Santi Kulprathipanja, James E. Rekoske, Daniel Wei, Robert v. Slone, Trung Pham, Chunqing Liu - Modern Petrochemical Technology - Methods, Manufacturing and Applications-Wiley-VCH (2021)

WILEY END USER LICENSE AGREEMENT
Go to www.wiley.com/go/eula to access Wiley’s ebook EULA.

Modern Petrochemical Technology
Modern Petrochemical Technology
Methods, Manufacturing and Applications
Santi Kulprathipanja
James E. Rekoske
Daniel Wei
Robert V. Slone
Trung Pham
Chunqing Liu
Authors All books published by WILEY-VCH
are carefully produced. Nevertheless,
Dr. Santi Kulprathipanja authors, editors, and publisher do not
Honeywell UOP warrant the information contained in
25 East Algonquin Road these books, including this book, to
IL be free of errors. Readers are advised
United States to keep in mind that statements, data,
illustrations, procedural details or other
Dr. James E. Rekoske items may inadvertently be inaccurate.
Ecolab Inc.
1 Ecolab Place Library of Congress Card No.:
St. Paul, Minnesota 55102 applied for
United States
British Library Cataloguing-in-Publication
Dr. Daniel Wei Data
Honeywell UOP A catalogue record for this book is
25 East Algonquin Road available from the British Library.
IL
United States Bibliographic information published by
the Deutsche Nationalbibliothek
Dr. Robert V. Slone The Deutsche Nationalbibliothek lists
Honeywell Advanced Mateirals this publication in the Deutsche
25 East Algonquin Road Nationalbibliografie; detailed
IL bibliographic data are available on the
United States Internet at <http://dnb.d-nb.de>.
Dr. Trung Pham © 2021 WILEY-VCH GmbH, Boschstr.

Honeywell UOP 12, 69469 Weinheim, Germany
25 East Algonquin Road
IL All rights reserved (including those of
United States translation into other languages). No
part of this book may be reproduced in
Dr. Chunqing Liu any form – by photoprinting,
Honeywell UOP microfilm, or any other means – nor
25 East Algonquin Road transmitted or translated into a
IL machine language without written
United States permission from the publishers.
Registered names, trademarks, etc.
Cover used in this book, even when not
Cover Image: specifically marked as such, are not to
© littlewormy/Shutterstock.com be considered unprotected by law.
Print ISBN: 978-3-527-34522-9

ePDF ISBN: 978-3-527-81819-8
ePub ISBN: 978-3-527-81817-4
oBook ISBN: 978-3-527-81816-7
Typesetting SPi Global, Chennai, India

Printing and Binding
Printed on acid-free paper
10 9 8 7 6 5 4 3 2 1
v
Contents
Foreword xi
Preface xiii
Part I Introduction 1
1 Refinery and Petrochemical Processes 3

1.1 Introduction 3
1.2 Petroleum 3
1.2.1 Forms of Petroleum 4
1.2.1.1 Gaseous Petroleum: Gaseous Petroleum Can Be Defined as Natural Gas
or Shale Gas 4
1.2.1.2 Liquid Petroleum: Liquid Petroleum Can Be Separated into Light Crude
Oil and Heavy Crude Oil 4
1.2.1.3 Solid Petroleum: Bitumen 5
1.2.2 Composition of Petroleum 5
1.2.3 Petroleum Refinery 6
1.2.4 Petroleum Products 9
1.3 Petrochemical Building Blocks 10
References 16
2 Petrochemical Markets 17
2.1 Introduction 17
2.2 The Market for Ethylene 18
2.3 The Market for Propylene 22
2.4 The Market for C4 Olefins and Diolefins 26
2.5 The Market for Aromatic Compounds 32
2.5.1 The Market for Benzene 32
2.5.2 The Market for Toluene 34
2.5.3 The Market for Xylene 36
2.6 The Market for Synthesis Gas 37
vi Contents
2.6.1 The Market for Ammonia 38

2.6.2 The Market for Methanol 39
References 40
Part II Olefins and Synthesis Gas 43
3 Olefins and Synthesis Gas Production Technologies 45

3.1 Introduction 45
3.2 Steam Cracking 45
3.3 Light Paraffin Dehydrogenation 48
3.3.1 Process and Catalyst 48
3.3.1.1 Reactor Section 49
3.3.1.2 Product Recovery Section 50
3.3.1.3 Catalyst Regeneration Section 50
3.3.2 UOP C3 Oleflex Complex 51
3.3.3 UOP Iso-C4 Oleflex Complex 54
3.3.4 UOP C3 /C4 Oleflex Complex 56
3.3.5 UOP Mixed-C4 Oleflex Complex 56
3.4 Heavy Normal-paraffin Dehydrogenation 57
3.4.1 Production of Normal-paraffins 58
3.4.2 Production of LAB 59
3.5 Methanol to Olefins 59
3.6 Syngas Production 62
3.6.1 Removal of Impurities in Raw Natural Gas 63
3.6.2 Conversion of Natural Gas to Syngas 64
3.6.2.1 Steam Reforming 65
3.6.2.2 Partial Oxidation 65
3.6.2.3 Autothermal Reforming 66
3.6.3 Coal Gasification to Syngas 67
References 72
4 Uses of Olefins and Synthesis Gas 75

4.1 Introduction 75
4.2 Uses of Ethylene 75
4.2.1 Ethylene to Polyethylene 75
4.2.2 Ethylene to Polyvinyl Chloride 77
4.2.3 Ethylene to Monoethylene Glycol 78
4.3 Uses of Propylene 79
4.3.1 Propylene to Polypropylene 79
4.3.2 Propylene to Propylene Oxide 80
4.3.3 Propylene to Acrylonitrile 81
4.4 Uses of Syngas 82
4.4.1 Syngas to Hydrogen 82
4.4.2 Syngas to Ammonia 82
4.4.3 Syngas to Methanol 83
References 83
Contents vii
Part III Aromatics 85
5 Aromatic Production Technologies 87

5.1 Introduction 87
5.2 Aromatic Production 87
5.2.1 Reforming Technology 88
5.2.1.1 Semi-regenerative Reforming 91
5.2.1.2 Cyclic Reforming 91
5.2.1.3 CCR Reforming 91
5.2.1.4 High-pressure vs. Low-pressure Consideration 94
5.2.2 Aromatic Transalkylation and Disproportionation Technology 95
5.2.2.1 Selective Toluene Disproportionation Process 99
5.2.3 Isomerization of Aromatics 100
5.2.3.1 Reaction Mechanism 100
5.2.4 Toluene Methylation 103
5.2.5 Aromatization of Light Naphtha/Hydrocarbons 105
5.2.5.1 Mechanism 106
5.2.6 Catalytic Cracking 107
5.2.7 Production from Pyrolysis and Coking 111
5.2.7.1 Steam Cracking/Pyrolysis Gasoline 111
5.2.7.2 Coke-oven Benzole 113
5.2.7.3 Biomass Pyrolysis 115
5.3 Summary 119
References 119
6 Uses of Aromatics 121

6.1 Introduction 121
6.2 Polymer Production from Basic Aromatic Building Blocks 121
6.2.1 Ethylbenzene and Styrene Production 122
6.2.2 Cyclohexane to Nylon 6, Nylon 6,6 123
6.2.2.1 Adipic Acid Production 124
6.2.2.2 Caprolactam Synthesis 125
6.2.2.3 Properties of Nylon 6 vs. Nylon 6,6 128
6.2.3 Production of Cumene 128
6.2.3.1 Bisphenol A production 130
6.2.3.2 Production of Nitrobenzene 130
6.2.3.3 Production of Aniline 131
6.2.3.4 Production of MDI and Poly-MDI (PMDI) 131
6.2.3.5 Production of Toluene Diisocyanate (TDI) 131
6.2.3.6 Production of Polyurethane 132
6.2.3.7 Alkylbenzenes and Alkylbenzene Sulfonates 132
6.2.3.8 Maleic Anhydride from Benzene 133
6.2.3.9 Dichlorodiphenyltrichloroethane (DDT) from Benzene 133
6.3 Resins and Chemicals Production 133
6.3.1 Unsaturated Polyester Resins Synthesis 138
6.3.2 Alkyd Resin Synthesis: 139
viii Contents
6.3.3 p-Xylene Oxidation to Terephthalic Acid 139

6.3.4 Poly(ethylene terephthalate), PET, and Poly(ethylene naphthalate),
PEN 141
6.4 Summary 142
Acknowledgments 142
References 143
Part IV Industrial Separation 145
7 Adsorption 147
7.1 Principle of Adsorption 147
7.2 Adsorbents 148
7.2.1 Zeolite Adsorbents 148
7.2.2 Zeolite Synthesis 150
7.2.3 Forming Zeolite Powders into Usable Shapes 152
7.2.3.1 Pelletization of Zeolite Powder to Extrudate Adsorbent 153
7.2.3.2 Accretion of Zeolite Powder to Bead Adsorbent 153
7.3 Bulk Liquid-Adsorptive Separation 154
7.3.1 Modes of Liquid-Adsorptive Separation 154
7.3.1.1 Adsorption Isotherms 156
7.3.1.2 Pulse-Test Procedure 156
7.3.1.3 Breakthrough Procedure 157
7.3.1.4 Simulated Moving-Bed Operation: Sorbex Technology 158
7.3.2 Liquid-Adsorptive Separation Processes 159
7.3.2.1 Equilibrium-Selective Adsorption 159
7.3.2.2 Rate-Selective Adsorption 166
7.3.2.3 Shape-Selective Adsorption 168
7.3.2.4 Ion Exchange 169
7.3.2.5 Reactive Adsorption 170
7.4 Commercial Bulk Liquid-Adsorptive Process 170
7.4.1 Parex 170
7.4.2 MX SorbexTM 173
Acknowledgment 174
References 174
8 Distillation 179
8.1 Introduction to Distillation 179
8.2 Principles and Systems 181
8.2.1 Vapor–Liquid Equilibrium (VLE) Theory 181
8.2.2 McCabe–Thiele Diagram 182
8.2.3 Factors for VLE-Based Distillation Design 184
8.2.3.1 Relative Volatility 184
8.2.3.2 Activity Coefficient 185
8.2.3.3 Solubility 185
8.2.3.4 Azeotropes: Maximum Concentration Limit 185
8.2.3.5 Surface Area 186
Contents ix
8.2.4 Operating Modes of Distillation 186

8.2.4.1 Batch vs. Continuous 186
8.2.4.2 Tray vs. Packed Column 187
8.3 Distillation Columns and Trays 188
8.3.1 Flow Within the Distillation Column 188
8.3.2 Multicomponent Distillation 189
8.3.3 Tray Types: Sieve, Valve, Bubble-Cap 190
8.4 Distillation Packing Materials 192
8.4.1 Random Packing 193
8.4.2 Structured Packing 194
8.5 Commercial Examples 195
8.5.1 Crude Oil Distillation 195
8.5.2 Extractive Distillation 198
8.5.3 Azeotropic Distillation 198
8.5.4 Dividing Wall Column Distillation 198
8.5.5 Reactive Distillation 200
References 200
9 Membranes 203
9.1 Principle and Background 203
9.2 Types and Preparation of Membranes 204
9.2.1 Polymeric Membranes 205
9.2.1.1 Fundamentals of Polymeric Membranes 205
9.2.1.2 Preparation of Polymeric Membranes 207
9.2.2 Inorganic Membranes 213
9.2.2.1 Introduction to Microporous Crystalline Inorganic Membranes 214
9.2.2.2 Preparation of Microporous Crystalline Inorganic Membranes 215
9.2.3 Mixed-matrix Membranes 218
9.2.3.1 Introduction to Mixed-matrix Membranes 219
9.2.3.2 Preparation of Mixed-matrix Membranes 221
9.3 Membrane Modules 223
9.4 Membrane Reactors 225
9.5 Commercial Applications of Membranes 227
References 235
10 Absorption 241
10.1 Introduction to Absorption 241
10.2 Acid Gas Removal 243
10.3 Chemical Solvent 245
10.3.1 Amines Process 246
10.3.1.1 Monoethanolamine (Primary Amine) 248
10.3.1.2 Diethanolamine (Secondary Amine) 248
10.3.1.3 Methyldiethanolamine (Tertiary Amine) 248
10.3.2 BenfieldTM Process 249
10.4 Physical Solvent 251
10.4.1 Solvent Selection 252
x Contents
10.4.1.1 DMEPG (Dimethyl Ether of Polyethylene Glycol,

(CH3 O(C2 H4 O)n CH3 )) 252
10.4.1.2 MeOH (Methanol, H3 COH) 253
10.4.1.3 PC (Propylene Carbonate, C4 H6 O3 ) 254
10.4.2 Flow Schemes 254
10.4.3 SelexolTM Process 255
10.5 Additional Commercial Uses of Absorption 257
10.5.1 Glycol Dehydration 257
10.5.2 Organics from Air (Air Stripping) 259
References 259
11 Extraction Technology 261

11.1 Introduction 261
11.2 Extraction Processes 261
11.2.1 Liquid–Liquid Extraction (LLE) 261
11.2.2 Supercritical Fluid Extraction (SCFE) 263
11.2.3 Liquid–Liquid Extraction vs. Supercritical Fluid Extraction 264
11.3 Solvents 264
11.3.1 Selectivity 264
11.3.2 Recoverability 265
11.3.3 Immiscible 265
11.3.4 Density 265
11.3.5 Chemical Stability 265
11.3.6 Loading Capacity (La ) 265
11.4 Extractors 266
11.4.1 Simple Extractor 266
11.4.1.1 Sieve-tray Column 266
11.4.1.2 Packed Column 267
11.4.2 Mechanical Extractors 268
11.4.2.1 Rotary and Reciprocal Extractors 268
11.4.2.2 Centrifugal Extractors 268
11.4.3 Simple Extractors vs. Mechanical Extractors 269
11.5 Industrial Extraction 269
11.5.1 Hydrometallurgy Extraction 269
11.5.1.1 Zirconium/Hafnium Separation 270
11.5.1.2 Uranium/Plutonium Separation 272
11.5.2 Hydrocarbon Extraction 272
11.5.2.1 m-Xylene/Xylene Isomer Separation 272
11.5.2.2 Sulfolane Process 274
11.5.2.3 Merox Process 277
11.6 Summary 280
Acknowledgments 280
References 281
Index 283
xi
Foreword
Over the past 30 years, there has been a dramatic global expansion of the petro-
chemicals industry. The industrialization and economic growth of countries in
Asia, the Middle East, Latin America, and now Africa has raised the purchasing
power of billions of people and created surging demand for products that are
ultimately derived from petrochemical feedstocks. As a supplier of technology to
the petrochemicals industry, Honeywell UOP has been a willing participant in this
expansion and has worked with its customers around the world to deploy safe and
state-of-the-art processes for making the basic monomers and chemicals that go
into making the thousands of products that we all consume every day.
Dr. Santi Kulprathipanja has been at the heart of developing many of Honeywell
UOP’s technologies. In his very distinguished career, Dr. Kulprathipanja was named
as inventor of over a hundred patents, including many of the technologies described
in this book. While best known for his many inventions, Dr. Kulprathipanja was
always a strong advocate for developing close links between industry and academia,
and for many years he has also taught university classes around the world so that he
could share his knowledge with the next generation of engineers and chemists.
This book builds on Dr. Kulprathipanja’s highly regarded classes in petrochemicals
processing and separation technologies and contains additional contributions from
colleagues at Honeywell UOP who have joined him in his teaching and provided
additional insights from their own technical expertise. It should be a very useful
reference for chemical engineers or others who might be starting a career in the
petrochemicals industry, or in companies or institutions that serve that industry.
Gavin Towler
Chief Technology Officer and
Vice President of Research and
Development Honeywell UOP
xiii
Preface
This book will provide an overview of markets and uses for common petrochemical
building blocks and also a survey of the technology used to make the key petrochemi-
cal building blocks. The book will be written with a practical chemistry and chemical
engineering course work in mind, with the target audience being university profes-
sors, students, scientists, and engineers those that are in the field of petrochemicals
and petrochemical technology. The book will contain both critical information to
help the readers to understand how the technologies are used to produce petro-
chemicals, how the various petrochemicals are used, and examples and problems
designed to reinforce the learning about the basic science, engineering, and use of
petrochemicals.
The topics covered by the book include Chapter 1: Refining and Petrochemical
Process, Chapter 2: Petrochemical Markets, Chapter 3: Olefins and Synthesis Gas
Production Technologies, Chapter 4: Uses of Olefins and Synthesis Gas, Chapter 5:
Aromatic Production Technology, Chapter 6: Uses of Aromatics, Chapter 7: Adsorp-
tion, Chapter 8: Distillation, Chapter 9: Membranes, Chapter 10: Absorption, and
Chapter 11: Extraction. Highlights in each chapter can be found as given in the fol-
lowing content.
Refining and petrochemical process is introduced in Chapters 1 and 2. As we can
notice in the last 10 years or so, the combination of high demand for electric cars
and higher automobile engine efficiency in the future will mean less conversion of
petroleum into fuels. However, the demand for petrochemicals is forecast to rise due
to the increase in world population. With this, it is expected that modern and more
innovative technologies will be developed to serve the growth of the petrochemi-
cal market. Petroleum is refined/converted into intermediate refinery products. The
intermediate refinery products are then converted to fuels such as fuel gas, liquefied
petroleum gas, gasoline, jet fuel, kerosene, auto diesel, and other heavy products. In
addition, these petroleum intermediates can be further processed and separated into
products for petrochemical applications.
We start our journey toward a better understanding of petrochemical markets
through an exploration of the building blocks of the industry – those commodity
chemicals that are utilized in large volume – in Chapter 1. We introduce the
sources of these building blocks – predominantly from three separate but important
complexes: olefin complexes, aromatic complexes, and synthesis gas complexes.
xiv Preface
We also introduce the “interconnected” nature of these building blocks – the fact
that their sources and uses are frequently connected in which one building block is
co-produced with another. In Chapter 2, we begin to explore this interconnection in
more detail, providing some statistics for each building block. Many of the building
blocks have many different sources, are made as co-products with other building
blocks and chemicals, and the supply of these building blocks depends rarely just
on one of the building blocks. Because of this, the markets frequently get out of
balance – where either shortages or large surpluses of one building block are noted.
The beauty of these markets then comes through, as substitution and movements
between the building blocks occur to rebalance the supply and demand. The ability
to make many final products (different products) from one building block, as well as
the ability to use many building blocks to make one final product (different routes)
make the petrochemical markets extremely complex and extremely robust.
Olefins are key building blocks of chemical industry. Light olefins, ethylene and
propylene, are the key components used in the manufacture of many polymers
and chemicals. Butylene is also in high demand for further processing to pro-
duce methyl-tert-butyl ether (MTBE) and alkylate for gasoline blending. Heavy
normal-olefins (C10 –C14 ) are the main feedstock component of linear alkylbenzene
(LAB), the most common raw material in the manufacture of biodegradable
household detergents.
Production and uses of olefins are documented in Chapters 3 and 4. They
include the olefins productions using steam cracking, paraffin dehydrogenation,
and methanol to olefins technologies as well as the uses of the olefins to produce
downstream products such as polyolefins, polyvinyl chloride, mono-ethylene glycol,
propylene oxide, acrylonitrile, LAB, MTBE, and alkylate. Chapters 3 and 4 also
cover the manufacturing technologies for producing syngas, which is used in large
volumes for the production of commodity chemicals such as hydrogen, ammonia,
and methanol.
Aromatics are important building blocks for petrochemical intermediate and
derivative products. Chapters 5 and 6 describe various methods and commercial
processes to produce aromatics such as benzene, toluene, and xylene (BTX).
Naphtha reforming is still the most efficient and sensible method to produce aro-
matics since direct dehydrogenation of naphthenes is favorable provided sufficient
temperature while valuable hydrogen co-product can be obtained at the same time.
Other catalytic processes include fluidified catalytic cracking, aromatization of
light naphtha, toluene methylation, aromatic isomerization, transalkylation, and
disproportionation. Noncatalytic methods include steam cracking, pyrolysis, and
coke-oven benzole production.
Essential materials such as nylons, polyesters, plastics, pesticides, coatings, and
resins are primarily derived from basic aromatic building blocks, including BTX.
These materials have strong mechanical strength and thermal stability with aro-
matic ring in the backbone. Electrophilic aromatic substitution allows functional
groups to attach to the aromatic molecules that make them flexible monomers for
further production or synthesis of polymers for construction and automotive indus-
tries or consumer and pharmaceutical products.
Preface xv
The commercially established technologies can be reviewed and leveraged to

develop new methods, including new intermediates or advanced materials that can
be derived from renewable resources. A circular economy for plastics has started
and continues with aims to promote recycle and eliminate wastes; new technologies
will emerge to adapt to challenges and opportunities ahead.
Separation processes play a significant role in the chemical industry. These
processes include adsorption, distillation, membranes, absorption, extraction,
and crystallization. The most utilized technique is distillation, which is used in
90–95% of all processes in the chemical industry. However, due to technological
and economic limitations, distillation is not always feasible. For the remaining
applications, other techniques must be explored. Adsorption is one of the more
desirable options because of its extensive and flexible applications. Particularly,
in the bulk industrial liquid adsorption that utilizes zeolite as an adsorbent and
counter-current simulated moving-bed technology, which is known as “Sorbex tech-
nology.” Examples of these bulk liquid industrial adsorption are the separations of
para-xylene from other xylene-isomer, meta-xylene from other xylene-isomer, olefin
from paraffin, n-paraffin from other paraffin-isomer, fructose from other-sugar.
Because zeolite adsorbent and Sorbex technology are critical to the adsorption
separation performance processes, Chapter 7 will detail zeolites’ synthesis, zeolites’
features, and Sorbex technology operation mechanism. Besides, other factors that
are critical to the adsorption performance will be detailed.
As mentioned in Chapter 7, the most utilized separation technique is distillation,
which is used in 90–95% of all processes in the chemical industry. Distillation is
documented in Chapter 8. Chapter 8 covers the commercial development of distil-
lation as well as the theory behind distillation, including vapor–liquid equilibrium
(VLE) as well as the construction of McCabe–Thiele diagrams to aid in the design
of distillation columns. Our attention then turns to a comparison of both batch and
continuous distillation as well as the internals of distillation towers (tray and packed
columns).
The strengths and weaknesses of each type of column, tray, and packing are cov-
ered in detail. The criteria by which trays are selected vs. packed columns are delin-
eated. The performance and cost aspects of sieve, valve, and bubble-cap trays are
compared in terms of the operating range, operating capacity, cost, and pressure
drop. We then turn our attention to column-packing materials and discuss the evo-
lution of random column packings from simple Raschig rings to more elaborate
IntaloxTM designs. Finally, our summary of packing is completed with a survey of
structured packing with a discussion of the strengths, challenges, and opportunities
within this segment of distillation technology.
Chapter 8 concludes with a survey of commercialization examples starting with
crude oil distillation (both atmospheric and vacuum methods). While distillation
methods have been developing for thousands of years, innovation continues to
drive the introduction of new commercial methods. We finish Chapter 8 with
overviews of extractive distillation, azeotropic distillation, dividing wall column
technology, and reactive distillation. Each of these methods continues to evolve
from a cost–performance perspective.
xvi Preface
Membrane technology has been a critical enabler for effective and efficient gas,
vapor, and liquid separations for many years. The membrane technology is described
in Chapter 9. Polymeric membrane–based industrial separation processes have
evolved rapidly since the development of asymmetric integrally skinned polymeric
membranes by Loeb and Sourirajan in 1960s and the development of thin-film
composite (TFC) membranes by Cadotte in 1980s. Inorganic membranes have the
advantages of high thermal stability, high chemical resistance, and resistance to
harsh environments over polymeric membranes. Therefore, inorganic membranes
not only have been commercially successful in some liquid separation applications,
but also have been studied for catalytic membrane reactor, energy storage, and
other applications. Further advancement in reducing the cost and improving the
processability and reproducibility of the inorganic membranes is still critical for
inorganic membranes to be adapted for new commercial applications. Innovation
in membrane materials with long-term stability, chemical resistance and specific
to applications, membrane fabrication methods, module design, new membrane
process design, and process technology integration has been instrumental to the
widespread industrial adoption of membrane technology, and it will continue to
enable new membrane-based separations and reactions. Chapter 9 will discuss sev-
eral aspects of membranes, including principles, types and preparation, membrane
modules, membrane reactors, and commercial applications.
Absorption technology is given in Chapter 10 to detail method of separation com-
monly used in the oil and gas and chemical industries for removal of acid contami-
nants such as H2 S (hydrogen sulfide) and CO2 (carbon dioxide) from hydrocarbons;
this separation strategy is often employed in the purification, transport, and storage
of natural gas. As natural gas usage has increased over the past 20 years due to its
lower CO2 emissions vs. coal, the removal of acid gases has become more critical
and helps protect large capital investments from corrosion.
There are two primary classes of absorption we cover in Chapter 10: chemical
absorption and physical absorption. The strengths and weaknesses of each method
are discussed. At low partial pressures of the component (generally low concentra-
tion), a chemical solvent is generally preferred while at high partial pressures of
the component to be removed, physical solvent has advantage due to its high load-
ing capability. The criteria by which a specific amine is selected are reviewed: (i)
Required purity of cleaned gas stream for further processing or end use, (ii) Com-
position of feed gas, (iii) Utility requirements, process costs, and (iv) Corrosion and
solvent degradation/solvent losses.
We then turn our attention to CO2 removal using an inorganic chemical solvent
(K2 CO3 in the UOP BenfieldTM process). The Honeywell UOP Benfield Process is
a thermally regenerated cyclical solvent process using an activated, inhibited hot
potassium carbonate solution to remove CO2 , H2 S, and other acid gas components.
An overview of the physical solvent options is included: SelexolTM (DMPEG,
Dimethyl Ether of Polyethylene Glycol), RectisolTM (Methanol), and Fluor (PC).
Cost and performance factors such as solvent volatility, contaminant solubility,
and process conditions such as pressure and temperature often guide the choice of
solvent in physical absorption systems.
Preface xvii
Finally, we review a couple of key commercialization examples for absorption:

Glycol dehydration and the stripping of organic contaminants from air.
Though there are many extraction processes, Chapter 11 will focus on
liquid–liquid extraction (LLE) and supercritical fluid extraction (SCFE). LLE
is widely used in industry for hydrometallurgy and hydrocarbon extraction. Appli-
cations in gas–liquid extraction called absorption are used commercially to extract
carbon dioxide and sulfur-species (acid gas) from natural gas. Absorption is covered
in Chapter 10.
In Chapter 11, critical materials and processes that impact extraction per-
formance, including solvent selection and type of extractors, will be described.
Selecting the right solvent and type of extractors is key to achieving high purity
and efficiency in the extraction process. After introducing the fundamentals of
extraction, the advantages of extraction over other separation technology will
be demonstrated with the following commercial applications: (i) Hydrometal-
lurgy such as extraction of Zr from Hf, and U from Pu; and (ii) Hydrocarbon
separation such as aromatics extraction from nonaromatic, m-xylene from other
xylene-isomers, mercaptan removal, and SCFE of caffeine from coffee.
Honeywell UOP Santi Kulprathipanja

Des Plaines, Illinois, USA James E. Rekoske
17 June 2020 Daniel Wei
Robert V. Slone
Trung Pham
Chunqing Liu
1
Section I
Introduction
3
Refinery and Petrochemical Processes
1.1 Introduction
The combination of high demand for electric cars and higher automobile engine effi-
ciency in the future will mean less conversion of petroleum into fuels. However, the
demand for petrochemicals is forecast to rise due to the increase in world popula-
tion. With this, it is expected that modern and more innovative technologies will be
developed to serve the growth of the petrochemical market.
In a refinery process, petroleum is converted into petroleum intermediate prod-
ucts, including gases, light/heavy naphtha, kerosene, diesel, light gas oil, heavy gas
oil, and residue. From these intermediate refinery product streams, several fuels
such as fuel gas, liquefied petroleum gas, gasoline, jet fuel, kerosene, auto diesel, and
other heavy products such as lubricants, bunker oil, asphalt, and coke are obtained.
In addition, these petroleum intermediates can be further processed and separated
into products for petrochemical applications.
In this chapter, petroleum will be introduced first. Petrochemicals will be intro-
duced in the second part of the chapter. Petrochemicals – the main subject of this
book – will address three major areas, (i) the production of the seven cornerstone
petrochemicals: methane and synthesis gas, ethylene, propylene, butene, benzene,
toluene, and xylenes; (ii) the uses of the seven cornerstone petrochemicals, and (iii)
the technology to separate petrochemicals into individual components.
1.2 Petroleum
Petroleum is derived from the Latin words “petra” and “oleum,” which means “rock”
and “oil,” respectively. Petroleum also is known as crude oil or fossil fuel. It is a
thick, flammable, yellow-to-black mixture of gaseous, liquid, and solid hydrocarbons
formed from the remains of plants and animals. Over millions of years, this organic
mixture of plants and animals was subjected to enormous pressure and heat as layers
of earth further buried them over time. The organic mixture changed chemically and
broke down into hydrocarbons. Because of the time it takes to form petroleum, it is
referred to as a nonrenewable energy source.
Modern Petrochemical Technology: Methods, Manufacturing and Applications, First Edition.

Santi Kulprathipanja, James E. Rekoske, Daniel Wei, Robert V. Slone, Trung Pham, and Chunqing Liu.
© 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.
4 1 Refinery and Petrochemical Processes
1.2.1 Forms of Petroleum

Petroleum can be in gas, liquid, and solid forms of hydrocarbons. The forms of
petroleum are functions of pressure, temperature, and other surrounding conditions
such as rock and soil. Under low surface pressure and temperature conditions,
lighter hydrocarbons such as methane, ethane, propane, and butane occur as gases,
while heavier hydrocarbons such as octane, benzene, xylenes, and paraffin wax are
in the form of liquids or solids. However, in an underground oil reservoir where
pressures are high, the proportions of gas, liquid, and solid depend on subsurface
conditions. It is expected that there are higher percentages of solids and liquids
compared to gas at higher pressures.
1.2.1.1 Gaseous Petroleum: Gaseous Petroleum Can Be Defined as Natural Gas

or Shale Gas
Natural gas is a mixture consisting primarily of methane, but also commonly
includes varying amounts of other higher alkanes, and small percentages of carbon
dioxide, nitrogen, hydrogen sulfide, and helium [1]. Natural gas is found in deep
underground reservoirs with sandstone and associated with other hydrocarbon
reservoirs.
Shale gas has the same composition as natural gas, except that it is found trapped
within shale formations [2]. Shale gas wells depend on hydraulic fractures to allow
the gas to flow from the rock and have become an increasingly important source of
natural gas in the United States since the start of this century. In 2000, shale gas pro-
vided only 1% of U.S. natural gas production; by 2010 it was more than 20%. Today,
the United States is the largest shale gas producer in the world [3], while China is
estimated to have the world’s largest shale gas reserves [4].
1.2.1.2 Liquid Petroleum: Liquid Petroleum Can Be Separated into Light Crude
Oil and Heavy Crude Oil
Light crude oil has a low viscosity and is a liquid at room temperature. These prop-
erties make light crude oil easy to pump and extract. It is composed of short-chain
paraffins, which are straight and branched-chain hydrocarbons. Because virgin light
crude oil comprises these short chains, it does not have to be heavily refined to pro-
duce gasoline. Typically, the chain-length range in gasoline is 4–12 carbons, making
light oil a desirable crude to be processed in a refinery. Approximately 30% of the
world’s petroleum reserves is light oil crude.
Heavy crude has a higher viscosity, but is still a liquid at room temperature. It usu-
ally contains more sulfur, nitrogen, and other contaminants than light oil. To refine
heavy oil to gasoline, it must be cracked and treated to remove the contaminants.
It requires more energy input and cost. Beyond the need for additional refinement,
heavy crude also needs additional extraction techniques to recover oil from the wells.
These techniques include stream stimulation to make the oil less viscous and the
injection of air into the wells to create fires that burn heavier hydrocarbons and
degrade them into lighter, more easily pumped compositions.
The transport of heavy oil requires the addition of diluting agents, particularly
in pipelines. The other major drawback to heavy crude is its environmental impact
because it contains sulfur and heavy metals, both of which must be removed.
1.2 Petroleum 5
Heavy metals are often toxic and their removal from crude presents disposal issues.
Sulfur, which may be as high as 4.5 wt%, is corrosive to pipeline metal and refinery
components.
1.2.1.3 Solid Petroleum: Bitumen

Bitumen is a naturally occurring solid hydrocarbon that generally contains a mix-
ture of large polycyclic aromatic with exceptionally low hydrogen-to-carbon ratios.
Bitumen crude is found impregnated in sedimentary rock, which is mainly located
in Canada. This rock is often referred to as oil sand or tar sand. Bitumen will not flow
unless heated or mixed with lighter crudes to reduce viscosity. Most bitumen extrac-
tion processes require some level of physical mining as opposed to pump extraction
as with gaseous and liquid crude oil.
When crude oil prices increase, it can become profitable to upgrade bitumen to
synthetic crude for fuel. Bitumen can also be used for a number of other purposes,
including mortar between bricks, base material for statues, and waterproofing.
Today, the largest use of bitumen is roofing applications and at a lower concentration
in roadways.
1.2.2 Composition of Petroleum

In general, oil wells produce predominantly liquid petroleum (crude oil), with some
dissolved natural gas. Because the pressure is lower at the Earth’s surface than deep
underground, some of the natural gas will come out of the solution and be recovered.
Furthermore, at the higher deep underground temperatures and pressures, the dis-
solved gas may contain heavier hydrocarbons such as pentane, hexane, and heptane.
Once these heavier gases reach the lower pressure and temperature surface condi-
tions, they will condense out of the gas phase to form condensate, which has similar
appearance to gasoline and is similar in composition to some volatile light crude oils.
Petroleum is a mixture of a very large number of different hydrocarbons. The four
most commonly found categories of molecules are alkanes, cycloalkanes, aromatics,
and more complicated chemical compounds such as asphaltenes. Petroleum also
contains trace organic compounds of sulfur, nitrogen, oxygen, and metals such
as iron, nickel, copper, and vanadium. Each petroleum source has a unique mix
of molecules that define its physical and chemical properties, such as color and
viscosity. The molecular composition of crude oil varies widely depending on the
source. The chemical elements [5] in crude oil can vary as shown in Table 1.1.
Table 1.2 shows the percent ranges of the four key hydrocarbons in crude oil [6].
Alkanes – also called paraffins – are saturated hydrocarbons with straight or
branched chains containing only carbon and hydrogen. They have the general
formula Cn H2n+2 . Examples of alkanes in crude oil are hexane, decane, and
2,2,4-trimethylpentane (iso-octane).
Cycloalkanes – or, naphthene – are saturated hydrocarbons that have one or more
carbon rings to which hydrogen atoms are attached. The general formula of naph-
thene is Cn H2n . Examples of cycloalkanes are cyclopentane and cyclohexane.
Aromatics are unsaturated hydrocarbons that have one or more planar six-carbon
rings called benzene rings to which hydrogen atoms are attached. They have the
Table 1.1 Chemical elements in crude oil.
Element Percent
Carbon 83–85
Hydrogen 10–14
Nitrogen 0.1–2.0
Oxygen 0.05–1.5
Sulfur 0.05–6.0
Metals <0.1
Table 1.2 Hydrocarbon in crude oil.
Hydrocarbon Percent
Alkanes 15–60
Naphthenes 30–60
Aromatics 3–30
Asphaltenes Remaining
general formula Cn H2n−6 . Examples of aromatics are benzene, toluene xylenes,

ethylbenzene, 1,2-dimethylbenzene, naphthalene, and biphenyl.
Sulfur compounds in crude oil are typically found as mercaptans such as ethylmer-
captan and cyclic sulfides such as tetrahydrothiophen and benzothiophen.
There are two types of nitrogen-based compounds in petroleum. The first type is
basic nitrogen molecules that have an extra lone pair of electrons to facilitate reac-
tions. Basic nitrogen compounds are pyridine, quinolone, and phenanthridine. The
second type of nitrogen compounds is nonbasic. In this type of nitrogen, the lone
pair of electrons with the nitrogen is delocalized in the ring structure making the
molecule less reactive. Indole and carbazole are two nonbasic nitrogen compounds.
Some crude oils contain trace organic oxygen compounds such as benzoic acid
and phenol.
Metals such as Fe, Ni, Cu, V, Ca, Mg, Hg, As, and Na are generally found in
petroleum. Metals can be in the form of organic or inorganic salts. For example,
V can be found in an organic porphyrin structure and CaCl2 , MgCl2 , and NaCl
suspended or dissolved as entrained water inorganic salts. Metals in petroleum
must be removed or neutralized before processing, as they cause corrosion and
fouling to refinery equipment and poison of catalysts.
1.2.3 Petroleum Refinery

An oil refinery is an essential part of the downstream side of the petroleum
industry. Petroleum refineries are very large industrial complexes that involve many
1.2 Petroleum 7
processing units and auxiliary facilities such as utility units and storage tanks. Some
modern petroleum refineries process as much as 900 000 barrels per day of crude
oil. According to Oil and Gas Journal, a total of 636 refineries operated worldwide
at the end of 2014 for a total capacity of 87.75 million barrels per day of crude oil.
Reliance Industries’ Jamnagar Refinery in Gujarat, India, currently is the world’s
largest oil refinery. Each refinery has its own unique arrangement and combination
of refining processes largely determined by the refinery location, desired products,
and economic considerations.
Conversion of heavy oils to useful products requires breaking large molecules into
smaller ones. The breaking or cracking of large molecules can be accomplished by
one or more combinations of heat, pressure, and chemical reaction. In the process,
harmful or unwanted compounds such as metals, sulfur, nitrogen, and oxygen are
also removed. The conversion of heavy oils into useful products requires many sep-
aration and chemical reaction processes. These processes are briefly characterized
as follows:
● Desalination: This is the first unit in the refinery complex. Salts from crude oil are
extracted before it enters the atmospheric distillation unit.
● Crude oil distillation (atmospheric distillation): The desalted crude oil is separated
into various fractions for processing in downstream units.
● Vacuum distillation: The residue oil from the bottom of the crude oil distillation
unit is further distilled at a vacuum pressure well below atmospheric.
● Naphtha hydrotreater: The hydrotreater desulfurizes, denitrogenizes, and deoxyg-
enizes naphtha using hydrogen from the atmospheric distillation unit. The naph-
tha must be treated before sending the stream to the catalytic reformer unit.
● Catalytic reformer: The reformer converts the hydrotreated naphtha into refor-
mate, which has a higher content of aromatics and cyclic hydrocarbons. The end
products are high-octane gasoline and para-xylene aromatics, which is a critical
petrochemical in making PET.
● Distillate hydrotreater: This unit, similar to the naphtha hydrotreater, removes sul-
fur, nitrogen, and oxygen from distillates (such as diesel) and other units within
the refinery after atmospheric distillation.
● Fluid catalytic cracker (FCC): The FCC process is used to upgrade the heavier,
higher boiling-point fractions from the crude oil distillation by converting them
into more valuable lighter and lower boiling-point products.
● Hydrocracker: The hydrocracker uses hydrogen to upgrade heavy residual oils
from the vacuum distillation unit by thermally cracking them into more valuable
lighter and lower-viscosity products.
● Merox: The Merox process desulfurizes LPG, kerosene, and jet fuel by oxidizing
mercaptans to organic disulfides.
● Coking: Delayed cokers, fluid cokers, and flexicokers are used to crack very heavy
residual oils into gasoline and diesel fuel leaving petroleum coke as a residual
product.
● Alkylation: Alkylation is a process that uses sulfuric or hydrofluoric acid or an
ionic liquid as a catalyst to produce high-octane components for gasoline blending.
Sulfar plant Sulfur

Saturate Sulfur
vapour recovery removal Fuel gas
Liquefied
petroleum gas
Light straight-run naphtha
Normal paraffins
Sulfur removal Gasoline

Heavy
naphtha Naphtha Catalytic
hydrotreater reformer
Atmospheric Kerosene Hydrotreating

unit Kerosene/jet fuel
crude
distillation Diesel Hydrotreating
column unit Diesel
Atmospheric gas oil

Gas oil
Desalter
Crude oil Atmospheric residue

Fuel oil
Asphalt Asphalt
oxidation
Figure 1.1 Topping and hydroskimming refinery.
For example, iso-butane and butylenes can be converted into iso-paraffin alkylate,
which is a very high-octane gasoline.
● Dimerization: Dimerization is used to convert olefins into higher-octane gasoline.
For example, butenes are dimerized into isooctene, which is subsequently hydro-
genated to form isooctane.
● Isomerization: This process converts linear molecules such as normal pentane to
higher-octane branched molecules for blending into gasoline. Isomerization units
also are used to convert linear molecules such as normal butane into isobutane for
use in alkylation units.
Oil refining is the process where crude oil is transformed into more desirable and
valuable products using the processes described above. Refineries often are classified
by the number and type of process units available for transforming crude oil into
petroleum products.
Topping refineries are the least complex refineries and are used to separate the
crude oil into its constituent petroleum products by atmospheric distillation; naph-
tha is produced, but no gasoline. Hydroskimming refineries are also one of the sim-
plest types of refineries used in the petroleum industry and are equipped with atmo-
spheric distillation, naphtha reforming, and additional treating processes to produce
gasoline. Important to note, a hydroskimming refinery produces a surplus of fuel
with a relatively unattractive price and demand. Figure 1.1 shows schematic topping
and hydroskimming refineries.
The next refinery classification in complexity level is a medium-conversion
refinery. A medium-conversion refinery is shown in Figure 1.2. To design
a medium-conversion refinery, unit operations such as vacuum distillation,
thermocracking, fluid catalytic cracking, and asphalt oxidation are added to the
hydroskimming refinery. This added level of complexity allows conversion of fuel
oil to light distillates and middle distillates.
1.2 Petroleum 9
Sulfar plant Sulfur

Saturate Sulfur
vapor recovery removal Fuel gas
Hydrogen plant Hydrogen
Liquefied
petroleum gas
Hydrotreater C5/C6 Normal paraffins
isomerization

Heavy
naphtha Naphtha Catalytic
hydrotreater reformer
Kerosene Hydrotreating
Atmospheric Kerosene/jet fuel
crude unit
unit Diesel
column
Atmospheric Catalytic
condensation
gas oil
Fluid catalytic Unsaturated Sulfur
cracking unit vapor recovery removal
Alkylation
Visbreaking Sulfur
thermal cracking removal C4 isomerization
Desalter MTBE
Vacuum
distillation
Crude oil column
Residual fuel oil
Atmospheric
residue
Asphalt Asphalt
oxidation
Figure 1.2 Medium-conversion refinery.
Sulfar plant Sulfur

Saturate Sulfur
vapor recovery removal Fuel gas
Hydrogen plant Hydrogen
Liquefied
petroleum gas
Hydrotreater C5/C6 Isomer Normal paraffins
isomerization separation

Heavy
naphtha Naphtha Catalytic Aromatics Benzene/toluene/
hydrotreater reformer separation
xylene
Kerosene Hydrotreating
Atmospheric Kerosene/jet fuel
crude unit
unit Diesel
column
Hydrocracking Catalytic
Atmospheric unit condensation
gas oil
Fluid catalytic Unsaturated Sulfur
cracking unit vapor recovery removal
Alkylation
Visbreaking Sulfur
thermal cracking removal C4 isomerization
Desalter MTBE
Coking
Vacuum
distillation
Crude oil column
Demetallizing Residual fuel oil
Atmospheric
residue Coke
Dewaxing Lubricant Lubricants
compounding
Asphalt Asphalt
oxidation
Figure 1.3 High-conversion refinery.
The most complex type of refinery is the high-conversion refinery, which is shown
in Figure 1.3. The high-conversion refinery adds more inter-related processes such
as hydrocracking, coking, demetallizing, and dewaxing. In particular, a coking unit
adds cracking capability for conversion of fuel oil into distillates and petroleum coke,
allowing high-efficiency conversion of the crude oil feedstock into higher yields of
more valuable products.
1.2.4 Petroleum Products

Petroleum products are complex hydrocarbon mixtures, in contrast to petrochem-
icals, which are a collection of well-defined usually pure chemical compounds
Table 1.3 A breakdown of the petroleum products made from

a typical barrel of U.S. oil.
Petroleum products Percent
Gasoline 46
Jet fuel 9
Diesel and other fuel 26
Heavy fuel oil 4
Lubricants 1
Asphalt 3
Other products 11
Source: US Energy Information Administration, 31 May 2018.
derived from petroleum refineries. Depending on the composition of the crude oil
and the demands of the market, refineries can produce different types and pro-
portions of petroleum products. Petroleum products are usually grouped into four
categories: light distillates (LPG, gasoline, naphtha), middle distillates (kerosene,
jet fuel, diesel), heavy distillates, and residuum (heavy fuel oil, lubricating oils, wax,
asphalt). The largest share of petroleum products is fuels, which include single
or blended combinations of hydrocarbons to provide gasoline, jet fuel, diesel fuel,
heating oil, and heavier fuel oils. Heavier fractions also can be used to produce
asphalt, tar, paraffin wax, lubricating, and other heavy oils. Table 1.3 summarizes a
breakdown of the petroleum products made from a typical barrel of U.S. oil.
Refineries also produce chemicals that are used to produce polymers, detergents,
and other useful consumer products. Since petroleum contains sulfur-containing
molecules, elemental sulfur is often produced as a petroleum byproduct. Carbon,
in the form of petroleum coke, and hydrogen gas also are produced as petroleum
byproducts. For example, hydrogen often is used as an essential reactant for other
oil refinery processes such as hydrocracking and hydrodesulfurization.
1.3 Petrochemical Building Blocks

The industrial use of petrochemicals has focused throughout its history on roughly
seven main building blocks, often called intermediates, which are obtained from
both natural gas and petroleum processing. These seven basic building blocks are:
● Ethylene: Also known as ethene and possessing the formula C2 H4 , it is a color-
less flammable gas. Because it contains only two carbon atoms, it is the simplest
carbon–carbon double bond building block.
● Propylene: Also known as propene or methyl ethylene, propylene has the formula
C3 H6 . It also is an excellent C=C double-bond building block allowing unique
reactions and functionality because of the presence of both olefinic and aliphatic
carbon–carbon bonds.
● C4 mono- and diolefins: This grouping is composed of butenes (also called

butylenes) having the formula C4 H8 , and butadiene having the formula C4 H6 .
While we group these building blocks together for simplicity, they are often used
for vastly different chemistry purposes and final products.
● Benzene: The simplest of the aromatic ring structures having a formula of C6 H6
makes benzene a very important organic molecule in the petrochemical industry.
● Toluene: Also known as toluol, toluene has the same ring structure as benzene but
has one methyl group substituted on the ring. Its formula is C7 H8 .
● Xylenes: While strictly defined as the three dimethyl substituted benzene isomers
(1,2-, 1,3-, and 1,4-dimethylbenzene), the petrochemical industry often classes
ethylbenzene into the same grouping called xylenes. This is because the dimethyl
benzenes and ethylbenzene have the same formula (C8 H10 ) and very similar boil-
ing points. The result is that these compounds are nearly always found together
in raw material sources.
● Synthesis gas: The oddest grouping in the building blocks, synthesis gas is a mix-
ture of H2 and carbon monoxide (CO) in a molar ratio typically ranging from
around 1.0 to slightly higher than 2.0. Synthesis gas is used to produce many final
products and intermediates due to the flexible nature of the building blocks.
Chemical structures of these basic building blocks are shown in Figure 1.4.
From these seven building blocks, and adding abundant, readily available chemi-
cals such as air, pure oxygen, or pure nitrogen, the petrochemical industry has built a
wide variety of value-added compounds that have become essential to modern soci-
ety. For example, in a modern automobile, plastics and petrochemical compounds
H
H H3C CH3
H H H C CH3 CH3
C C H
C C CH2 H3C
H H
H H
H3C CH3
CH2
Ethylene Propylene 1-Butene 2-Butene Isobutene
CH3
CH2
H2C H2C C
H2 CO
1,3-Butadiene 1,2-Butadiene
Synthesis gas
C4 mono- and di-olefins
CH3
CH3 CH3
CH3
H3C CH3
CH3
Benzene Toluene p-Xylene m-Xylene o-Xylene Ethylbenzene
Xylene isomers
Aromatic hydrocarbons
Figure 1.4 Chemical structure of common petrochemical building blocks.

account for more than 50% of the volume of materials used. Despite this large frac-
tion of the volume, the petrochemical parts make up only about 10% of the vehicle
weight [7]. Clearly the trend of replacing automotive parts with lighter-weight pieces
based on petrochemicals will have to continue as lighter, more energy-efficient, and
electric vehicles become more popular.
The seven building blocks are mostly present as components of raw materials
(petroleum, natural gas) described earlier. Indeed, most of the building blocks had
been identified in the 1800s as components of these raw materials. For example,
Faraday [8] was reportedly the first to successfully isolate and characterize benzene
in 1825. However, these compounds were needed to be present in much larger quan-
tities, and isolated as intermediates for subsequent chemical reactions, for them to be
of use as building blocks. With petroleum and natural gas as starting materials, the
industry began to invent methods to obtain these building blocks in vast quantities.
Today, the unit operations (reactions, separations, heat transfer, etc.), which are
combined in systematic ways to make petrochemical building blocks in large quanti-
ties, often are referred to as petrochemical complexes. Although very few petrochem-
ical complexes are exactly alike, we can categorize these complexes into three basic
types.
● Olefin complexes: The complexes make predominantly the olefins ethylene, propy-
lene, and butenes. It is the primary technology for the product of olefins in steam
cracking [9], but other technologies such as dehydrogenation and methanol to
olefins technologies [10, 11] are gaining traction. In a steam cracking–based olefin
complex, the primary product is ethylene; the other products are formed in far
lower quantities. These complexes will sometimes be “fully integrated” – meaning
they have full recovery of propylene, butenes, and even the aromatics formed. Oth-
ers will recover only the ethylene and propylene and sell the remainder of the
liquid products to others for recovery. A sample block-flow diagram for an olefins
complex using steam cracking of hydrocarbons is shown in Figure 1.5.
● Aromatics complexes: These complexes make mostly aromatics hydrocarbons, and
often make most of the product as para-xylene for use in the synthesis of puri-
fied terephthalic acid, PTA [12]. These complexes usually employ a technology to
reform the naphtha cut of petroleum into aromatic rings, and then use a series
of technologies to interconvert the aromatic rings into the desired products, with
Ethane recycle
Methane
Feed Pyrolysis Quench and Cryogenic Demethanizer Deethanizer C2 splitter Ethylene

treatment section compression cooling
Feed
Fuel gas
Pyrolysis Hydrogen
gasoline Depropanizer C3 splitter Propylene
Fuel oil (to aromatics extraction)
Propane recycle
C4 products
Debutanizer (to separation or sales)
Pyrolysis gasoline
(to aromatics extraction)
Figure 1.5 Block flow diagram of typical naphtha-based olefins unit.

Raffinate
Sulfolane Benzene
extraction
H2 LE Toluene
C9 aromatics
H2 LE H2 LE
THDA
Naphtha CCR H2 LE
RS Clay BC TC HA
HDT platform
Tatoray
C10 +
ortho-Xylene
H2 LE Light ends
C7A
Clay XC O-X Parex Isomer

DC7
Clay para-Xylene
Figure 1.6 Typical aromatics complex.
para-xylene being typically the most desired. A typical plant-flow diagram for an
aromatics complex is shown in Figure 1.6.
● Synthesis gas complexes: In these complexes, coal, natural gas, or some other
hydrocarbon source is steam reformed into a synthesis gas mixture (H2 and CO).
This synthesis gas then is converted into a variety of different products, including
methanol or ammonia by reacting the hydrogen with nitrogen recovered from
air. While methanol is a very-large-volume commodity chemical, ammonia
production is even larger due to the use of ammonia in agricultural fertilizers.
While the types of reformers and the layout vary significantly based on feedstock,
product ratios, and reformer technology, a typical layout for a synthesis gas plant
making methanol from natural gas is shown in Figure 1.7.
Together, these three types of complexes form most of the conversion of raw materi-
als into the building blocks, which support the petrochemical industry.
Many authors have tried to point to the “birth” of petrochemicals – something
which has always resulted in controversy. This is because petrochemicals have
existed for as long as our natural resources have existed, but were often masked
Water
Water Steam
pretreatment generation Purge gas (fuel)
Purge gas (fuel)

De-
sulfurization
Natural gas Syngas Methanol
Methanol
Reformer cleanup/ Compression reactor Product methanol
purification
adjustment section
Water
(to waste water treatment)
Figure 1.7 Block flow diagram of typical methanol plant.

N2
CO2
Ammonia Urea
Hydrogen
Synthesis gas Poly(vinyl

acetate)
CO Acetic acid Vinyl acetate
Poly(vinyl
alcohol)
Methanol
Natural gas
H2
MTBE LDPE,HDPE,LLDPE
Methane Formaldehyde
Chlorine Poly(vinyl
EDC Vinyl chloride chloride)
Ethane Ethylene oxide Ethylene
Ethylene glycol Polyester
alpha-Olefins
Steam Propylene
Olefins
Propane cracking Ethanol
Polyisobutylene
Butylenes Isobutane
Polypropylene
Butanes Propylene oxide Polyols
Butadiene
Condensate Acrylonitrile Polyacrylonitrile

Pyrolysis
gasoline Polybutadiene
Cumene Phenol Butadiene-styrene

Polystyrene
Acetone
ABS
Ethylbenzene Styrene
Methyl methacrylate
n-Butane
Maleic
anthydride
SB latex and rubber
Benzene Cyclohexane Adipic acid Nylon 66
Benzene Caprolactam Nylon 6

Crude oil Naphtha Aromatics Toluene diisocyanate Polyurethane
Toluene Toluene diamine
Alkyd resins
ortho-Xylene Phthalic anhydride Plasticizers
meta-Xylene Isophthalic acid Unsaturated
Xylenes
polyester resins
para-Xylene Purified terephthalic acid
Polyester
Dimethyl terephthalate fibers and resins
Figure 1.8 Illustrative example of products from building blocks.

Chemicals
Ethylene dichloride Ethylene oxide
Vinyl chloride Ethylene glycol
Ethylbenzene Vinyl acetate
Styrene Ethanol
H H Oligomers α -Olefins
C=C Fatty alcohols
H H
Ethylene
Polymers Low-density polyethylene

High-density polyethylene
Linear low-density polyethylene
Poly vinyl acetate
poly vinyl chloride
Figure 1.9 Examples of products from ethylene.
or shrouded in the complexity of the complicated mixtures they were contained

in as described earlier. However, one of the first commercial complexes built
for the large-scale production of a petrochemical was in the early 1900s in West
Virginia. With its start in 1927, the Blaine Island plant of the Carbide and Carbon
Chemicals Corporation (later known as Union Carbide and recently acquired by
Dow Chemical) was the first truly integrated petrochemical plant. It produced
predominantly ethylene through a process, which we now call steam cracking of
hydrocarbons [13].
For any building block, intermediate or product to be useful, it must have an end
use, which adds values for consumers. For example, the largest end use of ethylene
(almost 70%) is in the manufacture of polyethylene. Polyethylene is broadly used in
food packaging, in engineered polymers, and in many other applications. With such
important and growing applications, it is no wonder that the production of ethylene
has been growing at slightly higher than GDP (about 1.5× the global GDP growth
rate on average from 1990 through 2017 [14]).
One aspect that makes the petrochemical building blocks particularly interest-
ing, however, is the versatility to make multiple different products from the same
molecule. This versatility is illustrated in Figure 1.8, which shows the major build-
ing blocks and only a few of the primary products, which can be produced from
them. For example, the same ethylene molecule used to make polyethylene can also
be used to make ethylene glycol, which is used as a coolant, as an ingredient in cos-
metics and foodstuffs, and as a key component of polyester fibers and bottles.
Indeed, dozens of other petrochemicals, fine and specialty products can be derived
from ethylene; the major products and further intermediates produced by ethylene
are shown in Figure 1.9. The same is true for the other olefins, and for most of the
other building blocks, making the evaluation petrochemical markets and economics
a particularly rich and exciting field. We will discuss this in depth in Chapter 2.
References
1 Background. Naturalgas.org (archived from the original on 9 July 2014).

2 U.S. Energy Information Administration. Eia.gov (retrieved on 6 August 2013).
3 Beschloss, M. U.S. now World’s leading natural gas producer. Desert Sun
(retrieved on 4 November 2014).
4 Organic origins of petroleum. US Geological Survey (archived on 27 May 2010).
5 Speight, J.G. (1999). The chemistry and technology of petroleum (3rd ed., rev. and
expanded ed.). New York: Marcel Dekker. pp. 215–216, 543. ISBN: 0824702174.
OCLC 44958948.
6 Hyne, N.J. (2001). Nontechnical Guide to Petroleum Geology, Exploration, Drilling,
and Production, 2e, 1–4. Tulsa, OK: Penn Well Corp. ISBN: 087814823X. OCLC
49853640.
7 American Chemistry Council, Economics & Statistics Department (July 2018).
Plastics and Polymer Composites in Light Vehicles.
8 Faraday, M. (1825). On new compounds of carbon and hydrogen, and on certain
other products obtained during the decomposition of oil by heat. Philosophical
Transactions of the Royal Society 115: 440–466.
9 (2017). CEH: Ethylene, Chemical Economics Handbook Series. IHS Markit.
10 IHS Markit Note (11 April 2018). The tight link between methanol and olefins.
https://cdn.ihs.com/www/blog/20180411-Chem-MTO-Infographic.pdf (accessed
03 September 2020).
11 IHS Markit: Methanol demand growth driven by methanol-to-olefins,
China demand. Hydrocarbon Process (13 June 2017). https://www
.hydrocarbonprocessing.com/news/2017/06/ihs-markit-methanol-demand-
growth-driven-by-methanol-to-olefins-china-demand.
12 (2018). CEH: Mixed Xylenes, Chemical Economics Handbook Series. IHS Markit.
13 Modern life made possible. West Virginia Department of Commerce, Press Cen-
ter. https://urethaneblog.typepad.com/my_weblog/2014/03/the-first-ethylene-
cracker-in-the-us-was-in-wv.html.
14 Eskew, B. (2018). US petrochemicals: the growing importance of export markets.
IHS Markit Presentation at EIA Energy Conference (04 June 2018). https://www
.eia.gov/conference/2018/pdf/presentations/blake_eskew.pdf.
17
Petrochemical Markets
2.1 Introduction
Describing the web of production and uses of petrochemicals often evokes a number
of different adjectives: complex, integrated, confusing are among the most common
descriptors. In our experience, however, through the introduction and explanation
of a few simple economic and technical concepts, the business end of petrochemical
production and use can be quite simple and exciting. The key to unpacking this com-
plexity is the concept of the core building blocks for modern petrochemicals, which
we introduced in Chapter 1. As a reminder, these core building blocks are:
● Ethylene
● Propylene
● C4 Mono- and Di-Olefins
● Benzene
● Toluene
● Xylenes
● Synthesis Gas
We call them the building blocks of the modern petrochemical industry because,
quite simply, nearly 90% of the volume of petrochemical products produced and
traded are derived from these seven compounds or mixtures [1].
The building blocks are often referred to as primary or commodity chemicals. Sim-
ply put, a commodity chemical is often described as a chemical material made at
large scale and which is used and traded globally. While it is true that the build-
ing blocks are, indeed, commodities, they are not the only commodity chemicals. So
why does this definition matter? Because commodities have commonalities, which
make them easy to trade – to sell between companies, and even between regions. For
commodity chemicals, these commonalities are often chemical specifications – what
concentration, what level of impurities, and what form (liquid, solid, solution, etc.)
are the typical specifications used in the commodity chemical industry. With this set
of properties to define the commodity, one company can buy the commodity from
one or more different suppliers at different times and have confidence that they can
use that commodity to make the end products they want to make.
18 2 Petrochemical Markets
In Chapter 1, we also briefly described how these building blocks are made. In
addition to how they are made, how much and where (geographically) the building
block is made is also important to understand. Collectively, this information makes
up what can be called sources of supply, or just source or supply in short. Of course,
there is no use making a chemical or building block if there is no use for it, or demand
for the chemical. Sometimes this use is also called a sink, where the term “sink”
represents a use or consumption of the chemical in question. The spatial and time
balance and trade of these chemical commodities is done through markets. While
often not physical locations like they historically have been, markets are the pro-
cesses and institutions through which goods change hands. Matching sellers (sup-
ply) with buyers (demand) is the key function of the market. Often times, today, this
is done electronically without the need for the buyer and seller to meet or, in some
cases, even know each other. Through this process, the commodities, which have
common specifications, can be traded between companies, regions, or even globally.
While not the intent of this chapter, it is important that the reader understand
some basic fundamentals regarding the economic principle regarding the balance of
supply and demand [2]. Briefly, the process by which supply and demand are bal-
anced is through the market-clearing price [3]. Generally speaking, suppliers will
want to supply more of a commodity if the price is high (subject to capacity and
other constraints, of course). Conversely, the higher the price of a commodity, the
less likely people who use the commodity (the demand side) will want to buy it
and the more likely they will seek alternatives. This opposite response to the price
of a commodity means that there will be a price at which supply and demand are
perfectly balanced – the so-called market-clearing price. This is important because
supply of and demand for chemical commodities: (i) follows these same principles;
and, (ii) will fluctuate throughout time dependent upon many variables, including:
● raw material availability
● labor availability
● success or problems with production operations
● success or problems with operations, which consume the chemical
● logistics problems
● and many other factors
Figure 2.1 shows how this simple yet powerful economic theory works in an illus-
trative example.
Now that we have introduced the building blocks, the concept of commodity mar-
kets, and the principles of supply and demand, we can start to take a closer look at
the market situation with respect to our major building blocks of the modern petro-
chemical industry.
2.2 The Market for Ethylene

One of the largest volume building blocks and commodity chemicals is also one of
the smallest in molecular size – ethylene. At more than 162 million metric tons per
Supply
Price (example, dollars per unit)
Surplus
Market-claring price
at Q0, P0
Shortage
Demand
Quantity (example, number of units)
Figure 2.1 Illustration of supply, demand curves, and market-clearing price.
annum (estimated) consumption in 2019, ethylene is the largest of the hydrocarbon

building blocks and second only to synthesis gas in terms of total annual consump-
tion [1]. Ethylene is an exceptionally versatile molecule, being the starting point for
hundreds, if not thousands, of final polymers and chemical compounds. Owing to
the size and complexity of the markets and manufacturing processes for ethylene,
production is concentrated in the hands of large, integrated refining and petrochem-
ical companies. In 2019, it has been estimated that approximately 46% of the world’s
ethylene production came from assets of the top 10 producing firms [4].
Figure 2.2 shows a standard capacity-demand graphic for ethylene. The bar
on the left-hand side illustrates the proportion of ethylene produced from the
primary feedstocks. The vast majority (more than 95%) of the world’s ethylene
results from the thermal conversion of hydrocarbons through the steam-cracking
process. We estimate about 40.4% of ethylene globally came from the cracking of
naphtha, while 37.6% from that of ethane in 2019 [4]. The balance of the thermally
derived ethylene came from LPG (Liquefied Petroleum Gas) (15.3%) and heavier
hydrocarbon fractions (2.7%) [4]. Historically, naphtha has been the preferred
feedstock for thermal cracking for several reasons. First, naphtha is readily available
and transportable, often as the byproduct of producing heavy transportation fuels
like diesel and jet fuel. Naphtha is too light to be converted into these heavier fuels
and, instead, is converted into gasoline, which is typically a lower-volume demand
transportation fuel outside of the United States. Second, until recently, the economic
value of light naphtha was quite low. This changed with the rapid expansion of
the polyester markets as light naphtha is the primary feedstock for the production
of xylenes, most notably, p-xylene. Third, there are a number of very significant
side-products formed in the thermal conversion of naphtha and heavier products
MTO/CTO – 4.0%
Gasoil – 2.7% Others – 9%
Global ethylene market – capacity and demand
Ethylbenzene – 5%
250 000 0.90
LPG – 15.3%
0.89 Ethylene dichloride – 9%
200 000 0.88

Ethylene oxide – 15%
Capacity, demand (kMT/a)

0.87
Operating rate (%)

150 000 0.86
Ethane – 37.6% 0.85
100 000 0.84
0.83
50 000 0.82 Polyethylene – 62%
0.81
Naphtha – 40.4% 0.80
08
09
20
21
22
23
24
25
12
13
14
15
16
17
18
19
10
11
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Capacity Demand Operating rate
2019 sources 2019 uses
Figure 2.2 Global supply and demand dynamics for ethylene. Source: Based on data from
IHS Markit and internal Honeywell UOP Estimates.
such as butadiene and benzene. Historically, these were a significant source of

revenue for an olefins complex; recently, the co-product value of these chemicals
has reduced, putting pressure on the economics of integrated naphtha crackers.
Ethane has always been used as a feedstock in places of abundance, such as the
Middle East. It has recently returned to prominence in the United States with large
ethane supplies becoming available as a result of the shale gas boom [5]. With the
rapid addition of ethane-based ethylene in the United States, it is likely that ethane
will equal or surpass naphtha as the predominant feedstock in the next five years
[5, 6].
The remaining amount of ethylene (about 4% in 2019) is produced primarily from
methanol or coal in the methanol/coal to chemicals process (M/CTO) [4]. While this
technology is a small fraction of the total global ethylene production, it has been
increasingly utilized in regions where coal and natural gas are more abundant than
liquid hydrocarbons. Indeed, there has been a substantial growth in this nonconven-
tional ethylene source in the past 10 years, driven by significant capacity additions
in China. A significant co-product of the M/CTO process is propylene.
The middle portion of Figure 2.2 shows the capacity and demand balance (the
bars) for ethylene globally. In 2019, the world’s capacity to produce ethylene
reached approximately 185.9 million metric tons per annum. This capacity has
been increasing at average rate of about 3–4% annually for the past several years.
Capacity addition is not smooth, however, as many years, even a decade, can be
required to properly plan, permit, construct, and begin operation of a large ethylene
complex. Indeed, the standard for a “world-scale” ethylene plant has increased
to 1.0 million metric tons per annum for a naphtha-based plant or 1.5 million
metric tons for an ethylene-based plant. About three to four of these plants are
needed annually to keep up with the demand increases alone. When one factors
in geographic growth due to abundant, cheap feedstocks (e.g. ethane in the United
States) and higher demand rates due to concentrated economic growth (e.g. India
and China), it is likely that between four and six plants will be added annually going
forward. Clearly, both the production of ethylene and the construction of ethylene
complexes are large and growing businesses.
As of 2019, the capacity to produce ethylene exceed the demand for ethylene by
about 12–13%. This is evident in Figure 2.2 by the difference in the size of the bars
representing capacity and demand. It is also shown in the line, which is the ratio of
demand to capacity and is frequently termed the operating rate. This is the global
operating rate needed for supply from the available capacity to meet the demand.
As such, it does not represent the operating rate at any individual plant; it is simply
a mathematical construct, which allows the balance of supply and demand. Despite
being grounded only in this theoretical balance, it is a useful number to watch.
Production plants of any kind, ethylene plants included, are rarely able to operate
at full rates for 365 days a year – plant upsets, weather, equipment breakdowns,
logistics challenges all conspire to reduce the theoretical limit to a practical limit.
As such, when the global operating rate approaches a critical value, it is frequently
a strong indication that more capacity – either through expanding existing facilities
or building new ones – must occur for supply and demand to remain balanced.
In the chemical industry, the “bellwether” operating rate value is typically around
90–92%. When operate rates are expected to exceed this value, supply disruptions
become more likely. As can be seen from Figure 2.2, the global operating rate for
ethylene has been rising for most of the past decade as the demand has recovered
from the global recession of 2008–2009, consuming some of the overcapacity, which
resulted from a construction boom in the Middle East [7] that occurred around the
same time. Our forward expectation is that the global operating rate will remain in a
tight band between 87% and 90%, which is a historically high level for the operating
rate. This is likely to mean regional supply pressures can occur more frequently due
to plant impairments as has been experienced in recent years [8].
The right-hand side of the Figure 2.2 illustrates the uses of ethylene. Production
of polyethylene in various grades dominates the demand for ethylene, with about
62% globally. A significant fraction of the polyethylene produced (over 50% by most
estimates) is used in nondurable customer packaging, including food packaging and
bottles; sadly, in the United States, less than 30% of this material was recycled in
2017 [9]. Single-use and nondurable uses of plastics were under significant scrutiny
in 2019 [10], and the continued trajectory for polyethylene is unclear. However, the
short-term offers few alternatives to the cost and convenience of polyethylene. It will
be interesting to see how the environmental and the economic challenges around
nondurable plastics are resolved in the coming years.
The production of ethylene for use in ethylene oxide manufacture follows at a
distant 15%. Ethylene oxide has three predominant uses: as a freezing-point depres-
sant and heat-transfer improver for mixtures in water in automotive antifreeze, in
surface-active nonionic surfactants, and as a key raw material in the production of
polyesters. The production of polyester represents the majority of the demand for
ethylene oxide. Ethylene dichloride (9%) and ethylbenzene (5%) account for most of
the remaining share of ethylene consumption, being used in the vinyls and styrenics
plastics chains, respectively. The remaining demand (9%) is spread across a number
of small products throughout the plastic and chemical value chains.
US Western Europe
Ethane Propane Butane Naphtha Gas oil M/CTO Other Ethane Propane Butane Naphtha Gas oil M/CTO Other
Middle East Northeast Asia
Ethane Propane Butane Naphtha Gas oil M/CTO Other Ethane Propane Butane Naphtha Gas oil M/CTO Other
Figure 2.3 Regional variation in feedstock sources for ethylene. Source: Figure created
using data adapted from IHS Markit [11].
Throughout this section, we have used global facts and figures to illustrate the fea-
tures of the market for ethylene. It is important to note, however, that the regional
markets for ethylene (and others of the building blocks) can be quite different from
geography to geography. While it is beyond the scope of this book to describe all of
these regional variations, we have constructed Figure 2.3 to illustrate the differences
that can occur. Figure 2.3 compares the sources of ethylene in four large produc-
ing regions – the United States, Western Europe, the Middle East, and Northeast
Asia – in 2016 [11]. The United States and the Middle East show large percentages of
ethylene produced from ethane because of the availability of the feedstock. Western
Europe and China show very low proportions from ethane, but quite high propor-
tions of ethylene produced from naphtha. In addition, China shows about 7% of its
ethylene produced from coal and methanol to olefins technology – a pathway not
used in any of the other regions. These significant regional differences lead to a rich
and varied global market.
2.3 The Market for Propylene
At almost 114 million metric tons estimated to be produced in 2019 [12], propylene
is one of the top commodity chemicals produced as an intermediate for a large
variety of chemicals compounds. Because of the existence of both aliphatic and
olefinic bonds in the molecule, the chemical versatility of this molecule is sub-
stantial; indeed, propylene is the smallest such molecule with both key types of
carbon–carbon bonds. The chemical reaction’s versatility and ability to synthesize
so many unique compounds stem directly from the presence of these highly varied
bonds. Though not as concentrated as ethylene, more than 35% of the world’s
propylene production comes from the top 10 producing companies. Because of the
High severity FCC/other – 4.0% Global propylene market – capacity and demand
MTO/CTO/MTP/CTP – 4.6% Acrylic acid/others – 9.5%
200 000 0.84
Metathesis – 3.8% Cumene – 5.9%
Propane dehydro (PDH) 12.1% 180 000 Oxo alcohols – 5.9%
0.82
160 000 Acrylonitrile – 5.6%

140 000 Propylene oxide – 8.2%
0.80
Operating rate (%)

Refinery sources – 28.7% 120 000
100 000 0.78
80 000
0.76
60 000
Polypropylene – 66.9%
40 000
0.74
Steam cracking – 46.8% 20 000
08 0.72
09
20
21
22
23
24
25
12
13
14
15
16
17
18
19
10
11
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
2019 sources 2019 users
Figure 2.4 Global supply and demand dynamics for propylene. Source: Based on data from
IHS Markit and internal Honeywell UOP Estimates.
complexity and size of propylene production and markets, these companies also
tend to be the world’s largest integrated oil refining and petrochemical companies.
Propylene is marketed typically in three major grades – minor, more specialized
grades do exist – referred to as refinery, chemical, and polymer grade. They are listed
in order of increasing purity. Refinery grade contain typically between 50% and 75%
propylene, while chemical grade is typically 92–98% propylene. Polymer grade is
always equal to or greater than 99.5% propylene, but there are multiple different
impurity specifications (on methylacetylene and propadiene, for example), which
alter the grades slightly. Our standard view for the capacity and demand balance
for all propylene grades is shown in Figure 2.4. Starting from the left-hand bar, we
can see that the sources of propylene are much more varied than the sources of
ethylene. Only about 47% of the propylene produced in 2019 was through steam
cracking, as described in Chapter 1 and in Section 2.2. This compares to more than
95% of the ethylene produced coming from the same process. Indeed, propylene
is the primary “co-product” of the thermal conversion of hydrocarbons to make
ethylene. However, as hydrocarbon feedstocks for the production of ethylene shift
toward abundant ethane from heavier hydrocarbons, the amount of propylene
produced as a co-product declines. This is illustrated in Table 2.1, which presents the
relative production of key intermediates from the thermal conversion of different
hydrocarbon species.
A significant portion of the world’s sources of propylene comes from refineries.
The so-called fluid catalytic–cracking process, or FCC unit, which focuses on the
production of naphtha for motor fuels, also produces an off-gas byproduct, which
contains a substantial quantity of propylene. From Figure 2.4, we can see that
this source accounts for about 29% of the world’s supply of propylene. Adding
the production of propylene from FCC (byproduct of gasoline manufacture) and
steam cracking (byproduct of ethylene manufacture) shows that slightly more
than 75% of the world’s propylene results in recovery of this value material as a
byproduct of manufacturing routes primarily focused on other compounds (gasoline
and ethylene). This is one manner in which the markets become complex and
linked – for example, a producer who is focused on the ethylene market must also
Table 2.1 Relative amounts of byproducts, including propylene, from thermal cracking of
various feedstocks.
Feedstock Ethane Propane Butane Naphtha
Feed amount 1243 2264 3361 3230
Products
Ethylene 1000 1000 1000 1000
Propylene 20 347 376 516
Butylenes — 64 126 320
Pyrolysis gasoline — 177 274 730
Fuel gas 150 719 623 813
Source: Adapted from Nexant Inc. [5].
watch and make assessments of the propylene market to understand the feasibility
of new capacity investment, or of increasing or decreasing operating rates.
The remainder of the source of propylene, slightly less than 25% of the world’s
supply, comes from processes, which are described as “on-purpose” propylene pro-
duction. As the name implies, these processes have the production of propylene as
their primary purpose. These processes have gained substantially in popularity in
recent years, driven by a separation in growth rate between propylene and ethylene.
The consumption of propylene has typically been growing at a slight multiple to the
production of ethylene – for example, in recent years, propylene demand growth
rates have been between 4.0% and 4.5%, while ethylene demand has been increasing
between 3.0% and 3.5%. Additionally, gasoline demand has been declining, resulting
in lower operating rates for FCC units and lower production of propylene. This has
put pressure on propylene markets as recently as 2012–2014, resulting in the need
for additional sources of propylene. These on-purpose methods are frequently seen
as more expensive routes, but often simplify the business of propylene as they result
in fewer (often times only one or two) products, which need to be marketed.
Most significant in the growth of these on-purpose methods has been propane
dehydrogenation (also called PDH) technology, which now accounts for more
than 10% of the world’s production of propylene. This process starts with propane,
an abundant and relatively cheap raw material, and makes only propylene as the
product, thereby simplifying the business operations and diversifying both the raw
materials and the production routes to propylene [13]. Significant new capacity
for PDH units has been built in China using important propane from the United
States and the Middle East. In addition to PDH, methanol to olefins (MTO) and
coal to olefins (CTO) units produce propylene besides ethylene. This production
technology has also seen a growth in adoption in China in recent years with its
large reserves of coal [14].
The middle portion of Figure 2.4 shows the global capacity and demand balance
(the bars) for propylene. The world’s capacity to produce propylene has increased
steadily in recent years between 4.0% and 4.5%, with capacity reaching approxi-
mately 138.1 million metric tons in 2019 [12]. As with ethylene, capacity addition
is not perfectly smooth due to the size of the plants and the timing of construction
decisions; however, the capacity curve is much smoother owing to the smaller capac-
ity of typical propylene plants, particularly the on-purpose plants. As with ethylene,
the difference in the size of the bars for the capacity to produce and the demand for
propylene is accounted for by the operating rates of the production plants, shown in
Figure 2.4 as the line. The operating rate for propylene has remained quite constant
over the last 20+ years, from a low of about 76% during the global financial crisis
of 2008–2009 to a high of about 83–85% from 2004 to 2007. Because of the demand
growth and the relatively stable operating rate, the environment for new investment
in propylene production has remained strong and the market has remained quite
healthy over this same time.
Finally, the bar on the right-hand side of Figure 2.4 shows the approximate dis-
tribution of uses of propylene in 2019. Like with ethylene, the large majority of
propylene produced (66.9%) gets consumed for the production of various grades of
polymer, in this case, polypropylene. Other significant uses of propylene include the
production of propylene oxide (8%), acrylonitrile (6%), and a variety of other chem-
ical compounds, including cumene/phenol, oxo alcohols, acrylic acid, and superab-
sorbent polymers. While some polypropylene ends up in single-use plastics (bottle
caps, closures, some food packaging), a larger percentage of polypropylene ends
up in durable goods (carpets, other fibers, engineered plastics, etc.). This makes
polypropylene likely to be less subjected to the scrutiny placed on single-use plastic
materials than, for example, polyethylene.
The production of propylene has seen dramatic shifts in raw material uses and
geographic distribution in recent years, and these shifts are likely to continue [12].
Most of these shifts have been caused by observable and predictable macroeconomic
trends. As we have previously mentioned, the shift in feedstocks for the steam crack-
ing of hydrocarbons from heavier materials (e.g. naphtha) to lighter hydrocarbons
(e.g. ethane) has resulted in a lowering of the propylene produced as a byproduct
of ethylene (see Table 2.1). Adding to this decline in traditional supply has been a
decline in gasoline demand, which results in less propylene produced as a byproduct
from the FCC unit. Occurring simultaneously has been the broad industrialization
of China and the increased demand of the materials produced from propylene (e.g.
polypropylene) to be available in China. The resulting increase in production of
propylene in China has been enormous, as seen in Figure 2.5. Indeed, since 2013,
China has overtaken North America as the region with the highest capacity to pro-
duce and, by 2021, will be nearly double the production capacity of the next largest
region (see Table 2.2, [15]) – all occurring in a span of less than 10 years. Much of
this capacity expansion has been enabled by the combination of increased propane
supply from shale exploration in the United States as well as the recent boom in ship-
ping availability [16]. It is these types of shifts that make the petrochemical markets
dynamic, exciting, and, at times, frustrating.
2013 2019
North America
South America
Europe (WE+CE)
Russia and CIS
108.9 M MT/a 138.1 M MT/a
MEA
SE Asia + India
China
NE Asia example, China
Figure 2.5 Regional distribution of propylene capacity – 2013 vs. 2019.
Table 2.2 Regional propylene capacity changes – 2013–2021.
2013 2016 2019 2021
North America 23 500 24 100 25 600 26 100

South America 4 000 4 000 4 000 4 400
Europe 19 500 19 500 19 500 19 900
Russia and CIS 2 200 2 700 3 400 3 700
MEA 10 900 12 500 13 800 13 800
SE Asia and India 11 900 12 500 14 500 17 100
China 19 400 30 100 38 400 44 100
Other NE Asia 17 500 17 800 18 900 19 300
2.4 The Market for C4 Olefins and Diolefins
Like the markets for ethylene and propylene, which are large and have widely var-
ied end products for which they are intermediates, the C4 olefins and dienes are
produced in slightly smaller quantities but also have large and widely varying uses.
However, there is one complication – while we group these C4 olefins and diolefins
for convenience, there are really three separate compounds we need to consider:
n-butenes, isobutene, and butadiene. While the end uses of the three separate com-
pounds are quite different, the sources of these compounds are often the same or at
least highly integrated as we will see in the discussion below. For simplicity, we will
split the three classes of compounds into two distinct markets: butylenes, comprising
normal and iso C4 olefins, and butadiene.
In 2019, the world’s capacity to produce butylenes was approximately 80.1 million
metric tons per annum, broken down into 46.3 million tons per year of n-butenes,
and 33.8 million tons per year of i-butenes [17]. These butylenes are produced
through three main refinery and chemical processes: FCC (as we saw with propy-
lene in Section 2.3), steam cracking (as with both ethylene and propylene), and the
direct conversion of isobutane to isobutene through the butane dehydrogenation
2.4 The Market for C4 Olefins and Diolefins 27
Global butylene market – capacity and demand

Chemical uses, 20.0%
Others, 21.2% 100 000 0.64
90 000 Other fuel uses, 0.7%

0.62
BDH, 15.0% 80 000

70 000
0.60 Ethers, 35.7%
Operating rate (%)

60 000
Steam cracking, 18.4%
50 000 0.58
40 000
0.56
30 000
FCC, 45.4% 20 000 Alkylate, 43.6%
0.54
10 000
0.52
12
13
14
08
09
20
20 1
22
20 3
24
25
15
16
17
18
20 9
10
11
2
1
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Figure 2.6 Global supply and demand dynamics for butylenes.
process (analogous to the PDH process for propylene production). In 2019, the
capacity to produce butylenes was roughly estimated as 45.4% FCC, 18.4% steam
cracking, 15% BDH (butane dehydrogenation), and 21.2% from a variety of other
smaller chemical processes (as byproducts) [17]. This information on the capacity
to produce butylenes is summarized in Figure 2.6.
The demand for butylenes is quite a bit smaller than the world’s capacity to
produce, with 2019 seeing a global butylene demand of approximately 49.0 million
metric tons [17]. This demand is split nearly equally between normal (23.9 million)
and isobutenes (25.1 million). Interestingly, we can also segment the butylene
demand into demand for fuels (approximately 39.7 million metric tons in 2019)
and demand for further petrochemical production (approximately 9.4 million
metric tons). Butylenes are used in fuels in predominantly two different ways. The
olefins can undergo alkylation or dimerization chemistry to produce a hydrocarbon
product of longer chain length and a higher degree of branching. This results in
a higher-octane value for the molecule, which is beneficial for gasoline internal
combustion engines [18]. The second method involves reacting the olefins with an
alcohol to make an ether, for example, methyl tertiary butyl ether (MTBE), ethyl
tertiary butyl ether (ETBE), or tertiary amyl methyl ether (TAME). These ether fuel
additives are prized for both their octane values, which are typically higher than
pure hydrocarbons, but also because they add oxygen to the gasoline, which has
been shown to improve combustion leading to lower levels of air pollutants [19].
In addition to fuel uses, the butylenes are used as a raw material for the pro-
duction of a number of chemical compounds, including propylene (through mixed
ethylene/butylene metathesis), 1-butene (for polymer additives), monomethyl acry-
late (MMA), polyisobutylene (PIB), sec-butyl alcohol/methyl ethyl ketone (MEK),
to name a few. Because the end use typically has a specific chemical requirement
(for example, MTBE is made from isobutylene, not n-butylenes), the estimation of
a capacity utilization is quite complex and depends strongly on the process and the
ratio of n- and isobutenes produced by the process. As a result, we have shown both
the fuel and chemical demand levels and the n- and isobutene demand breakdowns
on the right-hand side of Figure 2.6 [17].
It is tempting to attempt to separate out the different butylene compounds and do
a market balance on each butylene component. While this can be approximated at
Table 2.3 Butylenes produced from steam cracking and refinery sources (thousands of
metric tons per annum) for selected years.
Steam crackers FCC
Year n i Ratio n i Ratio
2000 2715 4124 0.658 11 056 7820 1.414

2005 3295 4935 0.668 9 875 8608 1.147
2010 2684 4825 0.556 8 051 8591 0.937
2015 2895 5376 0.539 12 845 9115 1.409
2020 4562 5745 0.794 14 746 9423 1.565
Source: Data and format adapted from IHS Markit [20].
the aggregate level, it is difficult to break down the specific sources of butylenes. For
example, what is the number of metric tons per annum of n-butene produced from
steam cracking? The answer to this, even at an approximate level, would require a
very detailed analysis involving factors (at a minimum) such as the feedstock and
capacity of all the steam-cracking units, and the extent to which the units recover
the butylenes produced. This is well beyond the intention of this chapter. Instead,
we can look at the aggregate production of butene isomers from the various sources
and derive some conclusions.
Table 2.3 shows the reported annual production of the two primary categories
of butylenes, n-butene and isobutene, for selected years (2000, 2005, 2010, 2015,
and an estimate for 2020) from both steam-cracking operations and from refiner-
ies [20]. A couple of interesting trends can be observed. First, the amounts of C4
olefins declined from 2000 to 2010 and have since begun to rebound. This is true
for both FCC and steam-cracking sources. The decline in C4 olefin production in
steam cracking corresponds to the period of time of a large shift in feedstock eco-
nomics favoring lighter feeds like ethane and propane. During this period, despite
increasing supply of ethylene from steam cracking, the amount of C4 olefins pro-
duced declined because the feedstocks got much lighter on average. However, once
all the units, which can change to lighter feedstocks, have done so, then additions
of ethylene production also mean additions of C4 olefin production, and this is seen
in the later years. For the FCC units, the production numbers were quite stable dur-
ing the period of 2000–2010 and then began to increase substantially from 2010 to
2020. The increase since 2010 corresponds to a period of time when alkylate for fuels
became increasingly valuable from refineries due to global clean fuel standards [21]
resulting in more alkylate (and therefore C4 olefin) production.
Our second observation is that the ratio of normal to iso olefins resulting from
the steam-cracking operation is much more consistent over the years shown than
the ratio from the FCC unit. Steam cracking is a thermal process and there are
limited choices to be able to adjust the ratio of normal to iso, and largest among this
is the feedstock. As a result, the ratio remains much more stable as not all plants
can adjust feedstock readily. As discussed above, the shift to lighter feedstocks is
consistent with the slight increase in normal to iso ratio toward the end of the
period shown. The ratio of normal to iso olefin produced from the FCC process
from 2000 to 2020 varies widely from a low of 0.937 to a high of 1.565. There are
likely multiple reasons for this observation, but one involves changes to the FCC
process. The FCC process uses a catalyst, one which has been engineered over
the last several decades to produce more isobutane and isobutene as the branched
isomers are more valuable to the refinery for alkylation. This is most likely the
reason for the significant shift from 2000 to 2010.
It is important to note that there is currently substantial overcapacity for the
production of butylenes, but it is not homogeneously distributed between n-butenes
and isobutene. The demand for n-butenes corresponds to a utilization of capacity of
slightly higher than 50%, while the same estimate for isobutene shows a utilization
of close to 75%. Despite isobutene being the less thermodynamically stable isomer,
there is a stronger demand growth for isobutene due to its significant use across
many compounds – alkylate, fuel additives, and chemical uses – than there is for
n-butenes.
Although produced in much lower volume than butylenes, butadiene is a much
more valuable compound used exclusively in chemical production. In 2019, it is esti-
mated that global demand for butadiene will reach almost 13 million metric tons
against a capacity of approximately 13.4 million metric tons. The production of buta-
diene is quite tight – meaning that producing assets operate a quite high utilization
level. Indeed, the utilization of butadiene production assets runs typically in the
range of 90–95% (see Figure 2.7), which is quite high for petrochemical produc-
tion assets. There are several reasons for this. First, butadiene is produced almost
exclusively as a byproduct of the production of ethylene in crackers; butadiene is
only economically recovered from heavy feedstocks used in crackers (e.g. naphtha
and heavier feeds). As a result of being a byproduct and being a sought-after mate-
rial, the production of butadiene is quite lucrative, resulting in asset owners want-
ing to operate the assets as much as possible. Second, the crackers, which provide
the source, also have typically a very high utilization ratio because these are very
capital-intensive operations and return on investment requires high utilization.
While the only appreciable source of butadiene currently available is as a byprod-
uct from naphtha and heavier feed crackers, there are multiple uses of butadiene
and multiple competitive options, depending on the end market. Butadiene demand
growth has been rather unsteady over the past decade with historical year-over-year
growth varying from 1% to almost 5% from 2008 to 2018. Key drivers for this are
the aforementioned tightness of the supply. Consumers view this tightness as risk
and therefore look for alternative raw materials, which can be substituted, and this
substitution creates a rather “lumpy” demand profile. Tight supply also increases
the price at which people sell butadiene in the market, which opens up opportuni-
ties for competition as well. In addition, one of the key substitute products (natu-
ral rubber, see below) is an agricultural product promoted by several governments
in Southeast Asia in recent years as a means for rural economy improvement. In
Figure 2.7, the forward growth curve for butadiene has been estimated at about 4%
Global butadiene market – capacity and demand

20.00 1.00
Others, 16%
18.00
SB latex, 9%
16.00 0.95
Nitrile rubber, 6%
14.00
ABS, 12%
Operating rate (%)

12.00 0.90
10.00
SB rubber, 27%
8.00 0.85
6.00
4.00 0.80 PB rubber, 30%

2.00
0.75
2019 uses
12
13
08
09
14
20
21
22
23
24
25
15
16
17
18
19
10
11
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Figure 2.7 Global supply and demand dynamics for butadiene.
growth year-over-year, but this will likely vary substantially for the same economic
and competitive reasons mentioned above.
Butadiene is used as raw material for a number of engineered polymers and elas-
tomers. Figure 2.7 also shows the global demand by use of butadiene. Note that
there have been substantial shifts over relatively short periods of time in the uses.
Butadiene rubber and styrene-butadiene rubber (SBR) remain the largest uses, sup-
porting 30% and 27% of the global demand, respectively, in 2019. However, over the
last 10 years, the consumption toward butadiene rubber has overtaken SBR because
of increased uses of butadiene rubber in high-performance tires for the automotive
market [22]. Within tires, butadiene rubber is always used in conjunction with SBR
and often with natural rubber, an agricultural product from the rubber tree [23].
Indeed, adjustments to the produces’ formulae for production of tires often change
as the relative cost of synthetic polymers and natural rubber changes. Other sig-
nificant uses for butadiene rubber include footwear, wire and cable insulation, and
conveyor belts [22]. Besides SBR and butadiene rubber, butadiene is key raw mate-
rial for the production of a co-polymer known as acrylonitrile-butadiene-styrene, or
ABS. This is rapidly growing engineered polymer used for high-performance appli-
cations such as household appliances, automotive parts and interiors, and other
demanding applications.
The recovery of butylenes and butadiene from refineries and steam crackers is an
elegant, complicated, and highly integrated process. Figure 2.8 shows a simplified
generic flow scheme for the recovery processes resulting in butadiene, isobutylene,
and n-butylenes and a few critical products of the C4’s as well. Crude unsaturated C4
streams from steam crackers and refinery sources are shown on the left along with
isobutane, which can be converted into isobutylene through a dehydrogenation
process similar to that described early for on-purpose propylene production. In
general, only the unsaturated C4 streams from steam crackers will have a high
enough concentration of butadiene to warrant recovery, as shown in Figure 2.8.
When butadiene recovery is warranted, it is passed to a butadiene extraction unit,
Isobutane
Butane dehydro
MTBE MTBE cracking
High-purity isobutylene
FCC-based
crude C4ʹs Raff-1
Selective
hydrogenation Etherification
Raff-2 (normal butylenes)

Butadiene Isobutylene
extraction polymerization
Steam cracker-
based crude C4ʹs
Acetylene Polyisobutylene
hydrogenation
Figure 2.8 Generic recovery scheme for C4 unsaturates.
which uses solvent extraction technology to selectively absorb the butadiene, and
then strip the butadiene from the solvent using heat [24]. Licensors of this process
technology include BASF, McDermott, Nippon-Zeon, LyondellBasell, and others
[25]. If butadiene is not desired, or not present in high enough concentration to
be economically recovered, the unsaturated C4 stream is typically passed through
a selective hydrogenation unit, which will convert the diene and acetylenes into
mono-olefins. The effluent from the selective hydrogenation units will now look
very much like the effluent from the butadiene extraction units. The industry has
termed this stream “Raffinate-1” or “Raff-1.” This Raff-1 stream will be a combi-
nation of isobutylene, normal butylenes, and butanes (saturated C4 compounds).
It is quite different from the effluent of an isobutane dehydrogenation unit, which
is predominantly isobutylene and butanes, but for the purposes of Figure 2.8, the
streams can be considered together from this point forward.
The Raff-1 stream then will generally pass to an isobutylene utilization or recovery
section. Isobutylene can be used to make MTBE (by condensation with methanol),
which is a high-octane gasoline additive used in many countries, or to make PIB.
If ultra-high purity isobutylene is desired, one can make MTBE first to recover the
isobutylene from the Raff-1 mixture, and then crack the MTBE back into isobuty-
lene and methanol. This reaction is equilibrium limited, but by separation of the
isobutylene you can recycle the MTBE to extinction and have high recovery of very
high isobutylene. This is frequently practiced to derive high-quality butyl rubber.
Following the recovery or consumption of the isobutylene, the industry has termed
this effluent stream as “Raffinate-2” or “Raff-2.” This stream is now predominantly
normal butylenes contained in saturated butanes and can be used further processed
to make pure normal butylenes, alkylate for motor fuel applications, solvents, or
other products [17]. There are a number of different options and processes, which
can utilize these unsaturated C4 streams, and Figure 2.8 should be considered
only one example configuration. The versatility of these unsaturated C4 streams
results in a wide number of important specialty compounds with critical uses in our
economy.
2.5 The Market for Aromatic Compounds

Our next “building block” is actually a number of building blocks, which encom-
passes carbon numbers ranging from 6 to 8 but have an important functional group,
which distinguished them: all of these building blocks contain an aromatic ring. An
aromatic ring, also known as an arene compound, is an organic compound, which
has a planar “cyclic” structure of atoms [26]. For the purposes of our building blocks,
we can further limit this to compounds containing only carbon and hydrogen, and
further limit it to four distinct chemical compounds and their isomers:
● benzene, the simplest hydrocarbon aromatic, with molecular formula C6 H6 ,
resulting in the atoms being arranged in a planar structure;
● toluene, having formula C7 H8 , where one of the hydrogens in benzene has been
replaced with a CH3 methyl group;
● xylene, having the formula C8 H10 , where two of the hydrogens of benzene have
been replaced with CH3 methyl groups. There are three isomers defined by the
positional relationship of the methyl groups – ortho-xylene, meta-xylene, and
para-xylene, and
● ethylbenzene, closely related to xylene and having the same formula, but where
the benzene ring is substituted one time by a C2 H5 ethyl group instead of two
methyl groups.
Each of these four is characterized as containing a ring composed of six carbon
atoms in a planar structure. The first three in the list above are the primary aromatic
building blocks, which we will consider in this section. We will first consider these
compounds separately, and in Chapter 5 we discuss their interconversion through a
process known as transalkylation.
2.5.1 The Market for Benzene

Benzene was originally a byproduct of the coking process for the steel industry [27]
and is still today produced from this source. However, the market for benzene has
grown large enough to support its own processes for its production, which has decou-
pled benzene market dynamics from those of the steel industry to a large extent. In
2019, the world’s capacity to produce benzene exceed 65.7 million metric tons per
annum [28], which comes from five key sources. Reformate, which accounted for
about 39% of the world’s benzene capacity in 2019, can be described as the ben-
zene that is extracted from the products of naphtha reforming. Naphtha reforming
is a key unit operation in the manufacture of both transportation fuels and aro-
matic petrochemicals like benzene, toluene, and xylenes [28]. The second-largest
source of benzene capacity is pyrolysis gasoline (32%), or pygas for short, which is a
byproduct of steam cracking of hydrocarbons to produce ethylene. A large amount of
benzene (16%) can be produced in processes, which can be generically described as
transalkylation, which is defined here as the interchange of alkyl groups among the
aromatic rings. This is a key process in the efficient manufacturing of p-xylene [29],
where as much of 50% of the p-xylene in a complex can be produced through the
Global benzene market – capacity and demand

HDA, 4.2% Others, 8.9%
250.000 0.900
Coke oven, 8.0%
0.890 Nitrobenzene, 9.0%
Transalkylation, 16.5% 200.000 0.880 Cyclohexane, 10.7%

0.870
Operating rate (%)

150.000 0.860 Cumene, 21.3%
Pygas, 32.2% 0.850
100.000 0.840
0.830
Ethylbenzene, 50.1%
50.000 0.820
Reforming, 39.1%
0.810
0.800
08
09
20
21
22
23
24
25
12
13
14
15
16
17
18
19
10
2019 sources 11 2019 uses

20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Figure 2.9 Global supply and demand dynamics for benzene.
transalkylation process. Extraction of benzene from liquids resulting from the cok-
ing process of coal used for steel manufacturing remains a significant contributor to
the capacity to produce benzene (8%). The remainder of benzene production capac-
ity is attributed to a process called hydrodealkylation (HDA), accounting for about
5% of global capacity. In the HDA process, toluene or (rarely) other alkyl-substituted
aromatic compounds are converted – usually in a thermal process – into methane
and benzene. This capacity usually runs intermittently as the economics of the con-
version of toluene into benzene are frequently not favorable.
The fraction of global capacity production given in the paragraph above represents
the total capacity amounting to a possible production of approximately 65.55 million
metric tons of benzene in 2019. In reality, only about 49 million metric tons of ben-
zene were produced. Figure 2.9 shows the sources of supply of benzene in 2019, and
the inset table in Figure 2.9 shows the approximate operating rate for the five dif-
ferent key sources of supply (e.g. the actual production divided by the capacity to
produce) in 2019. As mentioned above, the lowest operating rate is observed for the
swing capacity, HDA-based benzene, at about 31%. The highest estimated operating
rate in 2019 is observed for the production of benzene as a byproduct of p-xylene
production through the transalkylation process (almost 89%). This is because, as a
byproduct, production economics are less dependent on benzene alone. Indeed, the
amount of benzene produced by a p-xylene production facility has decreased dramat-
ically over the years driven by economics strongly favoring p-xylene over benzene.
The total production of benzene in 2019 was approximately 48.97 million met-
ric tons. The production of benzene has been increasing at an average rate close
to 2% for nearly a decade and this growth rate is expected to continue to the next
several years [28]. Benzene is used for a wide variety of end products, including
a number of monomers, which are essential for production of polymers such as
styrenics, phenolics, and nylon. Almost 50% of the world’s demand for benzene is
driven by the need to produce ethylbenzene (through benzene alkylation with ethy-
lene [28]); ethylbenzene is a critical precursor for the production of styrene, which,
in turn, is dehydrogenated into styrene – the starting monomer for polystyrene and
related styrenic materials. The next largest demand for benzene is through the pro-
duction of cumene (about 21%), which is similarly produced by the alkylation of
benzene with propylene [28]. Cumene is further converted into phenol and acetone
through an oxidation process [30]. Phenol is a critical monomer for the production
of the phenolic polymer chain, which includes phenolic resins (through conden-
sation with formaldehyde), nylon through hydrogenation to cyclohexanone, poly-
carbonates, and bisphenol-A through condensation with acetone. All these plastics
compounds are in wide use in household products as well as engineered materials.
Cyclohexane is a key intermediate in the production of adipic acid and caprolactam,
both of which are critical starting materials for the production of various grades of
nylon fibers [28]. In 2019, it is estimated that about 11% of the world’s benzene will
be hydrogenated to produce cyclohexane for these nylon polymers. Together, these
three value chains (styrenics, phenolics, and nylons) represent about 82% of the
demand for benzene production. The remainder of the benzene produced is used
in the synthesis of nitrobenzene (10%, intermediate in polyurethane), chloroben-
zene (2%, dyes and herbicides), maleic anhydride (2%, coatings and surfactants), and
other key intermediates for the vast petrochemical industry. The market for benzene
is described in Figure 2.9 using our common format.
2.5.2 The Market for Toluene

The market dynamics for toluene are very interesting because while toulene is
recovered in quite large quantities, it is produced in much larger quantities, which
are not isolated but are instead sold or further processed. Toluene production
is driven by demand for three very large commodities – gasoline, ethylene, and
p-xylene – and is isolated and produced as a byproduct of making these three large
commodities. As a result, the “capacity” to produce toluene is somewhat difficult to
assess – many millions of metric tons are produced in naphtha reforming for gaso-
line, for example, and are never isolated and produced; this toluene, however, could
be quite easily isolated for production and sale through fractionation, extraction, or
some combination of both. In our analysis of capacity, we only account for facilities
with demonstrated capability and practice of producing toluene. Toluene has a
number of uses – ranging from high-value uses in chemicals to modest-value use
as a gasoline-blending component. Indeed, much of the toluene, which is isolated,
actually ends up being blended back into the gasoline market where local needs for
higher octane exist [31].
Despite the complexities, the capacity to recover toluene can be estimated based
on the process units, which exist at refinery and petrochemical locations around the
world. Based on that analysis [31], it is estimated that the capacity to isolate toluene
in 2019 is approximately 43.5 million metric tons per annum (see Figure 2.10). The
capacity to produce will likely increase at a compound annual growth rate of about
3–4% over the next 5–10 years as new capacity comes onstream. Very little in the
way of capacity additions are likely in the United States and Western Europe, while
the majority of new capacity investment will be in China, Southeast Asia, and the
Middle East as these are the regions where more toluene is needed for production of
chemicals as well as blending into gasoline. The large majority of toluene is produced
in naphtha reforming, about 77% in 2019, with naphtha crackers for the production
Global toluene market – capacity and demand

Styrene/other, 2.1% Others, 1.9%
60.00 0.680
Coke oven, 2.9% Toluene diisocyanate, 4.8%
Pygas, 17.8% 50.00 0.670 Solvents, 18.0%
Operating rate (%)

40.00 0.660
Gasoline blending, 28.1%
30.00 0.650
Reforming, 77.2%
20.00 0.640
Transalkylation, 47.2%
10.00 0.630
0.620
21
08
09
10
11
12
13
14
15
16
17
18
19
20
22
23
24
25
20
20
20
20
20
20
20
20
20
20
2019 sources 2019 users
20
20
20
20
20
20
20
20
Figure 2.10 Global supply and demand dynamics for toluene.
of ethylene being the second largest source at about 18% [31]. The remainder of the
produced toluene comes from minor processes like extraction from coking ovens
and coal-based chemical processes.
The demand for toluene was approximately 27.8 million metric tons in 2019, grow-
ing at a rate between 2% and 3% annually [31]. This growth rate is expected to con-
tinue for the next several years as p-xylene demand continues to increase at about
4.5–5.5% growth per year. While toluene and the transalkylation reaction are a sig-
nificant source of xylenes for the production of p-xylene (see Chapter 5), most of the
new complexes being developed globally are moving toward net-zero isolation of
toluene [32], resulting in a slower growth rate for toluene demand. Also, large-scale
blending of toluene for gasoline property (vapor pressure and octane) improvement
is already practiced, and this has two significant consequences. First, the demand
growth for gasoline blending will be muted as most regions are now using toluene,
leaving growth equivalent of the overall gasoline demand growth only. Second, this
growth in gasoline blending has set a floor or minimum price for toluene in the mar-
ket as producers can simply sell toluene for gasoline blending. This has closed off a
traditional market of p-xylene production from toluene and mixed xylenes. Com-
plexes used to be able to make acceptable margins by purchasing toluene and mixed
xylenes on the market and converting them into p-xylene. In the past 10 years, this
has become a very difficult and low-margin business.
Toluene is isolated and produced for three basic reasons: the production of
benzene and xylenes (BX; usually related to p-xylene production); the production
of chemicals from toluene; and, gasoline blending. The largest demand of these
three is for the production of benzene and xylenes (BX), with demand reaching
about 13.1 million metric tons in 2019 [31]. The next largest demand is for blending
into gasoline, which reached approximately 7.8 million metric tons in 2019 [31],
though demand growth will slow in the coming years as indicated above. The final
category of toluene demand, chemicals production from toluene, includes a diverse
set of compounds, including toluene diisocyanate, benzoic acid, and common and
specialty solvents. This category accounted for about 7.0 million metric tons of
toluene demand in 2019 [31]. Over the next decade, it is likely that the highest
demand growth will remain aligned with the demand for p-xylene, while growth of
toluene-derived chemicals will start to approach a similar growth rate.
Other, 0.4% Global xylene market – capacity and demand Isophthalic acid, 1.0%
Pygas, 4.5% 120.00 0.800 Others, 1.4%
Solvents, 9.6%
0.780
100.00 Phthalic anhydride, 5.1%
Transalkylation, 25.6% 0.760
Operating rate (%)

80.00 0.740
0.720
60.00
0.700 PTA/PET,polyester, 82.9%
Reforming, 69.5% 40.00 0.680
0.660
20.00
0.640
0.620
21
22
08
09
10
11
12
13
14
15
16
17
20 8
19
20
23
24
25
1
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Figure 2.11 Global supply and demand dynamics for xylenes.
2.5.3 The Market for Xylene

Mixed xylenes are a key intermediate in the production of several polyester com-
pounds and have grown to become a very significant part of the petrochemicals
businesses. Because of the rapidly growing demand for polyethylene terephthalate,
in particular, the market for mixed xylenes has exploded by nearly a factor of 3
since 2000 [33]. All three of the xylene molecules have end uses in plastic mate-
rials, but more than 83% of the mixed xylenes end up as p-xylene [33]. However,
mixed xylenes typically contain about 23% p-xylene at equilibrium, and more than
50% m-xylene – for which the smallest market exists. As a result, much of the prod-
uct of mixed xylenes involves extracting xylene isomer of interest (usually p-xylene)
and then re-equilibrating the p-xylene depleted stream to approach the equilibrium
distribution again [34]. This isomerization-separation loop is the basis for a signif-
icant growth in the polyester industry and has become quite energy and material
efficient over the last three decades of research and development [34].
The world’s capacity to produce mixed xylenes is estimated to be approximately
84.7 million metric tons in 2019 and has been growing at a compound annual
growth rate of about 4.0–4.5% for several decades (see Figure 2.11). As consumer
use of polyester materials, PET (polyethylene terephthalate) bottles, and other
plastics involving polyesters has grown, the capacity to produce mixed xylenes – the
raw material source for these polyester materials – has continued to keep pace.
However, the majority of the new capacity to produce mixed xylenes has and will
continue to be added in Asia – specifically, China, Korea, India, and, to a lesser
extent, Southeast Asia. Indeed, since 2007, more than 94% of the capacity added to
produce mixed xylenes has been in Asia [29]. The underlying trend in the polyester
chain has been one of growth of production in China, and the development of
a supply chain of p-xylene and mixed xylene to serve that market development
has been a highly regional story. Indeed, while China had a capacity to produce
mixed xylenes of around 5 million metric tons in 2006, capacity has grown to
nearly 22 million metric tons in 2019 [33]. That being said, there have been some
developments outside of Asia, particularly in the Middle East, which will continue
as resource owners look to increase the value they gain from bringing those raw
materials to market.
Mixed xylenes come from a variety of sources, the most abundant of which
are naphtha reforming, pyrolysis gasoline from naphtha crackers, transalkylation
processes, and other sources, including coal-derived chemicals [33]. Almost 70%
of the world’s supply of mixed xylenes comes from naphtha reforming, with much
of this being on-purpose or petrochemical-oriented naphtha reforming. Some
continue to be extracted from motor fuel reformers, though this is becoming less
and less attractive due to the demands on octane in motor fuels. Only about 4–5%
of the mixed xylenes are supplied from naphtha crackers, and this percentage is
likely to decline over the next decade [33]. The increased demand for xylenes, and
the shift to lighter (ethane) feedstocks in crackers, virtually assures the importance
of pyrolysis gasoline will decline over the next decade. It will, however, remain
very relevant from specific geographic regions. About 25–30% of the xylenes are
supplied from the various transalkylation processes, which involve moving alkyl
groups, specifically methyl groups, across the aromatic rings. This will remain
roughly steady, and will grow with the naphtha-reforming capacity; in a modern
xylene complex today, it is very rare to build reformers without a transalkylation
unit.
As mentioned in an earlier paragraph of this section, the great majority of mixed
xylenes produced (83%) are used to make p-xylene, which, in turn, is used to make
purified terephthalic acid and, ultimately, polyester products (fibers, films, and bot-
tles). Slightly more than 5% of the demand for mixed xylenes is for the production
of o-xylene, a key raw material in the manufacture of phthalic anhydride (PA). PA is
uses in many PVC (polyvinyl chloride)-based products as a plasticizer. About 1% of
the demand for mixed xylenes is driven by m-xylene, despite the fact that this is the
isomer available in highest quantity when the xylenes are at equilibrium. m-Xylene
is used to add ductility and impact resistance to plastic bottles [33]. All of this means
the demand for mixed xylenes has been growing 3.5–4.5% per year and will con-
tinue to do so as long as polyester demand growth continues. In 2019, the demand
for mixed xylenes was approximately 60.2 million metric tons per annum [33], with
about 60% of this mixed xylene being consumed in Northeast Asia.
2.6 The Market for Synthesis Gas
Synthesis gas is a widely used fuel-gas mixture, which is primarily composed of car-
bon monoxide and hydrogen [35, 36] and has primary uses in power generation,
synthesis of liquid fuels, use as a gaseous fuel, separated and used as industrial gases,
and for the synthesis of chemicals. Our primary focus in this section is on the use of
synthesis gas for chemical synthesis, which is slightly larger than 50% of the over-
all synthesis gas market in 2019. While there are many different chemicals that are
derived from synthesis gas, there are two chemicals that make up the vast major-
ity of the synthesis gas chemicals industry – methanol and ammonia. We focus our
discussion on these two chemicals in Sections 2.6.1 and 2.6.2.
2.6.1 The Market for Ammonia

Nitrogen is an essential element for all life as we know it – both plants and ani-
mals – and is used as a fundamental building block or catalyst by most biological
organisms, providing fuel for growth. In essence, ammonia therefore fuels human,
animal, and plant growth on the plant to a large extent. With the population of
the planet expected to reach 9.8 billion people by 2050, it is estimated that the food
supply will need to increase by 70% from 2016 levels to provide the appropriate nutri-
tion [37]. An essential path to closing this gap is agricultural productivity, which
is enabled by higher levels of fertilizer and soil quality, which plays directly to the
increased demand for ammonia. While nitrogen is captured through air separation,
natural gas or coal are the primary sources of hydrogen for the synthesis of ammonia.
The emergence of a coal-to-chemicals industry in China and favorable economics
in North America driven by cheap natural gas created a strong environment for
ammonia growth beginning in the mid-2000s. The last decade has been one of strong
increase in capacity and demand, driven by these strong economics.
The capacity to produce ammonia had been quite stagnant for some years, and
then underwent a renaissance in the mid-2000s. In 2004, for example, capacity to
produce stood at about 156.8 million metric tons per annum; by the end of 2019,
that number reached nearly 255 million metric tons of capacity [37, 38]. Capac-
ity has increased at a growth rate of about 3.0–3.5% over this period of time, and
is expected to increase by the same or higher growth rate for the next 5–10 years.
Figure 2.12 shows the typical capacity and demand balance for ammonia. Natural
gas is used as the hydrogen source for about 70% of the world’s capacity of ammo-
nia [36, 37]. As a consequence, ammonia synthesis had broadly moved to areas with
cheap natural gas such as North Africa and the Middle East. The shale gas revolu-
tion in North America has changed this landscape, with US-based shale gas being
among the lowest-cost natural gas in the world. Construction of a number of ammo-
nia and ammonia-based chemicals plants has been started since 2008 and more are
planned. In spite of this increase in capacity in the United States, China has main-
tained its position with about one-third of the capacity to produce ammonia from
the late-2000s through to today. Going forward, it is expected that the proportion
of capacity in China will drop, driven by investment in the Middle East and North
Other hydrocarbons, 2.7% Global ammonia market – capacity and demand

350.00
Industrial uses, 8.1%
0.840
Coal reforming, 26.7% 0.820 Other fertilizers, 15.5%
300.00
0.800 Nitric acid, 10.1%

250.00
Operating rate (%)
0.780
Other uses (mixed), 13.8%
200.00 0.760
150.00 0.740
Natural gas reforming, 70.6%
0.720
100.00 Urea, 52.5%
0.700
50.00
0.680
0.660
20 1
20 2
09
10
08
11
12
13
14
15
16
17
18
19
20
23
24
25
2
2

20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Figure 2.12 Global supply and demand dynamics for ammonia.

America. Most of China’s ammonia capacity is based on coal as the hydrogen source,
and this capacity is coming under more and more environmental scrutiny.
The demand for ammonia has increased steadily across the last two decades at
about 2.0–2.5% with demand reaching 190.9 million metric tons per annum in 2019
[37, 38]. About one-third of the demand is driven by China, consistent with its
capacity – ammonia and ammonia-derived chemicals is one area where China has
achieved a strong level of self-sufficiency. At least 81% of the ammonia demand is
driven by fertilizers, which should be connected to population growth and increased
nutritional needs [37]. Slightly more than 50% of the ammonia produced ends up as
urea products for use directly or indirectly as fertilizers [36, 37]. After urea, there are
a number of smaller ammonia-based products like phosphate, nitrate, sulfate, etc.,
which share fertilizer and other markets as end-use applications. There are dozens
of other applications for ammonia, from refrigerants to cleaners to fermentation
feed, but the consumption for these uses is minor [39].
2.6.2 The Market for Methanol

Methanol has moved from a relatively modest market to a large, diverse, rapidly
growing chemical market over the last decade or so. With a production capacity
of approximately 28.9 million metric tons in 1995, methanol grew modestly to
about 39.5 million metric tons in 2004 for a compound annual growth rate of about
3.5% [40]. In 2019, the capacity to produce methanol had increased to more than
124.1 million metric tons per annum, an annual compound growth rate of greater
than 10% since 2004 [41]. The methanol industry is one of the most fragmented of
our building blocks, meaning that only 28% of the world’s capacity is controlled by
the top 10 producers [40]. Geographically, almost 54% of the capacity to produce
methanol is within Northeast Asia, which, in this case, is almost exclusively China.
This boom in capacity addition has been in response to a number of new uses for
methanol, but has also resulted in a lower operating rate compared to the historical
norms (see Figure 2.13).
The story around methanol demand over the past two decades is a very compli-
cated one with many different drivers as new uses and applications for methanol
have developed. In the late 1990s, the major drivers for methanol consumption
Global methanol market – capacity and demand

Other hydrocarbons, 8.4% 180.00 0.740
Other chemicals, 13.2%
160.00 0.720 Solvents, 4.1%
Coal reforming, 31.4%
140.00 0.700 Acetic acid, 8.2%
Operating rate (%)
120.00 0.680 MTO/MTP, 18.9%

100.00 0.660
80.00 0.640
Formaldehyde, 26.8%
60.00 0.620
Natural gas reforming, 60.2%
40.00 0.600
Fuel uses, 28.8%
20.00 0.580
0.560
20 1
20 2
09
10
08
11
12
13
14
15
16
17
18
20 9
20
23
24
25
2
2
1
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20

Figure 2.13 Global supply and demand dynamics for methanol.

were formaldehyde and a fuel oxygenate additive called MTBE at about 7.7 million
metric tons and 5.9 million metric tons, respectively. Formaldehyde is used to
produce resin products for the wood products and building products industry [40].
Since the late 1990s, despite a ban on MTBE use in the United States, the fuel usage
of methanol has grown dramatically reaching 25.7 million metric tons in 2019 [42].
This comes in the form or MTBE, another ether called tertiary amyl methyl ether
(TAME), dimethyl ether, and direct fuel blending; in 2019, direct fuel blending
into gasoline had reached almost 7 million metric tons, from 0 in the late 1990s
[41]. In addition, a new market for methanol – converting methanol into ethylene
and propylene – has emerged in China, consuming over 13 million metric tons
of methanol in 2019 [41]. In addition, formaldehyde use has continued to grow,
reaching 22.6 million metric tons per annum in 2019. Considering all these factors,
and the ups and downs of various different end uses, the demand for methanol is
estimated to be just slightly above 84 million metric tons in 2019 [41]. Since 2008,
the demand has nearly doubled in 11 years for a 6.4% annual growth rate.
Looking at the split of demand by use in 2019, we find that about 27% of the
demand is driven by formaldehyde and building products end uses, which is the
largest single chemical use. However, more than 28% of the methanol produced ends
up in the fuels sector through MTBE and other fuel ethers, direct gasoline blending,
and the biodiesel industry. The fastest growing use, and third largest, is the MTO
application (about 19%), where methanol is converted to ethylene and propylene
(see Chapter 3). The remainder of the methanol is used in acetic acid (8%), solvents
(4.1%), and many other small chemical compounds [42]. This is summarized in the
capacity-demand figure for methanol in Figure 2.13.
References
1 IHS Markit (15 August 2019). CEH Petrochemical Industry Overview.

2 The authors recommend Pindyck, R.S. and Rubinfeld, D.L. (2005). Microeco-
nomics, 6e. Pearson-Prentice Hall for a complete description.
3 Pindyck, R.S. and Rubinfeld, D.L. (2005). Microeconomics, 6e, 24.
Pearson-Prentice Hall.
4 (2017). CEH Ethylene. IHS Markit; internal estimates from Honeywell UOP.
5 Nexant Inc. (2014). The global petrochemical industry. Presentation, Amsterdam
(7 April 2014).
6 (2017). CEH Ethylene, 18. IHS Markit; internal Honeywell UOP estimates.
7 (2017). CEH Ethylene, 98. IHS Markit (Middle East Capacity for Ethylene
graphic).
8 Baumgarten, S. (17 October 2016). BASF confirms second death in Germany
explosion. https://www.icis.com/explore/resources/news/2016/10/17/10045568/
basf-confirms-second-death-in-germany-explosion/ (accessed 04 September
2020).
References 41
9 Toto, D. (17 December 2018). Plastic bottle recycling rate declines in 2017.
https://www.recyclingtoday.com/article/2017-plastic-bottle-recycling-rate/
(accessed 04 September 2020).
10 Loepp, D. (30 May 2019). Under pressure, plastics find place in the global circu-
lar economy. https://www.plasticsnews.com/blog/under-pressure-plastics-find-
place-global-circular-economy (accessed 04 September 2020).
11 CEH Ethylene. IHS Markit.
12 (2017). CEH Propylene. IHS Markit and internal Honeywell UOP estimates.
13 Myers, R.A. (2018). Handbook of Petrochemical Product Processes, 2e.
McGraw-Hill, Section 10.3.
14 EChemI Website (12 July 2018). 2018 China newly added methanol and
MTO/CTO capacity statistics. http://www.echemi.com/cms/13515.html (captured
08 November 2019).
15 (2017). CEH Propylene. IHS Markit.
16 Rekoske, J.E. (2016). Technoeconomic impacts of abundant natural gas liq-
uids on the chemical industry. Presentation at the 11th Natural Gas Conversion
Symposium, Tromso, Norway (6 June 2016). Natural Gas Conversion Board:
Washington, DC.
17 (2017). CEH Butylenes. IHS Markit and internal Honeywell UOP estimates.
18 U.S. Department of Energy, Energy Information Agency (13 February 2013).
Alkylation is an important source of octane in gasoline. https://www.eia.gov/
todayinenergy/detail.php?id=9971 (captured 08 November 2019).
19 U.S. Department of Energy, Energy Information Agency Document. MTBE, oxy-
genates and motor gasoline. https://www.eia.gov/outlooks/steo/special/pdf/mtbe
.pdf (captured 08 November 2019).
20 CEH Butylenes. IHS Markit.
21 Kranz, K., duPont STRATCO Alylation Training. (September 2008). Intro to
alkylation chemistry. https://silo.tips/download/intro-to-alkylation-chemistry.
22 Encyclopedia Britannica. Butadiene rubber. Online Edition. https://www
.britannica.com/technology/butadiene-rubber (captured on 23 October 2019).
23 Sethurai, M.R. and Mathew, N. (eds.) (1992). Developments in Crop Science:
Volume 23, Natural Rubber. Elsevier Science.
24 White, W.C. (2007). Chemico-Biological Interactions 166 (1–3): 10–14.
25 Nizamoff, A.J. (2013). On-Purpose Butadiene. Nexant ChemSystems PERP
Report.
26 Carey, F.A. and Sundberg, R.J. (1990). Advanced Organic Chemistry Part A:
Structure and Mechanisms, 499. New York: Plenum Press.
27 Spitz, P.H. (1988). Petrochemicals: The Rise of an Industry, 42. New York: Wiley.
28 (2017). Benzene World Analysis. IHS Markit.
29 (2018). Mixed Xylene. World Economics Handbook, 15. IHS Markit.
30 Matar, S. and Hatch, L.F. (1994). Chemistry of Petrochemical Processes, 265. Lon-
don: Gulf Publishing.
31 (2018). CEH Toluene. IHS Markit and internal Honeywell UOP estimates.
32 Internal Honeywell UOP analysis.
33 (2018). Mixed Xylene. World Economics Handbook. IHS Markit, and internal
Honeywell UOP estimates.
34 Myers, R.A. (2018). Handbook of Petrochemical Product Processes, 2e.
McGraw-Hill, Section 13.3.
35 Chauvel, A. and Lefebvre, G. (1989). Hydrogen, synthesis gases and their
derivatives. Petrochmical Processes, Part 1: Synthesis Gas Derivatives and Major
Hyxrocarbons, 27. Paris: Editions Technip.
36 Speight, J.G. (2019). Natural Gas, A Basic Handbook, 2e. London: Gulf Profes-
sional.
37 (2017). CEH Ammonia. IHS Markit.
38 Internal Honeywell UOP estimates.
39 Appl, M. (1999). Ammonia: Principle and Industrial Practice. Wiley-VCH Verlag
GmbH.
40 (2017). CEH Methanol. IHS Markit.
41 (2017). CEH Methanol. IHS Markit and internal Honeywell UOP estimates.
42 Olah, G.A., Goeppert, A., and Prakash, G.K.S. (2018). Beyond Oil and Gas: The
Methanol Economy, 3e. Wiley-VCH.
43
Section II
Olefins and Synthesis Gas

45
Olefins and Synthesis Gas Production Technologies
3.1 Introduction
Olefins are key building blocks of chemical industry. Light olefins, ethylene and
propylene, are the most important olefins used in the manufacture of a large number
of chemicals in commodities and specialties such as polyethylene, ethylene dichlo-
ride, ethylene oxide, polypropylene, propylene oxide, and acrylonitrile. Butylene
is also in high demand for further processing to produce methyl-tert-butyl ether
(MTBE) and alkylate for gasoline blending. Heavy normal-olefins (C10 –C14 ), derived
from straight run kerosene, are the main feedstock component of linear alkylben-
zene (LAB), the most common raw material in the manufacture of biodegradable
household detergents. In this chapter, the manufacturing technologies having com-
mercial significance for producing these olefins will be discussed, including steam
cracking, paraffin (light paraffin and heavy normal-paraffin) dehydrogenation, and
methanol to olefins. There are other processes in a refinery, e.g. fluid catalytic crack-
ing (FCC), that produce light olefins as byproducts. These technologies will not be
discussed in this chapter.
Synthetic gas (syngas), a mixture of carbon monoxide and hydrogen, is used in
large volumes for the production a wide variety of commodity chemicals such as
hydrogen, methanol, ammonia, olefins, and MTBE. Hence, it represents a suitable
alternative to produce base chemicals from resources other than crude oil. Syngas
manufacturing technologies discussed in this chapter include conversion of natural
gas to syngas and coal gasification to syngas.
3.2 Steam Cracking
In the steam-cracking process, a hydrocarbon feedstock is mixed with steam and

cracked at elevated temperatures in a tubular reactor. It can accept a variety of hydro-
carbons, ranging from gases (ethane, propane, and butane) to liquids (naphtha and
gas oils), as shown in Figure 3.1. The reactor effluent gases are first liquefied and
then separated by distillation columns as ethylene, propylene, a crude C4 –fraction,
pyrolysis gasoline, and heavies.

46 3 Olefins and Synthesis Gas Production Technologies
Ethane Ethylene
Gas propane
butanes Propylene
Gaseous
Naphtha Crude C4’s: products
Liquid
gas oils (Butadiene and
butylene)
Steam cracking
furnace Pyrolysis
800 – 900°C gasoline:
Low pressure (Aromatics
and C5’s) Liquids
No catalyst
Heavies
Figure 3.1 Steam-cracking technology.
There are three steps in the steam-cracking process: pyrolysis and cooling, com-
pression and acid gas removal, and cryogenic cooling and product separation.
Pyrolysis and cooling: The feedstock is initially mixed with a dilution steam and
heated to 815–900 ∘ C at 1.7–2.4 bars. The resulting products are immediately
cooled to 340–510 ∘ C and further cooled to 37–43 ∘ C in a water quenching tower.
Compression and acid gas removal: The cracked gas stream is compressed from atmo-
spheric pressure to 35–38 bars. During the compression, acid gases (H2 S and CO2 )
are removed by scrubbing the stream. The resulting stream is dried over a solid
adsorbent.
Cryogenic cooling and product separation: The dried gas is cooled to between −90
and −130 ∘ C in a refrigeration train. The resultant gas, consisting primarily of
hydrogen, methane, and carbon monoxide, is purified to recover hydrogen; the
remainder is burned as plant fuel. Low-temperature fractionation separates the
dried products. Ethane and propane are generally recycled, while hydrogenation
of acetylene, propyne, and propadiene yields ethylene and propylene.
For the conventional endothermic cracking process, ethylene yield is improved by
raising the cracking temperature and reducing residence time. The operating condi-
tions are constrained by the metallurgy of the furnace tubes and coking tendency in
the cracking coils. The formation of coke deposits in cracking furnace coils remains
one of major obstacles in steam-cracking operations. Coke formation and subse-
quent accumulation in furnace components result in significant energy and main-
tenance costs. Typically, a naphtha pyrolysis furnace is decoked every 15–40 days.
Maximum cycle time is 60–100 days.
Ethylene fractionation separates ethylene as a high-purity product (>99 wt%) from
ethane, which is combined with propane and recycled for cracking. Propylene is
separated from propane in the propylene fractionation section as a chemical-grade
product (typically 93–95 wt% min) or more frequently as a polymer-grade propylene
(>99.5 wt%).
A mixed C4 stream is produced from the cracking of naphtha and heavy feedstocks
providing an important source for butadiene and butylene. More than 95% of world
3.2 Steam Cracking 47
Table 3.1 Typical products from cracking various feedstocks in a 450 KMTA ethylene plant
(KMTA).
Vac
Products Ethane Propane N-Butane Naphtha gas oil
H2 -Rich gas 30 20 15 15 306

CH4 -Rich gas 40 300 250 220 200
Ethylene 450 450 450 450 450
Propylene 10 290 190 260 260
Butadiene 10 30 40 80 80
Butylenes/butanes 5 20 80 130 80
Pyrolysis gasoline + BTX + others 20 140 160 990 1200
Fuel oil 0 10 30 300 610
Ethylene yield (wt%) 81 43 40 37 24
Total 575 1260 1215 2445 3186
supply of finished butadiene is extracted from the mixed C4 produced in the steam
crackers. The production of mixed C4 stream is strongly dependent on cracking
severity and feedstock specification. Typically the amount of mixed C4 produced
per ton of ethylene will decrease when running the cracker at higher severity.
The selection of feedstock for steam cracking is considered as one of the most
important issues for a petrochemical complex. Steam crackers are designed for feed
flexibility so that they can process a predefined range of feedstocks. Product depen-
dence on feedstock is shown in Table 3.1.
An ethane cracker is preferred when the feedstock is available in sufficient
amounts. As a result of shale gas development in the United States producing high
abundance of ethane, ethane cracking has become more attractive due to its low
cost, but byproduct credits from naphtha-based crackers in Europe and Asia could
be enough to keep them profitable maintaining a large share of the olefin market [1].
In a naphtha-based cracker, the feedstock contains normal-paraffins, isoparaffins,
naphthenes, and aromatics. Ethylene and other olefins are formed primarily from
paraffins and naphthenes in the feedstock. Normal-paraffins are the preferred com-
ponents for high ethylene and propylene yields. Isoparaffins produce much smaller
yields of ethylene and propylene than normal-paraffins.
An option to maximize the profitability from a naphtha-based cracker is to inte-
grate it with a refiner by using UOP MaxEneTM . The MaxEne process can simul-
taneously increase the yield of high value products from a naphtha-based cracker
and catalytic reforming unit. This process is designed to extract normal-paraffins
from full-range naphtha, consisting of molecules in the C6 to C10 range. The separa-
tion uses a simulated countercurrent contacting scheme, in which normal-paraffins
are first adsorbed in a series of fixed beds of adsorbent and then desorbed using a
desorbent. The extract stream with high-purity normal-paraffins makes a superior
n-Paraffins
Steam 30% Higher
Naphtha cracking ethylene yield
( C6–C10)
MaxEne
5–6 higher RON for

Gasoline Blending
Reforming
i-Paraffins 2–3% Aromatics yield
naphthenes increase for aromatics
aromatics complex
Figure 3.2 UOP MaxEne process.
steam cracker feed, while the raffinate stream enriched with naphthene/aromatic is
optimal as a reforming feed (Figure 3.2). The process enables:
● 30% ethylene yield increase from existing naphtha cracker with no loss in propy-
lene,
● 5–6 RON increase of reformate from reforming unit to gasoline pool, and
● 2–3% aromatics yield increase for reformate from reforming unit to aromatics com-
plex.
3.3 Light Paraffin Dehydrogenation
3.3.1 Process and Catalyst

The majority of propylene and butylene are produced as byproducts of petroleum
refineries (FCC) and steam crackers. Thus, most of propylene and butylene
are byproducts of other products, specifically gasoline and ethylene. However,
when production capacity is not coupled with demand for those byproducts, a
supply/demand imbalance can occur.
The paraffin dehydrogenation process provides petrochemical producers with a
catalytic, on-purpose means for making propylene or butylene from their corre-
sponding paraffin, independent of demand for gasoline and ethylene.
Paraffin dehydrogenation for the production of light olefins has been in use since
the late 1930s. Catalytic dehydrogenation of butanes over a chromium catalyst was
practiced to produce butylenes, which were then dimerized to octenes and hydro-
genated to octanes to yield high-octane aviation fuel. Dehydrogenation of butanes
over a chromium catalyst was first developed and commercialized at Leuna in Ger-
many and was also independently developed by then Universal Oil Products (UOP)
in United States, together with ICI in United Kingdom.
In the dehydrogenation process using chromium catalysts, the catalyst is con-
tained in a fixed-bed reactor. A significant amount of coke is deposited on the
catalyst during the dehydrogenation; therefore, a number of reactors are used in
parallel – some for dehydrogenation while the rest are being purged or regenerated.
3.3 Light Paraffin Dehydrogenation 49
In the late 1980s, the application of chromium catalysts with cyclical operation was
extended by Houdry to the dehydrogenation of propane to propylene and isobutane
to isobutylene. The new process application was called CatofinTM [2].
In 1959, a fluidized-bed dehydrogenation technology with chromium catalyst
was developed and commercialized in the former Soviet Union. This reactor design
avoids the cyclical operation by circulating the catalyst from the reactor to the
regenerator. Circulating regenerated catalyst is used to provide the heat of reaction
in the riser, while spent catalyst is reheated by carbon burn in the regenerator.
The technology was commercialized by Yarsintez [3] and further developed and
commercialized by Snamprogetti [4, 5].
The Philips STAR process also features regeneration of the catalyst on a cyclic basis
using a fixed-bed fired-tube reactor. Steam is used to lower the partial pressure of
the hydrocarbons and to slow down the deposition of coke on the catalyst providing
for extended cycle times. For continuous operation, various furnace modules (banks
of tubes) can be operated in parallel with several modules operating in the process
mode, while one is in the regeneration mode.
In the early 1970s, UOP introduced continuous catalyst regeneration (CCRTM )
technology that enabled noble metal catalysts to remain at their most desirable
activity for several years without having to shut down the reactor for catalyst
regeneration. The combination of noble metal catalysts operating at high severity in
conjunction with CCR technology made it possible to design, build, and economi-
cally operate large catalytic dehydrogenation units that can produce light olefins, in
particular, propylene and butylene, at high selectivities. This technology is known
as the UOP OleflexTM process.
The UOP Oleflex process can dehydrogenate propane, isobutane, normal-butane,
or isopentane feedstocks separately or as a mixture of spanning two consecutive
carbon numbers. The UOP Oleflex process and uses of the Oleflex process to
produce propylene, MTBE, and alkylate are discussed in Sections 3.3.2 and 3.3.3 of
this chapter.
As shown in Figure 3.3, the UOP Oleflex process is separated into three different
sections:
● Reactor section,
● Product recovery section, and
● Catalyst regeneration section.
3.3.1.1 Reactor Section

Hydrocarbon feed is mixed with hydrogen-rich recycle gas. This combined feed is
heated to the desired reactor inlet temperature and converted at high olefin selectiv-
ity in the reactors.
The reactor section consists of several radial-flow reactors, charge and interstage
heaters, and a reactor feed-effluent heat exchange. The diagram shows a unit with
four reactors, which would be typical for a unit processing propane feed. Three reac-
tors are used for butane dehydrogenation. Three reactors are also used for blends of
propane and butane feed, supplying heat through interstage heaters, placed between
Catalyst flow
C
C
R
Dryer
Rx effluent
compressor
Heater cells
Cold box
To frac
section
H2 Recycle
Fresh and
recycle feed Net gas (H2)
Figure 3.3 UOP Oleflex process flow.
the reactors. The catalyst is collected on the bottom of the last reactor, transported
pneumatically to the regenerator. The effluent leaves the last reactor, exchanges heat
with the combined feed, and is sent to the product recovery section.
3.3.1.2 Product Recovery Section

The reactor effluent is cooled, compressed, dried, and sent to a cryogenic separation
system. The dryers serve two functions: (i) to remove trace amounts of water formed
from the catalyst regeneration, and (ii) to remove hydrogen sulfide. The treated efflu-
ent is partially condensed in the cold separation system and directed to a separator.
Two products come from the Oleflex product recovery section: separator gas
and separator liquid. The gas from the cold high-pressure separator is expanded
and divided into two streams: recycle gas and net gas. The net gas is recovered
at 85–93 mol% hydrogen purity. The impurities in the hydrogen product consist
primarily of methane and ethane. The separator liquid, which consists primarily of
the olefin product and unconverted paraffin, is sent downstream for processing.
3.3.1.3 Catalyst Regeneration Section

The regeneration section, shown in Figure 3.4, is similar to the CCR unit used in the
UOP PlatformingTM process. The CCR perform four functions:
● Burning the coke off the catalyst,

● Redistributing the platinum,
● Removing excess moisture, and
● Reducing the catalyst prior to returning to the reactors.
The slowly moving bed of catalyst circulates in a loop through the reactors and the
regenerator. The cycle time around the loop can be adjusted within broad limits but
is typically anywhere from 5 to 10 days. The catalyst activity can be restored with-
out interrupting the catalytic dehydrogenation process in the reactor and recovery
sections.
Collector
Disengaging
hopper Lift gas
blower
R
Regeneration
tower
R Flow control
Catalyst
lift lines hopper
Lock
hopper Lock
hopper Surge hopper
Lift engager
Lift
engager
Nitrogen H2
lift gas lift gas
Figure 3.4 Oleflex regeneration section.
The Oleflex process for light paraffin dehydrogenation uses a platinum catalyst
supported on alumina base. The key role of the dehydrogenation catalysts is to accel-
erate the main reaction while controlling other reactions. The alumina-supported
platinum catalysts are highly active but are not selective to dehydrogenation. In addi-
tion, the catalyst rapidly deactivates due to fouling by heavy carbonaceous materials.
Therefore, the properties of platinum and the alumina support need to be modified
to suppress the formation of byproducts and to increase catalyst stability. Arsenic,
tin, germanium, lead, and bismuth are among metals reported as platinum activity
modifiers [6]. Platinum is a highly active catalytic element and is not required in
large quantities to catalyze the reactions when it is dispersed on a high surface area
support. The high dispersion is also necessary to achieve high selectivity to dehy-
drogenation relative to undesirable side reactions, such as cracking. A typical high
surface area alumina supports employed having acidic sites that accelerate skeletal
isomerization, cracking, oligomerization, and polymerization of olefinic material,
and enhance “coke” formation. Alkali or alkaline earth metals assist in the con-
trol of the acidity [6]. Modified catalysts possess high activity and high selectivity
to mono-olefins. The major byproducts are controlled kinetically.
3.3.2 UOP C3 Oleflex Complex

The main reaction and side reactions for propane dehydrogenation are shown below.
Main reaction:
C3 H8 ↔ C3 H6 + H2
Side reactions:
C3 H8 ↔ C3 H4 + 2 H2
700
650
600
Temperature, ° C
550
500 40% Paraffin conversion
450
400
350 10% Paraffin conversion
300
2 3 4 5 6 7 8 9 10 11 12 13 14 15
Carbon number
Figure 3.5 Temperatures required to achieve 10% and 40% conversion of C2 –C15
normal-paraffins at 1 bar. Source: Bhasin et al. [6]. © 2001, Elsevier.
C3 H8 → C2 H4 + CH4
C2 H4 + H2 ↔ C2 H6
The main propane dehydrogenation is operated at equilibrium favored by high tem-
perature and low pressure. Figure 3.5 [6] shows the temperatures required for the
dehydrogenation of light paraffins are much higher than those for heavy paraffins.
For 40% conversion of propane requires a temperature of 600 ∘ C. It is advisable to
increase the conversion per pass as much as possible so that the cost of recycling
the unconverted reactant and that of separating the reactant and product is reduced.
However, high temperatures also enhance side reactions, such as coke formation,
causing fast catalyst deactivation and thermal cracking reactions that reduce the
selectivity to the desired product. Thus, all factors need to take into consideration
to optimize the process economics.
Oleflex process units typically operate in conjunction with fractionators and other
process units within a production complex. In a propylene complex (Figure 3.6),
a preheated propane-rich liquefied petroleum gas (LPG) feedstock is introduced
to a depropanizer to reject butanes and heavier hydrocarbons. The depropanizer
overhead is then directed to the Oleflex unit. The once-through propane conversion
closely approaches the equilibrium value defined by the Oleflex process conditions,
to selectively produce propylene and hydrogen with minimized side products. Two
product streams are created within the C3 Oleflex unit: a hydrogen-rich vapor
product and a liquid product rich in propane and propylene. The net hydrogen can
be exported directly, upgraded to PSA hydrogen, or used as fuel within the plant if
hydrogen is not in demand in the vicinity of the C3 Oleflex complex.
Trace levels of methyl acetylene and propadiene are removed from the Oleflex
liquid product by selective hydrogenation. The SHP process, consisting of a single
liquid-phase reactor, selectively saturates diolefins and acetylenes to mono-olefins
without saturating propylene.
Net gas C2–
Propylene
Oleflex
unit SHP
Deethanizer
Depropanizer
C3 LPG
P–P Splitter
C4+
Figure 3.6 UOP C3 Oleflex complex for propylene production.
Ethane and lighter material enter the propylene complex in the feed and are
also created by nonselective reactions within the Oleflex unit. These light ends
are rejected from the complex by a deethanizer column. The deethanizer bot-
toms are then directed to a propane–propylene (P–P) splitter. The splitter produces
high-purity propylene as the overhead product. Typical propylene purity ranges from
99.5 and 99.8 wt%. Unconverted propane from the Oleflex unit concentrates in the
splitter bottoms and is returned to the depropanizer for recycle to the Oleflex unit.
UOP’s latest Oleflex complex design with new generation of catalyst, DeH-26, con-
sumes less than 1.12 MT of propane per MT of propylene produced [7] (Figure 3.7).
Investment in a C3 Oleflex complex can be an excellent feedstock and end-product
diversification play to increase the operating flexibility of the refinery. The propane
feedstock from natural gas liquid (NGL) plants can be supplemented with propane
produced within the refinery and sent to the C3 Oleflex complex. The C3 Ole-
flex complex can also be used to upgrade lower-purity refinery-grade propylene
(RGP) streams produced by the FCC unit to chemical-grade propylene (CGP) or
polymer-grade propylene (PGP) in the propylene complex P–P splitter to realize
an additional product value uplift. Additionally, the hydrogen produced within the
Oleflex 1990
100% DeH-6 Catalyst
1.25 MTC3/MTC3=
Oleflex 2004
CCOP reduction from base
High
95% DeH-14
Performance
1.18 MTC3/MTC3=
Oleflex 2015
Oleflex 1997 Energy reduction Next generation
DeH-16
DeH-12
90% 1.14 MTC3/MTC3= Oleflex 2018
1.21 MTC3/MTC3= DeH-26
Energy
reduction <1.12 MTC3/MTC3=
Oleflex 2011 Further energy
85% DeH-16 reduction
1.16 MTC3/MTC3=
Energy reduction
80%
0 5 10 15 20 25 30 35 40
% ROI improvement from base
Figure 3.7 UOP Oleflex technology evolution.

Propane PGP
RGP C3 Oleflex
Refinery
complex
H2 CGP
Figure 3.8 UOP C3 Oleflex complex integration with an existing refinery.
C3 Oleflex complex can be used for hydroprocessing needs within the refinery or
exported for sale (Figure 3.8).
3.3.3 UOP Iso-C4 Oleflex Complex

A typical iso-C4 Oleflex complex configuration is shown in Figure 3.9 for the pro-
duction of methyl tertiary butyl ether (MTBE) from butane and methanol. MTBE
usage grew in the early 1980s in response to octane demand resulting initially from
phase-out of lead from gasoline and later from rising demand for premium gasoline.
Ethanol can be substituted for methanol to make ethyl tertiary butyl ether (ETBE)
with the same process configuration. Furthermore, isopentane may be used in addi-
tion to or instead of field butane to make tertiary amyl methyl ether (TAME) or
tertiary amyl ethyl ether (TAEE).
Three primary catalytic processes are used in an MTBE complex:
● Paraffin isomerization to convert normal butane into isobutane,

● Dehydrogenation to convert isobutane into isobutylene, and
● Etherification to react isobutylene with methanol to make MTBE.
Field butane, a mixture of normal-butane and isobutane obtained from natural gas
condensate, is fed to a deisobutanizer (DIB) column. The DIB column prepares an
isobutane overhead product, rejects any pentane or heavier material in the DIB bot-
toms, and makes a normal-butane side cut for feed to the butane isomerization unit.
H2 C3–
i-C4 Frac MTBE

Ethermax
Oleflex section
Methanol
Deisobutanizer
Field
butanes
Butamer
C5+
Figure 3.9 UOP iso-C4 Oleflex complex.

The DIB overhead is directed to the iso-C4 Oleflex unit, in which isobutane is selec-
tively converted to isobutylene and hydrogen. The main reaction and side reactions
are shown below:
Main reaction:
i-C4 H10 ↔ i-C4 H8 + H2
Side reactions:
i-C4 H10 ↔ n-C4 H10
n-C4 H10 → n-C4 H8 + H2
n-C4 H8 ↔ C4 H6 + H2
Cracking of butanes to lighter materials also occurs as side reactions. The main
isobutane dehydrogenation is operated at equilibrium favored by high temperature
and low pressure.
Two product streams are created within the iso-C4 Oleflex unit: a hydrogen-rich
vapor and a liquid product rich in isobutane and isobutylene. The iso-C4 Oleflex
liquid product is sent to an etherification unit, where methanol reacts with isobuty-
lene to make MTBE. Raffinate from the etherification unit is depropanized to remove
propane and lighter materials. The depropanizer bottoms are then dried, saturated,
and returned to the DIB column (Figure 3.9).
Although MTBE is a good gasoline-blending component, it has properties that
cause problems. MTBE is more soluble in water than most other components of
gasoline. If the gasoline containing MTBE is spilled on the ground or leaks out of
underground storage tanks, MTBE could contaminate groundwater. Even fairly
small amount of MTBE in water can give it an unpleasant taste and odor, making
the water undrinkable. MTBE also does not break down (biodegrade) easily, so it is
harder to clean up once contamination occurs. In the late 1990s, many community
drinking water suppliers in the U.S. were found to have detectable levels of MTBE.
As a result, MTBE has been phased out for gasoline blending in the United States
in late 1990s. U.S. refineries were then faced with the challenge of replacing the lost
volume and octane value of MTBE in the gasoline pool. Isooctane was identified
as a cost-effective alternative to MTBE. It utilized the same isobutylene feeds used
in MTBE production and offered excellent gasoline-blending value. Furthermore,
isooctane production can be achieved in a low-cost revamp of existing MTBE plant
to the UOP InAlkTM process unit.
As shown in Figure 3.10, the Oleflex dehydrogenation unit can be easily inte-
grated with downstream dimerization of isobutylene followed by hydrogenation
(UOP InAlk process). The hydrogen from the Oleflex can be used to hydrogenate
isooctene to isooctane, a high-octane gasoline-blending component. The UOP
process for this combination is the UOP InAlk complex.
i-C4 Dimerization
Oleflex
InAlk
H2
Deisobutanizer
Hydrogenation
Field
butanes
Iso-octane
Frac section
Butamer
C3–
C5+
Figure 3.10 UOP InAlk complex.
H2 C3–
Propylene
C3/i-C4 C3/C4
Oleflex separation
Methanol
Deisobutanizer
LPG Ethermax
Butamer MTBE
Frac
Section
C5+
Figure 3.11 UOP C3 /C4 Oleflex complex.
3.3.4 UOP C3 /C4 Oleflex Complex

Mixed dehydrogenation of propane and isobutane within the same reactor system
has also been of interest in the marketplace. UOP has developed and commercialized
co-feeding propane and isobutane to a dehydrogenation unit for co-production of
propylene and MTBE (Figure 3.11).
3.3.5 UOP Mixed-C4 Oleflex Complex

Demand for alkylate, nonaromatic source of octane has increased with more strin-
gent fuel specifications. Cheap, excess butane from shale gas (approximately 60%
normal-butane and 40% isobutane) provides an upgrading opportunity to meet the
need for high octane alkylate. Refiners are interested in butane dehydrogenation to
feed sulfuric acid alkylation. By leveraging its existing isobutane Oleflex experience,
UOP developed mixed-C4 Oleflex technology to provide suitable feedstock for sul-
furic acid alkylation. This was accomplished by maximizing the conversion while
managing the increased rate of heavies’ formation associated with normal-butane
3.4 Heavy Normal-paraffin Dehydrogenation 57
H2
Field Mixed -C4 Frac

Oleflex Alkylation section Alkylate
butanes
Butamer
Figure 3.12 UOP mixed-C4 Oleflex complex.
dehydrogenation. UOP mixed-C4 Oleflex complex (Figure 3.12) has the following
features:
● Capability to process any mixture of butanes, including 100% normal-butane,
● Field butanes fed directly to the mixed-butane Oleflex – elimination of the need
for deisobutanizer, and
● Alkylate octane increased by 1–2 numbers ((R+M)/2) using the feed from
mixed-butane Oleflex relative to the use of 100% isobutylene feed.
3.4 Heavy Normal-paraffin Dehydrogenation

Normal-paraffins are linear hydrocarbons of C10 –C14 chain lengths that are usually
separated from the kerosene of crude oil by close-cut distillation, followed by isola-
tion from branched-chain hydrocarbons and aromatics by selective adsorption with
molecular sieves. The normal-paraffins are dehydrogenated to produce correspond-
ing normal-olefins as a raw material for producing LAB.
In 1960s, UOP started development work on heterogeneous platinum catalysts
supported on an alumina base for the dehydrogenation of heavy normal-paraffins.
The resulting successful process, known as the UOP PacolTM process, was first
commercialized in 1968. The advent of the UOP Pacol process marked a substantial
transformation in the detergent industry and contributed to the widespread use of
LAB on an economical, cost-effective basis. More than 40 Pacol units have been
built or are under design or construction; practically all new LAB capacity built
on a worldwide basis over the last two decades makes use of UOP’s Pacol catalytic
dehydrogenation process.
In the Pacol process (Figure 3.13) [8], the normal-paraffins are dehydrogenated in
a vapor-phase reaction to corresponding mono-olefins the presence of hydrogen over
a selective platinum dehydrogenation catalyst. An adiabatic radial-flow reactor with
feed preheat is normally used to compensate for the endothermic temperature drop
and to minimize pressure drop within an efficient reactor volume. Relatively high
space velocities are used so that only a modest amount of catalyst is required. Hydro-
gen and some byproduct light ends are separated from the dehydrogenation reactor
effluent, and a part of this hydrogen gas is recycled back to the dehydrogenation
reactor to minimize coking and enhance catalyst stability. The separator liquid is an
equilibrium mixture of linear of olefins and unconverted normal-paraffins, which
are charged to a reactor for the selective conversion of diolefins to mono-olefins.
Hydrogen-rich
off-gas
Light
Hydrogen C end gas
recycle gas
Reactor
Light end
Separator liquid
Charge Stripper
heater
Benzene
Alumina UOP Linear

treater Detergent alkylate detergent
Linear process alkylate
paraffin Paraffin recycle
charge
Figure 3.13 UOP pacol dehydrogenation process.
The equilibrium conversion for the dehydrogenation reaction is determined by

temperature, pressure, and hydrogen partial pressure. As expected, the equilibrium
conversion increases with temperature and decreases with pressure and with
increasing hydrogen-to-hydrocarbon ratio. Kinetically, the overall conversion
depends on space velocity; excessively high space velocities do not allow for suffi-
cient conversions, and space velocities that are too low lead to lower selectivities
because of the onset of side and competitive reactions.
Since first commercial operation of the Pacol process in 1968, several advances
with modified platinum catalyst have been made showing significantly lower cok-
ing and doubling the catalyst stability. For longer carbon chain paraffins, diffusion
also becomes a critical parameter, and thus large-pore low-density support is needed
to access all platinum sites. Further catalyst performance improvement has been
realized by introducing finely dispersed thin platinum layer over an inert core [8].
Figure 3.14 illustrates the flow scheme of an integrated complex featuring the UOP
UnionfiningTM , MolexTM , Pacol, DeFineTM , PEPTM , detergent alkylation, and DetalTM
processes.
There are two major sections in a LAB complex: production of normal-paraffins
and production of LAB.
3.4.1 Production of Normal-paraffins

Kerosene prefractionation is often used to tailor the kerosene feed to the desired
carbon range. Kerosene is stripped of light ends and heavier components so that the
heart cut, containing the desired normal-paraffins for the production of LAB of a
certain range of molecular weight, is produced.
The distillate Unionfining process hydrotreats kerosene at sufficient severity to
remove sulfur, nitrogen, olefins, and oxygenate compounds, which might otherwise
poison the Molex adsorbent.
3.5 Methanol to Olefins 59
Kerosene Benzene
Light
Prefractionation ends Heavy
hydrotreating Aromatics alkylate
Hydrogen
Detal or
Pacol detergent
Purification PEP LAB
Define alkylate
Raffinate
return to Recycle Paraffin
refinery
Figure 3.14 Integrated LAB complex.
The Molex process is a liquid-state separation of normal-paraffins from branched

and cyclic components using SorbexTM technology. The simulated moving-bed
adsorptive separation results from using a proprietary multi-port rotary valve.
The extract stream is a high-purity normal-paraffins stream. The raffinate stream,
consisting mainly of iso- or cyclic-kerosene range components, is often blended into
jet fuel.
3.4.2 Production of LAB

In the Pacol process, the normal-paraffins are dehydrogenated in a vapor phase reac-
tion to corresponding mono-olefins over a highly selective and active catalyst.
The DeFine process is a liquid-phase, selective hydrogenation of Diolefins in the
Pacol reactor effluent to corresponding mono-olefins over a catalyst bed.
The PEP process allows the selective removal of aromatics in the feed to the Detal
or detergent alkylation unit.
Detergent alkylation is a process in which benzene is alkylated with mono-olefins
produced in the Pacol unit to LAB using HF acid as a catalyst. In 1995, the first
commercial Detal process unit using a solid-bed catalyst alkylation process was com-
missioned. This technology abolished the use of liquid acid in the plant, reducing
capital investments, maintenance costs, and waste treatment.
The LAB technology is the most economical technology available today, and more
than 70% of the world’s LAB is produced using UOP technologies.
3.5 Methanol to Olefins

Conversion of methanol to hydrocarbons was first developed by Mobil to produce
gasoline (MTG) in 1975 using ZSM-5 catalysts [9]. Selectivity of methanol to ethy-
lene and propylene over ZSM-5 was generally low, with selectivities favoring heavier
more highly branched hydrocarbons and aromatics.
Iso-parafins
2 CH3 OH ↔ CH3 -OH-CH3 → C2 − C5 Olefins → Aromatics
C6 + Olefins
SAP0-34 ZSM-5 Figure 3.15 Framework of

SAPO-34 and ZSM-5.
Ethylene
Source: Chen et al. [13].
3.8 A
5.5 A © 2005, Elsevier.
Propylene
Isobutylene
Benzene
Ethanol conversion SAPO 34 exhibited the following features:
Characteristics of ZSM (medium-pore zeolite) for conversion of methanol are sum-

marized below:
● Major olefin product is propylene,
● Significant C5 + /aromatics byproduct formation, and
● Slow deactivation by coking.
In the 1980s, a group of scientists at Union Carbide (the group later became part
of UOP LLC) discovered SAPO-34, a silicon-, aluminum-, and phosphorous-based
molecular sieve, which is an excellent catalyst for conversion of methanol to ethylene
and propylene [10–12].
As shown in Figure 3.15, SAPO-34 has a smaller pore size (about 3.8 Å) compared
to that of ZSM-5 (about 5.5 Å) along with the small sizes of certain organic molecules
are keys to the MTO process [13].
The cavity dimensions and pore opening of SAPO-34 influence the diffusion
of heavy and/or branched hydrocarbons, and this leads to high selectivity to the
desired small linear olefins. Another key feature of the SAPO-34 molecular sieve
is its optimized acidity relative to aluminosilicate-based zeolitic materials. The
mild acid function on SAPO-34 affects conversion of olefins to paraffins by hydride
transfer reaction, leading to lower paraffinic byproduct formation. A further advan-
tage of SAPO-34 is that the majority of the C4 –C6 fraction product is olefinic. As
discussed later in this section, these olefinic compounds make the heavy byproduct
suitable for upgrading by olefin cracking.
For methanol conversion, SAPO-34 exhibited the following features:
● Major olefin products are ethylene and propylene,
● Fast deactivation by aromatic coke, and
● SAPO molecular sieves more stable than corresponding zeolite structure.
Based on discoveries of SAPO-34, UOP developed a process based on SAPO-34 [14,
15]. Working with Norsk Hydro in Norway, UOP successfully developed UOP/Hydro
MTO process. In development of the process, UOP built on its in-house FCC expe-
rience for fluidized reactor and regenerator.
Unrelated to MTO at the time, Atofina was at work in the 1990s developing
olefin-cracking technology. Shortly after, in 2000, Atofina (which later became
part of Total Petrochemicals and nowadays of Total Refining and Chemicals) and
3.5 Methanol to Olefins 61
Regen gas
MTO process integrated with
Olefin cracking process (OCP)
MTO Light Ethylene

DME
recovery Olefin
recovery Propylene
C 4+
Air Water
OCP
Sep
Methanol section
By-products
Figure 3.16 UOP advanced MTO process.
UOP formed a joint alliance to further develop olefin-cracking technology. This

collaboration led to development of the Total Petrochemicals/UOP olefin-cracking
process (OCP).
The OCP has been integrated with UOP/Hydro MTO process – this combination
of processes is the basis for advanced MTO.
A major milestone for MTO commercialization was the start-up of the semicom-
mercial unit in 2009, fully integrated MTO demonstration unit in Belgium, which
successfully demonstrated the performance of the integrated UOP/Hydro MTO Pro-
cess with the OCP.
The UOP/Hydro MTO process utilizes a fluidized reactor and regenerator sys-
tem to convert methanol to olefins using the SAPO-34 catalyst. The UOP/Hydro
MTO process can be operated on “crude” or undistilled methanol as well as with
pure (grade AA) methanol. The choice of feedstock quality generally depends on
project-specific situations because there can be advantages in either case. Figure 3.16
illustrates a simple flow diagram for the UOP advanced MTO process.
The methanol feed is preheated and then introduced into the reactor. The
conversion of methanol to olefins requires a selective catalyst that operates at
moderate-to-high temperatures. The reaction is exothermic, so heat can be recov-
ered from the reaction. Carbon or coke accumulates on the catalyst and requires
removal to maintain catalyst activity. The coke is removed by combustion with air
in a catalyst regenerator system. A fluidized-bed reactor and regenerator system is
ideally suited for the MTO process because it allows for heat removal and CCM.
The reactor operates in the vapor phase at temperatures between 350 and 550 ∘ C
and pressures between 2 and 4 bars. A slipstream of catalyst is circulated to the
fluidized-bed regenerator to maintain high activity. The operation of the reactor
system is characterized as stable steady-state.
The reactor effluent is cooled and quenched to separate water from the product
gas stream. The product gas is compressed and then unconverted oxygenates are
100
Light Olefin carbon yields, wt%
90
80
70
1.0 1.2 1.4 1.6 1.8 20
Propylene/ethylene (P/E) product ratio, wt.
Figure 3.17 Propylene-to-ethylene ratio in UOP advanced MTO process.
recovered and returned to the reactor. The reactor provides very high conversion so
there is no need for a large recycle stream. After the oxygenate recovery section, the
effluent is further processed in the fractionation and purification section to remove
contaminants and separate the key products from the byproduct components. Ethy-
lene and propylene are produced as polymer-grade products and sent to storage. The
C4 –C6 fraction is sent to the OCP reactor where it is selectively converted to light
olefins, majority of which is propylene. Typically, the propylene-to-ethylene ratio in
OCP reactor effluent is about 4. The OCP product is depropanized, the C3 and lighter
fraction is sent to MTO product recovery section, and residual C4 plus fraction is
taken as byproduct fuel.
As shown in Figure 3.17, the advanced MTO process, which is the integrated
MTO-OCP processes, can produce propylene to ethylene product ratio between
1.2 and 1.8 [16] to help meet the growing demand for propylene and additional
flexibility is achievable with the technologies, if desired.
3.6 Syngas Production

Syngas is a gas mixture consisting primarily of hydrogen, carbon monoxide, and very
often some amount of carbon dioxide. Syngas can be produced from many sources,
including natural gas, coal, biomass, or virtually any hydrocarbon feedstock, by reac-
tion with steam (steam reforming), carbon dioxide (dry reforming), or oxygen (par-
tial oxidation). Historically, natural gas has been the primary feedstock for syngas
production, but over the last 10 years or so, coal has been used as the feedstock for
making syngas in China since an abundance of source of natural gas is not available
in the country.
Syngas production technologies are the focus of the sequent sections, including
removal of impurities in raw natural gas, conversion of natural to syngas, and coal
gasification to syngas.
Pipeline spec:
Acid gas
Gas • CO2 < 2 – 8%
CO2
H2O N2 Hg C2+ • H2S < 4 ppm
composition H2S
• Hg < 0.01 μg/Nm3
• H2O < 2 – 8%
CH4, C2+,
H2O, H2S, Pipeline gas
Raw CO2, N2, Hg
natural
gas
LNG
LNG spec:
Acid gas N2 Hg C2+ • CO2 < 50 ppm
H2O
CO2 • H2S < 2 – 4 ppm
H2S • Hg < 0.01 μg/Nm3
• H2O < 0.1 ppmv
Figure 3.18 Requirements for removal of natural gas impurities.
3.6.1 Removal of Impurities in Raw Natural Gas

Raw natural gas typically consists of methane, heavier gaseous hydrocarbons, acid
gases (CO2 , H2 S), water, and mercury. Acid gases in combination with water are
highly corrosive and can rapidly destroy pipeline and equipment. Furthermore, H2 S
is hazardous for human beings and CO2 reduces natural gas stream heating value.
Thus, the concentration of impurities in the raw natural gas must be reduced to meet
the quality specification for pipeline transmission or producing liquefied natural gas
(LNG) (Figure 3.18).
The technologies that are commonly used to remove acid gases include chemical
and physical adsorption processes, cryogenic condensation, and membranes.
Chemical adsorption: The amine-based process is commonly used for chemical
adsorption of acid gases. Figure 3.19 shows the UOP Amine GuardTM FS process
[17]. CO2 and H2 S react chemically with the amine solvent (USARSOLTM ) and
form dissolved chemical compounds in the amine adsorber. The solvent is
regenerated in the amine stripper by application of heat. For natural gas treating,
the Amine Guard is capable of selective removal of H2 S and partial removal
of CO2 for pipeline specification. Another commonly used process for acid
gas removal from natural gas is UOP BenfieldTM process [18]. It is a thermally
regenerated cyclical solvent process using an activated, inhibited hot potassium
carbonate solution to remove CO2 and H2 S (Figure 3.20).
Physical adsorption: Physical adsorption processes are used when the feed gas is
characterized by high partial pressures of acid gases and low temperatures. In
a physical adsorption process, the solvent interacts only physically with the dis-
solved gas with thermodynamic properties such that the relative adsorption of
CO2 and H2 S is more favored over the other components of the gas mixture. CO2
and H2S are then removed by the strippers. Figure 3.21 shows the UOP SelexolTM
technology for H2 S removal and CO2 capture [19].
Membranes: Membranes are thin, semipermeable barriers that selectively separate

certain components. They are characterized by permeability or capacity (flux)
and selectivity. Feed gas enters along the side of the membrane and as the feed
gas travels the membrane, CO2 , H2 S, and H2 O permeate through the membrane
and are collected in the permeate tube. The driving force for transport is the
high-feed and low-permeate pressures. The gas on the feed side does not per-
meate leaves through the side of the element opposite the feed position. In high
H2 S applications, most polymers have been shown to have very little resistance
to H2 S plasticization (“softening” of the membrane) and as a consequence can
only be used at very low H2 S partial pressure limits. UOP’s cellulose acetate
members in contract have been shown to be applicable at extreme H2 S partial
pressure conditions [20]. Figures 3.22 and 3.23 show one-stage and two-stage
UOP SeparexTM processes, respectively. The simplest membrane-processing
scheme is the one-stage process. A feed gas is separated into a permeate stream
rich in acid gas and hydrocarbon-rich residual stream. In high H2 S-removal
applications, hydrocarbons permeating with the acid gases may be rejected and
therefore are not lost to future gas sales. If re-injection is not a problem, then
multistage system can be used to achieve higher hydrocarbon recoveries.
Mercury removal: The primary reason for removing mercury from natural gas is
to protect downstream equipment and catalysts. UOP HgSIVTM regenerative
system is designed to dry the natural gas and remove mercury (Figure 3.24)
[21]. Since the sorption sites for mercury removal are separate from additive to
the dehydration sites, mercury removal is accomplished by replacing a portion
of the dehydration-grade molecular sieve with HgSIV adsorbents. The dryer
bed size does not have to be increased to remove both water and mercury.
Mercury is removed from HgSIV adsorbents at conventional dryer regeneration
temperatures. Properly regenerated HgSIV adsorbents can be disposed of as
nonhazardous waste. Mercury can be removed from the regeneration gas by
condensation when the mercury level in the feed gas is high. Treatment of the
HgSIV spent regeneration gas is usually not needed for low-to-moderate mercury
feed gas levels.
3.6.2 Conversion of Natural Gas to Syngas

The catalyst systems used for syngas production from natural gas and subsequent
processes are susceptible to deactivation by sulfur. To protect the catalysts, there is
a natural gas purification section to convert sulfur species to hydrogen sulfide, fol-
lowed by a zinc oxide–based adsorption process to remove the hydrogen sulfide [22].
Hydrodesulfurization∶ R − S + H2 → R − H + H2 S
H2 S Adsorption∶ ZnO(s) + H2 S(g) → ZnS(s) + H2 O(g)

There are three types of process widely used by industry for conversion of natural
gas to syngas: steam reforming, partial oxidation, and autothermal reforming.
Acid gas
Sweet Acid gas
gas KO drum
Make-up
water Acid gas
cooler
Amine Lean Lean
absorber solution solution Reflux
Flash gas pump cooler pump
to fuel
header Filtration
system Amine Make-up
Feed Make-up stripper water
gas water
Rich
flash drum
Amine
reboiler
Lean/rich Lean solution
exchanger booster pump
Figure 3.19 UOP amine guard FS process (1-stage) for acid gas removal. Source: Based on
“UOP Amine GuardTM FS Technology for Acid Gas Removal”, https://honeywelluop
.chinacloudsites.cn/wp-content/uploads/2011/02/UOP-Amine-Guard-Technology-for-Acid-
Gas-Removal-tech-presentation.pdf [17].
3.6.2.1 Steam Reforming

Steam reforming of natural gas is one of the most common processes for syngas
production. In this catalytic process, methane reacts with steam at 800–900 ∘ C and
moderate pressure (around 40 bars) in presence of a catalyst to produce syngas [23].
The H2 /CO ratio produced in the reforming reaction is close to 3.
CH4 + H2 O → 3H2 + CO
The steam reforming of methane is an endothermic reaction. The reaction heat is
provided by sending the feed stock consisting of a mixture of desulfurized natural
gas and steam (typically up to three steam to carbon ratio) to a tubular reactor. The
reformer tubes are packed with catalyst and heated externally by burners to achieve
high temperatures needed for considerable syngas yields, making steam reforming
a major energy consumer. Several different transition metals can be used as catalyst
in the reforming process, but nickel is the most widely used because it is usually the
cheapest and leads to satisfactory results [24].
Because this process of steam reforming leads to highest H2 /CO ratio, it is consid-
ered ideal for hydrogen production.
3.6.2.2 Partial Oxidation

The partial oxidation of methane is a process whereby methane reacts directly with
oxygen or air in the presence of a catalyst, and product of this reaction is syngas with
a good H2/CO ratio. The partial oxidation reaction is shown below.
CH4 + 1/2 O2 → CO + 2H2
Acid gas Acid gas

condenser
Acid gas
Pure gas knock-out
Reflux Benfield
drum
pump Regenerator
Benfield
absorber Condensate
pump
Lean
solution
filter
Feed
gas Hydraulic
turbine
(options)
Lean
Carbonate Condensate
solution
reboiler reboiler
flash
Lean solution
pump
Figure 3.20 UOP benfield process for acid gas removal.
In a noncatalytic partial oxidation, the production of syngas is carried out at

temperatures of 1500–1200 ∘ C without a catalyst. A noncatalytic partial oxidation
process was developed by Texaco and Shell, which results in high syngas yields at
high temperature and pressures [25]. The problems related to the homogeneous
process are high temperatures, long residence time as well as excessive coke
formation, which reduce the controllability of the reaction [26]. The use of a catalyst
in the production of syngas lowers the reaction temperature to around 800–900 ∘ C
[27]. Methane is converted with oxygen (or air) over a noble metal (Pt, Rh, Ir,
Pd) or a non-noble metal (Ni, Co) catalyst to syngas in a single-step process. The
catalytic partial oxidation can be used only if the sulfur content of natural gas is
below 60 ppm [26]. Higher sulfur content would poison the catalyst and therefore
noncatalytic partial oxidation should be used for such fuels. Many studies have
been conducted for the catalytic partial oxidation on reaction mechanism, reactor
configuration, reactor simulations, and catalyst systems [26].
3.6.2.3 Autothermal Reforming

The autothermal reforming of methane is a combination of both steam reforming
and partial oxidation. Thus, in the steam reforming, there is contact with a flow of
gaseous oxygen in the present of a catalyst. The autothermal reforming of methane
process was designed to save energy, because the thermal energy required is gener-
ated by the partial oxidation of methane [28]. The H2 /CO ratio of the syngas obtained
in the autothermal reforming is a function of the gaseous reactant fractions intro-
duced in the process input. Typically, the H2 /CO ratio ranges between 1 and 2 [29].
Treated
gas
CO2
absorber
H2S Acid
CO2
Lean stripper gas
solution Reflux
filter accumulator
Sulfur Makeup
absorber H2S Reflux water
pump Export
Feed concentrator
gas water
Stripper
Packinox reboiler
exchanger
Figure 3.21 UOP selexol technology for H2 S removal and CO2 capture. Source: Based on
“UOP SelexolTM Technology Applications for CO2 Capture”, https://www.uwyo.edu/eori/_
files/co2conference09/mike%20clark%20revised%20selexol_wyoming_conference.pdf [19].
Figure 3.22 One-stage flow scheme. Residue

(Low acid
gas)
Feed
Permeate
(High acid
gas)
Figure 3.23 Two-stage flow Residue

scheme.
Feed
Permeate
Permeate
Stage 1
The autothermal reforming is also utilized as a “secondary reformer” (for lowering

the CH4 residue), and it is placed after a primary steam reformer in syngas plants
integrated with ammonia synthesis reactors [30].
3.6.3 Coal Gasification to Syngas

Coal gasification is an old process for converting coal partially or completely to com-
bustible gases. After purification, these gases can be used as fuels or as raw materials
Liquid
H2O
Feed gas
Liquid
Hg
Adsorption Adsorption Regeneration
Dry and Hg-Free

product gas
Product slipstream regeneration gas
Figure 3.24 UOP mercury removal and recovery system. Source: Based on Corvini et al.
[21].
Table 3.2 Major chemical reactions in goal gasification.
Gasification with oxygen C + 1/2 O2 → CO
Combustion with oxygen C + O2 → CO2

Gasification with carbon dioxide C + CO2 → 2CO
Gasification with steam C + H2 O → CO + H2
Gasification with hydrogen C + 2H2 → CH4
Water gas shift CO + H2 O → CO2 + H2
Methanation CO + 3H2 → CH4 + H2 O
for chemical manufacturing. In recent years, production of chemicals by coal gasi-

fication has had a significant growth led by China. China does not have sufficient
sources of crude oil and natural gas and coal is the dominant energy source, account-
ing for about 64% of China’s total primary energy consumption [31]. China has con-
structed coal gasification plants in a major way to produce syngas for manufacturing
of a variety of downstream chemicals.
Coal gasification is a process that converts coal into a gaseous mixture composed
primarily of carbon monoxide, carbon dioxide, hydrogen, and methane at high tem-
perature in the presence of oxygen and steam. The chemistry of gasification is quite
complex. Some of the major chemical reactions are shown in Table 3.2.
As the coal is exposed to rising temperature in the gasifier, devolatilization and
breaking of weaker chemical bonds occurs, releasing volatile gases such as tar
vapors, methane, and hydrogen. The volatile products and some of the carbon react
with limited oxygen to form carbon dioxide and carbon monoxide, releasing heat
needed for the subsequent gasification reactions. The remaining carbon reacts with
carbon dioxide, steam, and hydrogen to produce carbon monoxide, hydrogen, and
methane. Water gas shift and methanation are reversible reactions taking place
simultaneously based on the gasifier conditions.
There are three major types of gasifiers: moving-bed (also called fixed-bed),
fluidized-bed, and entrained-flow. Each type is defined by how the reactor brings the
coal in contact with the reacting gas and the temperature profile inside the gasifier.
Moving-bed reactor: In a moving-bed reactor, large particles of coal move slowly
down through the gasifier while reacting with gases moving up (Figure 3.25) [32]
at moderate pressures (25–30 bars). Reactions within the gasifier occur in differ-
ent “zones.” In the “drying zone” at top of the gasifier, the entering coal is heated
and dried, while cooling the product gas before it leaves the reactor. The coal is fur-
ther heated and devolatized by the high-temperature gas as it descends through
the “carbonization zone.” In the next zone, the “gasification zone,” the devola-
tized coal is gasified by reaction with steam and carbon dioxide. Near the bottom
of the gasifier, in the “combustion zone,” which operates at the highest tempera-
ture, oxygen reacts with the remaining char. Due to the countercurrent flow and
the inherent recuperation of the sensible energy in the gas through devolatiliza-
tion and coal drying, this type of gasifier gives the highest cold gas efficiency of
any of the gasifiers.
Fluidized-bed reactor: Fluidized-bed reactors employ back-mixing and efficiently
mix feed coal particles with coal particles already undergoing gasification in the
reactor vessel. Coal of small particles (<6 mm) is supplied through the side of
the reactor, and oxidant and steam are supplied near the bottom (Figure 3.26)
[32]. Due to the thorough mixing within the gasifier, a constant temperature is
sustained in the reactor bed. The gasifiers normally operate at moderately high
temperature to achieve an acceptable carbon conversion (e.g. 90–95%) and to
decompose most of the tar, oils, phenols, and other liquid byproducts. However,
the operating temperatures are usually less than the ash fusion temperature so as
to avoid clinker formation and the possibility of de-fluidization of the bed. This,
in turn, means that fluidized-bed gasifiers are best suited to relatively reactive
coals, low-rank coals, and other fuels such as biomass [32].
Entrained-flow reactor: In entrained-flow systems, fine coal feed and oxidant (air
or oxygen) and/or steam are fed co-currently to the gasifier (Figure 3.27) [32].
Entrained-flow gasifiers operate at high temperature and pressure with turbulent
flow, which causes rapid feed conversion and allows high throughput. The
gasification reactions occur at a very high rate (typical residence time is on the
order of few seconds), with high carbon conversion efficiencies (98–99.5%).
Entrained-flow gasifiers have the ability to handle practically any coal feedstock
and produce a clean, tar-free syngas. Given the high operating temperatures, this
type of gasifier melts the coal ash into vitreous inert slag. Different systems may
use different coal feed systems (dry or water slurry) and heat recovery systems.
A comparison of moving-bed, fluidized-bed, and entrained-flow gasifiers is shown
in Table 3.3 [32].
g g
Coal Gasifier
bottom
Gas
Coal
Gas
Steam,
oxygen
or air
Ash
Gasifier
Steam, bottom
oxygen 0 250 500 750 1000 1250 1500
or air Ash Temperature - °C
Figure 3.25 Moving-bed gasifier.
Gas
Gasifier
bottom
Coal Gas
Coal
Steam,
oxygen
or air
Steam, Gasifier Ash

oxygen bottom
0 250 500 750 1000 1250 1500
or air
Ash Temperature - °C
Figure 3.26 Fluidized-bed gasifier. Source: Based on U.S. Department of Energy,

Gasification Technology [32].
Raw syngas should be purified and the H2 /CO ratio should be adjusted before it
is used as feed for downstream processing. Desired H2 /CO values for various syngas
applications are shown below:
● H 2 production: Pure H2
● Ammonia: H2 /CO = 3
● Methanol: H2 /CO = 2
To meet the downstream process requirements, the syngas is passed through a

multistage, fixed-bed reactor containing shift catalyst to convert carbon monoxide
and water into additional H2 and carbon dioxide according to the following reaction
Steam,
Coal Gasifier
oxygen
bottom
or air
Coal Steam,
oxygen
or air
Gas
Gas Slag
Gasifier
bottom
0 250 500 750 1000 1250 1500
Temperature - °C
Slag
Figure 3.27 Entrained-flow gasifier. Source: Based on U.S. Department of Energy,

Gasification Technology [32].
Table 3.3 Comparison of gasifier characteristics.
Moving bed Fluidized bed Entrained flow
Ash condition Dry Slagging Dry Agglomerate Slagging
Coal feed ∼2 in. ∼2 in. ∼1/4 in. ∼1/4 in. ∼100

Mesh
Fines Limited Better Good Better Unlimited
than dry
ash
Coal rank Low High Low Any Any
Gas 800–1200 800–1200 1700–1900 1700–1900 >2300
temperature
(∘ F)
Oxidant Low Low Moderate Moderate High
required
Steam required High Low Moderate Moderate Low
Issues Fines and hydrocarbon liquids Carbon conversion Raw gas
cooling
Source: Based on U.S. Department of Energy, Gasification Technology [32].
known as the gas water shift reaction (GWSR):

CO + H2 O ↔ CO2 + H2
GWSR is a moderately exothermic reversible reaction. Therefore, with increasing
temperature, the reaction rate increases, but the conversion of reactants to products
becomes less favorable.
To take advantage of both the thermodynamics and kinetics of the reaction,
the industrial-scale GWSR is conducted in multiple adiabatic stages consisting
of a high-temperature shift followed by a low-temperature shift with intersystem

cooling [33]. The initial high-temperature shift at 310–450 ∘ C results in incomplete
conversion of carbon monoxide and a 2–4% carbon monoxide exit composition.
To shift the equilibrium toward hydrogen production, a subsequent low tempera-
ture shift reactor at 200–250 ∘ C is employed to produce a carbon monoxide exist
composition of less than 1%. Due to the different reaction conditions, different
catalysts must be employed at each stage to ensure optimal activity. The commercial
high-temperature shift catalyst is the iron oxide-chromium oxide catalyst and the
low-temperature shift catalyst is a copper-based catalyst. The copper-based catalyst
can be easily poisoned by sulfur compounds, whereas the iron-based catalyst is
sulfur tolerant. If sulfur is present in the feed, a guard bed should be used to remove
the sulfur compounds to protect the copper-based catalyst [33].
References
1 Foster, J. (2013). Can Shale Gale Save the Naphtha Crackers? London:
Platts.Platts special report: petrochemicals
2 Graig, R.G. and Spence, D.C. (1986). Catalytic dehydrogenation of liquefied
petroleum gas by the Houdry Catofin and Catadiene processes. In: Hand-
book of Petroleum Refining Processes (ed. R.A. Meyers). New York (Chapter
4.1: McGraw-Hill.
3 Sanfilippo, D., Buonomo, F., Fusco, G., and Miracca, I. (1998). Paraffins activa-
tion through fluidized bed dehydrogenation: the answer to light olefins demand
increase. Studies in Surface Science and Catalysis 119: 919–924.
4 Iezzi, R. Bartolini, A. (1997). Process for dehydrogenating light paraffins in a flu-
idized bed reactor. US Patent 5,633,421.
5 Luckenbach, E.C. Zenz, F.A. Papa, Giovanni Bertolini, A. (1997). Fluidized bed
reactor and process for performing reactions therein. US Patent 5,656,243.
6 Bhasin, M.M., McCain, J.H., Vora, B.V. et al. (2001). Dehydrogenation and
oxydehydrogenation of paraffins to olefins. Applied Catalysis A: General 221:
397–419.
7 Shakur, M. (2018). New generation of OleflexTM technology. Presented in UOP
International Technology Conference, Beijing, China.
8 Vora, B.V. (2012). Development of dehydrogenation catalysts and processes. Top-
ics in Catalysis 55: 1297–1308.
9 Chang, Silvestri, Smith, R.L. Aromatization reactions. US Patents 3,894,103;
3,928,483.
10 Kaiser, S.W. Production of light olefins. US Patent 4,499,327; 1985.
11 Kaiser, S.W. and Arabian, J. (1985). The Sciences and Engineering 10: 361.
12 Lewis, J.M.O. (1998). Methanol to olefins process using silicoalumino-phosphate
molecular sieve catalysts. Catalysis (ed. J.W. Ward). Amsterdam: Elsevier.
13 Chen, J.Q., Bozzano, A., Glover, B. et al. (2005). Recent advancements in ethy-
lene and propylene production using the UOP/hydro MTO process. Catalysis
Today 106: 103–107.
References 73
14 Kaiser, S.W. Production of light olefins. US Patents 4,499,327; 4,677,242; 1987.

15 Vora, B.V., Marker, T.L., Barger, P.T. et al. (1997). Economic route for natu-
ral gas conversion to ethylene and propylene. Studies in Surface Science and
Catalysis 107: 87–98.
16 Gregor, J.H. (2012). Maximize profitability and olefin production with UOP’s
advanced MTO technology. HIS word Methanol Conference, Madrid, Spain
(27–29 November 2012).
17 “UOP Amine GuardTM FS Technology for Acid Gas Removal”; https://www.uop
.com/tech-library.
18 Echt, W.; (2013). “Technologies for efficient purification of natural and synthesis
gases”; www.uop.com (accessed 4 September 2020).
19 “UOP SelexolTM Technology for acid gas removal”; https://www.uop.com/tech-
library.
20 Cnop, T. Dortmundt, D. Schott, M.; Continued development of gas separation
membranes for highly sour service. https://www.uop.com/tech-library.
21 Corvini, G. Stiltner, J. Clark, K.; Mercury removal from natural gas and liquid
streams. https://docplayer.net/20958934-Mercury-removal-from-natural-gas-
liquid-streams-john-markovs-uop-des-plaines-illinois-jack-corvini-uop-houston-
texas.html.
22 Liu, K., Song, C., and Subramani, V. (2010). Hydrogen and Syngas Production
and Purification Technologies. Hoboken, NJ: Wiley.
23 Saumel, P. (2003). GTL technology challenges and opportunities. Catalysis
Bulletin of the Catalysis Society of India 2: 82–99.
24 Neiva, L.S. and Gama, L. (2010). A study on the characteristics of the reforming
of methane: a review. Brazilian Journal of Petroleum and Gas 4 (3): 119–127.
25 Pena, M., Gomez, J., and Fierro, J.L. (1996). New catalytic routes for syngas and
hydrogen production. Applied Catalysis A: General 144: 7–57.
26 Saleh Al-Sayari, A. (2015). Recent development in the partial oxidation of
methane to syngas. The Open Catalysis Journal 6: 17–28.
27 Liu, J.A.; (2006). Kinetics, catalysts and mechanism of methane steam reform-
ing. Master Thesis of Science in Chemical Engineering, Worcester Polytechnic
Institute.
28 Ayabe, S., Omoto, H., Utaka, T. et al. (2003). Catalytic autothermal reforming
of methane and propane over supported metal catalysts. Applied Catalysis A:
General 241: 261–269.
29 Palm, C., Cremer, P., Peters, R., and Stolten, D. (2002). Small-scale testing of a
previous metal catalyst in the autothermal reforming of various hydrocarbon
feeds. Journal of Power Sources 106: 231–237.
30 Ghoneim, S.A., El-Salamony, R.A., and El-Temtamy, S.A. (2016). Review on
innovative catalytic reforming of natural gas to syngas. World Journal of Engi-
neering and Technology 5: 116–139.
31 National Bureau of Statistics (2016, 29 February). National Economy and Social
Development Statistic Bulletin 2015 (Internet). Beijing: National Bureau of
Statistics of the People’s Republic of China http://www.stats.gov.cn/english/
PressRelease/201602/t20160229_1324019.html.
32 Overview of DOE’s gasification program; U.S. Department of Energy, Gasifica-

tion Technology. Available online: http://www.netl.gov/technologies/coalpower/
gasification/pubs/pdf/DOE%20Gasification%20Program%Overview%202009%2009-
03%20vls.pdf.
33 Byron, S.R.J., Loganthan, M., and Shantha, M.S. (2010). A review of the water
gas shift reaction. International Journal of Chemical Reactor Engineering 8: 1–32.
http://www.netl.doe.gov/technologies/coalpower/gasification/pubs/pdf/DOE
%20Gasification%20Program%20Overview%202009%2009-03%20v1s.pdf (accessed
on 20 September 2009)
75
Uses of Olefins and Synthesis Gas
4.1 Introduction
As discussed in Chapter 3, ethylene and propylene are the main petrochemical build-
ing blocks. Ethylene is used as a feedstock to manufacture polyethylene, ethylene
chloride, ethylene oxide (EO), and other petrochemicals (Figure 4.1). These mate-
rials are very useful for the packaging, plastic processing, construction, houseware,
and textile industries. Most propylene is used to make polypropylene, but it is also a
feedstock to produce propylene oxide, acrylic acid, cumene, and many other chemi-
cal derivatives (Figure 4.2). Not only plastic processing, the packaging industry, and
the furnishing section, but also the automotive industry are users of propylene and
its derivatives.
In addition to ethylene and propylene, other olefins also find uses in fuel and petro-
chemical chemical applications. As discussed in Chapter 3, the leading applications
of butylene are for production of gasoline-blending components (alkylate, MTBE,
and ETBE) while heavy normal-olefins are primarily used for producing linear alkyl-
benzene (LAB), a dominant precursor of biodegradable detergents. Syngas, on the
other hand, represents a suitable alternative, from resources other than crude oil, to
produce chemicals with commercial significance such as hydrogen, ammonia, and
methanol. In this chapter, the uses of olefins (ethylene and propylene) and syngas
as raw materials to produce downstream products will be discussed.
4.2 Uses of Ethylene
4.2.1 Ethylene to Polyethylene

Polyethylene is the most widely used plastic, made by repeating –CH2 – unit.
There are three major types of polyethylene produced with different methods: low-
density polyethylene (LDPE) referred to as high-pressure polyethylene, high-density
polyethylene (HDPE) known as low-pressure polyethylene, and linear low-density
polyethylene (LLDPE).

76 4 Uses of Olefins and Synthesis Gas
Polyethylene
HDPE Containers, crates, drums

bottles, housewares
LDPE
Packaging, film bags,
toys, housewares
LLDPE
Ethylene Ethylene PVC film, window frames,

Vinyl chloride
dichloride pipes
Ethylene glycol PET, fibers, anti-freeze,

Ethylene oxide
Ethanol amines Detergents, cosmetics
Others
Figure 4.1 Derivatives of ethylene.
Polypropylene Plastic films, fibers,

packaging, etc.
Propylene Polyurethane forms,

oxide antifreeze, etc.
Propylene
Acrylic fibers, ABS
Acrylonitrile
resins, nylon, etc.
Others
Figure 4.2 Derivatives of propylene.
The high-pressure method produces LDPE in tubular or autoclave reactors [1, 2] at

a pressure between 1000 and 3000 bars and a temperature of 200–300 ∘ C with oxygen
or peroxide as initiator for the free radical polymerization. Unreacted ethylene at
reactor outlet is removed and recycled. The resulting LDPE has a density range of
0.91–0.93 g/cc.
There are three processes developed for low-pressure ethylene polymerization:
solution process, slurry process, and gas-phase process. Most of HDPE is made by
the slurry process [3] with a co-monomer (butene-1, hexene-1, or octene-1), in the
presence of a chromium on silica or Ziegler–Natta catalyst, at a pressure 15–35 bars
and a temperature 80–110 ∘ C. The HPPE has a density range of 0.94–0.96 g/cc.
LLDPE is mostly produced by low-pressure gas-phase and solution processes by
polymerization of ethylene with alpha olefins (butene-1, henxene-1, octene-1) with
Ziegle-Natta catalyst. LLDPE has a density of 0.91–0.94 g/cc.
The density difference is a result of variation in molecular structures. LPPE has
low density due to its extensive brancking structure. It is softest and least crystalline
of the polythylenes. HDPE is a denser material, more ridged, and less permeable
since its polymer chains include less bracking showing high crystallinity. LLDPE
has a basic linear structure with short branchings. When compared with LDPE, the
4.2 Uses of Ethylene 77
Ethylene
Ethylene Pyrolysis Vinyl Polymerization Polyvinyl
dichloride chloride chloride
(EDC) (VCM) (PVC)
Chlorine
Figure 4.3 Major steps for manufacturing of PVC.
short chains in LLDPE are able to move against on another upon elongation without
entangling together, resulting in improved strength and impact properties.
4.2.2 Ethylene to Polyvinyl Chloride

Polyvinyl chloride (PVC) is a commonly used plastic favored in construction, pip-
ing, and many other industries due to the properties of the polymer, including light
weight, chemical resistance, and versatility. Manufacturing of PVC involves the steps
shown in Figure 4.3.
Vinyl chloride (VCM) production involves the following reactions:
Direct chlorination CH2 = CH2 + Cl2 → ClCH2 − CH2 Cl
Oxychlorination CH2 = CH2 + 2 HCl + 1/2 O2 → ClCH2 − CH2 Cl + H2 O
Pyrolysis 2ClCH2 − CH2 Cl → 2CH2 = CHCl + 2HCl
Overall 2CH2 = CH2 + Cl2 + 1/2 O2 → 2 CH2 = CHCl + H2 O

Ethylene dichloride (EDC) is produced by direct chlorination or oxychlorination
using hydrogen chloride as the chlorination agent. The direct chlorination process
involves a liquid- or vapor-phase reaction of ethylene and chlorine using an FeCl3
catalyst. This exothermic reaction is controlled by mass transfer and can be operated
at 50–120 ∘ C to minimize the undesired by-product formation. Oxychlorination is
used in vinyl chloride production because it consumes the hydrochloric acid (HCl),
a major byproduct of vinyl chloride production. The purified EDC is sent to a thermal
cracker (pyrolyzer), which outputs VCM and HCl.
PVC is produced by polymerization of VCM as shown in Figure 4.4 [4].
There are three types of PVC manufacturing processes: suspension polymeriza-
tion, emulsion polymerization, and bulk polymerization. PVC made from the sus-
pension polymerization is by far the most common. The suspension polymerization
affords particles with average diameters of 100–180 μm. VCM and water are intro-
duced into the reactor along with a polymerization initiator and other additives. The
Figure 4.4 Polymerization of VCM to PVC. Source: H CI

Chanda and Roy [4]. H CI
n C
C C C n
H H
H H
contents of the reaction vessel are pressurized and continually mixed to maintain
the suspension and ensure a uniform particle size of the PVC resin. The reaction is
exothermic and thus requires cooling. As the volume is reduced during the reaction
(PVC is denser than VCM), water is continually added to the mixture to maintain
the suspension [5].
4.2.3 Ethylene to Monoethylene Glycol

Monoethylene glycol (MEG) is mainly used for two purposes: (i) for antifreeze for-
mulations, and (ii) as a raw material in the manufacture of polyethylene terephtha-
late (PET), which is subsequently used in the production of polyester fibers, resins,
films, and other consumables. MEG is conventionally produced through the hydrol-
ysis of EO, which is obtained via ethylene oxidation.
Nearly all the EO production capacity is based on direct oxidation using a silver
catalyst.
Ag
CH2=CH2 + ½ O2 H2C CH2
O
Air or high-purity oxygen is mixed with recycled gas and fresh ethylene [6].
Vapor-phase oxidation inhibitors such as EDC or vinyl chloride or other halo-
genated compounds are added in the inlet of the reactor to retard carbon dioxide
formation [7]. The gas is fed to s multi-tube reactor in which the reactor temperature
is controlled at 220–300 ∘ C and pressure is set at 10–30 bars. The oxygen-based EO
process is preferred over the air route in large plants due to higher yields and less
catalyst required [7].
EO is further processed to produce MEG by a liquid-phase, noncatalytic hydrolysis
reaction operated at an elevated temperature.
H2C CH2 + H2O HOCH2CH2OH
O
This conventional process was started in 1937 by Union Carbide [6]. In this reac-
tion system, MEG inevitably reacts with remaining EO generating diethylene glycol
and triethylene glycol.
Selectivity to the different glycols is controlled by varying the ratio of water to EO
with large excess of water promoting MEG selectivity. Removing the excess water is
energy intensive and requires capital investment in evaporators. These factors limit
the amount of excess water, which can be used for control of the selectivity to MEG.
To maximize the MEG selectivity, many different systems have been studied. Mit-
subishi and Shell developed a two-step process [8] called OMEGA process.
C2 H4 O + CO2 → C3 H4 O3
C3 H4 O3 + H2 O → HOC2 H4 OH + CO2
In the first step, EO is reacted with CO2 to form ethylene carbonate. In the second
step, the ethylene carbonate is catalytically hydrolyzed to MEG and CO2 released in
this second reaction step is recycled back to the first reaction step.
4.3 Uses of Propylene

4.3.1 Propylene to Polypropylene
Polypropylene (PP) is a thermoplastic polymer used in a wide variety of applications.
It belongs to the group of polyolefins and is partially crystalline and nonpolar. Its
properties are similar to polyethylene, but it is slightly harder and more heat resis-
tant. It is a white, mechanically rugged material and has a high chemical resistance
[9]. Polypropylene is the second-most widely produced commodity plastic (after
polyethylene) and it is often used in plastic films, fibers, and packaging.
Polypropylene is produced by the chain-growth polymerization of monomer
propylene.
H
H2C =C Polymerization CH2 CH
CH3 CH3 n
Polypropylene can be categorized as atactic polypropylene, syndiotactic

polypropylene, and isotactic polypropylene. In case of atactic polypropylene,
the methyl group (–CH3 ) is randomly aligned, alternating for syndiotactic
polypropylene, and evenly for isotactic polypropylene (Figure 4.5). This has an
impact on the crystallinity and thermal properties.
The term tacticity describes how the methyl group is oriented in the polymer
chain, usually indicated in percent, using the isotactic index. As the methyl group
is in isotactic propylene consistently located at the same side, the higher the iso-
tacticity, the greater the crystallinity, and thus also the softening point, rigidity, and
hardness [10]. Only the isotactic form is of commercial interest as a thermoplastic.
Atactic polypropylene, due to its lack of crystallinity, is too soft and rubber-like to
be of wide commercial use as a thermoplastic, while syndiotactic polypropylene is
more difficult to prepare using metallocene catalysts.
Atactic polypropylene
Isotactic polypropylene
Syndiotactic polypropylene
Figure 4.5 Molecular structures of polypropylene.

There are three different types of polypropylene: homopolymer, random copoly-

mer, and impact copolymer. Homopolymer contains only propylene monomers
having high stiffness, low density, good chemical resistance, and relatively
high-temperature resistance. However, homopolymer has poor impact resistance,
especially at low temperature. Polypropylene-containing ethylene as a comonomer
in the polypropylene chains at levels in the range of 1–8%, and this is referred to as
a random copolymer. Random copolymers have higher impact resistance, higher
clarity, and more flexibility than homopolymers. Impact copolymer has an ethylene
content of 45–46% giving better impact strength at lower temperatures, as well as
higher melting points, than random copolymers.
The industrial production process can be grouped into gas-phase polymerization,
bulk polymerization, and slurry polymerization. All state-of-the-art processes use
either gas-phase or bulk reactor systems [11].
Gas-phase polymerization: The gas-phase polymerization is carried out in a
fluidized-bed reactor, propylene is passed over a bed containing the heteroge-
neous (solid) catalyst, and the formed polymer is separated as a fine powder.
Unreacted gas is recycled and fed back into the reactor.
Bulk polymerization: In bulk polymerization, propylene monomer is fed into the
reactor loop in liquid phase. Insoluble polymer formed on the suspended cata-
lysts precipitates out to form a slurry. The formed polymer is withdrawn and any
unreacted monomer is flashed off.
Slurry polymerization: In the slurry polymerization, typically C4 –C6 alkanes are uti-
lized as inert diluent to suspend the growing polymer particles. Propylene is intro-
duced into the mixture as a gas.
Ziegler–Natta catalysts are used in the polymerization process. These catalysts
are generally produced by interaction of titanium and magnesium chlorides and
aluminum alky compounds. New metallocene catalysts open the possibility to
synthesize the polypropylene with tighter control of polymer molecular weight,
molecular weight distribution, and crystalline structure as well as long-chain
branched or blocky copolymers with good properties.
4.3.2 Propylene to Propylene Oxide

Propylene oxide production is the second-largest propylene outlet, after polypropy-
lene, accounting for about 7–8% of the global propylene consumption. About
60–70% of propylene oxide is converted to polyether polyols. These polyols are
building blocks in the production of polyurethane plastics. About 20% of propylene
oxide is hydrolyzed into propylene glycol for further processing to make unsatu-
rated polyester resins or for antifreeze and solvent applications. Other products
derived from propylene oxide are polypropylene glycol, propylene glycol ethers,
and propylene carbonate.
There are two major processes for the manufacture of propylene oxide: propylene
chlorohydrin process and propylene oxidation process using peroxides.
The chlorohydrination process consists of formation of propylene chlorohydrin by
the reaction between hypochlorous acid and propylene. The propylene chlorohydrin
is epoxidized to propylene oxide by a 10% solution of lime or NaOH.
Cl2 + H2O HOCl + HCl
CH3–CH=CH2 + HOCl CH3CH–CH2

OH Cl
2 CH3CHOH–CH2Cl + Ca(OH)2 or NaOH 2 CH3CH–CH2 + CaCl2 or NaCl + 2 H2O
O
Byproducts formed during the reactions are 1,2-dichloropropane and chlorinated

di-isopropyl ether.
Some of the disadvantages of the chlorohydrination process include use of costly
chlorine, production of weak calcium chloride byproduct, and corrosion problem
due to chlorine handling.
The other route to produce propylene oxide involves oxidation of propylene with
an organic peroxide.
CH3 CH-CH2 + RO2 H → CH3 CHCH2 O + ROH
The process is practiced with four hydroperoxides [12]: tert-butyl hydroperoxide,
ethylbenzene hydroperoxide, cumene hydroperoxide, and hydrogen peroxide.
One of the main benefits of the process using hydrogen peroxide to produce
propylene oxide (HPPO process) is that it eliminates the need for additional
infrastructure or marketing of coproducts that are produced by using other types of
hydroperoxides.
4.3.3 Propylene to Acrylonitrile

Acrylonitrile is a commodity petrochemical produced primarily from propylene.
The primary use of acrylonitrile is as the raw material for the manufacture of
acrylic and modacrylic fibers. Other major uses include the production of plastics
acrylonitrile-butadiene-styrene (ABS) and styrene-acrylonitrile (SAN), nitrile
rubber, adiponitrile, and acrylamide.
Acrylonitrile is produced by catalytic ammoxidation of propylene, also known as
the SOHIO process.
2 CH3 − CH = CH2 + 2 NH3 + 3 O2 → 2 CH2 = CH-C𝛯N + 6 H2 O
In the SOHIO process, propylene, ammonia, and air are passed through a
fluidized-bed reactor containing catalyst at 400–510 ∘ C and 1.5–3 bars. Acetonitrile
and hydrogen cyanide are byproducts that are recovered for sale [13].
4.4 Uses of Syngas

4.4.1 Syngas to Hydrogen
Growth in the hydrogen market is driven primarily by regulations pertaining to
desulfurization of fuel used in transportation and recent interest in the so-called
hydrogen economy putting the hydrogen as energy carrier. Use of syngas from steam
reforming of hydrocarbons is the most common way to produce hydrogen for refiner-
ies because of its reliable technology and its low production cost [14, 15].
As discussed in Chapter 3, the gas produced by the steam reforming reaction first
passes through a shift reactor system, where carbon monoxide converts to carbon
dioxide and hydrogen. Leaving the shift reactor, the gas contains high concentra-
tion of hydrogen, with carbon dioxide, carbon monoxide, and some unconverted
methane (plus nitrogen if this was present in feed to the steam reformer). The pres-
sure swing adsorption (PSA) system adsorbs the carbon monoxide, carbon dioxide,
and other impurities in a fixed-bed adsorber at feed (high) pressure. The impuri-
ties desorb from the bed upon “swinging” the adsorber from the feed to the tail gas
(low) pressure, and by using a high-purity hydrogen purge. The adsorbent does not
adsorb the hydrogen. Apart from the pure hydrogen product (up to 99.999 vol%), the
PSA system produces a low-pressure gas stream, the tail gas. It contains all of the
impurities present in the feed gas and some of the hydrogen used for regeneration
of the adsorbent. It is used as fuel gas for the steam reformer furnace to provide the
heat input. Figure 4.6 shows a simplified diagram of UOP PolybedTM PSA system for
purifying the hydrogen to meet downstream process requirements.
To purify the hydrogen in syngas for ammonia production, the syngas from the
shift reactor is first washed with a solvent such as monoethanolamide or hot potas-
sium to remove most of the carbon dioxide. Residual carbon dioxide is removed by
catalytic methanation, which is conducted over a nickel catalyst. Exit gas from the
methanator is cooled and compressed for ammonia synthesis.
4.4.2 Syngas to Ammonia

The manufacture of ammonia is crucial for the world’s agricultural industry. The
ammonia produced is used by far for manufacture of fertilizers, including urea and
ammonium salts (ammonium nitrate, ammonium phosphates, calcium ammonium
nitrate).
Hydrogen in syngas is purified, as discussed in Section 4.4.1, and then reacted
with nitrogen (usually from air) at a hydrogen to nitrogen ratio of 3, over a catalyst,
Figure 4.6 Integration of

Steam High POLYBED steam reforming with PSA
fule Reformer temp PSA H2 system.
el shift plant
Tail
gas
References 83
to produce ammonia.
N2 + 3 H2 ↔ 2 NH3
A wide range of reaction conditions are used to produce ammonia, temperature
between 400 and 500 ∘ C and pressure between 150 and 250 bars. As the reaction is
exothermic, cool reactants (nitrogen and hydrogen) are added to maintain stable
temperatures of the reactors. On each pass, only about 15% conversion occurs, but
any unreacted gases are recycled, and eventually an overall conversion of 97% can
be achieved [16].
Improvements have been made in ammonia synthesis catalysts. Traditionally,
ammonia catalysts have been ion-based materials with promoters. Recent research
has shown a ruthenium-based catalyst is much more active to enable the process to
take place at lower pressures.
4.4.3 Syngas to Methanol

The natural gas–based technology for making syngas gives higher carbon efficiency
and lower production cost. For methanol production, the ideal syngas molar ratio
is just over 2 : 1 (H2 :CO). The steam-reforming process produces a hydrogen-rich
syngas stream where excess hydrogen is typically purged from the system in the
methanol synthesis loop and is usually used as a fuel in the reformer.
The syngas from coal has a low H2 :CO ratio for methanol synthesis, requiring a
water shift plant to increase the H2 :CO ratio, which has a negative impact on the
overall carbon efficiency.
Methanol is produced via the following exothermic reactions:
CO + 2H2 → CH3 OH
CO2 + 3H2 → CH3 OH + H2 O

The methanol synthesis reactions are carried out in a water-cooled reactor at pres-
sure of 50–100 bars and temperature of 220–260 ∘ C in presence of a copper-based
catalyst.
References
1 Schuster, C.E. (2005). ExxonMobil high-pressure process technology for

LDPE. Handbook of Petrochemicals Production Processes (ed. R.A. Meyers),
pp. 14.45–14.58. NY: McGraw-Hill.
2 Mirra, M. (2005). Polimeri Europa polyethylene high-pressure technologies.
Handbook of Petrochemicals Production Processes (ed. R.A. Meryers), pp.
14.59–14.70. NY: McGraw-Hill.
3 Smith, M. (2005). Chevron Phillips slurry-loop-reactor process for polymerizing
linear polyethylene. Handbook of Petrochemicals Production Processes (ed. R.A.
Meryers), pp. 14.31–14.44. NY: McGraw-Hill.
4 Chanda, M. and Roy, S.K. (2006). Plastics Technology Handbook, 1–6. CRC Press.
5 Allsopp, M.W. and Vianello, G. (2012). Poly(vinyl chloride). In: Ullmann’s Ency-
clopedia of Industrial Chemistry (eds. Barbara Elvers, Stephen Hawkins, and
William Russey), pp. 717–742. Weinheim: Wiley-VCH.
6 McKetta, J.J. (1984). Encyclopedia of Chemical Processing and Design, vol. 20.
New York: Marcel Dekker Inc.
7 Othmer, K. (2004). Encyclopedia of Chemical Technology, 4e, vol. 9. New York:
Marcel Dekker Inc.
8 Shell Magazine 2006. Spring; https://www.shell.com/business-customers/
catalysts-technologies/licensed-technologies/petrochemicals/ethylene-oxide-
production/omega-process.html.
9 Whiteley, K.S., Geoffrey Heggs, T., Koch, H. et al. (2005). Polyolefins. In: Ull-
mann’s Encyclopedia of Industrial Chemistry (ed. Claudia Ley), pp. 36–69.
Wiley-VCH: Weinheim.
10 Devesh, Tripathi (2002). Practical guide to polypropylene. UK: Shawbury.
https://www.worldcat.org/oclc/568032693
11 Gahleitner, M. and Paulik, C. (2014). Polypropylene. In: Ullmann’s Encyclopedia
of Industrial Chemistry, 1–44. Wiley-VCH Verlag GmbH & Co. KGaA. https://doi
.org/10.1002/14356007.021_o04.pub2. IBBN 9783527306732.
12 Nijhuis, T.A., Makkee, M.M., Moulijn, J.A., and Weckhuysen, B.M. (2006).
The production of propene oxide: catalytic processes and recent developments.
Industrial & Engineering Chemistry Research 45 (10): 3447.
13 Brazdil, J.F. Acylonitrile. In: Ullmann’s Encyclopedia of Industrial Chemistry (ed.
Claudia Ley), pp. 1–10. Weinheim: Wiley-VCH.
14 Ngoh, S.K. and Njomo, D. (2012). An overview of hydrogen gas production from
solar energy. Renewable and Sustainable Energy Reviews 16: 6782–6792.
15 Cho, W.C., Lee, D.Y., Seo, M.W. et al. (2014). Continuous operation character-
istics of chemical looping hydrogen production system. Applied Energy 113:
1667–1674.
16 Appl, M. (2006). Ammonia. In: Ullmann’s Encyclopedia of Industrial Chemistry
(ed. Claudia Ley), pp. 1–33. Weinheim: Wiley-VCH. https://doi.org/10.1002/
14356007.a02_143.pub2.
85
Section III
Aromatics
87
Aromatic Production Technologies
5.1 Introduction
Aromatics are cyclic hydrocarbons with delocalized π electrons forming a planar
ring system with alternating double and single bonds. With the resonance of the
conjugated bonds in a ring system, aromatics are stable molecules and yet highly
reactive with excellent ability for electrophilic and nucleophilic aromatic substitu-
tions. Historically, aromatics were originally found as a lighting gas in the eighteenth
century. Benzene was later discovered in 1825 by Michael Faraday by recovering
the condensed part of a lighting gas found in whale oil. Benzene in coal tar was
later found in 1845 by August Wilhelm von Hofmann. The early findings show that
coal tar possesses a variety of aromatic components that can be used as fuels. In
the 1920s, the primary sources of aromatics came from coal tar and coke-oven ben-
zole, which derive from thermal processing of coal. With excellent octane as a fuel
property, aromatics were in high demand for fuel supply during World War II. In
1940, Standard Oil Company and M.W. Kellogg Limited developed the hydroform-
ing process to upgrade virgin naphthas. Toluene was extracted to produce explosives
(trinitrotoluene). The process used molybdenum oxide on alumina as a catalyst. In
1949, UOP innovated and commercialized the first platinum-based reforming pro-
cess, the PlatformingTM process. It provided better yields and higher-octane products
for motor fuels. The catalytic reforming process also has contributed to the petro-
chemical industry as the primary source of industrial aromatics, overtaking conven-
tional coal tar and coke-oven production.
5.2 Aromatic Production

Aromatics can be produced via thermal or catalytic processes. While thermal conver-
sion processes involve higher temperature and nonselective cracking/dealkylation
making various byproducts, catalytic processes make use of selective catalysts for
on-purpose production of aromatics and tailor rearrangements of the aromatic
molecules to improve para-xylene yield. Both thermal and catalytic processes
complement each other with mature technologies and reliability to meet market
demands.
88 5 Aromatic Production Technologies
C5/C6
Isomerization
Naphtha splitter
Naphtha
feed Gasoline
NHT
pool
C7+
Reforming
Figure 5.1 Typical naphtha complex configuration.
5.2.1 Reforming Technology

One of the major technologies that contribute to the production and consumption
of BTX in the world is catalytic reforming. Historically, catalytic reforming was first
developed in 1930s by UOP to produce high-octane aviation and commercial gaso-
line replacing lead additives for octane booster. The UOP Platforming process can
produce high-octane gasoline, aromatics that are raw materials for further process-
ing downstream in petrochemicals, and valuable byproducts hydrogen and liquefied
petroleum gas (LPG). For gasoline production, naphtha-reforming product is a great
octane contributor with high concentration of aromatics directly correlated with
high-octane value. A typical reformate (C5+ liquid) includes 50–85 vol% aromatics,
10–40 vol% paraffins, less than 1 vol% olefins, and less than 1 vol% naphthenes.
For a traditional naphtha complex, a naphtha feed stream is hydrogenated in the
naphtha hydrotreating (NHT) unit and the hydrotreated naphtha is sent to a naph-
tha splitter (Figure 5.1). The light naphtha components C5/C6 paraffins are routed
to the isomerization unit, while the heavier C7 + are sent to the Platforming unit.
For fuel production and blending requirements, the benzene level in the gasoline
pool is an important parameter that impacts the naphtha complex configuration.
Benzene and benzene precursors can be routed to the isomerization unit if the mini-
mum benzene level is required in fuel specifications. On the other hand, if maximum
benzene or aromatics is desired, then benzene precursors would be sent to the Plat-
forming unit. Optimization of the C6 components can be considered when sending
these components to the isomerization to offload the Platforming unit. There are
several advantages: more capacity can be pushed through the Platforming unit with
much less C6 components resulting in increased aromatic production with reduced
benzene concentration in the reformate. Another important advantage is that the
Platforming unit can be operated at lower severity since C6 components are among
the hardest to convert.
The isomerization unit couples hand in hand with a Platforming unit to help con-
vert C5 and C6 components to higher-octane components or isomerate while reduc-
ing benzene level at the same time. This allows the refineries to recover lost octane if
benzene is removed elsewhere by other processes. If high level of Benzene is present

(∼10–20% volume) in the isomerization feed, the UOP Penex Plus® process has been
a workhorse to convert high-level benzene to isomerate product. This process has
a Bensat® reactor, where benzene is hydrogenated, followed by Penex® isomeriza-
tion reactors. Note that Euro-V gasoline requirements allow maximum content of
aromatic at 35 vol%, with 1.0 vol% max benzene.
The role of the NHT is key to protect and maintain the life of catalysts in the
isomerization and Platforming units. Many types of naphtha can be processed in
the NHT unit, including crude straight-run naphtha (SRN), fluid catalytic crack-
ing (FCC) naphtha, hydrocracker, and/or coker naphtha. While shale oil or tight oil
crudes have low contaminants, Canadian bitumen or heavier crudes possess high
levels of sulfur, nitrogen, and metals. In the NHT unit, the sulfur and nitrogen impu-
rities in the naphtha feed are converted to hydrogen sulfide (H2 S) and ammonia
(NH3 ) in the presence of the hydrotreating catalyst. Olefin and diolefin components
are also hydrogenated to paraffins. The hydrogenation reaction exothermicity is gen-
erally small for straight-run naphtha while higher with coker naphtha or naph-
tha processed from heavy vacuum residues. The operating severities can be quite
different and need to be evaluated for increased amount of coker naphtha in the
feed. The hydrotreating catalyst is typical bimetallic catalyst, including group VIII
metals such as nickel, cobalt, and molybdenum. For straight-run, low N naphtha,
UOP’s HYT-1118 catalyst is optimized to increase hydrodesulfurization (HDS) and
hydrodenitrification (HDN) activity while being more than 30% less dense than the
proven performance S-120 catalyst. For cracked naphthas with high nitrogen level
in the feed, UOP’s H-1119 is a more robust catalyst for selective sulfur and nitrogen
removal. The catalyst can also be regenerated with standard regeneration methods
to minimize catalyst changeout costs.
Some features and benefits [2] of HYT-1119:
● Improved metal dispersion
● Increased surface area (20% higher for higher silicon pickup) and stability
● Lower aromatic saturation
● Lower fill cost with reduced loaded density
● Higher HDS and HDN activity
The heart of the naphtha complex is the reforming unit where a heavy naphtha cut
is converted to a reformate product that is rich in aromatics. The feed naphtha gen-
erally contains C6 to C8 to target the production of benzene, toluene, and xylenes.
The feed is typically characterized by hydrocarbon types (paraffins, naphthenes, and
aromatics – P, N, A). The naphthene-rich feed can easily convert to aromatics due to
the ease of aromatic ring formation from already-cyclic hydrocarbons. Thus, naph-
thene dehydrogenation and dehydrocyclization are important pathways to get to the
desired aromatics requiring both metal and acid functions (Figures 5.2 and 5.3). The
metal is mainly for dehydrogenation, while the acid function is for isomerization or
ring cyclization. The reforming catalyst needs to be optimized to have the balance
of metal and acid functions to preserve yields and avoid byproducts such as light
ends from cracking or dealkylation. Paraffin hydrocarbon is the most difficult type
in the feed to convert to a ring due to competitive pathways between hydrocrack-

ing and isomerization and dehydrocyclization. While certain steps or reactions can
occur reversibly due to equilibrium, hydrocracking is not reversible. Once the light
ends or LPGs are made, they cannot recombine back to form higher chain length
hydrocarbons. The dealkylation and demethylation reactions are also not reversible
(Figure 5.4). Over the years, Platforming catalysts have been developed and opti-
mized to make maximum liquid yield (or C5+ yield) and to maintain stability even
at high operating severities.
N-Paraffins
M or A M/A
A M
Cracked Cyclopentanes Cyclohexanes Aromatics
products
M/A
M or A
Isoparaffins III
Naphthene Naphthene
isomerization dehydrogenation
I II III
I Hydrocracking and demethylation Catalyst predominant active sites

A: acid
II Paraffin isomerization
M: Metal
III Paraffin and naphthene dehydrogenation and dehydrocyclization
Figure 5.2 General platforming reaction mechanism.
Naphthene
R R Figure 5.3 Major reactions happening in the
dehydrogenation S + 3H2
reforming reactor.
C
Paraffin R C C C C R C C C
isomerization
R Rʹ
Naphthene S
isomerization
S
Rʹ
S + H2
Paraffin
R C C C C
dehydrocyclization
Rʹ
S + H2
C C
Hydrocracking R C C C + H2 RH + C C C
H
R Rʹ
Aromatics H2 Rʹʹ
+ +
dealkylation
Figure 5.4 Undesirable reactions in the reforming reactor.

5.2.1.1 Semi-regenerative Reforming

Catalytic reforming process has been a go-to process for refineries around the world
to upgrade low-octane naphtha streams to higher-octane motor fuel–blending com-
ponents. The naphtha materials can be from various sources, including straight-run
naphtha from crude units and cracked naphtha from thermal processes, coking, or
hydrocracking units. Significant milestones in catalyst development and commer-
cialization were achieved in the late 1960s with the advent of bimetallic Platforming
catalysts. UOP developed R-16, R-18, R-50, R-60, and R-80 series catalyst with opti-
mization of activity and stability over the years. With further development by adding
a proprietary promoter metal to enhance catalyst activity and selectivity, UOP intro-
duced R-98 and R-500. The most recent addition is R-560 (best in class for stability),
which is a modified base catalyst that is more robust to harsh operation conditions
and has a longer cycle length. These catalysts have been in operation and proven
to give better activity, stability, and selectivity than the all-platinum catalysts. They
enabled refiners to extend their Platforming operations to higher charge rates and
higher product octane levels with existing constraints and equipment.
As of today, more than 600 UOP fixed-bed Platforming units are onstream. For
motor fuel production, catalytic reforming can be coupled with isomerization units
to produce targeted octane grades to meet gasoline-blending specification.
5.2.1.2 Cyclic Reforming

Cyclic reforming also includes fixed beds of catalytic reactors. In addition, there is
a spare same maximum size reactor, which allows swing operation when one of
the reactors is taken out for service or regeneration. The cyclic operation is more
complex with valve switches and other safety precautions; however, the reforming
process can be maintained continuous and may allow operators to run at lower or
moderate pressures to achieve higher reformate yield and hydrogen production.
5.2.1.3 CCR Reforming

In 1971, the first continuous catalyst regeneration (CCR) Platforming unit was
started up, which revolutionized the reforming process. Extremely high severities
can be achieved without frequent shutdowns with catalyst deactivation. As of
1999, there were 156 CCR Platforming units onstream. Since the introduction
of the UOP CCR PlatformingTM process in 1971, more than 250 units have been
commissioned and in operation. With lower-pressure operation and higher severity,
more hydrogen yield can be achieved. The hydrogen co-product in the reforming
process plays a critical role for other processes in the refinery to produce other key
products such as high-quality, low-sulfur diesel that meets stringent regulations.
Over the years, UOP has developed a wide range of CCR Platforming catalysts to
meet customers’ demands on better stability and higher performance, especially on
yields, octane, or maximization of aromatics production. The evolution of various
catalysts is summarized in Figure 5.5 [2]. The R-130 series catalyst was introduced
in 1993 with high activity (lower reactor inlet temperatures), good surface area sta-
bility, and chloride retention. Low-coke R-234 catalyst came about in 2000 that gives
20–25% lower coke make and higher C5+ yield vs. R-134 catalyst while maintaining
R-134 R-234 R-264 R-254 R-284 R-334 R-384
1993 2000 2004 2010 2011 2013 2017
140+ Units 100+ Units 115+ Units 25+ Units 4 Units 15 Units 10+ Units
High stability, long life
Low Coke
High activity Highest
High yeild Highest
Figure 5.5 UOP CCR platforming catalysts in commercial operations.
Table 5.1 UOP CCR Platforming catalyst properties.
Catalyst R-234 R-254 R-334
Shape Sphere
Pill diameter 1/16 in. (1.6 mm)
Surface area 180 m2 /g
Density (ABD) 35 lb/ft3 (560 kg/m3 )
Pt level 0.29 wt%
Crush strength 50+ N
Promoter No Yes No
Source: Wier et al. [2]. © 2014, Curran Associates.
excellent surface area stability and chloride retention as the R-130 series. R-264
catalyst was introduced to markets that offer high yield, so customers can increase
their throughput. With higher density, this catalyst has increased reactor pinning
margin. With high-throughput benefit, the refiners can dial up their feed rate, and
with higher activity, maximum BTX aromatics production can be achieved. R-254
and R-284 are promoted catalysts that can offer higher yields. UOP introduced 300
series products with R-334 catalyst in 2013 and the latest R-364 catalyst in 2017 pro-
viding highest yield, long life, and low coke. These catalysts, however, do not include
a promoter and are primarily optimized for high performance by using proprietary
base and manufacturing technique (Table 5.1). The latest-generation R-364 builds
on the success of UOP’s proven R-264 catalyst with over 100 loads sold. R-364
provides great flexibility in operation for the refiner with two different modes, high
activity or high yield.
The first catalytic reforming unit was installed in 1949. The unit was the
first commercial UOP designed semi-regenerative process, which includes three
fixed-bed reactors. The use of a noble metal catalyst has enabled catalytic reforming
a workhorse to produce high-octane gasoline. Typically, the semi-regenerative
reforming unit can operate from 12 to 24 months until the catalyst loses its activity
and selectivity. Catalyst is then regenerated in situ. Thus, higher-pressure operation
generally favors cycle length, keeping the unit onstream longer.
The catalytic process has evolved over times from semi-regenerative, cyclic
reforming to continuous catalytic regeneration (https://www.uop.com/processing-
solutions/refining/gasoline/ https://www.uop.com/processing-solutions/refining/
gasoline/#naphtha-reforming).
Process Variables:
● Reactor temperature
● Reactor pressure
● Space velocity
● Hydrogen/hydrocarbon ratio (H2 /HC ratio)
● Feedstock
Reactor Temperature
This is an important variable that directly controls the severity of the process to
affect activity, overall yields, or product quality (RON). High severity is referred
to as high-temperature operation while keeping other conditions constant. At high
severity, high-octane reformate can be achieved at the expense of catalyst activity.
Typical range for reactor inlet temperature is 490–550 ∘ C. Higher-range temperature
increases the rate of catalyst deactivated with rapid coke laydown leading to more
frequent regeneration.
Reactor Pressure
A reforming reactor can typically run at 3.5–30 bar pressure. Lower pressures favor
higher reformate and H2 yield, while higher pressures increase longer life of cata-
lyst with lesser coke buildup with lower reformate yield compared to lower-pressure
operation. The lowest operating pressures are only run in reforming units with CCR.
Note that high-pressure operations tend to give lower reformate yield and more light
end byproducts. This is a direct result of undesirable hydrocracking of paraffin com-
ponents favoring light end formation rather than aromatization or ring formation.
Space Velocity
Space velocity is defined as the ratio of the volumetric feed rate to the amount of cat-
alyst. A common acronym for space velocity is liquid hourly space velocity (LHSV)
in units of h−1 . It measures how fast the feed passes the catalyst. Lower space velocity
means larger catalyst volumes (more active sites available), and severity or temper-
ature can be lower to achieve the same activity or product quality. However, higher
space velocity (by pushing more feed flow) requires higher temperature to increase
reaction rates for the same product target. With fixed design of the reactor for a given
capacity, LHSV can only be varied by changing the naphtha charge rate to the unit.
Hydrogen/Hydrocarbon Ratio (H2 /HC Ratio)

Hydrogen is required to maintain catalyst stability, prolong catalyst life by keep-
ing coke formation under control. Thus, hydrogen/hydrocarbon ratio, or H2 /HC,
is defined as the ratio of the moles of hydrogen co-fed with the moles of feed naph-
tha. The H2 /HC can be adjusted by varying the rate of the recycled gas going back to
the reactor. The hydrogen purity from the separator overhead gas also has a direct
impact on the H2 /HC. To maintain a low catalyst deactivation rate (or less severity),
higher partial pressure of hydrogen is required and thus higher H2 /HC. However,
utility consideration must be considered when recycling more recycled gas to main-
tain H2 /HC.
Feedstock
Feedstocks to a reforming unit are characterized by distillate cut points and com-
ponents, including paraffins, naphthenes, and aromatics (PNA). Feeds with high
paraffin components are harder to convert to aromatics due to slow reaction rates.
Naphthene components convert at a much faster rate and get almost consumed in
the very first reactor. Thus, rich naphthene feedstocks would require less severity to
reach a product octane target and aromatic concentration. On the other hand, lean
naphthene feedstocks (more paraffin concentration) require more severe operating
conditions to achieve the product quality target. Regarding boiling range, feedstocks
with a higher end point or a wide boiling range require more severe operating con-
ditions. If the target product is fed to an aromatic complex, the boiling range of the
feedstock to the reformer is limited to a more narrow cut range so that the desired
aromatics, such as xylenes (and less C9 aromatics), can be obtained.
5.2.1.4 High-pressure vs. Low-pressure Consideration

Without continuous catalytic regeneration, reforming units were operated at higher
pressure with high H2 partial pressure to suppress coke formation and increase
run length. However, the reaction equilibrium is limited at higher pressure, and
thus lower overall yield of C5+ product is typically observed. With UOP introduc-
tion of Continuous Catalytic Regeneration in 1971, CCR Platforming process allows
units to operate at lower pressure with higher yields of aromatics while maintain-
ing catalytic activity by continuous catalyst circulation. CCR helps Platforming units
achieve better yields approaching toward equilibrium limit, producing high quanti-
ties of aromatics with maximum octane barrels (Figure 5.6).
It is noticeable to see the reverse effect of pressure on C5+ reformate yield ranging
from 75–91 LV% (Figure 5.8). While CCR Platforming allows low-pressure operation
without having to shut down the unit for regeneration, semi-regenerative reforming
or fixed-bed reforming cannot afford to operate at high severity and such low pres-
sure. An ideal reformer is the one that can operate at highest severity and lowest
pressure as possible to maximize yields. The catalyst should have high activity for
max octane product or aromatic concentration. Besides product quality, the stabil-
ity or onstream factor needs to be high, so the units continuously operate without
the need for shutdown losing productivity. In this direction, CCR reforming > Cyclic
reforming > Semi-regenerative reforming (Figure 5.7).
CCR
platforming
C5+yield
Advantage
Fixed-bed
platforming
Pressure
Figure 5.6 Reformate yield-pressure relationship for CCR and fixed-bed platforming.
UOP CCR Platforming is the leading catalytic technology with more than 250
units licensed in the world, producing rich aromatic reformates for either refining
or petrochemical uses. UOP RZ Platforming is another technology that converts
C6 and C7 materials in the raffinate stream rejected from extractive distilla-
tion to produce additional aromatics. With RZ-100 catalyst used in a fixed-bed
reactor system, RZ Platforming offers a way to convert difficult components to
aromatics, thus reducing the naphtha feed requirements for a target aromatics
capacity.
5.2.2 Aromatic Transalkylation and Disproportionation Technology

A catalytic reforming process can produce relatively high yields of BTX aromatics.
However, the selectivity or distribution of B, T, X aromatics cannot be controlled or
manipulated easily in a reformer alone but rather depending on the feed composition
or cut points allowed to enter the unit.
Among benzene, toluene, and xylene aromatics, benzene and xylene are of
stronger demand with higher-value aromatic derivatives than toluene. Aromatic
transalkylation offers the ability to convert toluene or C9 aromatics to more xylene
product. Thus, this technology allows specific target conversion of C7 and C9
aromatics components to more desired xylene to meet the market demand while
unloading the C7 and higher aromatics traffic that can build up in an aromatic
complex. In older aromatic complexes, thermal dealkylation (THDA) process is
very common for direct dealkylation of toluene to produce more benzene. This
process can be inefficient in that the methyl group is rejected as methane, and
thus the methyl group loss cannot be recovered. However, with transalkylation
technology, the methyl group is retained and not wasted by alkylating with another
aromatic molecule to further produce xylene, which is a highly desired product.
Furthermore, the methyl group can also be introduced from an outside source such
Net gas
Fuel gas
Rx Rx Rx C S
F E
E P
LPG
H H H D
E
B
Feed H
Legend
Rx Reactor Reformate
H Heater
CFE Combined feed exchanger
SEP Separator
DEB Debutanizer
(a)
Flue gas
H2-rich gas
Swing
Reactor Reactor Reactor
reactor
Overhead
Air
Inert gas Furnace
Separator
Pretreated
naphtha
Furnace Furnace Furnace
Reformate
(b)
Figure 5.7 Typical process diagrams of commercial reforming processes, including

(a) semi-regenerative, (b) cyclic regenerative, and (c) continuous catalyst regeneration.
as methanol injection or in a dedicated process unit such as toluene methylation

process with the same purpose of increasing xylene yield.
In 1969, the UOP TatorayTM transalkylation process was commercialized from the
joint development of UOP with the inventor, Toray company [4]. UOP then became
the licensor of the process. Figure 5.9 shows the process-flow diagram for the UOP
Tatoray unit. ExxonMobil also developed a similar transalkylation process in 1976.
The reactor is a fixed-bed type reactor converting toluene or a mixture of toluene
and C9 aromatics in the presence of a catalyst and H2 at a ratio of 3–12 H2 /toluene.
Hydrogen flow is required or co-fed in the reactor to maintain stable operation pre-
venting deactivation or coke formation.
The catalyst for this reaction was developed in the form of extrudates with a
supported noble metal on alumina or zeolite, or a rare-earth catalyst [4]. Operating
Stacked
reactor
Naphtha feed
from Net H2 rich gas
Rx Net gas
treating
CCR compressor Fuel gas
regenerator Rx Recovery
section
Rx
Light ends
C S Stabilizer
Rx
F E
E P
H H H H
Fired heaters
H Aromatics
Spent catalyst
rich reformate
Legend
CFE Combined feed exchanger
SEP Separator
(c)
Figure 5.7 (Continued)
90 Theoretical yield
C5 yied, Lv%
1990s
86 1970s 1980s
1960s
1950s
82
78
86 90 94 98 102 106
RON clear
Figure 5.8 Reformate – octane relationship
conditions are reported at 350–530 ∘ C, and 20–50 bar [4]. The xylene products
obtained are in equilibrium with 23–25% p-xylene, 50–55% meta-xylene, and
23–25% ortho-xylene. Also, note that the equilibrium xylenes produced in the
Tatoray process contain low ethyl benzene (EB). This low EB concentration product
makes Tatoray process attractive and ready for isomer separation using ParexTM
unit or para-xylene crystallization unit.
For reaction mechanism, the two main pathways are toluene disproportionation
and transalkylation of toluene and C9 aromatics. With methyl group ability to
jump over from one ring molecule to another in the presence of the catalyst, the
methylated aromatic products are limited by equilibrium. Figure 5.10 illustrates the
equilibrium compositions of aromatic products for any given feed methyl/phenyl
ratio. For 100% feed, the methyl to phenyl ratio is 1, while all C9 aromatic feed would
have a 2 : 1 of methyl to phenyl ratio. The more methyl groups available in the feed,
the more chances or favorable equilibrium to proceed toward alkylated aromatic
products. However, two high methyl/to methyl ratios also lead to production of
heavier alkyl aromatics such as C10 and C11 aromatics. Thus, a typical 50 : 50
100%
Benzene
90%
80%
Composition, (wt%)
70%
Toluene
60%
50%
40%
30%
C8 aromatics
20%
C10 aromatics
10% C9 aromatics
0%
1.00 1.17 1.30 1.40 1.49 1.55 1.63 1.72 1.84 2.01 2.28
Feed methyl/phenyl ratio
Figure 5.9 Equilibrium compositions of aromatics with various feed methyl/phenyl ratios
at 700 K.
Feed surge Heater Reactor Separator Stripper

drum Purge
gas to Fuel gas
Isomar unit
Toluene from
toluene column
Cg aromatics
from Ag column
Overhead liquid
Toluene from
Parex unit
Recycle gas
Product
Figure 5.10 UOP Tatoray flow diagram.
mixture feed of toluene and C9 aromatics is a good feed to produce xylenes. Con-
version per pass is from 46% to 59% [4]. Higher conversion reduces the amount of
recycle materials through the benzene-toluene fractionation section of the complex
for equipment cost and utility consumption considerations. However, pushing
high conversion may result in benzene and xylene selectivity drop, faster catalyst
deactivation, and increase rates of ring loss or formation of heavy aromatic by
products.
Disproportionation
2 +
Toluene Benzene Xylenes
Transalkylation
+
Toluene C9 aromatics Xylenes
Transalkylation mechanism shows a conversion pathway for C9 aromatics to addi-

tional xylenes. This allows more C9 aromatic precursors to be fed to a reforming unit,
and thus the end cut point for the naphtha feed is not restricted to exclude C9 com-
ponents. Also, having C9 components together with toluene increases the methyl to
phenyl ratio that favors equilibrium formation of xylenes. For older petrochemical
complex, thermal hydrodealkylation provides a simple solution to convert toluene to
additional benzene. However, this process does not allow additional production of
xylenes, which may be more valuable than benzene. Thus, the amount of C9 compo-
nents or precursor materials is strictly limited in the original naphtha feed entering
the reforming unit and subsequent aromatic complex since circling these compo-
nents around in the complex consumes utilities and is thus generally not favored.
As of 2013, UOP has licensed a total of 60 Tatoray units with a wide range of
catalyst portfolio, including TA-4, TA-5, TA-9, TA-20, TA-30, and TA-32. The latest
generation of catalysts provides high-stability catalyst with improved performance
and lower overall ring loss to light ends or benzene co-boiler products. The feed-
stocks can vary widely from light feed 100 wt% toluene to heavy feed mixture of
30 wt% toluene and 70% C9 aromatics depending on feedstock availability upstream
of the reforming unit and overall complex capacity.
5.2.2.1 Selective Toluene Disproportionation Process

While transalkylation process can produce xylenes at equilibrium composition, with
about 25% p-xylene, selective toluene disproportionation can produce highly selec-
tive product, p-xylene, at about 90%, much higher than equilibrium composition.
UOP PX-Plus selectively disproportionates toluene to benzene and xylenes with a
selective catalyst that includes a porous structure allowing benzene and p-xylene to
escape while inhibiting the diffusion of o-xylene and m-xylene. The reaction mecha-
nism of disproportionation of toluene occurs via a biomolecular intermediate. Upon
cleaving of the intermediate to benzene and xylene, the shifting of methyl groups
can occur on the xylene molecule in the presence of the catalyst.
The process equipment for selective toluene disproportionation are similar to

those for the transalkylation process; however, the feed is more exclusively toluene.
The stoichiometry of the disproportionation reaction shows equal molar of benzene
and xylene made. Thus, the selective toluene disproportionation process is highly
suitable for producers who desire to make significant quantities of benzene together
with p-xylene.
5.2.3 Isomerization of Aromatics

Xylene mixture produced from a reforming unit typically includes equilibrium
isomers such as para-xylene, ortho-xylene, meta-xylene, and ethylbenzene. With
equilibrium composition of xylenes, it is desirable to isomerize ethylbenzene and
other xylenes to para-xylene, which is highly desired to produce polyethylene
terephthalate (PET). The isomerization process is usually accompanied with a
selective separation process, either by crystallization of selective adsorption to
remove para-xylene from the equilibrium mixture. For adsorptive separation, UOP
Parex process can separate out highly pure para-xylene product up to 99.9 wt%
purity and 97 wt% recovery. The para-xylene product is referred to as the extract,
while the rest of the xylene isomers are retained in the raffinate stream. The
raffinate stream can then be contacted with an isomerization unit, such as a UOP
IsomarTM unit, to convert to more para-xylene and the equilibrium product can be
again separated in the downstream fractionation and Parex unit. The side-by-side
Parex and Isomar units are typically referred to as Parex–Isomar loop and the
primary intention is to break the equilibrium composition of xylenes by reaction
coupling with separation. Having the separation unit (Parex) is convenient for
product removal (para-xylene) to get it out of the equilibrium system and unload
the material flows or traffic in the reaction-separation loop. Theoretically, all the
xylene isomers can be cycled to extinction and converted to all para-xylene with
such as a system like this. However, for such highly integrated aromatic complex
shown in Figure 5.11, ortho- and meta-xylene can also be produced or extracted by
boiling-point fractionation or separate adsorptive process (MX Sorbex), respectively.
5.2.3.1 Reaction Mechanism

The reactions for C8 aromatic isomerization process proceed with two main cate-
gories: ethylbenzene dealkylation and ethylbenzene isomerization, the first pathway
provides valuable byproduct benzene, while the second pathway preferably isomer-
izes ethylbenzene to xylenes. Other xylene isomers can isomerize by re-rearranging
the methyl group to an equilibrium composition the reaction condition. The pro-
cess chemistry requires both acid site and metal site for the reactions to proceed.
Xylene isomerization takes place on acid sites, while metal sites are required for aro-
matic ring hydrogenation and naphtha ring dehydrogenation. To a certain extent or
increased operating severities, ring loss may occur by ring opening followed with
cracking to light ends. Dealkylation also occurs, but this does not lead to ring loss.
The byproduct again can be valuable such as benzene or toluene that can be recov-
ered and converted in the Tatoray unit.
Raffinate
Sulfolane Benzene
PlatformingTM
B T
TatorayTM
NHT Unit
A11+
p-Xylene
Naphtha
ParexTM IsomarTM Light ends
unit unit
Figure 5.11 Typical UOP aromatic complex.
CH3 CH3 CH3

CH3
Acid Acid
Xylene
isomerization
CH3
CH3
C2H5 C 2H 5 CH3 CH3

CH3 CH3
Metal Acid Metal
I-400TM catalyst
EB isomerization
C2H5
TM
I-300 catalyst H2
EB dealkylation + C2H4 + C2H6
The Isomar flow diagram is shown in Figure 5.12. The feed usually comes from
the Parex unit or referred to as Parex raffinate with depleted para-xylene. The feed
is then combined with hydrogen-rich recycle gas and make-up gas. Small amount
of hydrogen is consumed across Isomar reactor. The combined feed is heated and
vaporized to reactor operating temperature. The reactor is a fixed-bed radial-flow
reactor where hot feed is passed radially through the catalyst bed. The reactor efflu-
ent stream is heat exchanged with the feed and sent to the product separator where
liquid and gas are recovered. A small gas stream is purged to avoid light ends buildup
in the recycle gas. The liquid effluent is stabilized in the deheptanizer where over-
head gas is taken as fuel gas to supply fuel in the complex. The overhead liquid is
usually sent back to sulfolane unit for benzene recovery. The deheptanizer bottoms
product is sent to xylene splitters for separation and eventually makes its way back
to the Parex and Isomar units.
Parex
raffinate
Fuel Gas
Charge
heater Purge
Reactor gas
Separator
To benzene
recovery
Deheptanizer
Steam
Makeup
hydrogen To xylene splitter
Figure 5.12 UOP Isomar flow diagram.
Based on the reaction mechanism highlighted for C8 aromatics, the catalysts

can be tailored to maximize xylene product vs. xylene and benzene co-product.
UOP Isomar catalysts I-350 and I-400 can be different in that I-350 can selectively
convert EB to benzene with a selectivity more than 90 mol%, while ring reten-
tion is high with more than 99 mol% of aromatic rings conserved. Also for I-350
catalyst, byproduct such as toluene can be re-converted back to xylenes with the
presence of a Tatoray unit in the complex. The EB dealkylation pathway is not
equilibrium limited and thus conversion can be as high as 65–90 wt% per pass
that can substantially unload the recycle materials in the Parex–Isomar loop. The
I-400 type catalyst is specifically designed for isomerization of xylenes, which is
equilibrium limited reaction. The conversion of EB is limited to 30 wt% per pass due
to overall xylene equilibrium. The overall para-xylene yield is slightly higher for
the I-400 catalyst as shown in Figure 5.13 due to more selective catalyst and lower
operation severity. Both catalysts have commercially shown very high aromatic ring
retention corresponding with high yields of xylene and benzene and low hydrogen
consumption.
With classical chemical engineering reaction and separation to overcome equi-
librium, UOP Isomar and Parex continue to be workhorse for the petrochemical
industry. With highly energy-efficient design for aromatic complexes, UOP intro-
duced the Isomar and Parex, which is a selective adsorption process that can deliver
para-xylene with high recovery, 97 wt% and high purity, up to 99.9 wt%. The high
recovery of the adsorption process tremendously helps debottleneck the flow traffic
and recycle from the distillation and Isomar–Parex loop. The energy consumption
or utilities are thus lowered compared to other separation process, including crys-
tallization, which is based on the difference in melting points of xylene isomers. The
eutectic composition of mixed xylene has 13 wt% of p-xylene at −52 ∘ C, which limits
the recovery of the separation, and thus can be as high as 60–70% per pass. For a
100
EB dealkylation EB isomerization
Product yield, (wt%)
90 6.4 0.2
80 86.3 87.6
70
60
50
I-350 I-400
PX Benzene
Figure 5.13 Overall Parex–Isomar yields for various isomerization catalysts.
given xylene column and isomerization unit, Parex process can deliver more than
50% more p-xylene with higher recovery of separation per pass.
5.2.4 Toluene Methylation

Toluene methylation has recently gained traction as an emerging technology for
on-purpose para-xylene production. While other technologies revolve around
re-arranging methyl group from an internal feed or mixture to increase overall
yields of xylenes, toluene methylation directly introduces a methyl group addition
from an external source to a C7 aromatic. The methyl group is introduced via a
methanol molecule, which is a flexible intermediate for petrochemical production
either for olefins or aromatics.
Many licensors in the world have developed and commercialized this technology,
including UOP, ExxonMobil, SABIC, GTC, and Sinopec. The catalyst choice is
typical MFI zeolite or modified MFI base for selective para-xylene production. The
reaction pathway is direct alkylation of toluene with methanol in the presence of
a catalyst. While the main desired reaction can selectively produce para-xylene
with 85–95% concentration in mixed xylene (>23% equilibrium composition),
methanol and toluene also participate in side reactions. This leads to a certain limit
on methanol utilization since methanol is highly reactive and can self-react to form
dimethylether (DME) or olefins such as ethylene and propylene. The methanol to
olefins can occur at the operating condition with an acid catalyst. Once the olefins
are formed, they can further alkylate with aromatic rings present in the system
to form alkylated aromatics such as ethylbenzene, propylbenzene, and heavy
aromatics. To limit the side reactions from taking place, toluene and methanol
feed ratio may be controlled at higher than 1 : 1 stoichiometric molar ratio, thereby
increasing the rate of methanol utilization toward methylation reaction. Typical
conversion of toluene per pass is less than 36%, while methanol overall conversion
per pass is less than 95%. Note that toluene can undergo subsequent methylation
to form heavier aromatics such as trimethylbenzene and tetramethylbenzene.
The operating temperatures for toluene methylation can be from 400 to 700 ∘ C
depending on the operating target and choice of reactor configurations.
Desired reaction:
CH3 CH3 H3C CH3

CH3
Acid
+ CH 3OH + +
–H2O
CH3
CH3
Side reactions:
Methanol dehydration: 2CH 3OH CH 3OCH 3 + H2O
CH3 CH3 CH3
+ CH3OH
+
Acid
Sequential methylation: 2 CH 3OH CH3 CH3
–2H2O –H2O
CH3 CH3 CH3
Transalkylation:
CH3 CH3 CH3
Acid Acid
+ +
CH3
CH3
1
R
MTO and alkylation: olefins alkylation

n CH 3OH
(ethylene, propylene, etc.) with rings
R
Toluene methylation unit can be integrated into existing aromatic complex or

new complex to increase para-xylene production capacity. For a new aromatic
complex, certain design considerations can be made to take advantage of toluene
methylation for its selective para-xylene production with reactor effluent xylene
composition, including more than 90% para-xylene, whereas typical equilibrium
xylenes only include 22–24% of para-xylene. One of the design considerations is to
maintain a high-quality toluene feed stream dedicated to the toluene methylation
unit. Separate benzene and toluene columns can be combined into a single dividing
wall column where toluene can be separated in a side draw product. Capital and
operating costs savings can be realized without compromising the overall target
and performance for the whole complex. A typical UOP design for new aromatic
complex with toluene methylation is shown in Figure 5.14. It is also desirable to
drive benzene conversion to more profitable products. Benzene can also methylate
to toluene and subsequently para-xylene. A competitive pathway is to have benzene
transalkylate with heavier aromatics to directly produce toluene and xylenes. UOP
has considered and designed different scenarios to optimize the equipment invest-
ment, profit margins, and internal rate of return for respective options. The common
benefit with toluene methylation unit is that the stream going to the Parax and
Isomar loop is concentrated in para-xylene resulting in reduction in equipment size
Sulfolane Raffinate
TatorayTM
unit Heavy aromatics
Toluene
methylation
Methanol
A11+
TM
Platforming p-Xylene
ParexTM IsomarTM Light
unit unit ends
NHT
Naphtha
Figure 5.14 Aromatic complex with toluene methylation integration.
and utility consumption. The same can be said for existing units using crystallizer
for para-xylene separation instead of adsorptive separation. For existing aromatic
complexes, more than 20% capacity of para-xylene can be achieved using toluene
methylation with little changes in the process units. Toluene methylation is thus a
viable option for on-purpose para-xylene production particularly for regions with
cheap and abundant feedstocks such as China and Middle East.
5.2.5 Aromatization of Light Naphtha/Hydrocarbons

BP-UOP Cyclar process was first developed by British Petroleum (BP) in 1975. The
purpose is to valorize LPG or light hydrocarbons to valuable petrochemical prod-
ucts. As tight oil and lighter crudes become more abundant nowadays, cheap or
stranded feedstocks are readily available for viable conversion to petrochemical base
aromatics. Leveraging UOP’s expertise in CCR system and Platforming technology,
BP turned to UOP to jointly develop the technology and a high-strength formula-
tion of catalyst that can be used in the radial-flow or stack-bed reactors as previously
designed for UOP Platforming process. The catalyst activity can be maintained and
regenerated using the UOP CCR process.
In contrast to production methods via thermal or catalytic cracking, aromatics
can be re-constructed from smaller-chain molecules (C3/C4 hydrocarbons). Thus,
LPG is expanded from conventional use as fuels to modern use of turning to
petrochemical base aromatic production. Process-flow diagram for Cyclar is shown
in Figure 5.15. The overall process and equipment are very similar to a typical
catalytic reforming unit. The process has three major sections. The reactor section,
which consists of several adiabatic, radial-flow reactors in series (reactors can also
be stacked), charge heater, and feed-effluent heat exchanger. Next to reactor section
is the CCR section, including the regenerator equipment and catalyst transfer
system. And the last section is product recovery section, which consists of product
Reactor Regeneration
section section
Tail
Hydrogen gas
Gas recovery
C
C
R
Light ends
Heater cells
Fresh and
recycle feed
Stripper
Fresh
LPG feed
C6+
aromatics
Figure 5.15 Cyclar flow diagram.
separators, the main compressor, hydrogen recovery equipment (pressure swing

adsorption – PSA, or cold box, or absorber-stripper system), and product stripper.
The overall reaction is endothermic. The process stream leaving a reactor is thus
re-heated to raise to the reaction temperature between reactors. At normal operating
conditions, the rate of carbon deposition is slow and relatively small, accumulating
less than 0.02 wt% of the feed processed. This low level of carbon on spent catalyst
results in relatively mild regeneration conditions that prolong the life of the cata-
lyst. The catalyst is formulated to have high thermal stability and good resistance
to common feedstock contaminants. Proper regeneration can fully restore catalyst
activity and selectivity to the original performance as observed with fresh catalyst.
Good mechanical strength and low attrition property translates to less fines made
and less makeup cost when circulating the catalyst in the system.
5.2.5.1 Mechanism
The reaction mechanism for LPG and condensate aromatization process is driven by
dehydrocyclodimerization chemistry, which favors at 425–500 ∘ C temperature. The
reaction pathway includes the following steps:
● Dehydrogenation of paraffin to olefin – this is also the rate-limiting step to generate

reactive olefin intermediates
● Dimerization of the olefin
● Cyclization of the dimer to a naphthenic ring
● Dehydrogenation of the naphthene to an aromatic
Beside aromatics, a significant amount of valuable byproduct present is hydrogen

that is generated from the dehydrogenation step. A small amount of heavy hydro-
carbons and coke is also present a result of side reactions from olefin dimerization
and ring formation. This can be managed by coke burns and catalyst regeneration
in the CCR system.
Light
paraffins Olifins Oligomers Naphthenes Aromatics
Light ends Hydrogen
The aromatics yield increases with the carbon number in the original feedstock
(butane vs. propane). Low-pressure operation with high temperature also favors
high yield of aromatics. The liquid product contains greater than 90% BTX aromatics
depending on the feed composition. The overall aromatics yield for propane or
butane-rich feed can be 60–67%. Again, butane-rich feed would produce higher
overall aromatics yield. Hydrogen is a significant byproduct and up to 5 wt% of the
feed can be recovered as high-purity H2 with standard PSA technology
Also, with design flexibility, Cyclar unit can be configured or revamped to turn to
an olefin production unit such as propane dehydrogenation unit (PDH or UOP Ole-
flex) or repurposed to a naphtha-reforming unit. Starting from a small-chain-length
paraffin (C3–C4 in LPG), the molecule dehydrogenates to an olefin, which has more
active functional group that can then oligomerize to a dimer olefin. The ring clo-
sure can result in an aromatic molecule and at the same time produce hydrogen as
a byproduct. The aromatization of the dimer molecule is an acid-catalyzed reaction.
The commonly used catalyst is ZSM-5 as a base with a promoted metal.
5.2.6 Catalytic Cracking

Fluid catalytic cracking also produces a major amount of mono-aromatics that are
present in the gasoline cut. With the use of zeolite targeting high production of
olefins (propylene and butylenes), the gasoline cut from the main fractionation con-
tains high concentration of aromatics, which makes it a great gasoline blend due to
its high-octane value. However, with strict regulation on aromatic limits in the final
gasoline product, refiners need a way to extract the high-value aromatics out from
the FCC unit. Thus, more and more refiners are integrated with their petrochemical
counterparts to expand their aromatics production to maintain high profits, reduce
the price swing vulnerability in fuel market, and still meet all environment con-
straints at the same time.
Why is LCO a good stream or intermediate to upgrade? For component analysis in
Figure 5.16, LCO has a very high aromatics content from 75 to 85 wt% depending on
the crude source or the feed and the operating conditions in the FCC unit. However,
Gravity 963 kg/m3

aromatics 76.4 wt%
3-Ring aromatics
ene
2-Ring aromatics
hal
pht
Higher carbon #
Na
1-Ring aromatics
lefins
enes, o
naphth
r p a raffins,
ola
Non-p
Figure 5.16 LCO characterization showing rich concentration of multiring aromatics.
high aromatic content (usually means good octane) makes LCO a low cetane value
stream, with typical cetane from 15 to 25. In the US, significant amounts of LCO can
be blended to the diesel pool in up to 20% with similar boiling range with diesel – 95%
cut point at 360 ∘ C. LCO also has high sulfur and nitrogen content, up to 1.5 wt%
and 750 wt ppm, respectively. Note that major aromatic components are multiring
aromatics, 2-ring, and 3-ring components, shown by quantitative two-dimensional
GC × GC analysis. For aromatics production, these multiring components need to
be selectively converted or ring opened to single-ring aromatics. This is a suitable
strategy for the current trend of highly integrated refinery and petrochemical com-
plexes that can operate their FCC to maximize not only propylene but also aromatics
to maximize profits.
LCO upgrading mainly includes hydroprocessing for contaminant removals,
cetane improvement, or cut-point adjustments to meet fuel specifications.
Hydrotreating
(diesel, fuel oil)
LCO
Full conversion hydrocracking
(full range naphtha)
LCO UnicrackingTM process
(diesel, gasoline)
LCO-X process
(BTX, Lt. naphtha, LPG)
For a typical 60 000-barrel-per-day (bpd) FCC feed capacity, LCO yield can
be obtained at 14 000 bpd and sent to LCO-X process, which can produce
215 000 tons/year of mixed xylenes and 80 000 tons/year of benzene as petrochemical
LCO
feed
HT Light
LPG naphtha
HC
Aromatics Benzene
Off
maximization
gas
Mixed
xylenes
ULSD
blend
stock
Figure 5.17 UOP LCO-X process for maximization of aromatics.
products. Other byproducts include fuel gas, LPG, light naphtha, and ultra-low
sulfur diesel (ULSD). The amount of xylenes produced can easily be picked up by the
demand growth of para-xylene. For economy-of-scale considerations, LCO inter-
mediates from other FCCs can be routed together to a larger LCO-X unit to produce
more xylenes to gain healthy margin profits and quick return on investment.
60 KBPD 14 KBPD 215 KMTA

FCC feed LCO LCO-X xylenes
FCCU
process 80 KMTA
benzene
A general process diagram for UOP LCO-X process is shown in Figure 5.17. The
reactor includes a multizone hydrotreating combined with hydrocracking catalysts.
The reactor effluent is sent to a hot high-pressure separator where recycle gas is
taken off in the overhead and returned to the reactor and the hot liquid is sent to
a low-pressure separator. Offgas is separated out and the separator liquid is routed
to fractionation section for product recovery. Several products can be obtained from
LCO-X process, including fuel gas, LPG, light naphtha, aromatics, and ultra-low
sulfur diesel.
From process chemistry point of view, the catalyst is designed to convert multiring
aromatics to single-ring components. The simplified reaction pathway is depicted in
Figure 5.18. Di-alkyl naphthalene is desired to cleave to 2 monoaromatic molecules.
Some saturation of the ring, whether complete or partial saturation, can occur at a
certain extent to accommodate the subsequent ring opening or dealkylation. Thus,
some naphtha range materials may be formed as intermediate products and aromatic
maximization reactor can further convert the alkyl-cyclohexanes to corresponding
aromatics (Figure 5.18).
R R
R
Aromatics
R maximization
reaction
ed
R sir n R
De ctio
rea
R Hy R
dro R R
ge
na R
tio
n
R
Naphtha range
Figure 5.18 Simplified pathway for multiring conversion in LCO-X process.
LCO hydrocarbon distribution by cuts Coker gas oil hydrocarbon distribution by cuts
% Hydrocarbon type
% Hydrocarbon type
100% 100%
3 Ring
80% aromatics 80%
2 Ring
60% 60%
aromatics
40% 1 Ring aromatics 40%
20% 20%
0% 0%
Boiling point Boiling point
Product quality
Benzene ● 90–99.9%
Mixed xylenes ● Suitable for Parex unit feed for p-xylene
● Exceeds ASTM D-5211 specification
● ∼1% ethylbenzene
Light naphtha ● 80–90% C5 –C7 paraffins, 10–20% naphthenes
● 76–82 RON, high Iso content
LPG ● <0.5% olefin
● 60% C4 , 40% C3
Diesel ● <10 ppm sulfur, Ultra-Low-Sulfur Diesel (ULSD)
● 10–15 cetane increase over LCO
Refiners can re-configure or integrate their FCC to export aromatics for petro-
chemical production. BTX can be extracted using extractive distillation processes
with commercial proprietary solvents such as Carom, Udex, or Sulfolane (developed
by UOP and Shell).
UOP licensed LCO-X technology levering UnicrackingTM technology to convert
the byproduct LCO from FCC to convert multiring aromatics to single aromatics.
Figure 5.19 Hydrocracking Diesel

integration with reforming LCO Hydroprocessing Diesel
for petrochemicals VGO
production.
Heavy CCR
naphtha Platforming
Aromatics para-Xylene
complex benzene
As future trend of gasoline decreases with the emergence of higher-efficiency

engines, hybrid and electric vehicles, refineries will adjust and re-purpose their FCC
units to convert or extract more aromatics out of their intermediate product streams.
The LCO stream is rich in aromatics, including both single and double rings. UOP
LCO-X process offers a pathway to convert double-ring aromatic to single-ring aro-
matics to maximize the aromatic yields, thus improving the value of LCO stream,
which can typically be hydrotreated and blended with diesel fuels (Figure 5.19).
For hydrocracking-based refineries, as the fuel demand decreases and petrochem-
ical demand increases in the near future, the key factor to remain profitable while
keeping process flexibility is to integrate with reforming process to produce aromat-
ics for aromatic complex. The hydrocracking unit will adjust to run at higher severity
and produce higher yields of naphtha at the expense of diesel. The naphtha in turn
is fed to the reformer to produce aromatics. The hydrocracking unit also needs to
be able to convert to naphtha with a variety of feedstocks ranging from straight-run
diesel to LCO and VGO. The hydrogen management will also be critical and balanced
prudently among various units with the main source of hydrogen coming from the
reformer beside any hydrogen from steam reforming of natural gas.
5.2.7 Production from Pyrolysis and Coking

5.2.7.1 Steam Cracking/Pyrolysis Gasoline
Steam cracking has matured over the years and is the main technology to produce
olefins in the world. With the simple design, high reliability, and flexibility to pro-
cess a wide range of feed, steam cracking continues to be chosen for high-demand,
high-capacity petrochemical complexes. Historically, with naphtha range and heav-
ier feeds, beside olefin products, steam cracking also produces pyrolysis gasoline, a
valuable aromatic-rich product.
In the United States, lighter feedstocks such as LPG or ethane are preferred feed for
steam cracking. As a result, high yield of ethylene and olefins is produced from light
feedstocks. Very low yields of higher carbon chain products are produced. Pyrolysis
gasoline yield, including benzene, toluene, and xylene, is thus very low.
With the evolution of shale gas production in the United States, more and more
steam crackers will come onstream to produce olefins from light feedstocks such as
cheap ethane recovered from shale gas.
Feed;
ethane
propane
butane
naphtha
gas oil Oil and
Cracking Cracked gas precooling C2
water Deethanizer Demethanizer C2 splitter
furnaces compression drying hydrogenation
quench
Fuel Oil Hydrogen Ethylene

C3+
Pyrolysis gasoline
C3/C4 splitter Debutanizer C4s
Propane recycle
C5+
Figure 5.20 Steam-cracking flow diagram for petrochemical production.
Steam cracking is a highly energy-intensive process. Naphtha feed enters the

coiled tubes and vaporized at the convection section of the furnace and is preheated
∘
with steam to 600 C. About half a ton of steam is added per ton of naphtha
feed. The mixed feed is then passed to the cracking tubes, which are positioned
vertically in a cracking furnace. The cracking or pyrolysis reaction temperature is
high at 800–900 ∘ C. The reactions take place in 0.1–0.5 seconds. The effluent gas
leaving the cracking furnace at 850 ∘ C is rapidly quenched to stop further reactions
or undesired secondary products. The high heat associated with the product is
recovered to produce high-pressure steam (before the quench unit), which is used to
drive the cracked gas compressors. At the quench unit, pyrolysis tar or gasoline can
be recovered in the main column and the effluent gas is further washed in a water
wash column. The cracked gas is compressed to 30–40 bar pressure, treated and
dried to remove contaminants such as CO2 , H2 S, and H2 O before entering olefin
recovery section. High purity of ethylene and propylene, 99.9% polymer grade can
be achieved using C2 and C3 splitters with 90–150 trays with high overhead reflux
rate. Ethane and propane at the bottom of the columns are recycled back to the
cracking furnaces. Methane and hydrogen can also be recovered by using cryogenic
∘
separation or low-temperature distillation (<-100 C) with refrigeration and power
recovery system using Joule–Thompson effect. Note that the deethanizer overhead,
including mainly ethylene and ethane, is contacted with a selective hydrogenation
reactor to remove acetylene from the system. Ethane is further separated and
recycled. The pyrolysis gasoline (pygas) can be recovered from the quench oil in
a decanter after the cracking reactor and from the debutanizer column bottom
(Figure 5.20).
Steam-cracking Yields
Steam crackers are a major workhorse to produce ethylene, a critical monomer for
the production polyethylene and other polymer products. Other olefins such as
propylene and butenes are byproducts. Table 5.2 summarizes typical yields from
steam cracking from different feedstocks with the basis on the same amount of
ethylene production. Pyrolysis gasoline includes high-octane aromatic components
such as benzene, toluene, and xylene. Pyrolysis gasoline yields increase with heavier
hydrocarbon feeds. However, due to the nature of thermal nonselective cracking,
Table 5.2 Typical steam cracker yields for different feedstocks.
Product Feedstock
Ethane Propane Butane Naphtha Gas oil
Ethylene 0.79 0.43 0.40 0.34 0.31

Propylene 0.03 0.17 0.22 0.16 0.19
Crude C4s 0.03 0.04 0.10 0.10 0.11
Pygas 0.02 0.07 0.07 0.20 0.19
Gas oil (6 oil) 0.00 0.01 0.01 0.03 0.11
Hydrogen 0.02 0.01 0.01 0.01 0.01
Fuel 0.27 0.11 0.19 0.15 0.08
benzene concentration tends to be higher than toluene and xylene concentration in

pyrolysis gasoline. Thus, if xylene is the preferred product or component extract, a
more selective production method such as catalytic reforming of naphtha should be
used. With shale gas evolution in the US, lighter feedstocks become much cheaper
and petrochemical producers tend to reserve more expansive naphtha feedstocks
for production of aromatics, while lighter feedstocks are for production of olefins.
Product yields from a steam cracker can vary widely depending on the type of
feeds. Typical yields for steam crackers are summarized from Platts Analytics [5] in
Table 5.2. Ethane is particularly good and selective for ethylene production. With
cheap and abundant feedstock, ethane steam cracking is the most efficient and
economical way to produce ethylene for polymer and petrochemical intermediates.
However, the yields of other byproducts are reasonably low when ethane is the feed.
The higher yields of byproducts such as gas pyrolysis gasoline (pygas) increased
with increased carbon chain length in the feedstock. However, the pygas yield tops
out at 20% when naphtha is fed to the steam cracker. Thus, steam crackers are not
considered as the base aromatic production engine, whereas naphtha reformers
continue to be the main driver for BTX aromatics.
5.2.7.2 Coke-oven Benzole

The industrial aromatics started in early 1920s with tar coal and coke-oven benzole
being the main sources for aromatics production. Coal tar has a wide range of aro-
matic components, including benzene, toluene, naphthalene, anthracene, pyrene,
styrene, and indene. Other valuable compounds that can be recovered from coal
consist of phenols, anilines, pyridines, and quinoline.
Heating or pyrolysis of hard coal can produce a mixture of rich aromatics
consisting of high concentration of benzene and toluene. Coal pyrolysis process
occurs without air and product effluents may vary with different degrees of heating.
At low heating temperatures, less than 150o C, light components such as carbon
dioxide, water, and C2 to C4 hydrocarbons are emitted. Above 180 ∘ C, higher
molecular weights such as aromatics start to appear. At 350–550 ∘ C, degasification
Table 5.3 Composition of a typical

coke-oven benzole.
Composition %/ppm
Benzene 67%
Toluene 16%
pX, mX, oX, EB 6%
Styrene 1.3%
C9+ aromatics 7%
Nonaromatics 2%
rate increases to produce semicoke. Higher temperatures, 600–800 ∘ C, lead to

further coke formation accompanied by methane and hydrogen evolution. The
volatile aromatic product from coal pyrolysis is referred to as coke-oven benzole.
However, the yield of xylene compared to benzene and toluene is very low. Because
of high-temperature carbonization of coal, benzene produced is much higher than
the catalytic reforming and pyrolysis gasoline. The typical composition obtained
from coke-oven benzole is shown in Table 5.3.
Industrial scale of hard coal pyrolysis or carbonization is operated at very high
temperatures of 1000–1200 ∘ C [1]. The main product is coke. Typical yield from one
ton of coal is 750 kg of coke, 35 kg crude tar, 11 kg benzole, 2.4 kg ammonia, 150 kg
water, and 370 m3 coke-oven gas. In fact, coal tar and benzole are just byproducts
from coal pyrolysis with very low overall yields, total less than 5 wt%. As the demand
for industrial aromatics ramped up, pyrolysis of petroleum and other production
methods such as naphtha reforming became viable to supply the market with cheap
abundant feedstocks.
Figure 5.21 shows the coal pyrolysis process where coal is heated up in the coke
ovens at high temperature (1000–1200o C). The main product includes coke, but
many valuable co-products can be obtained, including tar, ammonium sulfate,
and coke-oven benzol. The crude benzole is extracted from the coke-oven gas by
counter-current absorption in a column using a sponge oil that selectively absorbs
benzol. The sponge oil is a heavy tar-oil fraction with boiling range 220–330o C
and can be recovered in a stripper column and returned to the absorption tower.
The coke-oven benzol stream can be further treated in a hydroprocessing unit and
extractive distillation in the refinery for BTX aromatics production.
Figure 5.22 shows the relative distribution of benzene, toluene, and xylene using
different technologies. While catalytic reforming produces lowest yield of benzene,
coke-oven benzole has highest yield for benzene. Catalytic reforming selectively pro-
duces highest xylene compared to steam cracking or coal pyrolysis processes. Thus,
catalytic reforming is the most efficient production method of xylenes. In fact, cat-
alytic reforming is highly integrated with aromatic complex for extraction or further
conversion to maximize para-xylene, which is a building block for PET.
Clean
coke oven gas
H2SO4
Coke oven gas Gas

Ammonium Benzol
condensation
sulfate unit recovery unit
unit
Coal Flushing
Coke ovens Liquor Ammonium
Benzol
sulfate
distillation
Coke Drag liquor

Coke oven benzol
Tar distillation
plant
Tar product
Figure 5.21 Coke-oven flow diagram with coke-oven benzol recovery unit.
80
70
60
Relative yield (%)
50
Benzene Toluene
40
Xylenes
30
20
10
0
Reforming Pyrolysis Benzole
Figure 5.22 Relative yields of aromatics for various technologies.
5.2.7.3 Biomass Pyrolysis

One of the most important uses of biomass is the production of wood derivatives
or chemicals such as natural rubber, cellulose and ethanol and organic chemicals
such as acetic acid or phenolic compounds. It is also well known that biodiesel and
ethanol products can be produced from renewable resources and these can be used
in fuel blend applications. From woody biomass, thermal pyrolysis or torrefaction
can yield a wide range of organic compounds that can be similarly obtained from
coal pyrolysis. The liquid product yield obtained from a fast pyrolysis can be as high
By-product
Biomass gas
Condenser
train
Feed
Reheater
handling
Reactor
Bio-oil
Blower Exhaust
gases
Blower
By-product
ash
Figure 5.23 RTPTM flow diagram. Source: Based on RTPTM [8].
Table 5.4 RTP green fuel typical properties.
Property Typical values
Gross calorific value, MJ/kg 16–19

Net calorific value, MJ/kg 15–17
Water content, wt% 15–30
Solids content, wt% 0.5–2.5
Ph 2–3
Specific gravity 1.1–1.3
Ash content, wt% 0.1–0.25
Flash point, ∘ C 45–99
Pour point, ∘ C −23 to −30
Table 5.5 Comparative fuel properties for various fuels.
Fuel MJ (l) Btu (US Gallon)
Heavy fuel oil (No. 6) 40.4 144 800

Light fuel oil (No. 2) 39.3 141 000
Ethanol 23.5 84 000
RTP green fuel 19.9 71 500
as 70–75 wt%. Other products include char and light gas. The liquid product con-
tains high concentration of oxygenated hydrocarbons. Nonoxygenated hydrocarbon
products or BTX aromatics are also present but in much small quantities and can be
very difficult to recover. The liquid product obtained from biomass pyrolysis can be
referred to as pyrolysis oil, bio-oil, or bio-crude.
A UOP joint venture commercial pyrolysis process is shown in Figure 5.23. This
is a rapid thermal process referred as UOP RTPTM leveraging experiences from UOP
fluid catalytic cracking (FCC) technology. Biomass, in the absence of oxygen, is
pyrolyzed in a fluidized reactor at 500 ∘ C and quickly quenched to yield 65–75 wt%
RTP green fuel (Table 5.4). Utility is minimized by circulating hot sand inside the
fluidized reactor. Char and sand are disengaged in a cyclone and can be reheated
and cycled back to the reactor. Lighter char is lifted and disengaged in another
cyclone and discharged as byproduct ash.
The energy content per volume basis for RTP green fuel is about half of typical
fuel oil (No. 6) (Table 5.5). Thus, for a boiler or furnace requirement, about twice the
volume of RTP green oil is needed to substitute for conventional fuel oil. This can
be economical if the biomass feedstocks are readily abundant and cheap. However,
for transportation fuel market, the bio-oil needs to be further processed or refined
to lower the oxygen content while meeting the other fuel specifications required for
motor fuels.
CH 2OH CH 2OH
HC HC
HC HC
CH 2OH CH 2OH
CH 2OH CH 2OH
CH CH
CH CH CH
CH H3CO
H3CO
O CH O CH
CH 2OH CH 2OH OCH 3

OCH 3
HC O H3CO HC O
H3CO
O CH O CH
CH 2OH
OCH 3 OCH 3
OH HC O
CHOH
OCH 3
OH
Tetralignol Pentalignol
Typical woody biomass contains about 50% cellulose, hemicellulose, and lignin.
Lignin component (as high as 25%) has high aromaticity and complex struc-
tures, which include multiple linkages of three basic monolignol building-block
molecules: p-coumaryl alcohol, coniferyl alcohol, and sinapyl alcohol. These
monomers have a common phenolic backbone with alkyl- or oxygen groups
Table 5.6 Product yield summary for catalytic hydrogenation of

lignin.
Product Weight percentage of organic lignin (%)
H2 O 17.9
C1–C5 25.2
o
C6–150 C 8.3
150–240 o C 5.7
150–240 o C phenols 37.5
240–260 o C 8.7
Residue 2.4
Hydrogen consumption 5.7
attached to it. Theoretically, these monomers can be deoxygenated or dealkylated

to yield corresponding aromatics or phenols.
Catalytic upgrading of bio-oils has been recently focused on to produce
higher-value products such as aromatics or specialty chemicals. The Hydrocarbon
Research Institute had shown that catalytic hydrogenation of lignin can yield very
high yield of phenolic molecules (37.5 wt%) as shown in Table 5.6 [7]. The aromatics
and naphthene components in the neutral oils can be as high as 14 wt%. However,
the overall yield on the biomass feed is low since the typical lignin fraction in woody
biomass is about 25 wt%.
The total aromatic hydrocarbon yield is quite low with less than 1 wt%, detect-
ed and quantified by GC × GC-FID/TOF-MS for detailed composition of the bio-oil
[6]. The aromatic oxygenates concentration is higher, but these molecules are highly
oxygenated. Economic considerations need to be evaluated whether to recover the
aromatic oxygenates (phenolic branch compounds) by groups for specialty chem-
icals or to process the molecules by deoxygenating them in hydrogen atmosphere
with mild or severe operating conditions. The catalysts can be quickly deactivated by
strong adsorption of oxygenates that further repolymerize and form heavier species
that depress the catalyst activity.
Thermal pyrolysis or fast pyrolysis of biomass produces a mixture of hydrocar-
bons and oxygenates. The product from the fast pyrolysis process is also referred
to as bio-oil or bio-crude due to the nature of highly oxygenated compounds. The
low boiling components may contain high concentration of light oxygenates such as
aldehydes, alcohols, and ketones. The higher boiling point fraction includes furanic
compounds and phenolics. The phenolic fraction includes low concentration of aro-
matic hydrocarbons, including benzene, toluene, and xylene.
References 119
5.3 Summary
The aromatic production technologies continue to evolve. Catalytic reforming
provides the major production of aromatic building blocks. Industrial licensors
continue to develop catalysts with better performance and selectivity for aromatics
while minimizing byproducts such as light ends or LPG. At the same time, design
improvements, data analytics, and process intensification will help refineries utilize
their resources at the highest efficiency, increase profits/margins, and minimize
wastes and downtime.
On-purpose and secondary conversion processes continue to provide pathways to
fine-tune the distribution of products to high yields of desired products. To achieve
this, reaction and separation engineering will continue to be integrated to overcome
equilibrium limitation. Novel production methods taking advantage of in situ reac-
tive separation will emerge as a breakthrough challenging the current/traditional
technologies that have been used over the past 70 years. This will require new tech-
nologies or hybrid systems making use of new catalysts, adsorbents, and membrane
materials to improve the performance and yields while lowering costs and invest-
ment barriers.
References
1 Franck, H.-G., Stadelhofer, J.W., (1988) Industrial Aromatic Chemistry. Springer.

2 Wier, J.J., Sioui, D., Metro, S. et al. (2014). Optimizing naphtha complexes in the
tight oil boom, AFPM Annual Meeting AM-14-35 (March 2014).
3 https://www.uop.com/processing-solutions/refining/gasoline/#naphtha-reforming
4 Rase, H.F. (2000). Handbook of Commercial Catalysts: Heterogeneous Catalysts.
CRC Press.
5 Johnson, J.A., Frey, S., Thakkar, V. “Unlocking high value xylenes from light cycle
oil”, AM-07-40, Annual NPRA Meeting 2007.
6 Negahdar, L., Gonzalez-Quiroga, A., Otyuskaya, D. et al. (2016). Characteriza-
tion and comparison of fast pyrolysis bio-oils from pinewood, rapeseed cake,
and wheat straw using 13 C NMR and comprehensive GC × GC. ACS Sustainable
Chemistry & Engineering 4 (9): 4974–4985.
7 Derk, T.A.H. and Huge J.P. (1983). Lignin hydrocracking process to produce
phenol and benzene, US4420644.
8 RTPTM (18 May 2015). Technology Overview, RTP Process. http://www.ensyn.com/
technology.html (accessed 23 October 2020).
121
Uses of Aromatics
6.1 Introduction
Aromatics make excellent monomers that can be crosslinked to synthesize valuable
polymers. With excellent substitution ability, high reactivity, good solvency, and
increased reactivity with substituent groups, aromatics continue to be the key
materials for polymer syntheses and development in material science. Essential
materials such as nylons, polyesters, plastics, pesticides, coatings, and resins are
primarily derived from basic aromatic building blocks, including benzene, toluene,
and xylene.
6.2 Polymer Production from Basic Aromatic Building

Blocks
Benzene
Styrene Polystyrene
SBR elastomer
Ethylbenzene
Adipic acid Nylon

caprolactam
Cyclohexane
Phenolic resin
Phenol Caprolactam
acetone Bisphenol A
Methyl methacrylate
Methyl isobutyl ketone
Cumene
Caprolactam - Nylon-6 production; BPA - polycarbonate plastics

122 6 Uses of Aromatics
Table 6.1 World consumption of benzene by

major end use – 2016.
Derivatives Percentage of use
Ethylbenzene 49
Cumene 21
Cyclohexane 12
Nitrobenzene 9
Detergent alkylation 3
Chlorobenzene 2
Maleic anhydride 2
Other 2
Source: Data from IHS Markit [2].
The world consumption of benzene in 2016 was 44 206 thousand metric tons and
summarized in Table 6.1 (IHS Markit [2]). It is projected to grow at 2.6% average
annual growth rate.
6.2.1 Ethylbenzene and Styrene Production

Currently, direct or straight recovery of ethylbenzene from an aromatic mixture,
including mixed xylenes and ethylbenzene, is not economical due to high required
energy and equipment/process cost. The production of ethylbenzene is typically
carried out via Friedel-Crafts process with direct alkylation of benzene and ethylene
using AlCl3 with co-catalysts such as ethyl chloride, BF3 , FeCl3 , ZrCl4 , SnCl4 , and
H3 PO4 . The typical process performance includes 40% conversion and 98 mol%
selectivity [1]. The reactant ratio of benzene and ethylene is typically maintained
more than required stoichiometric ratio so that ethylene does not self-react and
polymerize to other products.
CH3
+ H2C CH2
Styrene and Polystyrene:

Next to ethylene and vinyl chloride, styrene is the most important monomer
building block in the production of plastic materials and synthetic rubber such
as styrene-butadiene (SB) copolymer and polystyrene-divinylbenzene. Styrene
production is almost exclusively via catalytic dehydrogenation of ethylbenzene.
Similarly, divinylbenzene can be produced from diethylbenzene by dehydrogena-
tion at a high temperature of 600 ∘ C with superheated steam. The main application
of divinylbenzene is the manufacture of ion-exchange resins by co-polymerization
with styrene.
CH3 CH2
+ H2
Polymer from Styrene:

Polystyrene:
CH2 CH
Cross-linked polystyrene-divinylbenzene
CH2 CH
CH2
CH2 CH
CH2
CH2
n
Nylon 6, Nylon 6,6, PET, Polyurethane, Polystyrene, Polycarbonate…
6.2.2 Cyclohexane to Nylon 6, Nylon 6,6

Cyclohexane is an important intermediate for nylon production. Cyclohexane can
be produced from catalytic hydrogenation of benzene. Note that the benzene purity
is very critical since trace amount of sulfur can poison the catalyst.
From cyclohexane, different derivatives can be produced to monomer molecule
for different nylons, namely nylon 6 and nylon 6,6 [1].
Cyclohexane
Adipic Acid Cyclohexanone
Hexamethylene
Caprolactam
diamine
Nylon 6,6 Nylon 6

Table 6.2 World consumption of cyclohexane in

2017.
Derivatives Percentage in use
Caprolactam 55
Adipic acid 22
Cyclohexanol/cyclohexanone 20
Other 3
Source: Data from IHS Markit [2].
The total world consumption of cyclohexane in 2017 was 5223 thousand metric
ton. The major use is for nylon synthesis. The relative demand is summarized in
Table 6.2.
nH2N–(CH2)6–NH2 + nHOOC–(CH2)4–COOH [H2N–(CH2)6–NHCO–(CH2)4–CO]n
Hexamethylenediamine Adipic Acid Nylon 6,6
H
N
–H2O –H2O
C= O nH2N–(CH2)5–COOH [H2N–(CH2)5–CO]n
Caprolactam 6-Aminocaproic acid Nylon 6
6.2.2.1 Adipic Acid Production

Adipic acid can be synthesized from cyclohexane oxidation or from benzene via phe-
nol intermediate.
O OH
Co salt
+
Cyclohexane Cyclohexanone Cyclohexanol
HNO3Cu
Ammonium vandate
HOOC–(CH2)4–COOH
Adipic Acid
Or Benzene Phenol
Production of Hexamethylenediamine
DuPont Synthesis O O
2NH3 –H2O
HOOC (CH2)4 COOH H4NOOC (CH2)4 COONH4 H4NC (CH2)4 CNH4
Adipic acid Ammonium adipate Adipamide
H2 –2H2O
H2N (CH2)6 NH2 NC (CH2)4 CN
Hexamethylenediamine Adiponitrile
Cl2 NaCN
H 2C CH CH CH2 ClH2C CH CH CH2Cl NC CH2 CH CH CH2 CN
Butadiene 1,4-Dichloro-2-butene 1,4-Dicyano-2-butene
H2
HCN
H2C CH CH CH2 H3C CH CH CH2 CN
1,4-Addition
Butadiene Ni cat. with ligands
HCN
Ni cat. with ligands
Monsanto synthesis
H2, eletrochydro– H2
NC CH CH2 + H2C CH CN NC (CH2)4 CN H2N (CH2)6 NH2
dimerization
Acrylonitrile Adiponitrile Hexamethylenediamine
6.2.2.2 Caprolactam Synthesis

There are seven major methods to synthesize caprolactam:
● Conventional
● Toray
● DuPont
● Union Carbide
● SNIA Viscosa
● DSM
● ENI
Conventional
OH
O N
H
N
NH2OH∙H2SO4 Beckmann
C O
Hydroxylamine Rearrangement
Cyclohexanone Cyclohexanone Caprolactam

oxime
NH4NO2 + NH3 + H2O HON(SO3NH4)2
HON(SO3NH4)2 + 2H2O NH2OH·H2SO4

Note that for the conventional method, the byproduct is ammonium sulfate. For
every pound of caprolactam produced, 4.4 pounds of ammonium sulfate is obtained.
The high amount comes from sequential steps: manufacture of hydroxylamine
sulfate, manufacture of the oxime, and Beckman rearrangement.
Toray
NOH·HCl
HCl
+ NOCl
Light
Nitronyl Cyclohexanone
Cyclohexanone
chloride oxime hydrochloride
2H2SO4 + N2O3 2HNOSO4 + H2O
HNOSO4 + HCl NOCl + H2SO4
DuPont
NO2 NOH
H
N
HNO3 Zn+Cr C O
Cyclohexanone Nitrocyclohexane Cyclohexanone Caprolactam

oxime
Union Carbide
O
H
O N
CH3COOH NH3
C O C O
Peracetic acid
Cyclohexanone Caprolactone Caprolactam
SNIA Viscosa
CH3 COOH COOH
NH. H2SO4
O2 H2 NOHSO4 C O
+ CO2
160 °C 170 °C in H2SO4
Toluene Benzoic acid Cyclohexane Caprolactam

carboxylic acid sulfate
DSM
A buffered solution of NH2 OH, NO3 − + 2H+ → NH3 OH+ + 2H2 O
The buffered NH3 OH+ in H2 O is reacted with cyclohexanone in toluene to pro-
duce the oxime, which undergoes catalytic Beckmann rearrangement to yield final
caprolactam product. The spent aqueous solution is recycled back to previous reac-
tion steps.
ENI
OH
O N
+ NH3 + H2O2 + HO2
Cyclohexanone Cyclohexanone
oxime
H
N
C O
Caprolactam
DuPont – DSM process eliminates completely ammonium sulfate formation.

The caprolactam is produced from a different starting material, butadiene, rather
than cyclohexane or benzene derivative.
CH2 CH2COOCH3 CH2COOCH3

HC CO, CH3OH H 2C CO, H2 H2C
HC Catalytic H2 C Co2(CO)8 H2 C
carbonylation Hydroformylation
CH2 CH3 CH2CHO
with double bond shift
CH2COOCH3
H
N –CH3OH H2C NH3, H2
C O –H2O
Ring closure H2C
Catalytic
CH2CH2NH2 reductive
Caprolactam
amination
A different approach for nylon synthesis from butadiene: DuPont-BASF

HDMA-Caprolactam 2-for-1 synthesis process
CN CH2NH2 CH2NH2
CH2
2HCN H2
+
CN CN CH2NH2
CH2
Butadiene Dinitrile Ω-Cyanopentylamine Hexamethylenediamine
(HMDA)
H2O
–NH3
H Nylon 66
N
C O
Caprolactam
Nylon 6
6.2.2.3 Properties of Nylon 6 vs. Nylon 6,6

Both nylon 6 and nylon 6,6 are compatible materials. The major differences are:
● Nylon 6,6 has higher melting point. It is preferred for some engineering thermo-
plastics applications and also temperature-resistant packing films.
● Nylon 6 is used where flexibility and oxygen barrier properties are required or
more critical.
6.2.3 Production of Cumene

Isopropylbenzene or cumene can be produced by benzene alkylation with propy-
lene in either gas phase or liquid phase. The alkylation reaction is the classic
Friedel–Craft reaction, which takes place in the presence of sulfuric acid or AlCl3 as
a catalyst at 30–40 ∘ C. The benzene/propylene reaction molar ratio is kept at about
5-to-1 to control the selectivity and keep propylene from polymerization.
CH3
AlCl3
CH3
+ H3C CH CH2
Cumene to Phenol
CH3 CH3
O OH
CH3 + O2
CH3
Oxidation
Cumene
[H+] Acid hydrolysis
OH O
+ H3C CH3
Phenol Acetone
The phenol market is globally driven by the demand for bisphenol A (BPA), which
consumed 47% of global phenol production in 2017 (Table 6.4). BPA is driven by
strong demand for polycarbonate products, which outpaced the GDP growth rate
and is expected to continue in the future. Polycarbonate is used in automotive, OEM,
construction, optical media, and appliance industries. It directly competes with glass
and acrylic resins in glazing/sheet and acrylonitrile-butadiene-styrene (ABS) appli-
cations. BPA is also driven by strong demand of epoxy resins, which are used to
make surface coatings, circuit boards, and adhesives. The global use for phenol is
summarized in Table 6.3, which is adapted from 2018 IHS Markit report [3].
Table 6.3 Uses of nylon.
Application Examples
Carpet fibers Commercial , residential carpets

Textiles Hosiery, lingerie, wearing apparel, sport wear
Tire cords Truck, automobile, airplane tires
Industrial fibers Soft luggage, backpacks, marine ropes, conveyor belts, military
webbings, airbags
Engineering Various automobile components, wire % cable coating, molded
thermoplastics components
Packaging films Flexible food packaging
Table 6.4 World consumption of phenol by end use – 2017.
Derivatives Percentage of use
Bisphenol 47
Phenol formaldehyde (PF) resin 27
Nylon 15
Alkylphenol 3
Xylenol 2
Other 6
6.2.3.1 Bisphenol A production

Bisphenol A is produced by acid-catalyzed reaction of acetone with phenol using
hydrochloric acid (HCl) or sulfonated crosslinked polystyrene resin (H+ ) at
50–90 ∘ C. The reaction requires excess phenol with a high molar ratio of 15 : 1
of phenol to acetone to control product selectivity and prevent acetone from
dimerization or trimerization reaction in the presence of acid catalyst. Bisphenol A
product is separated from the reaction mixture by crystallization.
HO
O
[H+] CH3
+ H3C CH3 HO OH
CH3
Bisphenol A
6.2.3.2 Production of Nitrobenzene

The industrial production of nitrobenzene is achieved by isothermal nitration of ben-
zene. The nitration medium consists of 40% HNO3 , 40% H2 SO4 , and 20% water. Due
to safety standpoint, stainless steel equipment is used due to high corrosive nature
of the catalyst or nitration medium. The reaction proceeds heterogeneously between
aqueous and organic phases with nitration occurring in acid phase.
NO2
H2SO4
+ HNO3 + H2O
Nitrobenzene is mainly used for the production of aniline to further pro-

duce 4,4-diphenylmethane diisocyanate (MDI) and poly(methylene dipheny-
lene isocyanate) (PMDI). Nitrobenzene is also used for the production of
m-nitrochlorobenzene or m-chloroaniline for pharmaceutical applications.
Other uses of nitrobenzene include the production of m-nitrobenzenesulfonic acid
and m-aminophenol also for pharmaceutical applications.
6.2.3.3 Production of Aniline

The most important method to produce aniline is from gas-phase reaction of
nitrobenzene using iron (Fe) and HCl catalyst that gives very high yield of 99%
aniline.
NO2 NH2
+ 9 Fe + 4 H2O + 3 Fe3O4
Another method to produce aniline is from gas-phase ammonolysis of phenol at

400 ∘ C, 200 bar using SiO2 /Al2 O3 with tungsten and vanadium.
HO
NH2
+ NH3 + H2O
6.2.3.4 Production of MDI and Poly-MDI (PMDI)
NH2
HCHO HCl
+ NH CH2 NH2
p-aminobenzylaniline
Isomerization
2COCl2 + H2N CH2 NH2
O C N CH2 N C O + 4HCl
MDI
(+ poly-MDI)
6.2.3.5 Production of Toluene Diisocyanate (TDI)

CH3 CH3 CH3
NO2 NH2
HNO3 H2
NO2 NH2
Dinitrotoluene Diaminotoluene
CH3
NCO
+ 4HCl
NCO
Toluene diisocyanate (TDI)
Table 6.5 Total US 2014 demand for TDI, pure MDI,

polymeric MDI (2603.5 million pounds).
Application Percentage-%
Rigid foam 40.7

Binders 18.5
Flexible foam slabstock 16.8
Flexible foam molded 7.5
Elastomers 4.8
Coatings 3.8
Adhesives 3.5
Others 4.4
6.2.3.6 Production of Polyurethane

Toluene diisocyanate is used to produce polyurethane by reacting with MDI or PMDI
and polyols
Toluene Methylene Poly (methylene

diisocynanate + diphenylene + diphenylene
(TDI) diisocyanate (MDI) diisocyanate) (PMDI)
+ Polyols
Polyurethane
Polyurethane or TDI, MDI/PMDI are used in a variety of applications, from fur-

niture, automobiles, building insulation to bedding, refrigeration/packaging, and
nonfoam uses. The total US 2014 market demand for TDI, MDI, and PMDI is sum-
marized in Table 6.5 [4].
6.2.3.7 Alkylbenzenes and Alkylbenzene Sulfonates

Monoalkylbenzenes with long alkyl chains (>C8 hydrocarbons) in the form of alkyl-
benzene sulfonates are the most important raw materials for the production of ionic
detergents. There are two types of alkylbenzene sulfonates: linear-alkylbenzene and
branched-alkylbenzene
CH3
H3C CH2 CH2 CH3
(CH2)4 (CH2)5
H3C CH C CH
CH3
CH3 CH3
SO3H SO3H
Linear alkylbenzene sulfonate (LAS) Branched alkylbenzene sulfonate (BAS)

6.2.3.8 Maleic Anhydride from Benzene
O
Air
+ 41/2O2 O + 2CO2 + 2H2O
350–400 °C
V2O5 O
Maleic
Glycols anhydride O
Unsaturated O
polyester resins
O
Phthalic anhydride
6.2.3.9 Dichlorodiphenyltrichloroethane (DDT) from Benzene
Cl
HNO3
+ 1/ O2
HCl + 2
–H2O + Trace DCB
Cl
–H2O
2 + CCl3-CHO Cl CH Cl
CCl3
Dichlorodiphenyltrichloroethane (DDT) was one of the most important insecti-

cides. It is no longer used due to its poor biodegradability and is further banned from
usage due to links with cancer and birth defects (Figure 6.1). In summary, Benzene
and Toluene are original feedstocks to produce intermediates and polymers which
are shown in Figure 6.1.
6.3 Resins and Chemicals Production

Plastics and chemicals are produced from xylene derivatives. The important inter-
mediates include phthalic anhydride, isophthalic acid, and terephthalic acid, which
are derived from o-xylene, m-xylene, and p-xylene via oxidation reactions [1]. The
xylene derivatives are used to manufacture important chemicals such as plasticizers,
unsaturated polyester resins, alkyd resins, and synthetic fibers. Most of the xylenes
used in the United States and Western Europe come from catalytic reformate. The
world supply and demand for p-xylene are shown in Figure 6.2.
Significant investments in xylene production have been made over the past
decades to increase overall capacity due to high demand of mixed xylenes. These
investments have been focused in Northeast Asia region with highly developed and
integrated polyester industry.
Among xylene isomers, p-xylene is the most widely used and produced product,
accounting for about 83% of total mixed xylenes’ consumption in 2017 (IHS Markit
[3]). PX is the raw material to produce purified terephthalic acid (PTA) and dimethyl
Ethylbenzene Styrene Polystyrene
Benzenesulfonic acid
Phenolic resins
Chlorobenzene Phenol
Bisphenol A
Cumene
Benzene Nylon 6,6
Cyclohexane Adipic acid Polyesters

for urethanes
Caprolactam Nylon 6
Polymeric
Nitrobenzene Aniline isocyanates Urethanes
Maleic anhydride Phthalic anhydride
Toluene Dinitrotoluenes Toluene diisocyanate Urethanes
Figure 6.1 Polymers from benzene and toluene.
80 000 100%
Demand growth CAGR = 6.0%
95%
70 000
90%
60 000
% Capacity utilization
85%
50 000
80
KMTA of pX
40 000 75%
Total capacity
30 000
Total demand 70%
% Utilization
65%
20 000
60%
10 000
55%
0 50%
04
0
00
02
06
08
20
12
14
16
18
10
9
20
20
20
20
20
20
19
19
19
19
19
20
20
20
20
20
Figure 6.2 World supply and demand balance for p-xylene.
terephthalate (DMT), which are key ingredients for polyethylene terephthalate

(PET) polymer, which is, in turn, used for the production of polyester fibers, PET
resins, and PET film. o-Xylene is about 5.5% global consumption of mixed xylenes
and used for production of phthalic anhydride. This intermediate is key ingredient
for alkyd resin or plasticizer to be used in the production of PVC products, including
pipe, flooring, cable, and packaging film. m-Xylene is only about 1% of mixed
xylenes, which is used to make isophthalic acid, a key ingredient to produce plastic
bottles (Figure 6.2).
Table 6.6 Properties of xylene isomers.
Freezing Boiling Catalytic reformate

C8 -aromatics point (∘ C) point (∘ C) content range (%)
p-Xylene 13.0 138 18–22

m-Xylene −47.9 139 40–45
o-Xylene −25.2 144 17–22
Ethylbenzene −95.0 136 14–21
CH3
H3C
H3C
CH3
CH3
CH3
The freezing points among xylene isomers are different enough that certain
processes can take advantage for separation, for example, crystallization process.
Other separation processes such as adsorptive separation take advantage of different
adsorptive properties of the isomers (Table 6.6). Once the isomers are separated,
they can be oxidized to give acid or anhydride intermediates to produce polyester
fiber, plastic, plasticizer, and resin materials (Figure 6.3). The worldwide con-
sumption of mixed xylenes in 2017 is shown in Table 6.7 with predominantly high
percentage use of p-xylene, 83% of global xylene demand (IHS Markit [3]).
Table 6.7 World consumption of mixed xylenes in 2017 (thousands of metric tons).
Isomer Capacity (thousands Percentage

consumption of metric tons) of use
p-Xylene 46 523 83
m-Xylene 598 1
o-Xylene 3 077 5
Ethylbenzene — 0
Solvents and others 6 118 11
C8-aromatice
Distillation
p-Xylene m-Xylene Ethylbenzene o-Xylene
To purification
Crystallization
or
or zeolite separation
Isomerization
p-Xylene m-Xylene To purification

Ethylbenzene or
o-Xylene (minor) Isomerization
Figure 6.3 Overall C8 -aromatics separation.
The major use of p-xylene is to produce PTA, which is then used to produce PET
polymer. The PET polymer is in turn used to make PET solid-state resins and PET
films. The overall consumption for mixed xylenes is strongly tied with the global
polyester industry by the high growth of demand, consumption, and development
for polyester.
The total world producers for mixed xylenes consist of about 170 companies;
among those, the top 10 largest companies account for 45% of the market with
major region based in Asia (Figure 6.3). The top 10 producers are listed in Table 6.8.
Table 6.8 Top 10 world producers of mixed xylenes in 2017.
Average annual capacity Percentage

Ranking Producer (thousands of metric tons) of total
1 SINOPEC 7 126 9.2

2 ExxonMobil Corp. 5 575 7.2
3 CNPC 4 386 5.7
4 Reliance Industries 4 217 5.4
5 JXTG Nippon Oil & Energy 3 075 4.0
6 Total 2 597 3.3
7 SK Innovation 2 531 3.3
8 Saudi Aramco 2 047 2.6
9 CNOOC 1 943 2.5
10 Formosa Group 1 815 2.3
Others 42 320 54.5
Total 77 633 100
Oxidation reactions of Xylenes

O
CH3
O Phthalic anhydride
CH3
o-Xylene O O
CH3 COH
Isophthalic acid
CH3 COH
O
m-Xylene
O
CH3
COH
Terephthalic acid
H3C
HOC
p-Xylene O
o-Xylene oxidation to phthalic anhydride

COOH
O
CH2OH COH COOH
O
CH3 CH3 CH3 CH3
O
CH3 H 3C
o-Xylene O O
O O
O
O
O
O Phthalic anhydride
The oxidation reaction of o-xylene is in gas phase using V2 O5 catalyst and

K2 SO4 + TiO2 as modifier at 350–450 ∘ C in air. The reaction mechanism is com-
plicated with risk of explosion. The presence of byproducts requires distillation
for final purification. Note that phthalic anhydride can also be produced by direct
oxidation of naphthalene using V2 O5 , which is a strong oxidizing agent.
O
V2O5
+ 4.5 O2 O + 2CO2
O
Naphthalene Phthalic anhydride
Table 6.9 World consumption of phthalic anhydride in 2016.
Applications Capacity, thousands of metric tons Percentage of use
Plasticizers 1,931 50.3

Unsaturated polyesters 507 13.2
Alkyd resins 842 21.9
Others 561 14.6
The largest consumption component for phthalic anhydride (PA) is for the pro-
duction of phthalate plasticizers, which consumed about 50% of total use of PA in
2016 (IHS Markit [2]) (Table 6.9). China is the world’s biggest consumer of PA for
plasticizers followed by the Western Europe and the United States.
O O C2H5
C OCH2 CH C4H9
O + 2H C CH CH2 OH + H 2O
9 4
C2H5 C OCH2 CH C4H9

O
O C2H5
Phthalic anhydride 2-Ethylhexanol Di-2-ethylhexyl phthalate
or di-octyl-phthalate (DOP)
Di-2-ethylhexyl phthalate or di-octyl-phthalate (DOP) is widely used (80%) in

PVC. DOP is an increasingly important plasticizer for auto manufacturing industry.
DOP tested positive in rats for carcinogenicity and has been phased out from toy
manufacturing industry.
6.3.1 Unsaturated Polyester Resins Synthesis

Unsaturated polyester resins can be produced by reacting phthalic anhydride with
maleic anhydride and diethylene glycol. The polymer product is then cross-linked
with styrene. Unsaturated polyester resins can be used in reinforced and nonre-
inforced applications, including glass, cloth, fiber, or metal replacement for boat,
storage tanks, and automobile manufacturing.
O
O O O
HO OH
O + O +
Diethylene glycol O Maleic anhydride

Phthalic anhydride
HC CH HC CH
OOC C O O COOC O O O O O O COOC O
O H C R CH CH2 R CH2 O O
2 2 H2C R CH2 CH2 R CH2 O O H C R CH
2 2
n
*R: CH2 O CH2

6.3.2 Alkyd Resin Synthesis:

Alkyd resins are the most important vehicles for oil-based paints. However, their
growth has been static for years because more resistant coatings based on vinyl,
epoxy, and urethane resins have come into use. Also, the growth of alkyd resins
has been inhibited by the popularity of water-based or latex paints derived from
poly(vinyl acetate) and acrylic resins.
O
H2C OH H2C OH
O
O HC OH + HC O R
+
H2C O C R H2C O C R
O
O O
Phthalic anhydride Monoglyceride Diglyceride
O O O O O O
H2C CH CH2 O C O CH2 CH O C O CH2 CH O C O CH2 CH CH2
O O CH2 H2C O O
O O O O O C C O
R R C O C O R R
R R
*R: alkyl, e.g. CH3 (CH2)4 CH CH CH2 CH CH (CH2)7
6.3.3 p-Xylene Oxidation to Terephthalic Acid:
CH3 CH3
Catalyst
+ 1.5 O2 No further oxidation
–H2O
CH3 COOH
p-Xylene
p-Toluic acid
CH3 CH
Acetic acid
+ HBr
Br( from NH4Br or CoBr2)
Co and Mn Acetates
COOH COOH
CH2. COO. COOH
COOH COOH COOH

Terephthalic acid
Table 6.10 Global supply of PET polymer in 2017.
Region Percentage of distribution
Asia/Pacific 80
North America 7
Europe/Africa/Middle East 11
Latin America 2
Table 6.11 Worldwide PET consumption by end-use applications.
End-use applications Percentage of distribution
Polyester fibers 63
Polyester solid-state resin 19
Polyester film 4
PET chip 12
Other 2
For byproduct removal, metal-catalyzed hydrogenolysis is carried out followed by

p-toluic acid removal by crystallization.
COOH CH3
H2, Pd
COOH
COOH
p-Carboxybenzaldehyde p-Toluic acid
Poly(ethylene terephthalate) is produced from reacting terephthalic acid with

polyol
p-Xylene Dimethyl
terephthalate (DMT)
Ethylene glycol
Purified terephthalic Terephthalic acid
acid (PTA) (PA)
Poly(ethylene
Ethylene glycol terepthalate) (PET)
Poly(ethylene
terepthalate) (PET)
6.3.4 Poly(ethylene terephthalate), PET, and Poly(ethylene

naphthalate), PEN
PET is one of the most important and widely used commercial polyesters. Tables
6.10 and 6.11 show global supply and worldwide consumption of PET in 2017 [2].
It is an amorphous thermoplastic polymer that is solidified upon rapid cooling
and can be easily molded for various applications. PET has high strength, good
abrasion and heat resistance, and excellent dimensional stability that is good when
reinforced in glass-fiber applications. Another important polyester is poly(ethylene
naphthalate) (PEN). PEN polyester has lower oxygen permeability than PET, mak-
ing it more oxygen resistant for high-barrier performance applications (Table 6.12).
The glass transition temperature (T g ), a critical property – temperature region
where the polymer transitions from a hard, glassy material to a soft, rubbery
material, is higher for PEN (120 ∘ C) than for PET (75 ∘ C) (https://polymerdatabase
.com/polymer%20index/polyesters.html). Thus, PEN has higher tensile strength
and service temperature with reduced elongation and shrinkage due to higher glass
transition temperature (T g ). However, PEN resin is more expensive, about 3–4 times
the cost of PET (Table 6.12). To recap, xylenes are the major feedstock to produce
chemical intermediates and polymers that are shown in Figure 6.4.
Terephthalic
p-Xylene Poil(ethylene terephthalate)
acid
Poil(butylene terephthalate)
Poly(p-Xylene)
m-Xylene Isophthalic Unsaturated polyesters

acid
Xylene
Alkyd resins
Polyamide resins
Diphenyl
Polybenzimidazoles
isophthalate
Phthalix
o-Xylene Alkyd resins
anhydride
Unsaturated polyesters
Polyester polyol Urethanes
Figure 6.4 Polymers and chemical resins from xylenes.
Table 6.12 PET and PEN polymer general characteristics.
Temperature O2 permeation
Material resistance resistance Uses
PET Up to 75 ∘ C Low to medium Water, liquid, dressing, peanut butter,

soft drink
PET–PEN Up to 100 ∘ C Low to medium Jam, tomato sauces, sport drinks
copolymer
PET–PEN Up to 100 ∘ C Medium to high Cutups, carbonated drinks
blends
PEN Up to 120 ∘ C High Vegetables, meats, baby food, beer, wine
O O
H O C O CH2 CH2 OH
Poly(ethylene terephthalate) – PET
CH2 OH
O CH2
O
H C
O n
Poly(ethylene naphthalate) – PEN
6.4 Summary
Aromatic applications have been contributing to the society by improving the qual-
ity and every aspect of life. Industrial organic chemistry has evolved from the field
of aromatic chemistry giving ways to the development of material science. From
building-block BTX aromatics, many applications have been developed ranging from
pharmaceuticals, dyes, detergents, pesticides, plastics, synthetic fibers, and chemical
resins that can be used in construction and automotive industries.
Polymer science continues to develop with the fundamental chemistry using BTX
aromatics. With liquid-crystallinity characteristic, aromatic-based polymers have
been developed and continually invented to have better performance properties at
lower costs and reduced environmental impacts. The more biodegradable chemicals
or intermediates are the current challenges for development to combat wastes and
environmental footprints. The same is true for pesticides and pharmaceutical
products using aromatic-based molecules. They will be more environmentally
benign, nontoxic and can be used effectively at minimal dosage to protect our
environment and human wellbeing.
Acknowledgments
The author of Chapters 5 and 6 would like to dedicate the work to his wife, Linh
Le, and his son, Bryan Pham, for their love and support. He deeply appreciates
his wife’s tolerance and sacrifice for the family with her everyday works, no mat-
ter how big or small. She is always there with him and his son to take care of them
References 143
for the fun days and the bad days. The author, Trung Pham, is grateful and in debt
to Dr. Santi Kulprathipanja for his leadership and professional advice. With Dr. Kul-
prathipanja’s deep wealth of knowledge and expertise, he has taught a great deal
over the years, and the author feels very fortunate to have him as a colleague and
mentor. Trung thanks many colleagues at UOP, including Stanley Frey, Mary Joe
Wier, Clayton Sadler, James Johnson, Patrick Whitchurch, Patrick Silady, Mark Lap-
inski, Dave Lafyatis, Steve Philoon, Patrick Bullen, Aronson Huebner, Gavin Towler,
David Wegerer, and many others, who have discussed and contributed many ideas
to him over the years. He would also like to thank Mark Goldberg and John Simley
for reviewing these book chapters and providing valuable feedback.
References
1 Franck, H.-G., Stadelhofer, J.W. (1988). Industrial Aromatic Chemistry. Springer

2 (2017). Chemical Economics Handbook. IHS Markit.
3 (2018). Chemical Economics Handbook. IHS Markit.
4 (2014). 2014 End-Use Market Survey on the Polyurethanes Industry in the United
States, Canada and Mexico. American Chemistry Council.
145
Section IV
Industrial Separation
147
Adsorption
7.1 Principle of Adsorption

Adsorption process consists of two main pathways: adsorption and desorption.
Adsorption is the process whereby a feed gas or feed liquid components adsorb at the
surface of a microporous solid. The mixture of adsorbed feed components is called
the “adsorbate,” and the microporous solid is called the “adsorbent.” Adsorption is
dictated by the characteristics of the adsorbate and adsorbent interaction. If a certain
species A has a greater affinity for the adsorbent than another species B in the
mixture, the preferentially adsorbed species can in principle be separated from the
other molecules in the gas or liquid mixture. Desorption is the process whereby
the adsorbate is removed or recovered from adsorbent so that the adsorbent is avail-
able for reuse in the next cycle in the process. In gas-phase adsorption, the adsorbed
material is most often removed by changing the temperature and/or the pressure of
the system along with a carrier or sweeper gas. For liquid systems, a solvent must be
found that preferentially displaces the desired product species from the adsorbent.
The solvent that displaces the desired product from adsorbent is called “desorbent.”
Characteristics of desorption depend on specific interactions of desorbent with both
the adsorbent and adsorbate. It is these interactions that dictate and complicate the
chemistry of adsorption. However, because of the significant number of possible
combinations of adsorbents and desorbents, there are few reasons why any two or
more components cannot be separated by adsorption. The desorbent itself should
be easily separated from the product in another separation step, usually distillation.
This chapter will highlight the impact of adsorption technology for two industrial
applications: impurity removal and liquid bulk separation.
There are two types of adsorption, chemical adsorption (chemisorption) and
physical adsorption (physisorption). Chemisorption takes place when an adsorbed
component reacts chemically with the adsorbent. Desorption or recovery of the
adsorbed product is generally difficult or not possible. This type of adsorption is
generally used for trace impurities’ removal from liquid hydrocarbon or natural gas.
Because of the nature of the chemical interaction between adsorbent and adsorbate,
the adsorbent is not regenerable. On the other hand, physisorption does not involve
chemical reaction between adsorbate and adsorbent. With this type of adsorption,
the adsorbate is adsorbed through interaction with the adsorbent by electrostatic
148 7 Adsorption
or van der Waals forces. In liquid physisorption, the adsorbate can be desorbed by
a desorbent as mentioned earlier. In gas-phase separation, the adsorbate can be
recovered by raising the temperature, reducing the partial pressure of a component,
or in some cases, by using a vapor-phase desorbent. This type of adsorption is
generally used for bulk liquid or gas-phase separation. Liquid adsorption will be
the main subject of this chapter’s discussion. This is because bulk liquid adsorption
has a great impact in refining and petrochemical separations. To demonstrate the
impact of bulk liquid-adsorptive separation, two industrial hydrocarbon separation
technologies, the UOP ParexTM and UOP MX SorbexTM processes, will be discussed
at the end of the chapter.
7.2 Adsorbents
An adsorbent is typically a solid porous material that allows molecules of a gas
or liquid mixture to adhere to its surface, and this process is called adsorption. In
separation processes via adsorption, the role of the adsorbent is to provide a large
surface area for the adsorption of certain molecules. The strength with which these
molecules are bonded to the surface is determined by the nature of the interaction
between the molecule and the surface of the adsorbent. Adsorbents are made from
many types of materials such as zeolites, silica gel, activated alumina, activated clay,
polymeric resin, and activated carbon. General examples of adsorbent applications
are listed in Table 7.1. The selection of adsorbents is dictated by the physical and
chemical characteristics of adsorbents.
As shown in Table 7.1, zeolites are used in industrial bulk-scale separations such as
xylene separation and separation of n-paraffins from i-paraffins. Zeolites are micro-
porous crystalline aluminosilicate minerals with pores of uniform size from 3 to
10 Å, which are uniquely determined by the unit structure of the crystal. These pores
will completely exclude molecules that are larger than their diameter. In contrast,
activated carbon, activated alumina, silica gel, and activated clay do not possess an
ordered crystal structure and consequently the pores are nonuniform. The distri-
bution of the pore diameters within these adsorbent particles may be as narrow as
20–50 Å, or it may range as wide as 20 Å to several thousand Angstrom as is the case
for some activated carbons. In this case, almost all molecular species may enter the
pores. This is the reason why silica gel, activated alumina, activated clay, and acti-
vated carbons are not suitable for industrial bulk separation. Their uses are most
beneficial in the areas of polar species removal from vapor or liquid. The pore size
distribution for zeolite molecular sieve, a typical silica gel, and activated alumina are
illustrated schematically in Figure 7.1.
7.2.1 Zeolite Adsorbents

Zeolites are microporous crystalline aluminosilicates with a three-dimensional
microporous framework of [SiO4 ]4− and [AlO4 ]5− tetrahedra. Isomorphous sub-
stitution of Al3+ for Si4+ causes a negative charge of the zeolite framework, which
7.2 Adsorbents 149
Table 7.1 General uses of adsorbents.
Adsorbent Industrial applications
Zeolites Separate individual PX, MX, or ethyl benzene (EB) from C8-aromatics
Separate n-paraffins from i-paraffins and cyclic
Separate olefins from paraffins
Separate fructose from glucose in corn syrup
Separate individual N2 or O2 from air
Remove acid gas from natural gas
Remove iodide from nuclear off-gas or acetic acid
Remove water from azeotropes, or organic liquids
Purify H2
Silica gel Desiccant in packings and double glazing
Drying of gases, refrigerants, organic solvents
Dew point control of natural gas
Removal of polar impurities from hydrocarbons
Activated alumina Removal of SOx and NOx
Removal of vinyl chloride monomer (VCM) from air
Removal of odors from gases
Clean-up of nuclear off-gases
Purification of helium
Activated clays Removal of organic pigments
Refining of mineral oils
Removal of polychlorinated biphenyls (PCBs)
Removal of dienes/trienes from hydrocarbons
Polymeric resins Water purification
Recover and purify steroids, amino acids
Separate fatty acids from water and toluene
Separate fructose from glucose in corn syrup
Separate aromatics from aliphatic
Recover proteins and enzymes
Removal of colors from syrups
Removal of organics from hydrogen peroxide
Activated carbon Water treatment: portable, industrial, municipal
Sweetener decolorization
Chemical processing
Pharmaceuticals
Hydrocarbon recovery
Industrial off-gas control
150 7 Adsorption
100
3A
90
4A
80
5A
Percentage of pores
70 13X
60 Silica gel
50 Activated alumina
40
30
20
10
0
1 10 100 1000 10000
Angstroms
Figure 7.1 Distribution of pore sizes in adsorbents.
is compensated by charge balancing cations. Zeolites have uniform micropores,

with size-selective adsorption and catalytic properties. Zeolite minerals were first
described in 1756 by the Swedish mineralogist A.F. Cronstedt [1]. The zeolite
structure may be represented by the formula Mx/n [(SiO2 )y (AlO2 )x ]⋅wH2 O, where
M is a cation of valence n, y/x is the Si/Al ratio, and w is the number of water
molecules. For a given framework, it is possible to tailor the properties of the
zeolite for a desired application, by varying the Si/Al ratio and the counterion.
Examples of zeolite frameworks that have an impact in industrial separations are
LTA, MFI, and FAU. The zeolite A has an LTA framework. Both silicalite and
ZSM-5 have an MFI framework but differ greatly in the Si/Al ratio. There is no
formal convention about the Si/Al ratio of the two zeolites; however, silicalite has a
very high Si/Al ratio (basically no Al in the structure), and ZSM-5 has a Si/Al ratio
from several hundreds to 10 [2, 3]. The MFI framework has a three-dimensional
pore structure with straight pores running in the b-direction with pore openings
of 0.53 × 0.56 nm. and zigzag pores running in the a-direction with pore openings
of 0.51 × 0.55 nm. The zeolites X and Y have the same FAU framework but differ
only in the Si/Al ratio, which varies between 1 and 2.0 for zeolite X, and between
1.5 and 5.6 for zeolite Y [2, 3]. FAU has a three-dimensional pore structure with
pores running perpendicular to each other in the a-, b-, and c-directions. The pore
diameter is 0.7 nm and there is also a cavity with a diameter of 1.2 nm where the
channels intersect. Structure of zeolites A, Y, and silicalite, along with the general
characteristics of zeolites, are shown in Figure 7.2.
7.2.2 Zeolite Synthesis

Many high-quality books and monographs, as well as publications from the
International Zeolite Association, have documented the important aspects of
7.2 Adsorbents 151
Zeolite-A Zeolite-Y Silicalite
Figure 7.2 Distribution of pore sizes in adsorbents.
zeolite synthesis [4–7]. Additional information is available in the patent literature,

although the experimental detail given in patents sometimes is lacking. Still, more
formulation and synthesis details are present in the patent literature than the
open literature, which makes these patent documents a necessity when a complete
picture of this field is desired. Since there is enormous zeolite synthesis information
documented in these sources, only a brief description of zeolite synthesis will be
described here as a reference to this part of chapter.
The preparation of zeolite materials began in the late 1940s based on the pioneer-
ing work of R.M. Barrer. In 1948, Robert M. Milton at Union Carbide Corporation ini-
tiated research into zeolite production, building on the work of Barrer, and because
of their efforts, Union Carbide entered the zeolite market in 1954. Since that time,
Union Carbide, ExxonMobil, and numerous independent researchers have devel-
oped hundreds of zeolites. Despite the wide variety of zeolite microporous crystal
structures that can be synthesized, virtually all zeolite synthesis procedures share
the same basic steps as shown in Figure 7.3. That is, raw materials, which are com-
posed of SiO2 , Al2 O3 , Na2 O, and water, are introduced to form a caustic aqua-gel
solution of silica and alumina. The aqua-gel, containing the charge of nutrients nec-
essary for zeolite synthesis, is sealed in a vessel and exposed to elevated temperatures
(usually between 60 and 180 ∘ C) in a batch process. The solubility of the gel phase
is higher than the solubility of the newly formed zeolite phase and so the resulting
equilibrium between the amorphous gel phase, supersaturated aqueous phase, and
newly formed zeolite crystal phase at high temperature and autogenous pressure
provides the driving force for zeolite homogeneous nucleation. Under these same
conditions, subsequent crystal growth of formed nuclei competes with continued
nucleation until all nutrients are depleted and the batch synthesis is stopped. Due to
the variety of parameters involved in crystallization, batch times can vary from hours
to days [8]. Several different silica sources can be used for hydrothermal synthesis
such as colloidal silica, sodium silicate, or tetra ethyl ortho silicate. Aluminum can
be from sources such as aluminum hydroxide, alumina, or sodium aluminate. Other
parameters, which can be varied to produce different zeolite frameworks and crys-
tal sizes, are the silica-to-aluminum ratio, the water content, the pH, and the pres-
ence of any cations in solution. Organic templates such as triethanolamine (TEA)
and diethanolamine (DEA) can be used as structure-directing agents (SDAs). These
152 7 Adsorption
Raw materials:
(SiO2, Al2O3, Na2O, H2O)
Gel make-up
Heat
Crystallization
Solid/liqiud
separation
Figure 7.3 Zeolites synthesis process.
parameters have a strong influence on not only the identity of the framework which
is synthesized, but also on the crystal size distribution, the yield of solids, and the
crystallinity of the yielded solids.
7.2.3 Forming Zeolite Powders into Usable Shapes

Zeolites and other molecular sieves are important materials in catalysis, adsorp-
tion, and many other applications, both emerging and well established. Currently,
no more than a few dozen crystalline microporous structures are widely manufac-
tured for commercial use, in comparison to the hundreds of structures that have
been made in the laboratory. The highest-volume zeolites manufactured are among
the earliest-discovered materials: zeolite A (used extensively as ion exchangers in
powdered laundry detergents and n-paraffin separation) and zeolites X and Y (used
in xylene separation and catalytic cracking of gas oil).
Zeolites are generally in a powder form with a typical particle size ranging from
1 to 5 μm, and, in some cases, as small as a few nanometers. At this size, the
zeolites are usually too small to be used as an adsorbent or catalyst. Therefore,
forming zeolite powders into useful shapes is needed. There are several books
and monographs that outline equipment and manufacturing procedures that
are relevant to zeolitic adsorbents and catalysts [9, 10]. Many technologies are
available to industry for manufacture of zeolitic adsorbents and catalysts. The unit
operations and equipment that are used can therefore be found throughout the
catalyst and adsorbent-manufacturing industries. There are two adsorbent shapes
that industry normally produces – extrudates and beads. Pelletization and accretion
7.2 Adsorbents 153
Clay
Filter cake
Dryer exhaust
H2O
Dust recycle Gramulator
Weigh
Mixing Vent
Natural gas
Screen Calcination Screen

Dryer 1200 °F
700 °F
Extruder
Natural gas Reject
Activated project < 1.5% H2O
Figure 7.4 Adsorbent forming by extrusion process.
are the two techniques that are used to manufacture extrudate and bead adsorbents,
respectively.
7.2.3.1 Pelletization of Zeolite Powder to Extrudate Adsorbent

The pelletization of zeolite powder usually requires the addition of inorganic par-
ticulate binders to make the paste moldable and to impart the adsorbent physical
strength and attrition resistance after thermal treatment. Commonly used binders
are clays (aluminosilicates), amorphous aluminophosphate, alumina, silica, titania,
zirconia, or a combination of two or more of these components [11]. Clay binders are
popular as they make the paste plastic and at the same time impart a relatively high
mechanical strength to the extruded bodies enabling handling and post-treatment
processes such as cutting. Kaolin, attapulgite, sepiolite, and bentonite clays are com-
monly used binders for zeolites [12, 13]. Various types of organic additives are added
to the pastes to serve as thickening agents, lubricants, wetting agents, and temporary
binders to improve the plasticity of the paste. Commonly used plasticizing agents
for extrusion pastes are various types of modified water-soluble cellulose products,
polyethylene glycol, polyvinyl alcohol, and glycerine [14–16].
A method of forming the zeolite powder into an adsorbent is shown in Figure 7.4.
This includes the addition of zeolite (filter cake), clay, and water into the mixer.
The blended clay zeolite mixture is extruded through the die. The extrudates are
subsequently dried, granulated, screened, calcined, and activated to become the
final adsorbent product. The calcination temperature of a zeolite-binder adsorbent
is always lower than the thermal stability of the zeolite. The temperature and binder
are selected to achieve a mechanically strong adsorbent while keeping the surface
area and crystallinity of the zeolites intact. The thermally induced increase in
strength is probably related to the creation of new and stronger bonds both between
the binder phase and the porous zeolite powders and between the porous powders
themselves.
7.2.3.2 Accretion of Zeolite Powder to Bead Adsorbent

As in the case of pellitization, the process of accretion of zeolite powder to bead
adsorbent also requires the addition of inorganic particulate binders to make the
paste moldable and to impart the adsorbent physical strength and attrition resis-
tance after thermal treatment. The preparation of the zeolite/clay paste for accretion
154 7 Adsorption
is like that which is used in palletization. Only minor differences are expected, such
as the moisture content to suit to each system in making a strong adsorbent. The
method of forming the zeolite powder into a bead adsorbent is shown in Figure 7.5.
This includes adding zeolite (filter cake) and clay into a ribbon blender and trans-
ferring the mixture of zeolite and clay to an accretion drum with water addition.
The blended clay zeolite mixture is accreted to form various sizes of bead particles.
The beads are subsequently screened to the desired size particles as product, while
the over- and undersize beads are recycled back to the accretion drum. The bead
adsorbents are then calcined and finally activated to be the adsorbent product.
7.3 Bulk Liquid-Adsorptive Separation
Liquid adsorption includes two main events: adsorption and desorption. Adsorp-
tion of an adsorbate onto zeolite adsorbent is dictated by the characteristics of the
adsorbate–adsorbent interaction. A zeolitic adsorbent is a crystalline porous solid.
When adsorbent is immersed in a liquid mixture, the porous solid has the capability
to adsorb specific species from the liquid. At equilibrium, the composition of mate-
rial inside the pores differs from that of the liquid surrounding the porous solid. The
amount of adsorbate that can be taken up by the adsorbent (i.e. the adsorbent capac-
ity) is dependent on the nature of the adsorbate–adsorbent interaction, surface area,
and pore volume of the adsorbent.1
In general, a higher adsorbent capacity is seen with higher adsorbent selectivity,
higher adsorbent surface area, and higher pore volume. This phenomenon creates
a higher concentration profile of the more selectively adsorbed component in the
pores of the adsorbent than in the surrounding liquid. The process of adsorbing the
adsorbate into the zeolitic adsorbent is known as the adsorption step.
To desorb the adsorbate from the zeolitic adsorbent, a desorbent is added. A des-
orbent is a suitable liquid that is capable of displacing – or desorbing – the adsorbate
from the selective pores of the adsorbent. The process of recovering or desorbing the
adsorbate from the adsorbent is known as the desorption step. In a chromatographic
liquid separation process, the adsorption and desorption steps must ideally occur
simultaneously. After the desorption step, both the rejected product (product with
lower selectivity, resulting in less adsorption by adsorbent) and the extracted product
(product with higher selectivity, resulting in strong adsorption by adsorbent) contain
desorbent. In general, the desorbent is recovered by fractionation or evaporation and
recycled back into the system.
7.3.1 Modes of Liquid-Adsorptive Separation

Zeolitic liquid separation mechanisms are complex, due to the interaction between
the zeolitic adsorbent, adsorbate, and desorbent. To develop an adsorptive separation
1 Parts of 7.3 are reproduced from Kulprathipanja [14], Copyright Wiley-VCH Verlag GmbH & Co.
KGaA. Reproduced with permission.
Clay Dry cake
Ribbon blender
Blower
Dust collector
n
retio
Acc m
d r u
Water
Screen
Hardening
drum
Screen
Over
size Hardening
Beads
drum
Figure 7.5 Accretion system for bead adsorbent manufacture.

156 7 Adsorption
process, an understanding of the zeolite adsorbent, adsorbate, and desorbent inter-

action is required. This has led to the development of appropriate techniques such
as fundamental adsorption isotherms, pulse tests, and dynamic breakthrough tests.
These techniques allow scientists and engineers to select the most suitable adsorbent
and desorbent combinations, and optimize operating parameters such as temper-
ature, liquid-flow circulation and adsorbent moisture. However, isotherms, pulse
tests, and dynamic breakthrough tests do not offer enough information for scaling
up the commercial unit. The scale-up information must be obtained from the simu-
lated counter current – the Sorbex technology. Other very important considerations
must be addressed: flow distribution across the chamber diameter, plug flow within
the adsorbent bed, hydraulics, optimal configuration of adsorption, purification and
desorption zones, flow and pressure measurement and control, desorbent recovery
and energy integration, to name a few.
7.3.1.1 Adsorption Isotherms

In general, adsorption isotherms are generated using a batch experiment at a fixed
temperature and a fixed feed composition. These experiments include exposing a
known amount of adsorbent to a known concentration of adsorbate at a constant
temperature. Once equilibrium is established, the net adsorbate concentration in
the liquid is measured. This process is repeated at multiple adsorbate concentrations
and temperatures. A plot of adsorbate loading (g adsorbate/g adsorbent) vs. adsor-
bate concentration reveals the adsorption isotherm with the shape of the isotherm
determining the suitability of an adsorbent for a system [17].
7.3.1.2 Pulse-Test Procedure

The pulse test is a procedure used to search for a suitable adsorbent and desorbent
combination for a liquid-adsorptive separation. The properties of the suitable adsor-
bent, such as a type of zeolite, exchange cation, and adsorbent moisture content, are
a critical part of the study. The operating temperature and liquid-flow circulation
are also critical parameters that can be set in the pulse test. A pulse-test procedure
[18] begins with an injection of a small pulse of the feed mixture to be separated into
a desorbent stream that is flowing through a packed adsorbent column at a fixed
flow rate and temperature. The column effluent composition is then determined
as a function of time or volume of desorbent passed through the adsorbent bed by
using gas or liquid chromatography. Particularly important is the sequence and time
when each of the feed components exits the packed adsorbent column because these
characteristics describe the specific adsorbate and adsorbent interactions. By deter-
mining the interactions using the pulse test, the separation process can be optimized.
An example of results from a pulse test separating arbitrary components A and
B along with a tracer is shown in Figure 7.6. An inert tracer is selected that will not
be adsorbed by the system being studied. Each component elution time or volume,
shown in Figure 7.6, is taken at the maximum peak height or concentration of a
specific component. Furthermore, the net retention volume of each component
is measured, based on the maximum peak height or concentration of the tracer
as the zero origin. Because the net retention volume of any component is ideally
RB
Net retention volume of A = RA
RA Net retention volume of B = RB
Selectivity (B/A) = B = RB/RA
Envelope half-width tracer = Wt
Relative concentration
Envelope half-width, B = WB
Rate parameter = WB−Wt = DW
Tracer
A B
Wt WB
Volume (time)
Figure 7.6 Schematic pulse test: two components and paraffin tracer.
proportional to its distribution coefficient (its concentration in the adsorbed phase

divided by its concentration in the unadsorbed phase), the adsorbent selectivity
for the more strongly adsorbed component B over component A can be calculated
from the ratio of the net retention volumes of component B to component A. In
Figure 7.6, component B is more strongly adsorbed in the adsorbent packed bed
since it elutes after component A.
7.3.1.3 Breakthrough Procedure

The adsorbent and adsorbate interaction obtained using the pulse test takes place at a
high level of dilution. Thus, the interaction information might not fully represent the
actual interaction since the commercial feed concentration is normally much higher.
For a more representative concentration, the dynamic breakthrough test technique
is introduced. The breakthrough procedure is like the pulse-test procedure except a
large amount of high concentration feed is used. The breakthrough procedure can
be described as follows:
(1) Introduce solvent (such as desorbent) into a packed adsorbent column at a fixed
flow rate, pressure, and temperature.
(2) Introduce a feed mixture to be separated until the effluent reaches the feed com-
position.
The feed breakthrough comprises the components to be separated, the tracer,
and the desorbent.
(3) Introduce solvent (such as adsorbent) until no component in the feed mixture is
detected in the effluents.
A plot of column effluent composition as a function of time or volume of solvent

and feed reveals the adsorption and desorption behaviors of the components in the
feed mixture and illustrates whether the adsorption system is suitable for separating
the components. Particularly important is the sequence and time when each of the
feed components exits the packed adsorbent column because these characteristics
describe the specific adsorbate and adsorbent interactions. From the breakthrough
158 7 Adsorption
FRC
UOP rotary valve
1 D 13
12 PRC PRC 24
FR
Figure 7.7 Sorbex: adsorption section.
plot, zeolitic adsorbent selectivity, capacity, mass transfer rate, and desorbent
strength can be calculated.
7.3.1.4 Simulated Moving-Bed Operation: Sorbex Technology

Batch-type liquid-adsorptive separations can be carried out using the principle
of pulse test, as mentioned earlier. The technique is still practiced in the field of
small-scale separation such as in the pharmaceutical area. However, a batch-type
liquid-adsorptive separation is not applicable to refining and petrochemical since
it would require a very large volume of adsorbent and adsorbent. For this reason,
a continuous liquid-adsorptive separation technology, the UOP Sorbex process,
was developed. The Sorbex process operates on a liquid–solid countercurrent
contacting principle. The basic layout for the Sorbex process is shown in Figure 7.7.
The primary process consists of two adsorbent chambers, liquid-flow distributors,
piping, and pumps directing the circulating flow through the chambers and from
the bottom of one chamber to the top of the other. There is a bed line connecting
each distributor grid to a port on the rotary valve. The rotary valve steps the net
streams (feed and desorbent in, extract and raffinate out) around the 24 distributor
grids. A control system regulates the flows, pressures, and rotary valve movement
in the adsorption section. The adsorbent is contained in each of the 24 beds, with a
flow distributor located above and below each bed. For practical reasons, in most
cases, the adsorbent is divided between two chambers. There are some applications
whereon 8 or 12 adsorbent beds are needed, and these will utilize a single chamber.
The UOP rotary valve has been used in more than two hundred Sorbex units across
a variety of applications such as p-xylene (PX) separation, m-xylene (MX) separation,
olefin separation, and n-paraffin separation. The purpose of the rotary valve is to
move the inlet and outlet ports of the net streams (feed, desorbent, raffinate, extract)
around the 24 beds in stepwise fashion, creating a semicontinuous simulated coun-
tercurrent flow of adsorbent relative to the liquid phase. This allows simultaneous
adsorption, purification, and desorption to take place with more than an order of
magnitude efficiency over batch adsorption [19].
7.3.2 Liquid-Adsorptive Separation Processes

The degrees of freedom in liquid-phase adsorption are very large due to the number
of possible ways to modify the zeolite adsorbent characteristics. Key adsorbent vari-
ables include the zeolite framework structure, silica-to-alumina ratio, particle sizes,
chemical composition, counter exchange ion, water content, and binder. These vari-
ables can be carefully modified to selectively adsorb one component over others. The
adsorbed component can then be removed or desorbed from the adsorbent using a
suitable solvent functioning as a desorbent. To optimize separation by liquid-phase
adsorption, two opposing forces must be balanced: the adsorptive force of the adsor-
bent to the adsorbate and the desorptive force of the desorbent to the adsorbate.
Ideally, the adsorbent should have a lower selectivity for the desorbent than the
adsorbed component. However, the adsorbent should have a higher selectivity for
the desorbent than the rejected components in the mixture. This is to ensure that the
desorbent itself does not utilize most of the adsorbent capacity and, at the same time,
can effectively desorb the adsorbate from the adsorbent. There are many adsorption
mechanisms that zeolite separation processes can utilize. This depends on many
factors such as zeolite structures, zeolite pore sizes, zeolite Si/Al ratios, exchanged
cations, and the physical and chemical properties of adsorbates and desorbents. In
general, the mechanisms present in zeolite adsorption can be classified as following:
(1) Equilibrium-selective adsorption;

(2) Rate-selective adsorption;
(3) Shape-selective adsorption;
(4) Ion exchange;
(5) Reactive adsorption.
7.3.2.1 Equilibrium-Selective Adsorption

The foundation of equilibrium-selective adsorption is based on differences in the
equilibrium selectivity of the various adsorbates with the adsorbent. While all the
adsorbates have access to the adsorbent sites, the specific adsorbate is selectively
adsorbed, based on differences in the adsorbate–adsorbent interaction. This, in
turn, results in higher adsorbent selectivity for one component than the others. One
important parameter that affects the equilibrium-selective adsorption mechanism
is the interaction between the acidic sites of the zeolite and basic sites of the adsor-
bate. Specific physical properties of zeolites, such as framework structure, choice of
160 7 Adsorption
exchanged metal cations, SiO2 /Al2 O3 ratio, and water content, can be manipulated
to influence the acidity of zeolites, which affects separation performance. The
degree of separation is characterized by the zeolite selectivity and capacity, as well
as the choice of desorbent and operating conditions such as temperature, feed,
and desorbent flow rates. Therefore, to achieve a meaningful separation using the
equilibrium-selective adsorption mechanism, the following zeolite characteristics
and operating conditions are to be considered:
(1) Zeolite framework structure;
(2) Metal cation exchanged in zeolite;
(3) Zeolite SiO2 /Al2 O3 molar ratio;
(4) Moisture content in zeolite;
(5) Characteristic of desorbent;
(6) Operating temperature.
These variables are described in the following sections.
Zeolite Framework Structure

One of the most significant variables affecting zeolite adsorption properties is the
framework structure. Each framework type (e.g. FAU, LTA, MFI, MOR) has its own
unique topology, cage type (α, β), channel system (one-, two-, three-dimensional),
free apertures, preferred cation locations, preferred water adsorption sites, and
kinetic pore diameter. Some zeolite characteristics are shown in Table 7.2. More
detailed information on zeolite framework structures can be found in Breck’s
book entitled Zeolite Molecular Sieves [11]. The variety of the different frame-
work structures result in different adsorbent characteristics: acid strength, size of
molecule adsorbed, adsorption/desorption rate of different molecules, capacity, and
stability. Thus, these differences characterize the adsorbent’s selectivity to a specific
molecule and adsorbent–adsorbate interactions. Take, for example, the difference
in selectivity of BaY and Ba-Mordenite [20] to PX, MX and o-xylene (OX):
BaY selectivity: PX > OX > MX
Ba-Mordenite selectivity: PX > MX > OX
Ba-Mordenite’s selectivity to MX is higher than OX, but the opposite is true for
BaY. This reversal in selectivity is a result of differences in adsorbent framework
characteristics: mordenite has higher acid strength compared to Y zeolite. Adsorp-
tion and desorption rates of xylenes are expected to be faster in BaY compared to
Ba-Mordenite because Mordenite is a one-dimensional channel system, while Y
zeolite is a three-dimensional channel. With the reason stated, a three-dimensional
channel zeolite is the preferred mass separating agent of choice compared to one-
or two-dimensional channels for the liquid adsorption separation.
Metal Cation Exchanged in Zeolite

One of the parameters in the broad class of equilibrium-selective adsorption
mechanisms is the interaction between the acidic and basic sites of the adsorbent
and the adsorbate. Zeolites can be ion-exchanged with a variety of metal cations
Table 7.2 Typical properties of common zeolites.
Theoretical ion exchange

Channel Pore openings, Å Typical SiO2 /Al2 O3 capacity, meq/g (Na form,
Zeolite type system (hydrated form) mole ratio anhydrous)
Analcime 1D 2.6 4 4.9

Chabazite 3D 3.7 × 4.2, and 2.6 4 4.9
Clinoptilolite NK 4.0 × 5.5, 4.4 × 7.2, and 10 2.6
4.1 × 4.7
Erionite 3D 3.6 × 5.2 6 3.8
Ferrierite 2D 4.3 × 5.5, and 3.4 × 4.8 11 2.4
Phillipsite 3D 4.2 × 4.4, 2.8 × 4.8, and 4.4 4.7
3.3
Zeolite A 3D 4.2 into α-cage; 2.2 into 2 7.0
β-cage
Zeolite L 1D 7.1 6 3.8
Mordenite 2D 2.9 × 5.7 10 2.6
Zeolite omega 1D 7.5 7 3.4
Silicalite-1 3D 5.7-5.8 × 5.1-5.2 50 0.63
Zeolite X 3D 7.4 into supercage; 2.2 2.5 6.4
into β-cage
Zeolite Y 3D 7.4 into supercage; 2.2 4.8 4.4
into β-cage
Note: 1D = one-dimensional, 2D = two-dimensional, 3D = three-dimensional.

Source: Beck [11], Neuzil [18], and Broughton and Gerhold [19].
to alter their acidity and other properties. There is a strong correlation between
the total acidity of a zeolite (the sum of both Brönsted and Lewis acids) and the
ionic radius of the cation as well as the valence charge of the exchanged cation
[21–23]. Exchanged cations with lower ionic radii have higher zeolite acidity. The
correlation between zeolite acidity and ionic radius or exchanged cation valence
is measured by titration with n-butylamine and a Hammett indicator, methylred.
Zeolite acidity increases for monovalent exchanged cations from Cs < K < Na < Li
and increases from Ba < Sr < Ca < Mg and Ba < Pb < La < Ca < Zn < Co for divalent
exchanged cations. For cations of similar ionic radii, divalent cations have higher
zeolite acidity than monovalent cations. For example, Ba-zeolite has a higher acidity
than K-zeolite, and Ca-zeolite has a higher acidity than Na-zeolite and Li-zeolite.
Furthermore, strong acidic Brönsted properties can be observed in hydrated zeolites
due to polarization of adsorbed water in the strong electrostatic field between the
exchanged cations and the AlO4 – anions [23]. For hydrated zeolites, a higher
moisture content results in higher zeolite acidity.
The importance of zeolite acidic strength with zeolite Y for C8-aromatics (xylenes)
systems is illustrated below. In the presence of strong acids, xylene isomers have
varying basicity (Table 7.3), with MX being the most basic and PX the least basic
162 7 Adsorption
Table 7.3 Base strengths of aromatic hydrocarbons relative to HF.
Compound Relative basicity at 0.1 molar
Benzene 0.09
Toluene 0.63
p-Xylene 1.0
o-Xylene 1.1
m-Xylene 26
Durene (1,2,4,5-tetramethylbenzene) 140
Isodurene (1,2,3,5-tetramethylbenzene) 16 000
PX
MX
EB OX
Eluted volume
Figure 7.8 Separation of m-xylene from C8 -aromatics using NaY adsorbent.
among the C8-aromatics [24]. Based on the basicity of the xylenes, the acidity of Y
zeolite can be properly adjusted to selectively adsorb MX or PX. As demonstrated in
Figure 7.8, a more acidic zeolite such as NaY will selectively adsorb MX from other
C8-aromatics [25, 26]. In contrast, Figure 7.9 shows that a weaker acidic zeolite such
as KY will selectively adsorb PX from other C8-aromatics [27, 28]. In both systems,
toluene is used as the desorbent. However, acid-based interactions between zeolitic
adsorbents and adsorbates do not always correctly predict the trend of adsorbent
selectivity. This is illustrated by the adsorptive separation of durene from isodurene.
Pulse-test experiments indicated that the adsorbent selectivity for durene/isodurene
increases from KX < NaX < LiX, shown in Table 7.4 [29]. Because isodurene is a
stronger base than durene (Table 7.3), one would expect that the results for adsor-
bent selectivity shown in Table 7.4 would be the reverse order (LiX < NaX < KX).
This counterintuitive trend may be explained by steric hindrances caused by the dif-
ferences in ionic radii of the cations, which could play a role in the adsorbent and
adsorbate interaction.
OX
MX
EB
PX
Eluted volume
Figure 7.9 Separation of p-xylene from C8 -aromatics using KY adsorbent.
Table 7.4 Durene/isodurene selectivity as a

function of X-zeolite exchanged cation.
Cation Durene/isodurene selectivity
K 0.59
Na 0.97
Li 3.50
Zeolite SiO2 /Al2 O3 Molar Ratio

As mentioned earlier, multiple factors determine zeolite acidity, including the
SiO2 /Al2 O3 molar ratio. Zeolite acidity increases in strength as the molar ratio of
SiO2 /Al2 O3 decreases [22] due to the increase in AlO4− sites, which strengthens the
electrostatic field in the zeolite and increases the number of acid sites. However, the
wide array of cage and channel arrangements and electrochemical properties that
result from various crystalline structures and different SiO2 /Al2 O3 ratios also affect
zeolite acid strength. In certain conditions, a high density of AlO4− in the zeolite
framework could lower the acid strength of the adsorbent. The reverse in acid
strength can be explained by the dipolar repulsion of the AlO4− groups outweighing
the increase in polarizability [21, 30]. For the reasons stated, an increase in zeolite
acidity as the molar ratio of SiO2 /Al2 O3 increases is normally observed. Manipulat-
ing both the exchanged cations and the SiO2 /Al2 O3 ratio offers a great flexibility in
tailoring adsorbents for a specific application. However, the more variables that are
altered, the more difficult the adsorption behavior becomes to predict.
The adsorptive separation of durene and isodurene is used here to illustrate the
effect of zeolite acidity. As shown in Figure 7.10, LiX is more selective toward durene
than isodurene [29]. However, LiY is more selective toward isodurene than durene
(Figure 7.11). These results are consistent to what is expected since zeolite Y (Si/Al
of 2.5) is more acidic than zeolite X (Si/Al of 1.25).
164 7 Adsorption
Isodurene Durene
Eluted volume
Figure 7.10 Separation of durene and isodurene using LiX adsorbent.

Durene
Isodurene
Eluted volume
Figure 7.11 Separation of isodurene and durene using LiY adsorbent.
Figure 7.12 shows separation of fructose and glucose using adsorbents CaY, CaX,
and KX [31]. The combination of the zeolite framework structure and exchanged
cations is instrumental to this intricate mechanism as seen with the separation of
fructose and glucose. Ca2+ forms complexes with fructose and, thus, CaY shows
a high selectivity for fructose over glucose. In contrast, CaX does not exhibit high
selectivity for fructose. However, KX zeolite shows a good selectivity for glucose over
fructose. No simple explanation can be offered for this elaborate system. In the UOP
SarexTM process, Ca-Y zeolite and, later, Ca-exchanged resin were successfully used
as adsorbents that are selective for fructose.
Moisture Content in Zeolite

Adsorbed water molecules on a zeolite adsorbent are polarizable due to the strong
electrostatic field between the exchanged cations and alumina framework [23].
Hence, water molecules enhance the acidic properties of the zeolite’s Brönsted
acids. Adsorbate–adsorbent interactions and, therefore, adsorbent selectivity and
adsorbate mass transfer rates are altered due to water polarization. When devel-
oping an adsorbent to be used in a commercial adsorptive separation process, the
Figure 7.12 Fructose/glucose separation on zeolite
Concentration
G F
adsorbents. (a) Ca-Y adsorbent, (b) Ca-X adsorbent,
and (c) K-X adsorbent.
(a) Eluted volume
Concentration
GF
(b) Eluted volume
F G
Concentration
Eluted volume
(c)
water content of the adsorbent is adjusted to balance adsorbent selectivity and the
component mass transfer rate.
Characteristics of the Desorbent

In liquid-phase adsorption, some components of the feed steam are selectively
adsorbed or extracted by a solid zeolitic adsorbent. At the same time, other
components of the feed stream are rejected by the adsorbent. At equilibrium, the
liquid composition within the zeolite pores differs from that of the liquid surround-
ing the zeolite. In the process, a second liquid component, the desorbent, is also
introduced into the system. The function of the desorbent is to desorb and remove
any weakly adsorbed impurities and to recover the extracted feed components
from the adsorbent. For the desorbent to perform well in the process, a suitable
interactive force is needed between the desorbent and the components that are
extracted by the adsorbent. If the selectivity of the extracted component is too high,
it requires a high desorbent volume to desorb the extracted components from the
adsorbent. If the selectivity of the extracted component is too low, the desorbent
tends to compete with extracted components for capacity of adsorbent.
A basic guideline for choosing a desorbent is to match the chemical properties
of the extracted components with those of the desorbent. Another guideline is
an appropriate selection of boiling point differences to allow recovery from the
feed components from the desorbent after the adsorption section. The follow-
ing examples illustrate this concept: (i) for a n-paraffin separation, a n-paraffin
desorbent is preferable, (ii) for an aromatic separation, an aromatic desorbent is
preferable, (iii) for a highly polar adsorbate, a desorbent should be selected from a
class of alcohol, ketone, acid, or water. Desorbent features are illustrated in Table 7.5
166 7 Adsorption
Table 7.5 Separatio of p-xylene using KY and BaX adsorbents with various desorbents
[38, 39] by pulse tests.
PX net
retention PX/EB PX/MX PX/OX PX stage
Adsorbent Desorbent volume (ml) selectivity selectivity selectivity time (s)
BaX Phenyldecane 107 1.84 3.02 2.83 22.2

KY Phenyldecane No PX
desorption
KY Diphenylmethane 10.5 1.68 3.08 1.89 29.7
BaX Diphenylmethane 0 1 1 1
BaX 1,3-Diisopropylbenzene No PX
desorption
BaX 1,4-Diisopropylbenzene 56.4 2.28 3.45 3.34 19.8
KY 1,4-Diisopropylbenzene No PX
desorption
BaX 1,3,5-Triethylbenzene 77.6 1.3 3.0 2.32 43.9
KY 1,3,5-Triethylbenzene 61.4 2.33 7.58 5.23 40.8
BaX 5-Tertiarybutyl-m-xylene 73.2 1.94 3.28 2.87 23.4
KY 5-Tertiarybutyl-m-xylene 47.1 2.05 3.72 3.37 54.5
Source: Kulprathipanja [38] and Bland and Davidson [39].
using C8 -aromatic adsorbates with BaX and KY adsorbents. The results in Table 7.5
further emphasize the desorbent characteristic requirement mentioned above.
For instance, phenyldecane is a suitable desorbent for PX separation using BaX
adsorbent. However, phenyldecane is too weak to desorb PX from KY adsorbents.
In contrast, diphenylmethane offers good separation of PX with KY adsorbent, but
not with BaX. With BaX adsorbent, PX is separated from other C8-aromatics using
1,4-diisopropylbenzene, but not with other isomers of di-isopropylbenzene, such as
1,3-di-isopropylbenzene.
Operating Temperature
Another critical variable in liquid-phase adsorptive separation is the operating tem-
perature. Liquid-phase adsorption must be operated at a temperature that optimizes
selectivity and mass transfer rates. Generally, selectivity increases, and transfer rates
decrease at lower temperatures.
Operating Pressure
In general, operating pressure does not affect liquid-phase adsorption. However,
enough pressure must be applied to maintain the system in the liquid phase during
the entire process.
7.3.2.2 Rate-Selective Adsorption

Although most of the commercial adsorptive separation processes are operated
under the selective-equilibrium adsorption mechanism, adsorptive separation
may also be based on diffusion rates through a permeable barrier, which are
designated as “rate-selective adsorption” processes. In some instances, there may
Table 7.6 Energetics and predictions from modeling molecular diffusion in silicalite.
Energy barrier Maximum energy Diffusion

Organic molecules (kcal/mol) (kcal) prediction
n-Decane 2 −31 Fast

2-Methylnonane 8 −29 Moderate
2,6-Dimethyloctane 15 −17 Slow
3,3,5-Trimethylheptane 45 +35 Excluded
be a combination of equilibria as well as rate-selective adsorption. A rate-selective

adsorption process yields good separation when the diffusion rates of the feed
components through the permeable barrier differ by a wide margin.
Examples of rate-selective adsorption are demonstrated using silicalite adsorbent
for separation of C10 –C14 n-paraffins from non-n-paraffins [32, 33] and C10 –C14
mono-methyl-paraffins from non-n-paraffins [34–37]. Silicalite is a 10-membered
ring zeolite with a pore opening of 5.4 Å × 5.7 Å [18]. In the case of n-paraffins/
non-n-paraffins separation [32, 33], n-paraffins enter the pores of silicalite freely,
but non-n-paraffins such as aromatics, naphthenes, and iso-paraffins diffuse into
the pores more slowly. However, the diffusion rates of both normal-paraffins and
non-n-paraffins increase with temperature. So, one would expect to see minimal
separation of n-paraffins from non-n-paraffins at high temperatures but improved
separation at lower temperature.
For the rate-selective separation of C10 –C14 mono-methyl-paraffins from
non-n-paraffins [34–37], diffusion simulations were carried out using the Solids
Diffusion module in the Accelrys Insight II molecular modeling package [36]. The
modeling results from the diffusion simulations of four paraffins of varying carbon
numbers in silicalite are summarized in Table 7.6.
Note that the predicted diffusivity of the molecules, based on the magni-
tude of the energy barrier, agrees with the experimental data (see Table 7.6).
The retention volumes and hence diffusivities increase in the following order:
3,3,5-trimethylheptane < 2,6-dimethyloctane < 2-methylnonane < n-decane. The
increase in retention volumes of the paraffins is consistent with the decrease in
barriers energies. Both the calculations and pulse-test experiments indicate that
the trisubstituted paraffin, 3,3,5-trimethylheptane, is too large to fit in silicalite
pores. The disubstituted paraffin, 2,6-dimethyloctane, has limited diffusivity and is
mostly excluded from the pores. The monosubstituted paraffin, 2-methylnonane,
has good diffusion but is slower than the n-paraffin, n-decane. These results suggest
that normal and monomethyl paraffins can be separated from di-/trisubstituted
paraffins, naphthenes, and aromatics using silicalite adsorbent. Laboratory pulse
tests have, in fact, demonstrated these separations, as shown in Figure 7.13.
Separation of mono-methylparaffin from depleted kerosene (removal of
n-paraffins) was also demonstrated in a simulated moving-bed (SMB) pilot
plant with >90% mono-methylparaffin purity at >70% recovery [36, 37]. Another
example of rate-selective adsorption is the separation of di-isopropylbenzene
isomers using a silicalite adsorbent. Figure 7.14 shows the adsorption rates of
1,3-diisopropylbenzene and 1.4-di-isopropylbenzene into silicalite adsorbent. It
168 7 Adsorption
80
3,3,5-TM-C7
70 2,6-DM-C8
60 1,3,5-TMB
50
2-M-C9
40
30
20
10
0
20 30 40 50 60 70 80 90 100 110 120
Retention volume (CC)
Figure 7.13 Chromatographic separation of mono-methyl paraffin.

Concentration in liquid phase
1,3-Diisopropylbenzene
1,4-Diisopropylbenzene
Time
Figure 7.14 Rate-selective adsorption of di-isopropylbenzene isomers on silicalite

adsorbent.
illustrates the more rapid adsorption of 1,4-di-isopropylbenzene compared to

1,3-di-isopropylbenzene.
7.3.2.3 Shape-Selective Adsorption

Shape-selective adsorption, also known as molecular sieving, is a process that
separates molecules based on inclusion or exclusion from specific zeolite
pores. In contrast, the equilibrium- and rate-selective mechanisms are based
on adsorbate–adsorbent interactions and molecular diffusion rates through zeolite
pores. Thus, for shape-selective adsorption, the theoretical selectivity between
molecules that can and cannot enter a specific zeolite pore can be infinite. How-
ever, in most cases, adsorbate–adsorbent and/or adsorbate–adsorbate interactions
are also involved in the shape-selective mechanism. Figure 7.14 illustrates the
separation of n-C5/6 and non-n-C5/6 in CaA molecular sieves or 5A. The separation
mechanism is obvious when the kinetic diameter of the molecules and molec-
ular sieve pore size opening is compared. n-C5/6 have kinetic diameters of less
than 4.4 Å, which can diffuse freely into the 4.7 Å pores of the CaA molecular
sieve, while non-n-C5/6 have kinetic diameters of 6.2 Å. A commercial example
of shape-selective adsorption is the UOP MolexTM process, which uses a Ca form
A zeolite adsorbent as a molecular sieve to separate C10 –C14 n-paraffins from
non-n-paraffins (aromatics, branched, naphthenes).
7.3.2.4 Ion Exchange

Zeolitic adsorbents are composed of many ionic (or potentially ionic) sites. Zeolites
are crystalline, hydrated aluminosilicates containing most commonly Na+ , K+ , and
H+ cations. For most zeolites, the aluminosilicate structure is a three-dimensional
open framework of Al2 O4 and SiO4 tetrahedras linked to each other by oxygen
molecules. The framework contains channels and interconnected voids occupied
by cations and water molecules. The cations are mobile and can be exchanged with
other cations to varying degrees. In ion-exchange separation, cations in liquid are
reversibly exchanged with cations on a solid adsorbent. More specifically, cations
are interchanged with other cations without changing the structure of zeolites.
At equilibrium, electroneutrality in both zeolite and liquid phases is maintained.
Liquid separation based on ion exchange is important in industrial applications.
Critical factors to consider in developing an ion exchange separation process are
adsorbent capacity, selectivity, and kinetics. These factors are described in the
following sections.
Ion Exchange Capacity

Ion exchange capacity is a measure of the quantity of cations adsorbed or removed
by the zeolitic adsorbent. Theoretical ion exchanged capacities of some common
zeolites are calculated and summarized in Table 7.2.
Ion Exchange Selectivity

In general, zeolites have higher ion exchange selectivity for higher-charged cations.
For cations having the same valence, the ion exchange selectivity often depends
on the hydrated ionic radius. This is seen from the zeolite ion exchange selectivity
decreasing with ionic radii [19].
Cs+ > Rb+ > K+ > Na+ > Li+
Ba++ > Sr++ > ++

Ca > ++
Mg
Kinetics
Ion exchange involves the formation and breakage of bonds between ions in solu-
tion and exchange sites in a zeolitic adsorbent. The reaction equilibrium of the ion
exchange process depends most significantly on contact time, operating tempera-
ture, and ionic concentration.
170 7 Adsorption
7.3.2.5 Reactive Adsorption

In conventional chemical process designs, a chemical reactor is typically sequenced
with a downstream separator. The reaction is carried out initially to convert raw
materials into value-added products that are then isolated and recovered in the sep-
arator. The operating conditions of the reactor and separator can be varied to achieve
product yield and purity. However, the variations of operating conditions to achieve
optimum performance are subject to prevailing constraints. In many cases, recycle
streams are incorporated into the process to reprocess unreacted raw materials or
intermediate byproducts back through the reactor and separator to increase overall
process yields.
To overcome some of these sequential process constraints, process intensification
via reactive separation is an alternative. The reactive separations combine the unit
operations of reaction and zeolitic adsorption into a single process operation with
simultaneous reaction and separation. In combining sequential processing steps into
an integrated process, one can eliminate one or more recycle streams that are associ-
ated with optimizing performance of the original sequential process configuration.
Besides eliminating some of the recycling streams, the integration may lead to the
design of a separation process that cannot be achieved with separate reactor and
separator process flow elements. Other advantages of reactive adsorption over the
sequential reaction and adsorption include reducing energy and capital costs, over-
coming the equilibrium-limited conversion, operating at less severe conditions, and
increasing reaction efficiency. More detailed information on reactive adsorption can
be found in a book entitled Reactive Separation Process [38].
Developments in reactive separation date back to applications of simple chemical
treatments, including the use of acidic clays for the removal of olefins in hydrocar-
bon streams, acid-catalyzed polymerization over clay beds, and the use of solid KOH
to remove sulfur from various hydrocarbon streams [39]. More recent developments
have followed on the heels of the discovery and development of better-engineered
synthetic adsorbents and catalysts for applications in the food, biotechnology, phar-
maceutical, chemical, refining, environmental, and nuclear industries. For example,
Ag-substituted molecular sieves and ion exchange resins have been used to remove
trace iodides by reaction and precipitation from vapor and liquid streams to facil-
itate safe operation of nuclear reactors [40] and to purify acetic acid produced by
methanol carbonylation [41–45]. Many reactive adsorption processes developed to
date use traditional fixed-bed and fluidized/moving-bed absorber designs. However,
much of the recent development efforts have been focused on improving purification
processes by reactive chromatographic methods and SMB technologies.
7.4 Commercial Bulk Liquid-Adsorptive Process

7.4.1 Parex
Recognizing the potential importance of polyester to the synthetic fiber market,
UOP looked at many different separation mechanisms and methods to recover
purified para-xylene from the hydrocarbons that were available in gasoline frac-
tions from refineries. Up to that point, fractional crystallization was being used
for purification of para-xylene, with the recovery limited by eutectic composi-
tion to a little over 60%. The UOP Molex process had already been successfully
commercialized for production of high-purity normal paraffins that were used for
production of biodegradable detergents. Unfortunately, the molecular dimensions
of para-xylene and the other feed molecules did not allow molecular sieving to
be used for para-xylene separation. However, the separation stage efficiencies
that could be provided by zeolites were inherently very attractive. When Union
Carbide showed that X- and Y-faujasites could be produced economically at high
purities, UOP intensified its research on competitive adsorption (where essentially
all feed components can access the active sites). Richard Neuzil [46] discovered
that a para-xylene-selective adsorbent could be made by exchanging barium and
potassium onto X- or Y-faujasite, and that it would be possible to use a different
aromatic hydrocarbon to desorb the para-xylene.
This seminal discovery in 1969 led to the UOP Parex process for recovery of
high-purity para-xylene from a mixture of C8 hydrocarbons. The first Parex process
unit was commissioned in 1970 at URBK in Wesseling, Germany, and was designed
to produce approximately 80 000 metric tons per year of 99% para-xylene, with recov-
ery that exceeded 90%. This plant used a mixture of diethylbenzene isomers as the
desorbent. Thus, the extract and raffinate columns had a fairly simple split to make
between the C8 aromatic components of the feed and the C10 aromatic components
of the desorbent. These plants are referred to as “heavy-desorbent Parex” because
the desorbent has a higher boiling point than the feed components. This original
Parex plant was still on stream in 2019. With advances in adsorbent and other PX
technologies from UOP, the capacity of this plant, using the same hardware, is now
more than 250 000 tons/yr. Today, the UOP Parex process accounts for more than
70% of the world’s 50 000 000 tons/yr of para-xylene, as shown in Figure 7.15.
The growth in Parex process PX production between 1974 and 2014 was all with
heavy-desorbent Parex, with steadily higher-capacity plants being built, and with
numerous advances in adsorbent, desorbent, equipment, heat integration, and
control systems being introduced by UOP over this time. Single-train Parex plants
of up to 3 million tons/d of PX are now offered by UOP. Toray tried to use adsorptive
separations for PX with its aromizing process, but that technology suffered from the
use of horizontal adsorbent chambers where bed settling and flow maldistribution
became problematic. In the 1990s, Axens developed the EluxylTM process for PX
production, which uses >100 switching valves to accomplish the SMB that a single
UOP rotary valve can accomplish. Sinopec developed a similar multiple valve
technology in 2010, primarily for deployment within the Sinopec units in China.
Neither of these technologies has garnered the acceptance that has been achieved
by the UOP Parex process.
In the early days of introducing the Parex process to the industry, some prospec-
tive customers voiced concerns about the availability and cost of diethylbenzene as a
desorbent. UOP’s research had identified an alternative adsorbent formulation that
made toluene effective as a desorbent. Four plants of this type were sold. They were
172 7 Adsorption
55
Nameplate capacity, MTA (millions) 50
45
40
35
30
25
20
15
10
5
0
2002
2000
2004
2006
2008
1970
1972
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
1976
2012
2014
2016
2018
2010
1974
Crystallizer capacity Other capacity Parex capacity
Figure 7.15 The UOP Parex process accounts for more than 70% of the world’s
50 000 000 tons/yr of para-xylene.
called “light -desorbent Parex” because the desorbent has a lower boiling point than
the feed and were licensed. Two of these plants are still operating in this mode, while
the other two were revamped to the heavy-desorbent Parex configuration. For these
light-desorbent Parex plants, the fractionation in the extract and raffinate columns
is more difficult since there is only one carbon number between feed and desorbent,
and the desorbent must be boiled overhead. As a result, the utility consumption
in the extract and raffinate fractionators was much higher for the light-desorbent
compared to the heavy-desorbent Parex plants. For this reason, most customers pre-
ferred the heavy-desorbent Parex configuration, particularly since p-diethyl benzene
is readily available.
In recent years, there has been renewed interest in the energy input per ton of PX.
The regions that show the highest growth rates in demand for PX have also tended
to be the regions where energy costs are relatively high, for example, fuel gas val-
ued above US$10 or even US$15 per million BTU. Many of these locations also have
relatively high costs for electricity. These driving forces have led to significant inno-
vations in PX flowschemes by UOP, with the goal of significantly reducing the fuel
fires and the electricity consumed per ton of PX.
These advances have been achieved through a “holistic” view of the flowscheme
for converting naphtha to PX. To understand the potential for holistic energy inte-
gration (outside the battery limits of the Parex unit), a simplified flowscheme for PX
production is shown in Figure 7.16. Naphtha is fed to a CCR PlatformingTM process
to produce aromatic rings ranging from C6 to C11 in a stream called “reformate.”
The reformate is fractionated into a C7− and a C8+ fraction. The light fraction can
be sent to a SulfolaneTM extractive distillation unit where nonaromatics are rejected
and purified benzene and toluene are recovered. The C8+ fraction is sent to a xylene
rerun column where the C8 aromatics are taken overhead and sent through the UOP
Parex unit, which recovers purified PX. The raffinate from the Parex unit is then sent
R-254/284/334
ED Raffinate
SulfolaneTM
unit Benzene
CCR
PlatformingTM Reformate TA-32
unit splitter
B T Heavy
aromatics
TatorayTM column
NHT Unit
ADS-47 I-500
A10+
ADS-50 I-600
p-Xylene
Naphtha
ParexTM IsomarTM Light ends
unit unit
Xylene
column Deheptanizer
column
Figure 7.16 The latest-generation UOP Aromatics Technologies product portfolio.
through a UOP IsomarTM unit that re-equilibrates the xylenes, and the isomerate is
then sent to the xylene rerun column for removal of heavy byproducts. These heavy
aromatic byproducts, along with the toluene and A9+ in the reformate, are sent to a
UOP TatorayTM unit that uses transalkylation reactions to create additional xylenes
and benzene. UOP has offered this integrated aromatic technology since the late
1970s with many plants built throughout the world.
Taking this flowscheme as a basis, UOP identified several process intensification
changes that could be brought together to provide more than a 30% reduction in
energy input per ton of PX. This was done without increasing the capital expendi-
ture of the plant. A key element in this solution was a return to light-desorbent Parex,
but with a significantly improved adsorbent, innovative Parex process configuration,
and significantly different fractionation overhead and bottom composition limits
compared to earlier designs. This new technology has been well accepted by the PX
industry, accounting for more than 20 million tons of additional new PX capacity by
2019, a large fraction of the world’s demand for para-xylene [47].
7.4.2 MX SorbexTM
Of the four C8 aromatic isomers, MX is in highest concentration in most feed
streams, amounting to more than 50% of the C8 aromatics. As discussed in Section
7.4.1, MX is sent to the raffinate of the Parex process, and from there it is sent to
the Isomar process where it is converted back to additional PX. Despite being the
dominant C8 aromatic isomer, the end uses for purified MX are comparatively small
relative to those for PX. Whereas PX is primarily used for polyethylene terephthalate
174 7 Adsorption
(PET) fiber, resin, and film in relatively large amounts, the dominant use of MX
is to produce a co-monomer for PET resin called purified isophthalic acid (PIA)
where the co-monomer content is only 2–3% of the resin. Other uses for MX would
be for meta-xylene diamine (MXDA), which can be combined with adipic acid to
produce coatings and gas barriers for containers. The production of high-purity
MX was highly influenced by Mitsubishi and Amoco, which used HF/BF3 adduct
technology to recover and purify MX for many years.
UOP had known how to purify and recover MX for many years using adsorptive
separations [47] and worked steadily to improve the efficiency of this separation. In
the 1990s, the applications for PET resin quickly expanded, with sustained growth
rates above 10%, for carbonated and noncarbonated beverages. The key to this expan-
sion was the abundance or relatively low-cost MX derived from the UOP MX Sorbex
process. Beginning in 1998, UOP has now licensed several of these plants to pro-
ducers in the United States, Asia, and Europe, even to the two companies that had
originally been using HF/BF3 technology.
The MX Sorbex process is very similar to the light-desorbent Parex process in the
sense that the product MX is selectively adsorbed and extracted from the mixed
xylene feed. Toluene is used as the desorbent, with the desorbent recovered in the
overheads of the extract and raffinate columns. Many producers are also in the PX
market and can pull a slipstream of feed from their PX production plants to feed
their (relatively smaller) MX Sorbex plants. In contrast to the 50 million tons/yr of
PX produced in 2018, the worldwide production of high-purity MX is in the range of
1 million tons/yr.
Acknowledgment
The author of the chapter would like to thank James A. Johnson of Honeywell
UOP for writing the “Commercial Bulk Liquid-Adsorptive Process,” which includes
ParexTM and MX SorbexTM in this chapter.
References
1 Li, L., Xue, B., Chen, J. et al. (2005). Direct synthesis of zeolite coatings on
cordierite supports by in situ hydrothermal method. Applied Catalysis A: General
292: 312.
2 Sawad, J.A., Alizadeh-Khiavi, S., Roy, S., and Kuznicki, S.M. (2005). High density
adsorbent structures. W.I.P.O. Patent WO2005032694.
3 Golden, T.C., Golden, C.M.A., and Battavio, P.J. (2005). Multilayered adsorbent
system for gas separations by pressure swing adsorption. US Patent US6893483.
4 Breck, D.W. (1974). Zeolite Molecular Sieves. New York: Wiley.
5 Barrer, R.M. (1982). Hydrothermal Chemistry of Zeolites. London: Academic
Press.
References 175
6 van Bekkum, H., Flanigen, E.M., Jacobs, P.A., and Jansen, J.C. (eds.) (2001).
Introduction to Zeolite Science and Practice, Studies in Surface Science and Catal-
ysis, vol. 137. Amsterdam: Elsevier Science Publishers B.V.
7 Xu, R., Pang, W., Yu, J. et al. (2007). Chemistry of Zeolites and Related Porous
Materials: Synthesis and Structure. John Wiley & Sons (Asia) Pte. Ltd.
8 Thompson, R.W. (1998). Recent advances in the understanding of zeolite synthe-
sis. In: Molecular Sieves Science and Technology: Synthesis, vol. 1 (eds. H.G. Karge
and J. Weitkamp), 1–33. Berlin: Spriger-Verlag.
9 Stiles, A.B. (1987). Catalyst Supports and Supported Catalysts. Boston: Butter-
worths.
10 Stiles, A.B. and Koch, T.A. (1995). Catalyst Manufacture, 2e. New York: Marcel
Dekker.
11 Beck, D.W. (1974). Zeolite Molecular Sieves: Structure, Chemistry, and Use. New
York: Wiley.
12 Shams, K. and Mirmohammadi, S.J. (2007). Preparation of 5A zeolite mono-
lith granular extrudates using kaolin: investigation of the effect of binder on
sieving/adsorption properties using a mixture of linear and branched paraffin
hydrocarbons. Microporous and Mesoporous Materials 106: 268–277.
13 Jasra, R.V., Tyagi, B., Badheka, Y.M. et al. (2003). Effect of clay binder on
sorption and catalytic properties of zeolite pellets. Industrial and Engineering
Chemistry Research 42: 3263–3272.
14 Kulprathipanja, S. (2010). Zeolites in Industrial Separation and Catalysis. New
York: Wiley.
15 Serrano, D.P., Sanz, R., Pizarro, P. et al. (2009). Preparation of extruded catalysts
based on TS-1 zeolite for their application in propylene epoxidation. Catalysis
Today 143: 151–157.
16 Gordina, N.E., Prokof’ev, V.Y., and Il’in, A.P. (2005). Extrusion molding of sor-
bents based on synthesized zeolite. Glass and Ceramics 62 (9–10): 282–286.
17 Ruthven, D.M. (1984). Principles of Adsorption and Adsorption Processes. USA:
Wiley.
18 Neuzil, R.W. (1982). Separation of m-xylene. US Patent 4,326,092.
19 Broughton, D. and Gerhold, C. (1961). Continue separation process employing
fixed bed of sorbent and moving inlets and outlet. US Patent 2,985,589.
20 Namba, S., Kanai, Y., Shoji, H., and Yashima, T. (1984). Separation of p-isomers
from disubstituted benzenes by means of shape-selective adsorption on morden-
ite and ZSM-5 zeolites. Zeolite 4: 77–80.
21 Barthomeuf, D. (1996). Basic zeolites: characterization and uses in adsorption
and catalysis. Catalysis Reviews - Science and Engineering 34: 521–612.
22 Seko, M., Miyake, T., and Inada, K. (1979). Economical p-xylene and ethylben-
zene separated from mixed xylene. Industrial & Engineering Chemistry Product
Research and Development 18 (4): 263–268.
23 Ward, J.W. (1968). The nature of active sites on zeolites. III. The alkali and
lkaline earth ion-exchanged forms. Journal of Catalysis 10: 34–46.
176 7 Adsorption
24 Kilpatrick, M. and Luborsky, I.E. (1953). The base strengths of aromatic hydro-
carbons relative to hydrofl uoric acid in anhydrous hydrofluoric acid as the
solvent. Journal of the American Chemical Society 75: 577.
25 Kulprathipanja, S. (1995). Process for the adsorptive separation of metaxylene
from aromatic hydrocarbons. US Patent 5,382,747.
26 Kulprathipanja, S. (1983). Separation of bi-alkyl-substituted monocyclic aromatic
isomers with pyrolyzed adsorbent. US Patent 4,423,279.
27 Shimura, M., Wakamatsu, S., and Shirato, Y. (1996). Separation of p-xylene by
using zeolitic adsorbents. Japanese Patent 08,217,700.
28 Zinnen, H.A. (1989). Zeolitic p-xylene separation with tetralin heavy desorbent.
US Patent 4,886,930.
29 Kulprathipanja, S., Khune, K.K., and Patton, M.S. (1993). Process for separating
durene from substituted benzene hydrocarbons. US Patent 5,223,589.
30 Robo, J.A. and Gajda, G.J. (1989-1990). Acid function in zeolites: recent process.
Catalysis Reviews - Science and Engineering 31 (4): 385–430.
31 Johnson, J.A. and Kulprathipanja, S. (1989). Proceedings of the International
Conference on Recent Development in Petrochemical and Polymer Technologies
(12–16 December). Bangkok, Thailand: Petroleum and Petrochemical College,
Chulalongkorn University.
32 Kulprathipanja, S. and Neuzil, R.W. (1984). Low temperature process for separat-
ing hydrocarbons. US Patent 4,455,444.
33 Kulprathipanja, S. (2001). Monomethyl paraffin adsorptive separation process. US
Patent 6,222,088 B1.
34 Kulprathipanja, S. (2001). Process for monomethyl acyclic hydrocarbon adsorp-
tive separation. US Patent 6,252,127 B1.
35 Sohn, S.W., Kulprathipanja, S., and Rekoske, J.E. (2003). Monomethyl paraffin
adsorptive separation process. US Patent 6,670,519 B1.
36 Kulprathipanja, S., Rekoske, R.E., Gatter, M.G., and Sohn, S.W. (2003). Proceed-
ings of the Third Pacific Basin Conference on Adsorption Science and Technology,
Kyongju, Korea (25–29 May).
37 Kulprathipanja, S. (1991). Adsorptive separation process for the purification
of heavy normal paraffins with non-normal hydrocarbon pre-pulse stream. US
Patent 4,992,618.
38 Kulprathipanja, S. (2002). Reactive Separation Process. New York, USA: Taylor &
Francis.
39 Bland, W.F. and Davidson, R.L. (1967). Petroleum Processing Handbook. New
York, USA: McGraw-Hill.
40 Pence, D.T. and Macek, W.J. (1970). Silver zeolite: iodide adsorption studies.
The U.S. Atomic Energy Commission, Idaho Operations Office, Under Contract
#AT (10-1)-1230, Nov. 49 Hilton, C.B. (1986) Removal of iodide compounds from
nonaqueous organic media. US Patent 4,615,806.
41 Kulprathipanja, S., Spehlmann, B.C., Willis, R.R. et al. (1999). Method for
treating an organic liquid contaminated with an iodide compound. US Patent
5,962,735.
References 177
42 Kulprathipanja, S., Vora, B.V., and Li, Y. (1999). Method for treating a liq-
uid stream contaminated with an iodide-containing compound using a solid
absorbent comprising a metal phthalocyanine. US Patent 6,007,724.
43 Kulprathipanja, S., Lewis, G.J., and Willis, R.R. (2001). Method for treating
a liquid stream contaminated with an iodide-containing compound using a
cation-exchanged crystalline manganese phosphate. US Patent 6,190,562 B1.
44 Kulprathipanja, S., Sherman, J.D., Napolitano, A., and Markovs, J. (2002).
Method for treating a liquid stream contaminated with an iodide-containing
compound using a cation-exchanged zeolite. US Patent 6,380,428 B1.
45 Kulprathipanja, S., Vora, B.V., and Leet, W.A. (2003). Combination pre-
treatment/adsorption for treating a liquid stream contaminated with an
iodine-containing compound. US Patent 6,506,935 B1.
46 Neuzil, R.A. (1971). Aromatic hydrocarbon separation by adsorption. US Patent
3558730 A.
47 Frank (Xin X.) Zhu, James A. Johnson, David W. Ablin, and Gregory A. Ernst
(2020). Efficient Petrochemical Processes – Technology, Design and Operation.
Hoboken NJ: Wiley.
179
Distillation
8.1 Introduction to Distillation

Distillation is the oldest, yet still most commercially important, method of chemical
separation and purification. Advances in distillation have been critical to increas-
ing the standard of living for humanity over the past few thousand years [1, 2]. The
origins of distillation technologies trace to the third-century BCE in Mesopotamia.
Distillation knowledge was initially spread by word of mouth and apprenticeship
from its birthplace in the Middle East. Slowly, distillation methods spread to Asia
(especially China, Japan, and India) and Europe; these first steps in the develop-
ment of distillation methods were made largely to enable perfumes, medicines, and
alcohols to be isolated.
With the invention of mass printing of scientific information, distillation, and
other scientific methods could be shared and taught much more efficiently. These
advances in disseminating scientific knowledge were made in China via mechanical
wood block prints of the eighth century AD and then Europe with the invention of
the mechanical printing press in the fifteenth century AD. The start of the Industrial
Revolution in United Kingdom in 1760 enabled step-changes to be made in distil-
lation technologies as the energy required to commercialize these processes could
now be harnessed via steam engines. It was around 1800 that Benjamin Thompson,
Count Rumford, described some of the benefits of steam-heated distillation, includ-
ing rapid heating of feeds as well as greater glass vessel life versus direct heating.
The Napoleonic Wars forced continental Europe to develop new methods for
the isolation of sugar as the supply of cane sugar from the East and West Indies
was cut off by a blockade. One of the scientists tackling this problem was Jean
Baptist Cellier-Blumenthal (1768–1840). Blumenthal’s key contribution to the
forward march of distillation technology was the development of the rectifying,
or fractionation, column. Rectification is used today for the fractionation of crude
oil and consists of a multistage, continuous distillation process accomplished via
counter-current distillation (rectification) in a column. The crude liquid to be
separated is fed to the bottom of the column and brought to its boiling point. The
vapor produced from boiling moves upwards inside the column, exits at the top,
and is condensed. Part of the condensate is removed as the distillate product; the

180 8 Distillation
rest flows back into the column and moves downwards as liquid in the opposite
direction of the rising vapor. The upper portion of Blumenthal’s column incor-
porated bubble-cap trays, while the lower part included conical metal caps for
contacting vapor and liquid. These advances in distillation not only helped supply
the refined sugar Europe needed on the less volatile side of this process but also
launched, on the distillate side, the French rum industry.
The coal industry also drove innovations in distillation technology through the
nineteenth century. In 1823, Friedlieb Runge discovered phenol and aniline in coal
tar and, through further distillation, enabled the use of these tar byproducts in col-
orants and pharmaceutical product development. In addition, ammonia, benzene,
and other aromatic compounds beyond phenol were isolated from the coke gas and
were separated by washing, absorption, and distillation. Column and tray designs
continued to improve throughout the nineteenth century as a wider range of dis-
tilled products were produced. The residue from these processes found use in street
paving by the mid-1800s.
The next great leap in distillation technology arrived with the birth of the
petroleum industry. While the first generation of crude oil distillation was a simple
batch process to produce lamp oil, the outbreak of World War I required larger
volumes of oil and fuel to be isolated. With no chemical conversion, only 20–25%
of crude oil could be distilled into fuel and oil products. In the 1910s, thermal
cracking was utilized to generate lighter oils from the longer-chain hydrocarbons
in the crude. In 1916, steam-assisted distillation of petroleum was accomplished,
and the 1920s ushered in a switch in gasoline production from batch to continuous
distillation. Technologies such as thermal cracking via the Dubbs process [3]
allowed much higher yields of gasoline and other smaller-chain hydrocarbons to be
derived from crude oil.
These distillation process advances were paired with a greater understanding of
the theory and design of distillation. The measurement and prediction of mixture
properties, especially vapor–liquid equilibrium (VLE), are crucial to the accurate
prediction of separation performance.
For example, Rayleigh pioneered the use of Henry and Raoult’s laws to compare
test results to data calculations for simple batch distillations. Hausbrand [4] then
advanced this work using graphical representations of the VLE properties for non-
aqueous and aqueous mixtures. In this manner, the beginnings of VLE theory were
established, based on relative volatility (RV) differences in the materials being dis-
tilled. Finally, McCabe and Thiel developed their diagram to simplify the correlation
of distillation column design with VLE properties and mass balance data [5].
For more than 5 000 years, distillation has served as the primary method for the
separation of complex mixtures into refined, purified products. Even today, distilla-
tion is the most widely practiced separation technology and is responsible for approx-
imately 50% of capital and operating costs in industrial chemical processing. Dis-
tillation accounts for 50% of the annual process energy needs of the chemical and
refining industries [6].
Although the refineries of today – as well as those in the future – are far more
complex, with higher yields, higher value creation, lower regrettable emissions, and
more complex petrochemical production schedules, distillation has maintained its

position as the first and most critical separation step within the refinery. In this
chapter, we explore the basic principles of crude oil distillation, selection of dis-
tillation columns and packing, use of reactive distillation, and explore commercial
examples of distillation. Through a basic understanding of the distillation process,
we expect the reader will build an appreciation of the criticality of this separation
method to the cost and reward of operating a modern petrochemical refinery.
8.2 Principles and Systems

8.2.1 Vapor–Liquid Equilibrium (VLE) Theory
Distillation is a separation process based on VLE properties of a mixture. Whether
under atmospheric or vacuum conditions, distillation is based on understanding and
taking advantage of the properties of VLE of mixtures. Distillation is fundamentally
a separation method in which the variation of equilibrium composition between the
vapor phase and liquid phase provides the driving force of the separation. Thermo-
dynamic VLE compositions can be correlated and calculated with the appropriate
thermodynamic data and models.
For the successful separation of feed components via distillation, it is required that:
(1) The equilibrium distributions of mixture component compositions in liquid
phase must be different from the vapor phase.
(2) The relative volatility of mixture components one to the other must be either
greater than or less than one.
To begin, we define the K value of component a as:
Mole fraction of component a in vapor phase y

Ka = = a
Mole fraction of component a in liquid phase xa
In this way, K a represents the tendency of component a to volatilize into the vapor
phase and is a function of temperature, pressure, and, to a lesser degree, composi-
tion. If K a equals 1, there is an equal amount of component a in the vapor and liquid
phases, while K a > 1 means more a in the vapor phase and K a < 1 means more a in
the liquid phase [7].
To separate components via distillation, we must understand their relative volatil-
ity. The relative volatility of components a and b is defined as K a /K b . This ratio of
volatility is a measure of how easily two components can be separated by distillation.
If the ratio is high, component a is much more volatile than component b, and the
distillation is straightforward. As the ratio (K a /K b ) approaches a value of one, it is
more difficult to separate each of the components. In the extreme case of the value
of K a /K b being exactly one, separation by distillation is not possible other than by
azeotropic or reactive distillation methods. It is not possible to have a relative volatil-
ity ratio less than one as we always define the ratio as the K value of more volatile
component over the K value of the less volatile component.
182 8 Distillation
Vapor, Va, ya Figure 8.1 Binary separation of components a and b.
Feed, xF
Liquid, La, xa
The McCabe–Thiele design method provides a means to model the vapor and
liquid composition (mole fraction) of a component of a binary mix within a distilla-
tion column. It affords a graphical visualization of distillation principles while also
providing a solution to the material balance and equilibrium relationships. This rel-
atively simple approach allows the designer to build on the fundamentals of VLE
theory and clarify design basics such as the number of trays required to achieve a
given binary component separation as well as where to locate the feed within the
tray structure.
Figure 8.1 depicts a binary distillation of a feed of composition xF into a vapor
phase where the mole fraction of the more volatile component, a, is designed as ya
and the mole fraction of a in the liquid phase is xa .
Figure 8.2 shows the equilibrium of a single stage of the distillation of this feed
(labeled stage 1). Because the system is in equilibrium, there are two conditions that
are required to satisfy VLE:
(1) The amount of component a leaving the stage (La xa + V a ya ) is equal to the
amount of vapor and liquid entering the stage from the trays positioned before
(V a−1 ya−1 ) and after stage 1 (La+1 xa+1 ), respectively.
(2) The amount of component a in the vapor phase at equilibrium is equal to the K
factor times the mole fraction in the liquid phase, ya = K a (xa ).
8.2.2 McCabe–Thiele Diagram

The VLE conditions described above allow us to construct the McCabe–Thiele dia-
gram [8]. Shown in Figure 8.3 is a McCabe–Thiele diagram for the distillation of
components a and b (i.e. a binary mixture).
Figure 8.2 Vapor–liquid

equilibrium (VLE) concept.
Va La+1 Material balance stage 1 (Out = In)
ya xa+1 La xa + Va ya = La+1 xa+1 + Va–1ya–1
Stage 1 Equilibrium (VLE): ya = Ka(xa)
Va–1 La
ya–1 xa
Figure 8.3 McCabe–Thiele 1.0

diagram. T1
2
Feed T2
0.8
Tray
T3 4
Vapor mole fraction

1
0.6 T4 3
0.4 5
T5
0.2
ya = 0.9
xa = 0.1 xF = 0.5
0
0 0.2 0.4 0.6 0.8 1.0
Liquid mole fraction
To assemble this diagram, we:
(1) Label the axis as the mole fraction of the lighter component (a) in the vapor
phase (y-axis) versus the mole fraction of the same lighter component (a) in the
liquid phase (x-axis).
(2) Plot Line 1 with slope 1 and intercept of 0 across the diagram.
(3) Plot Curve 2: the VLE curve. This curve is derived from primary thermodynamic
reference values. Recall that according to VLE, ya = K a (xa ).
(4) Determine the feed composition: In this example, the feed contains 50% compo-
nent a, that is, xF = 0.5.
(5) Plot Line 3: the quality line or q-line, which is shown in Figure 8.3 as a mixture
with equal mole of component a as vapor and liquid in the feed. The line, there-
fore, has a slope of −1. M = q/(q − 1), where q = feed quality (liquid mol fraction
of the feed).
(6) Plot Line 4: the rectifying line. For a distillate composition of 90% component a,
we start this line at 0.9 on the x-axis and intersect Line 1. The slope of this line is
L/(L + D), where L = the molar flow rate of reflux and D is the molar flow rate of
the distillate product. The line is extended until it intersects Line 3 (the q-line).
(7) Plot Line 5 (stripping line). For a bottoms composition of 10% component a,
we start this line at 0.1 on the x-axis and intersect the point where Lines 1 and
3 meet.
(8) Now, we can plot the tray structure for this distillation by starting at the xa = 0.1
point and creating a stairstep structure, which stops when it intersects either the
stripping or rectifying lines.
The picture that emerges is a five-tray distillation structure capable of deliver-

ing 90% component a in the distillate (ya = 0.9) from a feed with 50% component
a (xF = 0.5). Additionally, the McCabe–Thiele diagram allows the feed to be located
184 8 Distillation
Figure 8.4 Five-tray binary

Condenser separation distillation column.
More
volatile
T1 product, ya
T2 Rectifying
section
Feed, xF
T3
T4
Stripping
T5 section
Reboiler
Less
volatile
product, xa
within the tray structure. The intersection point of the rectifying line and stripping
line falls with the Tray 3 step; therefore, Tray 3 is the best location for the feed in this
distillation setup. The resulting column and feed setup is shown in Figure 8.4.
8.2.3 Factors for VLE-Based Distillation Design

The successful design of a distillation using the VLE method involves not only the
steps covered above but also more detailed information on several factors, which
must be accounted for by the design engineers: relative volatility, activity coefficient,
solubility, maximum concentration limit, and surface area.
8.2.3.1 Relative Volatility

When the vapor pressure of a liquid is equal to the pressure exerted by the atmo-
sphere, we can say the liquid has reached its boiling point. As discussed previously,
distillation relies largely on differences in the volatility of components of a mixture
to separate and purify those materials.
The feed composition influences the boiling point of the feed mixture as the over-
all feed boiling point is a function of the vapor pressures of each component in the
mixture. As the feed mixture is heated, the most volatile components evaporate first
and are concentrated in the distillate with the less volatile materials being left behind
in the lower range of the distillation column. By knowing the boiling points and rela-
tive volatility of each component, the distillation system design engineer determines
how difficult a separation will be and how relative volatility needs to be accounted
for in this design.
Boiling point is not a static value. Rather, the boiling point of a mixture will vary
as temperature, pressure, and composition change across the distillation column.
As required by VLE theory, the vapor and liquid within each stage reach equilibrium
and no further separation is possible; each stage reaches a unique equilibrium con-
centration of the mixture components according to their relative volatility. Further
purification of the distillate (more volatile components) or bottoms (less volatile
components) is possible through the introduction of either more distillation columns
or other separation methods such as absorption or filtration.
Mixtures with a relative volatility (RV) value greater than 1.5 are described as a
simple distillation and require no special measures to accomplish the desired sep-
aration. By contrast, mixtures with relative volatility (RV) values less than 1.2 are
described as extractive distillation; these separations can be eased by the addition
of heavier components. This method will be described in more detail later in the
chapter [9].
8.2.3.2 Activity Coefficient

The separation of mixed feeds through distillation is a function of the relative
volatilities of the components, their activity coefficients, the relative ratios of the
liquid-phase to vapor-phase flow rates, and the ratio of surface area to liquid within
the distillation column. Activity coefficients help account for nonideal behaviors
of feed components during distillation such as behavior caused by surface tension
(the attraction force of the particles in the surface layer by the bulk of the liquid,
which tends to minimize surface area).
As noted above, the presence of more vapor molecules creates a higher surround-
ing vapor pressure on the mixture within the distillation column. The substance
with greatest relative volatility, the highest vapor pressure, will boil more at a given
temperature because less energy is required to release molecules from the surface
tension. However, the actual amount of a component that boils may differ from ide-
ality, and it is the activity coefficient for the component, which helps correct for this
variation from ideal behavior.
8.2.3.3 Solubility
Solubility is a function of temperature and pressure, so the relative solubility of com-
ponents in the feed as well as throughout the distillation column can make it more
or less difficult to separate those components. Differences in solubility can be used,
for example, to make possible the separation of materials that have the same rela-
tive volatility through the addition of a third component with a higher solubility for
one of the components; this method is referred to as azeotropic distillation and is
described in more detail later in this chapter.
8.2.3.4 Azeotropes: Maximum Concentration Limit

As purification is carried out, it is common to reach a limit where the vapor and
liquid phases include the same concentration of each component. If the liquid in
186 8 Distillation
this example is volatilized under the same pressure and temperature, it will yield the
same concentration of components a and b. In this case, it is not possible to further
separate these components as they have formed an azeotropic mixture that has hit
a constant, maximum concentration limit. One simple example of this maximum
concentration behavior is isopropyl alcohol in water; the concentration of isopropyl
alcohol cannot rise above 91% as the boiling points for the two components are equal
at this maximum concentration limit.
8.2.3.5 Surface Area

As a feed mixture is exposed to an increasing amount of surface area, it is easier to
volatilize the feed and separate the components by relative volatility. Equilibrium
can be established more quickly and efficiently in systems with larger surface area
as the components are vaporized and condensed more rapidly within a given stage
of the distillation column. Due to their increased surface area, both random and
structured packed columns increase the surface area of a distillation process versus
trays alone. A more complete description of column packing materials as well as
structured packings is included later in this chapter.
In addition to the five factors above, the distillation system design engineer needs
to also account for:
● Tray hydraulics: Tray hydraulics and component rheology dictate the flow behav-
ior of a liquid through a tray distillation system. These material properties, along
with the tray design properties, will have a significant impact on the ability of the
distillation apparatus to create the surface area needed to drive separations.
● Feed component stability: The feed boiling point needs to be compared to the indi-
vidual component’s decomposition temperature. Additionally, there is a need to
understand any possible reactions that may occur between components within the
distillation column as relative volatilities for these components can vary greatly
from the feed mixture components.
8.2.4 Operating Modes of Distillation

There are many different distillation strategies possible, and each brings its own
unique strengths and weaknesses as well as selection criteria for their use in different
separations. In this chapter, we will compare three different modes of distillation:
(a) Batch versus continuous

(b) Tray versus packed column
8.2.4.1 Batch vs. Continuous

Batch Distillation
Batch distillation, used mainly in laboratory and fine chemical separations, involves
the charging of the feed to the column in one shot with the distillation then being run
with no additional feed being introduced. This mode of operation was the first and
only means of distillation available for thousands of years. Today, batch distillation is
selected when there are drivers other than cost to consider for a separation process.
Some of the advantages of batch distillation include:
(1) Simplicity: Multiple components can be separated using only one feed and only
one distillation column.
(2) Flexibility: One batch distillation column can be used to handle a wide range of
feed materials with varying quality and composition.
(3) Capacity: Batch distillation is well suited for small volume separations.
(4) Feed difficulty: If a feed has high viscosity or high solids levels, it is easier to run
a batch distillation than use continuous processing methods.
(5) Quality control: Batch distillations can be quality-verified easily based on both
feed and product specifications.
Continuous Distillation
Continuous distillation processes complement batch processing in that they are
lower cost and are tuned to one specific type of feed. The emphasis for continuous
distillations is on stability and the production of a high volume of product stream(s)
at consistent quality.
When constructing a continuous distillation strategy, there are several column
options to consider:
(1) Binary column: Feed contains two components.
(2) Multicomponent column: Feed contains more than two components
(e.g. crude oil).
(3) Multiproduct column: Column has more than two product streams located at
different points along the height of the column.
(4) Extractive distillation: Where feed appears in the bottom product stream.
(5) Azeotropic distillation: Where feed appears at the top product stream.
8.2.4.2 Tray vs. Packed Column

The selection of an appropriate contactor is critical to success in every distillation.
There are two general types of contactors available: tray and packed columns.
Tray column: Trays of specified designs (e.g. sieve, valve, and bubble-cap) hold the
liquid phase as it flows across the surface and allows for contact between the vapor
and liquid phase through openings in the trays to drive enhanced separation. In a
crude oil refinery, a 30-tray stack is common in the atmospheric distillation unit
column.
Packed column: Packing material (either random or structured) are used to enhance
contact between vapor and liquid.
○ Structured packing provides high surface area with high void fraction to pro-
mote efficient vapor–liquid contact.
○ Random packing is commonly used in absorption, stripping, and fractiona-
tion operations in gas, refinery, and chemical plants
The strengths and weaknesses of each mode of column operation are included in
Table 8.1.
188 8 Distillation
Table 8.1 Tray columns versus packed columns.
Tray columns Packed columns
Horizontal flow across the trays Vertical flow through the packing
(counter-current)
Large diameter columns and with high Small-diameter (D < 0.6 m) columns with
liquid flow rates (>30 m3 /(m2 h)) low liquid flow-rate applications
(<50 m3 /(m2 h))
Metal Metal, ceramic, and plastic
High liquid residence time Low liquid residence time (i.e. for
heat-sensitive materials)
High-pressure drop (7 mbar per Low-pressure drop (0.1–0.5 mbar per
equilibrium stage) equilibrium stage) as required for vacuum
distillation
Easy to clean Not typically cleaned/reused
8.3 Distillation Columns and Trays

8.3.1 Flow Within the Distillation Column
Now that we have described the basics of distillation theory using the VLE approach
and detailed several key factors in distillation strategy and execution, we turn our
attention to the structure and function of distillation towers [8]. Figure 8.5 depicts
Figure 8.5 Flow within the distillation

Condenser tower.
Production 1
Reflux distillate
Rectifying
section
Feed
Stripping
section
Reboiler
Product 2
bottoms
a continuous distillation tower containing nine trays with the binary feed mixture
brought into contact with Tray 6. Starting at the top of this column, we see that part
of the condensate is returned to the column as reflux material to provide liquid flow
downward through the column. Likewise, a portion of the liquid bottoms fraction is
returned through the reboiler to the column and provides the upward flow of vapor
from the bottom of the column. The stripping section (below the feed inlet) primarily
removes the more volatile components from the liquid, while the rectifying section
(above the feed inlet) includes an enriched level of volatile components. Once the
targeted level of separation is achieved, this binary, continuous distillation results in
two products, the distillate (more volatile components) and the bottoms (less volatile
component).
As described earlier, binary systems in which the relative volatility is greater than
1.5 fall in the category of simple distillations and, in some cases, the rectifying section
of the column in Figure 8.5 can be removed to yield a simpler stripping column.
8.3.2 Multicomponent Distillation

While we have focused primarily on binary (two component) distillations thus far,
it is important to note that most commercial feeds (e.g. crude oil) consist of many
components and require the use of multicomponent, fractional distillation columns
and methods. As shown in Figure 8.6, it is possible to introduce both multiple feeds
and withdraw multiple products within a single distillation tower. The function
of the distillation column does not fundamentally change with this increase in
complexity, and the same key factors described earlier (relative volatility, activity
Figure 8.6 Multicomponent

continuous distillation column. Condenser
Product 1
Reflux
Feed 1
Product 2
Feed 2 Product 3
Reboiler Product 4
Product 5
190 8 Distillation
coefficient, solubility, maximum concentration limit, and surface area) determine

the best approach to distillation of these complex mixtures.
While pure components a and b distillate and bottoms products are possible in
binary distillations, multicomponent distillations conducted in a single column will
yield products, which are, themselves, multicomponent mixtures that are enriched
with cuts of differing boiling point ranges. To separate multicomponent feeds into
pure component product streams, more complex, multicolumn distillation systems
are required.
8.3.3 Tray Types: Sieve, Valve, Bubble-Cap

The distillation of binary and multicomponent feeds begins when the mixed feed is
introduced to a contactor within the distillation tower. Both tray and packed-bed
contactors are in widespread industrial use. In a tray contactor, the liquid phase
flows across the tray as in Figure 8.7, and the vapor moved upward through gaps
in the tray (shown in this figure as a simple sieve-style tray). The liquid is retained
on the tray by an elevated section called the exit weir and is transferred from stage
to stage within the distillation towers through vertical channels called downcom-
ers and then flows through the clearance onto the next stage’s tray. While the liquid
flows across the tray, the more volatile component(s) are taken into the vapor phase
in the bubbling area. This vapor can then pass upward from stage to stage through
the perforations in the tray. As the vapor moves upward through the above stage’s
tray openings, the structure of that tray will play a key role in both the cost and per-
formance of the overall distillation tower. There are three main types of cross-flow
trays: sieve, valve, and bubble-cap.
There are a few factors that influence the selection of the type of tray to employ:
(1) Operating range: The range of vapor and liquid transfer rates over which the tray
will efficiently separate components of the feed mix. The operating range is often
described in terms of the turndown ratio: the ratio of the highest to lowest flow
rates at which the tray will perform its intended separation. Bubble-cap trays
have the highest turndown ratio, while sieve trays have the lowest with valve
trays in between.
Downcomer Figure 8.7 Flow pattern in

A
cross-flow plate.
Bubbling area
Exit weir Clearance
B
A: Liquid de-entraining zone

B: Vapor disengaging from liquid
C: Degassed liquid
(a) (b) (c)
Figure 8.8 Common distillation tray types. (a) Sieve, (b) Valve, and (c) Bubble-cap.
(2) Operating capacity: The diameter of the column required to provide a give flow
rate is largely invariant across the three types of trays. Sieve trays have a slightly
higher capacity, then valve, and finally bubble-cap.
(3) Cost: Bubble-cap trays are more complex and, therefore, more expensive than
sieve or valve trays.
(4) Pressure drop: As might be imagined, sieve trays have the lowest pressure drop
followed by valve and then bubble-cap trays. Given the liquid low properties of
the feeds involved, vacuum columns need to account more for pressure drop
than atmospheric distillation columns.
Shown in Figure 8.8 are the three types of distillation trays we are most interested
in (sieve, valve, and bubble-cap). Each of the three types is described below:
(a) Sieve tray: This tray is simply a flat plate with 5–6-mm-diameter holes (or slots)
punched in it. The liquid flowing across the holes and tray is retained by vapor
upflow from below. Due to its simplicity, sieve trays are the least expensive, have
the lowest pressure drop, and are most resilient to solids and corrosion (espe-
cially when large holes are used). For efficient operation, the velocity of vapor
through the sieve tray holes must be sufficiently high to prevent liquid on the
tray deck from passing through the perforations to the tray below (weeping).
Conversely, if the velocity of the vapor through the sieve tray perforations is too
high, liquid from the tray below will transfer to tray above (entrainment) and
will cause lower column efficiency. Because of these limitations, the turndown
performance for sieve trays is relatively poor at 2 : 1.
(b) Valve tray: A valve tray is similar in construction to the simpler sieve tray, but
includes a moving valve (e.g. disk), which closes when vapor-flow rates are low
to avoid weeping of the liquid to the stage below. For this reason, valve trays are
characterized by higher turndown ratings (5 : 1 range) due to improved perfor-
mance at low vapor-flow rates but are more costly than sieve trays (about 1.2
times more expensive).
(c) Bubble-cap tray: A bubble-cap tray has riser or chimney fitted over each opening,
and a cap that covers the short riser pipe. The cap is carefully placed to maintain
a space between riser and cap, which allows the passage of vapor. This design
is the oldest type of cross-flow tray and ensures that liquid is maintained on the
tray at all vapor flow rates. While bubble-cap trays bring obvious advantages in
turndown performance (10 : 1 and higher), they suffer from high pressure drops,
high cost (2× sieve trays), and poor corrosion resistance.
A comparison of these three tray types is included in Table 8.2 below.
192 8 Distillation
Table 8.2 Tray comparison.
Tray type Capacity Efficacy Pressure drop Entrainment Turndown Cost
Sieve High High Medium Medium 2:1 1.0

Valve High High Medium Medium 4:1 1.2–1.5×
Bubble cap Medium Medium High 3× Sieve 15 : 1 2.0–3.0×
8.4 Distillation Packing Materials
While distillation itself dates back thousands of years, packed columns only came
into wide use in the 1930s with the introduction of random, but regular, packing
materials such as Raschig rings. Shown in Figure 8.9 is a packed column such as
that used in distillation or absorption; liquid and vapor flow counter-current within
this simple yet effective design.
Packed columns are commonly used for distillation, absorption, stripping, and
heat-exchange applications. The design of these systems is a function of their capac-
ity, pressure drop, and resistance requirements (both mechanical and chemical). The
packing materials loaded with the column establish the capacity and separation effi-
ciency of these distillation systems and have been steadily increasing in performance
over the past several decades to include not only high-performance random packings
but also structured packing materials [10].
Liquid Vapor Figure 8.9 Packed distillation column.

in out
Reflux
Packing
Support
Packed bed
Feed
Packing
Support
Liquid Vapor
out in
8.4 Distillation Packing Materials 193
Figure 8.10 Raschig ring column.

Source: Raschig [11].
8.4.1 Random Packing

Packed column distillation was introduced in the early nineteenth century to purify
alcohol as well as for sulfuric acid absorption. The first packing materials were of
relatively low capacity and consisted of glass beads, stones, and coke. Packing mate-
rial technology advanced greatly with the invention of on-purpose packing materials
such as Raschig rings, which were patented in 1915, to improve acid gas absorp-
tion processes (Figure 8.10) [11]. These rings were made of glass, porcelain, copper,
or iron, were one inch in both diameter and height, and provided improvements
in pressure loss as well as mass transfer properties. Engineered random packings
were long-lasting and provided intimate vapor–liquid contact while minimizing the
holdup of material within the packing and are still used in distillation applications
today.
In the words of Fritz Raschig, “By making these cylinders of uniform size it is
impossible for them to telescope together or fit one inside of another, but they
arrange themselves with only point or line contact with each other and with
their axes either horizontal or inclined and with the axes and surfaces of adjacent
cylinders, at various angles and inclinations with each other. Since any appreciable
surface contact of these cylinders with each other is impossible, and since each
cylinder presents a practically unobstructed passage therethrough, it results that
pocketing of the liquid is minimized or entirely avoided while the various inclina-
tions and arrangements of the adjacent cylinders are such as to cause the ascending
gases or vapors and the descending liquid to pass in a most varied course through
the tower, the course being a zig-zag or multivaried course with frequent interrup-
tions and changes in direction. A most intimate and promiscuous intermixing and
194 8 Distillation
A: Rasching ring (metal, ceramic)

A B C B: Berl saddle (ceramic)
C: Intalox® saddle (ceramic)
D: Pall ring (metal, ceramic)
E: Intalox® metal tower packing, IMTP (metal)
D E
Figure 8.11 Distillation packing elements (random).
intermingling of the ascending gases and vapors and the descending liquid is thus
affected” [11].
Raschig rings were just the start of a new, improved suite of distillation options.
Shown in Figure 8.11 are a range of packing materials used for packed column dis-
tillations and absorption processes today. Pall rings (D) represent a variant of the
original Raschig ring cylinder (A) design with strips of material folder into the inte-
rior of the cylinder; this change provides more even liquid distribution within the
packing phase.
Subsequent generations of packing materials included Berl and Intalox® saddles
(B and C), which are typically made of ceramics and exhibit superior wetting charac-
teristics paired with high corrosion resistance. The saddle packings were developed
to provide improved liquid distribution versus Raschig rings, and the Intalox sad-
dles were easier to manufacture than the original Berl design. While the saddles
themselves are more expensive than Raschig rings, the overall efficiency of Berl and
Intalox saddles as well as Pall rings has led to them being favored for most applica-
tions today.
Developed in the late 1970s, the Intalox IMTP random packing (E) repre-
sents a third-generation packing design, which combines the advantages of the
saddle-shaped packing with that of modern high-performance ring-type packings.
This shape delivers a lower-pressure drop at the same vapor and liquid loads
compared to previous-generation packings as it is one of the most popular random
packings in use today.
8.4.2 Structured Packing

The drive to increase separation capability and capacity of distillation and absorp-
tion columns led to the development in the 1930s of structured packing materials.
Structured packings offer a variety of improvements over trayed systems, includ-
ing superior separation efficiency, shorter column heights, and more continuous,
intimate mixing of vapor and liquid phases throughout the column. Additionally,
structured packing option provides higher capacity and lower-pressure drop than
random packing, but cost can be a barrier to adopting structured packing.
One of the first structured packing designs was the Stedman column from
General Electric for isotope separation [12]. As shown in Figure 8.12, the Stedman
packing consisted of wire cloth, which was punched and embossed to form a series
Metal
mesh
Vapor
openings
Separator
sheet Vapor
openings
(separator)
Figure 8.12 Stedman structured packing.
of triangular cells of regular shape and angle. The Stedman column design provided
lower holdup, low pressure drop, shorter column height, and low weight. These
same advantages drive the selection of structured packing today.
A second generation of structured packing designs and materials was introduced
in the 1970s and consisted of corrugated sheet packing, which further raised col-
umn capacity while lowering costs and increasing plugging resistance of the system.
These second-generation options are now widely adopted as attractive options for
column revamps and have in common design elements such as open honeycombed
structures with 45∘ inclined packing angles in combination with perforations to
maximize capacity and minimize pressure drop (Figure 8.12). The purpose of the
angles and design specifications in a structured packing is to force liquids to take
tortuous paths through the column, thereby creating a large surface area for contact
between different phases and increasing the separation efficiency.
Some common materials used in the design of structured packing are described
below:
Sheet metal: Can tolerate modest levels of fouling and handle a wide range of vapor
and liquid rates. Process applications range from low-pressure drop vacuum and
atmospheric distillation services to medium high-pressure absorbers.
Gauze: Preferred packing for deep vacuum and low liquid rate applications as it has
the lowest pressure drop per theoretical stage, which helps in the purification of
temperature sensitive materials such as those found in specialty chemicals and
pharmaceuticals.
Grid: Both corrugated metal and stamped blades are grid-type packing and both
bring with them a high level of tolerance to fouling as they incorporate open vol-
ume to facilitate flow even in the presence of fouling. Corrugated grid packing has
a smooth surface to minimize fouling and is constructed of much thicker material
than the normal sheet metal packing.
8.5 Commercial Examples

8.5.1 Crude Oil Distillation
The distillation of crude oil is depicted in Figure 8.13 and is performed as a contin-
uous distillation involving both atmospheric and vacuum distillation towers [13].
196 8 Distillation
Crude
oil Condenser
LPG (C3–C4)
Reflux
Desalter
Pentane / Hexane
(C5–C6)
Naphtha (C6–C11)
Heater
Kerosene Light
(C11–C14) vacuum
gas oil
Diesel
(C11–C20)
Heavy
vacuum
Atoms gas oil
gas oil
(C14–C20)
Reduced
crude Wax
Vacuum
Product 5 Heater cilumn
bottoms
Figure 8.13 Crude oil distillation.
Before distillation can proceed, there is typically a need to clean up the crude oil
as crude can include many problematic materials such as sand, inorganic salts, poly-
meric impurities, and more. In most crudes, NaCl, MgCl2 , and CaSO4 form droplets
dispersed throughout the crude. If not removed, these impurities can cause issues
downstream in the refining process such as abrasion from sand, as well as acidic
gases (e.g. HCl) upon heating of the impure crude.
The most common method to desalt crude oil is electrostatic precipitation
wherein the crude is first heated to decrease its viscosity. Fresh water is introduced
to the heated crude and forms small droplets within the continuous oil phase.
This water-in-oil dispersion is then introduced to a pressurized desalter vessel,
which includes a high-voltage electrical field that accelerates separation of the salt
containing water droplets from the desalted oil. As the separation continues, the
water droplets sink within the oil phase and are removed continuously from the
bottom of the desalter, along with any collected sediment, for disposal. Desalted
oil flows continuously from the top of the pressurized vessel and can then be
introduced to the distillation units. Desalting can typically achieve 99% removal
rates for common salts such as NaCl, MgCl2 , and CaSO4 .
Once the desalting process has removed inorganic salts, the resulting crude is
heated to approximately 360–400 ∘ C in tube furnaces and fed into the distillation por-
tion of the refinery. Most refineries employ two types of distillation columns in series:
atmospheric distillation and vacuum distillation. A pre-flash column can optionally

be used prior to the atmospheric column to remove the most volatile materials and
increase system capacity. As their names imply, the atmospheric distillation column
is intended to remove the most volatile components, while the vacuum distillation
unit allows for an increase in gas oil yield under reduced pressures while avoiding
high temperatures that could lead to cracking reactions and degrade the hydrocar-
bon feed.
To avoid thermal cracking, it is common for vacuum distillation units to run at
pressures below 0.05 atm. This reduction in pressure likewise reduces the boiling
point of the vacuum distillation components by as much as 100 ∘ C and helps avoid
the cracking reactions, which commonly occur if the feed experiences a temper-
ature of 350 ∘ C for extended periods of time [14]. The vacuum distillation unit is
fed from the bottoms, or reduced crude, of the atmospheric crude distillation unit
(Figure 8.13). Although designs vary greatly in accordance with crude feeds and sep-
aration goals, it is common for atmospheric units to utilize tray contactors (e.g. 30
stages) and for vacuum units to employ random or structured packings due to their
lower-pressure drop levels.
The major hydrocarbon components of crude oil are:
● Paraffins (alkanes): Straight carbon chains. Formula: Cn H2n+2
● Isoparaffins (iso alkanes): Branched carbon chains. Formula: Cn H2n+2
● Olefins (alkenes): Carbon chains that include at least one double bond. Formula:
Cn H2n
● Naphthenes: Saturated ring carbon structures with five or six carbon atoms in the
ring. Formula: Cn H2n+2 − 2Rn (where Rn = # of rings)
● Aromatics: Six carbon atoms in aromatic ring with varying levels of alkylation
around the ring (e.g. benzene, toluene, xylene, ethylbenzene).
As shown in Figure 8.13, there are many possible product cuts that can be isolated
from the atmospheric and vacuum distillation columns. Those highlighted in the
figure can be valorized in the following ways:
(1) Liquefied petroleum gas (LPG) (C3 –C4 ) can either be used for home heat-
ing/cooking or can be sent to dehydrogenation units such as OleflexTM from
UOP for the production of higher-value petrochemicals such as propylene.
(2) Pentane–Hexane (C5 –C6 ) can be isomerized (e.g. using PenexTM processing) to
increase octane values and is then blended into the gasoline pool.
(3) Naphtha (C6 –C11 ) is a versatile feed for either gasoline or petrochemicals pro-
duction; heavy naphtha is often reformed into aromatics that ultimately find
their way into polymer end uses.
(4) Kerosene (C11 –C14 ) serves as a primary feed for jet fuel and can also be utilized
in the manufacture of biodegradable linear alkylbenzene (LAB) detergents.
(5) Diesel (C14 –C20 ) is a high-value fuel commonly used for transportation.
In similar fashion, the vacuum column products can be converted to
higher-value materials using catalytic cracking (e.g. fluid catalytic cracker [FCC]TM ,
UnicrackingTM ) as well as slurry hydrocracking methods (e.g. UniflexTM ) to raise the
198 8 Distillation
value of the bottoms products from atmospheric distillation in end markets such as
olefins and aromatic petrochemicals as well as fuels.
8.5.2 Extractive Distillation

As discussed earlier, the relative volatility difference between components in a feed
determines how difficult separation of those components will be. A relative volatility
figure >1.5 means simple distillation will be effective. However, once the relative
volatility falls below 1.2, extractive distillation becomes a preferred option.
Extractive distillation involves the addition of a miscible, low-volatility solvent,
which alters the volatility of the other feed components such that separation
becomes easier. The solvent does not form an azeotrope with any of the feed
components but can interact within them, for example, through dipole interactions.
The net result is that the resulting relative volatility is changed to allow easier
separation by distillation.
For example, the separation of isobutane (i-C4 , −12 ∘ C boiling point) and 1-butene
(1-C4 = , −6 ∘ C boiling point) falls in the extractive range as the relative volatility of
these components is 1.16. With the addition of 80% furfural (162 ∘ C boiling point),
relative volatility increases to a value of 2.0 (1-butene is soluble in the higher boiling
furfural component due to dipole interactions), and the distillation can be carried out
in a straightforward manner. Other commercial examples of extractive distillation
include propane-propylene (via acrylonitrile), toluene-nonaromatics (via phenol),
and butane-butadiene (via furfural).
8.5.3 Azeotropic Distillation

When the relative volatility of different feed components is equal to 1, simple and
extractive distillations are no longer possible. Instead, separation must be accom-
plished via azeotropic distillation. In this strategy, an entrainer (volatile solvent) is
added to the feed mixture to form an azeotrope with one of the feed components. As
one component interacts more with the volatile solvent, relative volatility of the feed
components is increased and thereby makes it easier to accomplish a separation.
As an example, ethanol and water form an azeotrope at 94.6/5.4 wt% (89/11 mol%)
ethanol/water (relative volatility = 1). The addition of benzene as solvent/entrainer
serves to increase the relative volatility of water (more polar component), while
ethanol interacts more strongly with benzene. Therefore, the relative volatility of
ethanol/water decreases below a value of 1, and separation can be achieved in the
system. Other examples of azeotropic distillations include acetic acid-water (via
ethyl acetate), pyridine-water (vis bisphenol), and cumene-phenol (via trisubstituted
phosphate).
8.5.4 Dividing Wall Column Distillation

While process intensification has garnered more attention in recent years, dis-
tillation approaches, which incorporate the principles of process intensification,
Thermally Mechanically
integrate integrate
A A
A B
A/B
Dividing
wall
ABC ABC ABC PF
C1 C2 PF MC B B
B/C C B/C
C C
Simple sequence Pre-fractionation Dividing wall column

+ main column
Figure 8.14 Dividing wall column basics.
have been known for many decades. For example, dividing wall columns (DWC)
appeared in patent literature as early as 1946 [15]. Deployment of DWC technology
took place starting in the mid-1980s as control of these systems advanced greatly
with the advance of computer-aided process management tools. The roughly 30%
savings in capital investment and operating expenses have led to more than 200
dividing wall column installations globally in 2020.
Shown in Figure 8.14 is a representation of the sequential separation of a
three-component feed in two columns alongside the integrated dividing wall
column design. In the pre-fractionator section of the dividing wall column, a
rough separation is carried out between components A and C. Component A is
concentrated at the top of the prefractionator and C at the bottom with component
B distributed between the top and the bottom. The pre-fractionator overhead then
feeds the upper half of the main fractionator for the binary separation of A and B,
while the bottom of the pre-fractionator feeds the lower half of the main column for
the separation of B and C.
The main fractionator liquid from the top section refluxes the pre-fractionator top
and the main fractionator on the other side of the dividing wall, while the vapor from
the bottom section similarly strips the bottom of the pre-fractionator and the main
fractionator. The middle component B is withdrawn at the stage where its concen-
tration is at a maximum, avoiding re-mixing losses, which can lead to energy savings
of up to 25–30% or more.
There are now many relatively well-known potential applications for dividing wall
columns, including straight-run naphtha fractionation upstream or downstream of
reformers, FCC naphtha fractionation, C4 isomer separation, and benzene, toluene,
and xylene (BTX) separation. A dividing wall column also is used for UOP’s Pacol
enhancement process (PEP). In this process, A is pentane, B is benzene, C consists
of C7 + olefins, and D comprises C7 + aromatics. To prevent mixing of C and D, a
novel trap tray is applied and excessive B is added [16]. The next logical step in
DWC performance is to include multiple dividing walls within the DWC column,
200 8 Distillation
and designs have been developed to produce six high-purity products from a single
column [17].
8.5.5 Reactive Distillation

Reactive distillation is defined as the simultaneous distillation and catalytic
chemical reaction within a single process unit and, like the dividing wall column,
is another application of process intensification principles. Reactive distillation
is classified as a two-phase, counter-current, fixed-bed catalytic process [18, 19].
There are several benefits of reactive distillation, including:
(1) Lower capital investment: Catalysis and distillation combined in one
process unit.
(2) Lower energy costs: Exothermic reaction heat can be used to drive distillation
vaporization (lower reboiler heat demand).
(3) Longer catalyst life: Especially if the reaction zone is located above the feed.
(4) Entrainers: Azeotropes can be broken as reactants or products can serve as
entrainers.
(5) Higher conversion: Product removal from the reaction phase allows equilibrium
to be re-established at higher conversion.
The EthermaxTM process from UOP uses reactive distillation technology and
employs the KataMaxTM structural packing system from Koch-Glitsch; this packing
is specifically designed to expose a solid catalyst to a reacting liquid stream. The feed
consists of MeOH or EtOH and hydrocarbon streams containing reactive tertiary
olefins such as isoamylene and isobutylene. Ethermax catalyzes the conversion
of isobutylene and methanol into a high-octane methyl-tert-butyl ether (MTBE)
blending agent that contains no benzene or aromatics; the reactive distillation
process delivers a high MTBE yield and is currently used in 44 units worldwide.
References
1 Kockmann, N. (2014). History of distillation. In: Distillation: Fundamentals and

Principles (eds. A. Gorak and E. Sorensen), 1–43. London: Academic Press.
2 Forbes, R.J. (1948). A Short History of the Art of Distillation. Leiden: EJ Brill.
3 Dubbs, C.P. (1921). Process of converting hydrocarbons. US Patent 1,392,629,
filed 19 March 1919 and issued 4 October 1921.
4 Haubrand, E. (1916). Die Wirkungsweise der Rektifizier- und Destillier-Apparate
mit Hilfe einfacher mathematischer Betrachtungen, 3e. Berlin: Springer.
5 McCabe, W.L. and Thiele, E.W. (1925). Graphical design of fractionating columns.
Industrial & Engineering Chemistry 17 (6): 605–611.
6 Gorak, A. and Sorensen, E. (2014). Distillation: Fundamentals and Principles.
London: Academic Press.
7 Kister, H.Z. (1992). Distillation Design. New York: McGraw-Hill.
8 Towler, G.P. and Sinnott, R. (2012). Chemical Engineering Design, 2e. Burlington:
Elsevier.
References 201
9 Gerster, J.A., Mertes, T.S., and Colburn, A.P. (1947). Ternary systems
n-butane–1-butane–furfural and isobutane–1-butane–furfural. Industrial & Engi-
neering Chemistry 39 (6): 797–804.
10 Chilton, T.H. and Colburn, A.P. (1935). Distillation and absorption in packed
columns. Industrial & Engineering Chemistry 27 (3): 255–226.
11 Raschig, F. (1915). Absorption and reaction tower for acids. US Patent 1,141,266,
26 filed March 1914 and issued 1 June 1915.
12 Stedman, D.F. (1935). Packing for fractionating columns and the like. US Patent
2,047,444, filed 14 January 1935 and issued 14 July 1936.
13 Alfke, G., Irion, W.W., and Neuwirth, O.S. (2007). Oil refining. In: Ullmann’s
Encyclopedia of Industrial Chemistry (ed. James G. Speight), 7e, 208–260. Wein-
heim: Wiley-VCH.
14 Speight, J.G. (2007). The Chemistry and Technology of Petroleum, 4e. New York:
CRC Press.
15 Wright, R.O. (1946). Fractionation apparatus. US Patent 2,471,134, filed 17 July
1946 and issued 24 May 1949.
16 Yildirim, O., Kiss, A.A., and Kenig, E.Y. (2011). Dividing wall columns in chemi-
cal process industry. Separation and Purification Technology 80: 403–417.
17 Piszczek, R., Hergenrother, M., and Wang, Z. (2020, 2020). Introduction to
dual-dividing-wall columns. Chemical Engineering Progress 4: 22–25.
18 Tower, G.P. and Frey, S. (2002). Reactive distillation. In: Separation Processes
(ed. S. Kulprathipanja), 18–50. New York: Taylor & Francis.
19 Sakuth, M., Reusch, D., and Janowsky, R. (2007). Reactive distillation. In:
Ullmann’s Encyclopedia of Industrial Chemistry (ed. James G. Speight), 7e,
264–275. Weinheim: Wiley-VCH.
203
Membranes
Membrane technology has been an important enabler for effective and efficient
gas, vapor, and liquid separations and a wide range of other potential applications
such as for catalytic membrane reactors, membrane distillation, and energy storage.
This chapter covers perspectives of membranes from a combination of both
academic research and industrial development.
9.1 Principle and Background

A membrane is a selectively permeable barrier between feed and permeate phases
that can be used to separate the components of a gas, vapor, or liquid mixture
based on their relative permeation rate through the membrane material. Membrane
separation is a nonequilibrium process and determined by both the physical and
chemical properties of the mixtures to be separated as well as the membrane
materials and membrane structures. Membranes are commonly prepared from
organic, inorganic, or organic/inorganic composite materials. They are generally
not limited by azeotropes or eutectics, less energy intensive, and have lighter weight
and smaller footprint than other traditional separation processes such as cryogenic
distillation, absorption, and adsorption processes. In addition, membranes have
the advantages of modular design, no regeneration requirement, simple and rapid
installation, quick start-up, ease of operation, high reliability, low capital and
operating cost, ideal for offshore applications such as natural gas purifications on
floating, production, storage, and offloading (FPSO) vessels [1].
The state of the art of polymeric membrane-based industrial separation processes
has evolved rapidly since the development of asymmetric integrally skinned
polymeric membrane by Loeb and Sourirajan [2] and the development of thin-film
composite (TFC) membranes by Cadotte [3]. Membranes are competitive or poten-
tially competitive in a variety of separations such as gas, vapor/gas, vapor, aqueous
liquid, and organic liquid separations. Membrane technologies have been success-
fully used for large-scale water purification applications at low cost and minimum
environmental impact. Polymeric gas separation membranes are of special interest to

204 9 Membranes
petroleum producers and refiners, chemical companies, and industrial gas suppliers.
Several applications of gas separation membranes have achieved commercial
success, such as nitrogen enrichment from air, carbon dioxide removal from natural
gas and from enhanced oil recovery, separation and recovery of organic vapors, and
in hydrogen removal from nitrogen, methane, and argon in ammonia purge gas
streams. However, membrane technologies with limitations for organic vapor/vapor
or liquid feed separations, such as for olefin/paraffin, aromatic/nonaromatic, and
aromatic/aromatic separations, have not been commercially successful so far.
Polymeric membranes often suffer from thermal stability and plasticization or
swelling of the organic polymer matrix by the sorbed penetrating molecules such
as high concentration of carbon dioxide or olefins, which results in the decrease
of membrane selectivity and therefore lower product recovery and/or higher com-
pression cost due to higher recycle rates. Inorganic membranes such as microp-
orous zeolitic membranes can overcome some of the challenges facing polymeric
membranes. For example, zeolite membranes have the potential for separations with
high efficiency and high productivity under conditions where polymeric membranes
cannot survive by taking advantages of their superior thermal and chemical stabil-
ity, good erosion resistance, and high plasticization resistance to condensable gases
or vapors. The main challenges facing inorganic membranes for large-scale indus-
trial applications are their high cost, reproducibility, difficulty of making very thin
(<100 nm), grain boundary-free, pinhole-free, and crack-free selective layer.
The significant growth of membrane technologies in refinery, petrochemical,
and natural gas markets requires breakthrough development of new membranes,
modules, and processes that overcome the challenges facing the current membrane
technologies. Innovation in membrane materials, equipment, and process technol-
ogy integration will play a key role to enable new large-scale, low-energy-intensity
membrane-based separations.
9.2 Types and Preparation of Membranes

Three types of membranes, including polymeric, inorganic, and polymeric/inorganic
mixed-matrix membranes (MMMs), have been studied extensively in academia and
industries for a wide range of potential applications. These membrane applications
include gas separations, gas/vapor separations, organic vapor/vapor separations,
pervaporation, reverse osmosis (RO), organic liquid RO, forward osmosis (FO),
pressure-retarded reverse osmosis (PRO), nanofiltration (NF), organic solvent
nanofiltration (OSNF), ultrafiltration (UF), microfiltration (MF), electrodialysis,
fuel cell, batteries, membrane contactor, membrane distillation, membrane reac-
tors, etc. Polymeric membranes are the most common membranes for commercial
separation applications. For example, almost all of the gas separation and RO
commercial membranes are made from polymers. Inorganic microporous zeolite
membranes have been applied to commercial-scale pervaporation processes for
organic solvent dehydration. Inorganic ceramic membranes have commercial
applications in battery, NF, and UF processes.
9.2.1 Polymeric Membranes

9.2.1.1 Fundamentals of Polymeric Membranes
Polymeric membranes are the first membranes commercialized in large industrial
scale for separations and have been operated successfully in gas and liquid sepa-
rations, fuel cells, and some electrochemical processes. Polymers, including both
glassy and rubbery polymers, provide a range of properties, including low cost, good
permeability, high mechanical stability, and ease of processability that are important
for membrane fabrication and applications. For example, cellulose acetate (CA), and
cellulose triacetate (CTA) were used to manufacture commercial SeparexTM spiral
wound membrane for acid gases removal from natural gas [4, 5]. Polysulfone poly-
mer was used to produce PRISM® hollow-fiber membrane for hydrogen separation
and air separation [6] and porous polysulfone (PSf) support membrane for the fab-
rication of some gas separation membranes, RO, and NF membranes. Polyimides
and polyaramides were used for the fabrication of gas separation membranes for
acid gases removal from natural gas, air separation, hydrogen purification, or helium
recovery from natural gas. Silicone rubber polymer was used for the manufacture of
rubbery membranes for organic vapor/gas or nitrogen removal from natural gas [7].
The separation layer of most of the current commercial RO and NF membranes was
formed from the in situ polymerized polyamides (PAs).
Glassy polymers have stiffer polymer backbones than rubbery polymers and there-
fore allow smaller molecules such as hydrogen, helium, and water to pass through
faster, while larger molecules such as hydrocarbons and salts pass through slower
as compared to polymers with less stiff backbones. Gas separation, RO, and perva-
poration processes using polymeric membranes are based on a solution-diffusion
mechanism [8]. This mechanism involves molecular-scale interactions of the per-
meating molecule with the polymer. The permeating molecule is first sorbed at the
feed side surface of the membrane, then transported by a concentration gradient
of the permeating molecule, and finally desorbed at the permeate side surface of
the membrane. For a separation process using a polymeric membrane based on the
solution-diffusion separation mechanism, the membrane performance in separating
a given mixture of gases, vapors, or liquids comprising a more permeable compo-
nent A and a less permeable component B is determined not only by the intrinsic
permeabilities for components A and B (abbreviated as PA and PB ) and selectivity
of A over B (abbreviated as 𝛼 A/B ) of the polymeric membrane material, but also by
the membrane selective layer thickness and membrane stability. The permeability of
component A or B (PA or PB ), measured in Barrer units (1 Barrer = 10−10 cm3 (STP)
cm/cm2 s (cm Hg)), is the product of the flux of permeating component A or B and
the selective skin layer thickness of the membrane, divided by the partial-pressure
difference of permeating component A or B across the membrane. The 𝛼 A/B is the
ratio of the permeabilities of permeating component A and B (𝛼 A/B = PA /PB ), where
PA is the permeability of the more permeable component A and PB is the perme-
ability of the less permeable component B. The permeabilities of components A and
B through the membrane are determined by solubility coefficient (abbreviated as
S) and diffusivity coefficient (abbreviated as D) (PA = SA × DA and PB = SB × DB ).
206 9 Membranes
The solubility coefficient equals the ratio of sorption uptake normalized by some
measure of uptake potential, such as partial pressure. The diffusivity coefficient is a
measure of the mobility of component A or B passing through the voids between
the polymer chains in the polymeric membrane. Therefore, the selectivity of the
polymeric membrane for component A over component B (𝛼 A/B ) is the product of
the solubility selectivity of component A over component B (abbreviated as SA /SB )
and diffusivity selectivity of component A over component B (abbreviated as DA /DB )
(𝛼 A/B = SA /SB × DA /DB ). A polymeric membrane can have a high PA because of high
SA , high DA , or both, and the membrane can have a high 𝛼 A/B because of high SA /SB ,
high SA /SB , or both. Normally the solubility coefficient increases and the diffusiv-
ity coefficient decreases with an increase in the molecular size of the permeating
component for a glassy polymeric membrane and vice versa for a rubbery polymeric
membrane. Both permeability and selectivity are the intrinsic properties of the poly-
meric membrane materials and ideally are constant with the change of the thickness
of the membrane selective layer, feed pressure, flow rate, and other separation pro-
cess conditions although they are temperature dependent.
Membrane permeance, measured in gas permeation units (GPUs,
1 GPU = 10−6 cm3 (STP)/cm2 s (cm Hg)), is the pressure-normalized flux and
determined by the intrinsic permeability of the polymer material and the selective
skin layer thickness of the membrane. Therefore, a commercially viable polymeric
membrane based on a solution-diffusion separation mechanism needs to have a
high intrinsic permeability and a thin selective layer to achieve a sufficiently high
permeance at the preferred separation conditions. Otherwise, extraordinarily large
membrane surface areas are required to achieve high treating capacity. For mem-
brane separation applications, both high permeance and selectivity are desirable
because higher permeance decreases the size of the membrane area required to
treat a given volume of feed mixture, thereby decreasing the capital cost of the
membrane system, and higher selectivity results in a higher-purity product.
Polymeric materials have been studied extensively for the development of mem-
branes for separation applications, particularly for solution-diffusion-mechanism
based separations. Many research articles and patents have reported the synthesis of
polymeric membrane materials, such as polyimides (PIs), polyethersulfones (PESs),
polyethers, polyamides, polybenzoxazoles, polybenzimidazoles (PBIs), and poly-
mers with intrinsic microporosities. Traditional polymeric membrane materials,
however, have shown a well-known tradeoff between permeability and selectivity
(or so-called polymer upper-bound limit). By comparing the permeation data of over
three hundred different polymers, Robeson reported in 1991 and 2008 that the selec-
tivity and permeability of polymeric gas separation membrane materials were insep-
arably linked to one another, in a relation where selectivity increases as permeability
decreases and vice versa [9, 10]. Substantial research effort has been directed to
design novel polymer structures or modify existing polymer structures to overcome
the permeability and selectivity limits imposed by the polymer upper-bound limit
in academia and industrial research labs. Fabrication of defect-free high selectivity
membranes with a thin selective skin layer, development of advanced membrane
(a) (b)
(c) (d)
(e)
Figure 9.1 Schematic illustration of membrane structures: (a) symmetric nonporous

structure; (b) symmetric isotropic porous structure; (c) asymmetric porous structure;
(d) asymmetric structure with a thin nonporous selective layer (<100 nm) supported on a
porous support layer; and (e) thin-film composite asymmetric structure with a thin
nonporous selective layer (<100 nm) supported on a porous support layer. (Photos provided
courtesy of Honeywell UOP with a current copyright.)
element/module, and design of novel membrane processes have also played an

important role in the advancement of membrane technologies for separations.
9.2.1.2 Preparation of Polymeric Membranes

Polymeric membranes can have symmetric or asymmetric bulk structures and each
of them can have a porous or nonporous substructure as illustrated in Figure 9.1,
depending on their applications. Polymeric membranes can be fabricated into
many different geometries such as flat sheet that can form spiral wound membrane
modules, tube, and hollow fiber. Asymmetric polymeric membranes normally
have two different structures, including asymmetric integrally skinned membrane
structure and TFC membrane structure. Figure 9.2 shows the cross-sectional
High-Resolution Scanning Electron Microscopy (HRSEM) images of an asymmetric
integrally skinned flat-sheet membrane, a TFC flat-sheet membrane, an asymmetric
integrally skinned hollow-fiber membrane, and a TFC hollow-fiber membrane.
Symmetric nonporous dense-film polymeric membranes can be prepared either
by dissolving the polymer in an organic solvent followed by solution casting using a
casting knife and evaporating the organic solvent or by a melt pressing process using
a high-temperature plate under high pressure. Symmetric nonporous dense-film
polymeric membranes not only can be used as the proton-conducting membranes
for fuel-cell applications, but also can be used to screen membrane materials by
measuring the intrinsic properties of membrane materials for gas separations.
Symmetric porous polymeric membranes can be prepared by solution precipita-
tion, irradiation, or physical biaxial stretching of a symmetric nonporous dense-film
membrane to form uniform pores. Symmetric porous polymeric membranes such as
polypropylene Celgard® and Gore-Tex® membranes prepared by biaxial stretching
method can be used for many applications such as separators for battery applica-
tions, MF, and membrane contactor applications.
Asymmetric integrally skinned porous polymeric membranes are commonly used
for UF, MF, membrane distillation, and membrane contactor applications, and can
208 9 Membranes
(a) (b)
(c) (d)
Figure 9.2 Cross-sectional HRSEM images of: (a) asymmetric integrally skinned flat-sheet
membrane; (b) TFC flat-sheet membrane; (c) asymmetric integrally skinned hollow-fiber
membrane; and (d) TFC hollow-fiber membrane.
be used as the support membranes for the fabrication of RO, NF, and gas separation
membranes. This chapter will not cover the preparation of this type of asymmetric
integrally skinned porous polymeric membranes in detail.
Asymmetric polymeric membranes for gas separation, RO, NF, and pervaporation
applications can have either an asymmetric integrally skinned membrane structure
or TFC membrane structure. Such membranes normally consist of a thin, dense,
nonporous, selective skin layer of <100 nm and a less-dense void-containing
(or porous), nonselective support layer, with pore sizes ranging from large in the
support region to very small proximate to the skin layer. The thin, dense, nonporous,
selective skin layer provides the membrane resistance and selectivity and the thick,
porous support layer provides the membrane with high mechanical strength.
Design and selection of proper polymeric materials with desired properties is
a critical initial step for the development of asymmetric polymeric membranes.
In addition to the consideration of high permeability and selectivity, the polymer
cost, purity, molecular weight, commercial availability or manufacturing possibility,
solubility of the polymer in organic solvents for the preparation of polymer solu-
tions, film-forming property and membrane processability, mechanical property,
post-treatment possibility, chemical and thermal stabilities, and hydrophilic-
ity/hydrophobicity are all the factors that need to be considered to select the best
candidate polymer for the development of high-performance asymmetric polymeric

membrane for a particular application.
Preparation of Asymmetric Integrally Skinned Polymeric Membranes

For asymmetric integrally skinned polymeric membranes, the thin, dense, non-
porous, selective skin layer on top of the thick, porous support layer is formed
simultaneously with the thick, porous support layer by a solution precipitation
method from the same polymer solution via a phase-inversion process first devel-
oped by Loeb and Sourirajan. Therefore, the thin, dense, nonporous, selective skin
layer and the thick, porous support layer are of the same polymer material. For
a phase-inversion process, a one-phase polymer solution (or so-called polymer
dope) is converted into a two-phase system consisting of a solid (polymer-rich)
phase, which forms the membrane structure with either a flat-sheet geometry or a
hollow-fiber geometry; and a liquid (polymer-lean) phase, which forms the pores in
the final membrane. The advancement of membrane fabrication equipment since
the invention of phase-inversion process has enabled better control of membrane
fabrication-processing conditions such as tension, humidity, casting speed, and
ventilation in the evaporation zone, which results in the development of new
membranes from high-performance new polymers.
Preparation of asymmetric integrally skinned flat-sheet membrane via the
phase-inversion process comprises many steps, including: (i) design and prepa-
ration of a polymer-casting solution: A polymer is dissolved in a mixture of
organic solvents to form a homogeneous polymer solution. The composition of
the polymer-casting solution is one of the key parameters that determine the
membrane property. The polymer-casting solution comprises one polymer or
a blend of two or more polymers, organic solvents that can dissolve the poly-
mers such as N-methylpyrrolidone (NMP), N,N-dimethyl acetamide (DMAC),
N,N-dimethylformamide (DMF), 1,3-dioxolane, dimethyl sulfoxide (DMSO), and
some other components. Organic nonsolvents that cannot dissolve the polymers
such as methanol, ethanol, isopropanol, glycerol, acetone, n-octane, citric acid,
and lactic acid and other additives such as lithium chloride and lithium nitrate are
added to the polymer-casting solution in some cases to achieve the best membrane
property; (ii) casting the polymer-casting solution on a highly porous substrate such
as a polyester or nylon cloth substrate to form a controlled layer of polymer solution
on the surface of the substrate using a casting knife under controlled casting speed,
temperature, and humidity; (iii) allowing evaporation of a controlled amount of
solvents and nonsolvents from the surface of the layer of polymer solution on the
substrate under controlled humidity, temperature, air flow, and time to facilitate the
polymer densification and formation of the thin, dense, nonporous, selective skin
layer. This step can be avoided for the preparation of asymmetric porous integrally
skinned polymeric membranes; (iv) coagulating the layer of polymer solution on
the substrate by immersing the partially dried nascent membrane layer together
with the substrate into a low-temperature controlled coagulation medium such as a
0–5 ∘ C cold water bath when water is a nonsolvent for the polymer. An asymmetric
porous support layer underneath the thin, dense, nonporous, selective skin layer
210 9 Membranes
is formed after the organic solvents and nonsolvents diffuse out of the interior of
the layer of polymer solution and the nonsolvent liquid in the coagulation medium
such as water diffuses into the interior of the layer of polymer solution; (v) washing
the nascent asymmetric integrally skinned membrane comprising a thin, dense,
nonporous, selective skin layer and a highly porous support layer in a nonsolvent
medium such as a water bath to remove the leftover organic solvents, nonsol-
vents, and additives from the nascent asymmetric integrally skinned membrane;
(vi) annealing the nascent asymmetric integrally skinned membrane comprising a
thin, dense, nonporous, selective skin layer and a highly porous support layer in an
annealing medium with a controlled temperature such as a hot water-annealing tank
controlled at 55 ∘ C. The trace amount of leftover organic solvents, nonsolvents, and
additives from the coagulating medium will be further removed during this anneal-
ing step. In addition, both the thin, dense, nonporous, selective skin layer and the
highly porous support layer will be further stabilized with certain degree of shrink-
age. The stabilization of the membrane structure is critical to provide the membrane
long-term performance stability for high-pressure applications. This annealing step
is unnecessary for fabricating certain asymmetric integrally skinned membranes;
and finally, (vii) drying the wet asymmetric integrally skinned membrane to remove
the coagulation and annealing fluids from the membrane without causing changes
to the substructure of the membrane at certain temperature from about 50 to about
100 ∘ C. Some membranes need to be kept wet without drying before use and this
drying step can be avoided and replaced with an additional conditioning step by
placing the wet membrane in a conditioning medium that may contain water and
conditioning agent such as glycerol or polyethylene glycerol. The conditioning
agent diffuses into the pores of the membrane to prevent the collapse of the pores
during storage. In some cases, some defects on the thin, dense, nonporous, selective
skin layer of the membrane caused by gas bubbles, dust, or substrate imperfections
cannot be eliminated. These defects can be plugged by coating a thin layer of a highly
permeable polymer such as silicone rubber or high-permeability fluoropolymer on
the surface of the thin, dense, nonporous, selective skin layer of the membrane.
Preparation of asymmetric integrally skinned polymeric hollow-fiber membranes
via the phase-inversion process is similar to that of the asymmetric integrally
skinned flat-sheet membranes. One major difference is that a casting knife is
used to cast a thin layer of polymer solution, which will then be converted into
asymmetric integrally skinned flat-sheet membrane, but an annular spinneret is
used to extrude a hollow-fiber polymer solution, which will then be converted into
asymmetric integrally skinned hollow-fiber membrane. Another major difference
is that a highly porous substrate such as a polyester or nylon cloth, which will
provide the membrane mechanical strength, is normally used for the fabrication of
asymmetric integrally skinned flat-sheet membranes, but preparation of polymeric
hollow-fiber membranes does not need any substrate and the membranes are
freestanding membranes. In addition, normally the polymer spinning solution
for the preparation of hollow-fiber membranes has higher polymer concentration
and therefore higher solution viscosity than the polymer-casting solution for the
preparation of flat-sheet membranes. The composition of the polymer-casting
solution or spinning solution and the coagulation media are two key parameters
that determine the final membrane separation performance.
Overall, the thin, dense, nonporous, selective skin layer of the asymmetric
integrally skinned polymeric membranes will not be delaminated easily from
the thick, porous support layer because both layers are formed simultaneously
from a one-step phase-inversion process. However, fabrication of defect-free, high
selectivity asymmetric integrally skinned polymeric membranes sometimes is
challenging. The presence of nanopores or defects in the selective skin layer reduces
the membrane selectivity. In some cases, the high shrinkage of a high-performance
new polymeric membrane on the highly porous substrate during membrane casting
and drying process results in unsuccessful fabrication of asymmetric integrally
skinned flat-sheet membranes using the phase-inversion process. In some other
cases, novel polymeric materials have very limited solubility in common organic
solvents that are used for the fabrication of asymmetric integrally skinned polymeric
membranes, preventing the development of high-performance membranes.
Preparation of TFC Polymeric Membranes

Since the thin, dense, nonporous, selective skin layer and the thick, porous, non-
selective support layer of the asymmetric integrally skinned polymeric membrane
are formed simultaneously from the same polymeric material, the cost of a new
asymmetric integrally skinned polymeric membrane may be very high when an
expensive new polymeric membrane material is used. One approach to reduce the
cost of new membranes has been the development of TFC membranes that comprise
a thin, dense, nonporous, selective skin layer of a high-cost, high-performance poly-
mer deposited on an asymmetric integrally skinned, low-cost, porous, nonselective
support membrane [11]. The asymmetric integrally skinned, low-cost, porous,
nonselective support membranes used for the preparation of TFC membranes can
be formed from commercially available inexpensive polymers such as CA, CTA, a
blend of CA and CTA, polysulfone (PSf), PES, polyetherimide (PEI), polypropylene
(PP), or polyether ether ketone (PEEK). The thin, dense, nonporous, selective skin
layer on the TFC membranes can be prepared from relatively more expensive,
high-performance polymers such as polyimides, polymers with intrinsic microp-
orosities, and fluoropolymers. TFC polymeric membranes have the advantages of
low cost and high performance due to the use of inexpensive asymmetric porous
support membrane and the use of very small amount of expensive high-performance
polymer for the thin, dense, nonporous, selective skin layer. TFC polymeric mem-
brane also provides an option for new polymers that are difficult to be processed
as asymmetric integrally skinned membranes. TFC polymeric membranes with a
thin, dense, almost nonporous polyamide selective layer on top of an asymmetric
integrally skinned porous polymeric membrane are commercially available for RO
and NF applications. TFC polymeric membranes with a thin, dense, nonporous
selective layer have been used commercially for gas separations.
Most of the TFC polymeric membranes are bilayer membranes prepared from at
least two different polymers. For most of the bilayer TFC polymeric flat-sheet mem-
branes and some of the bilayer TFC polymeric hollow-fiber membranes, the thin,
212 9 Membranes
Thin selective layer

Gutter layer Asymmetric integrally
skinned porous layer
Asymmetric integrally
skinned porous Substrate
support membrane
Figure 9.3 Schematic illustration of trilayer TFC polymeric membrane.
dense, nonporous or almost nonporous, selective skin layer and the thick, porous
support layer are formed separately in two steps from different polymeric materials.
Similar to asymmetric integrally skinned membranes, the thin, dense, nonporous or
almost nonporous, selective skin layer is responsible for resistance and selectivity. It
is worth noting that the surface pore size control for the asymmetric porous support
membrane that will provide the TFC membrane mechanical strength and durability
will contribute to the TFC membrane separation performance even though the
asymmetric porous support membrane itself has no separation selectivity. The
two-step method for the preparation of bilayer TFC polymeric membranes typically
includes the fabrication of the asymmetric integrally skinned porous support poly-
meric membrane via the phase-inversion process followed by adding a thin selective
layer formed from a different polymer on the surface of the support membrane by a
coating method such as dip coating, spray coating, spin coating, laminating, inter-
facial polymerization, knife casting, or other methods. Some of the TFC polymeric
membranes have trilayers with an intermediate layer (or so-called gutter layer)
between the top thin selective skin layer and the bottom thick, porous, nonselective
support layer (Figure 9.3) to make a defect-free continuous selective skin layer [12].
Since the thin selective skin layer is formed separately from the thick, porous,
nonselective support layer in the TFC polymeric membranes, the thin selective
skin layer can be delaminated from the thick, porous, nonselective support layer in
some cases, which will result in significantly decreased selectivity for separations.
In addition, the porous, nonselective support layer needs to be insoluble in the
solvents that will be used for the preparation of the coating solution to form the
thin selective skin layer. Otherwise, the porous, nonselective support layer will
be damaged during the formation of the thin selective skin layer via the normal
coating process. To avoid these issues, TFC polymeric membranes have been
prepared via a one-step co-casting [13] or co-extruding [14] method in recent
years. TFC polymeric flat-sheet membranes can be fabricated by simultaneously
co-casting two different polymeric casting solutions containing the same polymer
or different polymers using two casting knives via a one-step phase-inversion
process. One solution at the bottom layer forms the low-cost, asymmetric, integrally
skinned, porous, nonselective bottom support layer and the other solution forms
another asymmetric integrally skinned high-performance selective layer on top
of the support layer. TFC polymeric hollow-fiber membranes can be fabricated
via a one-step co-extrusion phase-inversion process using two different polymeric
spinning solutions and a triple-annular spinneret. The core solution forms the
low-cost, asymmetric, integrally skinned, porous, nonselective core layer, and
the sheath solution forms the asymmetric integrally skinned high-performance
selective sheath layer. The bore fluid normally comprising an organic solvent and
water creates the center hollow of the hollow-fiber membrane.
Preparation of TFC polymeric membranes using one-step co-casting or
co-extruding approach allows the use of higher-cost, high-performance poly-
mer to form the asymmetric integrally skinned selective layer and the use of
inexpensive, commercially available polymer to form the asymmetric, integrally
skinned, porous, nonselective support layer without any potential delamination
issue and the issue of support membrane damage.
Similar to asymmetric integrally skinned polymeric membranes, TFC polymeric
membrane may also comprise a thin coating layer of a highly permeable polymer
such as silicone rubber or high-permeability fluoropolymer on the surface of the
thin, dense, selective skin layer of the membrane to plug the defects and further
improve the membrane selectivity for gas and vapor separation applications.
The development of new polymeric membranes from the starting polymers to
the large-scale commercial applications involves many critical steps. Not only poly-
meric materials, but also membrane engineering need to be studied in detail. The
selection or synthesis of the polymeric membrane material for a specific separa-
tion application is a key initial step. Membrane engineering, including membrane
fabrication, element/module preparation, process design, and simulation, is equally
important or even more important than membrane material study. Polymeric casting
or spinning solution study allows the successful conversion of a polymeric material
to a commercially viable membrane with either flat-sheet or hollow-fiber geome-
try and each of them can have an asymmetric integrally skinned or TFC structure.
The membrane element/module preparation involves not only advanced polymeric
membrane, but also unique design of the element/module such as permeate and feed
spacer for spiral wound element, as well as the glue that holds membrane leaves or
hollow fibers together to prevent leaks. To extract the best value from the membrane,
sometimes membrane post-treatment such as chemical or physical cross-linking is
used to further enhance the membrane performance. Membrane performance eval-
uation and technoeconomic analysis allow further membrane optimization. Overall,
polymeric material selection, design, and synthesis would be less than about 20% of
work in the whole membrane development process.
Consistent, long-term membrane research programs in academia and R&D
platforms in industry will continuously advance membrane technologies and lead
to breakthrough membranes for separations. These activities should include novel
membranes such as molecular design and synthesis of polymers, novel membrane
components, improved element/module materials, and application-guided concept
discovery, new membrane processes such as physical/computational modeling,
process intensification, invention of new membrane-based processes, as well as
advanced membrane characterization techniques.
9.2.2 Inorganic Membranes

Inorganic membranes such as microporous zeolite, metal–organic framework
(MOF), zeolitic imidazolate frameworks (ZIFs), crystalline organic framework
214 9 Membranes
(COF), carbon molecular sieve, ceramic, silica, metal, and stainless-steel membranes
have several advantages over polymeric membranes for a wide range of applications.
Inorganic membranes have high thermal stability, high chemical resistance and
resistance to harsh environments, high permeability, conductivity, and selectivity
for some applications. Therefore, inorganic membranes provide new separation
opportunities such as high-temperature catalytic membrane reactor applications.
Significant research was reported on four main types of inorganic membranes based
on their preparation methods in the early 1990s, including sol–gel membranes such
as alumina, titanium, zirconia, and silica membranes; chemical vapor deposition
(CVD) membranes such as silica, zirconia, and palladium membranes; pyrolysis
membranes such as carbon and carbon molecular sieve membranes; as well as
hydrothermal membranes, such as microporous crystalline zeolite membranes.
Many different microporous crystalline MOF and ZIF hydrothermal membranes
such as MOF-5, ZIF-8, and ZIF-67 have been studied in academia since the report
of the first MOF-5 membrane for gas separation in 2009 [15]. This chapter will focus
on the discussion of microporous crystalline zeolite, MOF, and ZIF membranes for
separations based on molecular sieving separation mechanism.
9.2.2.1 Introduction to Microporous Crystalline Inorganic Membranes

Microporous crystalline inorganic membranes, including traditional zeolitic
membranes such as NaA zeolite membrane, SAPO-34 molecular sieve membrane,
and the recent MOF and ZIF membranes are composite membranes with a
continuous, relatively thin, selective layer of microporous crystalline inorganic
material formed by in situ hydrothermal synthesis methods on the surface of a
highly porous, thick support membrane. A significant number of microporous
crystalline inorganic materials have been studied for the formation of the contin-
uous, relatively thin, selective layer of microporous crystalline inorganic material
in the microporous crystalline inorganic membranes since crystalline zeolite
membranes were first reported in late 1980s and early 1990s [16, 17]. For example,
aluminosilicates, titaniumsilicates, silicates, silicoaluminophosphates (SAPOs),
metalloaluminophosphates (MeAPOs), metallosilicoaluminophosphates (MeAP-
SOs), and their different ion-exchange forms such as NaA, MFI, SAPO-34, SSZ-13,
AlPO-18, DD3R, AlPO-17, SAPO-17, AlPO-5, T, NaY, ETS-4, mordenite, as well
as MOFs, ZIFs, and COFs such as MOF-5, ZIF-8, ZIF-67, ZIF-7, ZIF-95, MIL-53,
NH2 -MIL-53, HKUST-1, and Zn2 (bim)4 have been used for the preparation of the
microporous crystalline inorganic membranes. The highly porous, thick support
membranes used for the preparation of the microporous crystalline inorganic
membranes can be made from either polymeric or inorganic materials such as
poly(amide-imide) [18], nylon [19], polyvinylidene fluoride (PVDF) [20], stainless
steel [21], ceramic materials such as alumina [22], silica, titania, and glass [23]. The
highly porous, thick support membranes can have different geometries such as flat
sheet, hollow fiber, disk, tubular, or monolithic multiple channel.
Microporous crystalline inorganic membranes have the potential to overcome
some of the challenges facing polymer membranes for a wide range of separation
and catalysis applications because of their unique properties, such as uniform
micropores, tunable functionalities, superior thermal and chemical stability, good

erosion resistance, and high plasticization resistance. For example, SAPO-34
membrane has been synthesized and studied for high-temperature propane dehy-
drogenation membrane reactor application to significantly enhance the reaction
conversion and selectivity by continuously separating H2 during the reaction
process [24]. NaA zeolite membranes have been used commercially for solvent
dehydration applications [25].
The microporous crystalline inorganic membranes made from microporous
crystalline inorganic materials provide separation properties mainly based on
molecular sieving or competitive sorption mechanism. Competitive sorption–based
separation takes place for large-pore (≥6 Å) microporous crystalline inorganic
membranes when the pore size of the microporous crystalline inorganic material
in the continuous, relatively thin, selective layer is much larger than the molecules
to be separated. Molecular sieving will play the key role for small-pore (≤4 Å) and
some medium-pore (∼4–6 Å) microporous crystalline inorganic membranes when
the pore size of the microporous crystalline inorganic material is similar or larger
than the molecule to be separated in a mixture. Microporous crystalline inorganic
membranes need to have a continuous, very thin (<100-nm), grain boundary-free,
pinhole-free, and crack-free selective layer formed from a microporous crystalline
inorganic material for molecular sieving-based separations with high permeance
and high selectivity. However, some minor defects on the continuous, very thin
(<100-nm), selective layer formed from the small-pore microporous crystalline
inorganic material are acceptable for some applications such as the small-pore
microporous crystalline NaA zeolite membranes for solvent dehydration appli-
cations via a pervaporation process. The separation for this process is based on
preferential adsorption of water from the organic solvent/water mixtures rather
than on a real size-exclusion molecular sieving mechanism due to the hydrophilicity
of the high framework charge of the membrane; therefore, the membrane with a
defective selective layer still can have very high water/solvent selectivity.
9.2.2.2 Preparation of Microporous Crystalline Inorganic Membranes

For potential large-scale industrial separations, such as gas separations and solvent
dehydration, or catalysis applications, the microporous crystalline inorganic mem-
branes need to be cost effective, have high permeance, high selectivity, high-pressure
resistance for some applications, and high mechanical strength. Proper membrane
fabrication methods result in the formation of microporous crystalline inor-
ganic membranes with a continuous, very thin (<100 nm), grain boundary-free,
pinhole-free, and crack-free selective layer of a microporous crystalline inorganic
material on the surface of a highly porous, relatively smooth, high-surface-area,
nonselective support membrane. The highly porous support membrane provides
the membrane high enough mechanical strength. The continuous, very thin
(<100 nm), grain boundary-free, pinhole-free, and crack-free selective layer, having
only intracrystalline micropores and intercrystalline gaps with size smaller than
the intracrystalline micropores, provides the membrane high selectivity.
216 9 Membranes
Different from asymmetric integrally skinned polymeric membranes, microporous

crystalline inorganic membranes cannot be prepared directly from microporous
crystalline inorganic material to form asymmetric integrally skinned membrane
structure. A general method for the preparation of the microporous crystalline inor-
ganic membranes is to synthesize a continuous, relatively thin, defect-free, selective
layer of microporous crystalline inorganic material on the surface of a highly porous,
thick support membrane by an in situ hydrothermal synthesis method using either
an amorphous zeolite precursor sol or gel solution or homogeneous MOF or ZIF pre-
cursor solutions. Similar to the preparation of TFC polymeric membranes, the highly
porous, thick support membrane plays a critical role in the successful preparation of
high-performance microporous crystalline inorganic membranes. To make low-cost,
reproducible, high permeance, and high selectivity microporous crystalline inor-
ganic membranes with high mechanical strength, the porous support membrane
needs to have low cost, high surface area, high thermal and mechanical stability,
high chemical and hydrothermal stability, significantly higher permeance than
the final membrane (at least 10 times higher), uniform small pores on the surface,
high porosity with interconnected pore structure, thermal expansion coefficient
similar to that of the microporous crystalline inorganic materials, and high affinity
and compatibility with the selective layer of the microporous crystalline inorganic
material. Polymeric or inorganic hollow fiber and monolithic multiple channel
support membranes are more preferred geometries for the preparation of micro-
porous crystalline inorganic membranes to achieve high membrane area packing
density; therefore, smaller membrane area with reduced membrane, engineering,
and installation cost is required for large-scale membrane separation processes.
However, there have been numerous challenges for the preparation of
cost-effective, reproducible, high permeance, and high selectivity microporous
crystalline inorganic membranes using a cost-effective and environmentally benign
method. For example, the use of expensive porous support membrane, highly
basic, organic structure directing templates, and complicated synthesis procedure,
as well as the formation of large amount of synthesis waste such as the large
amount of water for membrane washing and poor reproducibility result in high-cost
microporous crystalline inorganic membranes. The formation of a very thick
selective layer of the microporous crystalline inorganic material, the low permeance
from the porous support membrane, and the poor orientation of the microporous
inorganic crystals provide the membrane with low permeance. The poor adhesion
and the thermal expansion coefficient difference between the selective layer of the
microporous crystalline inorganic material and the surface of the porous support
membrane, pinholes, grain boundary defects, intra- and intercrystalline cracks
result in the membrane with low selectivity.
Significant research efforts have been focused on the preparation of microporous
crystalline inorganic membranes with enhanced separation performance from
both academia and industry in the past two decades. High-quality NaA zeolite
membranes have been commercialized for alcohol dehydration application via a
microwave heating method, a secondary growth method, or a ball-milling-assisted
secondary growth method using high surface area monolith, porous, ceramic
support membranes [25]. Small-pore, high selectivity, microporous DD3R zeolitic

membrane was prepared using tubular or high surface area honeycomb-shaped,
porous, ceramic support membranes for CO2 /CH4 gas separation application
[26–28]. Small-pore, high selectivity, microporous CHA zeolitic membrane was
prepared using tubular, porous, ceramic support membrane for CO2 /CH4 and other
separation applications [29]. Medium-pore, microporous MFI zeolitic membranes
were synthesized on a porous, ceramic disk support membrane by microwave
heating–assisted secondary growth method without using organic structure
directing agents [30].
Several strategies have been evaluated to reduce the cost of the microporous crys-
talline inorganic membranes such as the use of a cheap, porous support membrane
or the use of a low-cost and simple membrane fabrication method that can avoid the
formation of extra microporous crystalline inorganic powders and the use of a large
amount of water for membrane washing. As an example, low-cost porous polymeric
support membranes, such as poly(amide-imide) [18], nylon [19], and PVDF [20],
were investigated to replace the expensive inorganic ceramic support membrane to
reduce the cost of the microporous crystalline inorganic membranes. Synthesis of
the low-cost, continuous, very thin (<100 nm), grain boundary-free, pinhole-free,
and crack-free selective layer of the microporous crystalline inorganic material
using simplified synthesis procedures is another approach to reduce the total cost of
the membrane. Structure-directing template-free synthesis [31], gel-free secondary
growth [32], dry-gel synthesis [33], and counter diffusion in situ synthesis [34]
have been reported for the preparation microporous crystalline zeolite membranes.
Structure-directing template-free synthesis method not only reduces the cost of
the membrane, but also allows the formation of a continuous, crack-free selective
layer of the microporous crystalline inorganic material without high-temperature
calcination step for template removal, which results in high membrane selectivity.
Controlling the orientation of the zeolite crystals via oriented secondary growth
[35–38], improving the adhesion between the selective layer and the surface of the
porous support membrane via counter diffusion method [39], synthesizing grain
boundary, pinhole, and crack-free continuous selective layer of the microporous
crystalline inorganic material via rapid thermal processing [40, 41], or using
high-aspect-ratio microporous crystalline inorganic material [42] has provided high
selectivity microporous crystalline inorganic membranes. Controlling the orienta-
tion of the zeolite crystals via oriented secondary growth to form vertically oriented
microporous zeolite channels with respect to the porous support membrane not only
improves the membrane selectivity, but also increases the membrane permeance.
In addition, the approach of using ultra-thin microporous crystalline inor-
ganic nanosheets to reduce the thickness of the continuous selective layer of the
microporous crystalline inorganic material and therefore enhance the membrane
permeance has attracted significant attention in most recent years. As an example,
ultrathin MFI nanosheets of about 3–100 nm in thickness have been used for the
preparation of thin MFI nanosheet membranes with a selective MFI layer of about
250 nm to 1 μm in thickness using different membrane fabrication methods such
as vacuum filtration deposition of thin sub-100 nm MFI nanosheet suspension
218 9 Membranes
on porous inorganic or polymeric support membrane [43, 44], gel-free secondary

growth using exfoliated MFI nanosheets of 3.2 nm in thickness as the seeds [32], or
direct synthesis of MFI nanosheet layer [45]. As another example, ultra-thin, highly
crystalline poly[Zn2 (benzimidazole)4 ] (Zn2 (bim)4 ) nanosheets of 1 nm in thickness
and large lateral area have been used to fabricate Zn2 (bim)4 nanosheet membrane
with a selective Zn2 (bim)4 nanosheet layer of several nm in thickness coated on
α-alumina porous support membrane via a hot-drop coating process [46]. However,
large-scale production of the cost-effective, reproducible microporous crystalline
inorganic membranes with a continuous, thin, defect-free, selective layer of the
microporous crystalline inorganic nanosheets is still very challenging. Synthesis of
the microporous crystalline inorganic nanosheets or the multistep membrane prepa-
ration procedures need to be significantly simplified. The selective layer thickness
of the microporous crystalline inorganic nanosheet layer still needs to be reduced
to provide high enough membrane permeance. The formation of the nonselective
defects on the continuous, thin, selective layer of the microporous crystalline inor-
ganic nanosheets needs to be avoided to achieve high selectivity for the microporous
crystalline inorganic nanosheet membranes. Very recently, ultra-thin ZIF-8 mem-
brane with a continuous, very thin, selective ZIF-8 layer of about 17 nm in thickness
deposited on low-cost, porous PVDF polymeric hollow-fiber support membrane
was fabricated via a gel-vapor deposition method [20]. The membrane showed
high permeances and high selectivities for gas separation applications, as well as
scaling-up potential due to the use of polymeric hollow-fiber support membrane.
Microporous crystalline inorganic membranes with uniform micropores and tun-
able functionalities have the potential to achieve “perfect” molecular sieving-based
separations featuring super high selectivity and permeance if a very thin (<100 nm),
grain boundary-free, pinhole-free, and crack-free selective layer of microporous crys-
talline inorganic material is formed. Overall, significant advancement in improv-
ing the quality, stability, permeance, and selectivity, as well as reducing the cost of
the microporous crystalline inorganic zeolite, MOF, and ZIF membranes for gas,
vapor, and liquid separations has been made from both academia and industry. How-
ever, the commercialization of microporous crystalline inorganic membranes for
large-scale industrial application has not been very successful except the success-
ful development of commercial-scale A-type of zeolite membranes for organic sol-
vent dehydration applications [47]. Further advancement in reducing the cost and
selective layer defects and thickness, as well as improving the processability and
reproducibility is still necessary for microporous crystalline inorganic membranes to
replace polymeric membranes for some commercial applications and to be adapted
for new commercial applications that are unsuitable for polymeric membranes.
9.2.3 Mixed-matrix Membranes

Both polymeric and inorganic membranes have their unique advantages and some
limitations. Up to now, most of the large-scale commercial membranes for gas and
liquid separations are polymeric membranes due to their relatively low cost, ease
of processability, good mechanical stability, and reasonably good selectivity and
permeance. Most of the polymeric membranes, however, have poor contaminant or

fouling resistance, low chemical and thermal stability, and low selectivity for some
applications. Inorganic membranes such as the microporous crystalline inorganic
membranes overcome these limitations of polymeric membrane and have shown
much higher selectivities than polymeric membranes for some gas separations
and are encouraging for some high-temperature membrane applications. However,
they are often much more expensive and more difficult to be fabricated than the
commercial polymeric membranes with the current state-of-the-art membrane
manufacturing processes, have low area packing density and poor mechanical
stability. Breakthrough development of disruptive membranes with the advantages
of both polymeric and inorganic membranes will create many new separation
opportunities for membrane technologies.
9.2.3.1 Introduction to Mixed-matrix Membranes

MMMs are composite membranes comprising a continuous organic polymeric
matrix, dispersed inorganic materials such as microporous zeolites, MOFs, ZIFs,
carbon molecular sieves, silica or alumina particles, activated carbons, graphene
oxides, and sometimes other liquid additives such as ionic liquids or salts. MMMs
are developed to overcome the limitations of polymeric membranes and inorganic
membranes and combine the advanced features of both membranes with the
potential of being fabricated using polymeric membrane manufacturing process.
MMMs have been studied for many potential applications such as gas, vapor, and
liquid separations, as well as fuel cell, battery, and catalysis applications.
MMMs comprising dispersed microporous crystalline inorganic materials such as
zeolite, MOF, or ZIF materials in a continuous polymeric matrix have been studied
most extensively in academia and industry. The separation mechanism of microp-
orous crystalline inorganic material/polymer MMMs is based on the combination of
the solution-diffusion separation of polymeric membrane with the molecular sieving
separation of microporous crystalline inorganic membrane. Therefore, MMMs have
the potential to achieve improved selectivity and enhanced permeance compared to
polymeric membranes when microporous crystalline inorganic materials with both
higher selectivity and permeability than those of the polymeric membrane materials
are incorporated into the MMMs. As illustrated in Figure 9.4, when using a MMM
comprising high selectivity and high-permeability microporous crystalline inor-
ganic materials dispersed in a continuous polymer matrix with inferior selectivity
and permeability to separate a given mixture of gases, vapors, or liquids comprising a
more permeable component A and a less permeable component B, the permeability
of component A in the MMM (abbreviated as PA, MMM ) is a combination of the per-
meability of A in the polymer matrix (abbreviated as PA, p ) and the permeability of A
in the dispersed microporous crystalline inorganic materials (abbreviated as PA, i ).
PA, MMM is greater than PA, p because PA, i is greater than PA, p . The less permeable
component B has larger size than the micropore size of the dispersed microporous
crystalline inorganic materials. Therefore, component B cannot permeate directly
through the micropores of the microporous crystalline inorganic materials and
needs to permeate through the polymer matrix via a much longer distance than the
220 9 Membranes
B A
Polymer matrix
MMM
Microporous crystalline inorganic material
Figure 9.4 Schematic illustration of mixed-matrix dense-film membrane for separating a

mixture of A and B.
polymeric material without dispersed microporous crystalline inorganic materials,

resulting in lower permeability of component B in the MMM (abbreviated as
PB, MMM ) than that in the polymer matrix PB, p . As a result, the selectivity of the MMM
for component A over component B (abbreviated as 𝛼 A/B, MMM ) is higher than that of
the polymeric membrane made from the same polymeric matrix material (abbrevi-
ated as 𝛼 A/B, p ). Ideally, the improvement of 𝛼 A/B, MMM and PA, MMM is determined by
the amount of high selectivity and high-permeability microporous crystalline inor-
ganic materials dispersed in the continuous polymer matrix and can be predicted
by using a Maxwell model [48]. Maxwell model has been used for the proper selec-
tion of the polymeric matrix material and the dispersed microporous crystalline
inorganic materials with known separation properties for a desired separation.
Some MMMs with good separation property and chemical property match
between the dispersed microporous crystalline inorganic materials and the con-
tinuous polymer matrix have exhibited separation properties superior to the
pristine polymeric membranes. Some of the MMMs have shown gas separation
performance well above Robeson’s polymer upper-bound limit. As an example,
UOP’s membrane research group has studied AlPO-14/polymer MMMs for
CO2 /CH4 separation [49, 50]. As shown in Figure 9.5, the CO2 /CH4 separation
performances of AlPO-14/P-1, AlPO-14/P-2, AlPO-14/P-3 dense-film MMMs,
comprising dispersed small-pore microporous AlPO-14 molecular sieve particles
and P-1, P-2, and P-3 continuous glassy polymer matrices, respectively, are different
due to the use of three polymers with different CO2 /CH4 separation properties
even though the same AlPO-14 dispersed particles are used. AlPO-14/P-1 MMM
with certain AlPO-14 loading has shown performance exceeded Robeson polymer
upper-bound limit for CO2 /CH4 separation. Most recently, rare-earth face-centered
cubic MOF/polyimide and ZIF-71/polyimide mixed-matrix dense films have been
prepared for CO2 /H2 S/CH4 and H2 /CO2 separations, respectively [51, 52].
However, some of the MMMs reported in the literature have shown nonideal
MMM performance due to the separation property mismatch or poor adhesion
between the polymeric and inorganic phases, poor dispersion of the inorganic
materials in the polymer matrix, or not small enough inorganic particles. Signifi-
cant research studies have been focused on mixed-matrix dense-film membranes.
Preparation of commercially viable asymmetric MMMs with integrally skinned or
TFC structure has been lack of investigation.
Robeson’s 1991
polymer upper bound
AlPO-14/P MMMs
𝛼CO2/CH4
P-3
P-2 P-1
PCO
2
Figure 9.5 CO2 /CH4 separation performance of AlPO-14/P-1, AlPO-14/P-2, AlPO-14/P-3

dense-film MMMs.
9.2.3.2 Preparation of Mixed-matrix Membranes

A variety of rubbery and glassy polymers have been studied as the continuous poly-
mer matrix and many different microporous crystalline inorganic materials have
been used as the dispersed phase for the preparation of MMMs. MMMs can have dif-
ferent structures such as mixed-matrix dense films, asymmetric integrally skinned
MMMs, and TFC MMMs. Asymmetric integrally skinned and TFC MMMs can have
either flat-sheet or hollow-fiber geometry.
As with polymeric dense films, mixed-matrix dense film has a symmetric structure
with a thickness of about 50–100 μm. Mixed-matrix dense films are not commer-
cially attractive for most of the applications, but are useful for measuring the intrin-
sic separation properties such as selectivity, permeability, and conductivity of the
membrane. Most of the mixed-matrix dense films are prepared from a mixed-matrix
casting dope comprising inorganic particles, polymer, and organic solvents via a
solution casting method. Preparation of mixed-matrix dense films with a thickness of
about 50–100 μm is much easier than those of asymmetric integrally skinned MMMs
and TFC MMMs. Different membrane preparation approaches have been reported to
solve the issues of agglomeration of the inorganic particles in the polymer matrix and
poor compatibility and adhesion at the interface of the inorganic particles and the
polymer matrix, which will result in MMMs with poor mechanical and processing
properties and poor separation performance. As an example, surface modification of
the inorganic particles to improve the interfacial adhesion between the two different
materials has been studied to improve the separation performance of mixed-matrix
dense films [53–55].
Mixed-matrix dense films need to be converted to either asymmetric integrally
skinned MMMs or TFC MMMs for large-scale gas, vapor, or liquid separations.
Same as asymmetric polymeric membranes, asymmetric integrally skinned MMMs
and TFC MMMs have a thin selective layer on a highly porous nonselective support
layer as shown in Figure 9.6. Most of the fabrication methods for asymmetric
MMMs are similar to those for asymmetric polymeric membranes. Both the thin
selective layer and the highly porous nonselective support layer are formed from the
222 9 Membranes
<100nm Figure 9.6 Schematic illustration of

asymmetric mixed-matrix membrane.
~100 μm
same mixed-matrix material comprising dispersed inorganic particles and the con-
tinuous polymer matrix for the asymmetric integrally skinned MMMs (Figure 9.7).
TFC MMMs comprise a thin mixed-matrix selective layer on a highly porous
nonselective support layer, which can be formed from a polymeric or inorganic
material different from the mixed-matrix material in the selective layer. For asym-
metric MMMs to achieve improved selectivity and permeance compared to the
asymmetric polymeric membranes, the formation of a thin (<∼100 nm), defect-free
mixed-matrix selective layer is critical.
Most of the asymmetric MMMs that have been reported so far are asymmetric inte-
grally skinned MMMs with either flat-sheet or hollow-fiber geometry prepared from
mixed-matrix casting or spinning dopes via the phase-inversion process [56–59]. The
compositions of the mixed-matrix casting or spinning dopes are different from those
for mixed-matrix dense films and normally comprise several different organic sol-
vents and nonsolvents as well as some additives to control the mixed-matrix selective
layer thickness and the morphology of the MMMs. Preparation of a stable, uni-
form mixed-matrix casting or spinning dope with the inorganic particles uniformly
dispersed in the polymer solution without particle agglomeration is critical for the
formation of a defect-free mixed-matrix selective layer, which will provide the MMM
with enhanced selectivity. Some of the asymmetric MMMs are TFC MMMs prepared
33065-414(2)
UOP HR-SEM COMPO 10.0kV ×850 10µm WD 8.1mm
Figure 9.7 Cross-sectional HRSEM image of an asymmetric mixed-matrix flat-sheet

membrane. (Photo provided courtesy of Honeywell UOP with a current copyright.)
9.3 Membrane Modules 223
by dip coating a mixed-matrix solution [60] or interfacial polymerization [61] to form

a thin mixed-matrix selective layer on top of a highly porous support membrane.
Different from thick mixed-matrix dense films that can be prepared relatively
easily and normally exhibit improved selectivity and permeability compared to
the pristine polymeric dense films, almost none of the asymmetric MMMs have
shown significantly enhanced selectivity and permeance or have performance
reproducibility issue compared to the pristine asymmetric polymeric membranes.
One difficulty is the synthesis of nanosized inorganic particles for the preparation
of asymmetric MMMs. To make asymmetric MMMs with much better separation
performance than asymmetric polymeric membranes, the thickness of the selective
mixed-matrix layer needs to be comparable to that of the selective polymeric layer,
which normally is ∼100 nm or less. The use of large inorganic particles with the
particle size in several micrometers range for the preparation of asymmetric MMMs
with a selective layer of ∼100 nm in thickness results in poor adhesion between the
inorganic particles and the polymer matrix and poor mechanical strength that can
easily form defects. In addition, the improper selection of the dispersed inorganic
material and the continuous polymer material, which results in permeability and
selectivity mismatch between the two materials, as well as the formation of interface
defects during the phase-inversion process also make defective asymmetric MMMs
lack of performance enhancement.
In summary, although MMMs have been extensively studied and significantly
advanced for over 30 years, very few MMMs have been commercialized for
large-scale industrial applications. Invention of new MMM fabrication techniques
to improve the dispersion of the inorganic particles in the polymer matrix, to
enhance the interaction, compatibility, and adhesion between the dispersed inor-
ganic materials and the polymer matrix, as well as the discovery of new synthesis
methods for inorganic materials particularly the microporous crystalline inorganic
material to produce nanosized inorganic particles without agglomeration, or the
discovery of new MMM fabrication methods that either allow the use of large
inorganic particles or form the inorganic materials in situ during the membrane
fabrication process will promote the commercialization of MMMs to advance the
membrane-based technologies. As an example, a flexible ZIF-8/PES MMM has been
prepared most recently using a new preparation method that allows the formation
of the microporous ZIF-8 material after the formation of the asymmetric ZnO/PES
composite membrane [62]. The growth of the ZIF-8 particles of <5 μm in diameter
is directed by the ZnO seed nanoparticles and restricted to the PES pores generated
during the phase-inversion process. The thickness of the effective mixed-matrix
selective layer and the defects in the selective layer, however, still need to be reduced
to make the membrane commercially viable.
9.3 Membrane Modules

Large-scale industrial membrane separation processes require large membrane
areas up to about 240 000 m2 to meet the unit treating capacity and product recovery.
224 9 Membranes
Residual
Feed spacer
Membrane Residual
Permeate spacer
Membrane
Feed spacer
Figure 9.8 Diagram of spiral wound membrane module.
These large membrane areas need to be packaged in cost-effective module config-

uration with high area packing density. The most common module configurations
for polymeric membranes and potentially for MMMs are spiral wound, plate and
frame, and hollow fibers. Asymmetric flat-sheet membranes can be incorporated
into spiral wound or plate and frame modules and asymmetric hollow fibers can
be formed into hollow-fiber modules. The common module configurations for
inorganic membranes include tube, hollow fiber, and multichannel monolith.
The areas of the asymmetric polymeric membranes in each spiral wound mod-
ule with a module diameter of 0.05–0.5 m can be in a range of 10–40 m2 . Spiral
wound membrane modules include membrane leaves, plastic sheets, a permeate
tube, and glue to hold the leaves together and each membrane leaf includes the
asymmetric flat-sheet membrane, permeate spacer, and feed spacer (Figure 9.8).
The performance of the spiral wound module is not only determined by the
advanced performance of the asymmetric flat-sheet membrane, but also the unique
design of feed and permeate spacer that can provide low-pressure drop, as well as
the design of the glue that is used to hold the membrane leaves together and prevent
module leaks. Spiral wound modules with the features of cross-flow, high-pressure
tolerance, high fouling resistance, high reliability, and ease of installation into
space-efficient, skid-mounted units have been used for large-scale RO, NF, UF, and
gas separation applications.
Asymmetric polymeric hollow-fiber membranes are freestanding membranes
and typically have an outside diameter of 100–1000 μm and a wall thickness of
50–200 μm. The areas of the asymmetric polymeric hollow-fiber membranes in
each hollow-fiber module can be in a range of 100–300 m2 . In addition, hollow-fiber
modules include hollow-fiber membranes and glue to hold the fibers together to
prevent module leaks and no feed spacer, permeate spacer, and permeate tube are
needed. Therefore, hollow-fiber module with the same diameter and length as those
of the spiral wound module can have much larger membrane areas than those in
the spiral wound module. Compared to spiral wound module, hollow-fiber module
has higher membrane area packing density, but has lower feed-pressure tolerance,
higher pressure drop, and more difficulty to repair defects. Hollow-fiber modules
9.4 Membrane Reactors 225
have been used commercially for RO, NF, UF, MF, and gas separations. There are
two types of hollow-fiber modules for these applications, including shell-side feed
modules and bore-side feed modules. Shell-side feed modules are generally used for
high-pressure applications up to about 1500 psig feed pressure such as for CO2 /CH4
separation and H2 purification with counter-current flow or cross-flow. Fouling on
the feed side of the membrane can be a problem with this kind of module design,
and pretreatment of the feed stream to remove particulates is required. Bore-side
feed modules are generally used for low- and medium-pressure feed streams up to
about 200 psig with counter-current flow, such as separation of N2 or water from
air, where good flow control to minimize fouling and concentration polarization on
the feed side of the membrane is desired.
Multichannel monolith module is the most preferred module configuration
for high-cost inorganic membranes such as the commercial ceramic membranes
and NaA zeolite membranes to achieve high membrane area packing density and
reduce membrane system cost for large-scale membrane separation processes with
high feed treating capacity and high product purity and recovery. The areas of the
inorganic membrane in each multichannel monolith module are about 10–15 m2 .
Several multichannel monolith zeolite membrane modules have been reported in
the literature [63, 64]. NaA zeolite membranes with multichannel monolith module
design have been used commercially in isopropanol and bioethanol dehydration
plants in China and India [25, 65].
9.4 Membrane Reactors

Membranes have been used for a wide range of gas, vapor, and liquid separations.
Membrane separation also has been integrated with a chemical conversion or
reaction in two-unit operations and membrane separation process is applied either
before or after the reaction process. Membrane reactor concept of combining a
chemical conversion process with a membrane separation process in one-unit oper-
ation was originally conceived by Alan Michaels 50 years ago [66]. The membrane
can have different functions in the membrane reactors, including retention of the
catalyst, selective or nonselective addition of the reactants, selective or nonselective
removal of the reaction products, or acting as part of the catalyst or catalyst support.
The advantages of the membrane reactor over the reaction/membrane separation
two-unit operations are integration of reaction and partial separation into one-unit
operation that lowers the Capital expenditure (CapEX), lower energy consumption
that lowers the Operating expenses (OPEX), enhancement of thermodynamically
limited or product-inhibited reactions that increases product yield or reaction rate,
decreases side reactions, or lowers the operating temperature and pressure to avoid
coking and reduce catalyst fouling, controlled reaction rates for selective reactions
that increases selectivity, as well as mediation and control of the contact between
two reactants. The industrial applications of membrane reactors, however, are still
facing some challenges due to the technical complexity of the process, difficulty in
optimization and design of a suitable reactor, poor sealing of the membrane reac-
tor chamber, high cost of the membrane and membrane module, limited thermal,
226 9 Membranes
chemical, and mechanical resistance of the membrane, poisoning or fouling of the

membrane, as well as low permeance of the membrane.
Both polymeric membranes and inorganic membranes such as metallic and
ceramic membranes have been studied for membrane reactor applications. Despite
the advantages of polymeric membrane reactors such as lower cost, commercially
mature membrane module design and fabrication, and ease of incorporating homo-
geneous or heterogeneous catalyst as compared to inorganic membrane reactors,
they are only suitable for reactions at relatively low temperatures of less than about
100 ∘ C due to the limited thermal resistance of the polymeric membranes. Some rep-
resentative reactions for polymeric membrane reactors are selective hydrogenation
[67, 68], hydrogenation [69], esterification [70], oxidation [71], and alkylation [72].
Inorganic membrane reactors are suitable for a wide range of high-temperature
reactions due to their high thermal stability. Some representative reactions for
inorganic membrane reactors are dehydrogenation such as ethane dehydrogenation
using Pt-impregnated alumina membrane [73], propane dehydrogenation using
SAPO-34 zeolite membrane reactor [24], oxidative coupling of methane using either
porous [74] or dense membrane [75], oxidative dehydrogenation [76], hydrogena-
tion [77], steam reforming and methane oxidation [78]. Among these reactions,
dehydrogenation reactions for olefin production and steam reforming and methane
oxidation for syngas production have been most extensively studied.
For syngas production in a membrane reactor, a mixed ion–conducting membrane
based on a dense ceramic inorganic membrane with perovskite structure was used
for a four-step O2 separation process. The first step is the diffusion of O2 to the sur-
face of the feed side of the membrane. In the second step, O2 dissociates to oxygen
ions and consumes electrons on the surface of the membrane. The third step is the
diffusion of the oxygen ions from the feed side to the permeate side of the membrane
and electrons moving in the opposite direction. In the last step, oxygen ions recom-
bine or react with CH4 gas to produce syngas in the presence of a reforming catalyst
on the permeate side of the membrane. The driving force for oxygen transport is the
chemical activity or oxygen partial-pressure gradient from the membrane feed side
to permeate side. The commercialization of this process has not been successful due
to several technical challenges such as cost, membrane fabrication and performance,
process integration, and membrane reactor scale-up.
For the catalytic dehydrogenation of propane to propylene, UOP has a commercial
OleflexTM process with four reactors used in series. Selective removal of part of the
H2 from the reaction mixture using inorganic ceramic membrane modules has been
studied [79]. Some of the requirements for the inorganic ceramic membrane to be
commercially viable for this application are low cost, high-temperature stability up
to 650 ∘ C, high H2 over propane and propylene selectivity, and high H2 permeance
at very low pressure of about 10–25 psig.
Despite the potential advantages of membrane reactors as compared to con-
ventional reaction processes, significant advancements in the high temperature
and long-term stability, new membrane module design that can provide high
membrane area packing density and good sealing at the operating temperature,
as well as large-scale membrane module fabrication feasibility are still required
for membrane reactors to find their large-scale industrial applications in the

petrochemical industry that requires high-temperature reaction processes.
9.5 Commercial Applications of Membranes
Membrane technologies have been used for a wide range of large-scale industrial
gas, vapor, liquid, and ion separation processes such as RO, natural gas upgrad-
ing, hydrogen recovery, organic vapor/gas and vapor/vapor separations, pervapora-
tion, NF, organic solvent NF, UF, MF, membrane contactor, electrodialysis, fuel cell,
and other electrochemical processing. Innovation in membrane materials, equip-
ment, and process technology integration has been instrumental in enabling the
widespread use of membranes in the chemical industry. Membrane technologies
have been widely adopted in semiconductor, automotive, water, food, pharmaceuti-
cal, biotechnology industries, and a wide range of environmental applications. Mem-
brane technology alone or its combination with other separation technologies in
integrated systems has been a successful approach for seawater desalination to meet
the freshwater demand in many regions of the world at lower cost and minimum
environmental impact. Worldwide sales of membrane products and systems have
seen a significant growth and increase in applications in the last two decades.
Reverse osmosis (RO): RO membrane technology was the first industrial-scale

commercial membrane technology and has been successfully used since 1960s
for brackish and seawater desalination and dairy processing. RO is one of
the fastest-growing membrane separation technologies. RO is a membrane
separation process based on solution-diffusion separation mechanism in which
properly pretreated seawater or brackish water is separated by the RO membrane
at moderate (200–400 psig for low-salt brackish water) to high (800–1000 psig
for high-salt seawater) pressures. The RO membrane rejects most solute ions
and molecules with high rejection rate, while allowing water to pass through
with high water flux to produce high-purity water up to 99.9% purity. Different
from gas separation and pervaporation membranes that have a thin, nonporous,
selective skin layer, RO membranes have a thin, relatively porous, selective skin
layer with ≤1 nm small pores. The first-generation RO membrane was CA asym-
metric integrally skinned membrane prepared from low-cost CA polymer based
on Loeb-Sourirajan phase-inversion process and the membrane was packaged
in spiral wound module for desalination process. The second-generation RO
membrane was cross-linked polyamide (PA) TFC flat-sheet membrane in spiral
wound modules and the thin selective layer of PA was formed by interfacial
polymerization on the surface of a highly porous support membrane. CA and
aromatic PA hollow-fiber RO membranes have also been developed in later years.
CA RO membranes have much higher chlorine and fouling tolerance than PA
RO membranes, but have lower solvent resistance and narrower pH range for
operation. The PA polymer on the polyamide TFC membrane has been modified
to further reduce the energy consumption and improve membrane stability and
228 9 Membranes
fouling resistance. Several advanced RO membranes have been commercialized

by Dow FILMTEC and Nitto Denko.
Gas separations: Gas separation membrane was first commercialized for large-scale
industrial hydrogen separation process by Permea (now part of Air Products) in
1980 and PRISM silicone rubber–coated polysulfone hollow-fiber membrane was
used in this process [80]. PRISM membrane has also been used for air separation.
Dried CA/CTA spiral wound membrane technology developed by Separex (now
part of Honeywell UOP) was first used commercially for acid gas removal from
natural gas in a small membrane unit in 1981. Cynara (now part of Schlumberger)
introduced the first dried CTA hollow-fiber membrane system also for acid gas
removal from natural gas in early 1980s. Generon produced the first N2 /air mem-
brane separation system in 1982. Since then, gas separation membranes have been
used for a wide variety of applications and N2 /air separation is the largest appli-
cation for gas separation membranes. Ube Industries commercialized polyimide
hollow-fiber membranes for nitrogen separation from air, H2 purification, and He
recovery. Air Liquide MEDAL commercialized polyimide and other hollow-fiber
membranes for air separation, H2 purification, and acid gas removal from nat-
ural gas. Membrane Technology and Research, Inc. (MTR) has several commer-
cial TFC spiral wound membranes for vapor/gas separation, N2 /CH4 , and other
applications. Typical polymers for the development of commercial gas separa-
tion membranes include glassy and rubbery polymers such as CA, CTA, polysul-
fone, polyimide, polyimide-aramide, polycarbonates, polyphenylene oxide, and
silicone rubber.
Large-scale H2 is produced by steam methane reforming. Refineries are
large-volume producers and consumers of H2 . The demand for H2 recovery
in refinery for existing and new hydrotreating and hydrocracking units has
grown dramatically. Refinery gases and various other hydrocarbon streams nor-
mally contain H2 , light hydrocarbons (C1 –C5 ), and light heteroatom-containing
gases. Improving the recovery of H2 from refinery gases could significantly
improve the economics of the H2 -valued refinery processes. The first large-scale
membrane system for H2 recovery is to recover H2 from ammonia purge stream
using PrismTM polysulfone hollow-fiber membrane. Since then, membrane tech-
nology has been extended to other H2 recovery applications such as adjustment
of H2 /CH4 or H2 /CO ratios for syngas production. UOP supplies PolysepTM
hollow-fiber membrane systems for H2 separation from a variety of feed sources.
H2 recovery from refinery streams to meet the increased demand of H2 is an
emerging field for gas separation membranes for hydrotreating, hydroprocessing,
or hydrodesulfurization processes.
Natural gas is expected to become the second-largest source of energy in the world
by 2035. With growing demand, the industry needs innovative solutions to effi-
ciently process gas reserves and capture valuable natural gas liquids. CO2 and H2 S
acid gases in raw natural gas, in combination with water, are highly corrosive and
can rapidly destroy natural gas pipelines and equipment. In addition, H2 S is toxic
and at relatively modest levels can be life threatening. CO2 reduces the heating
value of natural gas stream and wastes pipeline capacity. Amine absorption is the
widely used technology for CO2 and H2 S removal from natural gas. Amine plants,
however, suffer from problems such as high capital cost, complicated operation,
and expensive maintenance. Membrane separation is an alternative technology
for this natural gas upgrading application. Both spiral wound and hollow-fiber
polymeric membranes have been adopted commercially for this separation. Some
subquality natural gas reserves have high N2 content, which needs to be removed
to meet natural gas pipeline specifications. Cryogenic distillation and pressure
swing adsorption (PSA) technologies are currently used for this separation. MTR
has commercialized NitroSepTM membrane, which is more permeable to methane,
ethane, and other hydrocarbons than to N2 , for separating N2 from natural gas [7].
UOP’s Separex membrane technology has been an important enabler for effective
and efficient separation of contaminants such as CO2 and H2 S from natural gas
for onshore and offshore applications. Separex membrane technology was first
commercialized in small membrane units of <10 MMSCFD (million standard
cubic feet per day) treating capacities in early 1980s and it was used when
the remote locations of natural gas reserves required a simple and reliable
separation technology. Separex membrane systems became larger and larger with
the gas-treating capacity increased to 1000 MMSCFD in 2009 (Figure 9.9) and
have been extended from onshore applications to offshore platforms and most
recently offshore FPSO vessels that require lower-weight and smaller-footprint
membrane systems (Table 9.1). Figure 9.10 shows a single-lift Separex module
that offers small footprint and low weight for FPSO vessels. The simplest Separex
membrane-processing scheme is a one-stage flow scheme. For the one-stage flow
scheme (Figure 9.11), a natural gas feed is separated into a hydrocarbon-rich,
low CO2 and H2 S residual product stream and a CO2 and H2 S-rich permeate
stream. High hydrocarbon recovery (>95%) can be achieved using a two-stage
(Figure 9.12) or multistage flow scheme where the low-pressure, first-stage
permeate is compressed and processed in a second-stage membrane. For
example, a two-stage Separex membrane system was used for a large offshore
platform in Malaysia (Figure 9.13) to process 680 MMSCFD of natural gas. The
system reduced CO2 from 44 to 8 mol%. Another Separex membrane-processing
scheme is a two-step flow scheme (Figure 9.14) where the residue from the first
membrane is sent to the second membrane to further remove CO2 and H2 S and
the permeate from the second membrane is compressed and sent back to the
first membrane feed. The two-step membrane system has better hydrocarbon
recovery than that of the one-stage system, but is worse than that of the two-stage
system.
Separex commercial membrane systems include not only the membrane unit, but
also the membrane pretreatment unit. Proper membrane pretreatment design
is critical to the performance of the membrane systems. Many of Separex mem-
brane pretreatment systems include UOP MemGuardTM pretreatment system
(Figure 9.15), which is a regenerative system that can provide an absolute cutoff
of heavy hydrocarbons and can remove water, mercury, and other contaminants
along with the heavy hydrocarbons. MemGuard pretreatment systems have been
used for both onshore and offshore membrane systems. The invention of new
230 9 Membranes
1000
100
Feed flow (MMSCFD)
10
1 Avg size
0.1
1980 1985 1990 1995 2000 2005 2010
Year
Figure 9.9 Separex commercial membrane systems for natural gas upgrading with
significantly increased gas treating capacities from 1981 to 2009.
Table 9.1 Separex commercial membrane systems for offshore CO2 removal from natural
gas applications.
Treating capacity
Location (MMSCFD) CO2 removal (%) Stages
Thailand 32 50 to 20 1
Malaysia 680 44 to 8 2
FPSO Thailand 11 28 to 3 1
FPSO Brazil (14 units) 176–240 10–40 to 3–5 1–2
membrane materials and new membrane fabrication methods, as well as the

advances in membrane fabrication and engineering capabilities have enabled
significant improvement in the performance of Separex membranes, providing
reduced size and weight of the membrane system, as well as significantly
increased hydrocarbon recovery.
Vapor/gas and vapor/vapor separations: Extending gas separation membrane tech-
nology to large-volume organic vapor/gas and vapor/vapor separations such as
olefin recovery from polyolefin manufacturing plant and olefin/paraffin separa-
tions requires membranes with not only high selectivity and permeance, but also
high contaminant resistance and hydrocarbon plasticization resistance, as well
as long-term performance stability in the presence of organic vapors. Most of the
gas separation membranes are not suitable for these vapor/gas and vapor/vapor
separations. As with membranes for gas separations, membranes for vapor/gas
separations with a thin, nonporous, dense, selective layer on the surface of a
Figure 9.10 Single-lift Separex module for FPSO

applications. (Photo provided courtesy of Honeywell
UOP, with a current copyright.)
Figure 9.11 One-stage membrane Residue

flow scheme. (Low CO2, H2S)
Feed
Permeate
(CO2 and H2S enriched)
Figure 9.12 Two-stage Residue

membrane flow scheme.
Feed Membrane unit
Permeate
Membrane unit
Figure 9.13 Separex two-stage membrane

system in Malaysia. (Photo provided courtesy
of Honeywell UOP, with a current copyright.)
highly porous support layer are also based on solution-diffusion separation mech-
anism. Rubbery polymers such as silicone rubber with high solubility selectivity
for vapor over N2 or other gases have been used for the development of commer-
cial membranes for vapor/gas separations. Vapor/gas separation membranes are
commercially available from MTR, Dalian Eurofilm Industrial Ltd., and BORSIG
Membrane Technology GmbH.
232 9 Membranes
Residue
Feed
Permeate
Figure 9.14 Two-step membrane flow scheme.
Optional
Regenerable
Chiller and Coalescing
Feed MemGuardTM
separator filter
system
Particle Membrane
Heater
filter
Sales gas
Figure 9.15 Separex membrane system, including MemGuard system and membrane unit.
There is still no commercially available membrane for vapor/vapor separations such

as olefin/paraffin and paraffin isomer separations for petrochemical and refining
industries. Expensive and energy-intensive cryogenic distillation separation
technology has been used for many years for the separation of olefin and paraffin
components. Large capital expense and high energy cost have created incentives
for extensive research in developing low energy-intensive membrane technol-
ogy for this separation. Facilitated transport membranes (FTMs), polymeric
membranes, MMMs, and inorganic membranes have been studied in academia
and industry. FTMs, comprising metal ions as the complexing agent for olefins,
have very high olefin/paraffin separation selectivity and high potential for
commercialization. FTMs for olefin/paraffin separation are based on facilitated
transport mechanism. Olefins form reversible metal cation complexes such as
silver(I) complexes with the metal cations via the pi bonds on the surface of the
high-pressure feed side of the FTM, whereas no interaction occurs between the
metal cations and the paraffins (Figure 9.16). The reversible metal cation/olefin
complexes will transport through the selective layer of the FTM via complex
association/dissociation in the presence of water vapor and finally the olefins will
be dissociated from the reversible metal cation/olefin complexes and released at
Figure 9.16 Facilitated transport mechanism of High pressure R R

facilitated transport membrane (FTM) for
olefin/paraffin separation.
[Ag+] R
R
FTM
[Ag+]
R
[Ag+]
Ligh pressure R
the low-pressure permeate side of the FTM. FTMs have been prepared in lab scale
and some of them have been evaluated in demo unit for propylene/propane sep-
aration [81–83]. Membrane performance stability needs to be further improved
and the issue of complexing agent poisoning by contaminants needs to be solved
for commercial applications of FTMs for olefin/paraffin separations.
Pervaporation: Pervaporation is a membrane separation process to separate liquid
mixtures that combines evaporation of the volatile liquid component with
permeation of the vapor-phase component through a pervaporation membrane.
Vacuum is normally applied to the permeate side of the membrane to facilitate
the separation and recover the permeate. Both polymeric and inorganic mem-
branes have been used for commercial-scale pervaporation processes to separate
azeotropic mixtures such as alcohol and organic solvent dehydration, but sales
of pervaporation systems are much lower than other membrane separation pro-
cesses. Pervaporation using polymeric membranes is based on solution-diffusion
separation mechanism same as that for polymeric gas separation membranes.
As with polymeric gas separation membranes, polymeric pervaporation mem-
branes prepared from polymers such as polyvinyl alcohol or silicone rubber
have a thin, nonporous, selective skin layer on the surface of a highly porous
support layer. Pervaporation using small-pore microporous inorganic zeolite
membranes such as NaA membrane is based on preferential adsorption and/or
molecular sieving separation mechanism and the membranes have very high
selectivity. Most of the pervaporation membranes are commercially available
from GKSS (polymeric membrane), MTR (polymeric membrane), Jiangsu Nine
Heaven (NaA), Mitsui-BNRI ZeoSepA (NaA), Mitsubishi KonkerTM (NaA), and
Mitsubishi ZEBREXTM (high Si/Al ratio CHA).
Nanofiltration (NF): NF membrane is a loose RO membrane with an average pore
size of 1–2 nm in the thin selective skin layer and has been widely used for water
softening and treatment of high color and high organic content feed water. NF
separation is based on sorption-sieving separation mechanism. NF is superior to
both RO and UF for the treatment of bleaching effluents from pulp and paper
manufacturing plants due to its higher water flux than RO and higher rejection of
bleaching agent than UF. NF membrane can be operated at lower pressure than
RO membrane and has high rejection for divalent ions (>95%) and organics, but
has lower rejection for NaCl (10–50%) than RO membrane. Most of the commer-
cially available NF membranes are PA TFC membranes offered by Dow FILMTEC,
hydranautics, and several other companies. OSNF membranes are also loose RO
type of membranes with an average pore size of 1–2 nm in the thin selective skin
234 9 Membranes
layer and have high resistance to organic solvents. The most common polymers
for OSNF membranes are polyimides. OSNF-based MAX-DEMAX process devel-
oped by W. R. Grace and ExxonMobil has been used commercially for crude oil
dewaxing. OSNF is still an emerging membrane technology and there have been
very limited large-scale commercial sales for OSNF systems.
Ultrafiltration (UF): UF membrane is a porous membrane with an average pore size
of 1–100 nm in the surface skin layer and has been widely used for surface water
treatment, RO pretreatment, colloidal silica and iron removal from water, color
reduction, oil/water separation, milk processing, fruit juice clarification, automo-
bile paint recovery, biomolecule separation, as well as therapeutic proteins, DNA,
and RNA purification and filtration. UF membrane separation is based on not
only the pore size of the membrane, but also the shape and size of the molecules
to be rejected, as well as the solute/membrane interactions. The most common
UF membranes are prepared from CA, PSf, PES, polyacrylonitrile (PAN), PVDF,
PEI, or ceramic inorganic materials. Most recently, several approaches have been
studied to enhance the flux, reduce fouling, and improve solvent resistance of UF
membranes such as polymer chain modification, use of solvent-resistant polymers
such as PEEK and PBI, addition of hydrophilic polymer, blending of two different
polymers, modification of membrane surface, and coating of ceramic membrane
with silica or zirconia.
Microfiltration (MF): MF membrane is another porous membrane with an average
pore size of 0.1–10 μm in the surface skin layer and has been widely used for bio-
engineering to remove microorganisms from the fermentation products or remove
yeast from alcoholic beverages, removing particulates, colloidal materials, and
complexed heavy metals from liquid suspensions, and cell harvesting. MF mem-
brane separation is more complicated than simple sieving due to the pore size
reduction by the particles or macromolecules to be separated. The most common
MF membranes are prepared from CTA, PSf, PVDF, PP, polyethylene (PE), poly-
carbonate, polyester, PI, polytetrafluoroethylene (PTFE), nylon 6, nylon 6,6, or
ceramic inorganic materials.
UF and MF membranes are commercially available from a significant number
of companies such as Hydranautics, Suez, Hyflux, Pall, Tianjin MOTIMO
Membrane Technology, Litree Purifying Technology, Beijing Scinor Membrane
Technology, Beijing OriginWater Technology, Mitsubishi, and Jiangsu Jiuwu
Hi-Tech.
Membrane contactor: Membrane contactor is a membrane-based device that is used
to achieve gas/liquid mass transfer by allowing a gaseous phase and a liquid
phase to come into direct contact with each other without dispersion of one phase
within another. In a membrane contactor, the membrane separation is integrated
with a conventional phase contacting operation such as absorption or extraction
to exploit the benefits of both technologies. The key features for membrane
contactors are nonsensitivity to flooding, channeling, or back-mixing, modular
design, much higher surface area per unit volume as compared to conventional
columns or reactors, and use of vacuum and inert gas to create the driving force
for mass transfer. Common membranes used for membrane contactors are UF
References 235
or MF hollow-fiber membranes prepared from hydrophobic PP, PVDF, or PTFE

polymers. Membrane contactors have been commercialized for the removal of
gases from water such as stripping dissolved oxygen from ultra-high-purity water
for semiconductor industry, CO2 removal in beer production, carbonation of
beverage, and blood oxygenation.
Ion-exchange membrane applications: Ion-exchange membranes (IEMs) are elec-
trically conductive, semipermeable membranes that are permeable to either
positively or negatively charged ions. IEMs are made of hydrophobic organic
polymeric or inorganic ceramic materials with charged ionic side groups and
movable counterions. IEMs can be classified into cation exchange membranes
with negatively charged fixed ions such as sulfonic or carboxyl groups and
anion exchange membranes with positively charged fixed ions such as quater-
nary ammonium groups. A most recent review discussed about the materials,
preparation, and applications of IEM in detail [84]. IEMs have been used in
electrodialysis, diffusion dialysis, continuous deionization, or bipolar membrane
electrodialysis in commercial scale for seawater desalination, industrial wastew-
ater treatment of highly scaling waters, food and beverage production, and other
industrial wastewaters. IEMs have also been studied for energy storage and
transformation device applications such as reverse electrodialysis, fuel cell, and
redox flow batteries. Some representative companies that offer commercial
IEMs for fuel cell or redox flow battery applications are Asahi Glass (Flemion),
Asahi Chemical Industry, Tokoyana Soda (NEOSEPTA-F), Kashima-Kita Electric
Power, Sumitomo, and Evonik.
Since the discovery of the phase-inversion process to make asymmetric integrally
skinned polymeric membrane by Loeb-Sourirajan, membrane technology has been
an important enabler for effective and efficient separations and is widely used in gas,
water, and chemical industries. Innovation in membrane materials with long-term
stability, chemical resistance, and specific to applications, fabrication methods, mod-
ule design, new process design, and process technology integration has been instru-
mental to the widespread use of membrane technology, and it will continue to enable
membrane growth for new applications such as low-energy intensive olefin/paraffin
separations, aromatic/nonaromatic separations, aromatic isomers separations, and
paraffin isomer separations.
References
1 (2012, January). Honeywell UOP technology is used to clean natural gas on

FPSO vessels. Membrane Technology: 5.
2 Loeb, S. and Sourirajan, S. (1964). High flow porous membranes for separating
water from saline solutions. US Patent 3,133,132. Issued: 12 May 1964.
3 Cadotte, J.E. (1981). Interfacially synthesized reverse osmosis membrane.
US Patent 4,277,344. Issued: 7 July 1981.
4 Tang, M., King, W.M., Wensley, C.G. (1989). Air dried cellulose acetate mem-
branes. US Patent 4,855,048. Issued: 8 August 1989.
236 9 Membranes
5 Chiou, J. (2002). Epoxysilicone coated membranes. US Patent 6368382. Issued: 9

April 2002.
6 Henis, J.M.S. (1994). Polymeric Gas Separation Membranes (eds. D.R. Paul and
Y.P. Yampolskii), 441. Boca Raton, FL: CRC Press.
7 Bernardo, P., Drioli, E., and Golemme, G. (2009). Membrane gas separation: a
review/state of the art. Industrial and Engineering Chemistry Research 48: 4638.
8 Matsuura, T. (1994). Synthetic Membranes and Membrane Separation Processes,
Chapter 2. Boca Raton, FL: CRC Press.
9 Robeson, L.M. (1991). Correlation of separation factor versus permeability for
polymeric membranes. Journal of Membrane Science 62: 165.
10 Robeson, L.M. (2008). The upper bound revisited. Journal of Membrane Science
320: 390.
11 Riley, R.L., Hightower, G.R., and Lyons, C.R. (1973). Thin-film composite mem-
brane for single-stage seawater desalination by reverse osmosis. Applied Polymer
Symposia 22: 255.
12 Chiou, J. (1994). Composite gas separation membrane having a gutter layer com-
prising a crosslinked polar phenyl-containing-organopolysiloxane and method for
making the same. US Patent 5,286,280. Issued: 15 February 1994.
13 Kools, W. (2001). Process of forming multilayered structures. WO 01/89673.
14 Yates, S. F., McGuirl, M. C., Tonev, T. G., et al. (2012). Photo-crosslinked gas
selective membranes as part of thin film composite hollow fiber membranes.
US Patent 8,337,598. Issued: 25 December 2012.
15 Liu, Y.Y., Ng, Z.F., Khan, E.A. et al. (2009). Synthesis of continuous MOF-5
membranes on porous alpha-alumina substrates. Microporous and Mesoporous
Materials 118: 296.
16 Suzuki, H. (1987). Composite membrane having a surface layer of an ultra-
thin film of cage-shaped zeolite and process for production thereof. US Patent
4,699,892. Issued: 13 October 1987.
17 Haag, W.O. and Tsikoyiannis, J.G. (1991). Membrane composed of a pure molec-
ular sieve. US Patent 5,019,263. Issued: 28 May 1991.
18 Brown, A.J., Brunelli, N.A., Eum, K. et al. (2014). Interfacial microfluidic pro-
cessing of metal-organic framework hollow fiber membranes. Science 345: 72.
19 Yao, J., Dong, D., Li, D. et al. (2011). Contra-diffusion synthesis of ZIF-8 films on
a polymer substrate. Chemical Communications 47: 2559.
20 Li, W., Su, P., Li, Z. et al. (2017). Ultrathin metal–organic framework membrane
production by gel–vapour deposition. Nature Communications 8: 406.
21 Martínez Galeano, Y., Cornaglia, L., and Tarditi, A.M. (2016). NaA zeolite mem-
branes synthesized on top of APTES-modified porous stainless steel substrates.
Journal of Membrane Science 12: 93.
22 Ma, J., Shao, J., Wang, Z., and Yan, Y. (2014). Preparation of zeolite NaA mem-
branes on macroporous alumina supports by secondary growth of gel layers.
Industrial and Engineering Chemistry Research 65: 6121.
23 Dong, W.Y. and Long, Y.C. (2000). Preparation of an MFI-type zeolite membrane
on a porous glass disc by substrate self-transformation. Chemical Communica-
tions 12: 1067.
References 237
24 Kim, S.J., Liu, Y., Moore, J.S. et al. (2016). Thin hydrogen-selective SAPO-34 zeo-
lite membranes for enhanced conversion and selectivity in propane dehydrogena-
tion membrane reactors. Chemistry of Materials 28: 4397.
25 Rangnekar, N., Mittal, N., Elyassi, B. et al. (2015). Zeolite membranes-a review
and comparison with MOFs. Chemical Society Reviews 44: 7128.
26 Nakayama, K., Suzuki, K., Yoshida, M., et al. (2006). Method for preparing DDR
type zeolite membrane, DDR type zeolite membrane, and composite DDR type
zeolite membrane, and method for preparation thereof. US Patent 7,014,680.
Issued: 21 March 2006.
27 Uchikawa, T., Yajima, K., Nonaka, H., and Tomita, T. (2013). Method for produc-
tion of DDR type zeolite membrane. US Patent 8,377,838. Issued: 19 February
2013.
28 Teranishi, M., Miyahara, M., Ichikawa, M., and Suzuki, H. (2016). Porous body
and honeycomb-shaped ceramic separation-membrane structure. US Patent
9,321,016. Issued: 26 April 2016.
29 Imasaka, S., Masaya, I., Asegawa, Y., et al. (2016). Zeolite membrane, production
method therefor, and separation method using same. WO Patent Application
2016/006564. Published on 14 January 2016.
30 Tang, Z., Kim, S.J., Gu, X., and Dong, J. (2009). Microwave synthesis of MFI-type
zeolite membranes by seeded secondary growth without the use of organic struc-
ture directing agents. Microporous and Mesoporous Materials 118: 224.
31 Shi, H. (2015). Organic template-free synthesis of SAPO-34 molecular sieve mem-
branes for CO2 –CH4 separation. RSC Advances 5: 38330.
32 Agrawal, K.V., Topuz, B., Pham, T.C.T. et al. (2015). Oriented MFI membranes by
gel-less secondary growth of sub-100 nm MFI-nanosheet seed layers. Advanced
Materials 27: 3243.
33 Truter, L.A., Ordomsky, V.V., Nijhuis, T.A., and Schouten, J.C. (2012).
Preparation of ZSM-5 zeolite coatings within capillary microchannels.
Journal of Materials Chemistry 22: 15976.
34 Kwon, H.T. and Jeong, H.K. (2013). In situ synthesis of thin zeolitic-imidazolate
framework ZIF-8 membranes exhibiting exceptionally high propylene/propane
separation J. American Chemical Society 135: 10763.
35 Bein, T. (1996). Synthesis and applications of molecular sieve layers and mem-
branes. Chemistry of Materials 8: 1636.
36 Lai, Z., Bonilla, G., Diaz, I. et al. (2003). Microstructural optimization of a zeolite
membrane for organic vapor separation. Science 300: 456.
37 Pham, T.C.T., Nguyen, T.H., and Yoon, K.B. (2013). Gel-free secondary growth of
uniformly oriented silica MFI zeolite films and application for xylene separation.
Angewandte Chemie, International Edition 52: 8693.
38 Pham, T.C.T., Kim, H.S., and Yoon, K.B. (2011). Growth of uniformly oriented
silica MFI and BEA zeolite films on substrates. Science 334: 1533.
39 Kwon, H.T., Jeong, H.K., Lee, A.S. et al. (2015). Heteroepitaxially grown zeolitic
imidazolate framework membranes with unprecedented propylene/propane
separation performances. Journal of the American Chemical Society 137: 12304.
238 9 Membranes
40 Choi, J., Jeong, H.K., Snyder, M.A. et al. (2009). Grain boundary elimination in a
zeolite membrane by rapid thermal processing. Science 325: 590.
41 Schillo, M.C., Park, I.S., Chiu, W.V., and Verweij, H. (2010). Rapid thermal
processing of inorganic membranes. Journal of Membrane Science 362: 127.
42 Huang, Y., Wang, L., Song, Z. et al. (2015). Growth of high-quality,
thickness-reduced zeolite membranes towards N2 /CH4 separation using
high-aspect-ratio seeds. Angewandte Chemie, International Edition 54: 10843.
43 Varoon, K., Zhang, X., Elyassi, B. et al. (2011). Dispersible exfoliated zeolite
nanosheets and their application as a selective membrane. Science 334: 72.
44 Zhang, H., Xiao, Q., Guo, X. et al. (2016). Open-pore two-dimensional MFI zeo-
lite nanosheets for the fabrication of hydrocarbon-isomer-selective membranes
on porous polymer supports. Angewandte Chemie, International Edition 55: 7184.
45 Jeon, M.Y., Kim, D., Kumar, P. et al. (2017). Ultra-selective high-flux membranes
from directly synthesized zeolite nanosheets. Nature 543: 690.
46 Peng, Y., Li, Y., Ban, Y. et al. (2014). Metal-organic framework nanosheets as
building blocks for molecular sieving membranes. Science 346: 1356.
47 Morigami, Y., Kondo, M., Abe, J. et al. (2001). The first large-scale pervaporation
plant using tubular-type module with zeolite NaA membrane. Separation and
Purification Technology 25: 251.
48 Mahajan, R. and Koros, W.J. (2000). Factors controlling successful formation
of mixed-matrix gas separation materials. Industrial and Engineering Chemistry
Research 39: 2692.
49 Liu, C., Chiou, J.J., Wilson, S.T. (2009). Cross-linkable and cross-linked mixed
matrix membranes and methods of making the same. US Patent 7,485,173.
Issued: 3 February 2009.
50 Liu, C., Tang, M. W., Wilson, S. T., Lesch, D. A. (2010). Method of making
high performance mixed matrix membranes using suspensions containing poly-
mers and polymers stabilized molecular sieves. US Patent 7,815,712. Issued: 19
October 2010.
51 Liu, G., Chernikova, V., Liu, Y. et al. (2018). Mixed matrix formulations with
MOF molecular sieving for key energy-intensive separations. Nature Materials 17:
283.
52 Japip, S., Liao, K.S., and Chung, T.S. (2017). Molecularly tuned free volume of
vapor cross-linked 6FDA-durene/ZIF-71 MMMs for H2 /CO2 separation at 150 ∘ C.
Advanced Materials 29: 1603833.
53 Miller, S.J., Kuperman, A., Vu, D.Q. (2007). Mixed matrix membranes with small
pore molecular sieves and methods for making and using these membranes.
US Patent 7,166,146 B2. Issued: 23 January 2007.
54 Kulkarni, S.S., David, H.J., Corbin, D.R., Patel, A.N. (2003). Gas separation
membranes with organosilicone-treated molecular sieve. US Patent 6,508,860.
Issued: 21 January 2003.
55 Marand, E., Pechar, T.W., Tsapatsis, M. (2006). Mixed matrix membranes.
US Patent 7,109,140 B2. Issued: 19 September 2006.
References 239
56 Kulkarni, S.S. and Hasse, D.J. (2005). Novel polyimide based mixed matrix
membranes. US Patent Application 2005/0268782 A1. Published on 8 December
2005.
57 Ekiner, O.M. and Kulkarni, S.S. (2003). Process for making mixed matrix hollow
fiber membranes for gas separation. US Patent 6,663,805. Issued: 16 December
2003.
58 Jiang, L.Y., Chung, T.S., and Kulprathipanja, S. (2006). An investigation to
revitalize the separation performance of hollow fibers with a thin mixed matrix
composite skin for gas separation. Journal of Membrane Science 276: 113.
59 Ismail, A.F., Kusworo, T.D., and Mustafa, A. (2008). Enhanced gas permeation
performance of polyethersulfone mixed matrix hollow fiber membranes using
novel Dynasylan Ameo silane agent. Journal of Membrane Science 319: 306.
60 Jia, M., Peinemann, K.V., and Behling, R.D. (1992). Preparation and character-
ization of thin-film zeolite-PDMS composite membranes. Journal of Membrane
Science 73: 119.
61 Jeong, B.-H., Hoeka, E.M.V., Yan, Y. et al. (2007). Interfacial polymerization
of thin film nanocomposites: a new concept for reverse osmosis membranes.
Journal of Membrane Science 294: 1.
62 Hess, S.C., Grass, R.N., and Stark, W.J. (2016). MOF channels within porous
polymer film: flexible, self-Supporting ZIF-8 poly(ether sulfone) composite mem-
brane. Chemistry of Materials 28: 7638.
63 Liu, W., Post, P., Williams, J. L., et al. (2007). Multi-channel cross-flow porous
device. US Patent 7,169,213. Issued: 30 January 2007.
64 Fekety, C. R., Kinney, L. D., Liu, W., Song, Z. (2009). Zeolite membrane
structures and methods of making zeolite membrane structures. WO Patent
Application 2009005745A1. Issued: 8 January 2009.
65 Caro, J. and Noack, M. (2009). Zeolite membranes – status and prospective.
In: Advances in Nanoporous Materials, vol. 1 (ed. S. Ernst), 1. Elsevier.
66 Michaels, A.S. (1968). Separation techniques for the CPI (chemical process
industry). Chemical Engineering Progress 64: 31.
67 Liu, C., Xu, Y., Liao, S. et al. (1997). Selective hydrogenation of propadiene and
propyne in propene with catalytic polymeric hollow-fiber reactor. Journal of
Membrane Science 137: 139.
68 Liu, C., Xu, Y., Liao, S., and Yu, D. (1998). Mono- and bimetallic catalytic hollow
fiber reactors for the selective hydrogenation of butadiene in 1-butene. Applied
Catalysis A: General 172: 23.
69 Brandao, L., Fritsch, D., Mendes, A.M., and Madeira, L.M. (2007). Propylene
hydrogenation in a continuous polymeric catalytic membrane reactor. Industrial
and Engineering Chemistry Research 46: 5278.
70 Kita, H., Tanaka, K., Okamoto, K.I., and Yamamoto, M. (1987). The Esterifica-
tion of oleic acid with ethanol accompanied by membrane separation. Chemistry
Letters 16: 2053.
71 Solovieva, A.B., Belkina, N.V., and Vorobiev, A.V. (1996). Catalytic process of
alcohol oxidation with target product pervaporation. Journal of Membrane Sci-
ence 110: 253.
240 9 Membranes
72 Binning, R.C. and Kelly, J.T. (1959). Hydrocarbon conversion with dialytic sep-
aration of the catalyst from the hydrocarbon products. US Patent 2,913,507.
Issued: 17 November 1959.
73 Champagnie, A.M., Tsotsis, T.T., Minet, R.G., and Webster, A.I. (1990). A high
temperature catalytic membrane reactor for ethane dehydrogenation. Chemical
Engineering Science 45: 2423.
74 Lafarga, D., Santamaria, J., and Menendez, M. (1994). Methane oxidative
coupling using porous ceramic membrane reactors - I. reactor development.
Chemical Engineering Science 49: 2005.
75 Lu, Y., Dixon, A.G., Moser, W.R. et al. (2000). Oxidative coupling of methane
using oxygen-permeable dense membrane reactors. Catalysis Today 56: 297.
76 Capinelli, G., Carosini, E., Cavani, F. et al. (1996). Comparison of the catalytic
performance of V2 O5 /γ-Al2 O3 in the oxidehydrogenation of propane to propy-
lene in different reactor configurations: (i) packed-bed reactor, (ii) monolith-like
reactor and (iii) catalytic membrane reactor. Chemical Engineering Science 51:
1817.
77 Torres, M., Sanchez, J., Dalmon, J.A. et al. (1994). Modeling and simulation
of a three-phase catalytic membrane reactor for nitrobenzene hydrogenation.
78 Udovich, C.A. (1998). Ceramic membrane reactor for the conversion of natu-
ral gas to syngas. In: Natural Gas Conversion V, Studies in Surface Science and
Catalysis, vol. 119 (eds. A. Parmaliana, D. Sanfilippo, F. Frusteri, et al.), 417.
79 van Veen, H.M., Bracht, M., Hamoen, E., and Alderliesten, P.T. (1996). Funda-
mentals of Inorganic Membrane Science and Technology (eds. A.J. Burggraaf and
L. Cot), 641.
80 Baker, R.W. (2002). Future directions of membrane gas separation technology.
81 Pinnau, I., Tpy, L.G., and Casillas, C. (1997). Olefin separation membrane and
process. US Patent 5,670,051. Issued: 23 September 1997.
82 Herrera, P.S., Feng, X., Payzant, J.D., and Kim, J.-H. (2008). Process for the sepa-
ration of olefins from paraffins using membranes. US Patent 7,361,800. Issued: 22
April 2008.
83 Feiring, A.E., Laareri, J., Majumdar, S., and Shsnggusn N. (2018). Membrane
separation of olefin and paraffin mixtures. US Patent 10,029,248. Issued: 24 July
2018.
84 Ran, J., Wu, L., He, Y. et al. (2017). Ion exchange membranes: new developments
and applications. Journal of Membrane Science 522: 267.
241
10
Absorption
10.1 Introduction to Absorption

Absorption is a method of separation commonly used in the oil and gas and
chemical industries for removal of acid contaminants such as H2 S (hydrogen
sulfide) and CO2 (carbon dioxide) from hydrocarbons; this separation strategy is
often employed in the purification, transport, and storage of natural gas. There are
two primary classes of absorption: chemical absorption and physical absorption.
Chemical absorption involves a liquid absorbent, typically an aqueous solution of
amines, reacting with an acidic, non-hydrocarbon contaminant (e.g. CO2 or H2 S).
In physical absorption, there is no reaction between the liquid absorbent and CO2 ,
and the method of desorption tends to be through pressure or temperature swing
depending on the absorption liquid employed [1]. Through the chemical or physical
removal of H2 S and CO2 , natural gas producers are able to better protect their
pipelines and production infrastructure from damage due to corrosion.
With a stronger focus over time on reducing the CO2 emissions associated with
energy generation and usage, natural gas has risen as a percentage of overall energy
consumption. Table 10.1 compares the CO2 emissions (normalized to natural gas)
for a variety of hydrocarbon fuels. While coal emits nearly double the level of CO2
per unit energy produced, gasoline, heating oil, and diesel are also significantly less
efficient than natural gas from a CO2 emissions’ perspective. Traditionally, natu-
ral gas extraction was achieved through the drilling of vertical wells into natural
gas–bearing formations. In conventional natural gas deposits, the hydrocarbon gas
can flow easily up through the wells to the surface.
In North America, natural gas now can be produced efficiently from shale and
other types of sedimentary rock formations by forcing water, chemicals, and sand
(or other particles) down the well under high pressure. This process, called hydraulic
fracturing, or fracking, breaks up the rock formation, releases the natural gas from
the rock, and allows the natural gas to flow upward to the surface. At the top of the
well on the surface, natural gas can then be transported via pipelines and sent to
natural gas–processing plants.
The combination of low CO2 emissions relative to other hydrocarbons as well as
advances in extraction technologies such as hydraulic fracturing have led to a rapid
growth in the importance of natural gas as a source of energy both in North America
242 10 Absorption
Table 10.1 Normalized CO2 emissions per unit of energy for

various hydrocarbon fuels.
Fuel CO2 emissions
Coal (anthracite) 1.95

Coal (bituminous) 1.76
Coal (lignite) 1.84
Coal (subbituminous) 1.83
Diesel fuel and heating oil 1.38
Gasoline (without ethanol) 1.34
Propane 1.19
Natural gas 1.00
Source: US Energy Information Administration [2].
35
30
25
20
15
10
0
1950 1960 1970 1980 1990 2000 2010 2020
Figure 10.1 US Natural Gas Consumption (1950–2018) in trillion cubic feet. Source: US
Energy Information Administration [3].
as well as around the world (Figure 10.1). Natural gas production has risen steadily
over the past 50 years. As a result, the importance of absorption methods to the oil
and gas and chemical industries has been, likewise, increasing steadily. Proven natu-
ral gas reserves are around 190 trillion m3 globally, but total reserves are much larger
than this figure [4]. The Middle East and Russia hold the largest total natural gas
reserves, while the largest current national producer and consumer of natural gas is
the United States.
The main constituent of natural gas is methane (Table 10.2), but it can contain a
range of other higher hydrocarbons. While these hydrocarbons represent the main
attraction of natural gas extraction due to their energy value through combustion,
there also are many non-hydrocarbon materials present in natural gas such as H2 S
10.2 Acid Gas Removal 243
Table 10.2 Composition of natural gas (US Energy

Information Administration).
Component Molar fraction
Methane (CH4 ) 0.75–0.99

Ethane (C2 H6 ) 0.01–0.15
Propane (C3 H8 ) 0.01–0.10
Butane (C4 H10 ) 0.00–0.02
Nitrogen (N2 ) 0.00–0.15
Carbon dioxide (CO2 ) 0.00–0.30
Hydrogen sulfide (H2 S) 0.00–0.30
Helium (He) 0.00–0.05
Source: Hammer et al. [5]. © 2006, John Wiley & Sons.
and CO2 , which can interfere with the transport and storage of natural gas and also
can damage expensive energy infrastructure if not properly treated via methods such
as chemical and physical absorption. In addition, the presence of CO2 can lead to a
reduction in energy value of the natural gas as there is no energy value unlocked
with the attempted combustion of this contaminant.
Given its importance in the energy market, the application of gas separation meth-
ods to the removal of acidic contaminants such as H2 S and CO2 from natural gas will
be the focus of this chapter. It is important to note that many of these same chemi-
cal and physical absorption methods also are effective in the treatment of synthesis
gas (CO + H2 ), as both H2 S and CO2 require sweetening for the same reasons as
natural gas.
Gas absorption can be described as the removal of one or more pollutants from a
contaminated gas stream via intimate contact of that contaminated gas with a liq-
uid (absorber) that enables the pollutants to dissolve within the absorption liquid.
The principal factors dictating absorption performance are the solubility of the con-
taminants in the absorbing liquid as well as the ability of the absorption units to
efficiently mix gas and absorption liquid through the creation of a large amount of
interfacial surface area.
10.2 Acid Gas Removal

As described previously, natural gas has been growing in importance over the last
few decades and brings with it concerns regarding how to properly treat natural gas
for transport and storage. H2 S and CO2 are often found in natural gas streams and
bring specific risks and challenges. H2 S (hydrogen sulfide) is a highly toxic gas that
is corrosive to carbon steels and, if allowed to remain in natural gas, will oxidize into
SO2 , which, when emitted, can harm the environment. CO2 is, likewise, corrosive
to equipment and reduces the energy value of gas when combusted as part of the
natural gas stream.
244 10 Absorption
The corrosion chemistry of H2 S and CO2 is described in the equations below. Mild
grades of steel (e.g. carbon steel) often are pitted through the action of acid gas con-
taminants. Removal of these materials via chemical or physical absorption helps
preserve energy infrastructure, including wells and pipelines. In addition to H2 S and
CO2 levels, corrosion also is influenced by temperature, water chemistry (esp. pH),
flow velocity, oil or water wetting, and composition and the surface condition of the
steel.
CO2 corrosion reactions:
CO2 (g) → CO2 (aq) (10.1)
CO2 (aq) + H2 O → H2 CO3 (10.2)
H2 CO3 → H+ + HCO3 − (10.3)
HCO3 − → H+ + CO3 2− (10.4)
Fe0 (s) + CO2 + H2 O → FeCO3 + H2 (10.5)
H 2 S corrosion reactions:
H2 S(aq) ⇌ HS− (aq) + H+ (aq) (10.6)
HS− (aq) ⇌ H+ (aq) + S2 − (aq) (10.7)
Fe0 (s) + S2 − (aq) ⇌ FeS(s) (10.8)
Conventional processes for removing acid gases typically involve their countercur-
rent absorption from the contaminated natural gas or syngas stream using a regen-
erative solvent in an absorber column. This approach of contacting a gas stream to a
liquid absorber for the removal of acid gases is common across a wide range of pro-
cess industries, including natural gas production, refining, and chemical production.
Because each of these industries requires different degrees of acid gas removal, the
absorption solvent employed will vary significantly.
Shown in Table 10.3 are some common solvents (both chemical and physical)
employed in acid gas absorption/separation. The criteria for solvent selection in
chemical and physical absorption will be covered in more detail later in this chapter.
For now, please note that primary, secondary, and tertiary amines all find use in
chemical absorption processes. In addition, non-amine systems are used (e.g. potas-
sium carbonate in the removal of CO2 via the BenfieldTM process).
Prior to selecting either a chemical of physical absorption approach, the strengths
and weaknesses of each need to be considered. While the loading of a physical sol-
vent is proportional to the partial pressure of a component in natural gas, the loading
of a chemical solvent is determined by the stoichiometric reaction between a com-
ponent in natural gas and chemical solvent (i.e. it is a weak function of pressure).
At low partial pressure of the component (generally low concentration), a chemical
Table 10.3 Common solvents for acid gas absorption.
Chemical Physical
MEA (methanolamine) SelexolTM (dimethyl ethers of polyethylene glycol)

DGA (diglycolamine) SulfolaneTM (tetrahydrothiophenedioxide)
DEA (diethanolamine) Rectisol® Methanol
DIPA (diisopropylamine) Purisol® N-methyl-2-pyrrolidone
MDEA (methyldiethanolamine)
TEA (triethanolamine)
Potassium carbonate
Figure 10.2 Physical vs. chemical

absorption. 2
Solvent loading
1
Physical solvent
Chemical solvent
Partial pressure
absorption system generally is preferred (such as condition 1 in Figure 10.2). At high

partial pressure of the component to be removed (as in condition 2 in Figure 10.2),
physical solvent has an advantage due to its high loading capability.
In selecting a solvent, the primary consideration is establishing how low the levels
of H2 S and CO2 need to be, post absorption processing. A secondary consideration is
the impact of solvent selection on capital (i.e. equipment installation) and ongoing
operating expenses. Solvent selection will affect capital expenses for the absorption
unit and the design of the regeneration system. It is estimated that 50–70% of the
initial capital investment for an amine-based acid gas absorption unit is associated
with the magnitude of the solvent circulation rate, and 10–20% of the initial capex
depends on the regeneration energy required for the selected amine solvent system
[6]. In addition, about 70% of operating costs, excluding labor, result from the regen-
eration process.
10.3 Chemical Solvent
As demonstrated in 10.2, when impurity concentrations are low (e.g. 1–10% H2 S and
CO2 ), chemical absorption is favored over physical absorption. The ideal solvent
246 10 Absorption
Table 10.4 Qualitative characteristics of the amines for the removal of H2 S and CO2 .
Amine type Primary (MEA) Secondary (DEA) Tertiary (TEA, MDEA)
Enthalpy of reaction High Medium Low

Enthalpy of evaporation High Medium Low
Reaction rate High Medium Low
Corrosivity High Medium Low
Loading Low Medium High
Concentration (wt%) <32% 25–30% 35–55%
Source: Yildirim et al. [7]. © 2012, Elsevier.
for this process is capable of a fast, reversible chemical reaction with the gaseous
impurity. When impurity levels are even lower, solvents that react irreversibly are
employed, but such irreversible systems bring drawbacks such as solid waste treat-
ment and added cost. Chemical solvents include aqueous solutions of primary, sec-
ondary, and tertiary amines (see Table 10.4). We will separately consider inorganic,
alkaline salts such as potassium carbonate (K2 CO3 ) in our summary of the Benfield
process below.
Through acid-base reactions, aqueous solutions of these basic compounds capture
and remove acid gases and other impurities by forming weak chemical bonds with
dissolved acid gases (H2 S and CO2 ) in the absorber. The bonds are broken by heat in
the regenerator to release the acid gases and regenerate the solvent for reuse.
10.3.1 Amines Process

The first industrial use of amines for chemical absorption separation of acid gas from
hydrocarbon was patented by R.R. Bottoms of the Girdler Corporation, Louisville,
Kentucky, in 1930; claim 1 from this patent is included below [8]. As indicated in the
claim, the need for intimate mixing of liquid amine with hydrocarbon gas for effec-
tive absorption processes has been known since the beginning of industrial chemical
absorption.
1. The process of separating an acidic gas from gaseous mixtures, which includes
effecting intimate contact of a gaseous mixture with an absorbent agent in liquid
form including an amine selected from the group consisting of aliphatic and
cycloparaffin amines, and which is free from carboxyl or carbonyl groups, and
which has a boiling point not substantially below 100 ∘ C. – Bottoms [8]
The basic design of a chemical absorption unit is included in Figure 10.3. The flow
of gas upward and liquid amine downward encourages the sort of intimate mixing
required of a high efficiency chemical (or physical) absorption process. The packing
material within the column bed plays a role in creating the mixing needed to effi-
ciently remove acid gas impurities from natural gas and other hydrocarbon streams.
Liquid Gas
in out
Packed bed
Liquid Gas
out in
Figure 10.3 Packed

absorption column.
The first commercial chemical absorption systems used triethanolamine (TEA)

for acid gas removal. While TEA has been largely replaced due to its low capacity,
low reactivity, and poor stability, amines have remained the absorption solvent of
choice to the present day. The selection of an amine solvent is centered on four key
factors:
(1) Required purity of cleaned gas stream for further processing or end use
(2) Composition of feed gas
(3) Utility requirements, process costs
(4) Corrosion and solvent degradation/solvent losses.
A comparison of these selection criteria is included in Table 10.4. Choosing the

best amine solvent for a specific absorption end use will affect all aspects of perfor-
mance and payback, including absorption unit safety, capital requirements, unit life,
and operating cost. In many cases, simply changing the amine solvent can unlock
large gains in efficiency.
The concentration of amine in the solution is an important operating parame-
ter. High amine concentrations will reduce the amine solvent circulation rate and
thereby reduce the operating and capital cost of the gas plant. Two major factors
prevent high concentration of amine:
(1) Corrosion at high amine concentration.

(2) Temperature rise in the absorption column due to chemical reaction:
high-concentration amine (low circulation) will cause high-temperature
rise, which will reduce the equilibrium loading.
248 10 Absorption
We will now dive a bit deeper into the strengths of each main category of amine
solvent for chemical absorption.
10.3.1.1 Monoethanolamine (Primary Amine)

Monoethanolamine (MEA, H2 NCH2 CH2 OH) is a common solvent for chemical
absorption [9]. It is often used at a concentration of 10–20 wt% in water (i.e. below
32%). While MEA itself is not considered to be highly corrosive, degradation prod-
ucts of MEA are extremely corrosive. Specifically, MEA will react with oxidizing
agents such as COS, CS2 , SO2 , SO3 , and oxygen to form soluble products, which
must then be removed from the chemical absorption unit to avoid serious corrosion
damage. Degradation or deactivation of MEA also lowers the effective amine
concentration, but regeneration of the MEA is possible to avoid the expense and
complexity of having to add fresh MEA.
As a result, acid gas loading levels are typically limited to 0.3–0.35 mol acid gas
per mole of amine for carbon steel equipment. In stainless steel, it is possible to
increase loading levels up to 0.7–0.9 mol acid gas per mole of amine without harming
corrosion performance.
MEA is a primary amine with a high pH, which enables MEA solutions to produce
a sweetened gas product containing less than 4 ppm (by volume) of acid gas at very
low H2 S partial pressures. For MEA systems, nearly all the CO2 will be absorbed to
produce gas that meets the H2 S ppm specification. Because the enthalpy of reaction
for MEA with CO2 is high, a feed gas containing high concentrations of CO2 will
cause either extremely high reboiler duty or poor acid gas stripping. For all amines,
the enthalpy of reaction is a function of loading and other conditions.
10.3.1.2 Diethanolamine (Secondary Amine)

Diethanolamine (DEA, HN(CH2 CH2 OH)2 ) is often used in the 25–30 wt% range.
Like MEA, the total acid gas loading for DEA is also limited to 0.3–0.35 mol acid gas
per mole of amine for carbon steel surfaces. For stainless steel, the DEA loading can
be raised to 1 mol acid gas per mole of amine. The degradation products of DEA are
much less concerning than MEA degradation products. Unlike MEA, DEA is not
reclaimable in most units because it decomposes below its boiling point at ambient
pressure.
Since DEA is a secondary alkanolamine, its affinity for H2 S and CO2 is lower
and may not be capable of producing pipeline specification natural gas for some
low-pressure streams. As gas pressure is reduced, more stripping steam must be used
or a split-flow design needs to be employed. DEA is selective toward the removal of
H2 S and will permit a significant level of CO2 to remain in the product gas stream.
10.3.1.3 Methyldiethanolamine (Tertiary Amine)

Methyldiethanolamine (MDEA, H3 CN(CH2 CH2 OH)2 ) is used in the range of
35–55 wt%. Lower weight % solutions of MDEA are often used in low-pressure,
high-selectivity applications such as the cleanup of tail gas in Shell Claus off-gas
treatment (SCOT) units. Acid gas loadings as high as 0.7–0.8 mol acid gas per mole
of amine in carbon steel equipment are possible with MDEA due to their relatively
Table 10.5 Four typical AGFS (amine guard formulated solvent) flow schemes.
Reboiler duty
Desired product Typical (MBtu/lbmol
Scheme Feed gas quality gas quality application CO2 removed)
Flash only Very high acid gas >2% CO2 Pipeline NG 8–10
(>12%)
Conventional Low acid gas (<7%) 50 ppm CO2 LNG 45–60
1-stage High acid gas 50–1000 ppm CO2 LNG, NGL 32–40
(7–12%)
2-stage Very high acid gas 500 ppm CO2 Ammonia 12–18
(>12%)
low corrosivity. Oxidation of MDEA forms corrosive acids, which can result in the
buildup of iron sulfide in the absorption system. [10].
MDEA offers several unique advantages over primary and secondary amines,
including lower vapor pressure, lower heat of reaction, higher resistance to degra-
dation, lower corrosivity, and higher selectivity toward H2 S in the presence of CO2 ;
being a weaker base, MDEA reacts much more quickly with H2 S than CO2 . Its
lower heat of reaction allows MDEA to be used in pressure swing units for bulk
CO2 removal. In a pressure swing plant, the rich amine is flashed at or around
atmospheric pressure and little or no heat is required for stripping.
Once the correct amine for chemical absorption has been selected, the design of
the flow scheme must be considered. There are four basic flow scheme options for
an amine guard formulated solvent (AGFS) chemical absorption unit (Table 10.5).
The first scheme (flash) is a simple flow scheme and is relatively inexpensive, has
a low energy requirement, is ideal for bulk removal of CO2 and partial removal of
H2 S to generate pipeline quality natural gas. The second flow scheme (conventional,
depicted in Figure 10.4) is suitable for both LNG and natural gas liquids (NGL); it can
achieve CO2 levels below 50 ppm, H2 S levels below 4 ppm, and requires less solvent
than the flash system. The third option (1-stage) is capable of the same CO2 and H2 S
removal as a conventional unit but at lower energy usage. Finally, the 2-stage system
is suitable for LNG and ammonia and provides the flexibility to tradeoff solvent flow
rate for thermal regenerator duty (at much lower regeneration duty requirements
than the 1-stage system).
10.3.2 BenfieldTM Process

Originally developed at the U.S. Bureau of Mines in the early 1950s, the hot potas-
sium carbonate (HPC) process was modified and optimized by Benson and Field
[11, 12] and, subsequently, Honeywell UOP. The chemical solvents available for acid
gas removal in the 1950s were water, monoethanolamine (MEA) and other amines.
These solvents require operating temperatures below 125 ∘ F (50 ∘ C) to prevent sol-
vent degradation. Water could be used to remove CO2 , but it is very inefficient and
250 10 Absorption
Acid gas
Acid gas
Sweet KO drum
gas
Make-Up
Water
Acid gas
cooler
Amine Lean Lean
absorber solution solution Reflux
pump cooler pump
Flash gas
to fuel
header Filtration
Amine
system Make-Up
Feed Make-up stripper
Water
gas water
Rich
flash drum
Amine
roboiler
Lean/rich
Lean solution
exchanger
booster pump
Figure 10.4 Honeywell UOP conventional chemical solvent – amine process (suitable for
LNG and NGL).
does not provide acceptable treated gas purity. Caustic solutions, either NaOH or
KOH, could remove the acid gases very effectively, but could not be regenerated.
Using a solution of KOH to first pick up CO2 would generate potassium carbon-
ate in solution, which could absorb still more acid gases. Thus, the “hot potassium
carbonate” process was born.
The UOP Benfield Process is a thermally regenerated cyclical solvent process using
an activated, inhibited hot potassium carbonate solution to remove CO2 , H2 S, and
other acid gas components. More than 700 Benfield units have been put into com-
mercial service. In addition to its wide use in ammonia and hydrogen production,
the Benfield process has been applied in 50+ natural gas plants and in direct iron
ore reduction plants.
The absorption of CO2 into an aqueous potassium carbonate (K2 CO3 ) system
means there are two reaction paths available. Either hydroxide ions react with CO2
to directly form bicarbonate ions:
CO2 (aq) + OH− (aq) ⇌ HCO3 − (aq)
or water can hydrate CO2 to form carbonic acid, which subsequently deprotonates
to form bicarbonate:
CO2 (aq) + 2 H2 O(l) ⇌ HCO3 − (aq) + H3 O+ (aq)
The resulting drop in pH from either reaction is then buffered by the

carbonate-bicarbonate equilibrium, thus giving the overall reaction as:
CO2 (g) + CO3 2− (aq) + H2 O(l) ⇌ 2 HCO3 − (aq)

Acid gas
Purified gas
Feed gas
Absorber
Regeneration
2KHCO3 ⇔ K2CO3 + CO2 + H2O K2CO3 + CO2 + H2O ⇔ 2KHCO3
Figure 10.5 BenfieldTM process: chemical absorption based on K2 CO3 (conventional flow
scheme).
Potassium carbonate chemical absorption processes possess a number of advan-

tages over the amine-based absorption liquid processes, including the ability for
absorption to occur at higher temperatures, which makes the regeneration process
more efficient and economical [13]. Potassium carbonate also has a low cost, is less
toxic, nonvolatile, and less prone to degradation effects that plague amines at high
temperatures and in the presence of oxygen and other minor gas components such
as SOx and NOx .
A typical Benfield chemical solvent composition is:
Potassium carbonate (K2 CO3 ), in water (30%)

Activators – DEA and ACT-1
Corrosion inhibitor – V2 O5
Shown in Figure 10.5 is the conventional flow scheme for the Benfield process,
including the thermal regeneration system. With over 700 units in commercial
operation, the Benfield process offers efficiency coupled with low capital invest-
ment. This method of separation achieves very low CO2 concentrations (<50 ppm)
in treated gas as well as low H2 S levels (<4 ppm)
10.4 Physical Solvent
As shown in Figure 10.2, physical solvent absorption systems offer the advantage of
operating at high loading. Indeed, solvent loading increases linearly with the partial
pressure of the contaminant to be removed. Therefore, when the component to be
removed (e.g. CO2 or H2 S) is present at a high partial pressure, physical absorption
252 10 Absorption
systems are favored over chemical absorption systems. Additionally, contaminants

can often be removed from physical solvents simply by reducing pressure on the
physical solvent and with a minimum of heat applied. Finally, physical solvents
tend to be less corrosive than their chemical absorption amine-based counterparts.
Due to these advantages, physical solvents have grown in popularity for the removal
of contaminants in coal gasification applications as well as synthesis gas (CO + H2 )
treatment.
10.4.1 Solvent Selection

Solvent selection remains a key performance determinant for physical absorption
systems, and physical absorption system span a range of chemistries ranging from
propylene carbonate and glycol-based solvents to alcohols, alcohols, phosphates,
and pyrrolidones (Table 10.6). Cost and performance factors such as solvent volatil-
ity, contaminant solubility, and process conditions such as pressure and temperature
often guide the choice of solvent in physical absorption systems.
Table 10.7 shows the relative solubility of typical gas stream components in three
commercial solvents for physical absorption systems: SelexolTM (DMPEG), RectisolTM
(methanol), and Fluor (PC). The data have been normalized based on CO2 solubility.
The relative strength of each of these three physical absorption solvents is detailed
below. [15]
10.4.1.1 DMEPG (Dimethyl Ether of Polyethylene Glycol, (CH3 O(C2 H4 O)n CH3 ))
DMEPG is used in UOP’s Selexol process and is a mixture of dimethyl ethers of
polyethylene glycol where n falls in the range of two to nine. DMPEG is often used to
physically absorb H2 S, CO2 , and mercaptans from gas streams. As can be seen from
the data in Table 10.7, DMEPG can be used for selective H2 S removal, which requires
stripping, vacuum stripping, or a reboiler. Depending on downstream unit configu-
ration, the H2 S-rich feed can be sent onward to the Claus unit. It also is possible to use
Table 10.6 Physical solvent processes.
Process name Solvent Process licensor
SELEXOLTM Dimethyl either of polyethylene glycol UOP/DOW

(DMPEG)
Fluor solvent Propylene carbonate (PC) Fluor Daniel
Sepasolv MPE Methyl isopropyl ether of polyethylene BASF
glycol (MPE)
Purisol N-methyl-2-pyrrolidone (NMP) Lurgi
Rectisol Methanol Linde AG and Lurgi
Ifpexol Methanol IFP
Estasolvan Tributyl phosphate (TBP) IFP/Uhde
Methylcyanoacetate Methylcyanoacetate Unocal
Table 10.7 Relative solubility of chemical species at 25 ∘ C (CO2 = 1).
Selexol Rectisol Fluor (Propylene

Component (DMPEG) (Methanol) carbonate)
Carbon dioxide, CO2 1 1 1

Hydrogen sulfide, H2 S 8.82 3.29 7.06
Hydrogen, H2 0.013 0.0054 0.0078
Methane, CH4 0.066 0.051 0.038
Ethane, C2 H6 0.42 0.42 0.17
Propane, C3 H8 1.01 2.35 0.51
n-Butane, C4 H10 2.37 — 1.75
n-Pentane, C5 H12 5.46 — 5
n-Hexane, C6 H14 11 — 13.5
n-Heptane, C7 H16 23.7 — 29.2
Carbonyl sulfide, COS 2.3 7.06 3.29
Water, H2 O 730 — 300
Source: Stewart [14]. © 2014, Elsevier.
DMEPG for selective H2 S removal with deep CO2 removal typically accomplished
via a two-stage process with two absorption and regeneration columns. The first
column selectively removes H2 S, while CO2 is removed in the second absorber. The
second-stage solvent can be regenerated with air or nitrogen for deep CO2 removal,
or using a series of flashes if bulk CO2 removal is required.
Compared to the other two solvents (methanol and propylene carbonate), DMEPG
is of higher viscosity. This difference in rheology reduces mass transfer rates and tray
efficiencies and increases packing (or tray number) requirements. These differences
will become even more important when operating at reduced temperatures. DMEPG
service temperatures range from −18 ∘ C up to 175 ∘ C.
10.4.1.2 MeOH (Methanol, H3 COH)

Methanol is used in the Rectisol process from Linde and was the first commercial
physical absorption process based on an organic solvent. Methanol often is used in
synthesis gas applications (especially purification of syngas derived from the gasifi-
cation of heavy oil and coal). As the vapor pressure of methanol is relatively high,
low-temperature operations that require refrigeration are employed to prevent high
solvent losses. The need for refrigeration tends to drive the cost of operating a Rec-
tisol unit higher, and the solvent flow rate often is reduced to counteract this cost
disadvantage. Water washing of effluent streams often is used to recover the physical
solvent in methanol absorption systems. Methanol service temperatures in Rectisol
processes range from −62 ∘ C up to −40 ∘ C.
Like DMEPG, methanol offers a high degree of selectivity for H2 S over CO2 ,
combined with the ability to remove COS. While the need for methanol refrigeration
can bring added cost, acid gas solubility in physical solvents increases significantly
254 10 Absorption
as the temperature decreases, so low-temperature operation may be of interest.

Low temperature also reduces solvent losses by lowering the vapor pressure of the
methanol in the product streams. On balance, if H2 S is to be removed from a gas
with CO2 remaining in the treated gas, Selexol tends to be a preferred option versus
Methanol/Rectisol.
10.4.1.3 PC (Propylene Carbonate, C4 H6 O3 )

The Fluor solvent process uses propylene carbonate, a cyclic carbonate ester, since
the late 1950s. Propylene carbonate is favored when H2 S levels are low and CO2
removal is the focus of the physical absorption taking place. As seen in Table 10.7, the
solubility of light hydrocarbons (methane through propane) is lower than DMEPG
and methanol; this lower solubility results in lower recycle gas compression require-
ments and lower hydrocarbon loss in the CO2 vent gas stream.
The vapor pressure of propylene carbonate is higher than DMEPG; however, sol-
vent losses tend to be low. DMEPG is, by contrast, more selective than PC for H2 S
removal from gases containing CO2. PC cannot be used for selective H2 S treating
because high temperatures degrade the solvent. For this reason, Selexol is preferred
for selective H2 S removal. The operating temperature for PC ranges from −18 ∘ C up
to 65 ∘ C.
10.4.2 Flow Schemes

There are basically three sections to physical absorption unit:
(1) Absorber: High pressure and/or low temperatures

(2) Regeneration unit: Low pressure and/or higher temperature stripper
(3) Intermediate section: for heat exchange and solvent concentration. This section is
tailored to achieve a certain acid gas quality, suitable for the sulfur recovery unit.
For each project, a specific design is required to achieve the desired product qual-
ity and the optimum acid gas quality at the same time. There are three general
flow schemes for physical absorption, which differ primarily in their regeneration
approach (Table 10.8):
A flash regeneration system involved the enriched solvent being regenerated solely
through pressure let-down; there is no thermal regeneration involved. This method
is best suited for bulk CO2 removal to meet pipeline specs (<3% CO2 ) and is not
applicable for meeting low CO2 (e.g. treated gas) specifications or for high sulfur
Table 10.8 Three typical physical absorption flow schemes.
Scheme Desired product gas quality Typical application
Flash Regen <3% CO2 Bulk CO2 removal

Thermal Regen <1 ppm S Sulfur removal only (no CO2 )
Flash and thermal <1% CO2 ; <1 ppm S Sulfur and CO2 removal
Treated
gas
CO2
absorber
CO2 Acid
gas
Stripper
Lean solution Reflex
filter accumulator
Makeup
Sulfur water
Reflex
absorber
H2S pump
Feed Export
concentrator
gas water
Stripper
Packinox reboiler
exchanger
Figure 10.6 Physical solvent process flow scheme 3: Thermal and Flash Regeneration.
(H2 S) feeds. Because heat energy is not used, flash regen is a low-energy approach
to physical absorption.
Thermal regeneration has been applied to sulfur removal (not CO2 ) and allows a
low (<1 ppm) sulfur specification to be realized. This method works best with high
thermal stability solvents like DMEPG; it is not used with propylene carbonate and
that solvent would degrade under high-temperature regeneration. The use of ther-
mal regen does, of course, add energy cost to the physical absorption process, so the
performance needs of a process need to be balanced with cost considerations.
The third physical absorption scheme includes both thermal and flash distilla-
tion and can accomplish sulfur and CO2 removal. While more complex, a primary
advantage of this approach is that the product acid gas levels can be adjusted to meet
specific customer requirements. This flow scheme is shown in Figure 10.6 and is rep-
resentative of the process design for Selexol physical absorption unit, which will now
be described in more detail.
10.4.3 SelexolTM Process

Selexol from UOP is a commercial physical absorption process consisting of a mix-
ture of dimethyl ethers of propylene glycol (DMEPG), which is selective for both
CO2 and H2 S and can also remove COS and RSH contaminants as well. The prod-
uct streams from Selexol can be made essentially sulfur free with CO2 levels tunable
by need. For this reason, a wide range of end uses have been identified for Selexol,
including natural gas treatment, synthesis gas purification, acid gas removal with
simultaneous dehydration, and landfill gas treating. Selexol is not well suited for
applications that involve low acid gas partial pressures (use chemical absorption
methods below 50 psi) and hydrocarbon-rich streams.
256 10 Absorption
Table 10.9 SelexolTM relative selectivity.
Methane Hydrogen
H2 S 135 680
CH3 SH 340 1700
CO2 9 45
CO 0.43 2.2
COS 35 175
Table 10.10 SelexolTM advantage summary.
Selexol characteristic Advantage
Low vapor Pr Low solvent loss

Non fouling Clean service
High on-stream efficiency High availability
Low CW requirements Lower OPEX
Low power requirements Lower OPEX
100% organic and non-toxic Safe operation
The selectivity profile of Selexol is summarized in Table 10.9. The high selectivity
of DMEPG of H2 S over CO2 allows efficiency recovery of CO2 in the flash drum (and
minimizes CO2 in acid gas) and saves energy and money in the solvent recovery
process. Likewise, the selectivity of Selexol for CO2 over H2 minimizes downstream
H2 purification requirements.
Table 10.10 summarizes a few of the key advantages of Selexol versus other phys-
ical absorption methods. While most of these advantages speak to lower operating
costs (e.g. less solvent loss, higher availability, and low energy/power requirements),
other advantages such as lower toxicity, lower corrosion (no formation of heat-stable
salts), minimal process effluent, and protection of downstream equipment through
the capture of metal carbonyls add to the benefits of Selexol versus other absorption
options.
Shown in Figure 10.6 is a standard flow scheme for a Selexol unit tasked with the
removal of H2 S, COS, CO2 , HCN, and metal carbonyls. Given the low corrosivity of
DMEPG, a Selexol unit is made mostly of carbon steel.
The first column (sulfur absorber) removes most of the H2 S (and a limited
amount of CO2 ) from the feed syngas, which then flows to the second column
(CO2 absorber), which removes most of the CO2 . The rich solvent leaving the
CO2 absorption column is flashed in drums, from which relatively pure CO2 is
recovered. The solvent in the sulfur absorber column must be stripped in a column
with a reboiler to remove the high H2 S content gases.
10.5 Additional Commercial Uses of Absorption 257
Table 10.11 SelexolTM commercial experience summary.
Feed gas composition, %
Applications Number of units CO2 H2 S
Natural gas
Bulk CO2 removal 12 28–65 0–0.01
Selective H2 S removal 12 2–90 0.7–1.0
Synthesis and fuel gases

Steam reforming 6 14–19 —
Partial oxidation 3 33–35 0.02–0.09
Coal gasification 12 33–42 0.7–1.0
Landfill gases
Bulk CO2 removal 9 ∼50 —
Table 10.11 provides some perspective on the number and variety of Selexol units
in operation. As can be seen, Selexol’s combination of selectivity, energy efficiency,
and low corrosivity allows it to be used across a range of gas treatment end uses.
10.5 Additional Commercial Uses of Absorption

10.5.1 Glycol Dehydration
In natural gas systems, water vapor must be removed to a low level to avoid corro-
sion problems and pipeline blockage concerns caused by the formation of hydrates.
To avoid free water from forming within the pipeline, the temperature cannot fall
below the water dewpoint. The acid gas removal methods described previously in
this chapter can result in water-saturated gas, which requires dehydration before
the treated natural gas can be placed into the pipeline. Product specifications
require the maximum quantity of water in the gas to be 4–7 lbm/MMcf. To meet this
specification, glycol dehydration using diethyleneglycol (DEG) or triethyleneglycol
(TEG) is used. [16]
Glycol dehydration systems are efficient in the removal of water from natural
gas streams, and are also effective in the removal of volatile organic compounds
(VOCs) such as benzene, toluene, ethylbenzene, and xylene (BTEX). The absorp-
tion solvents used in this process, DEG and TEG, have a chemical affinity for
water and remove water from the natural gas stream (i.e. DEG serves as a liquid
desiccant). The glycol solution is intimately mixed with the wet gas stream in
a glycol contactor (Figure 10.7). While in the contactor, the glycol solution will
absorb water from the wet gas. The heavier water-bearing glycol moves to the
bottom of the contactor where it is removed. The natural gas, having been stripped
258 10 Absorption
Water vapor
Dry gas out + stripping gas
Flash gas
Still
Glycol
contactor
Flash Reboiler
tank
Stripping
Separator
gas
Wet Surge drum
gas in
Water
Filter
Pump
Figure 10.7 Glycol dehydration unit flow scheme. Source: Bentley [17]. © 1991, American
Institute of Chemical Engineers.
of most of its water content, then is transported out of the top of the contactor
dehydrator.
The glycol solution now contains the water removed from the natural gas and is
processed in a reboiler designed to remove only the water from the solution. The
boiling-point differential between water (100 ∘ C) and glycol (204 ∘ C) makes it rel-
atively easy to remove water from the glycol solution, allowing it be reused in the
dehydration process.
In addition to water from the wet natural gas stream, the glycol solution may
carry with it small amounts of methane and other compounds found in the wet
gas. Previously, this methane was simply vented out of the boiler. In addition to
losing a portion of the natural gas that was extracted, this venting contributes to
greenhouse gas emissions. To decrease the amount of methane and other com-
pounds that are lost, flash tank separator-condensers are employed to remove these
compounds before the glycol solution reaches the reboiler. The flash tank separator
reduces the pressure of the glycol solution stream, which allows the methane and
other hydrocarbons to vaporize (flash). The glycol solution then travels to the
boiler, which may also be fitted with air or water-cooled condensers to capture
any remaining organic compounds in the glycol solution. In total, absorption
systems are capable of recovering 90–99% of the methane that would otherwise
be flared into the atmosphere. The regeneration (stripping) of the glycol is limited
by temperature as both DEG and TEG decompose at or below their respective
boiling points.
References 259
Figure 10.8 Air stripping of VOCs from

water. Air + VOC Water + VOC
Air Water
10.5.2 Organics from Air (Air Stripping)

Air-stripping columns commonly are used for the removal of VOCs from water
(Figure 10.8). By forcing air upward through a descending mist of contaminated
water that then flows downward through a packed-bed column, VOCs are evapo-
rated from the water as air passing through the column absorbs the contaminated
vapors. The air then can be removed for proper treatment of the VOC emissions
(e.g. carbon absorption or thermal oxidation).
The two most common types of air strippers are packed-columns and sieve-tray air
strippers, though diffused-aeration tanks often can act as an effective means of air
stripping as well. The design of air stripping columns is governed by differences in
volatility between the VOCs in the contaminated water. This method of VOC abate-
ment can achieve 99+% removal of VOCs. Periodically, the bed-packing material
must be removed and regenerated with an acid wash.
References
1 Abu-Zahra, M.R.M., Sodiq, A., and Feron, P.H.M. (2016). Commercial liquid
absorbent based PCC processes. Absorption-Based Post-combustion Capture of
Carbon Dioxide (ed. Paul H.M. Feron), 757–778. Cambridge: Woodhead Publish-
ing.
2 US Energy Information Administration (2020). How much carbon dioxide is pro-
duced when different fuels are burned?, https://www.eia.gov/tools/faqs/faq.php?
id=73&t=11 ()
3 US Energy Information Administration (2019). Natural gas. Monthly Energy
Review (ed. Michael Kopalek), Table 4.3, November 2019, https://www.eia.gov/
totalenergy/data/monthly/.
4 Breeze, P. (2016). The Natural Gas Resource, Gas-Turbine Power Generation, 9–19.
Cambridge, MA: Academic Press.
5 Hammer, G., Lübcke, T., Kettner, R. et al. (2006). Natural Gas. In: Ullmann’s
Encyclopedia of Industrial Chemistry. Weinheim: Wiley-VCH.
260 10 Absorption
6 Asmara, Y. (2018). The roles of H2 S gas in behavior of carbon steel corrosion in

oil and gas environment: a review. Jurnal Teknik Mesin 7 (1): 37–43.
7 Yildirim, O., Kiss, A.A., Hüser, N. et al. (2012). Reactive absorption in chemical
process industry: a review on current activities. Chemical Engineering Journal
231: 371–391.
8 Bottoms, R.R. (1930) Process for separating acidic gases. US Patent 1,783,901,
filed 7 October 1930 and issued 2 December 1930.
9 Abu-Zahra, M.R.M., Schneiders, L.H.J., Niederer, J.P.M. et al. (2007). CO2 cap-
ture from power plants: Part I. A parametric study of the technical performance
based on monoethanolamine. International Journal of Greenhouse Gas Control 1:
37–46.
10 McCann, N., Maeder, M., and Attalla, M. (2008). Simulation of enthalpy and
capacity of CO2 absorption by aqueous amine systems. Industrial & Engineering
Chemistry Research 47 (6): 2002–2009.
11 Benson, H.E., Field, J.H., and Haynes, W.P. (1956). Improved process for CO2
absorption used hot carbonate solutions. Chemical Engineering Progress 52 (10):
433–439.
12 Tosh, J.S., Field, J.H., Benson, H.E., and Haynes, W.P. (1959). Equilibrium Study
of the System Potassium Carbonate, Potassium Bicarbonate, Carbon Dioxide, and
Water (ed. U.S.Bureau of. Mines), 5484. Pittsburgh.
13 Smith, H., Nicholas, N.J., and Stevens, G.W. (2016). Inorganic salt solutions for
post-combustion capture. In: Absorption-Based Post-Combustion Capture of Car-
bon Dioxide (ed. P.H.M. Feron), pp. 145–166. Cambridge: Woodhead Publishing.
14 Stewart, M.I. Jr., (2014). Gas sweetening. In: Surface Production Operations Vol-
ume 2: Design of Gas-Handling Systems and Facilities, 3e, vol. 2 (ed. M.I. Stewart
Jr.,), 433–539. Houston: Gulf Professional Publishing.
15 Burr, Barry & Lyddon, Lili. (2008). A comparison of physical solvents for acid
gas removal. GPA Annual Convention Proceeding, January 2008.
16 Speight, J.G. (2014). The Chemistry and Technology of Petroleum, 5e. Boca Raton,
FL: CRC Press.
17 Bentley, M.T. (1991). The basics of operating glycol dehydration units. In: 1991
AIChE Spring Annual Conference. TX: Houston 1–11 April 1991.
261
11
Extraction Technology
11.1 Introduction
Though there are many extraction processes, this chapter will focus on liquid–liquid
extraction (LLE) and supercritical fluid extraction (SCFE). Critical factors that con-
trol and affect extraction performance, including solvent selection and type of extrac-
tors, will be described. Selecting the right solvent and type of extractors are key to
achieving high purity and efficiency in the extraction process. After introducing the
fundamentals of extraction, the commercial extraction examples of hydrometallurgy
and hydrocarbon will be introduced.
11.2 Extraction Processes
11.2.1 Liquid–Liquid Extraction (LLE)

LLE is a method to separate components based on their relative solubility in two
different immiscible liquids such as water and isooctane. LLE is a technology used
to extract one or more components from one liquid (feed) phase into a second liq-
uid (solvent) phase. The transfer of the components from the feed to the solvent
is controlled by the solubility of each component in the two separate phases. The
two resulting phases from the extraction process are the enriched extract phase and
the depleted raffinate phase. After extraction, the solvent can be regenerated using
another separation technology such as distillation.
Figure 11.1 shows the four streams involved in the LLE process. Heavy Feed A + B
flows down from the top of extractor. Light solvent flows up the extractor. Since Feed
A is more soluble in the solvent, Feed A will be extracted by the solvent and enter
the extract separator to recover Product A. The solvent is then recycled back to the
extractor. The unextracted Feed B flows down the extractor with some solvent and
enters the raffinate separator to recover Product B. The solvent is then recycled back
to the extractor.
The extraction process is commonly operated at ambient temperature and
near-atmospheric pressure. For LLE to work, the component being removed from
the feed phase must preferentially distribute itself into the solvent phase from the
262 11 Extraction Technology
Product A
Extract A
(contain some
solvent)
Extract
separator
Feed (A+B)
Extractor
Solvent
Raffinate B
Raffinate B separator
(contain some
solvent)
Figure 11.1 Liquid–liquid extraction and solvent recovery system.
bulk feed. In addition, the solvent phase must be substantially immiscible from
the feed phase. Extraction efficiency can be enhanced by maximizing dispersion
of solvent phase into the feed phase. Extraction columns normally contain trays,
packing, or mechanical agitators to enhance the solvent dispersion. The function
of trays or packing is also to enhance solvent drop formation in the disperse
phase. Decreasing size of solvent drop increases mass transfer efficiency of feed
components to dissolve into the solvent phase. Mass transfer fundamentals in
extraction are like those in distillation.
11.2 Extraction Processes 263
11.2.2 Supercritical Fluid Extraction (SCFE)

SCFE is a method that uses a supercritical fluid such as carbon dioxide as the extrac-
tion solvent. SCFE can be used when the components to be separated have over-
lapping boiling points and distillation cannot be used. SCFE is also used to recover
thermally sensitive components. Carbon dioxide is a nontoxic solvent that is often
used in the food and pharmaceutical industries. In environmentally sensitive areas,
SCFE is used as substitutes for organic chemicals.
One application of SCFE involves replacing chlorofluorocarbon with supercritical
carbon dioxide in the extraction of caffeine from coffee bean. This process is demon-
strated in Figure 11.2. In the process, caffeine will be extracted from coffee bean by
liquid carbon dioxide in the extractor. After the extraction, caffeine will be washed
from liquid carbon dioxide using water in the scrubber. Liquid carbon dioxide will be
recycled back to the extractor. Caffeine will be recovered from water in the separator
and water is recycled back to the scrubber.
Another application of SCFE is the replacement of organic solvent in spray
painting with carbon dioxide. In the spraying process, unlike organic solvents,
carbon dioxide flashes, leaving no organic contaminates in the atmosphere. Propane
Water
Separator
CO2
Extractor
Caffeine
Coffee Scrubber
beam
CO2
CO2 and
coffeine
CO2 recovery
Figure 11.2 Supercritical extraction of caffeine from coffee beans using carbon dioxide
solvent.
Table 11.1 Characteristics: liquid–liquid extraction vs. supercritical fluid extraction.
LLE SCFE
1 Mass transfer between Slow Fast

solute (feed) and solvent
2 Solvent recovery By conventional such By simply dropping pressure or
distillation increasing temperature
3 Operation pressure Low High: 1000–5000 psig
4 Capital cost Low High: only high-value products
will bear the cost
5 Operating procedure Simpler Complicated
is another supercritical solvent used in SCFE to recover benzene, toluene, and

xylene (BTX) from paraffins and naphthenes, recover lubricating oils from residue,
and remove oils from sledges.
11.2.3 Liquid–Liquid Extraction vs. Supercritical Fluid Extraction

LLE has advantages and disadvantages compared to SCFE. For example, SCFE sys-
tem offers faster mass transfer rate for components to be separated and easier solvent
recovering than the LLE system. However, LLE system is less complicated to operate,
and needs lower operating pressure and capital cost. Table 11.1 summarizes these
features.
11.3 Solvents
The most important component affecting extraction efficiency is the solvent. There-
fore, solvent selection is a critical factor in developing an effective extraction system.
Criteria in solvent selection are described in this section.
11.3.1 Selectivity
Selectivity is calculated from the ratio of the distribution coefficients at equilibrium
of components A and B to be separated. The distribution coefficient of component A
is the ratio of concentration A in the extract phase to concentration A in the raffinate
phase. The distribution coefficient provides a measure of the affinity of component
A in the two phases. For the two-component separation, A and B, the separation
factor or the selectivity of A to B can be calculated as:
Ka
Separation factor or selectivity of A to B =
Kb
Xa
Ka = = distribution coefficient of A at equilibrium
Ya
11.3 Solvents 265
Xb
Kb = = distribution coefficient of B at equilibrium
Yb
Xa = amount of component A in extract phase
Ya = amount of component A in raffinate phase
Xb = amount of component B in extract phase
Yb = amount of component B in raffinate phase
11.3.2 Recoverability
After the extraction, the ability to recover solvent from the product is critical to the
efficiency of the process. Typically, this can be done by distillation. Therefore, when
selecting the solvent, the boiling point and thermal stability of the solvent need to
be taken into consideration when designing the system.
11.3.3 Immiscible
An extraction unit, a liquid stream (feed) containing the component(s) to be recov-
ered, is fed into an extractor, where it contacts a solvent. The solvent must be immis-
cible with feed liquid stream. This allows solvent to be dispersed as droplets in the
continued feed liquid phase.
11.3.4 Density
Mass transfer occurs between the droplets (solvent dispersed phase) and the
surrounding liquid (continuous-feed phase). For the two liquids to be subsequently
separated, solvent selected must have different density from the feed liquid stream.
The droplets then accumulate above or below the continuous phase, depending on
the liquids’ relative densities.
11.3.5 Chemical Stability

The selected solvent should be physically and chemically stable. The solvent should
have no chemical reaction with any of the components to be separated.
11.3.6 Loading Capacity (La )

To achieve high extraction capability, the solvent selected should have a high capac-
ity to retain the solute
(Solute)
La = = mole ratio of solute to solvent.
(Solvent)
11.4 Extractors
There are two types of liquid–liquid extractors: simple extractors and mechan-
ical extractors. Examples of simple extractors include spray, sieve trays, and
packed column. These extractors do not have any moving components. On the
other hand, mechanical extractors, such as mixer-settlers, rotary-disc columns,
reciprocating-plate columns (Karr), and centrifugal extractors, achieve separation
by mechanical agitation. Table 11.2 summarizes the several types of liquid–liquid
extractors.
11.4.1 Simple Extractor

Figure 11.3 shows the fundamental designs of three simple extractors: sieve-tray
column, spray column, and packed column. The dispersion of simple extractors is
achieved by vertical countercurrent flow of two phases: heavy and light under the
influence of gravity.
11.4.1.1 Sieve-tray Column

Sieve-tray columns have a simple geometry design and are used because of their
reliability. The columns are used widely in LLE as well as in supercritical extraction.
The simple geometry of sieve-tray columns is illustrated in Figure 11.4.
As shown in Figure 11.4, sieve-tray extractor columns are like distillation columns
except that weirs are not needed with sieve-tray extraction. The continuous-feed
phase moves across the tray and contacts the droplets of dispersed solvent phase,
which are generated at the tray perforations. The droplets of dispersed solvent move
vertically through the compartments between the trays. The mass transfer between
the continuous-feed phase and dispersed solvent-phase droplets to separate the feed
components occurs between the trays.
Table 11.2 Liquid–liquid extractors.
Extractor Characteristics
Simple extractors
● Spray, sieve, packed column Simple dispersion mechanism; suitable for
only a few stages of separation; low capital cost
Mechanical extractors
● reciprocating-plate column (Karr) Used with high viscosity liquids; system
becomes complex and costly for higher stage of
separation
● centrifugal extractor Used with liquids containing suspended solids;
easy-to-disperse components
Used with emulsified materials and systems
with low-density component differences; short
contact time for unstable materials; high
capital cost
11.4 Extractors 267
Light phase Light phase Light phase
Interface
Heavy
Heavy Heavy
phase
phase phase
Packing Packed
Holddown Sections
grid Coalesced
dispersed
Light phase
phase Light Packed Light
phase redistributor phase
Heavy phase Heavy phase Heavy phase
Spray extractor Packed extractor Sieve tray extractor
Figure 11.3 Simple extractor.
Extract
Feed - heavy
(continuous)
Continuous phase
(feed)
Tray
spacing
Coalesced Dispersed phase

solvent drop (solvent)
Solvent - light
(continuous)
Raffinte
Figure 11.4 Sieve-tray extractor.
In the extraction mechanism described above, the smaller the solvent droplets
generated, the higher the surface area of the dispersed solvent phase. The
smaller droplets result in better mass transfer between the feed and solvent. Mass
transfer efficiency of sieve trays can be predicted using published models and
correlations [1].
11.4.1.2 Packed Column

This type of column is used when only a few equilibrium stages are needed.
The packing enhances mass transfer by increasing droplets’ residence time and
Light phase
Heavy phase Light phase
Heavy phase
Centrifugal extractor
Rotary-disk column Reciprocating-plate

column (kar)
Figure 11.5 Mechanical added extractors.
dispersed-phase holdup. Packing also reduces back mixing of the continuous phase
by promoting the breakup of dispersed-phase droplets.
11.4.2 Mechanical Extractors

Mass transfer between the continuous-feed phase and solvent dispersed phase can be
aided by the use of mechanical agitation. Example schematics of mechanically aided
extractors are given in Figure 11.5: rotary-agitated columns, reciprocating columns,
and centrifugal extractor.
11.4.2.1 Rotary and Reciprocal Extractors

Advantages of mechanical extractors are that dispersion and interfacial area can be
significantly increased by mechanically aided agitation. Baffling is necessary to pre-
vent back mixing. Agitation in the rotating disk column is achieved by rotary motion
combined with baffling. In the reciprocating plate column, agitation is generated by
the vertical motion of the plate assembly.
11.4.2.2 Centrifugal Extractors

With the application of centrifugal force instead of gravity, extraction residence time
can be reduced, and phase separation enhanced. Because of their complex construc-
tion design, the capital cost of centrifugal extractors tends to be higher than that
of other types of extractors. Centrifugal extractors also have greater maintenance
requirement and therefore higher operational costs. However, they are compact and
offer relatively high throughput. These extractors are useful in applications where
contact time must be short. For example, in chemically unstable systems (e.g. extrac-
tion of antibiotics) when product inventory must be kept at a minimum, liquids tend
to emulsify, and components are difficult to separate.
Table 11.3 Features between simple and mechanically added extractors.
Features Simple extractors Mechanical extractors
1 Mass transfer Low High

2 Capacity High Low (except for centrifugal
extractor)
3 Efficiency Low High
4 Energy consumption Low High
5 Capital cost Low High
6 Service cost Low High
11.4.3 Simple Extractors vs. Mechanical Extractors

Mechanically aided extractors have many advantages and disadvantages compared
to simple extractors. Table 11.3 compares the features of simple and mechanically
aided extractors.
11.5 Industrial Extraction
LLE is widely used in industry for hydrometallurgy and hydrocarbon extraction.

Applications in gas–liquid extraction called absorption are used commercially to
extract carbon dioxide and sulfur-species (acid gas) from natural gas. Absorption was
covered in Chapter 10. In this next section, the impact of LLE will be demonstrated
with the following applications:
(1) Hydrometallurgy such as extraction of Zr from Hf, and U from Pu; and
(2) Hydrocarbon extraction such as aromatics from nonaromatic, m-xylene from
other xylene isomers, mercaptan removal, and SCFE of caffeine from coffee.
11.5.1 Hydrometallurgy Extraction

Hydrometallurgy extraction is a process used to recover metals from aqueous ore
leachates. However, because of the thermodynamic favorability of ionic species with
the polar aqueous phase, simple extraction of metal cation or anion from aqueous
phase into a nonpolar organic phase is unlikely to occur. For this reason, additional
chemicals into solvent are required to form complexes with the metal ions during
the contact of solute and solvent phases. These metal-complex formations enhance
the metal to be extracted from an aqueous phase to the nonpolar phase. More specif-
ically, properly designed chemical complexation reactions for specific metal cations
or anions enable them to be separated from a mixture of other species. Description
of both metal extraction fundamentals and specific applications have been docu-
mented and are referenced at the end of this chapter [2].
Table 11.4 Types of extraction solvents.
Reaction
Types of extraction solvents mechanism
Acid extractants: Complexation

1. Phosphoric acid (D2EPA) formation
2. Di(2-ethylhexylphosphoric acid) (D2EPHA)
3. Carboxylic acid (Versatic)
4. Sulfonic acid (DNNSA)
Basic extractants: Ion-pair
1. Primary amines/1,1,3,3,5,5,7,7,9,9-decamethyl decamine formation
2. Secondary amines/ditridecylamine
3. Tertiary amines/trioctylamine (TOA), triisooctylamine (TIOA)
4. Quaternary ammonium salts/trioctylmethylammonium chloride
Chelating extractants: Chelate
1. Hydroxyoximes (LIX 64) formation
2. Beta-Diketones (LIX 54)
3. 8-hydroxyquinolines (LIX 34)
Solvation: Solvate
1. Tributyl phosphate (TBP) formation
2. Phosphine oxide (TOPO)
In the hydrometallurgy reaction mechanism described above, metal extractions

rely on an extractant solvent compound. The solvent compound consists of a
diluent that provides the bulk physical properties and at least one other chemical
capable of reversibly complexing the solute. The reaction between the solute, in this
case the anion or cation metal, and the extractant is usually stoichiometric. Types
of extraction solvents for metal extraction are summarized in Table 11.4. These
extraction solvents form compounds with specific metals reversibly. This reversible
reaction–enhanced extraction is used as the basis for recovery of most metals from
leachate. There are four types of extraction solvents: acidic, basic, chelating, and
solvation extractants.
11.5.1.1 Zirconium/Hafnium Separation

Zirconium is an ideal material for use in nuclear reactors due to its low absorption for
thermal neutrons. Hafnium, on the other hand, with its strong neutron-absorption
properties is often a contaminant that can decrease the performance of the zirco-
nium. Because of the need to purify the zirconium, tributyl phosphate (TBP) solvent
was developed to extract hafnium from zirconium [3]. Separation of zirconium from
hafnium is demonstrated in a typical solvent extraction cycle in Figure 11.6.
In Figure 11.6, zirconium is separated from hafnium by using a solvent composed
of TBP in a hydrocarbon diluent such as dodecane. The feed from an acid ore
leachate includes both +4 cations of zirconium and hafnium in aqueous nitric acid
Scrub:
NaNo3 Product
HNO3 ZrNO3
Feed:
HfNO3 Strip:
ZrNO3 H2O
NaNo3
HNO3
HNO3
Solvent:
TBP
hydrocarbon
Raffinate: (dodecane)
HfNO3 HNO3
NaNo3
HNO3
Key:
organic phase
aqueous phase
Figure 11.6 Hafnium/zirconium separation utilizing extraction, scrubbing, stripping, and

regeneration.
and sodium nitrate. The feed is separated into a raffinate stream containing almost
all the hafnium and an extract stream that contains almost all the zirconium. In the
extraction unit, most of the zirconium is extracted into the solvent by complexation
with the TBP solvent. The hafnium is then back extracted from the solvent into the
aqueous phase in the scrubbing section. The scrubbing section is analogous to the
rectification section in a distillation column, and the scrub solution is analogous to
the distillation column’s reflux. The zirconium solute is removed from the solvent in
the stripping section by using a water strip solution. Finally, the solvent is purified
in the solvent regeneration section and then returned to the extraction unit.
11.5.1.2 Uranium/Plutonium Separation

Metals in nuclear waste can exist in the oxidation states of +1 to +6. More specifi-
cally, plutonium and uranium exist in the +3 or +4 state and +4 or higher oxidation
state, respectively. In the separation of uranium and plutonium, TBP is used as a sol-
vent as it only forms complexes with +4 and higher oxidation state metals [4]. With
this exclusive reaction between TBP and metal cations, uranium can be separated
from plutonium by utilizing the difference between the extractability of +3 and +4
cations in TBP.
The process for uranium and plutonium separation in nuclear waste is shown in
Figure 11.7. In the process, the solvent TBP is first used to extract +4 plutonium
and uranium while rejecting other +1, +2, +3 oxidation state metals from the feed.
The +4 plutonium is then reduced to +3 by adding an irreversible aqueous-phase
reductant such as hydroxylamine nitrate (HAN) or ferrous sulfamate (FeSAM). This
then allows the separation of plutonium from uranium by once again using TBP as
the solvent.
11.5.2 Hydrocarbon Extraction

Hydrocarbon extraction is based on the relative solubility of the components to be
separated in the solvent phase. The commercial examples that will be described next
are m-xylene separation from other xylenes using HF–BF3 solvent, aromatics sepa-
ration from nonaromatic using sulfolane solvent (Sulfolane process), and mercaptan
removal (Merox process).
11.5.2.1 m-Xylene/Xylene Isomer Separation

m-Xylene is an aromatic hydrocarbon. It is one of the three isomers of
dimethyl-benzene collectively referred to as xylene. The other two isomers
are o-xylene and p-xylene. The major use of m-xylene is in the production of
iso-phthalic acid, which is used as a copolymerizing monomer to alter the prop-
erties of polyethylene terephthalate (PET). Other uses of m-xylene are in the
production of unsaturated polyester resins. m-Xylene and other xylene isomers can
be produced using heavy naphtha feed in the reforming process. Because of the
closed boiling point of the xylene isomer, m-xylene cannot be separated from other
xylene isomers using distillation. Development of the m-xylene extraction described
below was the first technology developed for m-xylene separation.
Figure 11.7 U/Pu separation utilizing Scrub: Extract:

the difference between the HNO3 TBP complexed
extractability of +4 and +3 cations in U product
the TBP solvent. diluent
Aqueous
Reductant
Pu product
Feed: U+6 ,Pu+4

Pu+1,Pu+2,Pu+3
Raffinate: Solvent:
Pu+1,Pu+2,Pu+3 TBP diluent
Herman Pines [5] reports that the basicity of aromatics increases as the number
of alkyl groups bound to the benzene ring increases. Pines’ finding is summarized
in Table 11.5, which shows the relative basicity of methylbenzenes in HF–BF3 sol-
vent. It is clear all the xylenes are more basic than toluene. Furthermore, among the
xylenes, m-xylene is the most basic.
Based on the basicity of the different xylene isomers, Japan Gas–Chemical devel-
oped a LLE process using HF–BF3 solvent to extract m-xylene from another xylene
isomer [6, 7]. In this process, m-xylene (MX) reacts reversibly to form complexes with
boron trifluorides (BF3 ) in the presence of liquid hydrogen fluoride (HF). MX com-
plex (MXHBF4 ) is soluble in HF solvent and then can be separated from the other
feed stream isomers. The MX complex formation with HF–BF3 can be expressed by
Table 11.5 Base strengths of methylbenzenes relative to

HF–BF3 .
Relative
Methylbenzene basicity in HF–BF3
Toluene 0.63
p-Xylene 1
o-Xylene 2
m-Xylene 20
Pseudocumene (1,2,4-) 40
Hemimellitene (1,2,3-) 40
Durene (1,2,4,5-) 120
Prehnitene (1,2,3,4-) 170
Mesitylene (1,3,5-) 2 800
Isodurene (1,2,3,5-) 6 500
Pentamethylbenzene 8 700
Hexamethylbenzene 89 000
Source: Pines [5].
the equation below:
MX + HF (liquid) + BF3 (gas) → MXHBF4 (11.1)
To illustrate the extraction of MX through the formation of the MXHBF4 complex,

a simple laboratory experiment is introduced. Hydrocarbon-phase xylene isomers
are added into aqueous-phase HF. HF has a higher density and will stay at a lower
layer. BF3 gas is then charged into the extractor under mechanical agitation. After
the MXHBF4 complex is formed, the hydrocarbon and HF layers are completely sep-
arated from each other. Most importantly, the MXHBF4 complex is soluble in HF and
remains in the HF phase. The MX in the HF phase is recovered by the reversibility
of reaction (11.1), the decomposition of the complex MXHBF4 to MX + HF + BF3 .
One of the disadvantages of the MX extraction process is that the HF–BF3 is a cor-
rosive solvent. In practice, the MX LLE process is unfriendly and not economically
attractive. With this, a friendlier adsorptive separation process, MX Sorbex [8, 9], was
developed that replaced the MX extraction process just described. The MX Sorbex
development was detailed in Chapter 7.
The MX extraction process demonstrates that aromatics can be extracted from
other isomers based on their differences in basicity. Therefore, it should be expected
that other isomers in Table 11.5 can be separated using the same theory. For example,
isodurene should be able to be extracted from durene due to their basicity difference.
11.5.2.2 Sulfolane Process

p-Xylene is a critical petrochemical raw material for PTE production. p-Xylene (PX)
is one of the aromatics that is co-produced using catalytic reforming technology.
Besides aromatic, nonaromatic is also generated. Sulfolane is a process to recover

aromatics from Reformate. After sulfolane, aromatics product that contains BTX iso-
mer will be fractionated to BTX isomer. High-purity PX will then be recovered from
xylene isomer for the PTE application.
Sulfolane is a process used to recover high-purity aromatics from hydrocar-
bon mixtures such as reformed petroleum naphtha (reformate). The process is
also used to recover aromatics from pyrolysis gasoline (pygas) and coke-oven
light oil. The Sulfolane process is named after the solvent used, sulfolane or
tetrahydrothiophene-1,1-dioxide. Sulfolane is a five-membered ring structure
containing one atom of sulfur and four atoms of carbon with two oxygen atoms
bonded to the sulfur atom in the ring. Sulfolane was developed by Shell in the early
1960s as an industrial solvent and is used in aromatics extraction. Universal oil
products (UOP) has been the exclusive licensor of Sulfolane since 1965.
Sulfolane is used for extraction aromatics from nonaromatics because of its unique
properties in high aromatic/nonaromatic selectivity and hydrocarbons solubility.
A study of hydrocarbon solubility in sulfolane highlights its properties:
(1) With hydrocarbons containing the same number of carbon atoms, the solubility
decreases in the following order: aromatics > naphthenes > olefins > paraffins.
(2) Within the same hydrocarbon homologous series, the solubility decreases as
molecular weight increases.
Due to the unique sulfolane selectivity and solubility properties for hydrocarbons,
the Sulfolane process combines both LLE and extractive distillation (ED) in the same
process unit. This combination of technologies results in the effective processing
of feed stocks with a much broader boiling range than would be possible by either
extraction process alone. This combined extraction process has advantages for aro-
matic recovery in:
(1) LLE systems where light nonaromatic components are more soluble in the sol-
vent than heavy nonaromatics. Thus, LLE is more effective in separating aro-
matics from the heavy contaminants than the light ones.
(2) ED where light nonaromatic components are more readily stripped from the
solvent than heavy nonaromatics. Thus, ED is more effective in separating aro-
matics from light contaminants than heavy ones.
A simple commercial Sulfolane process [10, 11] flow diagram is shown in

Figure 11.8. Feed reformate enters the extractor and flows upward while the
Sulfolane solvent flows counter currently downward. The raffinate stream, which is
low in aromatics, is withdrawn from the top of the extractor. At the other end of the
unit, the Sulfolane solvent loaded with aromatics is withdrawn from the bottom of
the extractor and enters the extractive distillation column. In the ED unit, the light
nonaromatic components, having volatilities higher than aromatics, are separated
from the Sulfolane solvent. The overhead stream from the ED, light nonaromatics,
is recycled to the extractor to remove the heavy nonaromatics from the solvent
phase.
Raffiante
Sulfonate Benzene
extractor
Toluene
H2 LE
C9 aromatic
H2 LE H2 LE
THDA
Naphatha CCR RS H2 LE
HDT platformer Clay BC TC HA
Tatory
C10+
ortho-xylene
Light ends
H2 LE
Clay
XC O-X Parex Isomer DC7

Clay
para-xylene
Figure 11.8 UOP aromatics complex flowsheet.
The bottoms stream, which is substantially free of nonaromatic impurities

from the extractive distillation column, is sent to the aromatic’s recovery column,
where the aromatic products are separated from the solvent. This separation is
accomplished with minimal energy input, because of the large difference in boiling
point between the Sulfolane solvent and the heaviest aromatic components. To
minimize temperature exposures, the recovery column is operated under vacuum.
The Sulfolane solvent from the bottom of the recovery column is then returned
to the extractor. Next, the extract is recovered overhead and sent to distillation
columns downstream for recovery of the individual aromatic: BTX products.
To complete the process, the raffinate stream exits the top of the extractor and
enters the water wash column, where the dissolved solvent is removed with water.
The water in the solvent-rich phase is then vaporized in the water stripper. Accumu-
lated solvent from the bottom of the water stripper is pumped back to the recovery
column as the raffinate product exits the top of the raffinate wash column.
As described above, Sulfolane is a process to recover aromatics such as BTX isomer
from Reformate. p-Xylene, a critical petrochemical raw material for PTE produc-
tion, is then separated from xylene isomer by the Parex process. Because of the high
PX demand, other processes such as isomar and Tatoray are introduced to enhance
the PX product. To demonstrate the critical demand of PX, as of 2014, UOP has
licensed over 100 aromatic complexes and more than 700 individual process units to
produce aromatics. This includes more than 300 continuous catalyst regeneration
(CCR) PlatformingTM , 155 SulfolaneTM , 80 IsomarTM , 65 TatorayTM , and 100 ParexTM
process units. Figure 11.9 shows UOP’s aromatic complex flow sheet. In the com-
plex, aromatics will be produced using the CCR Platformer. After CCR Platformer,
the stream is fractionated into the C8 overhead. The C8 overhead is composed of
Unwashed raffinate (paraffins)
Light paraffins +
water
Sulfolane recovery column

Extractive distillation
Extractor column
Water wash column

Water
column
Sulfolane +aromatics +
Extract
light paraffins
Oil feed
(Aromatics
Sulfolane +aromatics
+ paraffins)
Water
Sulfolane
Backwash Water + sulfolane
Recycle sulfolne
Figure 11.9 Sulfolane process flowsheet.
aromatics and nonaromatic components, which are separated using the sulfolane
extraction process.
11.5.2.3 Merox Process

Global sulfur specifications for gasoline have been tightening over the last decade.
For example, European Union fuels for on-road and non-road vehicles have been
limited to 10 ppm sulfur since 2009 and 2011, respectively. In the United States,
ultralow sulfur diesel with a maximum of 15 ppm sulfur has been the requirement
since 2006 for on-road vehicles, 2010 for non-road vehicles, and 2012 for locomotives
and marine vessels.
Refiners rely heavily on sulfur extraction and sweetening process units, such as
the Merox process to meet gasoline and diesel product specifications. Merox is an
acronym for mercaptan oxidation. The technology was introduced to the refining
industry more than 50 years ago. The process is a proprietary catalytic chemical
process developed by UOP used in oil refineries and natural gas processing plants
to remove mercaptans from liquefied petroleum gas (LPG), propane, butanes, light
naphtha, kerosene, and jet fuel by converting them to liquid hydrocarbon disulfides.
The chemistry of catalytic oxidation of mercaptans (RSH) to disulfides (RSSR) in
an alkaline environment can be described as follows:
Extraction∶ RSH + NaOH → NaRS + H2 O
1 1 1
Caustic regeneration∶ NaRS + O2 + H2 O → NaOH + RSSR
4 2 2
1
Overall reaction∶ 2RSH + O2 → RSSR + H2 O
2
Table 11.6 Solubility of mercaptans in water and their

distribution between isooctane and 0.5-N NaOH aqueous
phase at 20 ∘ C
K q – Overall
Mercaptans extraction constant
Ethyl-mercaptan 77.9
n-Propyl-mercaptan 15.6
n-Butyl-mercaptan 3.59, 3.56
Tert-Butyl-mercaptan 2.46
n-Amyl-mercaptan 0.753
Tert-Amyl-mercaptan 0.362
n-Heptyl-mercaptan 0.0342
The extraction efficiency of the Merox process can be characterized by the extrac-
tion coefficient (K q ) in the equilibrium reaction and extraction equation below:
RSH (oil) ↔ RSH (aqueous) ↔ RS− (aqueous)
(RSH (aq) + RS− (aq))

Kq =
RSH (oil)
The extraction coefficients (K q ) of different mercaptans are summarized in
Table 11.6 [12]. The table shows that the lighter mercaptans such as C3 , C4 , LPG,
and naphtha gases have higher K q values. This indicates that the lighter mercaptans
can be extracted with a caustic solution. The resultant disulfides from the mercaptan
oxidation are separated and the caustic reused for extraction. Hence, the sulfur
content of the extracted hydrocarbon feed is reduced.
Heavier hydrocarbon fractions such as naphtha, kerosene, jet fuel, and diesel con-
taining mercaptans with low K q values cannot be readily extracted from hydrocar-
bons. Therefore, a sweetening Merox process is used where the mercaptan oxidation
reaction takes place in the presence of air and caustic solution. In the environment,
mercaptans are converted to disulfides, which remain in the sweetened hydrocarbon
product. The overall sulfur content, therefore, remains the same.
Figure 11.10 shows the Merox process applications for different types of hydro-
carbons [13]. It is confirmed that over 700 extraction Merox units are being applied
to the light mercaptan removal. It is also showed that over 700 sweetening Merox
units are being used to convert heavier mercaptans to disulfides. Merox LLE and
sweetening Merox will be described in the next sections.
Merox Liquid–Liquid Extraction

The flow diagram for the Merox LLE unit is shown in Figure 11.11. The light
naphtha feed flows up the extractor column and is counter-currently extracted with
a circulating caustic solution. If hydrogen sulfide (H2 S) is present in the feed, H2 S
Suitable for extraction merox process
LSR
naphtha FCC Kerosene Distillate
Gas up to 350 °C
LPG Pentane & Naphtha & Jet
phase EP
Condensate
Suitable for fixed-bed sweetening merox process
Figure 11.10 Merox process applications.
Extractor
Treated
product
Caustic
prewash
Oxidizer Disulfide
separator Spent
air
Disulfide
Air oil
Feed
Caustic settler
Regenerated caustic
Mercaptide-rich caustic
LP stream
Figure 11.11 Extraction Merox process.
is removed by passing it through a caustic prewash extractor. Removal of the H2 S

increases process efficiency by minimizing waste of the circulating caustic. After
the extractor column, the product passes to the caustic settler to further reduce
caustic carryover. The treated naphtha then flows through a sand filter coalescing
system to reduce the sodium content of the naphtha to <1 wt-ppm Na+ .
The Merox unit also contains a caustic regeneration section where the
mercaptan-rich caustic solution from the bottom of the extraction column
flows to an oxidizer column. Here, the extracted mercaptans are oxidized to
disulfides by air injected upstream of the oxidizer. The caustic stream leaving the
disulfide separator is then circulated back to the extraction column.
Sweetening Merox Process

The sweetening Merox process is used with pentane and heavier hydrocarbon feed
streams (Figure 11.10). As an example, the MinalkTM fixed-bed sweetening Merox
process is shown in Figure 11.12. The primary application for the Minalk process
Reactor
Dilute Air
caustic
Drain pot
Naphtha Product
To sewer
Figure 11.12 Sweetening Merox process.
is for treating naphtha. The Minalk process consists of a single reactor filled with a
fixed bed of activated charcoal impregnated with a Merox proprietary catalyst. Dilute
caustic is continuously injected into the naphtha feed stream along with air. The
naphtha is then passed downward through the fixed bed where the mercaptan is
oxidized to disulfide. The disulfide remains in the naphtha-phase product.
11.6 Summary
LLE, or solvent extraction, is one of the oldest and widely used techniques in indus-
trial separation. Supercritical extraction was developed in the later days with more
uses being developed. This chapter summarized the key components of the process,
including how to choose solvents and type of extractors. In addition to the basic prin-
ciples of extraction, the chapter highlighted some of the key commercial extraction
processes in metal extraction and hydrocarbon separation.
Acknowledgments
The author would like to thank Dr. Ames Kulprathipanja for reviewing the chapter.
Thanks also to Dr. Katipot Inkong and Ms. Orawee Lamoonkit of Chulalongkorn
University in drawing some of the figures.
References 281
References
1 Rocha, J.A., Humphrey, J.L., and Fair, J.R. (1986). Mass transfer efficiency of
sieve tray extractors. Industrial and Engineering Chemistry Process Design and
Development 25: 286.
2 Ritcey, G.M. and Ashbrook, A.W. (1984). Solvent Extraction, Principles and Appli-
cations to Process Metallurgy. New York: Elsevier.
3 Cox, R.P., Peterson, H.C., and Beyer, G.H. (1958). Separating hafnium from zir-
conium. Solvent extraction with tributyl phosphate. Industrial and Engineering
Chemistry 50 (2).
4 Schultz, W.W. and Navratil, J.D. (1984). Science and Technology of Tributyl Phos-
phate. Boca Raton, FL: CRC Press.
5 Pines, H. (1981). The Chemistry of Catalytic Hydrocarbon Conversions, 26.
Academic Press.
6 Ueno, T. (1970). Japan Gas-Chemical Xylene Separation Process, vol. 12. Bulletin
of The Japan Petroleum Institute.
7 Talbot, J.L. (1956). Separation of xylene isomers. US Patent 2,738,372.
8 Neuzil, R.W. (1982). US Patent 4,326,092.
9 Kulprathipanja, S. (1999). Process for adsorptive separation of metaxylene from
xylene mixtures. US Patent 5,900,523.
10 Kelly, M.F., and Uitti, K.D. (1970). Extractive distillation of aromatics with
sulfolane solvent. US Patent 3,551,327.
11 Meyers, R.A. (2016). Handbook of Petroleum Refining Processes, Fourthe. UOP
Sulfolane Process: McGraw-Hill Professional.
12 Yabroff, D.L. Extraction of mercaptans with alkaline solution. Industrial and
Engineering Chemistry 32 (2).
13 Sullivan D., The Role of the Merox Process in the Era of Ultra Low Sulfur Trans-
portation Fuels. Fifth EMEA Catalyst Technology Conference, 3 and 4 March
2004
283
Index
a zeolite 148
absorption zeolite powders 152–153
acid gas removal 243–245 zeolite synthesis 150–152
air stripping 259 adsorption
chemical 241 adsorbent 148–150
chemical solvent 245–251 bulk liquid adsorptive separation
CO2 emissions 241 154–170
gas separation 243 commercial bulk liquid adsorptive
glycol dehydration 257–258 process 170–174
natural gas 241, 242 principle of 147–148
physical 241 adsorption isotherms 156
physical solvent 251–257 adsorption step 154
accretion of zeolite powder 153–154 air stripping 259
acid gas removal 46, 63, 65, 66, 67, 228, alkanes 4, 5, 80, 197
243–245, 247, 257 alkyd resins 133, 139
acrylamide 81 alkylation 7, 8, 27, 29, 33, 34, 56, 58, 59,
acrylic acid 25, 75 103, 122, 128, 197, 226
acrylic resins 129, 139 alkylbenzene sulfonates 132–133
acrylonitrile 25, 30, 45, 81, 129, 198, 234 alkylbenzenes 132–133
Acrylonitrile–Butadiene–Styrene (ABS) AlPO-14/polymer MMMs 220
30, 81, 129 α-alumina porous support membrane
activated alumina 148 218
activated carbon 148, 219 alumina-supported platinum catalysts
activated clay 148 51
activity coefficients 185 aluminosilicates 60, 148, 153, 169, 214
adipic acid 34, 124–127 Amine Guard Formulated Solvent (AGFS)
adipic acid production 124–128 249
adiponitrile 81 Amine GuardTM FS process 63
adsorbate 147, 148, 154, 156, 157, 159, amines process 246–248
160, 162, 164, 166, 168 ammonia 13, 37, 38–39, 45, 67, 75, 81,
adsorbents 82–83, 89, 114, 180, 204, 228, 249,
general uses of 148–149 250
pore size distribution 148 aniline 113, 130, 131, 180

284 Index
aqua-gel 151 auto diesel 3

aqueous potassium carbonate (K2 CO3 ) auto-thermal reforming of methane 66
system 250 azeotropes 185–186, 200, 203
arene compound 32 azeotropic distillation 185, 187, 198
aromatic compounds
benzene 32–34 b
toluene 34–36 Ba-Mordenite selectivity 160
xylenes 36–37 batch distillation 186–187
aromatic production BaY selectivity 160
aromatic transalkylation and Ba-zeolite 161
disproportionation technology “bellweather” operating rate value 21
95–99 BenfieldTM process 63, 244, 249–251
aromatization of light naphtha/ BenSat® reactor 89
hydrocarbons 105–107 benzene 3, 6, 11, 17, 32, 121
catalytic cracking 107–111 aromatic compounds 32–34
isomerization of aromatics 100–103 rings 5
pyrolysis and coking 111–118 benzene and xylenes (BX) 35, 99
reforming technology 88–90 benzene, toluene, ethylbenzene, and
toluene methylation 103–105 xylene (BTEX) 257
aromatic transalkylation and benzoic acid 6, 35
disproportionation technology benzothiophene 6
95–100 Berl and Intalox® saddles 194
aromatics 5, 47, 87, 167, 197 binary column 187
complex 12, 13, 48, 276 bio-crude 117, 118
oxygenates 118 biodegradable household detergents 45
aromatization of light naphtha/ biomass pyrolysis 115–118
hydrocarbons 105–107 bio-oil 117
asphalt 3, 5, 8, 10 biphenyl 6
asphalt oxidation 8 Bisphenol A (BPA) 34, 121, 129, 130,
asphaltenes 5 198
asymmetric integrally skinned flat sheet Bisphenol A production 130
membrane 207–211 bitumen 5, 89
asymmetric integrally skinned hollow blended clay zeolite mixture 153, 154
fiber membrane 207, 208, 210 Blumenthal‘s column 180
asymmetric integrally skinned membrane borontrifluorides (BF3 ) 122, 174, 272,
structure 207, 208, 216 273, 274
asymmetric integrally skinned polymeric BP-UOP Cyclar process 105
membranes 209–211, 213, 216 branched-alkylbenzene 132
asymmetric integrally skinned porous breakthrough procedure 157–158
polymeric membranes 207, 208 BTX separation 199
asymmetric polymeric membranes 207, Bubble-Cap trays 180, 190, 191
208, 221–224 bubbling area 190
atactic polypropylene 79 bulk liquid adsorptive separation
atmospheric distillation 7, 8, 187, 191, adsorption isotherms 156
195, 197, 198 breakthrough procedure 157–158
Index 285
equilibrium-selective adsorption asymmetric integrally skinned

159–166 membrane 227
ion exchange 169 cellulose triacetate 205
pulse test procedure 156–157 centrifugal extractors 266, 268
rate-selective adsorption 166–168 ceramic materials 214, 235
reactive adsorption 170 ceramic membranes 204, 225, 226
shape-selective adsorption 168–169 chemical absorption 241, 244, 245, 246,
Sorbex process 158–159 247, 248, 249, 251, 252, 255
bulk polymerization 77, 80 chemical adsorption 63, 147
bunker oil 3 chemical grade propylene (CGP) 53
butadiene 11, 19, 26, 29, 30, 31, 46, 47, chemical solvent 245
81, 122, 127, 129, 198 amines process 246–248
butadiene rubber 30 BenfieldTM process 249–251
butenes 3, 8, 11, 12, 26, 27, 28, 29, 76, chemicals production 35, 133–141
112, 198 chemisorption 147
1-butene 27, 198 chlorobenzene 34, 130
butylenes 8, 11, 26, 27, 28, 29, 30, 31, ,
circulating regenerated catalyst 49
45, 48, 107
C4 isomer separation 199
C4 mono- and di-olefins 11, 17
c
CO2 corrosion reactions 244
C8-aromatics 136, 161–163, 166
CO2 emissions 241, 242
C8 aromatic isomerization process 100
coal gasification 43, 62, 67–72, 252
Ca-zeolite 161
coal pyrolysis process 113, 114, 180
caffeine 263, 269
coal tar 87, 113, 180
caprolactam 34, 121, 125–128
coal to olefins (CTO) 20, 24
synthesis 125–128
coatings 34, 121, 129, 139, 174
caprolactam- nylon-6 production; BPA-
coke 3, 9, 10, 46, 48–52, 60, 61, 66, 87,
polycarbonate plastics 122
91–94, 96, 106, 107, 113–115, 180,
carbazole 6
carbon dioxide 4, 62, 68, 69, 70, 78, 82, 275
113, 204, 241, 263, 269 coke-ovenbenzole 87, 113–115
carbonization zone 69 coking 7, 9, 32, 33, 35, 46, 57, 58, 60, 91,
Carom 110 111–118, 225
catalyst regeneration section 49, 50–51 C4 olefins 26–31
catalytic cracking 8, 9, 23, 45, 89, 105, commercial bulk liquid adsorptive process
107–111, 117, 152 meta-xylene (MX) SorbexTM 173–174
catalytic reformer 7 Parex 170–175
CatofinTM 49 commodity chemicals 17, 18, 22, 45
C3/C4 hydrocarbons 105 coniferyl alcohol 117
C10 -C14 mono-methyl-paraffins 167 continuous catalyst regeneration (CCRTM )
C10 -C14 n-paraffins 167, 169 technology 49
C3 /C4 Oleflex complex 56 continuous catalyst regeneration (CCR)
CCR PlatformingTM process 91, 172 reforming 91–94, 96
Celgard® membrane 207 continuous catalyst regeneration (CCR)
cellulose acetate (CA) 64, 205 system 93, 105
286 Index
continuous distillation 179, 180, 187, diethylbenzene 122, 171

189, 195 di-ethylene glycol 78
continuous organic polymeric matrix di-2-ethylhexylphathalate 138
219 diffusivity coefficient 205, 206
cross-linked polystyrene- di-isopropylbenzene 166–168
divinylbenzene 123 1,3-di-isopropylbenzene 166–168
crude oil distillation 7, 180, 195–198 1,4-di-isopropylbenzene 166–168
crystalline organic framework (COF) di-isopropylbenzene isomers 167, 168
213, 214 2-dimensional GCxGC analysis 108
cumene 25, 33, 34, 75, 81, 128–130, 198 dimerization 8, 27, 55, 106, 130
cumene/phenol 25 dimethyl benzenes 11
Cyclar flow diagram 106 1,2- 1,3- and 1,4-dimethylbenzene, 6, 11
cyclic reforming 91, 93, 94 dimethyl ether (DME) 103, 252, 255
cyclicsulphides 6 dimethyl ether of polyethylene glycol
cyclic-regenerative reforming 96 (DMEPG) 252–256
cycloalkanes 5 2,6-dimethyloctane 167
cyclohexane 5, 34, 109, 123–124, 127 dimethyl substituted benzene isomers
cyclopentane 5 11
dimethyl sulfoxide (DMSO) 209
d dimethyl terephthalate (DMT) 134
decane 5, 167 di-octyl-phthalate (DOP) 138
DeFineTM 58 diolefins 11, 26–31, 52, 57, 59
dehydrogenation 12, 24, 26, 30, 31, 45, 1,3-dioxolane 209
48, 49, 51, 52, 55–58, 89, 100, 106, 4,4-diphenylmethane diisocyanate (MDI)
107, 122, 197, 215, 226 130, 132
deisobutanizer (DIB) 54, 55, 57 dispersed inorganic materials 219, 223
delayed cokers 7 distillatehydrotreater 7
demetallizing 9 distillation
desalination 7, 227, 235 activity coefficients 185
desorbent 47, 147–148, 154–160, 162, azeotropes 185–186
165–166, 171–174 azeotropic distillation 198
characteristics of 165–166 batch distillation 186–187
desorption 147, 154, 156, 157, 159, 160, Bubble-Cap trays 191
241 continuous distillation 187
desorption step 154 crude oil 195–198
DetalTM process 58 dividing wall columns (DWC) 199
detergent alkylation 58, 59 extractive distillation 198
dewaxing 9, 234 flow within the distillation column
di-/tri-substituted paraffins 167 188–189
di-alkyl naphthalene 109 introduction to 179–181
dichlorodiphenyltrichloroethane (DDT) McCabe-Thiele diagram 182–184
133 multi-component distillation
diesel 3, 7, 10, 19, 40, 91, 108, 109, 111, 189–190
115, 197, 241, 277, 278 random packing 193–194
diethanolamine 151, 248–249 reactive distillation 200
Index 287
relative volatility 184–185 polyvinyl chloride (PVC) 77–78

sieve tray 191 ethylene chloride 75
solubility 185 ethylene dichloride (EDC) 21, 45, 77, 78
structured packing 194–195 ethylene fractionation 46
surface area 186 ethylene oxide (EO) 21, 45, 75, 78
tray contactor 190 ethyl mercaptan 6
tray vs. packed columns 187–188 exit weir 190
valve tray 191 extraction coefficients 278
vapor-liquid equilibrium (VLE) theory extractive distillation (ED) 95, 110, 172,
181–182 185, 187, 198, 275, 276
di-substituted paraffin 167 extractors
dividing wall columns (DWC) distillation mechanical 268
199 simple 266
divinylbenzene 122, 123 simple vs. mechanical 269
downcomers 190 ExxonMobil 96, 103, 151, 234
dried cellulose acetate (CA)/cellulose
triacetate (CTA) spiral wound f
membrane technology 205, 211, facilitated transport membranes (FTMs)
228, 234 232, 233
dried CTA hollow fiber membrane system FCC naphtha fractionation 199
228 FCCTM 197
DuPont synthesis 125 feed component stability 186
feedstocks 19, 20, 23, 24, 25, 28, 29, 36,
e 46, 47, 49, 94, 99, 105, 111–114, 117
electrostatic/Van der Waals forces 148 ferrous sulfamate (FeSAM) 272
elemental sulfur 10 field butane 54, 57
EluxylTM process 171 fixed-bed fired-tube reactor 49
entrained-flow reactor 69 flash regeneration system 254
equilibrium-selective adsorption flexible ZIF-8/polyethersulfone MMM
159–166 223
essential materials 121 flexicokers 7
ethane 4, 19, 20, 22, 23, 25, 28, 36, 45, floating, production, storage, and
46, 47, 50, 111, 112, 113, 118, 226, offloading (FPSO) vessels 203,
229 229, 231
ethane cracker 47 flow within the distillation column
ethene 10 188–189
ether fuel additives 27 fluid catalytic cracker (FCC) 7, 23–29,
EthermaxTM process 200 45, 48, 53, 60, 89, 107–111, 117,
ethyl tertiary butyl ether (ETBE) 27, 54, 197, 199
75 fluid catalytic cracking (FCC) process 8,
ethylbenzene 6, 11, 21, 32, 33, 81, 100, 23, 26, 45, 89, 107, 117
103, 122–123, 197, 257 fluidcokers 7
ethylene 3, 10, 17, 18, 45, 75 fluidized-bed dehydrogenation
monoethylene Glycol (MEG) 78 technology 49
polyethylene 75–77 fluidized-bed reactors 69
288 Index
Fluor 252, 254 hot potassium carbonate (HPC) process

free-standing membranes 210 249
fuel gas 3, 37, 82, 101, 109, 172 H2 S corrosion reactions 244
hydraulic fracturing 241
g hydrocarbon extraction 269, 272–280
gas phase adsorption 147 hydrocracker 7, 89
gas separation membranes 203, 204, hydrocracking 9, 10, 90, 91, 93, 109,
205, 228, 231, 233 111, 197, 228
gas water shift reaction (GWSR) 71 hydrodealkylation (HDA) 33, 99
gas-phase polymerization 80 hydrodenitrification (HDN) 89
gaseous petroleum 4 hydrodesulfurization (HDS) 10, 64, 89,
gasification zone 69 228
gasoline 3 hydrogen 57, 82, 96, 107
high octane gasoline 7, 8, 31, 55, 88, hydrogen/hydrocarbon ratio (H2 /HC
93 Ratio) 93, 94
pyrolysis gasoline (pygas) 32, 36–37, hydrometallurgy extraction 269–272
45–47, 111–114, 275 hydroskimming refineries 8
hydrotreater 7
glassy and rubbery polymers 205, 228
hydroxylamine nitrate (HAN) 272
glassy polymers 205, 221
glycol dehydration 257–258
i
Gore-Tex® membranes 207
i-butenes 26
immiscible solvent 265
h
impact copolymer 80
heavy aromatics 103
indole 6
heavy crude oil 4–5
industrial extraction
Heavy Desorbent Parex 171, 172
hydrocarbon extraction 272–280
heavy distillates 10
hydrometallurgy extraction 269–272
heavy fuel oil 10
Intalox® IMTP random packing 194
heavy gas oil 3
SiO2 /Al2 O3 molar ratio 160, 163–164
heavy normal-olefins 45, 75 ion exchange 122, 169, 170, 214, 235
heavy normal-paraffin 45, 57–59 ion exchange capacity 169
heavy normal-paraffin dehydrogenation ion exchange selectivity 169
45, 57–59 ion-exchange membranes (IEMs) 235
hexane 5, 197 ionic detergents 132
HF-BF3 solvent 272, 273 iso-butane 8
HgSIVTM regenerative system 64 iso-C4 Oleflex complex 54
high conversion refinery 9 iso-octane 5
high octane gasoline 7, 8, 31, 55, 88, 93 iso-paraffin alkylate 8
high-density polyethylene (HDPE) 75, iso-paraffins 8, 47, 167, 197
76 isobutene 11, 26–29
high-pressure polyethylene 75 isodurene 162–164, 274
homo-polymer 80 Isomar flow diagram 101, 102
Honeywell UOP Conventional Chemical isomerization 8, 51, 100
Solvent - Amine Process 249, 250 isomerization of aromatics 100–103
Index 289
iso-paraffins 47, 167, 197 liquid hydrogen fluoride (HF) 59, 162,
isophthalic acid 133, 134, 174 174, 272–274
isopropylbenzene 128, 166–168 liquid petroleum 4–5
isotactic polypropylene 79 liquid-liquid extraction (LLE) 261–262,
I-400 catalyst 102 264, 275, 278–279
vs. supercritical fluid extraction (SCFE)
j 264
jet fuel 3, 10, 19, 59, 197, 277, 278 Li-zeolite 161
Joule-Thompson effect 112 low pressure polyethylene 75
low-density polyethylene (LDPE) 75, 76
k LPG 7, 10, 19, 52, 88, 105, 107, 109, 111,
KataMaxTM structural packing system 119, 197, 277, 278
200 lubricants 3, 153
kerosene 3, 7, 10, 45, 57, 58, 59, 167, lubricating oils 10, 264
197, 277, 278
K-zeolite 161 m
maleic anhydride 34, 133, 138
l m-aminophenol 130
LAB technology 59 MaxEne process 47, 48
light crude oil 4–5 Maxwell model 220
Light DesorbentParex 172–174 McCabe-Thiele design method 182
light distillates 8, 10 McCabe-Thiele diagram 182–184
light gas oil 3 m-chloroaniline 130
light naphtha 19, 88, 105–107, 109, 277, MDI 130–132
278 mechanical extractors 266, 268
light olefins 45, 48, 49, 62 medium conversion refinery 8, 9
light paraffin 45, 48–57 medium-pore zeolite 60
light paraffin dehydrogenation membrane contactor 204, 207, 227, 234,
iso-C4 Oleflex complex 54–56 235
process and catalyst 48–51 membrane permeance 206, 217, 218
UOP C3 Oleflex Complex 51–54 membrane separation 203, 206, 211,
UOP C3 /C4 Oleflex complex 56 212, 216, 223, 225, 227, 228, 229,
UOP mixed-C4 Oleflex complex 56–57 233, 234
light solvent 261 membranes 64
light/heavy naphtha 3 gas separation 228
linear alkylbenzene (LAB) 45, 57–59, inorganic 213–218
75, 197 ion-exchange 235
linear low-density polyethylene (LLDPE) membrane contactor 234–235
75–77 microfiltration 234
linear-alkylbenzene 132 mixed matrix 218–223
liquefied natural gas (LNG) 63 modules 223–225
liquefied petroleum gas 3, 88, 277 nanofiltration 233–234
liquid adsorption 148, 154, 160 pervaporation 233
Liquid Hourly Space Velocity (LHSV) polymeric 205–207
93 principle and background 203–204
290 Index
membranes (contd.) microporous MFI zeolitic membranes

reactors 225–227 217
reverse osmosis 227–228 microporous zeolite 204, 213, 217, 219
ultrafiltration 234 microporous zeolites 219
vapor/gas and vapor/vapor separations microporous zeolitic membranes 204
230–233 microporous ZIF-8 material 223
MemGuardTM pretreatment system 229 middle distillates 8, 10
mercaptans 6, 7, 252, 277, 278, 279 Minalk process 279, 280
mercury removal 64, 68 MinalkTM fixed bed sweetening Merox
Merox liquid-liquid extraction 278–279 process 279
Merox process 7, 272, 277–280 mixed C4 stream 46, 47
meta-xylene (MX) SorbexTM 173 mixed ethylene/butylene metathesis 27
meta-xylene diamine (MXDA) 174 mixed ion conducting membrane 226
meta-xylene/xylene isomer separation mixed matrix membranes (MMMs)
272–274 introduction 219–221
metal cation exchanged in zeolite preparation of 221–223
160–166 mixed-C4 Oleflex complex 56–57
metal-organic framework (MOF) 213, m-nitrobenzenesulfonic acid 130
214, 218, 219 m-nitrochlorobenzene 130
metalloaluminophosphates (MeAPOs) moisture content in zeolite 160,
214 164–165
metallosilicoaluminophosphates molecular sieving 168, 171, 214, 215,
(MeAPSOs) 214 218, 219, 233
methane 3, 4, 33, 46, 50, 63, 65, 66, 68, MolexTM 58, 169
69, 82, 95, 114, 204, 226, 228, 229, molybdenum oxide 87
242, 254, 258 2 mono-aromatic molecules 109
methanol (MeOH) 39, 83, 200, 253–254 mono-methylparaffin 167
methanol plant 13 mono-substituted paraffin 167
methanol to olefins (MTO) 12, 22, 24, monoethanolamine (MEA) 248, 249
45, 59–62, 103 monoethylene Glycol (MEG) 78
methyl ethyl ketone (MEK) 27 monomethyl acrylate (MMA) 27
methyl ethylene 10 Monsanto synthesis 125
2-methylnonane 167 moving-bed reactor 69
methyl-tert-butyl ether (MTBE), 27, 31, multi-channel monolith module 225
40, 45, 49, 54, 55, 56, 75, 200 multi-component column 187
methylacetylene 23 multi-component distillation 189–190
methyldiethanolamine 248–249 multi-product column 187
microfiltration (MF) membranes 204, Mx/n [(SiO2 )y (AlO2 )x|i ].wH2O 150
234 m-xylene (MX) 36, 37, 99, 133, 134, 159,
microporous crystalline inorganic 160, 162, 269, 272–274
membranes m-xylene separation 159, 272
introduction 214–215
preparation of 215–218 n
microporous DD3R zeolitic membrane N2 /air membrane separation system
217 228
Index 291
Na-zeolite 161 ethylene 12, 45

NaA zeolite membranes 214–216, 225 separation 159
n- and iso-butenes 27 Olefin Cracking Process (OCP) 61, 62
nanofiltration (NF) membrane 233 Oleflex complex 51–57
naphtha 19, 197 Oleflex process 49–52
complex configuration 88 Oleflex product recovery section 50
hydrotreater 7 OleflexTM process 49, 226
reforming 8, 32, 34–37, 88, 107, 114 OMEGA process 78
naphtha hydrotreating unit (NHT) 88, on-purpose propylene production 24, 30
89, 101 organic non-solvents 209
naphtha-based cracker 47 organic solvent NF (OSNF) membranes
naphthalene 6, 109, 113, 137 204, 227, 233, 234
naphthenes 5, 47, 88, 89, 94, 167, 169, ortho- and meta-xylene 100
197, 264, 275 oxidation reactions of xylenes 137
natural gas 4 oxo alcohols 25
extraction 241, 242 o-xylene (OX) 37, 99, 133, 134, 137, 160,
production 4, 242, 244 272
natural rubber 29, 30, 115
n-butenes 26, 28, 29 p
n-decane 167 packed column 186–188, 192, 193, 259,
N,N-dimethyl acetamide (DMAC) 209 266, 267–268
N,N-dimethylformamide (DMF) 209 Pacol 57–59, 199
N-methylpyrrolidone (NMP) 209 Pacol dehydrogenation process 57–58
n-paraffin 152, 159, 165, 167 Pacol Enhancement Process (PEP) 59,
n-paraffin desorbent 165 199
n-paraffin separation 152, 159, 165 Pall rings 194
n-paraffins/non-n-paraffins separation para xylene aromatics 7
167 para-xylene 7, 12–13, 32, 87, 97,
nitrile rubber 81 100–105, 114, 171–173, 274, 276
nitrobenzene 34, 130, 131 paraffin dehydrogenation process 48
non-n-paraffins 167, 169 paraffins 5, 45, 60, 89, 94, 197
non-catalytic partial oxidation 66 paraffins, naphthenes and aromatics
normal and iso C4 olefins 26 (PNA) 89, 94, 167
normal-paraffins 47, 52, 57–59, 167 Parex 100–103, 110, 170–174
nylon 33, 34, 76, 121, 123, 124, 127, 128, partial oxidation 62, 64–66
129, 209, 210, 214, 217, 234 p-coumaryl alchohol 117
Nylon 6 121, 123–128, 234 pelletization of zeolite powder 153
Nylon 66 123–128 Penex® isomerization reactors 89
nylon fibers 34 pentane–hexane 197
PEPTM 58
o pervaporation 204–205, 215, 227, 233
oil refinery 6, 7, 10, 187 pesticides 121, 142
oil refining 8, 23 PET films 134, 136
olefins 45, 75, 197 PET solid-state resins 136
complexes 12, 232 petrochemical building blocks 10–15
292 Index
petrochemical complexes 12, 47, 99, poly(methylene diphenylene isocyanate)

108, 111 (PMDI) 130
petrochemical market poly(vinyl acetate) 139
ammonia 38 poly-MDI (PMDI) 131
benzene 32–34 poly[Zn2 (benzimidazole)4 ] (Zn2 (bim)4 )
C4 olefins 26–31 nanosheets 218
diolefins 26–31 polyamide (PA) TFC flat sheet membrane
ethylene 18–22 227
methanol 39 polyamides 206
propylene 22–26 polyaramides 205
toluene 34–36 polybenzimidazole (PBI) 206, 234
xylenes 36–37 polybenzimidazoles 206
petrochemical-oriented naphtha polybenzoxazoles 206
reforming 36 polyester 21, 210
petroleum polyester fibers 134
composition of 5–6 polyesters 121
gaseous 4 polyether ether ketone (PEEK) 234
liquid 4–5 polyethers 206
products 9–10 polyethersulfones 206
refinery 6–9 polyethylene 15, 45, 75
solid 5 polyethylene terephthalate (PET) 36,
phenanthridine 6 78, 100, 272
phenol 6, 25, 34, 124, 129–131, 180, 198 polymer 136
phenolics 33–34, 118 polyimides 205–206, 220, 228
phenyldecane 166 polyisobutylene (PIB) 27, 31
Phillips STAR processTM 49 polymer grade propylene (PGP) 46, 53
phthalic anhydride (PA) 37, 133–134, polymer production from basic aromatic
137, 138 building blocks
physical absorption 241, 243–246, adipic acid production 124–125
251–256 alkylbenzene sulfonates 132
physical adsorption 63, 147 alkylbenzenes 132
physical solvent 251 aniline 131
flow schemes 254–255 Bisphenol A production 130
SelexolTM 255–257 caprolactam synthesis 125–128
solvent selection 252–254 caprolactam- nylon-6 production; BPA-
physisorption 147–148 polycarbonate plastics 122
planar six-carbon rings 5 cross-linked polystyrene-
plasticizers 37, 133–135, 138 divinylbenzene 123
plastics 11, 21–22, 25, 34, 36, 80, 81, Cumene 128–130
121, 133, 142 cyclohexane 123–124
PlatformingTM process 50, 87, 91, 172 DuPont synthesis 125
polar adsorbate 165 MDI 131
poly(ethylene naphthalate) (PEN) 141 Monsanto synthesis 125
poly(ethylene terephthalate) 140 nitrobenzene 130
poly(ethylene terephthalate) (PET) 141 Nylon 6 123–124, 128
Index 293
Nylon 66 123–124, 128 propane dehydrogenation (PDH)

poly-MDI (PMDI) 131 technology 23–24, 27, 107
polystyrene 122, 123 propane-rich liquefied petroleum gas
polyurethane 132 (LPG) feedstock 52
styrene 122 propene 10
toluene diisocyanate (TDI) 131
propylbenzene 103
polymer-grade propylene 46, 53
propylene 3, 10, 12, 17, 22, 27, 45, 75,
polymeric gas separation membranes
107
203, 206, 233
polymeric membrane-based industrial acrylonitrile 81
separation processes 203 polypropylene (PP) 79–80
polymeric membranes propylene oxide 80–81
asymmetric integrally skinned propylene carbonate (PC) 80, 252–255
209–211 propylene complex 52–53
fundamentals of 205–207 propylene fractionation 46
preparation of 207–213 propylene oxide 25, 45, 75, 80–81
TFC 211–213 pulse test procedure 156–157
polymeric resin 148 purified isophthalic acid (PIA) 174
polypropylene (PP) 25, 45, 75, 79–80, purified terephthalic acid (PTA) 12,
207, 211
133, 136
polypropylene Celgard® membrane 207
p-xylene (PX) 19, 32–37, 97, 99–103,
PolysepTM hollow fiber membrane
110, 133–137, 139, 159, 160, 163,
systems 228
polystyrene 122, 123 166, 274, 276
polystyrene-divinylbenzene 122–123 oxidation to terephthalic acid 139
polyurethane 34, 123, 132 separation 159
polyvinyl chloride (PVC) 77–78 pyridine 6, 111, 198
microporous inorganic zeolite membranes pyrolysis gasoline 320, 36–37, 45–47,
233 111–114, 275
potassium carbonate chemical absorption pyrolysis and coking 111–119
processes 251 pyrolysis gasoline (pygas) 32
pressure swing adsorption (PSA) system pyrolysis oil 117
82, 106, 229
primary amine 248 q
primary/commodity chemicals 17
quinolone 6
PRISM® polysulfone hollow fiber
membrane 228
r
PRISM® hollow fiber membrane 205
Raffinate-1 (Raff-1) stream 31
PRISM® silicone rubber-coated
random copolymers 80
polysulfone hollow fiber
membrane 228 random packing 187, 192–194
propadiene 23, 46, 52 Raschig rings 192–194
propane 4, 24–25, 28, 45–46, 49, 51–56, rate-selective adsorption 159, 166–168
107, 112, 198, 215, 226, 233, 254, reactive adsorption 159, 170
263, 277 reactive distillation 181, 200
294 Index
reactor effluent gases 45 sieve tray 191

rectification 179, 272 columns 266
RectisolTM 252 extractor 266–267
refinery 3–10, 23, 26, 28–30, 34, 45, silica gel 148
53–54, 59, 91, 108, 114, 181, 187, silicalite 150, 167–168
196, 204, 228 silicates 214
refinery-grade propylene (RGP) 53 silicoaluminophosphates (SAPOs) 60,
reformate 7, 32, 48, 88–89, 91, 93–97, 214
133, 172–173, 275–276 silicone rubber polymer 205
reformate–octane relationship 97 simple extractors 266–268
reformed petroleum naphtha 275 sinapyl alcohol 117
reforming technology 88–95, 274 [SiO4 ]4- and [AlO4 ]5- tetrahedra 148
relative volatility 180, 181, 184–186, slurry polymerization 80
189, 198 sol-gel membranes 214
residuum 10 solid petroleum 5
resins 34, 40, 78, 80, 121–122, 129–130, solubility 151, 184–185, 190, 205–206,
133–142 208, 211, 243, 252–254, 261, 272,
reverse osmosis (RO) 204, 227
275, 278
rotary and reciprocal extractors 268
solubility coefficient 205–206
RTPTM flow diagram 116
solvent selection 244–245, 252, 261, 264
RTP green fuel typical properties 116
solvents
chemical stability 265
s
density 265
SAPO-34 60
immiscible 265
membrane 215
loading capacity 265
molecular sieve membrane 60, 214
recoverability 265
sec-butyl alcohol 27
selectivity 264–265
secondary amine 248–249
Sorbex process 158, 174
selective toluene disproportionation
99–100 SorbexTM technology 59, 156
SelexolTM (DMPEG) 252 sources of supply 18, 33
SelexolTM technology 63, 67, 252, space velocity 58, 93
255–257 steam cracking process 19, 26, 28,
semi-regenerative reforming 91, 93, 94, 45–46, 111
96 steam cracking yields 112–113
SeparexTM membrane technology 205, steam cracking-based olefin complex 12
229 steam reforming 62, 64–66, 82–83, 111,
SeparexTM processes 64 126
SeparexTM spiral wound membrane 205 Stedman structured packing 195
shale gas 4, 20, 38, 47, 56, 111, 113 structure directing agents (SDA’s) 151,
shape-selective adsorption 159, 168–169 217
sheet metal 195 structured packing 186–187, 192,
Shell Claus Off-gas Treatment (SCOT) 194–195, 197
248 styrene 33, 113, 122–123, 138
short-chain paraffins 4 styrene-acrylonitrile (SAN) 81
Index 295
styrene-butadiene (SB) copolymer 30, tertiary amyl methyl ether (TAME) 27,
122 40, 54
styrene-butadiene rubber (SBR) 30 tetrahydrothiophen 6, 275
styrenics 21–22, 33–34 tetrahydrothiophene-1,1-dioxide 275
sulfolane process 110, 272, 274–275, 277 tetramethylbenzene 103
SulfolaneTM extractive distillation unit thermal dealkylation process (THDA)
172 95
sulfur compounds 6, 72 thermal regeneration 251, 254–255
sulfur-containing molecules 10 thermally-derived ethylene 19
superabsorbent polymers 25 thermocracking 8
supercritical fluid extraction (SCFE) thermodynamic vapor-liquid equilibrium
261, 263–264 (VLE) compositions 181
liquid-liquid extraction (LLE) 264 thin film composite (TFC) membranes
surface area 51, 89, 91–92, 148, 203, 207
153–154, 184–187, 190, 195, 206, flat sheet membrane 207–208, 227
hollow fiber membrane 207–208
215–217, 234, 243, 267
polymeric membranes 211–213
sweetening Merox process 278–280
structure 207
symmetric nonporous dense film
titanium silicates 214
polymeric membranes 207
toluene 3, 11, 17, 32–36, 87, 89, 95–105,
symmetric porous polymeric membranes
111–115, 118, 121, 127, 134, 162,
207
171–174, 197–199, 257, 264, 273
syndiotactic polypropylene 79
toluene diisocyanate (TDI) 35, 131–132
syngas 75
toluene methylation 96, 103–105
ammonia 82–83
toluene methylation unit 104
hydrogen 82
toluene xylenes 6, 197
methanol 83 toluol 11
syngas production topping refineries 8
coal gasification 67–72 transalkylation 32–33, 36–37, 95–100,
natural gas conversion 64–67 173
removal of impurities in raw natural gas Tray hydraulics 186
63–64 tray vs. packed columns 187, 188
synthesis gas 3, 11, 17 tri-ethylene glycol 78
ammonia 38–39 2,2,4-trimethylpentane 5
methanol 39–40 3,3,5-trimethylheptane 167
synthesis gas complexes 13 tri-substituted paraffin 167
synthetic fibers 133, 142, 170 tributyl phosphate (TBP) 270–273
synthetic gas (syngas) 45 solvent 272–273
triethanolamine (TEA) 151, 247
t trimethylbenzene 103
tacticity 79 trinitrotoluene 87
TatorayTM transalkylation process 96 typical naphtha-based olefins unit 12
terephthalic acid 12, 37, 133, 139
tertiary amines 244, 246, 248–249 u
tertiary amyl ethyl ether (TAEE) 54 Udex 110
296 Index
ultra-low sulfur diesel (ULSD) 109 virgin light crude oil 4

ultra-thin ZIF-8 membrane 218 volatile organic compounds (VOCs)
ultrafiltration (UF) membrane 204, 234 257, 259
ultrathin MFI nanosheets 217
UnicrackingTM technology 110, 197 w
UniflexTM 197 wax 4, 10
UnionfiningTM 58 weaker acidic zeolite 162
unsaturated polyester resins 80, 133, “world-scale” ethylene plant 20
138, 272
UOP C3 Oleflex Complex 51–54 x
UOP C3 /C4 Oleflex complex 56 X- or Y-faujasite 171
UOP CCR PlatformingTM process 91 xylenes 3, 11, 17, 36, 121
UOP Fixed Bed Platforming units 91 isomers 133
UOP InAlkTM process 55 oxidation reactions of 137
UOP Iso-C4 Oleflex Complex 54–56
UOP IsomarTM unit 100, 173 z
UOP MaxEneTM 47 zeolites 148
UOP Mixed-C4 Oleflex complex 56–57 acidity 161
UOP MX SorbexTM 148 adsorbents 148–150
UOP OleflexTM process 49 framework structure 159–160, 164
UOP PacolTM 57 membranes 204
UOP ParexTM 100, 148 synthesis 151
TM
UOP Penex Plus 89 zeolite powder 152
UOP PlatformingTM process 50 accretion of 153
UOP TatorayTM unit 96, 173 pelletization 153
UOP’s Pacol catalytic dehydrogenation zeolitic adsorbent 152, 154, 158, 162,
process 57 165, 169
UOP/Hydro MTO Process 61 zeolitic imidazolate frameworks (ZIF)
uranium and plutonium separation 272 213–214, 216, 218–219
zeolitic liquid separation mechanisms
v 154
vacuum distillation 7–9, 195, 197 Ziegler–Natta catalysts 76, 80
valve tray 190, 191 zirconium/hafnium separation
Vapor Liquid Equilibrium (VLE) theory 270–272
180–182 Zn2 (bim)4 nanosheet membrane 218
vapor/gas and vapor/vapor separations
227, 230–233

Santi Kulprathipanja, James E. Rekoske, Daniel Wei, Robert v. Slone, Trung Pham, Chunqing Liu - Modern Petrochemical Technology - Methods, Manufacturing and Applications-Wiley-VCH (2021)

Uploaded by

Copyright:

Available Formats

Santi Kulprathipanja, James E. Rekoske, Daniel Wei, Robert v. Slone, Trung Pham, Chunqing Liu - Modern Petrochemical Technology - Methods, Manufacturing and Applications-Wiley-VCH (2021)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Santi Kulprathipanja, James E. Rekoske, Daniel Wei, Robert v. Slone, Trung Pham, Chunqing Liu - Modern Petrochemical Technology - Methods, Manufacturing and Applications-Wiley-VCH (2021)

Uploaded by

Copyright:

Available Formats

WILEY END USER LICENSE AGREEMENT

Go to www.wiley.com/go/eula to access Wiley’s ebook EULA.

Methods, Manufacturing and Applications

Dr. Trung Pham © 2021 WILEY-VCH GmbH, Boschstr.

Print ISBN: 978-3-527-34522-9

Typesetting SPi Global, Chennai, India

Printed on acid-free paper

1 Reﬁnery and Petrochemical Processes 3

2.6.1 The Market for Ammonia 38

Part II Oleﬁns and Synthesis Gas 43

3 Oleﬁns and Synthesis Gas Production Technologies 45

4 Uses of Oleﬁns and Synthesis Gas 75

Part III Aromatics 85

5 Aromatic Production Technologies 87

6 Uses of Aromatics 121

6.3.3 p-Xylene Oxidation to Terephthalic Acid 139

Part IV Industrial Separation 145

8.2.4 Operating Modes of Distillation 186

10.4.1.1 DMEPG (Dimethyl Ether of Polyethylene Glycol,

11 Extraction Technology 261

The commercially established technologies can be reviewed and leveraged to

Finally, we review a couple of key commercialization examples for absorption:

Honeywell UOP Santi Kulprathipanja

Reﬁnery and Petrochemical Processes

Modern Petrochemical Technology: Methods, Manufacturing and Applications, First Edition.

1.2.1 Forms of Petroleum

1.2.1.1 Gaseous Petroleum: Gaseous Petroleum Can Be Deﬁned as Natural Gas

1.2.1.3 Solid Petroleum: Bitumen

1.2.2 Composition of Petroleum

Table 1.1 Chemical elements in crude oil.

Table 1.2 Hydrocarbon in crude oil.

general formula Cn H2n−6 . Examples of aromatics are benzene, toluene xylenes,

1.2.3 Petroleum Reﬁnery

Sulfar plant Sulfur

Sulfur removal Gasoline

Atmospheric Kerosene Hydrotreating

Atmospheric gas oil

Crude oil Atmospheric residue

Figure 1.1 Topping and hydroskimming reﬁnery.

Sulfar plant Sulfur

Sulfur removal Gasoline

Figure 1.2 Medium-conversion reﬁnery.

Sulfar plant Sulfur

Sulfur removal Gasoline

Figure 1.3 High-conversion reﬁnery.

1.2.4 Petroleum Products

Table 1.3 A breakdown of the petroleum products made from

Petroleum products Percent

Source: US Energy Information Administration, 31 May 2018.

1.3 Petrochemical Building Blocks

● C4 mono- and diolefins: This grouping is composed of butenes (also called

Benzene Toluene p-Xylene m-Xylene o-Xylene Ethylbenzene

Figure 1.4 Chemical structure of common petrochemical building blocks.

Feed Pyrolysis Quench and Cryogenic Demethanizer Deethanizer C2 splitter Ethylene

Figure 1.5 Block ﬂow diagram of typical naphtha-based oleﬁns unit.

Clay XC O-X Parex Isomer

Figure 1.6 Typical aromatics complex.

Purge gas (fuel)

Figure 1.7 Block ﬂow diagram of typical methanol plant.