribution of toluene nitration versus temperature.
Filled symbols: this work, open symbols literature (17–19).
34
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

Where:
S = starting sulfuric acid (wt%), T = average temperature (°C) and N = starting
nitric acid (wt%), with the fitted ranges shown in Figure 7.

From the coefficients of the fitted equations 2 to 4, it can be seen that

increased sulfuric acid strength decreases the percent of 2-MNT while increasing
the percent of 4-MNT, with little clear effect on the percent of 3-MNT. Increased
temperature decreases the percent of 2-MNT and increases the percent of 3-MNT,
with a smaller increase in the percent of 4-MNT. Increased nitric acid strength
decreases the percent of 2-MNT, increases the percent of 3-MNT and increases the
percent of 4-MNT. Though, the high relative uncertainty on the fitting coefficient
for 3-MNT with nitric acid strength suggests that further work would be required
to refine this result. Other correlations are available in the literature which report
partial fits for only one isomer or only the 2-MNT/4-MNT ratio, but give similar
trends (16, 20). Data is also available at higher and lower sulfuric acid strengths
(18, 21), however this was not included because when looking at a wider set
of conditions it appears that non-linear fits would be required (see for example
Barnett et. al Figure 3 (18)).
Based on these results, changing from an isothermal to an adiabatic nitration
process, resulting in a higher nitration reaction temperature, will cause an increase
in 3-MNT. This would detract from the value of the energy savings possible with
adiabatic production of MNT if DNT was the desired final product. Thus, the
operation of such an adiabatic MNT process would need to be designed to operate
with as low a temperature as possible, essentially requiring as low a vacuum as
possible in the acid re-concentration stage to produce a satisfactory strength of
sulfuric acid at lower temperatures. From the correlation the use of a minimal
amount of nitric acid might also be helpful to minimize 3-MNT. Another option
might be to replace the sulfuric acid with a different type of acid. It has been
reported that use of phosphoric acid (21) or acetic acid (16) changes the isomer
distribution. However, most other acids would be less practical than sulfuric acid
for an industrial nitration process.

35
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

Figure 7. Correlation plots for fitting 4-MNT as a linear function of sulfuric acid,
temperature and nitric acid over the indicated ranges.

Toluene Nitration By-Products

For toluene, the presence of the methyl group on the benzene ring makes the
ring easier to oxidize. In DNT production, it is reported that the isothermal toluene
to MNT stage can generate on average about 0.72 wt% nitrocresols (4) as compared
to about 0.2 wt% nitrophenols for MNB production (10). Further, the methyl
group itself can be oxidized leading to benzoic acid products. This is an important
difference because this makes MNT less stable towards oxidation than MNB. This
is a factor that needs to be considered in the design of the nitration reactor system.
Some of the reaction routes and oxidation by-products from the nitration of toluene
36
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 are shown in Figure 8. These compounds can be further oxidized and different
isomers of these compounds will also be produced. Quantification of such a large
number of oxidation by-products is a considerable analytical challenge. These
oxidized, more water soluble by-products will also tend to partition into the spent
acid phase during the product decanting. The large amount of such by-products
generated during toluene nitration therefore raises issues about the build-up of such
by-products in the acid loop with repeated acid recycling.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

Figure 8. Diagram showing some toluene oxidation by-products. Further

oxidation of these by-products will lead to more possible compounds. Also, other
isomers will be produced.

Quantification of the by-products from benzene nitration is commonly done

by HPLC. The organic phase can be extracted with caustic solution to recover
oxidation by-products for analysis and the spent acid phase can be neutralized
37
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 and analyzed directly. A typical MNB extract chromatogram is shown in Figure 9
where the main mono-, di- and tri- nitrophenols can be separated and quantified.
The much larger variety of oxidation by-products for toluene nitration is apparent
in the lower chromatogram in Figure 9. A limited number of the expected
oxidation by-product compounds resulting from the nitration of toluene can be
obtained and used to prepare standards to identify and quantify some peaks.
Using these available standards to track different families of possible by-products
can provide a “window” into the oxidation reactions, but falls short of properly
quantifying the total amount of oxidation occurring.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

Figure 9. Chromatograms showing the increased complexity of monitoring

by-products for toluene versus benzene nitration. (UV detection at 270 nm
shown in these plots).

The importance of measuring the total amount of oxidation occurring can be

seen when comparing results from our 500 mL reactor done with different starting
temperatures. As the nitration temperature is increased from 80 to 100 °C, the
resulting organic and spent acid phases become progressively darker indicating
a likely increase in oxidation by-products. However, simply measuring some of
the cresols and benzoic acids shows a peak production at 90 °C (see Figure 10,
peaks in the 6 to 11 minute range). However, one can see a steady increase in
unresolved peaks of more water soluble compounds in the 1 to 4 minute range.
Thus, it appears that for the test starting at 100 °C, the rate of decomposition of
cresols has increased faster than their rate of formation. This would result in a
decrease in the cresol concentrations being tracked by the HPLC, but not because
of decreased cresol formation.
38
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

Figure 10. Chromatograms from the analysis of the alkaline extract of the crude
MNT product for various nitration starting temperatures. (UV detection at 320
nm shown in these plots).

One approach that has been used to measure the total amount of the oxidative
side reactions is to measure the production of nitrous acid (22), which is generated
from the nitric acid by the oxidation reactions. An alternative method evaluated in
our work was to titrate the crude MNT product with caustic. This involved placing
the crude MNT sample in a beaker with an equal volume of water and vigorously
mixing to create a well dispersed two-phase solution. This was then slowly titrated
with an aqueous NaOH solution, allowing the two phases to equilibrate after each
addition of caustic. This results in a titration curve as shown in Figure 11. From
such a curve one can measure the total amount of water soluble acid compounds
and different classes of compounds can be crudely quantified. The approach is
shown in Figure 11 where the shallow plateaus in the titration curve have been
assigned to certain classes of oxidation products. The assignments are based on
the pKa’s of the different families of oxidation by-products, and because this is an
extractive titration, the plateau positions also include approximate corrections for
the compound’s acid form partition coefficient (~1 pH unit for the nitrobenzoic
acids and ~2 pH units for the nitrocresols). Note, however, that the large number
39
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 of possible isomers will give a range of pKa values and extraction coefficients for
each family of compounds, and so the identification of the by-products by this
method is only approximate. However, it does provide a useful cross-check and
compliment to the HPLC analysis.

Black Acid

A second issue with toluene nitration is the possibility of producing a dark

colored spent acid referred to as “black acid” (4, 23, 24). Black spent acid can
cause foaming in the acid re-concentrator, when the re-concentrated acid is re-used
it can cause emulsions in the decanter and, eventually, after repeated re-use can
cause tar deposits in the acid loop. In isothermal MNT production, black acid is
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

formed if excess toluene occurs (i.e. the moles of toluene present are sufficiently
greater than the moles of nitric acid so that the nitric acid is essentially consumed).
This is one reason why isothermal MNT production typically uses excess nitric
acid (1, 3, 4). Though, even with excess nitric acid nitration, a follow-up toluene
solvent extraction of the spent acid is sometimes used to recover dissolved MNT
and some of the unreacted nitric acid, which if not carefully operated can lead to an
excess toluene situation and black acid. This is different from benzene nitration. In
modern adiabatic benzene nitration, the nitration reactor is run with excess benzene
in order to react essentially all the nitric acid. The unreacted excess benzene can
be easily recovered and recycled, resulting in very good utilization of the feed
chemicals, and no problems with black acid occur (2, 10).

Figure 11. Example titration curve for crude MNT product with plateaus
tentatively labeled. The plateau pH depends on the compound’s pKa
and its partition coefficient between MNT and water. The plateau width
depends on the compound’s concentration. NBAs = nitrobenzoic acids,
NHBAs = nitrohydroxybenzoic acids, DNCs = dinitrocresols and MNCs =
mononitrocresols.
40
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The formation of black acid is believed to be related to reactions of nitrous
acid with organic material. Black acid formation can be avoided by the removal
of nitrous acid from the spent acid (23). Black acid is also suppressed in the
presence of nitric acid (24). In considering the possible reactions of nitrous acid,
a key intermediate is likely the nitrosonium ion-toluene complex. This is one of
a family of dark orange to deep red colored complexes that are formed on mixing
aromatic compounds with nitrosonium ion (25, 26). In our testing, mixing benzene
or toluene with 0.5 mM NaNO2 in 70 wt% sulfuric acid immediately produces a
dark orange-red color. If nitric acid is also present, the solution initially turns
orange but then lightens to a yellow color over a few minutes. It is thought that
the nitrosonium ion-aromatic compound complex can be oxidized by nitric acid to
a nitronium ion-aromatic compound, that then would react to form a nitroaromatic
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

(25, 26). Under excess organic conditions, where the nitric acid is fully consumed,
the nitrosonium ion-aromatic compound complex can presumably react via other
pathways leading to the highly polar, sulfuric acid soluble, surface active, deeply
colored compounds that are characteristic of black acid. One such route would
be the oxidation of the aromatic compound by the nitrosonium ion which would
lead to nitric oxide and oxidized by-products such as cresols. A second reaction
route may be via the formation of nitrosotoluene (27). In the absence of nitric
acid the nitrosotoluene might react further through disproportionation reactions
or through redox reactions with easily oxidized organic compounds to produce a
variety of compounds such as hydroxyl amines and azo compounds (28). As azo
compounds tend to be highly colored compounds their formation, if it occurred,
would be consistent with the highly colored nature of black acid. These possible
reactions are diagrammed in Figure 12.

Figure 12. Possible products formed via the nitrosonium ion-toluene complex
with and without nitric acid present.

41
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 In our adiabatic nitration tests, when using excess toluene it was observed that
the initial cloudy white reaction mixture produced when the toluene was injected
turned to a pale yellow within one minute of toluene addition. The solution then
became progressively more orange between 2 to 4 minutes and finally turned red
between 4 to 5 minutes. (Conditions: a mixture of 2.5-3.0 wt% nitric acid and 64-
68 wt% sulfuric acid in water with a starting temperature of 80-100 °C and a molar
ratio of toluene to nitric acid of 1.0-1.1). However, if the reaction was run with
excess nitric acid (moles nitric acid to moles toluene of 1.05-1.2) the pale yellow
color also appeared after 1 minute but then no further color change occurred. While
black acid was not formed in any of the tests, the appearance of a red color near
the end of the excess toluene reaction (where the last traces of nitric acid would
have been completely reacted away), and no red color when excess nitric acid was
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

present, seems consistent with the observed results for mixing toluene with nitrous
acid without nitric acid present (which produced a dark orange-red color) and with
nitric acid present (which resulted in a light yellow color).
The toluene results, however, contrast with the results for benzene nitration.
Using excess benzene with similar nitration conditions, the mixed reaction
solution went from a cloudy white when the benzene was introduced to a
greenish-yellow after 1 to 2 minutes with no further color change after that time.
Excess nitric acid tests produced the same result. This is, of course, consistent
with the ability to run excess benzene in commercial adiabatic MNB production
without any issues with spent acid quality. A key reason for this difference is
likely related to the much lower amount of oxidation by-products with benzene
nitration and so a corresponding lower concentration of nitrous acid in the spent
acid. A second reason may be the greater resistance to oxidation of benzene versus
toluene which could act to slow the rates of black acid formation reactions and
in doing so lower the amount of nitric acid required to suppress them (assuming
competitive reactions as speculated in Figure 12). The appearance of red spent
acid in these excess toluene adiabatic nitration tests as opposed to black acid
reported for industrial isothermal toluene nitration may then be due to the lower
amount of reaction using 2.5-3.0 wt% nitric acid in these tests versus 17-32 wt%
nitric acid for typical isothermal toluene nitration (1). This would result in less
total reaction and so less total oxidation by-products and less nitrous acid in the
spent acid. The lower total reaction per pass typical of adiabatic nitration may
therefore be the reason that only “red acid” rather than “black acid” is observed
for our experiments that ran with excess toluene.

Acid Recycling
If adiabatic nitration using excess toluene conditions still produces black acid
type compounds, but simply fewer of them because less total reaction is carried
out, this raises the question of what happens when the spent acid in such a process is
repeatedly re-concentrated and re-used. A series of four tests was carried out where
the spent acid from each nitration test was re-concentrated under vacuum at about
80 °C at 20 mbar, and then re-used for the subsequent nitration. The experiments
used a mixture of 65.1 wt% H2SO4 and 2.5 wt% HNO3 in water, 1.10 mole toluene/
mole nitric acid, stirring at 1200 rpm (a mixing power of ~16 W/L) and a starting
42
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 acid temperature of 90 °C. After each test the spent acid was progressively darker
and HPLC analysis showed cresols in the organic phase increased from 6400 ppm
to 8900 ppm, with the cresol concentrations in the acid phase showing a similar
trend (see Figure 13).
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

Figure 13. The increase in HPLC measured cresols with repeated re-use of the
sulfuric acid. (Test 1 used fresh acid and the acid was then re-concentrated
and re-used for each additional test).

This build-up of by-products in the re-used acid may reach a steady state over
time. If not, some method to control the build-up of contaminants may be required
such as bleeding off some acid, extracting or stripping the acid, or running the
process with excess nitric (with a resultant slight loss in chemical utilization). In
isothermal nitration (with 15 to 25 wt% nitric acid in the starting mixed acid), it has
been reported that toluene nitration with excess toluene can be successfully carried
out without black acid causing problems through the use of steam stripping of the
spent acid at around 160-180 °C integrated with the acid re-concentration (29).
This same patent also provides an example showing problems found with the build-
up of organic compounds when recycling sulfuric acid during adiabatic toluene
nitration under excess toluene conditions where the acid was being re-concentrated
using flash evaporation at 90 °C at 60 mbar. Thus, to properly develop a usable
adiabatic toluene nitration process, testing must consider the effect of the nitration
conditions on the quality of the resulting re-concentrated acid. This means that
testing must include the full acid recycle loop rather than a single nitration reaction
in isolation.
43
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 A design for a small laboratory scale acid loop test system to perform such
experiments is shown in Figure 14 with a photo of the completed system in Figure
15. This system incorporates the key process components to allow testing with
continuous acid recycling. One difference from a full scale adiabatic acid recycle
loop (as diagrammed in Figure 1) is the use of air stripping for re-concentrating
the acid. In this case a metered flow of warm, dry air is used to remove water
vapor from the sulfuric acid re-concentration vessel (i.e. to lower the water
vapor pressure) as opposed to using a vacuum system. This avoided the need
for vacuum rated equipment and allowed for a compact system that could fit in
a fume hood (see Figure 15). The small size (production rate of ~1 kg MNT/h)
also minimizes chemical consumption and waste generation. This design of
experimental equipment therefore provides a simple system to assemble and
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

operate for the proper testing of adiabatic nitration in a way that considers the full
acid loop and the impact of acid recycling, which is key to a successful adiabatic
nitration process.

Figure 14. Diagram of a nitration acid loop test system.

44
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

Figure 15. Photo of the nitration acid loop test system.

Conclusions
As has been demonstrated in MNB production, adiabatic nitration has the
potential for significant energy savings versus isothermal nitration and so it would
be desirable to use adiabatic nitration for a wider variety of compounds. One
potential candidate compound for adiabatic nitration is toluene.
The effective reaction rate for toluene nitration under industrial conditions
is similar to that for benzene nitration and so nitrator designs for adiabatic
MNT would likely be similar to those used for MNB. However, the higher
temperatures encountered with adiabatic nitration leads to a shift in the product
isomer distribution, with a clear correlation between increased average nitration
temperature and increased 3-MNT. This is an issue if adiabatic toluene nitration
was to be used as a first step in the production of DNT for polyurethane
manufacture. In such a case, the adiabatic process should be designed so as to
minimize the average nitration temperature and possibly also to minimize the
amount of nitric acid in the nitrator feed.
Toluene is more easily oxidized than benzene and results in a greater variety
and a larger total amount of oxidation by-products. This makes the complete
identification and full quantification of these by-products by HPLC impractical.
It is therefore suggested that while HPLC can be used to track key representative
45
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 compounds, it should be combined with some other method to quantify the total
oxidation by-products. One possible method is a base titration of the organic
product which gives a quantitative measure of the total number of acid groups
present, and from the pH of their neutralization, a possible indication of the family
of compounds present.
Black acid did not immediately form with excess toluene adiabatic nitration,
possibly due to the much lower amounts of nitric acid used for adiabatic nitration
resulting in lower amounts of nitrous acid in the spent acid. However, significant
amounts of acid-phase soluble colored by-products are formed, which may build
up in the acid loop with continuous acid re-use. This issue highlights the need
for any study of adiabatic toluene nitration to include within its scope the full acid
recycle loop. The equipment to do such a study is not overly complex and a design
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

has been presented that can be made small enough to fit within a fume hood.
Overall, some of the key issues related to implementing adiabatic nitration
of toluene have been identified. These issues may result in some compromise in
the product quality in moving from isothermal to adiabatic nitration, or possibly
some additional process steps (for example for acid clean-up), however none of
the issues appear to completely prevent adiabatic toluene nitration. Further, while
additional costs would be associated with these issues, this must be balanced
against the large energy savings possible with adiabatic nitration. Future work
using a full adiabatic nitration loop test system will hopefully provide clear
answers to some of these questions.

References
1. Booth, G. Aromatic Nitro Compounds. In Ullmann’s Encyclopedia of
Industrial Chemistry; Wiley-VCH: Weinheim, 2005.
2. Dugal, M. Nitrobenzene and Nitrotoluenes. In Kirk-Othmer Encyclopedia of
Chemical Technology, 5th ed.; John Wiley & Sons, Inc., 2005; Vol. 17.
3. Guenkel, A. A. Nitrobenzene and Nitrotoluene. In Encyclopedia of Chemical
Processing and Design; McKetta, J. J., Cunningham, W. A., Eds.; Marcel
Dekker, Inc.: New York, 1990; pp 165−188.
4. Hermann, H.; Gebauer, J.; Konieczny, P. Industrial Nitration of Toluene
to Dinitrotoluene: Requirements of a Modern Facility for the Production
of Dinitrotoluene. In Nitration: Recent Laboratory and Industrial
Developments; Albright, L. F., Carr, R. V. C., Schmitt, R. J., Eds.; ACS
Symposium Series 623; American Chemical Society, Washington, DC,
1996; Chapter 21, pp 234−249.
5. Zaldivar, J. M.; Hernandez, H.; Nieman, H.; Molga, E.; Bassani, C. The
FIRES project: Experimental study of thermal runaway due to agitation
problems during toluene nitration. J. Loss Prev. Proc. Ind. 1993, 6 (5),
319–326.
6. Castner, J. B. Nitration of Organic Compounds. U.S. Patent 2,256,999, 1941.
7. Alexanderson, V.; Trecek, J. B.; Vanderwaart, C. M. Adiabatic Process for
Nitration of Nitratable Aromatic Compounds. U.S. Patent 4,021,498, 1977.

46
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 8. Alexanderson, V.; Trecek, J. B.; Vanderwaart, C. M. Continuous Adiabatic
Process for the Mononitration of Benzene. U.S. Patent 4,091,042, 1978.
9. Guenkel, A. A.; Rae, J. M.; Hauptmann, E. G. Nitration Process. U.S. Patent
5,313,009, 1994.
10. Guenkel, A. A.; Maloney, T. W. Recent Advances in the Technology of
Mononitrobenzene Manufacture. In Nitration: Recent Laboratory and
Industrial Developments; Albright, L. F., Carr, R. V. C., Schmitt, R. J., Eds.;
ACS Symposium Series 623; American Chemical Society, Washington, DC,
1996; Chapter 20, pp 223−233.
11. Coombes, R. G.; Moodie, R. B.; Schofield, K. Electrophilic aromatic
substitution. Part 1. The nitration of some reactive aromatic compounds
in concentrated sulphuric and perchloric acids. J. Chem. Soc. B 1968,
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

800–804.
12. Cox, P. R.; Strachan, A. N. Two-phase nitration of toluene. Part II. Chem.
Eng. J. 1972, 4, 253–261.
13. Giles, J.; Hanson, C.; Ismail, H. A. M. A Model for Rate of Nitration of
Toluene under Heterogeneous Conditions. In Industrial and Laboratory
Nitrations; Albright, L. F., Hanson, C., Eds.; ACS Symposium Series
22; American Chemical Society: Washington, DC, 1976; Chapter 12, pp
190−209.
14. Zaldivar, J. M.; Molga, E.; Alos, M. A.; Hernandez, H.; Westerterp, K.
R. Aromatic nitrations by mixed acid. Fast liquid−liquid reaction regime.
Chem. Eng. Proc. 1996, 35, 91–105.
15. Cerfontain, H.; Telder, A. The solubility of toluene and benzene in
concentrated aqueous sulphuric acid; implications to the kinetics of aromatic
sulfonation. Rec. Trav. Chim. Pays-Bas 1965, 84, 545–550.
16. Urbanski, T. Chemistry and Technology of Explosives; Pergamon Press: New
York, 1964; Vol. 1, Chapter 8, pp 265−344.
17. Konig, B. M.; Judat, H.; Blank, H. U. Process for the Adiabatic Preparation
of Mononitrotoluenes. U.S. Patent 5,648,565, 1997.
18. Barnett, J. W.; Moodie, R. B.; Schofield, K.; Weston, J. B. Electrophilic
aromatic substitution. Part 13: Kinetics, isomer yields, and the consequences
of ipso-attack in the nitration of toluene and polymethylbenzenes in aqueous
sulphuric acid, and their significance for the mechanism of aromatic nitration.
J. Chem. Soc. Perkin II 1975, 648–654.
19. Milligan, B. Isomer distribution in mixed-acid nitration of toluene. Evidence
for mass-transfer effects on selectivity. Ind. Eng. Chem. Fundam. 1986, 25,
783–789.
20. Molga, E. J.; Barcons, C.; Zaldivar, J. M. Mononitration of toluene and
quantitative determination of the isomer distribution by gas chromatography.
Afinidad 1993, 50, 15–20.
21. Harris, G. F. P. Isomer Control in the Mononitration of Toluene. In
Industrial and Laboratory Nitrations; Albright, L. F., Hanson, C., Eds.;
ACS Symposium Series 22; American Chemical Society: Washington, DC,
1976; Chapter 20, pp 300−312.
22. Hanson, C.; Kaghazchi, T; Pratt, M. W. T. Side Reactions during Aromatic
Nitration. In Industrial and Laboratory Nitrations; Albright, L. F.,

47
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Hanson, C., Eds.; ACS Symposium Series 22; American Chemical Society:
Washington, DC, 1976; Chapter 8, pp 132−155.
23. Milligan, B.; Huang, D. S. Process for Refining Aqueous Acid Mixtures
Utilized in Nitration of Aromatics. U.S. Patent 4,257,986, 1981.
24. Pohl, M. C.; Carr, R. V. C.; Sawicki, J. E. Recovery of Nitric Acid from
Nitration Spent Acid by Toluene Extraction. U.S. Patent 4,650,912, 1987.
25. Milligan, B. Some Aspects of nitration of aromatics by lower oxidation states
of nitrogen. J. Org. Chem. 1983, 48, 1495–1500.
26. Skokov, S.; Wheeler, R. A. Oxidative aromatic substitutions: Hartree-Fock/
density functional and ab initio molecular orbital studies of benzene and
toluene nitrosation. J. Phys. Chem. A 1999, 103, 4261–4269.
27. Kim, E. K.; Kochi, J. K. Oxidative aromatic nitration with charge-transfer
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch003

complexes of arenes and nitrosonium salts. J. Org. Chem. 1989, 54,

1692–1702.
28. Zuman, P.; Shah, B. Addition, reduction, and oxidation reactions of
nitrosobenzene. Chem. Rev. 1994, 94, 1621–1641.
29. Demuth, R.; Dobert, F.; Petersen, H.; Ronge, G.; Weber, H. M.;
Wurminghausen, T.; Zirngiebl, E. Continuous Isothermal Process for
Preparing Mononitrotoluenes. U.S. Patent 6,583,327, 2003.

48
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 4

Nitrophenolic By-Products Quantification in

the Continuous Benzene Nitration Process
Tiago J. G. Costa,1 Anabela G. Nogueira,1,2 Dulce C. M. Silva,2
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

Alejandro F. G. Ribeiro,2 and Cristina M. S. G. Baptista*,1

1Department of Chemical Engineering, University of Coimbra,
Pólo II, Rua Sílvio Lima, 3030-790 Coimbra, Portugal
2CUF - Químicos Industriais, Quinta da Indústria,

Beduído, 3860-680 Estarreja, Portugal

*E-mail: [email protected]

The High Pressure Liquid Chromatography (HPLC) method

for nitrophenols measurement in the benzene nitration process
has been improved. A new eluent mixture was studied using
two different compositions and the 30/70 (%v/%v) ratio of
acetonitrile and aqueous potassium dihydrogen phosphate
solution proved sufficient in accurately quantifying every
nitrophenolic compound. The influence of the pH upon peak
resolution was assessed and pH 7.0 provided the best results.
Once the chromatography parameters were set a reproducibility
study was carried out confirming the accuracy of the method.

Introduction
The nitration of benzene yields mononitrobenzene (MNB), a precursor to
aniline. Aniline is used in the manufacture of pharmaceuticals, dyes, pigments,
4,4′-methylenedianiline, and solvents. The nitration is usually carried out in
mixed acid media composed of sulfuric acid, nitric acid and water (1–3) via
an electrophilic substitution reaction (4–6). The reaction is accomplished in a
two-phase system: the aqueous acid phase and the organic phase. Of interest to
technologists involved in the manufacture of MNB via the mixed acid liquid-liquid
exothermic reaction are the challenges of mass transfer and reaction kinetics (1–3,
7–11). The benzene nitration has a well-established mechanism (5), accepted by

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 many experts in the field, where it is believed that mononitrobenzene formation
takes place in the aqueous phase. Industrially, the yield of MNB is high, but
significant side reactions occur during nitration. Benzene nitration by-products
mainly consist of dinitrated and nitrophenolic compounds. Several authors (7,
12) posit that the dinitrated compounds are formed through a consecutive reaction
after the mononitration of benzene while nitrophenolic by-products are formed via
a reaction parallel to the main reaction route. The nitrophenols (NPs) produced
are mono-, di- and tri-nitrophenol (MNP, DNP and TNP, respectively).
By-product formation is a concern in most industrial processes. In the
particular case of nitrophenols, mononitrobenzene producers are dealing with
toxic compounds requiring well designed remediation processes that are heavily
regulated. For these reasons minimization of their formation is desired. A
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

considerable amount of work has been dedicated to understanding and reducing

by-products formation (12, 13). This effort is challenging, mainly due to the
lack of extensive knowledge of by-product formation mechanisms and adequate
kinetic and mass transfer data.
Most research in this field has been carried out at the lab scale (5, 6, 10,
11) using operating conditions that do not always match those of the industrial
adiabatic process, and only traces of nitrophenols are produced. The amount of
literature concerning their formation is sparse. Mass transfer limitations between
the organic and acid phase are pointed out by Burns and Ramshaw (12) as partly
responsible for nitrophenol formation. It has been suggested that the benzene
oxidation agent is undissociated nitric acid that dissolves in the organic phase
resulting in the formation of phenol, the precursor of the nitrophenolic compounds
(12). The rate of phenol transfer to the acid phase is very fast (3), and a sequence
of sulfonation and nitration reactions (14) leads to the nitrophenolic by-products.
Once formed, the nitrophenolics (NPs) partition back to the organic phase. An
increased interfacial area between the two reacting phases is thought by many
to promote benzene mass transfer to the aqueous phase where nitrobenzene is
formed, increasing process selectivity and yield, and the design of the reactors has
for many years been driven by this principle. In the first patent on the adiabatic
nitration process (15), the dispersion of the organic phase is achieved by means
of continuous stirring of the reacting mixture. In the 1990s a new and different
configuration has been patented (16) and this makes use of newly designed
impingement elements which provide finely dispersed benzene drops improving
interfacial area. A different strategy that employs microreactors to improve the
contact between the two phases has been tested more recently (7, 17).
Interfacial area is one of the operating parameters influencing process
selectivity and it is known that mixed acid composition (3), feed flow ratio
between benzene and nitric acid and reaction temperature (13) are also relevant.
Nevertheless, the information currently in the literature is insufficient to predict
levels of nitrophenols production in industrial processes. To gather the desired
data, studies can be run in an actual industrial process. Operating factors that
govern nitrophenol production are chosen, and then changed in the process to
determine their impact on the formation of the impurities. Frequently however,
the degree of change in each factor is limited so as to not impact the rate of
production of the product.

50
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The success of such a plant study is highly dependent on the ability to sample
process streams and analytically determine the impurity profile with accuracy and
precision. Although most of the nitrophenolic impurities can be quantified by gas
chromatography (GC), trinitrophenol cannot, as it is not easily volatilized and can
be adsorbed on the GC column due to its polarity. A literature survey (18–21)
confirmed HPLC as the best analytical technique for quantifying phenol and its
nitrated derivatives. Several HPLC methods have been developed and studied,
many for identifying and quantifying phenols in waste water (18, 21). Another
related analytical study concerned a medical application (23).
In previous studies on nitration (9, 13) HPLC was used to measure NPs,
and the methods employed proved to be robust for quantifying the predominant
by-products. The literature survey summarized above showed that the current
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

methods could be improved to afford better quantification of additional

by-products. The subject of this paper is the optimization of analytical methods
to quantify the nitrophenolic by-products resulting from the nitration of benzene.

Experimental
Materials

The chemicals used to prepare samples, analytical standards, and HPLC

eluents were purchased from different suppliers.
The 1M sodium hydroxide solution (analytical reagent grade) and HPLC
grade acetonitrile were supplied by Scharlau Chemie S.A.. Phosphoric acid
(85% w/w) and potassium dihydrogen phosphate (analytical-reagent grade) were
purchased from Panreac.
Standard solutions, used for HPLC calibration, were prepared with:
4-mononitrophenol (98% w/w, where w/w is taken to mean the weight of
4-nitrophenol divided by the total weight of material) and 2,4-dinitrophenol (97%
w/w) purchased from Aldrich, 2-mononitrophenol (98% w/w) was provided by
Fluka, 2,6-dinitrophenol (98% w/w) was supplied by TCI and 2,4,6-trinitrophenol
(99% w/w) was purchased from BDH.
Water used in this study was obtained from a Milli-Q purification system from
Millipore.

Preparation of Standard Solutions

Standard solutions were prepared by adding known quantities of each analyte

to ultra-pure water. The nitrophenols were first mixed with approximately half
of the volume of water required, then about 1 g of sodium hydroxide (1 M) was
added to promote the complete dissolution of the analytes at room temperature.
To complete the preparation, ultra-pure water was added until the final desired
concentration of analyte was obtained. As a result of this study, the pH of the final
standard solution was adjusted to 7.0 by adding a small amount of aqueous solution
of phosphoric acid (1M). Each standard solution contained the five nitrophenols
used in this study at different concentrations which ranged from 2 – 10 ppm by
51
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 weight for 2- and 4-mononitrophenol and 2,6-dinitrophenol, and 11-160 ppm by
weight for both 2,4-dinitrophenol and 2,4,6-trinitrophenol.

Sampling Procedure

The plant samples analyzed were collected in different locations in the

nitration train at a CUF-QI nitrobenzene plant. The chemical compounds used
in the benzene nitration process are known to be hazardous, requiring particular
sampling procedures ensuring operator safety. Moreover, in the particular
continuous process studied the operating temperature was greater than 100 ºC,
and in order to stop the reaction, the samples were cooled as soon as collected
from the reactor. Benzene nitration is a heterogeneous liquid-liquid reaction. The
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

samples are allowed to stand and cool, where upon the phases separated. Due
to the different matrices, different analytical procedures were used to measure
the composition of each phase. This work focuses on the organic phase and, in
particular, on measuring the nitrophenols in that phase.

Nitrophenols Identification

Nitration samples were collected from the plant and allowed to separate. A
portion of the organic phase was treated with calcium carbonate to neutralize
and remove any trace acid present. The sample was filtered and then analyzed
by GCMS. A GC-MS from Agilent, models GC-7890A and MS-5975C, with
a HP-5MS column (30 m x 0.25 mm x 0.25 μm), was used for identifying the
nitrophenols present in the organic phase collected from the nitrobenzene plant.
The volume of sample injected onto the GCMS was 1 μL, oven conditions were:
40 ºC (1 min), to 100 ºC (15 ºC/min), 20 ºC/min to 240 ºC, and 10 ºC/min to 310
ºC, using helium as carrier gas at constant flow (30 mL/min). The main objective
of this work was to better quantify the compounds identified by GCMC by HPLC.

Sample Preparation for HPLC Analysis

The sampled organic phase is less dense and usually darker than the aqueous
phase, enabling for an easy identification of each separated phase after a short
settling time. The literature indicated that the nitration reactions leading to
nitrophenols occur in the acid phase, and these by-products transfer into the
organic phase. The samples injected onto the HPLC were aqueous based.
Therefore, the nitrophenols must be extracted from the organic phase. This is
performed by adding 1 mL of the organic phase to 2 mL of sodium hydroxide
solution (1 M), shaking this mixture vigorously and then separating the two phases
in a centrifuge. The extraction yield was found to be 94% and the efficiency
of the process can be confirmed by the change in color of organic and aqueous
phases. The aqueous extract, containing the extracted NPs, was diluted (13) and
pH adjusted. This was performed by adding 20 mL of ultra-pure water to 1 mL
of the extracted aqueous phase, followed by a careful addition of phosphoric acid
(1 M) until a pH of 7 was obtained.
52
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 HPLC Conditions and Instrumentation

The reversed-phase HPLC system consisted of a HPLC pump model K-1001

and an ultraviolet detector model K-2501, both from Knauer. Quantification was
done at 360 nm, and a UV spectrum of each analyte could also be obtained in
the ‘stop flow’ mode. Data processing employed chromatographic software from
DataApex.
The literature was surveyed to choose the HPLC columns best suited for
this study. The Mediterranean Sea 18, 5 μm, 250 mm x 4 mm column, reference
TR-010038 from Teknokroma packed with C18 silica matched the features
mentioned by Belloli et al. (19) as suitable for obtaining the separation of all the
NPs in the samples. The column was placed in an ELDEX oven at 30 ºC. The
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

flow-rate was 1 mL/min, and the isocratic mobile phase consisted of a mixture of
acetonitrile and an aqueous solution of KH2PO4 (50 mmol), as buffer (18). Two
eluent compositions were prepared for study. The first eluent was a 50%/50%
(v/v) mixture of acetonitrile and the aqueous KH2PO4, and the second was a
30%/70% (v/v) acetonitrile and aqueous KH2PO4. Before use, the solutions were
filtered through a 0.45 μm nylon membrane and degassed. The volume of sample
injected was 10 μL.

Results
The first stage of the study consisted in identifying the nitrophenols present
in organic phase in the samples collected from the plant. The goal was the
quantification of the predominant nitrophenols as well as the trace by-products.
We wanted to confirm the presence and amounts of each nitrophenol reported
in the literature. GCMS was used to determine the nitrophenols present in the
samples, which were identified according to the chromatographic retention times
of authentic material, Table I (see the Experimental Section for details of the
analysis).
A separate standard of each analyte was injected onto the HPLC to determine
retention times. Preliminary work suggested an adequate isocratic solvent system
to start with was 30% volume acetonitrile and 70% volume aqueous solution of
KH2PO4 (50 mmol), at pH 7.0 at a 1 mL/min flow rate. A chromatogram of a
sample containing the 5 nitrophenolic compounds is shown in Figure 1.

53
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table I. Nitrophenols present in the organic phase of the samples – GC-MS
information
Retention time
Abbreviation Compound CAS Number
(min)
2 MNP 2 Mononitrophenol 88-75-5 5:07
4 MNP 4 Mononitrophenol 100-02-7 6:13
2,6 DNP 2,6 Dinitrophenol 573-56-8 10:51
2,4 DNP 2,4 Dinitrophenol 51-28-5 12:31
TNP Trinitrophenol 88-89-1 21:05
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

Figure 1. HPLC run with reference compounds: 2,6 DNP (1), 2,4 DNP (2),
TNP (3), 4 MNP (4) and 2 MNP (5).

Eluent Composition
It is known (21, 23) that eluent composition is one of the parameters
influencing the separation of the analytes. The eluent used was a solution
composed of acetonitrile (21, 24) and aqueous KH2PO4 (50 mmol); two
volumetric ratios of these two components were assessed in this study: 50/50
(%v/%v, acetonitrile/aqu) according to Alber et al. (18), and 30/70 (%v/%v,
acetonitrile/aqu) with the hope of improving peak separation as suggested by
Preiss et al. (21). The pH of the two eluents assessed was 7.0. Figure 2 shows the
overlay of the two chromatograms obtained using the two eluent systems.
An increase in the aqueous component of the eluent resulted in longer
retention times and better separation of the analytes. The hydrophobic analytes
interact with the C18 more as the percentage of the aqueous component of the
54
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 eluent is increased, resulting in the longer retention times and better overall
separation. Moreover, this improved separation enabled quantification of the
five nitrophenols in the sample, while only three components were detected with
the 50/50 acetonitrile/aqueous eluent (peaks 1 and 5 do not show in the 50/50
(%v/%v) chromatogram). The 2- and 4-mononitrophenol and 2,6-dinitrophenol
amounts in the standard were low, thus the small peak height in the chromatogram.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

Figure 2. Chromatograms using different volumetric ratios of acetonitrile and

aqueous potassium dihydrogen phosphate as mobile phase: 2,6 DNP (1), 2,4
DNP (2), TNP (3), 4 MNP (4) and 2 MNP (5).

The chromatograms in Figure 2 were further assessed by calculating Rs,

a parameter that expresses the degree of separation of the peaks. According
to Pombeiro (25), the Rs should be greater than 1.5 for complete separation of
two neighboring peaks. Other important chromatographic calculations (24, 25)
were done to access the efficiency of the separation of the analytes, namely
determining the retention factor (k′), the selectivity α and the plate number
(N). The chromatographic data was analyzed using the Clarity chromatographic
Station Software to provide the chromatographic efficiency parameters listed in
Table II for the two eluent compositions studied.
The optimal range for retention the factor k′ is 1 to 10 (25); higher values
do little to improve resolution of the peaks and only results in longer analysis
time. Selectivity values (α) in Table II are greater than 1.0, as expected, as they
refer to retention factors of neighboring peaks. The magnitude α does not indicate
the advantage of one eluent system over the other. As shown in Table II, the
HPLC separation parameters with the 30/70 (%v/%v) eluent composition led to
an improvement of the resolution, Rs, for each analyte. The use of 30/70 (%v%/v)
acetonitrile/aqueous-buffer gave superior peaks separation and resolution.

55
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table II. HPLC separation efficiency parameters for two mobile phase
compositions
Nitrophenols 2,6DNP 2,4DNP TNP 4MNP 2MNP
Eluent – acetonitrile/ Aqu
50/50 (%v/%v)
KH2PO4 (50 mmol)
k′ * nd 4.3 5.9 11.6 * nd
α - 1.4 1.9 -
N * nd 665 1658 3242 * nd
Rs - 2.0 5.4 -
Eluent – acetonitrile/ Aqu
30/70 (%v/%v)
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

KH2PO4 (50 mmol)

k′ 0.9 1.3 3.1 3.7 7.0
α 1.5 2.4 1.2 1.9
N 3088 3324 3114 6670 8606
Rs 2.3 5.7 2.4 8.7
*nd- not detected

Adjustment of pH

The pH of the samples, standard solutions and eluent is an important factor

that influences both accuracy and the precision of the method (22). To assess the
influence of the pH of the sample on the separation, the pH of the samples were
varied from 1.0 to 12.0. Figure 3 shows over-layed chromatograms for samples
with a pH of 1, 7, and 12 adjusted with phosphoric acid (1M). It can be seen
that the retention times and peak shapes of the analytes, and the ability to detect
4-MNP and 2-MNP are significantly impacted by the pH of the sample. This is
most likely due to the degree of ionization of the nitrophenols in solution (22).
The pH of the sample dictates the extent of ionization of the nitrophenols in the
sample, as does the pH of the eluent. The UV absorbance value at the wavelength
used for quantification will also be affected by the extent of ionization.
The study of sample pH impact on the HPLC chromatography used a standard
that contained the nitrophenols at the targeted concentrations shown in Table
III. The pH of the nitrophenolic mixture standard was adjusted with phosphoric
acid (1M) to provide samples with pHs of 1, 5, 7, 10, and 12. The pH adjusted
samples were then analyzed to determine the best sample pH for the optimization
of analysis time and accuracy of the quantification of each analyte. These results
are summarized in Table III, and it is clear that a sample pH of 7 gave the best
quantification result. Recall that the pH of the eluent was 7. Varying the pH of
the eluent was not a subject of this study.

56
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

Figure 3. Influence of pH of the sample upon the chromatograms obtained from

the same sample with 2,6 DNP (1), 2,4 DNP (2), TNP (3), 4 MNP (4) and 2
MNP (5). The eluent composition was 30/70 (%v/%v) and the eluent flow rate 1
mL/min.

Table III. Influence of sample pH upon the quantification of nitrophenols.

The eluent composition was 30/70 (%v/%v) and the eluent flow rate 1
mL/min.
Composition (ppm)
2,6 DNP 2,4 DNP TNP 2 MNP 4MNP
Target composition 7.5 81.8 65.2 7.5 8.2
pH
1 8.3 87.2 70.6 8.1 8.7
5 12.5 86.8 69.7 8.0 8.9
7 7.4 81.9 64.7 7.2 8.3
10 11.1 86.7 68.1 7.9 8.7
12 8.8 84.4 66.5 7.6 8.4

57
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Standard Curves

Having established the best operating conditions for the HPLC analysis,
calibration curves were determined. Five standard solutions with different
concentrations of each of the five nitrophenols were prepared. Two concentration
ranges were considered: 11-160 ppm by weight for the predominant analytes
(2,4-dinitrophenol and trinitrophenol) and 2-10 ppm by weight for 2- and
4-mononitrophenol and 2,6-dinitrophenol. These standard solutions were injected
on the HPLC and results used to determine calibration curves for each nitrophenol.
The equations of the regression lines for each analyte had an R2 (coefficient of
determination) of 0.99, Table IV. This study demonstrated the precision of the
method within the concentration range of interest.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

Table IV. Validation range of the linear response for each solute
Validation range Equation of line a
Compound
(ppm) (slope and intercept; R²)
y = 27.137x - 2.5873;
2 Mononitrophenol 2-10
R² = 0.9997
y = 24.307x + 5.4361;
4 Mononitrophenol 2-10
R² = 0.9999
y = 13.43x - 3.2565;
2,6 Dinitrophenol 2-10
R² = 0.9964
y = 73.377x + 5.9492;
2,4 Dinitrophenol 11-160
R² = 0.9999
y = 81.13x + 11.402;
Trinitrophenol 11-160
R² = 1
a x= concentration of solute (ppm) and y=Peak area (mV.s)

Reproducibility of the Analytical Method

Ten replicate injections were made for each standard solution used in this
study. The coefficient of variation (CV) was calculated to assess intra-assay
precision. The calibration was carried out for five consecutive days using the
five standard solutions and the CV calculated to assess inter-day assay variation.
The results of this study for one of the standard solutions are presented in Table
V. The mean value (x̅) the standard deviation (σ), CV, and the percent error were
calculated. The error was consistently below 10% and the lower errors (0.71 and
1.18 %) correspond to the predominant by-products confirming the adequacy of
the method. Inter-assay precision was shown to be less than 4%.

58
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table V. Reproducibility parameters for a standard solution
x̅ σ CV Error
Compound ppm
ppm (%) (%)
2,4 DNP 80.9 0.37 0.46 1.18
TNP 78.3 0.21 0.27 0.71
2,6 DNP 8.1 0.27 3.38 9.27
2 MNP 8.3 0.06 0.68 2.33
4 MNP 8.7 0.07 0.82 2.18
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

Conclusion
The main goal of this study was to improve the HPLC method used to quantify
the nitrophenolic by-products produced in the nitration of benzene. The challenge
included measuring the mononitrophenol isomers and 2,6-dinitrophenol, at
low concentrations levels, where quantification has been difficult in the past.
The work carried out began with the selection of the HPLC column and
optimization of the mobile phase. The composition of the new mobile phase,
acetonitrile and aqueous potassium dihydrogen phosphate solution dramatically
impacted the chromatography. Two eluent compositions were assessed: 50/50
(%v/%v, acetonitrile/aqu) and 30/70 (%v/%v), and the satisfactory identification,
separation and quantification of the five nitrophenols in the samples were only
achieved with the 30/70 (%v/%v) mixture. The best quantification of the analytes
was obtained with a sample pH of 7.0. In spite of the improved performance of
this HPLC method, it is important to note that the analysis time is now longer,
although less than the 30 minutes, commonly considered as an acceptable limit.
The use of an ultra-high pressure liquid chromatograph would undoubtedly result
in a much lower analysis time. In addition to the improved chromatography, the
new method also allows using a lower oven temperature (30 ºC), and a longer
column lifetime is expected, in part due to the change in the eluent.

Acknowledgments
This study was carried out in the frame of the project CUFTECH-CUF-
Technology (Project nº5438) financed by QREN -20072013 and EU “Fundo
Europeu de Desenvolvimento Regional”. Financial support from Fundação
para a Ciência e Tecnologia (FCT) for Ph.D. Grant SFRH/BDE/33907/2009, is
gratefully acknowledged.

References
1. Alexanderson, V.; Trecek, J. B.; Vanderwaart, C. M. U.S. Patent 4,021,498,
1977.
2. Alexanderson, V.; Trecek, J. B.; Vanderwaart, C. M. U.S. Patent 4,091,042,
1978.
59
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 3. Guenkel, A. A.; Rae, J. M.; Hauptmann, E. G. U.S. Patent 5,313,009, 1994.
4. Coombes, R. G.; Moodie, R. B.; Schofield, K. J. Chem. Soc. B 1968,
800–804.
5. Olah, G. A.; Malhotra, R.; Narang, S. C. Nitration. Methods and
Mechanisms; Organic Nitro Chemistry Series; VCH Publishers, Inc.: New
York, 1989; Volume 109.
6. Hughes, E. D.; Ingold, C. K.; Reed, R. I. J. Chem. Soc. 1950, 2400–2440.
7. Dummann, G.; Quittmann, U.; Gröschel, L.; Agar, D. W.; Wörz, O.;
Morgenschweis, K. Catal. Today 2003, 79–80, 433–439.
8. Modak, S. Y.; Juvekar, V. A. Ind. Eng. Chem. Res. 1995, 34, 4297–4309.
9. Quadros, P. A.; Oliveira, N. M. C.; Baptista, C. M. S. G. Chem. Eng. J.
(Amsterdam, Neth.) 2005, 108, 1–11.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch004

10. Zaldivar, J. M.; Molga, E. J.; Alos, M. A.; Hernandez, H.; Westerterp, K. R.
Chem. Eng. Process. 1995, 34, 543–559.
11. Zaldivar, J. M.; Molga, E. J.; Alos, M. A.; Hernandez, H.; Westerterp, K. R.
Chem. Eng. Process. 1996, 35, 91–105.
12. Burns, J. R.; Ramshaw, C. Chem. Eng. Res. Des. 1999, 77, 206–211.
13. Quadros, P. A.; Castro, J. A. A. M.; Baptista, C. M. S. G. Ind. Eng. Chem.
Res. 2004, 43, 4438–4445.
14. Urbanski, T. Chemistry and Technology of Explosives; Pwn-Polish Scientific
Publishers: Warszawa, Poland, 1964; Vol. I.
15. Castner, J. B. U.S. Patent 2,256,999, 1941.
16. Rae, J. M.; Hauptmann, E. G. U.S. Patent 4,994,242, 1991.
17. Burns, J. R.; Ramshaw, C. Chem. Eng. Commun. 2002, 189, 1611–1628.
18. Alber, M.; Böhm, H. B.; Brodesser, J.; Feltes, J.; Levsen, K.; Schöler, H. F.
Fresenius’ J. Anal. Chem. 1989, 334, 540–545.
19. Belloli, R.; Barletta, B.; Bolzacchini, E.; Meinardi, S.; Orlandi, M.;
Rindone, B. J. Chromatogr., A 1999, 846, 277–281.
20. Hofmann, D.; Hartmann, F.; Herrmann, H. Anal. Bioanal. Chem. 2008, 391,
161–169.
21. Preiss, A.; Bauer, A.; Berstermann, H. M.; Gerling, S.; Haas, R.; Joos, A.;
Lehmann, A.; Schmalz, L.; Steinbach, K. J. Chromatogr., A 2009, 1216,
4968–4975.
22. Sun, Y.; Fang, N.; Chen, D. D. Y.; Donkor, K. K. Food Chem. 2008, 106,
415–420.
23. Thompson, M. J.; Ballinger, L. N.; Cross, S. E.; Roberts, M. S. J.
Chromatogr., B: Biomed. Sci. Appl. 1996, 677, 117–122.
24. Skoog, D. A.; Leary, J. J. Principles of Instrumental Analysis; Saunders
College Publishing: Philadelphia, PA, 1992; pp 579−604.
25. Pombeiro, A. J. L. O. Técnicas e operações unitárias em química
laboratorial; Fundação Calouste Gulbenkian: Lisbon, Portugal, 1983; pp
433−615.

60
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 5

Improved Processes for the Recovery of

Nitric and Sulfuric Acids from Nitration Plants
Manfred Pertler,* Gottfried Dichtl, Ulrich Walter,
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch005

Theo Hessenius, and Joachim Korbach

De Dietrich Process Systems GmbH (formerly: QVF Engineering GmbH),

55122 Mainz, Germany
*E-mail: [email protected]. Web site: www.qvf.com

De Dietrich Process Systems has a long history in the

recovery of nitric and sulfuric acids for all different types of
nitration processes. Most of these recovery operations are
tailor-made designs and provide the most efficient solutions to
our customers. We fabricate the equipment needed for these
operations from highly corrosion resistant borosilicate glass
and glass-lined steel. An improved system to recycle acid in a
DNT manufacturing facility is discussed.

Introduction
The QVF® BAT (Best Available Technology) allows the optimized
concentration and recovery of combined spent acid recycle streams from a
Dinitrotoluene (DNT) facility, specifically the spent DNT sulfuric acid (~70 wt%
sulfuric acid), nitric acid (~ 1-2%) and the pre-concentrated acidic yellow water
(from the first water wash of the DNT product).
A two distillation column process has been designed for acid recovery. The
first column is fed with the combined recycle acid streams, where the nitric acid is
almost completely separated from the spent sulfuric acid, to afford an overheads
product of nitric acid of up to 99 wt% (weight percent). In the second column
the remaining organics present in the spent sulfuric acid, such as MNT/DNT,

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 are recovered by a highly efficient stripping process using live stream injection.
The organics are steamed distilled overhead. The remaining sulfuric acid bottoms
product (~83 wt% sulfuric acid) from this stripping column has a content of less
than 10 ppm organics. This pre-concentrated sulfuric acid is concentrated up to
89 wt% or even above 97 wt%, in subsequent sulfuric acid concentration (S.A.C.)
units and then re-used for nitration.
With this combined QVF® process, we tremendously reduce the operating
energy costs (almost 30% less), lower cooling water consumption, and achieve
lower investment costs, while also reducing fugitive emissions.

History of De Dietrich / QVF®

Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch005

A pictorial history of the company is shown in Figure 1.

Figure 1. Since 1684 De Dietrich has designed industrial systems and products.
(Courtesy of Manfred Pertler). (see color insert)

In 2000 De Dietrich acquired QVF Engineering GmbH including all the

existing QVF® and SCHOTT® acid plant technologies. All former glass
and process systems activities from Quickfit, QVF Corning, EIVS, SCHOTT
Engineering and QVF Glastechnik are now concentrated in the De Dietrich
Process Systems Group.
62
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The photos shown in Figure 2 include De Dietrich® glass-lined equipment,
Rosenmund® filter dryers, and a QVF® borosilicate glass column packed with the
structured DURAPACK. These systems are used world-wide in highly corrosive
applications. De Dietrich also provides service when called upon. In 2009, the
year De Dietrich celebrated its 325th anniversary, all the De Dietrich subsidiaries
worldwide, changed their names to De Dietrich Process Systems. Hence, QVF®
has been established as a brand of De Dietrich Process Systems. De Dietrich
Process Systems GmbH, Germany, has two main tasks worldwide: Borosilicate
Glass and Process Systems design.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch005

Figure 2. De Dietrich Process Systems Group. (Courtesy of Manfred Pertler).

(see color insert)

Process Systems Expertise includes:

- development of in-house technology,

- reliance of De Dietrich patented technology,
- access to a comprehensive experience base in the industry,
- lastly, fabrication of equipment that is designed in-house, marrying the
expertise of process design and materials of construction (Figure 3).

For the development and optimization of different process systems we use our
own experimental facility, providing a full range of service from the feasibility
study to commissioning of equipment, to after-market maintenance, on a global
basis.
63
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch005

Figure 3. Process Systems Design and Process Expertise. (Courtesy of Manfred

Pertler). (see color insert)

Recovery of Acids in a DNT Nitration Facility

The author assumes that the reader is familiar with DNT manufacture (1). The
1st generation process to treat weak spent sulfuric acid from DNT manufacture is
shown in Figure 4.
This simplified flow sheet shows only the main units of operation of the
process, namely the denitrification (DENI) column C1 and the nitric acid –
pre-concentration (NA-PC) column C2. Both columns operate at atmospheric
pressure. The subsequent S.A.C. stages as well as a NOx absorption unit, which
are usually included in the full scope of the plant design, are not shown in Figure
4. The weak spent sulfuric acid (~70 wt%), containing nitric acid and according to
the solubility a DNT content of 0.5 wt%, is fed into the denitrification column C1.
Heat to the column is provided by a steam heated horizontal evaporator E1. DNT
and nitric acid are taken off over-head of C1 and the sulfuric acid, now at 72 to 75
wt% H2SO4, leaves the evaporator and contains between 0.1 and 0.4 wt% DNT.
The overheads are condensed in a condenser. The bulk of the condensed DNT is
separated and recovered in a horizontal separator. The aqueous phase collected
in the separator is then mixed with the acidic yellow wash water from the first
water wash of the DNT product, containing 10-20 wt% nitric acid, 5-12 wt%
sulphuric acid and < 0.5 wt% DNT. This mixture is fed to the NA-PC column C2.
An energy penalty is paid for the condensation of the overheads to recover the
DNT, as the nitric acid must be re-evaporated in the NA-PC (C2). The nitric acid
and DNT are taken overhead of C2 and condensed. The condensate containing
~1 wt% nitric acid and less than 0.15 wt% of DNT is fed to an additional phase
64
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 separator (not shown in Figure 4) before it is forwarded to waste water treatment
or is reused as wash water.
This 1st generation technology only provided for about 80% recovery of the
DNT in the feed to C1, complicating the operation of C2 and subsequent S.A.C.
units. The bulk of the DNT is not stripped in C1 and remains in the effluent of the
horizontal evaporator (E1).
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch005

Figure 4. DENI / NA-PC: 1st Generation Acid Recovery Process Design for a
DNT Manufacturing Facility. (Courtesy of Manfred Pertler). (see color insert)

The subsequent S.A.C. units concentrate the 72 to 75 wt% sulfuric acid from
C1 (from E1) up to 89 or even above 97 wt%. These additional S.A.C.s are
run under vacuum using cooled condensers. DNT can foul or even block the
condensers. One way to overcome these issues is to add MNT to the S.A.C.
overheads to keep the DNT from fouling the condensers.
There is almost no organic stripping inside of the NA-PC Column C2, so that
the soluble DNT in the C2 feed, that does not leave the C2 overhead, leaves via
the horizontal boiler E2 at the bottom of the C2. The concentration of DNT in
the E2 effluent is between 2 and 5 wt%, together with ~50 wt% nitric acid, and ~
10% sulfuric acid. The continuous increase of the DNT concentration inside of the
C2 increases the risk that a second DNT phase or at least some small aggregated
droplets of DNT could be established at the bottom of C2 or in the horizontal
boiler E2. Under normal operating conditions, the temperature in the bottom of
C2 and of E2 can be as high as 130 °C, especially if the sulfuric acid concentrations
increase above 10 wt%. This is a temperature at which DNT can decompose, the
consequences of which can be severe.
65
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The disadvantages of the 1st generation acid recovery process can be
summarized as follows:

- inefficient stripping of DNT in C1 (only 80% of the DNT in the feed to

C1 is taken overhead),
- risks with a DNT phase in the NA-PC C2, which can decompose,
- an additive (e.g. MNT) is needed to avoid fouling in the condensers in
the S.A.C. units.

A first improvement to this system was made decade ago, which yielded to
the 2nd generation acid recovery process, Figure 5.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch005

Figure 5. 2nd Generation Acid Recovery Unit for a DNT Nitration Process
(Patent No. DE 10356499B3). (Courtesy of Manfred Pertler). (see color insert)

The main advantage of the 2nd generation process is that the vast majority of
the DNT is stripped under vacuum conditions (380 mbar) out of the spent sulfuric
acid using live steam injection in the first column C1 (Figure 5). Heat is provided
to C1 with a steam heated horizontal heat exchanger (E1). The acid leaving E1
contains less that 20 ppm DNT. Since the acid leaving E1 (78 to 83 wt% H2SO4)
is virtually devoid of DNT, there are no operational issues with the condensers in
subsequent S.A.C. units and MNT does not need employed to prevent fouling by
DNT deposits in condensers.
The partial spray condenser on top of C1 is fed with acidic yellow wash water
to effect the partial condensation of a mixture of DNT, sulfuric acid and nitric
acid. This aqueous condensate recovered from the condenser is sent back to the
nitration, via a storage tank. The nitric acid and DNT that is not condensed then
66
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 feeds the NA-PC column, C2. The DNT that is fed to C2 is partially removed
over-head. The overheads product is composed of 1 to 2 wt% nitric acid and a
small amount of DNT ~ 0.15 wt%. Since the NA-PC column C2 is also operated
under the same vacuum conditions (~ 380 mbar), the temperature of the effluent
of E2 is below 100 °C. The composition of the effluent is ~ 52 wt% nitric acid, ~
8 wt% sulfuric acid, and 1-3 wt% DNT. The cooler temperature of the E2 effluent
and the lower amounts of DNT in the effluent of E2 reduces the potential for the
formation of a separate DNT phase and for the decomposition DNT, both of which
are safety concerns.
The recovery of DNT in the acid recovery system is >95%, since the recovered
sulfuric acid only contains less than 20 ppm DNT, and the condensate from C2
contains little DNT (1500ppm, water saturated with DNT).
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch005

To summarize the 2nd generation acid recovery system:

- high efficiency of DNT stripping is achieved,

- low operational cost since cooling water and steam consumption has been
reduced,
- lower investment costs are achieved.

The best available acid recovery technology is the 3rd generation of QVF®
DENI-S.A.C. process, where separate units for the recovery of nitric acid, sulfuric
acid and condensate have been designed (Figure 6).

Figure 6. 3rd Generation Acid Recovery System (Best Available Technology) for
a DNT Manufacturing Facility provided by QVF® (EP2295375A1; US Patent
Application No. 13/395810). (Courtesy of Manfred Pertler). (see color insert)

67
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 C1 is fed with a mixture of spent sulphuric acid (~70 wt%), containing nitric
acid (1-2 wt%), and DNT (< 0.5 wt%) and acidic yellow washing water. It is also
fed 88 to 93 wt% H2SO4, so that more concentrated nitric acid can be obtained
as an overheads product. The purpose of C1 is to strip-off all of nitric acid, and
HNO2, as well as a portion of the DNT that is present in the feed. This packed
stripping column is operated at ambient pressure and is designed as a combined
stripping and concentration unit operation. The overheads of C1 are condensed.
The condenser is cooled with 55 °C water to avoid fouling of the condenser with
solid DNT. The condensate is about 60 wt% nitric acid containing a small amount
of DNT. This nitric acid can be used back in the nitration process.
The 60 wt% nitric acid collected overhead of C1 can be concentrated further in
a N.A.C./S.A.C. (nitric acid concentrator/sulphuric acid concentrator) if desired.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch005

The effluent of C1 still contains considerable amounts of DNT. It is composed

of ~73 wt% H2SO4, traces of nitrogen species, and up to 4000 ppm DNT. This
stream is fed to the C2, the denitrification column, run at 380 mbar. Live steam
is injected into C2 to facilitate the removal of the DNT from the C2 feed. The
flow of the live steam must be strictly controlled to strip the DNT and any MNT
from the feed without overly diluting the system with water, resulting in a more
dilute effluent from the bottom of C2. Heat is supplied to C2 with a steam heated
horizontal evaporator (E2). The overheads of C2 is condensed in a condenser
cooled with 55 °C cooling water, to protect against DNT fouling of the condenser.
Condensate from the condenser of C2 is fed by gravimetric flow to liquid/liquid
separator vessel D3 for separating organics from aqueous condensate. The
organic rich phase is fed to a receiver vessel outside the battery limits for re-use in
customer’s nitration processes. The acidic aqueous phase from D3 is discharged
to customer’s waste water treatment system.
The effluent of E2 contains <10 ppm DNT and a maximum of 83 wt% H2SO4.
This acid is fed to a third stage S.A.C. for further concentration (up to ~97 wt%).
The low amounts of DNT in the E2 effluent protect down-stream equipment from
fouling and the associated safety issues.

Technical Improvements of the Equipment

The equipment used in the 3rd generation acid recovery process utilizes the De
Dietrich patented distributer and packing support CORE Tray® (Figure 7), and the
optimized porous-free and highly corrosion-resistant structured DURAPACK®
packing made from borosilicate glass (Figure 8).

68
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch005

Figure 7. Corrosion Resistant Tray (Patent No. US 7234692B2). (Courtesy of

Manfred Pertler). (see color insert)

Figure 8. Corrosion-resistant Structured Packing (Patent No. US 6726188B2).

(Courtesy of Manfred Pertler). (see color insert)
69
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Summary
Benefits of the 3rd generation of the QVF® acid recovery system are:

- efficient stripping of DNT from the recycle sulfuric acid to lower than 10
ppm,
- no separate nitric acid concentration unit requirement,
- nitric acid concentration up to 99% is possible if desired,
- minimal nitrogen containing species in the condensate of C2, simplifying
water treatment,
- no sulfuric acid in the nitric acid overheads of C1,
- optimized design for capital and operation costs.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch005

Plants with a capacity of up to 150 MT/hr (metric ton per hour) feed rates
have been demonstrated by the De Dietrich Group. The entire process, from design
engineering, equipment fabrication, and service can be offered from a single entity.

References
1. For the reader that is not familiar with the manufacture of DNT in mixed
acid media, see Hermann, H.; Gebauer, J.; Konieczny, P. Industrial Nitration
of Toluene to Dinitrotoluene: Requirements of a Modern Facility for
the Production of Dinitrotoluene. In Nitration: Recent Laboratory and
Industrial Developments; Albright, L. F., Carr, R. V. C., Schmitt, R. J., Eds.;
ACS Symposium Series 623; American Chemical Society, Washington, DC,
1996; pp 234−249.

70
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 6

Nitration Technology for Aromatics

As Described in the Patent Literature
Johannes Duehr*
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

Am Holzbruch 19, 47802 Krefeld, Germany

*E-mail: [email protected]

Several trends in nitration technology have emerged as

ascertained by examination of the patent literature. The
adiabatic process has become state of the art for benzene
nitration. Nitration waste water treatment has moved from
extraction processes to thermal processes (for safety reasons).
Though many patents claim adiabatic toluene nitration,
analogous to the adiabatic nitration of benzene, most new
dinitrotoluene plants continue to use isothermal processes;
this is most likely to control isomer ratios and to avoid high
temperatures for safety concerns. Several patents focus on the
improvement of the washing area to isolate the byproducts
as well as to recover nitric acid. The field of gas-phase
nitration is active; however this process is not competitive
with liquid-phase processes using sulfuric acid. The challenge
remains to find an efficient catalyst for the gas-phase nitrations.

Introduction

The nitration of organic molecules has been practiced for almost 200 years.
For example, the first nitration of the benzene molecule to mononitrobenzene
(MNB) was recorded by Mitcherlich in 1834. Today MNB, mononitrotoluene
(MNT), and dinitrotoluene (DNT) have become versatile precursors to many
other commodity chemicals such as aniline which is used to produce dyes, rubber
chemicals, and polyurethanes.
The production of these nitrated species continues to increase to meet market
needs. There has been a continuous effort to improve the nitration processes during

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 the periods of growth. The process development has focused on safety concerns,
economical aspects (costs), and environmental challenges.
The number of patents in the field of nitration is large. The most relevant were
selected to review here, and in particular the industrially relevant patents related
to MNB and MNT/DNT technologies will be emphasized. This paper discusses
how these patents have shaped the current MNB and MNT/DNT industries, and
what they suggest for future process trends. These are very high volume products
and are mostly feed materials for the production of polyurethanes.

Process Overview and Trend Areas for MNB and MNT/DNT

The methods to manufacture these materials are comprised essentially of the
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

same major process steps, as shown in Figure 1.

Figure 1. Scope and Definition of the Trend Areas.

The first step is the nitration reaction, which provides the crude product.
The reaction is followed by a product purification process, which involves
a liquid-liquid extraction step of the organic product phase with a water or
basic aqueous wash. The aqueous extract must be treated before discharge to
the environment; the quality of the discharged water must meet strict purity
specifications as mandated by governmental agencies. Similarly, the vent gases
from the plant have to be treated. In most cases, where a mixture of sulfuric acid
and nitric acid are used as the nitrating media, the spent acid is diluted by the
water created in the reaction, and must be reconcentrated before being fed back
to the nitration step.
The author has assumed that the reader is familiar with the basic chemistry
and process engineering for the mixed acid nitration of benzene and toluene to
afford nitrobenzene and MNT/DNT, respectively (1).

Patents
The patents reviewed tend to focus in a particular key area. Some focus
on heat integration, using the reaction enthalpy to supply the energy demand for
reconcentrating spent sulfuric acid. Other patents focus on lowering by-product
formation, such as nitrophenols, nitrocresols, dinitrobenzene (DNB), and in some
72
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 cases DNT. Some patents show methods of maximizing favorable ratios of desired
isomeric products. There are also patents concerning the reduction of emissions,
and there are others which describe gas-phase nitration technology.

Summary of Patents Related to the Nitration of Benzene

A review of the patent literature for MNB production suggests that the
adiabatic nitration process has almost completely replaced the isothermal process,
making the adiabatic reaction the state of the art. The first full-scale plants for
adiabatic mixed acid (sulphuric acid and nitric acid) nitration of benzene to
produce MNB were constructed in the U.S. in the late 1970’s and were equipped
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

with cascades of stirred reactor vessels. The flow of the reaction media in
the reactors was upwards so that the discharge from the reactors was under
lower pressure. The nitration vessels were pressurized with nitrogen to avoid
evaporation of residual benzene at the higher temperatures present at the tail end
of the reaction vessel train.
The concept of the continuous MNB process was first presented by Castner
(2) and further refined by Alexanderson et al. (3, 4). The reaction rate of nitration
of benzene with nitric acid is limited by mass transfer. The reaction itself occurs
at the surface of the organic (benzene) droplets as nitric acid reaches the interface
between the mixed acid phase and the organic phase. The surface area, over which
the reaction takes place, is maximized by creating an emulsion between the mixed
acid continuous phase and the organic (benzene) dispersed phase as small droplets.
A design of a mixing device for a plug flow reactor was described by Evans
(5), with the intention to increase the interfacial surface area between the mixed
acid and organic phases using high shear forces and turbulence. An aromatic
substrate and/or mixed acid can be forced through an annulus to form fine droplets
of each phase when the substrate and mixed acid are contacted with one another
in a reaction chamber.
Guenkel et al. (6) claim a nitration feed condition, for the concentration of
nitric acid in the mixed acid, which is outside the operating range previously
mentioned in the literature. Maintaining the nitric acid concentration below
3% favours the complete dissociation of nitric acid to nitronium ions and thus
accelerates the reaction rate. Mixing elements are used to generate a fine emulsion
of organic droplets in the mixed acid. Thus the mass transfer is enhanced and the
overall reaction rate increases. A significant benefit of the rate increase was that
the desired nitration reaction was favoured and the production of impurities was
minimized. An additional advantage of these operating conditions was improved
safety due to lower operating temperatures and lower nitric acid concentrations.
The reactor type was a plug flow reactor.
The invention of Larbig (7) relates to a process for treating the basic aqueous
extracts from the nitrated product washing step that contains nitrophenolic
compounds. The nitrophenolics are dissolved in the water as their corresponding
phenate salts. These salts are bactericides and cannot be released untreated into a
conventional biological plant. Larbig claimed a process to treat the effluent water
under pressure at temperatures above 170 °C resulting in the decomposition of
73
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 the nitrophenolics to carboxylic acids, which can then be successfully treated in
a subsequent biological wastewater treatment plant. The treatment is done in a
liquid filled system without the addition of air.
Lailach et al. (8) published a patent for a SAC (sulfuric acid concentration)
plant in an isothermal nitration plant, using a horizontal evaporator using a steam
heated exchanger constructed with tantalum. The evaporator is operated under
vacuum conditions.
Rae et al. (9) introduced a plug flow reactor with built-in devices to form and
regenerate the organic droplets in an emulsion from mixed acid and organics. The
reactor has been called the jet impingement reactor, because of the formation of
jet streams which impinge at a wall to afford fine droplets of the dispersed phase.
McCall (10) describes another technology to use the heat of the nitration
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

reaction in the process. Benzene was vaporized and fed to the nitration reactor
along with the mixed acid, where the reaction takes place in the liquid phase at
a temperature of about 120 – 160 °C. The heat of reaction is used to evaporate
the water formed in the reaction as well as the excess benzene as an azeotropic
vapour phase. The reaction resembles an adiabatic process. The vapour leaving
the reactor is condensed. The condensate forms 2 phases, a lighter benzene phase
and a water phase.
Gas-phase Nitration is the subject of the patent of Sato et al. (11). It is only one
example of numerous patents describing a gas-phase nitration. The motivation to
switch to gas-phase nitration is to avoid the large quantities of waste sulfuric acid
generated in a mixed acid nitration. The nitrating agents are nitrogen oxides NO2
or N2O4. The reaction occurs in the presence of a solid catalyst comprised of acidic
mixed oxides such as WO3, MO3, TiO2 and optionally SiO2 or ZnO. Other patents
of Sato mention catalysts comprised of acidic clay minerals ion exchanged with
polyvalent metals. The process has been tested on a laboratory scale.
Brereton et al. (12) integrated the vent gas treatment with the nitration plant.
He describes a pressurised absorption column for NOx gases to produce a weak
nitric acid which is recycled to the reactor as a separate feed. Thus the yield of
nitric acid can be improved. The NOx content remaining in the vent gas is small,
and in some instances can be discharged without further treatment. If organic is
present in the gas stream off the adsorption column, then the stream can be sent to
a volatile organic oxidizer before being discharged to the atmosphere.
Hermann et al. (13) describe a mixing apparatus used to mix nitric acid,
sulphuric acid and an aromatic organic to achieve a rotating main flow in a central
mixing tube at the entrance region of the reactor to form an emulsion.
Gillis et al. (14) present a tubular plug flow reactor for nitration having built-
in static mixing elements separated by coalescing zones. The mixing elements
provide for efficient nitration in a two-phase system (organic and mixed acid) due
to the formation of small droplets in the mixing zones. It is claimed that lower
levels of impurities result in this process configuration.
Knauf et al. (15) claim the use of an electrophoresis device to facilitate the
separation of organics from the wash water. The washed organic phase is passed
through an electric field created by the electrophoresis device. The organics still
contain small droplets of conductive water which are electrically charged and
migrate to charged metal plates where they coagulate and can be separated. The

74
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 purity of the organic product can be improved. This is an electrostatic coalesce of
sorts.
Boyd (16) describes a method to treat waste water from mixed acid nitration.
A concentrated alkaline aqueous extract of the organic product, containing
dissolved sodium nitrophenates is treated in a process involving supercritical
water oxidation.
Eiermann et al. (17) describes a liquid phase nitration of an aromatic
hydrocarbon using NOx and oxygen gas in the presence of a heterogeneous oxide
catalyst, in the presence of at least 0.1 mol% water (with respect to the aromatic
hydrocarbon). Use of the catalyst in the liquid phase reaction avoids the corrosive
medium of hot sulfuric acid.
Berretta (18) defines other process conditions for the adiabatic reaction to
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

form MNB in mixed acid media. A process is claimed wherein the nitric acid
concentrations is kept as low as reasonably possible (less than 3% in the mixed
acid), and to start the reaction at a low temperature (60°C – 96°C). The temperature
rise in the reactor is limited to about 20°C. Limiting the temperature was the main
factor identified that limited the formation of the typical by-products (<1200 ppm
of nitrophenolics, < 80 ppm dinitrobenzene).
Berretta (19) describes a process to reduce the amount of dinitrobenzene
formed in the adiabatic nitration of benzene in a plug flow reactor. Sulfuric acid
and benzene are fed to the front end of the tubular reactor. The nitric acid is fed
into the reactor at multiple points. Some of the nitric acid is combined with the
incoming sulphuric acid, and the balance of the nitric acid is fed a one, two, or
more points further down the reactor. Splitting the nitric acid feed in this manner
results in less DNB formation.
Gattrell (20) describes a wet air oxidation method to treat wastewater
generated in the manufacture of nitrated aromatic hydrocarbons in mixed acid
media. Wastewater containing nitrophenolic by-products containing at least three
equivalents of sodium hydroxide with respect to the total amount of nitrophenolic
compounds was heated in a reactor at 300 °C at ~450 psig for 90 minutes to
provide a water effluent that was amenable to convention biological wastewater
treatment.
Berretta (21) teaches a method to treat a nitrobenzene product, that has already
been washed with water and with alkaline water, with an acidic wash water with a
pH of <6. The acid used to acidify the water is typically nitric acid, but other acids
are described. The washed product is then fed to a stripper column or distillation
column to remove the volatile acid, the starting organic reactant, and some of the
neutralized by-products, to ultimately afford a more pure product.

Summary of Patents Related to the Nitration of Toluene

The major difference between the mixed acid nitration of toluene and the
mixed nitration of benzene is that the process to nitrate toluene, to produce mono
and dinitrated products, is commercially practiced using isothermal reaction
systems. Adiabatic systems, however, are described in the patent literature.

75
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Hoek (22) describes a process for the manufacture of trinitrotoluene (TNT),
executing the reaction in 3 steps (from toluene to MNT, MNT to DNT, DNT to
TNT). The reactions were done in a batch mode, but a continuous system is also
described.
Rowland (23) patented improvements for the continuous nitration process for
the production of MNT in a first step, and DNT from MNT in a second step. Thus
he could influence the isomer ratio in the product in favour of the 2,4-DNT isomer.
Samuelsen (24) patented a continuous isothermal nitration process of toluene
using an arrangement of several stirred tank reactors by feeding toluene in a
counter-current fashion with a nitrating agent (typically a mixed acid system of
sulfuric acid and nitric acid). NOx recovery was integrated into the process by
conveying NOx streams back to the reactors.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

Milligan et al. (25) describe an improved method to purify the spent acid from
the manufacture of DNT in three steps. The first step involves treating the acid that
had been separated from the product with an oxidizing agent (such as hydrogen
peroxide) to convert nitrous acid to nitric acid. Nitrous acid content in the recycled
sulfuric acid was linked to the color of the DNT product. The treated sulfuric acid
was then extracted with the toluene feed to remove the nitric acid. Finally, the
sulfuric acid was again treated with an oxidizing agent to destroy residual organics
present in the acid.
Gerken et al. (26) describe a fully-integrated SAC unit for the isothermal
two-step nitration of toluene to form MNT and DNT. The sulfuric acid is fed to
the process in a counter current manner, from the DNT stage to the MNT stage.
The acid leaving the MNT nitration section of the plant is then fed to a horizontal
evaporator, equipped with a steam-heated reboiler (tube and shell). The evaporator
is run under vacuum (40 to 100 mbar) at 170 to 185 °C. The overheads of the
evaporator contain mostly water, MNT and DNT. The oveerheads is condensed.
Because of the low temperature of the coolant on the condenser, the 2,4-DNT
can crystallize in the condenser, impairing the heat transfer or even plugging the
equipment. This can be avoided by injecting MNT or a mixture of MNT/DNT,
preferably from the mono nitration step, into the evaporator or directly into the
vapour off the evaporator prior to the condenser. The DNT is kept in solution at
the condensation temperature.
Adams (27) et al. claim a solvent extraction method to recover nitrophenolic
by-products from the wash water. This is accomplised by adjusting the pH of
the waste water according to the distribution coefficients of the by-products so
that extraction can be done selectively. For example, the basic wash water from
a nitrobenzene facility contains trace nitrobenzene and polynitrated phenolics,
such as dinitro- and tri-nitrophenol. The basic wash water is extracted with a
hydrocarbon solvent to remove nitrobenzene. The pH of the basic water can be
adjusted to 2 to 3 with sulfuric acid, and then extracted with ethylbenzene to
recover high purity marketable dinitrophenol, after removal of the ethylbenzene.
The aqueous phase can then further be acidified with sulfuric acid to a pH of 0.8
to 1.2, and again extracted with a hydrocarbon solvent. The hydrocarbon solvent
is removed to afford picric acid with a purity of 99%.
Quakenbush (28) described the reaction between toluene and nitric acid
in the liquid phase, and in the presence of a selected hydrated nitrate metal

76
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 salt. Sulfuric acid is not used. The molar ratio of toluene:HNO3:metal salt was
approximately 1.0:20.0:0.9. An example of the hydrated nitrate system was a
mixture of Mg(NO3)2 and Zn(NO3)2. Nitration in the presence of the metal salt
reduced the shock sensitivity of the product reaction mixture. The reactor had
to be cooled to control the heat of reaction. The process used a large excess of
nitric acid (20 moles HNO3 per mole of toluene). The formation of phenolic
by-products was minimized in this process, and was less than 350 ppm in total.
Mason (29) describes a process for producing DNT in the liquid phase using
anhydrous nitric acid and toluene at low reaction temperatures but without using
the hydrated nitrate salts. The product is free from nitrocresols.
Schieb et al. (30) propose one single adiabatic step to produce DNT where
the reaction heat is used to reconcentrate the spent acid.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

Schieb’s invention was elaborated by Klingler et al. (31). Toluene is nitrated

in mixed acid media using weak nitric acid in sulfuric acid to afford MNT in an
adiabatic process. The acid and MNT are separated in a decanter. The acid is sent
to a flash unit to remove a portion of water. Strong nitric acid is then added to
the effluent acid from the flash vessel. This mixed acid is then contacted with the
MNT product from the first stage nitration in a second nitration reaction vessel.
The reaction mixture from the second stage nitrator is sent to a decanter. The
acid phase from the decanter is sent to another flash unit and a portion of water is
removed overhead and condensed. The overheads contain DNT. MNT is added to
the vapour stream to avoid fouling of the condenser off the second flash unit.
Hermann et al. (32) optimize the recovery of acids in the washing area of
DNT manufacture. He describes a counter current flow of diluted aqueous solution
containing nitric, sulfuric, and nitrous acid against the flow of crude DNT product.
The resulting aqueous extract can be fed directly back to the nitration vessel, or
concentrated first and then returned to the nitration vessel.
Zhang et al. (33) describe the manufacture of DNT in a nitric only process.
Toluene is fed to a loop reactor along with 65 to 75 wt% nitric acid at 35 to 65 °C,
with a short contact period to afford a solution of MNT in nitric acid. The ratio of
acid to organic is 20:1 to 30:1. The solution is cooled in a separator to <30 °C to
provide an organic product phase and an acid phase. The acid phase is recycled
back to the loop reactor with fresh acid. The MNT phase is then fed to a second
continuously stirred tank reactor (CSTR) loop reactor along with ~94 wt% nitric
acid at 30 to 65 °C. The second nitration reaction mixture is then treated with an
alkali metal nitrate to effect phase separation of the product from the acid phase
(see Mason, R.W., US Patent 5,001,272). The product phase is then purified in the
conventional manner.
The Klingler et al. (34, 35) patented aspects of the adiabatic mononitration/
dinitration of toluene in mixed acid media. In one example toluene is nitrated in
a continuous single nitration step to afford DNT that contains residual MNT. The
molar ratio of HNO3:toluene is about 1.9:1.0. The heat of nitration and optionally
additional heat is used to distill a porton of the water present from the reaction
mixture. The presence of MNT in the overheads prevents fouling of the condenser.
Additional nitric acid can then be added to the reaction mass, to convert any
residual MNT to DNT, before separation of the organic product phase from the
acid phase in a separation vessel.

77
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Münnig et al. (36, 37) combine the alkaline and the acidic water streams
resulting from the washing of a nitrated aromatic, along with the aqueous
condensate from the SAC. The pH of the combined stream is adjusted to 5.
The organic present in the combined aqueous solution is allowed to settle as a
separate phase and removed. This combined aqueous stream is then extracted
with toluene. The toluene extracts nitrophenolics, MNT, and DNT. The toluene
phase is then fed to the first nitration reactor.
Hermann et al. (38) describe a process to remove and to recover all acids and
NOx from a crude nitrated aromatic product by means of a continuous extraction
process using a cross current washing process coupled with a counter current
multistage extraction process. Simplistically, the nitration mixture is allowed to
phase separate to provide an organic product phase and an acid phase (composed
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

of sulfuric acid, residual nitric acid, dissolved to produce concentrated sulfuric

acid, which is sent back to nitration. The organic phase is then washed with water
in a cross current process in a stirred tank and the resulting emulsion is sent to a
settler. The organic phase from the settler is washed again with water in a cross
current manner in another stirred tank. The emulsion from this stirred tank is sent
to another settler where the organic and aqueous phases separate. The wash waters
from the settlers can either feed the SAC unit or be sent directly back to nitration.
The organic product from the last settler is then washed in a conventional counter
current manner. Vent gases from the wash vessels is sent to a NOx scrubbing unit
to recover the NOx as dilute nitric acid, which can then be sent back to nitration.
Fritz et al. (39) describes an electrolytic cell to treat the alkaline waste water
from washing a nitrated aromatic product. In one example, the anode compartment
of an electrolytic cell was charged batch-wise with the 100 mL of brown basic
water from washing of a nitrated product that was composed of 1.5 g/L of NaNO2,
20 g/L of Na2CO3, 10 mg/L of MNT, 1 g/L of DNT, and 0.5 g/L of picric acid, with
a chemical oxygen demand of 6500 mg/L. The cathode compartment was charged
with 1M sodium hydroxide. Current was passed through the cell for 4 hours at 60
°C. The majority of the organic material was electrochemically oxidized to afford
a clear aqueous phase, the chemical oxygen demand of which was 150 mg/L.
Bae et al. (40) claim mixing the alkaline and acidic water washes produced
during the purification of MNB or DNT with the condensate from a SAC unit. The
resulting solution is pH adjusted to <5 and cooled to <50 °C. Any nitroaromatic
that phase separates from the cooled aqueous phase is recovered. The water phase
is then concentrated by vacuum distillation, and the bottoms product is incinerated.
Denissen et al. (41) describe methods for continuously preparing a
mononitrated aromatic, such as MNB, in an adiabatic process. A plug flow reactor
design is described that employs internal mixing elements.
Mackenroth et al. (42) describe the nitration of aromatic hydrocarbons
(toluene and benzene) using aqueous metal nitrates as the nitrating media at ~100
°C, wherein the metal is in the +3 oxidation state, M(NO2)3. The molar ratio of
water to metal nitrate salt is on the order of 4:1 to 9:1. Sulfuric acid is not used in
the process. Nitric acid can also be added to the metal nitrate media to effect the
desired nitration reaction.

78
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The patents reviewed above are a selection of a large number of published
patents. The author selected the patents above to review, taking into account their
novelty and relevance to the nitration industry.

Trends
The histograms in Figure 2 (benzene nitration) and in Figure 3 (toluene
nitration) give the patent frequency over the years by assigned process areas.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

Figure 2. Patent Frequency by Area - Nitration of Benzene.

Figure 3. Patent Frequency by Area - Nitration of Toluene.

79
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Analysis
There is a recent spike in patent activity for the nitration of benzene. Gas phase
reaction technology has been a relatively active area of study since the 1940’s,
peaking around the 70’s to 80’s, and declining since. This decline may be due to
the emergence of adiabatic processes. SAC technology patents are numerous in
the 70’s to 80’s. Waste water treatment patents peaked between 2000 and 2009.
Reaction patents are numerous in the last 3 decades, following the trend to conduct
the reactions adiabatically.
There are a few waste gas treatment related patents. This could potentially
change, as more strict NOx regulations are on the horizon globally.
The patents surveyed concern the manufacture of MNB, MNT/DNT or both.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

Most patents are from companies that have been making these products for many
years. There are few patents from small independent companies.

Summary
Following the historic development of the nitration technology over time as
described in the patent literature, one can state the following:

• The adiabatic process has become state of the art for benzene nitration.
• Waste water treatment technologies are moving from extraction processes
to thermal processes (for safety reasons).
• There are many patents claiming adiabatic processes for the nitration
of toluene. However, to the author’s knowledge, new plants are still
designed using isothermal processes. The reason may be in the shift
of isomer ratios or because of safety concerns that come with higher
operating temperatures.
• Some patents focus on improvement of the washing area to isolate the
by-products as well as to recover nitric acid.
• The field of gas-phase nitration has some activity, but to date is not
competitive with the liquid-phase process using sulfuric acid. It is still a
challenge to find a convenient catalyst to assist in gas-phase nitration.

References
1. For the reader that is not familiar with the manufacture of DNT in mixed
acid media, see Hermann, H.; Gebaurer, J.; Konieczny, P. Industrial Nitration
of Toluene to Dinitrotoluene: Requirements of a Modern Facility for the
Production of Dinitrotoluene. In Nitration: Recent Laboratory and Industrial
Developments; Albright, L. F., Carr, R. V. C., Schmitt, R. J., Eds.; ACS
Symposium Series 623; American Chemical Society, Washington, DC, 1996; pp
234−249.
2. Castner, J. B. Nitration of Organic Compounds. U.S. Patent 2,256,999, 1941, Du
Pont.

80
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 3. Alexanderson, V.; Trecek, J. B.; Vandervaart C. M. Adiabatic Process for
Nitration of Nitratable Aromatic Compounds. U.S. Patent 4,021,498, 1977,
Cyanamid.
4. Alexanderson, V.; Trecek, J. B.; Vandervaart C. M Continuous Adiabatic Process
for the Mononitration of Benzene. U.S. Patent 4,091,042, 1978, Cyanamid.
5. Evans, C. M. Manufacture of Organic Nitro Compounds. U.S. Patent 4,973,770,
1990, C-I-L.
6. Guenkel, A. A.; Rae, J. M.; Hauptmann, E. G. Nitration Process. U.S. Patent
5,313,009, 1994, NORAM.
7. Larbig, W. Process for Working Up Effluents. U.S. Patent 4,230,567, 1980,
Bayer.
8. Lailach, G.; Gerken, R., Schultz, K.-H.; Hornung, R.; Boeckmann, W.; Larbig,
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

W.; Dietz, W. Process for the Production of Nitrobenzene. U.S. Patent 4,772,757,
1988, Bayer.
9. Rae, J. M.; Hauptmann, E. G. Jet Impingement Reactor. U.S. Patent 4,994,242,
1991, NORAM.
10. McCall, R. Azeotropic Nitration of Benzene. U.S. Patent 3,981,935, 1976, Du
Pont.
11. Sato Hiroshi I.; Shuzo Nakamura T. Process for Nitration of Benzene. U.S. Patent
4,551,568, 1985, Mitsubishi.
12. Brereton Clive, M. H.; Guenkel A. A. Nitration Process. U.S. Patent 5,963,878,
1999, NORAM.
13. Hermann G.; Gebauer J. Process for Nitration of Aromatic Compounds. U.S.
Patent 5,736,697, 1998, Meissner.
14. Gillis P. A.; Braun, H.; Schmidt, J.; Verwijs, J. W.; Velten, H.; Platkowski, K.;
Tubular Reactor Having Static Mixing Elements Separated by Coalescing Zones.
U.S. Patent 7,303,732, 2007, Dow.
15. Knauf, T.; von Gehlen, F.-U.; Schmiedler, J.; Pilarczyk, K.; Drinda, P. Process
for the Production of Nitrobenzene. U.S. Patent 7,326,816, 2008, Bayer.
16. Boyd, D. A.; Stuart, A. G.; Guenkel, A. A. Integrated Effluent Treatment Process
for Nitroaromatic Manufacture. U.S. Patent 6,288,289, 2001, NORAM.
17. Eiermann, M.; Ebel, K. Nitration of Aromatic Hydrocarbons. U.S. Patent
6,362,381, 2002, BASF.
18. Berretta, S. Adiabatic Process for Making Mononitrobenzene. WO Patent
2010051616, 2010, NORAM.
19. Berretta, S.; Boyd, D. A. Method for Reducing the Formation of By-Product
Dinitrobenzene in the Production of Mononitrobenzene. WO Patent
2010054462, 2010, NORAM.
20. Gattrell, M. Subcritical Partial Oxidation for Treatment of Nitration Wastes. WO
Patent 2010130049, 2010, NORAM.
21. Berretta, S. Washing System for Nitroaromatic Compounds. WO Patent
2011021057, 2011, NORAM.
22. Hoek, T. J. J. Toluene Nitration Process. U.S. Patent 2,475,095 1949,
Staatsmijnen NL.
23. Rowland, K. A. Two-Zone Mononitration of Toluene. U.S. Patent 2,947,791,
1960, Du Pont.

81
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 24. Samuelsen, E. Manufacture of Nitrotoluene. U.S. Patent 3,204,000, 1965,
Aktiebolaget Chematur.
25. Milligan, B.; Huang D. S. Process for Refining Aqueous Acid Mixtures Utilized
in Nitration of Aromatics. U.S. Patent 4,257,986, 1981, Air Products and
Chemicals.
26. Gerken, R.; Lailach, G.; Becher, D.; Witt, H. Process for the Production of
Dinitrotoluene. U.S. Patent 4,663,490, 1987, Bayer.
27. Adams, G.; Bayer, A. C.; Farmer, A. D.; Brenda, J. H. Selective Recovery of a
Nitrophenolic By-Product from Nitration Waste Water by Extraction. U.S. Patent
4,986,917, 1991, First Chemical.
28. Quakenbush, A. D. Process for Preparing Dinitrotoluene. U.S. Patent 5,302,763,
1994, Olin.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch006

29. Mason, R. W. Process for the Production of Dinitrotoluene. U.S. Patent

5,354,924, 1994, Olin.
30. Schieb, T.; Wiechers, G.; Sundermann, R.; Zarnack, W. Process for the
Preparation of Dinitrotoluene. U.S. Patent 5,345,012, 1994, Bayer.
31. Klingler, U.; Schieb, T.; Wiechers, G.; Zimmermann, J. Process for the
Production of Dinitrotolune. U.S. Patent 5,679,873, 1997, Bayer.
32. Hermann, H.; Gebauer, J. Recovery of Nitric Acid from Nitration Process. U.S.
Patent 5,756,867, 1998, Meissner.
33. Zhang, C.; Pennington, T.; Baird, J. W.; Quakenbush, A. B.; Goldstein, S. L.;
Lickei, D. L.; Whitman, P. J. Two-Stage Dinitrotoluene Production Process. U.S.
Patent, 5,948,944, 1999, Arco Chemical.
34. Klingler, U.; Schieb, T.; Wastian, D.; Wiechers, G.; Zimmermann, J. Process for
the Production of Dinitrotoluene. U.S. Patent 6,258,986, 2001, Bayer.
35. Klingler, U.; Pirkl, H. G.; Schieb, T.; Wastian, D.; Adiabatic Process for
Producing Dinitrotoluene. U.S. Patent 6,528,690, 2003, Bayer.
36. Münnig, J.; Wastian, D.; Lorenz, W.; Keggenhoff, B. Process for Working Up
the Waste Water Obtained in the Preparation of Dinitrotoluene. U.S. Patent
6,936,741, 2005, Bayer.
37. Münnig, J.; Wastian, D.; Lorenz, W.; Keggenhoff, B. Process for Working
Up Secondary Components in the Preparation of Dinitrotoluene. U.S. Patent
6,953,869, 2005, Bayer.
38. Hermann, H.; Gebauer, J.; Konieczny, P.; Haendel, M. Recovery of Nitrating
Acid Mixtures from Nitration Processes. U.S. Patent 7,470,826, 2008, Meissner.
39. Fritz, R.; Haase, S.; Allardt, H.; Zoellinger, M.; Reetz, M.; Friedrich, H-J.
Process and Apparatus for Separating Nitroaromatics from Wastewater. U.S.
Patent 20110284391, 2011, BASF.
40. Bae, C. P. Method for Treating Wastewater Produced during the Manufacture
of Nitro Compounds. WO Patent 2011068317, in Korean Language, 2011,
HuChems.
41. Denissen, L.; Stroeter, E.; Arndt, J.-D.; Mattke, T.; Heinen, K.; Leschinski, J.
Process for Preparing Mononitrotoluene. WO Patent 2011023638, 2011, BASF.
42. Mackenroth, W.; Buettner, J.; Stroefer, E.; Voigt, W.; Bok, F. Process for
the Preparation of Nitrated Aromatics and Mixtures Thereof. U.S. Patent
20110306795, 2011, BASF.

82
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 7

Advances in Water Treatment of

Effluents fromMononitrobenzene (MNB)
Production Facilities
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

Steven Buchi,*,1 Sergio Berretta,2 and Tony Boyd1

1NORAM Engineering and Constructors Ltd., Vancouver, B.C., Canada
2B.C. Research Institute, Burnaby, B.C., Canada
*E-mail: [email protected]

Nitrophenolic compounds are formed as by-products in the

industrial production of MNB. Normally, these by-products
must be removed from the product MNB and end up in the
water effluent of industrial facilities. These compounds are
bio-toxic and must be treated before the effluent is discharged to
a biological treatment facility or the environment. A short list of
treatment options includes the following proven technologies,
representing current industrial best practices: thermal
destruction, wet oxidation, ozonation, solvent extraction, and
incineration. Biological treatment is a secondary treatment step
for further purification of aqueous waste streams. The above
technologies are used by different MNB manufacturers, but no
comparative analysis of these technologies is available in the
open literature. In this paper, the authors try to fill this void
by presenting a comparison of these treatment options, using
public literature, with reference to three main criteria: ability
to meet current environmental standards, economic targets, and
process robustness.

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Introduction
In the production of Mononitrobenzene (MNB), as with many other nitrations
of aromatic compounds, nitrophenolic species are formed as a by-product of
the nitration reaction. In the case of industrial MNB production these are
predominantly di-nitrophenols but the mono- and tri- (more commonly known as
picric acid) nitrophenol isomeric species are also formed to some degree.
Typically, these nitrophenolic by-products need to be removed prior to
the MNB being processed in the downstream processes. This is achieved
by contacting the crude MNB with an alkaline water stream in a multi-stage
counter-current washing system. The nitrophenols, which are organic acids, are
neutralized and extracted into the alkaline water phase. This nitrophenol rich
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

water effluent stream is generally known as “strong effluent” or “red water”.

In most MNB plants, caustic soda is used as the washing base to provide
the required alkalinity, although ammonia is also used. The drawbacks to using
ammonia are that it is a weaker base than caustic and as such results in an inferior
washing efficiency. Also ammonia introduces the possibility of forming unstable
ammonium nitrate/nitrite salts which require experience and careful system design
to ensure a safe and reliable plant. However, ammonia is preferred as the washing
base if incineration is chosen as the strong effluent treatment option, which is
discussed later.
Nitrophenols are bio-toxic and therefore this effluent stream normally
necessitates pre-treatment to reduce its toxicity before it can be discharged to a
biological treatment facility or to the environment.
There are several different strong effluent treatment technologies currently
available and practiced industrially, with the preferred treatment technology often
dictated by site-specific conditions and clients’ preference. The objective of this
paper is to provide a comparative analysis of the different effluent treatment
technologies with regards to capability to meet current environmental standards,
economic targets, and process robustness.

Background
Thermal Destruction

Thermal destruction (1) is the process whereby the aromatic ring of the
nitrophenol molecule is thermally cracked at elevated temperature and pressure
in the absence of oxygen. The process operates at sub-critical conditions with
a temperature of approximately 540 °F to 640 °F (280 °C to 340 °C) and at a
sufficient pressure to ensure that the water effluent remains in the liquid state.
Under these conditions, and given sufficient residence time, the nitrophenols are
broken down into short chain hydrocarbons, mostly comprised of organic acids.
Nitrogen bound to the nitrophenols is typically converted to ammonia (NH3) as
well as some amount of nitrate/nitrite (NO3-/NO2-). Figure 1 below is a graphical
representation of the process.

84
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

Figure 1. Simplified Schematic of the Thermal Destruction Process.

In general, the use of thermal destruction in the treatment of strong effluent

does not lead to an effluent suitable for discharge into the environment. However,
it sufficiently destroys the nitrophenols, reducing the bio-toxicity to a level that
the effluent can subsequently be treated in a biological treatment facility. It should
be noted that even though this technology reduces the toxicity of the effluent,
the Chemical Oxygen Demand (COD) and total nitrogen (TN) of the effluent
remains largely unchanged. Depending on the allowable discharge limits of
nitrogen species and the capability of the downstream biological treatment plant
a denitrification step may be required.
Two companies have developed and offer competing thermal destruction
technologies. These two technologies are in operation in several nitration plants
with proven safety records. With the many years of cumulative operating time
(>100 years), there has only been one safety incident that has been reported
involving a Thermal Destruction process. The design of the involved thermal
destruction technology has subsequently been improved to increase the layers of
protection against the process upset that led to the event.
Operating costs of thermal destruction units are relatively low as heat
integration allows for almost complete energy recovery.

85
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Wet Oxidation
In Wet Oxidation (2), the strong effluent is oxidized by the addition of air
or oxygen under sub-critical conditions. The process operates under conditions
similar to Thermal Destruction; that is, at high temperatures (between 540 °F and
640 °F) and at a pressure to ensure the effluent remains in liquid state. The addition
of an oxidant helps to oxidize the higher molecular weight compounds into lower
molecular weight molecules and thereby reduces the COD of the treated effluent.
Also, due to the oxidizing environment, the nitrogen bound to the nitrophenols is
expected to form mostly nitrate/nitrite (NO3-/NO2-).
Following the Wet Oxidation Reactor the treated liquid is sent to a flash drum
where the excess air/oxygen is separated from the liquid effluent. The resulting
off-gas typically requires further treatment with the normal method being a thermal
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

oxidizer or flare. Figure 2 below is a graphical representation of the process.

This technology should in principle produce an effluent which does not need
further treatment for denitrification and COD reduction. However, in practice
the COD reduction is generally lower than 80%. Therefore, further biological
treatment is usually required, although the loading will be much smaller than that
of the effluent treated by the Thermal Destruction process.

Figure 2. Simplified Schematic of the Wet Oxidation Process.

86
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 As opposed to Thermal Destruction, the Wet Oxidation process is normally
carried out under fairly acidic conditions. Therefore, the use of expensive
materials of construction is required, which results in a relatively high capital cost.
Wet Oxidation is currently being used to treat strong effluent from at least
one nitration plant in Europe. This process has similar safety concerns as with
Thermal Destruction. An additional potential operating problem experienced with
this technology is the formation of scaling due to precipitation of inorganic salts.
Partial Wet Oxidation (3), a variation to full Wet Oxidation, is a new
technology which has been piloted in the treatment of strong effluent, but has
yet to be commercialized. This technology attempts to gain some of the benefit
of reducing COD with air/oxygen addition while minimizing the aggressive
operating conditions such that less expensive materials of construction can be
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

used.

Solvent Extraction

Solvent Extraction (4) involves first acidifying the strong effluent to

precipitate the dissolved nitrophenolate salts in the strong effluent as their
conjugate organic acids. The free organic nitrophenols are then extracted by
contacting the aqueous effluent stream with an organic solvent, typically toluene,
which has a very low solubility in water. The aqueous effluent and organic solvent
streams are then separated. The water stream, saturated with the extracting
solvent, is then stripped to remove and return the solvent back to the extraction
step. The resulting effluent stream still contains significant amounts of COD,
mainly as toluene, such that direct discharge to the environment is not possible
and additional treatment in the form of a biological system or activated carbon
beds is required.
The extracting solvent, laden with nitrophenols, is sent to an evaporator
where the majority of the solvent is recovered. Complete solvent recovery is not
practical as nitrophenols will reach the solubility limit, precipitate and solidify
which causes operational and safety concerns. This concentrated nitrophenol
stream is considered a fuel and as such it is incinerated with negligible energy
disposal costs. Figure 3 below is a graphical representation of the process.
Solvent Extraction is known to be currently used to treat strong effluent from
at least two nitration plants on the US Gulf Coast.

87
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

Figure 3. Simplified Schematic of the Solvent Extraction Process.

Ozonation

Ozonation (5) is a process where the strong nitrophenol effluent is contacted

with ozone to oxidize and degrade organic compounds. The resulting effluent
contains significant levels of carbonate (CO3-2), nitrate (NO3-) and nitrite (NO2-).
The process operates at a temperature of approximately 140 °F (60 °C). The COD
of the stream is low and aromatic compounds are decomposed to very low levels.
However, depending on allowable nitrogen limits, the effluent may require a
denitrification step downstream of the ozonation unit.
Gas generated in the process contains small quantities of ozone and therefore
needs further treatment prior to being vented to the atmosphere. A thermal oxidizer
or flare to handle this small gas stream is the preferred treatment option.
Due to the relatively rapid decomposition of ozone, it is typically necessary
to generate the ozone at site using ozonators. Either dry air or oxygen can be used
as a feed to the ozonators with oxygen increasing the efficiency of the generator.
Figure 4 below is a graphical representation of the process.
Ozonation is known to be currently practiced in at least two nitration plants
on the US Gulf Coast.

88
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

Figure 4. Simplified Schematic of the Ozonation Process.

Incineration

Incineration (6) is a well-established process and is capable of treating many

industrial wastes including solids, liquids, sludges and gaseous waste streams. In
general, incinerators are considered off-the-shelf equipment with designs ranging
from a simple one stage incinerator to more complex arrangements including
recovery boilers to achieve better energy efficiency and flue gas scrubbers to
remove acid gases prior to discharge to atmosphere.
Chemicals within the waste streams are thermally decomposed to their
primary combustion products of carbon dioxide, water and ash. At temperatures
of 1300 °F (700 °C), most chemicals are broken down into lower molecular
weight compounds, and at temperatures of 1800 °F (1000 °C), nearly all wastes
are fully oxidized.
Nitrogen containing species should, in theory, be reduced to form N2;
however, in practice small amounts of NOx gases are present. However, the
NOx concentration should be below most regulatory emission levels. Sulfur
compounds in the strong effluent, such as sulfate, produce sulfur oxides (SOx) in
the incinerator. As a result, an acidic wash of the crude MNB can be done prior
to extracting the nitrophenols into the strong effluent to reduce sulfates and thus
give low SOx concentrations in the incinerator flue gas. Figure 5 below shows a
graphical representation of the process.
In general, inorganic salts produce ash in the incinerator and require an ash
removal and collection system, thereby complicating the design of the incinerator.
This is why ammonia is preferred over caustic as the washing base to extract
nitrophenols into the strong effluent, since ammonia is reduced to nitrogen (N2)
in the thermal oxidizer, whereas the ash left by the sodium greatly complicates the
oxidizer design and operation.

89
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

Figure 5. Simplified Schematic of the Incineration Process.

Incineration of strong nitrophenol waste effluent streams is known to have

been practiced in at least two adiabatic MNB plants in Europe with one of them
still operating. Both of these plants used ammonia-based washing to avoid residual
soda ash thereby simplifying the incinerator design and operation.
Operating costs of incineration are relatively high due to the large amount of
water that must be vaporized. However, energy efficiency can be increased with
the addition of a recovery boiler. Alternatively, the strong effluent stream can be
concentrated prior to incineration to minimize the amount of water required to
be evaporated as well as reducing the size of the thermal oxidizer. An additional
benefit to incineration is that no further treatment, such as biological treatment, is
required.

Deepwell/Biotreatment

There are other possible treatment options that have been used to treat the raw
strong effluent and are currently in operation, such as deepwelling and biological
treatment. These options either require strict environmental permitting and/or
large amounts of dilution water. As a result, these technologies are generally
viewed as not being available or practical and are not further discussed in this
paper.

Results and Discussion

The economics and performance of the treatment options considered above
were investigated as follows. Calculations were done for a 300,000 MTPY MNB
plant with a strong effluent stream given in Table I below.
90
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table I. Strong Effluent Flow and Composition
Flow rate 6000 kg/h
Temperature 50 °C

Nitrophenolates 1.5% wt%

MNB 10 ppm
Nitrates/Nitrites (as NO3-) 200 ppm
Sulfates (as SO4-2) 100 ppm
COD 12500 ppm
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

For each option the post-treated effluent quality was estimated based on the
available information provided in the references and is presented in Table II.

Table II. Expected Effluent Quality

Thermal Wet Solvent
Destruction Oxidation Extraction Ozonation Incineration
(ppm) (ppm) (ppm) (ppm) (ppm)
Residual
Aromatics
(e.g.
≤10 ≤10 600 ≤1 N/A
Nitrophenols,
MNB,
Toluene)
Total
Nitrogen (as 9000 9000 200 9000 N/A
NO3-)
Sulfates (as
100 100 100 100 N/A
SO4--)
COD 9300 2400 1700 low N/A

In general, Thermal Destruction only removes the bio-toxicity, whereas the

other technologies produce a more polished effluent with varying degrees of COD
reduction with Incineration producing little to no liquid effluent. Additionally,
Solvent Extraction is the only considered technology where the nitrogen species
are extracted, reducing the total nitrogen of the effluent.
Both operating and installed capital costs were also approximated for each
treatment option with the following basis for calculations. The results are shown
in Figure 6.

91
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Basis of Operating Cost Estimates

Direct operating costs for the various treatment options were estimated using
utility and chemical costs given in Table III. No allowance was made for additional
maintenance costs, extra operating staff, environmental permitting or extra lab
charges for effluent monitoring.

Table III. Utility and Chemical Costs

Description Cost (USD) Basis
Steam 20 US$/tonne
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

Cooling Water 0.028 US$/tonne

Power 0.06 US$/kWh
Natural Gas 0.3 US$/m3 NG
Caustic Soda 400 US$/tonne (100% Basis)
Sulfuric Acid 100 US$/tonne (100% Basis)
Toluene 1100 US$/tonne
Ammonia 450 US$/tonne (100% Basis)
Liquid Oxygen (LOX) 290 US$/tonne

Basis of Capital Cost Estimates

Equipment costs were estimated by pricing only the main equipment items
(pumps, heat exchangers, tanks, vessels, columns, internals, etc.) for the given
treatment technology. Auxiliary components such as piping items, instrumentation
and additional infrastructure are not included. The equipment cost was then scaled
using a Lang factor of 3 to obtain a total installed cost of the technology.
The additional costs for treatment technologies that require a thermal oxidizer
or flare for off-gas, denitrification stage to handle nitrate/nitrites and/or biological
treatment to further reduce COD, were not calculated. In many cases there will
be a site-wide biological treatment plant or a thermal oxidizer and the additional
loading will have a negligible effect on their size and operating costs.
Additionally fixed charges such as depreciation, taxes and technology licenses
were not taken into account.
Thermal Destruction is seen to have the smallest operating costs out of
all the considered technologies but has a relatively high capital cost. Whereas
Incineration with concentration has the smallest capital cost but has a relatively
high operating cost and requires significant time and cost to permit.
When comparing the installed cost plus 3 years of operating costs the most
economical treatment technologies appear to be Thermal Destruction, Solvent
Extraction and Incineration with concentration. Wet Oxidation should also be
considered.
92
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

Figure 6. Expected Operating and Capital costs for different treatment options
for a 300,000 MTPY MNB plant. (see color insert)

It should also be mentioned that the economics of Incineration is largely

dependent on the fuel costs of the region, which can vary widely. This could
lead to Incineration becoming more favorable in locations where fuel is relatively
inexpensive.

Conclusion
All the technologies discussed have proven track records and can be used
to treat strong effluent from a MNB plant. In most jurisdictions, Incineration
is currently the only technology which meets environmental discharge limits
without requiring addition treatment steps such as denitrification and/or biological
treatment.
If the site has biological treatment, Thermal Destruction is the preferred
choice. In fact most of the latest built MNB plants around the world have opted
for this option to treat their strong effluent.
In all cases selection of the chosen technology to treat strong effluent is
dependent upon the local facilities and conditions as well as environmental permit
requirements.

93
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 References
1. Larbig, W. Process for Working up Effluents Containing Nitro-Hydroxy-
Aromatic Compounds. U.S. Patent 4,230,567, 1980.
2. Sawicki, J. E.; Casas, B.; Huang, C-Y.; Killilea, W. R.; Hong, G. T. Wet
Oxidation of Aqueous Streams. U.S. Patent 5,250,193, 1993.
3. Gattrell, M. Sub-Critical Partial Oxidation for Treatment of Nitration Wastes.
WO Patent 2010/130049 A1, 2010.
4. Adams, E. G.; Barker, R. B. Process for Extracting and Disposing of
Nitrophenolic By-Products. U.S. Patent 4,925,565, 1990.
5. Schuster, L.; Stechl, H.-H.; Wolff, D. Cleanup of Nitroaromatics Containing
Wastewaters. U.S. Patent 6,245,242 B1, 2001.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch007

6. Hazardous Waste Disposal by Thermal Oxidation, John Zink Company.

http://www.johnzink.com/wp-content/uploads/hazardous-waste-disposal-
thermal-oxidation.pdf.

94
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 8

Incorporation of Oxidation Enhancement

through Hydrogen Peroxide Addition into a
Tested Mass Transfer/Reaction Model for an
Industrial NOx Absorption Process
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch008

Kyle G. Loutet,*,1 Andres Mahecha-Botero,1,2 Tony Boyd,1

Steven Buchi,1 and Clive M. H. Brereton1
1NORAM Engineering and Constructors Ltd., 200 Granville Street,
Suite 1800, Vancouver, BC, Canada V6C 1S4
2Department of Chemical and Biological Engineering,

University of British Columbia, 2360 East Mall, Vancouver, BC,

Canada V6T 1Z3
*E-mail: [email protected]

A comprehensive mass transfer/reaction model was developed

in the process simulator Aspen Plus to simulate an industrial
NOx absorption process. The model, recently published by the
authors, was tested by comparison with data collected at the
world-scale Wilton mononitrobenzene plant in Redcar, UK,
and the model was found to accurately predict NOx removal.
In the current study, an advanced NOx absorption technique,
namely the addition of hydrogen peroxide as an oxidizer to
an absorption process, was incorporated into the model in an
attempt to quantify improvements in NOx absorption. The use
of hydrogen peroxide to enhance NOx absorption processes
has been studied in the past, with specific focus on the ability
of hydrogen peroxide to oxidize nitric oxide (NO) to nitrogen
dioxide (NO2) and to oxidize nitrous acid (HNO2) to nitric
acid (HNO3). The current study, which focuses on the latter
oxidation, finds that the enhanced oxidation of HNO2 shows
significant potential for improving NOx absorption.

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Introduction
Industrial NOx absorption, utilized in the manufacture of nitric acid and
in the treatment of waste gases from the nitration and other industries, is a
complex absorption operation involving numerous gas- and liquid-phase chemical
reactions and mass and heat transfer phenomena. Designing NOx columns for
the purpose of pollution abatement, both air and water, is of particular interest
as industrialized and developing nations become increasingly conscious of their
environmental footprints. A reliable and accurate model of NOx absorption
operations represents a useful tool to improve designs and ultimately meet and
exceed environmental targets for both air and water.
In a past study, a rate-based model (i.e. a model that accounts for reaction
and mass transfer kinetics) was developed in the process simulator Aspen Plus
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch008

and tested using data gathered at the world-scale Wilton mononitrobenzene plant
in England (1). This model focused on the pressurized, counter-current absorption
of NOx into water, resulting in the production of dilute (<20 wt%) nitric acid as a
liquid product. In the current study, an alternative configuration is studied wherein
hydrogen peroxide is added to the system to enhance NOx absorption.

NOx Formation and Capture in Industrial Nitration Processes

The formation of NOx as a byproduct in nitration processes is an industry-

wide problem. In NORAM Engineering’s patented benzene nitration process (2),
in which nitrobenzene is continuously and adiabatically produced in a mixed acid
medium, the amount of NOx generated is predictable, based on knowledge of the
factors that control its formation. Two pathways to NOx formation are presented
below.

(a) Nitrous acid contained in the feed nitric acid decomposes to form nitric
oxide (NO)

(b) Nitrous acid produced in side reaction (2) decomposes to NO via reaction
(1)

The byproduct NOx gases produced in the nitration step are separated from the
liquid products in a three phase separator and sent to a NOx absorption column.
In the NOx column, the gases and air are fed to the bottom of the column and
contacted with demineralized water under pressure. The nitric acid produced in
the NOx column is collected and recycled to the nitration step, reducing the need
for more fresh nitric acid feed, lowering the NOx in the plant vent, and reducing
the nitrate/nitrite levels in the plant effluent (3). A portion of the liquid outlet is
typically cooled and circulated to the top and/or to an intermediate point in the
column. The column can contain packed or trayed sections or a combination of
both. Figure 1 shows a simplified version of the NOx absorption process. Also
96
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 shown in Figure 1 is a point where hydrogen peroxide might be introduced into a
NOx absorption system, that is, at the bottom of the column.
In order to understand the need for cooling and pressure in the NOx absorption
column, an understanding of the chemical reactions in both the gas and liquid
phases is required. The simplified reactions shown in Table 1 are adopted from
the literature.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch008

Figure 1. Schematic of industrial NOx absorption process with possible hydrogen

peroxide injection point identified. (see color insert)

The near-completion of reaction R1 is crucial to NOx absorption as NO is

marginally soluble in the liquid phase (6). Air is therefore added upstream of the
NOx column to initiate the oxidation of NO to NO2. Reaction R1 is unusual in that
it is kinetically favored by low temperature. A low temperature column is therefore
favorable for NO oxidation to NO2 and it has the added benefit of increasing the
solubility of NOx species in the liquid phase. Hence the need for a cooler as shown
in Figure 1. R1 is also third order in pressure and NOx absorption systems are
pressurized largely for this reason (6). Pressure has the added benefits of forcing
equilibrium reactions R2 and R3 to the product side and reducing the equilibrium
gas-phase concentrations of N2O3 and N2O4, which are known to be the primary
species that are absorbed from the gas phase to the liquid phase where they react
to form nitric and nitrous acid.
97
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table 1. Salient gaseous NOx and aqueous phase reactions
Reaction Stoichiometry Phase Reference
R1 2 NO + O2 → 2 NO2 Gas (4)
R2 2 NO2 ↔ N2O4 Gas (4)
R3 NO + NO2 ↔ N2O3 Gas (4)
R4 N2O3 + H2O ↔ 2 HNO2 Gas (4)
R5 N2O4 + H2O ↔ HNO3 + HNO2 Gas (4)
R6 2 NO2 + H2O → HNO2 + HNO3 Liquid (5)
R7 N2O3 + H2O → 2 HNO2 Liquid (5)
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch008

R8 N2O4 + H2O → HNO3 + HNO2 Liquid (5)

R9 3 HNO2 → HNO3 + H2O + 2 NO Liquid (5)

Once into the liquid phase, N2O3 and N2O4 react with water via reactions
R7 and R8, respectively, both of which produce HNO2. Nitrous acid in turn
decomposes via R9, producing NO which desorbs to the gas phase, bringing
the NOx absorption process full circle. The suppression of reaction R9 is the
focus of many NOx absorption enhancement systems, including enhancement by
hydrogen peroxide addition.

NOx Absorption Enhancement by Hydrogen Peroxide Addition

The suppression of reaction R9 has the potential to significantly improve the
NOx absorption process. Traditional designs have utilized bases such as caustic
soda to remove HNO2 as sodium nitrite. Although effective at reducing NOx
emissions, this method produces a liquid effluent containing sodium nitrites and
nitrates which must be treated and which cannot by recycled to the nitration plant.
The use of hydrogen peroxide to remove HNO2 represents an elegant solution to
this problem as the reaction products are nitric acid and water (R10):

The liquid waste from a peroxide-enhanced absorption process thereby avoids

costly effluent treatment and can potentially be recycled to nitration.
Hydrogen peroxide can also enhance NOx absorption by assisting the
oxidation of NO to NO2 on the gas side (R11). This would, however, require the
vaporization and gas phase injection of hydrogen peroxide into the system. The
enhancement of NO oxidation is not studied here, the reasons for which are given
in the following section.

The use of hydrogen peroxide to enhance NOx absorption has been studied in
the past (7–11) and results have shown that significant potential for improvements
in NOx absorption are achievable. Of particular interest is the observation that
98
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 NOx absorption increases with increasing nitric acid concentration when hydrogen
peroxide is present (8). In nitric acid solutions free of additives such as caustic
or peroxide, NOx absorption has been shown to decrease with increasing nitric
acid concentration (12). The ability to achieve higher nitric acid concentrations
translates to decreased water usage. If the nitric acid is to be returned to the
nitration reaction step, then increased energy efficiency is achieved since any water
added to nitration must ultimately be boiled off.
From a practical standpoint, the retrofit of an existing NOx absorption
system to allow for hydrogen peroxide introduction should be relatively simple.
The materials of construction that are commonly employed for NOx absorption
columns, such as 304 and 316 grades of stainless steel, are also compatible
with solutions of hydrogen peroxide. The number of peroxide-enhanced NOx
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch008

absorption units currently installed industrially is not known but a review of the
patent literature suggests that it has seen extensive application. One patent (13)
describes the use of hydrogen peroxide to scrub NOx from the off-gas from a
metal dissolution process while another patent (14) describes the use peroxide
to treat off-gas from a nitric acid etching step. The construction of a pilot-scale,
peroxide-enhanced NOx absorption column is also described in the literature (7).

Experimental Methods and Model Development

In the previous study (1), data from an operating NOx absorption system
were collected and analyzed. A rate-based model was developed in Aspen
Plus (utilizing the RateSep add-on) in an attempt to simulate plant conditions.
Model results were compared with plant data and the model was found to
accurately predict NOx removal. Details on the collection and analysis of plant
data, development of the model, including but not limited to reaction kinetics,
vapor-liquid equilibria, and selection of mass transfer correlations can be found
in the previous study (1). A schematic of the NOx absorption process is shown in
Figure 1.
In the current study, no additional plant data were collected. The modeling
results from the previous study are compared to the modeling results from the
current study, which now includes hydrogen peroxide addition to the absorption
system. All data presented herein represent modeling results.

Addition of Hydrogen Peroxide to the Rate-Based Model

Hydrogen peroxide is incorporated into the rate-based model in the liquid

phase via reaction R10. Although hydrogen peroxide can also enhance NOx
absorption in the gas phase via reaction R11, this is not accounted for in the
model. The reaction of NO with HNO2 in the gas phase is a complex reaction
mechanism involving free radicals and its incorporation into the model will
require an additional study.
It should also be noted that from an operability standpoint, the authors feel
that peroxide injection into the liquid phase is the more appealing option. Whereas
peroxide injection into the gas phase necessitates a vaporizer and potentially
99
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 increases the load on the compression system, injection to the liquid phase is
straightforward. The peroxide can simply be added to the absorber sump on flow
control.
The oxidation of HNO2 to HNO3 by H2O2 in the liquid phase is fast and the
rate expression shown below has been developed (11).

where k = 3012 L2 mol-2 s-1

Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch008

A number of simulations are presented in the Results and Discussion section.

The previous study (1) documented 17 experimental plant trials. Although
the original intention was to test different process conditions in each trial, the
experimenters were constrained due to plant limitations with regard to safety,
production, and process control. Details about the 17 trials can be found in the
published study (1).

Results and Discussion

Under normal conditions in the NOx absorption process, NOx gas enters the
column at a concentration on the order of 8-10% by volume. The gas passes
through a lower section of packing which operates with a large circulating flow
of liquid and the concentration drops to about 10,000 ppmv. Finally, the gas
passes through an upper packed bed in which demineralized water passes “once
through” and wherein the NOx concentration drops to about 300-400 ppmv in
the vent gas. In the previous study, model predictions were found to match plant
data to within about 3%. Under normal operation, the NOx column is operated
at approximately 2.7 bar(g). Note that in Figure 1 an effluent liquid stream is
shown as a side draw from the middle of the column. The side draw is in place to
allow the column’s upper section to be operated with caustic solution rather than
demineralized water. When the column upper section is run with demineralized
water, the effluent stream is simply combined with the nitric acid solution off the
column bottoms. When operated with caustic solution to the upper bed, the effluent
side draw becomes laden in sodium nitrate/nitrite, which then requires separate
treatment in a biological treatment system. Avoiding this biological treatment is
the key driver to operating the upper column section with demineralized water
rather than caustic solution.
Under conditions of NOx absorption with peroxide addition, the predicted
improvements in NOx capture efficiency are significant. Table II shows the
model-predicted NOx concentrations from the lower packed bed under normal
conditions and under conditions of peroxide addition. Model convergence issues
were encountered for Trials 13 and 14 and have hence been omitted from the
study. Results are represented graphically in Figure 2.
100
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Some interesting observations are noted from these results. Firstly, the
lower bed performance with peroxide addition approaches the total column
performance for normal conditions. Specifically, gas exiting the lower bed has
a NOx concentration on the order of 300-400 ppmv, whereas without peroxide
addition, the NOx concentration does not reach the 300-400 ppmv range until
after the upper bed. This implies that less packing material and a shorter column
might be feasible when peroxide addition is utilized.
It is also interesting to note that the simulations were re-run at atmospheric
pressure with peroxide addition, whereas normal conditions in the column are 2.7
bar(g) and no peroxide addition. These results, too, are shown in Table II. What
is interesting is that at atmospheric pressure with peroxide addition, the predicted
NOx composition leaving the lower bed is similar to the predictions at the normal
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch008

(i.e. pressurized) conditions without peroxide addition. In systems where a

final NOx concentration of 5000-10000 ppmv is acceptable, operation of the
NOx absorption column at or slightly above atmospheric pressure with peroxide
addition could mean significant capital and operating cost savings (compared
to pressurized without peroxide) as it would eliminate or reduce the size of the
compressor, and potentially lower the mechanical design pressure of the column
and cooler. As will be demonstrated in Figure 3, however, achieving 300-400
ppmv at atmospheric pressure with peroxide addition remains a challenge.
The upper packed bed was also studied and results are shown in Table III and
represented graphically in Figure 3. Here, NOx concentrations refer to the gases
exiting the top of the column. The improvement in performance is not nearly as
pronounced as in the lower packed bed. This is in large part due to how peroxide
is added to the system. Under the proposed arrangement, peroxide is added to
the bottom of the column and is therefore only exposed to the bottom bed. The
suppression of the nitrous acid decomposition reaction is therefore not achieved in
the upper packed bed. Although the results are not presented herein, suppression
of R9 (i.e. nitrous acid conversion to water, nitric acid and NO) in the upper packed
bed is not nearly as effective as in the lower bed and the results are only marginally
better than those shown in Table III. One hypothesis to explain this is that nitrous
levels in the upper bed are much lower than in the lower bed since the bulk of the
NOx gas has been absorbed in the lower bed. The nitrous decomposition reaction
(R9) is therefore less prevalent in the upper bed and so its suppression has a less
marked effect on NOx absorption
The other obvious difference for the upper packed bed is the performance
at atmospheric pressure with peroxide addition. Performance at atmospheric
pressure with peroxide is significantly lower than at normal conditions (i.e.
pressurized without peroxide). This is likely due to the fact that the gas leaving
the lower bed contains significant quantities of nitric oxide (NO) and the kinetics
of the oxidation of NO to NO2 are third order in pressure. At atmospheric
pressure, very little of the NO escaping the lower bed is oxidized to NO2 and
since NO is sparingly soluble in the liquid phase, the overall absorption of NOx
is significantly reduced.

101
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table II. Lower packed bed performance with and without peroxide addition
NOx Outlet Concentration from Lower Packed Bed
Trial Normal Conditions With Peroxide With Peroxide Addition at
(ppmv) Addition to the column Atmospheric Pressure (ppmv)
bottom (ppmv)
1 6385 343 5015
2 6301 329 4245
3 7088 354 3869
4 6952 452 3575
5 9762 771 9174
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch008

6 14888 1537 12852

7 7265 395 4606
8 6875 333 4539
9 4307 433 6146
10 7997 544 6806
11 7870 500 5500
12 10423 808 6553
15 6088 397 5620
16 6236 409 5998
17 6451 425 5966

Figure 2. Graphical representation of lower packed bed performance with and

without peroxide addition. (see color insert)
102
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table III. Upper packed bed performance with and without peroxide
addition
NOx Outlet Concentration from Upper Packed Bed
Trial Normal Conditions With Peroxide With Peroxide Addition at
(ppmv) Addition to the column Atmospheric Pressure (ppmv)
bottom (ppmv)
1 1322 230 4280
2 423 164 3148
3 396 157 1804
4 329 147 2671
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch008

5 474 224 5209

6 501 267 3888
7 397 163 3227
8 426 165 3312
9 712 248 4781
10 556 223 4089
11 466 216 3595
12 932 356 3897
15 393 160 3512
16 395 164 3679
17 406 168 3713

Figure 3. Graphical representation of upper packed bed performance with and

without peroxide addition. (see color insert)
103
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Conclusions and Recommendations
As regulations on NOx emissions to the atmosphere and nitrite/nitrate level
in effluents become more stringent, scientists and engineers will need to devise
and implement new NOx abatement technologies. These new technologies must
be energy efficient. The addition of a chemical agent such as hydrogen peroxide
appears to have great potential to reduce NOx emissions without producing
nitrite/nitrate-laden effluents. In certain applications, such as when final NOx
concentrations of 5000-10000 ppmv can be tolerated, its use may also allow for
lower pressure NOx absorption systems to be adopted, meaning decreased energy
consumption and increased safety. The viability of atmospheric pressure NOx
absorption will greatly depend on the target outlet NOx concentration from the
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch008

column. It has been shown that achieving low NOx concentrations at atmospheric
pressure is challenging, even with the addition of peroxide.
The economics of incorporating hydrogen peroxide addition to the liquid
phase of an existing NOx absorption are not expected to be prohibitive.
Since the materials of construction often used in NOx absorption columns are
typically compatible with peroxide solutions, a significant retrofit will likely
be unnecessary. Rather, the retrofit could be as simple as adding a system to
deliver peroxide, along with the requisite provisions for flow control. Additional
equipment, such as storage tanks and offloading stations, might be required at
sites not already using hydrogen peroxide, which could significantly increase the
capital cost of the retrofit. From a variable cost standpoint, a typical price point
for 50 wt% industrial grade hydrogen peroxide is about US$750 per metric tonne.
For the Wilton plant investigated herein, the peroxide consumption is predicted
to be about 240 tonnes/year for an annual cost of US$180,000. This would have
to be compared to the variable cost of a competing technology, such as selective
catalytic reduction (SCR). Note that fixed costs associated with the addition of an
SCR would likely be much higher than those with the peroxide retrofit.
In the NOx absorption system shown in Figure 1, the compressor typically
makes up in excess of 50% of the bare equipment cost for world-scale plants
such as Wilton, and typically consumes between 75 and 150 kWe. Depending
on the cost of electricity, the variable cost for the pressurized system could be
very comparable to that of the peroxide system, which combined with the potential
50% reduction in fixed cost, could make an atmospheric pressure, peroxide system
economically attractive.
The drawbacks of adding hydrogen peroxide would include increased
chemical consumption, the challenges that come along with introducing a new
chemical into an established chemical process (especially at existing plants) and
the increased absorption system complexity.
Additional work is required before a peroxide-enhanced system is
implemented on the industrial scale in a mononitrobenzene plant. Perhaps most
importantly, the implication of recycling nitric acid containing excess peroxide
to nitration needs to be carefully studied. The excess peroxide will likely
oxidize benzene to phenol which will ultimately be nitrated and lead to higher
nitrophenol content in the effluent. Similarly in non-nitration applications, and
depending on the use of the nitric acid produced in the NOx absorption operation,

104
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 the implications of having excess peroxide in the nitric acid would have to be
considered. Also, the optimization of how peroxide is added to the absorption
system is required. Finally, experimental tests are required to assess the validity
of the simulation results.

References
1. Loutet, K. G.; Mahecha-Botero, A.; Boyd, T.; Buchi, S.; Reid, D.;
Brereton, C. M. H. Experimental measurements and mass transfer/reaction
modeling for an industrial NOx absorption process. Ind. Eng. Chem. Res.
2011, 50, 2192–2203.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch008

2. Guenkel, A. A.; Rae, J. M.; Hauptmann, E. G. U.S. Patent 5,313,009, 1994.

3. Brereton, C. M. H.; Guenkel, A. A. U.S. Patent 5,963,878, 1999.
4. Patwardhan, J. A.; Joshi, J. B. Unified model for NOx absorption in aqueous
alkaline and dilute acidic solution. AIChE J. 2003, 49, 2728–2748.
5. Hupen, B.; Kenig, E. Y. Rigourous modeling of NOx absorption in tray and
packed columns. Chem. Eng. Sci. 2005, 60, 6462–6471.
6. Joshi, J. B.; Mahajani, V. V.; Juvekar, V. A. Invited review: Absorption of
NOx gases. Chem. Eng. Commun. 1985, 33, 1–92.
7. Thomas, D.; Vanderschuren, J. The absorption-oxidation of NOx with
hydrogen peroxide for the treatment of tail gases. Chem. Eng. Sci. 1996,
51, 2649–2654.
8. Thomas, D.; Vanderschuren, J. Modeling of NOx absorption into nitric acid
solutions containing hydrogen peroxide. Ind. Eng. Chem. Res. 1997, 36,
3315–3322.
9. Thomas, D.; Vanderschuren, J. Effect of temperature on NOx Absorption into
nitric acid solutions containing hydrogen peroxide. Ind. Eng. Chem. Res.
1998, 37, 4418–4423.
10. Thomas, D.; Vanderschuren, J. Removal of tetravalent NOx from flue gases
using solutions containing hydrogen peroxide. Chem. Eng. Technol. 1998,
21, 975–981.
11. Thomas, D.; Vanderschuren, J. Analysis and prediction of the liquid phase
composition for the absorption of nitrogen oxides into aqueous solutions.
Sep. Purif. Technol. 2000, 18, 37–45.
12. Miller, D. N. Mass transfer in nitric acid absorption. AIChE J. 1987, 33,
1351–1358.
13. Gubanc, D. M.; Liston, L. J.; Zimmerman, J. M. U.S. Patent 5,151,258, 1991.
14. Osborne, W. E.; Bomber, A. J.; Gee, M. L.; Pesklak, B. C.; Dick, F. A.; Park,
E.; Yetter, C. A.; Boyce, A. R. U.S. Patent 5,637,282, 1996.

105
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 9

Bench-Scale and Pilot Plant Nitration

Experiments – Practical Considerations
Evan Hobenshield*
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

BC Research Institute, Burnaby, British Columbia, Canada

*E-mail: [email protected]

The development of new nitration processes usually starts at

the bench or pilot plant scales. The path of development is very
specific to the particular nitration process to be investigated.
Some nitration processes can be particularly hazardous; others
can be hard to scale up or control. Many of these challenges
can be difficult to manage, especially when considering a new
nitration process, or new operating conditions for a known
process, where little or no published data is formerly available.
Researchers at BC Research have found through over 20
years of nitration research that a well-designed experimental
methodology can ease many of these challenges. This
paper presents a well-established nitration reaction research
methodology. In particular, experimental design, apparatus
design, safety evaluation, scales up issues, and sampling and
analytical techniques are discussed.

Overview

The first nitration to be reported was that of benzene itself. Mitscherlich in

1834 prepared nitrobenzene by treating benzene with fuming nitric acid (1). Since
then, nitration has been the subject of continuous study. Many means have been
used through the years to effect the nitration of organic substrates. For example,
nitronium salts in solution in inert organic solvents have been used to nitrate a wide
range of aromatic compounds (2). Benzene and its derivatives have been reacted
with solutions of mercuric nitrate in concentrated nitric acid to give nitrophenols.
Today, most high volume nitrations are done in mixed acid media (sulfuric acid and

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 nitric acid). Nitrobenzene, nitrotoluene, dinitrotoluene, and nitrochlorobenzene
are all manufactured by nitrating an aromatic starting material in mixed acid.
Though nitration technology is a mature industry, specific nitrations still
require development, especially that involving high production volume. For
example, the nitration of benzene over solid acid catalyst (3, 4) is an active area of
research. The nitration of benzene in mixed acid media continues to be optimized
(5–7).
In general, nitration reactions are exothermic, and resulting nitration mixtures
can generate a large amount of energy upon decomposition. Nitrations are
potentially dangerous even on the gram scale. In addition, by-products are
typical in the nitration of aromatic substrates, which creates challenges in plant
design and operation. The corrosive nature of the mixed acid must be carefully
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

considered at the lab, pilot and industrial scale. In a research laboratory, these are
some of the many challenges that the researcher must consider.

Introduction
Through the years, BC Research has conducted numerous nitration studies,
and in the process has developed a practical methodology for undertaking nitration
research programs. The methodology includes the consideration of the following
steps:

(1) setting clear objectives,

(2) performing a thorough literature search,
(3) designing the appropriate experimental plan,
(4) designing the experimental set-up,
(5) conducting a safety audit,
(6) developing a sampling and analysis plan,
(7) and having a chemical spill cleanup plan and a chemical disposal plan.

There are many aspects of experimental design to consider in each of these

steps, and careful planning will spare the researcher of delays, of chemical releases
and of more serious accidents.

Defining the Program Objectives

A research program typically originates from a specific product or process

need. Research objectives must be defined. Frequently this effort is not given
enough consideration. Research programs should start with defined objectives
that are concise and limited in scope. Most importantly, they should be within
the constraint of available physical and financial resources. The process of
defining the objectives should follow an established method, and requires the
up-front involvement of all technology functions, management (client), and
researchers. As an example, the following established iterative procedure has
been successfully used by BC Research through many projects (Figure 1)
108
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

Figure 1. Research Program Planning.

Research programs with well-defined objectives typically lead to better test

programs with lower R&D costs, shorter schedules, and better results.
Programs with a number of poorly defined objectives typically result in
overly complicated experimental designs and laboratory equipment. On the
experimental side, the number of experiments increases substantially to cover all
boundaries of the poorly defined objectives. Experimentally, vessels and auxiliary
equipment is configured for maximum flexibility to meet all objectives, which
usually means that the system is not an optimal design for a specific objective,
leading to compromised results.

Literature Search

A thorough literature search is required, bounded by the defined objectives of

the research program. A proper literature search usually helps steer the project in
the right direction from the onset, and potentially saves time and resources. There
is considerable literature concerning nitration reactions, and it is not unusual that
some of the research objectives are simply met by finding the appropriate literature
source.
A literature search, bounded by a nitration research program, should have
three main objectives:

• to find relevant data and information concerning objectives,

• to collect relevant physical properties, thermodynamic data, and kinetic
data for the nitration system to be studied,
109
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 • to collect all relevant hazardous information, and health and safety data,
• to gather of relevant information on appropriate experimental set-ups,
• and to gather information on appropriate analytical techniques.

A typical search would include sources such as academic journals, handbooks,

patents and the internet. In general, this is a very broad set of sources, so where
should one start? A good starting point is provided with the following sources:

• Nitration and Aromatic Reactivity - J.G. Hoggett

• Aromatic Chemistry - John D. Hepworth
• Handbook of Reactive Chemical Hazards - Bretherick
• Benzene and Its Industrial Derivatives - E.G. Hancock
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

• Unit Processes in Organic Synthesis - P.H. Groggings

• Handbook of Organic Chemistry - Beilstein
• Handbook of Industrial Chemistry - Riegel’s
• A Comprehensive Treatise on Organic and Theoretical Chemistry -
Mellor
• Gmelin Handbooks
• Landolt-Börnstein
• Kirk-Othmer
• Ullmann’s Handbook
• CRC Handbooks
• Solubilities of Inorganic and Organic Compounds - Seidell & Linke
• International Critical Tables
• Journal of Physical and Chemical Reference Data
• Journal of Chemical and Engineering Data
• NIST Database
• Patent Databases

Experimental Design Plan

Once the research program objectives are determined and the literature
search is completed, it is time to define the experiments required to meet the
objectives (the experimental design stage). Experimental design deals with the
science of collecting the maximum amount of relevant data or information with a
minimum amount of time and resources (i.e., minimum number of experiments).
An experiment is a system composed of independent input variables (factors) and
dependent output variables (results).
The methodologies used to define the experimental plan typical follow one of
two approaches: the traditional (classical) experimental method or the statistical
design method.
In the classical experimental method the factors affecting the result are
identified, and then experiments are designed to investigate each relevant factor,
one at the time. The factor investigated is manipulated while all other factors are
held constant. To better understand the effect on the result, three values of a given
factor are chosen (i.e., low and high end values of the range of interest, and a
value in the middle of the range). With this approach the number of experiments
110
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 to complete the study is particularly large. For instance, if the effect of three
factors is to be tested, the minimum number of experiments to complete the
work is estimated as 33, or 27. Of course, the number doubles when repeating
experiments to assess the reproducibility of the results.
In general, running a test program according to the classical scientific method
requires thoroughness and time. An important “weakness” of the classical
experimental method is that it is not possible to single out interaction effects
among factors.
The statistical design method, or so-called factorial design method, provides
a more efficient experimental design tool. It is less time consuming than the
classical method, and more importantly it allows the study of interactions among
factors. It provides the means to predict how changing more than one factor
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

simultaneously could affect the result. A factorial design includes experimental

runs for all combinations of values for each factor (high and low). It is helpful to
visualize the settings of three variables (factors), for example, as the corners of a
cube (factorial cube). This is called a 23 design. For example, if the three factors
of interest are temperature, pressure, and time, the factorial cube representing the
experimental conditions of the research program would look as follows (Figure 2).

Figure 2. Experimental Design for Three Factors.

On the above factorial cube, effects on the result due to temperature, pressure
and time are shown along the Z, X, and Y axes respectively. Starting at the origin
of the axes, a change on the “time” factor would take us to corner “A” of the
cube. The experimental conditions covered between the origin of the axes and
corner “A” would test how “time” affects the result. Similarly, a change on the
“temperature” factor would take us from the origin of the axes to corner “B” of the
cube. A simultaneous change in the “time” and “temperature” factors would take
us to corner “C” of the cube. The experimental conditions of corner “C” would test
how the simultaneous change of “time” and “temperature” affects the result, which
111
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 would allow us to understand interactions between the “time” and “temperature”
factors on the result.
A concise paper on the details of setting up a factorial experimental plan is
provided by Murphy (8). A more detailed explanation on the subject of factorial
testing is provided by Antony. (9).
The first phase of experimental design is to choose the factors to be studied,
and the high, medium, and low target value for each factor. This is an important
step that, if performed properly, can reduce both the cost and schedule of the
program. This step is usually accomplished with a set of screening experiments.
To understand the importance of screening experiments, let us look at a test
program that investigates the effect of temperature, stoichiometric ratio of the
reactants, energy input, catalyst concentration, and system pressure on the
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

reaction kinetics of a particular nitration reaction. In a factorial design study, the

minimum number of experiments required to complete the work would be 25,
or 32. However, a few screening experiments would quickly establish that the
“system pressure” and “stoichiometric ratio of reactants” factors have little or no
effect on the reaction kinetics, simplifying the program to a 23 factorial test (i.e.,
minimum of eight experiments).
As a starting point, the typical factors significantly affecting the most generic
aspects of mixed acid nitration reactions are provided in the following Table I.

Table I. Factors affecting mixed acid nitration reactions

Reaction
Selectivity Yield
Kinetics
Temperature √ √ √
Power Input to -- --
Agitator √

Sulfuric Acid
Concentration √ √ √

Reactants --
Stoichiometry √ √

Time -- -- √
-- Does not significantly affect the variable.

Experimental Apparatus (Configuration of Equipment)

The next step in the research program is the selection and configuration of the
laboratory equipment to be used in the study. In the research of nitration reactions,
the experimental apparatus design typically revolves around the reactor.
Some idea of the reaction kinetics and heat of reaction must first be
understood. The next step is the selection of reactor type and appropriate scale
for the study.

112
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chemical kinetic studies attempt to determine the rates of pertinent reactions,
and understanding factors that affect the rates. The desired kinetic data may
already be available in the open literature. Alternatively a few screening beaker
tests can be done to gather sufficient data. If screening tests are required, a first
step is to define a parameter that can be used to infer the degree of conversion of
the reactants. Given that nitration reactions are highly exothermic, this parameter
is usually the increase in temperature.
Kinetics looks at the time / conversion relationship. It deals with the
determination of how many moles of reactants have been converted to product
in a specific time frame, and establishing rate equation. The rate expression
can then be used to determine the required reactor residence time for a specific
yield. In many reaction systems, the extent of the reaction cannot be inferred by
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

temperature, and therefore laboratory analysis is required. In addition, the extent

of the reaction may be highly affected by the thermodynamic equilibrium of the
system. Fortunately, mixed acid nitration reactions are classified as “irreversible”.
In irreversible reactions, the thermodynamic equilibrium for the reaction highly
favors formation of products, and only an extremely small quantity of the limiting
reagent remains in the system at equilibrium. If the extent of reaction is not
limited by thermodynamic equilibrium constraints, then the limiting reagent is the
one that determines the maximum possible value of the extent of reaction. Under
these conditions, which apply to mixed acid nitration reactions, temperature is
a good measure of the “extent of reaction”, as is conversion of reactant(s) to
products.
Researchers should be careful to distinguish between the problem of
determining the reaction rate function and the problem of determining the
mechanism of the reaction. The latter involves a determination of the exact series
of molecular processes involved in the reaction. It is by far the more difficult
problem. From the chemical engineer’s viewpoint, who is interested in the design
of the scale of the reactor, knowledge of the reaction mechanism is useful, but
not essential. The engineer is more concerned with the problem of determining a
reaction rate expression for use in design calculations.
Thermodynamics, on the other hand, is concerned with equilibrium systems,
systems that are undergoing no net change with time. Thermodynamics of
chemical reaction looks at the conditions at which the reactants are converted to
product at the same rate as products are converted back to reactants. On reaction
systems with strong reversible reactions, total required product throughput and
yield would, for example, affect the reactor size and configuration. Reaction
systems that are highly reversible generally require recycle loops and are
configured differently than for irreversible reactions.
Choice of the nitration vessel is next important step in the program. There
are primarily three types of reactors to consider for nitration reactions, the batch
reactor, a plug flow reactor (PFR) or a continuous stirred tank reactor (CSTR).
Other types of reactor are used for industrial nitrations. Two other types of
reactors are a forced convective tube reactor and a high speed mixing reactor
(such a nitration inside a pump). These alternative reactors will not be discussed
here.

113
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The batch reactor is the simplest configuration. Reactants are charged to
the system and rapidly mixed. Mixing is typically carried out with a mechanical
agitator. The reaction rate varies with time but it is always uniform throughout
the vessel. The reactor can be configured to perform adiabatic or isothermal
reactions. For isothermal reactions, a cooling jacket with a temperature controller
is attached to the reactor. Batch reactors are not limited by residence time and
therefore provide the highest possible conversion for most reactions, and therefore
the highest yield.
The PFR and CSTR are continuous flow reactors, feed enters the inlet of
the reactor and product leaves the exit of the reactor. However, these two types
of reactors typically behave very differently with respect to conversion and
selectivity.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

The PFR behaves like a batch reactor. It is usually visualized as a long tube
where discrete volumes of reactants enter the reactor at time = 0. Each discrete
volume mixes as it moves through the reactor, but it does not mix with other
discrete volumes upstream or downstream. Reactants in each volume ‘age’ and
react as they flow down the tube. A discrete volume that has been in the reactor for
“t” seconds would have the same composition as if it had been in a batch reactor for
“t” seconds. The composition of a batch reactor varies with time. The composition
of a discrete volume flowing through the PFR varies with time in the same way.
A CSTR is continuously feed regents while the reaction mixture continuously
leaves the reactor. The level in the reactor is typically controlled as desired. The
reactants are mechanically agitated in the reactor. Mixing is sufficiently fast so
that the entering feed quickly disperses throughout the vessel and the composition
at any point is approximately the average composition. The concentration of the
reactor outlet is the same as the reactor internal composition. A set of CSTR in
series models a PFR. In general, both the PFR and CSTR are easier to scale up.
The typical design intent of a reactor is to maximize yields and selectivity
(i.e., operating cost of an industrial process) in the smallest possible equipment
volume (i.e., capital cost of an industrial reactor). The objective of maximum
reaction in a minimum of volume is achieved in a PFR rather than a single CSTR.
In a single CSTR the concentrations of reactants in the reactor is equal to the
concentrations of reactants in the effluent of the reactor. However, in a PFR
the concentrations of reactants change throughout the length of reactor. As the
concentrations of reactants decrease, the reaction rate decreases. Therefore a
PFR requires a significantly smaller volume than a CSTR for the same extent of
reactant conversion.
In industrial mixed acid nitration reactors, a single CSTR is rarely used. If
a CSTR is the preferred reactor design, a series of CSTRs is typically employed,
which simulates a PFR. In general, a number of CSTRs connected in series are
used for isothermal nitration reactions, wherein control of temperature is easily
accomplished. PFRs are the preferred choice for adiabatic nitration reactions.
Any mixed acid nitration reaction can be operated adiabatically or
isothermally. The selection of an adiabatic or isothermal reaction is dependent
upon the reaction’s desired yield, selectivity, design, and safety constraints. In
the adiabatic process, the heat generated by the reaction is typically used to
re-concentrate the sulfuric acid from the nitration. However, in the adiabatic

114
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 process the reaction temperature is not controlled, and since the reaction
temperature has a significant effect on the isomeric distribution of some nitration
products, then it follows that it also has an effect on the process selectivity and
yield. For example, nitrobenzene is typically produced adiabatically, because
nitrobenzene has no isomers. Toluene is nitrated isothermally to control the
isomeric distribution of the three isomers of the nitrotoluene product, as well as
safety concerns.
In general, batch reactors are better suited to perform the screening
experiments for a nitration research program. The scale of the reactor is not very
important. However, the size should be minimized (e.g. <500 mL), especially
when dealing with new nitration reactions or new conditions of a known reaction.
Nitration reactions are very exothermic and can be dangerous. Additionally, some
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

nitration products are thermally unstable.

The selection of a PFR or a series of CSTRs for the experimental study is
somewhat dependent on whether the nitration will be conducted adiabatically or
isothermally. If the study involves a nitration reaction with a product that has an
isomer distribution, then a set of cooling jacketed CSTRs in series should be used
for maximum research flexibility. This would be a more expensive set up than a
PFR, but it would allow the investigation of both isothermal and adiabatic reaction
conditions if the reactors are properly designed.
The scale of the apparatus for any research work is an important consideration.
To properly scale the apparatus one must first define the relevant non-dimensional
numbers. For example, when considering the importance of work impinging the
reaction mixture through mixing, one important factor is the Reynolds number
(Re). It can be interpreted as the ratio of the parameters that contribute to
turbulent flow to parameters that contribute to laminar flow. Another relevant
non-dimensional number is the Weber number (We) which is a measure of the
relative importance of the fluid’s inertia compared to its surface tension. It is a
useful quantity in understanding the formation of droplets.
There are a number of papers and general literature on the subject of reactor
scale-up and dimensional analysis (10, 11). Although helpful, the available
literature fails to clearly indicate how small an experimental apparatus should be
and still provide predictive behavior of a full–scale industrial reactor. However,
several research programs at BC Research have shown that a reactor scale-up of
1:100 provides an acceptable experimental reactor size for mixed acid nitration
reactions.
Once the kinetic data is in hand, and the selection of the reactor type is made,
and the scale factor is defined, the exact reactor design can be addressed. We have
learned that two design issues must be carefully considered before completing the
reactor design: mixing and the design of the reactant feed system(s).
The “mixed acid” or “nitrating acid” is a mixture of nitric and sulfuric acids
with sufficiently strong acidity to generate the nitronium ion, which is the active
nitrating species. In general the organic compound to be nitrated (e.g., benzene,
toluene, etc.) has a low solubility in the mixed acid and forms a second phase
when mixed with the acid media. Because of the low organic solubility in the
mixed acid and the rapid rate of the nitration reaction, the reaction is thought
to take place in the acid phase close to the interface of the organic phase and

115
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 acid phase (12). Therefore reaction vessels are well mixed to generate a large
interfacial area between the two phases to facilitate a rapid predictable rate of
nitration. However, generating a large interfacial area is not sufficient. The
reactants must also be evenly distributed throughout the reactor. For example, in
the production of nitrobenzene, unintended locations in reactor systems containing
high concentrations of nitric acid relative to the benzene can lead to over-nitrated
product (i.e., formation of dinitrobenzene). These unintended areas are typically
due to poor mixing or to a poor design for the introduction of materials to the
reactor system. The reactants must be introduced in the best possible location
within the reactor and at the right flow rate to avoid these macro effects. Poor
mixing is also a significant safety concern. Poorly mixed locations that suddenly
contact the acid phase with a large amount of organic phase can result in a rapid
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

evolution of heat that cannot be dissipated in a timely manner. This can lead to
thermal decomposition, the consequences of which can be severe (13).
The target criteria for ‘perfect’ mixing in a batch reactor or CSTR is that t1/2 is
at least 8 times greater than tmix. The definition of t1/2 is the half-life of the organic
substrate (14). A practical way to determine tmix for a specific reactor geometry,
is to charge the vessel with a weak acid containing an indicator that changes color
when the solution is neutralized with a base. A small excess of concentrated base
is added quickly to the reactor with mixing. The mixing time, tmix, corresponds to
the time the color completely changes throughout the solution.
For a given reactor configuration, the tmix depends primarily on the density,
viscosity, and Re number of the reaction mixture, as well as the geometrical
configuration of the reactor.
Researchers at BC Research have found that for reactors with reasonable
aspect ratios, the mixing required to obtain fast nitration reactions is attained when
the power input into the reactor is in the range of 8 to 12 Watts/kg.
How fast the reactants will mix also depends on where and how the reactants
are introduced into the reactor. If the reactants are introduced too fast or are
added in a low mixing area (e.g., behind a reactor baffle plate), then localized
concentration inhomogeneities may lead to non-optimal reaction conditions. In a
typical nitration experiment, the nitric acid and sulfuric acid are pre-mixed before
the organic phase is charged. In a PFR, the organic phase should be introduced
through a manifold that creates even distribution through a cross-sectional region
at entrance to the reactor. In a batch reactor or CSTR, the organic phase should
be introduced into the suction side of the agitator and close to the high shear point
of the impeller (i.e. at the impeller tip). The rate of reactant introduction is also
important. The ratio of the mixed feed rate to the substrate feed rate is typically
25:1 to 30:1.
In a PFR or CSTR, the flow of mixed acid to the reactor can be readily
measured. However in a batch reactor the mixed acid flow which is established by
the recirculation flow within the reactor cannot be easily measured. On the other
hand, the internal flow rate can be estimated using the following correlation (9),

116
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 where:
q: volumetric flow rate, m3/s
n: rotational speed of the impeller, revolutions/s
D: diameter of impeller, m
Nq: Flow number
Nq numbers for different impeller geometries can easily be found in the
literature. Several Nq values are shown below.
Nq = 0.5 , for marine propellers (square pitch)
Nq = 0.87, for a four-blade 45°C turbine
Nq = 1.3, for a disk turbine
Nq = 0.47 for a HE-3 high-efficiency impeller
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

Lastly, when employing a PFR or CSTR reactor, one must decide whether the
sulfuric acid, once separated from the product, will be reconcentrated and fed back
to the nitration vessel. Typically, recycling the sulfuric acid more closely mimics
the contemplated commercial facility.
A study where the sulfuric acid is not recycled is substantially lower in cost
but garners considerably less information about a process that wishes to recycle
the acid industrially. The quality of the sulfuric acid can impact experimental
results. For example, in a commercial process the recirculated sulfuric acid can
contain product and by-products, which can impact diffusion of fresh substrate
into the mixed acid media. Further, traces metals can accumulate in the sulfuric
acid recycle loop and impact aspects of the chemistry.

Safety Audit

A hazards evaluation of the experimental equipment configuration should

be undertaken, along with scrutiny of the operating procedures. The same
precautions taken in an industrial nitration facility should be taken in the
design of the laboratory or pilot plant equipment used for the study. These
precautions include, but are not limited to: knowledge of detonable compositions,
knowledge of compositions subject to a run-away exothermic decomposition,
elimination of low points where NOx, organics, nitric acid, nitrous acid, and
other species can collect, heat removal management, and estimation of energy
content of reaction mixtures (compared to TNT). In some cases the energy
required to initiate a deflagration or detonation of reaction mixtures should be
determined. Management of gas evolution must be taken into account, especially
NOx management. The experimental configuration should not allow for the
nitration reaction mixture to become isolated in a confined area, such as being
double blocked in a transfer line. The mixing of the organic substrate with
mixed acid must be efficient to eliminate the possibility of local concentration
inhomogeneities in the reactors. The ratio of HNO3:H2SO4:H2O:organic must
strictly be controlled throughout the system to avoid unintended reactive mixtures.
Any friction point in the process should be evaluated in terms of the amount
of energy that can impinge on process streams (e.g., valves, pumps, agitator
foot-bearings). Materials of construction must be compatible with strong acids.

117
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Studies may use indirect materials in the process, such as oils, greases,
protective devices, etc. The compatibility of these material with the acids and
reaction mixtures must be assessed. Material safety data sheets of all reagents are
reviewed.
Emergency shut-down procedures are reviewed. Personal protective
equipment such as clothing, gloves, and respirator are chosen. Plans and
procedures are developed to address chemical spills and waste disposal.
If the chemical stability of a compound is in question, and no reliable data
is available, further work is required before proceeding with the experimental
work. Stability of nitration compounds can be assessed using differential scanning
calorimetry (DSC) and accelerated rate calorimetry (ARC).
Other safety related items to consider are vent gas composition (presence
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

of carbon monoxide, N2O), shutdown and startup modes of operation (many

accidents happen at these times for various reasons), no flow situations, reverse
flow (nitration up stream of the nitrators), presence of organic impurities that
lower the onset temperature of decomposition, failure to properly separate acid
and organic phases, clean out procedures, unintended nitration in uncooled areas
within a process, and impact of fire.

Sampling and Analysis

Part of the experimental plan should include a well-defined sampling and

analysis strategy. Mixed acid nitrations are fast reactions, making analysis of
unreacted starting material a challenge. Partition coefficients of products and
by-products between organic and aqueous acid phases add to the complexity of
sampling.
Samples taken at the exit of the experimental reactor, or from the reactor, must
always be assumed to have unconverted reactants. Such samples must be quenched
to stop the nitration so that the starting material can be quantified. Quenching can
be done by rapidly cooling the sample, or by dilution of the sample with cold water.
A cooled sample loop can also be employed.
Once samples are taken, the organic and acid phases should be quickly
separated, even if the sulfuric acid has been cooled and diluted. Any residual
reactant, especially nitric acid, will keep reacting if the phases are not separated.
Refrigeration of samples that have been prepared for analysis should be kept
refrigerated until performing the analysis.
Typically, both the acid phase and organic phase are analyzed for products and
by-products. It is common that several isomers of a particular species are present,
and each may require quantification.
Sample preparations that require extraction steps need to be validated with
known concentrations of by-products to establish the recovery for each analyte
from the compositional matrices to be analyzed.
The analysis of organic components that are present as free acids are typically
done via gas chromatography (GC). Neutralized organic compounds (e.g., sodium
nitrophenols or sodium cresols) are typically analyzed via liquid chromatography
(HPLC). Finally, to assess the completion of reaction where nitric acid is the

118
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 limiting reactant, residual nitric acid in the spent mixed acid is measured. This is
done by measuring the nitrate & nitrite in the acid phase using ion chromatography.
Analytical standards made with volatile solvents are refrigerated to reduce
evaporation, and should be prepared in the same compositional matrix as
the sample. For example analytical results are improved for dinitrobenzene
analyses in mononitrobenzene when the dinitrobenzene standards are prepared in
nitrobenzene rather than in dichloromethane.

Conclusion
The experimental methodology and recommendations presented in this paper
are based on the know-how gathered by BC Research over the course of 20 years
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

of experimental nitration work. However, the reader should understand that this
summary is by no means a comprehensive methodology to use when conducting
nitration experiments. Rather it provides a number of practical considerations for
those about to undertake a nitration research program.

References
1. (a) Mitscherlich, E. Annln. Phys. Chem. 1834, 31, 625. (b) Mitscherlich, E.
Annln. Pharm. 1834, 12, 305.
2. Hoggett, J. G.; Moodie, R. B.; Penton, J. R.; Schofield, K. Nitration and
Aromatic Reactivity; Cambridge University Press: Cambridge, 1971.
3. Sato, H.; Hirose, K; Nagai, K.; Yoshioka, H.; Nagaoka, Y. Vapor phase
nitration of benzene over solid acid catalyst II. Appl. Catal., A 1998, 175
(1-2), 201–207.
4. Sreedhar, I.; Suresh Kumar Reddy, K; Ramakrishna, M.; Kulkarni, S. J.;
Raghavan, K. V. Studies of para-selectivity and yield enhancement in zeolite
catalyzed toluene nitration. Can. J. Chem. Eng. 2008, 86 (2), 219–227.
5. Knauf, T; Racoes, A.; Dohmen, W.; Rausch, A. U.S. Patent Application
2010/0076230 A1, 2010.
6. Quadros, P. A.; Oliveira, N. M. C.; Baptista, C. M. S. G. Continuous adiabatic
industrial benzene nitration with mixed acid at a pilot plant scale. Chem. Eng.
J. 2005, 108, 1–11.
7. Quadros, P. A.; Castro, J. A. A. M.; Baptista, C. M. S. G. Nitrophenol
reduction in the benzene adiabatic nitration process. Ind. Eng. Chem. Res.
2004, 43 (15), 4438–4445.
8. Murphy, T. D. Design and analysis of industrial experiments. Chem. Eng.
1977, June 6, 168.
9. Antony, J. Design of Experiments for Engineers and Scientists; Elsevier:
Burlington, MA, 2003, pp 54−92.
10. Paul, E. L.; Atiemo-Obeng, V. A.; Kresta, S. M., Eds.; Handbook of
Industrial Mixing Science and Practice; Wiley-Interscience: Hoboken, NJ,
2004.
11. Donati, G.; Paludetto, R. Scale up of chemical reactors. Catal. Today 1997,
34, 483–533.
119
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 12. Rahaman, M.; Mandal, B.; Ghosh, P. Nitration of nitrobenzene at high
concentrations of sulfuric acid: Mass transfer and kinetic aspects. AIChE J.
2010, 56 (3), 737–748.
13. Gygax, R. Chemical reaction engineering for safety. Chem. Eng. Sci. 1988,
43 (8), 1764.
14. Nauman, E. B. Chemical Reactor Design, Optimization, and Scaleup;
McGraw Hill: New York, 2001.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch009

120
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 10

Assessment of Chemical Reactivity Hazards for

Nitration Reactions and Decomposition
of Nitro-Compounds
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

Robert W. Trebilcock*,1 and Seshu Dharmavaram2

1E.I. du Pont de Nemours & Co., Building K-37, Chambers Works,
Deepwater, New Jersey 08023
2E.I. du Pont de Nemours & Co., 1007 Market St.,

Wilmington, Delaware 19898

*E-mail: [email protected]

Nitro-compounds made from the nitration of organic molecules

are highly reactive chemicals used for producing a variety
of commercial products throughout the world. However, the
potential for over-nitration and the inadvertent or unexpected
decomposition can lead to very high chemical reactivity
hazards with dangerous consequences. Such reactions are
exothermic leading to generation of heat that can accelerate
the reaction causing a rapid temperature and pressure increase
in a vessel. A detailed chemical reactivity hazard assessment
should include the estimation and/or measurement of several
parameters such as the heat of reaction, adiabatic temperature
rise, maximum self-heat rate, exothermal onset temperature and
oxygen balance.

Introduction
Nitro-compounds made from nitration of organic molecules are chemicals
used for producing a variety of commercial products throughout the world.
Nitration reactions are also done in the laboratory to synthesize new chemical
compounds of interest. Whether done in commercial plants or in the laboratory,
nitro-compounds can be highly reactive and a hazards assessment needs to be
done to evaluate the potential for over-nitration and inadvertent or unexpected
decomposition, which can lead to very high chemical reactivity hazards with

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 dangerous consequences. Nitration and decomposition reactions are exothermic
leading to generation of heat that can accelerate the reaction causing a rapid
temperature and pressure increase in a vessel. In addition to heat released, gas
generation can also occur that leads to significant pressure increases.
A detailed chemical reactivity hazard assessment should include the
estimation and/or measurement of several parameters such as the heat of reaction,
decomposition energy, oxygen balance, adiabatic temperature rise, maximum
self-heat rate, and exothermal onset temperature. Once a hazards assessment
is completed, for a specific situation or system, then the consequences of a
catastrophic event can be analyzed to illustrate the potential impact, which in turn
can then drive good risk reduction decisions. The hazards assessment needs to
define the expected reaction for the process being evaluated considering reactants,
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

products, byproducts and impurities. It needs to consider both standard operating

conditions and unusual conditions.
Moving outside of safe operating conditions has resulted in many accidents
both in the laboratory and in commercial plants. The consequences of incidents
involving nitro-compounds or nitration can include fires, decomposition,
deflagrations, detonations and toxic gas generation.
Over the years there have been many accidents in commercial plants and
laboratories involving nitro-compounds and nitration including several in the past
few years. As an example, a deflagration occurred on May 19, 1967 in a batch
plant making dinitrobenzene at DuPont’s Chambers Works plant in Deepwater, NJ
(1). In normal operation, nitrobenzene would be slowly fed into the sulfuric/nitric
acid mixture in the agitated reactor while cooling water was used to remove the
heat generated by the nitration reaction. However in this case, it was determined
that the nitrobenzene was fed to the batch nitrator without sufficient agitation. The
agitator was restarted and the unreacted ingredients reacted suddenly resulting in a
temperature rise above 200 °C. The heat of reaction exceeded the cooling capacity
of the reactor causing thermal runaway and the material in the reactor deflagrated
resulting in a fireball of 20-40 ft in diameter. Extensive damage occurred to the
plant. Fortunately, there were no serious injuries from this accident.
This example illustrates just one of the many ways that energy can be released
when using nitro-compounds or performing nitration reactions. An accident such
as this could occur both in commercial units and in the laboratory.

Chemical Reactivity Hazards Assessment

The first step in completing a hazards assessment for existing or potential
process is to define the expected reaction and process. The evaluation needs to
consider the reactants, products, byproducts (including gases) and impurities that
could be generated by the process. Consideration needs to be made for both
planned operating conditions and for potentially unusual conditions.
For handling of nitro-chemicals or for nitration processes some initial hazard
screening tools include calculation of the heat of nitration reaction, decomposition
energy of nitro-compounds, and oxygen balance of the compounds or mixtures.
Some of this information can be found in the literature or calculated from
122
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 the expected reactions or process. Depending on the results of the screening,
additional calorimetric testing may be needed.

Heat of Nitration Reaction

Many incidents involving nitro-chemicals or nitration ultimately result from
the unexpected generation of heat and generation of gas. One source of heat can be
the heat of reaction since most nitration reactions are exothermic. High exothermic
heat of reaction is a potential measure of the chemical reactivity hazard, since heat
generation can lead to decomposition, deflagration or detonation if not removed
by the process equipment.
The heat of reaction can be determined theoretically by using known heats of
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

formation if the overall chemical reactions are known. Testing in calorimeters can
also be used to determine the heat of reaction.
Typical heats of reaction for nitration are exothermic. As an example, the
heats of reaction for nitration of toluene are listed below in Table I. The high
exothermic heats of reaction are typical for nitration reactions and contribute to
the high chemical reactivity hazards.

Table I. Heats of Reaction for Toluene Nitration

Reaction Heat of Reaction
Toluene → Nitrotoluene -185 cal/g
Nitrotoluene → Dinitrotoluene (DNT) -163 cal/g

Many nitration processes can generate gases and other sources of heat in
addition to the heat of reaction from desired products. These need to be considered
and included in the design of the process equipment. Nitration reactor design will
need to consider heat generation from the following sources:

• Heat of reaction for the main reactions

• Heat of reaction for byproduct and impurity reactions
• Heat of dilution of acids
• Heat of mixing of acids

Heat generated by side reactions, dilution of acids and mixing of acids can
be significant depending on the chemistry practiced. The actual heat generated
in the reactor can be much higher than the heat generated solely by the desired
main reaction. Failure to consider all heat generation in the reactor could result in
insufficient cooling capacity in the reactor and increase the potential for thermal
runaway.
Once the total heat generated by the reaction is known this can be used to
calculate the adiabatic temperature rise in the reactor. In many cases this adiabatic
temperature rise can be quite large and can approach or exceed temperatures in
which the materials in the reactor begin to decompose.
123
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 In the incident described at the beginning of the chapter, the heat generated in
the reactor was typically controlled by slow addition of nitrobenzene to the acid
in the cooled and agitated reactor. However in the incident, there was insufficient
agitation and the unreacted feed materials began to react at once instead of over
time when the agitator was started. This caused the temperature in the reactor to
rise much faster than the cooling available, resulting in thermal runaway in the
reactor. The temperatures in the reactor reached the point where decomposition
reactions began to occur.

Some factors to consider in designing the nitration reactors are listed below:

• What is the adiabatic temperature rise on loss of cooling?

Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

The closer the normal reactor operating conditions are to the

decomposition temperatures of the reactor mass, the greater the
possibility of having an incident due to high temperature in the reactor.
In some cases backup cooling systems may be needed.

• How will the heat of reaction be removed from the reactor?

Many nitration reactors have internal cooling coils or jackets to remove

the heat of reaction. Others are designed with external cooling loops to
remove the heat generated. If the materials are thermally stable it may be
possible to evaporate some of the reactor mass to cool the reactor.

• How will the rate of heat generation be controlled?

Normally the rate of heat generation is controlled by limiting one or

more of the ingredients being fed to the reactor. The selection of which
ingredients to limit can influence decomposition temperatures and the
potential energy released if decomposition would occur.

Sometimes the rate of reaction is limited by mass transfer between phases

or limited by the reaction rate kinetics. In these cases, changes in agitation
or temperature can have a large effect on the reaction rate.

• How do changes in reactor temperature affect the reaction kinetics?

Increasing the reactor temperature will in most cases increase the reaction
rate. The reaction rate may increase exponentially with increased
temperature, while the heat removal capacity of the reactor will only
increase linearly with temperature. This can result in thermal runaway
if the reactor temperature is allowed to rise above normal operating
conditions. Increasing plant capacity without changing the reactor could
cause the reactor to approach limits in its cooling capacity, especially
during unusual conditions.

124
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 • What conditions in the reactor can cause upsets in the reactor that could
lead to thermal runaway or thermal decomposition?

Unsteady operation can change the rate of reaction or cooling required

causing upsets in the reactor that could lead to thermal runaway or thermal
decomposition. Some factors to consider are:

Wrong type or amount of material fed to the reactor

Incorrect time, order or rate of addition of materials to the reactor
Accumulation of reactants in the reactor
Contamination in the reactor
Operation at wrong temperature
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

Unsteady agitation (none, too little, too much, variable)

Loss of cooling (plugged tubes, loss of cooling water, fouling)
Excessive heating (control failure, initial heating on startup, self
or overheating on shutdowns, insufficient heating, heating from
external sources such as tracing, friction, mechanical energy,
fire)
Incorrect amounts of catalysts, inhibitors, sulfuric acid or water
Inadvertent addition of water to the system (heat of dilution)

Nitro-Compounds’ Decomposition Energy

Many nitro-compounds will decompose with a significant release of energy if
heated to the point where decomposition begins. Understanding the temperatures,
energy released and outcomes (decomposition, deflagration, detonation) of
thermal events is an important consideration in the design of commercial plants
or laboratory testing plans.
Decomposition energy is usually exothermic for nitro-compounds and can
generate gases. The decomposition energy can be significantly higher than the heat
of reaction generated by the expected process chemistry. Decomposition energy
of pure compounds or mixtures can be estimated using various thermodynamic
modeling programs. One of the programs that can be used to develop the
reaction thermochemistry is ASTM CHETAH® (Chemical Thermodynamic
and Heat [Energy] Release Program) developed by the American Society for
Testing and Materials (ASTM). The thermodynamic calculations described in
this chapter were done using HT-65, a DuPont high temperature thermodynamics
estimation program developed in the 1960’s for the formulation of explosives.
HT-65 estimates enthalpy and decomposition products using a Gibbs free energy
minimization method.
An illustration of the potential heat released from a decomposition reaction
and also combustion of dinitrotoluene is listed below in Table II.

125
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table II. Heat Released from Reactions of DNT
Reaction Heat Released
Decomposition Energy (DNT) -900 cal/g
Heat of combustion (DNT) -4,000 cal/g

The heat generated by the expected reaction of toluene to DNT (shown in

Table I) is about half the energy that could be released if DNT were to decompose.
The decomposition energy of DNT is for pure DNT. This energy is based on the
oxygen available in the DNT molecule. If pure DNT were to decompose, this is
the amount of heat that would be expected to be released. How DNT decomposes
(decomposition, deflagration or detonation) is determined by the conditions under
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

which the decomposition occurs such as rate of heating and confinement. DNT
is not completely oxygen balanced, but if extra oxygen is available during the
decomposition then the heat released could approach the heat of combustion.
High exothermic decomposition energy is a measure of the chemical reactivity
hazard of nitrochemicals since unexpected heat sources can lead to decomposition,
deflagration or detonation. Since most nitro-compounds will release heat upon
decomposition it is important to understand at what temperature the decomposition
begins. The closer the normal operation is to the decomposition temperature the
greater the risk. Impurities or mixtures could lower the decomposition temperature
of pure materials.
It is important to understand the decomposition properties of the chemical and
mixtures early in the development process to avoid the potential for unexpected
reactions both in laboratory development tests and in the actual commercial
facility. High calculated heat of reaction or decomposition energy increases the
need for actual calorimeter testing to determine parameters such as:

• Potential for thermal decomposition

• Adiabatic temperature rise
• Quantity and rate of heat release
• Gas evolution
• Decomposition rates vs. temperature
• Initial decomposition onset temperature

Many testing methods are available to determine the decomposition

temperature of nitrochemicals or mixtures of nitro-compounds. Only two will be
discussed in this chapter.
A quick screening method for determining heat of reaction or the onset
temperature of decomposition is Differential Scanning Calorimeter (DSC). The
advantages of using a DSC to develop decomposition temperatures is that the test
can be run very quickly and needs only a very small sample. The small sample
size and speed of the test is also a disadvantage since some thermal events can be
missed by the instrument.
Figure 1 shows a diagram of a DSC. A small sample is placed in the sample
cell and the sample is heated. The calorimeter compares the temperature in the

126
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 sample cell with the temperature in the reference cell and creates a plot showing
exothermic or endothermic reactions as a function of temperature. A typical DSC
output curve is shown in Figure 2.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

Figure 1. Differential Scanning Calorimeter (DSC).

Another calorimeter is the Accelerating Rate Calorimeter (ARC). This
calorimeter uses a larger sample size (1-50 ml). The sample size is large enough
that mixtures can be evaluated, the sample can be mixed during the test and
materials can be sampled after the test is complete. The ARC is a good instrument
for developing process data and evaluation of runaway reactions.

Figure 2. Typical Differential Scanning Calorimeter (DSC) Curve.

Figure 3 shows a schematic of a typical Accelerating Rate Calorimeter. The
sample is placed in the spherical bomb and the sample is heated in a heat-wait-
search mode to determine exothermic behavior in the calorimeter. Differences in
the heat input and the observed heat output are recorded. Pressure buildup in the
127
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 bomb is also recorded. The sample in the bomb can be agitated. One disadvantage
of the ARC testing is longer run times. Also the sample size is still fairly small
and can limit the accuracy of some of the measurements.

The ARC test can be used to develop the following information:

• Heat of reaction
• Onset temperature for decomposition reactions
• Adiabatic temperature rise
• Rate of change in pressure and temperature vs. time
• Self-accelerating decomposition temperature
• Detailed kinetics
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

• Post test sampling of gas evolution or reaction products is possible

An example of an ARC run is shown in Figure 4. The sample run in the ARC
test shown in Figure 4 is a mixture of two phases. The organic phase consists
primarily of nitrobenzene (80%) and dinitrobenzene (20%). The organic phase
also contains some reaction byproducts such as picric acid (0.5%). The second
phase is an acid phase consisting of 78% sulfuric acid. Excess nitric acid was
added to the acid phase to simulate an over-nitration condition. The acid phase
also contains some reaction byproducts such as nitrosylsulfuric acid. The two
phases were mixed in the ARC during the testing.

Figure 3. Accelerating Rate Calorimeter (ARC).

128
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The ARC test results shown in Figure 4 illustrates the advantages and
importance of performing calorimetric hazards screening of actual mixtures for
the collection of decomposition data. Literature values or DCS/ARC testing of
pure Nitrobenzene or Dinitrobenzene would have suggested a decomposition
temperature of around 300°C. However the ARC test in Figure 4 shows that the
under normal reaction conditions, with the presence of impurities in the organic
and acid phases, the actual point where decomposition reactions begin can be
considerably lower (130 °C). These initial self-heating rates are very low but
the rate of decomposition rises very quickly with increasing temperature and
significant decomposition could occur below 300°C.
The ARC test shows that actual decomposition safety margins could be
significantly lower than what would be predicted based only on testing of
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

pure compounds or literature values. Impurities can lower the decomposition

temperature and increase the rate of temperature rise. Some impurities to consider
are:

• Nitric Acid, oxidizers and other compounds containing oxygen

• Byproducts such as more highly nitrated impurities
• Impurities such as nitrophenols

Figure 4. ARC test nitrobenzene, dinitrobenzene and sulfuric acid with

impurities. (see color insert)

The ARC test provides information on the onset of self-heating. At the onset
of self-heating, the temperature rise may be small, much less than one °C per
minute. There may be conditions where reactive materials are held above the onset
of self-heating for long periods of time. The slow self-heating could be significant
129
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 if the heat generated by the self-heating is not allowed to be removed. Some of
these conditions could occur while the plant is not running. Some examples where
slow self-heating could occur and result in thermal runaway over time are listed
below:

• Material trapped under insulation can begin to self-heat and the insulation
can prevent loss of the heat causing thermal runaway
• Materials can be held under heat in process equipment or piping for long
periods of time during shutdowns causing thermal runaway
• Oxidation reactions (exposure to air or other oxidizers) can create
decomposition by-products that may have a lower decomposition
temperature than normal products
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

• Self-heating could increase the reaction rates of side reactions that

result in formation of impurities with lower decomposition temperature.
In some cases the decomposition products can autocatalyze the
decomposition reaction (2).

Any process upset that results in heat generation that exceeds the rate of
cooling can lead to decomposition. Some examples include:

• Unsteady operation
• Wrong ratio of ingredients
• Starting of agitation
• Unexpected leaks into or out of the process
• Backflow of reactor mass or other organics into nitric acid feed lines

The plant design needs to consider other sources of heat that could heat the
materials to the point of decomposition. Some potential sources of heat that could
lead to decomposition reactions are listed below:

• Heat buildup in deadheaded pumps or centrifuges. Decomposition

reactions (decomposition, deflagration or detonation) have occurred
when material has been deadheaded in pumps and centrifuges long
enough to reach the decomposition temperature. Temperature rise in
these conditions can be quite rapid.
• Friction caused by rotating equipment, seals or other mechanical action.
There have been cases where material has accumulated in packing of
agitator seals and decomposition reactions started when the packing was
tightened. Pump seals can also be a source of heat.
• External heat sources such as failure of temperature control in heat
exchangers or tracing.
• Impact can be considered a source of heating that can cause
decomposition reactions to occur. Need to consider impact due to metal
to metal contact, tramp (foreign) material or mechanical failure. Sand
or grit can increase impact sensitivity. Impurities (byproducts, nitric
acid, and oxygen containing compounds) can increase impact sensitivity.

130
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Some impurities such as picric and styphnic acid salts are very impact or
friction sensitive if isolated and allowed to dry.
• Adiabatic compression can generate very high temperatures if gases
are compressed due to pressure changes in the process. The heat
from the compressed gas can transfer to reactive materials and begin
decomposition reactions.
• Electrical shock, static electricity or electrical arcs could provide
sufficient heat to start decomposition reactions.

Oxygen Balance
Another important calculation when working with nitro-compounds is
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

calculation of the oxygen balance (OB) of the reactants, products and mixtures
formed in the process equipment.
The oxygen balance of a compound or mixture is a calculation used to indicate
if the compound/mixture is deficient in oxygen per the stoichiometric equation
below:

For calculating the oxygen balance in % oxygen for non-metallic compounds

the following formula can be used where MW is the molecular weight of the
material. If the molecule contains N, Cl or S atoms, they are ignored in the
calculation.

The oxygen balance of mixtures can be determined by taking the oxygen

balance of each component in the mixture and multiplying by the compounds
weight %. For example, see Table III below, the oxygen balance for nitrobenzene
is -162.6 and the oxygen balance for nitric acid is +63.5 using the above equation.
A mixture of 70 weight % NB and 30 weight % nitric acid would have an oxygen
balance of -94.8.

Table III. Example Oxygen Balance Calculation

Ingredient Weight % OB of Ingredient Oxygen Balance
Nitrobenzene 70% -162.6 -113.8
Nitric Acid 30% + 63.5 + 19.0
Oxygen balance of Mixture -94.8

A negative number indicates that the compound or mixture is deficient

in oxygen while a positive number indicates a compound/mixture with excess
oxygen. Mixtures with an oxygen balance of -120 to + 80 indicate materials
that are well oxygen balanced and may have higher potential chemical reactivity
hazards.
131
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Most commercial explosives are formulated to have oxygen balances close
to zero. The most energetic and sensitive explosives are oxygen balanced on a
molecular level. See Table IV for the oxygen balance for several high explosives.

Table IV. Oxygen Balance for Several High Explosives

High Explosive Oxygen Balance
Pentaerythritol tetranitrate (PETN) -10
Nitroglycerine (NG) +4
Picric Acid +45
Trinitrotoluene (TNT) -74
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

When oxygen balanced on a molecular level, all the oxygen needed to balance
the decomposition reaction is present without the need for mixing or material
transfer to supply additional oxygen. If the material begins to react due to the input
of heat (temperature, impact, friction) the decomposition or detonation energy can
be released very quickly.
Mixtures of materials containing oxygen deficient materials (fuels) and
oxygen rich materials (oxidizers) can result in oxygen balanced mixtures that are
more hazardous than the individual ingredients.
A common commercial explosive is an ammonium nitrate/fuel oil mixture
(ANFO) which is a mixture of 94% ammonium nitrate and 6% fuel oil.
Ammonium nitrate is positively oxygen balanced at +20 and will release energy
upon decomposition of 382 cal/g. By adding 6% fuel oil (OB = -343) to the
ammonium nitrate, the oxygen balance of the ANFO mixture is -2. By creating
an oxygen balanced mixture the energy released by decomposition or detonation
is greatly increased to 841 cal/g. In mixtures, the fuel and the oxidizers are not
chemically bonded but they are in sufficient proximity that once decomposition
occurs, the oxygen is available to react with the fuels, increasing the energy
released.
Table V compares the oxygen balance and energy release (detonation with
expansion) calculated using HT-65 for various forms of nitrated benzene and
compares the results to TNT.
Starting with benzene you have a material that acts as a fuel with a
very negative oxygen balance. In the pure form and in the condensed phase,
benzene would not be expected to be very reactive. If benzene is allowed to
vaporize and mix with oxygen it will release significant energy if exposed to
heat (ignition) sources. As more nitro groups are added forming nitrobenzene
(NB), dinitrobenzene (DNB) and trinitrobenzene (TNB) the materials become
more oxygen balanced and the amount of energy that could be released upon
decomposition increases.
As the oxygen balance approaches zero the required oxygen is available within
the condensed phase. Unlike benzene, which needed oxygen from the air to be
reactive, NB, DNB or TNB already has some oxygen present. If these materials are

132
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 heated to temperatures sufficient to begin decomposition, then significant energy
can be released without the need for additional oxygen from external sources.
Oxygen balance calculations need to be combined with energy calculations
and calorimeter testing to get a more complete understanding of the potential
energy released on decomposition. As an example, water has an oxygen balance
of zero so adding water to a mixture will bring the mixture oxygen balance closer
to zero. However, water acts as an inert and the addition of water will reduce the
total energy released on decomposition, by virtue of the fact that it can absorb heat
upon vaporization.

Table V. HT-65 Heat Released on Detonation with Expansion

Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

Compound Oxygen Equilibrium “Q” “Q” as % of

Balance Temperature cal/g QTNT
°K
2,4,6 Trinitrotoluene (TNT) -74 3536 937 100
Benzene -308
Nitrobenzene (NB) -163 2413 761 81
Dinitrobenzene (DNB) -95 3307 919 98
Trinitrobenzene (TNB) -56 4284 1047 112

Figure 5 below shows a HT-65 calculation done at a constant density of 0.5

g/cc for various mixtures of DNB and water. It shows that as water is added to
the DNB, the oxygen balance approaches zero but the total energy released also
falls. The drop off isn’t linear since HT-65 is predicting different equilibrium
products from the different mixtures. DNB/water mixtures are not single phase,
so the benefit of water addition may not be seen unless the mixture is kept agitated.
While the addition of water is usually a good way to reduce process safety
risk there are a few cases where excess water can cause some issues. Nitration is
very sensitive to acid strengths and excess water could stop the nitration and result
in accumulation of unreacted nitric acid. The unreacted nitric acid could react
quickly if the water is brought back to normal levels. Unreacted nitric acid or
organics could cause delayed reactions or more vigorous reaction in other vessels.
Unreacted organics could cause flashpoint issues if the unreacted organics are more
flammable than the final products. Water added to strong acid could generate
sufficient heat to start decomposition.
Excess water can also change the characteristics of the nitric acid ions. In
most nitration reactions, the reaction is run under conditions where the nitronium
ion (NO2+) is the reactive species. Depending on the nitric acid or nitric/sulfuric
acid mixture used for nitration, the concentration of the nitronium ion can change,
significantly changing the nitration kinetics. In some cases the nitronium ion can
drop to very low levels and the nitrate ion (NO3-) can begin to become more
important. The chemistry done by the nitrate ion may be very different than the
chemistry done by the nitronium ion, creating different byproducts which may

133
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 change the heat or reaction, thermal stability, decomposition temperature, gas
generation and other significant process parameters. The nitrate ion may cause
oxidation reactions in place of the expected nitration reactions. A good discussion
of ionic species in nitric acid and nitric/sulfuric acid mixtures can be found in the
literature (3).
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

Figure 5. HT-65 calculation for a mixture of DNB and water. (see color insert)

Sulfuric acid is another ingredient that is common in nitration mixtures.

Based on HT-65 calculations done at a constant density of 0.5 g/cc shown in
Figure 6, the sulfuric acid under equilibrium conditions may supply some oxygen
to oxygen deficient organics. Figure 6 shows the predicted oxygen balance and
energy released for decompositions of DNB and sulfuric acid mixtures. The
equilibrium products suggest that some of the oxygen in the sulfuric acid is
available to react with the organics increasing the energy released. The calculated
decomposition energy for the DNB/Sulfuric acid mixtures never exceeded the
decomposition energy for pure DNB. Sulfuric acid could be a source of oxygen,
but it doesn’t seem to increase the decomposition energy. Changes in sulfuric
acid strength could affect chemical reactivity, since the sulfuric acid strength
has a significant impact on reaction rates and the nitronium and nitrate ion
concentrations.
Table VI shows the results of some ARC testing for mixtures of DNB and
sulfuric acid (90% strength). The DNB and sulfuric acid used in these ARC tests
contained typical impurities found in nitration mixtures such as nitrophenols, nitric
acid and nitrosylsulfuric acid. ARC test conditions will influence the calculated
134
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 heat of decomposition “Q”. Small sample size may be needed to protect the ARC
equipment but will increase the error in the calculated “Q”. Low ARC cutoff
temperatures may be needed to protect the ARC equipment but may result in lower
calculated “Q” due to the ARC missing part of the decomposition reaction. High
self heat rates can exceed the ARC cooling capacity resulting in lower calculated
“Q”.
The ARC testing summarized in Table VI below is in general agreement with
the HT-65 heat of decomposition calculations shown in Figure 6 for DNB and
sulfuric acid mixtures. The ARC testing shows that significant energy can be
released from the decompositions of mixtures containing only small amounts of
reactive materials like DNB. However, the “Q” values shown in the table cannot
be directly used because of the limitations with the ARC testing discussed above.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

Figure 6. HT-65 calculation for a mixture of DNB and sulfuric acid. (see color
insert)

Nitric acid is another ingredient that could become part of nitration mixtures.
Since many nitro-compounds are negatively oxygen balanced and nitric acid is
positively oxygen balanced (OB= +63) mixtures containing nitric acid will affect
the oxygen and energy balances. Figure 7 below shows the HT-65 calculations
done at a constant density of 0.5 g/cc for mixtures of NB and nitric acid. Adding
extra nitric acid to NB greatly increases the energy released as the oxygen balance
approaches zero.

135
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table VI. ARC Test Results for DNB and Sulfuric Acid (H2SO4) Mixtures
Test Mixture “Q” Sample ARC Cutoff Maximum Self
Composition cal/g Size g Temperature °C Heat Rate °C/min
1 185 4.1 >351 739
20% DNB &
2 80% H2SO4 255 3.0 >393 1000
3 399 0.87 >300
4 305 0.82 307 302
60% DNB &
5 40% H2SO4 385 0.80 350 601
6 100% DNB 528 1.0 >430 >830
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

Figure 7. HT-65 calculation for a mixture of NB and nitric acid. (see color insert)

The detonability of nitrobenzene, nitric acid and water systems has been
described in the literature (4). In one inch diameter steel pipe, pure NB or pure
nitric acid was not detonable under the conditions studied. However mixtures of
NB containing from about 20% to 90% nitric acid were found to be detonable.
Excess nitric acid is also likely to increase impact and friction sensitivity,
lower the decomposition temperature, increase the likelihood of transition from
deflagration to detonation and reduce the critical diameter at which detonations
can propagate.
Some nitration reactions are done using a large excess of strong nitric acid in
the reactor. Reactor mixtures containing large amounts of nitric acid compared to
the amount of organics present and also containing very little water are likely to
be well oxygen balanced and may be quite reactive.

136
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 When designing processes using nitro-compounds, it is important to consider
that abnormal or unsteady feed rates can lead to more oxygen balanced and reactive
mixtures. Excess nitric acid could result in overnitration, oxidation reactions,
vapor phase nitration, side reactions and acid/base reactions.
Plant design needs to consider places where organics and oxidizers could
accumulate. One area to check is in the vent systems. Oxidizers in the vent system
could include nitric acid mists, nitrated organics, NOx or ammonium nitrate which
if allowed to come in contact with organics or CO could:

• React releasing heat that could cause decomposition

• Oxidize creating new compounds with lower decomposition
temperatures
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

• Form oxygen balanced mixtures that could react if heated by external

sources such as fire, steam or electric tracing
• Form flammable mixtures

Thermal stability of nitric acid when mixed with other process materials needs
to be considered. This includes not only process ingredients, but also materials
such as lubricants, gaskets and other materials used outside the process that could
accidentally come in contact with nitric acid.
Process or organic material back flowing into nitric acid feed lines could result
in oxygen balanced mixtures and possible reaction without cooling leading to
decomposition reactions in the piping.
Impurities or byproducts generated in the process often need to be removed
from the final products and in the process can be concentrated. Nitration often
generates nitrophenols such as picric and styphnic acids that are very oxygen
balanced and are also very impact sensitive when dry. Care must be taken
when isolating these compounds. Process leaks containing these compounds can
accumulate on equipment or insulation where the water could evaporate leaving
behind a impact sensitive residue.
Operations with processes and systems with higher decomposition energies
and those that are more oxygen balanced should consider additional testing to
determine safe operating conditions. Some of this testing could include:

• Critical Diameter: Consider running tests to determine the critical

diameter of products or mixtures to determine minimum pipe sizes
and confinement needed to sustain propagation of a detonation through
piping systems.
• Impact Sensitivity: Consider adding contaminants such as byproducts
and impurities. Also consider addition of materials such as sand or grit
that may increase the impact sensitivity.
• Thermal Stability: Consider testing products, byproducts and impurities
for thermal stability. Also may want to evaluate reaction mixtures for
decomposition onset temperatures.
• Card Gap Testing: Card gap testing is a way of measuring the sensitivity
of materials to propagation of detonation. The more distance between the
initiator and the sample the more sensitive the material is to detonation.

137
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 • Static Shock Testing: Some mixtures can be initiated by static
discharges.
• Vented Bomb Testing: Significant amounts of gases will be generated
during decomposition reactions. Depending on the rate of gas generation
it may be possible to vent some of this gas and minimize damage to
equipment. In some cases, venting can minimize the potential for
transition from deflagration to detonation. A typical vented bomb test is
the Koenan test, used to classify explosives for shipping.
• Cap Sensitivity: Mixtures or compounds can be tested for sensitivity to
blasting caps. Material detonable by blasting caps is very reactive.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

Summary
When evaluating existing process or potential new laboratory programs
involving nitro-compounds or nitration a chemical reactivity assessment should
be undertaken early in the endeavor and clearly documented.
Heat of reaction data for the expected and unexpected reactions is important
to define the cooling systems needed to remove the heat of reaction and to maintain
a safe operating distance away from temperatures that begin decomposition
reactions or thermal runaway.
Many nitration processes can generate gases and other sources of heat in
addition to the heat of nitration reaction in making the desired products. These
need to be considered and included in the design of the process equipment. For
nitration reactor and design of all other connected process equipment one needs to
consider:

• Heat of reaction for the main reactions

• Heat of reaction for byproduct and impurity reactions
• Heat of dilution of acids
• Heat of mixing of acids
• Heat of decomposition for nitro-compounds

The total heat generated in the reactor can be significantly higher than the
heat of reaction generated by the desired products. Heats of decomposition of
nitro-compounds can be an order of magnitude or more greater than the heats of
the nitration production reactions.
Knowledge of the decomposition temperature of reactants, products,
byproducts and impurities is important when determining the safe operating
temperatures of the process under scrutiny. Thermodynamic programs can be
used to predict the energy released if materials or mixtures begin to decompose.
Heat release during decomposition for nitro-compounds can be significant and
can result in deflagrations or detonations. Estimates can be made of the adiabatic
temperature rise and the maximum temperature. Calorimetric testing is likely to
be needed to determine key reaction parameters such as the onset temperature
and the rate of temperature rise for the decomposition reactions. Actual reactor
operating conditions may significantly lower the decomposition onset temperature
138
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 of pure materials since impurities, byproducts or reactants such as nitric acid may
be present in the reaction mixture.
The oxygen balance of the compounds and mixtures can be calculated and
mixtures with oxygen balances closer to zero are more likely to be reactive.
Oxygen balanced mixtures may have higher decomposition energy, have increased
impact and friction sensitivity, have lower decomposition onset temperature,
have increased potential for transition from deflagration to detonation and have
reduced critical diameter at which detonations can propagate.
A thorough chemical reactivity hazards assessment is necessary for safe
design and operation of all parts of a process where nitro-compounds are handled
and where nitration reactions are conducted.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch010

Acknowledgments

We would like to thank James E. Johnston a DuPont retiree who previously

worked at the DuPont Engineering - Explosion Hazards Laboratory for doing the
HT-65 calculations and conducting the ARC experiments used in this chapter.

References
1. Fritz, E. J. Anatomy of a nitration explosion. Loss Prev. 1969, 3, 41–45.
2. Bou-Diab, L.; Fierz, H. Autocatalytic decomposition reactions, hazards and
detection. J. Hazard. Mater. 2002, 93, 137–146.
3. Albright, L. F.; Sood, M. K.; Eckert, R. E Modeling Nitronium Ion
Concentrations in HNO3−H2SO4−H2O Mixtures. In Nitration: Recent
Laboratory and Industrial Developments; Albright, L. F., Carr, R. V. C.,
Schmitt, R. J., Eds.; ACS Symposium Series 623; American Chemical
Society: Washington, DC, 1996; pp 201−213.
4. Mason, C. M. Detonability of the system nitrobenzene, nitric acid, and water.
J. Chem. Eng. Data 1965, 10 (2), 173–175.

139
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 11

Safety Instrumented Systems for Process Safety

Subodh Medhekar*

Exponent, Inc., 320 Goddard Way, Suite 200, Irvine, California 92618
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

*E-mail: [email protected]

This paper presents a review of Safety Instrumented Systems

based approaches to prevent or mitigate hazardous events by
preemptively taking the process to a safe state. Various methods
to determine the appropriate Safety Integrity Levels for such
system applications are discussed along with a simplified
Nitration industry example.

Overview
Safety instrumented systems (SIS) are becoming increasingly popular for
safeguarding process plants in the oil/gas, petrochemical, and other process
industries. A SIS is a system specifically designed to prevent or mitigate
hazardous events by deliberately and preemptively taking a process to a safe state
when predetermined process conditions are violated. Safety Interlock Systems,
Emergency Shutdown Systems (ESD), and Safety Shutdown Systems (SSD) are
common names for such SIS. While the use and application of SIS has grown
substantially in the oil/gas, petrochemical, and other process industries over the
last decade, for many end users, SIS is still a new concept. SIS systems are now
being widely used in nitration plants around the world.
Figures 1 and 2 outline the basic concepts of a SIS. Take the case of a catalyst
bed reactor (simplified schematic as shown in Figure 1) whose temperature needs
to be controlled to be in a safe state. Normally a basic process control system
(BPCS) will monitor the temperature(s) within the reactor and through a variety of
ways (including reactant inlet flow control) will attempt to control the temperature
within the safe normal temperature bounds (see Figure 2). If for some reason the
reactor temperature cannot be controlled (say reactant inlet control valve fails
open or sticks open), an alarm may be sounded to alert an operator to take action.
However, if the process conditions are such that it is not possible for the operator
to respond quickly (or reliably) to avert the subsequent dire conditions (say

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 runaway reaction leading to overpressure, reactor damage and fire/explosion),
it may be necessary to implement an automated Safety Instrumented System to
take an automatic deliberate and pre-emptive action (upon detection of High
Temperature) and bring the system to a safe state. In case of this example, upon
detection of high temperatures, an automatic Programmable Logic Controller
(PLC) can take an action of independently closing the reactant feed stream to the
reactor. To be effective this SIS has to be reliable, specific and be independent of
the basic process control system.
SIS can be configured in many ways to meet a variety of process goals and
performance targets. Each SIS has one or more Safety Instrumented Functions
(SIF) with each SIF loop having a combination of logic solver(s), sensor(s), and
final element(s).
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Figure 1. Basic Concepts of SIS and BPCS.

A simple SIS system is shown in Figure 3 and comprises of one sensor, one
PLC, and one final element. In this case, upon detection of the process variable
being out of bounds the PLC will close the valve (the final element) and bring the
system to a safe state.
Figure 4 depicts a more complex system of two sensors (with 2oo2, 2 out of
2 voting logic) and two final control elements. In this configuration, both sensors
will have to detect the abnormal process condition and then the PLC will close both
of the valves. This configuration is a more complex (and also more expensive)
SIS than the one shown in Figure 2 but is also more reliable / effective in reducing
the inherent risks and also reducing the likelihood of spurious trips that can be
a nuisance to the process operations. A complex and more expensive, which is
highly reliable, system may be justified if the dire outcomes are very serious and
other means to reduce these risks are not available or are not very reliable.
142
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Figure 2. Event Scenario Progression Explaining the SIS Concept. (see color
insert)

Figure 3. SIF Loop Example (with Single Sensor, Logic Solver, and Single Final
Element).

Usually the focus of SIS is to prevent or mitigate dangerous failure

consequences. However, in addition to the consideration of dangerous failures,
the potential for spurious failures should also be considered when selecting a SIS
configuration as a spurious trip can disrupt production and result in significant
lost or deferred production costs.
143
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Figure 4. Another Example of SIF Loop (with Two Sensors, Logic Solver and
Two Final Elements).
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

SIS and Functional Safety Background

Functional safety, as defined by IEC standard 61508, is the safety that control
systems provide to an overall process or plant. Increasing use of programmable/
electronic/electric systems to carry out safety functions has raised the desire to
design safety systems in such a way as to prevent dangerous failures. The concept
of functional safety addresses the growing need for improved confidence in safety
systems. Industry experts formalized an approach for reducing risk in the process
plant environment through the development of standards IEC 61508, IEC 61511,
and ANSI/ISA 84.01.
Functional safety is a term used to describe the safety system that is dependent
on the correct functioning of the logic solver, sensors, and final elements to
achieve a desired risk reduction level. Functional safety is achieved when every
safety function is successfully carried out and the unprotected process risk
is reduced to the desired level. A safety instrumented system is designed to
prevent or mitigate hazardous events by taking a process to a safe state when
predetermined conditions are violated. Other common terms for SISs are safety
interlock systems, Emergency Shutdown Systems (ESD), and Safety Shutdown
Systems (SSD). Each SIS has one or more Safety Instrumented Functions (SIF).
Typical critical functional safety lifecycle elements include: Hazard / Risk
Assessment, Safety Integrity Level or SIL Determination, SIL Verification and
Maintenance. SIL1 or SIL2 or SIL3 relate to the levels of Safety Integrity that is
determined by the SIL classification analysis. SILa designation means no special
SIL requirement. A SIL2 level is more robust and will have a lower Probability of
Failure on Demand (PFD) than a SIL1 system. Table 1 defines PFD requirement
to meet the various SILs. The Risk Reduction Factor is the inverse of the PFD.
Thus, a SIL1 (with a PFD range from 0.1 to 0.01) implies a RRF requirement of
10-100 whereas a SIL3 would imply a RRF requirement of 1000 to 10000.
144
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table 1. Relation Between Safety Integrity Levels (SIL) and Required
Probability of Failure on Demand (PFD)
Safety Integrity Required PFD Risk Reduction Factor
Level (SIL) (RRF), 1/PFD
a no requirements (>10e-1 ) < 10
1 ≥10e-2 to <10e-1 10 – 100
2 ≥10e-3 to <10e-2 100 – 1,000
3 ≥10e-4 to <10e-3 1,000 – 10,000
4 ≥10e-5 to <10e-4 10,000 – 100,000
X Intolerable Intolerable
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Hazard/Risk Assessment

Hazard / Risk Assessments are usually performed prior to determination of the

SIL levels for the associated SIS. In the process industry, a HAZOP (Hazards and
Operability) study is performed to identify potential hazardous scenarios (and their
associated consequences and likelihoods). A FMEA (Failure Modes and Effects
Analysis), PHA (Process Hazards Analysis) or a HAZID (Hazard Identification
Study) are other hazard assessment studies that are also commonly performed.
Before determining the SIL of the SIS, the first step is to conduct a hazard
assessment to determine the functional safety needs and identify the tolerable
risk level. All mitigation and prevention safety layers (specific to the hazard
scenario) must be identified (see Figure 5). After all of the risk reduction and
mitigation impacts from the Basic Process Control System (BPCS) and other
layers of protection are taken into account, a user must compare the residual
risk against their risk tolerance (See Figure 6). If there is still an unacceptably
high level of risk, a Risk Reduction Factor (RRF) is determined and a SIS / SIL
requirement is calculated. Figure 7 depicts how the unprotected process risk
can be incrementally lowered using combinations of non-SIS and SIS layers of
protection. If just the non-SIS layers are able to reduce the risk to acceptable
levels then a SIS layer may not be needed. However, if despite all non-SIS layers
based protection, the risk still remains unacceptably high, then a SIS layer is
required. The degree to which this required SIS layer needs to provide additional
mitigation protection determines the integrity level of this SIS layer. This required
integrity level is assigned during a SIL Determination study. It should be noted
that the risk tolerance is subjective and site-specific. Each facility/ organization
must determine its own acceptable level of risk to personnel and capital assets
based on a variety of factors.

145
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Figure 5. Various Independent Layers of Protection. (see color insert)

Figure 6. Concepts of Process Risk, SIS and Residual Risk.

146
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Figure 7. Concepts of Risk Reduction with Various SIS and non-SIL Layers of
Protection. (see color insert)

Safety Integrity Level Determination

The classification of Safety Integrity Level (SIL) for a Safety Instrumented
Function (SIF) is determined by studying the consequences of failure on demand
against the following criteria (impact on reputation is not a criterion for SIL
determination):

• Impact of incident on Personnel

• Impact of incident on Environment
• Impact of incident on Asset

The following sections outline two SIL classification/determination

approaches: Risk Graph Method and Layer of Protection (LOPA) Method. The
risk graph approach is a conservative but a quick method, whereas the LOPA
approach tends to be more accurate but requires additional effort. It is possible to
use a combination of both methods when evaluating a very large number of SIF
loops. In this combination approach, a risk graph approach may be used to classify
all loops first and then the LOPA method may be used to re-evaluate/reclassify
the SIL 3 (and SIL 2) loops.
The risk graph and LOPA procedures are briefly described below.

Risk Graph Method

The risk graph method is a quick way to determine SIL requirements and is not
a detailed method of assessment. Typical SIL requirements for personnel, asset,
and environment using risk graph are shown in Tables 2, 3, and 4 respectively.
The risk graph method is intentionally conservative, and tends to over-estimate
the required risk reduction.
147
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Health and safety consequence ratings, loss consequence ratings, and
environmental consequence ratings are rated by the SIL determination team
for the various SIFs during the SIL determination study. Each hazard event is
ranked through a team consensus judgment by deciding the severity of the event
consequences based on its impact to personnel, assets, and environment.

Severity Criteria Based on Avoidance and Exposure

Table 5 shows the mitigation credits assigned to the severity category based on
the exposure (to the worker) and the possibility to avert danger. In order to evaluate
risk to personnel, a further consideration is given to exposure and probability of
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

avoiding the danger. Depending on these two factors, the severity to personnel as
originally evaluated (as per Table 2) can be reduced.
Occupancy at a location (and consequently exposure) is based on time for
which personnel are present in the hazardous area. The possibility of avoidance
of danger is based on following factors:

1. Operation of a process (supervised or unsupervised)

2. Speed of development of the hazardous event (quick, slow). For example,
an explosion is a sudden event
3. Ease of recognition of danger (seen immediately, detected by
instrumentation, deduced by operators)
4. Avoidance of hazardous events (escape routes available, engineered
protection)
5. Safety experience with similar situations

Probability of avoiding the hazardous event if protection system fails to

operate is finite (non-zero) if:

• Systems are provided to alert the operator that the SIS has failed
• The time between operator being alerted and a hazardous event occurring
exceeds one hour
• Independent systems are provided to avoid the hazard or the systems are
such that they provide enough time for personnel to escape the hazard

Demand Rate

The SIL determination team estimates the demand rate for the scenario(s)
being considered using the information presented in Tables 2, 3, and 4. Demand
rate is the likelihood of the SIS to be called upon to take protective action. The
demand rate is expressed in terms of Mean time between Demands (years). So
if a SIL loop is expected to be invoked to provide protection once in 3 years, the
Mean time between demands is 3 years and we would use the 1-10 year demand
rate category. Where the team is unable to accurately project a demand rate, a
conservative estimate is used.

148
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Process Safety Time

During the SIL classification, the process safety time for each SIF is also
determined. The process safety time is a function of the process dynamics and
is defined as the period of time that the process can be operated without protection
and with a demand present without entering a dangerous condition. The process
safety time determines the trip setting and the combined dynamic performance and
accuracy requirements for the SIF components, e.g. process measurement delay,
time between input state change and output state change in the SIF, and valve
stroke time.

Example: A trip is set at process pressure of 150 kPa. The consequences

Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

of overpressure will not be realized until the pressure reaches 170 kPa. Given
the dynamics of the process and the causes of overpressure, this will take at least
2-minutes. The process safety time therefore is 2-minutes.

Based on the demand rate and consequence severity, a target SIL is assigned.

Table 2. Risk Graph for Personnel

Mean time
between
Health / Safety (Requisite SILs)
demands
(Years)
<1 - SIL 1 SIL 2 SIL 3 SIL 4 X
1-10 - SIL 1 SIL 2 SIL 3 SIL 4 SIL 4
10-100 - SIL a SIL 1 SIL 2 SIL 3 SIL 3
100-1,000 - SIL a SIL a SIL 1 SIL 2 SIL 2
>1,000 - - SIL a SIL a SIL 1 SIL 1
Personnel No Slight Minor Major Permanent >3
Safety injury or injury or injury or injury or disability fatalities
health health health health or up to 3
effect effect effect effect fatalities
Severity
0 1 2 3 4 5
(Numeric)

149
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table 3. Risk Graph for Production / Assets (Economic)
Mean time
between
Production Losses / Asset Damage
demands
(Years)
<1 SIL a SIL a SIL a SIL 1 SIL 2 SIL 3
1-10 - SIL a SIL a SIL 1 SIL 2 SIL 3
10-100 - - SIL a SIL a SIL 1 SIL 2
100-1,000 - - - SIL a SIL a SIL 1
>1,000 - - - - SIL a SIL a
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Personnel 10-100K
<1K$ 1-10K $ 100K-1M $ 1-10M$ >10M$
Safety $
Severity
0 1 2 3 4 5
(Numeric)

Table 4. Risk Graph for Environment

Mean time
between
Environmental Impact
demands
(Years)
<1 SIL a SIL a SIL 1 SIL 2 SIL 3 X
1-10 - SIL a SIL 1 SIL 2 SIL 3 SIL 4
10-100 - SIL a SIL a SIL 1 SIL 2 SIL 3
100-1,000 - - SIL a SIL a SIL 1 SIL 2
>1,000 - - - SIL a SIL a SIL 1
Personnel Slight Minor Moderate Major Massive
Safety No effect effect Effect Effect Effect Effect
Severity
0 1 2 3 4 5
(Numeric)

150
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table 5. Mitigation Credits to Reduce Personnel Severity
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

In an another version of the Risk Graph approach, instead of using Tables 2-5
listed above, use is made of Risk Graph Figures such as the one shown in Figure
8. Using this approach the analyst simply selects consequence category (CA, CB,
CC, or CD), probability that the exposed area is occupied or not (Fa or Fb), the
probability that exposed persons are able to avoid the hazard or not (Pa or Pb) and
likelihood/demand category W, indicating the number of times per year that the
hazardous event would occur (W1, W2, W3) to determine the SIL level. The SIL
determination is usually documented in SIL determination worksheets as shown
in Figure 9.

Figure 8. Example Risk Graph for SIL Assignment. (see color insert)
151
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Figure 9. Example Risk Graph Template. (see color insert)

Layer of Protection Analysis (LOPA)

LOPA is a semi-quantitative method for assessing integrity levels, which
enables additional risk mitigation factors, not usually considered by the risk graph
method, to be taken into account and, thus, allows a more realistic estimation
of the required integrity level. This LOPA approach allows for a more accurate
evaluation of risk mitigation from the existing non-SIS layers. If additional
risk reduction is required and if it is to be provided in the form of a Safety
Instrumented Function (SIF), the LOPA methodology allows for a more accurate
and quantitative determination of the appropriate Safety Integrity Level (SIL) for
that SIF.

Initiating Event

The LOPA method starts with data developed in the HAZOP (or another other
form of Hazard Assessment) study and accounts for each identified hazard by
documenting the initiating cause. Each cause is reduced to a discrete failure event.
For example, a loss of cooling water to a reactor (as shown in the earlier reactor
example, Figure 1) as an initiating cause can be the result of an upstream CW pump
failure, a power failure, or a CW control valve failure. The frequency for each type
152
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 of failure would be different, and the layers of protections can be different for each
initiating event.
Once the initiating event occurs, various protective layers can prevent the
event from reaching a dire consequence (fire/explosion). The basic process control
can take action and shut the process down, the operator can take action to bring
the process to a safe state or a PSV can prevent the pressure buildup and relief
to a safe location. Figure 10 shows an Event Tree or a simplified LOPA diagram
for this example scenario. If these layers are not adequate to reduce the risk to an
acceptable level then a SIS can be considered for implementation.
Table 6 provides some generic values that can be selected by the analyst for
various typical initiating events encountered during a LOPA study.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Figure 10. Layers Of Protection Analysis (LOPA diagram)/Event Tree. (see

color insert)

Table 6. Generic LOPA Initiating Event Likelihoods

Initiating Event (IE) Frequency of Failure, yr-1
Analyzer failure all types 1e-1
Atmospheric Tankage leak 50 mm hole (acidic contents) 1.0E -03
Atmospheric Tankage leak 50 mm hole (non-acidic
contents) 1.0E – 04
Atmospheric Tankage leak rupture (acid contents) 1.0E -04
Atmospheric Tankage rupture (non-acidic contents) 1.0E -06
BPCS instrument loop failure 1.0E -01
Check valve failure external leak 1.0E -03
Compressor failure ( 10-50 mm leak) axial 1.0E -03
Compressor failure (10-50 mm leak) centrifugal 1.0E -03
Compressor failure (10-50 mm leak) reciprocating 1.0E -04
Compressor rupture all types (>50 mm leak) 1.0E -04
Cooling water failure diverse drivers/redundant pumps 1.0E -01
Continued on next page.

153
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table 6. (Continued). Generic LOPA Initiating Event Likelihoods
Initiating Event (IE) Frequency of Failure, yr-1
Expansion Joint failure large leak 1.0E -02
Expansion Joint failure rupture 1.0E -03
External impact crane load drop 1.0E -04
External impact 3rd party vehicle, backhoe, etc. 1.0E -02
Filter blockage or leak 1.0E -03
Filter rupture 1.0E -05
Flange leak blown gasket or flange separation 1.0E -04
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Heat exchanger external leak plate & frame 1.0E -02

Heat exchanger external leak shell & tube 1.0E -02
Heat medium failure 1.0E -01
General Utility Failure 1.0E 01
Instrument connection failure 1.0E -03
Lock Out Tag Out procedure failure 1.0E -03
Loss of power (redundant supplies) 1.0E -01
Operator failure – non routine procedure, stressful,
possible fatigue 1.0E -01
Piping – leak 10% section 1.0E -03
Piping – large leak 50mm 1.0E -04
Piping – guillotine fracture 1.0E -05
Pressure vessel <13mm leak in liquid 1.0E -04
Pressure vessel 13mm leak in liquid 1.0E -05
Pressure vessel > catastrophic failure 1.0E -05
Pressure vessel high temp – catastrophic failure 1.0E -05
Pump leak reciprocating 1.0E -01
Pump leak all other types 1.0E -02
Pump catastrophic failure 1.0E -04
Pump fail to run/loose flow 1.0E -02
Single Mechanical seal failure 1.0E -01
Refrigerated tank leak (50mm) 1.0E -03
Refrigerated tank rupture 1.0E -04
Regulator failure 1.0E -01
Relief value opens spuriously 1.0E -02
Continued on next page.

154
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table 6. (Continued). Generic LOPA Initiating Event Likelihoods
Initiating Event (IE) Frequency of Failure, yr-1
Spills 1.0E -01
Turbine/diesel engine speed with casing breach 1.0e -04
Valve failure external leak- motor or air operated valve 1.0E -03
Valve failure external leak – manual valve 1.0E -04
Develop using experience
Other initiating events of personnel
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Layers of Protection

After determining the initiating cause(s) and the consequence, various

independent protection layers (IPL) present are analyzed. Each protection layer
consists of a grouping of equipment and/or administrative controls that function
in concert with the other layers. Protection layers that perform their function
with a high degree of reliability may qualify as Independent Protection Layers
(IPL). Typical IPLs are alarms, pressure relief devices, etc. that help mitigate the
likelihood of an event.
Some mitigation layers may reduce the severity of the impact event but not
prevent it from occurring. Examples would be:

• Restricted access
• Deluge systems for fire
• Gas detection alarms
• Fire detection alarms; and
• Emergency evacuation procedures.

The LOPA team evaluates various IPLs present for every scenario and assigns
appropriate probability of failure on demands (PFDs).
Table 7 lists the multiple Protection Layers (PLs) that are normally provided
in the process industry. The Probability of Failure on Demand (PFD) is the
conditional probability of failure of these independent layers of protection. Thus,
if one assigns a Basic Process Control System (BPCS) layer in a LOPA study, we
can assign a PFD of 1E-01 (or 0.1) which implies that it will work successfully 9
out of 10 times but may fail in1 out of 10 demands.
Conditional probability of ignition (Table 8) is considered for scenarios
involving fire/explosion to estimate the likelihood of an event.
Presence of people (Table 9) is considered to modify the likelihood of
frequencies that can result in potential fatalities.

155
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table 7. PFD for Typical IPLs
Probability of Failure on
Independent Protective Layer (IPL)
Demand (PFD)
Basic process control system 1E-01
Critical Alarm & Human Intervention 1.0E -01
Simple, well documented action with clear reliable
indications that the action is required
Safety Instrumented Function
SIL 1 System 1.0E -01
SIL 2 System 1.0E -02
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

SIL 3 System 1.0E -03

Physical Protection
Double or tandem seal (leak only) 1.0E -01
Grounding & bonding 1.0E -01
Mechanical stop 1.0E -01
Orifice sized sonic flow 1.0E -01
Manual response to abnormal reading collected regularly 1.0E -01
on a check list
Sample collection & analysis to confirm nonhazardous 1.0E -01
conditions
Specific design, i.e., self draining lines 1.0E -01
Check valve 1.0E -02
Cylinder dome (during transport) 1.0E -02
Detonation/Flame arrestor 1.0E -02
Deflagration Vents 1.0E -02
Explosion Suppression System 1.0E -02
Fireproofing 1.0E -02
Housekeeping (dust explosion scenario only) 1.0E -02
Open vent 1.0E -02
Relief valve/ Rupture disk 1.0E -02
Rupture pin 1.0E -02
Blast wall 1.0E -03
Bunker 1.0E -03
Event pressure is less than1.0 times MAWP 1.0E -05
Event pressure is between 1.1 and 2.5 timer MAWP 1.0E -03
Event pressure is between 2.5 and 3.5 times MAWP 1.0E -01
Continued on next page.

156
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table 7. (Continued). PFD for Typical IPLs
Probability of Failure on
Independent Protective Layer (IPL)
Demand (PFD)
Event pressure is over 3.5 times MAWP 1.0E +00
Post release Protection 1.0E -01
Area hydrocarbon detectors 1.0E -01
Concrete Pad with slope 1.0E -01
Deluge System 1.0E -01
Fire Suppression System 1.0E -01
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Protective Clothing – specialized 1.0E -01

Restricted Area 1.0E -01
Safety Shower 1.0E -01
Slash Guard 1.0E -01
Bund / Dike 1.0E -02
Foam blanket vapor suppression system 1.0E -02
Underground Drainage System 1.0E -02
Other events Use experience of
personnel

Table 8. Conditional Frequency Modifier – Generic Ignition Probabilities

Conditional Modifiers Probability of Ignition
Vapors containing more than 10% Hydrogen (any quantity) 1.0
Hydrocarbon (M,I. > 0.3mJ) released less than 500kg 0.01
Hydrocarbon (M,I. > 0.3mJ) released between 500kg and
5,000kg 1.0
Hydrocarbon (M,I. > 0.3Mj) released less than 50 kg 0.01
Hydrocarbon (M,I. > 0.3Mj) released between 50 kg and 500
kg 0.1
Hydrocarbon release more than 5,000 kg 1.0
Due to static in closed underground vessel ordinary
hydrocarbon 0.1
Release of flammable material due to external collision ( crane 1.0
failure), impact or catastrophic failure

157
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table 9. Conditional Likelihood Modifier- People Present
Occupancy Levels Conditional Likelihood Modifier
People are present all the time 1.0
People are present for less than 12 hours per day 0.5
People are present for less than 1-2 hours per day 0.1
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Target Mitigated Event Likelihood (TMEL)

The likelihood of an event is determined by multiplying the initiating event

frequency and PFDs of the various IPLs present. The SIF under consideration is
not considered as an IPL. This intermediate event likelihood (initiating event, all
IPLs except SIS) is compared with the TMEL to determine the IL requirement for
the SIS.
TMEL for various undesirable outcomes (for an example facility) are
summarized in Tables 10, 11, and 12.
Figure 11 represents an example LOPA worksheet.

Table 10. Example TMEL Based on Personnel Risk

Severity
Safety Consequence Tolerable TMEL Target, yr-1
Level
1 Minor injury or First Aid Cases 1x10-2
Lost work injuries involving a few
2 days away from work. Restricted work 1 x 10-3
injuries
Injury resulting in prolonged absence
3 of work or permanent disability, Minor 1 x 10-4
injuries to three or more workers
One or more fatalities and/or
4 1 x 10-5
permanently disabling injuries
Three or more fatalities and/or
5 1 x 10-6
permanently disabling injuries

158
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table 11. Example TMEL Based on Environmental Risk
Severity
Environmental Consequence Tolerable TMEL Target, yr-1
Level
1 No offsite environmental impact 1x10-2
Minor offsite impact with no
2 remediation or costs less than 1 x 10-3
100,000
Moderate offsite environmental
3 impact or remediation costs 1 x 10-4
U$ 100,000 to 1-million
Significant offsite environmental
4 impact or remediation costs 1 x 10-5
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

U$ 1-million to 10-million
Extensive off site environmental
5 damage or remediation costs 1 x 10-6
> U$ 10-million

Table 12. Example TMEL Based on Financial Risks

Severity
Financial Consequence Tolerable TMEL Target, yr-1
Level
1 Less than U$ 100,000 1x10-2
Property damage or production loss
2 1 x 10-3
between U$ 100,000 to 1-million
Property damage or production loss
3 1 x 10-4
between U$ 1-million and 10-million
Property damage or production loss
4 1 x 10-5
between U$ 10-million to 100-million
Property damage or production loss
5 1 x 10-6
greater than U$ 100-million

Figure 11. Example LOPA Template/Example. (see color insert)

159
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 LOPA Study Example for the Nitration Industry
Let’s consider a common reactor configuration in the Nitration Industry.
Figure 12 depicts a continuously stirred reactor in which an organic substrate is
reacted with a mixture of H2SO4 and HNO3 acids. The reaction is exothermic and
the reactor is cooled using a continuous supply of cooling water (through coils).
The reactant and cooling water flow rates are controlled using flow controllers
which rely upon individual flow sensors and set point information from the BCPS
using the temperature and pressure information from the PI and TI sensors (the
BPCS logic is not shown). If the reactor temperature rises for any reason (say loss
of cooling water), we could have a runaway reaction leading to overpressures,
reactor damage and potential worker injury/fatality due subsequent fire/explosion.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Upon the detection of a high temperature the BCPS is programed to close the
organic and acid flows to the reactor. We need to determine if the current safety
layers are adequate to reduce the risk to acceptable levels. If not, we need to
determine what kind of SIS would be required.

Figure 12. LOPA Example: Nitration Industry Example Reactor.

160
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 A simplified Event Tree / LOPA diagram for the system can be constructed as
below (Figure 13).
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Figure 13. LOPA Tree / Event Tree for the Nitration Industry Example. (see
color insert)

We can determine from process operations that loss of cooling water (for
various reasons) can be expected once in a couple of years (say once in 2 years
or 0.5 times per year) which is our initiating event frequency. We can take credit
for the BPCS and assign a PFD of 0.1 (from Table 7). We understand from plant
operations, that even though operators are present at the site, they cannot be relied
upon to perform emergency action with a short period of time. Thus, we assign
the PFD for operator action to be 1.0 (no credit). We assume that the PFD for the
relief valve is 0.01 (Table 7) and ignition is guaranteed upon release (PFD=1.0,
Table 8). The expected outcome without any additional SIS is 5E-4
Based on continuous operator presence we understand that this explosion
event (should it occur) could lead to 1 or more fatalities and, thus, we determine
that the TMEL should be 1.0E-5 (Table 10). As the expected outcome without
any SIS layer is higher than this TMEL value, there is a need for an additional SIS
layer. This SIS layer would need to be configured as a separate independent layer,
which could detect the high temperature and take automatic mitigation action.
We calculate that the required RRF for this additional SIS needs to be 5E-4/
1E-5 = 50. Thus, we need to specify a SIL level of 2 which would have a minimum
RRF of 100.
Figure 14 depicts the Reactor Example with an example simple SIS
configuration. In this example configuration an independent set of TI sensors can
be used along a PLC controller to independently shut off flows of the organic
substrate and H2SO4 and HNO3 acids. We can also open the reactor drain valve
to divert the reactor contents (if appropriate). The configured SIS would need
to meet a specified SIL requirement. The exact configuration with appropriate
number of sensors (2oo2, 2 out of 2 voting logic, or 2oo3, 2 out of 3 voting logic),
PLC and final element combinations would have to be determined / confirmed
during the subsequent verification step to meet this SIL requirement.

161
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Figure 14. LOPA Example: Nitration Industry Example Reactor with additional
SIS.

SIL Verification
The SIL verification study follows the SIL determination study. The
objective of the SIL verification is to check if the SIL1 or higher SIF loops (as
actually configured) meet the required integrity level (IL) specified during the
SIL determination step. The verification is based on the failure rate data of the
individual components (initiator, logic solver, final element) of the proposed
as-configured SIF loop.
SIL verification is typically performed in accordance with IEC 61511 (1).
Exponent uses exSILentia (2) software available from Exida for SIL verification
purposes. Failure data for an element is used to estimate the probability of failure
on demand (PFD) for a given time period. PFDavg of sensors, logic solvers and
final elements is used to calculate the PFDavg for the SIF. As can be seen from the
equation below, the PFDavg value depends on the proof test interval (TI). For the
typical studies, a proof test interval of 6-12 months is used.
162
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The Risk Reduction Factor (RRF) is simply the inverse of the PFDave (RRF
= 1/PFDavg). The SIL verification calculation also estimates MTTFS (Mean Time
to Fail Spurious) for the SIF.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

In this example, since the final element PFD is more than an order of
magnitude higher than the other two elements, the PFD for the entire SIF will
essentially be the PFD of the final element.

Note, spurious trip rate for an element is 1/MTTFS for that element.
Besides the PFDavg requirement for a given SIL, IEC 61508 (3) and IEC
61511 impose architectural constraints on hardware components utilized in a safety
function. The architectural constraints represent hardware tolerance (minimum
faults that could lead to a loss of the SIF) for a given SIL, and depend upon the
hardware’s Safe Failure Fraction (SFF).
163
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Tables 13 and 14 illustrates hardware fault tolerance for Type A (e.g. valves,
relays, switches) and Type B (e.g. transmitters, PLCs) architectures as specified in
IEC 61508. A hardware fault tolerance of N implies that N+1 faults could cause
a loss of the safety function. For example, a single (1oo1) Type A control valve
with a SFF of 58% has a hardware fault tolerance of 0 and is allowed for SIL1
applications; however, for a SIL2 application a redundant, one-out-of-two (1oo2)
arrangement, or two valves in series, is required to meet architectural constraints.

Table 13. IEC 61508 Architectural Constraints for Type A Systems (see
color insert)
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Table 14. IEC 61511 Architectural Constraints for Type B Systems (see
color insert)

The architectural constraints thus impose additional requirements on the

hardware selection and arrangement to achieve a target SIL. The SIL for a safety
function presented in this report represents the lower of the two SILs – SIL based
on PFDavg and SIL based on architectural constraints.

164
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Failure Rate Data

The failure rate data used for SIL calculations are typically obtained from a
built-in instrument failure rate database (4) in the Exida software and from specific
failure data provided by the Company. Requests are made from the Company/
Vendor for all SIF components’ model/make prior to SIL verification. The failure
data available for each SIF is summarized and sent to Company/Vendor for review
prior to verification. Where specific failure data are not available, generic failure
rate data is used.
Figure 15 depicts a typical SIL verification Fault Tree Analysis (FTA) model
that is constructed for SIL verification purposes. Figure 16 depicts that example
SIL verification model results.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Figure 15. SIL Verification Example FTA Model. (see color insert)
165
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Figure 16. Example SIL Verification Model Results. (see color insert)

Maintenance
Proof testing is a requirement of safety instrumented systems to ensure that the
SIS is working and performing as expected. Testing must include the verification
of the entire system, logic solver, sensors, and final elements. The interval is the
period of time that the testing occurs. The testing frequency varies for each SIS and
is dependent on the technology, system architecture, and target SIL level. As can
be seen in Figure 17, it is possible to achieve higher SIL levels if testing frequency
is increased. The proof-test interval is an important component of the probability
of failure on demand calculation for the system.

• A typical mission time of 20-years is used based on average lifecycle of

most process facilities.
• A proof test interval (PTI) of 6-12 months is assumed for most sensors.
• PTI of 60-months (5-years) is assumed for logic solver.
• PTI of 6-12 months is used for typical final elements (e.g., safety relays
involved in final SIF action and valves).
• The proof test coverage (measure of effectiveness of proof test in
detecting failures undetected by automatic diagnostics) is usually
assumed to be 90% (based on industry accepted values of 80% to 95%).
• A common cause failure Beta of 10% is usually assumed based on
industry accepted values of 5% to 10%.
166
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 • Process conditions are assumed to be “clean”, i.e. there is no extreme
process environment resulting in failure rates higher than reported unless
the process conditions indicate otherwise.
• Mean time to Repair (MTTR) is assumed to be 24 hours for sensors, logic
solvers and final elements.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

Figure 17. Achieved SIL Based on Testing Frequency. (see color insert)

It is often assumed that (asking for and) getting a higher rated SIS is always
beneficial. However, one has to remember that as the SIL level increases, typically
the installation and maintenance costs and complexity of the system also increase.
Specifically, for the process industries, SIL 4 systems are so complex and costly
that they are not economically beneficial to implement. Usually, if process risks
are such that a SIL 4 system is required to bring the system to a safe state, then
this is usually an indication that the process design needs to be revised.
167
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 SIL based risk reduction is certainly not always the most cost-effective
solution for decreasing process risk. Implementing a SIL loop may require
increased equipment, which inevitably will require increased maintenance.
Higher SIL level will require more frequent proof testing which may ultimately
increase the amount of system maintenance. Thus, a SIL based risk reduction
is only recommended when process risk cannot be effectively reduced by other
methods.

Summary
An overview of Safety Instrumented Systems based approaches was
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

presented. The SIS systems are designed to prevent or mitigate hazardous

events by preemptively taking the process to a safe state. Risk Graph and
LOPA approaches to determine the appropriate Safety Integrity Levels for SIS
applications were presented. An example from the Nitration Industry (Reactor,
with Organic substrate & H2SO4/HNO3 reactants) was presented to demonstrate
a SIL determination using the LOPA methodology. SIS verification and testing
requirements were also discussed. Additional references for SIL determination
and verification approaches can be found in References (5–11).

References
1. Functional Safety: Safety Instrumented Systems for the Process Industry
Sector; IEC 61511; International Electrotechnical Commission (IEC):
Geneva, Switzerland, 2003.
2. exSILentia™, version 3.0.9.785; exida: Sellersville, PA.
3. Functional Safety of Electrical/Electronic/Programmable Electronic
Safety-Related Systems; IEC 61508; International Electrotechnical
Commission (IEC): Geneva, Switzerland, 2010.
4. Safety Equipment Reliability Handbook, 2nd ed.; exida: Sellersville, PA,
2005.
5. Classification and Implementation of Instrumented Protective Functions;
DEP 32.80.10.10-Gen; Shell: Hague, The Netherlands, January 2010.
6. Goble, W. M.; Cheddie, H. Safety Instrumented Systems Verification:
Practical Probabilistic Calculations; ISA - The Instrumentation, Systems
and Automation Society: Research Triangle Park, NC, 2005; ISBN
1-55617-909-X, [E].
7. Guidelines for Safe Automation of Chemical Processes; American Institute
of Chemical Engineers, Center for Chemical Process Safety (CCPS): New
York, 1993; ISBN 0-8169-0554-1.
8. Scharpf, E.; Hartmann, H.; Thomas, W. T. Practical Safety Integrity Level
(SIL) Target Selection; exida: Sellersville, PA, 2012; ISBN 978-1-934977-
03-3.
9. Lees, F. P. Loss Prevention in the Process Industries, 3rd ed.; Butterworth-
Heinemann: Oxford, U.K., 2005; ISBN 978-0-7506-7555-0.
168
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 10. Layer of Protection Analysis: Simplified Process Risk Assessment; American
Institute of Chemical Engineers, Center for Chemical Process Safety (CCPS):
New York, 2001; ISBN 978-0-8169-0811-0.
11. Perry, R. H.; Green, D. W. Perry’s Chemical Engineer’s Handbook, 7th ed.;
McGraw-Hill: New York, 1997; ISBN 0-07-049841-5.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch011

169
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 12

Lessons Learned from an Explosion in an

Ammonium Nitrate Neutralizer
Roland Huet* and Lawrence E. Eiselstein
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

Exponent Failure Analysis Associates, 149 Commonwealth Drive,

Menlo Park, California 94539
*E-mail: [email protected]

The factors leading up to the large explosion that occurred

in an ammonium nitrate neutralizer in Iowa in 1994 will be
discussed. Ammonium nitrate (in the form of 83 wt% solution
in water) was made in a neutralizer tank by reacting nitric acid
with ammonia. At the time of the accident, the neutralizer
had been shut down but the contents were kept hot by steam
injected through the nitric acid sparger. Several operational
factors combined to cause the explosion: the ammonium
nitrate was acidified when the nitric acid line was purged
into the neutralizer; the neutralizer contents were not being
circulated to maintain homogeneity; the neutralizer had been
contaminated with chlorides; the prolonged steam heating
raised the neutralizer temperature; and the steam bubbles
facilitated runaway decomposition.

Ammonium Nitrate Explosions

Ammonium nitrate is widely and safely used as a fertilizer and in many other
applications. However, it is well known that under certain circumstances, such as
contamination with hydrocarbons or excessive confinement, it can be a powerful
explosive. Mixtures of ammonium nitrate and fuel oil (ANFO) are routinely used
as explosives. Although the conditions leading to explosions of pure ammonium
nitrate are rarely observed in practice there have been a few large accidental
explosions of ammonium nitrate. Three such explosions will be briefly reviewed
here.

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Texas City, 1947

A large shipment of fertilizer grade ammonium nitrate was loaded and stowed
in two cargo ships (the S. S. Grand Camp and the S. S. High Flyer) in the Texas
City harbor in April 1947 for shipment to Europe (1) . The ammonium nitrate
fertilizer was packaged in paper bags; the fertilizer grains were lightly coated with
paraffin (3 wt. %) to avoid caking when exposed to humid conditions (1–13).
The S. S. Grand Camp was carrying 2200 tons of the wax-coated ammonium
nitrate in paper bags when a fire started in one of the ship holds and the crew tried
to smother the fire by shutting all hatches and injecting steam into the holds (1).
The hydrocarbons from the wax and paper, the heat from the fire and the injected
steam, and confinement by the closed holds provided conditions necessary for a
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

detonation. After several hours, the Grand Camp detonated on April 16, 1947
in the harbor, killing about 600 people in the surrounding dock area and causing
over $100 million of property damage (6). The fire spread to the S.S. High Flyer,
which detonated a few hours later on April 17 (4). A very similar accident occurred
a few months later in Brest, France, one of the European harbors receiving these
ammonium nitrate shipments (1, 6).
As a result of these accidents, safer anti-caking agents were developed.
Ironically this accidental explosion also led to the development of a new family
of industrial explosives, ANFO (6).

Port Neal, 1994

An ammonium nitrate fertilizer manufacturing plant located at Port Neal,

Iowa, owned by Terra, exploded on December 13, 1994. This explosion killed
four Terra employees, injured more than a dozen people, and devastated the plant.
This is the incident we will discuss in greater detail below. During a shutdown
of the ammonium nitrate production during freezing weather conditions, the
ammonium nitrate in the neutralizer vessel was kept hot by steam injection to
prevent solidification. In addition, the ammonium nitrate had inadvertently been
allowed to become very acidic and contaminated with chlorides; both conditions
are known to promote explosive decomposition. After several hours of steam
injection, the ammonium nitrate detonated.

Toulouse, France, 2001

An explosion of about 400 tons of ammonium nitrate occurred in Toulouse

at the AZF plant (Azote et Fertilisants) on 21 September 2001 (14–16). The
origin was in a warehouse used to collect ammonium nitrate that did not meet
specifications for finished product. The cause of this explosion is unclear, but the
leading theory is that a quantity of chlorinated compounds was inadvertently added
to the ammonium nitrate pile shortly before the explosion.

172
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The Accident at Port Neal, Iowa
At 6 am, on the 13th of December 1994, two process vessels (the neutralizer
and run-down tank) at Terra International’s Port Neal, Iowa ammonium nitrate
fertilizer plant detonated, killing four employees, injuring 18 others. All buildings
and equipment within a 200-ft radius were leveled (17–19). Although there were
numerous reasons proposed for this explosion, an independent court appointed
expert confirmed that the initial explosion occurred outside the nitric acid
sparger. The significance of this and other findings hold valuable lessons for plant
operators.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

Plant Layout

Terra Industries’ Port Neal complex manufactured anhydrous ammonia from

methane, water and air. The ammonia was split into four different streams: one
for urea manufacture, a second for the manufacture of 55 wt% nitric acid, a third
stream was combined with the nitric acid to make ammonium nitrate, and the
fourth ammonia stream was for direct sales. Thus the major output products of
the site were urea (granules, prill and liquor), ammonium nitrate and ammonia.

Ammonium Nitrate Neutralizer

The ammonium nitrate neutralizer is the vessel in which the ammonium nitrate
is made by reaction between ammonia gas and 55 wt% nitric acid. The process is
shown in Figure 1 and Figure 2. The neutralizer vessel is about 10 feet in diameter
and 15 feet high, it contains a hot 83 wt% solution of ammonium nitrate in water.
Ammonia gas is injected at the bottom of the neutralizer through a perforated
plate. Nitric acid is injected into the neutralizer through two titanium spargers,
which are two tubes 4 inches in diameter bent to form two semi-circles, with
two rows of holes to distribute the nitric acid. The ammonia and nitric acid react
together to form more ammonium nitrate solution and the heat of reaction keeps
the temperature elevated. The neutralizer vessel is not actively cooled; some of the
water made during the reaction evaporates naturally to balance the heat of reaction.
There is no mechanical stirring; the solution is mixed by the rise of the ammonia
bubbles up the center of the tank through a vertical draft tube open at both ends.
The vertical draft tube helps set up circulation of the liquid ammonium nitrate, up
inside the tube and down outside the tube.
An overflow line on the neutralizer directs the hot ammonium nitrate solution
into the rundown tank, from which it is pumped to storage tanks. Vapors in the
neutralizer head space, rich in ammonia, are sent to a scrubber where they are
neutralized with a weak nitric acid solution (pH above 1.4).

173
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

Figure 1. Schematic of the neutralizer process. The ammonium nitrate generated

in the neutralizer flows into the rundown tank before being pumped out. The
neutralizer off-gas, rich in ammonia, is treated in the scrubber. (Courtesy of
Roland Huet). (see color insert)

Figure 2. Interior view of the scrubber, neutralizer and rundown tank. Green
indicates the ammonium nitrate solution, blue the ammonia and yellow the nitric
acid. The spargers are on the outside of the draft tube. (Courtesy of Roland
Huet). (see color insert)
174
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The Explosion
On December 12, 1994 at approximately 9 pm in the evening, the neutralizer
at the Port Neal fertilizer plant was put in hot shutdown. Since there was no longer
any heat of reaction to keep the neutralizer hot, and the temperature outside was
10°F (-12°C) the operators were concerned that the ammonium nitrate would salt
out and freeze in the neutralizer. To avoid this possibility, they injected 200 psig
steam through a hose connected to the nitric acid spargers (17). Saturated steam
at this pressure is at about 387 °F (197 °C). The steam flow rate was not known
nor controlled, and the plant had not set a maximum or minimum temperature
for this abnormal operating condition. After about 9 hours of continuous steam
injection, conditions in the steam supply system changed, which may have led
to superheated steam of unknown temperature being injected into the neutralizer.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

About 30 minutes later, there was a detonation in the neutralizer and the rundown
tank. This resulted in four fatalities, 18 injuries and extensive damage to the plant
(17, 18, 20–22).

Controversy About the Cause of the Explosion

As may be expected in an accident of this magnitude, there was some
controversy between the plant operator and the equipment designer about the
causes of the explosion. The neutralizer design and the plant operation were the
subject of a contentious lawsuit; this fact and the extensive destruction of records
caused by the accident were such that many details of the plant layout, process
and operation were unclear. We will present the proposed theories, established
facts, controversy regarding the root cause, and the conclusions of an independent
court appointed expert.

Theories, Established Facts, and Open Questions

During the investigation of this accident, several theories were proposed for
the explosion. Some were quickly abandoned because they did not fit with the
observed evidence. For instance, there was speculation that some of the pumps
shown in Figure 1 may have been the cause of the explosion (17, 18). However,
the damage pattern indicated clearly that the first detonation occurred inside
the neutralizer (there was no damage to the neutralizer from an outside impact)
and none of the pumps that were recovered showed any sign of an internal
explosion(17).
Two competing theories remained in contention. Terra Industries and its
experts asserted that some ammonium nitrate had backed up into the titanium
nitric acid spargers and that the first explosion occurred inside this tube because
of the severity of the conditions at that location: increased confinement, increased
contamination by metal ions, and violent oxidation or burning of the titanium
tube that generated enough heat to initiate an explosion (23, 24). On this basis,
Terra sued Mississippi Chemical, the designer of the neutralizer, for defective
design of the sparger tube. Mississippi Chemical and its experts, including
Exponent, countered that poor operating conditions, specifically acidification
175
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 of the ammonium nitrate, chloride contamination and steam injection for many
hours, were the real reason for the explosion (21, 25).
The neutralizer was in hot shutdown at the time of the explosion, and steam
had been flowing through the nitric acid spargers for about 9 hours. Evidence
strongly suggested that the contents of the neutralizer were very acidic and
contaminated with chlorides; two conditions known to promote the explosive
decomposition of ammonium nitrate. Numerous tests, such as accelerating rate
calorimeter (ARC) tests, confirmed that acidification and chloride contamination
increased the rate of decomposition of ammonium nitrate. However, the severity
of the contamination was not clearly established, as discussed later, so there was
a considerable disagreement between the parties on the conditions inside the
neutralizer.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

The key question was whether the conditions inside the neutralizer itself
(heat, acidity, chlorides and confinement) were enough to lead to a detonation,
or if other factors present only inside the sparger tubes were necessary for the
detonation to occur. The other factors that could only affect ammonium nitrate
inside the spargers were confinement of the ammonium nitrate in a small volume,
and reaction of the ammonium nitrate with the titanium walls of the spargers. The
controversy thus narrowed to the location of the initial explosion: did it occur in
the neutralizer itself due to the heating and contamination of its content, or did it
initiate inside a nitric acid sparger, thus implying that confinement and reaction
with titanium were necessary for the explosion?

Independent Assessments

Two independent assessments concluded that the explosion occurred first in

the neutralizer itself, thus denying that the spargers played a role in the accident
(17, 22).
The first assessment was made by the Environmental Protection Agency
(EPA) in their investigation of the explosion (17). The main piece of evidence on
which they relied was the recorded level in the rundown tank, shown in Figure
3. The liquid level, measured with a pressure transducer, appeared to rise rapidly
to 100% (or higher) a few seconds before the explosion. Clearly, this is unlikely
to reflect an actual rise of the level in the rundown tank, but is more likely to
be an artifact of a change in pressure. Even though the pressure transducer was
believed to be a differential pressure transducer and should have been insensitive
to changes in the pressure inside the vessel, the only credible explanation for the
recorded level rise is a change in pressure inside the vessel (it is possible, for
instance, that the upper leg of the level transducer was plugged). This apparent
level rise was due to a large-scale decomposition of the ammonium nitrate inside
the neutralizer during the last few seconds before the detonation, which rules
out an initial detonation inside one of the spargers. If the initial explosion had
occurred inside a sparger and triggered a detonation of the entire neutralizer, this
detonation would have occurred immediately after the sparger explosion, and
there would not have been any pressure rise inside the neutralizer in the few
seconds before the detonation.

176
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

Figure 3. Recorded liquid level in the rundown tank just before the explosion.
(Reproduced from ref. (17)).

The second independent assessment was performed by an expert appointed

by the court in the trial between the plant operator and the equipment designer.
The court-appointed expert examined closely the few recovered fragments of
the spargers and determined, through careful analysis of their shape and some
computer modeling, that the initial explosion had occurred outside of the spargers,
not inside (22).

Accident Main Factors

The independent assessments indicated that the conditions inside the
neutralizer itself (acidity, chlorides, excessive heat, and confinement) were
sufficient to cause the detonation. The reasons for the presence of these four
factors are discussed below.

Acidity

Actions of the plant operators prior to and during the hot shutdown caused the
ammonium nitrate in the neutralizer to become very acidic, but the operators did
not realize it. Low pH is a condition known to promote explosive decomposition
of ammonium nitrate (26–32).

177
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

Figure 4. Key neutralizer operating parameters a few hours before the explosion.
(Courtesy of Roland Huet). (see color insert)

Figure 4 shows four key operating parameters for the neutralizer a few hours
before the explosion, which occurred at about 6 am on December 13, 1994. The
neutralizer was operating normally until about 8 am on December 12. At about
10 am, the neutralizer was placed in partial shutdown, with a reduction in the
ammonia flow (labeled offgas flow in Figure 4) but no corresponding reduction
in the acid flow rate. This led to an acidification of the neutralizer, where the pH
plunged to very low values at about 1 pm. The operators apparently became aware
of this situation and tried to remedy it starting at about 2 pm: they shut off acid
flow but maintained the ammonia flow. The neutralizer appeared to respond and
by about 5 pm the recorded pH was back to 6. However, the operators failed to
realize that the pH probe was not in the neutralizer itself, but in the overflow line
between the neutralizer and the rundown tank (see Figure 1). Since the neutralizer
was not in operation, no ammonium nitrate was flowing down this process line
and the probe was probably measuring the pH of a small pocket of product in
the line. This small quantity of ammonium nitrate was likely neutralized by the
ammonia flow into the neutralizer, but there was not enough ammonia injected
into the neutralizer to bring its pH back to neutral. Figure 5 shows the pH in the
neutralizer computed from the flows of acid and ammonia. The computed pH is
in good agreement with the measurement made on a grab sample taken just before
the partial shutdown, but becomes very acidic soon after and remains very acidic
up to the time of the explosion. The operators did not take any grab samples after
the start of the partial shutdown, so they were not aware of the extreme acidity in
the neutralizer.
178
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

Figure 5. Measured pH in rundown line and calculated pH in neutralizer.

(Courtesy of Roland Huet). (see color insert)

Chloride Contamination

The contentious nature of the lawsuit made it difficult to obtain detailed

information on the potential sources of chloride contamination. Nonetheless, it
appears that the neutralizer was contaminated by chlorides originating from plant
cooling water that leaked into the nitric acid manufacturing stream and from
there into the AN plant. The chloride levels in the nitric acid were apparently
not monitored, even though the nitric acid plant is a logical source of chloride
contamination: the EPA report on this incident stated that the level of chloride in
the nitric acid at the plant had not been analyzed since 1980 (17). Furthermore
maintenance records indicated that there was at least one serious cooling water
leak into a condenser in the nitric acid production plant, leading to chloride
contaminated nitric acid. This leak was recorded on operator logs from December
5-8, and led to a nitric acid plant shutdown on December 8. Increased chloride
concentrations are also likely during unstable acid plant operations, such as during
startup and shutdown, and the nitric acid plant had been started twice and shut
down three times during the week preceding the explosion, for various reasons.
Chlorides formed during acid plant operations concentrate in the weak acid trays
of the absorption column. Following the emergency shutdown of the nitric acid
plant on December 12, contents of the absorption column were pumped to the
nitric acid storage tank, introducing chlorides into the ammonium nitrate (AN)
process.
179
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The EPA investigation team concluded that chlorides were present in the nitric
acid used in the AN plant. Analytical data provided by Terra indicated chloride
concentrations in the west AN storage tank were 168 parts per million, and 557
parts per million in the nitric acid absorption column and the EPA investigation
team concluded that chlorides were present in the neutralizer, rundown tank, and
west AN storage tank at the time of the explosion.
At about 6 pm on December 12, the nitric acid absorption columns in another
part of the plant were drained for repairs. The drained nitric acid was highly
contaminated with up to 557 ppm of chlorides. Chlorides are another contaminant
well known for accelerating the explosive decomposition of ammonium nitrate
(14, 33–42), and action levels based on chloride concentrations range from
1 part per million to 80 parts per million (17). The operators meant to send
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

the contaminated nitric acid back to the nitric acid storage tank, but some
contaminated nitric acid was pumped into the neutralizer by mistake, as shown in
Figure 4: the nitric acid flowmeter recorded a fluid flowing into the neutralizer
between about 6:30 pm and 9 pm, Dec 12th, a time when operators denied adding
any acid into the vessel. A likely explanation for this recorded flow is that the
operators intended to send the contaminated nitric acid from the absorption towers
to the nitric acid storage tank, but they used for this operation a pump meant
to bring nitric acid into the neutralizer. Whether through improper positioning
of valves or misunderstanding of the flow characteristics of the various lines
used, some of the chloride-contaminated nitric acid was inadvertently sent to
the neutralizer, as shown in Figure 4 . However, the amount of flow into the
neutralizer, hence the chloride contamination level in the neutralizer, is not known
precisely. Apparently, the operators were not aware of the dangerous effects of
chlorides on ammonium nitrate stability.

Excessive Heat
With the neutralizer in shut-down, the operators were concerned about
keeping its contents hot during the freezing cold December night in Iowa. The
neutralizer was equipped with heating coils near its bottom, designed for this
purpose, but they could not be used because of disrepair. The operators instead
used a flexible hose to send steam at 200 °C (about 400 °F) directly into the
nitric acid sparger, to keep the neutralizer hot and preventing salting out of
solid ammonium nitrate. This steam injection lasted for about 9 hours, until
the explosion. Further, it is likely that the steam pressure and temperature were
increased about one hour before the explosion, in anticipation of a significant
steam demand to restart another part of the plant. The steam provided the heat
required for the explosive decomposition of the ammonium nitrate.
The temperature distribution inside the neutralizer before the explosion is not
known precisely. The neutralizer temperature trend is shown in Figure 4 up to 10
pm on December 12, about 8 hours before the explosion. The recorded temperature
remained approximately constant until the explosion. It is likely that the recorded
temperature did not capture some local heating within the neutralizer, which was
no longer effectively stirred: experiments showed that the steam bubbles collapsed
within a few inches of leaving the sparger. As opposed to ammonia bubbles, which
180
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 provided strong mixing in the neutralizer during normal service, the collapsing
steam bubbles did not provide any significant mixing of the ammonium nitrate
in the neutralizer, so the heat brought in by the steam may not have been evenly
distributed in the solution; therefore, the neutralizer thermocouple located away
from the titanium sparger may not have registered the highest temperature in the
neutralizer.

Confinement

Confinement refers to the situation where gases and other decomposition

products are not able to expand freely, thus increasing the pressure in the reaction
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

zone. In the case of ammonium nitrate, this increased pressure tends to favor the
rapid explosive decomposition path, which may lead to an explosion (16, 43–45).
In the Terra plant, the neutralizer itself was able to provide some confinement for
the ammonium nitrate: the neutralizer was an enclosed space, gases generated
at a sufficiently rapid rate cannot escape and therefore the pressure will rise.
Further, the large volume of the ammonium nitrate solution will also provide
pressurization and confinement through hydrostatic and inertial effects.

Conclusions
At its roots, the Port Neal explosion was due to insufficient monitoring during
abnormal plant conditions (hot shutdown), because of a lack of operator training
and awareness of process hazards. It was not clear that a proper and effective
process hazard analysis had been done at the plant, or that the potential hazards that
could lead to accelerated decomposition of ammonium nitrate had been identified.
As a result, the operators did not realize that the pH measurements were not reliable
when the neutralizer was not in production and did not take grab samples during
the shutdown. In addition, they were not aware that chlorides would accelerate
the explosive decomposition of ammonium nitrate and did not realize that they
had sent contaminated nitric acid from the absorption columns to the neutralizer,
instead of sending this nitric acid to a storage tank. Finally, they were not aware of
the danger of “cooking” the neutralizer with high-temperature steam for 9 hours.
All these conditions combined to cause this deadly explosion.

Acknowledgments
The authors would like to acknowledge that the finding regarding the
Terra/Mississippi Chemical incident were the result of many investigators
from both the government and various companies. The investigation that was
performed at Exponent involved many different consulting experts; however we
would specifically like to mention the efforts and contributions of Drs. C. A. Rau,
Jr., B. Ross and Mr. B. McGoran.

181
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 References
1. Greiner, M. L. Ammonium nitrate fertilizer: Exploding the myth. Ammonia
Plant Saf. 1986, 25, 1–9.
2. Burns, J. J.; Scott, G. S.; Jones, G. W.; Lewis, B. Investigation of the
Explosibility of Ammonium Nitrate; U.S. Department of the Interior, Bureau
of Mines: Pittsburgh, PA, 1953; pp 1−19.
3. Holtzscheiter, E. W. Hazards Analyses of Hydrogen Evolution and
Ammonium Nitrate Accumulation in DWPF, Revision 1; Westinghouse
Savannah River Co.: Aiken, SC, 1994.
4. Kintz, G. M.; Jones, G. W.; Carpenter, C. B. Explosions of Ammonium Nitrate
Fertilizer on Board the S.S. Grandcamp and S.S. High Flyer at Texas City,
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

Texas, April 16, 17, 1947; U.S. Department of the Interior, Bureau of Mines:
Pittsburgh, PA, 1948; p 16.
5. Scott, G. S.; Grant, R. L. Ammonium Nitrate: Its Properties and Fire and
Explosion Hazards (A Review with Bibliography); U.S. Department of the
Interior, Bureau of Mines: Pittsburgh, PA, 1948; p 32.
6. Meyers, S.; Shanley, E. S. Industrial explosives: A brief history of their
development and use. J. Hazard. Mater. 1990, 23, 183–201.
7. Fedoroff, B. T. Encyclopedia of Explosives and Related Items, Vol. I;
Picatinny Arsenal: Dover, NJ, 1960; pp A319−A333.
8. The Texas City Disaster, A Staff Report. The Quarterly, NFPA, 1947, Vol. 41,
Issue 1.
9. Medard, L. A. The Explosive Properties of Ammonium Nitrate. In
Accidental Explosions: Volume 2: Types of Explosive Substances; Marshall,
V. C., Ed.; Ellis Horwood Limited: Chichester, 1989; Chapter 24, pp
569−591.
10. Macy, P. F.; et al. Investigation of Sensitivity of Fertilizer Grade Ammonium
Nitrate to Explosion; Picatinny Arsenal: Dover, NJ, 1947.
11. Cox, G. W. Observations relating to the Texas City Disaster. Ind. Med. 1947,
16 (7), 352–354.
12. Ferling, J. Texas City Disaster. Am. Hist. 1996, 30 (6), 48–64.
13. Braidech, M. M. The Texas City Disaster. Facts and Lessons, The National
Board of Fire Underwriters: Paper presented at the 18th Annual Convention
of the Greater New York Safety Council, New York,April 16, 1948, p 1−16
(archived at Moore Library in Texas City, Texas).
14. Li, X. R.; Koseki, H. Study on the contamination of chlorides in ammonium
nitrate. Process Saf. Environ. Prot. 2005, 83 (1), 31–37.
15. Marlair, G.; Kordek, M.-A. Safety and security issues relating to low capacity
storage of AN-based fertilizers. J. Hazard. Mater. 2005, 123 (1–3), 13–28.
16. Dechy, N.; et al. First lessons of the Toulouse ammonium nitrate disaster,
21st September 2001, AZF plant, France. J. Hazard. Mater. 2004, 111 (1–3),
131–138.
17. Thomas, M. J.; et al. Chemical Accident Investigation Report: Terra
Industries, Inc., Nitrogen Fertilizer Facility, Port Neal, Iowa; U.S.
Environmental Protection Agency: Kansas City, KS, 1996.

182
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 18. Kirschner, E. Three lawsuits filed over Terra explosion. Chem. Eng. News
1995, September 11.
19. Hill, P. L.; et al., Expert Review of EPA Chemical Accident Investigation
Report: Terra Industries, Inc., Nitrogen Fertilizer Facility; National Institute
for Chemical Studies: Washington, DC, 1996.
20. Mannan, M. S.; West, H. H.; Berwanger, P. C. Lessons learned from recent
incidents: Facility siting, atmospheric venting, and operator information
systems. J. Loss Prev. Process Ind. 2007, 20 (4-6), 644–650.
21. Fleming, M. A. Following failure to success–An interview with Charles Rau.
J. Fail. Anal. Prev. 2001, 1 (1), 18–24.
22. Shockey, D. A.; Simons, J. W.; Kobayashi, T. Cause of the Port Neal
ammonium nitrate plant explosion. Eng. Fail. Anal. 2003, 10 (5), 627–637.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

23. Baker, Q. A.; et al. Investigation into the Origin of the Explosion at the Terra
Fertilizer Plant, Port Neal, Iowa; Limited Distribution Report, APTECH
Project AES 94122340-4; Aptech Engineering: Sunnyvale, CA, April 3,
2000.
24. Clark, K. J.; et al. A response to explosion case study, Letters to the Editor.
Pract. Fail. Anal. 2001, 1 (6).
25. Rau, C.; et al. Mississippi Chemical Corporation’s Position Paper for the
Court-Appointed Expert; April 5, 2000.
26. Bespalov, G. N.; Filatova, L. B.; Shidlovskii, A. A. Effect of nitric acid and
water on the thermal decomposition of an ammonium nitrate melt. Issled.
Obl. Neorg. Tekhnol. 1972, 31–35.
27. Kolaczkowski, A.;Biskupski, A. Nitric Acid and Ammonia Concentrations
as Factors Controlling the Thermal Decomposition of Ammonium Nitrate.
In the 6th International Conference on Thermal Analysis, Birkhauser Verlag:
Bayreuth, Germany, 1980.
28. Moshkovich, E. B.; et al. The effect of additions of acid on thermal
decomposition of concentrated ammonium nitrate solutions. Sov. Chem.
Ind. 1984, 16 (6), 732–740.
29. Smith, R. D. The thermal decomposition of ammonium nitrate: The roles of
nitric acid and water. Trans. Faraday Soc. 1957, 1341–1345.
30. Wood, B. J.; Wise, H. Acid catalysis in the thermal decomposition of
ammonium nitrate. J. Chem. Phys. 1955, 23 (4), 693–696.
31. Sinditskii, V. P.; et al. Ammonium nitrate: Combustion mechanism and the
role of additives. Propellants, Explos., Pyrotech. 2005, 30 (4), 269–280.
32. Sun, J.; et al. Catalytic effects of inorganic acids on the decomposition of
ammonium nitrate. J. Hazard. Mater. 2005, 127 (1–3), 204–210.
33. Rubtsov, Y. I.; et al. Kinetics of the influence of Cl- on thermal decomposition
of ammonium nitrate. Zhurnal Prikl. Khim. 1989, 62 (11), 2417–2422.
34. Rubcow, I. Y.; et al. Kineticheskoje zakonomernosti blijania chloride na
termicheskoje razlozenije ammiatchnoj selitry. Zurnal Prikl. Khim. 1989,
11, 2417–2422.
35. Rozman, B. Y. Mechanism of thermal decomposition of ammonium nitrate.
J. Appl. Chem. USSR (Zhurnal Prikl. Khim.) 1960, 33 (5), 1052–1059.
36. Petrakis, D. E.; Sdoukos, A. T.; Pomonis, P. J. Effect of the first row
transition metal cations on the mode of decomposition of ammonium nitrate

183
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 supported on alumina-aluminum phosphate and the final products obtained.
Thermochim. Acta 1992, 196, 447–457.
37. Pany, V. Study of the influence of the chloride on the decomposition of
ammonium nitrate by means of thermal analysis. Chem. Abstr. 1976, 85
(56018).
38. Oxley, J. C.; Kaushik, S. M.; Gilson, N. S. Thermal stability and
compatibility of ammonium nitrate explosives on a small and large scale.
Thermochim. Acta 1992, 212, 77–85.
39. MacNeil, J. H.; et al. Catalytic decomposition of ammonium nitrate in
superheated aqueous solutions. J. Am. Chem. Soc. 1997, 119 (41),
9738–9744.
40. Keenan, A. G.; Notz, K.; Franco, N. B. Synergistic catalysis of ammonium
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch012

nitrate decomposition. J. Am. Chem. Soc. 1969, 91 (12), 3168–3171.

41. Keenan, A. G.; Dimitriades, B. Mechanism for the chloride-catalyzed
thermal decomposition of ammonium nitrate. J. Chem. Phys. 1962, 37 (8),
1583–1586.
42. Colvin, C. I.; Fearnow, P. W.; Keenan, A. G. The induction period of the
chloride-catalyzed decomposition of ammonium nitrate. Inorg. Chem. 1965,
4, 173–176.
43. Russell, T. P.; Brill, T. B. Thermal decomposition of energetic materials 31:
Fast thermolysis of ammonium nitrate, ethylenediammonium dinitrate and
hydrazinium nitrate and the relationship to the burning rate. Combust. Flame
1989, 76, 393–401.
44. Ettouney, R. S.; El-Rifai, M. A. Explosion of ammonium nitrate solutions,
two case studies. Process Saf. Environ. Prot. 2012, 90 (1), 1–7.
45. Oxley, J. C.; et al. Ammonium nitrate: Thermal stability and explosivity
modifiers. Thermochim. Acta 2002, 384 (1), 23–45.

184
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 13

Process Design and Operational Controls To

Safeguard Strong Nitric Acid Recovery Systems
Thomas L. Guggenheim,* Roy R. Odle, and John Pace
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

SABIC Innovative Plastics, 1 Lexan Lane, Mt. Vernon, Indiana 47620

*E-mail: [email protected]

A nitric acid concentrator-sulfuric acid concentrator

(NAC/SAC) is a manufacturing process that produces strong
nitric acid, strong nitric being defined as >98 weight percent
(wt%). Operating procedures, compositional control, and
design of the system are all necessary to ensure a safe process.
Typically, 80 to 90 wt% sulfuric acid and 65 to 80 wt% nitric
acid are fed to a packed column to provide condensed strong
nitric acid overhead and 60 to 70 wt% sulfuric acid as a bottoms
stream. The mass balance of constituents across the column
during normal, startup, and shutdown conditions must be
controlled to allow for the proper operation of the system. This
paper will describe the operational and design changes made
to a NAC/SAC system after an event in which the strong nitric
acid condenser failed. The authors conclude that the condenser
failure, which occurred after unusual operating conditions, was
most likely caused by rapid gas formation, resulting from the
fast decomposition of a thermally unstable condensed phase
that had collected in low spots in the NAC overheads system,
which led to over-pressurization of the condenser.

Introduction
The Innovative Plastics business unit of SABIC operates a nitration
facility in Mt. Vernon, Indiana. The facility nitrates an N-alkylphthalimide
(1) derivative in 99 wt% nitric acid to afford primarily an isomeric mixture
of nitro-N-alkylphthalimides (2), used in the manufacture of an engineering
thermoplastic (equation 1) (1). A by-product of the nitration reaction is

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 tetranitromethane (3, abbreviated as TNM, bpt 126 °C). The majority of the
strong nitric acid (>97 wt% HNO3) is removed from the product by evaporation
under slightly reduced pressure, and condensed.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

The recovered nitric acid contains the majority of the TNM produced in the
process. The recovered strong acid is combined with weak nitric acid process
streams and clean commercial grade 69 wt% nitric acid to afford ~75 wt% nitric
acid (2). This combined weak nitric acid, stored in the ‘weak nitric acid tank’,
must be concentrated back to >99 wt% HNO3 for the efficient nitration of the
N-alkylphthalimide.
Production of 99 wt% nitric acid is accomplished by continuously feeding the
~75 wt% HNO3 from the weak nitric acid tank to a packed column (called a nitric
acid concentrator, NAC) along with ~85 wt% sulfuric acid (balance being water)
under slight vacuum. Strong nitric acid (99 wt% HNO3) is taken overhead of the
column and condensed in a water cooled condenser, while weak sulfuric acid (~65
wt%) is taken off the bottom of the NAC.
The exact concentrations of the NAC feed can vary. The sulfuric acid can be
80 to 90 wt%, and the nitric acid can be 65 to 80 wt%. Approximate concentrations
of acid streams in wt% are used throughout this paper, as the exact strengths are
not critical aspects of the discussion and conclusions.
While shutting down the nitration facility, the overheads condenser on the
nitric acid concentrator over-pressurized and failed. Fortunately, there were
no injuries resulting from the failure. The subject of this paper explains (i) the
investigation of the of the condenser failure, (ii) how the most likely cause of the
failure was determined, and (iii) the redesign of the condenser.

Brief Description of the NAC Equipment and Operation

A simplified process flow diagram of the NAC is shown in Figure 1 (3). It
is a glass column packed with structured glass packing. The ~85 wt% H2SO4 and
~70 wt% HNO3 are continuously fed to the NAC. The sulfuric acid’s strong affinity
for water effectively ‘breaks’ the maximum boiling azeotrope of HNO3/water (~68
wt% nitric acid), allowing for the distillation of >99 wt% nitric acid overhead of
the NAC and the collection of weak sulfuric acid (~65 wt%) off the bottom of the
186
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 NAC. A steam-heated reboiler provides heat to the NAC, and the 99 wt% HNO3
is taken overhead at about 660 mm pressure (b.pt. 80 to 81 °C). The 99 wt% nitric
acid is condensed in a horizontal condenser, and collected in a storage tank. The
two other significant species present in the overheads product are TNM (~0.5 wt%)
and dissolved oxides of nitrogen (NOx, 3-5 wt%). Some of the condensed 99 wt%
HNO3 is returned to the top of the column, through a flow control (FC) valve,
which controls the composition of the NAC overheads product and the overall
production efficiency of the NAC.
The material exiting the bottom of the NAC is primarily composed of ~65%
sulfuric acid containing some oxides of nitrogen, nitric acid, and nitrosyl sulfuric
acid. This bottom product of the NAC enters a glass denitrification tower (DEN),
where it is contacted with live steam. The steam hydrolyzes the nitrosyl sulfuric
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

acid to nitrous acid (HNO2), H2SO4, and NOx, some of which is oxidized to nitric
acid by the air present in the process. The nitric acid vapor, along with the oxides
of nitrogen present in the DEN, is ‘chased’ by air that is also fed to the DEN
back into the body of the NAC, and eventually overhead of the NAC. Finally,
the temperature of the vapor exiting the horizontal condenser to the vent system
is controlled by the flow rate of cooling water (CWS) through the U-tube bundle
inside the condenser. The ~65 wt% sulfuric acid from the bottom of the DEN is
concentrated back to 80 to 90 wt% sulfuric acid by simple evaporation of water.
The ~85 wt% sulfuric acid is eventually recirculated back to the NAC.

Figure 1. Simplified Schematic of the NAC.

187
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 A simplified schematic of the NAC overhead condenser is shown in Figure 2.
The overhead vapor line from the top of the NAC enters the top of the condenser.
The condenser is constructed with an internal U-tube bundle which is supported by
a plurality of evenly spaced Teflon® baffles. The metallurgy of construction of the
condenser was compatible with the chemical compositions that were potentially
present in the system.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

Figure 2. Simplified Schematic of the NAC Condenser. (see color insert)

One half of each Teflon® support is cut out to afford an open space. The
open spaces are alternated to the other side for each successive support. This
construction forces the vapor off the NAC to contact the tubes evenly throughout
the length of the condenser, increasing the heat transfer efficiency. The bottom of
each support is trimmed to allow the drainage of the condensate off the bottom of
the condenser.

Failure of the NAC Condenser and Damage Assessment

During a shutdown of the nitration facility, the NAC was running in normal
mode. The 99 wt% nitric acid and N-alkylphthalimide that feeds to the nitrators
had been ceased and the nitric acid was being recovered from the nitrated imide
product and stored in the weak nitric acid tank. For brevity, only the salient events
leading up to the condenser failure are detailed below.
At 9:45 am the nitrators were empty and the NAC was being fed ~70 wt%
HNO3 (from the weak nitric acid tank) and ~80 wt% H2SO4.
188
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 At 11:09 am, the nitric acid feed to the NAC was intentionally shut off as was
the steam to the NAC reboiler; the ~80 wt% sulfuric acid feed to the NAC and the
live steam to the DEN were continued.
At 11:10 am, the flow of condensate off of the NAC condenser to the nitric
acid storage tank stopped, but the flow from the condenser to the top of the NAC
continued. The NAC was said to be “on total reflux”.
From 11:10 am to 5:41 pm, the ~80 wt% sulfuric acid feed to the NAC and
the live steam to the DEN were continued. This condition produced ~65 wt%
sulfuric acid as the bottoms product of the DEN. The ~65 wt% sulfuric acid was
concentrated back to ~80 wt% sulfuric acid in the sulfuric acid concentrator. The
~80 wt% sulfuric acid was then returned to the NAC. The NAC was on total reflux
during this entire timeframe. The temperature at the top of the NAC at 11:00 am
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

was ~80 °C, 92 °C at 1200 pm, and ~114 °C at 1:00 pm. The top of the NAC
remained at ~114 °C up to the time of the event.
At 5:42 pm, the sulfuric acid feed to the NAC tripped off (ceased) due to a high
pressure interlock at the top of the column. Some steam flow was automatically
reestablished to the NAC reboiler. The live steam feed to the DEN had tripped
off. The temperature of the material in the reboiler was 128 °C at 5:42 pm. The
reflux return flow indication to the top of the column became erratic, most likely
indicating no flow. Finally, the temperature at the top of the NAC was 114 °C, and
the vapor off of the condenser was 52 °C.
At 6:02 pm the NAC overheads condenser exploded. At that time, the
temperature in the NAC reboiler had climbed to 153 °C because the steam to the
reboiler had been automatically reestablished at 5:42 pm, and the top of the NAC
was 114 °C. No personnel were in close proximity to the equipment damaged,
and there were no injuries.
The body of the condenser fragmented and the condenser end plates were
found intact 10 feet from their original location. The condenser tubes were crushed
and fragmented. The tubes near the tube sheet were fragmented. The tube sheet
is the plate where the ends of the tubes are secured (rolled to prevent leakage of
coolant into the process), and this plate is bolted to the CWS (cooling water supply)
and CWR (cooling water return) side of the condenser (see Figure 2). A one foot
diameter hole was found in the concrete floor directly below the tube sheet of the
condenser. The top of the NAC was destroyed and the DEN was cracked. The
NAC reflux return line was completely fragmented but the valve bodies in the line
were found intact. Failure mode experts from a consulting firm (4) found that
the shrapnel from the metals of construction failed in a ductile manner indicating
a rapid deflagration or low order detonation, as opposed to a brittle failure that
would be more consistent with a high-order detonation.
The survey of the damage suggested that the event took place in the condenser
and most probably in the reflux return line as well. The consulting firm determined
that the extent of the damage resulted from a low-order detonation which was the
equivalent of 2 to 4 pounds of trinitrotoluene (TNT) (5). The consulting firm also
estimated that the condenser could withstand a detonation equivalent of <2 pounds
of TNT, and that ~1200 psig was necessary to cause the shell of the condenser to
fail. The failure appeared to be the result of a rapid decomposition of a condensed
phase forming high pressure gas.

189
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The consulting firm also determined by a series of worst case heat of reaction
calculations, that the damage could not have been the result of a reaction of
hydrogen with air, or carbon monoxide with air. Further, it was concluded that
the failure was not the result of water leaking into the condenser, mixing with the
nitric acid present, and generating heat leading to the rapid vaporization (pressure
build) of nitric acid and or water.

Determination of the Probable Cause(s) of

the Condenser Failure
Hundreds of samples of process streams and debris were taken immediately
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

after the failure (6). Only the relevant findings are presented. The nitric acid
feeding the NAC on the day of the event was 80.6 wt% HNO3, 0.4 wt% TNM,
200 ppm of the nitro-N-alkylphthalimide (2), 4 ppm of nitrophthalic acid (4), 18
ppm of N-alkylphthalimide (1), 1000 ppm of nitroform (5, trinitromethane, bp at
12 mm 50 °C, mp 26.5 °C, freely soluble in water), 2.4 wt% oxides of nitrogen,
and 5 ppm calcium.

The sulfuric acid feeding the NAC (analyzed just after the failure) was 80
wt% sulfuric acid, 540 ppm of nitrophthalic acid, 3100 ppm methylamine, 80 ppm
TNM, 38 ppm nitroform, 25 ppm of nitro-N-alkylphthalimide, 3.7 wt% oxides
of nitrogen (some of which was in the form of nitrosylsulfuric acid (6) and nitric
acid), 100 ppm calcium, and about 300 ppm of other dissolved metals (primarily
chromium, iron, sodium, nickel, magnesium and potassium).

190
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Calcium sulfate was found partially fouling the packing of the NAC. Trace
amounts of nitroform and TNM were found in the remnants of the NAC reflux
return line.
Residual material was collected from the debris field. Remnants of the
overheads line, the condenser, and the reflux return line were swabbed to collect
residual chemical species that may have been present in the overheads system
before the failure (see footnote 6). Trace amounts of nitroform, tetranitromethane,
methylamine (as an acid salt), dinitrotoluene, compound 2, and nitrophthalic
acid were found in all three places. The amount of these species present in these
locations before the incident cannot be determined.
No evidence of corrosion was found among the remnants of the failed
condenser and associated piping. Hydrogen can be formed by the corrosion
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

of metal with dilute nitric acid and sulfuric acid (7). The absence of corrosion
suggested that hydrogen was not involved in the event.
No foreign organic material was found in any process stream after the event.
No unexpected metal contamination was found in the debris or process streams.
The damage pattern, knowledge of the process, and the ruling out of a gas-
phase explosion suggested that a condensed phase, present in the condenser and/
or the reflux return line, had rapidly decomposed generating high pressure gas
that then over-pressured the condenser. The investigation proceeded with the
hypothesis that the condensed phase was composed of an organic fuel and an
oxidant. The oxidizing and oxidizable constituents could have been present in
the same chemical compound. Some form of energy (e.g., thermal energy, even
ambient heat, friction, electrical discharge) was necessary to initiate the rapid
decomposition of a mixture of oxidant and oxidizer. We could not rule out friction
(from a valve, or moving fluid or gas) as the source of the initiator. The only
plausible source of an electrical discharge was the flow meter in the return reflux
line to the top of the NAC, but this meter was completely destroyed and could not
be examined.
The oxidants present in the NAC, the condenser, and the reflux return line
from the condenser to the top of the NAC just prior to the event were nitric
acid, oxides of nitrogen, oxygen, TNM, nitroform, and nitrosylsulfuric acid. The
fuels present were TNM, nitroform, trace amounts of nitro-phthalic acid and
N-alkylphthalimide, and methylamine. TNM and nitroform are compounds that
contain both fuel (carbon) and oxidant (nitro group) but they are more effective at
being an oxidant than a fuel. The cooling water fed temperature to the condenser
on the day of the event was 18 °C. The boiling point of N2O4 is ~20 °C.
One of the guiding principles when designing a nitration facility is that low
points and pockets are to be avoided, as these are places where unwanted material
can collect, stagnate, and subsequently and uncontrollably react. The horizontal
condenser was designed so that the condensate gravity-flowed to the drain of the
condenser (see Figure 2). It could not be verified, post failure, if condensate was
draining properly, since the condenser was destroyed. We also considered the
possibility that the condenser was fouled with calcium sulfate (the presence of
calcium resulting from a cooling water leak into the system), wherein a dam of
sorts formed and then allowed the pooling of condensed material in the bottom
of the condenser. There was no evidence found in the debris that suggested a

191
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 significant amount of calcium sulfate being present in the condenser, nor were we
able to determine if the condenser had been sloped properly to allow drainage.
A low point in the reflux return line was present, allowing for the collection and
stagnation of liquid, should the reflux return flow cease for any reason (see Figure
2).
An Aspen® Dynamics model of the NAC was developed and used to simulate
the NAC column during the time period starting when the column was placed
on total reflux (8). Physical properties for various constituents needed for the
model were gleaned from the literature, estimated, or determined experimentally.
Historical compositions of pertinent process streams were used in the model.
The electronically captured plant trend data for process stream flow rates, and
process temperatures and pressures were used in the Aspen model. The model
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

then provided calculated values for the temperature in the top of the NAC, the
DEN bottoms temperature and pressure, and the estimated composition of the
NAC distillate and NAC bottoms product as a function of time.
The model predicted that the majority of the TNM, nitroform, and oxides of
nitrogen (e.g., NO2, HNO3) present in the sulfuric acid feed to the NAC over the
timeframe of 11:09 am to 6:02 pm was stripped off overhead and was condensed
in the condenser. This condensate was returned to the top of the column and
essentially recirculated from the top of the column to the reflux condenser and
then back to the head of the column via the reflux return line. Some water was also
distilled overhead. As indicated previously, during this timeframe, only the 86%
sulfuric acid was being fed to the NAC, live steam was feeding the DEN, and the
column was on total reflux. Further, the model predicted that the material collected
overhead (the vapor and the condensate present in the condenser and reflux return
line) was 60 wt% HNO3 and contained at least two pounds of TNM with at least
one pound as a second phase, and at least 0.8 pounds of nitroform. The estimated
amount of phase-separated TNM was based on the known solubility limit of TNM
in weak nitric acid (see note 1). It is known that the rate TNM hydrolyzes to
nitroform in weak nitric acid is function of temperature (equation 2, footnote 1).

It was estimated, based on the dimensions of the reflux return line and reflux
condenser, that there were ~80 pounds of material present in the return line and as
much as 70 pounds of material present in the condenser, depending on the amount
of fouling in the condenser.
The amount of TNM and nitroform in the sulfuric acid feed to the NAC
had been determined before the condenser failure, and in mass balance studies
performed after the event. This led us to estimate that 6 to 12 pounds of TNM and
nitroform had been distilled out of the sulfuric acid feed from 11:09 am to 5:42
pm and was present in both the condenser and reflux return line at the time of the
192
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 event. The Aspen model also showed that the temperature of the distillate leaving
the NAC was 120 °C, in fair agreement with the actual observed temperature (114
°C).
Both the Aspen model and lab studies showed only a trace of other organics
(namely methylamine as a sulfate or nitrate salt, N-alkylphthalimide (1), nitro-N-
alkylphthalimide (2), nitrophthalic acid (4), each <50 ppm) were present in the
overheads at the time of the event.
We did not rule out the presence of significant amounts of NO2 and N2O4
(bp 20 °C) being present in the condenser and reflux return line at the time of the
failure. Recall that the cooling water on the day of the vent was running at 18 °C. It
is known that mixtures of organic material and N2O4 can be extremely hazardous
(9).
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

It is known that methylammonium nitrate is an explosive material (10) and we

decided to verify that it would not come overhead of the NAC. A lab distillation
of 80 wt% sulfuric acid containing 0.4 wt% methylamine at atmospheric pressure
with a pot temperature of 110 °C produced a distillate that contained <2 ppm of
methylamine (present as the sulfate salt). In all likelihood, the methylammonium
sulfate found overhead was the result of mechanical entrainment since the vapor
pressure of methylammonium sulfate is negligible.
Data collected at this juncture in the investigation led to the hypothesis that
a condensed phase in both the condenser and reflux return line was most likely
responsible for the condenser failure. Furthermore, it was surmised that the
condensed phase was most likely composed of 60 wt% nitric acid containing
oxides of nitrogen, considerable amounts of TNM and nitroform, and other trace
organic materials known to be present in the process.

Thermal Stability of Probable Condensed Phase Compositions

Responsible for the Condenser Failure
There was a possibility that TNM had phase separated from the condensed
phase present in both the condenser and reflux return line at the time of the event.
It is known that mixtures of TNM and organics can be powerful explosives (11).
It is also know that TNM can extract organics from weak nitric acid media. It
was thought that trace organic compounds and nitroform were also likely present
in the condensed phase present in both the condenser and reflux return line. With
all this in mind, a series of compositions were studied in the Advanced Reactive
System Screening Tool (ARSST), developed by Fauske and Associates (12).
This system is capable of heating a small amount of material (~1.0 gram) with
magnetic stirring in an open 10 mL cell that is contained in a 450 mL pressure
vessel equipped with a rupture disc (900 psig), Figure 3. The ARSST is an
Accelerated Rate Calorimeter (ARC) of sorts, where heat of reaction is not lost
to the test apparatus (13). The cell is insulated and heated at a controlled rate.
The software can also control the heating of the cell in either a heat-wait-search
mode or an isothermal operation. The open cell is equipped with a temperature
thermocouple and pressure transducer (not shown in Figure 3). The sample

193
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 container does not absorb a significant amount of reaction energy, leading to a
more accurate measurement of the adiabatic rate of rise of temperature.
The compositions tested in the ARSST are shown in Table 1. The exact
composition of the material present in the condenser and reflux return line will
never be known, but we were confident that we would learn something from the
thermal stability testing of the chosen compositions.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

Figure 3. Fauske and Associates ARSST.

The mixtures were heated at 2.0 °C/min. The heater power would decrease
when the sample began to self-heat in an attempt to maintain the desired heating
rate. The sample was kept to less than 1.5 grams to accommodate the estimated
peak pressure, should a detonation occur in the 450 mL high-pressure containment
vessel. All tests started at ambient pressure (0 psig). Both temperature and
pressure in the cell, containing the material, was measured over time. The
results of the first experiment are shown in Graph 1. The mixture of TNM and
194
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 organics was not reactive up to 160 °C. Almost identical results were obtained
in experiment 2 of Table 1 (graph not shown). The temperature oscillations
observed in Graph 1 were due to sub-optimal default parameters used to control
the heating of the sample. The parameters used were meant for a sample size of
10 grams. These parameters were adjusted in all subsequent experiments which
used a sample of ~1 gram.

Table 1. Composition Tested in the ARSST (1)

Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

Scheme 1. Experiment #1 (Table 1) ARSST: TNM in the Presence of Organic

Material. (see color insert)

195
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Thermal testing of a mixture of 79.2 wt% TNM (bp 126 °C) and 20.8 wt%
of nitro-N-alkylphthalimide (2) commenced at ~20 °C, experiment 3 of Table 1.
The material was heated to 115 °C and held at that temperature for 30 minutes.
No exotherm was observed during that time (Graph 2). The material was then
heated to 150 °C, where upon a mild exotherm at ~140 °C was observed, although
no rapid reaction occurred below 150 °C. It was observed that little pressure built
during the study.
The ARSST results for the thermal stability testing of pure nitroform (5) are
shown in Graph 3 (experiment 4 of Table 1). The onset of self-heating occurred
at 115 °C, resulting in a rapid temperature rise to 254 °C and a pressure rise of
17 psig. Recall that the cell in the reactor was open. It was apparent that the
material rapidly decomposed at 115 °C. This thermal stability data for nitroform
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

was not available in the open literature to our knowledge, although it was known
that nitroform will rapidly decompose.
Jim Dodgen provided technical information showing that nitroform is an
explosive in the pure state, and that a mixture of nitroform/water (90%/10%) was
detonable with a 50 gram charge of tetryl (N-methyl-N,2,4,6-tetranitroaniline) in
a one-inch pipe card gap test (14). It has also been reported that nitroform will
rapidly decompose upon distillation (15).
The ARSST behavior observed for a mixture of nitroform and organics
(experiment 5, Table 1) was similar to that of pure nitroform, however the peak
temperature (183 °C) and peak pressure (5 psig) were lower (Graph 4).

Scheme 2. Experiment #3 (Table 1) ARSST: TNM in the Presence of Compound

2. (see color insert)

196
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

Scheme 3. Experiment #4 (Table 1) ARSST: Pure Nitroform (Compound 5). (see

color insert)

Scheme 4. Experiment #3 (Table 1) ARSST: Nitrofrom (Compound 5) in the

Presence of Organic. (see color insert)

197
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Most Likely Cause of the Condenser Failure
From 11:09 am to 5:42 pm, the NAC column was fed only ~80% sulfuric
acid, containing low amounts of TNM and nitroform, while the column was on
total reflux. During this timeframe both the TNM (3) and nitroform (5) present
in the ~80 wt% sulfuric acid feed were stripped overhead and inventoried in the
condensed phase present in both the condenser and reflux return line. The TNM
was hydrolyzing to nitroform during this time frame since the overhead nitric acid
strength was low. The primary constituent in the overheads system was ~60 wt%
nitric acid. The condenser was cooling the overheads material during this time
frame.
At 5:42 pm on the day of the event, the reflux return flow stopped. As a
result, the condensed material in the condenser and reflux return line was no
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

longer being cooled by the water-cooled tubes in the condenser. The primary
two organics in the condensed phase were TNM and nitroform. TNM may have
phase-separated from solution. The temperature at the top of the NAC was 114
°C. The quiescent material in the overheads system self-heated through a slow
exothermic degradative process reaching ~115 °C: this is the point at which the
condensed phase rapidly decomposed to form a high pressure gas.
It is not clear whether the heat at the top of the NAC (the temperature indicator
in the NAC overheads line showed 114 °C at the time of the event) or the rapid rise
in the temperature of the reboiler at 5:42 pm (from 128 to 154 °C) contributed to
the heating of the quiescent material in the overheads system. The rapid formation
of the high-pressure gas caused the condenser to fail. The rapid decomposition of
the condensed phase, however, was classified as a low order detonation.
The ARSST results pointed to condensed nitroform as the primary material
being responsible for the condenser and reflux return line failure. Lab studies,
plant analyses and Aspen modeling all suggested that other organics (compounds
1, 2, 4, methylamine or an organic unknown) were not present in the overheads
system in significant amounts at the time of the event. This inferred that a mixture
of N2O4 and organics (other than TNM and nitroform) was not responsible for the
event.

Redesign of the NAC Condenser System

After considering numerous potential causes of the condenser failure, an
overheads system was designed that had no low spots where an unwanted
condensed phase composition could collect, Figure 4. The new vertical condenser
is a tube (process side) and shell (cooling side) design, equipped with a spray of
99% nitric acid to insure no material collects in the condenser.
The flow of condensate return to the top of the NAC is controlled with a flow
control valve and flow indication element. Also, the return line has no low spots
where material can collect should flow cease for any reason (it will self-drain to
the 99% nitric acid storage tank). The valve that controls the reflux return cannot
completely close by design. This design ensures that material will drain from the
reflux line to a safe location in the event of a process interruption. The condensate
flow from the bottom of the condenser is split between the reflux return line and
198
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 the nitric acid collection tank. This configuration limits the amount of unwanted
material, should it be present for any reason, that could potentially collect in both
the reflux return line and the bottom of the condenser. The new process was
subjected to a LOPA (Level Of Protection Analysis) and all the interlocks around
the system were reviewed and documented. The documentation included a full
narrative of the reasons for each interlock.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

Figure 4. New Vertical Condenser.

Summary
The NAC condenser failure was most likely caused by feeding ~80% sulfuric
acid to the NAC over a course of ~6 hours over which time both tetranitromethane
and nitroform were efficiently stripped from the feed, thereby allowing the
tetranitromethane and nitroform to collect and concentrate in the overhead
condenser and reflux return line as a condensed phase, while the column was
on total reflux. Some of TNM hydrolyzed to nitroform, a thermally unstable
compound at 115 °C, over this time frame. The material in the overheads
system lost cooling, likely self-heated, and then rapidly decomposed to form
high-pressure gas that failed the condenser.
The damage pattern and metallurgical analysis was not consistent with a gas
phase ignition leading to the condenser failure.

199
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The overheads system was redesigned with no low points where unwanted
potentially hazardous material could collect and become problematic. The actual
composition of the condensed phase which caused the failure will never be known.
The redesign of the new system that provides an overheads configuration that
prevents unsafe conditions considered the most credible causes for the explosion,
as well as many other mechanisms that involve concentration of reactive species
in the condenser or reflux return lines.

Experimental
Nitroform was synthesized by the method of Geckler (16).
TNM and 4-nitrophthalic acid were purchased from Aldrich Chemical and
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

used as received.
The methods used to analyze the sulfuric acid and nitric acid process streams
are documented in an accompanying contribution authored by R.R. Odle.
The method of using the ARSST is described in reference (11).

Acknowledgments
Aditya Kumar, Ganesh S, Karthik Venkataraman, and Ramesh Narayan
performed the ASPEN modeling. Jim Burelbach and Hans Fauske provided
guidance running the ARSST and interpreting the data. Jim Dodgen provided
technical information on reactive materials. Chet Grelecki, David Ross, Robert
Trebilcock, Michael Wenkoff, Colin Evans, Tracy Zhang, and Gary Davis offered
insight into the cause of the event. David Mongilio and Colin Evans contributed
to the redesign of the NAC condenser. Roy Odle, Mark DeLong, Karla Steele,
Roger Hurst, Jessica Jarman, David Zoller, Eugene Galperin, Jim Carnahan,
and Derek Lake contributed in the analysis of process samples and debris. The
authors participated in the analyses of debris and process streams, mass balances
of TNM and nitroform in the plant, thermal stability testing, condenser redesign,
and plant data collection.

References
1. Aspects of this process have been previously published, see Guggenheim, T.
L.; Evans, C. M.; Odle, R. R.; Fukuyama, S. M.; Warner, G. L. Removal and
Destruction of Tetranitromethane from Nitric Acid. In Nitration: Recent
Laboratory and Industrial Developments; Albright, L. F., Carr, R. V. C.,
Schmitt, R. J., Eds.; ACS Symposium Series 623; American Chemical
Society: Washington, DC, 1996; pp 187−200.
2. Several terms are used throughout this chapter. ‘Weak nitric acid’ means
nitric acid that is 30 to 75 wt% nitric acid where the balance of material is
water. ‘Strong nitric acid’ means >97 wt% nitric acid (containing 3 wt%
water). A process stream that is said to be 99% nitric acid is taken to mean
that the nitric acid present contains 1% water and ignores other substituents
200
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 in the 99% nitric acid, such as dissolved oxides of nitrogen (NO, NO2, and
higher oxides of nitrogen), TNM, and trace organic material.
3. For competitive reasons, the exact materials of construction (MOC) are not
disclosed. It was found that the MOC did not play a role in the condenser
failure.
4. Exponent® Engineering and Scientific Consulting was contracted to
determine the amount of energy released responsible for the overall damage
to plant and equipment.
5. For discussions on correlating energy released with extent of damage to plant
and equipment and the surrounding environment, see (a) Wharton, R. K.;
Formby, S. A.; Merrifield, R. J. Hazard. Mater. 2000, A79, 31−33. (b)
Mendonca-Filho, L. G.; Bastos-Netto, D.; Guirardello, R. J. Hazard. Mater.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

2008, 158, 599−604. (c) McIntyre, D. R.; Ford, E. Process Saf. Prog. 2009,
28 (3), 250. (d) Rodriguez, E. A.; Duffey, T. A. Weld. Res. Counc. Bull.
2000, 477, 31−60. (e) Tardif, H. P.; Sterling, T. S. J. Forensic Sci. 1967, 12
(2), 247−272.
6. Odle, R. R.; Guggenheim, T. L.; DeLong, L. M. Solubility, Equilibrium,
Behavior, and Analytical Characterization of Tetranitromethane,
Trinitromethane, Methyl Amine, and Ammonia in a Nitration Facility.
In Chemistry, Process Design, and Safety for the Nitration Industry;
Guggenheim, T. L., Ed.; ACS Symposium Series 1155; American Chemical
Society: Washington, DC, 2013; Chapter 14.
7. (a) Tousek, J. Collect. Czech. Chem. Commun. 1977, 42 (12), 3367−3374.
(b) Chilton, T. H. Strong Water: Nitric Acid: Sources, Methods of
Manufacture, and Uses); MIT Press: Cambridge, MA, 1968, Library of
Congress Card Number 67-16496.
8. The modeling was done by Aditya Kumar and Ganesh S. at the John Welch
Technology Center in Bangalore, and Karthik Venkataraman and Ramesh
Narayan of GE Plastics.
9. For a brief review of the hazards associated with N2O4 and organics see,
Bretherick’s Handbook of Reactive Chemical Hazards, 6th ed., Volume 1;
Urben, P. G., Ed.; Butterworth-Heinemann, Ltd.: Waltham, MA, 1999; pp
1792−1797.
10. Kurniadi, W.; Brower, K. R. J. Org. Chem. 1994, 59, 5502–5505.
11. Hager, K. F. Ind. Eng. Chem. 1949 41 (10), 2168. Detonable mixtures
of tetranitromethane and organic materials were known as Panclastite, see
Manheimer, V. Meml. Artillerie Fr., Sci. Tech. Armement 1954, 28, 505−516.
Also see Kaye, S. M. Encyclopedia of Explosives and Related Items, PATR
2700; 1978, p M83, and Kristoff, F. T.; Griffith, M. L.; Bolleter, W. T. J.
Hazard. Mater. 1983, 7, 199−210.
12. James Burelbach of Fauske and Associates assisted in performing the
experiments. See Burelbach, J. P. Advanced Reactive System Screening
Tool (ARSST); North American Thermal Analysis Society, 28th Annual
Conference, Orlando, FL, October 4−6, 2000. Also see Burelbach, J. P.
Quick Hazard Screening by Closed Cell, ARSST Using Standard ARCTM
Bombs; Asia Pacific Symposium on Safety (APSS), Osaka, Japan, October,
2009.

201
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 13. For a discussion of an ARC, see Townsend, D. I.; Tou, J. C. Thermochim.
Acta 1980, 37 (1), 1–30.
14. For a description of the card gap test, see Engineering Design Handbook,
Explosives Series, Explosive Trains; AMCP 706-179, U.S. Army Materiel
Command, 1974, section 12, p 10, and Macek, A. Chem. Rev. 1962, 62,
41−63.
15. Stull, D. R. Fundamentals of Fire and Explosion, AICHE Monograph Series
No. 10, 1977. Also see Marans, N. S. J. Am. Chem. Soc. 1950, 72, 5329.
16. Brown, L. H.; Geckler, R. D. Research in Nitropolymers and Their
Application to Solid Smokeless Propellants; Aerojet Engineering Corp.,
Quarterly Summary Report 371, April 15, 1949.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch013

202
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 14

Solubility, Equilibrium, Behavior,

and Analytical Characterization of
Tetranitromethane, Trinitromethane,
Methyl Amine, and Ammonia
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

in a Nitration Facility
Roy R. Odle,* Thomas L. Guggenheim, and L. Mark DeLong

SABIC Innovative Plastics, 1 Lexan Lane, Mt. Vernon, Indiana 47620

*E-mail: [email protected]

The safe operation of a nitration facility is very important. This

chapter focuses on a framework for operating a nitration facility
to inform the reader of some important safety considerations.
Some detailed chemistry relating to tetranitromethane (TNM)
and nitroform (NF) are discussed. More particularly, TNM
can be a by-product in the nitration of aromatics. TNM and
nitroform, which can be formed upon hydrolysis of TNM, can
form potentially explosive mixtures with organic compounds.
Furthermore, isolation of TNM and NF as separate phases is to
be avoided because they are thermally unstable. The solubility
of NF and TNM in nitric acid and the equilibrium between TNM
and NF are dependent on nitric acid strength. In the nitration
process described, methylamine and ammonia are produced
as by-products. The amine impurities, present as ammonium
nitrate and ammonium sulfate salts, are extremely difficult to
scrub and sequester. Of particular interest is the propensity
of methylammonium nitrate to form a solid in one section of
the vent system. The behavior of these various by-products,
both how they are formed and how they are eliminated, and
an overall mass balance are discussed. Additionally, the
analytical methodologies used to determine the mass balances
of by-products are described.

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Introduction
The authors assume that the reader is very aware that process, product, and
chemistry considerations in a nitration facility are of the utmost importance.
If the reader does not know this, hopefully, this chapter will help the reader
understand the importance of these considerations. However, a chapter in a book
is no substitute for many years of experience, seasoned professionals, proper
evaluation of hazards, knowledge of nitration chemistry, testing of hazards, and
many other aspects and capabilities necessary to reduce the risk of running a
nitration facility. The reader is encouraged to seek expert resources if they have
questions. This is not a task for the ill-informed!
SABIC’s Innovative Plastics division currently runs a nitration plant that uses
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

an “all-nitric” process. This means that the nitration of the substrate is run in only
nitric acid (+99%) with no other acids, solvents or reagents present. The all-nitric
process has been used in the past in laboratories and in manufacturing facilities.
However, there are reports that this type of nitration is unstable. This instability has
to do with the oxidative stability of certain species in the reaction. The oxidative
stability of the products and starting materials (used in the SABIC plant) makes
the all nitric process feasible for commercial production. This all-nitric process
has been practiced safely for well over 20 years. During this time, only one major
incident occurred where plant production was disrupted by operational anomalies.
The event prompted the identification of better design features of equipment to
avoid such disruptions (1). As with all nitration reactions, the yield of the desired
products of reaction is not perfect. There are some by-products that need to be
managed in the plant streams.
This chapter provides some insight into how the authors used mass balance
techniques to develop their understanding of the distribution of unwanted
by-products in the process. The chapter also discusses how the critical analyses
that were implemented are used and maintained.
The thermal stability of the desired nitrated product must also be well
understood. Nitration by-products, NOx, nitric acid and other species are known
to lower the on-set temperature of decomposition of nitrated products, and can
lead to a runaway reaction (2). Runaways are not the main subject of this chapter,
but one must recognize the common signs of a runaway. Those signs are an
unexplained temperature rise in the reaction mass, formation of bubbles and
gas, generation of NOx, mixed acid nitration reactions wherein excess nitric
acid was inadvertently added to the nitration reactor, a reaction mass color
change, or failure to observe the expected temperature rise that should accompany
the nitration of the substrate. This list is not comprehensive so the reader is
encouraged to study other causes on runaway reactions.
A runaway event can be rapid and have catastrophic consequences. The
thermal stability of potential mixtures present in a nitration process must be
evaluated to properly design and operate the facility.
The organization responsible for the operation of the plant applies a regular
review process to prompt self-assessment of all of the chemical analysis methods,
frequencies of those analysis, critical sample points, operating conditions, and
process procedures.

204
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Simple Goals for Understanding Plant Chemistry
The operation of a nitration plant is intended to produce a nitrated product.
However, it is well known that no chemical reaction is perfect and no reaction
yield is 100%. This is especially so in a nitration plant. Positional isomers can
be formed during the nitration step. Other impurities are formed from oxygen
versus nitrogen attack of the nitronium ion, or from oxidation reactions or other
types of side chemistry (3). There are three simple rules for assessing how certain
by-products (for this discussion, molecules containing C, H, N, O, and potentially
X, where X is halogen) in a plant could affect the safety of that plant: (i) location,
(ii) distribution, and (iii) exit.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

I. Location: Where Is the By-Product of Interest Formed?

This question raised by the “location” consideration takes one directly to the
heart of the plant. The reactor is usually the place where starting material contacts
the nitrating species. In most plants this place would be called the reactor. It may
be in the form of a tank, stirred tank reactor, tubular reactor or many variations.
This is the area of the plant where runaway reactions are probably the biggest
concern. This risk has probably been assessed and provided with some means to
prevent or abate such a runaway. Some of these techniques include strict control of
feed rates, strict control of mixing, elimination of hot spots in the process, quickly
draining the reactor into a quench basin, and means for aggressive temperature
control. It is not the intent of this chapter to direct the reader to favor any one
of these methods, just know that such means of control and risk abatement are
considered, designed and exist in many nitration facilities.
The desired product of the nitration is usually tracked quite well in the plant as
it has a direct effect on the efficiency of production. However, the isomers and side
products that may also be formed during the nitration are usually separated from
the main product. This separation as well as the configuration of any particular
nitration train can lead to distribution of small concentrations of different types
of by-products throughout the plant. These impurities may undergo reactions in
areas of the plant not designed for reaction.

II. Distribution
The second question raised by the “distribution” consideration is how do the
by-product compounds distribute in the plant? Other logical questions might be:
Where is the compound most concentrated? Or where is the greatest risk for the
compound to concentrate? The answers to these questions vary depending on
the nature of the compound. The physical properties of the compound must be
understood. What is its boiling point? What is its melting point? What is its vapor
pressure? What is the solubility in each of the streams in the plant?
Knowledge about the risk associated with each particular by-product must
be known. This is not always easy to determine. Some compounds may not be
reactive and are really of no concern as long as they do not contribute to an oxygen
balanced mixture somewhere in the plant. As earlier stated, some by-product
205
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 compounds will lower the onset temperature of product decomposition, which
can be rapid and catastrophic. Other compounds can be primary explosives!
Obviously one would want to know the concentration and distribution of those
kinds of risky compounds. Some compounds are not very soluble. Other
compounds are highly oxidative and therefore are dangerous if mixed with other
organic materials. And of course, in the real world, compounds can have varying
amounts of each of these attributes. The greatest risk is usually in the place where
the greatest concentrations of high risk compounds are found.

III. Exit

The third important question raised by the “exit” consideration is: How does
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

the by-product compound leave the plant?

This can happen in a couple of ways. The preferred option is for impurities
to be separated from the product and exit the plant in some sort of waste treatment
stream that is designed to capture and/or destroy those compounds. Sometimes the
molecule is transformed in the plant into a different form or species such that its
distribution must be controlled in order to remove it. There is also the possibility
that a compound can be oxidized to such an extent that it appears to be consumed
in the nitration facility. Usually this means that the organic compound is nearly
completely oxidized in the nitrating medium. An example of this would be an
aromatic ring that has reacted in an oxonium ion fashion to produce a phenolic
species. The phenolic species is quickly nitrated to form a picric acid type of
impurity. This type of molecule can be further oxidized, by oxides of nitrogen or
nitric acid, to finally give TNM, CO, CO2, and oxalic acid. There is a multitude
of intermediate products in this type of oxidation reaction. A molecule that is
“consumed” or oxidized in the nitration plant probably generates the most difficult
species on which to do a material balance.
As one can see, the goal for understanding plant chemistry becomes very
complicated as one has to understand both the chemistry and the physical nature of
the primary compounds as well as the isomers and the impurities. Even compounds
that are produced in the parts per million ranges can build up over time if there
is no way for that molecule to exit the plant. This can present a process risk if
the compound is a strong oxidizer or a primary explosive, and it concentrates
in one small area of the plant. Beware; this mechanism of concentration into a
small area has led to disastrous results through the history of nitration. Another
mechanism that increases risk is if impurities that stay with the primary product
act as sensitizers for thermal degradation.
A plant mass balance of products and by-products, with understanding of
reactions and physical properties, is required knowledge to aid in risk assessment.

Chemistry Considerations of Nitration: Where Are the

By-Products of Interest Formed?
The principles discussed in the above section will be illustrated with examples
and more discussion in this section of the chapter.
206
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 In Figure 1, the formation of the active nitrating species is shown, the
nitronium ion (4–7) .

Figure 1. Formation of Nitronium Ion.

In Figure 2, note a generic chemical equation for the nitration of an aromatic
species that is called the starting material (SM). Functional groups on this starting
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

material are designated by R groups. The method of nitration is not specified

except that nitric acid is being used as the nitrating agent. And of course water
from nitration is formed. The equation shows three different types of product
species. Two isomers are shown as a 4 isomer and a 3 isomer the yield is 91%
and 5% respectively (4-P-A, and 3-P-B). The OH-C compound represents the case
where the nitronium ion and has attacked from the oxo position to give a phenolic
compound. Once formed, it quickly nitrates two more times to give a species that
is beginning to look like picric acid. This type of reaction happens to some extent
in nearly all nitration reactions of aromatic compounds (8). In this example, it is
made at 4%.

Figure 2. Desired Product/Isomers and Typical Phenolic Formation. (see color

insert)
In other commercial nitration processes, this oxo-nitration species is usually
washed away in some basic aqueous stream or dilute acid stream and taken to
some sort of waste treatment. A lot of expense is often spent to treat and dispose
of streams containing these nitro phenolic compounds. Many times, if sulfuric
acid is used in conjunction with nitric acid is used (mixed acid nitration), these
streams also contain sulfonated impurities and are called either red water (9) or
black water.
It so happens that in the “all nitric” process example (Figure 3), heating the
phenolic impurity to 50°C in nitric acid gives nearly complete oxidation of the
compound. Nitric acid is an unusual reagent in that it can be a nitration medium
as well as an oxidative medium (10). This reaction is very useful because most of
207
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 the compounds formed from the oxidation, for example, carbon dioxide, carbon
monoxide, and oxalic acid are removed from the plant either as a gas or with
wastewater. However, one of the compounds formed by oxidation ends up being
per-nitrated to give tetranitromethane. Tetranitromethane in the presence of
water and nitric acid readily establishes equilibrium between itself and nitroform.
Nitroform can also continue to hydrolyze, or can decompose to eventually give
CO2 and NOx.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

Figure 3. Nitric Acid Oxidation and Formation of TNM. (see color insert)

The equilibrium between tetranitromethane and nitroform depends on the

strength of the nitric acid or one might think of it as the concentration of water
(11).
In this particular nitration example one of the side groups or R groups can
generate two different types of nitrogen compounds. In Figure 4, the top equation,
an oxidation reaction gives intermediate 4-P-D with pendant R′. This is followed
by hydrolysis to give a modified R′′ group (4-P-E) and ammonia (AM). In the
lower equation of Figure 4 simple hydrolysis of the R group gives methylamine
(mono-methyl-amine; MMA) and a different pendant group R′′′ (4-P-F).

Figure 4. Formation of AM and MMA from R Group Hydrolysis. (see color

insert)

The formation of these two nitrogen species was not well appreciated. Both
ammonia and MMA readily form salts with nitric acid. These salts can be unstable
or explosive when concentrated and subjected to certain conditions.
The equation in Figure 5 shows a mechanism where an aromatic compound in
a nitration reaction readily gives a certain amount of ammonia. Without belaboring
208
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 the mechanism, one observes that oxo-attack by the nitronium ion is followed by
nitrosation and oxidation to give a hydroxyl imine which decomposes in nitric
acid to eventually give ammonia (12). This reaction shows that ammonia can be
generated in a nitration facility without having any side groups or nitrogen bearing
moieties on the target starting material. Beware of ammonium nitrate salts, or
nitrite salts for that matter, building in the overhead vents of a nitration facility.
Consider the places where these salts might accumulate. This discussion in no way
can cover the myriad of possible routes that high risk compounds can be formed
and concentrated to give high risk situations.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

Figure 5. Side Reaction to Ammonia. (see color insert)

In the example shown in Figure 4 the primary source of methylamine is

from simple hydrolysis of the SM or products, and the source of ammonia is a
combination of oxidation and hydrolysis of the SM or products, as well as an
impurity in the methylamine. Considerable study has been carried out in order to
confirm those conclusions. Such answers do not come easy.

209
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 By-Products of Concern in Nitration Facilities
Every nitration reaction has its own unique chemistry as far as impurities or
side products that can be produced. The reader needs to be aware that the chemists
and engineers running a facility or piloting a new nitration must diligently sample
and analyze all process streams to understand these reactions. Again, the impact
of by-products on the thermal stability of the product must be evaluated.
There are some by-products that seem to be pervasive in the nitration
industry. The following is a partial list of those compounds; however, this is not
an exhaustive list! Variants of these compounds are likely in different process
situations. Those operating a nitration facility must be aware of these types of
compounds, Table 1.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

Table 1. High Risk Compounds, By-Products, and Mixtures Formed in

Nitration Plants
Reference
Compound or Mixtures Hazard Number

tetranitromethane strong oxidant (13)

nitroform oxidant, rapidly decomposes (14)
methylammonium nitrate
(nitrite) detonable in dry state (15)
ammonium nitrate (nitrite) detonable in presence of organics (16)
nitric esters of organic acids potentially detonable (17)
nitro-alkanes reactivity depends on oxygen balance (18, 19)
nitrated polyols high energy and sensitive (20)
mixtures of nitric acid
and organics with little oxygen balance determines
water present energy density (21)
oxygen balance determines
NOx and Organics energy density
NOx reacting with polymeric
materials in process sensitive polynitrated polymers form (22)
sensitive high energy density
polynitrated aromatics compounds present in process
excess nitric acid with organics
in mixed acid media increased risk of runaway reactions

Oxygen balance has been mentioned a few times in this chapter (21, 23). It is
simply the balance of oxidant and fuel. The more perfect the balance of oxidant
and fuel, the higher the energy that can result.

210
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 These are only a few examples of potential risk compounds and situations
in the nitration plant. There are many more that could be considered. Please,
know what you are doing and consult experts who can help you if you are new to
the nitration world. Being proactive and taking precautions is prudent and highly
advisable.

Analytical Challenges
The analytical challenges to complete a good mass balance in a nitration plant
are enormous! The matrices in which the compounds exist are corrosive, reactive,
and fuming. These present many challenges in sample preparation and analytical
method development. However, some very straightforward methods to measure
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

the compounds of interest have been developed. While these methods may not
directly transfer to all situations, they may give some direction as to techniques
that might be used to measure the by-products and compounds of interest in a
nitration plant or process.

Analyzing Tetranitromethane and Nitroform

The presence and concentrations of both tetranitromethane and nitroform
are very important so analyses of samples from the plant are needed.
Tetranitromethane is a strong oxidant as mentioned earlier and nitroform can be a
sensitizer or a detonable compound when concentrated. In one particular situation
the matrix is 40 wt% to 99 wt% nitric acid. The analytical procedure involved
the direct dilution of the sample with acetonitrile and water followed by analysis
via HPLC-PIC using a UV-Visible detector. The details will be given later in the
Experimental section. The same method can be used with either a weak or strong
sulfuric acid matrix. Note that the choice of the diluent is very important so as to
avoid highly reactive situations.

Analyzing MMA and Ammonia

MMA and ammonia are found in many streams in the example shown in
Figure 4. These species, which distribute between several process streams, can
be present in very low concentrations. The method to analyze for ammonia and
MMA is performed using ion chromatography. The initial sample matrix can be
sulfuric acid (30 wt% to 82 wt%) or nitric acid (20 wt% to 99 wt%). The presence
of the acid complicates the analysis. Figure 6 illustrates the sample prep method.
It was discovered that a sample can be diluted with water then passed through
an IC-OH sample prep cartridge column (center rectangle in Figure 6), to reduce
the amount or remove the acid anion. This is a way to remove or neutralize acid
without the introduction of a cation (like NaOH) that may cause interference
in the subsequent cation analysis. Passage of the diluted sample through the
column replaces the sulfate or nitrate with hydroxide. The sample prep column
is essentially an anion exchange resin in the hydroxide form. The column is first
filled with water, and then the diluted sample is passed through the cartridge. The
211
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 first material through the cartridge is discarded and additional sample is filtered
through the same cartridge and the filtrate collected. This filtrate is then analyzed
by Ion Chromatography.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

Figure 6. Ion Exchange Treatment in Nitric Acid Matrix.

Other Methods To Analyze Process Streams

Chromatography methods (like HPLC) are sufficient to quantify compounds

of interest. However, what if there are unknown organics in the stream? This
is a question that must be addressed while running a nitration facility. Total
organic carbon (TOC) has been successfully employed to augment and verify
chromatography (HPLC) methods. TOC works well in both nitric and sulfuric
acid streams as well as dilute water streams. The method is sensitive as long as
measures are taken to not overload the detector. This is easy to avoid by diluting
the sample with distilled water. This instrument is considered essential to keeping
a nitration plant safe.
The trustworthy methods are used daily to analyze plant streams and confirm
that nothing has changed and concentrations are of no concern. This is especially
true if something in the plant has changed. This change might have been a new
piece of equipment, use of a new lubricant or seal fluid in the system, a new
gasket in the system, or a new source of reactants or reagents. In many cases, an
extraction of plant samples is used to concentrate species for analysis by one of
the methods mentioned above. Occasionally, those same techniques of sample
preparation are used with one of the following analytical methods: GC-MS (Gas
Chromatography-Mass Spectroscopy), LC-MS (Liquid Chromatography-Mass
Spectroscopy), NMR (Nuclear Magnetic Resonance), FTIR (Fourier Transform
Infrared Spectroscopy), or ICP (Inductively Coupled Plasma). Note that ICP
is used to track metal concentrations in the plant streams. This can be an early
indication that corrosion is a problem in a certain area of the plant. A plant mass
balance of metals was not discussed but this type of mass balance is useful for
determining the aging and integrity of the plant.
212
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 To determine the behavior of compounds and by-products in the plant, all
streams need to have a method of analysis. The tough matrix is a challenge for
quick and easy analysis. However, with some thought and creativity the challenge
can be made manageable.

Mass Balance: Origin, Distribution, Level, and Exit of

Compounds of Concern
In this section a simple box diagram of a nitration plant, encompassing the
chemistry shown in Figures 2, 3, and 4, is depicted. It will be used to demonstrate
how a mass balance can show the point of origin of certain impurities, how
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

compounds distribute in the plant streams, and what the concentration levels
are in each area of the plant. In order to simplify the analysis and math, whole
sections of plant equipment have been included in a geometric shape to show the
principle of the mass balance analysis (Figures 7, 9, and 10 below).

Mass Balance of TNM and NF Around the Reaction Area

Figure 7 depicts the reaction area of the example process encompassed by

the chemistries shown in Figures 2, 3, 4, and 5. The mass balance of interest
is for TNM and NF. The concentration of TNM is shown in red letters, and the
concentration of trinitromethane (NF) is the second number in blue. The numbers
have been normalized for flow and are expressed in relative weight/time. There is
no need to do any extra math in order to understand the mass flow.

Figure 7. Mass Balance of the Reactor Area. (see color insert)

213
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Begin in the upper left-hand corner of Figure 7 with the oval labeled “Acid
Concentration”. This is the part of the plant where the nitric acid is contacted with
sulfuric acid in order to make strong nitric acid. For this discussion, strong nitric
acid has a concentration of greater than 99% nitric acid, HNO3, by weight. The
strong nitric acid is collected in a storage tank. Note the wt/time of 375, 10: TNM,
NF in the strong tank. Recall that the equilibrium of TNM and nitroform in nitric
acid is dependent on the nitric acid strength (11, 24). TNM is favored in strong
nitric acid where very little nitroform exists. This will be discussed further below
(Figure 8).
The strong nitric acid feeds the “Reactors” (Figure 7) where it contacts the
aromatic starting material (Ar-H). There is no TNM or NF in the pure starting
material. Flowing out of the reactor it the wt/time of 435, 10: TNM, NF, showing
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

an increase in the concentration of TNM. Indeed, TNM is formed in the reactor

system, as shown in Figure 3. Now this may not seem surprising to the reader,
however, it is very important in the grand scheme of things to know where the
TNM is being formed. There is no doubt that one point of origin for TNM is the
reactor system.
Now move to the upper right-hand oval of Figure 7. This represents the
“Product Isolation Area” of the plant. This area involves many different streams
and process steps to wash, purify and isolate the nitration product. The product
is essentially void of TNM or NF. Now to a chemist, there is never such a thing
as zero, but the levels are negligible and present no safety risk downstream. The
interesting observation is that the total mass flows in the acid recovery streams
account for the TNM and NF from the reactors.
Thus a point of origin for TNM is in the reactor system with a 16% increase
from the incoming mass flow. Analysis showed that in this strong nitric area TNM
is favored over NF. The product stream is an insignificant exit point for either
tetranitromethane or nitroform. Also of note is that even though the TNM and NF
were distributed in the complex product isolation area, the final feed stream to the
acid concentration system accounted well for both impurities of interest.

Equilibrium of TNM and NF and Solvolysis Behavior

As the concentration of nitric acid increases the ratio of TNM to NF also

increases, Figure 8. One may be tempted to calculate and equilibrium constant
from this data but that is not possible, as the nitroform decomposes via hydrolysis
to NOx, HNO3, nitrous acid, and presumably carbon dioxide (See Table 2). The
simple take away from Figure 8 is that nitroform is favored in weak acid and
tetranitromethane is favored in strong nitric acid.
Table 2 shows the solvolysis of tetranitromethane to nitroform in 40% nitric
acid at 90°C, and also shows the rate to equilibrium as well as the reality of the
hydrolysis in the TNM-NF system. Note that the mass balance does not close. This
is because nitroform further reacts in the system to give other hydrolysis products.
This experiment was performed at the solubility limit of TNM in 40 wt% nitric
acid.

214
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

Figure 8. Equilibrium of Tetranitromethane and Nitroform. (see color insert)

Table 2. Solvolysis of TNM in 40% Nitric Acid at 90°C (24)

Time(h) TNM (ppm by wt.) Nitroform (ppm by wt.)
0 4600 0
2,5 2600 270
21.5 700 800
43.5 275 1050

Mass Balance of TNM, NF, and MMA Around the Weak Nitric Acid
Recovery System
Figure 9 represents the weak nitric acid recovery section of the plant
encompassing the chemistry shown in Figures 2, 3, and 4. The mass balance
of interest now includes TNM, NF, and MMA. Again, the units are in relative
weight per time so that no complicated math needs to be done to interpret the
mass balance.
Start in the upper left-hand corner of Figure 9 with the oval which represents a
waste stream with organics and the wt/time of 0, 12, 0.15: TNM, NF, MMA which
215
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 is feeding into a weak nitric acid evaporator. Going overhead of the evaporator
is weak nitric acid containing: 0, 10.8, 0: TNM, NF, MMA. The evaporator
overheads eventually are fed to the nitric acid recovery unit.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

Figure 9. Mass Balance of Weak Nitric Acid Recovery. (see color insert)

Exiting the bottom of the evaporator (Figure 9) is weak nitric acid that contains
concentrated organics and 0, 0.6, 3.3: TNM, NF, MMA, which is feeding the
neutralizer on the bottom right-hand corner. If one does a little simple math, you
will deduce that MMA was formed in the nitric acid evaporator (a 21% increase).
Therefore, a point of origin for MMA is the weak nitric evaporator. This would
be an area to look for methylammonium nitrate in vent systems or other stagnant
places where it might form and collect. Methylammonium nitrate can collect as
crystalline material that is sensitive to impact.
The bottoms of the evaporator are neutralized with sodium hydroxide in the
neutralizer (Figure 9). Two process streams exit the neutralizer. The first is the vent
stream where the wt/time are 0, 0, 1.6: TNM, NF, MMA and the wt/time going
out the waste stream, which is an aqueous stream, containing 0, 0, 1.7: TNM, NF,
MMA. The conclusion around this unit operation is that nitroform is hydrolyzed
in the presence of sodium hydroxide. MMA is leaving the plant via the vent as
well as the aqueous stream. Once again, this would be a good place to make sure
that methylammonium nitrate is not being collected in the vent system.
Out of these very complex sample matrices the analytical results have proven
useful to show that MMA is formed in the evaporator system (a 21% increase), and
that nitroform is hydrolyzed and destroyed in the neutralizer, which is a sodium
hydroxide rich aqueous stream. MMA is leaving the plant in both a vent system
and a wastewater system.

216
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Mass Balance TNM, NF, and MMA Around Strong Acid Concentration

Figure 10 depicts a very complex system wherein weak nitric acid is mixed
with sulfuric acid and strong nitric acid is produced. In the upper left-hand corner
note the oval that represents a scrubbing system wherein sulfuric acid is used
to absorb nitrogen oxides or NOx. The sulfuric acid makes for very efficient
scrubbing of the NOx. First examine and consider the streams that leave the NOx
scrubbing system.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

Figure 10. Mass Balance of Acid Concentration. (see color insert)

Leaving the vent system of the scrubbers is a gas stream with the wt/time of
135, 6, 0: TNM, NF, MMA. The sulfuric acid feeds the nitric acid concentrator (the
center rectangle). The entering sulfuric acid stream has a wt/time of 12, T(trace),
23.4: TNM, NF, MMA. The complicated flows into the NOx scrubbing system
are not explained but the mass balance showed clearly that MMA was efficiently
absorbed into the sulfuric acid scrubbing system.
The bottom left-hand corner of Figure 10 has the oval labeled “Weak HNO3”.
Note that TNM is hydrolyzed to nitroform, CO2, and NOx to some extent in this
area of the plant. The complete mass balance that would let the reader draw this
conclusion is not shown, as the amount of equipment and streams is beyond the
scope of this chapter. However, take note that the wt/time weak nitric feed to the
nitric acid concentrator is 12.3, 16.8, 0: TNM, NF, MMA.
The nitric acid concentration step takes place in the orange rectangle in the
middle of the page. Here two streams leave this operation. The first stream is
strong nitric acid that moves to the left into the yellow strong nitric tank. Here the
wt/time composition is 47, 0, 0: TNM, NF, MMA. In the concentrated nitric acid,

217
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 99+ wt%, the equilibrium greatly favors TNM and NF is not detected. Also note
that MMA is absent in this stream.
The sulfuric acid stream (Figure 10) leaves the nitric acid concentrator to
the sulfuric acid concentrator process area, and contains a wt/time of 0, 0, 28.5
of TNM, NF, MMA. It is surprising that essentially neither TNM nor NF is in
the sulfuric acid stream. The equilibrium favors TNM, which is volatile and so
goes overhead with the strong nitric acid. The conclusion is that nitroform is
transformed to TNM in this concentrator. Note that all of the MMA leaves the
nitric acid concentrator with the sulfuric acid. The mass flow of MMA leaving the
nitric acid concentrator is a little more than expected. There is either an error in
the mass flow measurements or some mechanism of MMA formation is present
in the nitric acid concentrator. Indeed, there is a small concentration of organic in
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

this area that could give MMA upon hydrolysis, but this does not present a safety
risk.
Note the sulfuric acid concentrator which is the orange rectangle in the center
right of Figure 10. The concentrated sulfuric acid leaving the concentrator contains
wt/time of 0, 0, 24: TNM, NF, MMA. The concentrated sulfuric acid harbors the
MMA generated in the process. If the stream is monitored from the time of a
clean plant startup to equilibrium, it is observed that the concentration of MMA
goes from zero to a constant steady state level (~3000 ppm). Note that the MMA
present in the sulfuric acid is in the form of the methylammonium sulfate form,
not the nitrate form. The safety risks of MMA in sulfuric acid are much lower than
when MMA is present in a nitric acid stream.
The waste water stream leaving the sulfuric acid concentrator contains a wt/
time profile of 0, 0, 4.5 of TNM, NF, MMA. This stream provides an exit of MMA
from the process.
The mass balance calculations from this area of the plant (Figure 10) show
that MMA is scrubbed in the sulfuric acid scrubbers and eventually eliminated in a
sulfuric acid loop. Nitroform is effectively transformed to tetranitromethane which
stays with the strong nitric acid. The vent off the sulfuric acid NOx scrubbing unit
is sent to another NOx scrubbing unit, and this provides an exit point for TNM and
NF from the process. The technology of this final NOx scrubber is not discussed
here. However, it is the case that the TNM and NF that exit the sulfuric acid NOx
scrubbers are hydrolyzed in the final NOx scrubber. Thus Figure 10 shows the
distribution and reactions of the compounds of concern which are monitored in
this example.

Solubility of TNM and NF in Nitric Acid

Over the years the authors have gathered much information about the
solubility of tetranitromethane in various concentrations of nitric acid. They
have also observed the solubility as a function of temperature. The results are
summarized in the Table 3 (24). One should note that the solubility changes
dramatically with percent nitric acid, especially at the 60 wt% nitric acid level.
It should also be observed that temperatures between zero and 80 °C have little
effect on the solubility, Figure 11. The solubility of NF in water is quite high,
218
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 16.7 gr per 100 mL of water at 0 °C. The solubility of NF in different strengths of
nitric acid was not determined. This would be difficult because of the equilibrium
that was discussed earlier in the chapter.

Table 3. Solubility of TNM in Acid Media

Acid Composition 25 °C (wt%) 60 °C (wt%)

water 0.1 0.08

20% nitric acid 0.25 0.23
40% nitric acid 0.54 0.53
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

60% nitric acid 0.72 0.7

70% nitric acid 1.0 - 1.5 1.0 - 1.5
80% nitric acid 5.4 5.4
90% nitric acid 15.3 15.3
99% nitric acid >30 >30

98% sulfuric acid 0.8 - 1.0 2.5 - 3.1

86% sulfuric acid 0.7 - 0.8 not determined
86% sulfuric acid/90% nitric acid (2:1 by wt) 0.7 - 0.8 not determined

Figure 11. Solubility of TNM in Various Strengths (wt%) of Nitric Acid as a

Function of Temperature. (see color insert)

219
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 TNM and NF concentrations are consistent with the process knowledge of the
plant. The concentrations of TNM and NF in all process streams are well below
their solubility limits. Action levels are chosen (levels at which serious action is
taken in the plant which may include shutting the plant down) to be one third the
TNM solubility limit in the particular strength of nitric acid. This level is easily
detectable by an HPLC method (see Experimental).

Trapping Gaseous Amines in Acid Scrubbers

A strict and detailed material balance for the example plant employing the
general chemistry shown in Figures 2, 3, and 4 showed that MMA gas acts
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

differently in the presence of nitric acid than it does in the presence of sulfuric
acid. This type of behavior has also been observed for ammonia in certain plants.
This unusual behavior is shown in a dramatic way by the series of experiments
shown in Table 4.
The laboratory equipment used for the study is shown in Figure 12. Air was
passed through a flow meter at 2 SCFH (standard cubic feet per hour) into a small
impinger containing 9.0 grams of 40 wt% aqueous MMA. Later in the study, we
employed a bubbler instead of the jointed impinger as there was a concern that
the joint leaked. The gas flow from the impinger was conveyed via a Teflon®
tube to a large impinger (#1, 500 mL capacity) containing 400 mL of the desired
acid. The flow from this impinger was conveyed to the second impinger (500 mL
capacity) containing 400 mL of the desired acid. The exit of the second impinger
was vented to the laboratory hood. The impinger (or gas bubbler) containing the
aqueous MMA was heated to 45 °C using an external temperature controlled water
bath (not shown in Figure 12). The air flowed to the system for 2 hours, after which
it was found that the impinger containing the aqueous MMA no longer contained
MMA; some water remained in the impinger after 2 hours of air flow. The large
impingers were at ambient temperature. The frit at the end of dip tube in spargers
#1 and #2 were a C porosity frit (20 to 50 micron). The gas bubbles emanating
from the frit were small. In one experiment the frit was removed and the open end
of the tube had an internal diameter of ~5 mm. The gas bubbles emanating from
the open ended dip tube were large. In two experiments the frit was replaced with
a short piece of plastic tubing that fit snuggly over the end of the dip tube. The end
of the short piece of plastic tubing was plugged with a Teflon stopper. The sides of
the short piece of plastic tubing was pierced with a utility knife to create six slits
that were ~3 mm in length and ~0.5 mm wide. The gas bubbles emanating from
the slits were of medium size.
The level of MMA in the fluid contained in the 500 mL impingers was
measured, but that data is not presented here. However, the key observations are
shown in Table 4.

220
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Table 4. Trapping MMA Vapor (1)
MMAN Fog Break- Through
MMA Delivery Sparger #1 Sparger #2 Observed in of MAA to
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

Exp system Sparger Type Contents Contents Sparger #1? Sparger #2? Conclusions
C frit (small MMA is stripped from
A from impinger bubbles) water water no yes water with nitrogen sparge
white fog of MMAN
C frit (small 70% nitric formation in #1 and
B from impinger bubbles) 70% nitric acid acid yes yes traveled to #2
white fog of MMAN
C frit (small 32% nitric formation in #1 and
C from impinger bubbles) 32% nitric acid acid yes yes traveled to #2
221

open tube (large 86% sulfuric MMAS formation in #1,

D from impinger bubbles) 86% sulfuric acid acid no no no break-though to #2
MMAN once formed
spike MMA C frit (small 32% HNO3 spiked is not stripped out of
E into Sparger #1 bubbles) with MAA 32% HNO3 no no 32% nitric acid
MMAS once formed is
slit (medium 32% nitric not stripped out with
F from bubbler bubbles) 86% sulfuric acid acid no no nitrogen sparge
86% sulfuric acid
containing 10% MMAN fog formed in
slit (medium nitrosyl sulfuric 32% nitric sparger #1 and traveled
G from bubbler bubbles) acid acid yes yes to sparger #21
1MMAN is methylammonium nitrate. MMAS is methylammonium sulfate. In all experiments that air flow was 2 SCFH for 120 minutes. The content of
MAA in the spargers was determined (exact levels not shown).

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

Figure 12. Laboratory Equipment Used to Study Absorption of MAA Gas by

Aqueous Acids. (see color insert)

Experiment A
In this experiment 40% aqueous MMA was introduced into an impinger.
Sparger #1 was filled with water and fitted with a frit on the dip tube; in like
manner sparger #2 was also filled with water and fitted with a frit. Air was
allowed to bubble through the impinger and through each of the spargers at 2
SCHF for 120 minutes. MMA was stripped from the water in the impinger and
carried through sparger #1 and sparger #2. When sparger #2 was analyzed, MMA
was present. Water was not good at sequestering MMA.

Experiment B
This experiment was exactly the same as A except that 70% nitric acid was
placed in both sparger #1 and sparger #2. When the airflow was started, a white
fog was noted in the headspace of sparger #1 and also in the headspace of sparger
#2. The white fog in the headspace of sparger #1 actually flowed through the frit
and the liquid in sparger #2. The white fog was identified as methylammonium
nitrate. It was in a very finely divided liquid phase, mist, or sol. Unlike water
scrubbing the nitric acid produced a white fog that moved through the scrubbing
media.

Experiment C
This experiment was run as in B only 32% nitric acid was used in both
spargers. The observed results were exactly the same as in experiment B.
222
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Experiment D
The MMA delivery system used the small impinger however the spargers both
had open dip tubes rather than frits. This resulted in large bubbles emanating from
the dip tubes in the large spargers. The spargers were charged with 86% sulfuric
acid and the airflow was started. No fog was observed over the headspace of either
sparger nor was any MMA found in sparger #2. The sulfuric acid was a more
effective scrubber for vaporized MMA in this lab equipment.

Experiment E
The impinger in this case was empty and each sparger was filled with 32%
nitric acid. However, sparger #1 was spiked with the aqueous MMA. In this way,
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

the MMA was dissolved in the acid as the methylammonium nitrate salt. The
airflow was started through the impinger but no fog or sol was formed in either
sparger #1 or sparger #2 and no breakthrough of MMA was observed in sparger
#2. The methylammonium nitrate was formed and dissolved when the nitric acid
in sparger #1 was spiked with MMA, and it stayed there after being sparged with
air.

Experiment F
A bubbler was charged with the aqueous MMA. The dip tubes in the large
spargers were fitted a short plugged plastic tube. The plastic tube was pierced with
a utility blade to form short slits. The bubbles emanating from the slits were of
medium size. Sparger #1 was charged with 86% sulfuric acid while sparger #2 was
charged with 32% nitric acid. The gas flow was started and the system observed.
No white cloud or sol was observed above either sparger #1 or sparger #2 and
no breakthrough was observed in sparger #2. This again showed the efficacy of
sulfuric acid to scrub MMA.

Experiment G
A bubbler was charged with aqueous MMA. Sparger #1 was charged with
400 mL of 86 wt% sulfuric acid containing 10 wt% nitrosyl sulfuric acid. Sparger
#2 was charged with 32 wt% nitric acid. Air flow was established at 2 SCFH. A
white fog was observed in the head space of sparger #1 which flowed into sparger
#2. MMA was observed in sparger #2. It was not verified whether the fog was
methylammonium sulfate or methylammonium nitrate, but taking all the scrubbing
experimental results together, it was very likely that methylammonium nitrate was
formed in sparger #1 and carried forward to sparger #2 in experiment G.
These experiments were not repeated with ammonia but it is very likely that
similar behavior would be observed. The conclusions are that MMA vapor was not
scrubbed efficiently with nitric acid and MMA vapor was scrubbed efficiently with
clean sulfuric acid. Surprisingly, if the sulfuric acid contained nitrosyl sulfuric acid
(or presumably nitric acid), then methylammonium nitrate can form when MMA
vapor passes through the acid solution and escape as a sol. These findings are
223
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 very consistent with what was observed in the plant as shown by the mass balance.
While lab experiments seem to show the large difference in the scrubbing ability of
nitric versus sulfuric acid, processes are known where very minute amounts of mist
can defeat scrubbers and still allow small amounts of “mist” to travel and collect in
vent systems. In other words better does not mean perfect and the risk still needs
to be assessed in the actual specific equipment and configurations employed.

Summary
One must consider the unique safety concerns in all process areas when
operating a nitration plant. Compounds and by-products to watch for in a nitration
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

facility have been reviewed (Table 1).

The design of a nitration plant must avoid areas where material can collect and
stagnate. There should be low points that can trap material. All vessels, lines, and
vents should be self-draining. Material in the process should not be allowed to be
blocked in; there should always be a route to relieve pressure. A pump should
not be allowed to dead head through the use of proper sensors and interlocks.
Bubbles should not be allowed to form in process streams, as collapsing bubbles
can create micro environments of high temperature and pressure that can cause the
initiation of a fast uncontrollable reaction. Process streams should be kept moving
and not allowed to stagnate, especially without proper cooling. Many accidents
occur where nitration occurs in an area of the plant other than in the nitrator. One
must be cognizant of factors that contribute to runaway nitration reactions.
A mixture of an oxidant and organic contains more energy as it approaches
oxygen balance. These mixtures can detonate (the reaction front self-propagates
faster speed of sound) or deflagrate (the reaction front self-propagates at just under
supersonic velocity). Oxygen balanced mixtures of oxidant and fuel are more
sensitive (easily prone to reaction via heat, shock, static spark, or mechanical
energy input) than mixtures that are not. Oxygen balanced materials can be highly
sensitive or not sensitive, but they should be avoided whenever possible.
This is a partial summary of items to consider when designing and operating
a nitration plant. Each nitration facility presents unique challenges, and requires
expertise in design, hazards testing and evaluation, and operation. The intention
of this chapter is to urge you to be diligent in your quest to reduce risk in your
nitration facility. A proper Level of Protection Analysis (LOPA) and hazards of
operation review (Hazop) should be done when the plant is designed, and when
the plant and equipment change. Other primers and experts ought to be sought out
and used to plan your process design, operating rules, and risk abatement.
Good mass balances of substituents in a nitration plant are required to operate
a safe facility. Solubility limits of by-products in process streams must be known
to insure no phase separation occurs in the process. Phase separation can lead to a
host of process issues. Clean sulfuric and nitric acid streams should be analyzed for
carbon by TOC. Free amines must be tracked in process streams as nitrate/nitrite
salts are sensitive materials. We have shown that ammonium nitrate salts, present
in vapor streams as sols (fine solid particles), can defeat common scrubbers. Safety
critical analyses need to be performed every shift every day. Some analyses are
224
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 performed even when the plant is down to make sure nothing is changing in storage
vessels. Mass balances may vary over certain time frames, and in fact seldom does
the first mass balance ever close in a plant. It takes much work and diligence to
catalog the organic species to track, and to develop the best analytical methods to
do so. It also requires good measurement of mass flows throughout the plant.
In conclusion, the authors would like to emphasize again the importance of
understanding the risks involved in performing nitration reactions and operating a
nitration plant. Knowledge and experience are important tools to mitigate risk.

Experimental
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

TNM Analysis by HPLC

Product streams throughout the nitration plant are analyzed for

tetranitromethane (TNM). The acid streams are also analyzed for the nitration
starting material, product, and side reaction products. All samples are analyzed
by dissolving the sample in a large amount of acetonitrile (care must be taken to
avoid a safety issue with sample preparation). The diluted sample is then injected
onto a High Pressure Liquid Chromatograph (HPLC). Most results are reported
as parts per million (ppm) although a few are reported as weight %.
Samples from 15 to 20 plant locations are analyzed either once per shift or
once per day using this procedure. A sampling plan and a sampling matrix (with
sample preparation instructions unique for the sample) are utilized. Target values
of concern have been set for all streams. Appropriate standards are prepared, and
analyte recovery experiments performed to verify the method.

Safety Considerations/Hazards

Acetonitrile - flammable, harmful if inhaled or absorbed through skin

Nitric acid - highly toxic, oxidizer
TNM – oxidizer
Wear proper personal protective equipment when handling all chemicals
In case of skin contact rinse thoroughly with water
In case of eye contact rinse thoroughly with water and seek medical assistance

Equipment

Waters liquid chromatographic system consisting of the following: Waters

2695 Separations Module with Model 2489 Absorbance Detector (280 nm
detection), Waters Resolve C18 5 micron 8 x 100 mm Rad-Pac cartridge column,
2.5 mL/min flow rate, isocratic solvent (55/45, water/acetonitrile), 15 µL
injection, column oven temperature 35 °C.

225
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Percent Organics, NF, and Product Analysis by HPLC
Sample prep is the same as for the analysis of TNM in acid process streams
(appropriate dilution with acetonitrile). Appropriate standards are prepared, and
recovery experiments from each sample matrix were performed to verify the
method. Nitroform was prepared by the method of Geckler (25).

Samples Are Shot on an HPLC

Waters 2695 Separations Module; Model 2489 Absorbance Detector (280 nm

detection); using a Phenomenex Luna 5µ, C18, 150 mm by 3 mm column; isocratic
solvent system (prepared by dissolving 31.8 gr of tetramethylammonium chloride
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

in 120 mL of methanol, then adding that to 2400 mL of deionized water, then

adding to that 1080 mL of HPLC grade acetonitrile, then adding to that 0.8 mL
of glacial acetic acid, finally mixing and filtering before use); at 1.0 mL/min flow
rate; 5 µL injection; column oven temperature 30 °C.

References
1. Guggenheim, T. L.; Odle, R. R.; Pace, J. Process Design and Operational
Controls to Safeguard Strong Nitric Acid Recovery Systems. In Chemistry,
Process Design, and Safety for the Nitration Industry; Guggenheim, T.
L., Ed.; ACS Symposium Series 1155; American Chemical Society:
Washington, DC, 2013; Chapter 13.
2. Provided here are several articles describing studies of runway nitration
reactions. (a) Chen, C.-Y.; Wu, C.-W.; Duh, Y.-S.; Yu, S. W. An experimental
study of worst case scenarios of nitric acid decomposition in a toluene
nitration process. TransIChemE, 1998, 76, 211−216. (b) Kotoyori, T.
Investigation of a thermal runaway reaction involving a nitration process. J.
Loss Prev. Process Ind. 1991, 4, 120−124. (c) Lu, K.-T.; Luo, K.-M.; Lin,
P.-C.; Hwang, K.-L. Critical runaway conditions and stability criterion of
RDX manufacture in continuous stirred tank reactor. J. Loss Prev. Process
Ind. 2005, 18, 1−11. (d) Andreozzi, R.; Canterino, M.; Caprio, V., Somma,
D. I.; Sanchirico, R. Batch salicylic acid nitration by nitric acid/acetic acid
mixture under isothermal, isoperibolic and adiabatic conditions. J. Hazard.
Mater. 2006, A138, 452−458. (e) Somma, D. I.; Marotta, R.; Andreozzi,
R.; Caprio, V. Kinetic and safety characterization of the nitration process
of methyl benzoate in mixed acid. Org. Process Res. Dev. 2012, 16,
2001−2007. (f) Zaldivar, J. M.; Hernandez, H. N.; Molga, E.; Bassani, C.
The FIRES Project: Experimental study of thermal runaway due to agitation
problems during toluene nitration. J. Loss Prev. Process Ind. 1993, 6,
319−326. (g) Lunghi, A.; Alos, M. A.; Gigante, L.; Feixas, J.; Sironi, E.;
Feliu, J. A.; Cardillo, P. Identification of the decomposition products in an
industrial nitation process under thermal runaway conditions. Org. Process
Res. Dev. 2002, 6, 926−932. (h) Badeen, C.; Turcotte, R.; Hobenshield,
E.; Berretta, S. Thermal hazard assessment of nitrobenzene/dinitrobenzene
226
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 mixtures. J. Hazard. Mater. 2011, 188, 52−57. (i) Van Roekel, L. Thermal
Runaway Reactions: Hazard Evaluation. In Chemical Process Hazard
Review; Hoffman, J. M., Maser, D. C., Eds.; ACS Symposium Series 274;
American Chemical Society: Washington, DC, 1985; Chapter 8.
3. Hanson, C.; Kaghazchi, T.; Pratt, M. W. T. Side Reactions during Aromatic
Nitration. In Industrial and Laboratory Nitrations; ACS Symposium Series
22; Albright, L. F., Hanson, C.; Eds.; Washington, DC: American Chemical
Society, 1976; pp 132−134.
4. Schofield, K. Aromatic Nitration; Cambridge University Press: Cambridge,
1980; pp 23−25.
5. Gutsche, D. C.; Pasto, D. J. Fundamentals of Organic Chemistry; Prentice-
Hall: Englewood Cliffs, NJ, 1975; p 213.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

6. Olah, G. A. Preparative and Mechanistic Aspects of Electrophilic Nitration.

In Industrial and Laboratory Nitrations; ACS Symposium Series 22;
Albright, L. F., Hanson, C.; Eds.; Washington, DC: American Chemical
Society, 1976; p 23.
7. Urbanski, T. Chemistry and Technology of Explosives, Vol. 1; Pergamon
Press: Warsaw, 1964, 1985; pp 9−22.
8. Hanson, C.; Kaghazchi, T.; Pratt, M. W. T. Side Reactions during Aromatic
Nitration. In Industrial and Laboratory Nitrations; ACS Symposium Series
22; Albright, L. F., Hanson, C.; Eds.; Washington, DC: American Chemical
Society, 1976; pp 132−135.
9. Urbanski, T. Chemistry and Technology of Explosives, Vol. 1; Pergamon
Press: Warsaw, 1964, 1985; pp 389−91.
10. Urbanski, T. Chemistry and Technology of Explosives, Vol. 1; Pergamon
Press: Warsaw, 1964, 1985; pp 8−9.
11. Urbanski, T. Chemistry and Technology of Explosives, Vol. 1; Pergamon
Press: Warsaw, 1964, 1985; pp 589, 594.
12. Seyewetz, A. Compt. Rend. 1909, 148, 1110.
13. Kaye, S. M. Encyclopedia of Explosives and Related Items, PATR 2700
ed., Vol. 8; U.S. Army Research and Development Command, TACOM,
ARDEC, Warheads, Energetics and Combat Support Center, Picatinny
Arsenal: Picatinny, NJ, 1978; pp m83−m84.
14. Kaye, S. M. Encyclopedia of Explosives and Related Items, PATR 2700
ed., Vol. 8; U.S. Army Research and Development Command, TACOM,
ARDEC, Warheads, Energetics and Combat Support Center, Picatinny
Arsenal: Picatinny, NJ, 1978; pp m78−m81.
15. Kaye, S. M. Encyclopedia of Explosives and Related Items, PATR 2700
ed., Vol. 8; U.S. Army Research and Development Command, TACOM,
ARDEC, Warheads, Energetics and Combat Support Center, Picatinny
Arsenal: Picatinny, NJ, 1978; p m96.
16. Fedoroff, B. T.; et al. Encyclopedia of Explosives and Related Items,
PATR 2700 ed., Vol. 8; U.S. Army Research and Development Command,
TACOM, ARDEC, Warheads, Energetics and Combat Support Center,
Picatinny Arsenal: Picatinny, NJ, 1960, pp a311–a354.
17. Fedoroff, B. T.; et al. Encyclopedia of Explosives and Related Items,
PATR 2700 ed., Vol. 8; U.S. Army Research and Development Command,

227
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 TACOM, ARDEC, Warheads, Energetics and Combat Support Center,
Picatinny Arsenal: Picatinny, NJ, 1960; pp a25–a27.
18. Urbanski, T. Chemistry and Technology of Explosives, Vol. 4; Pergamon
Press: Warsaw, 1964, 1985; Chapter 8, pp 218−268.
19. Urbanski, T. Chemistry and Technology of Explosives, Vol. 1,Chapter 19,
Pergamon Press: Warsaw, 1964, 1985; pp 579−601.
20. Kaye, S. M. Encyclopedia of Explosives and Related Items, PATR 2700
ed., Vol. 8; U.S. Army Research and Development Command, TACOM,
ARDEC, Warheads, Energetics and Combat Support Center, Picatinny
Arsenal: Picatinny, NJ, 1978; p 341.
21. Akhavan, J. Chemistry of Explosives, 3rd ed.; Royal Society of Chemistry:
London, 2011; pp 85,99,107.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch014

22. Halle, R. Potential Hazards of Nitrogen Oxide Compound Accumulations

in Cryogenic Ethylene Recovery Facilities. Paper 15e, Ethylene Producers
Committee, AICHE Spring Meeting, 1993.
23. Guidelines for Safe Storage and Handling of Reactive Materials; Center for
Chemical Process Safety/AIChE, 1995; pp 68−70.
24. Guggenheim, T. L.; Evans, C. M.; Odle, R. R.; Fukuyama, S. M.; Warner,
G. L. Removal and Destruction of Tetranitromethane from Nitric Acid. In
Nitration: Recent Laboratory and Industrial Developments; Albright, L. F.,
Carr, R. V. C., Schmitt, R. J., Eds.; ACS Symposium Series 623; American
Chemical Society: Washington, DC, 1996; pp 187−200.
25. Brown, L. H.; Geckler, R. D., Research in Nitropolymers and Their
Application to Solid Smokeless Propellants; Aerojet Engineering Corp.,
Quarterly Summary Report 371, April 15, 1949.

228
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 15

Redesign for Safe Operation of

a Nitric Acid Recovery Unit
John Pace*,1 and Colin Evans2
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

1SABIC, Parkersburg, West Virginia 26101

2Chemetics (retired), Toronto, Canada
*E-mail: [email protected]

In an effort to improve the operability of a weak nitric acid

recovery unit operated by the Innovative Plastics division of
Saudi Basic Industries Corporation (SABIC), a technology
team from SABIC initiated a study to understand the root cause
of the occasional loss of adequate control of the operating
temperature and pressure of the unit. A review of the capacity
of the unit’s emergency pressure relief vent was a second
objective of the study. The emergency relief requirements were
determined based on calorimetric data on process samples.
The operating targets of the unit were modified to minimize
the frequency and risk of the process excursions and the relief
system was modified to protect against the estimated maximum
credible relief case scenario.

Background
An extended plant outage was taken by SABIC to address issues resulting
from an incident in a nitric acid concentration process (1). Several safety reviews
were conducted during the outage for other areas of the plant that were not involved
in the incident. The topic of this paper is the work done to address one of the
potential issues discovered in the review of a weak nitric acid recovery unit. This
unit recovers weak nitric acid from a process stream and recycles it to the nitric
acid concentrator.
The weak nitric acid recovery unit consists of an evaporator (Figure 1) which
receives a feed stream containing 1% to 3% (by weight) “organics” in a water/nitric
acid mixture of approximately 40% acid strength. Organics, in the sense used

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 in this discussion, are the main organic products of the nitration facility, which
employs the nitric acid recovery unit. There are other organic species present
in this stream, but they are not rountinely measured. These other organics are
byproducts of the process, and are present at a weight ratio of no more than one to
one, by-product to product organics. The term “organics”, from this point forward,
refers to the product organics only. “Acid strength” refers to the weight percent
nitric acid in the combined water and nitric acid portion of the stream content,
ignoring the organic content.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

Figure 1. Weak nitric acid evaporator system. (see color insert)

The acid strength of the recovered nitric acid that is taken overhead in this
evaporator is typically 35%. The bottoms stream is a waste, normally 12%
to 15% organics with an acid strength of approximately 60%. The evaporator
operates under vacuum, at approximately 100 mmHg absolute pressure, and
at approximately 65°C. Heat is supplied to the evaporator indirectly by steam,
through a thermosyphon reboiler. Pumping of the bottoms mixture is avoided to
prevent overheating of the bottoms material, since it is known that nitric acid can
exothermically oxidize the organics at elevated temperatures and generate NOx,
CO, and CO2. The vapor overheads of the evaporator are condensed and steam
jets are used to maintain the operating vacuum in the evaporator. The system
pressure is controlled by addition of steam to the suction of the jet, as a vacuum
break. A three inch emergency pressure relief vent was provided in the original
design of the evaporator, with a rupture disk having a burst pressure of 13.5
psig. The emergency pressure relief header is routed to a NOx removal system to
prevent the release of NOx to the atmosphere.
230
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 It was noted during the review that from time to time it was difficult for
operators to control the evaporator at a stable temperature and pressure. This
generally occurred when the organics content in the bottoms exceeded 18% to
20%, usually due to transitions in the composition of the feed to the evaporator
or to upsets in the rate of removal of the bottoms material from the evaporator.
Bottoms samples were routinely taken for analysis every 16 hours and the
operators typically adjusted the feed and steam rates to the evaporator to return
the bottoms organics content to the normal range of 12% to 15% if the analysis
indicated the organics were outside of that range. Upon examination of the past
operating data, several of these disturbances in the operating temperature and
pressure of the evaporator were found by the review team. The temperature and
pressure trends of a typical disturbance of this type are shown in Figure 2.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

Figure 2. Pressure and temperature trends during a typical evaporator

disturbance. (see color insert)

During these disturbances the operating pressure would typically increase

to as much as 700 mm Hg absolute, with the temperature reaching as high as
100°C. The review team decided to study these operating upsets in more detail,
particularly to see if the vessel was adequately protected from excessive pressure
in the event that the bottoms organic content became excessively high, and if the
usual methods of returning the evaporator to its target operating temperature and
pressure were not successful.

Discussion
An investigation was initiated to gain an understanding of the source of
temperature and pressure excursions in the operation of the nitric acid evaporator,
and to make recommendations to help avoid these events. A second objective was
to determine the adequacy of the existing emergency relief system and to make
231
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 recommendations for relief system upgrades if the existing system was found to
be inadequate.
To achieve these objectives, several samples were subjected to calorimetric
testing techniques, further described below. Use of these test results by the reader
for his or her processes should not be attempted since the test results are very
dependent on the nature of the organics present in the samples.
There was only one evaporator bottoms sample available for testing after
the nitric acid concentrator incident (1). The plant would not be re-started until
the changes in procedures and equipment design recommended during that safety
review were implemented. In order to have a wider range of samples available for
study, an effort was made to prepare and test “synthetic samples”. Synthetically
prepared mixtures were found to give calorimetry results that were not consistent
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

with the results on the available process sample. The organic content of the process
sample was 17%, on the upper end of the normal operating range. There was
enough of this sample to perform calorimetry for use in evaluation of the adequacy
of the emergency pressure relief system under normal operation. However the
quantity was insufficient to allow concentration of the material in the laboratory to
a higher organic content to enable evaluation of the relief requirements under more
extreme conditions. The available sample was tested in an Advanced Reactive
System Screening Tool (ARSST) calorimeter (2). Information on the behavior of
material with higher organic content would have to be to be inferred from ARC
calorimetry that had been performed on plant material several years earlier, on
samples ranging from 14% to 31% organics.
The 17% organics process sample that was available for calorimetric
evaluation was testing using procedures developed by the Design Institute for
Emergency Relief Systems (DIERS) program, sponsored by the American
Institute of Chemical Engineers. Methods were developed under DIERS
sponsorship for evaluating emergency pressure relief requirements for reactive
systems, where the relieving flow often consists of two phases. The ARSST is
one of a series of calorimeters that were developed specifically for generation of
data for use in prediction of relief behavior in large production scale equipment,
using the DIERS methodology. The ARSST allows generation of the required
data using relatively small samples (~10 grams). Characteristics of the ARSST
include:

1. A 10 cc agitated thin wall spherical glass test cell inside a 350 cc pressure
containment vessel.
2. The test cell is open to the containment vessel and pressure is equalized
between the two, allowing the use of the thin wall glass test cell.
3. A low phi-factor (discussed below), allowing collection of data at
effectively adiabatic conditions.
4. The test cell temperature can be ramped upward at a constant rate (or
alternately by a heat-wait-search routine) until an exotherm is detected.
5. A temperature controller and heater on the test cell allow maintenance of
the same temperature outside of the cell as inside, to prevent heat flow to
the containment vessel.

232
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 6. The 350cc pressure containment vessel can be vented and controlled at
a constant backpressure, or closed to allow the pressure to build as the
experiment proceeds, depending on which mode best fits the purpose of
the test. The use of the open test cell in the ARSST test method has been
confirmed to give results that compare very well with test results from
other DIERS developed apparatus that use closed test cells (such as the
larger VSP2 Vent Sizing Package calorimeter) (2).

Temperature vs. time and pressure vs. time data is collected during the
experiment and is later converted to rates of pressure rise and rates of temperature
rise vs. sample temperature. This rate of rise data is required for the relief system
design calculations.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

The phi factor mentioned above is a measure of the adiabatic capability of

the calorimeter. Large commercial equipment exhibits nearly adiabatic behavior
during thermal runaway events. The heat losses from the contents of large-scale
equipment are very slight compared to the heat being generated within the
equipment. These heat losses include heat loss to the surroundings as well as
heat that is absorbed in raising the temperature of the containing vessel. Failure
to conduct the small-scale experiments adiabatically leads to an understatement
of the heat released by the reaction, reducing the rate of temperature rise that is
measured. This error in turn leads to underestimation of the emergency pressure
venting requirements of the full-scale vessel. Many lab calorimeters, such as the
accelerating rate calorimeter (ARC), are not truly adiabatic in that their sample
cells have significant “thermal capacity” compared to that of the test cell contents.
Thermal capacity is defined as the product of the mass and the specific heat of the
material. The ARC sample cell, for example, is a relatively thick walled metal
container, designed so that it is able to withstand the high pressures that can be
generated during the experiments. The sample container of the ARSST is vented
into the containment vessel; equalizing the pressure inside and outside of the
sample container, allowing the use of thin wall glass with a much lower thermal
capacity.

The phi factor is calculated as:

where:
mS is the mass of the sample
mSC is the mass of the sample container
CPS is the specific heat of the sample
CPSC is the specific heat of the sample container

233
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The phi factor is 1.0 for a completely adiabatic system. Phi factors of ~1.05
are typical of large scale vessels, as well as of the ARSST sample cell and sample.
ARC calorimeters typically have phi factors of 2.5 to 3.0. ARC calorimeters are
not typically used for the calorimetry supporting emergency pressure relief designs
for this reason.
ARSST data, generated with the pressure containment vessel in the closed
mode of operation, for the 17% organics sample are shown in Figures 3 and 4.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

Figure 3. Temperature and pressure vs. time for the 17% organics sample. (see
color insert)

Figure 4. Rates of temperature rise and pressure rise vs. temperature for the
17% organics sample. (see color insert)
234
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The ARSST test on the sample containing 17% organics yielded a rate
of temperature rise of ~18°C/min and a rate of pressure rise of ~6 psi/min at
123°C, the temperature required for the bottoms composition to exert 10 psig
of pressure (the desired rupture disk bursting pressure) in a closed system under
boiling conditions. These values could be used for design for relief for up to
17% organics evaporator bottoms material. Considering that it is possible to
concentrate to above 17% in the bottoms, some additional safety margin had to
be designed into the relief system so that it could accommodate relief scenarios
starting from organic content in excess of 17%. The intent was for the relief
system to be capable of successfully relieving the evaporator in the case of the
maximum credible organics content.
Determination of the maximum credible organics content was accomplished
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

through statistical analysis of the bottoms organic data for a period of eight years
of operation prior to the nitric acid concentrator incident (1). The histogram of the
data is shown in Figure 5.

Figure 5. Histogram of the previous eight years’ percent organics.

The average percent organics for this period was 13.6%, with short term and
long term standard deviations of 2.7% and 4.1% respectively. From experience,
organic analyses on the very high end of the histogram (>25%) are actually
more likely to be due to analytical error. In order to provide a suitable design
margin for the emergency relief, the existing ARC test data for 31% organics was
converted to what would have been the expected behavior in a low phi factor
calorimetry test. In addition it would be necessary to shift the mean of the %
organics slightly lower, or to reduce the variability in the bottoms concentration,
in order to reliably avoid exceeding 31% organics in the evaporator (targeting
235
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 <1 time in a million samples). The probability of exceeding 31% organics based
on the statistics from the previous eight years performance was estimated as 28
times per million samples. The accuracy of this probability estimate however is
questionable, based on the long tail seen on the high end of the histogram, and on
the uncertainty of the reality of the values over 25%.
In order to have information at 31% organics similar to the low phi data of
Figure 4 for 17% organics, it was necessary to convert the ARC rate of temperature
rise data to a low phi basis. Conversion of ARC data for this purpose is only
recommended when the relative amount of “burn-up” of reactive species in the
sample is relatively low between the point of initiation of the test and the point
that the sample reaches the relief conditions (123°C in this case). The amount of
burn-up was calculated as ~15% for the previous ARC data for a sample with 15%
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

organics. The relative burn-up for the 31% organics ARC test would be much less
than 15%, confirming that scaling of the ARC data for the 31% organics relief case
is a reasonable approach in the absence of additional samples for ARSST testing.
The raw ARC temperature and pressure data for 14% and 31% organics are
shown in Figure 6 below.

Figure 6. ARC data (prior to conversion) for 14% and 31% organics. (see
color insert)
The ARC rate of temperature rise was calculated and then multiplied by the
phi factor for that specific ARC test to convert it to a low phi basis. The phi factor
adjusted rate of temperature rise is shown in Figure 7 for 15% and 22% organics
ARC data, along with the 17% organics ARSST data.
The phi adjusted 15% organics ARC data agrees very well with the ARSST
data for the 17% organics sample. When comparing the difference in organic
content of those two tests, the conversion of the ARC data yielded slightly
conservative results. This comparison of the temperature rise data from the two
different calorimeters led to confidence that the 31% organics ARC data could be
used for the relief system design. The phi adjusted rate of temperature rise and
the rate of pressure rise in the ARC experiments are given in Figure 8 below for
15%, 22%, and 31% organics.
236
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

Figure 7. Phi factor adjusted ARC data compared to ARSST data. (see color
insert)

This rate of rise information for temperature and pressure was used for the
design of the emergency pressure relief system for the evaporator. The rates of
rise used for the design were taken at 123°C, the temperature required for the
bottoms composition to exert 10 psig of pressure in a closed system under boiling
conditions.
The formula for calculation of the required venting area is given by:

where:
A is the vent area (square meters)
V is the reactant volume (cubic meters)
CD is the discharge coefficient
P is the venting pressure (psig)
is the self-heat rate at the relief pressure, °C/min
is the rate of pressure rise of the ARSST test cell at the desired relief pressure,
psi/min

The above equation is specific to the geometry of the ARSST, accounting for
the volume of the containment vessel when the measured rate of pressure rise is
used. In order to have a rate of pressure rise on an equivalent basis for the ARC
generated data, the same ratio of rate of temperature rise to the rate of pressure rise
of ~3 for the 17% organics ARSST test was used for the estimation of the correct
rate of pressure rise to use for the 31% organics case.
237
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The discharge coefficient is specific to the geometry of the emergency venting
system. A vessel discharging to atmosphere would have a discharge coefficient
close to 1.0, an ideal nozzle. In the case of the evaporator however, where the
vessel is relieved to a NOx scrubbing system remote from the evaporator, the
discharge coefficient was estimated to be ~0.38 based on the number of equivalent
lengths of piping between the evaporator and the ultimate discharge point.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

Figure 8. Phi adjusted rate of temperature rise and rate of pressure rise for 15%,
22% and 31% organics. (see color insert)

The calculated relief requirements at 17% organics are shown in Table 1.

These results indicate that the relieving capacity of the existing 3 inch emergency
vent is inadequate to control the pressure from a relieving event at 17% organics,
even if the vessel inventory was reduced by half and the relieving pressure reduced
to 5 psig.

238
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The calculated relief requirements at 31% organics are shown in Table 2. In
the case of 31% organics a relief diameter of approximately 14 inches is necessary
for relief of the normal liquid inventory at 10 psig. For the final design a new 16
inch relief line was installed, incorporating a rupture disk with a burst pressure
of 10 psig. The evaporator body was replaced, utilizing a design that allowed
normal operation at approximately 50% of the liquid inventory of the original
design. The 16 inch relief line provides protection for the case of 31% organics
content along with a higher than normal operating level, should they happen
simultaneously. The existing 3 inch relief line was retained, with a 5 psig rupture
disk, to allow protection of the main disk in the event of a pressure increase not
related to oxidation of the bottoms material.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

Table 1. Emergency pressure vent diameter vs. relief pressure and

evaporator inventory for 17% organics

Table 2. Emergency pressure vent diameter vs. relief pressure and

evaporator inventory for 31% organics

239
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 A word of explanation is needed to clarify the use of the equilibrium (i.e.
normal boiling) temperature and pressure of the sample mixture for determination
of the parameters for vent sizing. The equilibrium temperature/pressure
relationship is specifically valid only for open systems when gassy components
(i.e. NOx, CO, and CO2 in this case) are being generated. When the containing
vessel is closed, the gas components exert an additional pressure based on the
quantity present, over and above the equilibrium pressure of the components in
the liquid mixture. The degree to which the vessel is being vented prior to a relief
event, therefore, determines the actual temperature which will be attained at the
point of pressure relief. In cases where there is no venting or limited venting prior
to relief, the temperature attained will be lower than the equilibrium temperature.
The temperature at the point of relief determines the rate of temperature rise,
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

which is the most significant factor determining the relief area required. Closed
or partially vented vessels therefore will relieve at a lower temperature and a
lower rate of temperature rise than a fully vented one. Once the emergency relief
vent opens, the conditions in the vessel return quickly to conditions very close to
the equilibrium conditions as the excess gassy components are released. The vent
area determined, therefore, is generally conservative for the point of relief; and is
specifically sized for the flow required past the initiation point of the relief.
An additional use for the rate of pressure rise data from the ARC experiments
was to give insight into the loss of ability to control the pressure in the evaporator
when the organics content become too high. The rate of pressure rise data was
converted to kg/hr of gas generation, subtracting the rate of rise of the vapor
pressures of the water and nitric acid, and converting the resulting pressure rise
to an estimated gas (NOx, CO, CO2, average MW=44) generation rate in the
evaporator, given the sample size and volume of the ARC test cell (10cc) and the
normal inventory of the evaporator. The results are shown in Figure 9.

Figure 9. Rate of gas generation estimate for the evaporator as a function of the
temperature and organics content of the bottoms. (see color insert)
240
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 The maximum capacity of the steam jets used for pressure control of the
evaporator is 40 kg/hr. At its normal operating temperature of 65°C, the pressure
in the evaporator is easily controlled when the organic content of the bottoms is
~15%. At this organic content, the capacity of the steam jets would not be exceeded
unless the temperature of the evaporator is allowed to increase to ~72°C. At 22%
organics, however, this estimate suggests that the steam jet capacity would be
marginal at the normal operating temperature. This explains why the temperature
and pressure control of the evaporator becomes difficult if the organics content of
the bottoms increases beyond about 20%. A control target of a maximum of 16%
organics in the bottoms stream was established to help reduce the frequency of the
loss of temperature and pressure control of the evaporator.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

Conclusions

Through this investigation the SABIC team was (i) able to determine the
adequacy of the existing emergency relief system and to make recommendations
for relief system upgrades, and (ii) understand the causes of temperature and
pressure excursions in the operation of the nitric acid evaporator, and to make
recommendations to help avoid these excursions. Recommendations made from
the knowledge gained from the investigation resulted in the following actions
were taken to improve the safe operation of the evaporator:

1. A separate 16-inch relief vent line and rupture disk with a burst pressure
of 10 psig was installed on the evaporator.
2. The burst pressure of the existing 3-inch rupture disk was reduced to 5
psig from the original 13.5 psig burst pressure.
3. The normal evaporator bottoms operating inventory was reduced by 50%
through re-design of the evaporator body.
4. A maximum control limit for bottoms organics of 16% was introduced.
5. The frequency of bottoms organics testing was increased to every 8 hours
from every 16 hours.
6. The capability for addition of ambient temperature water to the
evaporator was added to help in control of any future significant
temperature/pressure excursions.
7. Operating data are presented in the Appendix for a full year of operation
after these modifications were completed. This information demonstrates
the improvement that was realized in control of the bottoms organics
content after the modifications were implemented.

241
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Appendix
Bottoms organics data is given in Figure 10 for one full year of operation,
seven years after the redesign and procedural changes. The average organic
content of the evaporator bottoms over that period of time was 12.2%, which is
1.4% lower than for the eight years preceding the changes. When compared with
the performance shown in Figure 5 prior to the redesign, the combined effects of
establishing the 16% upper target for organics and increasing the frequency of
testing have combined to essentially eliminate excursions above 17% organics.
In addition the variability has been significantly reduced, with short-term and
long-term standard deviations of 1.7% and 2.2% respectively. With this average
and standard deviation, no cases exceeding the 31% organics emergency relief
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

design basis are statistically expected per one million samples. There has been
only one instance of loss of control of the evaporator due to high bottoms organics
in the eight years since the re-start. In that instance, just after the re-start, water
was added to the evaporator to provide cooling in order to re-establish control.
The operating pressure has not reached the 10 psig burst pressure of the 16 inch
rupture disk in the eight years of operation after the modifications.

Figure 10. Histogram of organics data after the evaporator system modifications.
(see color insert)

242
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Acknowledgments
The authors would like to thank Roy Odle, Mark Delong, and Tim Allen, all of
SABIC; Hans Fauske of Fauske and Associates; and James Chan (consultant) for
their many contributions to this work. Appreciation is also expressed to Thomas
Guggenheim of SABIC, Roy Odle, and Patrick Gallagher, formerly of General
Electric, for the ARC data used in this work.

References
1. Guggenheim, T. L.; Odle, R. R.; Pace, J. Process Design and Operational
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch015

Controls to Safeguard Strong Nitric Acid Recovery Systems. In Chemistry,

Process Design, and Safety for the Nitration Industry; Guggenheim, T.
L., Ed.; ACS Symposium Series 1155; American Chemical Society:
Washington, DC, 2013; Chapter 13.
2. Burelbach, J. P. Advanced Reactive System Screening Tool (ARSST), North
American Thermal Analysis Society 28th Annual Conference, Orlando, FL,
October 4−6, 2000.

243
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Chapter 16

Materials Challenges in Strong Nitric and

Sulfuric Acid Service
L. Mark L. Delong*
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch016

SABIC Innovative Plastics, 1 Lexan Lane, Mt. Vernon, Indiana 47620

*E-mail: [email protected]

Understanding of a facility’s materials of construction is key to

the profitable operation of any facility where both strong (>95%)
nitric and sulfuric acid is used. Generally, one must be able
to ferret out specific corrosion mechanisms as a function of
acid concentration and temperature in order to ensure that the
proper materials of construction are used in key areas. In this
chapter, two corrosion events are discussed: (i) the suitability
of A611 stainless steel in nitric acid service and (ii) materials
of construction of steam ejectors used in the concentration of
sulfuric acid. The author also discusses the lifetime of a variety
of materials.

Introduction
The chemical and process conditions during the manufacture of
4-nitro-n-alkylphthalimide (1), can be extreme. The process employs a NAC/SAC
(nitric acid concentrator/sulfuric acid concentrator) to continuously produce
strong nitric acid. HNO3 strengths are typically >95% but can range from 20% to
100%, containing as much as 2.5% NOx with respect to the nitric acid present.
The concentration of H2SO4, used to break the 70% HNO3 azeotrope in the
distillation portion of concentrated HNO3 production and also used to scrub NOx
(thereby forming NOHSO4), can vary from 70% to 90%. The concentration of
NaOH, which is used to neutralize certain undesirable byproducts, can range from
20% and 50%. These corrosive chemicals exist in process conditions ranging in
temperature from ambient to 165 °C, and pressure from atmospheric to 65 mm
Hg.abs.

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Given the dynamic nature of the manufacturing process, the materials
must be both specific to its chemical and process environment as well as
sufficiently flexible to maintain corrosion resistance over a wide range of chemical
concentrations. Furthermore, cost and availability of the materials must be
considered and in line with financial objectives. When specifying the construction
materials, one must define the acceptable lifetime vs. replacement cost. Not all
materials need to be expensive exotic metals. For example, polytetrafluoroethyene
(PTFE)-lined 304L stainless steel eventually degrades in strong nitric acid.
However, when used in conjunction with the proper nitric service and routine
inspection, we have found that this piping can remain in use for more than five
years even in strong nitric acid service.
Table 1 describes features of construction materials used in the manufacturing
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch016

process. More particularly, Table 1 emphasizes proper use of construction

materials that can be employed in our process. This Table describes just a few
of many various material failure modes that can exist in this process. In our
PTFE-lined 304L example, its service life is significantly shorter than 5 years
when exposed to high concentrations of NOx. This is because NOx permeates the
liner eventually corroding the surrounding stainless steel shell (2). In cases where
the pipe is under vacuum, both liner and shell can eventually collapse leading to
a potential safety hazard.

Table 1. Various Failure Modes in a NAC/SAC Environment

246
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Other factors to consider are water quality and galvanic effects. For example,
when chloride levels exceed 200 ppm, 304L stainless steel is susceptible to
chloride stress cracking. For the high Si alloy described in the Table 1, loss in the
concentration of Si due to excessive cold work will lead to a galvanic effect. This
is because the steel in the cold work section lacks passivation compared to the
surrounding metal, thereby causing an anodic effect in the cold work section (3).
In this chapter, we concentrate on two incidents, highlighted in Table 1, that
we believe will be of general interest to those who operate a NAC/SAC. These
will be detailed below.

Experimental
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch016

The experimental results were obtained under contract with Hira Ahluwalia,
Material Selection Resources, Pennington, NJ. In the examination of the metal
failure mode of the SAC ejectors, optical microscopy, and to some extent,
trial and error were the only tools required. To understand weld failures in
strong nitric service, a combination of optical microscopy, scanning electron
microscopy (SEM),energy dispersive x-ray (EDX) spectroscopy and Auger
electron spectroscopy (AES) were employed (4, 5).

SAC Ejector Failure

A 2-stage steam ejector is used to provide vacuum in the concentration
of approximately 70% H2SO4 (recovered diluted product from the extractive
distillation of HNO3 that has had any HNO3 removed in a prior step) to yield
approximately ~85% H2SO4. This particular ejector is a the second stage ejector
used to provide additional vacuum needed in the main concentration step for the
purpose of using less overall steam.

Figure 1. Ejector Schematic.

For those unfamiliar with a steam ejector, it is most easily recognized as a form
of aspirator much like that used in a high school science lab. The difference is that
in an aspirator the maximum vacuum is limited to the vapor pressure of water while
the maximum vacuum from steam is much higher. Steam enters the ejector at the
inlet and exits the diffuser/discharge nozzle. The flow of steam creates a vacuum
on the overheads line from the sulfuric acid concentrator.
247
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Routine inspection found the diffuser and the weld at the discharge flange
suffered significant degradation (Figures 1-3).The nozzle is exposed to low
pressure steam with entrained H2SO4 droplets of unknown concentration at
approximately 100 °C at 400 mm Hg.
Figures 2 and 3 are pictures of the effect on Hastelloy. Various materials of
construction, including 316L, Hastelloy C, and Hastelloy C lined with Ultimet
(similar to Hastelloy, but contains Co for increased wear resistance) were
previously trialed. The lifetime of these materials were less than12 months. In
each case the head section remained pristine while the tail sections appeared as
shown in Figures 2 and 3.
Initially, the degradation mechanism was thought to be erosion, in part,
because the H2SO4 concentration entrained in the vapor was estimated at 2%. This
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch016

concentration is normally suitable for Hastelloy-type material. The degradation

seen in Figures 2 and 3, combined with the failure of wear-resistant Ultimet,
made it apparent that corrosion was a factor to consider. For corrosion to be an
issue with this material the entrained H2SO4 would have to be 10% or greater in
concentration.
The Merck Index (6) indicates that Zr is “Very slightly attacked by hot
concentrated sulfuric acid”. Furthermore, at the time (2008) Zr was approximately
30% cheaper than Hastelloy. Given this, a monolithic tailpiece made of nearly
pure Zr was fabricated and trialed. This ejector, with a Hastelloy head and a Zr
tailpiece, has now been in continuous service for five years. This trial tends to
validate the hypothesis that the entrained H2SO4 concentration is greater or equal
to 10% such that corrosion is the primary avenue of failure.

Figure 2. Diffuser Section (Hastelloy). (see color insert)

248
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch016

Figure 3. Heat affected zone of weld at discharge flange (Hastelloy). (see color
insert)

Heat Affected Zones in A611 for >95% HNO3 Service

A 611 is a proprietary austenitic stainless steel containing approximately
6% Si. Si is known to aid in corrosion resistance in strong, hot (>50°C) HNO3
acid because it is easily oxidized thereby providing a corrosion resistant layer.
Corrosion was observed in A611 welded pipe at the following locations of each
service temperature, with 70 °C showing the largest amount of degradation
(Figures 4 and 5):

1. Weld Metal
2. Weld fusion line
3. Heat affected zone of the welding activity

The behavior observed in Figures 4 and 5 is quite different than that seen when
solution annealed A610 (a slightly lower Si containing steel) is used in this facility.
In fact, the non-solution annealed material in Figures 4 and 5 had a service life that
was measured in months while the solution annealed material in nearly identical
service had lifespans of several years.
In solution annealing, steel is heated to 50 °C above the austenitic temperature
and held for sufficient time to allow the material to fully form austenite. The
austenitic temperature is dependent upon the alloy (primarily upon the amount
of Ni and Cr) and represents the transition of Fe from a body centered cubic
crystal structure to a face centered cubic configuration (7). After treatment at this
temperature, the material is usually quenched to form a homogenous equilibrium
microstructure. This process is commonly completed after welding and/or cold
249
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 working in order to insure that the steel is homogenous, i.e., carbon is fully
solubilized and the alloy is not segregated into domains of various compositions.
Given that the material in question had not gone through the solution annealing
process, we applied a combination of SEM and EDX to determine the size of
the grain boundaries and its composition, respectively. These techniques were
applied to areas that showed no apparent corrosion and areas that were visibly
corroded, particularly the heat affected zone of the welds. AES was also employed
to understand how the alloy may have changed as a function of depth through the
metal along the heat affected zones.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch016

Figure 4. Pipe in >95% HNO3 70 °C Service. Corrosion is seen in between and

along welds (white area). (see color insert)

Figure 5. Pipe in >95% HNO3 50 °C Service. Note the corroded areas are less
extensive. (see color insert)
250
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Figure 6 compares the SEM images of corrosion in a heat affected welding
zone to that in a non-heat affected area. Round pegs were placed at the exact
location of the analysis (the dark material behind the pegs held the pegs in place
for the purposes of taking the top left picture). The images are taken from a
pipe in the HNO3 service in question. Intergranular corrosion is seen in the heat
affected area at the field weld in 70 °C service (2nd peg from the left). The more
homogeneous manufacturer’s annealed fabrication weld, while not immune,
shows greater resistance to attack (top right image; though hard to see, the field
weld is the horizontal line in the middle of the image). The bulk pipe in the top
right image appears more compositionally homogenous than either weld section.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch016

Figure 6. Micrographic Data. Arrows indicate the location of the SEM image
(bottom) in relation to the optical image (top). Pictures grouped to the left
are from a corroded section of piping corresponding to the heat affected zone
(second dot—arrow point) or the field weld (middle dot, the weld is the vertical
divot extending below the peg, that has the approximate width of the peg). The
right group of pictures is from a non-corroded section of pipe where the left
arrow points to the bulk of the pipe and the right arrow points to the original
factory fabrication weld. (see color insert)

EDX data is presented in Figure 7. While the magnitude of the atomic

spectral lines does not indicate abundance between elements, it does provide
some indication of elemental abundance when comparing specific elements. In
the corroded section, it is clear that Si is more abundant on the surface than in
the non-corroded section. The white material seen along the sections of corroded
pipe is essentially “sand” (SiO2) that has migrated to the surface. Given that Si is
used in this grade of stainless to passivate the metal, it is clear that the depletion of
this element in the bulk leads to more rapid corrosion than is seen in the non-heat
affected zones or welds that had been solution annealed. Notice too the difference
in relative abundance of Fe, Cr, Ni and O between the corroded and non-corroded
areas. The non-corroded area is more homogenous and exhibits much less

251
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 oxidation as evidenced by the lack of an abundance of Si “sand” deposited on its
surface. It is apparent that inhomogeneity is influenced by the field weld activity.
Figure 8 is the elemental profile as a function of sputter depth from AES. There
are two notable features. First, the predominance of Si, O and C at the surface of
areas 1, 2 and 4 indicates that SiO2 forms at the surface. Given the data in Figure
7 and the overall obvious appearance of sand on the surface this is not surprising.
At slightly greater depths, C appears. This indicates that SiC forms. Cr is seen
along with C at slightly deeper sputter depths suggesting formation of CrC. This
in effect lessens the amount of both Si and Cr available for passivation. It is likely
that these materials form at the grain boundaries resulting in the appearance seen
in Figure 6 bottom left SEM image.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch016

Figure 7. EDX data in relation to location. Arrows indicate the location of the
EDX data (bottom) in relation to the optical image (top). The picture on the top
left is from a corroded section of piping corresponding to the heat affected weld
zone where the white “dust”, or “sand” seen in the optical image above (fourth
dot) corresponds to the “dust” or “sand” seen in Figures 4 and 5. The second
dot is in the same heat affected zone as the fourth dot; the only difference is that
the white dust was removed for the sake of comparison via EDX. The top right
picture is from a non-corroded section of pipe where the left arrow points to the
original factory fabrication weld. (see color insert)

The second feature of note, area 3, the weld, contains Ca. This element is not
inherent in the A611, but results from the weld slag. From a chemical compatibility
perspective, the weld is likely the weakest point in the pipe. Since the heat affected
zone is more inhomogeneous (as compared to solution annealed metal), attack at
the weld leads to attack at the heat affected zone.
252
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch016

Figure 8. AES data on the corroded sections. Areas correspond to the optical
image embedded in the figure, where area 3 is the weld. (see color insert)

Conclusion
Understanding the service environment of a manufacturing process is
critical when choosing metallurgy for strong acid service. Accurate in-process
measurements of temperature and concentration are invaluable. In the case of
the SAC jet material issue process measurements can avoid the trial and error
involved in specification of a construction material. However, in some cases,
obtaining these measurements may be extremely difficult. In the absence of good
in-process measurements, coupon testing is indispensable. Sometimes relatively
inexpensive metals (e.g., Zr) yield very good lifetimes.
Cost of materials is important. Acceptable lifetimes vs. replacement cost
must be defined: not all materials need to be expensive exotic metals such as Ta or
Nb. Teflon-lined 304L eventually degrades, but in the proper nitric service, routine
inspection demonstrates it can remain in use for >5 years.
Finally, as demonstrated in the A611 pipe example, a material is only as
good as its weakest point of attack. In general, liners are chemically resistant
(as opposed to impervious) and base metals/alloys are susceptible to stresses from
fabrication (microstructure). These considerations must be balanced to achieve
the most cost effective option to prevent unnecessary downtime.

253
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 References
1. Aspects of this process have been previously published, see Guggenheim, T.
L.; Evans, C. M.; Odle, R. R.; Fukuyama, S. M.; Warner, G. L. Removal and
Destruction of Tetranitromethane from Nitric Acid. In Nitration: Recent
Laboratory and Industrial Developments; Albright, L. F., Carr, R. V. C.,
Schmitt, R. J., Eds.; ACS Symposium Series 623; American Chemical
Society: Washington, DC, 1996; pp 187−200.
2. Carter, W. P. L. Measurement and Modeling of NOx Offgasing from FEP
Teflon Chambers; University of California: Riverside, CA; DuPont Technical
Bulletin: Permeation—Its Effects on Teflon®, Fluoropolymer Coatings.
3. Guidelines for Alloy Selection for Water and Waste Water Service; Vol. 28-2;
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ch016

Nickel Magazine, Nickel Institute, 2013.

4. Goldstein, J.; Newbury, D. E.; Joy, D. C.; Lyman, C. E.; Echlin, P.; Lifshin,
E.; Sawyer, L.; Michael, R. Scanning Electron Microscopy and X-ray
Microanalysis, 3rd ed.; Springer: New York, 2003.
5. Grant, J. T.; Briggs, D. Surface Analysis by Auger and X-ray Photoelectron
Spectroscopy; IM Publications: Chichester, 2003.
6. Windholz, M., Ed.; The Merck Index, 9th ed.; Merck: Rahway, NJ, 1976.
7. Verhoeven, J. D. Fundamentals of Physical Metallurgy; Wiley: New York,
1975

254
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Subject Index
A B

Adiabatic mononitrobenzene process, 1. Basic process control system (BPCS), 141

See Operational Issues Bench-scale and pilot plant nitration
background, 2 experiments
first-generation adiabatic MNB apparatus, 115
technology, 3 continuous flow reactors, 114
MNB market, 2 experimental apparatus (configuration of
operational issues equipment), 112
second-generation adiabatic MNB experimental design for three factors,
technology, 4 111f
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ix002

summary, 9 experimental design plan, 110

Adiabatic nitration for mononitrotoluene kinetics, 113
(MNT) production, 27 literature search, 109
acid recycling, 42 nitration reaction, 115
adiabatic nitration screening tests, 31f overview, 107
black acid, 40 program objectives, defining, 108
complexity of monitoring by-products, safety audit, 117
38f sampling and analysis, 118
different isomers, 34f thermodynamics, 113
extract of crude MNT product, 39f BPCS. See Basic process control system
fast 2 phase reaction of toluene, 32f (BPCS)
increase in HPLC measured cresols, 43f
isomer distribution, 33, 34f
nitration acid loop test system, 44f
nitrosonium ion-toluene complex, 41f
C
reaction rate, 32
reaction temperature, 35 Chemical reactivity hazards
re-concentration, 29 assessment, 122
temperature-time curve, 31f heat of nitration reaction, 123
toluene nitration by-products, 36 designing nitration reactors, factors,
toluene oxidation by-products, 37f 124
Advanced reactive system screening tool HT-65 calculation
(ARSST), 193 mixture of DNB and sulfuric acid,
Ammonium nitrate explosions, 171 135f
Port Neal, 1994, 172 mixture of DNB and water, 134f
Texas City, 1947, 172 mixture of NB and nitric acid, 136f
Toulouse, France, 2001, 172 nitration, impurities or byproducts, 137
Aromatics, nitration technology nitric acid, thermal stability, 137
analysis, 80 nitro-compounds’ decomposition energy,
continuous nitration process, 76 125
DNT production processes, 77 accelerating rate calorimeter (ARC),
gas-phase nitration, 74 128f
liquid phase nitration, 75 ARC test, 129f
MNB and MNT/DNT, process overview differential scanning calorimeter
and trend areas, 72 (DSC), 127f
patents, 72 DSC curve, 127f
reactor, 74 heat generation, 130
solvent extraction method, 76 heat released from reactions of DNT,
trends, 79 126t
ARSST. See Advanced reactive system other sources of heat, 130
screening tool (ARSST) self-heating, 129

261
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 oxygen balance, 131 F
HT-65 heat released on detonation
with expansion, 133t Fault tree analysis (FTA), 165
several high explosives, 132t FTA. See Fault tree analysis (FTA)
testing, 137
Continuous benzene nitration process
analytical method, reproducibility, 58
experimental H
acetonitrile and aqueous potassium
dihydrogen phosphate, 55 Heats of reaction for toluene nitration, 123t
adjustment of pH, 56
eluent composition, 54
HPLC analysis, sample preparation, I
52
HPLC conditions and instrumentation,
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ix002

Independent protection layers (IPLs), 155

53 Industrial nitration processes, NOx
nitrophenols identification, 52 formation and capture, 96
sampling procedure, 52 Industrial NOx absorption process, 95
standard solutions, preparation, 51 absorption enhancement, hydrogen
HPLC separation efficiency parameters, peroxide addition, 98
56t atmospheric pressure with peroxide
nitrophenolic by-products, 50 addition, 101
nitrophenols, 54t experimental methods and model
quantification of nitrophenols, 57t development, addition of hydrogen
standard curves, 58 peroxide to rate-based model, 99
Continuous stirred tank reactor (CSTR), hydrogen peroxide injection, 97f
113 lower packed bed performance with and
CSTR. See Continuous stirred tank reactor without peroxide addition, 102t
(CSTR) results and discussion, 100
salient gaseous NOx and aqueous phase
reactions, 98t
E upper packed bed performance with and
without peroxide addition, 103t
Explosion in ammonium nitrate neutralizer, IPLs. See Independent protection layers
171 (IPLs)
accident at Port Neal, Iowa
ammonium nitrate neutralizer, 173
explosion, 175 L
interior view of scrubber, neutralizer
and rundown tank, 174f Layer of protection analysis (LOPA)
neutralizer process, 174f generic ignition probabilities, 157t
plant layout, 173 generic initiating event likelihoods, 153t
accident main factors initiating event, 152
acidity, 177 layers of protection, 155
chloride contamination, 179 nitration industry example reactor, 160f
confinement, 181 with additional SIS, 162f
excessive heat, 180 PFD for typical IPLs, 156t
cause of explosion, controversy study example for nitration industry, 160
independent assessments, 176 target mitigated event likelihood
theories, established facts, and open (TMEL), 158
questions, 175 example based on environmental risk,
key neutralizer operating parameters, 159t
178f example based on financial risk, 159t

262
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 M TNM, NF, and MMA, weak nitric acid
recovery system, 215
MNB. See Mononitrobenzene (MNB) TNM and NF, reaction area, 213
Mononitration of benzene into weak nitric acid recovery, 216f
mononitrobenzene (MNB), 13 percent organics, NF, and product
actual reaction conditions for each run, analysis by HPLC, 226
19t solubility of TNM and NF in nitric acid,
DNB formation 218
effect of reaction average temperature, solubility of TNM in acid media, 219t
24f TNM analysis by HPLC
effect of sulfuric acid concentration, equipment, 225
25f safety considerations/hazards, 225
experimental procedure, 18 trapping gaseous amines in acid
experimental set up, 17 scrubbers, 220
nitrophenol formation experiment A, 222
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ix002

effect of reaction average temperature, experiment B, 222

22f experiment C, 222
effect of sulfuric acid concentration, experiment D, 223
23f experiment E, 223
process overview, 14 experiment F, 223
proposed experimental conditions for experiment G, 223
each run, 16t trapping MMA vapor, 221t
sulfuric acid ratio understanding plant chemistry, simple
DNB formation, 21f goals
nitrophenol formation, 20f distribution, 205
test program, description, 15 exit, 206
Mononitrobenzene (MNB), 1 location, 205
Nitration of benzene, 29
summary of patents, 73
Nitration of toluene, summary of patents,
N 75
Nitric acid concentrator-sulfuric acid
NAC/SAC. See Nitric acid concentrator (NAC/SAC), 185
concentrator-sulfuric acid concentrator
(NAC/SAC)
Nitration facility, 203
analytical challenges O
analyze process streams, other
methods, 212 Operation of nitric acid recovery unit
analyzing MMA and ammonia, 211 14% and 31% organics, ARC data, 236f
analyzing tetranitromethane and 17% organics sample
nitroform, 211 rates of temperature rise and pressure
by-products of concern, 210 rise vs. temperature, 234f
chemistry considerations of nitration, temperature and pressure vs. time,
206 234f
desired product/isomers and typical adiabatic capability, 233
phenolic formation, 207f background, 229
formation of nitronium ion, 207f characteristics of ARSST, 232
nitric acid oxidation and formation of emergency pressure vent diameter vs.
TNM, 208f relief pressure, 239t
side reaction to ammonia, 209f evaporator disturbance, pressure and
mass balance temperature trends, 231f
equilibrium of TNM and NF and organics data after evaporator system
solvolysis behavior, 214 modifications, 242f
TNM, NF, and MMA, strong acid phi adjusted rate of temperature rise and
concentration, 217 rate of pressure rise, 238f

263
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 phi factor adjusted ARC data compared corrosion-resistant structured packing,
to ARSST data, 237f 69f
previous eight years’ percent organics,
histogram, 235f
rate of gas generation estimate, 240f
weak nitric acid evaporator system, 230f
S
Operational Issues
crude MNB purification, 7 Safeguard strong nitric acid recovery
environmental, 9 systems
patents and technology advancement, 9 condenser failure, most likely cause, 198
reliability, 5 condenser failure, probable cause
safety determination, 190
ammonium nitrite, 6 ARSST, pure nitroform, 197s
benzene handling, 7 ARSST, TNM in presence of
exotherms in MNB distillation, 6 compound, 196s
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ix002

exotherms in nitration train, 6 Aspen® Dynamics model, 192

nitric acid/MNB, 6 composition tested in ARSST, 195t
waste treatment explosion, 191
aliphatics purge, 8 Fauske and Associates ARSST, 194f
dinitrobenzene purge, 9 nitroform, 192
nitrophenols, treatment, 7 oxidants, 191
NOX recovery, 8 probable condensed phase
sulfuric acid purge, 8 compositions, thermal stability, 193
residual material, 191
NAC condenser and damage assessment,
failure, 188
P NAC condenser system, redesign, 198
NAC equipment and operation, brief
PFD. See Probability of failure on demand description, 186
(PFD) simplified schematic of NAC, 187f
PFR. See Plug flow reactor (PFR) Safety instrumented functions (SIF), 142
Plug flow reactor (PFR), 113 Safety instrumented systems (SIS)
Probability of failure on demand (PFD), and BPCS, basic concepts, 142f
155 concept, 143f
concepts of risk reduction, 147f
and functional safety background, 144
hazard/risk assessment, 145
R maintenance, 166
reduce personnel severity, mitigation
Reaction vessels, 115 credits, 151t
Reactor configuration, 116 risk graph
Recovery of nitric and sulfuric acids from environment, 150t
nitration plants personnel, 149t
De Dietrich process systems group, 63f production/assets (economic), 150t
history of De Dietrich/QVF®, 62 risk graph method
2nd generation acid recovery unit, 66f demand rate, 148
process systems design and process process safety time, 149
expertise, 64f severity criteria, avoidance and
3rd generation acid recovery system, 67f exposure, 148
recovery of acids in DNT nitration safety integrity level determination, 147
facility, 64 SIF loop example, 143f
1st generation acid recovery process SIL assignment, example risk graph,
design, 65f 151f
technical improvements of equipment, various independent layers of protection,
68 146f
corrosion resistant tray, 69f

264
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 SIF. See Safety instrumented functions SAC ejector failure, 247
(SIF)
SIL verification, 162
example FTA model, 165f
failure rate data, 165
W
IEC 61508 architectural constraints for
type A systems, 164t Water treatment of effluents, 83
Strong nitric and sulfuric acid service, capital cost estimates, basis, 92
materials challenges deepwell/biotreatment, 90
diffuser section (Hastelloy), 248f expected effluent quality, 91t
experimental, 247 incineration, 89
heat affected zones in A611 operating cost estimates, basis, 92
corroded areas, less extensive, 250f ozonation, 88
corroded section of piping, 251f, 252f solvent extraction, 87
corrosion, in between and along strong effluent flow and composition, 91t
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ix002

welds, 250f thermal destruction, 84, 85f

NAC/SAC environment, various failure utility and chemical costs, 92t
modes, 246t wet oxidation, 86

265
In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.
 Editor’s Biography

Thomas L. Guggenheim
Thomas L. Guggenheim earned a B.S. degree in Chemistry from St. Olaf
Publication Date (Web): November 22, 2013 | doi: 10.1021/bk-2013-1155.ot001

College in 1978 and a Ph.D. in Physical Organic Chemistry from the University
of Minnesota (P.G. Gassman) in 1983. He began his career at the GE Corporate
Research Center, working in the area of engineering thermoplastics, before moving
to GE Plastics in Mt. Vernon, Indiana in 1989 (SABIC purchased GE Plastics in
2007). Since that time, he has worked on process chemistry optimization, new
process development, process safety analysis, wastewater management, analytical
method development, patent management, and is an expert in nitration chemistry
and processes to manufacture polyetherimides.

© 2013 American Chemical Society

In Chemistry, Process Design, and Safety for the Nitration Industry; Guggenheim, T.;
ACS Symposium Series; American Chemical Society: Washington, DC, 2013.