Carbon dioxide capture and acid gas injection 9781118938676, 1118938674, 9781118938683, 1118938682, 9781118938706, 1118938704, 978-1-118-93866-9, 113-115-115-1
429 79 15MB
English Pages [256] Year 2017
Polecaj historie
![The Three Sisters: Acid Gas Injection, Carbon Capture and Sequestration, and Enhanced Oil Recovery (Advances in Natural Gas Engineering) [7]
1119510066, 9781119510062](https://dokumen.pub/img/200x200/the-three-sisters-acid-gas-injection-carbon-capture-and-sequestration-and-enhanced-oil-recovery-advances-in-natural-gas-engineering-7-1119510066-9781119510062.jpg)
Table of contents :
Content: Cover
Title Page
Copyright Page
Contents
Preface
1 Enthalpies of Carbon Dioxide-Methane and Carbon Dioxide-Nitrogen Mixtures: Comparison with Thermodynamic Models
1.1 Introduction
1.2 Enthalpy
1.3 Literature Review
1.3.1 Carbon Dioxide-Methane
1.3.2 Carbon Dioxide-Nitrogen
1.4 Calculations
1.4.1 Benedict-Webb-Rubin
1.4.2 Lee-Kesler
1.4.3 Soave-Redlich-Kwong
1.4.4 Peng-Robinson
1.4.5 AQUAlibrium
1.5 Discussion
1.6 Conclusion
References
2 Enthalpies of Hydrogen Sulfide-Methane Mixture: Comparison with Thermodynamic Models
2.1 Introduction
2.2 Enthalpy 2.3 Literature Review2.4 Calculations
2.4.1 Lee-Kesler
2.4.2 Benedict-Webb-Rubin
2.4.3 Soave-Redlich-Kwong
2.4.4 Redlich-Kwong
2.4.5 Peng-Robinson
2.4.6 AQUAlibrium
2.5 Discussion
2.6 Conclusion
References
3 Phase Behavior and Reaction Thermodynamics Involving Dense-Phase CO2 Impurities
3.1 Introduction
3.2 Experimental
3.3 Results and Discussion
3.3.1 Phase Behavior Studies of SO2 Dissolved in Dense CO2 Fluid
3.3.2 The Densimetric Properties of CS2 and CO2 Mixtures
References
4 Sulfur Recovery in High Density CO2 Fluid
4.1 Introduction
4.2 Literature Review 4.3 Methodology4.4 Results and Discussion
4.5 Conclusion and Future Directions
References
5 Carbon Capture Performance of Seven Novel Immidazolium and Pyridinium Based Ionic Liquids
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
5.4 Results and Discussion
5.4.1 Density
5.4.2 Critical Properties
5.4.3 CO2 Solubility
5.4.4 The Effect of Changing the Cation
5.4.5 The Effect of Changing the Anion 5.4.6 Henry's Law Constant, Enthalpy and Entropy Calculations5.4.7 Thermodynamic Modeling of CO2 Solubility
5.5 Conclusion
Acknowledgements
References
6 Vitrisol® a 100% Selective Process for H2S Removal in the Presence of CO2
6.1 Introduction
6.2 Case Definition
6.3 "Amine-Treated" Cases by PPS
6.3.1 Introduction to PPS
6.3.2 Process Description
6.3.3 PFD
6.3.4 Results
6.3.4.1 Case 1
6.3.4.2 Case 2
6.4 Vitrisol® Process Extended with Regeneration of Active Component
6.4.1 Technology Description
6.4.2 Parameters Determining the Process Boundary Conditions 6.4.3 Absorption Section6.4.4 Regeneration Section
6.4.5 Sulphur Recovery Section
6.4.6 CO2-Absorber
6.4.7 PFD
6.5 Results
6.6 Discussion
6.6.1 Comparison of Amine Treating Solutions to Vitrisol®
6.6.2 Enhanced H2S Removal of Barnett Shale Gas (case 2)
6.7 Conclusions
6.8 Notation
References
Appendix 6-A: H&M Balance of Case 1 (British Columbia shale) of the Amine Process
Appendix 6-B H&M Balance of Case 2a (Barnett shale) of the Amine Process with Stripper Promoter
Appendix 6-C H&M Balance of Case 3 (Barnett shale) of the Amine Process (MEA)
Citation preview
Carbon Dioxide Capture and Acid Gas Injection
Scrivener Publishing 100 Cummings Center, Suite 541J Beverly, MA 01915-6106 Publishers at Scrivener Martin Scrivener ([email protected]) Phillip Carmical ([email protected])
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 1 Enthalpies 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 2 Enthalpies 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 3 Phase 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 4 Sulfur 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 5 Carbon 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 6 Vitrisol 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 8 Improved 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 9 Strategies 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 10 Modeling 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 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 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 5
1.4 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 NP
(H o H )exp (H o H )calc (H o H )calc
100%
(1.9)
1.4.1 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 7 600
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).
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 9 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).
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 13 700 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).
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 15 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).
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 21 800 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).
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 25 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).
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)
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
(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
1 NP
E E H exp H calc E H calc
(2.5)
100%
2.4.1 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 ( 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] [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 Experimental Work 5.2.1 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 Modeling 5.3.1 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
bi
i
i 0.45724
mi
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 )
5.4 Results and Discussion 5.4.1 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)
AAD (%)
[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 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
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
5.5 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 “Amine-Treated” Cases by PPS 6.3.1 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 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 n˙CO andn˙amine 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