(Advances in natural gas engineering) Carroll, John J. Wu, Ying Zhu, Weiyao-Carbon dioxide capture and acid gas injection-Wiley (2017)

Carbon Dioxide Capture and
Acid Gas Injection
Scrivener Publishing
100 Cummings Center, Suite 541J
Beverly, MA 01915-6106
Publishers at Scrivener
Martin Scrivener (martin@scrivenerpublishing.com)
Phillip Carmical (pcarmical@scrivenerpublishing.com)
Carbon Dioxide Capture and
Acid Gas Injection
Edited by
Ying Wu, John J. Carroll and Weiyao Zhu
This edition first published 2017 by John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
and Scrivener Publishing LLC, 100 Cummings Center, Suite 541J, Beverly, MA 01915, USA
© 2017 Scrivener Publishing LLC
For more information about Scrivener publications please visit www.scrivenerpublishing.com.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title
is available at http://www.wiley.com/go/permissions.
Wiley Global Headquarters
111 River Street, Hoboken, NJ 07030, USA
For details of our global editorial offices, customer services, and more information about Wiley products
visit us at www.wiley.com.
Limit of Liability/Disclaimer of Warranty
While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and
specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials, or promotional statements for this work. The fact that an organization,
website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the
understanding that the publisher is not engaged in rendering professional services. The advice and
strategies contained herein may not be suitable for your situation. You should consult with a specialist
where appropriate. Neither the publisher nor authors shall be liable for any loss of profit or any other
commercial damages, including but not limited to special, incidental, consequential, or other damages.
Further, readers should be aware that websites listed in this work may have changed or disappeared
between when this work was written and when it is read.
Library of Congress Cataloging-in-Publication Data
ISBN 978-1-118-93866-9
Cover image: Gas Drilling Machine | Cylonphoto | Dreamstime.com | Gas Storage Spheres |
Sasin Tipchai | Dreamstime.com | Na tural Gas Plant | Jevtic | Dreamstime.com
Cover design by Kris Hackerott
Set in size of 11pt and Minion Pro by Exeter Premedia Services Private Ltd., Chennai, India
Printed in
10 9 8 7 6 5 4 3 2 1
Contents
Prefacexiii
1Enthalpies of Carbon Dioxide-Methane and Carbon
Dioxide-Nitrogen Mixtures: Comparison with
Thermodynamic Models
1
Erin L. Roberts and John J. Carroll
1.1 Introduction
1
1.2 Enthalpy
2
1.3 Literature Review
2
1.3.1 Carbon Dioxide-Methane
4
1.3.2 Carbon Dioxide-Nitrogen
4
1.4 Calculations
5
1.4.1 Benedict-Webb-Rubin
6
1.4.2 Lee-Kesler
12
1.4.3 Soave-Redlich-Kwong
17
1.4.4 Peng-Robinson
23
1.4.5 AQUAlibrium
28
1.5 Discussion
33
1.6 Conclusion
36
References37
2Enthalpies of Hydrogen Sulfide-Methane Mixture:
Comparison with Thermodynamic Models
Erin L. Roberts and John J. Carroll
2.1 Introduction
2.2 Enthalpy
2.3 Literature Review
2.4 Calculations
2.4.1 Lee-Kesler
2.4.2 Benedict-Webb-Rubin
2.4.3 Soave-Redlich-Kwong
39
39
40
40
41
41
43
43
v
vi Contents
2.4.4 Redlich-Kwong
47
2.4.5 Peng-Robinson
47
2.4.6 AQUAlibrium
50
2.5 Discussion
50
2.6 Conclusion
52
References54
3Phase Behavior and Reaction Thermodynamics
55
Involving Dense-Phase CO2 Impurities
J.A. Commodore, C.E. Deering and R.A. Marriott
3.1 Introduction
55
3.2 Experimental
57
3.3 Results and Discussion
58
3.3.1 Phase Behavior Studies of SO2 Dissolved in
58
Dense CO2 Fluid
3.3.2 The Densimetric Properties of CS2 and
60
CO2 Mixtures
References61
4Sulfur Recovery in High Density CO2 Fluid
63
S. Lee and R.A. Marriott
4.1 Introduction
64
4.2 Literature Review
64
4.3 Methodology
65
4.4 Results and Discussion
66
4.5 Conclusion and Future Directions
67
References68
5Carbon Capture Performance of Seven Novel
Immidazolium and Pyridinium Based Ionic Liquids
Mohamed Zoubeik, Mohanned Mohamedali and Amr Henni
5.1 Introduction
5.2 Experimental Work
5.2.1 Materials
5.2.2 Density Measurement
5.2.3 Solubility Measurement
5.3 Modeling
5.3.1 Calculation of Henry’s Law Constants
5.3.2 Critical Properties Calculations
5.3.3 Peng Robinson EoS
71
71
73
73
73
73
76
76
76
76
Contents vii
5.4
Results and Discussion
77
5.4.1 Density
77
5.4.2 Critical Properties
77
5.4.3 CO2 Solubility
78
5.4.4 The Effect of Changing the Cation
81
5.4.5 The Effect of Changing the Anion
84
5.4.6 Henry’s Law Constant, Enthalpy and Entropy
Calculations85
5.4.7 Thermodynamic Modeling of CO2 Solubility
86
5.5 Conclusion
87
Acknowledgements88
References88
6Vitrisol a 100% Selective Process for H2S Removal in the
Presence of CO291
W.N. Wermink, N. Ramachandran, and G.F. Versteeg
6.1 Introduction
92
6.2 Case Definition
94
6.3 “Amine-Treated” Cases by PPS
95
6.3.1 Introduction to PPS
95
6.3.2 Process Description
96
6.3.3 PFD
97
6.3.4 Results
97
6.3.4.1 Case 1
97
6.3.4.2 Case 2
97
6.4 Vitrisol Process Extended with Regeneration of Active
Component99
6.4.1 Technology Description
99
6.4.2 Parameters Determining the Process Boundary
Conditions99
6.4.3 Absorption Section
101
6.4.4 Regeneration Section
102
6.4.5 Sulphur Recovery Section
104
6.4.6 CO2-Absorber105
6.4.7 PFD
105
6.5 Results
105
6.6 Discussion
110
6.6.1 Comparison of Amine Treating
Solutions to Vitrisol
6.6.2 Enhanced H2S Removal of Barnett
Shale Gas (case 2)
112
viii Contents
6.7 Conclusions
113
6.8 Notation
115
References115
Appendix 6-A: H&M Balance of Case 1 (British
Columbia shale) of the Amine Process
117
Appendix 6-B H&M Balance of Case 2a (Barnett shale)
of the Amine Process with Stripper Promoter
119
Appendix 6-C H&M Balance of Case 3 (Barnett shale)
of the Amine Process (MEA)
121
Appendix 6-D: H&M Balance of Case 1 (British
123
Columbia shale) of the Vitrisol process
Appendix 6-E H&M Balance of Case 2 (Barnett shale)
125
of the Vitrisol Process
7
New Amine Based Solvents for Acid Gas Removal
127
Yohann Coulier, Elise El Ahmar, Jean-Yves Coxam,
Elise Provost, Didier Dalmazzone, Patrice Paricaud,
Christophe Coquelet and Karine Ballerat-Busserolles
7.1 Introduction
128
7.2 Chemicals and Materials
131
7.3 Liquid-Liquid Equilibria
131
7.3.1 LLE in {methylpiperidines – H2O} and
{methylpiperidines – H2O – CO2}131
7.3.2 Liquid-Liquid Equilibria of Ternary
135
Systems {Amine – H2O – Glycol}
7.3.3 Liquid-Liquid Equilibria of the Quaternary
Systems {CO2 – NMPD – TEG – H2O}136
7.4 Densities and Heat Capacities of Ternary
137
Systems {NMPD – H2O – Glycol}
7.4.1 Densities
137
7.4.2 Specific Heat Capacities
137
7.5 Vapor-Liquid Equilibria of Ternary Systems
{NMPD – TEG – H2O – CO2}139
7.6 Enthalpies of Solution
140
7.7 Discussion and Conclusion
143
Acknowledgments143
References144
Contents ix
8Improved Solvents for CO2 Capture by Molecular
Simulation Methodology
147
William R. Smith
8.1 Introduction
147
8.2 Physical and Chemical Models
149
8.3 Molecular-Level Models and Algorithms for
Thermodynamic Property Predictions
150
8.4 Molecular-Level Models and Methodology for
MEA–H2O–CO2153
8.4.1 Extensions to Other Alkanolamine Solvents
and Their Mixtures
155
Acknowledgements157
References157
9Strategies for Minimizing Hydrocarbon Contamination
in Amine Acid Gas for Reinjection
161
Mike Sheilan, Ben Spooner and David Engel
9.1 Introduction
162
9.2 Amine Sweetening Process
162
9.3 Hydrocarbons in Amine
164
9.4 Effect of Hydrocarbons on the Acid Gas
Reinjection System
166
9.5 Effect of Hydrocarbons on the Amine Plant
167
9.6 Minimizing Hydrocarbon Content in Amine Acid Gas
171
9.6.1 Option 1. Optimization of the
Amine Plant Operation
171
9.6.2 Option 2. Amine Flash Tanks
176
9.6.3 Option 3. Rich Amine Liquid Coalescers
178
9.6.4 Option 4. Use of Skimming Devices
180
9.6.5 Option 5. Technological Solutions
182
References183
10Modeling of Transient Pressure Response for CO2 Flooding
Process by Incorporating Convection and Diffusion Driven
Mass Transfer
Jianli Li and Gang Zhao
10.1 Introduction
10.2 Model Development
10.2.1 Pressure Diffusion
10.2.2 Mass Transfer
10.2.3 Solutions
185
186
187
187
188
190
x Contents
10.3
Results and Discussion
191
10.3.1 Flow Regimes
191
10.3.2 Effect of Mass Transfer
192
10.3.3 Sensitivity Analysis
195
10.3.3.1 CO2 Bank
195
10.3.3.2 Reservoir Outer Boundary
196
10.4 Conclusions
196
Acknowledgments197
References197
11 Well Modeling Aspects of CO2 Sequestration
199
Liaqat Ali and Russell E. Bentley
11.1 Introduction
199
11.2 Delivery Conditions
200
11.3 Reservoir and Completion Data
201
11.4 Inflow Performance Relationship (IPR) and
Injectivity Index
201
11.5 Equation of State (EOS)
202
11.6 Vertical Flow Performance (VFP) Curves
205
11.7 Impact of the Well Deviation on CO2 Injection
208
11.8 Implication of Bottom Hole Temperature (BHT)
on Reservoir
209
11.9 Impact of CO2 Phase Change
213
11.10 Injection Rates, Facility Design Constraints and
Number of Wells Required
214
11.11 Wellhead Temperature Effect on VFP Curves
214
11.12 Effect of Impurities in CO2 on VFP Curves
216
11.13 Concluding Remarks
217
Conversion Factors
218
References218
12Effects of Acid Gas Reinjection on Enhanced Natural
Gas Recovery and Carbon Dioxide Geological Storage:
Investigation of the Right Bank of the Amu Darya River
Qi Li, Xiaying Li, Zhiyong Niu, Dongqin Kuang, Jianli Ma,
Xuehao Liu, Yankun Sun and Xiaochun Li
12.1 Introduction
12.2 The Amu Darya Right Bank Gas Reservoirs
in Turkmenistan
221
222
223
Contents xi
12.3
Model Development
223
12.3.1 State equation
224
12.3.1.1 Introduction of Traditional PR
State Equation
224
12.3.1.2 Modifications for the
Vapor-Aqueous System
224
12.3.2 Salinity
225
12.3.3 Diffusion
226
12.3.3.1 Diffusion Coefficients
226
12.3.3.2 The Cross-Phase Diffusion
Coefficients226
12.4 Simulation Model
227
12.4.1 Model Parameters
227
12.4.2 Grid-Sensitive Research of the Model
227
12.4.3 The Development and Exploitation Mode
230
12.5 Results and Discussion
230
12.5.1 Reservoir Pressure
230
12.5.2 Gas Sequestration
232
12.5.3 Production
235
12.5.4 Recovery Ratio and Recovery Percentage
238
12.6 Conclusions
239
12.7 Acknowledgments
240
References241
Index245
Preface
The Sixth International Acid Gas Injection Symposium (AGIS VI) was held
in Houston, Texas, in September 2016. As with previous Symposia, the focus
of AGIS VI was the injection of acid gas (CO2, H2S, and mixtures of these
components) for the purposes of disposal or for enhanced oil and/or gas
recovery. This book contains select papers from the Symposium in Houston.
The capture of carbon dioxide from flue gas and its disposal into a subsurface geological formation remains a viable option for the clean use of
hydrocarbon fuels. The related technology is acid gas injection. Here the
H2S and CO2 are removed from raw natural gas. This volume contains
papers directly related to these two topics ranging from the physical properties of the gas mixtures, evaluation of new and existing solvents, and
subsurface engineering aspects of the process. Furthermore, contributors
came from Canada, Europe, and China, as well as from the host country,
the United States. And this is reflected in the papers in this volume.
On a very sad note, Marco Satyro passed away on September 8, 2016,
just prior to the Symposium. Marco was a good friend of AGIS being an
active member of the Technical Committee for many years. He contributed
many papers and encouraged many others to participate. At the first AGIS
he presented the paper “The Performance of State of the Art Industrial
Thermodynamic Models for the Correlation and Prediction of Acid Gas
Solubility in Water” and this paper appeared in the first volume of the
Advances in Natural Gas Engineering. He also was the coauthor of several
other contributions to the Series and they are listed below. This volume is
dedicated to the memory of Dr. Satyro.
References – papers of M.A. Satyro from the Advances in Natural
Gas Engineering series.
M.A. Satyro, and J. van der Lee, “The Performance of State of the Art Industrial
Thermodynamic Models for the Correlation and Prediction of Acid Gas
Solubility in Water”, pp. 21–34, Acid Gas Injection and Related Technologies, Y.
Wu and J.J. Carroll (eds.), Scrivener Publishing (2011).
xiii
xiv Preface
H. Motahhari, M.A. Satyro, and H.W. Yarranton, “Acid Gas Viscosity Modeling
with the Expanded Fluid Viscosity Correlation”, pp. 41–52, Carbon Dioxide
Sequestration and Related Technologies, (2011), Y. Wu, J.J. Carroll, and Z. Du
(eds.), Scrivener Publishing (2011).
J. van der Lee, J.J. Carroll, and M.A. Satyro, “A Look at Solid CO2 Formation in
Several High CO2 Concentration Depressuring Scenarios”, pp. 117–128, Sour
Gas and Related Technologies, Y. Wu, J.J. Carroll, and W. Zhu (eds), Scrivener
Publishing (2012).
M.A. Satyro, and J.J. Carroll, “Phase Equilibrium in the Systems Hydrogen Sulfide
+ Methanol and Carbon Dioxide + Methanol”, pp. 99–109, Gas Injection for
Disposal and Enhanced Recovery, Y. Wu, J.J. Carroll, and Q. Li (eds.), Scrivener
Publishing (2014).
A.R.J. Arendsen, G.F. Versteeg, J. van der Lee,R. Cota, and M.A. Satyro,
“Comparison of the Design of CO2-capture Processes using Equilibrium and
Rate Based Models”, pp. 155–174, Gas Injection for Disposal and Enhanced
Recovery, Y. Wu, J.J. Carroll, and Q. Li (eds.), Scrivener Publishing (2014).
M.A. Satyro and H.W. Yarranton, “A Simple Model for the Calculation of Electrolyte
Mixture Viscosities”, pp. 95–104, Acid Gas Extraction for Disposal and Related
Topics, Y. Wu, J.J. Carroll, and W. Zhu (eds.), Scrivener Publishing (2016).
1
Enthalpies of Carbon Dioxide-Methane
and Carbon Dioxide-Nitrogen Mixtures:
Comparison with Thermodynamic Models
Erin L. Roberts and John J. Carroll
Gas Liquids Engineering, Calgary, Alberta, Canada
Abstract
The physical properties of acid-gas injection streams are important for use in
design considerations of the acid-gas scheme. One such property is the enthalpy
of the stream. As carbon dioxide is rarely pure, with methane and nitrogen being
common impurities in the stream, the effect of these impurities on the enthalpy is
also important to consider.
This study compares experimentally determined excess enthalpies and enthalpy
departures from literature to the enthalpy predictions of five different models,
Benedict-Webb-Rubin, Lee-Kesler, Soave-Redlich-Kwong, and Peng-Robinson
from VMGSim, as well as AQUAlibrium software. The mixtures studied are carbon dioxide-methane, as well as carbon dioxide- nitrogen mixtures at a wide range
of compositions.
The Soave-Redlich-Kwong model gave the most accurate predictions for
both the excess enthalpies and enthalpy departures, with Lee-Kesler frequently
giving the least accurate predictions for the mixtures.
1.1
Introduction
An increase in demand of natural gas has led producers to pursue poorer
quality reservoirs. These contain higher levels of carbon dioxide that then
must be responsibly disposed. Regulations prevent the flaring of the acidgas mixtures, therefore requiring an alternate means of disposal. One such
method is the injection of acid gas into subsurface reservoirs.
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (1–38)
2017 © Scrivener Publishing LLC
1
2 Carbon Dioxide Capture and Acid Gas Injection
An understanding of the physical properties of the stream is essential in
the design of the acid-gas injection scheme. The enthalpy of the stream is
required in the design of the compressor for injection. Common impurities
in the carbon dioxide include methane and nitrogen; therefore the effect of
these impurities on the enthalpy of carbon dioxide is required for design.
This paper investigates the accuracy of five different thermodynamic
models for predicting such mixtures. Four different equations of state,
Benedict-Webb-Rubin (BWR), Lee-Kesler (LK), Soave-Redlich-Kwong
(SRK), Peng-Robinson (1978) were used with VMGSim software, as well
as the AQUAlibrium model. BWR and LK are multi-constant equations,
and SRK and PR78 are cubic equations of state. The AQUAlibrium model
uses a variation of Peng-Robinson.
1.2
Enthalpy
The enthalpy of mixtures can be determined in a number of ways. One
method is to use excess enthalpy (enthalpy of mixing). Excess enthalpy is
defined as
HE
Hm
i
xi H i
(1.1)
where: HE – Excess enthalpy
Hm – Enthalpy of mixture
Hi – Enthalpy of component i
xi – mol fraction of component i
Alternatively, the enthalpy of the mixture can be represented as an
enthalpy departure, a difference between the enthalpy at a given pressure,
and the enthalpy at a reference pressure while keeping the temperature
constant.
Enthalpies can be expressed in J/mol, or for greater relevance to acidgas injection design, can be expressed in HP/MMSCFD. The conversion
between units is 1 HP/MMSCFD to 53.86 J/mol.
1.3
Literature Review
A review of literature was performed to compile experimental data for the
enthalpy of carbon dioxide-methane mixtures as well as carbon dioxidenitrogen mixtures. Table 1.1 summarizes the relevant data used in this study.
1.
2.
3.
4.
5.
6.
Lee & Mather(1972)
Barry et ale (1982)
Ng & Mather(1976)
Peterson & Wilson (1974)
Lee & Mather(1970)
Hejmadi et ale (1971)
31,40
3.5,6.5
0.1-0.9
1-12
40
0.5
0.7-13.8
-46-149
0.2-0.7
0.145, 0.423
3-13.7
0-90
0.1-0.9
0.1-0.9
1-11
0.5-4.6
Composition(mol% CO)
Pressure (MPa)
20,32,40
10-80
Temperature(DC)
Nitrogen
Nitrogen
Methane
Methane
Methane
Methane
Impurity
Table 1.1 Summaryof experimentaldata ofenthalpyof carbondioxide mixtures.
• Excess enthalpies
• All vapourenthalpies
• 27 data points
• Excess enthalpies
• All vapourenthalpies
• 108 data points
• Enthalpydeparture
• Liquid and vapourdensities
• 46 data points
• Enthalpydeparture
• Liquid and vapourdensities
• 42 data points
• Excess enthalpies
• All vapourenthalpies
• 60 data points
• Excess enthalpies
• All vapourenthalpies
• 646 data points
Comments
6
5
4
3
2
1
Ref.
Enthalpies of Carbon Dioxide-Methane
4
CARBONDIOXIDECAPTUREANDACID GAS INJECTION
1.3.1
CarbonDioxide-Methane
The most extensive study performed for enthalpies of carbon dioxidemethane mixtures was performedby Lee & Mather (1972). Their study
consisted of mol fractions of 0.1-0.9, taken at intervals of 0.1, for a
total of 9 different mol fractions. Measurementsof excess enthalpywere
reportedat 8 different temperatures from10-80 °C, with ranges of pressure of 1.0-4.4 MPafor 10°C, 1.0- 5.07 for 20 °C, 1.0-11.1 for 40°C, and
1.0-10.1 for 32 °C, 50°C, 60 °C, 70°C, and80 °C. In total,648 datapoints
were reported.Two typographicalerrors were found in the data set; they
are not included in the numerical error analysis but are represented in
the figures.
Anothersmaller study was performedby Barry et ale (1982), for excess
enthalpies of carbon dioxide-methanemixtures. Datawas taken at three
different temperatures,20°C, 32 °C, and 40°C. Seven different pressures
were used, rangingfrom 0.51 MPato 4.6 MPa,with pressureofover 2 MPa
onlybeing measuredfor 40°C. The mol fractionsmeasuredwere nottaken
in increments, instead were taken at a wide variety of fractions ranging
from 0.1 to 0.9.
Two other studies were done using enthalpy departures by Ng &
Mather(1976) and Peterson & Wilson (1974). Ng & Mather(1976) used
pressures of 3-13.7 MPa, and temperaturesof 0-90 °C for mol fractions
of 0.145 and 0.423. They used the ideal gas enthalpyas a reference point
to measure the enthalpydeparture.Peterson & Wilson (1974) only measured equimolarmixtures of carbondioxide and methanewith pressures
from 0.7-13.8 MPa and temperaturesof 255.4 K-422 K. The reference
enthalpyused was measuredat a pressure of 0.138 MPa. These twostudies were the only ones that measured both liquid and vapor enthalpies,
insteadofjust vapor.
1.3.2
CarbonDioxide-Nitrogen
Lee & Mather(1970) and Hejmadi et ale (1971) studiedthe excess enthalpies of carbondioxide-nitrogenmixtures. Lee & Mather(1970) looked at
mole fractionsfrom 0.1-0.9 at intervalsof0.1. Pressuresfrom 1.01 MPato
12.16 MPawere used, atonly a single temperatureof40°C.
Hejmadi et ale (1971) used only two differenttemperaturesof 31°C and
40 °C, andtwo differentpressuresof3.5 MPaand6.5 MPa. Theyused mole
fractionof nitrogenfrom 0.2-0.7.
Enthalpies of Carbon Dioxide-Methane
1.4
5
Calculations
The experimental enthalpies were compared to calculated enthalpies using
BWR, LK, SRK, and PR78 thermodynamic models from VMGSim software, as well as using AQUAlibrium software.
The six different mixtures (four with methane, two with nitrogen) as
summarized in Table 1.1 were evaluated. Four error functions for both the
excess enthalpies and the enthalpy departures were used to analyze the
accuracy of the prediction of each method.
For the excess enthalpies, the absolute average difference (AAD) was
defined as;
1
NP
AAD
E
E
H exp
H calc
(1.2)
where: NP – number of points
HEexp – experimental excess enthalpy
HEcalc – calculated excess enthalpy
and the average difference (AD) was defined as:
AD
1
NP
E
E
H exp
H calc
(1.3)
The absolute average error (AAE) in excess enthalpies was defined as:
AAE
1
NP
E
E
H exp
H calc
E
H calc
100%
(1.4)
and the average error (AE) was defined as:
AE
1
NP
E
E
H exp
H calc
E
H calc
100%
(1.5)
For enthalpy departures, the absolute average difference
AAD
1
NP
(H o H )exp (H o H )calc
(1.6)
6 Carbon Dioxide Capture and Acid Gas Injection
where H° – enthalpy of mixture at reference pressure
H – enthalpy of mixture at measured pressure
and the average difference was defined as:
AD
1
NP
(H o H )exp (H o H )calc
The absolute average error for enthalpy departure was defined as:
AAE
1
NP
(H o H )exp (H o H )calc
(H o H )calc
100%
(1.7)
(1.8)
and the average error was defined as:
AE
1.4.1
1
NP
(H o H )exp (H o H )calc
(H o H )calc
100%
(1.9)
Benedict-Webb-Rubin
For the Lee & Mather (1972) methane data of excess enthalpies, the AAD
was 78.1 J/mol and the AD was 2.6 J/mol. The AAE was 19.0% and the
AE was -14.6%. The maximum difference was 2113.2 Jlmol occurring
at 8.11 MPa and a mole fraction of 0.2. The maximum error was 131.7%
at the same conditions as the maximum difference. At lower pressures,
the enthalpies were overestimated, and at the higher pressures they were
underestimated. The greatest deviations occurred when there was a rapid
change in enthalpy with pressure. This occurred at around 7–10 MPa for
the 32°C and 40 °C temperatures.There was also avery large difference
between the calculated and experimental enthalpy for the 10.13 MPa isobar at 50 °C. Figures 1.1 through 1.8 show the experimental and calculated
enthalpies for the different temperatures.
The Barry et al. (1982) methane data of excess enthalpies had an AAD
of9.1 Jlmol, an AD of-8.3 Jlmol, an AAEof 14.2% and an AEof -11.0%.
The maximumdifference was 46.5 Jlmol at 4.6 MPa,40°C and 0.351 mole
fractionmethane.The maximumerrorwas 42.5% at 0.53 MPa,32°C and
0.63 mole fraction methane The deviations are smaller due to the lower
pressure range of the data.
The Lee & Mather (1970) nitrogen data of excess enthalpies taken at
40°C had similar results as the Lee& Mather (1972) methane data for
the 40 °C data,with the greatest difference occurring at 9.12 MPa. The
Enthalpies of Carbon Dioxide-Methane
600
500
Excess enthalpy (J/mol)
4.36
400
1.01 MPa
2.03 MPa
3.04 MPa
4.05 MPa
4.36 MPa
4.05
300
3.04
200
2.03
100
1.01
0
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.1 Experimental and calculated enthalpies at 10 °C using BWR (Lee & Mather,
1972).
700
600
Excess enthalpy (J/mol)
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
5.07
500
400
300
4.05
200
3.04
100
2.03
1.01
0
0
0.2
0.4
0.6
Mol fraction methane (J/mol)
0.8
1
Figure 1.2 Experimental and calculated enthalpies at 20 °C using BWR (Lee & Mather,
1972).
7
8 Carbon Dioxide Capture and Acid Gas Injection
4,500
8.61
4,000
10.1
3,500
Excess enthalpy (J/mol)
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
8.61 MPa
9.12 MPa
10.13 MPa
9.12
3,000
2,500
2,000
8.11
1,500
1,000
7.09
500
6.08
5.07
0
0
4.05
3.0
2.0
1.01
0.4
0.6
Mol fraction methane
0.2
1
0.8
Figure 1.3 Experimental and calculated enthalpies at 32 °C using BWR (Lee & Mather,
1972).
3,500
2,500
Excess enthalpy (J/mol)
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
11.15 MPa
11.15
3,000
10.1
2,000
9.12
1,500
8.11
1,000
7.09
500
0
0
0.2
6.08
5.07
4.05 3.04
2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.4 Experimentaland calculated enthalpies at40°C using BWR (Lee& Mather,
1972).
Enthalpies of Carbon Dioxide-Methane
2,000
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
1,800
Excess enthalpy (J/mol)
1,600
1,400
10.13
1,200
9.12
1,000
800
8.11
600
7.09
400
6.08
200
5.07
4.05
2.02
0
0
0.2
3.04
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.5 Experimental and calculated enthalpies at 50 °C using BWR (Lee & Mather,
1972).
1,200
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
1,000
Excess enthalpy (J/mol)
10.13
800
9.12
600
8.11
7.09
400
6.08
5.07
200
4.05
3.04
0
0
0.2
2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.6 Experimental and calculated enthalpies at 60 °C using BWR (Lee & Mather,
1972).
9
10 Carbon Dioxide Capture and Acid Gas Injection
800
600
Excess enthalpy (J/mol)
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
10.13
700
9.12
500
8.11
400
7.09
300
6.08
5.07
200
4.05
100
0
3.04
0
0.2
2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.7 Experimental and calculated enthalpies at 70 °C using BWR (Lee & Mather,
1972).
600
10.13
Excess enthalpy (J/mol)
500
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
9.12
400
8.11
7.09
300
6.08
200
5.07
4.05
100
3.04
2.02
1.01
0
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.8 Experimental and calculated enthalpies at 80 °C using BWR (Lee & Mather,
1972).
Enthalpies of Carbon Dioxide-Methane
11
AAD was 151.1 J/mol, the AD was 58.7 J/mol, the AAE was 15.0% and
the AE was –0.7%. The maximum difference was 969.8 J/mol at 9.1 MPa,
and 0.1 mole fraction nitrogen. The maximum error was 70% at the same
conditions as the maximum difference. Figure 1.9 shows the calculated and
experimentalenthalpiesfor the BWRmodel at 40°C.
The Hejmadi et al. (1971) nitrogen of excess enthalpies data had an AAD
of 26.1 J/mol, and AD of –11.0 J/mol, an AAE of 9.5% and an AE of –7.9%.
The maximum difference was 90.8 J/mol at 6.5 MPa, 31 °C, and 0.239 mole
fraction nitrogen. The maximum error was 14.1% at 3.4 MPa,40°C and
0.67 mole fraction nitrogen. As with the Barry et al. (1982) methane data,
the lower deviations are likely due to the lower pressure range used in the
measurements, as the highest pressure used was 6.5 MPa and the greatest
deviations typically occurred around 7–10 MPa for temperatures in the
30-40 -c range.
For the Peterson & Wilson (1974) methane data for enthalpy departures, the AAD was 56.4Jlmol, the AD was 26.2Jlmol, the AAE was 3.7%
andthe AE was 1.4%. Twopointswere omittedfrom the errorcalculations
4,000
1.01 MPa
2.03 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
11.15 MPa
12.16 MPa
3,500
11.1
Excess enthalpy (J/mol)
3,000
12.1
10.1
2,500
2,000
9.1
1,500
8.11
1,000
7.09
6.08
5.07
4.05
500
0
0
0.2
3.04
1.01
2.02
0.4
0.6
Mol fraction nitrogen (–)
0.8
1
Figure 1.9 Experimentaland calculatedenthalpiesat 40°C using BWR (Lee & Mather,
1970).
12 Carbon Dioxide Capture and Acid Gas Injection
due to BWR predicting a vapor/liquid mix. The Ng & Mather (1976) methane data for enthalpy departures had an AAD of 192.3 J/mol, an AD of
182.2 J/mol, an AAE of 3.8% and an AE of 3.0%
1.4.2
Lee-Kesler
The Lee & Mather (1972) methane data for excess enthalpies had an AAD
of 46.7 Jlmol, an AD of -43.2 [Zmol, an AAE of 20.1%, and an AE of
–19.7%. Figures 1.10 through 1.17 show the experimental and calculated
enthalpies for the 8 different temperatures. The greatest differences typically occurred at the highest pressure and at low methane mole fractions
for all temperatures. The maximum difference was 505.5 J/mol occurring
at 50 °C, 10.1 MPa and 0.1 mol fraction methane. The greatest errors always
occurred at a mole fraction of 0.1 and a pressure of 1.01 MPa. The maximum error was 98.0% occurring at 80 °C. For almost all data points, LK
overestimated the enthalpies. The only conditions where they were underestimated was at high methane mole fraction and high pressures.
For the Barry et al. (1970) methane data of excess enthalpies the AAD
was 12.0 Jlmol, the AD was -11.4 Jlmol, the AAE was 19.0%and the AE
600
4.36
500
Excess enthalpy (J/mol)
4.05
1.01 MPa
2.03 MPa
3.04 MPa
4.05 MPa
4.36 MPa
400
300
3.04
200
2.03
100
1.01
0
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.10 Experimental and calculated enthalpies at 10 °C using LK (Lee & Mather,
1972).
Enthalpies of Carbon Dioxide-Methane
700
5.07
600
Excess enthalpy (J/mol)
500
1.01 MPa
2.03 MPa
3.04 MPa
4.05 MPa
5.07 MPa
400
4.05
300
3.04
200
2.03
100
1.01
0
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.11 Experimental and calculated enthalpies at 20 °C using LK (Lee & Mather,
1972).
4,500
4,000
8.11
8.61
Excess enthalpy (J/mol)
3,500
9.12
10.13
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
8.61 MPa
9.12 MPa
10.13 MPa
3,000
2,500
2,000
1,500
7.09
1,000
6.08
5.07
500
0
0
0.2
4.05
1.01 2.02 3.04
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.12 Experimental and calculated enthalpies at 32 °C using LK (Lee & Mather,
1972).
13
Carbon Dioxide Capture and Acid Gas Injection
3,500
3,000
9.12
2,500
Excess enthalpy (J/mol)
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
11.15 MPa
10.13 11.15
2,000
1,500
8.11
1,000
7.09
500
0
0
0.2
6.08
5.07
4.05 3.04
2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.13 Experimentaland calculatedenthalpiesat 40°C using LK (Lee& Mather,
1972).
2,000
1,800
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
10.13
1,600
Excess enthalpy (J/mol)
1,400
9.12
1,200
1,000
800
8.11
600
7.09
6.08
400
200
0
0
0.2
5.07
4.05
3.04
2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.14 Experimental and calculated enthalpies at 50 °C using LK (Lee & Mather,
1972).
Enthalpies of Carbon Dioxide-Methane
1,200
10.13
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
1,000
9.12
Excess enthalpy (J/mol)
800
8.11
600
7.09
400
6.08
5.07
200
4.05
3.04
0
1.01
0
0.2
2.02
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.15 Experimental and calculated enthalpies at 60 °C using LK (Lee & Mather,
1972).
800
10.13
700
Excess enthalpy (J/mol)
600
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
9.12
500
8.11
400
7.09
300
6.08
5.07
200
4.05
3.04
100
2.02
0
1.01
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.16 Experimental and calculated enthalpies at 70 °C using LK (Lee & Mather,
1972).
15
16 Carbon Dioxide Capture and Acid Gas Injection
600
10.1
Excess enthalpy (J/mol)
500
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
9.1
400
8.1
7.0
300
6.0
200
5.0
4.0
100
3.0
2.0
0
1.0
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.17 Experimental and calculated enthalpies at 80 °C using LK (Lee & Mather,
1972).
was -16.6%. Themaximumdifference of 67.0 J/mol occurredat 4.6 MPa,
40°C, and a methane mole fraction of 0.649. The maximum error was
48.3% at 40 "C, 0.52 MPa,and 0.252 mole fractionmethane.
The Lee & Mather (1970) nitrogen data of excess enthalpies had an AAD
of231 l/mol, an AD of-226.4, an AAEof27.7% andan AE of -27.2%. The
maximum difference of 718.9 J/mol occurred at 12.16 MPa, and a methane
mole fraction of 0.2. The maximum errorwas 55.4% at 1.01 MPa,40°C,
0.1 mole fraction nitrogen. Figure 1.18 shows the calculated and experimentalenthalpiesfor the LKmodel at 40°C.
The excess enthalpy data for nitrogen from Hejmadi et al. (1971) had
an AD of –153.7 and an AE of –50.8%. All data points were overestimated
by LK, resulting in an AAD and AAE of the same magnitude as the AD
and AE. The maximum difference was 342.4 J/mol at 6.5 MPa, 31°C,
0.31 mole fraction nitrogen. The maximum error was 72.3% at 3.4 MPa,
31 °C, 0.228 mole fraction nitrogen.
Enthalpies of Carbon Dioxide-Methane
17
4,000
12.1
11.1
3,500
1.01 MPa
2.03 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
11.15 MPa
12.16 MPa
9.1
10.1
Excess enthalpy (J/mol)
3,000
2,500
8.11
2,000
1,500
7.09
1,000
6.08
5.07
4.05
500
0
0
0.2
3.04
2.02
1.01
0.4
0.6
Mol fraction nitrogen (–)
0.8
1
Figure 1.18 Experimentaland calculatedenthalpiesat 40°C using LK (Lee& Mather,
1970).
The Ng & Mather (1976) enthalpy departure data for methane had an
AAD of 152.8Jlmol, an AD of-151.37 Jlmol, an AAE of 4.0% and an AE
of -4.1 %. The only point where LKunderestimatedthe enthalpywas at
3 MPa,10°C and amethanemole fractionof 0.145. ThePeterson& Wilson
(1974) enthalpydeparturedata had an AAD of 149.3l/mol, and AD of
-145.9 Jlmol, and AAE of5.8% and an AE of -3.6%. The greatest
errorand
-45°C
°C
and
–20
°C.
difference occurredat temperaturesof
1.4.3
Soave-Redlich-Kwong
The Lee & Mather (1972) excess enthalpy methane data had an AAD
Figures 1.19 through 1.26 show the experimental and calculated excess
enthalpies as predicted by SRK for the different temperatures. The SRK
underestimated the excess enthalpies for the majority of the data points.
The greatest differences generally occurred at low methane mole fractions
was 98% occurring at a 1.01 MPa, 0.1 mole fraction methane and 50 °C.
18 Carbon Dioxide Capture and Acid Gas Injection
600
Excess enthalpy (J/mol)
500
400
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
4.36 MPa
4.36
300
4.05
200
3.04
100
2.03
1.01
0
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.19 Experimental and calculated enthalpies at 10 °C using SRK (Lee & Mather,
1972).
700
600
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
Excess enthalpy (J/mol)
500
5.07
400
300
4.05
200
3.04
100
2.03
0
1.01
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.20 Experimental and calculated enthalpies at 20 °C using SRK (Lee & Mather,
1972).
Enthalpies of Carbon Dioxide-Methane
19
4,500
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
8.61 MPa
9.12 MPa
10.13 MPa
4,000
Excess enthalpy (J/mol)
3,500
8.11
9.12
8.61
3,000
10.13
2,500
2,000
1,500
7.09
1,000
6.08
500
5.07
0
0
4.05
3.04 2.02
1.01
0.4
0.6
Mol fraction methane
0.2
0.8
1
Figure 1.21 Experimental and calculated enthalpies at 32 °C using SRK (Lee & Mather,
1972).
3,500
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
11.15 MPa
3,000
Excess enthalpy (J/mol)
2,500
10.13
11.15
9.12
2,000
1,500
8.11
1,000
7.09
500
4.05
0
0
0.2
6.08
5.07
3.04 2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.22 Experimentalandcalculatedenthalpiesat 40°C using SRK (Lee& Mather,
1972).
20 Carbon Dioxide Capture and Acid Gas Injection
2,000
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
1,800
Excess enthalpy (J/mol)
1,600
10.13
1,400
1,200
9.12
1,000
800
8.11
600
7.09
400
6.08
5.07
4.05
200
0
0
0.2
3.04 2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.23 Experimental and calculated enthalpies at 50 °C using SRK (Lee & Mather,
1972).
1,200
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
1,000
Excess enthalpy (J/mol)
10.13
800
9.12
600
8.11
400
7.09
6.08
200
0
5.07
0
0.2
4.05
3.04
2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.24 Experimental and calculated enthalpies at 60 °C using SRK (Lee & Mather,
1972).
Enthalpies of Carbon Dioxide-Methane
800
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
700
10.13
Excess enthalpy (J/mol)
600
500
9.12
400
8.11
300
7.09
6.08
200
5.07
4.05
3.04
100
0
0
0.2
2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.25 Experimental and calculated enthalpies at 70 °C using SRK (Lee & Mather,
1972).
600
Excess enthalpy (J/mol)
500
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
10.13
9.12
400
8.11
300
7.09
200
6.08
5.07
4.05
100
0
3.04
2.02
1.01
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.26 Experimental and calculated enthalpies at 80 °C using SRK (Lee & Mather,
1972).
21
22 Carbon Dioxide Capture and Acid Gas Injection
The Barry et al. (1982) methane data of excess enthalpy had an AAD
of 9.4 Jlmol an AD of 9.4 Jlmol an AAE of 22.9% and an AE of 22.8%.
The maximum difference of 22.3 Jlmol occurredat 4.6 MPa (the highest
pressure used), 40°C, and a methane mole fraction of 0.646. The maxi-mum errorof 54.8% occurredat 1.15 MPa,20°C and 0.883 mole fraction
methane.
The Lee & Mather (1970) nitrogen excess enthalpy data had an AAD
of 69.7 Jlmol, an AD of 61.3 Jlmol, an AAE of 12.4%, and an AE of
11.5%. Figure 1.27 shows the experimental and calculated enthalpies for
40°C using SRK. The maximum difference was 348.7 Jlmol occurring
at the same conditions as the Lee & Mather (1972) methane data. The
maximum error was 36.0% occurring at 9.12 MPa and 0.9 mol fraction
nitrogen.
For the Hejmadi et al. (1971) excess enthalpy nitrogen data, the AD was
45.2 Jlmol and the AE was 14.0%. Allthe enthalpieswere underestimated
4,000
1.01 MPa
2.03 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
11.15 MPa
12.16 MPa
3,500
12.16
3,000
Excess enthalpy (J/mol)
11.15
9.12
10.13
2,500
2,000
1,500
8.11
1,000
7.09
500
0
0
0.2
6.08
5.07
4.05
1.01
3.04
2.02
0.4
0.6
Mol fraction nitrogen (–)
0.8
1
Figure 1.27 Experimentaland calculatedenthalpiesat 40°C using SRK (Lee& Mather,
1970).
Enthalpies of Carbon Dioxide-Methane
23
by SRK, therefore the ADD and AAE were the same values as the AD and
AE. The maximum difference was 131.1 J/mol and occurred at 6.5 MPa,
31 °C, and 0.725 mole fraction nitrogen. The maximum error was 19.0%
occurringat 3.4 MPa,31°C and 0.729 mol fractionnitrogen.
For the enthalpy departure data for methane, the Ng & Mather (1976)
data had an AAD of 56.0 J/mol, an AD of –12.8 J/mol, an AAE of 2.5% and
an AEof 1.2%. The Peterson & Wilson (1974) datahadan AADof98.1, an
AD of -97.3 Jlmol, an AAEof3.7% and an AEof -4.8%.
1.4.4
Peng-Robinson
The Lee & Mather (1972) excess enthalpy methane data using PengRobinsonhad an AAD of 40.6 Jlmol, an AD of 36.3 Jlmol, a AAE of 9.9%
and an AE of 8.4%. Figures 1.28 through 1.35 show the calculated and
experimental enthalpies for Peng-Robinson. The maximum difference was
504.7 Jlmol occurringat a methane mole fraction of 0.1, 10.13 MPa and
600
Excess enthalpy (J/mol)
500
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
4.36 MPa
400
4.36
4.05
300
200
3.04
100
2.03
1.01
0
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.28 Experimental and calculated enthalpies at 10 °C using PR (Lee & Mather,
1972).
Carbon Dioxide Capture and Acid Gas Injection
700
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
600
Excess enthalpy (J/mol)
500
5.07
400
300
4.05
200
3.04
100
2.03
0
1.01
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.29 Experimental and calculated enthalpies at 20 °C using PR (Lee & Mather,
1972).
4,500
4,000
Excess enthalpy (J/mol)
3,500
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
8.61 MPa
9.12 MPa
10.13 MPa
8.11
9.12
8.61
10.13
3,000
2,500
2,000
1,500
1,000
7.09
500
6.08
5.07
4.05
0.2
0
0
3.04 2.02 1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.30 Experimental and calculated enthalpies at 30 °C using PR (Lee & Mather,
1972).
Enthalpies of Carbon Dioxide-Methane
3,500
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
11.15 MPa
3,000
Excess enthalpy (J/mol)
2,500
11.15
10.13
9.12
2,000
1,500
8.11
1,000
7.09
500
0
6.08
0
5.07
4.05 3.04
2.02 1.01
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.31 Experimentalandcalculatedenthalpiesat 40°C using PR (Lee& Mather,
1972).
2,000
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
1,800
1,600
Excess enthalpy (J/mol)
1,400
10.13
1,200
1,000
9.12
800
8.11
600
400
7.09
200
6.08
5.07
4.05 3.04
00
0.2
2.02 1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.32 Experimental and calculated enthalpies at 50 °C using PR (Lee & Mather,
1972).
25
26 Carbon Dioxide Capture and Acid Gas Injection
1,200
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
1,000
Excess enthalpy (J/mol)
10.13
800
9.12
600
8.11
400
7.09
6.08
200
0
5.07
0
0.2
4.05
3.04
2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.33 Experimental and calculated enthalpies at 60 °C using PR (Lee & Mather,
1972).
800
700
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
10.13
Excess enthalpy (J/mol)
600
9.12
500
400
8.11
300
7.09
6.08
200
5.07
4.05
100
0
3.04
1.01
0
0.2
2.02
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.34 Experimental and calculated enthalpies at 70 °C using PR (Lee & Mather,
1972).
Enthalpies of Carbon Dioxide-Methane
27
Excess enthalpy (J/mol)
600
500
10.13
400
9.12
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
8.11
300
7.09
200
6.08
5.07
100
0
4.05
3.04
2.02
1.01
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.35 Experimental and calculated enthalpies at 80 °C using PR (Lee & Mather,
1972).
40 "C. The maximumerror of 39.6% occurredat 0.1 mole fraction methane, 10.13 MPa, and 50 °C. The average errors in enthalpies decreased as
the temperatures increased, as well as with increasing pressure.
The Barry et al. (1982) methane excess enthalpy data had an AAD of
9.6, and AD of 9.5, an AAE of 23.2 and an AE of 23.2. The maximum difference was 22.3 J/mol at 4.6 MPa and 0.686 mole fraction methane.The
maximum error was 54.6% at 0.521 MPa, 20 "C, and 0.816 mole fraction
methane.
For the Lee & Mather (1970) excess enthalpy data for nitrogen, the AAD
was 79.4 J/mol, the AD was 73.3 J/mol, the AAE was 13.3%andthe AE was
12.6%. The maximum difference, of 378.6 J/mol, and the maximum error
of 38.9% occurred at pressures of 10.13 MPa and 9.12 MPa respectively.
Figure 1.36 shows the experimentalandcalculatedenthalpiesfor the 40 "C
nitrogen mixture.
The Hejmadi et al. (1971) excess enthalpy data for nitrogen had an AD
of45.6 J/mol andan AEof 14.3%. All the datapointswere underestimated
by PR78, therefore the ADD and AE were the same as the AD and AE.
The maximum difference of 127.9 J/mol occurred at 6.5 MPa, 31 °C, and a
28 Carbon Dioxide Capture and Acid Gas Injection
4,000
1.01 MPa
2.03 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
11.15 MPa
12.16 MPa
3,500
12.16
Excess enthalpy (J/mol)
3,000
2,500
11.1
10.13
9.12
2,000
1,500
8.11
1,000
7.09
500
0
0
0.2
6.08
5.07
4.05
1.01
3.04 2.02
0.4
0.6
Mol fraction nitrogen (–)
0.8
1
Figure 1.36 Experimentaland calculatedenthalpiesat 40°C using PR (Lee& Mather,
1970).
nitrogenmole fractionof0.239. The maximumerrorwas 19.5% at 3.4 MPa,
31 °C, and a mole fraction of 0.729 nitrogen.
For the enthalpy departure data, the Ng & Mather (1976) had an AAD
of 110.3 J/mol, an AD of -82.0 J/mol, a AAE of 5.2% and an AEof -4.4%.
The Peterson&Wilson (1974) datahad an AAD of 160.2 J/mo!, an AD of
160.2 J/mol, and AAE of 9.7% and an AE of –9.7%.
1.4.5
AQUAlibrium
The Lee & Mather (1972) excess enthalpy methane data had an AAD of
39.2 J/mol, an AD of 34.7 J/mol, an AAE of 9.8% and an AE of 8.3%.
Figures 1.37 through 1.44 show the experimentaland calculatedenthal--
pies for the different temperatures using AQUAlibrium. The maximum
difference was 512.2 J/mol at 11.15 MPa, 40°C, and 0.1 mole fraction
methane.The maximumerrorwas 41.8 at 10.13 MPa,50°C, and0.5 mole
fraction methane. The difference in enthalpies decreased as the temperatures increased.
For the Barry et al. (1982) methane data, the AAD was 9.6 J/mol, the AD
was 9.1 J/mol, the AAE was 23.4% and the AE was 22.1%. Themaximum
Enthalpies of Carbon Dioxide-Methane
29
600
Excess enthalpy (J/mol)
500
400
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
4.36 MPa
4.36
4.05
300
200
3.04
100
2.03
1.01
0
0
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.37 Experimental and calculated enthalpies at 10 °C using AQUAlibrium (Lee &
Mather, 1972).
700
600
Excess enthalpy (J/mol)
500
5.07
400
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
300
4.05
200
3.04
100
0
2.03
0
0.2
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.38 Experimental and calculated enthalpies at 20 °C using AQUAlibrium (Lee &
Mather, 1972).
30 Carbon Dioxide Capture and Acid Gas Injection
4,500
4,000
Excess enthalpy (J/mol)
3,500
3,000
8.11
8.61
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
8.61 MPa
9.12 MPa
10.13 MPa
9.12
10.13
2,500
2,000
1,500
7.09
1,000
500
00
6.08
5.07
4.05 3.04 2.02 1.01
0.2
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.39 Experimental and calculated enthalpies at 30 °C using AQUAlibrium (Lee &
Mather, 1972).
3,500
3,000
Excess enthalpy (J/mol)
2,500
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
11.15 MPa
10.13 11.15
9.12
2,000
1,500
8.11
1,000
7.09
500
0
0
6.08
0.2
5.07
4.05
1.01 3.04 2.02
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.40 Experimentaland calculatedenthalpiesat 40°C using AQUAlibrium(Lee &
Mather, 1972).
Enthalpies of Carbon Dioxide-Methane
31
2,000
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
1,800
1,600
10.13
Excess enthalpy (J/mol)
1,400
1,200
9.12
1,000
800
8.11
600
7.09
400
6.08
200
0
5.07
0
0.2
4.05
3.04
2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.41 Experimental and calculated enthalpies at 50 °C using AQUAlibrium (Lee &
Mather, 1972).
1,200
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
Excess enthalpy (J/mol)
1,000
10.13
800
9.12
600
8.11
400
7.09
6.08
200
0
5.07
0
0.2
3.04
1.01
4.05
2.02
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.42 Experimental and calculated enthalpies at 60 °C using AQUAlibrium (Lee &
Mather, 1972).
32 Carbon Dioxide Capture and Acid Gas Injection
800
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
700
10.13
Excess enthalpy (J/mol)
600
500
9.12
400
8.11
300
7.09
6.08
200
5.07
4.05
3.04
100
0
0
0.2
1.01
2.02
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.43 Experimental and calculated enthalpies at 70 °C using AQUAlibrium (Lee &
Mather, 1972).
Excess enthalpy (J/mol)
600
500
10.13
400
9.12
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
8.11
300
7.09
200
6.08
5.07
4.05
100
0
0
0.2
3.04
2.02
1.01
0.4
0.6
Mol fraction methane
0.8
1
Figure 1.44 Experimental and calculated enthalpies at 80 °C using AQUAlibrium (Lee &
Mather, 1972).
Enthalpies of Carbon Dioxide-Methane
33
difference was 22.3Jlmol at 4.6 MPa,40°C, and0.686 mole fractionmeth-ane. The maximum error was 54.6% at 0.52 MPa,20°C, and 0.477 mole
fraction methane.
The Lee & Mather (1970) nitrogen data had an AAD of 62.2, and AD of
62.2, an AAE of 13.0%,a nd an AE of 11.9%. Figure 1.45 shows theexperi-mentalandcalculatedenthalpiesfor the 40°C mixture.The maximumdif-ference was 368.6 J/mol at 10.13 MPa and 0.1 mole fraction nitrogen. The
maximum error was 38.0% at 1.01 MPa and 0.9 mole fraction nitrogen.
The Hejmadi et al. (1971) data for nitrogen had an AAD and AD of
44.1 Jlmol and an AAEand AE of 14.1%. The maximum difference was
119.2 J/mol at 6.5 MPa, 31 °C, and 0.239 mole fraction nitrogen. The maximum errorwas 19.5% at 3.4 MPa,31°C, and0.729 mole fractionnitrogen.
The Ng & Mather (1976) enthalpy departure data for methane had an
AAD of 116.0 Jlmol, an ADof -94.2 J/mo!, an AAEof 5.4% and an AE of
4.8%. The Peterson & Wilson (1974) enthalpydeparturedatahad an AAD
of 163.7 J/mol, an AD of –163.7, an AAE of 10.0% and an AE of 10.0%.
1.5
Discussion
Table 1.1 and Table 1.2 show the AAD and AAE for all excess enthalpy
data for all thermodynamic models, as well as a weighted average, based
on the number of data points used, of all mixtures for each model. For the
excess enthalpy data the SRK model provided the best overall AAE. The
AQUAlibrium model provided the best overall AAD with SRK obtaining
similar results. Overall, SRK, PR78 and AQUAlibrium all achieved similar
results, and predicted better than both LK and BWR, with LK being the
less accurate of the two.
Table 1.2 Absolute average difference in excess enthalpies for methane and
nitrogen mixtures using the different models.
BWR
LK
SRK
PR78
AQUA
78.1
46.7
38.6
40.6
39.2
9.1
12.0
9.4
9.5
9.6
Lee & Mather (1970)- Nitrogen
151.1
231.0
69.7
79.4
62.2
Hejmadi et al. (1971)- Nitrogen
26.1
153.7
45.2
45.6
44.1
Weighted Average
81.0
71.3
40.8
43.5
40.2
Lee & Mather (1972)- Methane
Barry et al. (1982)- Methane
34
Carbon Dioxide Capture and Acid Gas Injection
For the AAE, both the Lee & Mather (1970), (1972) data sets were best
predicted by SRK, while the Hejmadi et al. (1971) and Barry et al. (1982)
data were best predicted by BWR. The Hejmadi et al. (1971) and Barry
et al. (1982) data were taken at much smaller ranges of pressure, with maximums of 6.5 and 4.6 MParespectively, comparedto a maximumpressure
of 12.16 MPa and 11.15 of the Lee & Mather (1970), (1972) data, respec-
tively. In Figures1.3 and 1.4, showingthe predictionsfor BWR at32°C and
40 "C respectively, it can beseen thatwhen the enthalpyis changingrapidly
with increasing pressure, the BWR model provides very poor predictions.
This caused the greater error in the Lee and Mather (1970), (1972) data
compared to the Hejmadi et al. (1971) and Barry et al. (1982) data, where
pressures where rapid enthalpy change was happening were not measured.
Over wide ranges of pressures, the SRK model provided the more accurate
predictions.
For the AAD, the same trend was found as for the AAE, except for the
optimal model for the Lee & Mather (1970) data being the AQUAlibrium
model. Compared to the Lee & Mather (1972) data, the 1970 data was only
measured at single temperatureof 40°C. When comparing Figure 1.22
showing SRK at 40°C for Lee & Mather (1972) to Figure 1.45 showing
4,000
1.01 MPa
2.02 MPa
3.04 MPa
4.05 MPa
5.07 MPa
6.08 MPa
7.09 MPa
8.11 MPa
9.12 MPa
10.13 MPa
11.15 MPa
12.16 MPa
3,500
Excess enthalpy (J/mol)
3,000
2,500
11.15
10.13
12.16
9.12
2,000
1,500
8.11
1,000
7.09
500
0
0
0.2
6.08
5.07
4.05 3.04
1.01
2.02
0.4
0.6
Mol fraction nitrogen (–)
0.8
1
Figure 1.45 Experimentaland calculatedenthalpiesat 40°C using AQUAlibrium(Lee &
Mather, 1970).
Enthalpies of Carbon Dioxide-Methane
35
AQUA for Lee & Mather (1970), the AQUAlibrium model predicts the
excess enthalpy much more accurately, specifically at pressures of 9.12 MPa.
However, when comparingFigure 1.45 to Figure 1.40,b oth of which are
AQUAlibriumat 40°C, with Figure 1.45being for nitrogenandFigure 1.40
being for methane, the nitrogen mixture is much more accurate, whereas
Figure 1.45 closely resembles Figure 1.22. Therefore, it is likelythat the
optimal model for the Lee & Mather (1970) was AQUAlibrium due to the
impurity being nitrogen rather than methane.
The Lee & Mather (1972) data allows for a comparison of the accuracy
of each model at temperatures from 10 °C to 80 °C. For all models except
for LK,32°C and 40 °C gave the largest differences. For LK, thelargest
differences were at 32 °C and 50 °C. For SRK, PR78 and AQUAlibrium,
60 °C, 70 °C, and 80 °C gave smaller differences than the low temperatures
of 10 °C, and 20 °C. For BWR, the opposite trend occurred with the low
temperatureshaving smaller differences. For LK,1 0°C, 20 °C, 40°C, and
60–80 °C all had similar differences.
Tables 1.3and 1.4 show the AADand AAE for theenthalpydeparture
data for all 5 thermodynamic models as well as a weighted average of the
Table 1.3 Absolute average error in excess enthalpies for methane and nitrogen
mixtures using the different models.
BWR
LK
SRK
PR78
AQUA
Lee & Mather (1972)- Methane
19.0
20.1
9.7
9.9
9.8
Barry et al. (1982)- Methane
14.3
19.1
22.9
23.2
23.4
Lee & Mather (1970)- Nitrogen
15.0
27.7
12.4
13.3
13.0
Hejmadi et al. (1971)- Nitrogen
9.5
50.8
14.0
14.3
14.1
17.9
22.0
11.1
11.4
11.3
Weighted Average
Table 1.4 Absolute average difference in enthalpy departure for methane
­mixtures using the different models.
BWR
LK
SRK
PR78
AQUA
Ng & Mather (1976)
3.8
4.1
2.5
5.2
5.4
Peterson & Wilson (1974)
3.7
5.8
4.5
9.7
10.0
Weighted Average
3.8
5.0
3.5
7.5
7.8
36 Carbon Dioxide Capture and Acid Gas Injection
Table 1.5 Absolute average error in enthalpy departure for methane mixtures
using the different models.
Ng & Mather (1976)
Peterson & Wilson (1974)
Weighted Average
BWR
LK
SRK
PR78
AQUA
192.3
152.8
56.0
110.3
116.0
56.4
149.3
98.1
160.2
163.7
121.6
151.0
77.9
136.2
140.8
two mixtures for all models. As with the excess enthalpy data, the SRK
model performed well, with SRK giving the smallest overall AAE and
AAD. However, unlike the excess enthalpies, the PR78 and AQUAlibrium
models were much worse than SRK. For the AAD, PR78, AQUAlibrium,
BWR and LK all gave similar values, of which were considerably greater
than SRK. For the AAE, SRK had the smallest value, closely followed by
BWR, with PR78 and AQUAlibrium having the highest AAE.
Using both the AAE and AAD as criteria, Ng & Mather (1976) was
best predicted by SRK,while Peterson & Wilson (1974) was best pre-dicted by BWR. The Ng & Mather (1976) data may have been better
predicted by SRK compared to BWR because BWR was less accurate
at predicting high pressure enthalpies, and while both papers had similar pressure ranges, the Ng & Mather (1976) paper had more data at
higher pressures. Additionally, the method used to calculate the enthalpy
departures varied between the two methods, with Peterson & Wilson
(1974) using a reference enthalpyof 0.14 MPa, and Ng & Mather(1976)
uses the ideal gas enthalpy as the reference enthalpy. The difference in
calculation methods may affect the AAE and AAD comparison between
data sets.
1.6
Conclusion
The excess enthalpy data and the enthalpy departure data were overall predicted most accurately by the SRK model, with an exception of
AQUAlibrium giving the smallest AAD for excess enthalpies. For the
excess enthalpy data, the AQUAlibrium, SRK, and PR78 models all produced similar results, with BWR and LK giving much higher errors, with
LK giving the higher of the two. For the enthalpy departure data, when
using AAE as the criteria, LK, BWR and SRK all gave similar results, with
PR78 and AQUA giving much greater errors. When looking at AAD as the
Enthalpies of Carbon Dioxide-Methane
37
criteria, SRK was by far the best, with the other four giving comparative
results.
For the excess enthalpy data, the optimal methods of SRK for AAE and
AQUAlibrium for AAD, gave average errors of 11.1% and average differences of40.2 J/mo!. For use in acid gasinjection,a difference of40.2 Jlmol
equates to 0.75 HP/ MMSCFD. For enthalpy departures, SRK gave an overall average error of 3.5% and an average difference of 77.9 J/mol, or 1.5 HP/
MMSCFD. A difference of 0.75- 1.5 HP/MMSCFD is an acceptable margin
of error when considering the design of a compressor; however, the maximum difference for the SRKmodel was 504.4 Jlmol, or 9.4 HP/MMSCFD
under certain conditions, which may pose a problem in the compression
of the stream.
References
1. Barry, A., Kallaguine, S., and R. Ramalho, “Direct Determination of Enthalpy
of Mixing for the Binary Gaseous System Methane-Carbon Dioxide by an
Isothermal Flow Calorimeter,” J. Chem. Eng. Data 27, 258-264, 1982.
2. Hejmadi, A.V., Katz, D.L., and J.E. Powers, “Experimental Determination of the
Enthalpy of Mixing of N2 + CO2 Under Pressure,” J. Chem. Thermo., 3,483-496,
1971.
3. Lee, J.I., and Mather, A.E., “The Excess Enthalpy of Gaseous Mixtures of
Nitrogen and Carbon Dioxide,” J. Chem. Thermo., 2, 881–895, 1970.
Lee, J.I., and Mather, A.E., “The Excess Enthalpy of Gaseous Mixtures of Carbon
Dioxide with Methane,” Can. J. Chem. Eng., 50, 95–100, 1972.
5. Ng, H.J., and Mather, A.E., “Isothermal Joule-Thomson Coefficients in Mixtures
of Methane and Carbon Dioxide” J. Chem. Eng. Data, 21,291-294,1976.
6. Peterson, J.M., and Wilson, G.M., “Enthalpy and Phase Boundary Measurements
on Carbon Dioxide and Mixtures of Carbon Dioxide with Methane, Ethane
and HydrogenSulfide:' BrighamYoung University, Provo, Utah,
1974.
2
Enthalpies of Hydrogen SulfideMethane Mixture: Comparison with
Thermodynamic Models
Erin L. Roberts and John J. Carroll
Gas Liquids Engineering, Calgary, AB, Canada
Abstract
In the design of an acid gas injection scheme, the physical properties of the stream
are required to ensure successful injection into the subsurface reservoir. Common
impurities of a carbon dioxide acid gas injection stream are hydrogen sulfide and
methane. The excess enthalpies of these hydrogen sulfide-methane mixtures are
important in determining the compressor specifications in the acid gas injection
design to ensure proper injection to the subsurface reservoir.
This study compares the experimental data of excess enthalpies of a hydrogen
sulfide-methane mixture to the calculated excess enthalpies of six different thermodynamic models, Lee Kesler, Benedict-Webb-Rubin, Soave-Redlich-Kwong,
Redlich-Kwong, Peng-Robinson, and AQUAlibrium.
All models were found to have considerable error when predicting excess
enthalpies. The best model was Lee-Kesler with average absolute errors of 22.5%
and absolute average differences of 22.4 J/mol.
2.1
Introduction
Stricter regulations placed on the natural gas industry around carbon dioxide emissions have led to the disposal of the carbon dioxide by acid gas
injection to be a more favorable option. Common impurities in the carbon dioxide stream are methane and hydrogen sulfide. Traditionally, the
hydrogen sulfide was converted to elemental sulfur by the Claus process.
However, due to a decrease in demand of sulfur, injection into subsurface
reservoirs has become the more economical procedure for many gas plants.
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (39–54)
2017 © Scrivener Publishing LLC
39
40 Carbon Dioxide Capture and Acid Gas Injection
The excess enthalpies of these hydrogen sulfide-methane mixtures
are required for the design of the acid gas injection scheme. This study
uses six different thermodynamic models to predict the excess enthalpy
of a hydrogen sulfide-methane mixture. Five different equations of state,
Benedict-Webb-Rubin (BWR), Lee-Kesler (LK), Soave-Redlich-Kwong
(SRK), Redlich-Kwong (RK), and Peng-Robinson (1978) were used with
VMGSim software, as well the AQUAlibrium model. BWR and LK are
multi-constant equations, and SRK, RK and PR78 are cubic equations of
state. The AQUAlibrium model uses a variation of Peng-Robinson.
2.2
Enthalpy
Typically the enthalpy is expressed as a molar enthalpy, measured in J/mol,
or a mass enthalpy, measure in J/g. Other common units used for enthalpies include BTU/lb or BTU/lbmol.
The enthalpy of mixtures can be determined in a number of ways. One
method is to use excess enthalpy (enthalpy of mixing). Excess enthalpy is
defined as
HE
Hm
i
xi H i
(2.1)
where: HE – Excess enthalpy
Hm – Enthalpy of mixture
Hi – Enthalpy of component i
xi – mol fraction of component i
2.3
Literature Review
A review of the literature was performed to find all experimental data for
the enthalpies of the binary system of hydrogen sulfide and methane. Only
one experimental data set was found for this binary system, performed by
Barry et al. (1982). The enthalpy data was in the range of 0.18 to 0.85 mol
fraction hydrogen sulfide, taken at nominal pressures of 0.507 MPa,
1.013 MPa, and 1.52 MPa, and at nominal temperatures of 293.15 K,
305.15 K and 313.15 K. Only the nominal temperatures and pressures were
reported in the data. Another experimental data set was found, also by
Barry et al. (1983), but used a ternary system of carbon dioxide, hydrogen
Enthalpies of Hydrogen Sulfide-Methane Mixture 41
sulfide and methane. The data was at the same nominal temperatures and
pressures, and at a wide range of mol fractions.
2.4
Calculations
The experimental enthalpies from Barry et al. (1982) was compared to
calculated enthalpies calculated using Lee-Kesler, Benedict-Webb-Rubin,
Soave-Redlich-Kwong, Redlich-Kwong and Peng-Robinson thermodynamic models from VMGSim as well as AQUAlibrium software.
Four different error functions were used to compare the different thermodynamic models. The absolute average difference in excess enthalpies
(AAD) was defined as;
1
NP
AAD
E
E
H exp
H calc
(2.2)
NP– number of points
HEexp– experimental excess enthalpy
HEcalc– calculated excess enthalpy
and the average difference (AD) was defined as:
where:
AD
1
NP
E
E
H exp
H calc
(2.3)
The absolute average error (AAE) in excess enthalpies was defined as:
AAE
1
NP
E
E
H exp
H calc
E
H calc
100%
(2.4)
and the average error (AE) was defined as:
AE
2.4.1
1
NP
E
E
H exp
H calc
E
H calc
(2.5)
100%
Lee-Kesler
The Lee-Kesler model gave an AAD of 22.5 J/mol, an AD of 22.4 J/mol with
a maximum enthalpy difference of 60.3 J/mol occurring at 1.52 MPa and
42 Carbon Dioxide Capture and Acid Gas Injection
293.15 K. The AAE was 25.2%, and the AE was 24.8% with a maximum
error of 43.5% at 1.52 MPa and 293.15 K. Figures 2.1 through 2.3 show the
experimental excess enthalpies and calculated excess enthalpies using LK
for 293.15 K, 305.15 K, and 313.15 K. The errors and differences greatly
increased with increasing mole fraction of methane in the mixture, with
errors averaging around 15% for mole fractions from 0.2 to 0.3 methane,
200
0.507 MPa
1.013 MPa
1.52 MPa
180
Excess enthalpy (J/mol)
160
1.52
140
120
100
1.013
80
60
40
0.507
20
0
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.1 Experimental and calculated enthalpies at 293.15 K using Lee-Kesler (Barry
et al., 1982).
180
0.507 MPa
1.013 MPa
1.52 MPa
160
Excess enthalpy (J/mol)
140
1.52
120
100
80
1.013
60
40
0.507
20
0
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.2 Experimental and calculated enthalpies at 305.15 K using Lee-Kesler (Barry
et al., 1982).
Enthalpies of Hydrogen Sulfide-Methane Mixture 43
140
0.507 MPa
1.013 MPa
1.52 MPa
120
Excess enthalpy (J/mol)
1.52
100
80
1.013
60
40
0.507
20
0
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.3 Experimental and calculated enthalpies at 313.15 K using Lee-Kesler (Barry
et al., 1982).
and errors of around 40% for methane mole fractions of 0.8. The errors and
differences generally increased with increasing temperature and increasing
pressure, but only by 1–2% between the lowest and highest values.
2.4.2
Benedict-Webb-Rubin
The BWR model gave an AAD and an AD of 43.9 J/mol with a maximum
difference of 96.5 J/mol at 1.52 MPa and 293.15 K. The AAE and AE was
49.9% with a maximum error of 57.6% at 1.52 MPa and 293.15. Figures 2.4
through 2.6 show the experimental excess enthalpies and calculated enthalpies using BWR for the three different temperatures. Unlike the Lee-Kesler
model, there was no significant change in error with increasing mole fraction of methane, with all mole fractions having errors around 50%. There
was generally a small increase in error and difference with increasing pressure and temperature, but only by about 1–3% between the highest and
lowest values.
2.4.3
Soave-Redlich-Kwong
The SRK model gave an AAD and AD of 49.6 J/mol with a maximum
of 110.6 J/mol at a pressure of 1.52 MPa and a temperature of 293.15 K.
44 Carbon Dioxide Capture and Acid Gas Injection
200
0.507 MPa
1.013 MPa
1.52 MPa
180
Excess enthalpy (J/mol)
160
140
120
100
1.52
80
60
1.013
40
20
0
0.507
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.4 Experimental and calculated enthalpies at 293.15 K using Benedict-WebbRubin (Barry et al., 1982).
180
0.507 MPa
1.013 MPa
1.52 MPa
160
Excess enthalpy (J/mol)
140
120
100
1.52
80
60
1.013
40
0.507
20
0
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.5 Experimental and calculated enthalpies at 305.15 K using Benedict-WebbRubin (Barry et al., 1982).
Enthalpies of Hydrogen Sulfide-Methane Mixture 45
140
0.507 MPa
1.013 MPa
1.52 MPa
Excess eanthalpy (J/mol)
120
100
80
1.52
60
1.013
40
20
0
0
0.507
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.6 Experimental and calculated enthalpies at 313.15 K using Benedict-WebbRubin (Barry et al., 1982).
200
0.507 MPa
1.013 MPa
1.52 MPa
180
Excess enthalpy (J/mol)
160
140
120
100
1.52
80
60
1.013
40
0.507
20
0
0
0.2
0.4
0.6
0.8
Mol fraction fydrogen sulfide
1
Figure 2.7 Experimental and calculated enthalpies at 293.15 K using Soave-RedlichKwong (Barry et al., 1982).
The AAE and AE were both 57.2% with a maximum error of 65.1% at
0.507 MPa and 293.15 K. Figures 2.7 through 2.9 show the experimental
excess enthalpies and calculated excess enthalpies using SRK for the three
different temperatures. The mole fraction of methane in the mixture had
little effect on the errors and difference, as with the BWR model. There
was also very little difference in errors with changing pressure. However,
46 Carbon Dioxide Capture and Acid Gas Injection
180
0.507 MPa
1.013 MPa
1.52 MPa
160
Excess enthalpy (J/mol)
140
120
100
80
1.52
60
1.013
40
20
0
0.507
0
0.2
0.4
0.6
0.8
Mol faraction hydrogen sulfide
1
Figure 2.8 Experimental and calculated enthalpies at 305.15 K using Soave-RedlichKwong (Barry et al., 1982).
140
0.507 MPa
1.013 MPa
1.52 MPa
Excess enthalpy (J/mol)
120
100
80
1.52
60
40
1.013
20
0
0.507
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.9 Experimental and calculated enthalpies at 313.15 K using Soave-RedlichKwong (Barry et al., 1982).
Enthalpies of Hydrogen Sulfide-Methane Mixture 47
the errors and differences decreased slightly with increasing temperatures,
which was different than the LK and BWR models, though the difference
in errors was only about 3% between the lowest and highest values.
2.4.4
Redlich-Kwong
The RK model gave an AAD and AD of 62.1 J/mol with a maximum difference of 135.6 J/mol at 1.52 MPa and 293.15 K. The AAE and AE were both
72.1% with a maximum of 77.8% at 0.507 MPa and 293.15 K. Figures 2.10
through 2.12 show the experimental excess enthalpies and the excess
enthalpies calculated by RK for the three different temperatures. There was
a significant effect of methane mole fraction on the errors and differences,
with methane mole fractions around 0.2 having errors of 68% and methane
mole fractions around 0.8 having errors of around 78%. Unlike the LK,
BWR, and SRK model, for the RK model, errors decreased with increasing pressure, but only by about 3% between the highest and lowest values.
There was very little difference in errors with changing temperature.
2.4.5
Peng-Robinson
The AAD and AD for the PR78 model was 50.5 J/mol with a maximum
difference of 112.4 J/mol at 1.52 MPa and 293.15 K. The AAE and AE was
200
0.507 MPa
1.013 MPa
1.52 MPa
180
Excess enthalpy (J/mol)
160
140
120
100
80
60
1.52
40
1.01
20
0
0.507
0
0.2
0.4
0.6
Mol fraction hydrogen sulfide
0.8
1
Figure 2.10 Experimental and calculated enthalpies at 293.15 K using Redlich-Kwong
(Barry et al., 1982).
48 Carbon Dioxide Capture and Acid Gas Injection
180
0.507 MPa
1.013 MPa
1.52 MPa
160
Excess enthalpy (J/mol)
140
120
100
80
60
1.52
40
1.013
20
0.507
0
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.11 Experimental and calculated enthalpies at 305.15 K using Redlich-Kwong
(Barry et al., 1982).
140
0.507 MPa
1.013 MPa
1.52 MPa
Excess enthalpy (J/mol)
120
100
80
60
1.52
40
1.013
20
0.507
0
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.12 Experimental and calculated enthalpies at 313.15 K using Redlich-Kwong
(Barry et al., 1982).
Enthalpies of Hydrogen Sulfide-Methane Mixture 49
58.2% with a maximum of 66% at 0.507 MPa, and 293.15 K. Figures 2.13
through 2.15 show the experimental excess enthalpies and calculated excess
enthalpies using PR78 for the three different temperatures. The errors and
differences decreased by about 3% between the lowest temperature and
the highest temperature. Pressure did not affect the errors and differences,
200
0.507 MPa
1.013 MPa
1.52 MPa
180
Excess enthalpy (J/mol)
160
140
120
100
80
1.52
60
1.013
40
20
0.507
0
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.13 Experimental and calculated enthalpies at 293.15 K using Peng-Robinson
(Barry et al., 1982).
180
0.507 MPa
1.013 MPa
1.52 MPa
160
Excess enthalpy (J/mol)
140
120
100
80
1.52
60
1.013
40
20
0
0.507
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.14 Experimental and calculated enthalpies at 305.15 K using Peng-Robinson
(Barry et al., 1982).
50 Carbon Dioxide Capture and Acid Gas Injection
140
0.507 MPa
1.013 MPa
1.52 MPa
Excess enthalpy (J/mol)
120
100
80
1.52
60
40
1.013
20
0
0.507
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.15 Experimental and calculated enthalpies at 313.15 K using Peng-Robinson
(Barry et al., 1982).
and the methane mole fraction had only a small affect, with slightly greater
errors and differences with increasing methane mole fraction.
2.4.6
AQUAlibrium
The AQUAlibrium model had very similar results to the Peng-Robinson
model with an AAD and AD of 49.2 with a maximum difference of
109.9 J/mol at 1.52 MPa, and 293.15 K. The AAE and AE were 56.9% with a
maximum error of 64.9% at 0.507 MPa and 293.15 K. Figures 2.15 through
2.18 show the experimental excess enthalpies and calculated excess enthalpies using AQUAlibrium for the three different temperatures. As with the
Peng-Robinson model, the errors and difference decreased by about 3%
between the lowest and highest temperatures, and pressure and mole fraction did not have a significant effect.
2.5
Discussion
For predicting enthalpies, the six models all had considerable errors and
differences. The best model was LK, which predicted excess enthalpies
Enthalpies of Hydrogen Sulfide-Methane Mixture 51
200
0.507 MPa
1.013 MPa
1.52 MPa
180
Excess enthalpy (J/mol)
160
140
120
100
80
60
1.52
1.013
40
20
0
0
0.507
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.16 Experimental and calculated enthalpies at 293.15 K using AQUAlibrium
(Barry et al., 1982).
180
160
0.507 MPa
1.013 MPa
1.52 MPa
Excess enthalpy (J/mol)
140
120
100
80
1.52
60
40
1.01
20
0.507
0
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.17 Experimental and calculated enthalpies at 305.15 K using AQUAlibrium
(Barry et al., 1982).
within 25% or 22.5 J/mol, on average. The RK model had the highest
errors and differences with an average of 72.1% or 62.1 J/mol. The other
four models, BWR, SRK, PR78, and AQUAlibrium all had absolute average
errors around 50-60% and differences of 55-65 J/mol.
Both the LK and BWR models increased in error in increasing temperature. However, the SRK, PR78 and AQUAlibrium models all decreased
52 Carbon Dioxide Capture and Acid Gas Injection
140
0.507 MPa
1.013 MPa
1.52 MPa
Excess enthalpy (J/mol)
120
100
80
1.52
60
40
1.013
20
0
0.507
0
0.2
0.4
0.6
0.8
Mol fraction hydrogen sulfide
1
Figure 2.18 Experimental and calculated enthalpies at 313.15 K using AQUAlibrium
(Barry et al., 1982).
in error with increasing temperature. A change in temperature had little
effect on the errors for the RK model.
Similar to the trend with temperature, the LK and BWR models both
increased in error with an increase in pressure. The RK model decreased in
error with an increase in pressure. A change in pressure had little effect on
the SRK, PR78, and AQUAlibrium models.
For all models, an increase in mole fraction of methane increased the
error and difference. However, this effect was almost negligible in the
BWR, SRK, PR78 and AQUAlibrium models. A significant increase in
error with increasing mole fraction of methane occurred with the LK and
BWR models.
LK was found to be the best model, with an AAE of 25%, predicting
values 25–50% more accurate than the other five models. The errors were
greatly affected by the methane mole fraction in the mixture, as seen in
Figure 2.19. The errors varied from 2% at very low fractions to almost 45%
at very high fractions. If only data consisting of less than 0.3 mole fraction
methane is considered, the error of the LK model is reduced to 15%, or
16.1 J/mol on average.
2.6
Conclusion
Table 2.1 shows the various error for all enthalpy data for all thermodynamic models. The six different models, LK, BWR, SRK, RK, PR78 and
AQUAlibrium all lacked the ability to provide accurate predictions of the
Enthalpies of Hydrogen Sulfide-Methane Mixture 53
50
Absolute excess enthalpy error (%)
45
40
35
30
25
20
15
293.15 K
305.15 K
313.15
10
5
0
0
0.2
0.4
0.6
Mole fraction methane
0.8
1
Figure 2.19 Effect of mole fraction of methane on the error in excess enthalpy prediction
for Lee-Kelser at 293.15 K, 305.15 K, and 313.15 K.
Table 2.1 AAD, AAE, maximum difference and maximum error using the six
different thermodynamic models for the hydrogen sulfide-methane mixture
Model
AAD (J/mol)
LK
22.5
BWR
SRK
Max difference (J/mol) AAE (%) Max error (%)
60.3
25.2
43.5
43.9
96.5
49.9
57.6
49.6
110.6
57.2
65.1
RK
62.1
135.6
72.1
77.8
PR78
50.5
112.4
58.2
66.0
AQUA
49.2
109.9
56.9
64.9
experimental excess enthalpies of Barry et al. (1982). The best model was
LK, with an AAE of 25%, and an AAD of 22.5 J/mol. The RK model produced the least accurate prediction with an AAE of 72.1% and an AAD of
62.1 J/mol.
The pressure and temperature had a small effect on many of the models, with LK and BWR increasing in error with both increasing temperature and pressure. SRK, PR78 and AQUAlibrium decreased in error
with increasing temperature and the RK model decreased in error with
54 Carbon Dioxide Capture and Acid Gas Injection
increasing pressure. All models had an increase in error with an increase
in mole fraction; however, this effect was only significant for the LK and
RK models.
The most accurate predictions were achieved by LK. When only small
methane mole fraction of less than 0.3 are considered, the average absolute
error is reduced to 15%, with an absolute average difference of 16.1 J/mol.
In this analysis, only one data source was analyzed, as this was the only
experimental data set readily available. Further analysis of the models with
additional data sets is required to determine the validity of the models for
predicting excess enthalpies of hydrogen sulfide-methane mixtures.
References
1. Barry, A., Kallaguine, S., and R. Ramalho, “Excess Enthalpies of the Binary
System Methane-Hydrogen Sulfide by Flow Calorimetry”, J. Chem. Eng. Data,
27, 436–439, 1982.
2. Barry, A., Kallaguine, S., and R. Ramalho, “Ternary System Methane-Carbon
Dioxide-Hydrogen Sulfide. Excess Enthalpy Data by Flow Calorimetry”,
J. Chem. Eng. Data, 28, 375–381, 1983.
3
Phase Behavior and Reaction
Thermodynamics Involving Dense-Phase
CO2 Impurities
J.A. Commodore, C.E. Deering and R.A. Marriott
Department of Chemistry, University of Calgary, Calgary, Alberta, Canada
Abstract
High-density CO2 streams destined for subsurface reinjection contain multiple
impurities which can change the phase behavior, density or reaction chemistry.
Industrial streams include those aimed at Carbon Capture and Sequestration,
Enhanced Oil Recovery or sulfur/carbon management by Acid Gas Injection. The
purification, compression and injection processes for these streams involve fluids
over a large range of temperature (0–150 °C) and pressure (0.1–35 MPa). While
many chemical activity models for calculating complex high-pressure equilibria
have been reported for aqueous systems, CO2 rich systems have received very little
attention. Building on our work with H2O, H2S and COS in CO2, this new study
focuses on the volumetric effect of dissolving CS2 or SO2 into a high-pressure CO2
fluid at conditions up to p = 35 MPa and temperatures ranging from T = 50 to
125 °C. We describe our densimetric experiments and how those measurements
allow us to fully describe the fugacities of CS2 or SO2. These mixing coefficients
obtained and resulting fugacities can be directly incorporated into Gibbs Energy
Minimization routines, for calculation of high-pressure phase behavior and exploration of chemical reactivity.
3.1
Introduction
High-pressure CO2 streams can contain a variety of minor chemical species with potential to react to change phase behavior and/or chemical composition. Whether the intent of an injectate stream is Carbon Capture and
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (55–62)
2017 © Scrivener Publishing LLC
55
56 Carbon Dioxide Capture and Acid Gas Injection
Storage (CCS), Enhanced Oil Recovery (EOR) and/or conventional Acid
Gas Injection (AGI), our group has been interested in understanding how
the chemical equilibria can change under compression and injectate conditions, i.e., beyond the critical conditions for CO2 fluids (Tc = 30.94 °C and
pc = 7.38 MPa). While we do not develop marketable simulation tools, our
recent research has been aimed at density measurements to provide reference quality mixing parameters for the benchmarking of Gibbs Energy
Minimization based simulators. Impurities of interest include hydrogen
sulfide (H2S), sulfur dioxide (SO2), carbonyl sulfide (COS), oxygen (O2)
and carbon disulfide (CS2) among others.
Our approach has been to measure densities with an in-house built densimeter [1]. By measuring the change in density caused by dissolving a
small amount of impurity, we then calibrate mixing parameters for reduced
Helmholtz Energy reference equations. All measurements have been completed in the single phase region, where density changes are converted to
apparent molar volumes. Because apparent molar volumes are a type of
excess property, these properties are used to optimize mixing parameters
without the optimization being affected by imperfections in the pure-component equations of state (EOS), i.e., apparent molar volumes are more
sensitive to the intermolecular interactions versus the bulk density change.
In recent examples of this work, we have published results for H2O and
COS in CO2, which have allowed us to explore high-pressure water dewpoint and COS hydrolysis equilibria [1, 2]:
COS + H2O
H2S + CO2,(3.1)
Although the previous parameters were only calibrated with our volumetric measurements at a single low concentration, the resulting symmetric mixing coefficients were shown to better predict the phase behavior of
H2O + CO2 and COS + CO2, thus providing an external validation of our
measurements and optimization approach.
In this work we describe the measurements for the volumetric studies
of CS2 and SO2 in dense CO2 fluids up to p = 35 MPa and temperatures up
to T = 125 °C. Here CS2 can enter a CO2 injectate stream from incomplete
combustion (flue gas) or hydrocarbon production, whereas SO2 can only
enter an injectate stream through CCS or flue gas. The calibrated parameters from this work showed a significant improvement over the estimated
parameters when compared to measured dew and bubble points pressures
for SO2 + CO2 systems [3–5]. CS2 is different from the previous impurity
studies, because there are currently no high-accuracy Helmholtz energy
equations for CS2. Thus, we report an alternative method for calculating
fugacity, through a Fluctuation Solution Theory correlation.
Phase Behavior and Reaction Thermodynamics 57
3.2
Experimental
The densities of impurities in dense CO2 fluid were obtained using a vibrating tube densimeter (VTD) which was described in the previous work of
Deering et al. and shown to accurately measure density to an estimated
uncertainty of ±0.07 kg m–3 over a wide range of temperature and pressure [1]. With a VTD, the density of the fluid of interest can be related to
a reference fluid whose density is accurately known over the conditions of
interest by measuring the period of oscillation of the vibrating tube. The
fundamental equation relating the period of oscillation for both the fluid
of interest and reference fluid is given by equation 2,
ρ – ρo = KT(τ2 – τo2),(3.2)
where ρ and ρo are the fluid of interest and reference densities, respectively;
τ and τo are the vibrating tube’s period of oscillation of the fluid of interest and reference fluid; KT is the isothermal pressure dependent calibration constant (calibrated with a second well characterized fluid/calibration
fluid). Degassed H2O and N2 were used as the calibration and reference
fluids in this work. The density of the reference fluid (N2) was calculated
using Span et al. [4] and the density for the calibration fluid (H2O) was
calculated using Wagner and Pruß [5].
Mixtures were gravimetrically prepared in an evacuated 500 cm3 stainless steel vessels and agitated on a rocking table for two weeks for homogeneity. Each binary mixture was analyzed by a gas chromatograph (GC)
to verify the composition and/or identify any other impurities. The purity
from supplier was deemed sufficient and was used without any further
purification, see Table 3.1. The prepared mixtures were transferred to a
syringe pump (Teledyne-ISCO 260D) which can control pressure with a
precision of ±0.005 MPa.
Table 3.1 Chemical name, purities, source and analysis method.
Chemical name
Source
Analysis method Purity mol %
Carbon Dioxide
Praxair Inc.
GC-TCD/FID
>99.9995
Carbon Disulfide Praxair Inc.
GC-TCD/FID
>99.9
Sulfur dioxide
Praxair Inc
GC-TCD/FID
>99.98
Carbonyl sulfide
Praxair Inc.
GC-TCD/FID
>99.9
Nitrogen
Praxair Inc.
GC-TCD/FID
>99.998
Water
In-house, EMD Millipore Resistivity
18 MΩ∙cm–1
58 Carbon Dioxide Capture and Acid Gas Injection
Table 3.2 Impurity concentrations in CO2 and the conditions studied.
T/°C
p/MPa
x/mol%
Reference
COS
49.76–120.10
2.5–35
2.737
Deering et al. [2]
H2O
50–125
2.5–35
0.280
Deering et al. [1]
SO2
50.92–126.84
2.5–35
1.042
This work
CS2
50.94–127.35
2.5–35
1.011
This work
Solute
Density measurements were completed by charging the fluid from the
syringe pump into the vibrating tube to measure the period of the fluid
isothermally across the pressure range of interest. Table 3.2 shows the concentrations of solutes in dense CO2 (1) phase used in this work and the
previous studies.
3.3
Results and Discussion
3.3.1 Phase Behavior Studies of SO2 Dissolved in
Dense CO2 Fluid
The relative density measurements of the SO2(2) + CO2(1) system and pure
CO2 from this work were combined to calculate apparent molar volumes
–
(Vϕ,2) of the SO2:
–
Vϕ,2/(cm3∙mol–1) = M2/ρ – 1000Δρ/(mρρ1),(3.3)
where M2 is the molar mass of SO2, ρ and ρ1 are the densities of the mixture
and the pure CO2 respectively and m is the molality of the mixture.
The apparent molar volumes were then used to optimize the symmetric parameters found within the multi-fluid EOS through minimizing the
sum of a weighted sum of the squares difference between the volumes from
equation 3 and those calculated. The optimization began with the estimated
parameters of Kunz et al. [8] and the weighting was applied as the reciprocal
of the squared uncertainties of the apparent molar volumes. The calculation
of the mixture volumes using multi-fluid EOS requires the combination of
a pure fluid EOS and mixture contribution from the composition dependent reducing functions for supercritical density and temperature (δ and τ).
The description of the δ and τ functions which contain the binary parameters to be optimized uses the formulation by Kunz et al. [8]:
Phase Behavior and Reaction Thermodynamics 59
N
N
i
j
xi x j
v ,ij v ,ij
xi
xj
1
xj 8
v ,ij xi
1
1
1/3
c ,i
1/3
c, j
3
(3.4)
and
N
i
N
j
xi x j
T ,ij T ,ij
xi
T ,ij xi
xj
xj
Tc ,iTc , j , (3.5)
where Tc and ρc are the critical temperature and densities for the pure fluids, N is the number of components, x is the mole fraction of pure components in the mixture. βv,ij and γv,ij in equation 4 and βT,ij and γT,ij in equation 5
are used to fit symmetric and asymmetric portions of the mixing behavior.
In this work, only the symmetric parameters (γv,ij and γT,ij) were required
to adequately calculate the apparent molar volumes. The Helmholtz free
energy EOSs used to describe the pure fluid contribution are reported in
Table 3.3.
The optimized symmetric parameters obtained in this work were used
to calculate the dew and bubble point pressures for comparison to the
measured vapor-liquid equilibrium data for SO2(2) + CO2(1) systems (see
Figure 3.1) [3–5]. The results from the calculation with the optimized
parameter showed a better agreement with the literature vapor-liquid
equilibrium data and a significant improvement over the original estimated parameters by Kunz et al. [8]. We note that this improved agreement is based only on our apparent molar volumes which were measured
at a single concentration of 1.042 mol% SO2, again showing the utility in
the volumetric data for optimizing mixing parameters.
Table 3.3 The Helmholtz EOS model used to describe the pure component fluid.
Component
Formula
Pure fluid EOS used
Carbon dioxide
CO2
Reference equation of state by Span and
Wagner [9]
Sulfur disulfide
SO2
Short fundamental equation of state by
Lemmon and Span [10]
60 Carbon Dioxide Capture and Acid Gas Injection
12
SO2(2) + CO2(1)
10
p/MPa
8
6
4
60.06 °C
2
–10 °C
0
0
0.2
0.4
0.6
Mole fraction CO2(1)
0.8
1
Figure 3.1 The p-x diagram for SO2(2) + CO2(1) at T = –10 and 60.06 °C. , Lachet et al.
[3]; ,Caubet et al. [5]; ---, estimated binary mixing parameters combined with highaccuracy equations-of-state —, optimized binary mixing parameters from this work
combined with high accuracy reduced Helmholtz EOS.
3.3.2
The Densimetric Properties of CS2 and CO2 Mixtures
The relative density data for the CS2(3) + CO2(1) mixture was used in calculating the apparent molar volume for CS2 dissolved in dense CO2 fluid.
Because no reduced Helmholtz energy equation of state for pure CS2 was
found in the literature, the calculated apparent molar volumes were used
to optimize coefficients within a Fluctuation Solution Theory based correlation equation;11
cm3mol
V3
1
V1o
o
T ,1RT
a13 b13 exp c13 / V1o
1
V1o , (3.6)
–o
where V1 is the molar volume of the pure solvent (CO2), κoT,1, is the isothermal compressibility of the pure solvent and a13, b13, c13 are adjustable
parameters. The obtained adjustable parameters within the model show a
~ 50 cm3 mol–1 volume change upon increase in pressure to a dense CO2
region. No available other literature density data were found for comparison. The data are shown in Figure 3.2.
The adjustable parameters from equation 3 can then be used to calculate
fugacities for CS2 without employing an equation of state for CS2:
ln
3
o
1
a13 b13 V1o b13 exp c13 V1o
1 c13 , (3.7)
Phase Behavior and Reaction Thermodynamics 61
1000
CS2(3) + CO2(1)
127.353 °C
–
V3 /(cm3 mol–1)
600
101.632 °C
200
76.280 °C
–200
50.941 °C
–600
0
5
10
15
20
p/MPa
25
30
35
40
Figure 3.2 The apparent molar volume for CS2(3) dissolved in dense phase CO2(1)
investigated for p ≤ 35MPa, ( ), denotes experimental data from this work.
where 3∞ and 1o are infinite dilution fugacity coefficients of solute and
solvent respectively.
The above equation, in combination with the previous mixing parameters, can be utilized in Gibbs Free Energy minimization models for calculating chemical equilibria over a wide range of temperatures and pressures.
Future studies will include H2S and O2 in CO2, in addition to H2S rich fluids.
References
1. Deering, C. E., Cairns, E. C., McIsaac, J. D., Read, A. S., and Marriott, R. A.
The partial molar volumes for water dissolved in high-pressure carbon dioxide
from 318.28 K to 369.40 K and pressures to 35 MPa. The Journal of Chemical
Thermodynamics 93, 337–346, 2015.
2. Deering, C. E., Saunders, M. J., Commodore, J. A., and Marriott, R. A. The
Volumetric Properties of Carbonyl Sulfide and Carbon Dioxide Mixtures
from T = 322 to 393 K and p = 2.5 to 35 MPa: Application to COS Hydrolysis
in Subsurface Injectate Streams. Journal of Chemical and Engineering Data 61,
1341–1347, 2016.
3. Lachet, V., Bruin, T. de, Ungerer, P., Coquelet, C., Valtz, A., Hasanov, V.,
Lockwood, F., and Richon, D. Thermodynamic behavior of the CO2+SO2 mixture: Experimental and Monte Carlo simulation studies. Energy Procedia 1,
1641–1647, 2009.
62 Carbon Dioxide Capture and Acid Gas Injection
4. Blümcke, A., Ueber die Bestimmung der specifischen Gewichte und
Dampfspannungen einiger Gemische von schwefliger Säure und Kohlensäure.
Ann. Phys. Leipzig 270, 10–21, 1888.
5. Caubet, F., The liquifaction of gas mixtures. Z. Kompr. Fluess. Gase 8, 65, 1904
6. Span, R. A Reference Equation of State for the Thermodynamic Properties of
Nitrogen for Temperatures from 63.151 to 1000 K and Pressures to 2200 MPa.
Journal of Physical and Chemical Reference Data 29, 1361, 2000.
7. Wagner, W., and Pruß, A. The IAPWS formulation 1995 for the thermodynamic properties of ordinary water substance for general and scientific use.
Journal of Physical and Chemical Reference Data 31, 387–535, 2002.
8. Kunz, O., and Wagner, W. The GERG-2008 Wide-Range Equation of State for
Natural Gases and Other Mixtures: An Expansion of GERG-2004. J. Chem.
Eng. Data. 57, 3032–3091, 2012.
9. Span, R., and Wagner, W. A New Equation of State for Carbon Dioxide
Covering the Fluid Region from the Triple-Point Temperature to 1100K at
Pressures up to 800 MPa. J. Phys. Chem. Ref. Data. 25, 1509–1595, 1996
10. Lemmon, E. W., and Span, R. Short Fundamental Equations of State for 20
Industrial Fluids. J. Chem. Eng. Data. 51, 785–850, 2006.
11. O’Connell, J. P., Sharygin, A. V., and Wood, R. H. Infinite Dilution Partial
Molar Volumes of Aqueous Solutes over Wide Ranges of Conditions. Industrial
& Engineering Chemistry Research 35, 2808–2812, 1996.
4
Sulfur Recovery in High Density CO2 Fluid
S. Lee and R.A. Marriott
Department of Chemistry, University of Calgary, Calgary, Alberta, Canada
Abstract
After purification of natural gas through aqueous amines, several sources result in
low-quality low-pressure H2S acid gas mixtures (<1% H2S in CO2), where this lowlevel of H2S cannot sustain the front-end-furnace in a traditional Claus conversion process for commercial sulfur recovery. Other available methods to recover
low-level H2S normally produce poor quality non-commercial sulfur or rely on
disposable scavenging systems. Alternatively, several newer cryogenic separation
processes have been demonstrated, where the resulting CO2 rich acid gas liquids
do not require recompression. Despite the more favorable efficiency of these processes, the high-pressure CO2 fluids either need to be reinjected into zones which
can safely contain H2S, purified by complex multi-stage cryogenic separation or
H2S would need to be removed after the cryogenic separation. In this study, we
propose the direct oxidation of H2S to elemental sulfur in a dense-phase highpressure CO2 fluid. Through such a process, CO2 is purified and high-quality elemental sulfur is produced for commercial acid production. This study describes
some initial experimental measurement for sulfur solubility in CO2. Based on the
high-pressure experimental measurements, a solubility model for sulfur in CO2 is
offered, which has allowed us to estimate the conditions at which high-pressure
sulfur recovery may be possible. Our current work is introduced, where we are
now focusing on the catalytic conversion of H2S at the high-pressure conditions
provided through this work. The high-pressure catalytic conditions will be tested
along feed fluid impurities such as methyl thiol (CH3SH), carbonyl sulfide (COS),
and carbon disulfide (CS2).
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (63–70)
2017 © Scrivener Publishing LLC
63
64 Carbon Dioxide Capture and Acid Gas Injection
4.1
Introduction
Several unconventional natural gas streams contain a significant amount of
low-quality acid gas, e.g., Horn River in N.E. BC (10% CO2 and 500 ppm
H2S) [1]. The acid gas components must be removed from sour gas to produce sweet gas for environmental and safety reasons [2] and to meet pipeline requirements [3]. If CO2 is sold as a feed for other chemical processes,
enhanced oil recovery (EOR) or injected into storage/disposal reservoirs,
H2S will need to be removed. The primary available methods to remove
H2S include (i) scavengers, (ii) conventional aqueous amine separation and
the Claus process, or (iii) low-level oxidation (LO-CAT and others) [4].
Non-regenerable scavengers are a less economical choice when the volume
of natural gas is large, and for poor quality acid gas fluids, the overall concentration of H2S is not large enough to maintain Claus furnace temperature. Liquid redox processes often encounter problems such as absorber
plugging, bacterial contamination and poor-quality sulfur. This poor-­
quality sulfur is often landfilled due to impurity [5]; therefore, the commercial value can become an environmental liability. Finally, the purified
CO2 using these methods is discharged at low pressure, where CO2 would
need to be compressed at significant cost [6].
Alternative methods for the separation of high-pressure (HP) acid gas
from sour natural gas include the cryogenic processes [7–12]. The cryogenic
processes discharge the acid gas as a HP liquid that can be injected into
reservoirs for geo-sequestration or EOR without any further compression
(or minimal compression). Further distillation to remove the H2S is costly
and complicated [13], and there are currently no commercial methods to
oxidize low-level H2S to elemental sulfur in the HP CO2 fluid. Creating
a novel solution to convert the H2S into S8 while maintaining the densephase CO2, would not only allow a producer to capture the economic benefits of sulfur recovery, but also to conserve a HP CO2 stream that would
require minimal to no compression or purification to meet requirements
for sale (Figure 4.1). The first step in determining whether the oxidation of
H2S in a dense CO2 fluid is feasible is to explore if the solubility of sulfur in
CO2 is sufficient.
4.2
Literature Review
Our previous sulfur solubility model [14] can be used to predict sulfur
vapor pressure or sulfur dew-point in CO2; however, the model was calibrated for sour gases and was not fit specifically for the purpose of HP
Sulfur Recovery in High Density CO2 Fluid 65
CH4
to sales
Cryogenic
separation
O2
Catalytic
reactor
H2S/CO2 (l or sc)
+
minor impurities
H2S
CO2
COS
CH4
Separator
CO2 (l)
S8 (s)
Figure 4.1 Simplified schematic of the proposed sulfur recovery process.
CO2. Limited data were available in the literature for S8 solubility in CO2,
with no solubility data for T > 120 °C. Thus, measuring the solubility of
elemental sulfur in dense CO2 was necessary and was the initial focus of
this work. We note that the solubility of sulfur in pure CO2 data from other
reports [15, 16] were less consistent, which had been previously recognized
by Dowling et al. [14] and Serin et al. [17]. In addition, the solubilities calculated using the existing model by Dowling et al. [14] were consistently
less than the literature data for p > 15 MPa. These deviations were likely
a result of the model being calibrated for sour gas mixtures; therefore, a
better fit-for-purpose model was needed for pure CO2 and more accurate solubility calculations in the HP region where the best conditions for
catalysis are estimated to be based on theoretical equilibrium conversions
[18, 19]. Larger solubilities are desired for larger inlet H2S concentrations,
as reaching the sulfur dew-point could lead to deposition on the catalyst
and decreased conversion rates.
4.3
Methodology
The solubility of sulfur in CO2 was measured using a custom-built HP saturation column followed by cold and chemical traps shown in Figure 4.2.
66 Carbon Dioxide Capture and Acid Gas Injection
Data
logging
computer
Gas meter
HP
transducer
Platinum resistance
thermometer
Teledyne
Isco260D
syringe pump
Glass wool
U-glass traps
Sulfur equilibrium
vessel
Poppet
valve
Thermostated zone
Figure 4.2 Schematic of in-house built sulfur solubility measurement apparatus.
The column loaded with sulfur was pressurized with CO2 and was
shut in to reach equilibrium. Sulfur saturated CO2 was slowly released to
ambient pressure and temperature. The released fluid flowed through two
glass-wool u-traps before reaching a flow meter, for measuring the net
discharged CO2. Sulfur deposited on the transfer lines and the traps were
quantified using gas chromatography [20]. The total amount of deposited
sulfur was related to the net volume of CO2 that had exited the saturation
vessel to give the solubility of S8 in HP CO2.
4.4
Results and Discussion
Sulfur solubility in CO2 was measured at pressures of p = 10 and 20 MPa
and temperatures between T = 50 to 151 °C. The new experimental data
were consistent with Serin et al. [17]. While the experimental temperatures
were not the same, the literature data at T = 60 and 90 °C were aligned
between the experimental data of T = 50 and 100 °C. The experimental results indicated that the solubility of sulfur increased with pressure
(beyond p = 5 MPa) and temperature. The effect of pressure on sulfur
solubility can be explained by the increase of the CO2 fluid density. The
effect of temperature on sulfur solubility is due to the sulfur vapor pressure
increase.
Early indications from the modeling using a Mesmer-type equation did not allow for sufficient elemental sulfur to stay dissolved in CO2
Sulfur Recovery in High Density CO2 Fluid 67
50
10
p/MPa
40
Catalytic conditions
30
Separation
conditions
20
(0.1% H2S feed)
10
0
0.001
0.0001
0
50
0.1
0.01
100
[S8] = 1.432 g/m3
150
T/°C
200
250
300
Figure 4.3 Sulfur dew-point based on the Mesmer CO2 model.
([S8] < 1.432 g m–3, which corresponds to [H2S] < 1000 ppm, which is not
great for an inlet threshold) for the initially chosen catalytic conditions
T < 150 °C. Revised conditions at increased T and p, where the elemental
sulfur will remain dissolved in the dense-phase CO2 if 1.432 g m–3 < [S8]
< 10.0 g m–3 are still promising. This concentration range corresponds to
1000 ppm < [H2S] < 0.7%, which allows for a larger threshold for H2S inlet
concentration (Figure 4.3).
4.5
Conclusion and Future Directions
The Mesmer model showed that a non-sub-dew-point process may be possible at a higher temperature than initially considered. However, increasing T will decrease the total possible conversion of H2S to S8. Increasing
temperature to a point where the catalytic condition is above the expected
sulfur dew-point, but still low enough temperature for a sufficient conversion of H2S, will now be the focus of this project in association with
experimental catalysis. The parameters from the solubility models can be
used for Gibbs Energy Minimization routines to calculate the equilibrium
conversion in real fluid conditions. Upon selection of favorable pressures
and temperatures, the catalytic conversion of H2S to elemental sulfur in
dense CO2 over alumina and titania will be studied using a custom heterogeneous catalysis system. A further study will need to find temperature
ranges that work for different inlet H2S concentrations as well as the catalytic efficiency under the effect of feed fluid impurities.
68 Carbon Dioxide Capture and Acid Gas Injection
References
1. BC Oil and Gas Commission. Hydrocarbon and By-Product Reserves Report
January-December 2012. http://www.bcogc.ca/node/11111/download (accessed
Oct 17, 2016).
2. Tyndall, K.; McIntush, K.; Lundeen, J.; Fisher, K.; Beitler, C., When is CO2
more hazardous than H2S. Hydrocarb Process 90 (1), 45–48, 2011.
3. TransCanada, Gas Quality Specifications TransCanada and other pipelines.
2010.
4. Dalrymple, D. A.; Trofe, T. W.; Evans, J. M., Liquid redox sulfur recovery
options, costs, and environmental considerations. Environmental Progress
8 (4), 217–222, 1989.
5. Primack, H. S.; Reedy, D. E.; Kin, F. R., Method of stabilizing chelated polyvalent metal solutions. Google Patents: 1984.
6. Campbell, J. M., Gas conditioning and processing. Campbell Petroleum Series:
Vol. 4, 1982.
7. Kelley, B. T.; Valencia, J. A.; Northrop, P. S.; Mart, C. J., Controlled Freeze
Zone for developing sour gas reserves. Energy Procedia 4, 824–829, 2011.
8. Lallemand, F.; Lecomte, F.; Streicher, C., Highly Sour Gas Processing: H2S
Bulk Removal With the Sprex Process. In International Petroleum Technology
Conference, International Petroleum Technology Conference: Doha, Qatar,
2005.
9. Terrien, P.; Dubettier, R.; Leclerc, M.; Meunier, V. In Engineering of Air
Separation and Cryocap units for large size plants, Oxyfuel Combustion
Conference, Ponferrada, Spain, 9–13 Sep, 2013; Ponferrada, Spain, 2013.
10. Holmes, A. S.; Ryan, J. M.; Price, B. C.; Styring, R. E., Process Improves Acid
Gas Separation. Hydrocarb Process 61 (5), 131–136, 1982.
11. Ryan, J.; Schaffert, F., CO2 Recovery by the Ryan-Holmes Process. Chemical
Engineering Progress 80 (10), 53–56, 1984.
12. ZareNezhad, B.; Hosseinpour, N., An extractive distillation technique for
producing CO2 enriched injection gas in enhanced oil recovery (EOR) fields.
Energ Convers. Manage. 50 (6), 1491–1496, 2009.
13. Guvelioglu, G. H.; Higginbotham, P.; Palamara, J. E.; Arora, G.; Mamorsh, D.
L.; Fisher, K. S., H2S Removal from CO2 by Distillation. In Laurance Reid Gas
Conditioning Conference, Norman, Oklahoma, 2015.
14. Dowling, N. I.; Marriott, R. A.; Primak, A.; Manley, S., The Kinetics of H2S
Oxidation by Trace O2 and Prediction of Sulfur Deposition in Acid Gas
Compression Systems. In Sour Gas and Related Technologies, John Wiley &
Sons, Inc.: pp. 183–214, 2012.
15. Gu, M. X.; Li, Q.; Zhou, S. Y.; Chen, W. D.; Guo, T. M., Experimental and
Modeling Studies on the Phase-Behavior of High H2S-Content Natural-Gas
Mixtures. Fluid Phase Equilibr. 82, 173–182, 1993.
Sulfur Recovery in High Density CO2 Fluid 69
16. Kennedy, H. T.; Wieland, D. R., Equilibrium in the Methane-Carbon DioxideHydrogen Sulfide-Sulfur System. T. Am. I. Min. Met. Eng. 219 (7), 166–169,
1960.
17. Serin, J. P.; Jay, S.; Cezac, P.; Contamine, F.; Mercadier, J.; Arrabie, C.; LegrosAdrian, J. M., Experimental studies of solubility of elemental sulphur in
supercritical carbon dioxide. J. Supercrit. Fluid 53 (1–3), 12–16, 2010.
18. Gamson, B.; Elkins, R., Sulfur from hydrogen sulfide. Chemical Engineering
Progress 49 (4), 203–215, 1953.
19. Energy, A. A.; Resources, N.; Paskall, H. G.; Research, W.; Ltd, D., Capability
of the Modified-claus Process : a Final Report to the Department of Energy and
Natural Resources of the Province of Alberta. Western Research & Development:
1979.
20. Clark, P. D.; Lesage, K. L., Quantitative-Determination of Elemental Sulfur in
Hydrocarbons, Soils, and Other Materials. J. Chromatogr. Sci. 27 (5), 259–261,
1989.
5
Carbon Capture Performance of Seven
Novel Immidazolium and Pyridinium
Based Ionic Liquids
Mohamed Zoubeik, Mohanned Mohamedali and Amr Henni
Acid Gas Removal Laboratory, Clean Energy Technologies Research Institute
(CETRi), University of Regina, Regina, SK, Canada
Abstract
The objective of this study is to compare the solubility of carbon dioxide (CO2) in
seven ionic liquids, namely 1,2,3-Tris(diethylamino) cyclopropenylium dicyanamide, 1-Ethyl-3-methylimidazolium L-(+)- lactate, 3-Methyl-1-propylpyridinium
bis [(trifluoromethyl) sulfonyl]imide, Ethyldimethyl propylammonium
bis(trifluoromethyl sulfonyl)imide, 1,2,3-Tris(diethylamino) cyclopropenylium bis(trifluoromethanesulfonyl)imide, 1-(4-Sulfobutyl) -3-methylimidazolium
Bis(trifluoro methanesulfonyl)imide, 1-(4-Sulfobutyl)-3-methylimidazolium
hydrogen sulfate. Solubility measurements were performed at 313.15, 323.15 and
333.15 K, up to 20,000 mbar. CO2 solubility decreased in the following order:
[TCD][TF2N] > [PMPY][TF2N] > [EMMP][TF2N] > [emim][LACTATE] >
[TCD][DCN] > [(CH2)4SO3HMIm][TF2N] > [(CH2)4SO3HMIm] [HSO4]. The
three ionic liquids, [TCD][TF2N], [PMPY][TF2N] and [EMMP][TF2N], show
great potential for CO2 capture. Reported values of Henry’s law constants, enthalpies and entropies of absorption for CO2 were compared. The Peng-Robinson
Equation of state, with quadratic mixing rules, was capable of correlating all data
satisfactorily for all the ionic liquid systems.
5.1
Introduction
Climate change is considered one of the greatest environmental challenges
facing our civilization to date. The anthropogenic emissions of carbon dioxide (CO2) represent the greatest contribution to global warming and climate
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (71–90)
2017 © Scrivener Publishing LLC
71
72 Carbon Dioxide Capture and Acid Gas Injection
change. Human activities, especially those aimed at energy production, are
the principal sources of CO2 emissions, of which fossil fuel combustion represents the vast majority. Carbon dioxide is the main greenhouse gas and its
concentration in the atmosphere reached 400 ppm in 2013 [1, 2]. Therefore,
the deployment of environmentally benign, energy efficient and economic
CO2 capture technologies is becoming an important research topic [3]. There
has been continuous improvement in the field of capture technologies for CO2;
however, they are all associated with major drawbacks. Currently, the use of
absorption using chemical solvents, predominantly aqueous amine solutions
for CO2 separation, is considered the only commercially available technology
for the capture of CO2 from flue gases. However, amine absorption systems
are considered expensive due to the high energy required in the regeneration step, in addition to amine degradation during the thermal regeneration
process [4]. Amongst potential solvents for CO2 capture, ionic liquids (ILs),
as nonvolatile solvents, have been given much attention and are regarded
as promising candidates [5]. Ionic liquids are salts composed of cations and
anions and mainly exist as liquid at room temperature. They possess some
enviable characteristics such as low vapor pressure, high thermal and chemical stability rendering them as potential alternatives to the energy intensive
amine scrubbing ­process to achieve an environmentally and economically
viable CO2 separation [6]. In addition to their unique properties, ILs could
be systematically synthesized and tailored to fine-tune their final properties
toward more efficient gas separation characteristics with a proper combination of the cations and anions counterparts [7].
One of the most commonly studied ILs in the literature are the imidazolium based ILs along with other sulfonium, ammonium, and phosphonium
based solvents [7]. The availability of a wide range of choices of cations and
anions allows for more options for the optimization of the design and synthesis of new ILs to ultimately improve the CO2 solubility and reduce the
energy required in the regeneration step. A combined experimental and
theoretical approach was adopted by Yan et al. to understand the structureproperty relationship due to the addition of various functional moieties
such as aromatics, aliphatics and silane based groups attached to the triazolium based ILs [8]. The study reported on effective methods for the prediction of the modified ILs properties [8]. Anions substitution of ILs has
gained considerable attention in the past decades as a method to modify
the final properties of ILs. For instance, phosphonium hydroxide reacting
with substituted phenols resulted in ILs possessing new properties. Both
the viscosity and decomposition temperature of the ILs were significantly
impacted by the variations of the substituent group attached to the anion
part [9]. Several review papers are available in the literature to summarize
the numerous studies performed on using ILs for CO2 capture, and special
Carbon Capture Performance of Seven Novel 73
attention was dedicated to understanding the effects of varying the anion,
cation, substituent groups on the cation and anions [10–12].
In this work, we discuss the CO2 solubility in seven promising ILs measured in our lab using a gravimetric microbalance (IGA). The selection of
the ionic liquids was based on the fact that the results will allow us to better
understand the effect of varying the anion and cation parts, for instance,
the bis(trifluoromethylsulfonyl) imide anion, which is common among
some Ils, was reported to have high affinity toward CO2 due to its high
fluorination content. Furthermore, we also aimed at studying the effects of
changing the cation to compare different types of chemical functionality
on the cation with the same anion. The Peng-Robinson equation of state,
with a quadratic mixing rule, was used to correlate the experimental solubility data. Henry’s law constants, the entropy and enthalpy of the absorption process were also derived and reported.
5.2
5.2.1
Experimental Work
Materials
Ionic liquids used in this work were purchased from Sigma-Aldrich, io-litec and Solvionic, as reported in Table 5.1, with their acronyms and molecular weights. Research grade carbon dioxide (CO2) was purchased from
Praxair, with a purity of 99.99 wt.%.
5.2.2
Density Measurement
The densities of ionic liquids used in this research were measured at different temperatures using an Anton Paar DMA 4500 digital density meter.
The device allows for precision within 0.00001 g∙cm−3 and the uncertainty
of the measurements was estimated to be 0.00005 g∙cm−3. The apparatus
consists of a glass U-tube with a PT100 platinum resistance thermometer with an uncertainty of 0.01 K. The density meter was calibrated with
air and bi-distilled water. The U-tube was carefully cleaned and dried for
30 min at 353.15 K before injecting the ionic liquids. Approximately 2 mL
of a sample were slowly injected inside the glass U-tube of the apparatus.
When the desired temperature was reached, the density was measured. The
average of at least three measurements was reported.
5.2.3
Solubility Measurement
Solubility was measured using an Intelligent Gravimetric Analyzer (IGA
003) from Hiden Analytical (Figure 5.1). The gravimetric microbalance
Shorthand name
[TDC] [DCN]
[EMIM] [LACTATE]
[PMPY] [TF2N]
[EMMP] [TF2N]
[TDC] [TF2N]
[(CH2)4SO3HMIm][TF2N]
[(CH2)4SO3HMIm][HSO4]
Ionic liquid
1,2,3-Tris (diethylamino) cyclopropenylium dicyanamide
1-Ethyl-3-methylimidazolium
L-(+)-lactate
3-Methyl-1-propylpyridinium
bis[(trifluoromethyl)sulfonyl]imide
Ethyldimethylpropyl-ammonium
bis(trifluoro methylsulfonyl)imide
1,2,3-Tris(diethylamino)
cyclopropenylium bis(trifluoromethanesulfonyl)imide
1-(4-Sulfobutyl)-3-methylimidazolium
bis(trifluoromethanesulfonyl)imide
1-(4-Sulfobutyl)-3-methylimidazolium
hydrogen sulfate
Table 5.1 Characteristics of ionic liquids used in this work.
Structures
316.35
499.43
532.56
396.37
416.40
200.23
318.46
Molecular weight
74 Carbon Dioxide Capture and Acid Gas Injection
Carbon Capture Performance of Seven Novel 75
IGA
P
Red = CO2
Black = vacuum
Blue = water
MFC-A
MFC-B
Reactor and
sample cell
CO2 gas
cylinder
T
Water bath
Vacuum
pump
Figure. 5.1 Schematic of the gravimetric microbalance.
contains a sample bucket where the liquid is placed inside a pressurevessel that is able to operate up to 20 bar and 500 °C. For each experiment, a small amount of ionic liquid samples in the range of 60 to 90 mg
liquid sample was loaded into the sample container. Once the sample
was loaded, the chamber was sealed. After stability was attained, the
temperature was set at the degassing temperature of 348 K using an
external water jacket. The sample was then dehydrated and degassed
by completely evacuating the reactor using a diaphragm pump until
the pressure reached 20 mbar, followed by a turbo pump (Pfeiffer) to
achieve a vacuum of about 10 mbar. The degassing step was continued
for about 10 hours to remove all traces of water and other volatile contaminants until a stable weight was achieved for about one hour, at which
point the final weight was recorded. Temperature was then set at the
absorption temperature using a water bath (Polyscience) with accuracy
of 0.1 K. Temperature was measured with a type K platinum thermocouple (±0.1 K). When the set temperature was reached, at the desired
pressure value, parameters related to the mass relaxation behavior were
recorded by the IGASwin software. The absorption process was then
initiated by allowing CO2 via a mass flow controller (MFC) to reach a
pre-set ­pressure inside the microbalance chamber. Any real-time weight
change upon absorption was automatically recorded. Pressure and temperature were kept constant until equilibrium was reached. Then, the
76 Carbon Dioxide Capture and Acid Gas Injection
pressure was raised to the second data point of the isotherm, and this
process was repeated for all other pressure measurements. A sufficient
time of about 4 hours was given to reach equilibrium and allow for
weight stability.
5.3
5.3.1
Modeling
Calculation of Henry’s Law Constants
Henry’s law constant is calculated based on the definition given below
using the fugacity data obtained from PR-EOS, at near dilution conditions
Hi T , P
lim
xi
0
fi L
xi (5.1)
fiL represents the fugacity of the gas dissolved in the liquid phase. Applying
the equilibrium condition that the chemical potential is equal in the gas
and liquid streams, the following equation of Henry’s law can be deduced:
Pi
H i (T )xi (5.2)
where, P is the partial pressure of the gas and H (T) is Henry’s law ­constant.
i
i
In conclusion it related the equilibrium solubility with the partial pressure
of the gas [13].
5.3.2
Critical Properties Calculations
The classical Lydersen-Joback-Reid modified method [14] was used to predict the critical properties of the ionic liquids.
5.3.3
Peng Robinson EoS
The equilibrium CO2 solubility was correlated using Peng-Robinson equation of state (PR-EoS) [14]. The PR-EOS parameters are obtained by the
following equations:
P
RT
a(T )
V b V V b b(V b)
(5.3)
Carbon Capture Performance of Seven Novel 77
ai
i 0.45724
bi
i
mi
5.4
5.4.1
T
0.07780
R2Tci 2
Pci 2
(5.4)
RTci
Pci 1 mi 1 Tri 0.5
(5.5)
2
(5.6)
0.37464 1.54226wi 0.26992wi 2 a
xi x j ai a j
b
xi x j
0. 5
bi b j
2
(1 kij )
(5.7)
(5.8)
1
(5.9)
(1 Iij )
Results and Discussion
Density
Figure 5.2 shows the experimental measurements of the density for the
seven ionic liquids. The trend in the experimental density decreased linearly with increasing temperature, with [(CH2)4SO3HMIm] [TF2N]
­showing the highest density whereas the lowest density was noted for
[TDC][DCN] as shown in Figure 5.2.
The density-temperature data were modeled using a linear relationship
for all the ionic liquid as shown in Table 5.2. The average absolute deviations (AADs) between the linear model predictions and the experimental
data was found to be satisfactory as reported in Table 5.2.
5.4.2 Critical Properties
The predicted properties such as the critical points, normal boiling points,
and the acentric factors of the seven ionic liquids investigated in this report
are presented in Table 5.3 below.
78 Carbon Dioxide Capture and Acid Gas Injection
2.4
2.2
2.0
1.8
[TDC ][DCN]
[PMPY] [TF2N]
[EMMP] [TF2N]
[TDC] [TF2N]
[EMIM][LACATATE]
[(CH2)4SO3HMIm][TF2N]
[(CH2)4SO3HMIm][HSO4]
1.6
1.4
1.2
1.0
0.8
280
300
320
T(K)
340
360
Figure 5.2 Experimental density of pure ionic liquids.
Table 5.2 Temperature-dependent density correlations for the studied ionic
liquids.
Ionic liquids
Density (g/cm3)
[TCD][CN]
ρ (g/cm3) = 1.10193–0.0006 × [T(C)]
0.05
[EMIM] [LATATE]
ρ (g/cm3) = 1.1601–0.0007 × [T(C)]
0.02
[TCD][Tf2N]
ρ (g/cm3) = 1.2947–0.0009 × [T(C)]
0.17
[EMMP][TF2N]
ρ (g/cm3) = 1.4236–0.0009 × [T(C)]
0.06
[PMPY][TF2N]
ρ (g/cm3) = 1.1416–0.0009 * [T(C)]
0.06
[(CH2)4SO3HMIm][TF2N]
ρ (g/cm3) = 1.6016–0.0009 * [T(C)]
0.09
[(CH2)4SO3HMIm][HSO4]
ρ (g/cm3) = 1.4533–0.0006 * [T(C)]
0.14
5.4.3
AAD (%)
CO2 Solubility
The accuracy of measuring the solubility of CO2 using the IGA was first
verified by measuring the solubility in [bmim][PF6] at 323.15 K and comparing the results with previously published results by Shiflet [15] and
396.37
416.36
532.56
318.5
200.23
499.43
316.4
[PMPY][TF2N]
[TDC][TF2N]
[TDC][DCN]
[EMIM][LACTATE]
[(CH2)4SO3HMIm][TF2N]]
[(CH2)4SO3HMIm][HSO4]
MW (g/mol)
[EMMP][TF2N]
Ionic liquids
Table 5.3 Critical properties of ionic liquids.
1017.6
1097.6
693.4
858.6
938.1
839.8
715.4
Tb (K)
1433.0
1612.8
912.7
1073.7
1255.7
1234.2
1038.7
Tc (K)
25.88
32.7
28.24
16.15
18.03
27.55
25.88
Pc (bar)
744.8
1070.1
620.1
1115.9
1394.0
964.7
955.5
Vc (cm3/mol)
0.8437
0.377
0.9702
1.0726
0.5876
0.3070
0.3334
ω
0.3602
0.2615
0.2260
0.2019
0.2407
0.2591
0.2863
ZC
Carbon Capture Performance of Seven Novel 79
80 Carbon Dioxide Capture and Acid Gas Injection
Anthony [16]. The AAD% of the measured solubility and those reported
in the literature [15, 16] were 4 and 12%, respectively.
The reported CO2 solubility was at (313.15, 323.15 and 333.15) K with
pressures up to 20 bar, and the results are presented in Figures 5.3, 5.4, and
5.5, respectively.
40
CO2 mole fraction (%)
35
30
25
20
15
10
5
0
0
2000
4000
[TCD][DCN]
6000
8000 10000 12000 14000 16000 18000 20000
Pressure (mbar)
[PMPY][TF2N]
[EMMP][TF2N]
[TCD][TF2N]
[(CH2)4SO3HMIm][TF2N]
[EMIM][LACTATE]
[(CH2)4SO3HMIm][HSO4]
Figure 5.3 CO2 solubility in seven promising ionic liquids at temperature of 313.15 K
[17, 18].
35
CO2 mole fraction (%)
30
25
20
15
10
5
0
0
2000
4000
[PMPY][TF2N]
6000
8000 10000 12000 14000
Pressure (mbar)
[EMMP][TF2N]
[TCD][TF2N]
16000
18000 20000
[EMIM][LACTATE]
Figure 5.4 CO2 solubility in seven promising ionic liquids at 323.15 K [17, 18].
Carbon Capture Performance of Seven Novel 81
30
CO2 mole fraction (%)
25
20
15
10
5
0
0
2000
4000
6000
8000
10000
12000
14000
16000 18000
20000
Pressure (mbar)
[PMPY][TF2N]
[EMMP][TF2N]
[TCD][TF2N]
[EMIM][LACTATE]
[TCD][DCN]
Figure 5.5 CO2 solubility in seven promising ionic liquids studied in this work at
333.15 K [17, 18].
5.4.4
The Effect of Changing the Cation
As stated earlier, changing the anion structure of the IL significantly
impacts the CO2 capture abilities of the IL; the cation also can considerably influence the properties of the resultant IL [19]. We have
the opportunity to examine a set of ILs, with four ILs sharing the same
bis(trifluoromethyl sulfonyl)imide anion ([Tf2N]) and four different cations: 3-methyl-1-­
propylpyridinium, ethyldimethylpropylammonium,
1,2,3-tris(diethylamino) cyclopropenylium, and 1-(4-sulfobutyl)-3-methylimidazolium bis(trifluoro methane sulfonyl)imide. The CO2 solubility in
decreasing order is as follows: [TDC][TF2N]>[PMPY][TF2N]>=[EMMP]
[TF2N]> [(CH2)4SO3HMIm][TF2N]] as shown in Figure 5.6. It is evident
that [TDC] cation has the highest CO2 solubility when compared to the
other three cations. As apparent from the chemical structure shown in
Table 5.1, the [TDC] cation contains three nitrogen (amine) atoms each
surrounded by two ethyl radicals, [PMPY] has one and [EMPP] one. [TDC]
is expected therefore to have higher capacity to interact with CO2. Looking
closely at the structure of [(CH2)4SO3HMIm] [TF2N] with two nitrogen
atoms in the ring seem to reduce the accessibility of CO2 to the absorption
sites on these nitrogen atoms which are most stable in the ring structure
than with an ethyl group as in [TDC] leading to the lowest solubility shown
in [(CH2)4SO3HMIm][TF2N] cation. Another reason for the higher solubility of [TDC] cation is attributed to its lowest density as compared to the
82 Carbon Dioxide Capture and Acid Gas Injection
20000
18000
Pressure (M bar)
16000
14000
12000
10000
8000
6000
[PMPY ][Tf2N]
[EMMP][Tf2N]
[TCD][Tf2N]
[(CH2)4SO3HMIm][TF2N]
4000
2000
0
0
10
20
30
Mole fraction of CO2 in ionic liquids (%)
40
Figure 5.6 Comparison of solubility in four ionic liquids with the same anion to illustrate
the effect of the cation at 313 K.
other cations (see Figure 5.2), which corresponds to a higher free volume
within the structure of the [TDC][Tf2N] than in the other bulkier ILs, as
reported in the literature [19, 20]. Note also that the four cations reported
in this study are different in their basic structure; for instance [TDC] is a
propenylium based cation, [PMPY] is a pyridinium based cation, [EMMP]
is an ammonium based cation and [(CH2)4SO3HMIm][TF2N] is an imidazolium based cation. It is clearly observed that the cation with the highest number of nitrogen atoms possess strong basic strength, and that the
propenylium based cation is a stronger base than the imidazolium based
IL. Although the four cations have different performance in terms of
­equilibrium CO2 capacity, this has still less overall impact on the properties of ILs as opposed to the influence of the anion, as will be demonstrated
in the following section.
Another comparison has been carried out between [TDC][DCN] and
data for [bmim][DCN] published by Aki et al. in order to compare the performance of [TDC] cation with the widely used [bmim], which is an imidazolium based cation as shown in Figure 5.7 [21]. Again, [bmim] has two
nitrogen atoms; however, [TDC] has three, leading to an increased basicity.
One of the widely used anions in the field of CO2 capture using ionic liquids is [Tf2N], which received considerable attention and has shown good
performance due to the presence of two fluoroalkyl groups in its structure.
We have also compared ILs with [Tf2N] anions with some ILs published in
literature to get insights on the effects of the cation with [Tf2N] as an anion
as shown in Figure 5.8. This again shows the good performance of [TDC]
Carbon Capture Performance of Seven Novel 83
100
[BMIM][DCN] (Aki et al. 2004)
[TDC][DCN](this work)
Pressure (bar)
80
CH3
N+
60
N
CN
–
N
CN
CH3
40
H3C
H3C
CH3
N
20
+N
C
N
CH3
–
NΞC–N–CΞN
CH3 CH3
0
0.0
0.1
0.2
0.3
0.4
Mole fraction of CO2 in ILs
0.5
0.6
Figure 5.7 Comparison of CO2 solubility in [TDC] and [bmim] cations with same anion
at 313 K [18, 21].
0.30
P = 13 bar
P = 12 bar
P = 14.79 bar
P = 16 bar
Mole fraction of CO2
0.25
0.20
0.15
0.10
0.05
[C
ho
m
m
im
][T
Ak
f2
N]
][T
im
m
[b
[h
lin
e]
[T
f2
N]
M
ul
do
on
et
al
.(
20
07
ie
)
t
f2
al
N]
. (2
Ak
00
[E
M
ie
4)
M
[C
ta
P
6H
l. (
][T
4F
20
F2
[P
9m
04
N]
M
[C
)
im
PY
(th
8H
]
]
i
[T
[T
4F
s
w
f2
F2
13
or
N]
N]
m
k)
M
im
(
t
ul
hi
][T
d
s
o
w
f2
on
or
N]
k)
et
M
ul
al
. (2
do
on
00
[T
7)
et
DC
al
[o
][T
. (2
m
F2
00
im
N]
7)
][T
(
t
f2
hi
N]
sw
Ak
or
ie
k)
ta
l. (
20
04
)
0.00
Figure 5.8 Comparison of CO2 solubility at 60 °C with different cations paired with
[TF2N] anion at 333.15 K and about 12 to 14.97 bar.
84 Carbon Dioxide Capture and Acid Gas Injection
20000
18000
16000
Pressure (M bar)
14000
12000
10000
8000
[PMPY][Tf2N]
[EMMP][Tf2N]
[TCD][Tf2N]
[(CH2)4SO3HMIm][TF2N]
[TCD][DCN]
[EMIM][LACTATE]
[(CH2)4SO3HMIm][HSO4]
6000
4000
2000
0
0
10
20
30
Mole fraction of CO2 in ionic liquids (%)
40
50
Figure 5.9 Comparison of the reported solubility data of CO2 in the seven ionic liquids at
313.15 K.
as a promising cation for CO2 absorption at low pressure. As for [omim],
an imidazolium based cation, with very long alkyl chain length (C8H17), it
showed good performance (Figure 5.9) as compared to other imidazolium
based ILs as discussed by Aki and coworkers [21].
Finally, as presented in Figure 5.9, CO2 solubility decreases as follows:
[TCD][TF2N]>[PMPY][TF2N]>[EMMP][TF2N]>[emim]
[LACTATE]>[TCD][DCN]> [(CH2)4SO3HMIm][TF2N]>[(CH2)4SO3HM
Im][HSO4]. As mentioned earlier, at low pressure, [emim][LACTATE]
has the highest solubility, but behaves like [emim] [Ac] and other solvents that have a chemical interactions with CO2. The low solubility of
[(CH2)4SO3HMIm] [HSO4] can be explained by the possibility of a loss of
solvent due to a reverse reaction accelerated by the high vacuum and temperature used in the initial steps in the operation the microbalance.
5.4.5
The Effect of Changing the Anion
Evidence from experimental solubility measurements, and theoretical
molecular computations indicate that CO2 solubility in ILs is primarily
dependent on the anion side of the IL [22, 23]. The most common cations
Carbon Capture Performance of Seven Novel 85
investigated are [bmim] and [emim] with varying anions such as [Tf2N],
[PF6], [NO3] and [DCN]. The use of 1,2,3-tris(diethylamino)cyclopropenylium [TDC] as a cation with two different anions [Tf2N] and [DCN] is
discussed in this study. [Tf2N] based IL has shown better performance than
[DCN] based ionic liquids due to the presence of multiple fluoro groups
in [Tf2N] confirming the previous findings in literature [21]. This trend
is also observed in the case of changing the anion from [HSO4] to [Tf2N]
with the same cation [(CH2)4SO3HMIm]. It was found that the [Tf2N]
anion had 4 times higher CO2 solubility than the [HSO4] anion [24]. The
high CO2 solubility, as seen particularly with the [TF2N] anion, is ascribed
to the fluoroalkyl groups in [Tf2N], which are known to be highly reactive with CO2 [21]. This might be attributed to the favorable interactions
between the negative fluorine ions and the positively charged carbon in
CO2 [21, 25].
Furthermore, comparing solubility data for [emim][LACTATE],
reported by our group, with [emim][FAP] data reported by Althuluth et al.
[26] and [emim][TF2N] as reported by Schilderman et al. [25] shows that
the [FAP] anion has higher CO2 solubility at high pressures, most likely
due to the presence of a large number of fluorine atoms [FAP]. However
the CO2 uptake shown by the [LACTATE] cation is more pronounced at
low pressure, which could be attributed to the possible reaction between
the [Lactate] anion and CO2 during the absorption process, similar to the
[acetate] based anion as reported by Shiflett and Yokozeki [27] with the
added advantage of being more environmentally friendly due to the lower
number of fluoro groups involved.
The ionic liquid, [emim] [LACTATE], has a different solubility isotherm
than all other ionic liquids, possessing a noticeable high CO2 solubility at
low pressures, which could probably mean that [emim] [LACTATE] has
both physical and chemical interactions with CO2, similar to other carboxylic
anions such as [emim] [pivalate], [emim][Ac] and [emim] [benzoate] [28].
5.4.6 Henry’s Law Constant, Enthalpy and Entropy
Calculations
Henry’s law constants for CO2 in the ILs are given in Table 5.4. The experimental solubility data were fitted to a polynomial and then Henry’s law
constants were found by taking the slopes at low pressures. The ionic liquid
with the lowest Henry’s law constant is [TDC][TF2N], which has shown
the highest CO2 equilibrium capacity indicating an inversely proportional relationship between the temperature and solubility. Enthalpy and
entropy values for CO2, in the studied ILs, are also reported in Table 5.4.
86 Carbon Dioxide Capture and Acid Gas Injection
Table 5.4 Henry’s law constants and enthalpies and entropies of absorption for
CO2 in the studied ionic liquids [17, 18].
H (bar)
Ionic liquids
313.15 K 323.15 K
333.15 K
∆h (kJ/mol)
∆s (J/mol∙K)
[Emmp][TF2N]
36.1
53.0
61.0
–22.8
–69.4
[PMPY][TF2N]
43.7
52.1
60.1
–13.8
–42.9
[TDC][TF2N]
37.2
43.4
49.3
–12.2
–37.8
[TDC][DCN]
57.2
66.3
77.3
–13.0
–40.4
[EMIM][LACTATE]
46.2
54.4
64.8
–14.6
–45.7
[(CH2)4SO3HMIm]
[TF2N]
58.8
70.9
–
–15.7
–49.5
[(CH2)4SO3HMIm]
[HSO4]
274
301.4
–
–8
–25.2
The highest heat of absorption is observed with [Emmp][TF2N] followed
by [(CH2)4SO3HMIm][TF2N] and [EMIM][LACTATE] indicating strong
interactions with CO2. The negative values for entropy show a higher
degree of ordering as CO2 dissolved in these ILs [15].
5.4.7
Thermodynamic Modeling of CO2 Solubility
Several thermodynamic models have been proposed for modeling the
equilibrium solubility of CO2 in ionic liquids. The Peng-Robinson (PR) EoS
was used for the correlation of the data. The regression of the experimental
data to the models was performed to obtain the interaction parameters
between CO2 and the ionic liquid at different temperatures. The average
absolute deviations in percentage (AAD%) between the model estimations
and the experimental data, were obtained for all the ionic liquids at the
three different temperatures.
AAD%
100
N
Pi exp Pi calc
Pi exp
(5.10)
where, N represents the number of equilibrium data points at each temperature, Pexp and Pcalc are the experimental equilibrium pressure and
the calculated pressure, respectively. Table 5.5 summarizes the AAD%
obtained correlating the different IL systems averaged for the three
temperatures.
Carbon Capture Performance of Seven Novel 87
Table 5.5 Standard deviations PR-EoS for the ionic liquids + CO2 system.
Ionic liquids
5.5
AAD%
[EMMP][TF2N]
2.2
[PMPY][TF2N]
2.3
[TDC][TF2N]
3.1
[TDC][DCN]
2.0
[EMIM][LACATE]
0.5
[(CH2)4SO3HMIm][TF2N]]
1.2
[(CH2)4SO3HMIm][HSO4]
1.4
Conclusion
In this study, the CO2 solubility in seven novel ionic liquids is compared
to the best ionic liquid in the literature. CO2 solubility decreased in the
following order: [TCD][TF2N] > [PMPY][TF2N] > [EMMP][TF2N]
> [emim][LACTATE] > [TCD][DCN] > [(CH2)4SO3HMIm][TF2N] >
[(CH2)4SO3HMIm] [HSO4].
[EMIM][LACATE] showed the high capacity for CO2 but both the solubility curve shapes and the difficulty of the EoS to correlate the data suggest that the interaction with CO2 is much more than just a simple physical
absorption. [EMMP][TF2N] seems promising as it showed similar solubility trends to some ionic liquids that are well known for their high solubility,
such as [hmim][TF2N]. Both ILs are similar and show high capacity for
CO2 absorption due to a high degree of fluorination. However, high CO2
capacities were not found in the case of [TCD][TF2N], [PMPY][TF2N],
[EMMP][TF2N] when compared to [bmim][Ac], for example, as this IL
was shown to react chemically with CO2 to form a chemical intermediary
product which is responsible for its high CO2 solubility.
The most promising ionic liquid among the seven ionic liquids investigated is [TCD][TF2N], which is a propenylium based ionic liquid paired
with the well-known [Tf2N] anion, in addition to the three nitrogen atoms
in the [TCD] structure with two ethyl groups attached to each nitrogen. The
effect of fluorination and presence of S=O groups on the TF2N anion act
synergistically to increase the CO2 solubility by increasing the CO2-philicity
of molecules due to Lewis base-Lewis acid interactions with the carbon
88 Carbon Dioxide Capture and Acid Gas Injection
atom from CO2. However, the use of fluorinated ionic liquids is not without
its environmental and health drawbacks when used in high concentrations.
Acknowledgements
The authors acknowledge the support of the Acid Gas Removal Laboratory
at the University of Regina as well as the FGSR at the University of Regina.
In addition, the first author would like to thank the CBIE (Canadian
Bureau of International Education) and the Libyan government for a graduate scholarship.
References
1. G.M. Crawley, Fossil Fuels: Current Status and Future Directions, World
Scientific, 2016.
2. IPCC, Climate Change 2014–Impacts, Adaptation and Vulnerability: Regional
Aspects, Cambridge University Press, 2014.
3. N.J. Rosenberg, W.E. Easterling III, P.R. Crosson, J. Darmstadter, Greenhouse
warming: Abatement and adaptation, Routledge, 2016.
4. G.T. Rochelle, Amine scrubbing for CO2 capture, Science, 325, pp.1652–1654,
2009.
5. S. Babamohammadi, A. Shamiri, M.K. Aroua, A review of CO2 capture by
absorption in ionic liquid-based solvents, Reviews in Chemical Engineering,
31, pp. 383–412, 2015.
6. X. Zhang, X. Zhang, H. Dong, Z. Zhao, S. Zhang, Y. Huang, Carbon capture
with ionic liquids: overview and progress, Energy & Environmental Science,
5, pp. 6668–6681, 2012.
7. M. Ramdin, T.W. de Loos, T.J. Vlugt, State-of-the-art of CO2 capture with ionic
liquids, Industrial & Engineering Chemistry Research, 51, pp. 8149–8177, 2012.
8. F. Yan, M. Lartey, K. Damodaran, E. Albenze, R.L. Thompson, J. Kim,
M. Haranczyk, H.B. Nulwala, D.R. Luebke, B. Smit, Understanding the effect
of side groups in ionic liquids on carbon-capture properties: a combined
experimental and theoretical effort, Physical Chemistry Chemical Physics, 15,
pp. 3264–3272, 2013.
9. C. Wang, H. Luo, H. Li, X. Zhu, B. Yu, S. Dai, Tuning the physicochemical properties of diverse phenolic ionic liquids for equimolar CO2 capture by the substituent on the anion, Chemistry–A European Journal, 18, pp. 2153–2160, 2012.
10. M. Hasib-ur-Rahman, M. Siaj, F. Larachi, Ionic liquids for CO2 capture—
development and progress, Chemical Engineering and Processing: Process
Intensification, 49, pp. 313–322, 2010.
Carbon Capture Performance of Seven Novel 89
11. E. Torralba-Calleja, J. Skinner, D. Gutiérrez-Tauste, CO2 capture in ionic liquids: a review of solubilities and experimental methods, Journal of Chemistry,
2013, pp. 1–16, 2013.
12. S.D. Kenarsari, D. Yang, G. Jiang, S. Zhang, J. Wang, A.G. Russell, Q. Wei,
M. Fan, Review of recent advances in carbon dioxide separation and capture,
Rsc Advances, 3, pp. 22739–22773, 2013.
13. J.M. Prausnitz, R.N. Lichtenthaler, E.G. de Azevedo, Molecular thermodynamics of fluid-phase equilibria, Pearson Education, 1998.
14. D.-Y. Peng, D.B. Robinson, A new two-constant equation of state, Industrial &
Engineering Chemistry Fundamentals, 15, pp. 59–64, 1976.
15. M.B. Shiflett, A. Yokozeki, Solubilities and diffusivities of carbon dioxide
in ionic liquids:[bmim][PF6] and [bmim][BF4], Industrial & Engineering
Chemistry Research, 44, pp. 4453–4464, 2005.
16. J.L. Anthony, E.J. Maginn, 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, 106, pp. 7315–7320,
2002.
17. M. Zoubeik, M. Mohamedali, A. Henni, Experimental solubility and thermodynamic modeling of CO2 in four new imidazolium and pyridinium-based
ionic liquids, Fluid Phase Equilibria, 419, pp. 67–74, 2016.
18. M. Zoubeik, A. Henni, Experimental and thermodynamic study of CO2 solubility in promising [TF2N and DCN] ionic liquids, Fluid Phase Equilibria,
376, pp. 22–30. 2014.
19. H. Tokuda, K. Hayamizu, K. Ishii, M.A.B.H. Susan, M. Watanabe,
Physicochemical properties and structures of room temperature ionic liquids. 2. Variation of alkyl chain length in imidazolium cation, The Journal of
Physical Chemistry B, 109, pp. 6103–6110, 2005.
20. M.J. Muldoon, S.N. Aki, J.L. Anderson, J.K. Dixon, J.F. Brennecke, Improving
carbon dioxide solubility in ionic liquids, The Journal of Physical Chemistry B,
111, pp. 9001–9009, 2007.
21. S.N. Aki, B.R. Mellein, E.M. Saurer, J.F. Brennecke, High-pressure phase
behavior of carbon dioxide with imidazolium-based ionic liquids, The Journal
of Physical Chemistry B, 108, pp. 20355–20365, 2004.
22. J.L. Anthony, J.L. Anderson, E.J. Maginn, J.F. Brennecke, Anion effects
on gas solubility in ionic liquids, The Journal of Physical Chemistry B, 109,
pp. 6366–6374, 2005.
23. S. Seo, M. Quiroz-Guzman, M.A. DeSilva, T.B. Lee, Y. Huang, B.F. Goodrich,
W.F. Schneider, J.F. Brennecke, Chemically tunable ionic liquids with aprotic
heterocyclic anion (AHA) for CO2 capture, The Journal of Physical Chemistry
B, 118, pp. 5740–5751, 2014.
24. C. Cadena, J.L. Anthony, J.K. Shah, T.I. Morrow, J.F. Brennecke, E.J. Maginn,
Why is CO2 so soluble in imidazolium-based ionic liquids?, Journal of the
American Chemical Society, 126 pp. 5300–5308, 2004.
90 Carbon Dioxide Capture and Acid Gas Injection
25. A.M. Schilderman, S. Raeissi, C.J. Peters, Solubility of carbon dioxide in
the ionic liquid 1-ethyl-3-methylimidazolium bis (trifluoromethylsulfonyl)
imide, Fluid Phase Equilibria, 260, pp. 19–22, 2007.
26. M. Althuluth, M.T. Mota-Martinez, M.C. Kroon, C.J. Peters, Solubility of
carbon dioxide in the ionic liquid 1-ethyl-3-methylimidazolium tris (pentafluoroethyl) trifluorophosphate, Journal of Chemical & Engineering Data, 57,
pp. 3422–3425, 2012.
27. M.B. Shiflett, A. Yokozeki, Phase behavior of carbon dioxide in ionic liq­
uids:[emim][acetate],[emim][trifluoroacetate], and [emim][acetate]+[emim]
[trifluoroacetate] mixtures, Journal of Chemical & Engineer­ing Data, 54, pp.
108–114, 2008.
28. J. Blath, N. Deubler, T. Hirth, T. Schiestel, Chemisorption of carbon dioxide in
imidazolium based ionic liquids with carboxylic anions, Chemical Engineering
Journal, 181, pp. 152–158, 2012.
29. M.H. Al-Rashed, K.H. Alkhaldi, M.S. Al-Tuwaim, M.S. Fandary, A.S. Al-Jimaz,
Extraction of butylbenzene from dodecane using hexafluorophosphate-based
ionic liquids: Effect of cation change, Journal of Chemical & Engineering Data,
57, pp. 2907–2914, 2012.
6
Vitrisol® a 100% Selective Process for H2S
Removal in the Presence of CO2
W.N. Wermink, N. Ramachandran, and G.F. Versteeg
PROCEDE Gastreating, Enschede, The Netherlands
Abstract
Over recent years PROCEDE developed a solvent called Vitrisol that is 100%
selective for H2S removal from industrial gases in the presence of CO2. Examples of
possible applications are the removal of H2S from biogas, FPSO and associated gas.
Vitrisol is able to remove in one stage more than 99.9+% of the H2S present in
the gas phase and has the typical characteristics of very selective H2S scavengers.
However, a major difference of Vitrisol compared to the traditional scavengers
is that Vitrisol can be completely regenerated, resulting in a solvent with fully
restored activity and crystalline sulphur.
For the absorption process required for the removal of H2S, there is no real process conditional constraint and the operating pressures and temperatures can vary
at least between 0.1–10 MPa and 283–363 K, respectively. The Vitrisol regeneration process takes place at temperatures below 373 K. The process pressure can
vary from atmospheric up to 0.5–1 MPa. The Vitrisol process can be described
with the following overall reaction equation:
H2S + 0.5O2
H2O + So(s)
In the Vitrisol process no vast amounts of energy are required for the regeneration of the solvent; therefore this process has an extremely low energy footprint.
In the present contribution the performance of Vitrisol will be demonstrated
for applications in shale gas production with typical compositions of 10–1000 ppmv
of H2S and CO2 of about 1–10 vol.%. The Vitrisol process is also compared to
a standard amine treating process designed for selective H2S removal. From the
results it can be concluded that significant reductions can be achieved by using
the Vitrisol process for operational costs as depicted in the energy consumptions
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (91–126)
2017 © Scrivener Publishing LLC
91
92 Carbon Dioxide Capture and Acid Gas Injection
of the overall process. As the costs of energy (and cooling) are extremely location
dependent, no attempt was made to quantify the capital savings. Also it must be
noted that contrary to the amine processes for the Vitrisol process, no additional
treatment of the off-gas stream is required as the H2S is directly converted to crystalline sulphur and CO2 that can be emitted to the environment.
Moreover, this study also illustrates clearly that it is advantageous to first
remove H2S from a gas stream containing both H2S and CO2 prior to CO2 removal
to reduce operational costs.
6.1
Introduction
Hydrogen sulphide (H2S) is a highly toxic and corrosive gas. Removal of
H2S from acidic gas streams, such as natural gas, industrial gas or biogas, is
important for safety, health, environmental and economic reasons. Several
regenerative and non-regenerative H2S removal processes are readily available, which are economically viable only for specific gas compositions and
gas flow rates. Apart from non-regenerative H2S removal by the use of,
e.g., adsorbents, all the regenerative aqueous liquid redox desulphurization processes (e.g., THIOPAQ, LO-CAT, SulFerox) capture CO2 to varying extents besides H2S.
The conventional method of removing H2S from natural gas is using an
amine process. Subsequently, the H2S in the stripper gas is converted to
elemental sulphur by a consecutive Claus process. For natural gas fields,
usually containing more CO2 than H2S, this will result in an inlet acid gas
stream for the Claus process that is low in H2S and high in CO2 content.
The inlet gas stream should contain at least 20 mol% of H2S to be able to
produce a stable flame in a Claus furnace. Modification of the Claus process
is needed between 20 and 50 mol% H2S in the inlet acid gas stream. Above
50 mol% H2S content no modification of the Claus process is required
[1, 2]. Moreover, owing to the co-absorption of CO2 the regeneration costs
of the amine process are substantially increased.
The Vitrisol process [3] is a recently developed selective desulphurization process based on the removal of H2S by precipitation with copper sulphate (CuSO4) in an aqueous, acidic solution. Copper sulphide (CuS) and
sulphuric acid are formed in the gas treating process [4–6]:
H2S(g) + Cu2+ + SO42– + 2H2O +
CuS(s) + 2H3O+ + SO42–(6.1)
The Vitrisol process is able to remove H2S from acidic gas streams
without the co-absorption of CO2 [5, 7]. Because the precipitation reaction
Vitrisol a 100% Selective Process for H2S Removal 93
occurs rapidly, the removal of H2S is limited by mass transfer in the gas
phase. A Vitrisol pilot absorber was built to remove H2S from biogas,
obtain representative samples of CuS and to verify design rules. Operational
boundary conditions were determined with respect to continuous operation in the absorber and batch-wise operation of the absorption liquid.
The current status of the Vitrisol process is scavenger-like application.
Cu2+, the active compound in the absorption liquid, becomes depleted during H2S removal. It must be noted, however, that nowadays copper is an
expensive commodity; therefore increasing amounts of H2S lead to increasing operational costs. In order to reduce the operational costs for large
amounts of H2S and/or large-scale applications, a regeneration step was
developed to replenish Cu2+. The regeneration step is based on an operation encountered in copper ore processing, i.e., the dissolution of CuS with
ferric sulphate (Fe2(SO4)3) [8, 9]. Copper sulphate, elemental sulfur (S°)
and ferrous sulphate (FeSO4) are produced in this process:
Fe(SO4)3 + CuS(s)
to:
2FeSO4 + CuSO4 + S°(s)
(6.2)
Ferrous sulphate can be reoxidized to ferric sulphate with O2 according
4FeSO4 + 2H2SO4 + O2
2Fe2(SO4)3 + 2H2O(6.3)
Resulting in the overall net reaction for the removal of H2S:
H2S + 0.5O2
S°(s) + H2O(6.4)
For the development of the regeneration process, the reaction behavior of the parallel reactions occurring during the dissolution of CuS, i.e.,
Reactions 2 and 3, respectively, were investigated.
Wermink and Versteeg [7, 10] studied the oxidation of ferrous ions in
acidic sulphate solutions (Reaction 3), and proposed kinetic equations
derived by using both the data obtained for the initial reaction rates and
the experimentally determined Fe2+ concentration profiles, respectively.
Furthermore, Wermink and Versteeg [11] investigated the behavior of
the oxidation of ferrous ions in acidic sulphate solution, in the presence of
Cu2+. It was concluded that Cu2+ enhanced the oxidation rate of Fe2+; however, some of the experiments were affected by the rate of mass transfer of
oxygen.
Besides Fe2+ oxidation, Wermink and Versteeg [7, 12] studied the dissolution reaction of CuS with Fe3+ (Reaction 2). Representative samples of
94 Carbon Dioxide Capture and Acid Gas Injection
CuS, obtained from VitrisoI pilot absorber operations [13], were used in
the study. It was concluded that an increase in temperature increased the
rate of dissolution. Full conversion of CuS could be obtained, independent
of temperature.
From the above-mentioned investigations it was concluded that relatively mild conditions are required for the regeneration process, i.e., temperatures below 373 K and at pressures ranging from atmospheric to 1 MPa.
The conditions required in the regeneration process are case dependent,
e.g., on the amount of H2S to be removed.
In order to demonstrate the applicability of the Vitrisol technology,
two (conceptual) process designs of the Vitrisol process have been compared to standard amine treating processes of cases previously published
by Weiland and Hatcher [14].
6.2
Case Definition
The cases used to evaluate the process designs of the Vitrisol process and
standard amine treating processes are examples of shale gas from British
Columbia and an example of one of the gas plants built to process gas from
fields in the Barnett shale, as previously published by Weiland and Hatcher
[14] (see Table 6.1). In the simulations of the present study, the gases were
considered to be saturated with water.
Table 6.1 Case Specifications.
Case 1
Case 2
Gas
British Columbia shale
Barnett shale
H2S (ppmv)
26
750
CO2 (vol%)
1.1
2.5
CH4 (vol%)
balance
balance
T (°C)
31.8
32.2
P (MPa)
3.10
6.62
Flow (MMSCFD)
90
330
H2S removed (kg/h)
3.35
417
Vitrisol a 100% Selective Process for H2S Removal 95
In both cases the gases should be treated to pipeline quality, i.e., 4 ppmv
H2S and below 2 vol.% of CO2, according to Weiland and Hatcher [14].
For Case 1, the only compound required to be removed is H2S, because
CO2 is already below pipeline specifications. Therefore it is desirable to
select a process with the lowest possible removal of CO2. In Case 2, both
H2S and CO2 need to be treated to reach pipeline specifications. The
Vitrisol process is not able to remove CO2; therefore additional processing is required for Case 2 to obtain on spec gas. For this purpose also an
amine treating process is chosen.
6.3
6.3.1
“Amine-Treated” Cases by PPS
Introduction to PPS
Alkanolamines have been widely used for more than 80 years in the gas
treating industry, i.e., petrochemicals, refineries, natural gas processing.
Recently, formulated amines that comprise a promoter have been incorporated in gas treating and in large-scale post-combustion CO2 capture.
The acid gas treating industry mainly consists of processes where one or
more gaseous components are transferred from the gas phase to the liquid phase followed by a chemical reaction. Due to the complexity of the
absorption processes, modeling them requires very precise knowledge of
reaction ­kinetics, mass transfer, thermodynamics and physical ­properties.
In ­addition to the development of rigorous models that account for the
aforesaid phenomena, it is also important to incorporate the correct
description of vapor-liquid and liquid phase chemical equilibria including
the speciation of the various components.
A steady-state rate based flowsheeting software for the simulation of
the acid gas treating processes has been developed by Procede.15 The flowsheeting tool has models that can do the design, optimization and analysis
of acid gas treating processes including both pre- and post-combustion
CO2 capture, respectively. The process simulator consists of a user-friendly
graphical user interface (GUI) and a powerful numerical solver that can
handle rigorous simultaneous solutions of thermodynamics, kinetics,
rate-based mass transfer equations (also known as rate-based model) and
supports all unit operations involved in gas treating such as absorbers,
strippers, flash drums, heaters, pumps, compressors, mixers and splitters
as well as work flow tools such as automatic water and solvent makeup
calculators.
96 Carbon Dioxide Capture and Acid Gas Injection
The Procede Process Simulator (PPS) has extensive, carefully evaluated
databases of thermodynamic parameters, interaction coefficients, kinetics
that have been optimized to accurately predict vapor liquid equilibrium
(VLE), thermodynamic and physical properties and kinetically enhanced
mass transfer (both approximate and rigorous) for amine- and mixture
of amines-based capture processes. PPS is able to describe complete gas
treating processes involving complex flow schemes with multiple recycle
loops. Both absorber and stripper can be modeled as rate-based columns.
For optimal predictions of column performances, the program includes
databases of various commercially available tray types and a large collection of both dumped and structured packings; several mass transfer and
hydrodynamic correlations from open literature are implemented. PPS has
capabilities where users can include detailed characterization of proprietary amines, mixtures of amines, mixtures of amine and physical solvents
obtained from experiments used for the development of new gas treating
processes.
6.3.2
Process Description
The pipeline specifications of shale gas treating is to remove H2S to <4 ppmv
and CO2 down to 2.0 vol.%. As both acidic gases will be absorbed simultaneously, usually excessive amounts of CO2 are removed. The major challenge is
to treat the stripper gas, which also has considerable amounts of CO2 which
might be of inferior quality for a Claus plant.14 It is important to note that
the H2S absorption process is usually mainly gas-phase resistance controlled
and CO2 absorption is liquid-phase resistance controlled. Traditionally
N-methyldiethanolamine (MDEA) has been the solvent of choice in terms
of cost and effectiveness. The gas treating plant consisted a/o of an absorber,
stripper, lean-rich heat exchanger, lean solvent cooler, reboiler, condenser
and pumps. The absorber and stripper were modeled using the Higbiepenetration model for mass transfer. Structured packings were used in the
absorber and the stripper and the absorber section is operated counter currently. Aqueous 50 wt.% MDEA solution was selected as the solvent, for Case
2a and a stripper promoter was added. For Case 1, the solvent circulations
rate was 20 m3/h and for Case 2 and Case 2a, it was 100 m3/h.
The dimensions of the absorber and stripper columns are presented in
Table 6.2; structured packing Sulzer MellaPak Plus with a geometric surface of a = 250 m2/m3 was used. For the calculations of the mass transfer
parameters the correlations of Brunazzi and Paglianti [16] for the specific
area, Bravo et al. [17] for the liquid phase mass transfer coefficient and
Bravo et al. [18] for the gas phase mass transfer coefficient, respectively.
Vitrisol a 100% Selective Process for H2S Removal 97
Table 6.2 Case specifications.
Case 1
Case 2
Case 2a
British Columbia
shale
Barnett shale
Barnett shale
Absorber
Packing
Sulzer MellaPak
Plus
Sulzer MellaPak
Plus
Sulzer MellaPak
Plus
Height (m)
12
25
25
Diameter (m)
1
1.35
1.35
Sulzer MellaPak
Plus
Sulzer MellaPak
Plus
Sulzer MellaPak
Plus
3
10
10
1.22
1.5
1.5
Stripper
Packing
Height (m)
Diameter (m)
6.3.3
PFD
6.3.4
Results
6.3.4.1 Case 1
For the process conditions given in Table 6.1 and dimensions given in
Table 6.2 the H2S-specification could be met. As a consequence of the H2S
removal also CO2 was (partly) removed from 1.10 vol.% to 0.98 vol.%. The
stream with a flow of 128 Nm3/h leaving the stripper (stream 6 in Figure 6.1)
consists of 97 vol.% CO2 and 1.75 vol.% H2S and the stream that needs further processing. The H&M balances of Case 1 are given in Appendix A.
Next, the utility consumptions were calculated. For pumps, an e­ fficiency
of 80% was assumed. Heating can be performed with low-pressure steam.
Cooling is performed with either air (above 40 °C) or water (below 40 °C).
For the heat integration lean-rich heat exchanger (E-101), a temperature
approach 5 °C was assumed. Though MDEA has a low volatility, some
MDEA make-up is required. The utility consumptions are given in Table 6.3.
6.3.4.2 Case 2
For the process conditions given in Table 6.1 and dimensions given in
­Table 6.2, both the H2S-and CO2-specification could not be met. Therefore
98 Carbon Dioxide Capture and Acid Gas Injection
4
Gas outlet
CO2
6
10
Formulator-1
16
5
13
7
8
C-102
P-101
15
C-101
14
E-101
1
Gas inlet
Flash gas
9
12
11
18
V-101
3
Pump-1
Figure 6.1 PFD of the amine process with regeneration.
Table 6.3 Utility consumptions of cases 1 and 2 of the amine process.
Case 1
Case 2a
British Columbia shale
Barnett shale
Electrical power (kW)
22.4
239
Heating duty (kW)
569a
5191a
Cooling duty (kW)
371b, 0c
2892b, 0c
0.081
0.86
MDEA (kg/h)
Reboiler duty
Cooling duty above 40 °C
c
Cooling duty below 40 °C
a
b
a stripping promoter has been added. In this particular case, H3PO4 with
a concentration of 0.3 wt.% was used. It turned out that meeting the
H2S-specification was the process limiting step, in the simulations an
­outlet concentration was attained of 3.92 ppmv. The CO2-specification of
2.0 vol.% was easily realized, the simulations gave an outlet concentration
Vitrisol a 100% Selective Process for H2S Removal 99
of 1.83 vol.%, so about 35% more CO2 was removed that was demanded.
The stream (gas) with a flow of 2175 Nm3/h leaving the stripper consists
of 90.1 vol.% CO2 and 9.9 vol.% H2S, a gas outlet stream that certainly
demands further processing. The H&M balances of Case 2a are given in
Appendix B. The utility consumptions are also given in Table 6.3.
It can be concluded that amine gas treating processes can be used for the
upgrading of Cases 1 and 2 to meet the desired specifications. However,
the obtained off-gases need further processing. The composition of the offgases is such that treating with a Claus-operation is not feasible.
6.4 Vitrisol Process Extended with Regeneration of
Active Component
First an in-depth description of the Vitrisol process will be provided,
which is necessary to understand process designs, the H&M balances and
equipment designs for Cases 1 and 2, respectively.
6.4.1
Technology Description
For medium- and large-scale H2S removal operations, as e.g., the described
shale gas Cases in Table 6.1, the Vitrisol process will be operated with a
regeneration section in order to minimize operational costs (i.e., the consumption of copper). The Vitrisol process with regeneration basically
consists of three steps, i.e., the absorption section, the regeneration section,
and the sulphur recovery section, respectively.
6.4.2 Parameters Determining the Process Boundary
Conditions
The parameters that have a substantial impact on the process design of the
Vitrisol process with regeneration because of a/o issues with stable operation are:
1. The wt.% of copper sulphide (CuS) in the aqueous absorption liquid exiting the absorber,
2. The concentration of copper sulphate (CuSO4) in the aqueous absorption liquid exiting the absorber.
From experiments performed with the previously mentioned Vitrisol
pilot absorber13 it was concluded that the rheological and hydraulic
100 Carbon Dioxide Capture and Acid Gas Injection
behavior of the aqueous solution abruptly changed above concentrations
of 3 wt.% to 5 wt.% of CuS. Furthermore, it was observed that the addition
of Fe2+ to the absorption liquid (containing CuSO4 and H2SO4) did not
affect the H2S removal efficiency in the absorber.
In the conceptual design exit concentrations of CuS of 0.25 wt.% and
1.0 wt.% were chosen for Cases 1 and 2 respectively. Flexibility and turndown options are introduced into the design with respect to operation by
not designing the process near the maximum CuS concentration. E.g.:
1. Sudden spikes in H2S concentration will not affect the operability of the process significantly
2. A consistent, temporary increase in H2S concentration of
the gas in time will not require significant alteration of the
installed hardware (a phenomena often encountered with,
e.g., shale gas)
From the work by Ter Maat et al. [4–6] it can be concluded that the
reaction rate of H2S with Cu2+ is instantaneous compared to mass transfer rate, even at very low concentrations of Cu2+. Therefore, the removal
of H2S is not a function of the concentration of Cu2+ in the absorption
liquid.
Moreover, from the work by Wermink and Versteeg [7, 10, 11] regarding
Fe2+ oxidation experiments performed in the presence of Cu2+ it ­followed
that the presence of Cu2+ increased the conversion rate of Fe2+ with O2
to Fe3+ substantially compared to Fe2+ oxidation experiments ­performed
­without Cu2+. From the Fe2+ oxidation experiments performed in the presence of Cu2+ it was concluded that the presence of sulphuric acid (H2SO4)
did not affect the conversion rate of Fe2+ to Fe3+, and that an increase in Fe2+
concentration increased the conversion rate of Fe2+ to Fe3+ more significantly than an increase in Cu2+ concentration.
From the experimental work it was concluded that the dissolution of
CuS, in the presence of only Fe2+, Cu2+, H2SO4, and O2 (no Fe3+ initially),
­followed an initial conversion rate of CuS equal to the conversion rate
of Fe2+ with O2 in the presence of Cu2+ up to roughly 80% conversion.
Therefore it was concluded that:
1. Up to a CuS conversion of roughly 80% the oxidation of Fe2+
is the rate determining step in a batch reactor
2. It is desirable to increase the Fe2+ concentration instead of
the Cu2+ concentration to obtain a higher conversion rate of
Fe2+ in the oxidation step of the process
Vitrisol a 100% Selective Process for H2S Removal 101
Concentrations of CuSO4, FeSO4 and H2SO4 of 0.05 M, 0.75 M and
0.1 M were chosen, respectively, for the absorption liquid exiting the
absorber.
6.4.3
Absorption Section
The absorption of H2S is completely limited by mass transfer in the gas
phase and therefore first order with respect to the removal of H2S. This
implies that most of the CuS solids are precipitating in the first section of a
G/L contactor. Therefore in the selection of the absorber type a combination has been chosen of a packed bed and bubble column. Deep removal
is realized in the packed bed section while the gas-liquid disengagement
zone below the packed bed section is designed as a bubble column. This
bubble column section also acts as a liquid storage vessel.
The bubble column section was designed based on:
1. The specific interfacial areas of the bubble column for
Cases 1 and 2 were determined to be a = 190 m2/m3 and
a = 250 m2/m3, respectively, see Oyevaar [19].
2. The gas hold-ups of the bubble column for Cases 1 and 2
were determined to be = 0.35 and = 0.37, respectively, see
Oyevaar [19].
3. The gas phase mass transfer coefficient was calculated using
the approach of Colombet [20]; however, from this work
it is not evident how the effect of process pressure must be
taken into account as the gas phase diffusion coefficient is
inversely proportional to the pressure.
4. The gas phase mass transfer coefficient was corrected for the
pressure in two manners, i.e., according to the film theory
and Higbie’s penetration theory, respectively. According to
film theory, the gas phase diffusion coefficient is inversely
proportional with pressure, whereas the gas phase diffusion
coefficient is inversely proportional with the root of pressure
according to penetration theory.
5. The actual superficial velocity of the gas in the bubble column was set to vs,G = 0.15 m.s–1.
Although the absorption of H2S into the Vitrisol liquid is irreversible
and therefore no real preference exists for counter-current operation, this
mode has been selected in the present designs.
102 Carbon Dioxide Capture and Acid Gas Injection
The absorber column was designed based on:
1. As packing 1-inch Pall rings were selected. The effective
interfacial area was calculated according to the correlations
of Billet and Schultes [12].
2. The gas phase mass transfer coefficient was calculated
according to the correlations of Onda et al. [22].
3. The absorber operation was designed to reach H2S pipeline
specification independent of the bubble column operation.
In Case 2, the pressure of the Vitrisol liquid exiting the absorber is
flashed to the pressure in the oxidizer reactor. The flash vessel is designed
for a residence time of half an hour of the liquid. The flash vessel is stirred
to maintain a homogeneous solid-liquid solution. In Case 1 the flash vessel
is omitted, because CH4 losses are negligible.
Prior to entry in the oxidizer, heat is exchanged between the Vitrisol
liquid leaving the absorption section and regenerated Vitrisol liquid
entering this section.
6.4.4
Regeneration Section
From the work by Wermink and Versteeg12 it was concluded that at all the
temperatures investigated, i.e., temperatures ranging from 25 °C to 90 °C,
a conversion of CuS of approximately 100% could be obtained. Moreover,
with temperature the rate of dissolution of CuS increased. For the components Fe3+ and H2SO4 both zero order dependencies were observed.
Based on CuS dissolution experiments at 90 °C, in the presence of
CuS, FeSO4 and O2, it was concluded that the first 80% conversion of CuS
occurred at a rate identical to the initial conversion rate of Fe2+ with O2 in
the presence of Cu2+. The latter 20% could be fully converted, but required
an additional residence time of approximately one hour. Furthermore, at a
level of 80% conversion of CuS sufficient amounts of Fe3+ were produced to
convert the remaining 20% of the CuS. This was independently confirmed
experimentally in which the O2 supply was stopped at a conversion level of
80%. Therefore it was concluded that continuous air flow was not required
to fully dissolve the CuS and retain the Cu2+.
From oxidation experiments of Fe2+ with oxygen in the presence of Cu2+
it was concluded that some of the performed experiments were affected by
mass transfer of O2 in the liquid.11 Therefore intrinsic kinetics could not be
determined exactly. As the laboratory scale reactor was designed to have
very high mass transfer rates, it can be concluded that for process scale
Vitrisol a 100% Selective Process for H2S Removal 103
types, with substantially lower kL and a values, the mass transfer of O2 to
the liquid is the rate determining step of the CuS dissolution. Therefore the
design of this unit is based on the absorption rate of O2.
The dissolution process of CuS will be carried out in three steps; i.e., an
oxidation step, an extraction step and a candle filter operation, respectively.
In the oxidation step a continuous pressurized air flow is fed to a gasliquid contactor with high-intensity stirring at a temperature of 90 °C and
a pressure of 1 MPa. The heat generated by the compressor can be fully
integrated in the Vitrisol process. An excess amount of O2 is present to
ensure a stable O2 partial pressure. For compressors, pumps and stirrers an
efficiency of 80% was assumed. Build-up of water can occur in the process,
as the overall reaction is given by:
H2S + 0.5S2
H2O + So(S)
Preferably, the water produced in the regeneration section is evaporated
in the absorber. However, for Case 1 the liquid flow is relatively low compared to the gas flow, resulting in a negligible temperature increase of the
gas and therefore no water evaporation. Consequently, for Case 1 the compressor was designed to provide a larger flow to both ensure a stable O2
pressure and evaporate sufficient amounts of H2O. For Case 2 the process
has been designed to evaporate the excess water in the absorber; therefore
the compressor is designed to only ensure a stable O2 pressure.
The basis of design of the oxidizer is the relation by Van ‘t Riet23 derived
for kLa in stirred vessels with ionic aqueous solutions. A power to volume
ratio of the stirrer of 2000 W/m3 is selected as design parameter, resulting
in a kLa of 0.15 s–1 and 0.22 s–1 for Cases 1 and 2 respectively. The solubility
of O2 in the Vitrisol solution was determined with an adapted version of
the model by Weisenberger and Schumpe [11, 24].
Subsequently the Vitrisol liquid enters the extraction step, i.e., a liquid-liquid contactor with high-intensity stirring also at a tWr is dissolved
and the dissolution of CuS proceeds in the absence of air. The amount
of xylene entering the extractor is equal to 90% of the maximum solubility of sulphur in xylene between a temperature of 50 °C and 90 °C to
be able to dissolve an amount of sulphur equal to the sulphur present in
the CuS formed in the absorber. The residence time in the liquid-liquid
contactor of τ = 22.5 min was determined from experiments dissolving
sulphur in xylene. A residence time of approximately 1 h for full conversion of CuS is not required, because the extraction step is operated continuously instead of batch-wise, resulting in increased conversion rates of
CuS. A candle filter operation is placed subsequent to the extraction step
104 Carbon Dioxide Capture and Acid Gas Injection
to prevent possible slip of unconverted CuS to the crystallization section,
because modes of operation could be possible that result in a conversion of
CuS of below 100%.
Subsequent to the candle filter operation the solids-free Vitrisol liquid and xylene enter a settler to separate both phases. A residence time of
15 min is selected to separate both phases. The Vitrisol liquid is returned
to the absorber and xylene, partially saturated with dissolved sulphur,
enters the crystallizer in the sulphur recovery section at a temperature of
90 °C. The dissolved xylene amount in the Vitrisol liquid is determined
from the solubility of p-xylene in water.25 Xylene, dissolved in the Vitrisol
liquid, is stripped in the absorber. Xylene concentrations in the gas leaving the absorber in Cases 1 and 2 are determined to be 2.0 ppmv and
16.6 ppmv, respectively.
6.4.5
Sulphur Recovery Section
In the crystallizer the xylene with the dissolved sulphur is decreased in temperature from 90 °C to 50 °C, resulting in the formation of sulphur crystals.
Sulphur solubilities and dissolution times in various organic solvents, e.g.,
toluene, p-xylene, m-xylene and o-xylene, were experi­mentally determined
and used in the design of the crystallizer. Afterwards, the xylene containing
dissolved sulphur as well as crystallized sulphur is treated in a vacuum belt
filter to separate the sulphur crystals from the liquid. The dry matter content
of the sulphur crystals was assumed to be 70 wt.%. Xylene with dissolved
sulphur is returned to the extraction step, sulphur cake is further processed
in a sulphur melting operation. Sulphur cake is heated till the boiling point
of xylene in the sulphur melting operation, i.e., a temperature of 140 °C.
Xylene evaporates and is returned to the extraction step after condensation.
99+% liquid Sulphur is stored in a storage tank. A make-up stream of xylene
is fed to the extraction step because of xylene losses in the absorber.
It should be mentioned that the sulphur recovery section could be performed differently, when continuous removal of sulphur is not required,
and the amount of sulphur to be removed is rather low (like e.g., Case 1).
A saturation vessel could be included in the process design, i.e., a relatively
large stirred vessel containing xylene with a sulphur storage capacity equal
to a couple of weeks to months of operation. A possible mode of operation
would be to remove (part of) the saturated xylene, and crystallizing the
sulphur from the saturated xylene in vessel(s) in contact to the surroundings. Subsequently, the lean xylene can be returned to the Vitrisol process,
and the solid sulphur removed. In this type of process the use of continuous cooling and sulphur removal is not required.
Vitrisol a 100% Selective Process for H2S Removal 105
6.4.6
CO2-Absorber
For Case 2 the CO2 concentration is too high to meet the pipeline specifications (see Table 6.1). Therefore an additional CO2 removal technology
must be applied. For the CO2 removal step also an amine-based technology is chosen. It must be noted, however, that after the Vitrisol process
the H2S concentration is already 4 ppmv (or lower as will be discussed
below) and therefore intrinsically no need exists to use a selective amine
as MDEA. Moreover, the gas produced by the stripper is 99+% pure CO2
which can be reused or directly vented to the environment depending on
the local regulations. Initially, an aqueous 50 wt.% MDEA (Case 2b) solvent has been selected for the removal of the CO2 in order to meet the
specifications. Also, simulations by PPS have been carried out for an aqueous 30 wt.% MEA (Case 3) solution. For both Cases 2b and 3a, the solvent
circulation rate was 55 m3/h, respectively.
6.4.7
PFD
The PFD of the Vitrisol process with regeneration for Case 1 is shown in
Figure 6.2.
The PFD of the Vitrisol process with regeneration for Case 2 is shown
in Figure 6.3a.
The PFD of the CO2 removal process for Case 2 is shown in Figure 6.3b.
6.5
Results
The H&M balances of the Vitrisol process for Cases 1 and 2 are given in
Appendices D and E, respectively.
The utility consumptions are given in Table 6.4.
Because the Vitrisol process can be operated at relatively mild conditions, heating can be performed with low-pressure steam. Cooling can be
performed with cooling water and/or air, because cooling duties above a
temperature of 40 °C are required.
Xylene make-up is required because xylene losses arise from xylene
exiting the process in the absorber. Xylene exiting the absorber is not considered to be a loss, because xylene is a component frequently encountered
in natural gas. Xylene concentrations in the gas leaving the absorber in
Cases 1 and 2 are determined to be 2.0 ppmv and 16.6 ppmv, respectively.
The utility consumptions for the removal of CO2 from the Vitrisol treated gas stream of Case 2 are given in Table 6.5b as Cases 2b and 3,
1
Gas inlet
C-101
T-103
3
E-101
P-105
4
21
E-102
7
P-xylene
5
K-101
R-101
M-101
6
14
E-103
T-102
9
Air
10
Figure 6.2 PFD of the Vitrisol process with regeneration for Case 1.
2
Gas outlet
Chemical
makeup
8
P-104
11
19
E-105
V-101
M-102
Air
20
S-101
P-101
P-103
12
13
18
S-102
24
P-106
17
15
T-101
23
S-105
22
P-102
16
99+% sulphur
E-106
E-107
S-103
S-104
E-104
106 Carbon Dioxide Capture and Acid Gas Injection
1
Gas inlet
6
4
23
E-102
9
16
R-101
M-102
P-xylene
7
K-101
8
9
E-103
T-102
Air
10
8
P-104
13
21
E-105
V-102
M-103
Air
22
20
S-101
P-101
P-103
14
15
Figure 6.3a PFD of the Vitrisol process with regeneration for Case 2 (H2S removal section).
V-101
5
E-101
P-105
Flash gas
3
C-101
T-103
M-101
2
Gas outlet
Chemical
make up
S-102
26
P-106
19
17
T-101
25
S-105
24
P-102
16
99+% sulphur
E-106
E-107
S-103
S-104
E-104
Vitrisol a 100% Selective Process for H2S Removal 107
1
4
C-101
10
12
Flash gas
9
Formulator-1
Figure 6.3b PFD of the CO2 removal process for Case 2.
Gas inlet
Gas outlet
3
V-101
11
8
Pump-1
P-101
16
E-101
7
18
15
14
C-102
5
13
6
CO2
108 Carbon Dioxide Capture and Acid Gas Injection
Vitrisol a 100% Selective Process for H2S Removal 109
Table 6.4 Utility consumptions of Cases 1 and 2 of the Vitrisol process.
Case 1
Case 2
British Columbia shale
Barnett shale
Electrical power (kW)
11.4
784
Heating duty (kW)
18.9
520
Cooling duty (kW)
a
1.4
178a
0.94
28.3
Xylene (kg/h)
Cooling duty above 40 °C
a
Table 6.5a Absorber and stripper dimensions.
Case 2b
Case 3
Barnett shale
Barnett shale
Absorber
Packing
Sulzer MellaPak Plus
Sulzer MellaPak Plus
25
17
1.35
1.38
Sulzer MellaPak Plus
Sulzer MellaPak Plus
Height (m)
10
10
Diameter (m)
1.5
1.5
Height (m)
Diameter (m)
Stripper
Packing
Table 6.5b Utility Consumptions of Cases 2b and 3 of the Amine Process.
Electrical power (kW)
Barnett shale
Barnett shale
129
132
2140
2060a
Cooling duty (kW)
607b, 0c
626b, 0c
0.79d
0.50e
Reboiler duty
Cooling duty above 40 °C
c
Cooling duty below 40 °C
d
MDEA
e
MEA
b
Case 3
Heating duty (kW)
Amine (kg/h)
a
Case 2b
a
110 Carbon Dioxide Capture and Acid Gas Injection
respectively. From Tables 6.3 and 6.5b it can be observed that the utility
consumption, especially the reboiler duty, has considerably reduced when
H2S is removed upstream of the amine plant. Equipment specifications are
given in Table 6.5a.
From the simulations no real preference for either two amines can be
made based on operational costs. The dimensions of the MEA-based are
slightly smaller compared to MDEA. In the comparison the data obtained
for MDEA will be used.
6.6
6.6.1
Discussion
Comparison of Amine Treating Solutions to Vitrisol
The total utility consumptions for processing British Columbia shale to
pipeline quality with standard amine treating solutions and Vitrisol are
given in Table 6.6.
From Table 6.6 it can be concluded that significant reductions with
respect to utility consumptions can be achieved when Vitrisol is used to
remove H2S instead of a standard amine process. Furthermore, the H2S
Table 6.6 Utility consumptions of cases 1 and 2 of the amine process.
Standard amine process
Vitrisol® process
22.4
Heating duty (kW)
569a
Cooling duty (kW)
371 , 0
b
11.4
18.9
c
1.4b, 0c
0.081
Xylene (kg/h)
0.94
Xyleneabsorber gas (ppmv)
2.0
CO2,stripper gas (kg/h)
239
H2Sstripper gas (kg/h)
3.35
H2Sstripper gas (ppmv)
1.65 × 104
Reboiler duty
the cooling duty above 40 °C
c
the cooling duty below 40 °C
b
Case 1
Electrical power (kW)
MDEA (kg/h)
a
Case 1
Vitrisol a 100% Selective Process for H2S Removal 111
Table 6.7 Utility consumptions of cases 1 and 2 of the Amine process.
Electrical power (kW)
Case 2a
Case 2
Standard amine
process
Vitrisol process +
MDEA process
239
784 + 129
Heating duty (kW)
5191a
520 + 2140
Cooling duty (kW)
2892b, 0c
178b + 607b, 0c
0.86
0.79
MDEA (kg/h)
Xylene (kg/h)
28.3
CO2,stripper gas (kg/h)
4889
3634
H2Sstripper gas (kg/h)
417
1.79
H2Sstripper gas (ppmv)
9.24 × 104
625
Reboiler duty
Cooling duty above 40 °C
c
Cooling duty below 40 °C
a
b
is oxidized in the Vitrisol process, nullifying additional H2S removal
­ perations downstream. E.g., the stripper gas in the amine process still
o
contains 1.65 × 104 ppmv of H2S, as is shown in Table 6.6. Xylene make-up
is not considered to be a loss, because xylene is a component frequently
encountered in natural gas.
The total utility consumptions for processing Barnett shale to
­pipeline quality with standard amine treating solutions and a combination of Vitrisol with a standard amine treating solution are given in
Table 6.7.
From Table 6.7 it can be concluded that significant reductions with
respect to utility consumptions can be achieved when Vitrisol is used
to remove H2S prior to CO2 removal with a standard amine process.
Furthermore, though the H2S is oxidized in the Vitrisol process, an
additional H2S removal operation is required downstream to remove H2S
from the amine stripper gas. As described in Table 7, the quantity of H2S
to be removed from the stripper gas varies significantly between Barnett
shale treated by a standard amine process for both H2S and CO2 removal
and the combination of the Vitrisol process for H2S removal and standard
amine process for CO2 removal.
112 Carbon Dioxide Capture and Acid Gas Injection
6.6.2
Enhanced H2S Removal of Barnett Shale Gas (case 2)
As explained before, the Barnett shale gas coming from the Vitrisol process is required to be treated by a standard amine process to remove CO2.
However, the feed gas for the amine absorber, containing 4 ppmv of H2S,
will result in an increased H2S concentration coming from the amine stripper. If the H2S content of the feed gas for the amine absorber would be
decreased below 4 ppmv, the possibility exists to produce a stripper gas
which can be directly vented to the environment without additional H2S
removal. The only requirement would be to provide additional packing in
the Vitrisol absorber to enhance the removal of H2S.
Figure 6.4 shows the amount of H2S in the gas leaving the stripper for
varying of amounts of H2S in the feed gas.
Figure 6.5 shows the amount of H2S in the gas entering the amine
absorber and the gas leaving the amine stripper for varying packing heights
in the Vitrisol absorber.
From Figures 6.4 and 6.5 it can be concluded that it is advantageous to
remove more H2S upstream with the Vitrisol process than is required for
2.0
700
1.8
600
H2S out (kg/h)
H2S out (ppm)
500
H2Sout (kg/h)
1.4
1.2
400
1.0
300
0.8
0.6
200
0.4
100
0.2
0.0
0
0
1
2
3
4
H2Sfeed (ppm)
Figure 6.4 H2S content of the gas exiting the stripper as a function of the H2S
concentration in the feed gas.
5
H2Sout (ppm)
1.6
Vitrisol a 100% Selective Process for H2S Removal 113
10
1000
H2S feed
H2S out
1
100
2
4
6
8
10
0.1
10
0.01
1
0.001
H2Sout (ppm)
H2Sfeed (ppm)
0
0.1
Hpacking vitrisol absorber (m)
Figure 6.5 H2S content of gas entering the amine absorber (feed) and of gas exiting the
amine stripper (out) as a function of packing height in the Vitrisol absorber.
the gas specifications as this enables an easy handling downstream of the
amine stripper gas, e.g., venting CO2 to the environment. From Figure 6.5
it can be concluded that a packing height of 3.4 m is required to meet the
4 ppmv specifications, however, this will lead to a H2S concentration in the
CO2 off-gas from the amine stripper of 625 ppmv. In order to arrive at a
concentration of H2S of about 1 ppmv additionally about 4 m packing is
required (about 4 m3).
6.7
Conclusions
Over recent years PROCEDE developed a solvent called Vitrisol that is
100% selective for H2S removal from industrial gases in the presence of
CO2. Vitrisol is able to remove in one stage more than 99.9+% of the H2S
present in the gas phase and has the typical characteristics of very selective H2S scavengers. However, a major difference of Vitrisol compared
114 Carbon Dioxide Capture and Acid Gas Injection
to the traditional scavengers is that Vitrisol can be completely regenerated, resulting in a solvent with fully restored activity and crystalline
sulphur.
For the absorption process required for the removal of H2S, there is no
real process conditional constraint and the operating pressures and temperatures can vary at least between 0.1–10 MPa and 283–363 K, respectively. The Vitrisol regeneration process takes place at temperatures below
373 K. The process pressure can vary from atmospheric up to 0.5–1 MPa.
The Vitrisol process can be described with the following overall reaction
equation:
H2S + 0.5 O2
H2O + S°(s)
In the Vitrisol process no vast amounts of energy are required for the
regeneration of the solvent; therefore this process has an extremely low
energy footprint.
In the present contribution the performance of Vitrisol is demonstrated for two applications in shale gas production, as previously
described by Weiland and Hatcher, i.e., British Columbia gas (Case 1)
and Barnett gas (Case 2) respectively. The Vitrisol process is also compared to a standard amine treating process designed for selective H2S
removal.
From the results it can be concluded that significant reductions can be
achieved by using the Vitrisol process for the operational costs as depicted
in the energy consumptions of the overall process. E.g., in Case 1 the total
energy consumption of the Vitrisol process is 32 kW, whereas the total
energy consumption of the standard amine treating process is 962 kW.
In Case 2 the total energy consumption of the Vitrisol process in combination with a standard amine treating process is 4358 kW, whereas the
total energy consumption of the standard amine treating process is 8322
kW. Moreover, no additional treatment of the off-gas stream is required
for the process with Vitrisol upstream of the amine treating unit, as the
H2S is directly converted to crystalline sulphur. Depending on the local
governmental regulation the off-gas could be directly vented to the environment. This study also illustrates clearly that it is advantageous to first
remove H2S from a gas stream containing both H2S and CO2 prior to CO2
removal to reduce operational costs. As the costs of energy (and cooling)
are extremely location dependent, no attempt was made to quantify the
capital savings.
Vitrisol a 100% Selective Process for H2S Removal 115
6.8
a
D
e
H
k
P
T
vs
Notation
interfacial area
diameter
hold-up
height
mass transfer coefficient
pressure
temperature
superficial velocity
[m2/m3]
[m]
[-]
[m]
[m.s–1]
[Pa]
[°C, K]
[m.s–1]
Subscripts and superscripts
G
L
gas
liquid
References
1. Mcintyre, G. & Lyddon, L.; “Claus sulphur recovery options,” Petroleum
Technology Quarterly Spring, p. 57–61, 1997.
2. Perry, D. et al.; “Better acid gas enrichment,” Sulphur, vol. 326, p. 38–42, 2010.
3. Versteeg, G.F. and Ter Maat, H.; “Method and system for selective removal
of contamination from gas flows,” WO9855209, publication date 10-12-1998.
4. Maat, H. ter, “The development of a novel, selective desulfurization process,”
PhD thesis, University of Twente, 2006.
5. Ter Maat, H. et al.; “The removal of hydrogen sulfide from gas streams using
an aqueous metal sulfate absorbent. Part I. The absorption of hydrogen sulfide in metal sulfate solutions,” Separation and Purification Technology, vol. 43,
p. 183–197, 2005.
6. Ter Maat, H. et al.; “Theoretical and experimental study of the absorption rate
of H2S in CuSO4 solutions. The effect of enhancement of mass transfer by a
precipitation reaction,” Trans IChemE, Part A, Chemical Engineering Research
and Design, vol. 85(A1), p. 100–108, 2007.
7. Wermink, W.N. and Versteeg, G.F.; “The oxidation of Fe(II) in acidic sulphate solutions at elevated pressures. Part I. Kinetics above 1 M H2SO4,” paper
submitted.
8. Peacey, J. et al.; “Copper hydrometallurgy – current status, preliminary economics, future direction and positioning versus smelting,” Trans. Nonferrous
Met. Soc. China, vol. 14, no. 3, p. 560–568, 2004.
9. Dutrizac, J.E. and MacDonald, R.J.C.; “The kinetics of dissolution of covellite
in acidified ferric sulphate solutions,” Canadian Metallurgical Quarterly, vol.
13, no. 3, p. 423–433, 1974.
116 Carbon Dioxide Capture and Acid Gas Injection
10. Wermink, W.N. and Versteeg, G.F.; “The oxidation of Fe(II) in acidic sulphate
solutions at elevated pressures. Part II. Influence of H2SO4 and Fe(III),” paper
submitted.
11. Wermink, W.N. et al.; “The oxidation of Fe(II) with Cu(II) in acidic sulphate
solutions at elevated pressures,” paper submitted.
12. Wermink, W.N. and Versteeg, G.F.; “The dissolution of CuS with Fe(III) in
acidic sulphate solutions,” paper submitted.
13. Ter Maat et al., “Vitrisol a new, 100% selective technology for the removal of
H2S,” Proc. 8th International Conference of Sour Oil & Gas Advanced Technology
(SOGAT), Abu Dhabi, UAE, 2012.
14. Weiland, R.H. and Hatcher, N.A.; “Overcome challenges in treating shale gas,”
Hydrocarbon Processing, vol. 91, no. 1, p. 45–48, 2012.
15. van Elk, E.P. et al.; “A new flowsheeting tool for flue gas treating,” Energy
Procedia, vol. 1, no. 1, p. 1481–1488, 2009.
16. Brunazzi, E. and Paglianti, A.; “Liquid-film mass-transfer coefficient in a
column equipped with structured packings,” Ind. Eng. Chem. Res., vol. 36,
p. 3792–3799, 1997.
17. Bravo, J.L. et al.; “A comprehensive model for the performance of columns containing structured packings,” IChemE Symposium Series 128, p. A439–A457,
1992.
18. Bravo, J.L. et al.; “Mass transfer in gauze packings,” Hydrocarbon Processing,
vol. 64, no. 1, p. 91–95, 1985.
19. Oyevaar, M.H. and Westerterp, K.R.; “Interfacial areas and gas hold-ups in
gas-liquid contactors at elevated pressures from 0.1 to 8.0 MPa,” Chemical
Engineering Science, vol. 46, no. 5/6, p. 1217–1231, 1991.
20. Colombet, D. et al.; “Mass or heat transfer inside a spherical gas bubble at
low to moderate Reynolds number,” International Journal of Heat and Mass
Transfer, vol. 67, p. 1096– 1105, 2013.
21. Billet, R. and Schultes, M.; “Prediction of mass transfer columns with dumped
and arranged packings,” Trans. AchemE, vol. 77, Part A, p. 498–504, 1999.
22. Onda, K. et al.; “Mass transfer coefficients between gas and liquid phases in
packed columns,” Journal of Chemical Engineering of Japan 1, p. 56–62, 1968.
23. Van ‘t Riet, K.; “Review of measuring methods and results in nonviscous gasliquid mass transfer in stirred vessels,” Ind. Eng. Chem. Process Des. Dev., vol.
18, no. 3, p. 357–364, 1979.
24. Weisenberger, S. and Schumpe, A.; “Estimation of gas solubilities in salt solutions at temperatures from 273 K to 363 K,” AIChE Journal, vol. 42, no. 1,
p. 298–300, 1996.
25. Knauss, K.G. and Copenhaver, S.A.; “The solubility of p-xylene in water as a
function of temperature and pressure and calculated thermodynamic quantities,” Geochimica et Cosmochimica Acta, vol. 59, no. 12, p. 2443–2448, 1995.
Vitrisol a 100% Selective Process for H2S Removal 117
Appendix 6-A: H&M Balance of Case 1 (British
Columbia shale) of the Amine Process
Summary
Inlet Streams
Stream #
–
1
Stream Name
–
Stream
From Block
–
Gas Inlet
Pressure (absolute)
bar
31.0000
Temperature
°C
31.8000
Vapour fraction (molar)
–
1.0000
Flow (molar)
kmol/hr
4481.4208
Flow (mass)
kg/hr
7.3292E+04
Flow (Volume)
m3/hr
3439.7550
Carbon dioxide
mol%
1.0977
Hydrogen sulfide
mol%
2.5946E-03
Water
mol%
0.2078
MDEA
mol%
0.0000
Nitrogen
mol%
0.0000
Methane
mol%
98.6919
Ethane
mol%
0.0000
Propane
mol%
0.0000
n-Butane
mol%
0.0000
n-Pentane
mol%
0.0000
118 Carbon Dioxide Capture and Acid Gas Injection
Outlet Streams
Stream #
–
4
6
9
Stream Name
–
Stream
Stream
Stream
To Block
–
Gas Outlet
CO2
Flash Gas
Pressure (absolute)
bar
31.0000
1.5000
1.7500
Temperature
°C
36.0907
45.0000
32.2132
Vapour fraction
(molar)
–
1.0000
1.0000
1.0000
Flow (molar)
kmol/hr
475.5915
5.9776
0.7981
Flow (mass)
kg/hr
7.3045E+04
250.7511
12.9991
Flow (Volume)
m3/hr
3494.9809
104.6327
11.5333
Carbon dioxide
mol%
0.9773
91.1097
0.6938
Hydrogen sulfide
mol%
3.9259E-04
1.6469
2.3215E-02
Water
mol%
0.2201
6.4914
2.3448
MDEA
mol%
1.4902E-05
1.2999E–08
7.2078E-05
Nitrogen
mol%
0.0000
0.0000
0.0000
Methane
mol%
98.8022
0.7520
96.9381
Ethane
mol%
0.0000
0.0000
0.0000
Propane
mol%
0.0000
0.0000
0.0000
n-Butane
mol%
0.0000
0.0000
0.0000
n-Pentane
mol%
0.0000
0.0000
0.0000
Vitrisol a 100% Selective Process for H2S Removal 119
Appendix 6-B H&M Balance of Case 2a (Barnett shale)
of the Amine Process with Stripper Promoter
Summary
Inlet Streams
Stream #
–
1
Stream Name
–
Stream
From Block
–
Gas Inlet
Pressure (absolute)
MPa
6.6200
Temperature
°C
32.2000
Vapour fraction (molar)
–
1.0000
Flow (molar)
kmol/hr
1.6432E+04
Flow (mass)
kg/hr
2.7536E+05
Flow (Volume)
m3/hr
5534.3376
Carbon dioxide
ppmv
2.4964E+04
Hydrogen sulfide
ppmv
748.9186
Water
ppmv
1441.8694
MDEA
ppmv
0.0000
H3PO4
ppmv
0.0000
Nitrogen
ppmv
0.0000
Methane
ppmv
9.7285E+05
Ethane
ppmv
0.0000
Propane
ppmv
0.0000
n-Butane
ppmv
0.0000
n-Pentane
ppmv
0.0000
120 Carbon Dioxide Capture and Acid Gas Injection
Outlet Streams
Stream #
–
4
6
9
Stream Name
–
Stream
Stream
Stream
To Block
–
Gas Outlet
CO2
Flash Gas
Pressure (absolute)
MPa
6.6200
0.1500
0.1750
Temperature
°C
43.5532
45.0000
33.2011
1.0000
1.0000
1.0000
Vapour fraction (molar) –
Flow (molar)
kmol/hr 1.6313E+04 132.0003
6.2467
Flow (mass)
kg/hr
2.7013E+05 5461.7403
113.2937
Flow (Volume)
m3/hr
5806.9475
90.5366
Carbon dioxide
ppmv
1.8309E+04 8.4166E+05
6.8778E+04
Hydrogen sulfide
ppmv
3.9253
9.2417E+04
6841.0857
Water
ppmv
2109.6663
6.4919E+04
2.3575E+04
MDEA
ppmv
0.4414
1.2144E–02
0.7804
H3PO4
ppmv
0.0000
0.0000
0.0000
Nitrogen
ppmv
0.00000
0.0000
0.0000
Methane
ppmv
9.7958E+05 10008.3982
9.0081E+05
Ethane
ppmv
0.0000
0.0000
0.0000
Propane
ppmv
0.0000
0.0000
0.0000
n-Butane
ppmv
0.0000
0.0000
0.0000
n-Pentane
ppmv
0.0000
0.0000
0.0000
2309.9574
Vitrisol a 100% Selective Process for H2S Removal 121
Appendix 6-C H&M Balance of Case 3 (Barnett shale)
of the Amine Process (MEA)
Summary
Inlet Streams
Stream #
–
1
Stream Name
–
Stream
From Block
–
Gas Inlet
Pressure (absolute)
bar
66.2000
Temperature
°C
32.2000
Vapour fraction (molar)
–
1.0000
Flow (molar)
kmol/hr
1.6432E+04
Flow (mass)
kg/s
76.4265
Flow (Volume)
m /hr
5535.7196
Carbon dioxide
mol%
2.4964
Water
mol%
0.1441
MEA
mol%
0.0000
Methane
mol%
97.3595
3
Outlet Streams
Stream #
–
4
6
Stream Name
–
Stream
Stream
To Block
–
Gas Inlet
CO2 outlet
Pressure (absolute)
bar
66.2000
1.5000
Temperature
°C
32.2000
40.0000
Vapour fraction (molar)
–
1.0000
1.000
Flow (molar)
kmol/hr
1.6432E+04
77.7385
Flow (mass)
kg/s
76.4265
0.9216
Flow (Volume)
m3/hr
5535.7196
1339.1137
Carbon dioxide
mol%
2.4964
94.8935
Water
mol%
0.1441
4.9996
MEA
mol%
0.0000
1.3785E-08
Methane
mol%
97.3595
0.1069
122 Carbon Dioxide Capture and Acid Gas Injection
Formulated Streams
Stream #
–
18
Stream Name
–
Stream
From Block
–
Formulator-1
Pressure (absolute)
bar
66.2000
Temperature
°C
35.9375
Vapour fraction (molar)
–
0.0000
Flow (molar)
kmol/hr
2507.0022
Flow (mass)
kg/s
16.4953
Flow (Volume)
m /hr
55.0007
Carbon dioxide
mol%
3.9649
Water
mol%
85.2589
MEA
mol%
10.7762
Methane
mol%
2.2505E-22
3
11
35
29
413
70957
2165
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
H2SO4
CuSO4
FeSO4
Fe2(SO4)3
H2S
CuS
S
CH4
CO2
O2
N2
Xylene
4.0
168
264
0.9
2165
70955
9.4
0.61
45
168
n.a.
n.a.
1.62
9.4
413
45
36
3308
n.a.
0.9
413
29
45
3265
1.62
1.8
3308
n.a.
0.25
kg/hr
0
[ppm]
0.25
H2O
90
L+S
[H2S]
0
L
37.5
0
L+S
31.8
10
3.79
3.76
[wt.%]
G
32.2
31
3.62
3.76
Conc.solids
31.8
31
3.62
3.76
7
8
9
10
35
11
407
42
36
0.5
n.a.
0
G
15
1.01
39
0.047
9
0.5
n.a.
0
G
90
10
5.0
0.047
35
9
2.5
3308
n.a.
0.11
L+S
90
10
3.79
3.80
35
413
45
36
2
n.a.
0
G
90
10
5.4
0.049
Compressed
Compressed
air
Vitrisol air
Water Air
6
G
[°C]
31
3440
5
LP HT CuS
rich Vitrisol Air
4
Phase
Temperature
31
3440
[act m3/h]
[bara]
Flow
Pressure
73.3
[ton/h]
Flow
73.3
HP CuS rich HP lean
Vitrisol
Vitrisol
NG outlet
NG inlet
3
2
Stream number 1
n.a.
L
90
1.01
0
0
Vitrisol &
xylene
11
2
n.a.
0
G
90
10
54
0.049
12
55
5.1
3308
0
L
3.9
3.9
Appendix 6-D: H&M Balance of Case 1 (British Columbia shale) of the Vitrisol process
Vitrisol a 100% Selective Process for H2S Removal 123
[ppm]
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
[H2S]
H2O
H2SO4
CuSO4
FeSO4
Fe2(SO4)3
H2S
CuS
S
CH4
CO2
O2
N2
Xylene
55
5.1
413
45
36
3308
n.a.
0
90
L
[°C]
Temperature
1
3.9
[wt.%]
[bara]
Pressure
Conc.solids
[act m3/h]
Flow
0.94
413
45
36
3308
n.a.
0
L
90
31
3.8
3.8
Vitrisol
3.9
14
13
Vitrisol &
xylene
Phase
[ton/h]
Flow
Stream number
54
5.1
n.a.
0
L
90
1
0.07
0.059
Xylene
15
54
5.1
n.a.
5.5
L+S
50
1
0.07
0.059
Xylene
16
1.3
3.2
n.a.
70
S+L
50
1
3.09E-03
4.44E-03
Sulphur cake
17
19
20
21
22
52
2.0
n.a.
0
L
50
1
0.06
0.054
2.2
n.a.
0
L
15
10
2.56E-03
2.22E-03
55
2.0
n.a.
0
L
90
10
0.068
0.057
n.a.
0
L
0
0
1.3
3.2
n.a.
0
G
140
1
0.41
0.0013
Xylene
Vitrisol
Xylene make-up Xylene make-up Xylene
18
23
Xylene
24
n.a.
0
L
140
1
1.3
n.a.
0
L
140
1
1.75E-03 1.71E-03
3.15E-03 1.28E-03
Liquid
sulphur
124 Carbon Dioxide Capture and Acid Gas Injection
275.4
[ton/h]
[act m3/h]
[bara]
Flow
Flow
Pressure
G
0
[wt.%]
[ppm]
Phase
Conc.solids
[H2S]
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
kg/hr
CuS
S
CH4
CO2
O2
N2
Xylene
18050
28.3
18050
256470 256346
2.24
kg/hr
Fe2(SO4)3
H2S
419
884
kg/hr
124
1170
12619
kg/hr
FeSO4
99998
CuSO4
427
2287
427
kg/hr
kg/hr
n.a.
1
L+S
32.3
66.2
110.8
117.0
H2SO4
4
0
G
35.4
66.2
5524
3
4
5
6
7
28.3
12619
2838
1086
100074
n.a.
0
L
38.0
66.2
110.8
116.6
107.1
0.75
n.a.
0
G
32.1
10
13.4
0.108
16.7
1170
12619
884
2287
99997
n.a.
1
L+S
32.1
10
110.7
116.9
16.7
1170
12619
884
2287
99997
n.a.
1
L+S
90
10
113.4
116.9
HP CuS
LP CuS LP HT
rich
HP lean
rich
CuS rich
Vitrisol Vitrisol Flash gas Vitrisol Vitrisol
H2O
749
32.2
Temperature [°C]
66.2
5534
NG
outlet
NG
outlet
NG inlet
274.8
2
1
Stream
number
3195
979
45.0
n.a.
0
G
15
1.01
3513
4.28
Air
8
10
11
12
3195
979
45.0
n.a.
0
G
90
10
448.6
4.28
317
225
11903
2462
1086
100072
n.a.
0.46
L+S
90
10
113.5
117.0
3195
783
16.7
942
191
n.a.
0
G
90
10
457.8
4.24
n.a.
0
L
0
0
Compressed
Compressed
air
Vitrisol
air
Water
9
Appendix 6-E H&M Balance of Case 2 (Barnett shale) of the Vitrisol Process
3195
783
16.7
191
n.a.
G
90
1.01
4518
4.24
Air
13
Vitrisol a 100% Selective Process for H2S Removal 125
14
[°C]
Temperature
kg/hr
kg/hr
kg/hr
kg/hr
CO2
O2
N2
Xylene
kg/hr
CuS
kg/hr
kg/hr
H2S
kg/hr
kg/hr
Fe2(SO4)3
CH4
kg/hr
FeSO4
S
2838
12619
kg/hr
CuSO4
6708
639
1086
100072
kg/hr
kg/hr
H2SO4
n.a.
H2O
[H2S]
0
[wt.%]
[ppm]
Conc.solids
L
Phase
90
10
121.9
[act m /h]
[bara]
3
Flow
124.0
Vitrisol &
xylene
[ton/h]
Pressure
Flow
Stream
number
15
16
17
6708
639
12619
2838
1086
100072
n.a.
0
L
90
1
121.9
124.0
28.3
12619
2838
1086
100072
n.a.
0
L
90
66.2
113.6
116.6
6680
639
n.a.
0
L
90
1
8.35
7.32
Vitrisol
& xylene Vitrisol Xylene
6680
639
n.a.
5.5
L+S
50
1
8.19
7.32
Xylene
18
19
160
393
n.a.
70
S+L
50
1
0.38
0.55
Sulphur
cake
21
6520
246
n.a.
0
L
50
1
7.80
6.77
188
n.a.
0
L
15
10
0.22
0.19
Xylene
Xylene make-up
20
23
24
25
26
6708
246
n.a.
0
L
90
10
8.39
6.95
n.a.
0
L
0
0
160
n.a.
0
G
140
1
51.0
0.16
393
n.a.
0
L
140
1
0.22
0.39
160
n.a.
0
L
140
1
0.21
0.16
Vitrisol
Liquid
Xylene make-up Xylene sulphur Xylene
22
126 Carbon Dioxide Capture and Acid Gas Injection
7
New Amine Based Solvents for Acid Gas
Removal
Yohann Coulier1,2,3, Elise El Ahmar3, Jean-Yves Coxam1, 2, Elise Provost4,
Didier Dalmazzone4, Patrice Paricaud4, Christophe Coquelet3 and
Karine Ballerat-Busserolles1,2,3
Clermont Université, Université Blaise Pascal, Institut de Chimie de ClermontFerrand, Clermont-Ferrand, France
2
CNRS, UMR 6296, Institut de Chimie de Clermont-Ferrand, Clermont-Ferrand,
France
3
MINES ParisTech, PSL – Research University, CTP – Centre of Thermodynamics
of Processes, Fontainebleau, France
4
UCP, ENSTA ParisTech, Université Paris-Saclay, Palaiseau cedex, France.
1
Abstract
Treatment and separation of multicomponent gases using absorption/desorption
cycles in aqueous solutions is a very well-known and efficient method, used for
natural gas and biogas purification or greenhouse gas mitigation. More specifically, aqueous solutions of amine are used with efficiency for CO2 removal from gas
mixtures. However, the large energetic cost involved with carbon capture ­processes
is a critical downside. To address this issue, a new class of amine is considered to
decrease the cost of the regeneration: the demixing amines. These amines present a lower critical solution temperature that can be used with benefit in post-­
combustion processes. Precedent studies conducted on methylpiperidines in water
have shown the difficulty of obtaining “ideal” absorbent systems. The aim of this
chapter is to evaluate the influence of a physical absorbent, namely triethylene
glycol, on the thermodynamic properties (such as liquid-liquid equilibria, vaporliquid equilibria, heat capacities, densities and heat of solutions) of aqueous solutions of demixing amine in order to design new operation units for carbon capture
process.
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (127–146)
2017 © Scrivener Publishing LLC
127
128 Carbon Dioxide Capture and Acid Gas Injection
7.1
Introduction
Chemical absorption of acid gases by amine based solvents has found
applications in a wide variety of industries including gas processing and
the removal of CO2 from synthesis gas in the production of hydrogen or
ammonia. Other applications of this technology are the purification of
biogas [1] or CO2 removal from post-combustion gases in power plants
[2]. The principle of the technology applied to carbon capture processes is
based on selective absorption/desorption cycles of CO2 in aqueous absorbents. A schematic view of the process is shown in Figure 7.1. Although the
process is well adapted and extensively used for natural gas treatment, the
composition of the industrial effluents leads to a loss of energetic efficiency.
CO2storage
or valorisation
Lean solvent
purety: 99%
Stripper
Treated fume
CO2 Elimination > 90%
Heat exchanger
Absorber
Industrial effluent
p = 1 atm/T = 40 °C
3–30% CO2
T = 100–140 °C
Reboilling
CO2 Rich solvent
(a)
CO2
CO2 lean amine
Treated gas
Water +
amine
Decanter
(b)
Water +
amine +
CO2
Water
Stripper
Flue gas
Absorber
Water + CO2
Figure 7.1 Schematic representation of the CO2 separation process [3]; (a) classical
alkanolamine based absorbents; (b) demixing solvents.
New Amine Based Solvents for Acid Gas Removal 129
The major problem is the cost of the regeneration step that requires a lot of
energy to be efficient. It is thus necessary to adapt this process in order to
reduce the energetic cost of the desorption step.
Demixing solvents were proposed as an option for CO2 capture to reduce
the energy consumption involved in the regeneration of the absorbent [4].
These new absorbent solutions are constituted of amines that are partially
miscible with water, under specific conditions of temperature and gas loading [5]. As shown in Figure 7.1b, in the absorber, the aqueous solution of
amine remains monophasic and a large quantity of CO2 is absorbed similarly to the process using monoethanolamine (MEA). By increasing the
temperature in the decanter, the solution separates in two liquid phases, one
amine phase containing almost no CO2, and one aqueous phase containing chemically and physically absorbed CO2. Since the solubility of the gas
in the aqueous phase is smaller than in the original monophasic solution
due to composition and temperature changes, the CO2 in excess desorbs
from the solution while the remaining CO2 is contained in the water rich
phase. As a result, only this part of the solvent is heated in the regeneration
step of the separation process. The excess CO2 from the decanter and the
separated CO2 from the stripper are then compressed and transported for
being used or stored in safe conditions.
In order to apply such a process, two important parameters have to be
considered:
The liquid-liquid phase separation should only occur in the
decanter and has to be avoided in the absorber. For that purpose, the temperature of phase separation needs to be bigger
than the maximum temperature in the absorption column.
The amine rich phase should contain as little water as p
­ ossible.
The process is efficient if most of the CO2 not released
in the decanter remains dissolved in the aqueous phase. Thus,
the amine phase to be directly recycled in the absorber.
The DACOOTA project presented by Ballerat-Busserolles et al. [6] and
Fandino et al. [7] deals with the understanding of thermodynamic equilibria in {amine + H2O} and {CO2 + amine + H2O} systems which exhibit partial miscibility with water. This project is simultaneously supported by the
French National Agency of Research (ANR, [ANR-12-IS09-0001]) and the
Natural Sciences and Engineering Research Council of Canada (NSERC).
The goal of this research project is to elucidate the structure-property relationships for the potential amines under investigation, determine phase
diagrams with or without dissolved CO2, develop thermodynamic models,
130 Carbon Dioxide Capture and Acid Gas Injection
and evaluate the capabilities of the selected solvents for CO2 absorption. In
this project, methods to determine liquid-liquid equilibria (LLE) in mixtures containing a well-controlled quantity of gas dissolved were developed
in order to elucidate part of the questions concerning this process.
In recent years, the addition of a physical solvent in aqueous solutions
of amines was considered to optimize some steps of the process [8]. For
example, in order to prevent equipment corrosion in processes of CO2 capture with aqueous amines solutions, the use of a co-solvent such as glycol
has already been explored [9]. Benefits due to the replacement of a part of
the water by a physical solvent are the reduction of the specific heat capacity of the absorbent, together with the decrease of amine degradation and
the reduction of evaporation, lowering the cost of the separation process.
In order to design new operation units for CO2 removal or to evaluate
the retrofits of existing processes, it is important to investigate the thermophysical properties of the new demixing solvents containing physical
co-solvents. This includes phase equilibrium measurements (vapor-liquid
and liquid-liquid equilibria), as well as the study of transport and energetic
properties. The knowledge of these thermophysical properties will allow the
evaluation of the impact of addition of physical solvent on CO2 mass transfer. Moreover, CO2 gas stream is not pure and contain other chemicals such
as N2, Ar, NOx, and SO2 in the case of post-combustion capture process or
H2 and SO2 in case of pre-combustion process, and H2S and mercaptans
in case of gas processing or biogas purification. The impacts of these other
chemicals on the thermophysical properties and phase diagram need also
to be investigated.
In this work, the thermodynamic properties of a new ­
demixing
solvent composed of an aqueous solution of piperidines, namely
N-methylpiperidine (NMPD) or 2-methylpiperidine (2MPD), and a
physical solvent, triethylene glycol (TEG), are reported. Relying on the
thermodynamic representation of the process [6], the benefit of adding a
co-solvent were investigated as follows:
For the decantation step, the liquid-liquid equilibria of
{Amine – H2O – TEG} systems with dissolved CO2 were
studied.
For transport properties in the lines and energy cost of
the heating, densities and heat capacities of solutions were
investigated at different temperatures.
For solvent recycling and evaporation concerns, vapor-­
liquid equilibra (VLE) measurements for different CO2 loadings were performed on {Amine – H2O – TEG} systems.
New Amine Based Solvents for Acid Gas Removal 131
For energetic aspects of absorption and regeneration, the
enthalpies of solution of CO2 in {Amine – H2O – TEG} were
determined.
A comparative and comprehensive study to determine the positive
effects coming from the addition of a physical solvent on the demixing
solvent is proposed for all the investigated properties.
7.2
Chemicals and Materials
N-methylpiperidine, 2-methylpiperidine, and triethylene glycol were used
without further purification. Water was distilled and degassed before use
(resistivity 18.2 MW·cm). Solutions were prepared by mass; uncertainty
in mass fraction (w) is estimated to be less than ± 10-4. The solutions were
stored in glass bottles in an opaque cabinet to prevent any photo-degradation. Suppliers, purities and CAS numbers of all chemicals used in this
study are given in Table 7.1.
7.3
Liquid-Liquid Equilibria
7.3.1 LLE in {methylpiperidines – H2O} and
{methylpiperidines – H2O – CO2}
The LLE of the binary systems {NMPD – H2O} and {2MPD – H2O} have
previously been studied by Coulier et al. [10] and Stephenson et al. [11].
An experimental technique recently developed by Coulier et al. [12] allows
the determination of liquid-liquid equilibria with controlled quantities of
Table 7.1 Suppliers, CAS numbers and stated purities (mass fraction w) of
chemicals used in this study.
CAS Number
w
Sigma-Aldrich
626-67-5
99.9%
a
2-methylpiperidine (2MPD)
Sigma-Aldrich
109-05-7
98.3%
Triethylene glycol (TEG)
Sigma-Aldrich
112-27-6
99.0%
Carbon dioxide (CO2)
Air Products
124-38-9
99.995%
Chemical
Suppliers
N-methylpiperidine (NMPD)
racemate
a
132 Carbon Dioxide Capture and Acid Gas Injection
dissolved CO2. The LLE data were measured by Coulier et al. [12] using the
cloud point method. It consists of determining the temperature at which a
second liquid phase appears or disappears in a liquid system.
For solutions containing dissolved CO2, two different apparatuses using
the visual determination of the temperature of phase separation were set
up depending on the range of temperatures investigated. The first apparatus is a visual phase equilibrium cell SPM20 from Thar instruments. The
equipment features a high-pressure chamber provided with pressure and
temperatures sensors and a thick sapphire window that allows the visualization of the cloud point through a camera connected to a computer.
The second cell, supplied by CTP Mines ParisTech, is fully made of sapphire, allowing the visualization of the entire sample, instead of a limited
zone. This cell is immerged in a silicon oil cooling bath to extend measurements to temperatures below 273 K. The detailed characteristics of both
­apparatuses are given in Table 7.2.
Aqueous solutions of amine loaded with controlled quantities of CO2
are prepared in a custom-made flow mixing cell. The overall experimental
arrangement of the two systems is depicted in Figure 7.2. The mixing cell
Table 7.2 Characteristics of the visual cells used for cloud point measurements.
Equilibium cell
Sapphire cell
T (K)
Room T – 393
270–393
Control of T
Heat tape
Thermostatic bath
p (MPa)
1–400
1–80
Control of p
Buffer volume
Buffer volume
Inner volume (mL)
10–20 adjustable
5
Visualization of the sample
sapphire window
Full sample
Mixing cell
High p
pump
Water + amine
Equilibrium
cell
P
High p
pump
CO2
Figure 7.2 Overall experimental setup of liquid-liquid equilibrium cells for solutions
containing dissolved gas.
New Amine Based Solvents for Acid Gas Removal 133
is built with the same structure as the one developed at ICCF for enthalpies of solution measurements [13]. The mixing point consists of a Y piece,
where two 1/16" stainless steel tubes are soldered on the top branches of
the Y, while a unique tube containing the final mixture goes out from the
bottom branch of the mixing point.
The two fluids, CO2 and the aqueous amine solution, are injected into
the mixing cell supplied by two ISCO model 100 DM high-pressure syringe
pumps. As the syringe pumps deliver constant volumetric flow rates, they
were regulated at a constant temperature of 298.15 K using a thermostatic
bath in order to calculate accurately the composition of the aqueous solutions containing dissolved gas. The system pressure is maintained constant
at 0.02 MPa using a buffer volume of 1 dm3 equipped with a back pressure
regulator and placed at the end of the flow line. The gas loading α (mol CO2
/ mol amine) of the mixture leaving the mixing unit was determined using
the molar flow rates delivered by the two syringe pumps (Eq 7.1).
nCO2
namine (7.1)
where ṅCO andṅamine are the molar flow rates of CO2 and aqueous solution
2
of amine respectively. To calculate the molar flow rates, the densities of the
aqueous solution of amine and CO2 are needed at the experimental conditions of temperature and pressure. The densities of the solution as a function of the pressure were measured using an Anton Paar densimeter DMA
HP. The densities of CO2 were calculated using the equation of state from
Span and Wagner [14]. Details on the calculation of the loading charge and
its uncertainty are found in Arcis et al. paper [13]. The relative uncertainty
on loading charge using this method is estimated to be less than 4%.
The same devices are used to measure temperature of phase separation for solutions without dissolved gas. In that case, the solutions are
directly injected in the visual cell, without using the mixing cell prior to
the entrance of the visual cell.
The procedure for the cloud point determination is the same independently of the system measured (visual isochoric method). Once the cell is
entirely filled with the homogeneous solution (without any vapor phase), it
is isolated from the pumps. Then, the temperature in the cell is increased at
a definite scanning rate (0.2 to 1 K/min) to find the tightest possible temperature interval in which the second phase appears. During this procedure,
the cell is still connected to the buffer volume to avoid pressure increasing due to thermal expansion. The change in turbidity is detected visually.
The uncertainty on the temperature of the cloud point was estimated from
380
380
360
360
340
340
T/K
T/K
134 Carbon Dioxide Capture and Acid Gas Injection
320
300
300
280
0.0
(a)
320
0.2
0.4
xa
0.6
280
0.0
0.8
(b)
0.2
0.4
xa
0.6
0.8
Figure 7.3 Phase diagram, temperature versus mole fraction, for ternary mixtures of (a),
{CO2 –NMPD – H2O} and (b), {CO2 –2MPD – H2O}, at constant loading charges: opened
circle, α = 0 [10, 11] and filled circle, α = 0.2. Solid lines are smooth fitting lines.
reproducibility tests and is less than u(T) = 2K, while uncertainty on such
temperature determination for one experiment is u(T) = 0.5 K.
The phase diagrams of the binary systems {NMPD – H2O} and {2MPD
– H2O} were previously determined [10, 11] and the lower critical solution
temperatures were found to be 318 K for xNMPD = 0.07 and 339 K for x2MPD
= 0.05, respectively.
Concerning the liquid-liquid phase diagrams of the binary systems
illustrated in Figure 7.3, the behavior of the two methylpiperidines with
water is very different. For example at 353 K, without CO2, the water rich
phase of the {NMPD – H2O} system is poor in amine (xNMPD = 0.005) and
the water content of the amine rich phase is rather small (xw = 0.2). While
at the same temperature, the water rich phase of the {2MPD – H2O} system
is rather poor in amine (x2MPD = 0.017) but the amine rich phase is highly
rich in water (xw = 0.82). Without CO2, the phase diagrams of the binary
systems show that using NMPD instead of 2MPD is more favorable for the
demixing process.
At a constant gas loading charge of 0.2, the temperatures of phase
separation decrease significantly with the addition of NMPD and reach
280 K for a composition of amine solution xNMPD = 0.11. Measurements
were not feasible for more concentrated solutions, solutions due to the
limits of temperatures of our techniques (270 K–393 K). With 2MPD, the
phase diagram with dissolved CO2 is similar to the one without CO2 up to
x2MPD = 0.046. We do not observe any significant change of the lower critical end point. Moreover, a significant shrinkage of the immiscibility gap
is observed. Finally, we can also notice that the “amine phase” is very rich
in water. Those differences are mainly due to different chemical reactions
occurring in the solution in the presence of CO2 [12].
New Amine Based Solvents for Acid Gas Removal 135
Considering those phase diagrams, none of these amines can reach
the requirements of the proposed process with CO2. The ideal system
considering these methylpiperidines would be a compromise between the
large phase diagram of NMPD and the temperature of phase separation
obtained with 2MPD.
7.3.2 Liquid-Liquid Equilibria of Ternary Systems
{Amine – H2O – Glycol}
The addition of a physical solvent, triethylene glycol (TEG) was considered
to increase the temperatures of phase separation of the mixtures, without
changing the shape of the curve. A test study was then realized in the ternary liquid system {(N- or 2-)MPD – H2O – TEG} to verify the influence
of the TEG on the LLE.
The visual technique previously described was used to evaluate the influence of the glycol on the LLE at atmospheric pressure. For that purpose,
increasing amounts of TEG were added to aqueous solutions of NMPD
and 2MPD with a starting amine composition wa = 0.2. The temperatures
of phase separation for both systems are presented in Figure 7.4.
In an aqueous solution of 2MPD (w2MPD = 0.2), the addition of small
amounts of TEG leads to a sharp increase of the temperatures of phase splitting, limiting the amount of TEG to wTEG = 0.075 due to the temperature
370
360
T/K
350
340
330
320
310
0.00
0.05
0.10
0.15
0.20
0.25
0.30
wTEG
Figure 7.4 LLE of the ternary systems:
– TEG}.
, {NMPD – H2O – TEG} and
, {2MPD – H2O
136 Carbon Dioxide Capture and Acid Gas Injection
range of the technique. For the ternary system {NMPD – H2O – TEG}, phase
separation temperatures are also rising while adding TEG. Nevertheless,
these temperatures stay low enough with a reasonable amount of physical
solvent to be undertaken in the demixing process.
7.3.3 Liquid-Liquid Equilibria of the Quaternary Systems
{CO2 – NMPD – TEG – H2O}
The influence of CO2 on the phase diagram was then evaluated in mixtures containing NMPD and TEG. The liquid-liquid equilibrium data were
determined at 0.5 MPa for two mixtures, {NMPD (20) – TEG (20) –H2O
(60)} and {NMPD (20) – TEG (30) –H2O (50)}. Numbers in brackets
denote the weight percent of each mixture component. Figure 7.5 compares the phase diagrams of these two systems as a function of CO2 loading
charge with the one without TEG determined by Coulier et al. [12].
As shown in section 7.3.2, adding TEG to an aqueous solution of NMPD
yields to an increase of the temperatures of phase separation. The shape of the
LLE curves investigated with TEG is similar to the one obtained by Coulier
et al. [12] without TEG. The main difference concerns the temperature of the
lower critical end point which increases while adding TEG. However, it is a
very valuable benefit for the process with demixing solvent since temperatures of phase separation can be controlled by the quantity of physical solvent.
373
353
T/K
333
313
293
273
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
/(mol CO2/mol amine)
Figure 7.5 Phase diagram, temperature versus loading charge, for quaternary mixtures of
{CO2 – NMPD – TEG – H2O}: , {CO2 – NMPD (20) – TEG (0) – H2O (80)}; : {CO2 –
NMPD (20) – TEG (20) – H2O (60)} and , {CO2 – NMPD (20) – TEG (30) – H2O (50)}.
Numbers in brackets denote the weight percent of each mixture component. Dash lines
are smooth fitting lines.
New Amine Based Solvents for Acid Gas Removal 137
As the LLE regions are well controlled by adding TEG, measurements
of the other thermodynamic properties of the mixtures were carried out,
to provide additional information in case of a process development and to
compare the capacity of such solvents with ones without TEG. From the
previous results, NMPD appears to be the most promising amine for this
application. Thermodynamic properties such as densities and heat capacities of mixtures containing this amine were determined.
7.4 Densities and Heat Capacities of Ternary Systems
{NMPD – H2O – Glycol}
Densities and heat capacities are two essential thermodynamic properties that need to be measured to optimize separation processes. Indeed
the densities drive part of the transport properties of the solutions, and
heat capacities control the energetic cost resulting from heating during the
process. Densities are also needed for any calculations of molar properties
from volumetric measurements (solution composition in LLE or enthalpies of solutions containing CO2, heat capacities…).
7.4.1
Densities
The densities of the ternary solutions {NMPD – H2O – TEG} were measured at atmospheric pressure using an Anton Paar DMA 5000 density
meter and the experimental procedure is given by Coquelet et al. [15].
Compositions in amine, water and TEG and the ranges of investigated
temperatures are resumed in Table 7.3. The range of temperature studied is
limited by the LLE as measurements have to be realized for homogeneous
one phase solutions. The range of studied temperature is then larger when
adding TEG, as explained in section 7.3.2.
The influence of glycol on the densities of aqueous solution of NMPD is
shown in Figure 7.6. An increase of the densities is observed with the addition of TEG at all studied temperatures. The density of solutions decreases
also when the temperature is increased. The curves are mostly shifted to
the highest values of densities when TEG is added to the solution.
7.4.2
Specific Heat Capacities
The specific heat capacities of aqueous amine solutions were determined by
using a differential scanning microcalorimeter (microSC) from SETARAM,
France, equipped with liquid Cp cells of 1 mL inner volumes. The detection
138 Carbon Dioxide Capture and Acid Gas Injection
Table 7.3 Composition of the ternary systems {NMPD – H2O – TEG} and
­temperature range investigated in the density study.
NMP
TEG
H2O
T
wt %
K
20
20
60
283–333
20
30
50
283–343
30
20
50
283–338
30
30
40
283–343
20
80
0
283–343
30
70
0
283–343
0
100
0
283–343
100
0
0
283–343
20
0
80
283–313
30
0
70
283–313
1.08
1.06
/g.cm–3
1.04
1.02
1.00
0.98
0.96
280
290
300
310
320
330
340
350
T/K
Figure 7.6 Densities of the ternary systems {NMPD (w %) – TEG (w %) – H2O (w%)}.
, {NMPD (20) – TEG (0) – H2O (80)}; , {NMPD (20) – TEG (20) – H2O (60)};
, {NMPD (20) – TEG (30) – H2O (50)}; , {NMPD (20) – TEG (80) – H2O (0)}.
Numbers in brackets denote the weight percent of each mixture component. Dash
lines are smooth fitting.
New Amine Based Solvents for Acid Gas Removal 139
5.0
4.5
Cp/J.g–1.K–1
4.0
3.5
3.0
2.5
2.0
283
293
303
313
323
333
T/K
Figure 7.7 Specific Heat Capacities as a function of temperature for the ternary systems
{NMPD (w %) – TEG (w %) – H2O (w %)}. , {NMPD (20) – TEG (0) – H2O (80)};
, {NMPD (20) – TEG (20) – H2O (60)}; , {NMPD (20) – TEG (30) – H2O (50)};
, {NMPD (20) – TEG (80) – H2O (0)}. Numbers in brackets denote the weight percent
of each mixture component. Dash lines are smooth fitting lines.
is based on the Calvet principle. The experimental procedure is given by
Coulier et al. [16]. First, a blank experiment is performed by filling both
the sample and reference cells with nitrogen (N2). Then, the sample cell is
filled with the studied mixture while the reference cell is filled with N2. An
experimental run is made of a 20 min isothermal step at 278.15 K followed
by temperature scanning (0.5 K·min–1) up to 333.15 K. Experiments were
carried out at constant pressure (0.1 MPa) in both the sample and reference
cells. The influence of the physical solvent on the specific heat capacities is
shown in Figure 7.7.
As expected, TEG reduces the heat capacities of the absorbent solution.
The heat capacity is close to 2 J.g–1.K–1 when water is replaced by glycol as a
solvent for the NMPD. This decrease is highly valuable for process design
as the cost for heating the mixtures is drastically reduced with TEG.
7.5 Vapor-Liquid Equilibria of Ternary Systems
{NMPD – TEG – H2O – CO2}
A specific description of the experimental device used in this work to
measure VLE data has been reported by Zhang et al. [17]. Shortly, the
140 Carbon Dioxide Capture and Acid Gas Injection
technique of measurements is based on the “static-analytic” method
described by Laugier and Richon [18] and experimental procedure
is fully described in Coquelet and Richon [19]. With this apparatus
both the liquid and vapor phases can be sampled under pressure using
ROLSI capillary samplers [20, 21]. The equilibrium cell is immersed in
a thermo-regulated liquid bath. In order to ensure accurate temperature
measurements in the equilibrium cell and to check for thermal gradients,
the temperature is measured at the top and bottom flanges through two
100 Ω platinum resistance thermometer probes. A variable-speed stirrer
inside the cell accelerates the mass transfer between phases and reduces
the time needed to achieve equilibrium. Pressures are measured by three
pressure transducers of which the maximum absolute pressures are
0.35 bar, 1 bar and 10 bar, respectively. Sample analysis is carried out
by a gas chromatograph equipped with a thermal conductivity detector
(TCD). After calibration the uncertainty on CO2 composition in liquid
phase is lower than 0.04.
Before measuring VLE, the equilibrium cell and its loading lines were
first evacuated. About 30 mL of the mixture {NMPD (14) – TEG (17) –
H2O (69)} was introduced via a press at room temperature. The solution
was then heated to 313 K. Meanwhile, an adequate stirring was maintained
inside the cell. Phase equilibrium was assumed to be achieved while temperature and pressure readings stabilized for at least 30 min. The first pressure measurement gave the vapor pressure of the mixture investigated.
Carbon dioxide was then loaded from a gas tank with controlled temperature and pressure. For each equilibrium condition, at least six samples
of the liquid phase were withdrawn and analyzed to ensure composition
repeatability within ±1%. CO2 was then further introduced to measure the
next equilibrium condition.
The solubility of CO2 in a solution of {NMPD (14) – TEG (17) – H2O
(69)} was determined at 313 K. Experiments were conducted for different
CO2 loading charges (α), up to the saturation of the absorbent solution and
are illustrated in Figure 7.8.
7.6
Enthalpies of Solution
The experimental setup used in this study has been carefully reported elsewhere [13]. Briefly, the enthalpy of solution of CO2 in the ternary system
{NMPD – H2O – TEG} was measured by using a custom-made flow-­mixing
cell adapted to a Setaram BT2.15 heat conduction differential calorimeter.
Experiments were carried out at constant temperature and pressure. The
New Amine Based Solvents for Acid Gas Removal 141
7
6
p/bar
5
4
3
2
1
0
0.0
0.2
0.4
0.6
0.8
1.0
/(mol CO2/mol amine)
Figure 7.8 Equilibrium pressure as a function of CO2 loading charge for the system
{NMPD (14) – TEG (17) – H2O (69)} at 313 K. Numbers in brackets denote the weight
percent of each mixture component.
two fluids to be mixed (CO2 and ternary solution) were injected into the
flow lines by two high-pressure syringe pumps, thermo-regulated at near
ambient temperature. Experiments were carried out at different loadings α
(moles CO2/mol amine). The gas loading charge is determined as described
in the previous section.
Enthalpies of solution of CO2 in solutions of {NMPD – H2O – TEG}
were measured at 313 K at pressure of 1 MPa, for two absorbent mixtures
(wNMPD = 0.20, wTEG = 0.20). Experiments were conducted for different
loading charges (α), up to the saturation of the absorbent solution. As an
example, experimental enthalpies measured for {NMPD – H2O – TEG}
and expressed in kJ.mol−1 of CO2 (Figure 7.9a) and of NMPD (Figure 7.9b)
have been plotted versus loading charge α (mol of CO2/mol of amine). In
Figure 7.9a, the enthalpies of solution for CO2 are exothermic and equivalent, up to a loading charge of 0.5. The average enthalpy values ΔsolHav, for
α < 0.5 is found to be –71.5 kJ.mol−1. These values were not determined
for the binary system {NMPD – H2O} with wa = 0.2 because phase separations would occur while adding CO2 in this experimental condition of
temperature (Figure 7.5). In Figure 7.9b, experimental enthalpies of solution expressed in kJ.mol−1 of NMPD show two different domains. In the
first domain (0 < α < 1), ΔsolH increases linearly with the loading charge.
The value of the slope in this domain is equal to ΔsolHav obtained previously
142 Carbon Dioxide Capture and Acid Gas Injection
80
70
50
40
30
–
–1
solH/K.J.mol
60
20
10
0
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
1.4
1.6
1.8
/(mol CO2/mol amine)
(a)
80
70
60
–
–1
solH/K.J.mol
50
40
30
20
10
0
0.0
(b)
0.2
0.4
0.6
0.8
1.0
1.2
/(mol CO2/mol amine)
Figure 7.9 Enthalpy of solution (−ΔsolH) versus CO2 loading charge for an aqueous
solution {NMPD (20) – TEG (20) – H2O (60)}at T = 313 K and p= 1.0 MPa. (a) ΔsolH/ (kJ.
mol−1 of CO2), straight lines show the average values for the enthalpies of solution at low
loadings (α < 0.5); (b) ΔsolnH/(kJ.mol−1 of NMPD).
(Figure 7.9a). The second domain where the enthalpy of solution stays
constant is characteristic of a saturated solution. The intersection between
unsaturated (enthalpy increase) and saturated (plateau) domains yields the
solubility limit (s). The experimental solubility limit of CO2 in the ternary
system {NMPD – H2O – TEG} was graphically determined at 313 K and
1.0 MPa
New Amine Based Solvents for Acid Gas Removal 143
7.7
Discussion and Conclusion
A part of the thermodynamic properties required (phase diagram, heat
capacity, heat of absorption, density) for the design of the new operation
units for CO2 removal have been determined in this study. These preliminary results show the substantial benefits of the addition of a physical solvent to an aqueous solution of amine. Among these advantages, the specific
heat capacities of liquid phases can be lowered allowing energy savings, and
the temperatures of the phase separation in the presence of CO2 can be controlled. Nevertheless, additional thermodynamic data (such as Henry’s law
constant) are needed to complete this study, in order to develop a reliable
thermodynamic model that takes into account the mechanism of ­reaction
of CO2 with the amine and the formation of electrolytes species. For that
purpose, an original device developed by Provost et al. [22] allowing the
simultaneous measurements of the pressure and liquid phase composition,
as a function of time will be used. The liquid phase composition is evaluated through the analysis Fourier Transform InfraRed (FTIR) spectrum,
recorded in situ with an Attenuated Total Reflection (ATR) accessory.
In addition to this, thermodynamics models developed by Paricaud
et al. [23] will be applied to these systems and could run within process
simulation software. The first step will be the development of a thermodynamic model for electrolyte solutions, which considers the most important
chemical species in the aqueous and amine solutions, and is able to predict
the liquid-liquid immiscibility as well as the chemical and phase equilibria.
This model will be used to describe both the phase equilibria and energetic
properties such as heat capacities. It will be implemented into a code that is
compatible with the CAPE-OPEN interface of Simulis Thermodynamics
and prosim plus software developed by the Prosim company. Through
this interface, we will be able to simulate the main three elements of the
separation process (decanter and absorption and desorption columns) in
either the Prosim plus or Aspen one environment, and estimate the cost
and energy requirement for the CO2 capture.
Acknowledgments
This research work was part of collaboration between the Centre
Thermodynamics of Processes (CTP) of MINES ParisTech and the
Institute of Chemistry of Clermont-Ferrand (ICCF). The recommendations and concerns from Dr. John Carroll (Gas Liquid Engineering) concerning industrial applications of demixing amines are an inspiration of
144 Carbon Dioxide Capture and Acid Gas Injection
this work. Constructive advice and considerable expertise given by Pascal
Théveneau and Alain Valtz (CTP) are gratefully acknowledged. Financial
support allowing the acquisition of the microcalorimeter for heat capacities measurements from Contrat d’Objectif Partagé CNRS-UBP-FederRégion Auvergne is also acknowledged.
References
1. Z
hao, Q., Leonhardt, E., MacConnell, C., Frear, C., and Chen, S., Purification
technologies for biogas generated by anaerobic digestion. p. 1–24, 2010.
2. Rochelle, G.T., Amine Scrubbing for CO2 Capture. Science, 325(5948),
p. 1652–1654, 2009.
3. Raynal, L., Bouillon, P.-A., Gomez, A., and Broutin, P., From MEA to demixing
solvents and future steps, a roadmap for lowering the cost of post-combustion
carbon capture. Chemical Engineering Journal, 171(3), p. 742–752, 2011.
4. Raynal, L., Briot, P., Dreillard, M., Broutin, P., Mangiaracina, A., Drioli,
B.S., Politi, M., La Marca, C., Mertens, J., Thielens, M.-L., Laborie, G., and
Normand, L., Evaluation of the DMX Process for Industrial Pilot Demonstration –
Methodology and Results. Energy Procedia, 63, p. 6298–6309, 2014.
5. Raynal, L., Alix, P., Bouillon, P.-A., Gomez, A., le Febvre de Nailly, M., Jacquin,
M., Kittel, J., di Lella, A., Mougin, P., and Trapy, J., The DMX process: An
original solution for lowering the cost of post-combustion carbon capture.
Energy Procedia, 4(0), p. 779–786, 2011.
6. Ballerat-Busserolles, K., Lowe, A.R., Coulier, Y., and Coxam, J.Y., Calorimetry
in Aqueous Solutions of Demixing Amines for Processes in CO2 Capture, in
Acid Gas Extraction for Disposal and Related Topics. John Wiley & Sons, Inc.
p. 69–80, 2016.
7. Fandiño, O., Yacyshyn, M., Cox, J.S., and Tremaine, P.R., Speciation in LiquidLiquid Phase-Separating Solutions of Aqueous Amines for Carbon Capture
Applications by Raman Spectroscopy, in Acid Gas Extraction for Disposal and
Related Topics. John Wiley & Sons, Inc. p. 81–94, 2016.
8. Dash, S.K. and Bandyopadhyay, S.S., Studies on the effect of addition of piperazine and sulfolane into aqueous solution of N-methyldiethanolamine for CO2
capture and VLE modelling using eNRTL equation. International Journal of
Greenhouse Gas Control, 44, p. 227–237, 2016.
9. Zheng, C., Tan, J., Wang, Y.J., and Luo, G.S., CO2 Solubility in a Mixture
Absorption System of 2-Amino-2-methyl-1-propanol with Ethylene Glycol.
Industrial & Engineering Chemistry Research, 52(34), p. 12247–12252,
2013.
10. Coulier, Y., Ballerat-Busserolles, K., Rodier, L., and Coxam, J.Y., Temperatures
of liquid–liquid separation and excess molar volumes of {N-methylpiperidine–
water} and {2-methylpiperidine–water} systems. Fluid Phase Equilibria, 296(2),
p. 206–212, 2010.
New Amine Based Solvents for Acid Gas Removal 145
11. Stephenson, R.M., Mutual solubility of water and pyridine derivatives. Journal
of Chemical & Engineering Data, 38(3), p. 428–431, 1993.
12. Coulier, Y., Lowe, A.R., Moreau, A., Ballerat-Busserolles, K., and Coxam, J.-Y.,
Liquid-Liquid Equilibria in {Amine - H2O - CO2} Mixtures: New Methods for
Key Data. Fluid Phase Equilibria, 2016. accepted.
13. Arcis, H., Ballerat-Busserolles, K., Rodier, L., and Coxam, J.-Y., Enthalpy of
Solution of Carbon Dioxide in Aqueous Solutions of Monoethanolamine at
Temperatures of 322.5 K and 372.9 K and Pressures up to 5 MPa. Journal of
Chemical & Engineering Data, 56(8), p. 3351–3362, 2011.
14. Span, R. and Wagner, W., A New Equation of State for Carbon Dioxide
Covering the Fluid Region from the Triple‐Point Temperature to 1100 K at
Pressures up to 800 MPa. Journal of Physical and Chemical Reference Data,
25(6), p. 1509–1596, 1996.
15. Coquelet, C., Valtz, A., Courtial, 7., and Richon, D., Volumetric properties of
the water + 3-(dimethylamino) propylamine (DMAPA) mixture at atmospheric
pressure from 283.15 to 353.15 K. Thermochimica Acta, 448(1), p. 47–51, 2006.
16. Coulier, Y., Ballerat-Busserolles, K., Mesones, J., Lowe, A., and Coxam,
J.-Y., Excess Molar Enthalpies and Heat Capacities of {2-Methylpiperidine–
Water} and {N-Methylpiperidine–Water} Systems of Low to Moderate Amine
Compositions. Journal of Chemical & Engineering Data, 60(6), p. 1563–1571,
2015.
17. Zhang, F., Théveneau, P., El Ahmar, E., Canet, 7., Soo, C.-B., and Coquelet,
C., An improved static–analytic apparatus for vapor–liquid equilibrium (PTxy)
measurement using modified in-situ samplers. Fluid Phase Equilibria, 409,
p. 425–433, 2016.
18. Laugier, S. and Richon, D., New apparatus to perform fast determinations of
mixture vapor–liquid equilibria up to 10 MPa and 423 K. Review of Scientific
Instruments, 57(3), p. 469–472, 1986.
19. Coquelet, C. and Richon, D., Chapter 14 Solubility of BTEX and Acid Gases in
Alkanolamine Solutions in Relation to the Environment, in Developments and
Applications in Solubility. The Royal Society of Chemistry. p. 219–235, 2007.
20. Guilbot, P., Valtz, A., Legendre, H., and Richon, D., Rapid on-line samplerinjector: a reliable tool for HT-HP sampling and on-line GC analysis. Analusis,
28(5), p. 426–431, 2000.
21. Théveneau, P. and Legendre, H., Dispositif pour prélever des micro-échantillons
d’un fluide à l’état liquide contenu dans un containeur. 2014, French Patent
Application No. 1460309.
22. Diab, F., Provost, E., Laloué, N., Alix, P., Souchon, V., Delpoux, O., and Fürst,
W., Quantitative analysis of the liquid phase by FT-IR spectroscopy in the system
CO2/diethanolamine (DEA)/H2O. Fluid Phase Equilibria, 325, p. 90–99, 2012.
23. Fukumoto, A., Sales Silva, L.P., Paricaud, P., Dalmazzone, D., and Fürst, W.,
Modeling of the dissociation conditions of H2 + CO2 semiclathrate hydrate
formed with TBAB, TBAC, TBAF, TBPB, and TBNO3 salts. Application to
CO2 capture from syngas. International Journal of Hydrogen Energy, 40(30),
p. 9254–9266, 2015.
8
Improved Solvents for CO Capture by
Molecular Simulation Methodology
2
William R. Smith
Department of Mathematics and Statistics and Dept. of Chemistry, University of
Guelph, Guelph, Canada, and Faculty of Science, University of Ontario Institute of
Technology, Oshawa, Canada
Abstract
The goal of this paper is to describe a strategy for implementing molecular simulation methodology to model CO2 capture systems by combining advanced
molecular-level modeling and experimental measurement methodologies, with
the goal of discovering new alkanolamine solvents that yield improved CO2 solubility. The project has been funded for a 3-year period ending December 2018
by the Natural Sciences and Engineering Research Council of Canada under its
International Strategic Partnership Program. A collaborating partner, funded by
l’Agence Nationale de la Recherche of France, is a team at Blaise Pascal University,
Clermont-Ferrand, France. The University of Guelph is primarily responsible for
the theoretically based work described in this paper, and the French team is primarily responsible for related experimental research. The research of both groups
will be tightly integrated to achieve the project goals.
8.1
Introduction
Although alternatives exist for carbon capture in other contexts, we focus
here on the relatively mature industrial methodology of CO2 capture
from a post-combustion gas stream that can be effected by its preferential
absorption in a liquid solvent. The use of aqueous alkanolamine as a solvent to absorb CO2 was patented in 1930 [17], and is expected to be the
dominant technology for US coal-fired plants by 2030 [4]. The amine is
subsequently regenerated by stripping with steam and condensation of the
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (147–160)
2017 © Scrivener Publishing LLC
147
148 Carbon Dioxide Capture and Acid Gas Injection
T
T
A
b
s
o
r
b
e
r
P
P
S
t
r
i
p
p
e
r
C
o
m
p
r
e
s
s
o
r
Reboiler
T
Figure 8.1 CO2 capture by amine scrubbing. (Reproduced with permission from the PhD
thesis of Micka¨el Simond [19].)
H2O from the stripper vapor, leaving pure CO2 that can be compressed for
purposes such as enhanced oil recovery or for geological sequestration.
The process is represented in Figure 8.1. It was evaluated in 1991 [18] for
commercial use in coal-fired power plants and found to have unacceptably high energy requirements and costs, motivating the need for improved
methodologies.
Developing solvents with improved solubility and energy properties
would decrease the cost, but the number of potential solvent alternatives
is vast. Some have been developed commercially and demonstration facilities have been built [7] (including the Shell–CanSolv facility at Estevan
Saskatchewan, opened in October 2014), but one of the impediments in
the quest to discover improved solvents is the considerable time and cost
entailed in the design and construction of experiments to test alternative
solvents.
Although macroscopic thermodynamic models for predicting CO2 solubility in candidate solvents have been and continue to be developed (e.g.
[20, 21]), this is a challenging task, since the solvents are very nonideal, and
the process complexity is well beyond that of normal VLE modeling due
to the solvent’s chemical reactions with CO2. The predominantly empirical
basis of these models precludes confidence in the accuracy of their predictions beyond the ranges of experimental conditions from which their
parameters have been determined. In addition, the typically large number
of parameters in such models requires large (and costly to obtain) sets of
experimental data for their determination for each proposed solvent.
Improved Solvents for CO2 Capture 149
A potentially more cost-effective way to screen potential solvent candidates and to guide experiments is to develop more fundamentally based
and highly accurate mathematical models containing smaller parameter
sets to predict the properties of the systems involved, and this strategy is
at the core of the project. The models and algorithms based on molecularlevel simulation technology proposed herein can also supplement the macroscopic models and provide a means to extend their predictions to a much
wider range of operating conditions. Such models also allow the prediction of multiple thermodynamic and transport properties using the same
parameter set, furthering their advantage over the macroscopic approach,
which requires different models for each property and corresponding
additional parameter sets that must be determined from experimental
data. Significant progress in computational technology has now made possible the relatively low-cost implementation of increasingly sophisticated
molecular-based models and algorithms like those described in this paper.
8.2
Physical and Chemical Models
In general, one may describe the thermodynamic properties of solutions
like those considered here that are characterized by strong intermolecular interactions, by means of a physical model [22, 23], a chemical model
[24–26], or a combination of the two approaches for different system components. A physical model assumes the existence of only the original system species, whereas a chemical model assumes the existence of additional
species accompanied by chemical reactions among them. Although the
additional species are sometimes justified by spectroscopic evidence [27],
their introduction often has an empirical basis. Another distinguishing
model characteristic concerns whether it employs a macroscopic thermodynamic description via either an equation of state (EOS) and/or a chemical potential model, or a molecular-level description, by which take here
to mean a mathematical model of the molecular-level interactions in terms
of a force field (FF).
A simple example of a macroscopic chemical model is one for H2O that
assumes the existence of the distinct species {H3O+, OH–, H2O} and the
accompanying chemical reaction
2H2O = H3O+ + OH– (8.1)
On the other hand, a macroscopic physical model for the system such
as SAFT (for a recent review, see [20]) employs an EOS for H2O that mimics the reaction’s effects by means of a relatively simple molecular picture
150 Carbon Dioxide Capture and Acid Gas Injection
that incorporates orientationally dependent association sites on the water
molecules that result in the formation of hydrogen bonds. Similarly,
­
a molecular-level physical model for H2O employs a detailed FF for the H2O
molecule with strong orientationally dependent attractive forces; TIP4P
[28] and SPC/E FF [29] are typical water FFs often used, although there are
many others (see the review of Guillot [30]). However, when compounds
chemically react, rather than form relatively weak association bonds, models of the system must take this into account as a separate feature.
Macroscopic chemical models have considerable flexibility in fitting
experimental data, since the introduction of each new species introduces
corresponding additional parameters that can be adjusted to fit the data;
however, this flexibility comes at the cost of requiring extensive experimental
data to determine the increased number of parameters required to describe
their empirical temperature and pressure dependence. Macroscopic chemical models include NRTL [31] and UNIQUAC [32, 33], augmented by
semi-empirical Pitzer chemical potential expressions [34], the electrolyte
EOS of Fuürst and Renon [35] and the CPA (cubic-plus-association) [36,
37] EOS. Physical models have the advantage of generally requiring fewer
parameters, but unless they adequately reflect the underlying intermolecular interactions, their predictions may not be sufficiently accurate.
The SAFT EOS (see, for example, [20, 38, 39]) is an example of a physical
model that has both both macroscopic and molecular-level characteristics.
It has become routinely used by chemical engineers for various applications and as a component of chemical process simulation software packages
such as AspenPlus [40] and PROSIM [41]. The practical utility of the SAFT
approach lies in the fact that although its underlying FF is a relatively crude
approximation, its thermodynamic consequences can be expressed in analytical form. (Its historical origin is related to the simple Smith-Nezbeda
model [42] of associating hard-sphere fluids [43].) It is important to note
that SAFT parameters depend only on the interactions between molecular
sites and are independent of temperature. Notwithstanding, SAFT retains
a somewhat empirical flavor, in view of its approximate molecular picture.
8.3 Molecular-Level Models and Algorithms for
Thermodynamic Property Predictions
Molecular-level models describe each species molecule by means of a
mathematical model of its structure and its interactions with other system molecules, and in a sense are a molecular-level analogue of a macroscopic physical model. Chemical reactions among the molecules of the
Improved Solvents for CO2 Capture 151
system can be implemented in conjunction with such models by means of
the Reaction Ensemble Monte Carlo algorithm [10, 11], discussed below.
Molecular-level models have the following general advantages over macroscopic models:
a. Their fundamental basis enables them to be used for predictions
beyond the range of thermodynamic conditions in which the
parameters have been obtained by fitting to experimental data;
b. They are capable of systematic improvement, by developing
improved descriptions of the molecular interactions;
c. They have many fewer parameters than is the case for macroscopic models for a given level of accuracy; for example
(apart from the SAFT model), although temperature- and
pressure-dependent parameters are typically required for macroscopic models, this dependence is automatically accounted
for in the mathematical structure of molecular-level models.
This results in a reduced need for experimental data to fit the
parameters;
d. They require only a single (relatively small) parameter set to
predict (given appropriate algorithms) all thermodynamic
and transport properties of a substance; in contrast, different macroscopic models (and corresponding experimental
data to fit their parameters) are often required for different
properties.
The primary disadvantage of molecular-level models is their computational complexity, making it infeasible for them to be embedded in software
such as chemical process simulators. Nonetheless, the time and cost required
for computing data points using molecular-level models is orders of magnitude less than the cost of experiments to obtain the much larger data sets
required to fit the larger parameter sets of macroscopic models. Provided
they are sufficiently accurate, this also makes molecular-based models a
valuable tool for generating pseudo-experimental data, to which empirical
macroscopic models can be fitted for use in such simulation software.
Although molecular-level models promise increased predictive capability, their use has been historically impeded not only by their computational complexity, but also by the lack of suitable algorithms to directly and
efficiently implement certain thermodynamic properties and processes
of industrial interest. The problem of computational complexity has been
greatly alleviated by advances in computer technology, but the latter problem has continued to exist.
152 Carbon Dioxide Capture and Acid Gas Injection
Both Molecular Dynamics and Monte Carlo methods have been used
for over 40 years for molecular-level thermodynamic property predictions
(see, for example, [44, 45]), and each technique has advantages for different types of problems. Monte Carlo methods are generally more suitable
for calculating properties and simulating processes related to chemical
potentials, which are arguably the most important solution properties
underlying reaction and phase equilibria. However, although methods
based on Monte Carlo algorithms have a long history of use for phase
equilibrium calculations since the development in 1987 of the Gibbs
Ensemble Monte Carlo (GEMC) algorithm [46], their use for chemically
reacting systems is more recent. This has been enabled by the development of the Reaction Ensemble Monte Carlo (REMC) algorithm [10, 11]),
which has been applied to a much smaller number of systems. Space does
not permit its detailed description here; see the original papers or the
review of Turner et al. [12]. It can be considered to be the molecular counterpart of macroscopic methods that augment the ideal-gas part of the
species chemical potential value with a nonideal contribution modelled
by an empirical equation or derived from an EOS model. The REMC algorithm treats the total species chemical potentials similarly, by separately
incorporating the ideal-gas part of the partition function and accounting
for the nonideality by simulating the configurational part of the partition
function.
Finally, we mention in passing that other methods for molecular-level
modeling include the DFT-based ab-initio molecular dynamics [47]
and COSMO-RS [48] approaches and the “reactive force field (ReaxFF)
approach [49], all of which we deem to be less suitable for studying the
thermodynamics of the complex systems considered here, and which are
outside the scope of our study.
We remark that the implementation of REMC methodology is not
straightforward, particularly for dense liquid systems containing complex
molecules, and has technical challenges similar to those encountered by
other simulation methodology related to chemical potentials (such as the
GEMC method for phase equilibrium). The challenges include the requirement for the delicate treatment of the transition probabilities for reaction
steps involving flexible molecular FFs [50] and the general need for computational efficiency enhancements for dense complex systems, the latter
implemented by means of various “gradual particle insertion methods”
[51–55]. In addition, special techniques will be necessary to handle species
present in small amounts, in order that system sizes not be too large; we
expect that such species can be treated separately from the simulations by
means of a Henry–Law type of model.
Improved Solvents for CO2 Capture 153
8.4 Molecular-Level Models and Methodology for
MEA–H2O–CO2
We will initially consider the monoethanolamine (MEA) solvent and
MEA-H2O–CO2 solutions. This system has traditionally been described
at the macroscopic level by means of a chemical model, for example in
conjunction with a UNIQUAC model [32], the NRTL EOS [31, 56], or a
cubic-plus-association (CPA) EOS [37]. The SAFT EOS has also recently
been employed to model this system [39].
A first step in modeling the chemical equilibrium composition of a
­system typically entails the construction of a set of chemical reactions
modeling the chemical changes. In this regard, it is important to emphasize
that the equilibrium composition (assuming the absence of any kinetic or
other restrictions (see [9, 57], Sections 2.4 and 9.6) is independent of any
particular kinetic mechanism, and is governed only by thermodynamic
considerations for the set of species present at equilibrium (the “species
list”). The assumption of a particular species list is an important part of the
thermodynamic analysis, and may be motivated by a number of considerations. Spectroscopic analysis of the solution may provide some indication
of the species present, but this may also indicate association between molecules rather than intramolecular chemical bonds resulting from distinct
chemical species. In order to obtain an economical model entailing a small
species list, it is important to treat association by means of a FF whenever
possible.
If one assumes that the 9 species {H2O, H3O+, OH−, CO2, HCO3−,
CO32–, RNHCO2−, RNH2, RNH3−} comprise the system, where R = HO –
CH2 – CH2−, the rank of the system formula matrix is 4, and a “complete”
set of chemical reactions required to describe all possible chemical change
consists of 5 linearly independent reactions (a more appropriate term is
“chemical equations (see [9], Chapter 2). The choice of the particular set
is arbitrary, but the choice affects the computational efficiency of an equilibrium calculation. For example, the following is a possible complete set,
based primarily on mechanistic considerations (see [31, 39, 58] for other
sets)):
2H2O = H3O+ + OH−(8.2)
CO2 + 2H2O = H3O+ + HCO3−(8.3)
H2O + HCO3− = H3O+ + CO32−(8.4)
CO2 + 2RNH2 = RNHCOO– + RNH3+ (8.5)
154 Carbon Dioxide Capture and Acid Gas Injection
H3O+ + RNH2 = H2O + RNH3+ (8.6)
Reaction (8.2) represents H2O dissociation, (8.3) bicarbonate formation, (8.4) carbonate formation, (8.5) carbamate formation and (8.6) MEA
protonation.
The reaction set (8.2)–(8.6) is problematic in a computer simulation
because it contains reactions that have net charges on one or both sides
of the reactions; as a result, special procedures must be added in order
to ensure that the simulation box always contains a zero net charge.
Additionally, experience from macroscopic reaction equilibrium problems
suggests that it is advantageous to have the most abundant species on the
left sides of the reactions. With the additional assumption that the concentrations of OH−and CO32– are small, the following may be a computationally convenient reaction set (the last of which arises from the combination
(8.3) + (8.6):
CO2 + RNH2 + H2O = RNH3+ + HCO3– (8.7)
The system is typically specified by its temperature, T, and pressure, P,
the initial solvent composition weight fraction ω, and a solution CO2 loading parameter L. ω and L are defined by
n0 RNH2 M NH2
n0 RNH2 MRNH2
L
n0 H2O MH2O
n0 CO2
n0 RNH2
(8.8)
(8.9)
where n0 denotes a molar amount and M denotes the molecular weight.
These define the initial amounts of CO2 and RNH2 per mole of H2O solvent. The composition of the vapour phase species H2O, CO2 and RNH2 in
equilibrium with the calculated composition of the solution phase may be
calculated from the equality of their solution phase chemical potentials.
This may be performed using an equation of state, or by means of a Grand
Canonical Ensemble (GCE) simulation in the vapour phase.
To implement the REMC approach for a given system, for each species
we require a FF and an ideal-gas standard chemical potential, µ0(T). The
latter can be obtained from thermochemical tables (e.g. [59, 60]), or calculated by means of quantum mechanical software such as Gaussian09 [61].
At all stages, we will develop force fields that are transferable, in the sense
that the same FF parameters can be used to describe a given interaction
Improved Solvents for CO2 Capture 155
site in different molecules and that the FF is applicable for the prediction of
different properties (e.g., thermodynamic, structural, or transport) across
a wide range of state points (e.g., pressure, temperature, or composition).
This approach is similar in philosophy to the Group Contribution approach
of macroscopic thermodynamic models (e.g., [62]). As necessary, a small
number of pure fluid FF parameters may be fitted to appropriate experimental data. For mixtures, either Lorentz–Berthelot or geometric mean
combining rules will be used for the cross-species FF parameters, or we
will adjust them to experimental data. We will use the TIP4P [28] FF for
H2O, the FF of Simond et al. [63] for the MEA-H2O subsystem, the FF of
Vácha et al. [64] for H3O+ and a TraPPE FF for CO2 [65]. OPLS FFs are
available for many of the remaining species at http://virtualchemistry.org/
moldb.php.
Once all FF parameters have been determined for the constituent
binary systems by comparing against binary thermodynamic data, predictions will be made for a range of thermodynamic properties of the ternary
MEA–H2O–CO2 system without the need for any additional parameters,
and the results will be compared with available experimental data. Chief
among the properties of interest will be enthalpies, densities, vapor-liquid
equilibria, CO2 solubility and the T and P dependence of these properties.
In the course of model development, we will compare our results with
those obtained from available thermodynamic models in the literature and
with calculations using the ASPEN [40] and/or PROSIM [41] process simulation software packages.
8.4.1 Extensions to Other Alkanolamine Solvents and Their
Mixtures
Once the strategy has been finalized for the MEA–H2O–CO2 system, we
will apply the same approach to other alkanolamine–H2O–CO2 systems
of increasing complexity. The first of these involves the family of primary
alkanolamines based on the N–C–C–O backbone considered by Simond et al.
[19,63, 66], illustrated in Figure 8.2. This group has developed FFs for their
binary mixtures with H2O.
Simond first developed FFs for the pure alkanols [63] based on the
OPLS–AA functional form [67] using Gaussian09 [61] quantum mechanical calculations, based on a training set involving the simplest molecules
of the family, MEA and 2A1P, using experimental heat of vaporization and
liquid density data to refine the parameters. They were then extended to
the test set consisting of the remaining members of the family, and the
densities and enthalpy of vaporization predictions were found to be in
156 Carbon Dioxide Capture and Acid Gas Injection
N
O
C
C
MEA
2A1P
ABU
AMP
MIPA
AMP2
1A2B
Figure 8.2 Primary alkanolamines based on the structure N–C–C–O. (Reproduced with
permission from the PhD thesis of Micka¨el Simond [19].
good agreement with the experimental values. They then used these FFs
to study predictions of the excess enthalpy of their binary mixtures with
water. The TIP4P [28] FF for water was used, and geometric combining
rules were used for all the cross Lennard–Jones parameters between nonbonded sites of the components. They found that this did not produce
quantitative agreement with the heats of mixing as a function of concentration. By introducing a Lennard–Jones cross interaction site (involving
2 parameters, E and σ) corresponding to the hydrogen bond between the
OH group of the alkanolamine and the oxygen atom of water, they could
quantitatively reproduce the experimental excess enthalpies of mixing for
the MEA+H2O and ABU+H2O mixtures as a function of concentration.
They then found that incorporating the same site for the other mixtures
produced similarly quantitative agreement with the heat of mixing data.
The next stage of our study will be to extend the MEA system methodology for constructing FFs to the binary mixtures of this family with CO2.
As in the case of MEA, we will perform pure predictions of the ternary
properties using the FFs determined from the underlying binaries. Finally,
as time permits, we will use the same FFs to perform pure predictions for
mixtures of solvents with H2O and CO2. We do not anticipate that major
adjustments may be required in the case of these quaternary mixtures,
Improved Solvents for CO2 Capture 157
although the worst-case scenario is that adjustments may be required for
the binary solvent mixtures.
Also as time permits, we will extend the foregoing methodology to solvents selected from other (secondary and tertiary) alkanolamine families,
and finally to their mixtures. In all cases, we will perform predictions of the
CO2 solubilities and of other thermodynamic properties (CO2 solubility,
densities, enthalpies and vapor-liquid equilibria, and their T and P dependence), and compare these with experiment. In many cases, experimental
results will not be available, and these will be measured by the ANR group.
Experience with the MEA–H2O–CO2 ternary system will guide our strategies to determine FF parameters and conduct experiments to measure
required properties of these systems. Similarly, as for the simpler systems,
we will compare with the predictions of chemical process simulation software such as ASPEN [40] and PROSIM [41] where these are available.
A final result of the project will be a set of force fields for molecular segments by means of which the force field for an arbitrary molecule of interest
in a large family of primary, secondary and tertiary amine solutions with
H2O and CO2 may be constructed, and from which the CO2 solubility and
other thermodynamic properties may be calculated by means of molecular simulation. These results can be fitted to macroscopic thermodynamic
models for use in chemical process simulators involving such solutions.
Acknowledgements
The financial support of the Natural Sciences and Engineering Research
Council of Canada under Grant STPGP 479466, and computational facilities provided by SHARCNET (Shared Hierarchical Academic Computing
Network), a partner organization of Compute Ontario and the Compute/
Calcul Canada national advanced research computing platform, are gratefully acknowledged.
References
1. Prospects for CO2 capture and storage, 2004, IEA-International Energy
Agency. Head of Publication Service, IEA/OECD, Paris, France.
2. H. Yang et al., J. Env. Sci. 20, 14, 2008.
3. J. D. Figueroa, T. Fout, S. Plasynski, H. McIlvrie, and R. D. Srivastava, Int. J.
Greenhouse Gas Control 2, 9, 2008.
4. G. T. Rochelle, Science 325, 1652, 2009.
5. A. Schäffer, K. Brechtel, and G. Scheffknecht, Fuel 101, 148, 2012.
158 Carbon Dioxide Capture and Acid Gas Injection
6. M. Ramdin, T. W. de Loos, and T. J. Vlugt, Ind. Eng. Chem. Res. 51, 8149, 2012.
7. M. R. Abu-Zahra, Z. Abbas, P. Singh, and P. Feron, Carbon dioxide post–
combustion capture: Solvent technologies overview, status and future directions, in Materials and processes for energy: Communicating current research
and technological developments, edited by A. Méndez-Vilas, pages 923–934,
Formatex Research Center, Bardajo Spain, 2013.
8. S. Kumar, J. H. Cho, and I. Moon, Int. J. Greenhouse Gas Control 20, 87, 2014.
9. W. R. Smith and R. W. Missen, Chemical Reaction Equilibrium Analysis:
Theory and Algorithms (ISBN 0-89464-584-6, Reprint with corrections, of
1982 Wiley–Interscience publication of same name, (ISBN 0-471-09347-5)),
Krieger Publishing, Malabar, Florida, 1990.
10. W. R. Smith and B. Tr´ıska, J. Chem. Phys. 100, 3019, 1994.
11. J. K. Johnson, A. Z. Panagiotopoulos, and K. E. Gubbins, Molec. Phys. 81, 717,
1994.
12. C. H. Turner et al., Molec. Simulation 34, 119, 2007.
13. F. Mouˇcka, I. Nezbeda, and W.R.Smith, J. Chem. Theory Comput. In press,
DOI:10.1021/acs.jctc.5b00018, 2015.
14. W. R. Smith, S. Figueroa-Gerstenmaier, and M. Skvorova, J. Chem. Eng. Data
59, 3258, 2014.
15. H. Arcis, K. Ballerat-Busserolles, L. Rodier, and J.-Y. Coxam, J. Chem. Eng.
Data 56, 3351, 2011.
16. H. Arcis, L. Rodier, K. Ballerat-Busserolles, and J.-Y. Coxam, J. Chem.
Thermodynamics 41, 783, 2009.
17. R. R. Bottoms, US Patent 1783901, 1930, Girdler Corp.
18. F. Daniel, Engineering and economic evaluation of CO2 removal from fossil–
fuel–fired power plants (IR–7365), December, 1991, Electric Power Research
Institute, Palo Alto, CA.
19. M. R. Simond, E´tude des interactions mol´eculaires dans les solvants d’int´erˆet
pour le captage des gaz acides, PhD thesis, Institute of Chemistry, Un. Of
Clermont-Ferrand, 2013.
20. N. MacDowell, F. Llovell, C. S. Adjiman, G. Jackson, and A. Galindo, Molec.
Phys. 110, 1325, 2012.
21. A. Chremos et al., Chem. Eng. Transactions 35, 427, 2013.
22. J. J. van Laar, Z. Phys. Chem. 72, 723, 1910.
23. J. J. van Laar, Z. Phys. Chem. 83, 4599, 1913.
24. F. Dolazalek, Z. Phys. Chem. 64, 727, 1908.
25. F. Dolazalek, Z. Phys. Chem. 71, 191, 1910.
26. F. Dolazalek, Z. Phys. Chem. 83, 40, 1913.
27. L. M. S. G. A. Applegarth, C. R. Corbeil, D. J. W. Mercer, C. C. Pye, and
P. R. Tremaine, J. Chem. Thermodynamics 87, 153, 2014.
28. W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey, and M. L. Klein,
J. Chem. Phys. 79, 926, 1983.
29. H. J. C. Berendsen, J. R. Grigera, and T. P. Straatsma, J. Phys. Chem. 91, 6269,
1987.
Improved Solvents for CO2 Capture 159
30. B. Guillot, J. Molec. Liq. 101, 219, 2002.
31. D. M. Austgen, G. T. Rochelle, X. Peng, and C. C. Chen, Ind. Eng. Chem. Res.
28, 1060, 1989.
32. L. Faramarzi, G. M. Kontogeorgis, K. Thomsen, and E. H. Stenby, Fluid Phase
Equilib. 282, 121, 2009.
33. K. Thomsen, P. Rasmussen, and R. Gani, Chem. Eng. Sci. 54, 1787, 1999.
34. K. S. Pitzer, J. Am. Chem. Soc. 102, 2902, 1980.
35. W. Furst and H. Renon, Am. Inst. Chem. Eng. J. 39, 335, 1993.
36. A. S. Avlund, G. M. Kontogeorgis, and M. L. Michelsen, Ind. Eng. Chem. Res.
47, 7441, 2008.
37. A. T. Zoghi and F. Feyzi, J. Chem. Thermodynamics 87, 153, 2013.
38. G. N. I. Clark, A. J. Haslam, A. Galindo, and G. Jackson, Molec. Phys. 104,
3561, 2006.
39. N. MacDowell, F. Llovell, C. S. Adjiman, G. Jackson, and A. Galindo, Ind. Eng.
Chem. Res. 49, 1883, 2010.
40. Aspen Plus software, 2015, Aspen Technology, Inc., 20 Crosby Drive, Bedford,
Massachusetts.
41. ProSim software, http://www.prosim.net/.\
42. W. R. Smith and I. Nezbeda, J. Chem. Phys. 81, 3694, 1984.
43. M. S. Wertheim, J. Chem. Phys. 88, 1145, 1988.
44. M. P. Allen and D. J. Tildesley, Computer Simulation of Liquids, Oxford
University Press, Oxford, 1987.
45. D. Frenkel and B. Smit, Understanding Molecular Simulation: From Algorithms
to Applications (2nd Edition), Elsevier, Amsterdam, 2002.
46. A. Z. Panagiotopoulos, Molec. Phys. 61, 813, 1987.
47. D. Marx and J. Hutter, Ab Initio Molecular Dynamics: Basic Theory and
Advanced Methods, Cambridge U. Press, Cambridge, UK, 2009.
48. A. Klamt, COSMO–RS: From Quantum Chemistry to Fluid Phase
Thermodynamics and Drug Design, Elsevier, Amsterdam, 2005.
49. A. C. T. van Duin, S. Dasgupta, F. Lorant, and W. A. Goddard, J. Phys. Chem.
A 105, 9396, 2001.
50. T. W. Rosch and E. J. Maginn, J. Chem. Theory Comput. 7, 269, 2011.
51. N. Hansen, S. Jakobtorweihen, and F. J. Keil, J. Chem. Phys. 122, 164705,
2005.
52. S. Jakobtorweihen, N. Hansen, and F. J. Keil, J. Chem. Phys. 125, 224709, 2006.
53. A. Torres-Knoop, S. P. Balaji, T. J. H. Vlugt, and D. Dubbeldam, J. Chem.
Theory Comput. 10, 942, 2014.
54. W. Shi and E. J. Maginn, J. Chem. Theory Comput. 3, 1451, 2007.
55. W. Shi and E. J. Maginn, J. Comput. Chem. 29, 2520, 2008.
56. Y. Zhang and C.-C. Chen, Ind. Eng. Chem. Res. 50, 163, 2011.
57. G. W. Norval, M. J. Phillips, R. W. Missen, and W. R. Smith, Can. J. Chem. Eng.
69, 1184, 1991.
58. P. M. M. Blauwhoff, G. F. Versteeg, and W. P. M. van Swaaij, Chem. Eng. Sci.
39, 207, 1984.
160 Carbon Dioxide Capture and Acid Gas Injection
59. M. W. Chase, Jr., NIST–JANAF Thermochemical Tables, J. Phys. Chem. Ref.
Data, Monograph No. 9, Am. Chem. Soc. and Am. Inst. of Physics, Woodbury,
New York, 1998.
60. NIST Chemistry Webbook, http://webbook.nist.gov, 2011.
61. M. J. Frisch et al., Gaussian09 Revision D.01, 2009, Gaussian Inc. Wallingford
CT 2009.
62. J. Gmehling, Fluid Phase Equilib. 144, 37, 1998.
63. M. R. Simond, K. Ballerat-Busserolles, J.-Y. Coxam, and A. A. Pádua, Chem.
Phys. Phys. Chem. 13, 3866, 2012.
64. R. Vácha, V. Buch, A. Milet, J. P. Devlind, and P. Jungwirth, Phys. Chem. Chem.
Phys. 9, 4736, 2007.
65. J. J. Potoff and J. I. Siepmann, Am. Inst. Chem. Eng. J. 47, 1676, 2001.
66. M. R. Simond, K. Ballerat-Busserolles, J.-Y. Coxam, and A. A. H. Pádua,
J. Chem. Theory Comp. 10, 2471, 2014.
67. R. C. Rizzo and W. L. Jorgensen, J. Am. Chem. Soc. 121, 4827, 1999.
9
Strategies for Minimizing Hydrocarbon
Contamination in Amine Acid Gas for
Reinjection
Mike Sheilan, Ben Spooner and David Engel
Amine Experts – division of Sulphur Experts International Inc. Calgary,
Alberta, Canada
Abstract
Amine units are used by gas processing operations around the world to remove
acid gases such as H2S and CO2 from gas and liquid streams. Once regenerated
from the amine, the H2S and CO2 is then sent to reinjection.
There are many problems associated with acid gas compression when hydrocarbons are present in the gas. Excess hydrocarbons will be present in the acid gas
stream from the amine unit if the amine unit is not designed or operated properly.
Undesirable hydrocarbons exist in the inlet gas of most amine absorbers, in
either gas or liquid phase. Either way, the amine solution can and will absorb these
hydrocarbons to a certain degree. Various types of amines are somewhat soluble
in hydrocarbons depending on the conditions, which explains the absorption in
part. Entrainment of free hydrocarbons and emulsification of hydrocarbons in the
amine solution are also possible.
Hydrocarbons can create numerous operational and performance problems
in the amine unit or the reinjection system and must be minimized. This paper
discusses how to minimize the level of hydrocarbons in rich amine streams, and
thereby protect and enhance the efficiency of the amine units and reinjection
equipment.
Several strategies that can be implemented by gas processing operations to mitigate hydrocarbon contamination of the rich amine stream will be outlined. These
strategies range from reducing hydrocarbons in the gas feed, to operation of the
amine absorbers, to using hydrocarbon separation technologies on the rich amine
stream feeding the regenerator.
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (161–184)
2017 © Scrivener Publishing LLC
161
162 Carbon Dioxide Capture and Acid Gas Injection
The problems immediately resulting from hydrocarbon contamination of the amine acid gas being routed to the acid gas compressor
and their effect on the plant operation and efficiency are also identified
and discussed.
9.1
Introduction
There is a global trend of increasingly stringent environmental air quality legislation. For instance, in the United States, the Clean Air Act limits the emissions of Volatile Organic Compounds (VOCs) to 250 tons per
year and the total amount of Benzene, Toluene, Ethylbenzene and Xylene
(BTEX) emissions to 25 tons per year for a facility. Flaring sour gas is also
prohibited by strict SO2 emission guidelines. For these reasons, flaring or
venting acid gas streams is falling out of favor and reinjection of lower
level H2S acid gas streams is growing in popularity. The process of removing H2S and CO2 from natural gas streams using amine systems is well
established; however, on occasion heavier hydrocarbons are also removed.
Hydrocarbons in amine acid gas are known to cause process issues in
downstream Acid Gas Injection Systems (AGIS).
Thus, there is strong operational and legislative motivation to minimize
the amount of hydrocarbons in the acid gas and in the amine system.
Reducing hydrocarbons in amine systems is advantageous to not only
the AGIS, but also within the amine system. Hydrocarbons in amine can
result in foaming, fouling, pipe vibration, and destruction of gaskets in
plate and frame lean/rich exchangers.
Hydrocarbons in amine also represent lost production for the plant.
Amines are not meant or designed to remove hydrocarbons, which are the
components of natural gas (or Liquefied Petroleum Gas (LPG)), which are
used as fuel or sold.
9.2
Amine Sweetening Process
Utilizing alkanolamines for acid gas removal is a process that has been in
use since 1931. Today, amine treating is used for H2S and CO2 removal
worldwide in gas plants, refineries, steel plants and power plants.
Simplistically, an amine system absorbs CO2 and H2S contaminants
out of a gas or LPG stream in a contactor unit (absorber) at high pressure
and low temperature, and binds the contaminants to an amine molecule
Strategies for Minimizing Hydrocarbon Contamination 163
through a set of chemical reactions. The binding reaction is then reversed
in a stripper unit (regenerator) at low pressure and high temperature. The
CO2 and H2S loaded amine (referred to as “rich” amine) is thus regenerated
for reuse, and the CO2 and H2S exit the system via the acid gas stream and
may be further processed (see Figure 9.1).
Other than some specialty applications, the inlet gas to most amine
absorbers will contain a variety of hydrocarbons. When this occurs, most
gas-phase hydrocarbons will remain as gas and flow harmlessly up though
the amine and exit out of the top of the tower. In the event of liquid-­
liquid treating, liquid hydrocarbons in treater will mix with the amine as
intended, but then separate if a proper amount of residence time is given
to the amine in the bottom of the treater. In a gas-liquid contactor, very
little separation time is included in most designs and therefore, all liquid
hydrocarbons will travel with the rich amine. If the rich amine entering
the regenerator contains hydrocarbons, they will then vaporize and likely
travel with the acid gas (or be condensed and recirculated to the regenerator, increasing the foaming risk). If the acid gas is routed to an acid
gas injection system, the hydrocarbons can create problems, both operationally and mechanically. In the experience of Sulphur Experts, Amine
Acid Gas (AAG) containing more than 2% hydrocarbon is an indication of
problems upstream of the regenerator.
This paper discusses how to mitigate hydrocarbons in the rich amine
feeding the regenerator, thus reducing fouling, foaming, corrosion and
many associated operating problems both in the Amine Regeneration Unit
(ARU) and the AGIS.
Amine gas-treating plant
Sweet gas
Acid gas
Amine storage
tank
Regenerator
Absorber
Feed gas
Rich
solution
Figure 9.1 Generic Amine Plant.
Lean
solution
164 Carbon Dioxide Capture and Acid Gas Injection
9.3
Hydrocarbons in Amine
Hydrocarbons first enter the amine system in the contactor. Any hydrocarbons that end up in the rich amine will arrive there either by absorption (solubility), entrainment, condensation or emulsification. Amines are
organized into three categories: primary, secondary and tertiary. A paper
published by Critchfield et al. [4] explained how hydrocarbon solubility in
amines relates to the molecular weight of the amine. The order of hydrocarbon solubility, in order of lowest to highest is as follows:
Monoethanolamine
Diglycolamine
Diethanolamine
Methyldiethanoamine
Diisopropanolamine
There are two main types of hydrocarbons that can enter an amine
plant: polar and non-polar. Non-polar hydrocarbons are generally free
hydrocarbons, which do not blend well with amines and can be relatively
easily separated and removed. Polar hydrocarbons have a unique chemical characteristic, whereby part of the molecule is hydrophobic and the
other part is hydrophilic. This basically results in an emulsion of amine and
hydrocarbon. If the level of contamination is high enough in the amine,
significant amounts of hydrocarbons (being pulled along by the surfactant)
are carried with the solution to the flash tank.
Although slightly different in chemistry, aromatics (i.e., BTEX) exhibit
characteristics similar to polar hydrocarbons, and are also harmful to the
AGIS. The thesis of Borda [1] provides a good review of the available data
on BTEX and VOC solubility in amine solutions. Bullin & Brown [2] proposed theoretical conditions for minimizing the pick-up of hydrocarbons
and BTEX. This was based on simulation. The accuracy of these simulations
has been called into question by Borda [1]. Borda [1] presents new BTEX
data fitted to a different correlation and compares the results from this, and
results from the same simulator used by Bullins and Brown [2], to plant data.
The simulator significantly overpredicted the actual hydrocarbon solubility.
Overprediction of hydrocarbons in rich amine by a simulator is hard to
rationalize. It is expected that real plant results would always have more
hydrocarbons than predicted by a simulator. It is important to note that
most data for hydrocarbon solubility in amine solutions is based on lean
amine solutions that are not loaded with aqueous CO2 and H2S. Hatcher
et al. [5] explain the overprediction of hydrocarbon solubility in rich
Strategies for Minimizing Hydrocarbon Contamination 165
Figure 9.2 Amines Contaminated with Varying Degrees of Hydrocarbon.
amine through a “salting-out” phenomenon, as described further in the
“operational parameters” section of this report. Hatcher et al. [5] use an
­alternative simulator to Borda [1] and it is based on mass transfer rather
than equilibrium models.
Figure 9.2 is an example of different levels of hydrocarbon contamination that can occur within various amine systems.
Hydrocarbons in rich amine streams can be present essentially in the
following three forms, as previously detailed by Spooner and Engel [9]:
Free Hydrocarbons. These non-polar hydrocarbons will float
on top of the amine solution within a few minutes if given the
chance (see Figure 9.3). This typically occurs in the flash tank.
Soluble Hydrocarbons. All hydrocarbons will have certain
solubility in amine solutions (see Figure 9.4). The extent of
the solubility will depend on the following:
Type and concentration of amine (common hydrocarbons such as C1/C2/C3 are two to three times more soluble in amine compared to pure water [3]),
pH of the amine,
Amine contactor pressure and temperature, and
Type of hydrocarbon and polar functional groups such as
carboxylic acids and alcohols. Aromatics are included in
this group, as is explained further in this paper.
Emulsified Hydrocarbons. When surfactants are present, hydrocarbon contaminants can form very small droplets in the amine solutions (see Figure 9.5). These droplets
are stabilized by molecular surfactants (similar to soaps
or detergents) and also by small-size suspended solids.
Emulsion droplet sizes can range from a few microns to
about 500 microns. Micro-emulsions, which are the most
stable emulsions available (and can take weeks to separate)
are typically found when droplet sizes are <10 microns.
166 Carbon Dioxide Capture and Acid Gas Injection
Figure 9.3 Free Hydrocarbons in Amine.
9.4 Effect of Hydrocarbons on the Acid Gas
Reinjection System
Traditionally, acid gas streams were vented, flared, or sent to a sulphur plant.
The modern approach to low-level H2S acid gas management is to reinject the gas stream leaving the amine system back downhole. It is advised
to minimize hydrocarbons in reinjection systems for several reasons:
Light hydrocarbon reduce the density of the acid gas which
increases the required injection pressure. This in turn means
more compression power.
Non-acid gas components in the mixture means more volume to compress. The power needed by the compressor is
almost directly proportional to the volume.
Strategies for Minimizing Hydrocarbon Contamination 167
Figure 9.4 Soluble Hydrocarbons Concentrated in Reflux Water.
Heavier hydrocarbons (C4+) increase the dew point of the
acid gas mixture and may lead to unwanted condensation of
the acid gas (especially on the interstage).
The presence of too much methane in the acid gas may result
in a gas bubble forming in the well which impairs and may
even prevent injection.
9.5
Effect of Hydrocarbons on the Amine Plant
Hydrocarbons do not chemically bond to the amine, so there is no direct
harmful effect on the actual amine solution. Given the right amount of
time, or use of technology, any hydrocarbons mixed into the amine can
be removed and the amine can then be reused. In the meantime, however,
while a mixture of hydrocarbon and amine is circulating through the plant,
many negative consequences can result, the most common being as follows:
Foaming. Probably the first and foremost concern when
amines are contaminated with hydrocarbons is foaming.
Hydrocarbons have a lower surface tension than amine,
168 Carbon Dioxide Capture and Acid Gas Injection
Figure 9.5 Emulsified Hydrocarbons in Amine.
which allows the surface of the liquid to expand quite easily. When gas or steam is bubbled through the amine, the
bubble reaches the surface of the liquid, but does not “pop”.
The bubble remains, and more and more bubbles build up
on top of it until the entire vapor space is filled with this
foam, as illustrated in Figure 9.6. Foaming can be a concern
in both the contactor and the regenerator. When amines are
in a foam state, they do not remove H2S in the contactor and
cannot be regenerated in the regenerator.
Fouling. Hydrocarbons contribute to the “black shoe polish” that commonly fouls amine filters, lean/rich exchangers
and packed towers (see Figure 9.7). A carbon bed is used to
remove hydrocarbons, but because the carbon is generally
on the lean side, the amine has to flow through many pieces
of equipment beforehand, where hydrocarbons can form a
Strategies for Minimizing Hydrocarbon Contamination 169
Figure 9.6 Foaming Tendency of Hydrocarbon Contaminated Amine.
Figure 9.7 Black shoe polish on rich amine filters.
170 Carbon Dioxide Capture and Acid Gas Injection
matrix along with iron sulphides, salts, degraded amine, and
antifoam. This is what often fouls amine systems.
Gasket Destruction of Plate and Frame Exchangers. There
are several different types of gaskets available for separating the
plates in plate-and-frame exchangers. None of these gaskets
are immune to the harmful effects of hydrocarbons flashing
across the exchanger. Liquid hydrocarbons can cause polymerization of the gasket material and flashing hydro­carbons can
erode the gaskets. Operators must rely on the flash tank to
minimize the hydrocarbon content of the rich amine entering
the exchanger. Otherwise the risk of gasket failure and leaking
substantially increases, as shown in Figure 9.8.
Figure 9.8 Leaking exchanger gaskets.
Strategies for Minimizing Hydrocarbon Contamination 171
Loss of Treated Product. Hydrocarbons in amine represent
a loss of hydrocarbons in the treated gas (i.e., result in lower
volumes of sales gas or LPG). These losses directly impact
the profitability of the gas processing system, and are obviously undesirable.
9.6 Minimizing Hydrocarbon Content in Amine
Acid Gas
The best way to minimize hydrocarbons in amine acid gas is to ensure that
excess hydrocarbon does not enter in contact with the amine solution in
the first place. This requires a comprehensive and thorough evaluation of
the inlet gas stream to each amine contactor in the system. Typical amine
plant designs will include at least an inlet separator before the contactor to
knock out free liquids. However, many of these inlet separators are inadequate for complete hydrocarbon liquid removal.
Other options for minimizing the hydrocarbon content of the amine
acid gas include the following:
Option 1. Optimizing the operation of the actual amine
plant.
Option 2. Optimizing the amine flash tank operation.
Sometimes design changes to the tank interior may be necessary to ensure minimal hydrocarbon breakthrough.
Option 3. Using filtration and/or coalescing technology on
the rich amine to remove hydrocarbons.
Option 4. Using potential skimming capabilities on contactor, flash tank and reflux water.
Option 5. Technological solutions.
9.6.1
Option 1. Optimization of the Amine Plant Operation
Once the plant is designed, built and operating, there are two strategies
operators can employ to minimize the level of hydrocarbon pick up by the
amine: (i) preventing hydrocarbon entering the contactor with the sour gas
in the first place, and (ii) operating the plant at conditions that minimize
hydrocarbon solubility in amine, noting that examples include choice of
amine type and strength, circulation rate, rich loading, differential temperature between lean amine and inlet gas, and reflux operations.
172 Carbon Dioxide Capture and Acid Gas Injection
Preventing Hydrocarbon Ingress. Assessing the amount of
hydrocarbon entrainment in gas can be directly analyzed.
Several companies perform entrainment testing at moderate costs. When done correctly, the associated optimizations
that are possible as a result of the new data rapidly pays back
the cost of the testing. The resulting changes that are made to
process conditions can ensure reliability, integrity, capacity
and energy/chemical utilization, thereby resulting in valuable cost savings to the plant.
Bulk liquid hydrocarbons are meant to be removed from
inlet gas streams by the inlet separator. More precision
removal can be done using a cyclone separator and/or a
coalescing filter. The inlet separator is the most important
piece of equipment as far as hydrocarbons in amine acid
gas are concerned—if the inlet separator fails, there will be
serious consequences for the sulphur plant. Cyclones, centrifuges and coalescers are primarily used to prevent foam
promoting contaminants from entering the amine contactor,
which is of course important, but will not necessarily make
much of an impact on the level of hydrocarbons in the AAG.
Inlet separators rely on four basic parameters which
­determine the effectiveness of liquid separation from gas:
Density Difference between the Liquid and Gas. Higher density liquids will be removed easier from gas than
lighter ones because of the lack of gas solubility in the
hydrocarbon. The two phases want to separate.
Flow Directional Change. Flow direction changes are
possible for the gas, but not so much for the droplets of
liquid. Forcing the gas around an impingement plate followed by a demister pad in the top of the vessel creates
something similar to an obstacle course. The gas can go
through it, but liquid droplets impinge on the surface
of the obstruction and eventually build in size until the
droplets fall to the bottom of the separator. Figure 9.9 is a
typical inlet gas separator, with a diverter plate and demister pad for flow directional change.
Velocity. Velocity has a large effect on the volume of liquid hydrocarbon in a gas stream. The velocity of the gas
stream imparts drag force on each liquid droplet, pulling the droplet along the pipe. The only opposing force
Strategies for Minimizing Hydrocarbon Contamination 173
Gas
outlet
Mist
eliminator
Diverter plate
Inlet gas
Liquid
Liquid
outlet
Figure 9.9 Typical Inlet Gas Separator.
to counter this is gravity. It is important that gas velocities not be so high that they overcome gravity. Gas flow,
as well as gas pressure, determines the velocity through
the piping. This is why a separator has to be carefully designed to minimize pressure drop across the vessel, since
a drop in pressure results in the gas expanding and a corresponding increase in velocity and drag force.
Time. Time is the final separation parameter. It takes time
for gravity to pull droplets out of a gas stream. Therefore,
a larger separator tends to remove more liquids than a
smaller vessel, assuming proper design of each.
Proper operation of an inlet separator involves ensuring the liquid level
is kept low at all times. The frequency of the level control valve opening
should be noted, since frequent dumping of the vessel could mean an
excessive amount of liquids having to be removed from the gas stream.
This could indicate that that there is a possible problem upstream that
requires investigation.
Separators should always operate with a low and consistent pressure
drop. Low pressure is required to prevent excessive drag forces on liquid
174 Carbon Dioxide Capture and Acid Gas Injection
droplets and consistent pressure is to ensure that there is no fouling or
plugging of the demister pad within the vessel.
Plant Operating Conditions. There are several operational
parameters that directly affect the amount of hydrocarbon
in amine acid gas:
Amine Circulation Rate. Because of the inevitable solubility of hydrocarbons in amine (described earlier in this
paper), a higher amine circulation rate of amine will carry more hydrocarbons into the circulating rich solution.
Furthermore, higher amine circulation rates decrease the
flash tank residence time, lowering the hydrocarbon removal efficiency.
Amine Rich Loading. By increasing the loading (aqueous CO2 and H2S amine salt concentration) of the amine
solution, less amine and water is available for interaction
with the charged part of the hydrocarbon. This renders the
hydrocarbon-hydrocarbon interactions stronger than the
amine-hydrocarbon interactions, causing the hydrocarbon molecules to coagulate through hydrophobic interactions with one another. The fact that higher rich loadings
reduce the amine-hydrocarbon solubility levels is even further reason to lower the amine circulation rate if possible.
Differential Temperature between Lean Amine and
Inlet Gas. Before gas streams enter the amine contactor, the gas streams pass through an inlet gas separator,
which allows for liquid hydrocarbons to be separated.
Therefore, the gas leaving the separator should be at the
hydrocarbon dewpoint, meaning if the gas pressure were
to increase or the temperature to decrease, hydrocarbons
would condense out and form droplets in the gas line. It
is important to not allow this to occur as it will result in
liquid hydrocarbons in the rich amine.
For this reason, it is recommended the inlet gas separator be located within 10 m of the amine contactor, which
will minimize the risk of condensation of hydrocarbons
along the pipeline. It is also recommended that this line
be insulated.
It is also possible for hydrocarbons to condense inside
the actual amine absorber, which can happen if the gas is
cooled while travelling up the contactor. This will hap-
Strategies for Minimizing Hydrocarbon Contamination 175
pen if the lean amine being injected into the contactor is
cooler than the inlet gas stream. In fact, because the hydrocarbon dewpoint of the gas changes as acid gases are
removed (because the removal of acid gases, the hydrocarbon dewpoint temperature will be higher at the top of
contactor as compared to the bottom), it is recommended
that operators maintain a minimum five Celsius or Fahrenheit degree temperature differential between the lean
amine and the inlet gas. On especially rich gas streams, a
greater differential temperature may be required.
Reflux Operation. Despite optimizing amine circulation rates, rich loadings and flash tank operations, there
will still likely be hydrocarbons entering the regenerator. The last area where hydrocarbons can be removed
before leaving with the acid gas is in the reflux system.
Proper condensing of the regenerator overhead stream
will minimize the level of hydrocarbon vapor and the hydrocarbons will instead circulate with the reflux. The recommended reflux temperature is between 35 and 45 °C,
or 95 to 115°F. Within the amine industry, there is some
variation of which end of this temperature range is recommended. Refiners tend to run on the higher end of
this range to minimize the risk of ammonia salt precipitation. Gas plants, or systems with no ammonia ingress
(these are the systems who would be utilizing an AGIS),
should target lower reflux temperatures.
Since hydrocarbons (including methanol) are condensed in the reflux, the concentration will increase if there
is no reflux purge. An increase in hydrocarbon content in
the reflux is undesirable, as this increases the risk of hydrocarbon carryover with the acid gas, and can also cause
foaming and other operational problems in the regenerator. Hydrocarbon contaminated reflux streams should be
either continuously purged to a sour water stripper or disposal tank. If purging is not available, completely emptying the reflux tank to disposal is acceptable. This should be
done based on visual or laboratory analysis of reflux water
(or better yet reflux skimming) for hydrocarbons.
It is important to recognize the implications of operating outside the recommended reflux temperature range.
176 Carbon Dioxide Capture and Acid Gas Injection
Specifically operating at less than 35 °C/95 °F will not only
have limited effect on the amount of water and hydrocarbon in the acid gas, but will also increase the necessary
reboiler duty since the cold reflux will be condensing an
inordinate amount of steam traffic in the upper section
of the regenerator. Operating at higher than the recommended reflux temperature will allow excessive water
and hydrocarbon to escape with the acid gas, negatively
affecting acid gas compressor operations as well as increasing the make-up water demand on the amine plant.
9.6.2
Option 2. Amine Flash Tanks
Separation technologies based on pressure drop, velocity changes and
residence time are among the most common separation systems used in
oil and gas operations. All these technologies have the common theme of
using simple concepts to attempt to solve a separation problem. One such
example is the amine flash tank. This device removes off-gases by reducing
the rich amine pressure downstream of an amine contactor.
If designed correctly, these systems also provide limited liquid-liquid
separation capabilities for free hydrocarbon removal since these hydrocarbon liquids will float to the top of the amine solution within 30 minutes and
can then be skimmed. No emulsified or dissolved contaminant is affected.
For a flash tank to effectively separate hydrocarbons from amine, there
must be sufficient residence time as well as sufficient pressure reduction.
Entrained hydrocarbon gases will flash off within three to five minutes,
noting that the lower the flash tank pressure, the faster and more efficient
hydrocarbon gases will flash.
Flash tank pressure is determined by the necessary flash gas and rich
amine pressures. Common flash gas destinations include the following:
Low pressure fuel gas absorber inlets
Flare
Incineration
Acid gas
The rich amine must leave the flash tank and flow through the following:
Rich amine filters (if they exist)
Lean/rich exchanger
The vertical piping up to almost the top of the regenerator
Strategies for Minimizing Hydrocarbon Contamination 177
Generally, 70 psig/475 kPag is enough pressure to push both the gas and
amine to their respective destinations. It is important to not pressure the
flash tank any higher than necessary since this will reduce the amount of
hydrocarbon flashing. If the flash gas is routed to an extremely low pressure flare or incinerator, some plants will operate the flash tank at virtually
atmospheric pressure and will install rich amine pumps immediately after
the flash tank in order to push the amine through the exchanger and into
the regenerator.
Flash tank residence time is a function of the size of the vessel, the level
at which the amine is maintained inside the vessel, and the circulation rate
of the amine. If the tank is designed only for two-phase separation, the
tank will simply be an open tank with a gas outlet in the top and liquid
drain at the bottom. Operators should set the level at 50 to 60% full. This
maximizes residence time, while still allowing for vapor disengaging space,
noting that when hydrocarbons flash from liquid to gas, the hydrocarbons
expand in size by up to 300% and can therefore carry gas upwards as a
result of high velocities. Having at least 40% of the tank as vapor space will
allow for gravity to pull the droplets of that amine back out of the flash gas
and into solution.
For designers, sizing a 30-minute residence time is often achieved by
assuming some variables in the Stokes Law calculations, such as hydrocarbon density and hydrocarbon droplet size. The typical values for these
tend to be a specific gravity of 0.6 to 0.7 and droplet sizes of 150 microns
and larger. It is important to note that these values do not account for the
possibility of heavier hydrocarbons and micro-emulsions that exist in rich
amine streams.
The Stokes Law calculations also do not account for any possible surfactant contaminants, which aid in stabilizing micro-emulsions. Granted,
designers must balance the size and cost of the flash tank with expected
performance. Designing a flash tank for 10 micron droplet removal would
make the flash tank so large that it would not be economical to build or fit
within the battery limits of the unit.
Assuming the residence time is adequate, free hydrocarbons floating on
the amine surface can and should be skimmed. Depending on the flash
tank design, these hydrocarbons may flow over a weir or into an internal
bucket that is attached to the inside of the vessel walls. When skimming,
it is preferential to be able to sample the skimmed liquid to ensure it is
indeed hydrocarbon and not amine. Some skimming sections have site
glasses where the interface between amine and hydrocarbon can be seen.
These make it very simple for operators to prevent the skimming of amine
to disposal.
178 Carbon Dioxide Capture and Acid Gas Injection
A number of different flash tank designs are available, some certainly
more effective than others. A growing trend is to incorporate metal mesh
internals to promote coalescence, which can compensate for low residence
times. These promising systems so far have provided marginal results
because of poor understanding of highly fouling rich amine streams. Most
flash tanks, whether two- or three-phase separators, are often sized based
on correct parameters, but can lack the understanding of liquid and solid
contaminant loading. To use coalescing mesh-pads correctly, these have to
be designed not only according to the gas velocity across the pad (using the
modified Souders-Brown equation), but should also consider liquid and
solid properties in addition to internal flow geometry. Any disregard of
these aspects will invariably lead to element flooding and liquid carry-over
or fouling with an increase in differential pressure.
Case Study
In one case study used to verify the strategies presented in this paper,
several operational changes were made to a gas plant in Wyoming.
The effects of the hydrocarbon content of the acid gas were measured.
This plant had an inlet gas composition involving:
614 kmol/h hydrocarbons
11.8 kmol/h BTEX
5% H2S
8% CO2
System pressure of 5740 kPa(g)
As is shown in Table 9.1, as the amine circulation rate, flash tank
pressure and reflux temperature were dropped, a significant decrease
in hydrocarbon and BTEX content of acid gas was seen.
9.6.3
Option 3. Rich Amine Liquid Coalescers
Presently, many amine units only separate hydrocarbons in the flash tank. It is
commonly believed that flash tanks will be able to separate any hydrocarbon
in the rich amine streams. This is correct to some extent. However, the reality
is that emulsions in rich amine streams are very stable, with droplet sizes near
or less than the micro-emulsion range (10 microns and smaller). If Stokes
Law is used to calculate the required residence time for the separation of a
15 micron emulsion, the result will indicate approximately two days. If the
Strategies for Minimizing Hydrocarbon Contamination 179
Table 9.1 Hydrocarbon reduction case study.
Optimization strategies (in
order of implementation)
Acid gas
hydrocarbon
content
Btex content
Treated gas
Original operational
conditions:
50% MDEA
160 m3/h
Lean amine temperature 50 °C
Inlet gas temperature 40 °C
Rich loading 0.14 mol/mol
Flash tank pressure 758 kPa
Reflux temperature 60 °C
1.36 kmol/h
12.24 kmol/h 2.4 ppm H2S
0.174% CO2
Decrease Amine to 70 m3/h
(0.314 mol/mol rich loading)
0.44 kmol/h
5.19 kmol/h
<1 ppm H2S
0.479% CO2
Flash tank pressure dropped to
480 kPag
0.275
5.10
<1 ppm
0.48 % CO2
Flash tank pressure dropped to
35 kPag, rich pump installed
0.017
3.16
<1 ppm
0.48 % CO2
Decrease reflux to 35°C
0.016
3.14
<1 ppm
0.48 % CO2
particle size is slightly smaller, then the separation time can be in the order of
weeks or even months. Based on the low efficacy of residence tanks to properly separate these emulsified contaminants to the level required for feed into
the regenerator, it is necessary to use secondary systems, such as coalescers.
Coalescence is the recombination of two or more small liquid droplets to
produce a single droplet that is larger in size. This phenomenon also takes
advantage of Stokes Law, which relates the velocity of separation of a particle or droplet in a medium to the diameter of the contaminant, densities,
viscosity and gravitational pull. As coalescence takes place, small micronsize contaminant droplets coalesce into fairly large droplets, resulting in an
almost immediate separation from the continuous phase.
Mechanical coalescing systems are basically comprised of the following
three technologies:
Inclined plates
Metal mesh
Microfiber based
Because of the particle size and the high fouling properties of the solids
and emulsified hydrocarbons in rich amine steams, only disposable filters
180 Carbon Dioxide Capture and Acid Gas Injection
and microfiber-based coalescers are able to provide the proper particle
removal and emulsion separation. Other systems, such as back-flushable
metal-based filters, do not work because the adhesion of solids to the surface is too strong, and prevents a backwash from being effective.
Suspended solids removal upstream of the coalescer is mandatory, as
this will protect the coalescer elements and will also help destabilize the
emulsion. If solids are introduced into a coalescing filter, it will plug off
almost immediately and become ineffective.
Liquid-liquid coalescers are devices designed to separate small emulsified liquid contaminants in a liquid stream. These devices are segregated
into the following two general categories:
Low efficiency systems with metal-based internals
Coalescers with microfiber internals
The two types of coalescers perform rather differently and should be
used for different objectives and specific cases. While metal-based internal coalescers are effective for separating free liquids and macro-emulsions
(~100 microns and larger), coalescers with microfibers are more suited for
separating micro-emulsions (100 microns and smaller).
9.6.4
Option 4. Use of Skimming Devices
Hydrocarbons which float to the surface of amine or reflux water may be
skimmed, assuming the design of the vessel allows for skimming. Most flash
tanks have skim connections or hydrocarbon carryover weirs included in the
design. Some contactors, reflux accumulators and regenerators also have skim
connections. Whenever possible, hydrocarbons should be skimmed out of the
system in order to prevent any chance of evaporation. Areas with higher residence times, such as flash tanks and regenerator bottoms have the potential
for particularly high hydrocarbon volumes, as shown in Figure 9.10, which
presents a side-by-side comparison of regenerator skimmings and regenerator bottoms, taken less than one minute apart from the same regenerator.
Although not commonly thought of being a “filter”, the regenerator
absolutely will drive hydrocarbons out of the amine, in effect filtering
the solution. The vaporized hydrocarbons enter the reflux system where
the hydrocarbons are possibly condensed in the reflux and where if not
removed, will eventually be recycled back to the regenerator with the reflux
water. Reflux water that is contaminated should therefore be partially
purged, or better yet skimmed. Figure 9.11 shows hydrocarbons floating
on reflux water, a common sight in hydrocarbon-contaminated amine
systems.
Strategies for Minimizing Hydrocarbon Contamination 181
Figure 9.10 Regen skimmings (left), Regen bottoms (right).
Figure 9.11 Reflux water with skimmable hydrocarbon.
182 Carbon Dioxide Capture and Acid Gas Injection
9.6.5
Option 5. Technological Solutions
More stringent legislation has led to additional treatments being proposed
in order to meet current and future environmental specifications. Morrow
[7] proposed the addition of a stripping column for hydrocarbons on the
rich amine line (see Figure 9.12). This design uses a portion of the sweet
gas to strip VOC and BTEX from the amine. The design has been patented
and is described by Morrow [7], Morrow & Wallender [10], Morrow &
Lunsford [6] and McIntyre et al. [8]. A VOC and BTEX removal of greater
than 75% is apparently possible, Morrow & Wallender [10]. Bullin &
Brown [2] modelled the performance of a hypothetical hydrocarbon stripping column on an Methyldiethanolamine (MDEA) plant and found it to
remove 70% of the benzene (although 10% of the CO2 in the stream was
also liberated, which could certainly be a problem).
Bullin & Brown [2] also model a hot flash vessel (see Figure 9.13) that
is located downstream of the lean-rich heat exchanger on the rich amine
stream. Essentially, a higher temperature flash leads to more hydrocarbons
being removed. The simulations show that this is more effective at removing VOCs and does not have a significant impact on BTEX. Unfortunately,
substantial amounts of CO2 and H2S are liberated at these temperatures
with the VOCs requiring an additional amine contactor for these gases.
Sweet gas
Acid gas
to SRU
Hydrocarbons
Absorber
Regenerator
Lean/rich
exchanger
Flash
gas
Feed gas
Rich amine
flash
Hydrocarbon
stripper
Lean amine pump
Figure 9.12 Amine plant with hydrocarbon stripping column.
Strategies for Minimizing Hydrocarbon Contamination 183
Sweet gas
Acid gas
to SRU
Hot flash
gas
Absorber
Lean/rich
exchanger
Flash
gas
Regenerator
Hot flash &
amine treating
Feed gas
Rich amine
flash
Lean amine pump
Figure 9.13 Amine system with hot flash vessel.
References
1. Borda, R.T. 2011. Experimental Measurement of Multi-component BTEX
Solubility in Amine Solutions. M.Sc. Thesis, University of Oklahoma, 2011.
2. Bullin, J.A., and W.G. Brown. 2004. Hydrocarbon and BTEX Pickup and Control
from Amine Systems. Proceedings of the 83rd Gas Processors Association
Convention, 2004.
3. Carroll, J.J., J. Maddocks, and A.E. Mather. 1998. The Solubility of Hydrocarbons
in Amine Solutions. Laurence Reid Gas Conditioning Conference, March
1998.
4. Critchfield, J., P. Holub, H. Ng, A. E. Mather, F. Jou, and T. Bacon, “Solubility
of Hydrocarbons in Aqueous Solutions of Gas Treating Amines”, Proceedings
of the 2001 Laurance Reid Gas Conditioning Conference.
5. Hatcher, N.A., C.E. Jones, and R.H. Weiland. 2013. Hydrocarbon Solubility
in Amine Treating Solvents: A Generalized Model. Laurence Reid Gas
Conditioning Conference, February 2013.
6. Morrow, D.C., and K.M. Lunsford. 1997. Removal and Disposal of BTEX
Components from Amine Plant Acid Gas Streams. Pages 171 to 173 of the
Proceedings of the 76th Annual GPA Convention. San Antonio, Texas,
1997.
7. Morrow, D.C. 1996. Removal/Disposal of BTEX Components in Amine Systems.
Permian Basin Regional GPA Meeting, 1996.
184 Carbon Dioxide Capture and Acid Gas Injection
8. McIntyre, G.D., V.N. Hernandez-Valencia, and K.M. Lunsford. 2001. Recent
GPA Data Improves BTEX Predictions for Amine Sweetening Facilities.
Proceedings of the 80th Gas Processors Association Convention, 2001.
9. Spooner, B.H., and D.L. Engel. 2012. Reducing Hydrocarbons in Sour Water
Stripper Acid Gas. Sulphur 2012 Conference, November 2012.
10. Wallender, J.W., and D.C. Morrow. 1999. Reduction of BTEX Emissions From
Amine Plant Acid Gas Streams. GRI Gas Industry Air Toxics Conference, San
Antonio, Texas, May 24–26, 1999.
10
Modeling of Transient Pressure Response
for CO2 Flooding Process by Incorporating
Convection and Diffusion Driven
Mass Transfer
Jianli Li and Gang Zhao
University of Regina, Regina, Saskatchewan, Canada
Abstract
Traditionally well testing models simulate the CO2 flooding process as tworegion or three-region composite models, which normally neglect the dynamic
mass transfer process and over-simplify the transient viscosity in transition zone.
This leads to insufficient technical capacity to deal with more complicated field
­situations where reservoir heterogeneity largely affects CO2 injection process.
Aiming at eliminating this restriction/limitation, this study proposes a comprehensive transient pressure model for CO2 flooding. It consists of two submodels
for the pressure and mass transport processes, respectively. The proposed reservoir physical system is actually an enhanced three-region composite model which
includes CO2 bank, transition zone, and oil zone. Pressure change in each region
are ­properly modeled and CO2 concentration change in the transition zone is also
reasonably described. The entire model is solved in a semi-analytical manner
and a trapezoidal approximation scheme is used for variable flow velocity and oil
­viscosity. Type curves of the proposed comprehensive transient pressure model are
plotted. Four flow regimes are identified: the early radial flow, the transition flow,
the pseudo-radial flow, and the boundary-dominated flow regimes. The investigation of the impact of mass transfer process in the transition zone concludes that
it mainly affects the transition flow regime with much slower slop change and the
pseudo-radial flow regime with lower straight line compared with the case without
considering the mass transfer process in the transition zone. This is caused by a
gradual change in properties, such as total compressibility and viscosity, from a
CO2 bank to a transition zone rather than an abrupt change for properties from
a CO2 bank to an untouched oil zone. Sensitivity analysis shows that an injection
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (185–198)
2017 © Scrivener Publishing LLC
185
186 Carbon Dioxide Capture and Acid Gas Injection
scheme with smaller injection rate and longer injection period is better for viscosity reduction due to time dependent dynamic process than that of a larger injection rate and a shorter injection period in terms of viscosity reduction with the
same amount of injected CO2.
10.1
Introduction
As one of the most plentiful compounds found on this planet, the idea of
using CO2 to remove oil from underground reservoirs originated as early
as 1952 when a patent for an oil recovery method with CO2 was proposed.
Through six decades’ development, CO2 flooding has been widely applied
to enhance oil recovery all over the world. In United States, all the largest
CO2 flooding projects were miscible displacements of medium-to-heavy
oils with a number of injection and producing wells, except for an immiscible Yates Sand flood in West Texas. There have been 120 CO2 flooding
projects by 2012. About three billion cubic feet of CO2 is supplied every
day for the flooding projects, leading to an incremental oil production of
about 350,000 bbl/d. It is estimated that 42 billion barrels of recoverable
reserves has been added [Winslow, 2012].
CO2 flooding is carried out by injecting a large amount of CO2 into a
reservoir. This process can be either an immiscible or a miscible process
depending on whether the operating pressure achieves the minimum
miscibility pressure (MMP) or not. Immiscible processes were studied in
the 1950s and 1960s, and the recovery mechanisms were identified to be
reduction in oil viscosity, oil swelling, and dissolved-gas drive. Miscible
displacement can be achieved when the operating pressure is greater than
MMP. It is much more effective than an immiscible displacement in terms
of oil recovery. Production mechanisms in a miscible process involve:
CO2–hydrocarbon miscible drive, hydrocarbon vaporization, direct
­miscible CO2 drive, and multiple-contact dynamic miscible drive [Holm
and Josendal, 1974; 1987].
Traditionally, well testing models simulate the CO2 flooding process as
two-region or three-region composite models, which normally neglect the
dynamic mass transfer process and oversimplify the transient viscosity in
the transition zone. This leads to insufficient technical capacity to deal with
more complicated field situations where reservoir heterogeneity largely
affects the CO2 injection process.
This study proposes and develops a comprehensive well testing model for
CO2 flooding process by incorporating convection–diffusion driven mass
transfer process in modeling transient pressure response. The pressure model
Modeling of Transient Pressure Response for CO2 Flooding 187
covers three regions (gas bank, transition zone, and oil zone) while the mass
transfer model describes fluid dynamics in a transition zone. The two models
are reasonably coupled together. Bottom-hole pressure is calculated and type
curves are plotted. Effects of reservoir and fluid properties and operating
parameters on dimensionless pressure and its derivative are analyzed.
10.2
Model Development
10.2.1
Pressure Diffusion
A reservoir undergoing a CO2 miscible flooding process consists of three
zones: a CO2 bank, a miscible zone, and a crude oil bank (Figure 10.1).
CO2 and crude oil banks are assumed to be filled purely with carbon dioxide and crude oil, respectively. The mixing of crude oil and CO2 happens
only in the miscible zone. Also, water phase is eliminated in this study and
the reservoir has a radial shape with uniform porosity and permeability.
The entire model contains two submodels: pressure diffusion over all the
three regions and mass transfer in the transition zone.
CO2 bank is considered to be filled with gas and residual oil. In the transition zone, the mixture of crude oil and CO2 is assumed to be in liquid
state. The untouched oil zone is treated as an extension of the miscible
zone in which concentration equals zero. The pressure distribution over
the entire model is described by a diffusivity equation [Lee, 1996]:
1
P P
r
r r
z r
P Ct P
,r
z k t w
Well
rw rBK
CO2 bank
rMZ
Miscible zone
Oil zone
Figure 10.1 Schematics of a reservoir undergoing CO2 flooding
r
re
r1 (10.1)
188 Carbon Dioxide Capture and Acid Gas Injection
Ct P
,r
k
t 1
1
P
r
r r
r
r
re
(10.2)
where r is the radial distance from an injector, m; P is the pressure, Pa; µ is
the viscosity, Pa.s; z is the compressibility factor, dimensionless; is the
porosity, dimensionless; k is the permeability, m2; Ct is the total isothermal
compressibility, Pa–1; r1 is the location of a preset boundary between a CO2
bank and a miscible zone, m; and t is the time variable, s. The viscosity,
compressibility factor, and total isothermal compressibility are all functions of pressure. At the inner boundary (injector borehole radius), CO2 is
injected at a constant rate. The outer boundary is considered to have three
types: a closed boundary, a constant pressure boundary, and an infinite
boundary. Initially, the gas reservoir has a uniform pressure distribution.
r
P
r
P re , t
q Bg
P
r
(10.3)
2 kh rw
0 (Closed boundary)
r re
(10.4)
Pi (Constant pressure boundary)(10.5)
P r
,t
P r ,0
Pi (Infinite reservoir)(10.6)
Pi
j 1, 2,
,N
(10.7)
where rw is the wellbore radius, m; T is the temperature, K; h is the reservoir
thickness, m; and subscript “sc” denotes the standard condition.
10.2.2
Mass Transfer
The mixing of CO2 and crude oil is caused by not only molecular diffusion but also convective dispersion across the CO2–oil miscible zone. The
­governing equation of a mass transfer process in a radial system:
c
t
D
2
c
r
2
1 c
r r
V
c
V
c
,r
r
r 1
r
re
(10.8)
Modeling of Transient Pressure Response for CO2 Flooding 189
where c is the concentration, dimensionless; D is the diffusivity, m2/s; and
V is convection velocity, m/s. The convection velocity is calculated by using
Darcy’s law:
k
V
P
(10.9)
The inner boundary of the miscible zone is assumed as completely
­saturated with CO2 all the time, which indicates a Dirichlet BC. The outer
boundary of the miscible zone is regarded as a no-flow boundary and a
Neumann BC is applied. Initially, the model is free of CO2. The BCs and
IC are described:
c*
cr
1
cr
0
e
ct
(10.10)
(10.11)
(10.12)
0
0
Viscosity
r
i
1
P
r
ra1
P1
P
r
N
ri
Pi
P
r ran
Pn
Figure 10.2 Schematics of a trapezoidal approximation of the viscosity profile in the
miscible zone
190 Carbon Dioxide Capture and Acid Gas Injection
where c* is the saturation concentration under certain operating ­conditions,
dimensionless.
10.2.3
Solutions
A pseudo-pressure and pseudo-time method is applied to accurately solve
nonlinear governing equation, Eq. (10.1), in which the pseudo-terms are
defined as
m(P ) 2
t ap (t )
P
P0
t
0
P
dP
z (10.13)
1
dt
Ct (10.14)
For the miscible zone, due to the complex relationships among
v­ iscosity, pressure, and concentration, it is difficult to solve Eqs. (10.2)
and (10.8) simultaneously. Thereby, this study compute the pressure and
­concentration stepwise. At a step, the true curved viscosity profile is
simplified as a trapezoidal profile. A piecewise linearization scheme and
Laplace transformation are applied to obtain the semi-analytical solutions to each segment of the transition zone. Then the solutions are coupled together through reasonable two conditions to ensure the ­continuity
and smoothness on the pressure profile at the interface of neighbouring
segments:
Pj (rj , t ) Pj 1 (rj , t )
k Pj
r
rj
k Pj 1
r
(10.15)
(10.16)
rj
here the subscript of pressure, “i” (i = 1, 3 … n), denotes the pressure of the
ith segments of the miscible-zone. At last, a matrix form of equations can be
formed which is solved by using the Guess Elimination algorithm [Muller,
2001]. Detailed solution theory can be found in a previous study [Jia et al.,
2013]. The mass transfer model is solved in a similar manner as that for
the pressure model. Two coupling conditions are applied at the interface in
between two neighboring segments to ensure the continuity and smoothness on the concentration profile:
Modeling of Transient Pressure Response for CO2 Flooding 191
Input
D, Ct, k, , h, Bg, q
Pressure transfer model
G.E., BCs, IC
Pressure, pressure gradient
P, P
Convection velocity
V
Mass transfer model
G.E., BCs, IC
Viscosity
Concentration
c
No
Termination
condition
Yes
Finish
Figure 10.3 Flowchart for calculating the solutions to the pressure diffusion and mass
transfer models.
c j (rj , t ) c j 1 (rj , t )
cj
cj
r
r
rj
1
(10.17)
(10.18)
rj
Figure 10.3 shows a flowchart for calculating pressure and concen­tration
profiles in the above-developed CO2 miscible model.
10.3
Results and Discussion
10.3.1
Flow Regimes
The dimensionless pressure and its derivative for the transient pressure
model of CO2 flooding considering mass transfer process are ­calculated.
The main flow regimes are identified according to the type curves.
Figure 10.4 shows the pressure characteristics of a CO2 miscible flooding with the consideration of mass transfer process. It can be seen that
192 Carbon Dioxide Capture and Acid Gas Injection
1000
4
3
pD, dpD/dln(tD)
100
2
10
1
1 The first radial flow regime
1
2 Transition flow regime
3 Pseudo radial flow regime
4 Boundary dominated flow regime
0.1
1.0E + 02 1.0E + 03 1.0E + 04 1.0E + 05 1.0E + 06 1.0E + 07 1.0E + 08 1.0E + 09
tD
Figure 10.4 Type curves of the transient pressure model for CO2 miscible flooding
considering mass transfer process (Ra = 10 m, Re = 150 m, Q = 0.01 m3/s, Ct = 1e-8,
k = 10 mD, µ = 4.7 cP).
there are mainly four flow regimes on the type curves: the first radial
flow regime, transition flow regime, the pseudo-radial flow regime, and
­boundary-dominated flow regime. Figure 10.5a shows the relationship
between CO2 concentration and distance as time goes. It can be seen that
the CO2 concentration in the miscible zone increases and expands outwards with the continuous injection of CO2. Figure 10.5b shows the visco­
sity of CO2–oil mixture versus distance as time goes. It is an obvious trend
that the larger the concentration of CO2 in the mixture, the greater the
viscosity will be reduced.
10.3.2
Effect of Mass Transfer
Viscosity reduction through CO2–oil mass transfer is a key production mechanism of CO2 miscible flooding processes. The varying visco­
sity in the miscible zone, in return, influences the pressure propagation.
Figure 10.6 compares the transient pressure responses with and without
considering the mass transfer process in the miscible zone. The differences
in dimensionless pressure and pressure derivative caused by the viscosity
profile in the miscible zone are quite obvious. This means that the mass
transfer process should not be ignored for a reliable bottom-hole pressure
analysis.
The first flat section of pressure and pressure derivative curves with
solid lines is due to a radial flow at the early time period. The following
Modeling of Transient Pressure Response for CO2 Flooding 193
1
0.9
t = 1e3 seconds
t = 1e4 seconds
t = 1e5 seconds
t = 1e6 seconds
Concentration, Vol. %
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
10
15
20
25
30
35
Distance, m
(a)
0.005
0.0045
0.004
Viscosity, Pa.s
0.0035
0.003
0.0025
0.002
0.0015
t = 1e3 seconds
t = 1e4 seconds
t = 1e5 seconds
t = 1e6 seconds
0.001
0.0005
0
10
(b)
15
20
25
Distance, m
30
35
Figure 10.5 (a) CO2 Concentration and (b) Oil Viscosity distributions across the miscible
zone (Ra = 10 m, Re = 150 m, Q = 0.01 m3/s, Ct = 10–8 Pa–1, k = 10 mD, µ = 4.7 cP).
sharp increase of pressure and pressure derivative curves is caused by
a sudden change in fluid properties, such as total compressibility and
viscosity, from a CO2 bank to a miscible zone, which is defined as a
transition flow period. With the decrease of the viscosity in miscible
zone, the pressure and pressure derivative curves go down slightly. Then
the flow arrives at the pseudo-radial flow period followed by boundary-­
dominated flow indicated by the unit slop in the pressure derivative
curve.
194 Carbon Dioxide Capture and Acid Gas Injection
1000
pD, dpD/dln(tD)
100
10
1
With mass transfer
With mass transfer
Without mass transfer
Without mass transfer
0.1
1.0E + 02 1.0E + 03 1.0E + 04 1.0E + 05 1.0E + 06 1.0E + 07 1.0E + 08 1.0E + 09
tD
Figure 10.6 The effect of mass transfer process on dimensionless pressure and
dimensionless pressure derivative.
Table 10.1 Physical properties and operating conditions of the base case.
Properties
Symbol
Value
Unit
Total Compressibility
Ct
1 × 10-8
Pa
Initial reservoir pressure
Pi
15´106
Pa
Well radius
Rw
0.091
m
CO2 bank radius
Ra
10
m
Reservoir outer radius
Re
150
m
Thickness
h
1
m
Permeability
k
1 × 10–14
m2
Porosity
f
0.26
1
Injection rate
Q
0.01
m3/s
Temperature
T
333.15
K
Pressure at standard condition
Psc
1 × 105
Pa
Temperature at standard condition
Tsc
293
K
Critical Pressure
Pc
7.377 × 106
Pa
Critical Temperature
Tc
304.13
K
Modeling of Transient Pressure Response for CO2 Flooding 195
10.3.3
Sensitivity Analysis
10.3.3.1 CO2 Bank
This section investigates the sensitiveness of four reservoir and fluid
­properties and one operating condition to the pressure responses during
a CO2 miscible flooding process. In a CO2 miscible flooding reservoir, the
radius of each region indicates the drainage area and has a direct effect on
transient pressure responses. Figure 10.7 shows the effect of CO2 bank size
on dimensionless pressure and pressure derivative responses. It is found
that when the radius of CO2 bank equals to 5 and 10 m, there are two radial
flow regimes which are characterized by a 0.5 slope-line at an early time
period and a horizontal straight line before the occurrence of boundary
effect, respectively. A larger CO2 bank radius leads to a longer first radial
flow regime at the early time period and a shorter pseudo-radial flow
regime at a later time period.
For the case of Ra = 1 m, the first radial flow at the early time period is
missing for two reasons. One reason is the big difference in the properties of
the CO2 bank and the miscible zone. The viscosity and total c­ ompressibility
of the CO2 bank are much lower and larger compared with those in the
miscible zone, respectively. The other reason is the small radius (Ra = 1 m)
of the CO2 bank, because of which the flow reaches to the radial flow
period very quickly, as shown on the green curves of the dimensionless
pressure and pressure derivative. In this case, although the first radial flow
regime does not exist at the early time, it has the longest pseudo- radial
flow regime compared with other cases (Ra = 5, 10, 20 m). For the case of
1000
pD, dpD/dln(tD)
100
10
Ra = 1 m
Ra = 5 m
Ra = 10 m
Ra = 20 m
1
Ra = 1 m
Ra = 5 m
Ra = 10 m
Ra = 20 m
0.1
1.0E + 02 1.0E + 03 1.0E + 04 1.0E + 05 1.0E + 06 1.0E + 07 1.0E + 08 1.0E + 09
tD
Figure 10.7 The effect of CO2 bank size on dimensionless pressure and pressure derivative.
196 Carbon Dioxide Capture and Acid Gas Injection
10000
pD, dpD/dln(tD)
1000
Re = 100 m
Re = 150 m
Re = 200 m
Re = 100 m
Re = 150 m
Re = 200 m
100
10
1
0.1
1.0E + 02 1.0E + 03 1.0E + 04 1.0E + 05 1.0E + 06 1.0E + 07 1.0E + 08 1.0E + 09
tD
Figure 10.8 The effect of reservoir outer boundary size on the dimensionless pressure and
pressure derivative curves.
Ra = 20 m, the first radial flow regime is the longest. But the closed outer
boundary effect occurs soon after the appearance of the pseudo-radial flow.
10.3.3.2 Reservoir Outer Boundary
Three outer boundary radii of Re = 100, 150 and 200 m are tested to
­analyze the effect of reservoir outer boundary sizes on the transient pressure responses. It can be seen from Figure 10.8 that the longer the outer
boundary radius, the longer the pseudo-radial flow regime and the later the
appearance of the boundary effect. It is obvious that the bigger the reservoir,
a longer time is needed for pressure to reach the outer boundary and thereby
leads to a longer pseudo-radial flow regime. It is also found that before the
appearance of the pseudo-radial flow regime, the dimensionless pressure
and pressure derivatives of the three cases are exactly the same. Therefore,
the size of the reservoir can be clearly reflected on the pressure responses.
10.4
Conclusions
This thesis developed a comprehensive transient pressure model for a CO2
flooding miscible process. Through this study, some conclusions are made
as following:
1. A convection–diffusion mass transfer model is developed to
evaluate the dynamic mixing process between oil and CO2 in
Modeling of Transient Pressure Response for CO2 Flooding 197
the miscible zone. It is then incorporated into the pressure
model through the viscosity profile in the miscible zone.
2. Four flow regimes can be identified in the dimensionless
pressure and pressure derivative profiles: the first radial
flow regime, transition flow regime, the pseudo-radial flow
regime, and boundary-dominated flow regime.
3. A comparison of two pressure responses with or without
considering CO2–oil mass transfer in the miscible zone
shows that mass transfer process has a significant effect on
the transition and the pseudo-radial flow regimes. Thus, it is
imperative to take the mass transfer process into account for
a reliable modeling CO2 flooding process.
4. The size of the CO2 bank strongly influences the length of
the first and the pseudo-radial flow regimes. A larger radius
of the CO2 bank leads to a longer first radial flow regime and
a shorter pseudo-radial flow regime. A trivial radius of CO2
bank can cause the missing of the first radial flow regime but
a much longer the pseudo-radial flow period. Similarly, the
size of the reservoir outer boundary can also be identified in
the pressure responses. A bigger radius of the reservoir outer
boundary results in a longer pseudo- radial flow period.
Acknowledgments
Funding support from the NSERC Discovery grant is highly appreciated.
Thanks also go to KAPPA engineering for offering software licenses for
academic use.
References
1. Annual Production Reports, Oil and Gas J. 5 April, 1982.
2. Baviere, M. Basic Concepts in Enhanced Oil Recovery Processes. Published for
SCI by Elsevier Applied Science, London and New York, 1991.
3. Holm, L. W. and Josendal, V. A. Mechanisms of Oil Displacement by Carbon
Dioxide. Journal of Petroleum Technology, 26(12), 1427–1438, December 1974.
SPE-4736-PA.
4. Holm, L.W. CO2 Flooding: Its Time Has Come. Journal of Petroleum Technology,
34(12), 2739–2745, December 1982. SPE-11592-PA.
5. Holm, W. L. Evolution of the Carbon Dioxide Flooding Processes. Journal of
Petroleum Technology, 39(11), 1337–1342, November 1987. SPE-17134-PA.
198 Carbon Dioxide Capture and Acid Gas Injection
6. Jia, X., Zeng, F., and Gu, Y. Semi-Analytical Solutions to a One-Dimensional
Advection, 2013.
7. Diffusion Equation with Variable Diffusion Coefficient and Variable Flow
Velocity. Appl. Mathe. & Comput., 221(2), 268−281
8. Lee, J. and Wattenbarger, R. A. Gas Reservoir Engineering. Society of
Petroleum Engineers. Richardson, Texas, 1996.
9. Martin, D. F. and Taber, J. J. Carbon Dioxide Flooding. Journal of Petroleum
Technology, 44(4), 396–400, 1992. SPE-23564-PA.
10. Moritis, G. CO2 and HC injection Lead EOR Production Increase. Oil & Gas
Journal, 49–82, 23 April, 1990.
11. Muller, K.E. Computing the confluent hypergeometric function, M(a, b, x),
Numer. Math. 90(1), 179–196, 2001.
12. Mungan, N. Carbon Dioxide Flooding as an Enhanced Oil Recovery Process.
Journal of Canadian Petroleum Technology, 31(09), 13–15, November 1992.
PETSOC-92-09-01.
13. Stalkup, F. I. Carbon Dioxide Miscible Flooding: Past, Present, and Outlook
for the Future. Journal of Petroleum Technology, 30(8), 1102–1112, August
1978. SPE-7042-PA.
14. Winslow, D. Industry Experience with CO2 for Enhanced Oil Recovery.
Workshop on California Opportunities for CCUS/EOR, 27 June, 2012.
11
Well Modeling Aspects of CO2
Sequestration
Liaqat Ali and Russell E. Bentley
WSP | PARSONS BRINCKERHOFF, 16200 Park Row, Suite 200, Houston, TX
77087, USA
Abstract
The study presents the results of well modeling for CO2 sequestration wells completed into a deep porous brine aquifer reservoir. In CO2 injection projects, it is
essential to design an interface between the reservoir and the pipeline. Inflow performance relationship (IPR) and vertical flow performance (VFP) analyses help
design this interface. We present the results of well modeling of a vertical well and
an extended reach directional (ERD) well in this study for a typical CO2 sequestration field. We also briefly discuss the effect of impurities in CO2 stream. The results
showed that the temperature change at lower wellhead pressures along the wellbore was greater than that at higher wellhead pressures. The results also showed
that injection rate requirements dictate the number of wells required.
11.1
Introduction
In this study we describe CO2 flow through a vertical wellbore into a
deep porous brine aquifer reservoir. Like any other oil and gas production and injection field, CO2 sequestration projects also require designing an interface between the pipeline and the reservoir. Therefore, the
first and last nodes are at the wellhead and at the well total depth (TD)
respectively. The IPR and VFP help design this interface. The IPR for a
well is the relationship between the flow rate of the well and the flowing pressure of the well. The VFP relates to estimating the pressurerate ­relationship in the wellbore as the fluids move through the tubing.
The CO2 flow through a vertical wellbore could be single phase or two
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (199–220)
2017 © Scrivener Publishing LLC
199
200 Carbon Dioxide Capture and Acid Gas Injection
phase depending upon the temperature and pressure condition in the
wellbore. The phase change in the wellbore directly affects the well injection pressure. It is also important to know whether for the given tubing size, wellhead temperature, and pressure the required rates can be
achieved. The IPR and VFP curves provide great insights into designing
an optimal injection program.
We have used PROSPER software to develop IPR and VFP curves and
investigate the efficacy of the project’s requirements. The main objectives
of the study were (1) to verify if the minimum wellhead pressure (WHP)
of 110 Bara and a wellhead temperature (WHT) of 35 degrees C would
be sufficient for a steady state injection rate of 3 million tons per annum
(Mtpa) [4.5 million S m3/d (MSm3/d)] and (2) to quantify the injection
capacity of vertical well as well as the ERD well under this scenario for
pure CO2.
11.2
Delivery Conditions
Due to variations in rates and temperatures and corresponding changes
in wellhead pressures, various cases have been performed in this study.
The pressure and temperature conditions, along with the pipeline delivery
conditions for the VFP analyses, are shown on the phase diagram for pure
CO2 (Figure 11.1).
Phase envelope
160
Analyses performed:
Pressure: 70–160 bara
Temperature: 11–35 °C
120
Pressure (bara)
Supercritical
region
Dense phase region
140
100
Delivery conditions
(110 bara & 35 °C)
80
Liquid region
60
Gas Region
40
0
Critical temperature: 30.94 °C
Critical pressure: 73.98 bara
Vapor region
20
0
5
10
15
20
25
30
Temperature (°C)
35
40
45
50
Figure 11.1 Phase envelope of pure CO2 with delivery conditions and the conditions on
which the sensitivities were performed.
Well Modeling Aspects of CO2 Sequestration 201
11.3
Reservoir and Completion Data
Reservoir and completion data are given in Table 11.1. The reservoir pressure and temperature have been calculated using a normal pressure gradient of 0.433 psi/ft at an average injection depth. Please note that the ERD
well has much greater completion length compared to the vertical well.
11.4 Inflow Performance Relationship (IPR) and
Injectivity Index
The IPR analysis was conducted in PROSPER software to establish the
required flowing bottom hole pressure to inject at a given rate for any
given reservoir pressure. The multi-rate Forchheimer IPR model with
pseudo-pressure was used to fit a set of rates and pressures obtained from
reservoir simulations for vertical as well as deviated wells (Figure 2). The
Forchheimer equation accounts for non-Darcy pressure losses in high flow
rates situations as given below.
(Pr2–Pwf2) = aQ2 + bQ(11.1)
where:
Pr = Reservoir pressure
Pwf = Flowing bottom hole pressure
Q = Flow rate
Table 11.1 Reservoir and completion data.
Well
Vertical well
ERD well
TVD, mss
1500
1330
Completion length, m
140
960
7(0.1778)
7(0.1778)
Average injection depth, m
1400
1300
Turbing size, in (m)
BHP, Bara
138
126
BHT °C
64
60
Reservoir permeability, md
650
550
Reservoir porosity
0.17
0.17
Reservoir thickness, m
200
200
Wellbore thickness, m
200
200
Wellbore radius, in (m)
9 5/8 (0.22425)
9 (5/8) 0.22425
202 Carbon Dioxide Capture and Acid Gas Injection
160
Injectivty index:
Vertical well: 0.5 (MSm3/d)/Bara
ERD well: 0.3 (MSm3/d)/Bara
Flowing BHP pressure (Bara)
155
150
145
140
135
Simulation data-vertical well
IPR-vertical well
Simulation data-ERD well
IPR-ERD well
130
125
120
0
2
4
6
Gas rate (MSm3/day)
8
10
Figure 11.2 Comparison of the reservoir simulation data and the calculated IPR using
Prosper for the vertical well and ERD well.
a = Darcy pressure loss coefficient
b = Non-Darcy pressure loss coefficient
Figure 11.2 also presents the corresponding injectivity index for each
well. The injectivity index is defined as the gas rate divided by the difference of flowing BHP and the reservoir pressure. The reservoir pressures
of 138 Bara and 128 Bara for the vertical well and the ERD well are used
respectively. The IPR relationships were obtained by matching the simulation data with the Forchheimer IPR model of PROSPER software. The IPR
curves calculated using PROSPER are shown by the grey and black dotted
curves for the vertical well and ERD well respectively. The simulation data is
also presented as open circles and squares to show the fitness of the curves.
11.5
Equation of State (EOS)
Two equations of state were available in PROSPER software: SoaveRedlich-Kwong (SRK) and Peng-Robinson (PR). Recently, Du et al. (2014)
compared different equation of states including SRK and PR. They found
that PR equation is the most precise EOS to compute density and predict
CO2 phase behavior. Furthermore, Boyle and Carroll (2002) performed a
study on predicting acid-gas (pure H2S, pure CO2 and mixture of the two)
densities for the pressure and temperature ranges specific to acid-gas injection applications. They found that the SRK equation was unsatisfactory
Well Modeling Aspects of CO2 Sequestration 203
for predicting densities over the entire range of pressure and temperature
conditions used. Mazzoccoli et al. (2013) also reached a similar conclusion that the SRK equation showed higher errors in calculating densities
for pure CO2 as well as a binary mixture of CO2 with Ar, O2, N2 and CH4,
respectively. Based on this review, the PR equation was used for this study.
Consistency checks were made to ensure that the PROSPER PR equation reproduces the experimental results in the pressure and temperature
conditions of the well modeling for VFP analyses. These checks included
comparing the PROSPER PR calculations of phase diagrams and d
­ ensity
for pure CO2 and CO2 mixtures. The checks also included comparing
viscosity calculated using Lohrenz-Bay-Clark with that calculated using
Freng-Wakeham correlation.
For pure CO2, the Span & Wagner (SW) EOS (1996) is based on extensive experimental data and is considered to be the most accurate EOS.
Comparison of phase diagrams of PR and SW equations showed that the
difference in the critical point is less than 1% (Figure 11.3).
Comparison of PR and SW equations showed that the error in densities is less at lower temperatures and higher at higher temperatures at a
pressure of 110 Bara for both PR and PR with volume shift (Figure 11.4).
However, PR yields better results than that of PR with volume shift for
operating conditions of this study. The error for PR ranges from 0.4 at 11
degrees C (dense phase region) to as high as 7.7% at 35 degrees C (supercritical region) at a pressure of 110 Bara.
Phase envelope
80
Pressure (Bara)
70
60
50
EOS
Temperature °C Pressure bara
Peng-Robinson
30.94
73.9777
Span & Wagner
31.1282
73.773
Difference, %
0.60
0.28
40
30
20
Span & Wagner
Peng-Robinson
0
5
10
15
20
25
30
Temperature (°C)
35
40
45
Figure 11.3 Comparison of phase diagrams of Peng-Robinson and Span & Wagner
equations.
50
204 Carbon Dioxide Capture and Acid Gas Injection
1000
Span & Wagner
PR
PR with volume shift
900
800
12.0
600
Absolute error (%)
Density (Kg/m3)
700
500
400
300
200
10
8.2
7.4
6.0
4.8
7.7
4.2
4.0
2.0
11.3
PR
PR with volume shift
8.0
0.0
100
0
10.0
2.7
0.4
11
20
30
Temperature (°C)
15
20
35
25
Temperature (°C)
30
35
40
Figure 11.4 Comparison of errors and densities calculated from PR, PR with volume shift
and Span & Wagner equations for WHP pressure of 110 Bara for pure CO2.
0.12
Fenghour
Lohrenz (PROSPER)
0.08
0.06
Absolute error (%)
Viscosity (mPa.s)
0.10
0.04
0.02
0.00
0
Errors in viscosity of pure CO2
20
18
16
14
12
10.8
10.110.2 10.5 10.6 10.8 10.5 10.4
9.3
10
8.0 8.7
8.0
8
6
4
2
0
0 4 8 12 16 19 23 27 31 35 Min Max
Temperature (ºC)
10
20
Temperature (°C)
30
40
Figure 11.5 Comparison of viscosities and errors associated for pure CO2 for WHP
pressure of 110 Bara.
Viscosity was calculated using the Lohrenz-Bray-Clark correlation which
was chosen among several correlations available in PROSPER software. A
comparison of viscosities calculated using this correlation and the FenghourWakeham correlation (1998) was made for pure CO2. The Fenghour-Wakeham
correlation was developed based on extensive experimental data with pure
CO2 and is considered to be more accurate for pure CO2. Lohrenz-Bray-Clark
correlation is based on reservoir fluids containing methane through heptanes
plus, H2S, N2 and CO2. The results are presented in Figure 11.5. The error for
Lorenz viscosities is in the range of 8 to 10.8% for viscosities.
Well Modeling Aspects of CO2 Sequestration 205
1400
225 K
245 K
265 K
285 K
300 K
320 K
350 K
400 K
450 K
Density (Kg/m3)
1200
1000
800
600
400
Operating conditions (table 1)
Temperature: 284–308 K
Pressure: 7–16 Mpa
200
0
0
10
20
30
40
Pressure (MPa)
50
60
70
Figure 11.6 Density of a mixture: CO2 (91%) and N2 (9%) showing experimental data
(Carroll and Roberts, 2013) and predictions from PR EOS of PROSPER. The continuous
lines are the calculated values.
Experimental data from Roberts and Carroll (2013) was used to compare the densities calculated by PROSPER using the same mixture: CO2
(91%) and N2 (9%) to further check robustness of PR of PROSPER software. Figure 11.6 presents the results that confirm the conclusions arrived
at by Roberts and Carroll (2013) in their paper that the greatest errors
occur at lower temperatures. Low temperatures are not typically of interest in carbon sequestration applications. Figure 11.6 shows that the match
between experimental and PROSPER PR equation density calculations is
very good in the range of the operating conditions being considered for
this study.
11.6
Vertical Flow Performance (VFP) Curves
Several sets of VFP curves were developed for the vertical and ERD wells.
These curves cover a wide range of flowing bottom hole (BHP) and WHP for
wellbore geometries, wellhead conditions and reservoir pressures. We present the results of IPR-VFP analyses for reservoir pressures of 138 Bara and
128 Bara for the vertical well and ERD well respectively. A WHP of 110 Bara
and a WHT of 35 degrees C were used to develop the VFP curves. The VFP
curves were also developed for WHT of 11 degrees C, 20 degrees C (only for
the vertical well), 30 degrees C and 35 degrees C for each well. These curves
are not shown here; however, a summary of their results are presented.
206 Carbon Dioxide Capture and Acid Gas Injection
Table 11.2 Critical points of the cases considered.
Critical
­temperature °C
Case
Critical
­temperature Cricondentherm Cricondenbar
Bara
°C
bara
Pure CO2
30.9400
73.9777
30.9400
73.9777
CO2 with 5%
contaminants
30.0089
77.7632
30.0347
72.3256
Phase envelope
80
75
Pressure (bara)
70
65
60
55
50
45
40
Pure CO2
CO2 with 5% contaminants
35
30
0
5
10
15
20
Temperature (°C)
25
30
35
Figure 11.7 Phase diagram of the compositions used in well engineering analysis for pure
of CO2 and CO2 with 5% contaminants.
In addition to pure CO2, CO2 mixtures (5% contaminants) were also
used to perform VFP analyses. The graphical results of VFP analyses are
not presented here; however, they are briefly discussed. The contaminants
mainly consisted of H2O, N2, H2, Ar, O2, CH4, CO, H2S, C2H6, SO2 and NOx.
For the worse case of NOx, only NO2 was used in the analysis. Table 11.2
presents the critical temperature and critical pressure for each case and
Figure 11.7 presents the phase diagram. Addition of impurities and their
molar concentration increase the two-phase region of phase diagram as
shown in Figure 11.7.
Figure 11.8 presents VFP curves for the vertical well for the delivery
conditions: WHP of 110 Bara, 35 degrees C and an initial reservoir pressure of 138 Bara. Figure 11.8 shows that this well can handle gas rates up to
6.5 MSm3/d (4.3 Mtpa) at the required operating conditions which is well
above the designed gas rate of 4.5 MSm3/d (3 Mtpa) for this study.
WHT = 35 °C
At current reservoir
pressure, this well can make
6.5 MSm3/d (4.3 Mtpa)
Reservoir pressure
150 bara
138 bara
Pc
rit
60
1,000
ba
80
ra
ra
ra
3,000
2,000
8b
ara
10
0b
ba
ba
3.9
120 bara
110 bara
14
0b
ar
a
5 Mtpa
70
:7
3 Mtpa
260
250
240
230
220
210
200
190
180
170
160
150
140
130
120
110
100
90
80
70
60
50
40
30
20
1 Mtpa
Pressure (bara)
Well Modeling Aspects of CO2 Sequestration 207
4,000
5,000
6,000
Gas rate (1000 Sm3/d)
7,000
11
0b 1
ar 20 ba
a
ra
ar
a
8,000
9,000
10,000
At current reservoir
pressure, this well can make
5.45 MSm3/d (3.6 Mtpa)
WHT = 35 °C
Reservoir pressure
150 Bara
140 Bara
128 Bara
110 Bara
ra
2,000
3,000
ra
11
0B
4,000
5,000
6,000
Gas rate (MS3/d)
16
0B
ar
a
ara
Ba
ar
8B
90
0B
3.9
Ba
1,000
:7
10
rit
a
3 Mtpa
Pc
Bar
ar
a
5 Mtpa
80
70
270
260
250
240
230
220
210
200
190
180
170
160
150
140
130
120
110
100
90
80
70
60
50
40
1 Mtpa
Pressure (Bara)
Figure 11.8 The VFP curves for the vertical well showing the gas rate of 6.5 MSm3/d
(4.3 Mtpa) at WHP of 110 Bara, WHT of 35 degrees C and reservoir pressure of 138 Bara.
7,000
12
0
a
14
Ba
0B
ra
8,000
ar
a
9,000
Figure 11.9 The VFP curves for ERD well showing the designed rate of 3 Mpta (4.5
MSm3/d) and full potential of 3.37 Mtpa (5.45 MSm3/d) at WHP of 110 Bara, WHT of
35 degrees C and reservoir pressure of 128 Bara.
Figure 11.9 presents the results of the VFP analysis for the ERD well. This
well can make 5.47 MSm3/d (3.62 Mtpa) which is also above the designed
gas rate of 4.5 MSm3/d (3 Mtp) at the required operating conditions.
The overall results of the VFP analysis at the WHT of 11 degrees C,
20 degrees C, 30 degrees C, and 35 degrees C temperatures are presented in
Figure 11.10. For the vertical well, gas rate ranges from 4.3 to 6.0 Mtpa (6.5
to 9.1 MSm3/d) whereas it ranges from 3.6 to 4.8 Mtpa (5.4 to 7.2 MSm3/d)
for the ERD well. The results also indicate that the higher the injection
temperature, the lower the gas rates and vice versa. However, the vertical
well has higher rates than that of the ERD well at each temperature.
208 Carbon Dioxide Capture and Acid Gas Injection
7
6.0
5
4
5.2
5.6
4.8
4.3
3.6
4.1
0
Vertical well
11 °C
30 °C
35 °C
20 °C
1
11 °C
2
30 °C
3
35 °C
Gas rate (Mtpa)
6
ERD well
Pure CO2
Figure 11.10 Comparison of gas rate for the vertical and ERD wells for WHT
temperature of 11 degrees C, 30 degrees C and 35 degrees C and for WHP of 110 Bara.
11.7
Impact of the Well Deviation on CO2 Injection
The impact of the well trajectory and well deviation on CO2 injection was
investigated by studying the change in properties such as pressure, temperature and density profiles along the wellbore at three different temperatures
(11 degrees C, 30 degrees C and 35 degrees C) for the vertical and ERD
wells for pure CO2. Change in properties along the wellbore is a function
of pressure loss, gas rate, temperature and phase behavior of the injected
fluid. Figure 11.11 is an illustration of the pressure losses in the wellbore
for the vertical and ERD wells at an injection temperature of 11 degrees C.
In this figure, the 1st curve from the left is the friction pressure, the 2nd
curve is the hydrostatic pressure, the 3rd curve is the net loss/gain and the
4th curve is the pressure along the wellbore. As can be seen, the pressure
curve is the mirror image of the net loss/gain curve. Inflections in the pressure curves are caused by the pressure losses as seen clearly just below the
tubing/packer depth in each well. Temperature and pressure conditions also
change at this location causing a pronounced effect on the pressure losses.
Figure 11.12 presents the change in properties for the vertical and ERD
wells at an injection temperature of 35 degrees C and WHP of 110 Bara.
This figure shows that most pronounced change in properties occurs at
the tubing/packer depth for each well and at the heel of the ERD well. At
the tubing/packer depth, the configuration of the wellbore changes from
7"(0.1778 m) tubing to 9 5/8" (0.22425 m) casing.
Figure 11.13 shows the change in pressure and temperature along the wellbore at three different injection temperatures with respect to phase envelope.
Well Modeling Aspects of CO2 Sequestration 209
0
0
200
1000
Tubing/packer depth: 1100 m
400
2000
Measured depth (m)
Measured depth (m)
600
800
1000
3000
4000
Heel: 4810 m
1200
5000
1400
Tubing/packer depth: 1330 m
Vertical well
1600
Net loss/gain
Res. pressure = 138 Bara
Hydrostatic pressure
WHP = 110 Bara
Friction pressure
Gas rate = 9.1 (MSm3/d)
Wellbore pressure
1800
–150 –100 –50
0
50
100 150 200
Pressure (Bara)
6000
ERD well
Res. pressure = 128 Bara
WHP = 110 Bara
Gas rate = 7.31 (MSm3/d)
7000
–100
Friction pressure
Net loss/gain
Hydrostatic pressures
Wellbore pressure
0
100
Pressure (Bara)
200
Figure 11.11 Pressure losses in the vertical well and ERD well at an injection temperature
of 11 degrees C.
Figure 11.13 shows that the wells with wellhead temperatures of 11 degrees
C and 35 degrees C experienced greater change in temperature (13 degrees
C for 11 degrees C curve and 10 degrees C for 35 degrees C curve) along the
wellbore compared to the same wells at 30 degrees C (only 5 degrees C).
Figure 11.14 shows the effect of well deviation on VFP curves at an
injection temperature of 35 degrees C and WHP of 110 Bara for pure CO2.
This figure shows that at an injection temperature of 35 degrees C, the ERD
well has reduced injection capacity of 3.6 Mtpa (5.4 MSm3/d) which is 24%
less than that of the vertical well that had an injection capacity of 4.7 Mtpa
(7.1 MSm3/d).
11.8 Implication of Bottom Hole Temperature (BHT)
on Reservoir
Figure 11.15 presents BHT as a function of gas rate for different injection
temperatures for the vertical and ERD wells. This figure shows that for the
vertical well, from 1 to 5 Mtpa (1.5 to 7.6 MSm3/d), the decrease in BHT
210 Carbon Dioxide Capture and Acid Gas Injection
0
0
Tubing/packer depth: 1330 m
1000
Tubing/packer depth: 1100 m
1000
3000
4000
Heel
5000
1000
2000
Tubing/packer depth: 1100 m
Measured depth (m)
2000
Tubing/packer depth: 1100 m
Measured depth (m)
Measured depth (m)
2000
0
Tubing/packer depth: 1330 m
3000
4000
Heel
5000
6000
7000
100
120
140
Pressure (Bara)
160
3000
4000
Heel
5000
6000
Vertical well, gas rate = 6.5 mSm3/d
ERD well, gas rate = 5.4 mSm3/d
Tubing/packer depth: 1330 m
6000
Vertical well,gas rate = 6.5 mSm3/d
ERD well, gas rate = 5.4 mSm3/d
7000
30
35
40
45
Temperature (°C)
50
Vertical well, gas rate = 6.5 mSm3/d
ERD well, gas rate = 5.4 mSm3/d
7000
680
690
700 710 720
Density (Kg/m3)
730
Figure 11.12 Change in properties along the wellbore in vertical and ERD wells at an
injection temperature of 35 degrees C and WHP of 110 Bara.
180
Vertical well
ERD well
Phase envelope
160
WHT = 30 °C WHT = 35 °C
Pressure (Bara)
140
WHT = 11 °C
120
Supercritical
region
100
80
60
Critical point
40
20
0
0
5
10
15
20
25
30
Temperature (°C)
35
40
45
50
Figure 11.13 Change in pressure and temperature along the wellbore with reference to
phase envelope.
ranges from 5 degrees C to 10 degrees C. For the ERD well, the BHT reduction ranges from 15 degrees C to 18 degrees C depending upon the surface
temperature.
Figure 11.15 also shows that from 1 to 3 Mtpa (1.5 to 4.5 MSm3/d), the
decrease in temperature is only 1 degrees C for vertical well and 6 degrees C
Well Modeling Aspects of CO2 Sequestration 211
250
IPR-vertical well
VFP-vertical well
IPR-ERD well
VFP-ERD well
150
0
1 Mtpa
50
0
2000
5 Mtpa
100
3 Mtpa
Pressure (Bara)
200
4000
6000
8000
Gas rate (1000 Sm3/d)
WHT = 35 °C
10000
12000
Figure 11.14 Effect of well deviation on VFP curves for an injection temperature of
35 degrees C and WHP of 110 Bara.
for the ERD well for WHT of 35 degrees C. For both wells, the temperature change in the fluid’s BHT is about the same (5 to 6 degrees C) for a
WHT of 11 degrees C. There is no apparent change in the fluid’s BHT for
a WHT of 20 degrees C for the vertical well. For a WHT of 30 degrees C at
the vertical well, the fluid’s BHT temperature fluctuates at injection rates
between 1 Mtpa (1.5 MSm3/d) and 3 Mtpa (4.5 MSm3/d), however, 1 Mtpa
(1.5 MSm3/d) and 3 Mtpa (4.5 MSm3/d), the fluid’s BHT is the about the
same. For the ERD well, the change in the fluid’s BHT is about the same (5
to 6 degrees C) at a WHT of 30 degrees C.
Since the maximum designed injection rate per well was 3 Mtpa
(4.5 MSm3/d), the maximum difference between the injected fluid’s BHT
and the initial reservoir temperature is 18 degrees C (Initial reservoir temperature (Tr) = 67 degrees C) for the vertical well. For the ERD well, it is
15 degrees C (Initial Tr = 63 degrees C) for a WHT of 35 degrees C.
Impact of cold water in waterflood reservoirs is well documented
(Gadde and Sharma, 2001; Mitchell et al., 2013). Due to the cold fluid
­temperature at the bottom of the hole, CO2 has been suggested as a stimulation fluid in hydraulic fracturing of the tight shale reservoirs (Amro et al.,
2011; Enayatpour and Patzek, 2013). Goodarzi et al. (2010) presented their
work on thermal aspects of geomechanics and induced fracturing in CO2
sequestration applications for a vertical well. They concluded by comparing the results of the thermal model with those of the iso-thermal case
212 Carbon Dioxide Capture and Acid Gas Injection
70
Vertical well
WHP = 110 Basra
Reservoir pressure = 138 Bara
60
BHT (C)
50
40
35
30
°C
30 °C
20 °C
20
0
0
2000
5 Mtpa
1 Mtpa
10
3 Mtpa
11 °C
4000
6000
8000
Gas rate (1000 Sm3/d)
10000
12000
70
ERD well
WHP = 110 Basra
Reservoir pressure = 128 Bara
60
BHT (C)
50
40
35 °C
30 °C
30
20
0
3 Mtpa
0
1 Mtpa
10
2000
4000
5 Mtpa
11 °C
6000
8000
Gas rate (1000 Sm3/d)
10000
12000
Figure 11.15 BHT as function of gas rates for different temperatures for vertical and ERD
wells.
that the total minimum stress at the wellbore decreases with time and falls
below the injection pressure quite early during the injection. Also, thermal effects could increase the speed of fracture propagation in the storage
layer depending on the injection rate. The study cited was conducted for
Well Modeling Aspects of CO2 Sequestration 213
lower injection rates than this study; however, the temperature differences
between the BHT and reservoir temperature were very similar. Based on
the injection rates used in their study they concluded that injection of CO2
at a temperature lower than reservoir temperature reduces the fracture
pressure, which in turn reduces the injectivity. The results also pointed out
that it may not be practical to avoid induced fracturing in CO2 projects.
Therefore, they recommended optimizing the injection rate and temperature by maximizing the injection capacity, while maintaining the safety of
the storage reservoir by limiting the fracture length.
Based on the insights from the above-mentioned study, we recommend
that a geomechanical study be performed to study the impact of injection
rate and temperature on injectivity, fracturing and fracture length.
11.9
Impact of CO2 Phase Change
Transient flow occurs during events such as start-up, shutdown, planned
or uncontrolled depressurization of pipeline. Aursand et al. (2013) presented a review of current research challenges for modeling of transient
flow of CO2 mixtures in pipelines. Liljemark et al. (2011) concluded that
quick shutdown and load change caused two-phase flow in the vertical section of the pipeline (injection well).
Single-phase flow exists within the borehole above the critical point
(73.98 Bara and 30.94 degrees C) for pure CO2. At low injection rates (e.g.,
during shutdown/start-up of injection operations), the pressure could be
significantly lower than the critical pressure and two-phase flow will most
likely occur near the surface down to a depth at which the injectate goes
back into a dense state again with increasing hydrostatic pressure. There is
no particular concern for any of the subsurface components while operating in this state and it is not unusual to operate wells with two-phase flowing conditions. We have recent experience with an acid gas injection well
(15% H2S / 85% CO2) which, when going from a transient to steady state,
needed a somewhat higher wellhead pressure (a pressure “bump”) to overcome the two phase phenomena and return the injectate to a single phase
again. The lower the injection rate, the longer the transient period and the
lower the pressure bump. This would imply that should such phenomena
appear, a certain time period would need to be accounted for when going
from start-up to steady state. (A longer “ramp up” time would be needed
for a smaller pressure bump.)
The length of the transient period and the additional pressure necessary to resume steady state flow is apparently velocity dependent. At faster
214 Carbon Dioxide Capture and Acid Gas Injection
velocities the phenomena may be less pronounced or even non-existent.
A full transient study would need to be undertaken to properly gauge
whether or not pressure anomalies appear, how large they might be, and
how long they might last.
11.10 Injection Rates, Facility Design Constraints
and Number of Wells Required
Figures 11.8 and 11.9 present the VFP sensitivities for the vertical and ERD
wells, and show that designed injection rates of 3 Mtpa (4.5 MSm3/d) can
be achieved. To access the full capacity rate of 5Mtpa (7.6 MSm3/d), either
two wells are required or the injection pressure has to be increased from
the designed 110 Bara to 120 Bara for the vertical well or to 130 Bara for
the ERD well.
For a steady state injection of 1 Mtpa (1.5 MSm3/d), one vertical or
ERD well is sufficient. However, a second well will allow continued injection in the event the first well needs to be shut down for intervention
operations.
For steady state injection of 3 Mtpa (4.5 MSm3/d), two ERD wells are
recommended. For contingencies such as well interventions, one additional well will be required.
For steady state injection of 5 Mtpa (7.6 MSm3/d), two vertical wells or
two ERD wells are recommended. For contingencies such as well interventions, a third well will be required.
11.11
Wellhead Temperature Effect on VFP Curves
The effect of WHT on gas rates has been discussed briefly in the previous
sections. In this section, the effect of WHT on gas rates for the vertical and
ERD wells for temperatures of 11 degrees C, 30 degrees C and 35 degrees C
for pure CO2 will be discussed.
Figure 11.16 presents IPR-VFP curves for the vertical and ERD wells.
Figure 11.16 clearly shows that the gas rate is significantly affected as the
WHT is increased. For vertical well, at 11 degrees C, the well’s gas rate
of 5.98 Mtpa (9.1 MSm3/d) is 16% more than 5 Mtpa (7.6 MSm3/d). At
30 degrees C, the gas rate is 5.18 Mtpa (7.8 MSm3/d) which is a 13%
decrease. With increase of 5 degrees C in WHT (at 35 degree C), the gas
rate further decreases to 4.2 Mtpa (6.5 MSm3/d) which is a 9% decrease
and overall decrease of 21% from its full injection capacity at 11 degrees C.
Well Modeling Aspects of CO2 Sequestration 215
250
Vertical well
(WHP = 110 Bara)
138 Bara
150
11
C
100
C
35
Pressure (Bara)
200
30
0
2000
5 Mtpa
1 Mtpa
0
3 Mtpa
C
50
4000
6000
8000
Gas rate (1000 Sm3/d)
10000
12000
250
ERD well
(WHP = 110 Bara)
150
126 Bara
35
100
C
0
2000
5 Mtpa
3 Mtpa
1 Mtpa
0
C
C
50
11
30
Pressure (Bara)
200
4000
6000
8000
Gas rate (1000 Sm3/d)
10000
12000
Figure 11.16 Effect of injection temperature on VFP curves for the Vertical and ERD wells.
At an injection temperature of 11 degrees C, the ERD well has a full gas
injection capacity of 4.83 Mtpa ((7.31 MSm3/d) which is 38% more than
3 Mtpa (4.5 MSm3/d). At a WHT of 30 degrees C, this gas rate reduces to
4.14 Mtpa (6.26 MSm3/d) which is a 14% decrease in gas injection capacity.
At an increase of 5 degrees C (at 35 degrees C), the gas rate further reduces
to 3.62 Mtpa (5.48 MSm3/d) which is a further decrease of 13% in gas rate
and an overall decrease of 25% from the full gas injection capacity. A gas
rate of 3.6 Mtpa (5.4 MSm3/d) is still 17% higher than 3 Mtpa (4.5 MSm3/d).
216 Carbon Dioxide Capture and Acid Gas Injection
11.12
Effect of Impurities in CO2 on VFP Curves
Figure 11.17 shows the effect of impurities on wellhead pressures and
injection rates. Only cases for pure CO2 and CO2 with 5% contaminants
are plotted in this figure for vertical as well as ERD wells to show the maximum effect. This figure shows that at a given injection rate the WHP for
CO2 with 5% contaminants will be higher than that of pure CO2 for both
the wells. The vertical well shows higher wellhead pressures for CO2 with
5% contaminants than the ERD well.
The results of VFP curves for CO2 with 5% contaminants (not presented
in the paper) showed that the addition of impurities reduces the gas rates
from 5 to 10% at all temperatures. The higher the temperature, the higher
is the reduction in gas rates.
Vertical well
(Reservoir pressure = 138 Bara, WHT = 30 C)
180
Pure CO2
CO2 with 5% contaminants
WHP, Bara
160
140
120
100
80
60
0.0
2.0
3.0
4.0
Injection rate (MSm3/d)
5.0
6.0
7.0
ERD well
(Reservoir pressure = 128 Bara, WHT = 30 C)
200
Pure CO2
CO2 with 5% contaminants
180
WHP, Bara
1.0
160
140
120
100
80
60
0.0
1.0
2.0
3.0
4.0
Injection rate (MSm3/d)
5.0
6.0
Figure 11.17 Effect of impurities on wellhead pressures and injection rates.
7.0
Well Modeling Aspects of CO2 Sequestration 217
11.13
Concluding Remarks
The objective of this study was to design an interface between the reservoir
and the pipeline. Therefore, the first and last nodes were wellhead and well
bottom hole respectively. IPR and VFP help design this interface.
The Forchheimer IPR Model in PROSPER software was chosen and fitted with the simulation data to obtain the representative IPR. The IPR analyses were performed for reservoir pressures of 138 Bara for the vertical well
and 128 Bara for the ERD well. In addition, IPR curves were also generated
for pressures other than the reservoir pressure such as 110 Bara, 120 Bara
and 150 Bara for the vertical well; 110 Bara, 120 Bara, and 150 Bara for
the ERD wells. VFP analyses were performed for a wide range of wellhead
pressures.
Consistency checks were made and a PR EOS equation was selected
for the work performed. These checks indicated that for pure CO2, the
PR equation can have errors in density ranging from 0.4% at 11 degrees
C (dense phase region) to as high as 7.7% at 35 degrees C (supercritical
region) at a pressure of 110 Bara. Experimental data (Roberts and Carroll,
2013) was used to compare the computed densities using the same mixture: CO2 (91%) and N2 (9%) as the experimental data to further check
robustness of PR of PROSPER software. Results confirm the conclusions
arrived by Roberts and Carroll (2013) in their paper that the greatest errors
occur at lower temperatures. Low temperatures are not typically of interest
in carbon sequestration applications.
VFP analyses indicated that all pure CO2 cases showed higher gas rates
for the corresponding wells compared to CO2 with 5% contaminants.
Furthermore, lower temperatures showed higher gas injection capacities
for all the wells. For pure CO2 at a WHT of 11 degrees C, maximum gas
rates of 6 Mtpa (9.1 MSm3/d) and 4.8 Mtpa (7.2 MSm3/d) were observed
for the vertical well and ERD well respectively. Overall, the ERD well had
lower gas injection capacity than the vertical well at all temperatures.
Moreover, at a given injection rate the WHP for CO2 with 5% contaminants
will be higher than that of pure CO2 for both the wells. The vertical well
shows higher wellhead pressures for CO2 with 5% contaminants than the
ERD well. Addition of impurities and their molar concentration increase
the two-phase region of phase diagram. Also, the addition of impurities
reduces the gas rates from 5 to 10% at all temperatures. The higher the
temperature, the higher is the reduction in gas rates.
Pressure losses are more pronounced at the tubing/packer depth for all
the wells and at heel for the ERD well. The pressure along the wellbore
mirrors the net pressure losses experienced in the wellbore. The change in
218 Carbon Dioxide Capture and Acid Gas Injection
density correspondingly is more pronounced at the tubing/packer depth of
the ERD well.
Conversion Factors
tpa × 1.5 E+00
bar × 14.503773 E+00
m × 3.28 E+00
m × 39.37 E+00
m3 × 35.31 E+00
kg/m3 × 6.242796 E-02
Mpa × 145.038
= m3/d
= psi
= ft
= in
= ft3
= lb/ft3
= psi
References
1. Amro, M., Haefner, F. and Mueller, M., Synergetic CO2 Storage and Gas/Oil
Production from Tight Reservoirs. The 6th Jordanian International Mining
Conference, Amman, Jordan, Nov. 1–3, 2011.
2. Aursand, P., Hammer, M., Munkejord, S.T. and Wilhelmsen, O., Pipeline
Transport of CO2 Mixtures: Models for Transient Simulation. International
Journal of Greenhouse Gas Control 15, p. 174–185, 2013.
3. Boyle, T.B. and Carroll, J.J., Study Determines Best Methods for Calculating
Acid-Gas Density. Drilling & Production, Oil & Gas Journal, Jan. 14, 2002.
4. Du, B., Cheng, L. and Ding, J., Simulation Studies of Temperature and
Pressure Distribution in Carbon Dioxide Injection Well. EJGE Vol. 19, Bund.
F. p. 1467–1476, 2014.
5. Enayatpour, S. and Patzek, T., Thermal Shock in Reservoir Rock Enhances
the Hydraulic Fracturing of Gas Shales. SPE Unconventional Resources
Technology Conference. Denver, Colorado, August 12–14, 2013.
6. Fenghour, A. and Wakeham, W.A., The Viscosity of Carbon Dioxide. J. Phys.
Chem. Ref. Data, Vol. 27, No. 1, 1998.
7. Gadde, P.B. and Sharma, M.M., Growing Injection Well Fractures and Their
Impact on Waterflood Performance. SPE Annual Technical Conference and
Exhibition. New Orleans, Louisiana, Sept. 30-Oct. 3, 2001.
8. Goodarzi, S., Settari, A., Zoback, M. and Keith, D.W., Thermal Aspects of
Geomechanics and Induced Fracturing in CO2 Injection with Application to
CO2 Sequestration in Ohio River Valley. SPE 139706. New Orleans, Louisiana,
Nov. 10–12, 2010.
9. Liljemark, S., Arvidsson, K., McCann, M.T.P., Tummescheit, H. and Velut, S.,
Dynamic Simulation of a Carbon Dioxide Transfer Pipeline for Analysis of
Normal Operation and Failure Modes. Energy Procedia 4, p. 3040–3047, 2011.
Well Modeling Aspects of CO2 Sequestration 219
10. Mazzoccoli, M., Guido, G.D., Bosio, B., Arato, E. and Pellegrini, L.A., CO2mixture Properties for Pipeline Transportation in the CCS Process. Chem.
Eng. Trans., Vol. 32. p. 1861–1866, 2013. DOI: 10.3303/CET1332311
11. Mitchell, P., Smith, K. and Podgorney, R.K., Cold Water Injection Effects in
Fractured Reservoirs. The Geological Society of America, Denver, Colorado,
Oct. 27–30, 2013.
12. Roberts, E.L. and Carroll, J.J., Densities of Carbon Dioxide-Rich Mixtures:
Part 2: Comparison with Thermodynamic Models. 4th International Acid gas
Injection Symposium. Calgary, Canada, September 24–27, 2013.
13. Span, A. and Wagner, W., A New Equation of State for Carbon Dioxide
Covering the Fluid Region from the Triple-Point Temperature to 1100 K at
Pressure up to 800 Mpa. J. Phys. Chem. Ref. Data, Vol. 25, No. 6, 1996.
12
Effects of Acid Gas Reinjection on
Enhanced Natural Gas Recovery and
Carbon Dioxide Geological Storage:
Investigation of the Right Bank of the Amu
Darya River
Qi Li, Xiaying Li, Zhiyong Niu, Dongqin Kuang, Jianli Ma, Xuehao Liu,
Yankun Sun and Xiaochun Li
State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute
of Rock and Soil Mechanics, Chinese Academy of Sciences, Wuhan, China
Abstract
Due to the environmental pollution of H2S, greenhouse effect of CO2 and economic problems of sulfur recovery, the reinjection of acid gas (H2S + CO2) into
subsurface formations, which may dissolve some of the rock and a great deal of
the cementing material, is an effective option to treat acid gas. The Amu Darya
right bank, Turkmenistan, is one of the most important sources of China’s natural gas supply in Central Asia. However, most of the gas reservoirs contain CO2
and H2S, which increases both the difficulties of development and the financial
burdens of the developers. The paper simulates the reinjection of the produced
acid gas into the reservoirs using the ECLIPSE simulator and analyzes the effect
of the reservoir pressure, gas sequestration, production and recovery ratio on the
reservoir formations. The numerical results show not only that the reinjected acid
gas is sequestered by gas-gas displacement but also that there is residual storage
space. When the acid gas injection rate is set to 8 × 104 m3/day and the maximum daily production rate is set to 1 × 105 m3/day, more acid gas may be stored,
thereby enhancing the natural gas recovery. The injected acid gas breaks through
the production well in the seventh year, and the reservoir pressure maintains the
stability of the reservoir. The production time is overly shortened and the reservoir energy rapidly declines when the daily production is oversized, which is
adverse to long-term stability during gas reservoir production. Therefore, in this
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (221–244)
2017 © Scrivener Publishing LLC
221
222 Carbon Dioxide Capture and Acid Gas Injection
study, the reinjection of all acid gas produced in the research area is the reservoir
exploitation countermeasure, considering safety, the environment, and economic
efficiency. This can efficiently enhance the gas recovery and lead to acid gas storage
to solve the problems of acid gas reservoir exploration.
12.1
Introduction
Compared with conventional gas reservoirs, the difficulties of exploiting
acid gas reservoirs include economic, environmental, safety and market
issues [1, 2]. These issues are embodied in the capacity of the evaluation
techniques, fluid phase safety testing, and corrosion studies, among others
[3, 4]. The acid gas reservoir development strategy is related to the protection and storage of sulfur, acid gas reinjection, electricity generation by acid
gas and safety issues [1, 5]. Taking the Puguang gas reservoir as an example,
the difficulties of exploiting acid gas reservoirs include that there is no standard for reservoir development and that acid corrosion, sulfur deposit, and
safe mining occur [6, 7]. For acid gas reservoirs, during the processes of
gas development and exploitation, solid sulfur precipitation may result in
the plugging of pore throats when the reservoir temperature and pressure
decline. Injected acid gas may compensate for the energy loss, relieve the
temperature and pressure change, and dissolve part of the rocks [8, 9].
Several scholars have performed research related to Acid Gas
Reinjection [10]. In 2007, at the Z3Z oil reservoir in the main rivers in
Zama, Trivedi et al. (2007) studied the optimal injection strategies and
importance of developing operational parameters for acid gas sequestration during tertiary oil recovery using the ECLIPSE simulator [11]. In
2008, Pooladi-Darvish et al. studied CO2 injection for the enhanced gas
recovery and geological storage of CO2 in the Long Coulee Glauconite F
Pool, Alberta, Canada using the GEM simulator [12]. In 2012, Khan et al.
studied the effects of CO2 and acid gas injection on EGR and storage using
the TEMPEST simulator [13]. Lubas et al. (2012) investigated the effects of
acid gas reinjection on the CO2 concentration in natural gas produced in
the Borezcin reservoir [14]. Hou et al. (2012) thermo-hydromechanically
modeled the CO2 injection for EGR using the TOUGH-FLAC simulator
[15]. In 2014, Li and Elsworth studied the geochemistry associated with
CO2-enhanced shale gas recovery [16]. In 2015, Wang et al. studied the
effect of impurities on the CO2 storage capacity in geological formations
[17]. This paper investigated the impact of the acid gas storage capacity and
EGR on the parameters governing acid gas reinjection using the ECLIPSE
simulator.
Investigation of the Right Bank of the Amu Darya River 223
The Amu Darya right bank gas project in Turkmenistan is a major
overseas project of the China National Petroleum Company. Most of the
reservoirs are acid gas reservoirs, which causes many difficulties for development and exploitation [18]. Thus, based on the combination of security development strategies for acid gas reservoirs and the rational use of
resources, this paper suggests that acid gas reinjection can be considered a
development strategy for acid gas reservoirs.
12.2 The Amu Darya Right Bank Gas Reservoirs in
Turkmenistan
Most of the gas reservoirs are carbonate gas reservoirs containing H2S and
CO2. Some gas reservoirs in the study area, such as the BP gas reservoir, are
in the reservoir geological study stage [19]. Other gas reservoirs are in the
drilling stage, which uses the directional drilling technique [20]. Some gas
reservoirs are exploited. There are difficulties associated with old wells that
need to be repaired and the recovering capacity. For example, 33 old wells
were repaired in the Sa gas reservoir [21]. The highlights of this research
are the variation of the gas amount, the stability of the reservoir evaluation
and changes in production that occur when reinjecting acid gas from production wells into formations at a certain proportion of H2S and CO2 for
environmental protection and economic maximization.
It is well known from previous studies that acid gas components contain
methane, ethane, propane, n/iso-butane, n/iso-pentane, C5 above, H2S,
CO2, N2 and He [19–23]. According to the data from the Sam53-1 well,
the burial depth in the Callovian-Oxfordian formation is 3500~3600 m,
corresponding to a rock-forming temperature of 90~100 °C and a main
temperature range of 90~110 °C [24]. The pressure factor in western
Callovian-Oxfordian formation in the study was 0.85~1.08. The pressure
factor in eastern and southeastern Callovian-Oxfordian formation was
1.65~1.90 [25].
12.3
Model Development
At present, it is of economic and environmental importance to address the
acid gas from production wells by reinjecting acid gas into the reservoirs to
enhance the gas recovery. This paper uses numerical simulation using the
compositional simulator ECLIPSE [26].
224 Carbon Dioxide Capture and Acid Gas Injection
12.3.1
State equation
The model includes only two phases: the gas phase and the water phase.
The equation of state (EOS) used is the modified Peng-Robinson (PR) EOS.
12.3.1.1 Introduction of Traditional PR State Equation
The A coefficient in the PR state equation is: A
jk
X j X k A jk
In the formula, x is the component mole fraction, A jk
1 k jk
A j Ak
and kjk are binary interaction coefficients. Ajk is defined as: A j
1/2
j Prj
Trj2
,
.
The reduced critical temperatures and pressures are Tr = T/Tc and Pr = P/Pc,
where Tc is the critical temperature and Pc is the critical pressure.
Arj P
The simplified coefficient A is defined by T and P: A j
2
T2
ojTcj j
1/2
1/2
r
In the formula, A j
; j is defined by j
1 f w j 1 Trj1/2 ;
Pcj
is a constant; wj is the acentric factor; f(wj) is the polynomial of the
oj
acentric factor; and Trj is the simplified temperature.
There are three coefficients in the compositional simulator (ECLIPSE):
A1jk
A2jk
A3jk
2
ojTck
1 k jk
1 k jk
1 k jk
1/2
pcj
2
ojTck
2
okTck
1/2
p j pk pck
1/2
2
okTck
pck
pcj
2
ojTck
1/2
pcj
1/2
p j qk
2
okTck
pck
pk q j 1/2
q j qk p
.
T2
In the formula, P and T are the pressure and temperature in each grid.
In every grid, Ajk is simplified as Ajk
A1jk
A2jkT 1/2
A3jkT
12.3.1.2 Modifications for the Vapor-Aqueous System
Two parts of the traditional PR EOS were modified according to Soreide
and Whitson’s proposal.
Investigation of the Right Bank of the Amu Darya River 225
For the water component, α1/2 was modified and replaced by an
expre­
ssion with a more complex temperature dependence: 1/2 1
aq1 1 aq2Tr aq3 Tr 3 1 .
In the formula, aq1 = 0.4530, aq2 = 1–0.0103cs1.1, aq3 = 0.0034, cs is the
binary concentration expressed as a molality.
The remainders take the usual form. A general form for all components
is 1/2 = p + qTr1/2 + rTr + sTr–3
Different binary interaction coefficients are used for the gas phase and
the aqueous phases. This may cause problems close to the critical point,
but the critical temperature of a water-dominated system is in the region of
Tcw = 647 K, which should be well above the area of interest.
Soreide and Whitson proposed adding a temperature dependence to the
aqueous phase binary interaction coefficients:
k ajw
bq1 bq2Trj bq3Trj2 In the formula,
bq1 – A0(1+S0cs), bq2 = A1(1 + S1cs), bq3 = A2(1 + S2cs)
A0 = 1.112 – 1.7369wj–0.1, A1 = 1.1001 + 0.836wj, A2 = –0.15742 – 1.0988wj
and S0 = 0.017407, S1 = 0.033516, S2 = 0.011478.
Special values are used for the binary interaction coefficients of the
aqueous phase and N2 and those of CO2 and H2S:
kNa 2w
a
kCO
2w
1.70235(1 0.025587cS3/4 ) 0.44338(1 0.08126cs3/4 )TrN 2 0.31092(1 0.15587cS0.7505 ) 0.23580(1 0.17837cs0.979 )TrCO2
21.2566 exp( 6.7222TrCO2
kHa 2Sw
cs )
0.20441 0.23426TrH2S These fit into the general quadratic form. The modifications of α1/2 and
the binary interaction coefficients introduce a coefficient of the PR state
equation. The model uses the modified PR state equation.
12.3.2
Salinity
The salinity in the model is defined by the molality: cs = 1000·ns/mw, where
ns is the number of moles of salt and mw is the mass of the water.
226 Carbon Dioxide Capture and Acid Gas Injection
12.3.3
Diffusion
12.3.3.1 Diffusion Coefficients
There are two diffusion models that can be used in the simulator ECLIPSE
300. The first diffusion model is driven by the concentration gradient:
x
Ji
cDi i . The second diffusion model is driven by the gradient of the
d
chemical potential:
Ji
cDia xi
1
RT d
i
MiG h h0
Mi DiT In T
,
where Ji is the molar flux of the component i per unit area; C is the
total molar concentration given by c = 1/vm; vm is the molar volume of
the mixture; Di is the diffusion coefficient of the component i; Dia is the
activity-­corrected diffusion coefficient of the component i; DiT is the thermal diffusion coefficient of the component; xi is the mole fraction of the
is the gradient in the direction of flow; Mi is the molecucomponent i;
d
lar weight of the component i; G is the acceleration due to gravity; h is the
height; h0 is the reference height; T is the temperature; R is the gas constant;
and μi is the chemical potential of the component i.
12.3.3.2 The Cross-Phase Diffusion Coefficients
There are two models in cross-phase diffusion in ECLIPSE. One model
is the activity-driven diffusion. The diffusion between the cells is driven
by the chemical potential, resulting in the diffusion over phase boundaries.
The diffusion coefficients for the cross-phase diffusion (oil phase and gas
phase) are set by combining the diffusion coefficients given by DIFFAOIL
Doia Dg ia .
and DIFFAGAS, Dog ia
The other model is molecular-driven diffusion. The fluids are mixed
on the border of each cell, and the resulting composition is used to set
up the potential for molecular diffusion of each component. The water-gas
and gas-water diffusion coefficients of the GASWAT model are set by the
keywords DIFFCWG and DIFFCGW. The molecular diffusion coefficient
constructs a border composition by mixing the composition of neighboring cells. As an example, we consider a cell I containing oil and a cell J containing gas. An artificial border composition is constructed by m̃i = (bo·xi)I
+ (bg·yi)J, where bo is the oil molar density, xi is the oil mole fraction
of component i, bg is the gas molar density, and yi is the gas mole fraction of the component i. The border pressure and temperature are set to
Investigation of the Right Bank of the Amu Darya River 227
P
PI
PJ
and T
TI TJ
. The cross-phase diffusive flow between the
2
2
oil in cell I and the gas in cell J for component i is given by Fcross = FXog
+ FXgo + FYog + FYgo.
FXog
Diff Dog i min SoI , Sg J
bo max xi
xi ,0 FXgo
Diff Dgoi min SoI , Sg J
bo max xi
xi ,0 FYog
Diff Dog i min SoI , Sg J
bg max yi
yi ,0 FYgo
Diff Dgoi min SoI , Sg J
bg max yi
yi ,0 In the formula, SoI and SgI are the oil and gas saturation in cell I and J,
Diff is the diffusivity, DogI and DgoI and are the cross-phase oil-gas and
gas-oil diffusion coefficients.
12.4
Simulation Model
12.4.1
Model Parameters
Based on previous studies [27] ½C, this paper constructs the stratum data
of the Sam53-1 well to determine the horizons and thicknesses of the layers (Figure 12.1) [28, 29]. Combined with the simulator, the components
of the model include C1, C2, CO2, H2S and H2O. The initial porosity, initial permeability and relative permeability are shown in Table 12.1 and
Figure 12.2 [30]. The initial temperature is 101 °C [31]. The original rock
compressibility is 1.78 × 10–5 [32].
12.4.2
Grid-Sensitive Research of the Model
The size of the model is 2000 m × 2000 m. Based on experience, we choose
the gridding schemes 5×5, 10×10 and 20×20 for grid-sensitive research to
determine the most suitable gridding scheme. The paper analyzes the reservoir pressure and saturation of the gridding schemes by keeping the
other conditions the same (Figure 12.3). At the early stage, the reservoir
pressure of the 5×5 model is slightly low, and those of the 10×10 and 20×20
models are basically the same. In the middle and later stages, the reservoir pressure of the 5×5 model is moderate, and those of the 10×10 and
20×20 models are the highest and lowest, respectively. The saturation of
228 Carbon Dioxide Capture and Acid Gas Injection
Stratum system
System
Series
Stage
Formation
Upper jurassic
Kimmeridgian
HA
Depth
(m)
Simplified lithologic
section
2300
2310
2320
2330
2340
XVac
2350
2360
2370
2380
XVp
2390
2400
2410
2420
2430
2440
2450
Jurrasic
Medium-upper
jurassic
series
Callovianoxfordian
XVm
2460
2470
2480
2490
2500
2510
2520
XVhp
2530
2540
2550
XVal
2560
2570
Z
2580
2590
2600
2610
XVa2
2620
2630
2640
2650
2660
XV1
Medium-lower jurassic series
2670
2680
2690
2700
Figure 12.1 Synthetic histogram of the Callovian-Oxfordian of the Sam53-1 well in this
study [28, 29].
Investigation of the Right Bank of the Amu Darya River 229
Table 12.1 Physical reservoir properties.
Parameter
Caprock
Reservoir
Basement
Depth (m)
2270~2310
2310~2690
2690~2730
Porosity
0.00001
0.098
0.00001
Permeability (mD)
0.00001
54
0.00001
0.5
Krg
Krw
krg / Krw
0.4
0.3
0.2
0.1
0
0.2
0.4
0.6
Sw
1
0.8
Figure 12.2 The relative permeability relationship diagram for the Callovian-Oxfordian
in the study area [21].
3.10E + 02
20-FGSAT
5-FGSAT
2.50E – 02
2.70E + 02
5-FPR
2.30E + 02
1.50E – 02
10-FPR
2.10E + 02
Saturation
2.00E – 02
10-FGSAT
2.50E + 02
1.00E – 02
1.90E + 02
5.00E – 03
0.00E + 00
31/1/1
29/1/1
23/1/1
21/1/1
19/1/1
17/1/1
15/1/1
13/1/1
11/1/1
09/1/1
07/1/1
05/1/1
03/1/1
01/1/1
1.50E + 02
27/1/1
20-FPR
1.70E + 02
25/1/1
Reservoir pressure (bar)
2.90E + 02
3.00E – 02
Time(Y/M/D)
Figure 12.3 Time-dependent diagram of reservoir pressure and gas saturation.
Note: FPR is the reservoir pressure. FGSAT is the reservoir gas saturation. The time period is January
of the first year to December of the 30th year.
230 Carbon Dioxide Capture and Acid Gas Injection
the 5×5 model is somewhat high at the early stage, and those of the 10×10
and 20×20 models is basically the same. In the middle and later stages,
the saturation of the 5×5 model slowly decreases and stays in the middle,
that of the 10×10 model is the lowest, and that of the 20×20 model is the
highest. Theoretically, the 5×5 gridding scheme is better, but it can lead to
more grids, and additional calculation time is needed. This paper analyzes
the relative errors of the 5×5 and 20×20 grids compared to the 10×10 grid
to determine the optimal gridding scheme. Comparing the 5×5 grid to the
10×10 grid, the maximum relative reservoir pressure error is 0.000021, the
minimum is –0.008466 and the average value is 0.006404. The maximum
relative saturation error is 0.009934, the minimum error is 0.000256 and
the average value is 0.006505. Comparing the 20×20 grid to the 10×10 grid,
the maximum relative reservoir pressure error is 0.003596, the minimum
value is –0.003596 and the average value is –0.008175. The maximum relative saturation error is 0.015857, the minimum error is –0.004544 and the
average value is 0.008755. The relative error of the 5×5 grid and 10×10
grids is therefore very small. With respect to the whole, the small error may
be negligible. In general, based on experience, the operational efficiency
and acid gas reinjection, the grid scheme of 10×10 is optimal.
12.4.3
The Development and Exploitation Mode
According to the 10×10 grid scheme and the above-mentioned parameters, the paper establishes a geological model (Figure 12.4). The acid gas
that is reinjected into the reservoir is produced for the purposes of this
research. Nine development and exploitation modes are designed in nine
different models (Table 12.2). According to the produced acid gas, the gas
component is the total CO2, and the component ratios are 8:2 and 7:3.
12.5
Results and Discussion
12.5.1
Reservoir Pressure
The changes of the reservoir pressure can reflect reservoir energy losses
and are an important indicator for measuring the production mode. The
changes of the reservoir pressure are basically identical in models 1, 2, 3,
4, and 5, (i.e., the pressure decreases slowly, and the pressure lapse rate
decreases gradually). In model 6, during the flow period, the reservoir
pressure decreases to 197 bar, and after 35 years, the reservoir pressure
Investigation of the Right Bank of the Amu Darya River 231
2000 m
460 m
2000 m
(a)
–2270 m
–2310 m
Caprock
Reservoir
–2690 m Baserock
–2730 m
(b)
Figure 12.4 The geological model in the study area (a: 3D view; b: sectional view).
Table 12.2 The differences of the nine development models.
Mode
Maximum of
CO2/H2S Injection rate daily production
gas ratio
(m3/day)
(m3/day)
Productive life
without injection +
injection (year)
Model 1
7:3
187
10000
10+20
Model 2
7:3
187
10000
15+20
Model 3
7:3
187
10000
20+20
Model 4
8:2
281
10000
15+20
Model 5
10:0
261
10000
15+20
Model 6
7:3
80000
10000
15+20
Model 7
7:3
80000
1×105
15+20
Model 8
7:3
50000
1×106
20+20
Model 9
7:3
60000
1×107
15+20
232 Carbon Dioxide Capture and Acid Gas Injection
3.00E + 02
Reservoir pressure (bar)
2.50E + 02
2.00E + 02
1.50E + 02
1.00E + 02
5.00E + 01
Model 1
Model 2
Model 3
Model 4
Model 5
Model 6
Model 7
Model 8
Model 9
0.00E + 00
01/1/1 06/1/1 11/1/1 16/1/1 21/1/1 26/1/1 31/1/1 36/1/1
Time (Y/M/D)
Figure 12.5 Time-dependent reservoir pressure diagram.
increases to 227 bar. In model 7, during the flow period, the reservoir pressure drops to 27 bar, and during the gas injection period, the reservoir
pressure is approximately 27 bar. In model 8, during the flow period, the
reservoir pressure begins to drop quickly, with a change in reservoir pressure of approximately 25 bar after 7 years, which is small. After the acid
gas is injected, the change of the reservoir pressure is also small, approximately 26 bar. In model 9, the reservoir pressure first declines quickly and
then declines slowly; 5 years later, the reservoir pressure is basically stable.
After injecting the gas, the pressure increases slightly, and the increment
is very small (Figure 12.5). Due to the rapid rates of pressure losses, which
almost reach 1 in models 8 and 9, production models such as these are not
accepted because they counteract the stability of the reservoirs.
12.5.2
Gas Sequestration
Gas sequestration is an indicator for optimizing the production models.
Figures 12.6 and 12.7 show that the gas sequestration decreases linearly in
models 1, 2, 3, 4, and 5, and the injected acid gas amount does not affect the
acid gas amount in the reservoir because the injection amount is very small
compared with the original amount of gas (6.56 × 108 m3 on the ground
surface). The injection amount of these models should be stored in the
Investigation of the Right Bank of the Amu Darya River 233
1.20E + 09
Gas in place (m3)
1.00E + 09
Model 1
Model 2
8.00E + 08
Model 3
Model 4
6.00E + 08
Model 5
4.00E + 08
Model 6
Model 7
2.00E + 08
Model 8
0.00E + 00
01/1/1 06/1/1 11/1/1 16/1/1 21/1/1 26/1/1 31/1/1 36/1/1
Model 9
Time (Y/M/D)
Figure 12.6 Time-dependent diagram of the total amount of gas sequestration.
reservoir, although it is of no use for enhancing gas recovery. In models 6, 7,
8 and 9, during the flow period, the gas in place decreases, and the lowering
rates differ due to different maximum daily production amounts. The rate
of model 9 is the highest, followed by the rates of models 8 and 7. The gas in
place increases during the gas injection period. The rising rate is also different because of the different acid gas injection amounts. The rate of model 6
is the highest, followed by that of model 7. Based on a detailed analysis
of model 6, the CO2 sequestration in the reservoir slowly increases after
21 years, and the H2S sequestration in the reservoir slowly increases after
25 years. In model 7, CO2 and H2S sequestration drop slowly ­during the
flow period. The CO2 and H2S sequestration increases first at a fast speed
and later at a slower speed after the acid gas is injected. At the same time,
the CO2 sequestration is higher than the H2S sequestration. In model 8,
the CO2 and H2S amounts of the reservoir decrease slowly over a 20-year
flow period. The CO2 amount of the reservoir begins to quickly increase
in June of the 22nd year, and the H2S amount of the reservoir begins to
quickly increase in February of the 23rd year. In model 9, after acid gas
is injected, the gas in place exhibits a small increase, the CO2 and H2S
sequestration increase, and the rising rates gradually decrease. Based on
this phenomenon and the combination with the differences of the models,
the rising rate of the gas in place is the maximum during the gas i­njection
period of model 6, although the final total amount is greater than that of
the original gas in place, which counteracts the reservoir stability. The CO2
and H2S sequestration dramatically increase, which may result in reservoir “burdens” such as sulfur and hydrate blockage. Considering the gas
234 Carbon Dioxide Capture and Acid Gas Injection
Model 1 CO2
8.00E + 08
Model 2 CO2
Model 3 CO2
7.00E + 08
Model 4 CO2
Model 5 CO2
CO2, H2S sequestration
6.00E + 08
Model 6 CO2
5.00E + 08
Model 7 CO2
4.00E + 08
Model 9 CO2
3.00E + 08
Model 2 H2S
Model 8 CO2
Model 1 H2S
Model 3 H2S
Model 4 H2S
2.00E + 08
Model 5 H2S
1.00E + 08
Model 6 H2S
0.00E + 00
01/1/1 06/1/1 11/1/1 16/1/1 21/1/1 26/1/1 31/1/1 36/1/1
Time (Y/M/D)
(a)
Model 2 CO2
1.80E + 08
CO2, H2S sequestration
Model 8 H2S
Model 1 CO2
2.00E + 08
Model 3 CO2
1.60E + 08
Model 4 CO2
1.40E + 08
Model 6 CO2
Model 5 CO2
1.20E + 08
Model 7 CO2
1.00E + 08
Model 9 CO2
Model 8 CO2
Model 1 H2S
8.00E + 07
Model 2 H2S
Model 3 H2S
6.00E + 07
Model 4 H2S
4.00E + 07
(b)
Model 7 H2S
Model 5 H2S
2.00E + 07
Model 6 H2S
0.00E + 00
01/1/1 06/1/1 11/1/1 16/1/1 21/1/1 26/1/1 31/1/1 36/1/1
Model 8 H2S
Model 7 H2S
Time (Y/M/D)
Figure 12.7 (a) Time-dependent diagram of the amounts of CO2 sequestration and H2S
sequestration, (b) zooming into (a) by shortening the Y-axis.
Investigation of the Right Bank of the Amu Darya River 235
sequestration during the flow period, the gas sequestration of model 7
decreases steadily, and those of models 8 and 9 decrease quickly. The production modes of models 8 and 9 belong to the depletion development.
The storage amounts increase after acid gas injection, the CO2 rising rate
is greater and the amount of CO2 increases more rapidly because there is
more CO2 than H2S in the reservoir and because there is more CO2 than
H2S in the injected acid gas.
12.5.3
Production
In models 1, 2, 3, 4, 5 and 6, the daily production is always the maximum
value, which reveals that there is a possibility of daily production increase.
Therefore, changes in daily production are observed after it is enhanced by
one order of magnitude in models 7, 8 and 9. In model 7, the daily production reaches to 1×105 m3/d in the first 12 years and decreases until the acid
gas is injected. It then increases up to year 30, while the increasing speed
gradually decreases and inclines to stability, approximately 1×105 m3/d. In
model 8, daily production decreases from 1×106 m3/d to 411 m3/d rapidly
at first, but it later gradually slows down. When the acid gas is injected, it
begins to increase step by step within 2 years until it remains constant after
8 years. In model 9, the daily production declines rapidly in the first few
months and then slowly reduces until it equals zero in year 7. Following
the injection of acid gas, it again increases with a small value, as shown in
Figure 12.8. The daily production in model 6 remains constant for 35 years.
The natural gas in the reservoirs is much richer, which suggests that the
maximum daily production will be enhanced. The model 7 maintains the
maximum daily production for 12 years and increases slightly 15 years after
the acid gas injection, which indicates stable and decreasing production
stages, which is helpful for understanding the reservoirs and verifies the
enhanced function of the acid gas injection. Models 8 and 9 belong to the
depletion development with a much lower production speed. Therefore,
the enhancement of the production in the short term can use the depletion
development, although this method is not advisable for the enhanced acid
gas injection.
The breakthrough of the acid gas should be noted during the acid
gas injection period. In model 7, the CO2 amount begins to increase on
January 1 of the 17th year, and the H2S amount increases on June 1 of
the 17th year (Figure 12.9). When the acid gas is driven to the production
well, the injected acid gas will break through, as shown in Figures 12.10
236 Carbon Dioxide Capture and Acid Gas Injection
1.40E + 06
Daily production (m3/d)
1.20E + 06
1.00E + 06
Model 3
Model 4
8.00E + 05
Model 5
6.00E + 05
Model 6
Model 7
4.00E + 05
Model 8
Model 9
2.00E + 05
0.00E + 00
01/1/1 06/1/1 11/1/1 16/1/1 21/1/1 26/1/1 31/1/1 36/1/1
Time (Y/M/D)
(a)
Daily production (m3/d)
2.00E + 05
1.50E + 05
Model 3
Model 4
Model 5
1.00E + 05
Model 6
Model 7
Model 8
5.00E + 04
Model 9
0.00E + 00
01/1/1 06/1/1 11/1/1 16/1/1 21/1/1 26/1/1 31/1/1 36/1/1
(b)
Time (Y/M/D)
Figure 12.8 (a) Time-dependent diagram for the daily production, (b) zooming into (a)
by shortening the Y-axis.
and 12.11. Near the ground of the production well, there are CO2 and H2S
plumes, and the CO2 plume is larger than the H2S plume. In model 9, the
CO2 amount begins to increase on March 1 of the 17th year, and the H2S
amount increases on October 1 of the 17th year (Figure 12.12). When the
acid gas is driven to the production well, the injected acid gas will break
through, as shown in Figures 12.13 and 12.14. Near the ground of the production well, there are CO2 and H2S plumes, and the CO2 plume is larger
Investigation of the Right Bank of the Amu Darya River 237
Daily production(m3/d)
2.50E + 04
2.00E + 04
CO2
1.50E + 04
1.00E + 04
H2 S
5.00E + 03
15/1/1
15/2/1
15/3/1
15/4/1
15/5/1
15/6/1
15/7/1
15/8/1
15/9/1
15/10/1
15/11/1
15/12/1
16/1/1
16/2/1
16/3/1
16/4/1
16/5/1
16/6/1
16/7/1
16/8/1
16/9/1
16/10/1
16/11/1
16/12/1
17/1/1
17/2/1
17/3/1
17/4/1
17/5/1
17/6/1
17/7/1
17/8/1
17/9/1
17/10/1
17/11/1
17/12/1
18/1/1
18/2/1
18/3/1
18/4/1
18/5/1
18/6/1
18/7/1
18/8/1
18/9/1
18/10/1
18/11/1
18/12/1
0.00E + 00
Time (Y/M/D)
Figure 12.9 Time-dependent diagram of the acid gas production of the production well
over 15~18 years in model 7.
Note: WCO2 is the CO2 concentration of the production well. WH2S is the H2S content of the production well.
0
1000
Production well
–2300
–2400
–2500
–2600
–2700 m
0.000
0.241
0.482
0.723
2000 m
0.964
Figure 12.10 The CO2 plume in the 18th year of model 7.
0
1000
Production well
–2300
–2400
–2500
–2600
–2700 m
0.000
0.078
0.157
0.235
2000 m
0.313
Figure 12.11 The H2S plume in the 18th year of model 7.
than the H2S plume. The acid gas amount dramatically increases during the
production stage, which counteracts the acid gas storage because injected
acid gas breaks through the production well. The H2S rising rate in the production well is smaller than the CO2 rising rate because the H2S content of
the injected acid gas is very small.
1.40E + 04
1.20E + 04
1.00E + 04
CO2
8.00E + 03
6.00E + 03
4.00E + 03
H2S
2.00E + 03
0.00E + 00
15/1/1
15/2/1
15/3/1
15/4/1
15/5/1
15/6/1
15/7/1
15/8/1
15/9/1
15/10/1
15/11/1
15/12/1
16/1/1
16/2/1
16/3/1
16/4/1
16/5/1
16/6/1
16/7/1
16/8/1
16/9/1
16/10/1
16/11/1
16/12/1
17/1/1
17/2/1
17/3/1
17/4/1
17/5/1
17/6/1
17/7/1
17/8/1
17/9/1
17/10/1
17/11/1
17/12/1
18/1/1
18/2/1
18/3/1
18/4/1
18/5/1
18/6/1
18/7/1
18/8/1
18/9/1
18/10/1
18/11/1
18/12/1
Darily production(m3/d)
238 Carbon Dioxide Capture and Acid Gas Injection
Time (Y/M/D)
Figure 12.12 Time-dependent diagram of the acid gas production of the production well
over 15~18 years in model 9.
0
1000
Production well
–2300
–2400
–2500
–2600
–2700 m
0.000
0.238
0.475
0.713
2000 m
0.950
Figure 12.13 The CO2 plume in the 18th year of model 9.
0
1000
Production well
–2300
–2400
–2500
–2600
–2700 m
0.000
0.074
0.148
0.222
2000 m
0.295
Figure 12.14 The H2S plume in the 18th year of model 9.
12.5.4
Recovery Ratio and Recovery Percentage
In model 7, the recovery ratio increases linearly during the flow period and
decreases after 12 years. The recovery ratio first increases slowly during
gas injection and later increases linearly. The recovery ratio first increases
exponentially in model 9; later, the value is invariant. During the gas
Investigation of the Right Bank of the Amu Darya River 239
160
Recovery ratio (%)
140
120
100
80
Model 7
Model 9
60
40
20
0
01/1/1 06/1/1 11/1/1 16/1/1 21/1/1 26/1/1 31/1/1 36/1/1
Time (Y/M/D)
Figure 12.15 Time-varying diagram of recovery ratio of models 7 and 9.
injection period, the recovery ratio first increases slowly and later increases
linearly, but the rising rate is smaller than that in model 7 (Figure 12.15).
In model 7, the change trend of daily recovery percentage is the same
as that of the daily production. The daily recovery percentage is 0.015%
during the flow period. The value decreases until the 13th year, and the
minimum is 0.0035%. The value increases to 0.011% during the gas injection. The daily recovery percentage decreases first in model 9. The daily
recovery percentage increases until the gas is injected. However, the value
after 35 years is smaller than that in model 7 (Figure 12.16).
12.6
Conclusions
1. The natural gas production mode is usually the flowing production. On the right bank of the Amu Darya River, we put
forward a win-win approach, which is to inject acid gas from
the production wells of tertiary recovery.
2. Acid gas reinjection is feasible for the development of Amu
Darya blocks.
3. Using the optimal strategy can enhance the daily production and gas recovery with a suitable injection rate. Based
on the long-term and sustainable development in the study
area, the best development strategy includes an acid gas ratio
of 7:3, an injection rate of 8×104 m3/day, a maximum daily
production of 1 × 105 m3/day, a flowing 15-year period and
the injection of 20-year acid gas.
240 Carbon Dioxide Capture and Acid Gas Injection
0.2
Recovery percent/day (%)
0.18
0.16
0.14
0.12
0.1
Model 7
0.08
Model 9
0.06
0.04
0.02
0
01/1/1
06/1/1
11/1/1
16/1/1
21/1/1
26/1/1
31/1/1
36/1/1
Time (Y/M/D)
(a)
0.04
Recovery percent/day (%)
0.035
0.03
0.025
0.02
Model 7
Model 9
0.015
0.01
0.005
0
01/1/1
(b)
06/1/1
11/1/1
16/1/1
21/1/1
26/1/1
31/1/1
36/1/1
Time (Y/M/D)
Figure 12.16 (a) Time-dependent diagram of the recovery percentage of models 7 and 9,
(b) zooming into (a) by shortening the Y-axis.
12.7
Acknowledgments
We acknowledge financial support from the National Natural Science
Foundation of China (Grant No. 41274111) and the Hundred Talent
Program of the Chinese Academy of Sciences.
Investigation of the Right Bank of the Amu Darya River 241
References
1. H. Ye, J. Wang. Challenges and Solutions to Acid Gas Reservoir Development.
Oil Forum. vol. 28, no. 4, pp. 63–65, DOI: 10.3969/j.issn.1002-302x.2009.04.018.
2009.
2. X. Liu, Q. Li, L. Du, X. Li. Economic Comparison of Both Acid-gas Reinjection
and Sulfur Recovery in High-sour Gasfields. Natural Gas Technology and
Economy. vol. 6, no. 4, pp. 55–59, DOI: 10.3969/j.issn.2095-1132.2012.04.015.
2012.
3. Z. Du. Oversea Experiences of Acid Gas Reservoir Development and their
Enlightenments. Natural Gas Industry. vol. 26, no. 12, pp. 35–37, 2006.
4. Q. Li, G. Liu, X. Liu. “Development of Management Information System of
Global Acid Gas Injection Projects”. In: Y. Wu, J.C. Carroll, Q. Li, Eds., Gas
Injection for Disposal and Enhanced Recovery. Wiley. pp. 243–254, 2014.
5. Q. Li, X. Li, L. Du, G. Liu, X. Liu, N. Wei. “Potential Sites and Early Opportunities
of Acid Gas Re-injection in China”. In: Y. Wu, J.J. Carroll, W. Zhu, Eds., Sour
Gas and Related Technologies. Wiley Scrivener. pp. 131–140, 2012.
6. X. Wu, S. Zeng. On the difficulties and countermeasures of security development in Puguang high sour gas field. Chemical Enterprise Management 34,
pp. 147, 2015.
7. Q. Li, X. Liu, L. Du, B. Bai, Z. Fang, M. Jing, X. Li. Economics of acid gas reinjection with comparison to sulfur recovery in China. Energy Procedia. vol. 37,
pp. 2505–2510, DOI: 10.1016/j.egypro.2013.06.132. 2013.
8. Q. Li, X. Liu, L. Du, X. Li. “Review of Mechanical Properties Related Problems
for Acid Gas Injection”. In: Y. Wu, J.C. Carroll, Q. Li, Eds., Gas Injection for
Disposal and Enhanced Recovery. Wiley. pp. 275–292, 2014.
9. B. Yu, Y. Z. Liu, X. Q. Liang, L. Wang. Interfacial Evolution and Migration
Characteristics of Acid Gas Injected into a Saline Aquifer. Chinese Journal
of Chemical Engineering. vol. 21, no. 4, pp. 434–441, DOI: 10.1016/s10049541(13)60481-4. 2013.
10. Y. Wu, J. C. Carroll, Q. Li, eds., Gas Injection for Disposal and Enhanced
Recovery. Hardcover ed. Advances in Natural Gas Engineering. eds. Y. Wu,
J.C. Carroll, Wiley-Scrivener: New York. p. 400, 2014.
11. J. J. Trivedi, T. Babadagli, R. G. Lavoie, D. Nimchuk. Acid Gas Sequestration
During Tertiary Oil Recovery: Optimal Injection Strategies and Importance
of Operational Parameters. Journal of Canadian Petroleum Technology. vol. 46,
no. 3, pp. 60–68, 2007.
12. M. Pooladi-Darvish, H. Hong, S. Theys, R. Stocker, S. Bachu, S. Dashtgard.
CO2 Injection for Enhanced Gas Recovery and Geological Storage of CO2 in
the Long Coulee Glauconite F Pool, Alberta. In: 2008 SPE Annual Technical
Conference and ExhibitionSPE: Denver, Colorado, USA. pp. SPE 115789,
2008.
242 Carbon Dioxide Capture and Acid Gas Injection
13. C. Khan, R. Amin, G. Madden. Effects of CO2 and acid gas injection on
enhanced gas recovery and storage. Journal of Petroleum Exploration and
Production Technology. vol. 3, no. 1, pp. 55–60, DOI: 10.1007/s13202-0120044-8. 2012.
14. J. Lubas, W. Szott, P. Jakubowicz. Effects of Acid Gas Reinjetcion on CO2
Concentration in Natural Gas Produced from Borzecin Reservoir. NAFTAGAZ. vol. ROK LXVIII, no. 7, pp. 405–410, 2012.
15. Z. Hou, Y. Gou, J. Taron, U. J. Gorke, O. Kolditz. Thermo-hydro-mechanical
Modeling of Carbon Dioxide Injection for Enhanced Gas-recovery (CO2EGR): A Benchmarking Study for Code Comparison. Environmental Earth
Sciences vol. 67, no. 2, pp. 549–561, DOI: 10.1007/s12665-012-1703-2. 2012.
16. X. Li, D. Elsworth. Geomechanics of CO2 Enhanced Shale Gas Recovery.
Journal of Natural Gas Science and Engineering vol. 26 no. pp. 1607–1619,
DOI: 10.1016/j.jngse.2014.08.010. 2014.
17. J. Wang, Z. Wang, D. Ryan, C. Lan. A Study of the Effect of Impurities on CO2
Storage Capacity in Geological Formations. International Journal of Greenhouse
Gas Control. vol. 42, no. pp. 132–137, DOI: 10.1016/j.ijggc.2015.08.002. 2015.
18. Q. Li, D. Kuang, G. Liu, X. Liu. Acid Gas Injection: A Suitability Evaluation
for Sequestration Site in Amu Darya Basin, Turkmenistan. Geological Review.
vol. 60, no. 5, pp. 1133–1146, 2014.
19. M. Lie, S. Xu, L. Wu, J. He, H. Ming, R. Zhang. Features of Gas Reservoirs in
Sub Salit Faulted Reef Complex in the Right Bank of the Amu Darya River,
Turkmenistan, and their significance in Exploration: a Case Study of the BP
Gas Reservoir. Natural Gas Industry. vol. 35, no. 6, pp. 24–29, 2015.
20. S. Guo, K. Xu, X. Xu, G. Zhang, S. Xiong. Directional Drilling Technology in
the Right Bank of the Amu Darya River, Turkmenistan. Drilling & Production
Technology. vol. 38, no. 4, pp. 122–124, 2015.
21. H. Zou, X. Wang, J. Kang, H. Zhu, Q. Wang. Key Technologies of Restoring
the Productivity of Sour Gas Fields in Turkmenistan. Natural Gas Industry.
vol. 29, no. 9, pp. 94–96, 2009.
22. Q. Wang, X. Cheng, H. Fei, L. Wu, R. Chen. Reservoir-Forming Process of a
Block in the Right Bank of the Amu Darya River. Natural Gas Exploration &
Development. vol. 35, no. 2, pp. 1–4, 2012.
23. B. Liu, X. Wang, W. Li, S. Zhang, J. Dong, Y. Yang. The Problems and Solutions
in The Drilling at The Right Bank of Amu Darya in Turkmenistan. In: IADC/
SPE Asia Pacific Drilling Technology Conference and Exhibition: Tianjin,
China. pp. SPE 155825, 2012.
24. H. Wen, B. Gong, R. Zheng, H. Liu, L. Wu, R. Chen, S. Li, S. Chen. Deposition
and Diagenetic System of Carbonate in Callovian-Oxfordian of Samandepe
Gasfield, Turkmenistan. Journal of Jilin University (Earth Science Edition).
vol. 42, no. 4, pp. 991–1002, 2012.
25. W. Yan, H. Jiang, X. Wu, G. Xu, R. Chen. Geological Characteristics and
Drilling Design of the Amu Darya Right Bank Block, Turkmenistan. Natural
Gas Industry. vol. 30, no. 5, pp. 73–76, 2010.
Investigation of the Right Bank of the Amu Darya River 243
26. Schlumberger. ECLIPSE reservoir simulation software-Technical Description.
2012.
27. Q. Li, D. Kuang, G. Liu, X. Liu. Acid Gas Injection: A Suitability Evaluation for
the Sequestration Site in Amu Darya Basin, Turkmenistan. Geological Review.
vol. 60, no. 5, pp. 1133–1146, 2014.
28. Q. Wang, X. Yan, W. Xu, R. Zheng, F. Li, X. Wang, L. Wu. SequencePaleogeographic Characteristics and Evolution of Callovian-Oxfordian in
Amu Darya Basin, Turkmenistan. Geology and Exploration. vol. 50, no. 4,
pp. 795–804, 2014.
29. R. Zheng, H. Chen, Q. Wang, C. Cui, H. Fei, W. Xu. The Reservoir Characteristics
and their Controlling Factors of Callovian-Oxfordian Carbonates in Amu
Darya Basin. Acta Petrologica Sinica. vol. 30, no. 3, pp. 779–788, 2014.
30. T. Wang. The Research on Evaluating Carbonate Fractured Reservior of Amu
Darya Right Bank Block. Xi’an Shiyou University: Xi’an. 2011.
31. C. Liu, S. Zhou, L. Yan, S. Deng. Drilling Engineering Difficulties in HighSulfur Gas Fields in Turkmenistan and Countermeasures. Natural Gas
Industry. vol. 33, no. 1, pp. 79–84, 2013.
32. C. Li. The Relationship between Rock Compressibility and Porosity. China
Offshore Oil and Gas (Geology). vol. 17, no. 5, pp. 355–358, 2003.
Also of Interest
Check out these forthcoming related titles coming soon
from Scrivener Publishing
Also Available in the Advances in Natural Gas
Engineering Series:
Acid Gas Extraction for Disposal and Related Topics, edited by Ying Wu,
John Carroll, and Weiyao Zhu, ISBN 9781118938614. This fifth volume in
the series, Advances in Natural Gas Engineering, offers the most in-depth
and up-to-date treatment of acid gas extraction for disposal, an important
innovation in natural gas engineering. NOW AVAILABLE!
Gas Injection for Disposal and Enhanced Recovery, edited by Ying Wu, John
Carroll, and Qi Li, ISBN 9781118938560. This fourth volume in the series,
Advances in Natural Gas Engineering, offers the most in-depth and up-todate treatment of the disposal and enhanced recovery of natural gas. NOW
AVAILABLE!
Sour Gas and Related Technologies, edited by Ying Wu, John J. Carroll, and
Weiyao Zhu, ISBN 9780470948149. Written by a group of the most wellknown and knowledgeable authors on the subject in the world, volume
three focuses on one of the hottest topics in natural gas today, sour gas.
This is a must for any engineer working in natural gas, the energy field, or
process engineering. NOW AVAILABLE!
Acid Gas Injection and Related Technologies, edited by Ying Wu and John
J. Carroll, ISBN 9781118016640. Focusing on the engineering of natural
gas and its advancement as an increasingly important energy resource,
this volume is a must-have for any engineer working in this field. NOW
AVAILABLE!
Carbon Dioxide Sequestration and Related Technologies, edited by Ying Wu
and John J. Carroll, ISBN 9780470938768. volume two focuses on one of
the hottest topics in any field of engineering, carbon dioxide sequestration.
NOW AVAILABLE!
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (247–248)
2017 © Scrivener Publishing LLC
Other Titles from the Same Authors:
Acid Gas Injection and Carbon Dioxide Sequestration, by John J. Carroll,
ISBN 9780470625934. Provides a complete overview and guide on the hot
topics of acid gas injection and CO2 sequestration. NOW AVAILABLE!
Carbon Dioxide Thermodynamic Properties Handbook 2nd Edition, by Sara
Anwar and John J. Carroll, ISBN 9781119083580. This second edition of a
contemporary classic is the most comprehensive collection of carbon dioxide (CO2) data ever compiled, with new information added to the revision, a “one stop shop” for any engineer or scientist who works with carbon
dioxide in any industry. NOW AVAILABLE!
Index
Amine, 63, 64, 72, 91, 92, 94–99,
105, 109–113, 127–143,
147–157, 161–183
Benedict-Webb-Rubin
(see equation of state)
Bubble point, 56, 59
Capital cost, 92, 114
Caprock, 229, 231
Casing, 208
Chemical reaction, 85, 91–93, 95,
100, 103, 114, 134, 143,
148–154, 163
Compressibility factor,
185, 188, 193, 195
Compressor, 2, 37, 39, 95,
103, 162, 166, 176
Corrosion, 130, 163, 222
Crude oil, 187, 188
DEA (see amine)
DEG (see glycol)
Density, 55–60, 63–68, 73,
77–78, 81, 137–138, 143,
155, 166, 172, 177, 202–205,
208, 210, 217, 226
Dew point, 56, 59
Deviation factor, (see
compressibility factor)
Diethylene glycol, (see glycol)
DGA, (see amine)
EG, (see glycol)
Enhanced oil recovery, 55, 64, 148
Enthalpy, 1–36, 39–52, 85,
140–142, 155–156
EOR (see enhanced oil recovery)
Equation of state, 2, 40, 56, 59, 60,
71, 73, 76–77, 133, 149, 202, 224
Ethane, 117, 118, 223
Ethylene glycol, (see glycol)
Glycol, 127, 130–131, 135, 137, 139
Hydrostatic gradient, 208–209, 213
Injection pressure, 166, 200, 212, 214
Interfacial area, 101–102
Lee-Kesler (see equation of state)
Mass transfer, 93, 95, 96, 100–
103, 130, 140, 165, 185–188,
190–192, 194, 196, 197
MDEA (see amine)
MEA (see amine)
Mercaptan, 63, 130
Methanol, 175
Operating cost, 91–93, 99, 110, 114
Peng-Robinson (see equation of state)
Permeability, 187–188,
194, 201, 227, 229
Ying Wu, John J. Carroll and Weiyao Zhu (eds.) CO2 Capture and Acid Gas Injection, (245–246)
2017 © Scrivener Publishing LLC
245
246
Index
Pipeline, 64, 95, 96, 102,
105, 111, 174, 199, 200,
213, 217
Piperidine, 127, 130, 131,
134–135
Porosity, 187, 188, 194, 201,
227, 229
Pump, 57–58, 66, 75, 95–97,
103, 133, 141, 177, 179
Reaction (see chemical reaction)
Redlich-kwong (see equation of state)
Safety, 64, 92, 213, 222
Soave (see equation of state)
TEG (see glycol)
Thiol (see mercaptan)
Triethylene glycol (see glycol)
Tubing, 199–200, 208, 217
Viscosity, 72, 179, 185–186,
188–193, 195, 197, 203–204
z-factor (see compressibility factor)