Application of Ionic Liquids for Gas Sweetening

University of Calgary

PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2014-05-20

Application of Ionic Liquids for Gas Sweetening

Mortazavi Manesh, Soheil

Mortazavi Manesh, S. (2014). Application of Ionic Liquids for Gas Sweetening (Unpublished

doctoral thesis). University of Calgary, Calgary, AB. doi:10.11575/PRISM/24964

http://hdl.handle.net/11023/1531

doctoral thesis

University of Calgary graduate students retain copyright ownership and moral rights for their

thesis. You may use this material in any way that is permitted by the Copyright Act or through

licensing that has been assigned to the document. For uses that are not allowable under

copyright legislation or licensing, you are required to seek permission.

Downloaded from PRISM: https://prism.ucalgary.ca

UNIVERSITY OF CALGARY

Application of Ionic Liquids for Gas Sweetening

by

Soheil Mortazavi Manesh

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

CHEMICAL AND PETROLEUM ENGINEERING

CALGARY, ALBERTA

"May, 2014"

© Soheil Mortazavi Manesh 2014

iii

Abstract

In this study, general models were developed to predict the solubility of CO2, H2S, CH4

and C2H6 in ionic liquids (ILs) where no experimental data are available. These models

use fugacity functions based on asymmetric activity coefficients calculated using the

fundamental COSMO-RS method, the Peng-Robinson equation of state (PR-EOS), and an

empirical Henry’s constant of the solute in the IL. The Henry’s constant has been

correlated with the temperature and pressure of the system and physical properties of the

ILs. It was found that, for CO2 and CH4, the molecular weight (MW) of the ILs and for

H2S and C2H6, the surface area of the ILs are the best choice to correlate solubilities. 425

ILs were ranked based on their absorption capacity and selectivity of H2S and CO2 versus

CH4 and C2H6 absorption using the newly developed procedure. The top eight ILs were

selected and characterized for use in a commercial simulator. EOS’s binary interaction

parameters between solutes and IL were determined using the proposed model.

The performance of the ILs as solvents in gas sweetening plants is compared to

Morphysorb (a physical solvent) and MDEA (a selective chemical solvent) at similar gas

feed and product specifications. Among the candidate ILs, pmim-L appears to be the best

option for gas sweetening. ILs show better performance over MDEA and Morphysorb

when operating at high H2S compositions, that is for bulk removal of acid gases. In a case

study, the H2S concentration was reduced from 13% to 5% and the total heating and

pumping duty required for the pmim-L gas plant was 23 times less than the MDEA and 9

times less than the Morphysorb. Also, pmim-L required 89 times less cooling than

MDEA and 13 times less than a Morphysorb for the simulated gas plant. Furthermore, IL

iv

gas plants require negligible makeup solvent whereas MDEA plant requires 4.8 kg/hr

pure MDEA and 13 m3/hr makeup water. A Morphysorb plant required 84 kg/hr makeup

solvent. It was also shown that ILs are hydroscopic and can reduce the water content of

natural gas. With a few percent additional energy, the pmim-L gas plant can be converted

to a gas sweetening-dehydration plant which is able to meet the water content

specifications for natural gas pipelines. Alkanolamine plants require an additional

dehydration unit to produce sales gas. Based on the partial pressure of H2S in the feed

and product, guidelines have been provided to choose between MDEA and pmim-L gas

plants.

v

Acknowledgements

It is with heartfelt gratitude that I acknowledge the unconditional support I received from

my supervisor Dr. Robert Marriott through my time as a Ph.D. student. The studies in this

thesis simply would not have existed without his advice and inspiration. I will always

remember him for the friendly working atmosphere he provided for me.

I am grateful to Dr. Marco Satyro who contributed greatly to this project. His excellent

insights and recommendations were extremely helpful in this research. I feel incredibly

lucky that I have been able to learn from someone so insightful and knowledgeable.

I am in debt of Dr. William. Y. Svrcek, my M.Sc. supervisor and member of my Ph.D.

supervisory committee for his generous support and friendship during my M.Sc. and

Ph.D. at the University of Calgary.

I am also grateful to Dr. Harvey W Yarranton, member of my PhD supervisory

committee for his support.

My sincere appreciation to Dr. Amir Badakhshan for his inspirations. I am grateful to Dr.

Cyrus Ghotbi and Dr. Vahid Taghikhani my M.Sc. supervisors at Sharif University of

Technology in Iran.

This thesis was funded by several sources:

Department of Chemical Engineering award (2008). Dr. Satyro’s support (2008-2009).

Using this fund, I had trip to Germany to attend a short course in 2008. NSERC-IPS

scholarship in collaboration with Virtual Material Group (2009-2012). Using the funding

vi

from the NSERC ASRL IRC in Applied Sulfur Chemistry and Travel grant of University

of Calgary, I presented my work in IL conference in Spain 2011. Dr. Marriott’s support

(2012-2013). Also using this fund I presented my work in 62nd Canadian Chemical

Engineering Conference in Vancouver in 2012. The funding from my parents throughout

my PhD program made it possible to complete my research. I have to thank ASRL for

providing the office space (2011-2012).

I would also like to acknowledge Mr. Paul Davis, Dr. Norman Dowling and Dr. Peter

Clark for the discussions on absorption selectivity of solvents. Thanks to Virtual Material

Group for providing the VMGSim software.

I would like to acknowledge Dr. Andreas Klamt and Dr. Karin Wichmann for the

discussions about COSMO-RS calculations and COSMOtherm software.

I am also grateful to Dr. Jalel Azaiez for the excellent discussions in Advance

Mathematic course. I would also thank Dr. Brent Young for technical discussions in plant

design and advanced control. My gratitude also extends to Dr. Brij Maini, Dr. Ayodeji A.

Jeje, Dr. Mehran Pouladdi-Darvish, Dr. Michael Foley, Dr. Hasan Hasanzadeh, Dr. Jalal

Abedi and Dr. M. A. Trebble. Thanks to all of the staff of the Department of Chemical

and Petroleum Engineering for their support.

I gratefully acknowledge Shell Canada Energy for providing time and support to

complete this study.

I would like to acknowledge the staff of Alberta Sulfur Research Ltd. (ASRL) for their

support. Thanks to Krystyna Ciesluk, senior administrator of ASRL. She was an amazing

vii

help to me while I was in ASRL, answering my questions and helping me with all sorts of

things. Thank you Krystyna.

Thanks to Dr. Abazar Shamekhi for the scientific discussions and moral support.

Finally and for most of all, I would like to thank my wife, my sister and my parents, for

their love and their continual support of my work. Only because of them and their

inspirations I was able to keep going during the tough times.

viii

Dedication

To

my parents,

my sister

my wife

ix

x

Table of Contents

Approval Page ..................................................................................................................... ii

Abstract .............................................................................................................................. iii

Acknowledgements ..............................................................................................................v

Table of Contents .................................................................................................................x

List of Tables ................................................................................................................... xiii

List of Figures ....................................................................................................................xv

List of Symbols, Abbreviations and Nomenclature ........................................................ xxii

CHAPTER ONE: INTRODUCTION ................................................................................27

1.1 Background ..............................................................................................................27

1.1.1 Gas Quality Specifications ..............................................................................28

1.1.2 Gas Sweetening ...............................................................................................28

1.1.3 Absorption into a Solvent ................................................................................29

1.1.3.1 Physical Solvents ...................................................................................29

1.1.3.2 Chemical Solvents .................................................................................29

1.2 Ionic Liquids ............................................................................................................32

1.3 Objective ..................................................................................................................32

1.4 Thesis Overview ......................................................................................................33

1.5 Summary ..................................................................................................................34

CHAPTER TWO: IONIC LIQUIDS .................................................................................35

2.1 Introduction ..............................................................................................................35

2.2 IL Structure ..............................................................................................................35

2.3 Applications .............................................................................................................37

2.3.1 ILs as Electrolytes ...........................................................................................37

2.3.2 ILs as Solvents .................................................................................................38

2.3.3 ILs as Supported Liquid Membrane (SLM) ....................................................38

2.3.4 ILs in Gas Chromatography (GC) ...................................................................38

2.3.5 ILs as Natural Gas Solvent ..............................................................................39

2.4 Summary ..................................................................................................................39

CHAPTER THREE: INTRODUCTION TO COSMO AND COSMO-RS MODELS ......41

3.1 Introduction ..............................................................................................................41

3.2 Representation of Particles in Quantum Mechanics ................................................42

3.2.1 Basis Sets .........................................................................................................45

3.2.1.1 Minimum Basis Set ................................................................................46

3.2.1.2 Triple-Zeta (TZ) Basis Function ............................................................47

3.2.1.3 Split-Valence Basis Set (SV) ..................................................................47

3.3 Energy Calculations of Many Electron Systems .....................................................48

3.3.1 Hartree-Fock Theory (HF) ..............................................................................48

3.3.2 Density Functional Theory (DFT) ...................................................................51

3.4 Conductor-like-Screening-Mode (COSMO) Theory ...............................................55

3.5 COSMO for Real Solvents (COSMO-RS) ................................................................58

3.6 Generating Chemical Potential and Activity Coefficient of Solute Based on

COSMO-RS ............................................................................................................61

xi

3.7 Summary ..................................................................................................................65

CHAPTER FOUR: THE SOLUBILITY OF CO2, H2S, CH4 AND C2H6 IN IONIC

LIQUIDS...................................................................................................................66

4.1 Introduction ..............................................................................................................66

4.2 Database ...................................................................................................................68

4.3 COSMO Calculations ..............................................................................................75

4.4 Solubility Model ......................................................................................................76

4.4.1 Bubble Point Pressure as a Measure of the Solubility of a Gas into an IL ......76

4.5 Solubility of CO2 in ILs ...........................................................................................78

4.5.1 Maiti’s Model [124] ........................................................................................79

4.5.2 Mortazavi-Manesh et. al.’s Model-1 [117] ......................................................82

4.5.3 Mortazavi-Manesh et. al.’s Model-2 [116] ......................................................87

4.6 Solubility of H2S, CH4 and C2H6 in ILs .................................................................102

4.7 Screening ILs Based on the Solubility of CO2, H2S, CH4 and C2H6 .....................117

4.8 Selectivity of absorption ........................................................................................123

4.9 Summary ................................................................................................................130

CHAPTER FIVE: CONCEPTUAL DESIGN OF GAS PLANTS ..................................132

5.1 Introduction ............................................................................................................132

5.2 Defining ILs for the Commercial Simulators ........................................................133

5.2.1 Estimation of Critical Properties of ILs .........................................................134

5.2.2 Validation of the VMGSim’s Cp Calculations ..............................................136

5.3 Equation of State’s Set up for Mixtures .................................................................138

5.4 Kwoen Gas Plant ...................................................................................................145

5.4.1 IL Gas Sweetening Plant ...............................................................................147

5.4.1.1 Description of the units in the IL gas sweetening plant .......................148

5.4.2 Gas Plant with Morphysorb as the Absorbent ...............................................152

5.4.2.1 Description of the units in the Morphysorb gas sweetening plant .......153

5.4.3 Gas Plant with Chemical Solvent ..................................................................156

5.4.3.1 Description of the Units of the MDEA Gas Sweetening Plant ............157

5.5 Simultaneous Dehydration and Sweetening using ILs ..........................................174

5.5.1 Gas Sweetening and Dehydration using MDEA and TEG ...........................176

5.6 Shale Gas Case Study ............................................................................................185

5.6.1 Description of the IL Gas Sweetening Plant .................................................186

5.6.1.1 Inlet Separator and Absorber ...............................................................186

5.6.1.2 Recovering the Absorbed Hydrocarbons .............................................188

5.6.1.3 IL Regeneration ...................................................................................188

5.7 Summary ................................................................................................................200

CHAPTER SIX: SUMMARY .........................................................................................205

6.1 Introduction ............................................................................................................205

6.2 Geometry Optimization of ILs and Activity Coefficient Calculations ..................205

6.3 Solubility Models ...................................................................................................206

6.4 Conceptual IL-Gas Plant Simulation ....................................................................208

CHAPTER SEVEN: RECOMMENDED FUTURE WORK ..........................................212

xii

7.1 Introduction ............................................................................................................212

7.2 Melting Point .........................................................................................................212

7.3 Thermal Decomposition ........................................................................................212

7.4 VLE Measurements ...............................................................................................213

7.5 Chemical Reactions ...............................................................................................213

7.6 Mass Transfer Rate and Tray Efficiencies .............................................................213

7.7 Viscosity of ILs ......................................................................................................214

7.8 Corrosivity .............................................................................................................214

7.9 Heat Capacity and the Heat of Solvation ...............................................................214

7.10 Toxicity ................................................................................................................214

7.11 Cost of ILs ...........................................................................................................215

REFERENCES ................................................................................................................216

APPENDIX A: COSMO CALCULATIONS ..................................................................227

APPENDIX B: .................................................................................................................269

Peng-Robinson Equation of State[128] .......................................................................269

Soave–Redlich–Kwong equation of state [125] ..........................................................270

Advanced Peng Robinson Equation of State[152] ......................................................272

xiii

List of Tables

Table 1-1. Gas Quality Specifications[2] .......................................................................... 28

Table 2-1. Melting point of some common salts .............................................................. 35

Table 4-1 Experimental data and conditions for CO2-IL mixtures [115, 116] ................. 69

Table 4-1 Continued ......................................................................................................... 70

Table 4-1 Continued ......................................................................................................... 71

Table 4-2. Experimental data and conditions for H2S-IL mixtures .................................. 73

Table 4-3. Experimental data conditions for CH4-IL mixtures ......................................... 74

Table 4-4. Experimental data conditions for C2H6-IL mixtures ....................................... 74

Table 4-5. Methods for predicting the total pressure of 31 CO2-IL mixtures ................... 84

Table 4-6. Experimental total pressure of CO2-IL mixtures [126] vs. calculated

pressure. Comparison between Maiti’s model [123], Mortazavi-Manesh model-2

[115] and Mortazavi-Manesh model-1[116] using for the data which are not

included in the regression ......................................................................................... 85

Table 4-7. Predicting the equilibrium pressure of CO2-IL for 27 mixtures [54-59, 61-

66, 70, 72-74, 117, 118] using different parameters in Equations (4-14) and

(4-16) [115] ............................................................................................................... 93

Table 4-7. Continued ........................................................................................................ 94

Table 4-7. Continued ........................................................................................................ 95

Table 4-8. Different Parameter Combinations for Predicting the Total Pressure of

H2S-IL Mixtures ...................................................................................................... 104

Table 4-8. Continued ...................................................................................................... 105

Table 4-8. Continued ...................................................................................................... 106


CH4-IL Mixtures ..................................................................................................... 107

Table 4-9. Continued ...................................................................................................... 108

Table 4-9. Continued ...................................................................................................... 109


C2H6-IL Mixtures .................................................................................................... 110

xiv

Table 4-10. Continued .................................................................................................... 111

Table 4-10. Continued .................................................................................................... 112

Table 4-11. Recommended Parameters for CO2-IL, H2S-IL, CH4-IL and C2H6-IL ....... 113

Table 4-12. AAR% for the activity coefficients calculated between fitted NRTL model

and COSMO-RS model for different solutes in 425 ILs ......................................... 118

Table 4-13. ILs that are within the top 28th

percentile for five selectivities important

for sour gas treatment (4/2 CHSHS ,

2/2 COSHS , 62/2 HCSHS ,

4/2 CHCOS , 62/2 HCCOS ) ........ 127

Table 4-13. Continued .................................................................................................... 128

Table 4-13. Continued .................................................................................................... 128

Table 5-1. ILs chosen for gas processing. The CT Cp ω and

K298.15ρ were estimated

using Valderrama et. al. model[143-145] ............................................................... 136

Table 5-2. Comparison between the experimental data and VMGsim predictions of the

heat capacity of bmim-CH3SO4 .............................................................................. 138

Table 5-3. APR’s binary interaction parameters of Equation (5-7) for solute-IL

mixtures ................................................................................................................... 142

Table 5-3. Continued ...................................................................................................... 143

Table 5-3. Continued ...................................................................................................... 144

Table 5-4. Conditions considered for the gas plants ....................................................... 147

Table 5-5. Pressure of the absorber and flash tank for IL-gas plant ............................... 152

Table 5-6. Published[156] using Morphysorb versus calculated using NFM

composition of the upgraded gas and acid gas for Kwoen gas plant ...................... 153

Table 5-7. Pressure of the flash tanks for the Morphysorb-gas plant [156] ................... 154

Table 5-8. Typical shale gas property ............................................................................. 186

Table 5-9. Performance Summary of pmim-L, Morphysorb and MDEA gas

sweetening plants based on feed and product specifications of Kwoen case study 202

Table 5-10. Performance Summary of pmim-L and MDEA-TEG gas sweetening-

dehydration plants at fixed mole percent of H2S (5.33%) and water content

(4lb/MMscf) in the dry upgraded gas ...................................................................... 203

xv

List of Figures

Figure 1-1 Emission reductions achieved by using natural gas; : Displacing 1 kWh

of coal-based electricity with natural gas : Displacing 1 Gasoline Gallon

Equivalent (GGE) of vehicle fuel with natural gas[1] .............................................. 27

Figure 1-2. Schematic comparison of the loading of physical and chemical solvents at

different partial pressures of acid gas ....................................................................... 30

Figure 1-3. Different types of amines, a: primary; b: secondary; c: tertiary .................... 31

Figure 2-1. [1-ethyl-3-methylimidazolium][Bis(trifluoromethylsulfonyl)-imide] or

[emim][Tf2N] ............................................................................................................ 36

Figure 3-1 Charge distribution on the molecular cavity of water after COSMO/DFT

calculations[92]. Dark red represents higher electron density and dark blue

represents lower electron density. ............................................................................. 58

Figure 3-2. -profile of water based on COSMO calculation[92] .................................. 59

Figure 4-1. Schematic of the P-X diagram at constant temperature of solute A in two

ILs in which solubility of A in IL1 is higher than the solubility of A in IL2, ─:

IL1; ─: IL2 ................................................................................................................ 77

Figure 4-2. Schematic of the P-X diagram at constant temperature of solute A or B in

an IL in which solubility of A in IL is higher than the solubility of B in IL; ─: A-

IL; ─: B-IL ................................................................................................................ 78

Figure 4-3. Experimental vs. calculated total pressure of CO2-IL mixtures using the

Maiti’s model [123] : bmim-Tf2N [55, 56, 62], : emim-Tf2N [70], : hmim-

Tf2N [118], : omim-Tf2N [55], : b2Nic-Tf2N [54], : bmim-BF4 [55, 56], :

bmim-CH3SO4 [57], : bmim-EtGLEtGLeC2SO4 [54], : bmim-NO3 [55], :

bmim-PF6 [55, 58, 59, 61, 117], : bmim-TFA [54], : bmim-Triflate[55], :

bmmim-PF6 [59], :bmmim-BF4 [59], : C6H4F9mim-Tf2N [54], : emim-BF4

[63, 64], : emim-PF6 [66], : emim-Triflate [65], : emmim-Tf2N [59], :

hmim-Triflate [72], : hmpy-Tf2N [54], : MeBu3N-Tf2N [56], : MeButPyrr-

Tf2N [56], : N4111-Tf2N [54], : N4444-doc [54], : N-bupy-BF4 [73], :

pmim-Tf2N [74], : HOemim-BF4 [75], : HOemim-Tf2N [69], +: HOemim-PF6

[68], : emim-C2SO4 [67, 68] ................................................................................... 81

Figure 4-4. Experimental vs. calculated total pressure of CO2-IL mixtures Equations

(4-4) and (4-6) [116], : bmim-Tf2N [55, 56, 62], : emim-Tf2N [70], : hmim-

Tf2N [118], : omim-Tf2N [55], : b2Nic-Tf2N [54], : bmim-BF4 [55, 56], :

bmim-CH3SO4 [57], : bmim-EtGLEtGLeC2SO4 [54], : bmim-NO3 [55], :


bmmim-PF6 [59], :bmmim-BF4 [59], : C6H4F9mim-Tf2N [54], : emim-BF4

[63, 64], : emim-PF6 [66], : emim-Triflate [65], : emmim-Tf2N [59], :

xvi


Tf2N [56], : N4111-Tf2N [54], : N4444-doc [54], : N-bupy-BF4 [73], :

pmim-Tf2N [74], : HOemim-BF4 [75], : HOemim-Tf2N [69], +: HOemim-PF6

[68], : emim-C2SO4 [67, 68] ................................................................................... 86

Figure 4-5. Experimental total pressure of CO2-IL mixtures vs. calculated pressure

using Equation (4-14) [115] without any IL parameters, : bmim-Tf2N [55, 56,

62], : emim-Tf2N [70], : hmim-Tf2N [118], : omim-Tf2N [55], : b2Nic-Tf2N

[54], : bmim-BF4[55, 56], : bmim-CH3SO4 [57], : bmim-

EtGLEtGLeC2SO4[54], : bmim-NO3[55], : bmim-PF6 [55, 58, 59, 61, 117], :

bmim-TFA [54], : bmim-Triflate[55], : bmmim-PF6[59], : C6H4F9mim-Tf2N

[54], : emim-BF4 [63, 64], : emim-PF6[66], : emim-Triflate[65], :bmmim-

BF4[59], : emmim-Tf2N [59], : hmim-Triflate [72], : hmpy-Tf2N [54], :

MeBu3N-Tf2N [56], : MeButPyrr-Tf2N [56], : N4111-Tf2N [54], : N4444-

doc[54], : N-bupy-BF4[73], : pmim-Tf2N[74]....................................................... 91

Figure 4-6 Experimental total pressure of CO2-IL mixtures vs. calculated pressure

using Equation (4-14) and 6-parameter Equation (4-16) with MW as IL

parameter[115], : bmim-Tf2N [55, 56, 62], : emim-Tf2N [70], : hmim-Tf2N

[118], : omim-Tf2N [55], : b2Nic-Tf2N [54], : bmim-BF4[55, 56], : bmim-

CH3SO4 [57], : bmim-EtGLEtGLeC2SO4[54], : bmim-NO3[55], : bmim-PF6

[55, 58, 59, 61, 117], : bmim-TFA [54], : bmim-Triflate[55], : bmmim-

PF6[59], : C6H4F9mim-Tf2N [54], : emim-BF4 [63, 64], : emim-PF6[66], :

emim-Triflate[65], :bmmim-BF4[59], : emmim-Tf2N [59], : hmim-Triflate

[72], : hmpy-Tf2N [54], : MeBu3N-Tf2N [56], : MeButPyrr-Tf2N [56], :

N4111-Tf2N [54], : N4444-doc[54], : N-bupy-BF4[73], : pmim-Tf2N[74] .............. 96



parameter only included in [115], : bmim-Tf2N [55, 56, 62], : emim-Tf2N

[70], : hmim-Tf2N [118], : omim-Tf2N [55], : b2Nic-Tf2N [54], : bmim-

BF4[55, 56], : bmim-CH3SO4 [57], : bmim-EtGLEtGLeC2SO4[54], : bmim-

NO3[55], : bmim-PF6 [55, 58, 59, 61, 117], : bmim-TFA [54], : bmim-

Triflate[55], : bmmim-PF6[59], : C6H4F9mim-Tf2N [54], : emim-BF4 [63,

64], : emim-PF6[66], : emim-Triflate[65], :bmmim-BF4[59], : emmim-Tf2N

[59], : hmim-Triflate [72], : hmpy-Tf2N [54], : MeBu3N-Tf2N [56], :

MeButPyrr-Tf2N [56], : N4111-Tf2N [54], : N4444-doc[54], : N-bupy-BF4[73],

: pmim-Tf2N[74] ..................................................................................................... 97


using Equation (4-14) and 6-parameter Equation (4-16) with COSMO energy as

IL parameter[115], : bmim-Tf2N [55, 56, 62], : emim-Tf2N [70], : hmim-

Tf2N [118], : omim-Tf2N [55], : b2Nic-Tf2N [54], : bmim-BF4[55, 56], :

bmim-CH3SO4 [57], : bmim-EtGLEtGLeC2SO4[54], : bmim-NO3[55], :


bmmim-PF6[59], : C6H4F9mim-Tf2N [54], : emim-BF4 [63, 64], : emim-

PF6[66], : emim-Triflate[65], :bmmim-BF4[59], : emmim-Tf2N [59], :

xvii


Tf2N [56], : N4111-Tf2N [54], : N4444-doc[54], : N-bupy-BF4[73], : pmim-

Tf2N[74] .................................................................................................................... 98



IL parameter only included in [115], : bmim-Tf2N [55, 56, 62], : emim-

Tf2N [70], : hmim-Tf2N [118], : omim-Tf2N [55], : b2Nic-Tf2N [54], :

bmim-BF4[55, 56], : bmim-CH3SO4 [57], : bmim-EtGLEtGLeC2SO4[54], :

bmim-NO3[55], : bmim-PF6 [55, 58, 59, 61, 117], : bmim-TFA [54], :

bmim-Triflate[55], : bmmim-PF6[59], : C6H4F9mim-Tf2N [54], : emim-BF4

[63, 64], : emim-PF6[66], : emim-Triflate[65], :bmmim-BF4[59], : emmim-

Tf2N [59], : hmim-Triflate [72], : hmpy-Tf2N [54], : MeBu3N-Tf2N [56], :

MeButPyrr-Tf2N [56], : N4111-Tf2N [54], : N4444-doc[54], : N-bupy-BF4[73],

: pmim-Tf2N[74] ..................................................................................................... 99



IL parameter[115], : bmim-Tf2N [55, 56, 62], : emim-Tf2N [70], : hmim-

Tf2N [118], : omim-Tf2N [55], : b2Nic-Tf2N [54], : bmim-BF4[55, 56], :

bmim-CH3SO4 [57], : bmim-EtGLEtGLeC2SO4[54], : bmim-NO3[55], :


bmmim-PF6[59], : C6H4F9mim-Tf2N [54], : emim-BF4 [63, 64], : emim-

PF6[66], : emim-Triflate[65], :bmmim-BF4[59], : emmim-Tf2N [59], :


Tf2N [56], : N4111-Tf2N [54], : N4444-doc[54], : N-bupy-BF4[73], : pmim-

Tf2N[74] .................................................................................................................. 100

Figure 4-11. Correlation between AARArea/AARMW and polarizability for CO2 (■), H2S

(), CH4 (▲) and C2H6 (). .................................................................................... 103

Figure 4-12. Experimental total pressure of H2S-IL mixtures vs. calculated pressure

using Equations (4-13) , (4-14) and (4-16) with surface area as the IL parameter,

: bmim-BF4[77], : bmim-PF6 [76, 77], : bmim-Tf2N [77], : emim-C2SO4

[68], : emim-PF6 [80], : emim-Tf2N[80], : hmim-BF4 [78], : hmim-Tf2N

[78], : hmim-PF6 [78], : HOemim-BF4 [75], : HOemim-Triflate [79], +:

HOemim-PF6 [79], : HOemim-Tf2N [79], : omim-Tf2N [81] .......................... 114

Figure 4-13. Experimental total pressure of CH4-IL mixtures vs. calculated pressure

using Equations (4-13) , (4-14) and (4-16) with MW as the IL parameter, :

bmim-BF4[82], : bmim-CH3SO4 [83], +: bmim-PF6 [58, 85], : hmim-Tf2N

[84] .......................................................................................................................... 115

Figure 4-14. Experimental total pressure of C2H6-IL mixtures vs. calculated pressure

using Equations (4-13) , (4-14) and (4-16) with surface area as the IL parameter,

: bmim-Tf2N[56], : hmim-Tf2N[88, 120], +: bmim-PF6[56, 58, 85], : bmim-

BF4[82] .................................................................................................................... 116

xviii

Figure 4-15. Solubility of CO2 in ILs at T = 298.15 K and 2000 kPa for different

combinations of anions and cations. Henry’s constants are calculated based on

MW of IL [115] ....................................................................................................... 119

Figure 4-16. Solubility of H2S in ILs at T = 298.15 K and 2000 kPa for different


surface area of ILs. .................................................................................................. 120

Figure 4-17. Solubility of CH4 in ILs at T = 298.15 K and 2000 kPa for different


the MW of ILs. ........................................................................................................ 121

Figure 4-18. Solubility of C2H6 at T = 298.15 K and 2000 kPa for different


surface area of ILs. .................................................................................................. 122

Figure 4-19. Investigating the selectivity of different combinations of ILs at 298.15 K

and 2000 kPa. : ILs that are within the top 28th

percentile for five selectivities

important for sour gas treatment (4/2 CHSHS ,


4/2 CHCOS ,

62/2 HCCOS ). ............................................................................................................... 126

Figure 5-1. Cp of bmim-CH3SO4, : experimental [148]; ─: VMGSim[137]

predictions ............................................................................................................... 137

Figure 5-2. Comparison of total pressure in water-[emim][Tf2N] mixture. ─ :

COSMO-RS method verses experimental data. Calculations are done using

COSMOthermX software[92], ○:[153] ,▲:[154]. AAR = % 5.6 ............................ 140

Figure 5-3. Comparison of total pressure in water-[emim][C2SO4] mixture. ─:

COSMO-RS method verses experimental data [155]. Calculations are done using

COSMOthermX software, : 322.9 K, ■: 312.9 K, ▲: 302.9 K. AAR = %19.1 .... 140

Figure 5-4. Approximate locations of Kwoen and Pine River gas plants [157]. ............ 145

Figure 5-5. Gas sweetening plant designed in this study using IL as the absorbent.

The simulation is done using VMGSim[137]. T-1: inlet separator; T-2, T-3, T-4,

T-5: Flash tanks; H-1, H-2: Heaters, P-1: Pump; CP-1, CP-2, CP-3:

Compressors; AC-1, AC-2, AC-3 air-coolers. ........................................................ 150

Figure 5-6. The chemicals in Morphysorb[4, 156]; a: N-Formylmorpholine (NFM),

b:Acetylmorpholine (NAM) ................................................................................... 152

Figure 5-7. Gas sweetening plant using Morphysorb as the absorbent. The simulation

is done using VMGSim[137] . T-1: inlet separator; T-2, T-3, T-4, T-5: Flash

tanks; H-1, H-2: Heaters, P-1: Pump; CP-1, CP-2, CP-3: Compressors; AC-1,

AC-2, AC-3 air-coolers ........................................................................................... 155

xix

Figure 5-8. Gas sweetening plant using MDEA as the absorbent using VMGSim[137].

T-1: inlet separator; T-2: Flash tank; Hx-1: Heat exchanger; C-1: cooler; P-1:

pump ....................................................................................................................... 158

Figure 5-9. CO2 mole percent in the upgraded gas for different ILs, Morphysorb and

amine (45% wt MDEA) gas plants at fixed H2S mole percent in the upgraded gas

(5.33%) .................................................................................................................... 160

Figure 5-10. CH4 mole percent in the upgraded gas for different ILs, Morphysorb and


(5.33%) .................................................................................................................... 161

Figure 5-11. CH4 mass flow rate in the upgraded gas for different ILs, Morphysorb

and amine (45% wt MDEA) gas plants at fixed H2S mole percent in the

upgraded gas (5.33%) ............................................................................................. 162

Figure 5-12. Solvent flow rate for different ILs, Morphysorb and amine (45% wt

MDEA) gas plants at fixed H2S mole percent in the upgraded gas (5.33%) .......... 163

Figure 5-13. Loading of rich solvent from the absorber, Equation (5-9), for different

ILs, Morphysorb and amine (45% wt MDEA) gas plants at fixed H2S mole

percent in the upgraded gas (5.33%) ....................................................................... 164

Figure 5-14. Loading of lean solvent from the absorber, Equation (5-9), for different



Figure 5-15. Solvent make up for different ILs, Morphysorb and amine (45% wt


Figure 5-16. Water mole percent in upgraded gas for different ILs, Morphysorb and


(5.33%) .................................................................................................................... 167



(5.33%) .................................................................................................................... 167

Figure 5-18. Water flow rate in up-graded gas for different ILs, Morphysorb and

amine (45% wt MDEA) gas plants at fixed H2S mole percent in the upgraded

gas (5.33%) ............................................................................................................. 168

Figure 5-19. Water flow rate in upgraded gas for different ILs, Morphysorb and

amine (45% wt MDEA) gas plants at fixed H2S mole percent in the upgraded

gas (5.33%) ............................................................................................................. 168

Figure 5-20. Required heating energy for different ILs and Morphysorb gas plants at

fixed H2S mole percent in the upgraded gas (5.33%) ............................................. 169

xx

Figure 5-21. Required heating energy for different ILs Morphysorb and amine (45%

wt MDEA) gas plants at fixed H2S mole percent in the upgraded gas (5.33%) ..... 170

Figure 5-22. Power consumption of pumping for different ILs, Morphysorb and


(5.33%) .................................................................................................................... 171

Figure 5-23. Compression power for different ILs, Morphysorb and amine (45% wt


Figure 5-24. Cooling required for (Air coolers and other cooling units) for different



Figure 5-25. Simultaneous gas sweetening and dehydration plant using IL as the

absorbent. The simulation is done using VMGSim [137]. T-1: inlet separator; T-

2, T-3, T-4, T-5: Flash tanks; C-1: Cooler; H-1, H-2: Heaters, P-1: Pump; CP-1,

CP-2, CP-3: Compressors; AC-1, AC-2, AC-3 air-coolers. ................................... 175

Figure 5-26. Triethyleneglycol (TEG) ............................................................................ 176

Figure 5-27. Gas sweetening-dehydration plant using MDEA as the sweetening and

TEG is used for dehydration. The simulation is done using Unisim Design [163].

T-1: inlet separator; T-2: Flash tank; HX-1, HX-2: Heater exchangers, P-1, P-2:

Pumps; C-1, C-2, C-3: Coolers ............................................................................... 178

Figure 5-28. H2S mass flow rate in upgraded gas IL-sweetening-dehydration gas

plants and MDEA-TEG-sweetening-dehydration gas plant at fixed mole percent

of H2S (5.33%) and water content (4lb/MMscf) in the dry upgraded gas. .............. 180

Figure 5-29. CO2 mole percent in upgraded gas IL-sweetening-dehydration gas plants

and MDEA-TEG-sweetening-dehydration gas plant at fixed mole percent of H2S

(5.33%) and water content (4lb/MMscf) in the dry upgraded gas. .......................... 180

Figure 5-30. CO2 mass flow rate in upgraded gas IL-sweetening-dehydration gas



Figure 5-31. CH4 mole percent in upgraded gas IL-sweetening-dehydration gas plants

and MDEA-TEG-sweetening-dehydration gas plant at fixed mole percent of H2S

(5.33%) and water content (4lb/MMscf) in the dry upgraded gas. .......................... 181

Figure 5-32. CH4 mass flow rate in upgraded gas IL-sweetening-dehydration gas



Figure 5-33. Comparison energy consumption and solvent flow rate for in upgraded

gas IL-sweetening-dehydration gas plants at fixed mole percent of H2S (5.33%)

xxi

and water content (4lb/MMscf) in the dry upgraded gas; : Heating requirement;

: Cooling requirement; ■: power required for compressors; ◊: Pumping power

requirement; ▲: IL standard volume flow rate ...................................................... 183


gas IL-sweetening-dehydration gas plants and MDEA-TEG-sweetening-

dehydration gas plant at fixed mole percent of H2S (5.33%) and water content

(4lb/MMscf) in the dry upgraded gas; : Heating requirement; : Cooling

requirement; ■: power required for compressors; ◊: Pumping power requirement 184

Figure 5-35. IL gas plant for Shale gas sweetening. The simulation is done using

VMGSim[137]. T-1: inlet separator; T-2, T-3, T-3, T-4, T-5, T-6: Flash tank;

HX-1, HX-2: Heater exchanger, H-1: Heater, P-1: Pump; CP-1, CP-2, CP-3:

Compressor; AC-1, AC-2, AC-3, AC-4: Cooler. .................................................... 187

Figure 5-36. Solvent flow rate at different H2S content of upgraded gas for shale gas

sweetening plants; : pmim-L; :MDEA; ○: TEG ................................................ 189

Figure 5-37. Power consumption at different H2S content of upgraded gas for shale

gas sweetening plant using pmim-L; ◊: compression; : pumping; : cooling; ○:

reboiler .................................................................................................................... 190


gas sweetening plant using MDEA-TEG; ■: pumping; ▲: cooling; : heating .... 191

Figure 5-39. Mass flow rate of CO2 and CH4 in acid gas at different H2S content of

upgraded gas for shale gas sweetening plant using pmim-L; : CO2; ○: CH4 ........ 192

Figure 5-40, Indirect reaction of a tertiary amine with CO2; reaction (a) is slow and is

the controlling reaction ........................................................................................... 192

Figure 5-41. Mass flow rate of CO2 and CH4 in upgraded gas at different H2S content

of upgraded gas for shale gas sweetening plant using pmim-L; : CO2 in pmim-L

gas plant; ○: CH4 in pmim-L gas plant; : CO2 in MDEA-TEG gas plant; ○: CH4

in MDEA-TEG gas plant; ....................................................................................... 193

Figure 5-42. Mole percent of CO2 and CH4 in upgraded gas at different H2S content


gas plant; ○: CH4 in pmim-L gas plant; : CO2 in MDEA-TEG gas plant; ○: CH4

in MDEA-TEG gas plant; ....................................................................................... 194

Figure 5-43. H2S loading in the solvent at different H2S partial pressure;○: amine

(45% wt MDEA); : pmim-L; ■:Equivqlent to H2S feed composition presented

in Table 5-4 , (13.6 mole % H2S); ▲: Equivqlent to H2S composition in the

treated gas presented in Table 5-4 , (5.3 mole % H2S); : Equivalent to feed

composition in Shale gas case study, (500 ppm H2S); : Equivalent to treated

gas composition in Shale gas case study, (100 ppm H2S); ..................................... 196

xxii

List of Symbols, Abbreviations and Nomenclature

Symbol Definition

A Surface area

A Coefficient of Equation (4-12)

AAR Absolute average relative error

AAD Absolute average deviation

AM Coefficient of Equation (5-2)

AO Atomic orbital

APR Advanced-Peng-Robinson

B Coefficient of Equation (4-12)

b2Nic 1-butyl-nicotinic acid butyl ester

BF4 Tetraflouroborate

BM Coefficient of Equation (5-2)

bmim 1-butyl-3-methylimidazolium

bmmim 1-butyl-2,3-dimethylimidazolium

BP Becke-Perdew

Cl Chloride

CH3SO4 Methylsulfate

CH4 Methane

C2H6 Ethane

C2SO4 Ethylsulfate

C6H4F9mim 1-methyl-3-(3,3,4,4,5,5,6,6,6-nonafluorohexyl)imidazolium

C8SO4 Octylsulfate

CHB Hydrogen bond strength

Ccf3

COSMO Conductor-like-screening-model

COSMO-RS Conductor-like screening model for realistic solvation

CO2 Carbon dioxide

DBP Dibutylphosphate

DEA Diethanolamine

DEP Diethylphosphate

DFT Density functional theory

Doc Docusate

E Energy of the system

EHB Hydrogen bonding interaction energy

EM Coefficient of Equation(5-4)

emim 1-ethyl-3-methylimidazolium

emmim 1-ethyl-2,3-dimethylimidazolium

EOS Equation of state

EvdW Van der Waals interactions interaction energy

EtGLEtGLeC2SO4 2-(2-methoxyethoxy)ethyl sulfate

ETT S-Ethyl-tetramethylisothiouronium

EXC Exchange energy

xxiii

FEP Tris(pentafluoroethyl)trifluorophosphate v

if Fugacity of component i in vapor phase

l

if Fugacity of component i in liquid phase

G Gibbs free energy

GC Gas chromatography

GGE Gasoline Gallon Equivalent

GTO Gaussian type orbitals

HF Hartree-Fock

hexc Total energy operator

H Hamiltonian operator

Hi Henry’s constant

HOemim 1-(2-hydroxyethane)-3-methylimidazolium

H2O Water

H2S Hydrogen sulfide

hmim 1-hexyl-3-methylimidazolium

hmg Hexamethylguanidinium

hmpy 1-hexyl-3-methylpyridinium

IL Ionic liquid

l Angular momentum quantum number

L Lactate

KE Kinetic energy operator

kij Binary interaction parameter

0ijk Coefficient of Equation (5-7)



KS Kohn-Sham

m Mass of the particle

MDEA Methyldiethylethanol

MEA Monoethanol amine

MeBu3N Methyl-tributylammonium

MeButPyrr 1-butyl-1-methylpyrrolidinium

lm Angular momentum quantum number

MMscfd Million standard cubic feet per day

MO Molecular orbital

ms Spin quantum number

MW Molecular weight

n Principal quantum number

ni Moles of i

N Normalization constant

N2311 Ethyl-propyl-dimethylammonium

N4111 Butyltrimethylammonium

N4444 Tetrabutylammonium

NAM Acetylmorpholine

xxiv

N-bupy 1-butylpyridinium

NFM Formylmorpholine

NO3 Nitrate

OF Objective function

omim 1-octyl-3-methylimidazolium

P Probability

PFD Process flow diagram

pmim 1-pentyl-3-methylimidazolium

pmg Pentamethylguanidinium

pmeg Pentamethylethylguanidinium

pmpeg Pentamethylpropylguanidinium

PR Peng-Robinson

p Pressure

Cp Critical pressure

PF6 Hexaflourophosphate

po

Reference pressure

q Charge

r Position

R Universal gas constant

jiS / Selectivity of absorption

SLM Supported liquid membrane

SO2 Sulfur dioxide

STO Slater type orbitals

SV Split-valence

t Time

T Temperature

Tb Normal boiling point temperature

TC Crritical temperature

TCA Tricyanomethanide

TEG Triethyleneglycol

TFA Trifluoroacetate

Tf2N Bis(trifluoromethylsulfonyl)-imide

tmdeg Tetramethyldiethylguanidinium

tmdpg Tetramethyldipropylguanidinium

tmg Tetramethylguanidinium

Triflate Trifluoromethanesulfonate

TZ Triple-zeta

V Potential energy operator

Velst Exchange energy operator COSMO

elstV Solvent effect energy operator

VC Critical volume

VLE Vapor liquid equilibrium V Molar volume at infinite dilution

x Mole fraction

X Coefficient of Equation (4-16)

xxv

X1 Coefficient of Equation (4-16)

X2 Coefficient of Equation (4-16)

Yl,m Spherical harmonic functions

Z Compressibility factor

Z Atomic number

Z Total partition factor

ZC

Combinatorial factor

ZR

Residual factor

Coefficient of Equation (4-4) Coefficient of Equation (4-14) Coefficient of Equation (4-6)

effa Effective contact area

r Coefficient of Equation (5-6)

Interaction parameter

Polarizability volume

Coefficient of Equation (4-6)

Coefficient of Equation (4-14)

r Coefficient of Equation (5-6)

Dielectric constant Coefficient of Equation (4-14)

T Coefficient of Equation (5-2)

bMT Coefficient of Equation (5-2)

MT Coefficient of Equation (5-1)

MV Coefficient of Equation (5-4)

r Basis function

i Molecular orbital

i Fugacity coefficient

i Area fraction

i Volume fraction

Coefficient of Equation (4-16) Electron density Charge density Wave function

wavefunction of the many electron system 2 Laplacian

vdW Van der Waals interaction parameter

S Chemical potential of surface segment

*

i Pseudo-chemical potential

C

i Combinatorial contribution to the chemical potential

R

i Residual contribution to the chemical potential

xxvi

i Chemical potential of pure component

TC Coefficient of Equation (4-4)

crt Coefficient of Equation (4-6)

i Activity coefficient

i Asymmetric activity coefficient

i Infinite dilution activity coefficient

Acentric factor

27

27

Chapter One: Introduction

1.1 Background

Natural gas is a low carbon content fossil fuel which produces 25% less CO2 emissions

than oil and 50% less CO2 than coal [1]. This difference suggests that increasing natural

gas for power generations and transportation greatly reduces emissions. Figure 1-1

highlights different type of emission reduction achieved by switching to natural gas.

Figure 1-1. Emission reductions achieved by using natural gas; : Displacing 1 kWh

of coal-based electricity with natural gas : Displacing 1 Gasoline Gallon

Equivalent (GGE) of vehicle fuel with natural gas[1]

25.1%

78.4%

91.3%

99.9%

100%

64.8%

87.4%

99.8%

100%

100%

0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0%

Carbon Dixide

Nitrogen Oxide

Particulates

Sulfur Dioxide

Mercury

28

28

1.1.1 Gas Quality Specifications

Table 1-1shows the gas specification of trans Canada and US pipelines.

Table 1-1. Gas Quality Specifications[2]

Trans Canada Pipelines US Pipelines

Specs Canadian Mainline System Alliance USA

Hydrogen Sulphide Max 23 mg/m3 Max 1 grains/ccf

3

Total Sulfur Max 115 mg/m3 Max 5 grains/ccf

3

Carbon Dioxide Max 2% by volume Max 2% by volume

Oxygen Max 0.4% by volume Max 0.4% by volume

Temperature Max 50°C Max 122°F

Heating Value Min 36 MJ/m3

Max 41.34 MJ/m3

Min 962 BTU/ft3

Water Max 65 mg/m3

Max 4 lbs/MMcf

Hydrocarbon Dewpoint Min -10°C at

5500 kPa absolute

Min 14°F at opt. pres.

1.1.2 Gas Sweetening

After production of the natural gas, treatment is required to remove CO2 and H2S from

the natural gas. CO2 must be removed because it lowers the energy value of the natural

gas whereas H2S is a poisonous gas which could also contribute to SO2 emissions. There

are several methods for removing these gases, the most common one being the chemical

absorption of these gases using a basic solvent, notably alkanolamine based solvents.

29

29

1.1.3 Absorption into a Solvent

Absorption is one of the most common methods to purify natural gas. In this method, a

solvent is used in order to remove CO2 and H2S from the gas stream. The solvent must

dissolve CO2 and H2S and ideally none of the valuable components of the gas such as

CH4. The purification is done based on the relative solubility of CO2 and H2S versus CH4.

There are generally two types of solvents, physical and chemical solvents.

1.1.3.1 Physical Solvents

These solvents physically dissolve the components, i.e., there are no ionic or covalent

bonds formed between the solute and solvent molecules. Selexol [3] and Morphysorb [4]

are examples of physical solvents. Normally, physical solvents are used with high partial

pressures of acid gas and/or when a less pure product is required.

1.1.3.2 Chemical Solvents

Chemical solvents react with the unwanted compounds and remove them from the gas

stream by acid base type reactions or formation of covalent bonds. Acid-base type

reactions are reversible in order to regenerate the solvent and recycle it to the absorber

via a regenerator where heat is applied to reverse the chemical reactions and release the

acid gases for further processing, for example to reinjection or to a sulfur recovery plant.

Figure 1-2 is a schematic of the chemical and physical solvent behavior in terms of acid

gas partial pressure and shows that chemical solvents absorbs more acid gas than physical

solvents at lower acid gas partial pressures; however, at higher partial pressures physical

solvents absorb more acid gas than chemical solvents.

30

30

Aqueous solutions of alkanolamine are commonly used as chemical solvents. Among

these, monoethanol amine (MEA); diethanolamine (DEA), and methyldiethanol amine

(MDEA) are the most common. The major functional groups in alkanolamine molecules

Figure 1-2. Schematic comparison of the loading of physical and chemical solvents

at different partial pressures of acid gas

are the hydroxyl group and the amine groups. Hydroxyl groups increase the water

solubility and lower the amine vapor pressure, important for avoiding losses in the

solvent regeneration. Acid gas is absorbed because of the alkalinity of amine group in

31

31

aqueous solutions via acid-base reactions. Alkanolamines are classified as primary,

secondary and tertiary if there are two, one or no protons directly attached to the nitrogen

group, respectively, Figure 1-3.

Figure 1-3. Different types of amines, a: primary; b: secondary; c: tertiary

The major reactions involved in acid gas removal by amines are shown in reactions (1-1)-

(1-4) CO2 reacts via reactions (1-2) and (1-4). Only, primary and secondary amines

rapidly go through reaction (1-4). This reaction produces a carbamate species which is

not easily regenerated. With tertiary amines, CO2 reacts through the slow reaction (1-2)

[5, 6]. The different reaction rates at which carbon dioxide and hydrogen sulfide are

absorbed play an important role in the solvent formulation and the design of plants that

allow optimal slippage of carbon dioxide to the sales gas while keeping the amount of

hydrogen sulfide under specification.

1-1

1-2

32

32

1-3

1-4

1.2 Ionic Liquids

Ionic liquids (ILs) are low melting salts which form liquids that contain cations and

anions at low temperature, including many that are liquid at room temperature. Due to

their ionic nature, ILs have negligible vapor pressure which significantly reduces the

evaporation and loss of solvent even in low pressure regeneration. Certain ILs can absorb

CO2 and H2S and relatively low amounts of the valuable components of the natural gas

such as CH4.

1.3 Objective

This study focuses on the ILs as an alternative solvent for gas sweetening. Different ILs

are considered and ILs suitable for gas sweetening are selected to conceptually design IL-

gas sweetening plants. The advantages and disadvantages of using ILs in gas processing

are discussed. The outcomes of this project are:

A database of available solubilities for CO2, H2S, CH4 and C2H6 in ILs

Models to predict the solubility of CO2, H2S, CH4 and C2H6 in ILs

A Screening procedure for the different ILs and ranking ILs suitable for gas

sweetening

33

33

Conceptual design of sweetening gas plants using ILs

Comparison of the conceptual IL gas sweetening plants against typical gas

sweetening plants with either physical or chemical solvents

1.4 Thesis Overview

This study is divided into seven chapters.

Chapter One provides some background about gas processing and introduces the

objectives and structure of the.

Chapter Two introduces ILs along with their key properties such as their low melting

point and negligible vapor pressure. The industrial applications of the ILs are also

discussed in Chapter Two; for example, as an electrolyte, solvent, supported liquid

membrane, in gas chromatography, and as a gas absorbent.

Chapter Three includes a general introduction to the COSMO-RS theory and some of the

tools used to understand and implement COSMO-RS. References for a more detailed

description of those methods are provided.

Chapter Four compares different models available for generating solubility estimates for

sour gas mixtures with no experimental data followed by the introduction of COSMO-RS

method of calculation. The available experimental database is introduced, followed by

the proposed models for calculating the solubility of CO2, H2S, CH4 and C2H6. The

absorption selectivity is introduced and the ILs are screened based on appropriate criteria.

Chapter Five introduces the top ranked ILs suitable for gas treatment and gas sweetening

and plants are designed from a conceptual point of view. The process simulation results

are compared with the gas plants utilising physical and chemical solvents. Some of the

34

34

challenges of designing IL-gas plants are addressed. The advantages and disadvantages of

the new process are discussed.

Chapter Six includes the summary and conclusions.

Chapter Seven provides the recommended future work.

1.5 Summary

This chapter has provided some background related to this study. The difference between

physical and chemical solvents for gas sweetening was explained. Ionic liquids were

introduced as an alternative solvent for gas sweetening. Finally, the objectives and

structure of this thesis were introduced.

35

35

Chapter Two: Ionic Liquids

2.1 Introduction

As was discussed in chapter one, the purpose of this study is to investigate the use of ILs

as potential solvent for gas treatment. In this chapter the structures of various ILs are

introduced along with their key properties such as their low melting point and negligible

vapor pressure. Some other applications of ILs also are discussed.

2.2 IL Structure

Salts are formed by cations and anions. For an example of a common salt, NaCl is a

crystallized form of Na+ and Cl

- paired by ionic bonds. Due to the strong Coulombic

interactions between the cations and anions, pure salts often have very high melting

temperatures. Table 2-1 shows the melting temperature of some common salts.

Table 2-1. Melting point of some common salts

Salt Melting

point,

C

NaCl 800.80 [7]

NaBr 746.85 [8]

NaI 659.85 [9]

KCl 770.85 [8]

KBr 733.85 [9]

KI 680.85 [9]

ILs are low melting point salts which are often formed from organic cations and organic

or inorganic anions. The larger the molecular moieties, the more sterically destabilised

the solid compound becomes. This increases the relative liquid stability allowing ILs to

36

36

remain liquid at relatively low temperatures. Figure 2-1 shows an example of IL,

[emim][Tf2N] with melting point of -17 C [10].

Figure 2-1. [1-ethyl-3-methylimidazolium][Bis(trifluoromethylsulfonyl)-imide] or

[emim][Tf2N]

The first IL, ethylammonium nitrate [EtNH3][NO3] with a melting point of 13–14 C was

discovered by Paul Walden in 1914 [11, 12]. In 1992 the ILs with emim cations were

introduced which do not react or decompose in the presence of water [13].

Most ILs are formed from the combination of organic cations such as imidazolium based

ions and inorganic or organic anions such as sulfate or bis(trifluoromethylsulfonyl)-

imide.

Parameters that can affect the melting point of ILs are:

The size of the ions [11, 14]. The salts with larger ions have generally lower

melting point.

The asymmetry of the ions lowers the melting point.[11, 15-17]

Crystallization is inhibited in salts with ions that have different conformations

[11, 18], i.e. highly flexible molecular ions can lower the melting point

37

37

The factors above suggest that the properties of ILs can be tuned by changing the

chemical structure of the cations and anions that form the ILs. In the future, these tuned

ILs can be designed to exhibit convenient properties for solvent design such as melting

point, liquid viscosity, solubility of chemicals in ILs, and low vapor pressure amongst

other properties. Many different ILs can be synthesized having a wide range of physical

and chemical properties. This potentially large variability of chemical structures make the

development of predictive methods for the estimation of properties of ILs and associate

mixtures an important activity and a key part of this study.

2.3 Applications

2.3.1 ILs as Electrolytes

There is increased interest in ILs as electrolytes for lithium or lithium-ion batteries. Most

of these ILs are quaternary ammonium cations (such as [R4N], pyridinium, imidazolium,

pyrrolidinium anions such as [BF4], [PF6], triflate, Tf2N [19-23]).

Polymer-in-IL electrolytes are made by dissolving compatible polymers in ILs [24-29].

The conductivity of these polymer-in-IL electrolytes is higher than that of other classical

polymer electrolytes. For example the highest conductivity of the traditional polymer

electrolytes at room temperature is between 10-5

and 10-4

S.cm

-1; However, the

conductivity of P(VP-c-VA)/ emim-Tf2N is about 3106 S.cm

-1 at room temperature

[25].

38

38

2.3.2 ILs as Solvents

ILs exhibit negligible vapor pressure at most normal operating conditions and this makes

them attractive compared to conventional volatile organic solvents from an emissions

point of view. Negligible vapor pressure contributes to lower operating costs because the

solvent makeup is in theory negligible. The low volatility also potentially reduces the

environmental and human health concerns that accompany exposure to organic solvents.

Since ILs are able to dissolve a wide range of organic and inorganic compounds, they can

be used as solvents in homogeneous catalytic systems [30-36]. Since some chemicals are

immiscible in some ILs, ILs can be used for extraction purposes [37-42]. For example,

Wang et. al. studied the imidazolium based ILs, for the recovery of some amino acids

from aqueous media[43].

2.3.3 ILs as Supported Liquid Membrane (SLM)

Supported liquid membranes (SLMs) [44] use porous supports that are impregnated with

a solvent. For example, IL-membranes are used for CO2/N2 separation and CH4/N2

separation[45].

2.3.4 ILs in Gas Chromatography (GC)

The stationary phase of GC needs to have certain properties such as high thermostability,

low volatility and good wetting ability. Therefore, ILs are good candidates for the

stationary phase for GC [46-53].

39

39

2.3.5 ILs as Natural Gas Solvent

Usually gas solvents are used for purification and removal of unwanted compounds

present in natural gas to meet the sales pipeline specifications or address environmental

concerns. One of the first applications of molten salts as absorbent was developed in

1969 in which the mixture of sodium, potassium and lithium carbonate is used to remove

sulfur dioxide from flue gas [54]. This process is operates at 204 C. Due to the high

temperature this IL solvent is not a room temperature IL. In recent years, there has been

increased interest in ILs as gas treatment solvents. Because ILs structures can be tuned in

order to modify its properties such as capacity and selectivity of absorption. The

negligible vapor pressure of ILs decreases the potential for solvent emission and

mitigates the environmental concerns. The low volatility also reduces the cost for solvent

makeup. There are a number of studies which aim to measure the solubility of gases in

ILs. Significant to the purpose of this research, are studies on the solubility of carbon

dioxide [55-76], hydrogen sulfide [69, 76-82], methane [59, 83-86], ethane [59, 83-86],

oxygen [86-88], hydrogen [68, 89], sulfur dioxide[90, 91] in various ILs such as bmim-

Tf2N, emim-BF4 and HOmim-Triflate. The solubility of gases in ILs must be known in

order to choose an IL as gas solvent; however, at the present, not enough gas solubility

data in ILs are available.

2.4 Summary

In this chapter the ILs are introduced as organic salts with low melting points and

negligible vapor pressure. The parameters that affect the melting point are the size of the

ions, asymmetry of the ions, existence of different conformers and presence of hydrogen

40

40

bonding. Some applications of the ILs such as electrolyte, solvent, supported liquid

membrane, gas chromatography and gas absorbent were introduced.

41

41

Chapter Three: Introduction to COSMO and COSMO-RS models

3.1 Introduction

Progress in methods of calculation using quantum mechanics and statistical

thermodynamics along with the wide availability of the high speed of computers allows

for the first principle calculation for physical properties of different chemicals in a

reasonable time. First principle calculations require no empirical or experimental

parameters; therefore, a wide variety of chemical species can be investigated, where the

number of investigations are only limited by choice of computational theory level and/or

access to processor time.

These calculations help to reduce the amount of experimental work required to survey

large families of molecules with solutes of interest arising from a combinatorial

experimental program. This chapter includes a brief introduction to COSMO-RS theory

and some of the tools used to implement COSMO-RS are also discussed. The COSMO-RS

model is used in this thesis to calculate the activity coefficient of solutes in IL solutions.

It should be stressed that not all of these potential ILs have already been synthesized.

Prior to COSMO-RS calculations, the DFT/COSMO calculations must be done. The

DFT/COSMO calculation minimizes the energy, optimizes the geometry, and provides

the screen charge distribution of each component of the mixture. Since, this is a first

principle calculation, it does not require available experimental information, except for

fundamental physical constants such as the mass of electrons and speed of light. The

DFT/COSMO calculation was performed using the TmolX software[92]. The COSMO-RS

42

42

calculations were performed using the COSMOtherm software[93]. Again, the COSMO-

RS calculations provided thermodynamic properties of the mixture, such as the activity

coefficients of the components.

3.2 Representation of Particles in Quantum Mechanics

Classical physics can be used to understand and predict precise trajectories, locations and

momenta. Various phenomenon were left unexplained by classical mechanics at the

beginning of the 20th

century because classical mechanics applies only to macroscopic

particles. Explanations of these non-classical observations have led to the field of

quantum mechanics which studies microscopic particles, such as the electrons and nuclei

that make up the atoms within a molecule. These types of calculations rely only on the

fundamental physical constants and advancing theory versus the availability of

experimental quantities.

In order to understand quantum mechanics, one must first accept that “energy is

quantized” and that “all system information is contained within a wavefunction”. These

two statements require wave-particle duality and other various postulates, including:[94]

All dynamic information is contained in a system’s wavefunction, , where the

mathematical wavefunction is found by the Schrödinger equation:

3-1

43

43

EH ˆ ,

where, is the wave-function which is an eigenfunction of H . H is called the

Hamiltonian, an operator that returns the system energy; E is an eigenvalue that

represents the energy of the system.

The probability that a chemical system will be found within some region of a

multidimensional space is equal to the integral of d2

over that region of

space. (The Born interpretation).

Acceptable wave-functions are continuous, have a continuous first derivative, are

single-valued, and are square-integrable.

Hydrogenic atoms such as H, He+ and O

7+ contain only one electron. The Schrödinger

equations of the hydrogenic atoms can be solved exactly. The wavefunction describing a

hydrogenic atom is known as an atomic orbital and is given by:

3-2

,,, ,, lmlln YrRr ,

where, rR ln, is the radial portion and ,, lmlY is the angular momentum portion of the

orbital.

By Solving the Schrödinger equation for an orbital, three quantum numbers are obtained.

The first quantum number is shell or principal quantum number (n). It is the energy level

in which the electron is found. The value of n can be set between 1 to n (or n = K, L, M,

44

44

N, …). n is the outer most shell containing an electron. The second quantum number is

subshell or the angular momentum quantum number, l, which can describe the shape of

the atomic orbital. The values of l ranges from 0 to n-1 (or l = s, p, d, f, …). The third

quantum number is the magnetic quantum number ( lm ). For a given l, lm ranges between

–l to +l. The spin quantum number ( sm ) is the fourth quantum number which cannot be

obtained by solving the Schrödinger equation. It is based on the Pauli Exclusion Principle

that it is impossible to have two identical spin orbitals. For electrons (Fermions), the

value of sm can only be 2

1 or

2

1 ( or electrons).

Atoms with two or more electrons are called many-electron atoms. Since the electrons of

the many-electron atoms interact with each other, no analytic expression for orbitals and

energies can be given [94].

Modern computational quantum calculations begin with the hydrogen orbitals as

described above. Hydrogen type orbitals are linearly combined by assigning spin orbitals

which allow for two electrons to occupy a hydrogen type orbital. The Pauli Exclusion

Principle is implemented by determinants to combine these functions so that the orbital

vanishes when two electrons of the same spin are within the same orbital. Note that the

combination of one-electron orbitals to form a many-electron system has removed or

decoupled the electron correlations and relativistic effects.

Since the nuclei of each atom are much heavier than the electron, the Born-Oppenheimer

approximation is applied, where nuclei are treated as stationary particles governed by

45

45

coulomb interactions and the wave functions are probability functions of moving

electrons [95].

3.2.1 Basis Sets

In most molecular quantum mechanical methods the molecular orbital (MOs), i , are

formed using a linear combination of basis functions or basis sets r ,

3-3

r

rrii C ,

where, Cri is the coefficient. The basis functions have to be able to approximate the actual

radial wave function sufficiently well to produce meaningful results with a reasonable

computational cost. Integrals should be evaluated quickly and accurately. The radial

functions and how they are combined are normally referred to as basis sets. A more

flexible basis set (more basis functions and less coupling) results in a lower energy and

more accurate orbital functions, but is computationally more demanding. There are two

types of principal basis functions:

Slater Type Orbitals (STO) which mimic the hydrogen atomic orbitals given by

Equation 3-4,

3-4

rn

mlmln erNYr 1

,,,, ,,, .

46

46

The actual form of the hydrogen orbitals is through the STOs; however, the integration of

STOs are mathematically difficult.

Boys [96] proposed the Gaussian Type Orbitals (GTO) which allow for

mathematical ease and integrals which can be explicitly evaluated,

3-5

222

,,,, ,,, rn

mlmln erNYr .

The radial shape of the GTOs are less like the hydrogen orbitals but the same angular

momentum functions or spherical harmonic functions have been retained. Since GTOs do

not provide the correct radial decay shape, linear combination of GTOs are fit to

reproduce as accurately as possible an STO. In terms of computational efficiency GTOs

require much less calculation time than STOs [95]; therefore, the penalties for adding

more basis functions are more reasonable. Usually, instead of using the individual GTOs,

a linear combination of GTOs is fit to reproduce STOs. Hehre et. al. [97] have

systematically examined linear combination of 2 to 6 GTOs. Obviously more GTOs

describe STOs better, however the calculation become increasingly complicated.

3.2.1.1 Minimum Basis Set

A minimum basis set consists of one function for each atomic molecular orbital, AO. For

example for CH4 a minimum basis set includes 1S, 2S, 2Px, 2Py, 2Pz AOs for carbon and

1S AO for each hydrogen. In total the minimum basis set of CH4 consists of 9 basis

functions.

47

47

3.2.1.2 Triple-Zeta (TZ) Basis Function

A Triple-Zeta or TZ basis set replaces each AO of the minimum basis set by three basis

functions. This allows for more accurate calculations by better approximation of STOs.

For instance, for CH4 a TZ basis set consists of 27 basis functions.

3.2.1.3 Split-Valence Basis Set (SV)

SV uses a fixed number of basis functions for each inner-shell AO (proportioned to fit an

approximate STO shape) and two or more functions for the valence shell of AO. These

two basis functions are allowed to be optimized during the computational solution. The

most common examples are the basis sets of 3-21G, 6-31G and 6-31G(d,p) [98]. For 3-

21G, there is one basis function for each of the non-valence (core) orbitals composed of 3

fixed Gaussian functions, whereas the valence orbitals each contain two basis functions

(one with 2 fixed Gaussian functions and one with 1 Gaussian function). The split

valence allows for better optimization due to mathematical flexibility, because the core

orbitals are less affected by external electron environment than the valence orbitals.

Therefore, it is more efficient to have more flexibility in the valence basis functions than

those closer to the nucleus.

Polarized Basis Set (P)

The shape of the AOs (angular momentum) show different symmetries when they form

molecules, i.e., they can slightly distort. To allow for this polarization, the basis functions

with higher quantum number (angular momentum) than the maximum valence shell of

the ground state atom can be added. For the 6-31G(d,p) basis set, there is a split valence

48

48

(3 & 1 Gaussians) and the addition of polarization functions, d and p. For example, 2P

orbitals would be added to all hydrogen atoms. The polarization of hydrogen is required

for modeling hydrogen bonding interactions. Splitting functions or adding polarisation

requires more computational time, but yields better results.

As will be explained later, the triple zeta split valence plus polarization basis set (TZVP)

is implemented for the calculations in this study.

3.3 Energy Calculations of Many Electron Systems

It is impossible to provide an analytical solution to the Schrödinger equation for many

electron systems. In this section two major quantum mechanics theories for calculation of

the energy of the many electron atoms or molecules are briefly described, Hartree-Fock

theory (HF) and Density Functional theory (DFT).

3.3.1 Hartree-Fock Theory (HF)

Similar to Equation (3-1), the Schrödinger equation for a many electron system can be

described as follows:

3-6

EH .

In which H is the Hamiltonian operator and is defined as follows:

3-7

49

49

jiall ij

n

i i rr

ZH

,

2

1 i,all

2 1

2

1ˆ

,

where, 2 is the Laplacian, Z is the atomic number of the nuclei , ir is the distance

between electron i and nuclei , and ijr is the distance between electrons i and j. The

first term in Equation (3-7) represents the kinetic energy of electrons, for second term

represents the attraction potential energy between all binary combinations of electron-

nuclei, and the third term represents the repulsion potential energy of all binary

combinations of electron-electron.

The in Equation (3-6) represents the wavefunction of the many electron atom or

molecule. Since it is impossible to solve analytically the Schrödinger equation, in

Hartree-Fock theory, the wave function of the many electron system is constructed based

on single electron wavefunctions and their spin functions. The function that represents

the relation between single electron wavefunctions and the spins with the wave function

of the many electron system is a determinant called Slater determinant, given by:

3-8

nnnnf nn 22,22,,11,11 11 ,

where, f is the Slater determinant for a 2n-electron system, i is the single electron

wavefunction with alpha spin, and i is the single electron wavefunction with beta

spin.

50

50

Slater determinant is constructed in such a way to take the Pauli Exclusion Principle into

account. In this way, instead of considering the effect of multiple electrons in a single

wavefunction , the problem is simplified into considering the effect of single electron

cloud of i on single electron cloud

j .

The energy of system can be calculated as follows [98]:

3-9

d

dHE

ˆ,

where, dxdydzdd and is the spin function. Therefore, for a 2n-electron system,

the integral is 2n 4 fold.

For a normalized wavefunction we can rewrite Equation 3-9 as

3-10

dHE ˆ ,

or

3-11

|ˆ| HE .

51

51

The energy calculated from Equation (3-11) is the quantum-mechanical average of the

energy or the expectation value of the Hamiltonian operator. The expectation value of an

operator is the integral of wavefunction over the operator. The Variation theory states that

the energy calculated from Equation (3-11) must be greater than or equal to the true

ground-state energy of the molecule. Thus, the true energy has the minimum value.

Roothaan [99] and Hall [100] presented the single electron wavefunctions as a linear

combination of the basis functions. Minimizing the energy, E with respect to ’s and

considering variation theory results in an algorithm to iterate the ’s (coefficients of the

basis functions) for the given geometry of the many electron system until the energy

minimizes. This algorithm is called SCF (self consistent field).

Variation theory also means that the more basis functions in the HF equation the lower

and more accurate is the energy that can be calculated. If infinite basis sets are used in

HF the solution of the HF equation is called “HF limit”. The HF, limit properties can be

obtained by extrapolation[101].

The main limitation of the Hartree-Fock theory is its one-electron nature [101, 102].

3.3.2 Density Functional Theory (DFT)

Recall that the linear combination of orbitals (single electron functions) has removed or

decoupled electron correlation and relativistic effects. This means that even the most

flexible basis set can only result in a limited energy or accuracy, referred to as the

52

52

Hartree-Fock limit. Corrections to this limit can be made using various approaches, one

of which is Density Functional Theory.

Hohenberg and Kohn proved that any ground state property of a molecule is a functional

of the ground state electron density. Therefore the ground state electronic energy is

determined completely by the electron density and there exist a one-to-one

correspondence between the electron density of a system and energy [102, 103]:

3-12

00 EE

Note that functional means a function applies on another function and it is shown by a

bracket. For example in Equation (3-12) the functional E is applied on the electron

density 0 which itself is a function of wavefunction. The Hohenberg and Kohn theory is

an existence statement; however, the challenge of the DFT method is that the functional

connecting the density and ground state energy state energy is not known[102].

The foundation of DFT method was introduced by Kohn and Sham (KS)[104]. In KS

theory (1), the molecular energy is expressed as a summation of some terms from which

only one term involves the unknown functional. (2) An initial guess for electron density

is used to obtain the initial guess for the KS orbitals. This initial guess is used to

iteratively refine these orbitals (SCF). The final KS orbitals are used to calculate the

electron density and therefore the energy.

53

53

The ground state electronic energy of the real molecule is the sum of electron kinetic

energies, the nucleus-electron attraction potential energies and the electron-electron

repulsion energies:

3-13

][][][ 0000 eeNe VVKEE .

in which the angle brackets are the expectation values or quantum-mechanical average

values, and KE, NeV and eeV are the functional of the kinetic energy of electrons, the

nucleus-electron attraction potential energies, and the electron-electron repulsion

energies, respectively.

In order to calculate the kinetic energy of KF-DFT, an imaginary reference system is

defined with the same electron density as the real system. In this reference system the

electrons do not interact with each other. The electron kinetic energy is defined as the

kinetic energy of the non-interacting reference system plus any additional differences

between the reference kinetic energy and the real system. In this non-interacting

reference system, the kinetic energy is the summation over all single electron kinetic

energies. The wave function of a single electron system is available exactly from the

Schrödinger equation, therefore the kinetic energy of the reference system can be

calculated exactly. The real kinetic energy of the system, real

KE 0 is the kinetic

54

54

energy of the reference system, ref

KE 0 plus the difference between the kinetic

energy of the real and reference system 0KE ,

3-14

000 KEKEKErefreal

.

The attraction potential energy between the electrons and nuclei is readily available by

considering the nuclei as a point charge (Born-Oppenheimer approximation) and using

the classical coulomb interaction between nuclei and electron clouds [95].

The electron-electron interactions can be obtained by calculating the classical coulomb

electron-electron repulsion energy plus any difference between the classical and quantum

mechanics. The classical electron repulsion can be calculated by considering the coulomb

repulsion between each two electron clouds over the entire volume of the system [105].

Therefore Equation (3-13) can be rearranged as

3-15

][][][][][ 000000 eeclassicaleeNerefVKEVVKEE ,

or

3-16

XCclassicaleeNerefEVVKEE ][][][ 0000 ,

55

55

in which XCE is called the exchange energy and is the total difference between the real,

reference and classical systems for both kinetic energy and electron-electron interactions,

3-17

][][ 00 eeXC VKEE .

In principle, if a correct exchange functional is chosen, Equation (3-16) calculates the

exact energy of a system with the given electron density of 0 .

The KS equations are obtained by differentiation of E0 with respect to KS orbitals. An

iterative method along with the variation theory is used to solve the KS equation. It

begins with guessing the geometry of the system. Then the electron density of the system

0 is guessed. By solving the KS equations the coefficients of the KS orbitals (the

coefficients of the basis functions) are calculated. Therefore the modified electron density

of the system can be calculated based on the KS orbitals. This iteration is continued until

there is no significant change in the electron density and the energy is minimized. Then

the next geometry of the system is considered and its electron density is calculated using

the similar iterations. This procedure is repeated until the geometry of the system is

optimized and a minimum energy of the system is obtained.

3.4 Conductor-like-Screening-Mode (COSMO) Theory

The dielectric continuum models were developed by describing a solvent as a dielectric

medium with their dielectric constant, . In Conductor-like-Screening-Model (COSMO),

56

56

the molecules are assumed to be in a conductor solvent with a dielectric constant of

infinity. In this way the solvent can screen the solute molecules. The solutes are assumed

to be in a molecular shaped cavity with the screen charges on the surface of the cavity.

The electrostatic energy of the system can be written as [106]:

3-18

)(qeffectssolventenergycoulombphasegasqE .

in which, q is the screen charge on the surface of the cavity. The gas phase coulomb

energy is the net interaction caused by the electrons with electron density of r , and

nuclear charges, Z. The solvent effect consists of the interaction of the screen charges on

the surface of the cavity and the solute charges (i.e., electron density and nuclear

charges).

At given electron density and nuclear charges, Equation (3-18) can be minimized

analytically to calculate the optimized screen charge on the cavity[106],

3-19

qE = Gas phase coulomb energy + Solvent effects (q*).

The second term of Equation (3-19) expresses the solvent effects based on the optimized

screen charges of the cavity is called the screen energy [107].

57

57

As discussed earlier, density functional theory (DFT) was chosen to calculate the energy

for a molecule based on the electron density of the molecule, where the electron density

probability is expressed as a sum of squared molecular orbitals. The optimal molecular

orbitals are calculated by minimizing the total energy (Equation (3-16)), by solving the

Kohn-Sham equations. This involves an optimization of molecular geometry, i.e.,

location of nuclei.

The COSMO theory provides an additional term that describes the solvent effects,

COSMO

elstV which is based on the optimized screening charge of the cavity at the given

electron density of the molecule. This additional term can be introduced into Kohn-Sham

equation as

3-20

XC

COSMO

elstclassicaleeNerefEVVVKEE 00000 ][][][ .

The energy associated with COSMO

elstV is called COSMO Energy. By solving the Kohn-Sham

equation with the additional term, COSMO

elstV , the electron density of molecule associated

with a minimized the energy of the molecular conformation and the optimized screening

charges on the cavity will be available.

58

58

In this study the TmolX commercial software [92] was used to perform COSMO/DFT

calculations which utilizes the BP exchange functional [108-110] and a triple-zeta

valance polarized basis set (TZVP) [106].

As an example Figure 3-1 shows the charge distribution on the molecular cavity of water

after COSMO/DFT calculations[93].

Figure 3-1. Charge distribution on the molecular cavity of water after COSMO/DFT

calculations[93]. Dark red represents higher electron density and dark blue

represents lower electron density.

3.5 COSMO for Real Solvents (COSMO-RS)

In this theory, the molecular interaction of solvent and solute molecules are considered

and, the solvent is not just a dielectric medium. For this purpose, the molecules of solute

and solvent will be optimized separately using the COSMO method. Then the charge

density of different segments of the molecular cavity of solute and solvent (i.e., and

) can be obtained.

59

59

Figure 3-2. -profile of water based on COSMO calculation[93]

The histogram or -profile of the charge density of segments of each molecule, )(P ,

can be calculated. )(P indicates how much of a surface we find in a polarity interval [

- d /2, + d /2]. Figure 3-2 is an example which shows the -profile of water

based on COSMO calculation[93]. As it can be seen in Figure 3-2 there are two major

peaks that resulting from the negative electron-rich oxygen atom and positive polar

hydrogen atoms. The -profile of a mixed solvent can be calculated as

60

60

3-21

n

i

ii

n

i

i

i

S

Ax

px

P

1

1

,

in which, xi is the mole fraction of component i and Ai is the surface area of component i.

COSMO-RS treats the solution as an ensemble of pair-wise interacting surface segments

with charge densities of . The interaction energies between each pair of surfaces can be

described based on the charge densities of two surfaces and . In COSMO-RS theory

three types of interactions are considered:

Electrostatic (Emisfit):

3-22

2

2,

effmisfit aE ,

in which aeff is the effective contact area, ( Aaeff 25.6 ) and is the interaction

parameter (=5950 kJ/mole/A ).

Hydrogen bonding (EHB):

61

61

In order to have hydrogen bonding, the donor should be less than a threshold, - HB and

acceptorr should be larger than a threshold, HB ( HB = 0.085 kJ/mol/A ). Given this

assumption the energy of hydrogen bonding can be shown as [106]

3-23

HBacceptorHBdonorHBeffHB CaE ;0max;0min;0min ,

in which CHB is the hydrogen bond strength (CHB = 36.7 kJ/mol/A ).

Van der Waals interactions (EVdW):

In COSMO-RS the Van der Waals interactions are approximated as,

3-24

VdWVdWeffVdW aE , ,

in which VdW and VdW depend on the element involved in binary interaction (i.g.

AmolkJHVdW //0361.0 and AmolkJCVdW //0401.0 ).

3.6 Generating Chemical Potential and Activity Coefficient of Solute Based on

COSMO-RS

Considering an ensemble of molecules that represents a liquid, the partition sum, Z, of an

ensemble is divided into two contributions as

62

62

3-25 RC ZZZ ,

in which, ZC is the combinatorial factor that takes into account all size and shape of the

molecules. ZR is the residual contribution and that take into account the non-steric

interactions such as Van der Waals, electrostatic and hydrogen bonding [106].

The Gibbs free energy of the system can be calculated as[106]

3-26

ZRTG ln .

The Staverman-Guggenheim (SG) expression, Equation (3-27), [106, 111, 112] is used to

describe the combinatorial factor of Equation (3-25),

3-27

i i

iii

i

ii

C

SG qxz

xxZ ln

2lnln ,

where, xi, i and i represent the mole, area and volume fractions of the component i,

respectively. Note, Equation (3-27) has also been used as the combinatorial factor in the

UNIQUAC model[106, 113].

63

63

The chemical potential of surface segments S which is called -potential can be

calculated by solving series of non-linear equations[106],

3-28

.,,1

expln

dEEaRT

p

a

RT

HBmisfitSeffS

eff

S

The S or -potential is chemical potential per surface area of a piece of surface

polarity in a solvent characterized by -profile Sp .

The pseudo-chemical potential, *

i , is defined by Equation (3-29) [114],

3-29

iii xRT ln* .

where, i is the chemical potential.

The pseudo-chemical potential of compound i in the system S can be calculated by

integration of S over the surface of compound[106],

3-30

dpTSTSTSTS Si

C

i

R

i

C

ii ),(),(),(,* .

64

64

The activity coefficient of component i can be calculated based of pseudo-chemical

potential on compound i [106] ,

3-31

RT

TiTSTS ii

i

,,exp),(

** .

in which, Tii ,* is the pseudo-chemical potential of the component i in the pure solvent

i.

In this study the COSMO-RS calculation were performed with COSMOthermX

commercial software[93]

65

65

3.7 Summary

In this chapter the COSMO and COMO-RS are briefly introduced. Using the COSMO

model, one can calculate the energy of a molecule at the given density of the electron of

the molecule. This model takes into account the solute and solvent effects. The solvent in

the COSMO model is considered to be a perfect conductor. Energy of the sysytem is

minimized by optimizing the charge distribution on the molecular shaped cavity of the

molecule. The COSMO calculations can be implemented using DFT level quantum

calculations. The COSMO/DFT calculations are performed by utilizing the BP [108-110]

functional with and a triple-zeta valance polarized basis set (TZVP)[106].

In the COSMO-RS model, the COSMO calculations are done for both solute and solvent

molecules. The activity of the component i can be described based on the calculated

chemical potential of the system.

The COSMO-RS model is used in this study to predict the activity coefficient of

components in the mixture without the aid of experimental information. There will be

detailed discussion of this application in the coming chapters.

66

66

Chapter Four: The Solubility of CO2, H2S, CH4 and C2H6 in Ionic Liquids

4.1 Introduction

In addition to saturated light hydrocarbons and water, natural gas is often produced with

significant concentrations of acid gases such as CO2 and H2S. CO2 must be removed to

meet heating value specification for sales gas; whereas, H2S is removed due to its toxicity

and to reduce the overall release of SO2 to the atmosphere during combustion. As shown

in Table 1-1, based on the pipeline and transportation specification maximum H2S

concentrations are limited to 23 mg/m3 whereas maximum allowable CO2 is typically 2%

by volume [2]. Currently, absorption in alkanolamine solutions such as aqueous

solutions of diethanol amine (DEA) and absorption in physical solvents such as

polyethylene glycol dimethyl ether based, e.g. Selexol are the common commercial

processes used for treating gas streams containing H2S and/or CO2 [6].

During the past decade, ILs have received increasing attention as a class of non-

traditional solvents, with potential to be used for natural gas treatment [115]. Given the

potential of ILs to be used as a selective solvent for the removal of both CO2 and H2S, the

availability of some experimental data for mixtures relevant to gas conditioning provided

a useful starting point to search for effective ILs for use in industrial gas sweetening.

Robust correlation of the available experimental data can be used to extend information

to systems which have yet to be studied.

There are many questions that need to be addressed before beginning to design a gas

plant, such as: What are the capacity and selectivity of absorption of ILs? How does one

67

67

best calculate the thermodynamic equilibria required to design absorbers and strippers? Is

there sufficient experimental data to tune the required thermodynamic equations? Are

there enough solvent physical properties such as viscosity, density and heat capacity to

allow at least a rough design and estimation of efficiencies? By carefully considering

these questions, one can appreciate that the most important tool for designing a gas plant

from a conceptual point of view is to have a model that provides a reasonable

representation of the experimental data and is able to provide estimates for many ILs

(existing or yet to be synthesized) and solute mixtures.

Some of the challenges are:

Based on the choice of the cations and anions, many combinations of ILs are

possible, but not all of the combinations are suitable for gas processing.

Very little or no experimental data are available for many potentially interesting

combinations of ILs and solute.

Appropriate calculation methods are required to predict values for cases where no

experimental data are available. These estimated solubility values can be used to

tune the thermodynamic model (i.g. a simple cubic equation of state) that will

ultimately be used to design the gas plant. In addition these data can be used for

screening the ILs based on their absorption capacity and selectivity to find

potential ILs.

An available database of experimental solubility information is provided and the

proposed models for estimating the pressure dependent solubility of CO2, H2S, CH4 and

68

68

C2H6 are presented later within the COSMO-RS framework. The absorption selectivity is

calculated and the ILs are screened based on this criteria.

4.2 Database

The objective of this study is to find ILs for the separation of CO2 and H2S from a natural

gas stream containing CH4, C2H6, CO2 and H2S. At this time limited experimental

solubility data are available for these components. Therefore, a thermodynamic model is

required to correlate and then predict the solubility of these gases in different ILs at gas

processing conditions. To do so, the small amount of existing experimental data can be

used to:

Determine model parameters,

Determine the best model and the most significant parameters.

A database of experimental solubilities for 31 mixtures of CO2–IL systems was compiled

by Mortazavi-Manesh et. al. [116, 117]. Those data and up-dated data are summarized in

Table 4-1. A variety of cations were included in the database. For

69

Table 4-1 Experimental data and conditions for CO2-IL mixtures [116, 117]

Data Points Purity Tmin (K) Tmax (K) pmin (kPa) pmax (kPa) Loading (max) Reference

b2Nic-Tf2N 6 >99%a,b 333.3 333.3 2190 8700 2.6 [55]

bmim-BF4 62 b, c, d 283.1 333.3 0.03 8500 1.1 [56, 57]

bmim-CH3SO4 54 >98%wte 293.2 413.1 908 9805 0.8 [58]

bmim-EtGLEtGLeC2SO4 10 >99%a,f 313.31 333.36 2090 9120 1.0 [55]

bmim-NO3 17 g, d 298.2 333.2 1031 9316 0.9 [56]

bmim-PF6 174 d, h, i, j 282.05 348.25 10.2 13237 1.7 [56, 59, 60,

62, 118]

bmim-Tf2N 37 k, d, l, m 279.98 339.97 292 9444 2.6 [56, 57, 63]

bmim-TFA 15 >99%a,f 298.17 333.41 1170 8770 1.6 [55]

bmim-Triflate 26 n, d 298.2 333.3 1039 9752 1.7 [56]

bmmim-PF6 105 <1.4 ppm halide 283.15 323.15 60 1300 0.3 [60]

bmmim-BF4 98 <1.4 ppm halide 283.15 323.15 19.7 1300 0.3 [60]

C6H4F9mim-Tf2N 31 >99%a,b 298.16 333.13 4.76 1300 0.5 [55]

emim-BF4 34 >99wt%o, 97%p 298.15 343.2 251 4329 0.3 [64, 65]

emim-PF6 27 >97%q 313.12 366.03 1490 13100 0.8 [67]

emim-C2SO4 53 - 283.3 353.1 38 1546 0.2 [68, 69]

70

Table 4-1 Continued

emim-Tf2N 76 >97%r 313.15 453.15 639 10258 1.5 [71]

emim-Triflate 30 >98%s 303.2 343.2 180 5884 1.2 [66]

emmim-Tf2N 34 >99% 283.15 323.15 7.28 1300 0.3 [60]

hmim-Tf2N 15 >99%t 281.9 348.5 48.6 1976.4 1.2 [119]

hmim-Triflate 40 99.6wt%u 303.85 344.55 680 9400 2.2 [73]

hmpy-Tf2N 41 >99%a,f 283.18 323.15 4.81 1300 0.6 [55]

MeBu3N-Tf2N 8 - 298.1 298.1 49.8 550 0.1 [57]

MeButPyrr-Tf2N 44 - 283.1 323.1 19.9 1300 0.5 [57]

N4111-Tf2N 6 >99%a,b 333.23 333.23 1560 8090 2.2 [55]

N4444-doc 6 >99%a,b 333.46 333.46 2010 9170 3.9 [55]

N-bupy-BF4 21 - 313.1 333.1 1547 9580 1.4 [74]

omim-Tf2N 20 v, d 298.2 333.3 1326 9752 3.5 [56]

pmim-Tf2N 56 >99wt%w 293.42 363.29 618 10198 3.0 [75]

HOemim-BF4 44 >99.5x 303.1 353.1 114 1194 0.1 [76]

HOemim-Tf2N 40 - 303.1 353.1 97 1114 7.3 [70]

HOemim-PF6 44 - 303.1 353.1 133 1127 0.1 [70]

71

Table 4-1 Continued

a The purity determined by 1H NMR spectroscopy.

b

Water content was measured by Karl Fischer titration and was in the order of 200 ppm. c water: 0.12wt%, Br

-<10 ppm, NH4 = 18 ppm [56], Cl

- < 1.4 ppm, Br

- < 8 ppm, NH4 < 18 ppm [57].

d ILs were washed with water and dried at T = 348.15 K under vacuum [56].

e IL was further degassed and dried under vacuum for 2 days before the experiments.

f H2O < 1500 ppm.

g H2O ≈ 0.16 wt%, Br

-:550 ppm, Ag: 20 ppm.

h H: 50-100 ppm, Cl

- < 10 ppm.

i Cl

-: 3 ppm, dried and degassed to ca 10

-9 bar and T = 348.15 K[59].

j purity > 97%, Cl

-: 4.7 mgL

-1, dried and degassed at 348.15 K under vacuum [62, 118].

k Cl

- < 50 mgKg

-1 , dried at 330 K NMR determined purity 99.4%[63].

l H2O= 450 ppm, Br

- < 10 ppm [56].

m purity > 99% [57].

n H2O = 0.14 wt%, halide free.

o halogen < 20 and H2O < 100 ppm, [64].

p H2O = 0.9 wt%, degassed at T = 343.2 K under vacuum [65].

q Degassed by freeze-thaw cycling under vacuum.

r dried under vacuum at T = 298.15 K.

s H2O < 0.7wt%; degassed at 343.2 K under vacuum.

t H2O < 20 ppm

u H2O < 59×10

-4 wt%.

v H2O = 110 ppm, Br

- < 10 ppm.

w Br

- < 100 ppm, H2O = 20 ppm.

x H2O < 100 ppm [76]

72

72

example, the data include imidazolium based cations such as bmim and emim,

ammonium based cations such as N4111 and N4444, pyrrolidinium based cations and

pyridinium based cations. Examples of anions used in the database include methylsulfate,

nitrate, and anions with fluoride in the molecular structure such as Tf2N, PF6 and BF4.

Most temperatures within the database are between T = 280 K and 330 K and in a few

limited cases, such as emim–Tf2N and bmim–CH3SO4, the maximum temperature is

greater than T = 410 K. In the case of MeBu3N-Tf2N, the solubility information is limited

to T = 298 K.

The loading of an IL is defined here as the number of moles of dissolved solute in the

moles of IL (as per the cation-anion pair). Given the variety of ILs and conditions, the

maximum mole fraction shown by the compiled experimental data spans a wide range of

loadings between 0.1 to 7.3.

Experimental solubility data for 14 H2S-IL mixtures [69, 76-82]were collected by

Mortazavi-Manesh et. al.[120]. A variety of cations were included in the database such

as bmim, emim, hmim, HOemim and omim. The anions used in the database are ethyl

sulfate and anions with fluorine in the molecular structure such as Tf2N, PF6, BF4 and

trifluoromethanesulfonate. Experimental Conditions for these data are reported in Table

4-2.

The collected experimental data for 4 CH4-IL systems are shown in Table 4-3. The

cations included in the database are bmim and hmim. The anions used in the database are

73

73

BF4, CH3SO4, PF6 and Tf2N. A database of 4 ILs was used for the solubility of C2H6-IL

mixtures, Table 4-4.[59, 83-86] The cations included in the database are bmim and

hmim. The anions used in the database are BF4, PF6 and Tf2N.

Table 4-2. Experimental data and conditions for H2S-IL mixtures

Purity Data

Points

Tmin

(K)

Tmax

(K)

pmin

(kPa)

pmax

(kPa)

Loading

max

Reference

bmim-BF4 a 42 303.15 343.15 60.8 836 0.55 [78]

bmim-PF6 a,b 73 298.15 403.15 69.0 9630 7.00 [77, 78]

bmim-Tf2N >98%c 44 303.15 343.15 94.4 916 1.04 [78]

emim-C2SO4 - 36 303.15 353.15 111 1100 1.00 [69]

emim-PF6 >97% 40 333.15 363.15 144.9 1933 0.56 [81]

emim-Tf2N >99% 42 303.15 353.15 107.7 1686 1.56 [81]

hmim-Tf2N >99%d 30 303.15 353.15 97.4 1050 1.14 [79]

hmim-BF4 >98%c 33 303.15 343.15 111 1100 1.00 [79]

hmim-PF6 >98%c 34 303.15 343.15 138 1090 0.79 [79]

HOemim-BF4 >99.5%e 51 303.15 353.15 121 1066 0.33 [76]

HOemim-Triflate - 42 303.15 353.15 105.9 1839 1.21 [80]

HOemim-PF6 - 47 303.15 353.15 133.6 1685 0.86 [80]

HOemim-Tf2N - 41 303.15 353.15 156.2 1832 1.34 [80]

omim-Tf2N >99.95%d 47 303.15 353.15 93.5 1912 2.78 [82]

a purity >99%, H2O <10

-2 wt%.

b purity: 98%, H2O = 0.05–0.1 wt%[77].

c H2O < 1wt%.

d H2O <10

-2 wt%.

e H2O < 10

-2 wt%.

74

74

Table 4-3. Experimental data conditions for CH4-IL mixtures

Purity Data

Points

Tmin

(K)

Tmax

(K)

pmin

(kPa)

pmax

(kPa)

Loading

max

Reference

bmim-BF4 97% mol 13 283.5 343.09 46.5 97.6 0.001 [83]

bmim-CH3SO4 >98%wt 24 293.15 413.2 1363 8853 0.048 [84]

bmim-PF6 a, b 107 283.15 343.08 115 1399 0.011 [59, 86]

hmim-Tf2N >99%wt 24 293.3 413.25 886 9300 0.228 [85]

a Cl

- = 3 ppm[59].

b purity > 99.9%mol, H2O ≈ 1.50×10

-2 wt%[86].

Table 4-4. Experimental data conditions for C2H6-IL mixtures

Purity Data

Points

Tmin

(K)

Tmax

(K)

pmin

(kPa)

pmax

(kPa)

Loading

max

Reference

bmim-BF4 97% mol 12 283.02 343.22 42.4 93.6 0.004 [83]

bmim-PF6 a, b 100 283.1 343.12 1 1399 0.045 [57, 59, 86]

bmim-Tf2N >99% 63 283.1 323.1 2.1 1300 0.144 [57]

hmim-Tf2N >99.5% molc 90 283.3 368.4 100 13070 0.671 [89, 121]

a Cl

- = 3 ppm.

b purity > 99.9%mol, H2O ≈ 1.50×10

-2 wt%[86].

c H2O < 2×10

-3 wt%[89].

75

75

4.3 COSMO Calculations

The COSMO and COSMO-RS models were introduced in Chapter 3. The COSMO-RS

model is used in this study to calculate the activity coefficient of solutes in IL solutions.

Overall there were 25 cations and 17 anions considered in this study. Therefore, the

combinations of cations and anions make 425 distinct ILs. It should be stressed that not

all of these potential ILs have been synthesized, but some structures, even though at this

moment hypothetical, may prove to be ideal for gas treatment. The preliminary structure

of each ion is built using the ChemBio3D Ultra software[122]. Prior to COSMO-RS

calculations the DFT/COSMO calculations must be done. The DFT/COSMO calculation

minimizes the energy, optimizes the geometry and provides the screen charge distribution

of each component of the mixture. Since this is a first principle calculation, it does not

require available experimental information, except for the fundamental physical constants

such as the mass of electrons, speed of light, etc. The DFT/COSMO calculation was done

using the TmolX software[92]. The COSMO/DFT calculation was performed by utilizing

BP [108-110] functional with and a triple-zeta valence polarized basis set (TZVP)[106].

The COSMO-RS calculations were done using the COSMOtherm software[93]. The

COSMO-RS calculations provide thermodynamic properties of the mixture such as the

activity coefficients of the components. The detail results of the calculation are provided

in Appendix A.

The optimized ions in DFT/COSMO calculations are used to build the IL solvents. The

ILs are considered to be 50:50 molar mixtures of cations and anions, i.e. the solvent net

charge is zero as all ILs studied are 1:1 ion pairs. Therefore the mole fraction of a solute

in a solute-IL mixture must be converted to the mole fraction of a solute in a ternary

76

76

system of solute-cation-anion. In addition the activity coefficient of the solute in this

ternary system of solute-cation-anion also needs to be adjusted to obtain the activity

coefficient of solute in a binary solute-IL system as shown in Equation(4-1)[123],

4-1

ternary

ion

ternary

i

ternary

i

binary

i xx .

4.4 Solubility Model

The key information used to choose a specific IL for gas processing is the solubility of

different gases in the given ILs. Based on this information the absorption capacity of ILs

can be compared and used to select ILs that absorb more of the targeted acid gases, CO2

and/or H2S, and less of the hydrocarbon compounds such as CH4 and C2H6. If this

information is not available experimentally it should be calculated in a reliable way.

Usually thermodynamic models such as UNIQUAC [113] require adjustable parameters

that are tuned based on available experimental data. The objective of this study is to

develop a generalized model that is able to predict the solubilities in new systems which

have not been studied, i.e., as discussed there are a large number of ILs considered for

which literature data is not available.

4.4.1 Bubble Point Pressure as a Measure of the Solubility of a Gas into an IL

Bubble point pressure of a mixture at temperature T and composition xi is the maximum

pressure at which equilibrium between a vapor and a liquid phase can exist. The bubble

point pressure can be used as a measure to describe the solubility of a gas in an IL. As

77

77

discussed in Chapter 2, the vapor pressure of ILs is negligible relative to the bubble point

pressure of the solute. Hence in a binary system of a solute A and an IL, we can assume

that the vapor phase consists only of component A. If, at the same equilibrium

temperature and pressure, gas A is more soluble in IL1 than in IL2, (Figure 4-1), the mole

fraction of A in IL1 will be higher than in IL2; also, at the same temperature and

composition, the equilibrium pressure of A in the A-IL1 system will be lower than the

equilibrium pressure of A-IL2 system.

Figure 4-1. Schematic of the P-X diagram at constant temperature of solute A in two

ILs in which solubility of A in IL1 is higher than the solubility of A in IL2, ─: IL1;

─: IL2

If the solubility of gas A in an IL is higher than the solubility of gas B in the same IL, at

the same equilibrium pressure and temperature, Figure 4-2, the mole fraction of A in IL

will be higher than B in IL. In other words, at the same temperature and composition of A

and B in binary systems of A-IL or B-IL, the equilibrium pressure of A is lower than B.

0 1

Pre

ssu

re

x

PA-IL1

PA

xA-IL2 xA-IL1 xA

PA-IL2

78

78

Figure 4-2. Schematic of the P-X diagram at constant temperature of solute A or B

in an IL in which solubility of A in IL is higher than the solubility of B in IL; ─: A-

IL; ─: B-IL

4.5 Solubility of CO2 in ILs

Three different approaches using COSMO-RS are used to predict the solubilities of

solutes in ILs. In all cases, it is assumed that the vapor pressure of IL is negligible and the

vapor phase consists only of the solute.

Based on the COSMO-RS model, the pseudo chemical potential can be calculated as,

4-2

dp Si

C

ii ,

where

i is the pseudo chemical potential, Equation 3-29, which was defined in Chapter

3. C

i in the combinatorial contribution which takes into account the shape and size

0 1

Pre

ssu

re

x

PA

P1

xB xA x1

PB

79

79

differences of the molecules in the system. The histogram ip or -profile of the

charge density on the molecular shaped cavity of the molecule i, and S is the surface

-potential of solvent, Equation 3-28.

4.5.1 Maiti’s Model [124]

In this model the activity coefficient of component i is calculated by Equation 3-31,

3-31

RT

iii

*

exp .

In which, *

i is the pseudo-chemical potential calculated from COSMO-RS model, and

i is the chemical potential of pure component i at T and p.

Equation (4-3) is used to calculate the chemical potential of pure component i,

4-3

)()( CCTi TTT ,

where, CT and are the empirical parameters and TC is the critical temperature of the

solute. From Maiti’s work for i = CO2, CT = -18547.524 kJmol

-1 and = -83.736 kJmol

-

1K

-1.

80

80

Equation (4-4) is used to calculate the equilibrium pressure of the solute-IL mixtures,

4-4

RT

Tx

pT

pxTxp iiii

i

,exp

,),( .

Where, p is the equilibrium pressure, is the fugacity coefficient of solute. The Soave-

Redlich-Kwong equation of state [125] is used to calculate the fugacity coefficients.

81

81

Figure 4-3. Experimental vs. calculated total pressure of CO2-IL mixtures using the

Maiti’s model [124] : bmim-Tf2N [56, 57, 63], : emim-Tf2N [71], : hmim-Tf2N

[119], : omim-Tf2N [56], : b2Nic-Tf2N [55], : bmim-BF4 [56, 57], : bmim-

CH3SO4 [58], : bmim-EtGLEtGLeC2SO4 [55], : bmim-NO3 [56], : bmim-PF6 [56,

59, 60, 62, 118], : bmim-TFA [55], : bmim-Triflate[56], : bmmim-PF6 [60],

:bmmim-BF4 [60], : C6H4F9mim-Tf2N [55], : emim-BF4 [64, 65], : emim-PF6 [67],

: emim-Triflate [66], : emmim-Tf2N [60], : hmim-Triflate [73], : hmpy-Tf2N

[55], : MeBu3N-Tf2N [57], : MeButPyrr-Tf2N [57], : N4111-Tf2N [55], : N4444-

doc [55], : N-bupy-BF4 [74], : pmim-Tf2N [75], : HOemim-BF4 [76], :

HOemim-Tf2N [70], +: HOemim-PF6 [69], : emim-C2SO4 [68, 69]

82

82

Equations (4-3) and (4-4) have been evaluated by Mortazavi-Manesh et. al. [116, 117]

based on 31 CO2-IL mixtures and the results are summarized in Table 4-5 and Figure 4-3.

This approach estimates the saturation pressure with an average error of AAR = 43%.

Figure 4-3 also shows a significant bias in the model with an average of -711 kPa. The

bias suggests that the ILs solubility model would perform better from a gas processing

point of view than what actual performance would be, which is a problem in reliable

process design. Also, there are large errors at high pressures which are in the range of

practical industrial absorption conditions (8000 kPa).

4.5.2 Mortazavi-Manesh et. al.’s Model-1 [117]

The existing bias of the Maiti’s model shown in Figure 4-3 and Table 4-5 and the larger

errors at high pressures indicate that there is residual pressure dependency that has not

been properly captured/ modelled. To overcome this problem Mortazavi-Manesh et. al.’s

model-1 was developed[117]. It starts with the following expression for the equilibrium

criteria:

4-5

ii pxpTp ,.

In Equation (4-5) it is assumed that the vapor pressure of the IL is negligible. Therefore,

the vapor phase consists only of pure solute i. The non-ideality of the vapor phase, , can

be determined by an equation of state such as Soave-Redlich-Kwong EOS [125].

83

83

i is the activity coefficient of solute i in the IL that can be calculated from Equation

(3-31). The pseudo chemical potential is calculated from COSMO-RS calculations,

Equation (3-31), [93, 126]. An empirical equation is proposed to describe the chemical

potential of standard pure state solute i,

4-6

CCcrti ppTTpT ),( ,

in which, , and crt are the parameters of the equation. CT and Cp are the critical

temperature and critical pressure of solute i. T and p are the temperature and pressure of

the mixture. The equilibrium pressure is calculated from Equations (3-31), (4-4), (4-5)

and (4-6). For i = CO2, crt -19444 kJmol-1

, = -79 kJmol-1

K-1

, and 0.115

kJmol-1

Pa-1

.

Due to the conditions for natural gas piping and common gas plant absorber design, the

calibration data from the database (Table 4-1), has been limited to pressures below 8000

kPa and loadings below 1. The parameters of the Equation (4-6) were obtained by

minimizing the following objective function, Equation (4-7),

4-7

2

2

.exp

2

.

,

,ln

Txp

TxpOF

CO

CO

calc

84

84

where, pcalc.

is the calculated equilibrium pressure and pexp.

is the experimental pressure of

the mixture. Note, the IL vapor pressure is assumed to be zero, these pressures are both

the partial pressure of the solute and the total pressure of the system.

The performance of the model is shown in Table 4-5 and Figure 4-4. By including the

pressure term the average error (AAR) for CO2 solubility in ILs was significantly

decreased from 43% to 23%. Perhaps more significant is the reduction in the bias from

711 to -63 kPa.

Table 4-5. Methods for predicting the total pressure of 31 CO2-IL mixtures

Method AAR %

AAD kPa

Bias kPa

Maiti’s Model [124] 43 712 -711

Mortazavi-Manesh et. al.’s Model-1 [117] 23 326 -63

In order to further evaluate the predictions of the model, the solubilities of CO2 in hmim-

FEP, MeButPyrr-FEP and ETT-FEP which were not included in the optimization were

tested [117] and the results are shown in Table 4-6. The average AAR for the present

model was 21.4% which shows significant improvement comparing with Maiti’s model

with average AAR of 59.6 %.

85

85

Table 4-6. Experimental total pressure of CO2-IL mixtures [127] vs. calculated

pressure. Comparison between Maiti’s model [124], Mortazavi-Manesh model-2

[116] and Mortazavi-Manesh model-1[117] using for the data which are not

included in the regression

IL

Mortazavi-Manesh

Model-1[117]

Mortazavi-Manesh

Model-2 [116] Maiti’s Model [124]

AAR

(%)

AAD

(kPa)

Bias

(kPa)

AAR

(%)

AAD

(kPa)

Bias

(kPa)

AAR

(%)

AAD

(kPa)

Bias

(kpa)

hmim-FEP 10.6 77 -74 11.6 109 -107 52.1 358 -359

MeButPyrr-FEP 23.6 161 -161 13.9 114 -114 61.6 419 -419

ETT-FEP 30.1 155 -155 22.1 121 -121 65.0 366 -366

Average 21.4 131 -130 15.9 115 -114 59.6 382 -382

86

86

Figure 4-4. Experimental vs. calculated total pressure of CO2-IL mixtures

Equations (4-4) and (4-6) [117], : bmim-Tf2N [56, 57, 63], : emim-Tf2N [71], :

hmim-Tf2N [119], : omim-Tf2N [56], : b2Nic-Tf2N [55], : bmim-BF4 [56, 57], :

bmim-CH3SO4 [58], : bmim-EtGLEtGLeC2SO4 [55], : bmim-NO3 [56], : bmim-

PF6 [56, 59, 60, 62, 118], : bmim-TFA [55], : bmim-Triflate[56], : bmmim-PF6

[60], :bmmim-BF4 [60], : C6H4F9mim-Tf2N [55], : emim-BF4 [64, 65], : emim-

PF6 [67], : emim-Triflate [66], : emmim-Tf2N [60], : hmim-Triflate [73], :

hmpy-Tf2N [55], : MeBu3N-Tf2N [57], : MeButPyrr-Tf2N [57], : N4111-Tf2N

[55], : N4444-doc [55], : N-bupy-BF4 [74], : pmim-Tf2N [75], : HOemim-BF4

[76], : HOemim-Tf2N [70], +: HOemim-PF6 [69], : emim-C2SO4 [68, 69]

87

87

4.5.3 Mortazavi-Manesh et. al.’s Model-2 [116]

This model starts with the well known equilibrium criteria,

4-8 l

i

v

i ff .

ILs have negligible vapor pressure, therefore the fugacity of the vapor phase is given by

the solute concentration only, Equation (4-9)

4-9

i

v

i pf .

Equation (3-31) relates the activity coefficient to the chemical potential. There are two

conventions to define the activity coefficients and the reference state chemical potential.

The conventions are usually based on the assumption that the solute approaches the ideal

behavior when its mole fraction approaches unity or zero. If the symmetric convention is

used the activity coefficient of solute approaches unity (Raoult’s law ideal solution),

when the composition of the solute approaches to pure component. If the asymmetric

convention is used the activity coefficient of the solute approaches unity (Henry’s law

ideal solution), when the composition of the solutes approaches infinite dilution. If the

symmetric convention is used for a system where a pure substance is not in the same

physical state as the mixture, the reference state of that component is treated as a

hypothetical pure substance with the same physical state as the solution. This situation

88

88

can happen when a gas is dissolved in a liquid solvent. Usually, the choice of convention

depends on the state of the pure component at the temperature and pressure conditions of

the mixture. The asymmetric normalization scale is used to express the fugacity of the

liquid phase,

4-10

iii

l

i Hxf .

The asymmetric activity coefficient,

i , is obtained using Equation (4-11),

4-11

i

ii

,

where, i , is the activity coefficient of the solute and

i , is the activity coefficient of

component i at infinite dilution. COSMO-RS, Equations (3-29) and (3-30), [93, 126] is

used to calculate i and

i .

The fugacity coefficient of CO2 is calculated using the Peng-Robinson equation of

state[128],

4-12

BZ

BZ

B

ABZZ

414.0

414.2ln

22)ln(1ln ,

89

89

where, is the fugacity coefficient of the pure component (in this case i = CO2) and Z is

the compressibility factor. The parameters A and B are defined in the Appendix B

The total pressure of the solute-IL systems can then be calculated using Equation (4-13),

4-13

i

iii Hxp

*

.

The Henry’s constant is the remaining parameter in Equation (4-13) to be described. If a

meaningful relationship between the Henry’s constant of solute and a simple IL

molecular parameter is found, solubilities beyond the calibration set can be provided. To

do so, several different options were explored where Equation (4-14) proved useful,

4-14

T

p

TH i

ln .

Equation (4-14) is an empirical relation similar to the Krichevsky-Kasarnovsky

expression, Equation (4-15), [129] except that the activity coefficient is calculated from

the COSMO-RS model and it is assumed that the vapor pressure of the IL is negligible.

4-15

90

90

RT

pvH

x

f

1*

12

1

1 lnlnln .

Although Equations (4-15) resembles Krichevsky-Kasarnovsky expression, the

parameters , and are empirical correlating parameters.

Mortazavi-Manesh et. al. [116] used 27 sets of experimental data for CO2-IL mixtures

and minimized the objective function, Equation(4-7), in order to obtain the parameters of

Equation (4-14). Similar to the development of the previous model, the experimental data

are limited to pressures below 8000 kPa and a maximum loading of 1. These criteria are

based on the pressures and temperatures in natural gas pipelines and absorbers in gas

plants. The results are shown in Figure 4-5 and TABLE 4-7.

91

91


using Equation (4-14) [116] without any IL parameters, : bmim-Tf2N [56, 57, 63],

: emim-Tf2N [71], : hmim-Tf2N [119], : omim-Tf2N [56], : b2Nic-Tf2N [55], :

bmim-BF4[56, 57], : bmim-CH3SO4 [58], : bmim-EtGLEtGLeC2SO4[55], :

bmim-NO3[56], : bmim-PF6 [56, 59, 60, 62, 118], : bmim-TFA [55], : bmim-

Triflate[56], : bmmim-PF6[60], : C6H4F9mim-Tf2N [55], : emim-BF4 [64, 65], :

emim-PF6[67], : emim-Triflate[66], :bmmim-BF4[60], : emmim-Tf2N [60], :

hmim-Triflate [73], : hmpy-Tf2N [55], : MeBu3N-Tf2N [57], : MeButPyrr-Tf2N

[57], : N4111-Tf2N [55], : N4444-doc[55], : N-bupy-BF4[74], : pmim-Tf2N[75]

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Calc

ula

ted

Pre

ssu

re (

kP

a)

Experimental Pressure (kPa)

92

92

To improve the results, ideally an easily measured or calculated physical property of IL

can be used to relate the Henry’s constant of solute in the ILs to the physical property of

IL. The , and parameters of Equation (4-14) can be expressed as simple linear

functions of a physical property of IL, .

4-16

21 ,

21 ,

and

21 .

The empirical parameters ii , and i are constants and is a property of the IL such

as the molecular weight. The IL parameters investigated in this study were molecular

weight, surface area of the molecule and the COSMO energy of the ILs. “The COSMO

energy is obtained from solving the Kohn-Sham equations for the molecule under the

influence of an external electrostatic potential generated from a virtual conductor. The

area is the surface area of the cavity of the molecule”[130]. The surface area and the

COSMO energy of the ions were calculated using the COSMOtherm software [93] and

are reported in Appendix A. In order to determine the controlling parameter, the

parameters were added to Equation (4-14) systematically. The results are summarized in

TABLE 4-7. For instance, using molecular weight of IL as a parameter for in

93

93

Equation (4-16) while using the common values for parameters and , the average

error was reduced from 31.25% to

Table 4-7. Predicting the equilibrium pressure of CO2-IL for 27 mixtures [55-60,

62-67, 71, 73-75, 118, 119] using different parameters in Equations (4-14) and (4-16)

[116]

Trial AAR% AAD, (kPa) Bias, (kPa)

1. 14.8976

= -1868.6241

= -3.394210-2

31.25 581 -123

2. MW

1 = 17.4792, 2 = -8.578310-3

1 = -2388.3363, 2 =1.7161

1 = 6.386810-3

, 2 = -7.761210-5

12.54 258 -38

3. MW

1 = 15.6439, 2 = -3.392810-3

1 = -1787.4922

1 = -1.609210-2

13.55 280 -38

4. MW

1 = 14.5124

1 = -1438.0321, 2 = -1.0475

1 = -1.640010-2

14.42 301 -40

94

94

TABLE 4-7. Continued

5. MW

1 = 15.4507

1 = -2020.7933

1 = 1.220110-1

, 2 = -5.392710-4

19.52 359 -127

6. COSMO energy

1 = 17.5434, 2 = 6.501210-7

1 = -2465.0706, 2 = -1.404010-4

1 = -1.439910-2

, 2 = 3.178410-10

13.00 264 -41

7. COSMO energy

1 = 15.5312, 2 = 2.024310-7

1 = -1836.8851

1 = -1.531010-2

14.08 285 -41

8. COSMO energy

1 = 14.6335

1 = -1561.0652, 2 = 6.220810-5

1 = -1.559510-2

14.98 308 -43

95

95

TABLE 4-7. Continued

9. COSMO energy

1 = 15.2884

1 = -1986.4901

1 = 7.744610-2

2 = 2.852310-8

20.60 397 -98

10. Surface Area

1 = 16.82363, 2 = -6.37171017

1 = -2053.6215, 2 = 7.16831019

1 = 3.773810-2

, 2 = -1.70281016

15.13 313 -48

96

96



parameter[116], : bmim-Tf2N [56, 57, 63], : emim-Tf2N [71], : hmim-Tf2N [119],

: omim-Tf2N [56], : b2Nic-Tf2N [55], : bmim-BF4[56, 57], : bmim-CH3SO4 [58],

: bmim-EtGLEtGLeC2SO4[55], : bmim-NO3[56], : bmim-PF6 [56, 59, 60, 62,

118], : bmim-TFA [55], : bmim-Triflate[56], : bmmim-PF6[60], : C6H4F9mim-

Tf2N [55], : emim-BF4 [64, 65], : emim-PF6[67], : emim-Triflate[66], :bmmim-


MeBu3N-Tf2N [57], : MeButPyrr-Tf2N [57], : N4111-Tf2N [55], : N4444-doc[55], :

N-bupy-BF4[74], : pmim-Tf2N[75]

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Calc

ula

ted

Pre

ssu

re (

kP

a)


97

97


using Equation (4-14) and 4-parameter Equation (4-16) with MW as IL parameter

only included in [116], : bmim-Tf2N [56, 57, 63], : emim-Tf2N [71], : hmim-

Tf2N [119], : omim-Tf2N [56], : b2Nic-Tf2N [55], : bmim-BF4[56, 57], : bmim-

CH3SO4 [58], : bmim-EtGLEtGLeC2SO4[55], : bmim-NO3[56], : bmim-PF6 [56,

59, 60, 62, 118], : bmim-TFA [55], : bmim-Triflate[56], : bmmim-PF6[60], :

C6H4F9mim-Tf2N [55], : emim-BF4 [64, 65], : emim-PF6[67], : emim-

Triflate[66], :bmmim-BF4[60], : emmim-Tf2N [60], : hmim-Triflate [73], :

hmpy-Tf2N [55], : MeBu3N-Tf2N [57], : MeButPyrr-Tf2N [57], : N4111-Tf2N [55],

: N4444-doc[55], : N-bupy-BF4[74], : pmim-Tf2N[75]

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Calc

ula

ted

Pre

ssu

re (

kP

a)


98

98


using Equation (4-14) and 6-parameter Equation (4-16) with COSMO energy as IL

parameter[116], : bmim-Tf2N [56, 57, 63], : emim-Tf2N [71], : hmim-Tf2N [119],

: omim-Tf2N [56], : b2Nic-Tf2N [55], : bmim-BF4[56, 57], : bmim-CH3SO4 [58],

: bmim-EtGLEtGLeC2SO4[55], : bmim-NO3[56], : bmim-PF6 [56, 59, 60, 62,

118], : bmim-TFA [55], : bmim-Triflate[56], : bmmim-PF6[60], : C6H4F9mim-

Tf2N [55], : emim-BF4 [64, 65], : emim-PF6[67], : emim-Triflate[66], :bmmim-


MeBu3N-Tf2N [57], : MeButPyrr-Tf2N [57], : N4111-Tf2N [55], : N4444-doc[55], :

N-bupy-BF4[74], : pmim-Tf2N[75]

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Calc

ula

ted

Pre

ssu

re (

kP

a)


99

99


using Equation (4-14) and 4-parameter Equation (4-16) with COSMO energy as IL

parameter only included in [116], : bmim-Tf2N [56, 57, 63], : emim-Tf2N [71],

: hmim-Tf2N [119], : omim-Tf2N [56], : b2Nic-Tf2N [55], : bmim-BF4[56, 57],

: bmim-CH3SO4 [58], : bmim-EtGLEtGLeC2SO4[55], : bmim-NO3[56], : bmim-

PF6 [56, 59, 60, 62, 118], : bmim-TFA [55], : bmim-Triflate[56], : bmmim-

PF6[60], : C6H4F9mim-Tf2N [55], : emim-BF4 [64, 65], : emim-PF6[67], : emim-




0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Calc

ula

ted

Pre

ssu

re (

kP

a)


100

100

Figure 4-10. Experimental total pressure of CO2-IL mixtures vs. calculated

pressure using Equation (4-14) and 6-parameter Equation (4-16) with COSMO

energy as IL parameter[116], : bmim-Tf2N [56, 57, 63], : emim-Tf2N [71], :

hmim-Tf2N [119], : omim-Tf2N [56], : b2Nic-Tf2N [55], : bmim-BF4[56, 57], :

bmim-CH3SO4 [58], : bmim-EtGLEtGLeC2SO4[55], : bmim-NO3[56], : bmim-

PF6 [56, 59, 60, 62, 118], : bmim-TFA [55], : bmim-Triflate[56], : bmmim-

PF6[60], : C6H4F9mim-Tf2N [55], : emim-BF4 [64, 65], : emim-PF6[67], : emim-




0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Calc

ula

ted

Pre

ssu

re (

kP

a)


101

101

13.55% (Trial 3). A slightly better correlation obtained by making the parameters and

functions of the molecular weight Trial 2). Interestingly, the simplest molecular

parameter used in this study, the molecular weight, was also the most effective in

capturing the differences between different ILs in CO2-IL systems. From an engineering

point of view it is more wise to keep the correlating parameters as simple as possible,

hence the molecular weight was chosen as the correlating parameter when i = CO2. The

recommended final form of the correlation is Trial 2 of TABLE 4-7. The results are

shown Figure 4-6. The results for other trials are shown in Figure 4-7 to Figure 4-10.

In order to further evaluate the predictions of the different models, the solubility of CO2

in hmim-FEP, MeButPyrr-FEP and ETT-FEP which were not included in the

optimization were tested [117]. The results are shown in Table 4-6. The average AAR for

the present model[116] is 15.9% which shows a significant improvement when compared

with the Mortazavi-Manesh et. al.’s Model-1[117] with an AAR of 21.4% and Maiti’s

Model [124] with an average AAR of 59.6 %. This model also provided better results in

terms of bias, shown in Table 4-6 the average bias of the present model is -114 kPa

compared with the Mortazavi-Manesh et. al. Model-1 with an average bias of -130 kPa

and Maiti’s Model with average bias of -382 kPa. Thus, the suggested form of the

correlation for solubility of CO2-IL systems is the 6-parameter Mortazavi-Manesh et.

al.’s Model-2[116] with molecular weight as the physical property of the IL. This model

will be used for predicting the solubility of CO2 in ILs in this study.

102

102

4.6 Solubility of H2S, CH4 and C2H6 in ILs

The approach for modeling the solubility of CO2 in ILs was also used to develop models

describing the solubilities of H2S, CH4 and C2H6 in ILs. Equation (4-16) was used to

linearly relate the parameters of Equation (4-14) with a physical property of IL. The IL

parameters used in this study were the molecular weight, surface area of the molecule,

and the COSMO energy of the ILs calculated and reported in Appendix A. The

parameters were added to Equation (4-14) systematically. The results are summarized in

Table 4-8 to Table 4-10. The results suggest that, for the H2S-IL and C2H6 cases, using

the molecular surface area parameter creates lower AAR, AAD and bias compared with

MW and energy parameters and is the best parameter to model these mixtures. However,

for the CH4-IL case, using MW as the parameter results in the lower AAR, AAD and bias.

It was shown that MW of ILs better describes the solubility of CO2 in ILs. Figure 4-11

shows that AAR (IL’s molecular surface area)/AAR (IL’s molecular weight) is linearly

correlated with the available molecular polarizability [131] of the solutes. Although only

four solute species have been tested for correlation, the results indicate that the solutes

with larger dispersive interaction (larger polarizability) correlate better with the surface

area of the solution molecules than those with low dispersive interaction (low London

dispersion forces). Figure 4-11 indicates that surface area is a useful correlating property

for the model with solute polarizabilities greater than 3.6 10-24

cm3.

103

103

Figure 4-11. Correlation between AARArea/AARMW and polarizability for CO2 (■),

H2S (), CH4 (▲) and C2H6 ().

Table 4-11 summarises the recommended model parameters for the binary mixtures of

CO2, H2S, CH4 and C2H6 in ILs. The results are summarized in Figure 4-12, Figure 4-13

and Figure 4-14. The models also show that the Henry’s constants for CH4 and C2H6 do

not have a strong pressure functionality.

In this study it was assumed that there is no water in the system. There are studies that

show the presence of water has essentially no effect on the solubility of CO2 in ILs [56,

132, 133]. However there are other studies suggesting that water can change the

solubilities of CO2 in ILs [74, 134].

0.55

0.7

0.85

1

1.15

1.3

1.45

2 2.5 3 3.5 4 4.5 5

AA

RA

rea

/ A

AR

MW

Polarizability, ' 1024 (cm3)

104

104


H2S-IL Mixtures

Trial

AAR%

AAD,

(kPa)

Bias,

(kPa)

1. 14.7891

= -2117.5596

= -3.0001 10-1

33.4 176 -40

2. MW

1 = 15.8227, 2 = -4.722810-3

1 = -2120.5839, 2 = 3.645010-1

1 = -2.204510-1

, 2 = -4.724210-16

26.5 151 -33

3. MW

1 = 15.4689, 2 = -3.603510-3

1 = -2005.2480

1 = -2.206310-1

26.5 151 -33

4. MW

1 = 14.3318

1 = -1635.7108, 2 = -1.1693

1 = -2.214610-1

26.6 151 -33

105

105

Table 4-8. Continued

Trial

AAR%

AAD,

(kPa)

Bias,

(kPa)

5. MW

1 = 14.8207

1 = -2114.0880

1 = -2.073410-14

, 2 = -1.007610-3

30.5 166 -53.3

6. COSMO energy

1 = 17.3967, 2 = 7.227610-7

1 = -2588.6714, 2 = -1.354910-4

1 = -5.285310-1

, 2 = -7.145210-8

26.5 150 -33

7. COSMO energy

1 = 15.2868, 2 =1.958910-7

1 = -2053.4115

1 = -2.287510-1

27.6 157 -35

8. COSMO energy

1 = 14.4902

1 = -1795.5383, 2 = 6.324910-5

1 = -2.301210-1

27.7 157 -35.4

106

106


Trial

AAR%

AAD,

(kPa)

Bias,

(kPa)

9. Surface Area

1 = 13.5909, 2 = 1.88121017

1 = -1243.5823, 2 = -2.13001020

1 = -2.353610-1

, 2 = -1.1955106

25.8 144 -32

10. Surface Area

1 = 15.7923, 2 = -4.6458 1017

= 2 = -1962.3615

= 2 = -2.355510-1

25.8 145 -32

11. Surface Area

1 = 14.2247

1 = -1449.9632, 2 = -1.51781020

1 = -2.353510-1

25.8 145 -32

12. Surface Area

1 = 14.7575

1 = -2090.5313

1 = -3.669510-14

, 2 = -9.67091016

30.3 161 -53.2

107

107


CH4-IL Mixtures

Trial

AAR%

AAD,

(kPa)

Bias,

(kPa)

1. 12.0401

= -66.6689

= -4.876510-2

39.4 582 -85

2. MW

1 = 13.3728, 2 = -5.704710-3

1 = 2.9415, 2 =-8.542010-2

1 = -5.204810-14

, 2 = -2.009910-18

25.9 115 -27

3. MW

1 = 13.4527, 2 = -5.958110-3

1 = -23.7668

1 = -5.649510-14

31.8 248 -67

4. MW

1 = 11.5601

1 = 600.1159, 2 = -1.9815

1 = -9.848810-14

32.1 255 -72

108

108


Trial

AAR%

AAD,

(kPa)

Bias,

(kPa)

5. MW

1 = 12.2103

1 = -119.059

1 = -5.080610-14

, 2 = -1.452210-4

36.4 463 -148

6. COSMO energy

1 = 14.22300, 2 = 6.685310-7

1 = -483.8101, 2 = -1.229510-4

1 = -7.623210-3

, 2 =-1.686210-17

35.6 362 -78

7. COSMO energy

1 = 12.7882, 2 =3.210110-7

1 = 11.7532

1 = -3.668110-3

35.4 333 -67

8. COSMO energy

1 = 11.4624

1 = 443.8596, 2 = 1.060010-4

1 = -3.050610-3

36.1 356 -63

109

109


Trial

AAR%

AAD,

(kPa)

Bias,

(kPa)

9. Surface Area

1 = 14.3037, 2 =-6.15011017

1 = 58.8408, 2 = -8.29191019

1 = -6.198810-7

, 2 =-3.1438105

30.8 267 -153

10. Surface Area

1 =15.1671, 2 = -8.6357 1017

= 2 = -226.6758

= 2 = -3.788210-14

31.0 270 -151

11. Surface Area

1 =12.1519

1 = 760.4290, 2 = -2.85121020

1 = -2.410910-15

30.9 288 -151

12. Surface Area

1 = 13.2956

1 = -448.2667

1 = -3.712810-14

, 2 = -1.55671016

39.2 497 -114

110

110


C2H6-IL Mixtures

Trial AAR%

AAD,

(kPa)

Bias,

(kPa)

1. 12.3173

= -659.1238

= -1.2022 10-1

54.1 428 -144

2. MW

1 = 12.4890, 2 = 5.2197 10-4

1 = -31.9887, 2 = -2.3939

1 = -6.378010-14

, 2 =-1.592710-5

25.9 115 -27

3. MW

1 = 15.5058, 2 = -7.281310-3

1 = -954.3318

1 = -7.011410-3

26.0 131 -39

4. MW

1 = 12.6856

1 = -86.1368, 2 = -2.2517

1 = -6.453010-3

26.0 114 -25

111

111


Trial AAR%

AAD,

(kPa)

Bias,

(kPa)

5. MW

1 = 12.5663

1 = -733.4690

1 = -4.366210-14

, 2 = -3.009410-4

52.7 434 -178

6. COSMO energy

1 = 11.9549, 2 = -8.524910-8

1 = -53.3053, 2 = 1.536510-4

1 = -1.335410-2

, 2 =-1.0288 10-20

29.8 157 -29

7. COSMO energy

1 = 14.5074, 2 =4.125310-7

1 = -838.0676

1 = -1.347210-2

29.6 169 -41

8. COSMO energy

1 = 12.3930

1 = -187.2379, 2 = 1.27510-4

1 = -1.328010-2

29.7 158 -31

112

112


Trial AAR%

AAD,

(kPa)

Bias,

(kPa)

10. Surface Area

1 = 13.4868, 2 = -1.1321017

1 = -13.0335, 2 = -2.72151020

1 = -2.445110-16

, 2 = -2.1930103

21.4 67 -20

11. Surface Area

1 =16.9376, 2 = -9.86841017

= 2 = -1081.9206

= 2 = -1.401810-3

22.3 95 -41

12. Surface Area

1 =13.0326

1 = 125.4950, 2 = -3.06791020

1 = -5.019510-15

21.5 68 -18

13. Surface Area

1 = 12.5503

1 = -730.2244

1 = -2.022710-1 1

, 2 = -2.95911016

53.2 435 -169

113

113

Table 4-11. Recommended Parameters for CO2-IL, H2S-IL, CH4-IL and C2H6-IL

Component

Parameter AAR

%

AAD,

(kPa)

Bias,

(kPa) Value

CO2 MW 1 = 17.4792, 2 = -8.578310

-3

1 = -2388.3363, 2 =1.7161

1 = 6.386810-3

, 2 = -7.761210-5

12.54 258 -38

H2S Surface

Area

1 = 13.5909, 2 = 1.88121017

1 =-1243.5823, 2 = -2.13001020

1 = -2.353610-1

, 2 = -1.1955106

25.8 144 -32

CH4 MW 1 = 13.3728, 2 = -5.704710

-3

1 = 2.9415, 2 =-8.542010-2

1 = -5.204810-14

, 2 = -2.009910-18

25.9 115 -27

C2H6 Surface

Area

1 = 13.4868, 2 = -1.1321017

1 = -13.0335, 2 = -2.72151020

1 = -2.445110-16

, 2 = -2.1930103

21.4 67 -20

114

114

Figure 4-12. Experimental total pressure of H2S-IL mixtures vs. calculated pressure

using Equations (4-13) , (4-14) and (4-16) with surface area as the IL parameter, :

bmim-BF4[78], : bmim-PF6 [77, 78], : bmim-Tf2N [78], : emim-C2SO4 [69], :

emim-PF6 [81], : emim-Tf2N[81], : hmim-BF4 [79], : hmim-Tf2N [79], : hmim-

PF6 [79], : HOemim-BF4 [76], : HOemim-Triflate [80], +: HOemim-PF6 [80], :

HOemim-Tf2N [80], : omim-Tf2N [82]

115

115

Figure 4-13. Experimental total pressure of CH4-IL mixtures vs. calculated pressure

using Equations (4-13) , (4-14) and (4-16) with MW as the IL parameter, : bmim-

BF4[83], : bmim-CH3SO4 [84], +: bmim-PF6 [59, 86], : hmim-Tf2N [85]

116

116

Figure 4-14. Experimental total pressure of C2H6-IL mixtures vs. calculated

pressure using Equations (4-13) , (4-14) and (4-16) with surface area as the IL

parameter, : bmim-Tf2N[57], : hmim-Tf2N[89, 121], +: bmim-PF6[57, 59, 86], :

bmim-BF4[83]

117

117

4.7 Screening ILs Based on the Solubility of CO2, H2S, CH4 and C2H6

Using the anions and cations shown in Appendix A, the solubility of CO2, H2S, CH4, and

C2H6 were calculate for 425 IL ion-pairs at 298.15 K and a partial pressure of 2000 kPa.

Partial pressure of 2000 kPa was chosen to ensure gas solubilities were being compared;

since, H2S at 298.15 K would be liquid above 2032 kPa[135]. Most treatment

applications for producing gas would have methane partial pressures well-above 2000

kPa.

To solve the non-linear Equation (4-13) for solubilities at these conditions, either the

activity coefficients would have to be recalculated using COSMO-RS at each iterative

solution or a model can be used for rapid convergence. For this comparison, the

parameters of the NRTL model, Appendix B, [136] were fitted using the activity

coefficients calculated from COSMO-RS for the solubility range of gas-IL mixtures. The

NRTL model was then used to solve Equation (4-13) for xi. The AAR% between the fitted

NRTL and COSMOtherm results are shown in Table 4-12. The results indicate that there

is less than 0.7% difference between CO2, CH4 and C2H6 activity coefficients calculated

from fitted NRTL and COSMOtherm. This difference is 6.5 % for H2S activity

coefficients.

118

118

Table 4-12. AAR% for the activity coefficients calculated between fitted NRTL model

and COSMO-RS model for different solutes in 425 ILs

Component AAR%

CO2 0.67

H2S 6.51

CH4 0.58

C2H6 0.64

For comparison purposes at practical conditions, the solubilities for all combinations of

anions and cations of Appendix A were calculated at 298.15 K and 2000 kPa and shown

in Figure 4-15 to Figure 4-18. ILs containing doc, FEP and Tf2N anions show the highest

average CO2, H2S, CH4 and C2H6 solubilities; whereas, Cl, NO3, BF4 and Lactate anions

show the lowest average solubilities for all four of these gases. ILs containing

C6H4F9mim, N4444 and b2Nic cations show the highest average solubility of the four

gases. A large absorption capacity for CO2 or H2S alone cannot be used to suggest that

an IL has the potential for gas treatment applications, because it also likely that it will

have a high capacity of CH4 and C2H6. In previous studies[124, 127], selectivity was

overlooked and only the absorption capacity was used for screening ILs for separation of

CO2 from a gas stream.

119

119

Figure 4-15. Solubility of CO2 in ILs at T = 298.15 K and 2000 kPa for different

combinations of anions and cations. Henry’s constants are calculated based on MW

of IL [116]

120

120

Figure 4-16. Solubility of H2S in ILs at T = 298.15 K and 2000 kPa for different


surface area of ILs.

121

121

Figure 4-17. Solubility of CH4 in ILs at T = 298.15 K and 2000 kPa for different

combinations of anions and cations. Henry’s constants are calculated based on the

MW of ILs.

122

122

Figure 4-18. Solubility of C2H6 at T = 298.15 K and 2000 kPa for different


surface area of ILs.

123

123

4.8 Selectivity of absorption

For gas sweetening containing primarily CH4, C2H6, H2S and CO2, it is important to

know the relative absorption (selectivity) of the major gas stream components. Thus the

solubilities calculated for CO2[116] , H2S, CH4 and C2H6 for the 425 ILs were used to

estimate the absorption selectivity [120],

4-17

PTx

PTxPTS

j

iji

,

,,/ ,

where, Sij is the selectivity of the absorbing component i over j; xi and xj are the mole

fractions of component i and j in ILs, respectively. Note, selectivities are approximations

of the actual selectivities one would experience in industrial practice, because the

solubilities are calculated for binary systems of solute-ILs and it is assumed that there are

no tertiary solute effects. Our primary criterion for a H2S and CO2 separation is to

choose potential solvents which will absorb more H2S and CO2 over CH4 and C2H6. Fit-

for-purpose screening criteria can be chosen based on the application, feed gas and outlet

gas specifications.

In some cases processes require higher selectivity for H2S over CO2 or CO2 slip. For

example, (1) conventional sour gas streams with high CO2 where sulfur recovery

furnaces require high H2S for effective performance of the Claus process, (2) some shale

gas streams where CO2 is in the per cent concentration levels and H2S is present only in

the ppm concentration levels, and (3) separation of Claus tail gas streams where there is

124

124

need to recycle only the remaining H2S. Therefore, for this study we will also focused on

higher H2S/CO2 selectivity as a secondary criterion.

For all combinations of cations and anions presented in Appendix A, 4/2 CHCOS ,

2/2 COSHS ,

62/2 HCSHS , 4/2 CHCOS ,

62/2 HCCOS was calculated using Equation (4-17). For each i and j

thejiS /of all the combinations of cations and anions were ranked. Then, the top 28% ILs

(120 ILs) with highest jiS / were chosen and the observed

jiS / were summarized as

follows:

4-18

2/2 COSHS 2.6,

4/2 CHSHS >51.0, 62/2 HCSHS 10.8,

4/2 CHCOS 18.8, 62/2 HCCOS 4.0,

The 28% cut off corresponded to a reasonable short-list of top-ranking ILs. This cut-off

could be expanded or reduced, depending on the scope of a follow-up study or research

program. Note, the top 28% was chosen, because this cut off reduced the overall possible

ILs to approximately 15% based on this modeling (58 ILs), i.e., 15% of all ILs are ranked

in the top 28% for each of the five selectivity relationship. Figure 4-19 shows the ILs

which rank in the top 28% of all five selectivities required for a basis sour gas treatment.

This ranking of potential ILs for gas treatment produces results which are different than

those based on capacity alone. Although ILs containing anions doc, FEP and Tf2N were

estimated to have the highest CO2 and H2S capacity, they are not suitable for gas

125

125

sweetening applications because they have poor SH2S/CH4 and SCO2/CH4. At T = 298.15 K

and p= 2000 kPa, ILs containing the anions BF4, NO3 and CH3SO4 have the most

number of combinations that meet the selective conditions for the top 28th

percentile.

ILs containing the cations N4111, pmg and tmg have the most number of combinations

that meet the selective conditions. Table 4-13 shows the ILs that are within the top 28th

percentile for five selectivities important for sour gas treatment (4/2 CHSHS ,

2/2 COSHS ,


62/2 HCCOS ) at T = 298.15 K and p = 2000 kPa. The advantage of this

ranking and selection is that (a) it has been used for an estimate of the best group of ILs

to consider for further process development (see chapter 5), (b) it can be used for further

experimental thermodynamic exploration for a more focused group of ILs, i.e. 58 versus

425 ILs (c) it can be used to target future synthesis studies.

126

126

Figure 4-19. Investigating the selectivity of different combinations of ILs at 298.15 K

and 2000 kPa. : ILs that are within the top 28th

percentile for five selectivities

important for sour gas treatment (4/2 CHSHS ,


4/2 CHCOS , 62/2 HCCOS ).

127

127

Table 4-13. ILs that are within the top 28th

percentile for five selectivities important

for sour gas treatment (4/2 CHSHS ,


4/2 CHCOS , 62/2 HCCOS )

IL Selectivity, S

H2S/CO2 H2S/CH4 CO2/CH4 H2S/C2H6 CO2/C2H6

bmim-BF4 3.0 62.8 20.9 14.7 4.9

bmim-CH3SO4 2.7 51.8 19.4 10.9 4.1

bmim-Cl 2.8 66.2 23.5 17.6 6.3

bmmim-BF4 2.9 57.4 19.6 12.8 4.4

bmmim-NO3 2.7 58.4 21.4 13.3 4.9

emim-BF4 3.1 66.4 21.6 20.3 6.6

emim-C2SO4 2.7 53.8 19.8 12.8 4.7

emim-CH3SO4 2.8 56.1 20.1 14.5 5.2

emim-NO3 3.1 66.7 21.7 20.8 6.8

emim-PF6 2.9 55.3 19.2 18.8 6.5

emim-TCA 3.2 74.0 23.0 14.7 4.5

emim-TFA 2.6 53.2 20.4 14.9 5.7

emmim-BF4 3.1 61.1 19.6 17.8 5.7

emmim-CH3SO4 2.7 52.3 19.7 13.4 5.0

emmim-L 2.7 56.7 21.4 12.9 4.9

emmim-NO3 2.9 63.7 21.9 18.6 6.4

emmim-TCA 3.0 63.6 21.5 12.1 4.1

emmim-TFA 2.6 51.3 19.4 13.8 5.2

128

128



hmim-Cl 2.7 61.0 22.6 12.9 4.8

HOemim-BF4 3.0 62.9 21.1 19.1 6.4

HOemim-L 3.1 58.6 18.9 14.1 4.5

HOemim-NO3 3.0 62.9 21.3 19.6 6.6

HOemim-PF6 2.7 54.8 20.2 17.7 6.5

MeButPyrr-BF4 3.1 62.3 20.3 15.0 4.9

MeButPyrr-CH3SO4 2.7 51.6 18.9 11.2 4.1

MeButPyrr-Cl 3.0 70.2 23.6 19.2 6.5

MeButPyrr-NO3 2.9 65.5 22.3 16.2 5.5

N2311-BF4 3.1 67.8 22.0 18.7 6.1

N2311-C2SO4 2.8 54.9 19.8 11.7 4.2

N2311-CH3SO4 2.8 57.3 20.5 13.9 5.0

N2311-NO3 3.0 69.4 22.8 19.3 6.3

N2311-TFA 2.8 55.9 20.0 14.4 5.1

N4111-BF4 3.2 71.9 22.4 17.9 5.6

N4111-CH3SO4 2.9 59.6 20.6 13.3 4.6

N4111-Cl 3.2 75.5 23.8 21.7 6.8

N4111-L 2.9 64.7 22.1 12.9 4.4

N4111-NO3 3.2 72.3 22.8 18.7 5.9

N4111-TFA 2.9 59.0 20.3 13.8 4.8


129

129


N-bupy-BF4 3.1 65.2 21.3 16.5 5.4

N-bupy-CH3SO4 2.7 53.0 19.5 12.1 4.4

N-bupy-Cl 2.9 66.1 23.0 19.2 6.7

N-bupy-L 2.8 57.9 21.0 11.8 4.3

N-bupy-NO3 2.9 63.1 22.0 16.4 5.7

N-bupy-TFA 2.7 51.3 19.4 12.3 4.6

pmg-BF4 3.1 61.6 20.0 18.7 6.1

pmg-Cl 3.4 69.2 20.2 23.4 6.9

pmg-L 2.9 58.6 20.6 13.8 4.8

pmg-NO3 3.1 62.8 20.2 19.4 6.3

pmg-TCA 3.0 65.2 21.5 12.5 4.1

pmim-BF4 3.0 60.0 20.0 12.5 4.2

pmim-NO3 2.8 61.6 21.8 12.8 4.5

tmg-BF4 3.2 69.5 21.9 21.1 6.6

tmg-CH3SO4 2.7 57.4 21.4 15.5 5.8

tmg-L 3.3 65.2 20.0 15.6 4.8

tmg-NO3 3.1 68.1 22.2 21.3 6.9

tmg-PF6 3.1 62.0 19.8 20.1 6.4

tmg-TCA 3.3 72.6 22.3 13.8 4.2

tmg-TFA 3.1 60.2 19.3 16.6 5.3

130

130

4.9 Summary

In this chapter the DFT/COSMO calculations were used to optimize the cation and anion

geometry of IL solvents. A database off the available experimental data for the solubility

of CO2, H2S, CH4 and C2H6 was built. Different models were tested to describe the

solubility of solutes in ILs, assuming that there was no chemical reactions between the

solutes and ILs. The model parameters were calculated based on the available

experimental data. In addition models were evaluated based on their ability to predict the

solubility of binary mixtures of CO2, H2S, CH4 and C2H6 in ILs in mixtures that were not

included in the regression experimental set. The proposed model is based on the Peng-

Robinson equation of state for gas fugacity coefficient calculations, asymmetric activity

coefficient calculated from COSMO-RS method, and an empirical Henry’s constant

correlation that is a function of the temperature and pressure of the system and the

molecular a physical property of the ILs.

This chapter showed that the molecular surface area of ILs can better describe the

dispersion interaction based solubility of the more polarizable solutes H2S and C2H6. The

solubility of less polarizable molecules, CO2 and CH4, are best described by the MW of

ILs. The models suggest that, in general, larger ILs with higher molecular surface area or

MW have higher solubility capacity. The following factors affect the selectivity of solute

absorption in ILs: (1) the solubility of each solute changes differently by the size of ILs;

(2) the model shows that the solubilities of CO2 and H2S in ILs are more pressure

dependent than the solubilities of CH4 and C2H6; therefore, at higher pressures more CO2

131

131

and H2S are absorbed in ILs. (3) the activity coefficient of the solute in ILs depends on

the interactions between the solute and IL and affects the solubility of the solute.

Using these models one can screen the ILs based on their absorption selectivity and

capacity of H2S and CO2 versus CH4 and C2H6. 425 ILs were screened at 298.15 K and

2000 kPa. ILs containing anions doc and FEP show the highest average CO2, H2S, CH4

and C2H6 absorption, but it was noted that high pure component capacity does not lead to

good candidates for separation. The absorption selectivities of CO2 and H2S over CH4

and C2H6 in ILs at 298.15 K and 2000 kPa were also calculated. ILs containing the

anions BF4, NO3 and CH3SO4 and containing cations N4111, pmg and tmg showed the

most number of combinations that met the criteria for separation of H2S and CO2 from

natural gas at 298.15 K and 2000 kPa. In order to choose an IL for gas processing other

physical properties of the ILs must also be considered, such as melting point, viscosity,

corrosivity, decomposition temperature, diffusion constant of gases into ILs.

132

132

Chapter Five: Conceptual Design of Gas Plants

5.1 Introduction

Predictive thermodynamic models were developed in the previous chapter to describe the

solubility of CO2, H2S, CH4 and C2H6 in ILs. In this chapter the best candidate ILs have

been chosen based on their absorption capacity and selectivity. Simplified gas treatment

processes were designed from a conceptual point of view. The results are compared with

the similar gas processes using the physical solvent Morphysorb and the chemical solvent

MDEA. Some of the potential difficulties in designing IL gas plants are investigated and

the advantages and disadvantages of these absorption process are discussed.

Table 4-13 provides a list of potential ILs for gas processing that were within the top 28th

percentile for five selectivities important for sour gas treatment(4/2 CHSHS ,

2/2 COSHS ,


62/2 HCCOS ). In addition to selectivity criteria, other concerns have to

be addressed in order to choose an IL for gas processing. ILs with relatively higher

melting point require extra processing such as preheating to avoid solidification of IL. To

avoid operational issues and the possibility of operator exposure, the toxicity of the ILs

must also be considered. For this reason, ILs that contain cyanide (CN) or fluoride (F)

groups were eliminated from the list. Energy consumption, solvent flow rate, solvent

makeup due to thermal degradation or chemical degradation, stability in water, corrosion,

hydrocarbon carryover and complexity of the gas processing unit are also important

factors. Finally, the viscosity of ILs is a major potential issue. In addition to problems

related to pumping at cold temperatures, viscosity is inversely correlated with mass

133

133

transfer efficiencies and is a major contributor to low mass transfer efficiencies for gas-

liquid contactors [137].

In total 8 ILs were chosen from Table 4-13 as absorbents in the following gas

conditioning case studies. These ILs are shown in Table 5-1.

5.2 Defining ILs for the Commercial Simulators

In this study, the gas treatment plants were simulated using VMGSim process simulation

software [138]. An equation of state was used for calculation of the thermodynamic

equilibria and physical properties. It is assumed that the gas stream consists of CO2, H2S,

CH4, C2H6 and water, used as a prototype for actual natural gas streams. The composition

of these components is changed depending on the feed considered; however, in general,

CH4 is the major component. Albeit simple, this choice of compounds allows for the

investigation of multi-component effects on the solubility of gases in ILs. Since ILs are

not included in the VMGSim pure component database, they were added manually as

hypothetical compounds. The minimum information required to define a hypothetical

compound are MW, critical properties and the density of ILs at 298 K. An equation of

state (APR EOS) was used to calculate the thermodynamic properties of the system.

Equations of state predict very low (not zero) vapor pressure for ILs which is realistic.

The substance specific parameters of the cubic equation of states are defined based on the

critical properties and acentric factor of each component.

134

134

5.2.1 Estimation of Critical Properties of ILs

There are several methods available to estimate the critical properties of compounds. In

one approach, empirical equations can be used to predict the critical temperature of

compounds based on temperature dependency of surface tension and density [139-141].

In another approach [119, 142], the Vetere method [143] can be used, where a fixed

number is assumed for the critical compressibility of ILs and the critical temperature and

pressure can be calculated knowing the density of compound at two different

temperatures. Valderrama et. al.[144-146] proposed a group contribution method in

which, the critical property of a compound is calculated by summing the contributions of

defined group of atoms and their frequency of occurrence of each group. In Valderrama’s

method, the normal boiling point (Tb), critical temperature (TC), critical pressure (Cp ),

critical volume (VC), acentric factor ( ) and the density ( ) of ILs are estimated using

Equations (5-1) to (5-6) respectively,

5-1

bb TnT 2.198 ,

5-2

2

CCMM

bC

TnTnBA

TT ,

5-3

2

CM

C

pnC

MWp ,

135

135

5-4

CMC VnEV ,

5-5

1loglog43

log437.0

4343

b

C

b

C

bC

C

b

C

CbC

Cb

p

p

p

p

TT

T

p

p

TTT

TT ,

and

5-6

bC

b

r

rr

r

r

TT

TT

ln

7

2,

where, AM = 0.5703, BM = 1.0121, CM = 0.2573, EM =6.75, r = 0.5703 and r = 1.0121

are constants. The parameters Tb, Tc, Pc and VC are defined based on the functional

groups forming the molecules [144-146].

Since Valderrama’s method is the only method that estimates critical properties, acentric

factor, normal boiling point and density at 298.15 K of ILs, this method was used in this

study to estimate these properties. The results are shown in Table 5-1.

136

136

Table 5-1. ILs chosen for gas processing. The CT Cp ωand

K298.15ρ were estimated

using Valderrama et. al. model[144-146]

IL Melting

point

CT

K

Cp

bar

298.15K

calculated

298.15K

measured

bmim-NO3 309 [147] 954.79 27.33 0.6436 1.129 1.1565

[148]

bmim-CH3SO4 269.05

[149]

1081.64 36.10 0.4111 1.234 1.20956

[150]

hmim-L - 995.41 20.58 1.1060 1.096 -

MeButPyrr-CH3SO4 - 1023.74 30.93 0.4152 1.190 -

N4111-CH3SO4 - 853.39 32.32 0.4306 1.181 -

N-bupy-CH3SO4 - 1038.38 36.17 0.3566 1.228 -

omim-NO3 - 1029.75 20.08 0.8058 1.060 1.0642

[151]

pmim-L - 974.35 22.08 1.0738 1.113 -

5.2.2 Validation of the VMGSim’s Cp Calculations

One of the important factors in comparing processes is the energy requirement of the

process; i.e., the heating load calculation. In order to validate the heat calculations of

VMGSim, the available experimental heat capacity of the ILs was compared with the

VMGSim predicted heat capacities. For the selected ILs, only experimental heat

capacities of bmim-CH3SO4 were available. Table 5-2 and Figure 5-1 compare the

137

137

experimental verse predicted heat capacities of bmim-CH3SO4. As shown in Table 5-2,

the AAR% between the measured heat capacity and VMGSim prediction is 11%. This is a

fairly good prediction since no ideal gas CP data are available and the enthalpy and CP

were estimated by the APR equation of state.

Figure 5-1. Cp of bmim-CH3SO4, : experimental [149]; ─: VMGSim[138]

predictions

200

250

300

350

400

450

280 300 320 340 360 380

Cp

, k

J/m

ol.

K

T(K)

138

138

Table 5-2. Comparison between the experimental data and VMGsim predictions of

the heat capacity of bmim-CH3SO4

T(K) Experimental VMGsim

283.1 226.2 295.3

293.1 246.3 305.3

303.1 266.6 315.2

313.1 286.9 325.1

323.1 307.2 335.0

333.1 327.4 344.8

343.1 347.7 354.5

353.1 368.0 364.2

363.1 388.3 373.9

373.1 408.5 383.4

AAR% 11.3

5.3 Equation of State’s Set up for Mixtures

The Advanced-Peng-Robinson (APR) equation of state was chosen for this study,

Appendix B. The APR EOS has a volume shift correction which improves the calculation

of both liquid and vapor densities particularly in the vicinity of the critical point [152,

153] and specially determined parameters for important compounds and mixtures

typically encountered in natural gas processing.

139

139

VLE data for binary mixtures of CO2-IL, H2S-IL, CH4-IL, C2H6-IL and water with IL are

required to optimize the binary interaction parameters of the EOS. The VLE data for CO2-

IL and CH4-IL for bmim-CH3SO4 are available [58, 84]. The models proposed in Chapter

4 were used to obtain the VLE data for other IL-solute mixtures.

For the mixtures that have experimental water-IL data, COSMO-RS predictions were

found to be adequate at low water fractions. Figure 5-2 shows a comparison of the

experimental and the COSMO-RS calculations for water in bmim-Tf2N (AAR of 5.6%).

The experimental data indicates LLE at compositions greater than 0.5, but COSMO-RS

calculation does not show LLE at 0.5. Since in gas sweetening applications, the water

composition is significantly lower than 0.5, the COSMO-RS calculations were deemed

adequate. Figure 5-3 illustrates the experimental verses the COSMO-RS calculations for

water in emim-C2SO4 (AAR of 19.11%). Based on the available examples of water-IL

mixtures and the fact that the water composition of the gas plant applications is small, the

VLE data of the ILs-water mixtures chosen in this study were predicted using COSMO-

RS calculation without further modification.

140

140

Figure 5-2. Comparison of total pressure in water-[emim][Tf2N] mixture. ─ :

COSMO-RS method verses experimental data. Calculations are done using

COSMOthermX software[93], ○:[154] ,▲:[155]. AAR = % 5.6

Figure 5-3. Comparison of total pressure in water-[emim][C2SO4] mixture. ─:

COSMO-RS method verses experimental data [156]. Calculations are done using

COSMOthermX software, : 322.9 K, ■: 312.9 K, ▲: 302.9 K. AAR = %19.1

0

10

20

30

40

50

60

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

p(T

ota

l), k

Pa

xWater

0

2

4

6

8

10

12

14

0 0.2 0.4 0.6 0.8 1

p(T

ota

l), k

Pa

xWater

141

141

A temperature dependent relation, Equation (5-7) was chosen to describe the binary

interaction parameters of the APR EOS,

5-7

TkT

kkk ij

ij

ijij lnln2

1

0 .

In order to fit the parameters of Equation (5-7), the following objective function was

minimized.

5-8

2

.exp

.

,

,ln

Txp

TxpOF

calc

The advantage of this objective function, Equation (5-8), is that it weighs the calculated

values which are certain order of magnitude higher than the experimental values the same

as values that are the same order of magnitude lower than the experimental data. Also,

the errors at high pressure are treated the same way as the errors at low pressures. Table

5-3 shows the coefficients of the binary interaction parameter of APR EOS.

142

Table 5-3. APR’s binary interaction parameters of Equation (5-7) for solute-IL mixtures

IL CO2 H2S CH4 C2H6 Water

bmim-NO3 0kij 2.82 2.93482 0.14543 0.11315 0.009

1kij 0.14025 0.389698 0.359469 -0.23799 0.01667

2kij -0.47847 -0.504883 0.14543 0.11315 -0.055

bmim-CH3SO4 0kij 0.12666 2.99962 -0.423537 0.232871 0.006403

1kij 0.012213 -2.54187 -0.029870 0.248428 0.011738

2kij -0.010304 -0.52025 0.108237 -0.014449 -0.043154

hmim-L 0kij -0.409504 2.751150 -1.167199 -0.044772 0.009806

1kij 0.106802 -2.807958 0.721887 -0.004797 0.017583

2kij 0.076027 -0.477686 0.190347 0.030017 -0.061611

143



MeButPyrr-CH3SO4 0kij 0.0337 2.8319 0.620390 0.10583 0.009834

1kij -0.000836 -1.255704 -0.098520 -0.043305 0.017566

2kij 0.00119 -0.49097 -0.071520 0.010806 -0.047240

N4111-CH3SO4 0kij -0.724309 2.991965 -0.452163 0.004416 0.010784

1kij 0.428172 -2.43493 0.320204 -0.029297 0.018169

2kij 0.140545 -0.515268 0.134873 0.040803 -0.063665

N-bupy-CH3SO4 0kij -0.208591 2.93257 0.525705 0.003553 0.009

1kij 0.017178 0.187881 -0.039174 -0.02234 0.016667

2kij 0.043636 -0.507713 -0.053235 0.030822 -0.055

144



omim-NO3 0kij -0.736299 2.306833 -0.078233 -0.76876 0.010296

1kij 0.441199 13.679931 0.026757 -0.098336 0.017583

2kij 0.137309 -0.409957 0.067604 0.044431 -0.061611

pmim-L 0kij -0.473926 6.869017 -0.092917 -0.168027 0.009806

1kij 0.342533 -204.373095 -0.017044 -0.193476 0.017583

2kij 0.087540 -1.080694 0.053358 0.053955 -0.061611

145

145

5.4 Kwoen Gas Plant

At this point all the required modeling tools are available for the simulation of a gas

treatment plant using IL as the solvent. To make the study as realistic as possible, the

feed conditions of an existing gas sweetening plant were. In order to compare the

operating conditions of different gas plants, a fixed product composition (H2S % in the

treated gas) was used. In this way the solvent recirculation, hydrocarbon carryover,

energy consumption, etc. of different gas plant designs, used as case studies, can be

compared.

The Kwoen gas plant[157] located in North Eastern British Colombia, Figure 5-4, was

Figure 5-4. Approximate locations of Kwoen and Pine River gas plants [158].

146

146

designed to de-bottle neck the Pine River gas sour plant. The feed to the Pine River plant

is produced within the Grizzly Valley (ca. 200 km), which extend across the Alberta and

British Columbia provincial borders and include several very sour natural gas reservoirs.

The sour gas feed composition ranges from 5% to 40% acid gas and contributing

producers must dehydrate the sour gas prior transporting. As the main trunk line to

Kwoen is carbon steel and cannot handle condensable water.

The Pine River plant was expanded in 1994 to a capacity to process 560 MMscfd feed gas

and 94 MMscfd acid gas which then produces some 2000 Long Tons (LT/day) of

elemental sulphur. The current design plant feed for Pine River contains 16.8% total acid

gas, 9.3% H2S and 7.5 % CO2. The acid gas content from the Grizzly Valley trunk line

was higher than the Pine River design value at 21.0%. The plant capacity was therefore

fully utilized because the maximum acid gas into the plant had been reached. However,

the raw feed from the Grizzly Valley was 120 MMscfd less than the plant design;

therefore, Pine River could produce more gas if there was less H2S at the inlet. By the

year 2000 new sour gas in excess of 130 MMscfd had to be processed. To process the

additional sour gas, either an expansion in the Pine River gas plant had to be completed

or a new processing facility had to be installed to reduce the acid gas content of the feed

to the Pine River. The second option was selected and built at Kwoen. It is important to

note that the Kwoen gas plant does not produce sales gas. It is a bulk acid gas removal

unit and acid gas re-injection facility. The Kwoen plant removes 28 MMscfd acid gas

from the 130 MMscfd increased sour gas flow, thus allowing the Pine River gas plant to

operate at capacity. Table 5-4 shows the conditions encountered in the Kwoen plant.

147

147

In this study the conditions of Kwoen gas plant [157] were used and several IL-gas plants

were simulated and their performance compared with the Kwoen gas plant which uses

Morphysorb as the absorbing solvent. In addition, an Amine gas plant is designed based

on the conditions shown in Table 5-4. The performances of all these different gas plants

are compared.

Table 5-4. Conditions considered for the gas plants

Sour Gas, Feed

Gas flow, MMscfd 300

CO2 % 8.6

H2S % 13.6

CH4 % 77.5

C2H6 % 0.22

H2O % 0.08

Treated Gas

H2S % 5.33

5.4.1 IL Gas Sweetening Plant

The ILs of Table 5-1 were used as the gas conditioning solvent in the gas plants that

were compared using the conditions shown in Table 5-4. VMGSim software [138] along

148

148

with the optimized binary interaction parameters from Table 5-3 were used to simulate

the plants. Figure 5-5 shows a schematic of the IL-gas sweetening plant.

5.4.1.1 Description of the units in the IL gas sweetening plant

Inlet Separator

Before entering the absorber, the gas passes through an inlet separator where entrained

liquid and other contaminants such as corrosion inhibitors are removed. The inlet feed

gas then passes through a feed gas pre-heater prior entering the absorber. The heater is

only required for ILs with a melting point close to the feed temperature, i.e., to avoid

solidification of the solvent. For ILs with melting points far from the feed condition, this

heater need not be included in the gas plant design/simulation.

Absorber

The sour gas from the inlet separator enters the bottom of a trayed absorber and lean IL is

fed to the top of the absorber. As explained in Chapter Four, the solubility of H2S and

CO2 increases by increasing the pressure, but the solubilities of CH4 and C2H6 do not

change significantly with a change in pressure. For this reason, ILs remove CO2 and H2S

more selectively at higher pressures. The pressure of the absorber varies between about

74 and 75 bar (pressure drop of 1 bar). The composition of the lean gas is controlled by

manipulating the flow rate of the lean IL entering the absorber. The capacity of

absorption of ILs increases by decreasing the temperature. For this reason a relatively low

temperature of 288 K is used for the absorber. Other factors that control the operating

149

149

temperature of the absorber are the melting point temperature of ILs and is the primary

variable in the selection of the operating temperature of the absorber. The temperature of

the absorber has to be well above the melting temperature of the ILs to avoid solidifying

of ILs and plugging. The temperature dependency of the viscosity of the ILs is another

important factor, because at higher viscosities the ILs, pumping of the solvent is more

difficult and more importantly higher viscosity of the ILs will affect the efficiency of

absorption for each component. The heaters H1 and H2 are used to adjust the temperature.

The temperature of the absorber is assumed to be 288 K for ILs with low melting

temperatures such as bmim-CH3SO4 or ILs with no melting point or viscosity

information. Higher temperatures are required for ILs with higher melting points such as

bmim-NO3.

150

Figure 5-5. Gas sweetening plant designed in this study using IL as the absorbent. The simulation is done using VMGSim[138].

T-1: inlet separator; T-2, T-3, T-4, T-5: Flash tanks; H-1, H-2: Heaters, P-1: Pump; CP-1, CP-2, CP-3: Compressors; AC-1,

AC-2, AC-3 air-coolers.

151

151

Recovering the Absorbed Hydrocarbons

In the absorber, methane and ethane are dissolved in the liquid phase. In order to recover

the absorbed hydrocarbons, flash tanks are used (T-2, T-3 and T-4, Figure 5-5). The rich

IL from the absorber passes through these three flash tanks allowing the methane and

ethane to flash off due to depressurization. The pressure of each flash tanks was

optimized using VMGSim[138] to maximize the recovery of methane. These selected

pressure profile resulted in the highest CH4/H2S ratio in the vapor stream of the flash tank

and the lowest CH4 percentage in the acid gas stream. The pressures of the flash tanks

are reported in Table 5-1. The pressure and temperature of the recovered gases are

recompressed and, cooled and recycled back as feed to the absorber.

Regenerating the IL

In order to regenerate the rich IL, a final flash tank operating at near atmospheric pressure

can be used to remove the absorbed H2S and CO2 from the IL. Because the APR EOS

calculates very low, but not zero vapor pressure for ILs, the make-up IL is added to the

regenerated IL to satisfy the mass balance of the IL circulation at steady state conditions.

The pressure and temperature of the lean IL are adjusted and fed back to the absorber.

Figure 5-9 to Figure 5-24 show the results of the simulation for the IL gas plants.

152

152

Table 5-5. Pressure of the absorber and flash tank for IL-gas plant

T-2 T-3 T-4 T-5 Absorber

Pressure, bar 29.92 14.16 12.36 1.05 74.05 - 74.81

5.4.2 Gas Plant with Morphysorb as the Absorbent

Morphysorb is a physical solvent licensed by Uhde GmbH (KU) [157], It consists of N-

Formylmorpholine (NFM) and Acetylmorpholine (NAM) [4, 157]. Figure 5-6 shows the

structure of NFM and NAM. Only NFM has been included in the database of VMGSim.

The conditions of Table 5-4 are used and it is assumed that NFM represents Morphysorb

Figure 5-6. The chemicals in Morphysorb[4, 157]; a: N-Formylmorpholine (NFM),

b:Acetylmorpholine (NAM)

The APR EOS is again selected. Figure 5-7 shows the Process Schematic of the

Morphysorb gas plant. Table 5-6 compares published [157] versus calculated

compositions of the lean gas and acid gas of Kwoen gas plant. As shown in Table 5-6 the

published and calculated compositions of the lean gas and acid gas are in a very good

agreement. This would indicate that the assumption of representing Morphysorb by NFM

is adequate for the purposes of this study.

153

153

Table 5-6. Published[157] using Morphysorb versus calculated using NFM

composition of the upgraded gas and acid gas for Kwoen gas plant

Published

Mole%

Calculated

Mole%

Upgraded gas Acid gas Upgraded gas Acid gas

H2S 5.33 78.71 5.33 72.69

CO2 7.21 19.6 6.87 21.24

CH4 86.81 1.47 87.59 5.48

C2H6 0.23 0.09 0.19 0.39

5.4.2.1 Description of the units in the Morphysorb gas sweetening plant

Absorber

The sour gas from the inlet separator enters the bottom of the tray absorber and the

Morphysorb enters the top of the absorber. The pressure of the absorber varies between

74 and 75 bar. The composition of the lean gas is adjusted by manipulating the flow rate

of the lean Morphysorb to match the desired H2S concentration in the lean gas.

154

154

Recovering the Absorbed Hydrocarbons

The rich Morphysorb from the absorber is sent to two flash tanks that are used to recover

the methane and ethane. The pressures used in the simulation for each recovering-flash-

tanks were obtained from the published data of the Kwoen gas plant [157] and the

pressures are reported in Table 5-1. The pressure and temperature of the recovered gas

are controlled using compressors and air-coolers and the flashed gas is recycled back to

the absorber.

Regenerating the Morphysorb

Two flash tanks are used to regenerate the rich Morphysorb thus removing the absorbed

H2S and CO2 from the solvent [157]. The makeup Morphysorb is added to the

regenerated solvent. The pressure and temperature of the solvent are adjusted using a

pump and a heater and the lean solvent is then fed back to the absorber.

Table 5-7. Pressure of the flash tanks for the Morphysorb-gas plant [157]

T-2 T-3 T-4 T-5

P/ bar 29.30 12.76 4.48 1.76

155

Figure 5-7. Gas sweetening plant using Morphysorb as the absorbent. The simulation is done using VMGSim[138] . T-1: inlet

separator; T-2, T-3, T-4, T-5: Flash tanks; H-1, H-2: Heaters, P-1: Pump; CP-1, CP-2, CP-3: Compressors; AC-1, AC-2, AC-3

air-coolers

156

156

There are some differences between the Morphysorb gas plant [157] and the IL gas plant

simulations:

(1) The temperature of the feed gas in the IL gas plant is controlled and can be

increased if the melting point of the IL is not low enough (close to 288 K),

(2) The recovery of the IL is performed in 3 steps in which the pressure of each step

is chosen to maximize the amount of recovered hydrocarbons and the recovery of

the Morphysorb is done is two steps,

(3) A single regeneration step is used in the IL gas plat, whereas in the Morphysorb

gas plant the regeneration takes place in two stages.

5.4.3 Gas Plant with Chemical Solvent

Aqueous mixtures of alkanolamines are conventionally used as chemical solvents for gas

sweetening. An amine-gas plant was simulated as part of the thesis case studies, Table

5-4 present the data used for Methyl Diethanol Amine (MDEA) gas plant, simulated

using VMGSim software [138]. The amine property package was chosen for this

simulation since it rigorously takes into account the chemical reactions and also provides

the necessary estimates the efficiency of trayed contactors based on the different

chemical reaction rates and consequent separation efficiencies between H2S and CO2.

The strength of MDEA is commonly chosen between 20 to 50 %wt [159], but the

optimum MDEA concentration depends on the viscosity of the solution. The viscosity of

the MDEA solution increases at higher concentrations which impacts the tray efficiency.

At constant solvent flow rate the loading of the MDEA increases at higher

concentrations. Due to corrosion considerations, the maximum loading of about 0.5

157

157

mole/mole is practical in carbon steel equipment with no inhibitors. If the inhibitors are

added the loading can be increased to 0.7-0.8 mole/mole. Also at higher loading the

temperature of the solution in the absorber increases due to more absorption of acid gas

through the exothermic reactions taking place in the absorber.

5.4.3.1 Description of the Units of the MDEA Gas Sweetening Plant

Absorber

The field gas from the inlet separator enters the bottom of a trayed contactor. The

simulated absorber has 25 stages and 1 bar pressure drop. The lean amine enters at the

top of the contactor. CO2 and H2S react with the aqueous solution of MDEA. Low

temperature favors these reversible exothermic reactions. The lean gas exists at the top of

the absorber. The H2S percentage of the lean gas is controlled by the MDEA flow rate to

the absorber and was adjusted to produce same H2S concentration as that of the IL-gas

plants. In this way, the other parameters of the two types of gas plants can be easily

compare. The rich MDEA containing the reacted CO2 and H2S leaves the bottom of the

absorber.

158

Figure 5-8. Gas sweetening plant using MDEA as the absorbent using VMGSim[138]. T-1: inlet separator; T-2: Flash tank;

Hx-1: Heat exchanger; C-1: cooler; P-1: pump

159

159

Regenerating the MDEA

The rich amine from the contactor is passed through a flash tank to remove the absorbed

hydrocarbons. The reaction between CO2 and H2S with amine is reversible and increasing

the temperature favors the reverse reaction. The rich MDEA regenerator feed is preheated

in a feed bottoms heat exchanger using the hot regenerated amine (lean amine) from the

bottom of the regeneration tower. The re-boiler provides energy to reverse the reaction

between H2S and CO2 and amine. The over head acid gas steam is cooled down in the

condenser to recover some of the water and amine. The hot regenerated amine is further

cooled before being recycled to the absorber, thus completing the loop. Since water is lost

in the acid gas leaving the regenerator as well as small amount of MDEA, makeup water

and amine are added to the regenerated amine/lean MDEA. The main difference between

the Morphysorb gas plant and an amine gas plant is the regeneration step. In Morphysorb

gas plants the regeneration is accomplished via pressure reduction in series of flash

drums, while in an in amine gas plant a reboiled stripper is needed to regenerate the rich

amine.

Figure 5-9 to Figure 5-24 summarize the simulation results for the IL, Morphysorb gas

plant, and MDEA gas plants. The H2S content of the lean gas was kept constant for all

cases by adjusting the solvent flow rate. This was done in order to compare other key

variables of gas plants such as the amount of CO2 and H2S removed from the feed gas

stream, the amount of hydrocarbon carry over into the acid gas stream, the solvent flow

rate required, the total energy required, the solvent makeup rate, and the water content of

the lean gas.

160

160

Figure 5-9 shows the CO2 mole percent of the lean gas for ILs, Morphysorb and MDEA

gas plants. It can be seen that the CO2 mole percent is almost the same for all the cases.

Among ILs, pmim-L has the lowest CO2 of 7.3%, the Morphysorb has a CO2 of 6.8% and

for MDEA CO2 is 8.1%.

Figure 5-9. CO2 mole percent in the upgraded gas for different ILs, Morphysorb

and amine (45% wt MDEA) gas plants at fixed H2S mole percent in the upgraded

gas (5.33%)

Figure 5-10 shows the CH4 mole percent in the lean gas for the selected ILs, Morphysorb

and MDEA gas plants. The methane mole percent is similar for all the cases. The

maximum CH4 mole percent among ILs occurs in the case of pmim-L and N-bupy-

CH3SO4, 87%. For the case of Morphysorb and amine it is 87.6% and 85.9%,

respectively.

0

2

4

6

8

10

12

14

CO

2 m

ole

%

161

161

As explained in Chapter Four, a good gas processing solvent should absorb valuable

hydrocarbons as little as possible. Figure 5-11 shows that the mass flow rate of methane

in the lean gas is maximum (185.8 tons/hr) for the bmim-NO3 and pmim-L. The

minimum methane mass flow rate occurs in the case of Morphysorb, 184.2 tons/hr. This

indicates that IL gas plants can produce more methane than the Morphysorb plant.

Actually, these numbers are about the same.

Figure 5-10. CH4 mole percent in the upgraded gas for different ILs, Morphysorb


gas (5.33%)

84

85

86

87

88

89

90

CH

4 m

ole

%

162

162

Figure 5-11. CH4 mass flow rate in the upgraded gas for different ILs, Morphysorb


gas (5.33%)

In the case of MDEA gas plant, the CH4 flow rate is slightly lower than bmim-NO3,

185.7 tons/hr.

Figure 5-12 shows the solvent flow rate for different ILs, Morphysorb and MDEA gas

plants. Among the IL gas plants, hmim-L has the lowest solvent flow rate, 649 m3/hr.

Morphysorb requires a flow rate of 639 m3/hr. bmim-NO3 requires the maximum flow

rate of 1000 m3/hr. This occurs due to the melting point of bmim-NO3 being 309 K

(Table 5-1). To prevent solidification, the plant must operate at higher minimum

temperature than other ILs and this higher temperature does significantly impact the

absorption capacity of the solvent. The MDEA gas plant requires the lowest flow rate of

401 m3/hr.

183

183.5

184

184.5

185

185.5

186

CH

4 m

ass

flo

w r

tae

(ton

s/h

r)

163

163

Figure 5-12. Solvent flow rate for different ILs, Morphysorb and amine (45% wt

MDEA) gas plants at fixed H2S mole percent in the upgraded gas (5.33%)

Loading is defined as the moles of CO2 and H2S absorbed per each mole of pure solvent,

5-9

Solvent

S2H2COLoading

n

nn .

Figure 5-13 shows the loading of the rich solvent exiting from the absorber. MDEA

loading is highest (1.45) followed by hmim-L with the loading of about 0.81. This agrees

with hmim-L’s lowest flow rate among the ILs. Although MDEA has the maximum

loading of 1.45, this loading in not normally used in an MDEA gas treating plant due to

0

200

400

600

800

1000

1200

Solv

ent

flow

rate

(m

3/h

r)

164

164

corrosion considerations. The maximum loading of MDEA is typically limited to 0.45 to

0.5 [160, 161]

Figure 5-13. Loading of rich solvent from the absorber, Equation (5-9), for different

ILs, Morphysorb and amine (45% wt MDEA) gas plants at fixed H2S mole percent

in the upgraded gas (5.33%)

Figure 5-14 shows the loading of lean solvent in which MDEA has the lowest loading of

0.004 and is within the recommended range (0.004 – 0.01) [160, 161]. Among the ILs,

bmim-NO3 has the lowest loading of 0.028, due to its higher operating temperature.

Morphysorb has the highest loading of the lean solvents.

0

0.2

0.4

0.6

0.8

1

1.2

Load

ing

165

165

Figure 5-14. Loading of lean solvent from the absorber, Equation (5-9), for different



Figure 5-15 shows the flow rate of the makeup solvent. The ILs have very low makeup

are very low flows due to their negligible vapor pressures. Small losses are estimated

from the simulation since the APR EOS estimates very small but non-zero vapor

pressures for the ILs. The makeup flow rate of Morphysorb is 84.26 kg/hr, which

indicates its higher operating cost due to solvent makeup. This suggests that an advantage

of using ILs for gas processing is their negligible make-up requirement. The makeup flow

rate for pure MDEA is 4.8 kg/hr. Also, the MDEA gas plant requires 13.7 m3/hr water

makeup.

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

Load

ing

166

166

Figure 5-15. Solvent make up for different ILs, Morphysorb and amine (45% wt


Figure 5-16 to Figure 5-19 show the water content of the lean gas. The IL-gas plants

show lower water content in comparison with Morphysorb gas plant. Among ILs, bmim-

NO3 shows the lowest water content, 14.6 kg/hr because of its higher operating

temperature, followed by omim-NO3, 21.9 kg/hr. The water content of upgraded gas in

Morphysorb gas plant is 52.5 kg/hr. This suggests that another advantage of using IL-gas

plants is that they can be used for the dehydration of gas; whereas, the amine gas plants

increase the moisture of the gas since the solvent is an aqueous solution. Because the

amine gas plant increases the water content of the upgraded gas, a separate dehydration

process is required. Figure 5-17 and Figure 5-19 show the water content of the lean gas is

0.39% or 943.8 kg/hr water.

0

10

20

30

40

50

60

70

80

90

Mak

eup

(k

g/h

r)

167

167



(5.33%)



(5.33%)

0

0.005

0.01

0.015

0.02

0.025

Wate

r (m

ole

%)

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

Wate

r(m

ole

%)

168

168

Figure 5-18. Water flow rate in up-graded gas for different ILs, Morphysorb and


(5.33%)

Figure 5-19. Water flow rate in upgraded gas for different ILs, Morphysorb and


(5.33%)

0

10

20

30

40

50

60

Wate

r fl

ow

rate

(k

g/h

r)

0

100

200

300

400

500

600

700

800

900

1000

Wate

r fl

ow

rate

(k

g/h

r)

169

169

Figure 5-20 and Figure 5-21 compare the required heating power consumption of the

selected ILs, Morphysorb, and MDEA gas plants. N4111-CH3SO4 requires the least

energy, 0.41 MW followed by N-bupy-CH3SO4, 0.49 MW. bmim-NO3 requires the

highest energy input due to its higher operating temperature. The Morphysorb gas plant

requires 2.6 MW which is higher than all the IL gas plants except for bmim-NO3. As can

be seen from Figure 5-21 the energy required for MDEA gas plant is 72.63 MW which is

about 100 times higher than that of the IL gas plants. In a remote bulk separation facility

like Kwoen, a change in energy consumption of nearly 1/5th

from the current would be

significant.

Figure 5-20. Required heating energy for different ILs and Morphysorb gas plants

at fixed H2S mole percent in the upgraded gas (5.33%)

0

1

2

3

4

5

6

Pow

er C

on

sum

pti

on

MW

170

170

Figure 5-21. Required heating energy for different ILs Morphysorb and amine

(45% wt MDEA) gas plants at fixed H2S mole percent in the upgraded gas (5.33%)

Figure 5-22 compares the energy required for pumping, which is similar for all the IL and

Morphysorb plants. Among the ILs, omim-NO3 requires the least energy, about 2 MW,

while Morphysorb requires 2.2 MW. The energy required for pumping in the amine case

is the lowest, 1.08 MW, due to the lowest solvent flow rate. The lowest solvent flow rate

and pumping energy of the MDEA plant is due to the high MDEA loadings. This high

loading of MDEA would possibly cause corrosion which is to be avoided in gas plants. In

lower loadings of MDEA, the flow rate of the solvent is increased and the pumping duty

is higher.

Figure 5-23, shows the energy required for compression for IL, Morphysorb, and MDEA

plants. The MDEA gas plant does not require gas compression. The maximum energy is

0

10

20

30

40

50

60

70

80

Pow

er C

on

sum

pti

on

MW

171

171

required for the Morphysorb gas plant, 6.7 MW approximately 19 times more than for the

IL gas plants. Among IL gas plant bmim-NO3 followed by pmim-L require less

compression, 0.34 MW.

Figure 5-22. Power consumption of pumping for different ILs, Morphysorb and


(5.33%)

Figure 5-24 shows the total cooling energy required for the studied gas plants. Among the

ILs, pmim-L required the lowest cooling energy, 0.66 MW. Morphysorb requires about

12 times more than the pmim-L gas plant at 7.81 MW. The MDEA gas plant required the

most cooling energy, approximately. 89 times more than the IL gas plant, 58.42 MW.

0

0.5

1

1.5

2

2.5

3

3.5

4

Pow

er C

on

sum

pti

on

MW

172

172

Figure 5-23. Compression power for different ILs, Morphysorb and amine (45% wt


Figure 5-24. Cooling required for (Air coolers and other cooling units) for different



0

1

2

3

4

5

6

7

8

Pow

er C

on

sum

pti

on

MW

0

10

20

30

40

50

60

70

Cooli

ng p

ow

er

MW

173

173

The results of using the various solvents in the gas plant simulation would indicate the

following:

Among the IL gas plants, pmim-L required the lowest flow rate and requires less

energy for operation.

Assuming no degradation and entrainment, the IL gas plants require negligible

makeup solvent. Whereas, makeup solvent for Morphysorb (84 kg/hr) and for the

MDEA gas plants (4 kg/hr and water 13 m3/hr) would result in additional

operating costs due to make up.

The solvent flow rate of pmim-L is 4% more than Morphysorb and 37% more

than MDEA.

The pmim-L gas plant requires significantly less energy for operation. The total

energy consumption including heating, pumping and compression for

Morphysorb gas plant is 3.8 times and for the MDEA is 23.4 times more than the

pmim-L gas plant.

The total cooling requirement for Morphysorb gas plant is 13 times and MDEA

89 times more than the pmim-L gas plant.

The pmim-L gas plant reduces the water content of gas, whereas MDEA increase

the water content of gas. In this case study the water content of lean gas form

MDEA gas plant was about 42 times more than the pmim-L gas plant.

All of the above conclusions suggest that pmim-L is a good candidate for further

studies. In the next section, the dehydration effect of the pmim-L gas plant is

investigated in more detail.

174

174

5.5 Simultaneous Dehydration and Sweetening using ILs

Water in a gas stream can form a hydrate with the methane and plug the pipe [162]. In

Canada, the gas quality specification of natural gas in transportation pipe lines requires a

maximum water content of 4lb per MMscf (64 mg/m3) [163]. As was previously shown,

relative to the ILs, Morphysorb has lower affinity for water. On the other hand the

MDEA gas plant increased the water content of the treated gas, because the process

saturates the treated gas with water. One of the potential advantages of the IL-gas plants

is that ILs not only remove the H2S and CO2 from the gas stream but also they decrease

the water content of the gas. In this section, this idea is tested by designing a novel gas

plant. The natural gas composition of Table 5-4 is used. H2S is removed to the level

shown in Table 5-4 and the water content specification of 4lb per MMscf is to be met in

this novel gas treating process.

The proposed IL based gas-sweetening-dehydration plant is similar to the IL-gas

sweetening gas plant discussed in the previous section. The feed gas enters the bottom of

a tray contactor and the lean IL is fed to the top of the contactor. In order to remove the

absorbed water from the rich IL the temperature of IL is increase by a few degrees (from

about 14 C to 17 C) before the regeneration flash tank. The pressure and temperature of

the IL are adjusted prior to feeding the rich IL to the contactor. The H2S composition of

the lean gas is controlled by manipulating the IL flow rate to the contactor. The water

content of the lean gas is controlled by the temperature of the IL into the final flash tank.

175

Figure 5-25. Simultaneous gas sweetening and dehydration plant using IL as the absorbent. The simulation is done using

VMGSim [138]. T-1: inlet separator; T-2, T-3, T-4, T-5: Flash tanks; C-1: Cooler; H-1, H-2: Heaters, P-1: Pump; CP-1, CP-2,

CP-3: Compressors; AC-1, AC-2, AC-3 air-coolers.

176

176

The Process Schematic of the IL-sweetening-dehydration process is show in Figure 5-25.

5.5.1 Gas Sweetening and Dehydration using MDEA and TEG

In this section, a gas dehydration unit was added to the MDEA plant simulation in order

to meet the water content specification of the lean gas. The feed gas condition and

composition of this integrated gas plant are assumed to be the same as the IL-gas

sweetening-dehydration plant of the previous section. The same H2S and water content

specifications are used for this plant simulation. For dehydration triethyleneglycol (TEG)

is the solvent of choice. The structure of TEG is shown in Figure 5-26.

Figure 5-26. Triethyleneglycol (TEG)

For this simulation MDEA was used for gas sweetening and TEG was used for gas

dehydration. The simulation was done using Unisim Design software [164]. The Process

Schematic of the gas treating/dehydration plant can be shown in Figure 5-27.

177

177

The TEG used for dehydration contains some water impurity. The temperature of the

inlet wet gas does influence the water capacity of TEG. The temperature of the wet sweet

gas from the amine plant is cooled. This is necessary so as to increase the absorption

capacity of the TEG, which reduces the required circulation flow rate of the TEG. A

higher absorption capacity also allows operating at a lower TEG purity which reduces the

reboiler load of the TEG regenerator. For this reason the wet lean gas from the MDEA

sweetening plant is cooled and then fed to the TEG contactor. The dry lean gas leaves the

top of the contactor, while the rich TEG leaves the bottom of the contactor. The rich TEG

is pre-heated in the feed/bottoms heat exchanger and is fed to the TEG regenerator. The

lean TEG is cooled and pressured and fed again to the contactor

178

Figure 5-27. Gas sweetening-dehydration plant using MDEA as the sweetening and TEG is used for dehydration. The

simulation is done using Unisim Design [164]. T-1: inlet separator; T-2: Flash tank; HX-1, HX-2: Heater exchangers, P-1, P-2:

Pumps; C-1, C-2, C-3: Coolers

179

179

Figure 5-28 to Figure 5-34 compare the performance indicators of MDEA-TEG gas plant

versus different ILs as the solvent in the sweetening-dehydration process. Figure 5-30

shows the CO2 mass flow rate in the lean gas. N-bupy-CH3SO4 generates the lowest CO2

mass flow rate in the lean gas followed by pmim-L.

The operating costs of gas plants are an important variable that can be used to evaluate

different available solvent options for gas treatment. Theses operating costs would

include energy consumption and solvent makeup. In the case studies of this thesis, it was

assumed that there is no IL loss due to their negligible vapor pressure. The solvent flow

rate is another factor that influences the cost of the plant. Figure 5-33 and Figure 5-34

compare the energy requirement of the gas plant for heating, compression of the gas,

pumping the liquid and cooling the fluids. Also, the IL flow rates are shown in Figure

5-33 for different ILs. Omim-NO3 requires the lowest heat energy, 0.57 MW followed by

N-bupy-CH3SO4, 0.79 MW and pmim-L, 0.85 MW. On the other hand hmim-L has

relatively higher melting point than the other studied ILs and has to operate at higher

temperature; therefore, it requires the highest heating energy among the six ILs. The

cooling requirements of the ILs are shown in Figure 5-33 where, pmim-L requires the

minimum cooling, 0.73 MW followed by omim-NO3, 0.81 MW. hmim-L requires the

highest cooling of 3.7 MW. The compressions requirements of different gas plants are

compared in Figure 5-33. Pmim-L requires the lowest compression power, 0.36 MW and

hmim-L requires the highest compression of 1.42 MW. Also, Figure 5-33 shows that the

circulation flow rate of pmim-L is the lowest.

180

180

Figure 5-28. H2S mass flow rate in upgraded gas IL-sweetening-dehydration gas

plants and MDEA-TEG-sweetening-dehydration gas plant at fixed mole percent of

H2S (5.33%) and water content (4lb/MMscf) in the dry upgraded gas.

Figure 5-29. CO2 mole percent in upgraded gas IL-sweetening-dehydration gas



23.5

23.7

23.9

24.1

24.3

24.5

24.7

24.9

H2S

(Ton

/hr)

0

2

4

6

8

10

12

14

CO

2 (

mole

%)

181

181

Figure 5-30. CO2 mass flow rate in upgraded gas IL-sweetening-dehydration gas



Figure 5-31. CH4 mole percent in upgraded gas IL-sweetening-dehydration gas



40

41

42

43

44

45

46

47

48

49

CO

2(T

on

/hr)

84

85

86

87

88

89

90

CH

4 m

ole

%

182

182

Figure 5-32. CH4 mass flow rate in upgraded gas IL-sweetening-dehydration gas



680 m3/hr, followed by hmim-L 688 m

3/hr and omim-NO3706 m

3/hr. N4111-CH3SO4 and

N-bupy-CH3SO4 ,at 789 m3/hr, require the highest flow rate among the six studied ILs.

From Figure 5-28 to Figure 5-33 it can be concluded that pmim-L followed by omim-

NO3 not only absorb more H2S and CO2 and less CH4, but also the require the lowest

heating, cooling and compression. Also they require the lowest solvent circulation flow

rate. This study suggests that pmim-L and omim-NO3 are good candidates for

simultaneous dehydration and sweetening of the natural gas. The cost of using these ILs

for gas sweetening can be compared to conventional solvents such as amines or other

physical solvents with an additional dehydration plant.

184

185

186

187

CH

4 m

ass

flo

w r

tae

(ton

s/h

r)

183

183


gas IL-sweetening-dehydration gas plants at fixed mole percent of H2S (5.33%) and

water content (4lb/MMscf) in the dry upgraded gas; : Heating requirement; :

Cooling requirement; ■: power required for compressors; ◊: Pumping power

requirement; ▲: IL standard volume flow rate

The results of this section show that the IL gas plant with a simple heater can

theoretically control the water content of the upgraded gas. This modification, adds only

0.069 MW heating and 0.071 MW cooling in order to adjust the water content of lean

gas.

620

640

660

680

700

720

740

760

780

800

0

0.5

1

1.5

2

2.5

3

3.5

4

Flo

w r

ate

m3/h

r

Pow

er C

on

sum

pti

on

MW

att

184

184


gas IL-sweetening-dehydration gas plants and MDEA-TEG-sweetening-dehydration

gas plant at fixed mole percent of H2S (5.33%) and water content (4lb/MMscf) in the

dry upgraded gas; : Heating requirement; : Cooling requirement; ■: power

required for compressors; ◊: Pumping power requirement

In order to lower the water content of the lean gas from the MDEA gas plant a TEG

dehydration plant must be integrated with the MDEA unit, which then requires an

additional absorber, regenerator, heat exchanger, cooler and pump. Also, the integrated

gas plant requires additional 3.14 MW for heating and additional 14.71 MW for cooling.

By comparing the total energy requirement of pmim-L and MDEA-TEG, it can be seen

that the MDEA-TEG requires 23.8 times more heating and pumping energy than pmim-L

gas plant. Furthermore, the MDEA-TEG gas plant requires 86.2 times more cooling than

the pmim-L gas plant.

0

10

20

30

40

50

60

70

80

Pow

er C

on

sum

pti

on

MW

att

185

185

5.6 Shale Gas Case Study

The H2S content of the lean gas of the gas plants studied in sections 5.4 and 5.5 was 5%.

Negligible solvent make-up, simultaneous dehydration, less cooling and less heating are

the advantages the IL-gas plants when compared with the Morphysorb and MDEA gas

plants. The focus of this section is to investigate the potential of IL-gas plants to reduce

the H2S content of lean gas to lower levels and possibly down to sales gas specifications.

Since gas absorption in ILs is assumed to occur physically, the following issues have to

be taken into account to increase the H2S absorption:

1. Choosing ILs with high selectivity toward H2S to mitigate hydrocarbon

absorption and lower solvent circulation rates. Based on the studies in sections

5.4 and 5.5, pmim-L gas plant operates with the lowest energy and circulation

rate. For this reason pmim-L is chosen for this study.

2. Higher solvent circulation rates are expected to reduce the H2S content of the

lean gas.

3. To increase the absorption driving force between the IL and the gas in the

absorber, IL has under go deep regeneration. Since there is an inverse relation

between the gas solubility in the ILs and temperature, a stripper must be used to

regenerate IL.

Typical shale gas composition was chosen for this study and consists mostly of methane

with 10% CO2 and 500 ppmv H2S, Table 5-8.

186

186

Table 5-8. Typical shale gas property

Field gas property value

Gas flow, MMscf 300

P, psia 1085

CO2 % 10

H2S, ppmv 500

CH4 % 86

C2H6 % 4

H2O % 0

5.6.1 Description of the IL Gas Sweetening Plant

Figure 5-35 shows the process schematic of the IL-gas processing with deep

regeneration. The simulation of this gas processing plant was done using VMGSim[138].

5.6.1.1 Inlet Separator and Absorber

As described in section 5.4, the inlet separator, T-1 is used to remove entrained liquid,

solids, and heavy condensable hydrocarbons from the input field gas. The lean IL is fed

to the top of the absorber and the sour gas is fed to the bottom of the absorber. The

pressure of the absorber varies between about 74 and 75 bar.

187

Figure 5-35. IL gas plant for Shale gas sweetening. The simulation is performed using VMGSim[138]. T-1: inlet separator; T-2,

T-3, T-3, T-4, T-5, T-6: Flash tank; HX-1, HX-2: Heater exchanger, H-1: Heater, P-1: Pump; CP-1, CP-2, CP-3: Compressor;

AC-1, AC-2, AC-3, AC-4: Cooler.

188

188

5.6.1.2 Recovering the Absorbed Hydrocarbons

Similar to the IL-gas processing plant of section 5.4, three flash tank are used, T-1, T-2

and T-3, to recover any absorbed CH4 and C2H6. The pressure and temperature of the

recovered recycle gas are adjusted using compressors, CP1, CP2, CP3 and air-coolers,

AC1, AC2 and AC2.

5.6.1.3 IL Regeneration

The rich IL enters the flash tank T-5 and the bulk of acid gas is removed from the rich IL.

The H2S mole fraction of IL from T-5 is about 710-4

. A regenerator column is used to

further regenerate the IL. The rich IL from flash tank T-5 is pre-heated by the heat

exchangers HX1 and HX2 and then enters the top of the stripper. The acid gas from flash

tank T-5 is partially recycled to the regenerator as a stripping gas to strip the rich IL. The

mole fraction of IL decreases by about one order of magnitude to about 710-5

. The hot

lean IL from the regenerator is pre-cooled in heat exchanger HX2 and its pressure and

temperature is adjusted using the pump P1 and cooler C1. The hot acid gas from the top

of the stripper is pre-cooled in heat exchanger HX1 and then cooled in the air-cooler AC-

4. The cooled acid gas is mixed with the acid gas coming from the flash tank T-5 to

produce the acid gas stream leaving the gas plant. The operating temperature range could

extend up to the point where ILs thermally decompose, hence a study would be required.

This impacts the maximum allowable temperature in the reboiler and heat exchangers.

The required IL makeup can be influenced by the decomposition rate at high

temperatures.

189

189

The flash tank T-6 is used to recover any vaporized and/or entrained IL. This flash tank

can be removed if the amount of the entrained ILs is not significant. The gas plant was

simulated for different H2S content of the lean gas. The performance of the IL-gas plant,

Figure 5-37, is also compared with the MDEA-TEG plant, Figure 5-27 using the same

feed, Table 5-8.

Figure 5-36, shows the solvent circulation rate versus H2S content of the lean gas. As

expected higher flow rates are required to decrease the H2S content. The pmim-L

circulation flow rate is higher than the equivalent MDEA plant requirement. As can be

seen the pmim-L circulation rate increases linearly until the point that the lean IL and

upgraded gas composition are approaching equilibrium. From this point on the required

IL flow rate increases exponentially.

Figure 5-36. Solvent flow rate at different H2S content of upgraded gas for shale gas

sweetening plants; : pmim-L; :MDEA; ○: TEG

1

10

100

1000

10000

0.0 100.0 200.0 300.0 400.0 500.0

m3/h

r

H2S, ppmv

190

190

Figure 5-37 and Figure 5-38 show the energy consumption and cooling requirements of

pmim-L and MDEA-TEG gas plants. It can be seen that there is correlation between the

energy consumption and the solvent flow rate, Figure 5-36. Contrary to the case studies

of sections 5.4 and 5.5, the MDEA-TEG gas plant requires less energy to decrease the

H2S composition in the lean gas.


gas sweetening plant using pmim-L; ◊: compression; : pumping; : cooling; ○:

reboiler

Figure 5-39 shows the CO2 and CH4 flow rates in the acid gas for the pmim-L gas plant.

The CO2 flow rate is approximately 30 times more than the flow rate of CH4. The CO2

and CH4 absorption increases rapidly for lower H2S targets in the lean gas due to higher

IL flow rates.

0.01

0.1

1

10

100

1000

0.0 100.0 200.0 300.0 400.0 500.0

Pow

er C

on

sum

pti

on

MW

H2S, ppmv

191

191


gas sweetening plant using MDEA-TEG; ■: pumping; ▲: cooling; : heating

Figure 5-41 shows the CO2 and CH4 flow rates in the lean gas for the MDEA-TEG gas

plant. As H2S composition decreases, no significant change in the CO2 and CH4 flow

rates is observed. MDEA is a tertiary amine which does not have hydrogen attached to

the nitrogen to react directly with CO2 to form a carbamate [5, 6]. For this reason CO2

must first react with water to form bicarbonate. Then, the bicarbonate and MDEA go

through an acid-base reaction. The rate of formation of CO2 is slow and controls the

0.01

0.1

1

10

100

0.0 100.0 200.0 300.0 400.0 500.0

Pow

er C

on

sum

pti

on

MW

H2S, ppmv

192

192

Figure 5-39. Mass flow rate of CO2 and CH4 in acid gas at different H2S content of

upgraded gas for shale gas sweetening plant using pmim-L; : CO2; ○: CH4

overall reaction rate of CO2 with MDEA. Figure 5-40 illustrates the indirect reaction of

CO2 with MDEA.

Figure 5-40, Indirect reaction of a tertiary amine with CO2; reaction (a) is slow and

is the controlling reaction

At H2S compositions lower than 100 ppm, the flow rates of CO2 and CH4 decline slightly

due to physical absorption.

0

0.5

1

1.5

2

2.5

0

10

20

30

40

50

60

70

0.0 100.0 200.0 300.0 400.0 500.0

CH

4, T

on

/hr

CO

2,T

on

/hr

H2S, ppmv

193

193

Figure 5-41. Mass flow rate of CO2 and CH4 in upgraded gas at different H2S

content of upgraded gas for shale gas sweetening plant using pmim-L; : CO2 in

pmim-L gas plant; ○: CH4 in pmim-L gas plant; : CO2 in MDEA-TEG gas plant;

○: CH4 in MDEA-TEG gas plant;

Figure 5-42, shows the CO2 and CH4 mole percent in lean gas for both pmim-L and

MDEA-TEG gas plants at different H2S compositions in the lean gas. For MDEA-TEG

gas plant the CO2 and CH4 mole percents are almost the same, at about 9% and 82%

respectively. By dropping the H2S content of the upgraded gas in the pmim-L gas plant,

the CO2 mole percent declines from the 9% to 0.0%. Due to the CO2 and H2S absorption

by pmim-L, the CH4 of the treated gas increases from 86% to 96%.

203.5

204

204.5

205

205.5

206

206.5

0

10

20

30

40

50

60

70

0.0 100.0 200.0 300.0 400.0 500.0

CH

4, T

on

/hr

CO

2,T

on

/hr

H2S, ppmv

194

194

Figure 5-42. Mole percent of CO2 and CH4 in upgraded gas at different H2S content


gas plant; ○: CH4 in pmim-L gas plant; : CO2 in MDEA-TEG gas plant; ○: CH4 in

MDEA-TEG gas plant;

In general the MDEA-TEG plant removes the H2S more effectively for the case study in

this section. IL’s H2S absorption is physically and the partial pressure of H2S is the

driving force for this absorption. Therefore, the driving force is not high enough at lower

the H2S content and, significantly higher IL flow rates are needed to reduce the H2S level.

Although pmim-L has a large affinity toward H2S, at very low compositions of H2S, the

driving force for hydrocarbons absorption increases which leads to absorption of more

hydrocarbons. This reduces the overall efficiency of the pmim-L gas plant.

Comparing the case studies discussed in sections 5.4, 5.5 and the Shale gas case study,

the energy requirement for the given feed and product compositions, the IL gas plant is

84

86

88

90

92

94

96

98

0

2

4

6

8

10

12

0.0 100.0 200.0 300.0 400.0 500.0

CH

4, m

ole

%

CO

2, m

ole

%

H2S, ppmv

195

195

better than the MDEA gas plant. However at certain compositions such as the Shale gas

case study MDEA is more efficient than pmim-L gas plant. In order to further investigate

this phenomenon, it is useful to analysis the capacity of H2S absorption in pmim-L and

MDEA. This would provide guidelines to assess the operating conditions and to choose

the right sweetening method. Figure 5-43 compares the H2S loading of pmim-L and

MDEA (45 wt%) over H2S vapor pressures ( 0 to 1400 kPa). The MDEA graph was

generated using the VMGSim MDEA property package and the pmim-L graph was

generated using the procedure discussed in section 5.2. Figure 5-43 shows that there are

three operating zones for pmim-L gas plant relative to the MDEA gas plant.

Zone I:

Zone I is located at the H2S vapor pressures between 0 to 200 kPa. In this zone H2S

vapor pressure versus H2S loading in MDEA has the minimum slope. This indicates that

H2S is efficiently removed by MDEA because; the difference in loading per decrease in

H2S vapor pressure is maximum. However, the slope of the H2S vapor pressure versus the

loading of pmim-L is maximum or, the difference in the H2S loading per decease of H2S

vapor pressure is minimum. For example, in order to decrease the H2S vapor pressure

from 200 kPa to 0 kPa, the difference in the loading of H2S in MDEA is about 0.9.

However the difference in the loading of the pmim-L is 0.1. This means more equilibrium

stages are required for the IL to read the desired H2S composition.

196

196

Figure 5-43. H2S loading in the solvent at different H2S partial pressure;○: amine

(45% wt MDEA); : pmim-L; ■:Equivqlent to H2S feed composition presented in

Table 5-4 , (13.6 mole % H2S); ▲: Equivqlent to H2S composition in the treated gas

presented in Table 5-4 , (5.3 mole % H2S); : Equivalent to feed composition in

Shale gas case study, (500 ppm H2S); : Equivalent to treated gas composition in

Shale gas case study, (100 ppm H2S);

Another feature of this zone is that the loading of H2S in MDEA is higher than in pmim-

L, which results in a lower molar based MDEA than pmim-L to decrease to a certain

0

200

400

600

800

1000

1200

1400

0 0.2 0.4 0.6 0.8 1 1.2 1.4

Part

ial

Pre

ssu

re (

H2S

), k

Pa

Loading H2S/Solvent

Zone I

Zone II

Zone III

197

197

level of H2S vapor pressure. This affects the size of equipments such as absorber

diameter, pumps and heat exchangers.

In addition, in this zone the H2S loading in pmim-L is very low. To compensate for this,

the IL must be regenerated to very low concentration. For this reason, the simple flash

tanks cannot fully regenerate the IL and a stripper is recommended. This makes the gas

sweetening plant more complex and energy demanding.

The efficiency of MDEA in zone I is the highest hence a minimum solvent flow rate is

required. In contrast, pmim-L is least efficient in zone I with maximum solvent flow rate.

The Shale gas plant operates in this zone and this explains the noticeable difference in the

required solvent flow rate and energy for the pmim-L gas plant. Due to this fact the

MDEA showed better results over the pmim-L. Note, this evaluation is based on only

H2S absorption; however, if CO2 absorption is desired too, a chemical solvent with more

CO2 absorption capacity must should be used.

Zone II:

Zone II, the H2S vapor pressures is between 200 to 1260 kPa. In this zone, the slope of

H2S vapor pressure versus the H2S loading in MDEA increases rapidly which indicates

that the reaction between H2S and MDEA approaches equilibrium and that the H2S is

absorbed physically by MDEA solution. On the other hand, the slope of the H2S vapor

pressure versus H2S loading in pmim-L is lower than that of MDEA. For example in

order to reduce the vapor pressure of H2S from 1260 kPa to 200 kPa the loading

difference for pmim-L and MDEA are about 1 and 0.2, respectively.

198

198

In this zone as in zone I, the H2S loading in MDEA is higher than pmim-L, which results

in a higher pmim-L flow rate than the MDEA flow rate.

The higher loading difference in pmim-L per change in H2S vapor pressure and also the

higher loading in pmim-L in zone II results in two advantages. First, the extra

regeneration using the stripper might not be necessary which simplifies the gas plant.

This reduces the capital and operating cost. The second advantage of operating pmim-L

in zone II is that larger difference between the loading at feed and product H2S vapor

pressure allows the IL to be regenerated by only pressure reduction in the flash tanks. The

combination of these factors and the fact that the IL gas plants dehydrate the gas, makes

the pmim-L gas plant a competing option over the MDEA gas plant in this zone. For

example in the case study of section 5.5, although the total solvent flow rate of the pmim-

L gas plant is higher than MDEA-TEG gas plant (680 m3/hr versus 397.5 m

3/hr MDEA

and 25 m3/hr TEG), other factors such as simplicity of pmim-L gas plant, 24 times less

total energy consumption, lower energy consumption per solvent flow rate and negligible

solvent makeup make the pmim-L gas plant a contender.

Zone III:

Zone III: The H2S vapor pressure higher than 1200 kPa

The loading of H2S is higher in pmim-L is than in MDEA resulting is lower molar base

flow rate for pmim-L. Also, the loading difference of H2S in solvent per reduction of is

H2S vapor pressure is higher in pmim-L than MDEA. For example in order to reduce the

vapor pressure form 1400 kPa to 1200 kPa, the H2S loading difference in pmim-L and

199

199

MDEA is about 0.3 and 0.02 respectively. This provides more flexibility in the

absorption and regeneration. For example the regeneration can be done with lower

pressure differences in the flash tank.

It can be concluded that the choice of pmim-L or MDEA-TEG gas plants depends on the

feed and desired product compositions. The solvent flow rate and the total energy

requirement and the guidelines provided here will aide in selecting the appropriate

solvent for a specific application.

200

200

5.7 Summary

This chapter utilizes the thermodynamic model developed in Chapter Four:, and

compares the ILs for gas sweetening applications, both for bulk and sale gas

specifications. The candidate ILs are selected based on the ranking provided in Chapter

Four:. The ranking was based on the capacity and selectivity for the absorption of H2S

and CO2 versus CH4 and C2H6. Due to environmental concerns, ILs containing, fluoride

and cyanide groups were removed from the ranking. The Eight chosen ILs were: bmim-

NO3, bmim-CH3SO4, hmim-L, MeButPyr-CH3SO4, N4111-CH3SO4, N-bupy-CH3SO4,

omim-NO3 and pmim-L.

VMGSim was used to simulate the gas plants. Because, there is no pre-built IL available

in the database of the commercial simulators, ILs must be defined for the simulator. The

basic properties required for defining the hypothetical compounds were molecular

weight, densities at 298.15 K, critical properties and acentric factor. Only the densities

for three ILs are experimentally available at this time. Therefore, the unknown densities,

critical properties and acentric factor were predicted using a group contribution method.

The Advanced Peng-Robinson equation of state (APR) was selected for phase behavior

calculations for the mixture of IL, CH4, C2H6, CO2, H2S and water. The binary interaction

parameters for binary mixtures of IL-solute were calculated based on the available VLE

data or estimated VLE data using the models developed in Chapter 4. However, for the

IL-water mixtures with no measured data, the COSMO-RS method with no additional

adjustment proved to be satisfactory and used to generate the VLE data.

201

201

Using the tuned equation-of-state, several gas plant simulations were completed. The

performance of the IL-gas plants were compared to an industrial physical solvent

(Morphysorb) and a conventional chemical solvent (MDEA). The key comparison

parameters were the overall energy consumption (e.g. heating, cooling and pumping); the

solvent flow rate, which determines the equipment sizing, and makeup solvent and/or

water, which address some of the operating costs and environmental concerns. In order to

have a reasonably valid comparison, the same feed composition, and the same H2S

specification for lean gas were used in the case study simulations.

To add reality to the case studies the similar feed and product compositions of the Kwoen

gas plant which is a bulk removal process and operates using Morphysorb, were used in

the gas plant simulations. The feed pressure was 80 bar, about 13% H2S and the produced

gas was specified for 5% H2S. All eight ILs, Morphysorb and MDEA used the

compositions. Table 5-9 summarizes the key performance variables for the Kwoen case

studies. Among the tested ILs pmim-L performed better than the other ILs.

202

202

Table 5-9. Performance Summary of pmim-L, Morphysorb and MDEA gas

sweetening plants based on feed and product specifications of Kwoen case study

Relative to pmim-L Makeup,

Relative

Volumetric

Solvent Flow,

Relative Water

content,

upgraded gas

Relative Heating,

pumping and

compression power

Relative

Cooling

power

Solvent

kg/hr

Water

m3/hr

Pmim-L 1 1 1 1 0.0 0

MDEA 0.63 42 23.4 89 4.82 13.07

Morphysorb 0.96 2.3 3.8 13 84 0

The advantages of the IL-gas sweetening plants resulting from this case study are:

No solvent/water makeup because the studies assumed no degradation or

entrainment carry over, while Morphysorb required 84 kg/hr make up and

MDEA required 4.82 kg/hr MDEA and 13.07 m3/hr water makeup.

Lower energy consumption than MDEA and Morphysorb. MDEA consumes 23.4

times more combined heating/pumping energy and 89 times more cooling energy.

While Morphysorb required 3.8 and 13 times more heating/pumping and cooling

energy, respectively.

Mitigating the water content of gas, whereas MDEA increases the water content

of gas. The water content of upgraded gas from MDEA gas plant has 42 times

more water than pmim-L gas plant

203

203

Based on the dehydration potential of ILs, a simple modification, a heat exchanger before

the last flash tank was included in the plant simulation. Only few degrees increase in IL

regeneration temperature removes enough water from the solvent to lower the water

content of the lean gas. However, a TEG gas dehydration plant must be integrated with

the MDEA gas sweetening plant in order to meet the water content specifications of the

lean gas. Other than the complexity of the MDEA-TEG gas plant relative to modified

pmim-gas plant, the TEG dehydration unit required significant additional energy to meet

lean gas water specification. The modified pmim-L gas plant required 0.069 MW

additional heating and 0.071 additional cooling to meet the water specifications.

However, the MDEA-TEG gas plant required an additional 3.14 MW heating and 14.71

MW cooling. Table 5-10 shows a comparison of the key performance variables for the

modified pmim-L and the MDEA-TEG gas sweetening-dehydration plants.

Table 5-10. Performance Summary of pmim-L and MDEA-TEG gas sweetening-

dehydration plants at fixed mole percent of H2S (5.33%) and water content

(4lb/MMscf) in the dry upgraded gas

Relative to pmim-L Makeup,

Kg/hr

Relative

Volumetric

Solvent Flow,

Relative Heating,

pumping and

compression power

Relative

Cooling

power

Solvent

kg/hr

Water

m3/hr

Pmim-L 1 1 1 0 0

MDEA-TEG 0.62 23.8 86.2 MDEA: 4.9

TEG: 9.8

13.28

204

204

The last case study of this chapter was the composition of a typical sour shale gas. The

feed consisted mostly of CH4 and CO2 and 500 ppm of H2S. Because the physical solvent

must deal with low H2S composition, a stripper was used to regenerate the solvent to very

low levels of H2S in the lean solvent. Part of the acid gas was used as the stripping gas

was used to regenerate the IL to lower levels of H2S. This case study showed that the

pmim-L’s solvent flow rate and energy consumption are significantly higher than that of

the MDEA-TEG gas plant.

To explain the performance of pmim-L at different conditions, the H2S loading of pmim-

L and MDEA-TEG at different partial pressure of H2S were shown to be a useful

guideline. Three different operating conditions were defined. The result of this analysis

shows that pmim-L should only be used as the solvent when the operating condition is in

zone III or II. In these zones the H2S vapor pressure of feed is high or moderately high,

i.e. more than 200 kPa. For low H2S vapor pressures such as shale gas, the

MDEA/chemical solvent performance is superior to pmim-L.

205

205

Chapter Six: Summary

6.1 Introduction

In this thesis the application of ILs for gas sweetening and gas dehydration was

considered. It began with the Quantum level investigation and optimization of cations

and anions geometry of the individual ILs. Then solubility models were developed for

estimating the solubility of CO2, H2S, CH4 and C2H6 in a wide range of ILs including

where many ILs which have not yet been synthesized. The top ranked ILs were used for

simulation of several IL- gas processing plants. These simulation results were compared

to conventional chemical solvent as well as physical solvent gas plants. The advantages

and disadvantages of the IL-gas plants were discussed.

6.2 Geometry Optimization of ILs and Activity Coefficient Calculations

The COSMO-RS Quantum calculation model was used in this study to predict the activity

coefficient of components in the mixture without the aid of experimental information.

This model takes into account the solute and solvent molecular characteristics from first

principles. The solvent in the COSMO model is considered to be a perfect conductor.

Energy of the system is minimized by optimizing the charge distribution of the molecular

shaped cavity of the molecule. COSMO/DFT calculations were performed by utilizing the

BP [108-110] functional with and a triple- valence polarized basis set (TZVP)[106].

The DFT/COSMO calculations were used to optimize the geometry of ILs’ cations and

anions. The activity coefficients of the solute components were determined based on the

calculated chemical potential of the system.

206

206

6.3 Solubility Models

The next step involved in constructing general models for predicting the solubility of

different solutes in the ILs with no measured experimental data. These models consist of

three sources of information, (1) COSMO-RS calculations, (2) an equation of state, and

(3) an easily calculated physical property. In order to adjust the empirical parameters, a

database for the available existing experimental data for the solubility of CO2, H2S, CH4

and C2H6 was built.

Different models were introduced to describe the solubility of solutes in the ILs,

assuming no chemical reactions between the solutes and ILs and the solute is only

physically dissolved. The model parameters were calculated based on the available

experimental data. The models were evaluated based on their ability to predict the

solubility of binary mixtures of CO2, H2S, CH4 and C2H6 in ILs in mixtures that were not

included in the regression experimental set. The recommended model was based on the

Peng-Robinson equation of state for gas fugacity coefficient calculations, asymmetric

activity coefficient calculated from COSMO-RS method, and an empirical Henry’s

constant correlation that was a function of the temperature and pressure of the system and

a physical property of the ILs. The results of this modeling showed that the molecular

surface area of ILs can better describe the dispersion interaction based on the solubility of

the more polarizable solutes, H2S and C2H6. The solubility of the polarizable molecules,

CO2 and CH4, were better described by the MW of ILs.

207

207

The models suggest that, in general, larger ILs with higher molecular surface area or MW

have higher solubility capacity. The following factors affect the selectivity of solute

absorption in ILs: (1) the solubility of each solute changes depending on the size of IL;

(2) the model showed that the solubilities of CO2 and H2S in ILs are more pressure

dependent than the solubilities of CH4 and C2H6; therefore, at higher pressures more CO2

and H2S are absorbed in the ILs (3) the activity coefficient of the solute in the ILs

depends on the interactions between the solute and IL and affects the solubility of the

solute.

Using these models, the ILs were screened based on their selectivity and capacity of

absorption of H2S and CO2 versus CH4 and C2H6. 425 ILs were screened at 298.15 K and

2000 kPa. ILs containing anions doc and FEP show the highest average CO2, H2S, CH4

and C2H6 absorption capacity, but it was noted that high pure component capacity did not

lead to good candidates for separation. The selectivities for the absorption of CO2 and

H2S over CH4 and C2H6 in ILs at 298.15 K and 2000 kPa were also calculated. The ILs

containing the anions BF4, NO3 and CH3SO4 and containing cations N4111, pmg and tmg

showed the largest number of combinations that met the criteria for separation of H2S and

CO2 from natural gas at 298.15 K and 2000 kPa. Due to environmental concerns, the ILs

containing, fluoride and cyanide groups were removed from the consideration. The eight

selected ILs were: bmim-NO3, bmim-CH3SO4, hmim-L, MeButPyr-CH3SO4, N4111-

CH3SO4, N-bupy-CH3SO4, omim-NO3 and pmim-L.

208

208

6.4 Conceptual IL-Gas Plant Simulation

A commercial simulator, VMGSim, was used to simulate the gas treatment plants. had to

be input as new components into the simulators because there was no pre-built IL

available in the component database. The basic properties required for defining the

hypothetical compounds were molecular weight, densities at 298.15 K, critical properties

and acentric factor. Only the density data for three ILs are experimentally available at this

time. Therefore, the unknown densities, critical properties and acentric factor are

predicted using a group contribution method.

The Advanced Peng Robinson equation of state (APR EOS) was selected for the phase

behavior calculations for the mixtures of IL, CH4, C2H6, CO2, H2S and water. The binary

interaction parameters of binary mixtures of IL and solute were adjusted using available

VLE data. However, experimental data are available only for few IL-solute mixtures. For

the binary mixtures with no VLE data, models were used to estimate the VLE data. For

IL-water mixtures with no measured data, the COSMO-RS method with no additional

adjustment used to generate the VLE data.

Using the tuned equation of state, several treatment gas plants were simulated. The

performance of the IL-gas plants were compared against an industrial physical solvent

(Morphysorb) and a typical chemical solvent (MDEA). The key parameters were the

energy consumption (e.g. heating, cooling and pumping); the solvent flow rate, which

influences the equipment sizes of the equipments; and, makeup solvent and/or water,

which adds to the operating costs and environmental concerns. In order to compare the

209

209

different gas plants on the same basis, the same feed composition was used for and fixed

H2S composition for lean gas was also applied in all cases.

The first sets of tests were built based on the to feed and product composition of Kwoen

gas plant operated using Morphysorb. The feed pressure was 80 bar and contained about

13% H2S and the lean produced gas contained about 5% H2S. All eight ILs, Morphysorb

and MDEA were tested at on these conditions.

The advantages of the IL-gas sweetening plants for this case study are:

Negligible solvent/water makeup, compared with 84 kg/hr solvent make up for

Morphysorb and 4.82 kg/hr MDEA and 13.07 m3/hr water makeup for MDEA

Lower energy consumption than MDEA and Morphysorb. MDEA consumes 23.4

times more combined heating/pumping energy and 89 times more cooling energy.

Morphysorb requires 3.8 and 13 times more heating/pumping and cooling energy,

respectively.

Lower the water content of the produced gas, MDEA increases the water content

of gas. The water content of lean gas from MDEA gas plant has 42 times more

water than the pmim-L gas plant

Based on the dehydration potential of ILs and, a simple modification, namely an

exchanger prior to the last flash drum combined treatment and dehydration using the ILs.

Only few degrees increase in the IL regeneration temperature removes enough water

from the solvent to meet the water content specification of the lean gas. However, a TEG

210

210

gas dehydration plant must be integrated with the MDEA gas sweetening plant in order to

meet the water content specifications of the lean gas. Other than the additional equipment

of the MDEA-TEG gas plant relative to modified pmim-gas plant, TEG requires

significant additional energy to meet the water content of the lean gas. The modified

pmim-L gas plant requires 0.069 MW additional heating and 0.071 additional cooling to

meet the water specification. However, the MDEA-TEG gas plant required an additional

3.14 MW heating and 14.71 MW cooling.

The final gas plant simulation used the composition of a typical sour shale gas. The feed

consisted mostly of CH4 and CO2 and H2S composition at 500 ppm. Because the physical

solvent must deal with such a low H2S composition, a stripper was used to regenerate the

IL to very low H2S levels. In this case study, the pmim-L’s solvent flow rate and energy

consumption are significantly higher than the MDEA-TEG gas plant.

To explain the performance of pmim-L at different conditions, the H2S loading of pmim-

L and MDEA-TEG at different vapor pressure of H2S was shown to be useful guide line.

Three different operating conditions were developed. It was shown that the pmim-L

should only be considered when the plant operating condition was located in Zone III or

II. In these zones the H2S vapor pressure of feed is high or moderately high (more than

200 kPa). For low H2S vapor pressures such as shale gas, the chemical solvent gas

treating plant must be used.

211

211

There are several advantages for choosing ILs to the right gas processing application such

as, (1) simplicity of the IL-gas plant particularly the IL base sweetening-dehydration gas

plant provides a simpler design, lower maintenance demand, less control equipments and

lower capital cost. (2) compared to MDEA and MDEA-TEG plant reduces the operating

cost by consuming significantly less energy, lower solvent makeup and no water makeup

(3) the IL gas plant from environmental point of view requires less energy which leads to

less CO2 emission; (4) it has been shown that there is significantly lower solvent loss for

IL gas plant when compared to a MDEA or Morphysorb IL- gas plants; (5) the water

usage in IL gas plant is much lower than in the MDEA gas plant since less water is

consumed for cooling, steam generation and there is water loss in MDEA gas plant.

212

212

Chapter Seven: Recommended future work

7.1 Introduction

In this thesis it has been shown that ILs are useful solvents for gas sweetening and

dehydration. The ILs with high selectivity for CO2 and H2S under certain conditions were

superior to a MDEA or Morphysorb gas plants both economically and environmentally.

However, there are several areas that deserve careful attention before implementing the

IL gas plants which can be investigated in future studies. This chapter recommends a

number of these research topics.

7.2 Melting Point

Operating the absorber at low temperatures increases the absorption capacity which

results in lower IL flow rate and smaller sized equipment. However, it is important to

make sure the temperature of the IL in the gas plant is higher than its solidification

temperature. Therefore, the melting point of the IL must be known in order to design the

absorber. If the melting point of the ILs is not very low, heaters will normally be

required. The ILs with higher melting points might not be suitable for gas processing.

7.3 Thermal Decomposition

The operating temperature range can be extended up to the point where ILs thermally

decompose. As noted in Chapter 5, a stripper column can be used to regenerate IL to a

very low level. Before implementing these designs, the tolerance of the ILs to the

213

213

temperature should be known. The decomposition temperature of ILs dictates the

maximum temperature at which the reboiler can be operated. This may impact the flow

rate of IL makeup.

7.4 VLE Measurements

The thermodynamic models developed in this study were able to estimate the solubility

of solutes in ILs and predict VLE data. The benefit of this approach is to rank the ILs

based of their absorption capacity and selectivity. However, VLE measurements for the

candidate ILs must be conducted in order to fine tune the gas plant simulations.

7.5 Chemical Reactions

In this study it is assumed that no reaction occurs between the solute and ILs. This

assumption has to be confirmed experimentally. If there is any chemical reaction between

the solutes and ILs, the kinetics of the reactions, the species produced, regenerateability

of the IL and the regeneration need to be determined. ILs’ stability in the solutes and

water should be investigated (chemical decomposition). This may affect the makeup

requirements and the absorption rates.

7.6 Mass Transfer Rate and Tray Efficiencies

The mass transfer rate and the tray efficiencies are important design factors which affect

the required number of trays, solvent flow rate and the equipment sizing. In this thesis, it

is assumed that the tray efficiencies of the absorption columns are ideal. However, the

tray efficiencies should be obtained experimentally.

214

214

7.7 Viscosity of ILs

The viscosity of ILs impacts the pumping duty and also affects the absorber tray

efficiencies. The effect of temperature and solute concentration on the IL viscosity needs

to be determined experimentally.

7.8 Corrosivity

The corrosivity of ILs is an important factor which impacts the maximum allowable

temperature, maximum flow rate of IL and maximum acid gas loading into the IL and the

materials used in the processing equipment. Corrosion does have a large impact on the

capital cost, operating cost and the operability of the gas processing unit.

7.9 Heat Capacity and the Heat of Solvation

Heat capacity and the heat of solvation are important factors in gas processing design. It

affects the heat exchanger design and the temperature profile in the absorber which

indirectly affect the absorption capacity and selectivity, viscosity of ILs, tray efficiency

and rate of corrosion.

7.10 Toxicity

Since some of the ILs are new, little or no toxicity information is available. It is important

to determine the toxicity of these chemicals before any industrial application. This will

affect the solvent handling procedure in the plant. Also the impact of ILs on the

environment (e.g. the ILs’ biodegradability) must be investigated.

215

215

7.11 Cost of ILs

At the moment many of the ILs investigated in this study have either been synthesised in

laboratory or have not yet been synthesised. As a consequence the cost of ILs might be

very high at this time. If an industrial application of ILs is found, the ILs might be

produced on an industrial scale at lower costs.

216

216

References

1. ENCANA. Natural gas end use. 2011; Available from: www.encana.ca.

2. TransCanada, Gas Quality Specifications TransCanada and other pipelines, 2010.

3. UOP. UOP Selexol Technology for Acid Gas Removal. 2009; Available from:

www.uop.com.

4. Palla, R., D. Leppin, and A. Jamal, Status Report: Operation of the Morphysorb

Process at a Canadian Gas-Treating Plant. Gas TIPS, 2003. Summer: p. 8.

5. Jou, F.-Y., F.D. Otto, and A.E. Mather, The solubility of mixtures of H2S and

CO2 in an mdea solution. The Canadian Journal of Chemical Engineering, 1997.

75(6): p. 1138-1141.

6. Kohl, A. and R. Nielsen, Gas Purification. 5th ed1997, Texas: Gulf Publishing

Company.

7. Kaufmann, D.W. and D.W. Kaufmann, Physical properties of Sodium Chloride in

crystal, liquid, gas, and acqueous solution states1960, New York: Reinhold Pub.

Corp.

8. Stull, D.R., JANAF Thermochemical Tables. 2nd ed1971, Washington, D.C.:

NSRDS-NBS 37, U.S. Department of Commerce.

9. Chase, M.W., et al., JANAF Thermochemical Table third edition. J. Phys. Chem.

Ref. Data, 1985. 14(Supplement No. 1).

10. Fredlake, C.P., et al., Thermophysical properties of imidazolium-based ionic

liquids. Journal of Chemical & Engineering Data, 2004. 49(4): p. 954-964.

11. Plechkova, N.V. and K.R. Seddon, Applications of ionic liquids in the chemical

industry. Chem. Soc. Rev., 2008. 37(Copyright (C) 2012 American Chemical

Society (ACS). All Rights Reserved.): p. 123-150.

12. Walden, P., Molecular weights and electrical conductivity of several fused salts.

Bull. Acad. Imp. Sci. St.-Petersbourg, 1914(Copyright (C) 2012 American

Chemical Society (ACS). All Rights Reserved.): p. 405-22.

13. Wilkes, J.S. and M.J. Zaworotko, Air and water stable 1-ethyl-3-

methylimidazolium based ionic liquids. J. Chem. Soc., Chem. Commun.,

1992(Copyright (C) 2012 American Chemical Society (ACS). All Rights

Reserved.): p. 965-7.

14. Kapustinskii, A.F., Lattice energy of ionic crystals. Quart. Revs. (London), 1956.



15. Seddon, K.R., Ionic liquids for clean technology. J. Chem. Technol. Biotechnol.,

1997. 68(Copyright (C) 2012 American Chemical Society (ACS). All Rights

Reserved.): p. 351-356.

16. McFarlane, D.R., et al., High conductivity molten salts based on the imide ion.

Electrochim. Acta, 2000. 45(Copyright (C) 2012 American Chemical Society

(ACS). All Rights Reserved.): p. 1271-1278.

http://www.encana.ca/

http://www.uop.com/

217

217

17. Downard, A., et al., Structural Studies of Crystalline 1-Alkyl-3-

Methylimidazolium Chloride Salts. Chem. Mater., 2004. 16(Copyright (C) 2012

American Chemical Society (ACS). All Rights Reserved.): p. 43-48.

18. Holbrey, J.D., et al., Crystal polymorphism in 1-butyl-3-methylimidazolium

halides: supporting ionic liquid formation by inhibition of crystallization. Chem.

Commun. (Cambridge, U. K.), 2003(Copyright (C) 2012 American Chemical


19. Lewandowski, A. and A. Swiderska-Mocek, Ionic liquids as electrolytes for Li-

ion batteries: An overview of electrochemical studies. J. Power Sources, 2009.


Reserved.): p. 601-609.

20. Sugimoto, T., et al., Ionic liquid electrolyte systems based on

bis(fluorosulfonyl)imide for lithium-ion batteries. Journal of Power Sources,

2009. 189(1): p. 802-805.

21. Larush, L., et al., On the thermal behavior of model Li-Li xCoO 2 systems

containing ionic liquids in standard electrolyte solutions. Journal of Power

Sources, 2009. 189(1): p. 217-223.

22. Sirisopanaporn, C., A. Fernicola, and B. Scrosati, New, ionic liquid-based

membranes for lithium battery application. Journal of Power Sources, 2009.

186(2): p. 490-495.

23. Borgel, V., et al., On the application of ionic liquids for rechargeable Li

batteries: High voltage systems. Journal of Power Sources, 2009. 189(1): p. 331-

336.

24. Watanabe, M., S.-I. Yamada, and N. Ogata, Ionic conductivity of polymer

electrolytes containing room temperature molten salts based on pyridinium halide

and aluminum chloride. Electrochim. Acta, 1995. 40(Copyright (C) 2012


25. Tiyapiboonchaiya, C., et al., Polymer-in-ionic-liquid electrolytes.

Macromolecular Chemistry and Physics, 2002. 203(13): p. 1906-1911.

26. Forsyth, M., S. Jiazeng, and D.R. MacFarlane, Novel high salt content polymer

electrolytes based on high Tg polymers. Electrochim. Acta, 2000. 45(Copyright

(C) 2012 American Chemical Society (ACS). All Rights Reserved.): p. 1249-

1254.

27. Fuller, J., A.C. Breda, and R.T. Carlin, Ionic liquid-polymer gel electrolytes. J.

Electrochem. Soc., 1997. 144(Copyright (C) 2012 American Chemical Society

(ACS). All Rights Reserved.): p. L67-L70.

28. Fan, J., R.F. Marzke, and C.A. Angell, Conductivity vs NMR correlation times,

and decoupled cation motion in polymer-in-salt electrolytes. Mater. Res. Soc.

Symp. Proc., 1993. 293(Copyright (C) 2012 American Chemical Society (ACS).

All Rights Reserved.): p. 87-92.

29. Angell, C.A., et al., Li-conducting ionic rubbers for lithium battery and other

applications. Solid State Ionics, 1994. 69(Copyright (C) 2012 American Chemical


218

218

30. Olivier-Bourbigou, H. and L. Magna, Ionic liquids: perspectives for organic and

catalytic reactions. J. Mol. Catal. A: Chem., 2002. 182-183(Copyright (C) 2012


31. Dyson, P.J., Review: Synthesis of organometallics and catalytic hydrogenations in

ionic liquids. Applied Organometallic Chemistry, 2002. 16(9): p. 495-500.

32. Mehnert, C.P., N.C. Dispenziere, and R.A. Cook, Preparation of C9-aldehyde via

aldol condensation reactions in ionic liquid media. Chem. Commun. (Cambridge,

U. K.), 2002(Copyright (C) 2012 American Chemical Society (ACS). All Rights

Reserved.): p. 1610-1611.

33. Earle, M.J., et al., Friedel-Crafts reactions in room temperature ionic liquids.

Chem. Commun. (Cambridge), 1998(Copyright (C) 2012 American Chemical


34. Nara, S.J., J.R. Harjani, and M.M. Salunkhe, Friedel-Crafts Sulfonylation in 1-

Butyl-3-methylimidazolium Chloroaluminate Ionic Liquids. J. Org. Chem., 2001.


Reserved.): p. 8616-8620.

35. Kaufmann, D.E., M. Nouroozian, and H. Henze, Molten salts as an efficient

medium for palladium-catalyzed C-C coupling reactions. Synlett, 1996(Copyright

(C) 2012 American Chemical Society (ACS). All Rights Reserved.): p. 1091-

1092.

36. Mathews, C.J., P.J. Smith, and T. Welton, Palladium catalyzed Suzuki cross-

coupling reactions in ambient temperature ionic liquids. Chem. Commun.

(Cambridge), 2000(Copyright (C) 2012 American Chemical Society (ACS). All

Rights Reserved.): p. 1249-1250.

37. Le, B.V. and R. Gree, Wittig reactions in the ionic solvent [bmim][BF4] (1-butyl-

3-methyl-1H-imidazolium tetrafluoroborate). Chem. Commun. (Cambridge),


Reserved.): p. 2195-2196.

38. Dullius, J.E.L., et al., Selective Catalytic Hydrodimerization of 1,3-Butadiene by

Palladium Compounds Dissolved in Ionic Liquids. Organometallics, 1998. 17(5):

p. 815-819.

39. Arce, A., H. Rodriguez, and A. Soto, Use of a green and cheap ionic liquid to

purify gasoline octane boosters. Green Chemistry, 2007. 9(3): p. 247-253.

40. Seddon, K.R., Ionic liquids: a taste of the future. Nat Mater, 2003. 2(6): p. 363-

365.

41. Fischer, T., et al., Diels-Alder reactions in room-temperature ionic liquids.

Tetrahedron Lett., 1999. 40(Copyright (C) 2012 American Chemical Society


42. Bouquillon, S., et al., Heck arylation of allylic alcohols in molten salts. J.

Organomet. Chem., 2001. 634(Copyright (C) 2012 American Chemical Society


43. Wang, J., et al., Recovery of amino acids by imidazolium based ionic liquids from

aqueous media. Green Chem., 2005. 7(Copyright (C) 2012 American Chemical


219

219

44. Mulder, M., Basic Principles of Membrane Technology. 2nd ed1996, The

Netherlands: Kluwer Academic Publishers.

45. Scovazzo, P., et al., Gas separations using non-hexafluorophosphate [PF6]-

anion supported ionic liquid membranes. J. Membr. Sci., 2004. 238(Copyright

(C) 2012 American Chemical Society (ACS). All Rights Reserved.): p. 57-63.

46. Armstrong, D.W., L. He, and Y.-S. Liu, Examination of Ionic Liquids and Their

Interaction with Molecules, When Used as Stationary Phases in Gas

Chromatography. Anal. Chem., 1999. 71(Copyright (C) 2012 American Chemical


47. Anderson, J.L., et al., Characterizing Ionic Liquids On the Basis of Multiple

Solvation Interactions. J. Am. Chem. Soc., 2002. 124(Copyright (C) 2012


48. Qi, M. and D.W. Armstrong, Dicationic ionic liquid stationary phase for GC-MS

analysis of volatile compounds in herbal plants. Anal. Bioanal. Chem., 2007.


Reserved.): p. 889-899.

49. Anderson, J.L. and D.W. Armstrong, Immobilized ionic liquids as high-

selectivity/high-temperature/high-stability gas chromatography stationary phases.

Anal. Chem., 2005. 77(Copyright (C) 2012 American Chemical Society (ACS).

All Rights Reserved.): p. 6453-6462.

50. Lambertus, G.R., et al., Rapid determination of complex mixtures by dual-column

gas chromatography with a novel stationary phase combination and

spectrometric detection. J. Chromatogr., A, 2006. 1135(Copyright (C) 2012


51. Berthod, A., L. He, and D.W. Armstrong, Ionic liquids as stationary phase

solvents for methylated cyclodextrins in gas chromatography. Chromatographia,



52. Ding, J., T. Welton, and D.W. Armstrong, Chiral ionic liquids as stationary

phases in gas chromatography. Anal. Chem., 2004. 76(Copyright (C) 2012


53. Han, X. and D.W. Armstrong, Ionic liquids in separations. Acc. Chem. Res.,


Reserved.): p. 1079-1086.

54. Oldenkamp, R.D. and E.D. Margolin, The Molten Carbonate Process For Sulfur

Oxide Emissions. Chem. Eng. Prog., 1969. 65(November): p. 73-76.

55. Muldoon, M.J., et al., Improving Carbon Dioxide Solubility in Ionic Liquids. The

Journal of Physical Chemistry B, 2007. 111(30): p. 9001-9009.

56. Aki, S.N.V.K., et al., High-pressure phase behavior of carbon dioxide with

imidazolium-based ionic liquids. The Journal of Physical Chemistry B, 2004.

108(52): p. 20355-20365.

57. Anthony, J.L., et al., Anion effects on gas solubility in ionic liquids. The Journal

of Physical Chemistry B, 2005. 109(13): p. 6366-6374.

220

220

58. Kumełan, J., D. Tuma, and G. Maurer, Solubility of CO2 in the Ionic Liquids

[bmim][CH3SO4] and [bmim][PF6]. Journal of Chemical & Engineering Data,

2006. 51(5): p. 1802-1807.

59. Anthony, J.L., E.J. Maginn, and J.F. Brennecke, Solubilities and thermodynamic

properties of gases in the ionic liquid 1-n-butyl-3-methylimidazolium

hexafluorophosphate. The Journal of Physical Chemistry B, 2002. 106(29): p.

7315-7320.

60. Cadena, C., et al., Why is CO2 so soluble in imidazolium-based ionic liquids?

Journal of the American Chemical Society, 2004. 126(16): p. 5300-5308.

61. Shiflett, M.B. and A. Yokozeki, Separation of CO2 and H2S using room-

temperature ionic liquid [bmim][PF6]. Fluid Phase Equilibria, 2010. 294(1-2): p.

105-113.

62. Shiflett, M.B. and A. Yokozeki, Solubility and diffusivity of hydrofluorocarbons

in room-temperature ionic liquids. AIChE Journal, 2006. 52(3): p. 1205-1219.

63. Lee, B.-C. and S.L. Outcalt, Solubilities of Gases in the Ionic Liquid 1-n-Butyl-3-

methylimidazolium Bis(trifluoromethylsulfonyl)imide. Journal of Chemical &

Engineering Data, 2006. 51(3): p. 892-897.

64. Kim, Y.S., et al., Solubility measurement and prediction of carbon dioxide in

ionic liquids. Fluid Phase Equilibria, 2005. 228-229: p. 439-445.

65. Soriano, A.N., B.T. Doma, and M.-H. Li, Solubility of Carbon Dioxide in 1-Ethyl-

3-methylimidazolium Tetrafluoroborate. Journal of Chemical & Engineering

Data, 2008. 53(11): p. 2550-2555.

66. Soriano, A.N., B.T. Doma Jr, and M.-H. Li, Carbon dioxide solubility in 1-ethyl-

3-methylimidazolium trifluoromethanesulfonate. The Journal of Chemical

Thermodynamics, 2009. 41(4): p. 525-529.

67. Shariati, A. and C. Peters, High-pressure phase behavior of systems with ionic

liquids: II. The binary system carbon dioxide+1-ethyl-3-methylimidazolium

hexafluorophosphate. The Journal of Supercritical Fluids, 2004. 29(1-2): p. 43-48.

68. Jacquemin, J., et al., Thermophysical properties, low pressure solubilities and

thermodynamics of solvation of carbon dioxide and hydrogen in two ionic liquids

based on the alkylsulfate anion. Green Chemistry, 2008. 10(9): p. 944.

69. Jalili, A.H., et al., Solubility and diffusion of CO2 and H2S in the ionic liquid 1-

ethyl-3-methylimidazolium ethylsulfate. The Journal of Chemical

Thermodynamics, 2010. 42(10): p. 1298-1303.

70. Jalili, A.H., et al., Solubility of CO2 in 1-(2-hydroxyethyl)-3-methylimidazolium

ionic liquids with different anions. The Journal of Chemical Thermodynamics,

2010. 42(6): p. 787-791.

71. Schilderman, A., S. Raeissi, and C. Peters, Solubility of carbon dioxide in the

ionic liquid 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide. Fluid

Phase Equilibria, 2007. 260(1): p. 19-22.

72. Yokozeki, A. and M.B. Shiflett, Vapor–liquid equilibria of ammonia + ionic

liquid mixtures. Applied Energy, 2007. 84(12): p. 1258-1273.

73. Shin, E.-K. and B.-C. Lee, High-Pressure Phase Behavior of Carbon Dioxide

with Ionic Liquids: 1-Alkyl-3-methylimidazolium Trifluoromethanesulfonate.

Journal of Chemical & Engineering Data, 2008. 53(12): p. 2728-2734.

221

221

74. Blanchard, L.A., Z. Gu, and J.F. Brennecke, High-pressure phase behavior of

ionic liquid/CO2 systems. The Journal of Physical Chemistry B, 2001. 105(12): p.

2437-2444.

75. Carvalho, P.J., et al., High pressure phase behavior of carbon dioxide in 1-alkyl-

3-methylimidazolium bis(trifluoromethylsulfonyl)imide ionic liquids. The Journal

of Supercritical Fluids, 2009. 48(2): p. 99-107.

76. Shokouhi, M., et al., Solubility and diffusion of H2S and CO2 in the ionic liquid 1-

(2-hydroxyethyl)-3-methylimidazolium tetrafluoroborate. Journal of Chemical &

Engineering Data, 2009. 55(4): p. 1663-1668.

77. Jou, F.-Y. and A.E. Mather, Solubility of hydrogen sulfide in [bmim][PF6].

International Journal of Thermophysics, 2007. 28(2): p. 490-495.

78. Jalili, A.H., et al., Solubility of H2S in ionic liquids [bmim][PF6], [bmim][BF4],

and [bmim][Tf2N]. Journal of Chemical & Engineering Data, 2009. 54(6): p.

1844-1849.

79. Rahmati-Rostami, M., et al., Solubility of H2S in ionic liquids [hmim][PF6],

[hmim][BF4], and [hmim][Tf2N]. The Journal of Chemical Thermodynamics,

2009. 41(9): p. 1052-1055.

80. Sakhaeinia, H., et al., Solubility of H2S in 1-(2-hydroxyethyl)-3-

methylimidazolium ionic liquids with different anions. Fluid Phase Equilibria,

2010. 298(2): p. 303-309.

81. Sakhaeinia, H., et al., Solubility of H2S in ionic liquids 1-ethyl-3-

methylimidazolium hexafluorophosphate ([emim][PF6]) and 1-ethyl-3-

methylimidazolium bis(trifluoromethyl)sulfonylimide ([emim][Tf2N]). Journal of

Chemical & Engineering Data, 2010. 55(12): p. 5839-5845.

82. Jalili, A.H., et al., Solubility of CO2, H2S, and their mixture in the ionic liquid 1-

octyl-3-methylimidazolium bis(trifluoromethyl)sulfonylimide. The Journal of

Physical Chemistry B, 2012. 116(9): p. 2758-2774.

83. Jacquemin, J., et al., Solubility of carbon dioxide, ethane, methane, oxygen,

nitrogen, hydrogen, argon, and carbon monoxide in 1-butyl-3-methylimidazolium

tetrafluoroborate between temperatures 283 K and 343 K and at pressures close

to atmospheric. The Journal of Chemical Thermodynamics, 2006. 38(4): p. 490-

502.

84. Kumełan, J., et al., Solubility of the single gases methane and xenon in the ionic

liquid [bmim][CH3SO4]. Journal of Chemical & Engineering Data, 2007. 52(6): p.

2319-2324.

85. Kumełan, J., et al., Solubility of the single gases methane and xenon in the ionic

liquid [hmim][Tf2N]. Industrial & Engineering Chemistry Research, 2007. 46(24):

p. 8236-8240.

86. Jacquemin, J., et al., Low-pressure solubilities and thermodynamics of solvation

of eight gases in 1-butyl-3-methylimidazolium hexafluorophosphate. Fluid Phase

Equilibria, 2006. 240(1): p. 87-95.

87. Anderson, J.L., J.K. Dixon, and J.F. Brennecke, Solubility of CO2, CH4, C2H6,

C2H4, O2, and N2 in 1-Hexyl-3-methylpyridinium

Bis(trifluoromethylsulfonyl)imide: Comparison to Other Ionic Liquids. Accounts

of Chemical Research, 2007. 40(11): p. 1208-1216.

222

222

88. Husson-Borg, P., V. Majer, and M.F. Costa Gomes, Solubilities of Oxygen and

Carbon Dioxide in Butyl Methyl Imidazolium Tetrafluoroborate as a Function of

Temperature and at Pressures Close to Atmospheric Pressure†. Journal of


89. Costa Gomes, M.F., Low-pressure solubility and thermodynamics of solvation of

carbon dioxide, ethane, and hydrogen in 1-hexyl-3-methylimidazolium

bis(trifluoromethylsulfonyl)amide between temperatures of 283 K and 343 K.

Journal of Chemical & Engineering Data, 2007. 52(2): p. 472-475.

90. Anderson, J.L., et al., Measurement of SO2 Solubility in Ionic Liquids. The


91. Lee, K.Y., et al., Use of ionic liquids as absorbents to separate SO2 in SO2/O2 in

thermochemical processes to produce hydrogen. International Journal of

Hydrogen Energy, 2008. 33(21): p. 6031-6036.

92. TmoleX, 2008, COSMOlogic GmbH & CO. KG and Ryoka Systems Inc.

93. Eckert, F., et al., COSMOthermX, 2008, COSMOlogic GmbH & CO. KG and

Ryoka Systems Inc.: Leverkusen.

94. Atkins, P.W. and J.d. Paula, Physical Chemistry. 9 ed2010, New York: W. H.

Freeman and Co.

95. Levine, I.N., Quantum Chemistry. fifth ed2000, New Jersey: Prentice Hall.

96. Boys, S.F. Electronic wave functions. I. A general method of calculation for the

stationary states of any molecular system. in Proceedings of the Royal Society of

London. Series A. Mathematical and Physical Sciences. 1950. The Royal Society.

97. Hehre, W.J., R.F. Stewart, and J.A. Pople, Self Consistent Molecular Orbital

Methods. I. Use of Gaussian Expansions of Slater Type Atomic Orbitals. The

Journal of Chemical Physics, 1969. 51(6): p. 2657-2664.

98. Pople, J.A. and D.L. Beveridge, Approximate Molecular Orbital Theory1970,

New York: McGraw-Hill.

99. Roothaan, C.C.J., New Developments in Molecular Orbital Theory. Reviews of

Modern Physics, 1951. 23(2): p. 69-89.

100. Hall, G.G., The molecular-orbital theory of chemical valency. VIII. A method of

calculating ionization potentials. Proc. R. Soc. London, Ser. A, 1951.



101. Cramer, C.J., Essentials of Computational Chemistry: Theories and Models2002,

West Sussex: John Wiley & Sons, Ltd.

102. Jensen, F., Introduction to Computational Chemistry. 2nd ed2007, West Sussex:

John Wiley & Sons Ltd.

103. Hohenberg, P. and W. Kohn, Inhomogeneous Electron Gas. Physical Review,

1964. 136(3B): p. B864-B871.

104. Kohn, W. and L.J. Sham, Self-Consistent Equations Including Exchange and

Correlation Effects. Physical Review, 1965. 140(4A): p. A1133-A1138.

105. Lewars, E.G., Computational Chemistry Introduction to the Theory and

Applications of Molecular and Quantum Mechanics, 2011, Springer: New York.

106. Klamt, A., COSMO-RS: from quantum chemistry to fluid phase thermodynamics

and drug design. 1st ed2005: Amsterdam : Elsevier.

223

223

107. Andzelm, J., C. Kolmel, and A. Klamt, Incorporation of solvent effects into

density functional calculations of molecular energies and geometries. J. Chem.

Phys., 1995. 103(Copyright (C) 2012 American Chemical Society (ACS). All

Rights Reserved.): p. 9312-20.

108. Becke, A.D., Density-functional exchange-energy approximation with correct

asymptotic behavior. Physical Review A (General Physics), 1988. 38(Copyright

1989, IEE): p. 3098-100.

109. Vosko, S.H., L. Wilk, and M. Nusair, Accurate spin-dependent electron liquid

correlation energies for local spin density calculations: a critical analysis.

Canadian Journal of Physics, 1980. 58(Copyright 1980, IEE): p. 1200-11.

110. Perdew, J.P., Density-functional approximation for the correlation energy of the

inhomogeneous electron gas. Physical Review B (Condensed Matter), 1986.

33(Copyright 1986, IEE): p. 8822-4.

111. Guggenheim, E.A., Mixtures : the theory of the equilibrium properties of some

simple classes of mixtures, solutions and alloys 1952, Oxford: Clarendon Press.

112. Staverman, A.J., The entropy of high polymer solutions. Generalization of

formulae. Recueil des Travaux Chimiques des Pays-Bas, 1950. 69(2): p. 163-174.

113. Abrams, D.S. and J.M. Prausnitz, Statistical thermodynamics of liquid mixtures: A

new expression for the excess Gibbs energy of partly or completely miscible

systems. AIChE Journal, 1975. 21(1): p. 116-128.

114. Naim, A.B., Solvation thermodynamics1987, New York: Plenum Press.

115. Brennecke, J.F.G., IN), Maginn, Edward J. (Granger, IN), Purification of gas with

liquid ionic compounds, 2003, University of Notre Dame du Lac (Notre Dame,

IN): United States.

116. Mortazavi-Manesh, S., M. Satyro, and R.A. Marriott, A semi-empirical Henry's

law expression for carbon dioxide dissolution in ionic liquids. Fluid Phase

Equilibria, 2011. 307(2): p. 208-215.

117. Mortazavi-Manesh, S., M. Satyro, and R.A. Marriott, Modelling carbon dioxide

solubility in ionic liquids. Can. J. Chem. Eng., 2013. 91(4): p. 783-789.

118. Shiflett, M.B. and A. Yokozeki, Solubilities and diffusivities of carbon dioxide in

ionic liquids: [bmim][PF6] and [bmim][BF4]. Industrial & Engineering

Chemistry Research, 2005. 44(12): p. 4453-4464.

119. Shiflett, M.B. and A. Yokozeki, Solubility of CO2 in room temperature ionic

liquid [hmim][Tf2N]. The Journal of Physical Chemistry B, 2007. 111(8): p.

2070-2074.

120. Mortazavi-Manesh, S., M.A. Satyro, and R.A. Marriott, Screening ionic liquids as

candidates for separation of acid gases: Solubility of hydrogen sulfide, methane,

and ethane. AIChE J., 2013.

121. Florusse, L.J., S. Raeissi, and C.J. Peters, High-pressure phase behavior of ethane

with 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide. Journal of


122. ChemBio3D Ultra, 2007, Cambridgesoft.

123. Diedenhofen, M. and A. Klamt, COSMO-RS as a tool for property prediction of

IL mixtures-A review. Fluid Phase Equilibria, 2010. 294(Compendex): p. 31-38.

224

224

124. Maiti, A., Theoretical screening of ionic liquid solvents for carbon capture.

ChemSusChem, 2009. 2(7): p. 628-631.

125. Soave, G., Equilibrium constants from a modified Redlich-Kwong equation of

state. Chemical Engineering Science, 1972. 27(6): p. 1197-1203.

126. Klamt, A., Conductor-like screening model for real solvents: a new approach to

the quantitative calculation of solvation phenomena. The Journal of Physical

Chemistry, 1995. 99(7): p. 2224-2235.

127. Zhang, X., Z. Liu, and W. Wang, Screening of ionic liquids to capture CO2 by

COSMO-RS and experiments. AIChE Journal, 2008. 54(10): p. 2717-2728.

128. Peng, D.-Y. and D.B. Robinson, A new two-constant equation of state. Industrial

& Engineering Chemistry Fundamentals, 1976. 15(1): p. 59-64.

129. Prausnitz, J.M., R.N. Lichtenthaler, and E.G.d. Azevedo, Molecular

Thermodynamics of Fluid-Phase Equilibria. 3rd ed1999, New Jersey: Prentice

Hall International Series.

130. Wichmann, K., Email Communication, 2013, COSMOlogic specialist:

Leverkusen.

131. McQuarrie, D.A., Statistical Mechanics2000, Sausalito: University Science

Books.

132. Baltus, R.E., et al., Low-pressure solubility of carbon dioxide in room-

temperature ionic liquids measured with a quartz crystal microbalance. The


133. Scovazzo, P., et al., Regular Solution theory and CO2 gas solubility in room-

temperature ionic liquids. Industrial & Engineering Chemistry Research, 2004.

43(21): p. 6855-6860.

134. Fu, D., et al., Effect of water content on the solubility of CO2 in the ionic liquid

[bmim][PF6]. Journal of Chemical & Engineering Data, 2006. 51(2): p. 371-375.

135. Goodwin, R.D., Hydrogen Sulfide Provisional Thermophysical Properties from

188 to 700 K at Pressures to 75 Mpa, 1983, Chemical Engineering Science

Division, SBSIR83-1694, National Engineering Laboratory: Boulder, CO.

136. Renon, H. and J.M. Prausnitz, Local compositions in thermodynamic excess

functions for liquid mixtures. AIChE Journal, 1968. 14(1): p. 135-144.

137. Seader, J.D., E.J. Henley, and D.K. Roper, Separation process principles:

Chemical and biochemical operations. 3rd ed2011, Hoboken, NJ: Willey.

138. VMGsim, 2010, Virtual Materials Group Inc.: Calgary.

139. Rebelo, L.P.N., et al., On the Critical Temperature, Normal Boiling Point, and

Vapor Pressure of Ionic Liquids. The Journal of Physical Chemistry B, 2005.

109(13): p. 6040-6043.

140. Shereshefsky, J.L., Surface Tension of Saturated Vapors and the Equation of

Eötvös. The Journal of Physical Chemistry, 1931. 35(6): p. 1712-1720.

141. Guggenheim, E.A., The Principle of Corresponding States Journal of Chemical

Physics, 1945. 13: p. 253-261.

142. Shiflett, M., et al., Phase behavior of {carbon dioxide+[bmim][Ac]} mixtures.

The Journal of Chemical Thermodynamics, 2008. 40(1): p. 25-31.

143. Vetere, A., Again the Rackett equation. The Chemical Engineering Journal, 1992.

49(1): p. 27-33.

225

225

144. Valderrama, J.O. and P.A. Robles, Critical Properties, Normal Boiling

Temperatures, and Acentric Factors of Fifty Ionic Liquids. Industrial &

Engineering Chemistry Research, 2007. 46(4): p. 1338-1344.

145. Valderrama, J.O. and R.E. Rojas, Critical Properties of Ionic Liquids. Revisited.

Industrial & Engineering Chemistry Research, 2009. 48(14): p. 6890-6900.

146. Alvarez, V.H. and J.O. Valderrama, A modified Lydersen-Joback-Reid method to

estimate the critical properties of biomolecules. Alimentaria, 2004. 254: p. 55-66.

147. Strechan, A.A., et al., Thermochemical properties of 1-butyl-3-methylimidazolium

nitrate. Thermochimica Acta, 2008. 474(1–2): p. 25-31.

148. Mokhtarani, B., et al., Density and viscosity of 1-butyl-3-methylimidazolium

nitrate with ethanol, 1-propanol, or 1-butanol at several temperatures. The

Journal of Chemical Thermodynamics, 2009. 41(12): p. 1432-1438.

149. Fernández, A., et al., Thermophysical properties of 1-ethyl-3-methylimidazolium

ethylsulfate and 1-butyl-3-methylimidazolium methylsulfate ionic liquids. Journal

of Chemical & Engineering Data, 2007. 52(5): p. 1979-1983.

150. Fern ndez, A., et al., Volumetric, Transport and Surface Properties of

[bmim][MeSO4] and [emim][EtSO4] Ionic Liquids As a Function of

Temperature. Journal of Chemical & Engineering Data, 2008. 53(7): p. 1518-

1522.

151. Mokhtarani, B., et al., Densities and Viscosities of Pure 1-Methyl-3-

octylimidazolium Nitrate and Its Binary Mixtures with Alcohols at Several

Temperatures. Journal of Chemical & Engineering Data, 2010. 55(9): p. 3901-

3908.

152. Mathias, P.M., T. Naheiri, and E.M. Oh, A density correction for the Peng--

Robinson equation of state. Fluid Phase Equilibria, 1989. 47(1): p. 77-87.

153. Wei, Y., et al., Compositional Simulation Using an Advanced Peng-Robinson

Equation of State, in SPE Reservoir Simulation Symposium2011: The Woodlands,

Texas, USA.

154. Döker, M. and J. Gmehling, Measurement and prediction of vapor–liquid

equilibria of ternary systems containing ionic liquids. Fluid Phase Equilibria,

2005. 227(2): p. 255-266.

155. Kato, R. and J. Gmehling, Measurement and correlation of vapor–liquid

equilibria of binary systems containing the ionic liquids [EMIM][(CFSO)N],

[BMIM][(CFSO)N], [MMIM][(CH)PO] and oxygenated organic compounds

respectively water. Fluid Phase Equilibria, 2005. 231(1): p. 38-43.

156. Sumartschenkowa, I.A., et al., Experimental Study of Thermodynamic Properties

of Mixtures Containing Ionic Liquid 1-Ethyl-3-methylimidazolium Ethyl Sulfate

Using Gas−Liquid Chromatography and Transpiration Method. Journal of


157. Kowalsky, G., et al., Performance of Morphysorb solvent in a Commercial Acis

Gas Treating Plant, in 53rd Annual Laurance Reid Gas Conditioning

Conference2003: Norman, Oklahoma, USA.

158. Google Map. 2013; Available from: https://maps.google.ca.

159. Blanc, C., M. Grall, and D. G. The part played by degradation products in the

corrosion of gas sweetening plants using diethanolamine (DEA) and

226

226

methyldiethanolamine (MDEA). in Proceeding of the Gas Conditioning

Conference. 1982.

160. DuPart, M.S., T.R. Bacon, and D.J. Edwards, Understanding corrosion in

alkanolamine gas treating plants. Part 2. Case histories show actual plant

problems and their solutions. Hydrocarbon Process., Int. Ed., 1993. 72(Copyright

(C) 2013 American Chemical Society (ACS). All Rights Reserved.): p. 89-90, 92,

94.

161. Nielsen, R.B., et al., Controlling corrosion in amine treating plants. Proc.

Laurance Reid Gas Cond. Conf., 1995. 45th(Copyright (C) 2013 American

Chemical Society (ACS). All Rights Reserved.): p. 182-212.

162. Engineering data book: FPS version, Gas Processors Suppliers Association

(GPSA). 11 ed1998: Gas Processors Suppliers Association.

163. Younger, A.H., Natural Gas Processing Principles and Technology, 2004,

Univercity of Calgary: Calgary.

164. Honeywell, Unisim Design, 2012.

227

APPENDIX A: COSMO CALCULATIONS

Abbreviate Name: Tf2N*

Ion type: Anion CAS#: 98837-98-0

Full Name: Bis(trifluoromethylsulfonyl)-imide OR Methanesulfonamide, 1,1,1-trifluoro-N-[(trifluoromethyl)sulfonyl]

MW(g/gmol): 280.147 COSMO Energy (kJ/mol): 4802438.314

Surface Area× 1020

m2: 203.481

Structure: C2F6NO4S2

Molecular Shaped Cavity

-profile

* the ion is available in COSMOtherm database[93]

228

Abbreviate Name: PF6*


Full Name: hexaflourophosphate

MW(g/gmol): 144.964 COSMO Energy (kJ/mol): -2472349.663

Surface Area× 1020

m2: 108.624

Structure


-profile


229

Abbreviate Name: BF4*


Full Name: tetraflouroborate


Surface Area× 1020

m2: 90.570

Structure


-profile


230

Abbreviate Name: Triflate or OTF Ion type: Anion CAS#: 37181-39-8

Full Name: Trifluoromethanesulfonate OR Methanesulfonic acid, 1,1,1-trifluoro


Surface Area× 1020

m2: 127.799

Structure: CF3O3S


-profile

231

Abbreviate Name: CH3SO4 Ion type: Anion CAS#: 21228-90-0

Full Name: Methylsulfate OR Sulfuric acid, monomethyl ester, ion(1-)


Surface Area× 1020

m2: 118.640

Structure


-profile

232

Abbreviate Name: TFA Ion type: Anion CAS#: 14477-72-6

Full Name: Trifluoroacetate


Surface Area× 1020

m2: 111.870

Structure: C2F3O2


-profile

233

Abbreviate Name: C2SO4*


Full Name: Ethylsulfate


Surface Area× 1020

m2: 138.731

Structure: C2H5O4S


-profile


234

Abbreviate Name: EtGLEtGLeC2SO4 Ion type: Anion CAS#: 595565-53-0

Full Name: 2-(2-methoxyethoxy)ethyl sulfate or Diethyleneglycolmonomethylethersulfate OR Ethanol, 2-(2-methoxyethoxy)-, 1-

hydrogen sulfate, ion(1-)


Surface Area× 1020

m2: 221.057

Structure: C5H11O6S


-profile

235

Abbreviate Name: C8SO4*


Full Name: Octylsulfate OR Sulfuric acid, monooctyl ester, ion(1-)


Surface Area× 1020

m2: 255.555

Structure: C8H17O4S


-profile


236

Abbreviate Name: Doc Ion type: Anion CAS#: 10041-19-7

Full Name: Docusate or 1,4-bis(2-ethylhexyloxy)-1,4-dioxobutane-2-sulfonate


Surface Area× 1020

m2: 466.964

Structure: C20H38O7S


-profile

237

Abbreviate Name: NO3 Ion type: Anion CAS#: 14797-55-8

Full Name: Nitrate


Surface Area× 1020

m2: 76.396

Structure


-profile

238

Abbreviate Name: Cl*


Full Name: Chloride


Surface Area× 1020

m2: 52.810

Structure


-profile


239

Abbreviate Name: DEP Ion type: Anion CAS#: 48042-47-3

Full Name: Diethylphosphate OR Phosphoric acid, diethyl ester, ion(1-)


Surface Area× 1020

m2: 181.174

Structure: C4H10O4P


-profile

240

Abbreviate Name: DBP Ion type: Anion CAS#: 32288-01-0

Full Name: Dibutylphosphate OR Phosphoric acid, dibutyl ester, ion(1-)


Surface Area× 1020

m2: 260.266

Structure: C8H18O4P


-profile

241

Abbreviate Name: FEP Ion type: Anion CAS#: 429679-87-8

Full Name: Tris(pentafluoroethyl)trifluorophosphate, OR Phosphate(1-), trifluorotris(1,1,2,2,2-pentafluoroethyl)-


Surface Area× 1020

m2: 260.173

Structure: C6F18P


-profile

242

Abbreviate Name: TCA Ion type: Anion CAS#: 17997-24-9

Full Name: Tricyanomethanide OR Methanetricarbonitrile, ion(1-)


Surface Area× 1020

m2: 131.451

Structure: C4N3


-profile

243

Abbreviate Name: L Ion type: Anion CAS#:113-21-3

Full Name: Lactate OR 2-hydroxypropanoate OR Propanoic acid, 2-hydroxy-, ion(1-)


Surface Area× 1020

m2: 116.493

Structure: C3H5O3


-profile

244

Abbreviate Name: bmim*

Ion type: Cation CAS#: 80432-08-2

Full Name: 1-butyl-3-methylimidazolium OR 1H-Imidazolium, 3-butyl-1-methyl-


Surface Area× 1020

m2: 200.99376

Structure: C8H15N2


-profile


245

Abbreviate Name: emim*


Full Name: 1-ethyl-3-methylimidazolium OR 1H-Imidazolium, 3-ethyl-1-methyl-


Surface Area× 1020

m2: 161.21853

Structure: C6H11N2


-profile


246

Abbreviate Name: hmim*


Full Name: 1-hexyl-3-methylimidazolium OR 1H-Imidazolium, 3-hexyl-1-methyl-


Surface Area× 1020

m2: 241.04133

Structure: C10H19N2


-profile


247

Abbreviate Name: omim*


Full Name: 1-octyl-3-methylimidazolium OR 1H-Imidazolium, 1-methyl-3-octyl-


Surface Area× 1020

m2: 280.67573

Structure: C12H23N2


-profile


248

Abbreviate Name: emmim*


Full Name: 1-ethyl-2,3-dimethylimidazolium, 1H-Imidazolium, 3-ethyl-1,2-dimethyl-


Surface Area× 1020

m2: 175.26831

Structure: C7H13N2


-profile


249

Abbreviate Name: N-bupy Ion type: Cation CAS#: 45806-95-9

Full Name: 1-butylpyridinium OR Pyridinium, 1-butyl-


Surface Area× 1020

m2: 191.8613

Structure: C9H14N


-profile

250

Abbreviate Name: N4111 Ion type: Cation CAS#: 7685-30-5

Full Name: Butyltrimethylammonium OR 1-Butanaminium, N,N,N-trimethyl-


Surface Area× 1020

m2: 182.5554

Structure: C7H18N


-profile

251

Abbreviate Name: pmim*


Full Name: 1-pentyl-3-methylimidazolium OR 1H-Imidazolium, 1-methyl-3-pentyl-


Surface Area× 1020

m2: 220.9927

Structure: C9H17N2


-profile


252

Abbreviate Name: MeButPyrr Ion type: Cation CAS#: 223437-10-3

Full Name: 1-butyl-1-methylpyrrolidinium OR Pyrrolidinium, 1-butyl-1-methyl-


Surface Area× 1020

m2: 201.2503

Structure: C9H20N


-profile

253

Abbreviate Name: C6H4F9mim Ion type: Cation CAS#: 872672-60-1

Full Name: 1-methyl-3-(3,3,4,4,5,5,6,6,6-nonafluorohexyl)imidazolium OR 1H-Imidazolium, 1-methyl-3-(3,3,4,4,5,5,6,6,6-

nonafluorohexyl)-


Surface Area× 1020

m2: 280.9179

Structure: C10H10F9N2


-profile

254

Abbreviate Name: hmpy Ion type: Cation CAS#: 111398-60-8

Full Name: 1-hexyl-3-methylpyridinium OR Pyridinium, 1-hexyl-3-methyl-


Surface Area× 1020

m2: 254.1098

Structure: C12H20N


-profile

255

Abbreviate Name: MeBu3N or N1444 Ion type: Cation CAS#: 29814-63-9

Full Name: Methyl-tributylammonium OR 1-Butanaminium, N,N-dibutyl-N-methyl-


Surface Area× 1020

m2: 278.6938

Structure: C13H30N


-profile

256

Abbreviate Name: b2Nic Ion type: Cation CAS#: 1321655-39-3

Full Name: 1-butyl-nicotinic acid butyl ester OR Pyridinium, 1-butyl-4-[(2-methylpropoxy)carbonyl]-


Surface Area× 1020

m2: 300.5751

Structure: C14H22NO2


-profile

257


Full Name: Tetrabutylammonium OR 1-Butanaminium, N,N,N-tributyl-


Surface Area× 1020

m2: 326.0087

Structure: C16H36N


-profile

258

Abbreviate Name: bmmim*


Full Name: 1-butyl-2,3-dimethylimidazolium OR 1H-Imidazolium, 3-butyl-1,2-dimethyl-


Surface Area× 1020

m2: 215.00855

Structure: C9H17N2


-profile


259

Abbreviate Name: HOemim Ion type: Cation CAS#: 61755-32-6

Full Name: 1-(2-hydroxyethane)-3-methylimidazolium OR 1H-Imidazolium, 3-(2-hydroxyethyl)-1-methyl-


Surface Area× 1020

m2: 171.4358

Structure: C6H11N2O


-profile

260


Full Name: Ethyl-propyl-dimethylammonium, OR 1-Propanaminium, N-ethyl-N,N-dimethyl-


Surface Area× 1020

m2: 175.3617

Structure: C7H18N


-profile

261

Abbreviate Name: ETT Ion type: Cation CAS#: 83008-27-9

Full Name: S-Ethyl-tetramethylisothiouronium OR Methanaminium, N-[(dimethylamino)(ethylthio)methylene]-N-methyl-


Surface Area× 1020

m2: 207.0881

Structure: C7H17N2S


-profile

262

Abbreviate Name: tmg Ion type: Cation CAS#:

Full Name: Tetramethylguanidinium OR Guanidine, N, N',N',N'-tetramethyl-


Surface Area× 1020

m2: 165.1547

Structure: C5 H14N3


-profile

263

Abbreviate Name: hmg Ion type: Cation CAS#: 44872-05-1

Full Name: Hexamethylguanidinium


Surface Area× 1020

m2: 196.5339

Structure: C7H18N3


-profile

264

Abbreviate Name: pmg Ion type: Cation CAS#: 119543-23-6

Full Name: Pentamethylguanidinium OR Methanaminium, 1-(dimethylamino)-1-imino-N,N,N-trimethyl-


Surface Area× 1020

m2: 175.6711

Structure: C6H16N3


-profile

265

Abbreviate Name: pmeg Ion type: Cation CAS#:

Full Name: Pentamethylethylguanidinium


Surface Area× 1020

m2: 212.5013

Structure: C8H20N3


-profile

266

Abbreviate Name: pmpg Ion type: Cation CAS#: 1394900-47-0

Full Name: Pentamethylpropylguanidinium OR Methanaminium, N-[(dimethylamino)(methylpropylamino)methylene]-N-methyl-


Surface Area× 1020

m2: 230.9859

Structure: C9H22N3


-profile

267

Abbreviate Name: tmdeg Ion type: Cation CAS#: 227096-59-5

Full Name: Tetramethyldiethylguanidinium OR Ethanaminium, N-[bis(dimethylamino)methylene]-N-ethyl-


Surface Area× 1020

m2: 223.9279

Structure: C9H22N3


-profile

268

Abbreviate Name: tmdpg Ion type: Cation CAS#: 204865-16-7

Full Name: Tetramethyldipropylguanidinium OR 1-Propanaminium, N-[bis(dimethylamino)methylene]-N-propyl-


Surface Area× 1020

m2: 251.0080

Structure: C11H26N3


-profile

269

269

APPENDIX B:

Peng-Robinson Equation of State[128]

B- 1

B- 2

0231 32223 BBABZBBAZBZ

B- 3

22TR

apA

B- 4

RT

bpB

B- 5

RT

pvz

B- 6

C

CC

p

TRTa

22

45724.0

B- 7

C

CC

p

RTTb 07780.0

B- 8

307.0CZ

B- 9

)()(

)(

bvbbvv

Ta

bv

RTp

270

270

,rc TTaTa

B- 10

cTbTb

B- 11

5.05.0 11 rT

B- 12 226992.054226.137464.0

B- 13

BZ

BZ

B

ABZZ

414.0

414.2ln

22)ln(1ln

B- 14

BZ

BZ

b

b

a

ax

B

ABZZ

b

b

px

f ij

iji

i

i

ii

414.0

414.2ln

2

22ln1lnln

B- 15

i j

ijji axxa

B- 16

i

iibxb

B- 17

2/12/11 jiijij aaka

In which, kij is the binary interaction parameter.

Soave–Redlich–Kwong equation of state [125]

B- 18

bvv

Ta

bv

RTp

)(

271

271

B- 19

0223 ABBBAZZZ

B- 20

22TR

apA

B- 21

RT

bpB

B- 22

C

CC

p

TRTa

22

42747.0

B- 23

C

C

p

RTb 08664.0

B- 24

TTaTa C

B- 25

Z

BZ

B

ABZZ lnln1ln

B- 26

5.05.0 11 rT

B- 27 217.057.1480.0

B- 28

Z

B

b

b

a

a

B

ABZZ

b

b

px

f iii

i

i

i 1ln2ln1lnln5.0

5.0

B- 29

272

272

5.05.0

5.05.0

5.0

5.0

/

/

CiCiii

CiCiii

pTx

pT

a

a

B- 30

CiCii

CiCii

pTx

pT

b

b

/

/

Advanced Peng Robinson Equation of State[152]

B- 31

41.0

41.0c

corr fsvv

where s, is a component dependent molar volume correction term. v is the molar volume

predicted by Peng-Robinson equation of state and Corrv refers to the corrected molar

volume.

B- 32

Tv

p

RT

v

2

B- 33

sbvsvvf c

PR

ccc 946.3

B- 34

i

ii sxs

B- 35

i

ciiC vxv

273

273

Non-Random Two-Liquid Model (NRTL) [136]

B- 36

1212

1212

2121

212121

Gxx

G

Gxx

Gxx

RT

g E

B- 37

ijijijG exp

B- 38

TCT

BA ij

ij

ijij ln

274

Documents

Application of Ionic Liquids for Gas Sweetening