Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Absorption and desorption of CO2 and CO in alkanolamine systems Jamal, Aqil 2002-12-31

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_2002-75023X.pdf [ 11.63MB ]
Metadata
JSON: 1.0058971.json
JSON-LD: 1.0058971+ld.json
RDF/XML (Pretty): 1.0058971.xml
RDF/JSON: 1.0058971+rdf.json
Turtle: 1.0058971+rdf-turtle.txt
N-Triples: 1.0058971+rdf-ntriples.txt
Original Record: 1.0058971 +original-record.json
Full Text
1.0058971.txt
Citation
1.0058971.ris

Full Text

ABSORPTION AND DESORPTION OF C02 AND CO IN ALK AN OL AMINE SYSTEMS by AQIL JAMAL M.Sc, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia, 1992 M.Tech., Indian Institute of Technology, Madras, India, 1986 B. Tech., Harcourt Butler Technological Institute, Kanpur, India, 1984 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES Department of Chemical and Biological Engineering We accept this thesis^ as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA July 2002 © Aqil Jamal, 2002 In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada ABSTRACT Absorption and desorption of carbon dioxide (CO2) and carbon monoxide (CO) in aqueous alkanolamine solutions are modeled and important kinetic and physical property data are obtained using novel experimental methods. The model is based on the principle of diffusional mass transfer accompanied with fast to very slow chemical reactions in the liquid phase. Fast reactions are represented by CO2 absorption/desorption in aqueous alkanolamines and slow reactions are represented by CO absorption in aqueous diethanolamine (DEA). The experiments for CO2 absorption and desorption were conducted in a novel hemispherical contactor designed and developed in this work. The absorption experiments were conducted at near atmospheric pressure using pure CO2 saturated with water at 293 to 323 K with initially unloaded solutions. The desorption experiments were performed at 333 to 383 K for C02 loadings between 0.02 to 0.7 moles of CO2 per mole of amine using humidified N2 gas as a stripping medium. The experiments for CO absorption were carried out in a 660 mL batch autoclave reactor at 313 to 413 K with amine concentration between 5 to 50-wt% in distilled water. The partial pressure of CO in the reactor was varied from 800 to 1100 kPa. The data for C02 absorption and desorption in aqueous amine systems were analyzed using a new, rigorous mathematical model. The model predicts the experimental results well for all amine systems studied. The results indicate ii that the theory of absorption with reversible chemical reaction could be used to predict desorption rates. The kinetic data obtained show that desorption experiments could be used to determine both forward and backward rate constants accurately. The absorption experiments on the other hand could only be used to determine forward rate constants. The data for CO absorption in aqueous diethanolamine (DEA) solutions were analyzed using the model for mass transfer with extremely slow reactions. The data are consistent with a mechanism by which formyl-diethanolamine (DEAF) is predominantly formed by direct insertion of CO into DEA. The data also confirm that DEAF formation via the DEA-formate reaction is relatively slow and reversible. iii TABLE OF CONTENTS ABSTRACT ii LIST OF TABLES x LIST OF FIGURES xviiACKNOWLEDGEMENT xxix CHAPTER 1. INTRODUCTION 1 1.1 Background 1 1.1.1 Commercially Important Amines 2 1.1.2 Acid Gas Treating Process 5 1.2 Importance of Kinetic and Physical Property data in Design and Simulation of Gas Treating Systems 8 1.2.1 Kinetic Data on C02-Amine Systems 9 1.2.2 Kinetic Data on CO-Amine systems 10 1.3 Objectives and Scope of This Work 11 1.3.1 Organization of the Thesis 2 PART I. C02 ABSORPTION AND DESORPTION IN AMINE SYSTEMS CHAPTER 2. LITERATURE EVIEW 14 2.1 Mass Transfer with Chemical Reaction 5 2.2 Reaction Mechanisms of C02-Amine-Water Systems 20 2.2.1 C02 Reactions with Water 22.2.2 C02 Reactions with Amines 1 2.2.3 C02 Reactions with Aqueous Amine Blends 26 2.3 Kinetic Data for C02 Absorption/Desorption in Aqueous Amines 27 2.3.1 Kinetic Data for C02-MEA-Water System 22.3.2 Kinetic Data for C02-DEA-Water System 30 iv 2.3.3 Kinetic Data for C02-MDEA-Water System 34 2.3.4 Kinetic Data for C02-AMP-Water System 7 2.3.5 Kinetic Data for Amine Blends 42 2.4 Research Needs 46 CHAPTER 3. EXPERIMENTAL APPARATUS AND METHODS 48 3.1 Overview 43.2 Hemispherical Contactor 9 3.3 Experimental Setup and Procedure 52 3.4 Data Acquisition and Calibration 7 3.5 Chemicals 60 CHAPTER 4. MATHEMATICAL MODEL 61 4.1 Reaction Mechanism 62 4.1.1 Reactions for C02+AMP+H20 (Base Case) 62 4.1.2 Reactions for 02+MEA+H20 64 4.1.3 Reactions for C02+DEA+H204.1.4 Reactions for C02+MDEA+H20 65 4.1.5 Reactions for C02+MEA+MDEA+H20 66 4.1.6 Reactions for C02+MEA+AMP+H204.1.7 Reactions for C02+DEA+MDEA+H20 66 4.1.8 Reactions for C02+DEA+AMP+H20 7 4.2 Reaction Rates 64.2.1 Reaction Rates for C02+AMP+H20 System (Base Case) 67 4.2.2 Reaction Rates for C02+MEA+H20 System 68 4.2.3 Reaction Rates for C02+DEA+H20 System 69 4.2.4 Reaction Rates for C02+MDEA+H20 System 69 4.2.5 Reaction Rates for C02+MEA+MDEA+H20 System 70 v 4.2.6 Reaction Rates for C02+MEA+AMP+H20 System 70 4.2.7 Reaction Rates for C02+DEA+MDEA+H20 System 72 4.2.8 Reaction Rates for C02+DEA+AMP+H20 System 72 4.3 Reactive Gas Absorption/Desorption Model 73 4.3.1 Hydrodynamics of Liquid Film 74.3.2 Model Equations 76 4.3.3 Liquid Bulk Concentrations 81 4.3.4 Rate of Absorption or Desorption with Chemical Reaction 83 4.3.5 Rate of Absorption or Desorption without Chemical Reaction 83 4.3.6 Enhancement Factor 85 4.3.7 Overall Reaction Rate (rtotai) 84.4 Model Parameters 86 4.5 Numerical Implementation 87 4.6 Parameter Estimation 9 CHAPTER 5. RESULTS AND DISCUSSION 93 5.1 Model Verification 95.1.1 CO2 Absorption/Desorption in Aqueous Amines 93 5.1.2 Detailed Profiles in the Hemispherical Film 97 5.1.3 Numerical versus Analytical Solutions 102 5.2 Parametric Sensitivity Analysis 104 5.2.1 Effect of Operating Parameters 105.2.2 Effect of Physical Property Parameters 109 5.2.3 Effect of Kinetic Parameters 111 5.3 Henry's Constant and C02 Diffusivity in Amine Solutions 115 5.3.1 Correlation for Henry's Constant of C02 in Amine Solutions 115 5.3.2 Correlation for C02 Diffusivity in Amine Solutions 116 5.4 C02 Absorption and Desorption in Aqueous Amine Solutions 118 5.4.1 C02 Absorption/Desorption in Aqueous Solutions of MEA, DEA and AMP Solutions 120 vi 5.4.2 C02 Absorption/Desorption in Aqueous MDEA Solutions 134 5.4.3 C02 Absorption/Desorption in Aqueous Amine Blends 138 5.4.4 Comparison of the Parameter Estimates with Literature Data. ...142 5.4.5 Predicted Absorption/Desorption Rates Based on Correlations Developed in This Work 147 5.5 Conclusions 155 PART II CO ABSORPTION AND CO INDUCED DEGRADATION IN AMINE SYSTEMS CHAPTER 6. INTRODUCTION AND LITERATURE REVIEW 157 6.1 Background 156.2 Literature Review 8 6.3 Objectives 161 CHAPTER 7. EXPERIMENTAL APPARATUS AND EXPLORATORY EXPERIMENTS 163 7.1 Experimental Apparatus 167.2 Exploratory Experiments 6 7.2.1 CO Absorption in Aqueous DEA Solutions 166 7.2.2 Material Balance 168 7.2.3 CO Absorption in Pure DEA 170 7.3 Proposed Reaction MechanismCHAPTER 8. MATHEMATICALMODEL 172 8.1 Reaction mechanism 178.2 Reaction Rates 3 8.3 Mathematical Model 178.4 Model Parameters 7 8.5 Parameter Estimation 178 vii CHAPTER 9. RESULTS AND DISCUSSION 180 9.1 Absorption and Degradation in CO-DEA System 180 9.1.1 Determination of kLa 189.1.2 Determination of HCo-D by Nitrogen Analogy 184 9.1.3 Determination of k3 and k.3 188 9.1.4 Determination of ki, k2 and HCo-DEA 195 9.2 Absorption and Degradation in CO-MDEA System 200 9.3 Absorption and Degradation in CO-DEA+DEAF System 202 9.4 Model Predictions 203 9.5 Process Implications 6 9.6 Conclusions 208 CHAPTER 10. OVERALL CONCLUSIONS AND RECOMMENDATIONS 210 10.1 Conclusions on C02 Absorption and Desorption in Amine Systems Kinetics 2110.2 Conclusions on CO Absorption and CO Induced Degradation in Amine Systems 4 10.3 Recommendations for Future Work 216 NOMENCLATURE 218 REFERENCES 222 APPENDICES 235 A. Determination of C02 Loading in Amine Solutions 235 B. Calibration of Measuring Instruments 240 C. Analytical Solution for Physical Absorption/Desorption Model 252 D. Derivation of Rate Expression for Zwitterion Mechanism 259 E. Density and Viscosity of Aqueous Amine Solutions 262 viii F. Henry's Constant of C02 and N20 in Aqueous Amine Solutions 272 G. Diffusivity of C02 and N20 in Aqueous Amine Solutions 293 H. Equilibrium constants 31I. Determination of Gas-Side of Gas-Side Mass Transfer Coefficient 319 J. Determination of Formate and DEAF 324 K. Experimental Data 327 ix LIST OF TABLES Table 1.1: Molecular structure of commonly used alkanolamines 3 Table 2.1: Summary of the results reported on the C02-MEA-Water system 28 Table 2.2: Summary of the results reported on the C02-MEA-Water system 31 Table 2.3: Summary of results reported on the C02-MDEA-Water system 36 Table 2.4: Carbamate stability constants for MEA, DEA and AMP by C13-NMR (Sartori and Savage, 1983) 39 Table 2.5: Summary of results reported on the C02-AMP+Water system 40 Table 2.6: Summary of results reported on C02-MEA+MDEA+Water and C02-DEA+MDEA+Water systems 44 Table 4.1: Overall reaction rate (rtotai) • 86 Table 4.2: Parameters for absorption/desorption model 88 </ Table 5.1: Base case operating conditions for parametric sensitivity analysis 105 Table 5.2: Effect of deviation in physical property parameters on C02 absorption and desorption rates in aqueous systems of MEA, DEA, MDEA and AMP 110 x Table 5.3: Kinetic parameters for C02-aquous amine systems studied in this work 112 Table 5.4: Effect of reactions involving zwitterion deprotonation to water and hydroxyl ions on the rate of absorption (see rxns 4.3, 4.4, 4.12, 4.13, 4.18,4.19) 113 Table 5.5: Kinetic parameters for C02-aquous amine systems studied in this work (excluding reactions for zwitterion deprotonation to water and hydroxyl ions) 114 Table 5.6: Experimental conditions for C02 absorption/desorption in aqueous amine solutions 119 Table 5.7: Estimates of rate constants in eq. (5.4) from absorption and desorption data 128 Table 5.8: Estimates of rate constants in eq. (5.5) from absorption and desorption data 128 Table 5.9: Estimates of rate constants in eq. (5.6) from absorption and desorption data 129 Table 5.10: Estimates of rate constants in eq. (5.7) from absorption and desorption data 136 Table 5.11: Estimates of the combined rate constants for amine blends from desorption data 138 Table 5.12: Comparison of the activation energies of k^ k10, k16 and k22 146 Table 7.1: Overall CO balance as a function of temperature for CO absorption in 30 wt% aqueous DEA solutions 169 XI Table 7.2: Formate and DEAF concentrations of solutions resulting from exposing pure DEA to CO for 5 hours at p°0 = 690 kPa 169 Table 8.1: Model parameters 178 Table 9.1: Henry's constant of CO in 30 wt% aqueous DEA solution from N2-Analogy 186 Table 9.2: Henry's constants of CO and N2 in water and organic solvents (Fogg and Gerrad, 1991) 187 Table 9.3: Estimates of the rate and equilibrium constants for reaction (6.3) '. 193 Table 9.4: Estimates of ki, k2 and HC0_DEA 196 Table 9.5: CO reaction with aqueous DEA and MDEA 202 Table 9.6: Measured and predicted Formate and DEAF concentrations for CO absorption in 30 wt% aqueous DEA solution 204 Table 9.7: Effect of amine concentration on CO reaction with aqueous DEA solution at 393 K 205 Table 9.8: CO absorption in aqueous solution of 30 wt% DEA and 18.4 wt% DEAF 205 Table 9.9: CO absorption in 30 wt% aqueous MDEA solution 205 Table A.1: Calibration Data for Gastec Detector Tubes No. 2HH (Standard Solution: 10, 000 ppm (by weight) Na2C03) N2 Flow = 120 ml/min, Sparge Time = 40 min) 239 Xll Table B.1: GC operating conditions for gas phase N2O/CO2 measurement 248 Table B.2: Calibration Data for N20 measurement using GC 249 Table B.3: Calibration Data for CO2 measurement using GC 250 Table E.1: Binary Parameters of the Redlich-Kister Equation of the Excess Volume (eq. E.4 and E.5) 264 Table E.2: Parameters of the Density Equation for Pure Fluids (eq. E.6)....265 Table E.3: Density of Aqueous Amine Blends at 303 K and 323 K (Total Amine = 25 wt%, Water = 75 wt%) 266 Table E.4: Binary Parameters of the Redlich-Kister Equation for the Viscosity Deviation (eqs. E.8 and E.9) 269 Table E.5: Parameters of the Viscosity Equation for Pure Fluids (eq. E. 10).270 Table E.6: Viscosity of Aqueous Amine Blends at 303 K (Total Amine = 25 wt%, Water = 75 wt%) 271 Table F.1: Henry's constant of C02 in water 278 Table F.2: Henry's constant of N20 in water 279 Table F.3: Henry's constant of N20 in aqueous MEA solutions 280 Table F.4: Henry's constant of N20 in aqueous DEA solutions 281 xiii Table F.5: Henry's constant of N20 in aqueous MDEA solutions 283 Table F.6: Henry's constant of N20 in aqueous AMP solutions 285 Table F.7: Henry's constant of N20 in MEA+MDEA+H20 286 Table F.8: Henry's constant of N20 in MEA+AMP+H20 287 Table F.9: Henry's constant of N20 in DEA+MDEA+H20 288 Table F.10: Henry's constant of N20 in DEA+AMP+H20 289 Table F.11: Parameters in eq. (F.6) for Henry's constant of N20 in pure amines (Wang et al., 1992) 290 Table F.12: Parameters in excess Henry's constant of N20 in binary and ternary solvent systems (this work) 290 Table G.1: Diffusivity of C02 in water 301 Table G.2: Diffusivity of N20 in water 302 Table G.3: Diffusivity of N20 in aqueous MEA solutions 303 Table G.4: Diffusivity of N20 in aqueous DEA solutions 304 Table G.5: Diffusivity of N20 in aqueous MDEA solutions 305 Table G.6: Diffusivity of N20 in aqueous AMP solutions 307 xiv Table G.7: Diffusivity of N20 in MEA+MDEA+H20 308 Table G.8: Diffusivity of N20 in MEA+AMP+H20 309 Table G.9: Diffusivity of N20 in DEA+MDEA+H20 310 Table G.10: Diffusivity of N20 in DEA+AMP+H20 311 Table G.11: Constants in eq. (G.7) for interaction parameter B12 (this work) 311 Table G.12: Constants in eqs. (G.7) for interaction parameters An and Aj2 (this work 312 Table 1.1: Estimates of kg from experiments and correlation (CDEA = 25 wt%) 323 Table J.1: IC operating conditions for formate analysis 326 Table J.2: GC operating conditions for DEAF analysis 326 Table K.1: C02 absorption in pure water 327 Table K.2: N20 absorption in pure water 328 Table K.3: N20 absorption in 25 wt% aqueous MEA solution 329 Table K.4: N20 absorption in 25 wt% aqueous DEA solution 330 Table K.5: N20 absorption in 25 wt% aqueous MDEA solution 331 xv Table K.6: N20 absorption in 25 wt% aqueous AMP solution 332 Table K.7: N20 absorption in aqueous solutions of 12.5 wt% MEA plus 12.5 wt% MDEA 333 Table K.8: N20 absorption in aqueous solutions of 12.5 wt% MEA plus 12.5 wt% AMP 334 Table K.9: N20 absorption in aqueous solutions of 12.5 wt% DEA plus 12.5 wt% MDEA 33Table K.10: N20 absorption in aqueous solutions of 12.5 wt% DEA plus 12.5 wt% AMP 336 Table K.11: C02 absorption in aqueous MEA solutions 337 Table K.12: C02 absorption in aqueous DEA solutions 337 Table K.13: C02 absorption in aqueous DEA solutions for estimation of kg...338 Table K.14: C02 absorption in aqueous MDEA solutions 339 Table K.15: C02 absorption in aqueous AMP solutions 340 Table K.16: C02 absorption in aqueous blends of MEA+MDEA 342 Table K.17: C02 absorption in aqueous blends of MEA+AMP 342 Table K.18: C02 absorption in aqueous blends of DEA+MDEA 343 XVI Table K.19: C02 absorption in aqueous blends of DEA+AMP 343 Table K.20: C02 desorption from aqueous MEA solutions 344 Table K.21: C02 desorption from aqueous DEA solutions 345 Table K.22: C02 desorption from aqueous MDEA solutions 346 Table K.23: C02 desorption from aqueous AMP Solutions 347 Table K.24: C02 desorption from aqueous blends of MEA+MDEA 348 Table K.25: C02 desorption from aqueous blends of MEA+AMP 348 Table K.26: C02 desorption from aqueous blends of DEA+MDEA 349 Table K.27: C02 desorption from aqueous blends of DEA+AMP 349 xvii LIST OF FIGURES Figure 1.1: Typical absorber/stripper system for acid gas removal 7 Figure 2.1: Comparison of second-order rate constant for C02-AMP reaction 39 Figure 3.1: Schematic Drawing of the Absorption/Desorption with the Hemispherical Contactor 50 Figure 3.2: Experimental setup 53 Figure 3.3: Typical computer output from an absorption experiment (MEA = 14.3 wt%, T = 303 K, PT = 101.2 kPa, QL = 2.1 mL/s) 59 Figure 3.4: Typical computer output from a desorption experiment:(MEA = 20 wt%, a = 0.279 mol/mol, T = 378 K, PT = 202 kPa, QL = 2.2 mL/s) 5Figure 4.1: Schematic of the Liquid Film 74 Figure 4.2: Numerical scheme for model solution 91 Figure 4.3: Numerical scheme for parameter estimation 92 Figure 5.1: Predicted and experimental absorption rates of C02 in aqueous solutions of MEA, DEA and MDEA at 303 K (pC02 = 97.0 kPa) 94 Figure 5.2: Predicted and experimental desorption rates of C02 from aqueous solutions of MEA, DEA and MDEA at 373 K 95 xviii , Figure 5.3: Predicted and experimental desorption rates for CO2 desorption from aqueous MEA solution at 373 K 96 Figure 5.4: Predicted C02 profiles at different latitudes for C02 absorption in pure water (T = 313 K, QL = 2.0 mL/s, pco = 92.7 kPa) 98 Figure 5.5: Predicted C02 profiles at different latitudes for C02 desorption from aqueous AMP solution (T = 383 K, QL = 2.0 mL/s, CAMP = 2.11 kmol/m3, Ptotai = 220 kPa, a = 0.2 mole of C02/mole of amine) 98 Figure 5.6: Effect of amine concentration on CO2 profile at 8 = 90° for CO2 absorption in aqueous AMP solution (T = 313 K, QL = 2.0 mL/s, pco =92.7 kPa) 99 Figure 5.7: Predicted concentration profiles for C02 absorption in aqueous AMP solution at 6 = 45° (T = 313 K, QL = 2.0 mL/s, CAMP = 0.223 kmol/m3, pco = 92.7 kPa) 100 Figure 5.8: Predicted concentration profiles for C02 absorption in aqueous AMP solution at 0 = 90° (T = 313 K, QL = 2.0 mL/s, CAMP = 0.223 kmol/m3, Pco2 = 92.7 kPa) 100 Figure 5.9: Predicted concentration profiles for C02 desorption in aqueous AMP solution at 9 = 45° (T = 383 K, QL = 2.0 mL/s, CAMP = 2.11 kmol/m3, Ptotai = 220 kPa, a = 0.2 mole of C02/mole of amine)... 101 Figure 5.10: Predicted concentration profiles for C02 desorption in aqueous AMP solution at 6 = 90° (T = 383 K, QL = 2.0 mL/s, CAMp = 2.11 kmol/m3, Ptotai = 220 kPa, a = 0.2 mole of C02/mole of amine) 10Figure 5.11: Absorption rates of CO2 in pure water predicted from analytical and numerical solutions (QL = 2.0 mL/s, pco = 87-97 kPa) 103 xix Figure 5.12: Effect of amine concentration on CO2 absorption and desorption rates in aqueous AMP system 106 Figure 5.13: Effect of CO2 loading on C02 absorption and desorption rates in aqueous AMP system 107 Figure 5.14: Effect of temperature on C02 absorption and desorption rates in aqueous AMP system 107 Figure 5.15: Effect of C02 partial pressure on C02 absorption and desorption rates in aqueous AMP system 108 Figure 5.16: Effect of total pressure on CO2 absorption and desorption rates in aqueous AMP system 108 Figure 5.17: Effect of liquid flow rate on C02 absorption and desorption rates in aqueous AMP system 109 Figure 5.18: Predicted and experimental absorption rates for C02 absorption in aqueous MEA solution at 303 K 122 Figure 5.19: Predicted and experimental absorption rates for C02 absorption in aqueous DEA solution at 303 K 123 Figure 5.20: Predicted and experimental absorption rates for C02 absorption in aqueous AMP solution at 303 K 123 Figure 5.21: Predicted and experimental desorption rates for C02 desorption from aqueous MEA solution at 378 K 124 Figure 5.22: Predicted and experimental desorption rates for C02 desorption from aqueous DEA solution at 382 K 124 xx Figure 5.23: Predicted and experimental desorption rates for CO2 desorption from aqueous AMP solution at 378 K 125 Figure 5.24: Predicted and experimental desorption rates for CO2 desorption in aqueous MEA solution at 333 to 378 K using literature correlations 125 Figure 5.25: Predicted and experimental desorption rates for C02 desorption in aqueous DEA solution at 343 to 382 K using literature correlations 126 Figure 5.26: Predicted and experimental desorption rates for C02 desorption in aqueous AMP solution at 343 to 378 K using literature correlations 126 Figure 5.27: Arrhenius plot of the estimates for ki0 from absorption and desorption data 129 Figure 5.28: Arrhenius plot of the estimates for k10 /K^K^ from absorption and desorption data 130 Figure 5.29: Arrhenius plot of the estimates for k_10 /k„ from absorption and desorption data 130 Figure 5.30: Arrhenius plot of the estimates for k16 from absorption and desorption data 131 Figure 5.31: Arrhenius plot of the estimates for k16 /K16K17 from absorption and desorption data 131 Figure 5.32: Arrhenius plot of the estimates for k_16 /k17 from absorption and desorption data 132 XXI Figure 5.33: Arrhenius plot of the estimates for ki from absorption and desorption data 132 Figure 5.34: Arrhenius plot of the estimates for k., /K1K2 from absorption and desorption data 133 Figure 5.35: Arrhenius plot of the estimates for k_1/k2 from absorption and desorption data 133 Figure 5.36: Predicted and experimental absorption rates for CO2 absorption in aqueous MDEA solution at 303 K 135 Figure 5.37: Predicted and experimental desorption rates for C02 desorption from aqueous MDEA solution at 378 K 135 Figure 5.38: Predicted and experimental desorption rates for C02 desorption in aqueous MDEA solution at 343 to 378 K using literature correlations 136 Figure 5.39: Arrhenius plot of the estimates for k22 from absorption and desorption data 137 Figure 5.40: Arrhenius plot of the estimates for k.22 from absorption and desorption data 137 Figure 5.41: Arrhenius plot of the estimates for k.i0/k24 from desorption data 139 Figure 5.42: Arrhenius plot of the estimates for k.-|/k25 from desorption data 139 Figure 5.43: Arrhenius plot of the estimates for k.i0/k26 from desorption data 140 xxii Figure 5.44: Arrhenius plot of the estimates for k.i6/k27 from desorption data 140 Figure 5.45: Arrhenius plot of the estimates for k_i/k28 from desorption data 141 Figure 5.46: Arrhenius plot of the estimates for k.i6/k29 from desorption data 141 Figure 5.47: Comparison of the second-order rate constant (ki0) for C02-MEA reaction determined in this work with those reported in the literature 143 Figure 5.48: Comparison of the second-order rate constant (ki6) for C02-DEA reaction determined in this work with those reported in the literature 144 Figure 5.49: Comparison of the second-order rate constant (k-i) for C02-AMP reaction determined in this work with those reported in the literature 145 Figure 5.50: Comparison of the second-order rate constant (k22) for C02-MDEA reaction determined in this work with those reported in the literature 146 Figure 5.51: Predicted and experimental desorption rates CO2-MEA as a function of temperature and C02 loading 148 Figure 5.52: Predicted versus experimental desorption rates C02-MEA calculated from the correlation developed in this work 148 Figure 5.53: Predicted and experimental desorption rates for C02-DEA system as a function of temperature and C02 loading 149 xxiii Figure 5.54: Predicted versus experimental desorption rates CO2-DEA system calculated from the correlation developed in this work 149 Figure 5.55: Predicted and experimental desorption rates for CO2-AMP system as a function of temperature and C02 loading 150 Figure 5.56: Predicted versus experimental desorption rates CO2-AMP system calculated from the correlation developed in this work 150 Figure 5.57: Predicted and experimental desorption rates for C02-MDEA system as a function of temperature and CO2 loading 151 Figure 5.58: Predicted versus experimental desorption rates C02-MDEA system calculated from the correlation developed in this work 151 Figure 5.59: Predicted versus experimental absorption rates for CO2-MEA+MDEA system at 303 K based on the correlation developed in this work (pco =97 kPa) 153 Figure 5.60: Predicted versus experimental absorption rates for CO2-DEA+MDEA system at 303 K based on the correlation developed in this work (pco =97 kPa) 153 Figure 5.61: Predicted versus experimental absorption rates for CO2-MEA+AMP system at 303 K based on the correlation developed in this work (pC02=97 kPa) 154 Figure 5.62: Predicted versus experimental absorption rates for C02-DEA+AMP system at 303 K based on the correlation developed in this work (pco, =97 kPa) 154 Figure 7.1: Schematic diagram of the experimental setup 164 xxiv Figure 7.2: CO loading as a function of time and temperature (DEA= 30 wt%, p°0= 760-990 kPa) 167 Figure 8.1: Schematic diagram of chemical reaction autoclave ..174 Figure 9.1: Nitrogen absorption in 30 wt% aqueous DEA solution at 1050 rpm and 284-333 K 182 Figure 9.2: Nitrogen absorption in 30 wt% aqueous DEA solution at 1050 rpm and 353-413 K 182 Figure 9.3: Measured and predicted volumetric mass transfer coefficient 184 Figure 9.4: Measured and predicted formate ions and DEAF concentration..191 Figure 9.5: Arrhenius plot of the estimates for the forward rate constant of reaction (6.3) 193 Figure 9.6: Arrhenius plot of the estimates for the reverse rate constant of reaction (6.3) 194 Figure 9.7: Arrhenius plot of the estimates for the equilibrium constant of reaction (6.3) 195 Figure 9.8: Measured and predicted partial pressures for CO absorption in 30 wt% aqueous DEA solution 197 Figure 9.9: Arrhenius plot of the estimates for the second order rate constant of reaction (6.1) 198 Figure 9.10: Arrhenius plot of the estimates for the second order rate constant of reaction (6.2) 198 XXV Figure 9.11: Effect of reaction (6.3) on the formation of formate and DEAF at T = 393 K , = 1000 kPa and CDEA= 30 wt% 200 Figure 9.12: Measured and predicted concentration and partial pressure profiles for CO absorption in aqueous solution of 30 wt% DEA and 18.4 wt% DEAF at 393 K 203 Figure 9.13: Predicted concentration profiles for CO absorption in 30 wt% aqueous DEA solution atT = 313 K and pco = 1000 kPa 207 Figure 9.14: Predicted concentration profiles for CO absorption in 30 wt% aqueous DEA solution at T = 393 K and pco = 1,000 kPa 207 Figure A.1: Experimental setup for CO2 measurement in Liquid Samples 236 Figure A.2: Calibration Curve for Gastec Detector Tube No. 2HH for C02 Measurement in Liquid Samples 238 Figure B.1: Calibration Curve for Pressure Transducer on Absorption/Desorption Chamber (Omega Model PX202-030GV) 241 Figure B.2: Calibration Curve for Pressure Transducer on Heating Tank (Omega Model PX202-030GV) 241 Figure B.3: Setup for Mass Flow Meter Calibration using a Soap Film Meter 242 Figure B.4: Calibration Curve for Measuring Dilution N2 Flow Rate using Mass Flow Meter (Brooks Model 5700) 242 Figure B.5: Calibration Curve for Measuring Stripping N2 Flow Rate using Mass Flow Meter (Colepalmer Model GFM171) 243 XXVI Figure B.6: Calibration Curve for Measuring CO2 Flow Rate using Mass Flow Meter (Colepalmer Model GFM171) 243 Figure B.7: Calibration Curve for Measuring N20 Flow Rate Using Mass Flow Meter (Colepalmer Model GFM 171) 244 Figure B.8: Setup for C02 Analyzer Calibration 245 Figure B.9: Calibration Curve for Gas-Phase C02 Measurement using Infrared Analyzer (Model NOVA-300) 246 Figure B.10: Calibration Curve for Gas Phase N20 measurement using GC (Shimadzu Model GC 8A, Column: Chromosorb 102) 247 Figure B.11: Calibration Curve for Gas-Phase C02 Measurement using GC (Shimadzu Model GC 8A, Column: Chromosorb 102) 248 Figure E.1: Measured and calculated densities at 303 and 323 K 265 Figure E.2: Measured and calculated viscosities at 303 K 270 Figure F. 1: Henry's constant of C02 in water 291 Figure F.2: Henry's constant of N20 in water 291 Figure F.3: Measured and calculated values of Henry's constant of N20 in aqueous amine solutions between 288 to 393 K 292 Figure F.4: Measured and calculated values of Henry's constant of N20 in aqueous amine blends between 288 to 393 K 292 xxvii Figure G.1: Diffusivity of CO2 in water as a function of temperature 298 Figure G.2: Diffusivity of N20 in water as a function of temperature 298 Figure G.3: Stokes-Einstein plot for diffusivity of N20 in aqueous alkanolamine solutions 299 Figure G.4: Stokes-Einstein plot for diffusivity of N20 in aqueous alkanolamine blends 299 Figure G.5: Measured and calculated diffusivities of N20 in aqueous amine solutions 300 Figure G.6: Measured and calculated diffusivities of N20 in aqueous amine blends 300 Figure 1.1: Measured and calculated gas-side mass-transfer coefficients....322 xxviii ACKNOWLEDGEMENT I wish to express my sincere gratitude and admiration to my thesis supervisors, Dr. Axel Meisen and Dr. C. Jim Lim for their guidance and encouragement over the entire course of this work without which this work would not have been possible. I would also like to thank my thesis committee members Dr. Alfred Guenkel, and Dr. Bruce Bowen for their guidance and support. I am indebted to Mr. Peter Roberts in the workshop for his professional work in the construction of the experimental apparatus. Thanks to Mr. Horace Lam and Ms Qi Chen in the stores who were always very helpful and friendly. The financial support provided by the Natural Science and Engineering Research Council of Canada is gratefully acknowledged. Finally, I want to thank my parents, my wife and my son Cengiz for their love and understanding during the course of this work. xxix CHAPTER 1 INTRODUCTION i 1.1 Background The separation of acid gas impurities such as carbon dioxide (CO2) and hydrogen sulfide (H2S) from gas mixtures is an important operation in natural gas processing, petroleum refining, coal gasification and ammonia manufacture. The amount of CO2 and H2S in these gas streams may vary from a few hundred parts per million to more than 30 percent by volume. In most cases, the H2S is almost completely removed from the gas mixtures because it is highly toxic to humans and extremely corrosive to gas pipelines and other equipment. In natural gas, the C02 is removed to meet pipeline specifications, to enhance heating value and to reduce transportation costs. In ammonia plants, on the other hand, the C02 is removed from the synthesis gas mainly to prevent catalyst deactivation. Since C02 is widely regarded as a major greenhouse gas, potentially contributing to global warming, there has recently been considerable interest in developing technologies for capturing and sequestering the large quantities of CO2 produced from industrial sources such as fossil-fuel electric power generation facilities. Various technologies have been developed for C02 and H2S removal from gas streams. These include absorption by chemical and physical solvents, cryogenic separation and membrane separation. Among these methods, gas absorption by chemical solvents is one of the most popular and effective l methods. In general, aqueous solutions of alkanolamines are the most commonly used chemical solvents for the removal of acid gases. The basic idea in this type of process is to remove the acid component (e.g., C02 and/or H2S) from the gas-phase by using a basic species (i.e., aqueous amines) with which it reacts in a reversible manner. The reversibility of these reactions is extremely important as it enables continuous use of amine solution over extended periods. 1.1.1 Commercially Important Amines Table 1.1 lists some of the common alkanolamines used in gas treating applications. These solvents can be thought of as substituted ammonia molecules. The number of substitutions on the nitrogen atom determines the type of alkanolamine. In primary amines, one hydrogen atom on the nitrogen is replaced with a functional group, in secondary amines two hydrogen atoms are replaced and in tertiary amines all three hydrogen atoms are replaced. The chemical structure of alkanolamines is ideally suited for acid gas removal. The amine group provides the required basicity that allows it to react with acid gases reversibly and the hydroxyl group makes the amine more water-soluble thereby reducing its vapor pressure, so that less amine is lost at the top of the absorber and stripper. The reaction rates of H2S and C02 differ greatly in alkanolamine solutions because of the difference in their structure. As a Bronsted acid, H2S reacts directly with the amine function in the acid-base neutralization step. This neutralization is much faster than the time it takes for H2S to diffuse into the bulk liquids. Consequently, H2S can be considered in chemical equilibrium in the 2 liquid at all points in the contactor including the interfacial film and actual kinetic data are not generally required to model H2S transport. Table 1.1 Molecular structure of commonly used alkanolamines HO —C — C — N< HO—C—C HO—C—C N—H Monoethanolamine (MEA) Diethanolamine (DEA) -C—C—OH H2C H2C >N—H -C—C—OH HO—C—C HO—C—C N—CH, Diisopropanolamine (DIPA) Methyldiethanolamine (MDEA) CH3 I I M HO—C —C—N<,, I I ^ CH3 2-amino-2-methyl-1 -propanol (AMP) N XCH2—CH2—OH H 2-piperidine ethanol (PE) 3 The reaction of CO2 with a basic solvent is much slower than that of H2S. The slower reaction rate of C02 is due to its nature as a Lewis acid, which must hydrate before it can react by acid-base neutralization. It may also react directly with the amine to form a carbamate. The rate of hydration and carbamation are both slow and can be comparable to the rate of diffusion of C02. The reaction rate may therefore limit the overall absorption for C02. It is this fact which creates a need for reliable reaction rate data so that the acid gas contactor can be modeled accurately. Aqueous MEA and DEA solutions are generally used for bulk C02 removal when the partial pressure of C02 is relatively low and the product purity requirement is high. DIPA is used primarily in special applications where it is necessary to preferentially absorb H2S over C02. Both primary and secondary amines react fairly strongly with C02 to form stable carbamates and their heats of reactions are substantial. Consequently, these amines require substantial energy for regeneration. The energy is supplied in a stripper at elevated temperatures thereby making the amine susceptible to degradation. MDEA is the most commonly used tertiary amine and it is mostly used for the selective removal of H2S from natural gases. The most important difference between primary and secondary amines on the one hand and tertiary amines on the other is that tertiary amines do not directly react with CO2 to form carbamate. Instead, they are believed to catalyze the hydrolysis of C02 to carbonate and bicarbonate ions. This also means that regeneration of C02 rich MDEA solution requires less energy in comparison to primary and secondary amines. 4 In many gas-treating applications one finds that MDEA is too selective towards H2S (i.e., it allows too much C02 to slip through the absorber), while DEA and MEA are not selective enough. In MEA and DEA based plants, this lack of sufficient selectivity towards H2S causes too much C02 to be removed and pushes up solvent regeneration cost. Therefore, in recent years there has been considerable interest in using blends of MDEA and DEA or MDEA and MEA. The benefit of using such amine blends is that they combine the advantages of higher absorption rates of MEA and DEA and the better stripping characteristics of MDEA. Moreover, they provide an additional degree of freedom in the form of amine composition that can be manipulated to exercise some control on the selectivity of MDEA towards H2S over C02 by adjusting the gas-liquid contact time. Recently a new class of amines referred to as sterically hindered amines, has been introduced as a commercially attractive alternative to MEA, DEA and DIPA. Examples are 2-amino-2-methyl-1-propanol (AMP), which is the hindered form of MEA, and 2-piperidine ethanol (PE), which is the hindered form of DEA. In comparison to primary and secondary amines, hindered amines form less stable carbamate and have good selectivity for H2S. Exxon Research and Engineering Company have already commercialized a hindered amine based H2S-selective gas treating process (Goldstein et al., 1986). 1.1.2 Acid Gas Treating Process A typical process schematic for removing acid gases is shown in Figure 1.1. A feed gas containing typical hydrocarbons (CH4, C2H6, C3H8, etc.) along , 5 with acid gases is contacted countercurrently with the descending amine solution in a packed or plate column to provide favorable conditions for mass transfer and chemical reactions. The absorber operates at high pressure and low temperature (< 40 °C). The purified gas leaves the top of the absorber. The solution discharging from the bottom of the absorber is rich in acid gases and is heated in an exchanger after pressure reduction in a flash drum. The purpose of the flash is to liberate dissolved hydrocarbons. The rich solution is then fed to the stripper (or desorber) where acid gases are removed from the solution by steam stripping. The stripper operates at slightly above ambient pressure and high temperature (> 120 °C). The acid gases in the overhead accumulator are either incinerated or sent to a sulfur plant. The lean amine solution leaving the reboiler is cooled in the lean-rich heat exchanger, before being returned to the absorber. The major energy consumption in this process is in the stripper. The energy supplied to the reboiler is used for two reasons: (1) to produce enough water vapor so that the vapor phase partial pressure of acid gases is reduced to provide a driving force for desorption, and (2) to provide enough energy to reverse the reactions which occur in the absorber. In fact the reactions of CO2 with aqueous alkanolamine solutions are highly exothermic, releasing energy in the absorber and requiring energy in the stripper. 6 Purified Gas Absorber (High Pressure, Low Temperature) Raw Gas and Impuritie (C02,H2S...) Lean Amine Cooler Lean Amine, Solution Dissolved Hydrocarbons (Fuel Gas) Rich Amine Solution Flash Drum Lean-Rich | /^^Heat Exchanger • Pump ~ ,_ CO,,H,S.. Overhead z ^ Condenser Reflux Drum Stripper (Low Pressure, Boiling) Steam Reboiler I KeDOII IP Figure 1.1: Typical absorber/stripper system for acid gas removal The reboiler heat duty is the most significant operating cost of this type of system (Blauwhoff et al., 1985). It is, therefore, desirable to gain a better understanding of the the underlying principles behind the absorption/desorption process and to find solvents and/or operating modes, which reduce the reboiler heat duty. Stripping is also an area where amine solutions are exposed to high temperatures (-120 °C). This may accelerate certain undesirable irreversible reactions between amine and acid gases resulting in products from which the amine cannot be regenerated under typical operating conditions. As a result, significant amounts of valuable amine may be lost or rendered ineffective. Recently, amine degradation has been the subject of considerable research, but 7 it is still not much fully understood. In fact, amine degradation due to the presence of CO is one of the topics covered in this dissertation. 1.2 Importance of Kinetic and Physical Property Data in Design and Simulation of Gas Treating Systems Gas treating using alkanolamines has been practiced in industry for over half a century. However, it is only recently that substantial progress has been made in developing a fundamental understanding of these seemingly simple processes. Modern computer technology has made it possible to use sophisticated mass transfer rate-based models to design and simulate almost any type of gas treating systems (Katti, 1995; Katti and McDougal, 1986; Yu and Astarita, 1987; Krishnamurthy and Taylor, 1985; Sardar, 1985). These models have been particularly useful in the development and commercialization of mixed solvents, the composition of which can be tailored to specific applications (Katti, 1995). Comprehensive models require calculations of component and energy balances around each phase on every tray or section of packing. This in turn, involves determining the driving forces, mass and heat transfer coefficients, interfacial area, and interaction between mass transfer and chemical reaction. Thus a rate-based model consists of a vapor-liquid equilibrium (VLE) model, reaction kinetics model, hydrodynamic model and a model to describe the effect of reaction on mass transfer. Although these models are robust enough to treat gas-liquid contacting from fundamental principles and therefore have the potential of accurately predicting the performance of chemical absorbers and 8 strippers, the models are ultimately as accurate as the experimental data upon which they are based. While great advances have been made in developing increasingly complex models to simulate gas-treating units, similar advances are lacking for the basic kinetic and physical property data. 1.2.1 Kinetic Data for C02-Amine Systems A comprehensive literature review of basic kinetic data for commercially important amines is given in Chapter 2. Several general observations about the existing literature data are noted here. First, the great majority of data are gathered under absorption conditions. However, in designing and simulating real amine processes, data under stripping conditions are equally important. Second, the studies on absorption and desorption have been performed in isolation and there has never been a major attempt to investigate if data collected under absorber conditions can be utilized to predict desorption rates. Finally and most importantly, the experimental data on reaction kinetics have been mostly analyzed assuming highly simplified pseudo-first-order kinetics and pseudo-first-order or corresponding second-order rate coefficients have been reported. Whereas, this assumption greatly simplifies the mathematics, it does not adequately represent the parallel and reversible reactions that occur in typical acid gas-amine systems. In addition, the simplified approach provides no insight into the actual reactive mass transfer phenomena and the kinetic data have limited reliability for design and optimization and scale-up of industrial units. In real situations, rarely if ever is there a single mass transfer regime throughout column. 9 Therefore, it is highly desirable to develop a comprehensive model that includes all relevant possible reactions in an acid gas-amine system and that can be used to predict both absorption and desorption rates. The development of such a model is the primary objective of this thesis. 1.2.2 Kinetic Data on CO-Am ine Systems In gas treating plants, amine solutions are continuously regenerated and reused for extended periods. Thus, maintaining the quality of amine solution is absolutely critical for efficient and economic operation. Although acid gas-amine reactions are reversible, irreversible reactions may also occur resulting in products from which the amine cannot be regenerated under typical operating conditions. This phenomenon is called amine degradation. Numerous studies have been published on this subject (Dawodu and Meisen, 1991; Kim, 1988; Chakma, 1987; Chakma and Meisen, 1988; Kennard and Meisen, 1985; Kim and Sartori, 1984). However, all these studies are mainly concerned with amine degradation due to C02, carbon disulfide (CS2), carbonyl sulfide (COS) and H2S/C02 mixtures. However, some sour gas mixtures (e.g., gas mixtures in hydrogen plants and FCC units) may also contain significant concentrations of carbon monoxide that could also contribute to amine degradation. A literature review on this subject (see Chapter 6) reveals that no detailed scientific study has ever been done to investigate the kinetics of CO-induced degradation of alkanolamines. This problem is of considerable commercial importance because in many refinery gas-treating units, where CO partial pressures are high, large quantities of DEA can be lost or rendered ineffective. 10 Therefore, it could be considerable economic benefit to investigate the mechanism by which CO reacts with aqueous DEA solution and estimate relevant solubility and kinetic parameters. These data are essential for estimating DEA losses. Therefore, the second part of this thesis is devoted to investigating the kinetics of the CO-induced degradation of DEA. 1.3 Objectives and Scope of this Work The work presented in this thesis can be divided into two parts. Part 1 deals with the kinetics of C02 absorption/desorption in aqueous amines and Part 2 deals with the kinetics of CO-induced degradation of aqueous DEA. The main objectives of Part 1 were to obtain kinetic and physical properties data that are needed to predict C02 absorption/desorption rates in aqueous amine solutions and to investigate if the theory of absorption with reversible chemical reactions can be applied to predict desorption rates. To accomplish these objectives, a new hemispherical gas-liquid contactor was designed and absorption/desorption rates of C02 into aqueous solutions of MEA, DEA, MDEA, AMP and their mixtures were measured at various temperatures, amine concentrations and C02 loadings. The experimental data were interpreted by developing a comprehensive reactive mass transfer model. This model takes into account the reversibility of the chemical reactions, the effect of the diffusion of reactants and reaction products through reaction zone, and the gas-phase mass transfer resistance. The model is capable of predicting both absorption and desorption rates. ll In Part 2, the knowledge acquired in studying CCVamine kinetics (i.e., very fast reacting systems) was applied to investigate the kinetics of CO-induced degradation of aqueous DEA (i.e., very slow reacting system). This is the first detailed study ever done on this subject and the project grew out of discussion with Equilon Enterprises LLC, Westhollow Technology Center, Houston Texas and Shell Global Solutions International, Amsterdam, The Netherlands. The specific objectives of this study were to: (a) investigate the mechanism by which CO reacts with aqueous DEA, (b) identify major degradation products and (c) estimate the corresponding solubility and kinetic parameters over a range of temperatures and CO partial pressures so that DEA losses could be quantified and managed. To accomplish these objectives, a novel reaction mechanism was proposed and a mathematical model was developed. The model consists of a set of differential and algebraic equations that describes gas absorption with slow chemical reactions in a well-mixed batch reactor. The parameter estimation procedure presented in this work is based on a novel experimental approach that is fast, accurate and suitable for any gas-liquid reaction system. 1.3.1 Organization of the Thesis The thesis is divided into two parts. The first part (Chapters 2 to 5) covers experimental and modeling study of C02 absorption/desorption in aqueous alkanolamine solutions using a novel hemispherical contactor. The second part (Chapters 6 to 9) deals with the kinetics of CO induced degradation of DEA. Chapter 2 presents a comprehensive review of the literature on the kinetics of CO2 absorption and desorption in aqueous solutions of MEA, DEA, 12 MDEA, AMP and their blends. In this chapter, the theory of mass transfer with chemical reactions is presented to highlight the importance of the kinetics of these systems. Major gaps in knowledge are identified and future research needs that form the basis of most of the work presented are recommended. Chapter 3 describes the novel hemispherical contactor and experimental procedures to measure absorption and desorption rates of C02 in aqueous amine solutions. Chapter 4 presents the development of a comprehensive model for gas absorption and desorption over a hemispherical liquid film. This chapter also includes the algorithms for solving the model equations and for parameter estimation. Chapter 5 presents the results of the C02 absorptions/desorption studies. Chapters 6 to 9 present a comprehensive account of the work on CO-induced degradation of aqueous DEA. These chapters also present a novel technique (patterned on the so called "Nitrogen Analogy") to measure the physical solubility of CO in aqueous DEA solutions. Major conclusions and recommendations for future work arising from Parts 1 and 2 are summarized in Chapter 10. 13 CHAPTER 2 LITERATURE REVIEW CO2 absorption/desorption in aqueous amines involves mass transfer with multiple reversible chemical reactions. Under typical operating conditions, the rates of these reactions (particularly those involving molecular CO2) are of the same order of magnitude as rate of diffusion in the solution. Therefore, to model the CO2 removal process, an understanding of the mechanism and kinetics of these reactions is essential. A number of studies on the kinetics of C02-amine reactions in aqueous solutions have been reported in the literature. The experimental procedures in these studies typically involve measurements of C02 absorption rates in aqueous amines using laboratory apparatus with known hydrodynamics and interfacial areas. These rates are then interpreted using kinetic models to estimate rate constants. The purpose of this chapter is to emphasize the importance of kinetic data in predicting C02 absorption and desorption rates and review the results and methods reported in literature. The kinetic data for four important amines (MEA, DEA, MDEA and AMP) and their blends (MEA+MDEA, MEA+AMP, DEA+MDEA, DEA+AMP) have been compiled and research needs that are later studied in this work are identified. The discussion in this chapter is limited to reaction mechanisms and kinetics only. A review of important physical property data that are also needed for the process modeling of these systems is given in Appendices E to H. 14 2.1 Mass Transfer with Chemical Reaction From two-film theory (Lewis and Whitman, 1924) the rate of C02 absorption or desorption in aqueous amines can be represented by: NA=-ND = P;°2^C°2 (2.1) 1 , Hco2 kga Ek°a The term (pcc,2 -p£o2) 'n equation (2.1) represents the driving force for mass transfer, pcc,2 denotes the partial pressure of C02 in the bulk gas and pcc,2 denotes the equilibrium partial pressure of C02 corresponding to its concentration in the bulk liquid. For absorption pcc,2 > pcc,2 and for desorption Pco2 < Pco2 • The value of p°COz depends on the solubility of C02 in the liquid at the prevailing temperature and pressure. It can be calculated by solving thermodynamic equilibrium models. At low temperatures and high C02 partial pressures, the equilibrium favors absorption because most of the C02 in the liquid phase is chemically combined with the amine and water and the value of pCOz is much lower than pC02. At high temperatures and low C02 partial pressures, the equilibrium favors desorption because most of the C02 in the liquid is in molecular form and the value of pcc.2 is much higher than pcc.2. The equilibrium chemistry that governs the C02-Water-Amine system is therefore important in determining the mass transfer driving forces. The most commonly used models for calculating pco are those of Kent and Eisenberg 15 (1976), Deshmukh and Mather (1981) and Austgen (1989). The main difference among these models is in their description of liquid phase non-idealities. The model of Kent and Eisenberg (1976) does not account for liquid phase non-ideality and is therefore the simplest. The models of Deshmukh and Mather (1981) and Austgen (1989) properly account for thermodynamic non-idealities and, while more accurate, they are also mathematically complex and computationally time consuming. All three models have been used extensively, and comprehensive reviews have been provided by Weiland et al. (1993) and Austgen et al. (1991). The denominator in equation (2.1) represents the resistance to mass transfer and it consists of two terms: the gas-phase resistance (1/kga) and the liquid-phase resistance (Hcc.2 /Ek°a). Depending on the operating condition, one or both of these resistances are significant. The gas-side resistance to mass transfer is usually important for systems containing highly reactive amines such as MEA and piperazine or when the C02 partial pressure in the gas phase is low. Fortunately, the gas-side resistance to mass transfer can be easily estimated because the only major unknown parameters in evaluating the gas-film coefficient are the gas diffusivities, which are readily available in the literature (Reid etal., 1987). In C02 removal processes, the liquid-side resistance to mass transfer is generally dominant. The estimation of this resistance requires knowledge of the effect of chemical reaction on mass transfer. The effect is usually expressed in terms of the enhancement factor (E), which is defined as the ratio of the rate of 16 absorption or desorption with chemical reaction to the rate without chemical reaction. In most cases of practical importance, the enhancement factor is a complex function of a dimensionless parameter commonly referred to as the Hatta number (DeCoursey, 1982, 1992) given by: M=Vk2.CAMDc02 k° v ' Under the assumption that the amine concentration at the gas-liquid interface is not significantly different from that in the bulk solution, Danckwerts, (1970) derived the following expression: E = Vl + M2 (2.3) When the gas-side resistance to mass transfer is negligible because of the low gas solubility (i.e., Large HC02), then substitution of equations (2.2) and (2.3) into equation (2.1) gives: NA=-ND=J1 + (k c n ^ K2 k°a Hco, (Pco,-P°co2) (2-4) Equation (2.4) demonstrates that the kinetic coefficient, k2nd, is important in determining absorption/desorption rates. The effect of k2nd on the mass transfer rate can be examined with respect to the Hatta number (M). When M « 1, the rate of reaction is slow and it does not affect the mass transfer rate; consequently E ~ 1. This regime is called "slow reaction regime". In this regime the reaction kinetics have no significant effect on mass transfer, corresponding to the case of physical absorption or desorption. When M » 1, the enhancement factor is 17 approximately M and the mass transfer rate becomes independent of the liquid-film coefficient; i.e., This regime is called the "fast" or "pseudo-first-order regime" (Levenspiel, 1972). In this regime the reaction is fast enough that it enhances the mass transfer but not so fast that it depletes the amine concentration in the boundary layer significantly. Therefore, in this regime, CAM can be considered constant throughout in the boundary layer and can be taken as equal to the bulk liquid concentration, and an apparent or pseudo-first-order rate constant (kapp = k2NDCAivi) can be defined. The pseudo-first-order approximation is usually used in the literature to estimate kappfrom absorption rate data (see Tables 2.1 to 2.6). In these studies, the experimental conditions are set so that 3 < M « Eco, where Eoo is the maximum possible enhancement factor. For a CC>2-amine reaction, Danckwerts (1970) has derived the following expression for E^: where VAM is the amine stochiometric coefficient. As M -> oo, the implication is that the reactions are so fast that chemical equilibrium prevails everywhere. The transport rate becomes independent of the reaction rate and is limited by the diffusion of the liquid reactant to the interface. (2.5) 18 This regime is called the "instantaneous reaction regime" (Levenspiel, 1972). The enhancement factor in this regime represents an upper bound on the potential enhancement of mass transfer. In this regime the interface concentrations of liquid reactants and products may be calculated by solving the chemical equilibrium model. In the transition from the "fast" to "instantaneous regime" (1 « M < Eoo), the reactions occur largely near the interface and the amine concentration within the boundary layer may become significantly depleted. Consequently, the pseudo-first-order approximation given by equation (2.5) is no longer valid and the reaction rates at any point in the boundary layer must be calculated using the actual amine concentration at that point. This requires solving the mass balances (including chemical reaction terms) for each species throughout the liquid film. Analytical solutions for the enhancement factor (like eq. 2.3) are not possible and the equations must be solved numerically. An important objective of this thesis is to determine mass transfer rates in mixed solvents where the components may react at different rates. Therefore, the transition regime cannot be ignored. A set of conditions that gives the pseudo-first-order regime for one component may well give a transition regime for the other. For this reason we have developed a model that covers all reaction regimes and can be applied to interpret both absorption and desorption rate data. It is clear from the above observations that kinetics plays an important role in the CO2 absorption/desorption process using aqueous amines. A thorough understanding of the reaction mechanism and determination of reliable kinetic 19 data are essential for effective design and simulation of such systems. The remainder of this chapter is therefore devoted to developing an understanding of the mechanism of CO2 reactions with aqueous amines and a review of the literature data on absorption and desorption kinetics. Physical properties such as Henry's constant of C02 in amine solution, diffusivities of C02 and other chemical species in the aqueous solution and density and viscosity of amine solutions are equally important for process modeling. An extensive literature review of each physical property is given in Appendices E to H. 2.2 Reaction Mechanisms of C02-Amine-Water Systems 2.2.1 C02 Reactions with Water In aqueous solution, C02 exists in several forms. First it dissolves: C02(g)->C02(aq) and then it reacts with water and hydroxyl ions: C02 +H20<->H2C03 (2.6) C02+OH -<->HCO-3 (2.7) The bicarbonate ions quickly establish equilibrium with carbonate ions: HCO3 + OH' <-> CO*" + H20 (2.8) The direct reaction of C02 with water (reaction 2.6) is very slow (k = 0.026 s"1 at 298 K, Pinsent et al., 1956) and usually neglected in interpreting absorption rate data, as its contribution to mass transfer is insignificant (Rinker et al., 1996; 20 Danckwerts, 1979). This is appropriate because the absorption experiments are generally conducted in equipment with short contact time (xc < 1 s). The C02 reaction with hydroxyl ions (reaction 2.7) is fast and can enhance mass transfer even when the concentration of OH" ions is low. Pinsent et al. (1956) and Read (1975) have reported correlations for the forward rate constant and equilibrium constant of reaction (2.7). Reaction (2.8) is an equilibrium reaction and Read (1975) has reported a correlation for its equilibrium constant. These correlations are presented in Appendix H. 2.2.2 C02 Reactions with Amines Data for C02 reactions with aqueous amines have been interpreted either by the zwitterion mechanism, which was first proposed by Caplow (1968) and later reintroduced by Danckwerts (1979), or by the base-catalyzed hydration reaction mechanism proposed by Donaldson and Nguyen (1980). 2.2.2.1 Zwitterion Mechanism According to this mechanism, C02 in the liquid phase reacts with amine to form a zwitterion intermediate that is subsequently deprotanated by any base in the solution (Danckwerts, 1979; Blauwhoff et al., 1984). Most of the published kinetic data for MEA, DEA and AMP have been analyzed using this mechanism. For the CO2-AMP system, the zwitterion mechanism is described by the following reactions: 21 Zwitterion Formation: C02 +R4NH2 ^.R.NH+COO (2.9) -1 Zwitterion Deprotonation: R4NH^COO"+R4NH2 <-> R4NHCOQ-+R4NH^ (2.10) R4NH^COO" +H2O^R4NHCOO" +H30 (2.11) R4NH+COO" + OH' ^R4NHCOO" + H20 k-4 (2.12) Similar reactions have been reported for MEA and DEA. Danckwerts (1979) combined reactions (2.9) to (2.12) and derived the following expression for the C02 amine reaction: where [C02]e is the concentration of molecular C02 that is in equilibrium with the other ionic and non-ionic species present in the solution. Equation (2.13) suggests two limiting cases (Versteeg and van Swaaij, 1988): (a) when the termk^ /(k2[RNH2] + k3[H20] + k4[0H"])«1 (i.e., the zwitterion deprotonation is much faster then its formation), then the rate of C02 reaction with amine can be expressed in terms of simple second-order kinetics: -MR4NH2]([CQ2]-[CQ2]e) (2.13) co2 22 rCo2 =-MR4NH2]JC02]-[COj].} (2.14) (b)when k2[RNH2]» k3[H20] and k4[OH"]and the term k_,/k2[RNH2]»1 (i.e., the zwitterion deprotonates only with the amine and the zwitterion deprotonation is much slower than its formation), then the overall kinetics are third order: rCo2 =-^[R4NH2]2([C02]-[C02]e) (2.15) K-i The above analysis indicates that the zwitterion mechanism is able to cover the transition region where the reaction order with respect to amine lies between 1 and 2. It may be noted that equation (2.13) is applicable to both absorption and desorption. However, since the majority of research reported in the literature has been focused on absorption into C02-free amine solutions, [C02]e is generally taken as zero. Furthermore, in most cases it was assumed that during the absorption experiments the amine concentration did not change appreciably and the forward reaction dominated (Danckwerts and Sharma, 1966; Sada et al., 1976; Hikita et al., 1977a). As a result, equation (2.14) can be simplified further and the following pseudo-first-order rate expression results (Little et al., 1992a,b): rCo2 = -kappC02 (2.16) where kapp = k2nd[Amine]. In the studies where the full zwitterion mechanism rate expression represented by equation (2.13) was used and combined rate constants (ki, k-|k2/k-i, kik3/ k.-i and k-ikV k.-i) were reported, the methodology used in estimating 23 rate constants was essentially the same as for the pseudo-first-order approach. Again, [C02]ewas set to zero and equation (2.13) could be simplified to: rCo2=-kapp[R4NH2][C02] (2.17) where: k-=f+kk FT kk (218) Ki -i^[R4NH2] + —!-^-[H20] + ^-[OH-] k _^ k _^ k _ ^ In this approach, kapp is estimated by using equation (2.17) for various amine concentrations and, from equation (2.18), the individual and combined rate constants are estimated by non-linear regression (Blauwhoff et al., 1984; Versteeg and van Swaaij, 1988; Xu et al., 1996; Xiao et al., 2000; Ko and Li, 2000). The concentrations of amine and water are generally taken as their initial values whereas the concentration of hydroxyl ions is calculated from the dissociation constants of water and amine. In any given experiment, these concentrations were assumed constant. As shown in the next section, most of the recent studies on the kinetics of C02 with MEA, DEA and AMP have used this approach. For aqueous solutions, the contribution of OH" ions in the deprotonation of zwitterion is generally neglected (Blauwhoff et al., 1984; Versteeg and van Swaaij, 1988; Versteeg and Oyevaar, 1989; Bosch et al., 1990b; Rinker et al., 1996; Xu et al., 1996; and Xiao et al., 2000). This is justified because the concentration of hydroxyl ions in aqueous amine solutions is 2 to 3 orders of magnitude lower than that of water or amine. Furthermore, it falls significantly as a result of C02 absorption. 24 2.2.2.2 Base-Catalyzed Hydration Reaction Mechanism Donaldson and Nguyen (1980) proposed this mechanism for CO2 reaction with tertiary alkanolamines. According to these authors, the tertiary alkanolamines, such as TEA and MDEA, do not react directly with CO2 to form carbamate. Instead they catalyze the CO2 hydration reaction: k K C02 +R1R2R3N + H20 ^R^RgNhT + HC03 (2.19) k_22 The important point to note in this mechanism is that water must be present for the reaction to occur. Most of the data on C02-MDEA systems have been analyzed based on this mechanism. There is general agreement in the literature (see Table 2.3) that the reaction rate is first order with respect to amine and C02. Therefore, the rate expression for reaction (2.19) can be written as: rCo2 =-k22[R1R2R3N][C02] + ^-[R1R2R3NH+][HC03] (2.20) K22 Except for Rinker et al. (1995), almost all studies on the kinetics of the C02-MDEA system have assumed that the forward reaction is the dominant one and that the amine concentration does not change appreciably during the experiment. Consequently, equation (2.20) can be simplified as: rCo2 =-k22[R1R2R3N][C02] = -kapp[C02] (2.21) Equation (2.21) is identical to equation (2.16) and represents a pseudo-first-order rate expression similar to that for MEA, DEA and AMP. Therefore, the methodology described above was also used to estimate the rate constants for 25 the CO2-MDEA system (Barth et al., 1984; Little et al., 1990a; Ko and Li, 2000). Since the CO2 reaction with MDEA is much slower compared to its reaction with MEA, DEA and AMP, the pseudo-first-order approximation may well be justified. To analyze desorption data, the rate expression given by equation (2.20) must be used. Furthermore, the CO2-MDEA reaction is slow, and reaction (2.7), therefore, must also be considered for interpreting absorption/desorption rate data. Recently, Rinker et al. (1995) have shown that the effect of reaction (2.7) on the mass transfer enhancement of the CO2-MDEA system is quite significant. 2.2.3 C02 Reactions with Aqueous Amine Blends When more than one amine is present (e.g., AMP+MEA or MDEA+AMP), the zwitterion of carbamate-forming amines (such as MEA, DEA and AMP) will also be deprotonated by the second amine present in the solution. For the CO2+AMP+MEA system, the reaction is: In the literature (Glasscock et al., 1991, Xiao et al., 2000) the kinetic data for blended amines are treated in the same way as those for single amines and the approximate rate expression given by equation (2.17) is used. The only difference in this case is that the apparent rate constant (kapp) is modified to include reaction (2.22): R4NH2-COO-+R1NH2 <-> R4NHCOCT +R1NH3" (2.22) k 1 1 (2.23) app k.,k2 [R4NH2] + li^[H20]+l^[OH-]+l^[R1NH2] 26 As before, the term involving hydroxyl ions is usually neglected since it is negligible compared to other terms (Xiao et al., 2000) 2.3 Kinetic Data for C02 Absorption/Desorption in Aqueous Amines 2.3.1 Kinetic Data for C02-MEA-Water System MEA is the most commonly used alkanolamine in the gas treating industry and has been studied widely. The literature sources with kinetic data on the C02-MEA reaction are summarized in Table 2.1. Even though different experimental conditions have been used, the agreement between the published data is fairly good. As concluded by Blauwhoff et al. (1984) and later confirmed by Barth et al. (1986), the correlation of Hikita et al. (1977a) fits the data quite well over the temperature range of 288 to 353 K: log10(k2MEA) = 10.99-2152/T where k2lMEA is the second-order rate constant in units of m3 kmol"1 s"1 and T is the temperature in Kelvins. The activation energy from this correlation is 41.2 kJ/mol, which agrees well with the values reported by other investigators. Without exception, an overall second-order reaction (first-order with respect to MEA and first-order with respect to C02) was found regardless of experimental techniques and conditions (Blauwhoff et al., 1984). 27 Table 2.1: Summary of the results reported on the CCvMEA-Water system Reference T CAM k2nd @ 298 K Ea Modef Apparatus & (K) (kmol/ m3) (m3/ kmol s) (kJ/mol) Method Little et al. (1992b) 318, 333 1.0 (3703)++ — g-z stirred cell (A)+++ Alper (1990) 278-298 0.01-1.5 5545 46.7 ps-1 stopped flow (A) Bosch et al. (1990a) 313-323 1.0 - - ps-1 stirred cell (D) Penny & Ritter (1983) 278-303 0.01-0.06 4990 42.2 ps-1 stopped flow (A) Laddha & Danckwerts (1981) 298 0.5-1.7 5720 ps-1 stirred cell (A) Donaldson & Nguyen (1980) 298 0.03-0.08 6000 ps-1 aq. amine membrane (A) Alvarez-Fuster et al. (1980) 293 0.2-2.0 (5750) ps-1 wetted wall column (A) Hikita et al. (1977a) 278-315 0.02-0.18 5868 41.2 ps-1 rapid mixng (A) Sada et al. (1976) 298 0.2-1.9 8400, 7140 - ps-1 laminar jet (A) Leder et al. (1971) 353 7194 39.7 ps-1 Stirred cell (A) Groothius (1966) 298 2.0 6500,5720 ps-1 Stirred cell (A) Contd. 28 Table 2.1: Summary of the results reported on the C02-MEA-Water system (Contd.) Reference T CAM k2nd @ 298 K Ea Model Apparatus & (K) (kmol/ m3) (m3/ kmol s) (kJ/mol) Method Danckwerts & Sharma (1966) 291-308 1.0 7600, 6970 41.8 ps-1 laminar jet (A) Clarke (1964) 298 1.6-4.8 7500 - ps-1 laminar jet (A) Astarita (1961) 295 0.25-2.0 (6443) - ps-1 laminar jet (A) Jensen et al. (1954a) 291 0.1, 0.2 (6103) - - competition method with 0.1 and 0.2 M NaOH (A) + ps-1 = pseudo-first-order, ps-1-z = pseudo-first-order with zwitterion mechanism, g-z = general model with zwitterion mechanism ++ Values in the bracket indicate that k2nd was estimated using Ea= 41 kJ/moie +++ A = absorption, D = desorption Although the data generally agree well with regard to reaction order and activation energy, it should be noted that most data were acquired under absorption conditions in the temperature range of 293 to 308 K. These data were mostly interpreted using simplified pseudo-first-order models. The more recent data of Little et al. (1992b) were analyzed using a rigorous model based on the zwitterion mechanism but their study was again limited to absorption. An important conclusion from this study was that, for fast reversible reactions like 29 those in the CO2-MEA system, rigorous modeling is the only way to estimate the reaction rate constants from absorption experiments. There has been no significant work on the kinetics of the CO2-MEA reaction under desorption conditions. Bosch et al. (1990a) provided the only published data for C02 desorption rates from rich MEA solutions. However, their data were taken at low temperatures in the range of 313 to 323 K. These authors used pseudo-first-order kinetics with a simplified equilibrium model to predict their CO2 desorption rates. However, they were unsuccessful in interpreting their data. 2.3.2 Kinetic Data for C02-DEA-Water System DEA is the second most common alkanolamine used for bulk CO2 removal. Because of its wide use, the literature on C02-absorption is extensive. The principal results, mostly from Blauwhoff et al. (1984) and Rinker et al. (1996), are summarized in Table 2.2. There is general disagreement on the order and rate of reaction with respect to DEA. The reaction order for C02 is generally accepted to be one but the order for the amine varies between 1 and 2. Blauwhoff et al. (1984) explained some of the discrepancies in the reported results by means of the zwitterion mechanism that includes all bases (i.e., [Amine], [H20] and [OH-1]) for zwitterion deprotonation. However, recently Rinker et al. (1996) using a rigorous approach for data interpretation were unable to find any significant contribution of hydroxyl ions and water in zwitterion deprotonation. 30 Table 2.2: Summary of results reported on the C02-DEA+Water system Reference T (K) CAM (kmol /m3) k2nd@ 298 K (rrrVkmol s) Order w.r.t. Amine Ea (kJ/mol) Modef Apparatus & Method Rinker et al. (1996) 293-343 0.25-2.8 4089 1-2 14.1 9-z laminar jet (A)+++ Davis & Sandall (1993) 293-313 0.25-2.0 351 (no water) 1-2 - ps-1-z Wetted sphere (A) Little et al. (1992b) sos-sis 0.2-4.0 (1454)++ 1-2 - g-z stirred cell (A) Crooks & Donellan 298 0.1-1.0 2250 2 - - stopped flow (1989) (A) Versteeg & Oyevaar (1989) 298 0.09-4.40 3170 1-2 ps-1 -z stirred cell (A) Versteeg & van Swaaij (1988) 298 5790 1-2 ps-1-z stirred cell (A) Barth et al. (1986) 298 0.02 275 1 ps-1 stopped flow (A) Blauwhoff et al. (1984) 298 0.5-2.3 5800 1-2 ps-1 -z stirred cell (A) Blanc & Demarais (1984) 293-333 0.01-4.0 656 1 43.5 ps-1 wetted wall column (A) Contd. 31 Table 2.2: Summary of results reported on the COrDEA+Water system (Contd.) Reference T (K) CAM (kmol k2na@298 K (m3/kmol s) Order w.r.t. Amine Ea (kJ/mol) Model Apparatus & Method Laddha & Danckwerts (1982) 284 0.5-2.0 (1492) 1-2 ps-1-z stirred cell (A) Laddha & Danckwerts (1981) 298 0.46-2.88 1410 1-2 ps-1-z stirred cell (A) Donaldson & Nguyen (1980) 298 0.03-0.09 1400 1 ps-1 aq. Amine membrane (A) Alvarez-Fuster et al. (1980) 293 0.25 0.82 (3006) 2 ps-1 Wetted wall column (A) Hikita et al. (1977a) 278-313 0.17-0.72 3132 2 ps-1 rapid mixing (A) Coldrey & Harris (1976) 292 0.1-1.0 (1636) 1 ps-1 rapid mixing with 0.002-0.005 M NaOH (A) Sada et al. (1976) 298 0.25-1.92 1340 1 - ps-1 laminar jet (A) Leder (1971) 353 - (7870) 1 - ps-1 stirred cell (A) Groothius (1966) 298 2.0 1300 1 stirred cell (A) Contd. 32 Table 2.2: (Contd.) Summary of results reported on the C02-DEA+Water system Reference T CAM k2na@298 K Order Ea (K) (kmol (m3/kmols) w-rt (kJ/mol) /m3) Amine Model Apparatus & Method Danckwerts 308 &Sharma (1966) Sharma (1964) Jorgensen (1956) Jensen et al. (1954a) van Krevelen & Hoftizer (1948) 291 Jorgensen 273 (1956) 291 291 292-329 1.0 1.0 0.1-0.3 0.2-0.3 0.1-0.2 0.05-3.0 (1317) (1648) (3381) (6595) (8360) 650 1 ps-1 ps-1 laminar jet (A) laminar jet (A) competitive reaction with 0.1, 0.2, 0.3 M NaOH (A) competitive reaction with 0.2, 0.3 M NaOH (A) competitive reaction with 0.1, 0.2 M NaOH (A) packed column (A) + ps-1 = pseudo-first-order, ps-1 -z .= pseudo-first-order with zwitterion mechanism, g-z = general model with zwitterion mechanism ++ Values in the bracket indicate that k2nd was estimated using Ea= 14.1 kJ/mole +++ A = absorption, D = desorption 33 Crooks and Donnellan (1989) suggested a different mechanism for the reaction of CO2 with DEA by proposing a single step termolecular reaction. In this mechanism, DEA is postulated to react simultaneously with one molecule of C02 and one molecule of base. Their mechanism leads to an overall reaction rate, which has the same form as, that of the second limiting case of the zwitterion mechanism discussed above. However, as pointed by Little et al. (1992b) and Rinker et al. (1996), this mechanism does not explain the fractional orders with respect to DEA concentration observed in non-aqueous solvents (Sada et al., 1985; Versteeg and van Swaaij, 1988; Davis and Sandall, 1993). Nevertheless, there is no plausible explanation in the literature as to why a proton transfer step such as base proton extraction would be rate limiting. Since the zwitterion mechanism adequately covers the varying reaction order, it is fairly universally used in the literature (see Table 2.2) to interpret C02 absorption data. Perhaps a different way of confirming the applicability of this mechanism would be to conduct experiments under desorption conditions and see if the kinetic data acquired under absorption conditions and using a zwitterion mechanism, could be used to predict desorption rates. No such study has been reported in the literature. Critchfield and Rochelle (1988) provided the only study that reports C02 desorption rates from rich DEA solutions but their experiments were performed only at 298 K. 2.3.3 Kinetic Data for C02-MDEA-Water System MDEA is the most widely used tertiary amine with major applications in the selective removal of H2S from gases containing both C02 and H2S. The kinetics 34 of the C02 reaction with MDEA must be known to estimate C02 pickup in the absorber and in designing systems with mixed solvents where MDEA is blended with a fast reacting primary or secondary amine. Many studies dealing with the kinetics of C02-MDEA reactions have been reported in the literature. Important results from these studies are given in Table 2.3. All these studies confirm the base catalyzed C02 hydration reaction mechanism proposed by Donaldson and Nguyen (1980) for the C02-tertiary amine reaction. The reaction order with respect to both amine and C02 has always been found as one. There are some discrepancies regarding the activation energy and the reported values of second order rate constants. The second-order rate constants at 298 K vary by a factor of 2 depending on the authors (see Table 2.3). In most cases, pseudo-first order models were used to interpret the absorption rate data. Recently, Rinker et al. (1996) used the rigorous model to estimate kinetic coefficients and pointed out that in this system, the C02 reaction with hydroxyl ions must be taken into account. C02 desorption from rich MDEA solutions has been studied by Critchfield and Rocheile (1987), Bosch et al. (1990a) and Xu et al. (1995) in the temperature range of 303 to 343 K. In all these studies, the experimental desorption rates are in good agreement with those predicted by simple models for mass transfer with fast pseudo-first-order reaction. These results are important as they indicate that C02 desorption from MDEA solutions is controlled by mass transfer with fast reaction and the rate of desorption could be predicted based on absorption kinetic data. 35 Table 2.3: Summary of results reported on the C02-MDEA-Water system Reference T CAM k2nd @298 K Ea Model+ Apparatus & (K) (kmol/ m3) (m3/kmol s) (kJ/mol) Method Ko&Li (2000) SOS-SIS 1.0-2.5 5.41 44.9 ps-1 wetted wall column (A)++ Pacheco et al. (2000) 298-373 2.9-4.3 2.5 49.0 ps-1 wetted wall column (A) Pani et al. (1997) 296-343 0.8-4.4 5.2 44.0 ps-1 stirred cell (A) Rinker et al. (1995) 293-342 0.8-2.5 6.2 38.0 9 wetted sphere (A) Xu et al. (1995) 313-343 2.6-3.9 - - ps-1 Packed column (D) Rangwala etal. (1992) 298-333 0.8 -2.5 4.4 48.0 ps-1 stirred cell (A) Bosch et al. (1990) 298-323 1.0 - - ps-1 stirred cell (D) Little et al. (1990a) 298 0.2-2.7 5.5 - ps-1 stirred cell (A) Toman & Rochelle (1989) 298-308 4.3 5.5 ps-1 stirred cell (A) Tomcej & Otto (1989) 298-348 1.7-3.5 5.4 42 ps-1 wetted sphere (A) Contd. 36 Table 2.3: Summary of results reported on the C02-MDEA-Water system (Contd.) Reference T CAM k2nd @298 K Ea Model Apparatus & (K) (kmol/ m3) (m3/kmol s) (kJ/mol) Method Versteeg &van Swaaij (1988) 293-333 0.2-2.4 4.4 42 ps-1 stirred cell (A) Haimour et al. (1987) 288-308 0.9-1.7 2.4 72 ps-1 stirred cell (A) Critchfield & Rochelle (1987) 282-350 1.7 2.5 56 ps-1 stirred cell (A) Yu et al. (1985) 313-333 0.2-2.5 4.8 39 ps-1 stirred cell (A) Blauwoff et al. (1984) 298 0.5-1.6 4.8 - ps-1 stirred cell (A) Barth et al. (1984) 298 0.02-0.2 3.2 - ps-1 stopped flow (A) + ps-1 = pseudo-first-order, g = general model ++ A = absorption, D = desorption 2.3.4 Kinetic Data for C02-AMP-Water System AMP, a relatively new amine, is primarily used for the selective removal of H2S from natural gas (Goldstein, 1986). It is a primary amine in which the amino group is attached to the tertiary carbon atom. The reaction rates of C02 with AMP are significantly higher than those with MDEA and considerably less than those with MEA. Like MEA and DEA, AMP reacts directly with C02 to form carbamate, but its carbamate is highly unstable and quickly hydrolyzes to give free amine 37 and bicarbonate as evident from the value of its carbamate stability constant (Kc) given in Table 2.4 (Sartori et al., 1983). As a result, AMP can absorb up to 1 mole of C02/mole of amine as opposed to 0.5 mole of C02/mole of amine for MEA and DEA. Moreover, the heat of reaction for C02-AMP is less than those of C02-MEA and C02-DEA. It, therefore, requires less energy for regeneration. A number of studies on the kinetics of the C02-AMP reactions have been reported in the literature. Important results and experimental conditions from these studies are summarized in Table 2.5. A plot of the second-order rate constant of the C02-AMP reaction as a function of temperature as reported by various investigators is shown in Figure 2.1. Excluding, the data of Chakraborty (1986), who reported a k2nd value of 100 nvVkmol s at 313 K and Bosch et al. (1990b), who reported this value to be 10,000 m3/kmol s at 298 K, the agreement among the reported data is satisfactory. However, the values of k2nd at 298 K reported by various authors differ quite significantly (see Table 2.5). The activation energy obtained by regressing the reported data is 36.13 kJ/mole. It is interesting to note that the activation energy of 41.7 kJ/mole calculated from recently published data of Saha et al. (1995) is exactly the same as that reported by Alper (1990) even though the latter author used a different experimental technique. Recently, Xu et al. (1996) and Messaudi and Sada (1996) have reported significantly different activation energies, which are 24.6 and 51.5 kJ/mole, respectively. 38 Table 2.4: Carbamate stability constants for MEA, DEA and AMP by C13-NMR (Sartori and Savage, 1983) Amine Kc(m3/kmol)at313K MEA 12.5 DEA 2.0 AMP < 0.1 x Messaoudi and Sada (1996) oShou Xu (1996) A Sana et al. (1995) • Al per (1990) X Bosch et al. (1990) +Yih and Shen (1988) • Chakraborty et al. (1986) A Sharma (1965) 3.0 3.2 3.4 3.6 3.8 1000/T(1/K) Figure 2.1: Comparison of second-order rate constant for C02-AMP reaction. 10000 1000 100 -i n o A A ° o J I L_ 39 Table 2.5: Summary of results reported on the C02-AMP-Water system Reference T (K) CAM (kmol/ m3) k2nd@298 K (m3/ kmol s) Order w.r.t. Amine Ea (kJ/mol) Modef Apparatus & Method Messaudi and Sada (1996) 293-313 0.5-2.0 271.8 1 51.5 ps-1 stirred cell (A)++ Xu et al. (1996) 288-318 0.25-3.5 810.4 1.32-1.50 24.3 ps-1-z stirred cell (A) Saha et al. (1995) 294-318 0.5-0.2 563.5 1 41.7 ps-1 wetted wall column (A) Alper (1990) 278-298 0.5-2.0 502.0 1.14-1.15 41.7 ps-1 stopped flow (A) Bosch et al. (1990b) 298 0.2-2.4 10,000 1 - g-z stirred cell (A) Yih &Shen (1988) 313 0.26-3.0 1270.0 (at 313 K) 1 ps-1 wetted wall column (A) Chakraborty etal. (1986) 313 0.5-1.0 100.0 (at 313 K) 1 - ps-1 PD cell (A) Sharma (1965) 298 0.2-2.0 1048.0 1 - ps-1 stirred cell (A) + ps-1 = pseudo-first-order, ps-1-z = pseudo-first-order with zwitterion mechanism, g-z = general model with zwitterion mechanism h+ A = absorption, D = desorption 40 The reaction order with respect to C02 was generally found to be one (Alper, 1990; Saha et al. 1995; Xu et al., 1996). However, the order with respect to amine has been reported to be slightly more than one (Alper, 1990; Xu et al., 1996). This suggests that the zwitterion mechanism should describe the data better. There is some disagreement amongst the researchers about the right mechanism. The majority (Sartori and Savage, 1983; Alper, 1990; Bosch et al., 1990b; Xu et al., 1996) is of the view that C02 reacts with AMP to form carbamate by the zwitterion mechanism and that the carbamate is highly unstable, quickly hydrolyzing to bicarbonate. However, according to Yih and Shen (1988), the formation of carbamate in AMP is inhibited due to the bulkiness of the group attached to the tertiary carbon atom and the equilibrium favors bicarbonate formation via a zwitterion intermediate according to the following reaction: R4NH +COCT +H20 <-> HC03 +R4NH; Chakraborty et al. (1996) proposed yet another mechanism according to which AMP, like MDEA, does not participate directly in the reaction and acts only as a base catalyst for C02 hydration reaction. Finally, the data are limited to temperatures typical of absorbers (i.e., 298 to 318 K). No desorption data have been reported. Except for Bosch et al. (1990b), the experimental data have been analyzed using simple pseudo-first-order models. More data are required so that discrepancies reported in the literature regarding the mechanism and the reaction order with respect to amine can be resolved. The data should be collected under both absorber and desorber 41 conditions. To analyze the data, a comprehensive model that involves all possible reactions should be developed so that other mechanisms proposed in the literature can be evaluated. 2.3.5 Kinetic Data for Amine Blends Primary and secondary amines (such as MEA and DEA) are fast reacting amines that form stable carbamates by direct reaction with CO2. The tertiary amines, such as MDEA, on the other hand, are slow reacting amines and do not form carbamates. Hindered amines, such as AMP, have moderate reaction rates and can form carbamates, but the latter are highly unstable and quickly hydrolyze to give bicarbonate ions and free amine. Consequently, the regeneration energy requirements for MEA and DEA are quite high compared to those for MDEA and AMP. Blending of primary or secondary amines with tertiary or hindered amines is therefore an attractive option for designing less energy intensive processes for bulk CO2 removal. An additional advantage of blended amines is that, by varying the composition of the amine blend, the selectivity towards H2S can be adjusted. Although Chakraborty et al. (1986) introduced this concept over 15 years ago; little fundamental data are available for the rate of absorption/desorption of C02 in amine blends. Important kinetic studies involving blends of MDEA with MEA or DEA and blends of AMP with MEA or DEA are summarized in the next two sections. 42 2.3.5.1 Kinetic Data for Aqueous Blends of MEA+MDEA and DEA+MDEA Previous work on C02 absorption/desorption in MEA+MDEA and DEA+MDEA systems is summarized in Table 2.6. Most of the data reported are for total amine concentrations between 1 and 3 kmol/m3 and temperatures between 298 and 313 K. Critchfield and Rochelle (1988,1987) and Glasscock et al. (1991) have reported both C02 absorption and desorption rate data but they are limited to low temperatures (i.e., 288-313 K). The effect of tertiary amines on the reaction rates of primary or secondary amines has generally been taken into account by including an additional reaction where the zwitterion is also deprotonated by the tertiary amine. The results of Glasscock et al. (1991) and Hagewiesche et al. (1995) are important, as they have used rigorous diffusion-reaction and equilibrium models to interpret their data. The available data on C02 absorption/desorption from these blends are sparse and more data, especially near stripper temperatures, are need. 2.3.5.2 Kinetic Data for Aqueous Blends of MEA+AMP and DEA+AMP C02- MEA+AMP-Water System: In recent years, there has been a considerable interest in capturing and sequestering C02 from industrial point sources such as flue gases from power plants. Because of its high reaction rates with C02, MEA is generally considered as the best solvent for this application. However, there are many technological problems associated with MEA-based process that require further 43 Table 2.6: Summary of results reported on C02-MEA+MDEA-Water and C02-DEA+lv1DEA-Water systems Reference Blend T (K) Total Amine Cone. (kmol/m3) Modef Apparatus & Method Hagewiesche etal. (1995) MEA+MDEA 313 2.6-3.0 g laminar jet (A)++ Rangwala et al. (1992) MEA+MDEA 293 2.0-3.5 ps-1 stirred cell (A) Glasscock et al. (1991) MEA+MDEA DEA+MDEA 288-313 0.0-3.0 g-z stirred cell (A/D) Critchfield & Rochelle (1988) DEA+MDEA 298 2.0 ps-1 stirred cell (A/D) Critchfield & Rochelle (1987) MEA+MDEA 304 2.0 ps-1 stirred cell (A/D) + ps-1 = pseudo-first-order, ps-1-z = pseudo-first-order with zwitterion mechanism, g=general model, g-z = general model with zwitterion mechanism + A = absorption, D = desorption considerations. MEA is highly corrosive to packing and other equipment and its concentration in the aqueous solution cannot exceed 15-17%. Consequently, very high circulation rates are required. Since MEA carbamate is highly stable, the energy requirements for regenerating the MEA are high. MEA is also highly susceptible to degradation, especially when gas streams contain small quantities of oxygen or fly ash as observed by our own research group and recently by Supapetal. (2001). 44 The above problems can be effectively addressed by designing a novel solvent blend containing MEA and AMP. The AMP has appreciable reactivity with C02, but adding small quantities of MEA can further enhance it. Being the major constituent of the blend, the higher cyclic capacity and lower heat of reaction of AMP may significantly reduce the regeneration heat requirement of the process. Until the end of 1999, when the experimental work of this dissertation was being completed, no study on either absorption or desorption involving this system had been reported. Very recently, Xiao et al. (2000) have published absorption kinetic data for this system. They studied C02 absorption in aqueous mixtures of AMP and MEA in the temperature range of 303 and 313 K using a laboratory wetted wall column. The concentration of AMP was set at 1.7 and 1.5 kmol/m3 and the concentration of MEA was varied from 0.1 to 0.4 kmol/m3. A hybrid reaction rate model (consisting of a first-order reaction for MEA and a zwitterion mechanism for AMP) was used to model the data. Overall pseudo-first-order and corresponding second-order rate constants were reported. More theoretical and experimental work, especially under desorption conditions, is required. C02-DEA+AMP-Water System: No absorption or desorption studies on this system have been reported. This is probably because the reactivities of DEA and AMP with C02 are quite similar and apparently no advantage can be gained by blending these two amines. It will, however, be interesting to see if the capacity and desorption rates 45 of DEA can be increased by adding AMP as the latter does not form a stable carbamate. 2.4 Research Needs The major areas needing further research on C02 absorption/desorption in the eight aqueous amine systems discussed above may be summarized as follows: 1. Most of the available data have been obtained under absorption conditions in the temperature range of 298 to 313 K using lean amine solutions representing typical conditions near the top of the absorption columns. The kinetic data that represent the absorber bottom and stripper conditions (high temperature and high loading) are needed as a function of C02 loading. 2. The studies on absorption and desorption have been done in isolation and there has never been an attempt to investigate if the data collected under absorber conditions can be utilized to predict desorption rates. 3. The experimental data on reaction kinetics have been mostly analyzed assuming simplified pseudo-first-order kinetic models with pseudo-first-order or corresponding second-order rate coefficients being reported. This assumption greatly simplifies the complex mathematics, which govern the reaction-diffusion process in the mass transfer boundary layer. However, the disadvantage of using this approach is that it does not represent the actual concentration profiles of different chemical species in the boundary 46 layer near the gas-liquid interface and the kinetic data obtained may not be reliable. In real situations, it is rare for a single reaction regime to exist throughout the column. Moreover, this approach does not provide any insight into which reactions enhance mass transfer and which reactions do not. A comprehensive model is therefore needed which describes the diffusion-reaction process in the boundary layer and that takes into account all possible reactions under absorption and desorption conditions. 4. The kinetic data for MEA and MDEA from various sources are in good agreement and no further data under absorption conditions are needed. However, it is important to investigate if these data can be used to predict desorption rates. 5. There are discrepancies in the literature regarding the reaction order with respect to amine for the CO2-DEA and C02-AMP systems. More data, especially under stripping conditions, are needed to reconcile these differences. 6. The kinetic data on amine blends are scarce and more data under both absorption and desorption conditions are needed. Only a single study on C02 absorption in an MEA+AMP blend has been reported in the literature, and no absorption or desorption involving AMP+DEA have been published. It would be nteresting to investigate (from both theoretical and experimental perspectives) if the addition of MEA or DEA can substantially enhance AMP reaction rates. If this occurs, such blends would become potential solvents for bulk C02 removal from flue gases. 47 CHAPTER 3 EXPERIMENTAL APPARATUS AND METHODS 3.1 Overview In absorption/desorption studies on CCValkanolamine systems, kinetic data are typically obtained by conducting experiments using special types of laboratory contactors. The most commonly used contactors are stirred cells, single sphere units, wetted wall columns and laminar jets. These contactors are designed in such a way that the interfacial area is accurately known and the hydrodynamics are well defined so that physical mass transfer coefficients can easily be obtained from first principles. The advantage of knowing the interfacial area and mass transfer coefficient in advance is that the reaction kinetics and the mass transfer can be decoupled and the measured rate data can be easily interpreted to study the kinetics of gas-liquid reactions. In this chapter, a new laboratory gas-liquid contactor is described. The apparatus is, in principle, similar to the wetted sphere units used in many previous studies on CCValkanolamine kinetics (Davidson and Cullen, 1957; Savage et al., 1980; Tseng et al., 1988; Al-Ghawas et al., 1989; Tomcej and Otto, 1989; Richard and Sandall, 1993; Tamimi et al., 1994a,b; Dehouche and Lieto, 1995; Rinker et al., 1995). However, in this work the conventional design has been improved by replacing the full sphere by a hemisphere and by designing a better liquid feed and receiver system. 48 3.2 Hemispherical Contactor Figure 3.1 depicts the main features of the hemispherical contactor. All parts of the apparatus in contact with the amine solution were constructed from stainless steel, Teflon, or Pyrex glass. The hemispherical contacting surface was constructed from a 76-mm diameter solid, stainless steel sphere. The upper half of this sphere was machined to have shiny and smooth surface and the lower half was machined to fit into a 79-mm ID stainless steel funnel (included angle) leaving a uniform annular space of 1.5 mm. The hemisphere was centered by three equally spaced stainless steel inserts, which were fastened on the inside of the funnel by metal screws inserted from the outside. The hemispherical assembly was enclosed in two concentric cylinders that formed the absorption/desorption chamber. The total height of the chamber was 457 mm. The inner cylinder was made from 19-mm thick Pyrex glass that had a conical shape, 152 mm ID at the bottom and 76 mm ID at the top. The outer cylinder was made from two cylindrical parts of 203 mm ID. The upper part was 152 mm long and constructed from 13-mm thick stainless steel pipe. The lower part was 305 mm long and was made from 13-mm thick QVF Corning glass. The purpose of using Corning glass in the lower part was to permit viewing the film during the experiment. To prevent heat losses, the upper part of the outer cylinder was wrapped in glass wool insulation and the lower part was wrapped in thermolyne electrical heating tape. 49 I Gos In V o -1 Hemispherical Surface (SS 316) 2 Receiver Tube (SS 316) 3 Flexible Hose (SS 316) 4 Inner Cylinder (Pyrex Glass) 5 Top Outer Cylinder (SS 316) 6 Bottom Outer Cylinder (QVF Glass 7 Threaded Rod (Carbon Steel) 8 Top Flange (SS 316) 9 Bottom Flange (SS 316) t 1 Li qui d I n Figure 3.1 Schematic Drawing of the Absorption/Desorption with the Hemispherical Contactor 50 The seal on both ends of the absorption/desorption chamber was provided by two 12-mm thick and 305 mm diameter stainless steel flanges and Teflon O-rings between the flanges and the glass. The top flange was compressed on the glass cylinders by means of six equally spaced 10 mm diameter bolts attached to the bottom flange. To avoid breakage of the glass cylinders due to uneven thermal expansion of the glass and metal components, the inner cylinder was connected to the top flange by a flexible stainless steel hose. The hemispherical assembly with the attached receiver tube was installed at the center of the inner cylinder at 228 mm from the bottom flange. The receiver tube was sealed to the bottom flange by means of a compression fitting. The feed liquid is passed through the stainless steel hemisphere by means of a 4-mm ID stainless steel tube and is discharged at the pole of the hemisphere (see Figure 3.1). It flows as a well-defined liquid film over the surface of the steel hemisphere and is collected by the stainless steel funnel at the base of the hemisphere. The effective mass transfer contact area provided by the hemisphere is 98.03 cm2. The liquid is discharged to the storage tank via a 10 mm ID stainless steel receiver tube soldered to the base of the funnel. The liquid level in the funnel is maintained by means of a stainless steel constant head tank (76 mm ID and 76 mm height). The constant head tank is moved vertically up or down on a finely threaded rod by means of a jacking mechanism. The receiver tube is connected to a constant head tank by a 4 mm ID stainless steel flexible Teflon tubing. 51 In Figure 3.1, the gas is shown to enter the top of the chamber and exit at the bottom. However, there are also provisions for the gas to enter at the bottom and exit at the top. The down flow of gas is generally preferred to prevent surface rippling due to opposite gas and liquid flow directions. 3.3 Experimental Setup and Procedure Figure 3.2 shows the flow diagram of the experimental equipment. The amine solution was stored in a 40 L stainless steel tank (item number 9 in Figure 3.2) equipped with a stainless steel diffuser for mixing and loading the amine solution with CO2 in case of desorption runs. The amine solution and the gas stream were heated to the desired temperature in a carbon steel tank (item number 10 in Figure 3.2), which had the same dimensions as those of the amine tank. This tank was equipped with a diffuser, an immersion heater and two stainless steel heating coils, one for heating the amine solution and the other for heating the air that flowed through the annular space in the absorption/desorption chamber. Both coils were made from 12 m long, 4-mm ID stainless steel tubing. Inside the tank, these coils were immersed in the water that filled approximately half of the heating tank. The water temperature was raised to the experiment temperature by means of an immersion heater controlled using a PID type temperature controller (Omega CN76000). 52 CL ZI -*—> 0 CO "ro -4—» C 0 E i-0 CL X LU CN CO 0 i_ D CT) 1/1 a. : l V The experimental setup shown in Figure 3.2 was used for both the absorption and desorption experiments. The procedure in both cases was essentially the same. In the absorption experiments, pure CO2 or N20 was used, whereas in the desorption experiments pure N2 was used. All absorption experiments were done under atmospheric pressure. Desorption experiments below 343 K were performed at atmospheric pressure and those above 343 K were carried out at 203 kPa. Higher system pressure was required to prevent flashing of the feed solution due to boiling and excessive C02 desorption. The experimental setup was regularly tested for leaks by pressurizing it with N2 up to 250 kPa and monitoring the system pressure for several hours. The apparatus always held the pressure extremely well. In a typical absorption or desorption experiment, the amine tank was filled with approximately 20 L of freshly prepared solution and then sealed. The solution was circulated through the hemispherical unit kept under pressure by N2 from a gas cylinder. The solution was initially fed at high flow rates to ensure complete wetting of the hemisphere. When the liquid film was stabilized, the flow rate was reduced to the desired value. The solution flow rate was measured using a Cole-Palmer variable area flow meter (model H03229-31) and controlled manually by a precision needle valve. For each run, the flow rate was checked using the bypass line. The bypass line was also used to withdraw liquid samples (10 mL) to determine C02 loadings as described in Appendix A. Typically, the liquid flow rate ranged between 1.5 to 3.5 mL/s and the exposure time varied between approximately 0.3 to 0.6 s. The solution temperature was measured 54 using a 1.5-mm diameter, K type thermocouple, installed at the center of the liquid feed point. The feed solution temperature was maintained at the desired value by regulating the power input to the immersion heater attached to the heating tank and the electrical tape wrapped around the liquid feed line. After passing over the hemisphere, the liquid flowed into the receiving funnel where it was maintained at a desired level by means of a constant level device. Prior to entering the constant head tank (item 3 in Figure 3.2), the solution temperature was brought down to the ambient temperature by passing it through a water-cooled condenser (item 2). At this point, a liquid sample could also be withdrawn for analysis. From the constant head tank, the liquid ran down to a Pyrex glass drain tank (50 mm ID and 300 mm height). In the drain tank, the liquid level was maintained at a preset value by adjusting its outflow by means of a precision needle valve. The solution from the drain tank was collected in a storage tank for reuse. The CO2 content (loading) in the amine solution was determined using Gastec tubes. The detailed procedure for this method is given in Appendix A. The gas entering the absorption/desorption chamber was saturated with water vapor at the experiment temperature by bubbling it through the water in the heating tank (item 10). The stirring action of the bubbling gas also helped maintain a uniform water temperature in the heating tank. Saturating the gas feed with water vapor reduced water losses from the liquid film and heat imbalances in the chamber. At all times during an experiment, the water in the heating tank was kept saturated with the feed gas. 55 Gases were supplied from gas cylinders with a discharge pressure maintained at 35 kPa above the system pressure. The gas discharge pressure was controlled by means of low-pressure precision regulators (Skeans Model P-16-04-L00). Depending on the mode of operation, solution concentration and amine type, the flow rate of the gases ranged between 200-1200 mL/min. Typically, the gas stream entered the top of the chamber and exited at the bottom. After leaving the chamber the gas stream was diluted with a CO2 free N2 gas. The flow rate of dilution N2 typically ranged between 200-1000 mL/min. The gas mixture was cooled to ambient temperature by passing it through a water-cooled condenser. The cooled gas was then flashed into a Pyrex glass tank (100 mm ID and 300 mm height), to separate the gas from the water condensate. The moisture in the gas stream was further lowered by passing it through an acrylic tube (50 mm ID and 30 mm height) filled with anhydrous calcium sulfate beads. The dry gas mixture was then analyzed using an infrared C02 analyzer (Model NOVA 300, NOVA Analytical Systems Inc., Hamilton, ON). In absorption experiments with N20, the gas samples were collected in 1 L Tedlar gas sampling bags and the gas composition analyzed using a GC (Shimadzu Model GC 8A). In some runs, the gas phase C02 concentration was also measured using the GC to double check the accuracy of the infrared analyzer. The calibrations and procedure for gas phase C02 and N20 measurements using the infrared analyzer and GC are given in Appendix B. The feed and dilution gas flow rates were measured using Cole-Palmer (Model GFM 171) and Brooks (Model 5700) mass flow meters, respectively. The 56 gas flow rates were maintained at the desired value by means of needle valves installed upstream of the mass flow meters. The mass flow meters were calibrated using a soap film meter (see Appendix B). The temperatures of the gas stream in the heating tank and the absorption/desorption chamber were measured by type K thermocouples. In order to maintain the gas stream at the desired temperature, electrical heating tapes were wrapped around the gas feed line and the outer glass wall of the chamber. The power input to these tapes was regulated using two independent PID temperature controllers (Omega CN76000). All temperatures were controlled to within ±0.5 °C. The pressures in the chamber and heating tank were measured using Omega (Model PX-202-030GV) pressure transducers. In the desorption experiments, the system pressure was controlled manually by a precision needle valve installed on the gas line downstream of the flash tank. The transducers were precisely calibrated using a mercury manometer (see Appendix B). 3.4 Data Acquisition and Calibration The signals from the thermocouples, pressure transducers, mass flow meters and CO2 analyzer were read and recorded by means of an IBM 486 PC equipped with a CIO-DAS08 data acquisition board. The number of analog input channels in the A/D board were expanded using two externally mounted CIO-EXP16 signal-conditioning accessory boards, each with 16 input channels. The output signals from all sensors were calibrated as described in Appendix B. The 57 calibration equations were included in the data acquisition software written in Visual Basic specifically for this application. During data acquisition, the software also displayed (on a real time basis) all measurements as well as the absorption and desorption rates. The absorption or desorption rates at any given time were calculated by the program from the feed and dilution gas flow rates and the exit gas composition using the following mass balance equations: FNd" x % C02 in Exit Gas NA(mmol/s) = Fco ^ - (3.1) c°2 100-%CO2 in Exit Gas) (FN + FND" x % C02 in Exit Gas ND(mmol/s) = ^ ^ - (3.2) (100-%CO2 inExitGas) where, NA and No denote the absorption and desorption rates in units of mmol/s, and Fcc>2, F^ed and F^' denote the flow rates of feed C02, feed N2, and dilution N2 respectively in units of mmol/s. Figure 3.3 and 3.4 depict typical plots of exit gas compositions and mass transfer rates for C02 absorption and desorption for aqueous MEA. In the absorption experiments (Figure 3.3), the C02 concentration initially drops very rapidly and then increases to a steady value. This behavior results from the fact that the solution is fed initially at a high flow rate to ensure complete wetting of the hemisphere. This causes it to absorb more C02 and the exit C02 concentration consequently falls. However, when the liquid film stabilizes, its flow rate is 58 0 _J I I I I I I I I I I I I I I I I I I I I I I l_ _J L_l I I I 1 I 1_ 0 0 10 20 30 40 Time (min) Figure 3.3: Typical computer output from an absorption experiment (MEA = 14.3 wt%, T = 303 K, PT = 101.2 kPa, QL = 2.1 mL/s) 50 3 40 T—i—i—i—i—i—i—I—I—|—p—i—i—i—I—I—i—i—i—|—I—I—I—I—I—I—i—i—i—I—i—i—i—i—i—i—i—i—r 10 8 o E 6 w co CD * x 4 LU H 2 CM o o 0 I ' i • i • ' I I I I I i i i I I I I i I I I I 1 I I I I i I i i i i i i I i i I Q 0 10 20 30 40 Time (min) Figure 3.4: Typical computer output from a desorption experiment:(MEA = 20 wt%, a = 0.279 mol/mol, T = 378 K, PT = 202 kPa, QL = 2.2 mL/s) 59 reduced to the desired rate and the exit gas composition comes back to a steady value. An experiment was considered complete when the C02 concentration in the exit gas remained constant for at least 10 minutes. In the desorption experiments (Figure 3.4), the concentration of C02 in the stripping gas steadily increases until it reaches a constant value, The gradual increase in C02 concentration in the stripping gas occurs because at the start of a desorption run all lines are purged with pure N2 and it takes a while for the C02 to diffuse into N2 gas filling the lines downstream of the absorption/desorption chamber. The reproducibility of the absorption and desorption experiments was always within 2%. 3.5 Chemicals All gases had a purity of 99.9% and were supplied by Prax Air Company. The amines had a purity of more than 99% and were supplied by Sigma-Aldrich, van-Waters & Rogers and Travis Chemicals. The aqueous amine solutions were prepared using distilled water. 60 CHAPTER 4 MATHEMATICAL MODEL This chapter presents a mathematical model for C02 absorption and desorption in aqueous amine solutions falling as a laminar liquid film over a hemispherical surface. The model was developed for the general case where the aqueous solution may contain a binary mixture of amines (e.g., MEA+AMP or MEA+MDEA) and where C02 undergoes a series of reversible reactions. This model can be readily modified to accommodate a single aqueous amine (e.g., MEA or AMP) or water by setting the initial concentration of one or both amines to zero. The model reduces to the well-known pseudo-first-order kinetics when all reactions are lumped into a single, overall reaction by setting the rates, of all but one reaction to zero. The important model parameters and physical property data needed to solve the model equations are identified. Correlations are presented for parameters for which data are available in the literature. For cases where data were deficient or not available in the literature (e.g., Henry's constant and diffusivity of C02 in aqueous amine blends), new data were obtained and new correlations were developed. A methodology to solve the model equations and to estimate unknown model parameters using the non-linear regression package GREG is also presented in this chapter. 61 4.1 Reaction Mechanism The model for CO2 absorption and desorption applies to the following aqueous amine systems: 1. CO2+MEA+H2O 5. CO2+MEA+MDEA+H2O 2. CO2+DEA+H2O 6. CO2+MEA+AMP+H2O 3. CO2+MDEA+H2O 7. CO2+DEA+MDEA+H2O 4. CO2+AMP+H2O 8. CO2+DEA+AMP+H2O Since most of the data presented in this work refers to the C02+AMP+H20 system, this system was chosen as the base case and the reactions governing this system are presented first. For systems involving other amines, the additional reactions are given subsequently. In accordance with the convention used in the amine literature, MEA is represented as R1NH2, where R1 denotes -CH2CH2OH. DEA is represented as R1R2NH, where R1 = R2 = -CH2CH2OH. MDEA is represented as R1R2R3N, where R1 = R2 = -CH2CH2OH and R3 = -CH3. AMP is represented as R4iNH2, where R4 = -C(CH3)2CH2OH. 4.1.1 Reactions for C02+AMP+H20 (Base Case) When CO2 is absorbed or desorbed in aqueous AMP solutions, the following reactions may occur in the liquid phase: 62 AMP- Zwitterion Formation: k K C02+R4NH2<!^R4NH^COO- (4.1) k_, AMP-Zwitterion Deprotonation: k K R4NH^COO- + R4NH2 <V R4NHCOO" + R4NH3 (4.2) k_2 R4NH^COO" + H2O^R4NHCOO~ +H30+ (4.3) k.3 k K R.NH+COO' + OH" ^>R4NHCOO- + H20 (4.4) AMP-Carbamate Reversion: k5,K5 R4NHCOO-+H2O^R4NH2+HC03 (4.5) k-5 AMP Deprotonation: k K R4NH^+OH ^R4NH2+H20 (4.6) k-e Bicarbonate Formation: k K C02+OH\^HC03 (4.7) k-7 Carbonate Formation: HC03+OH-^CO^+H20 (4.8) k-8 Water Dissociation: 2H2O^OH+H30+ (4.9) k_8 In the above reactions, the k's denote rate constants and the K's denote equilibrium constants. 63 4.1.2 Reactions for C02+MEA+H20 When C02 is absorbed or desorbed in aqueous MEA solutions, reactions (4.7) to (4.9) are the same as those in the base case, but reactions (4.1) to (4.6) are replaced by the following reactions: MEA-Zwitterion Formation: k10,K10 COa+RJNHa ^R^H+COCT (4.10) k-10 MEA-Zwitterion Deprotonation: R1NH^COO-+R1NH2kl51R1NHCOO-+R1NH: (4.11) k K R.NH^COCT + H20 ^ R^HCOCT + H30+ (4.12) k.12 k K R^H+COCT +OH' ^R^HCOO" +H20 (4.13) MEA-Carbamate Reversion: k K R1NHCOO-+H2O^R1NH2+HC03 (4.14) k-i4 MEA Deprotonation: k K RJMH^+OH "^RJNHa+HaO (4.15) 4.1.3 Reactions for C02+DEA+H20 When C02 is absorbed or desorbed in aqueous DEA solutions, reactions (4.7) to (4.9) are once again the same as those in the base case, but reactions (4.1) to (4.6) are replaced by the following reactions: 64 DEA- Zwitterion Formation: k K C02 +R1R2NH ^6RJR2NH+COCT DEA-Zwitterion Deprotonation: R^NhrCOCr +R1R2NHk,S7 R^NCOO- +R1R2NH2 k_17 R1R2NH+COO-+H20 RJR2NCOO~ + H30+ ^-18 R1R2NH+COO"+OH" ^ RJR2NCOO"+H20 K-,9 DEA-Carbamate Reversion: ^ 20-^20 RJR2NCOO+H20 ^ RJR2NH + HC03 k-20 DEA Deprotonation: R^NH* + OH" ^ R^NH + H20 4.1.4 Reactions for C02+MDEA+H20 When C02 is absorbed or desorbed in aqueous MDEA solutions, reactions (4.7) to (4.9) are again unchanged, but reactions (4.1) to (4.6) are replaced by the following: MDEA-C02 Reaction: k K C02 + R^^N + H20 ^ R1R2R3NH+ + HC03 (4.22) k_22 MDEA Deprotonation: k K RJR2R3NH++OH" ^\R2R3N + H20 (4.23) 65 (4.16) (4.17) (4.18) (4.19) (4.20) (4.21) 4.1.5 Reactions for C02+MEA+MDEA+H20 When C02 is absorbed or desorbed in aqueous blends of MEA and MDEA, in addition to reactions (4.7) to (4.15) and (4.22) to (4.23), the following reaction occurs: Deprotonation of MEA-Zwitterion to MDEA: R^H+COO" +R1R2R3Nk2<i24R1NHCOO- +R1R2R3NH+ (4.24) k_24 4.1.6 Reactions for C02+MEA+AMP+H20 When C02 is absorbed or desorbed in aqueous blends of MEA and AMP, in addition to reactions (4.1) to (4.15), the following reactions must be included: Deprotonation of AMP-Zwitterion to MEA: k K R4NH^COO"+R1NH2 2^5R4NHCOCT +R1NH; (4.25) k-25 Deprotonation of MEA-Zwitterion to AMP: R1NH^COO-+R4NH2 <-> R^HCOCT + R4NH+ (4.26) ^-26 4.1.7 Reactions for C02+DEA+MDEA+H20 When C02 is absorbed or desorbed in aqueous blends of DEA and MDEA.in addition to reactions (4.7) to (4.9) and (4.16) to (4.23), the following reaction occurs: Deprotonation of DEA-Zwitterion to MDEA: R^Nr-rCOCr +R1R2R3Nk2i27R1R2NCOO" +R1R2R3NH+ (4.27) 66 4.1.8 Reactions for C02+DEA+AMP+H20 When C02 is absorbed or desorbed in aqueous blends of DEA and AMP, the following reactions occur in addition to reactions (4.1) to (4.9) and (4.16) to (4.21): Deprotonation of AMP-Zwitterion to DEA: k28,K28 R^H+COCT+R^NH <-» R4NHCOCr+R1R2NH2- (4.28) k_28 Deprotonation of DEA-Zwitterion to AMP: k29.K2g R1R2NH+COO-+R4NH2 <-> R1R2NCOO"+R4NH^ (4.29) 4.2 Reaction Rates In the above reaction schemes, all reactions are considered to be reversible. Some of these reactions proceed at finite rates while others (that involve only a proton transfer) occur almost instantaneously. This section presents the rate expressions for reactions with finite rates. For convenience, the chemical species in reactions (4.1) to (4.29) are renamed as follows: C, = C02, C2 = R4NH2,C3 = R4NH^, C4 = R4NHCOO", C5 -HC03, C6 = CO"2, C7 = OH", C8 = H30+, C9 = H20, C10 = R,NH2> C„ = R^H*. C12 = RJNHCOCT, C13 =R1R2NH, C14 =R1R2NHJ,C15 =R1R2NCOO"> C16 =R1R2R3N, C17 =R1R2R3NH+ (4.30) 4.2.1 Reaction Rates for C02+AMP+H20 System (Base Case) This case includes reactions (4.1) to (4.9). Among these, only reactions (4.1) to (4.4) and (4.7) have finite rates. The latter are expressed as follows: 67 where: ""l-4 = -k'i C,C2 ~ C4 1+ r7 = -k7C,C7 +—C5 K7 A = 'JO vk-i y K1K; + vk-i y K1K3 + vk-i y K1K4 (4.31) (4.32) (4.33) B = i\2 c2 + cq + ik_i , lk-1 > y Ik-1 J c7 (4.34) Equation (4.31) is based on the assumption that pseudo-steady state is valid for the zwitterion. The detailed derivation of the base case equations is given in Appendix D. The derivations for other systems involving zwitterion mechanisms are similar and therefore not presented separately. 4.2.2 Reaction Rates for C02+MEA+H20 System This case includes reactions (4.7) to (4.15), with reactions (4.7) and (4.10) to (4.13) having finite rates. The rate expression for reaction (4.7) is given by equation (4.32). The rate for reactions (4.10) to (4.13) is given by: "10-13 (AY _k10 IBJ. f4\ (4.35) 1 + vBj 68 where: A Vk-io y K10K11 • + k1 vx-io y K10K12 • + ^ k13 ^ C0 v,x-io y Ki0K13 (4.36) B = k ^ Vk-10 , C10 + ^k N vk-10 y C9 + vk-io y C7 (4.37) 4.2.3 Reaction Rates for C02+DEA+H20 System This case includes reactions (4.7) to (4.9) and (4.16) to (4.21), with reactions (4.7) and (4.16) to (4.19) having finite rates. The rate expression for reaction (4.7) is given by equation (4.32). The rate for reactions (4.16) to (4.19) is given by: (A) _k16 CiC13 -c15 •16-19 1 + B (4.38) where: A = 'k N K17 vk-ie y '14 K16K17 • + M8 vk-ie y K16K18 • + ki vk-ie y ^16^19 (4.39) B = vk-ie y C13 + /k ^ *18 vk-ie y |C9 + 1'k ^ *19 vk-ie y C7 (4.40) 4.2.4 Reaction Rates for C02+MDEA+H20 System This case includes reactions (4.7) to (4.9) and (4.22) to (4.23), with reactions (4.7) and (4.22) having finite rates. The rate expression for reaction 69 (4.7) is given by equation (4.32) and the rate for reaction (4.22) can be obtained from: r22 --k^Cp^ + — C5C17 K22 (4.41) 4.2.5 Reaction Rates for C02+MEA+MDEA+H20 System This case includes reactions (4.7) to (4.15) and (4.22) to (4.24), with reactions (4.7), (4.10) to (4.13), (4.22) and (4.24) having finite rates. The rate expressions for reactions (4.7) and (4.22) were already defined by equations (4.32) and (4.41), respectively. The rate of reactions (4.10) to (4.13) and (4.24) can be expressed as: 10 ^1^10 — '10-13,24 1 + B (4.42) where: A = vk-10 j K10K11 • + 1k ^ K12 vk-10 j K10K12 • + *13 VK-10 J Ki0K12 • + ^24 k-i '17 V,x-10 J K10K24 (4.43) B = f k A vk-10 j C10 + 'k ^ K12 k-10 !C9 + f k A *13 vk-10 j C7 + ,X24 vk-ioy '16 (4.44) 4.2.6 Reaction Rates for C02+MEA+AMP+H20 System This case includes reactions (4.1) to (4.15) and (4.25) to (4.26), with reactions (4.1) to (4.4), (4.7), (4.10) to (4.13) and (4.25) to (4.26) having finite 70 rates. The rate expression for reaction (4.7) is defined by equation (4.32). The rate of reactions (4.1) to (4.4), (4.25) and (4.10) to (4.13) and (4.26) are given by: where: and 11-4,25 1+ B (4.45) A r i. \ • + K,K3 [k_jK,K k25 25 (4.46) (k \ B = l\2 C2 + C9 + 4 c7 + ^25 ^k-i, Lk-1 J y lk-1 J ^k_i J '10 (4.47) -kio CiC10 -C12 '10-13,26 1 + B (4.48) where: (k ^ vk-10 j '11 • + v12 Vk-10 J K10K12 • + k13 Vk-10 J K10K13 • + ^26 Vk-10 J ^10^26 (4.49) B = f k A Vk-io J lc10 + 1k ^ *12 Vk-io J lc9 + Ak ^ Vk-10 J C7 + 'k ^ *26 vk-ioy (4.50) 71 4.2.7 Reaction Rates for C02+DEA+MDEA+H20 System This case includes reactions (4.7) to (4.9), (4.16) to (4.23) and (4.27), with reactions (4.7), (4.16) to (4.19), (4.22) and (4.27) having finite rates. The rate expressions for reactions (4.7) and (4.22) are defined by equations (4.32) and (4.41), respectively. The rate for reactions (4.16) to (4.19) and (4.27) is given by: -kie CiC13 "16-19,27 1 + '1^ vBy (4.51) where: ( k17 1 C14 k-16 7^16^17 + ^18 f I, \ Vk-16 J K16K18 + ki Vk-16 J •\27 Ki6K1s Vk-16 ) c17 K16K27 (4.52) B = fk ^ K17 vk-ie j C13 + vk-ie j IC9 + r k ^ ^19 vk-ie j rk ^ k-i V,x-16 J (4.53) 4.2.8 Reaction Rates for C02+DEA+AMP+H20 System This case includes reactions (4.1) to (4.9), (4.16) to (4.21) and (4.28) to (4.29), with reactions (4.1) to (4.4), (4.7), (4.16) to (4.19) and (4.28) to (4.29) having finite rates. The rate expression for reaction (4.7) is defined by equation (4.32). The rates for reactions (4.1) to (4.4), (4.28), and (4.16) to (4.19) and (4.29) are given by: 72 11-4,28 c.|C2 ~ C4 1+ (4.54) where: A = K K,K3 ^k_iy • + KiK-i + k28 v^-1 j '14 28 (4.55) flO (y \ B = 1\2 c2 + C9 + c7 + •^28 vk-1 J lk_i J lk-1 J lk-1 ) '13 (4.56) and M6 CiC13 -C15 LB J "16-19,29 (4.57) where: A = k17 Vk-16 J '14 K16K17 + 1k ^ *18 vk-ie y ^k A N19 K16K18 Vk-16 J K16K19 • + K29 vk-ie j ^16^29 (4.58) B = K17 Vk-16 J C13 + M8 Vk-iey CQ + *19 Vk-16 J :c7 + 'k A K29 Vk-16 J (4.59) 4.3 Reactive Gas Absorption/Desorption Model 4.3.1 Hydrodynamics of Liquid Film Figure 4.1 is a schematic diagram of the hemispherical film described in Chapter 3. The liquid descends in the form of a laminar film from the pole of the 73 hemisphere towards its equator where it is collected in the funnel. The gas, depending on the operating conditions, is either absorbed into the liquid film or is desorbed from it. Gas-Liquid Interface Figure 4.1: Schematic of the Liquid Film This system can be mathematically modeled to calculate the rate of gas absorption or desorption over the exposed area of the film provided the film thickness and the velocity profile are known as a function of position (0). Lynn et 74 hemisphere towards its equator where it is collected in the funnel. The gas, depending on the operating conditions, is either absorbed into the liquid film or is desorbed from it. Gas-Liquid Interface Figure 4.1: Schematic of the Liquid Film This system can be mathematically modeled to calculate the rate of gas absorption or desorption over the exposed area of the film provided the film thickness and the velocity profile are known as a function of position (9). Lynn et 74 al. (1955) have studied the hydrodynamics of laminar liquid films flowing over a sphere. They assumed that the thickness of the liquid film at any latitude on the sphere is the same as it would be for the same flow rate per unit length on a plane surface making the same angle with the vertical. Based on this assumption, the film thickness at latitude 9 (A9) is given by: Ae=A0(sin9)-2/3 (4.60) where A0 is the film thickness at the equator (9 = nl2) of the sphere and is given by: A0 = ' 3vQ ,27iRgy ,1/3 (4.61) It follows from the above assumption that a half parabolic velocity profile will exist at all latitude on the sphere so that the velocity distribution, Ve, in the film can be approximated by the following equation: Ve=V0[l-x2J (4.62) where x is the dimensionless distance from the gas-liquid interface, i.e, x = (R + Ae - r)/ Ae (see Figure 4.1). V0 is the velocity at the film surface: V0 = 3Q 47iRAoy (sin9)-1M (4.63) Note that at 9 = 0 (i.e., at the pole of the hemisphere), equation (4.60) is not valid because at that point the thickness of the liquid film becomes infinite. However, this is a physical impossibility because only a finite amount of liquid 75 (about 2 mL/s) flows through the liquid feed tube of 3 mm I.D. We calculated the thickness of the liquid film at different 9 values for water and amine solutions using equation (4.60) and we found that we get reasonable values at 9 very close to the pole. For numerical solution of the model equations, the liquid film in 0 direction was divided into 200 points to integrate the absorption rate over the entire hemispherical surface. This means that the first point at which the concentration gradient was calculated lies at 9 = 0.45° (0.008 radian). For water at 298 K and liquid flow rates of 2, 3 and 4 mL/s, the film thickness at 9 = 0.45° (near the pole) calculated from equation (4.60) are 3.31, 3.80 and 4.17 mm respectively. The corresponding values at 9 = 90° (at the equator) are 0.13, 0.15 and 0.17 respectively. The thinning of the liquid film from the pole to the equator occurs because the same amounts of liquid flows through the increasing cross sectional area as it descends from the pole to the equator. 4.3.2 Model Equations The mathematical model presented below is based on the concept of gas absorption or desorption accompanied by multiple reversible chemical reactions in a hemispherical liquid film. The main assumptions involved in the model derivation are: • The interfacial concentration of dissolved (molecular) C02 corresponds to the physical solubility as determined by Henry's law (i.e., p., = H,C\). 76 • Mass transfer in the direction of the liquid flow is dominated by bulk convection and diffusion in the 9 direction is negligible. • The flow field is uniform in the O direction (V<D= 0). Hence, due to symmetry, bulk transfer and diffusion in the O direction are non-existent. • The gas is saturated with water vapor at the system temperature and pressure and there is no transfer of amine or water from the liquid film into the gas phase. • The diffusion coefficients of various ionic species are equal so that there are no electrostatic potential gradients present in the liquid film. Rinker et al. (1995) and Hagewiesche et al. (1995) have found that a more rigorous approach of taking into account the electrostatic gradient terms resulting from unequal diffusion coefficients requires much more computational effort and has little effect on the predicted rates of absorption. In addition, the parametric sensitivity analysis performed in this work show that a large deviation (±50%) in the values of the diffusivities of ionic species do not have any significant effect on the calculated absorption and desorption rates (see Table 5.2). • The film is isothermal and the system is at steady state. The following equations govern the diffusion and reaction processes for C02 absorption/desorption in an aqueous solution containing MEA and AMP. The equations are based on material balances for elements of the liquid film. 77 C02 Balance: ae F^x.ejsinB d2C, _ F^x^JsineSC, ax' F2,x,e) ax ^Aox-ner^ (4.64) Tbfa/ Carbon (C02) Balance: dC= aCR aC ac, ac4 —- +—1 + —^ + ae ae ae ae 12 ae f*2 F,(x,0)sine a^c, | D4 a'c4 [ D5 a'c5 | D6 a2c6 | p12 a2c12 ax' D1 ax2 D, ax2 D, ax2 D, ax' 2F1(x,e)sin8['ac, | D4 ac4 | D5 ac5 [ D6 ac6 ( p12 ac12 D, ax D, ax D, ax D1 ax F2(x,e) v ax Tote/ AMP Balance: (4.65) ac, ac, ac. ae ae ae ^(x, 9) sin 9 D2 a2c2 D, a2c, D4 a2c4 D, ax' D, ax' D, ax' -2 F,(x,9)sin9 F2(x,9) D2 ac2+D3_ac3_+D4 ac4 D1 ax D, ax D, ax (4.66) Total MEA Balance: ac10 ac„ ac ae +- 11 ae +- ae ^2- = F1(x,e)sine D10 a2c10 +D!ia^c!i+D12 a2c12 D1 ax" D1 ax' .F^x.GJsine F2(x,6) Dio 3C10 +Pii5CJL+D12 ac12 ax2 D, <3x D., ax D, ax (4.67) 78 Electron Neutrality Balance: dC3 dC8 dCu 5C4 3C5 03C6 dC7 dC • + • + 12 ae ae ae ae ae F^x.ejsine ae ae 'D3 a2c3 ae D« a2c 8+Pl1 D4 d'c4 D, ax2 .^(x.ejsine D1 ax2 D1 ax2 ax2 D, ax2 p5a2c5 2D6a2c6 p7a2c7 D12 a2c12 D, ax^ DT ax' D, ax' 'D3 ac3 F2(x,e) +• DA ac 8+D11ac11 D4ac4 D, dx DT ax DT dx DT 3X p5ac5 2D6ac6 p7ac7 p12ac12 DT dx DT ax DT dx D, dx Equilibrium Reactions: (4.68) (4.69) K6= °2 K„ = C3C7 cfi C5C7 K9 = C7C8 (4.70) (4.71) (4.72) w C10C5 K14 =— '12 (4.73) K15 = CTTC7 (4.74) 79 where: Re = Q 2TIRV (4.75) D1 (4.76) Ga = R3g (4.77) G(9) Ga v3Rey (sine)2'2 (4.78) F^x.e): G(e) + (1-x)' 1-x2 3ReSc (4.79) F2(x,e) = G(9) + (1-x) (4.80) The above model consists of eleven partial differential and algebraic equations (4.64 to 4.74) and eleven unknowns. The equations can be solved numerically subject to the following initial and boundary conditions: Initial Conditions: At 6 = 0, Cj =C° for j =1-8,10-12 (4.81) Boundary Conditions: at 9 > 0 and x = 0, fdC^ and i Pi A ( c;-vkgHA (sine) 2/3 'ac ^ v dx j dx V Vj=2-8,10-12 = 0 (4.82) 80 at G > 0 and x = 1. <3x V ^=1-8,10-12 = 0 (4.83) For the case where the gas side resistance to mass transfer is negligible (kg -» oo), the boundary condition given by equation (4.82) reduces to: fdC } at 9 > 0 and x = 0, C\=^- and —>- = 0 (4.84) H1 vexy j=2-8,10-12 Note that the model equations (4.64) to (4.74) are general and applicable to any of the aforementioned systems by substituting corresponding chemical species, reaction rates and diffusivities. The model equations were solved numerically using a commercial software package called Athena Visual Workbench (see Section 4.5). 4.3.3 Liquid Bulk Concentrations In order to solve the model equations (4.64) to (4.74), the bulk concentrations (C°), must be known. They can be obtained by writing the following balances: Total C02 Balance: C° +C°4 + C° +C° +C°2 = ainitial([R4NH2]initial +[R1NH2]initial) (4.85) Total AMP Balance: C°+C°+C°4=[R4NH2]initial (4.86) 81 Total MEA Balance: C?0+C^+A2=[RiNHJinitial (4.87) Electron Neutrality Balance: C° + C° + C° - C° - C° - 2C° - C° - C°2 = 0 (4.88) Equilibrium Equations: *a=^ (4-89) Ke = c^" (4"90) 7 C°C° K8=-^- (4.92) c°c° K9=C?C° (4.93) /~»0 /->0 K14=^P (4.94) u12 ^15 r~*o (4.05) C°C° 82 4.3.4 Rate of Absorption or Desorption with Chemical Reaction The rate of gas absorption or desorption over the hemispherical film at any latitude 0 can be calculated by invoking the Fick's law of diffusion: NA(6) = -ND(9) = -[2n(R + Aj Sine]5if^-1 (4.96) Ae V <?X 7(0,6) The total rate of absorption or desorption is obtained by integrating equation (4.96) over the entire hemisphere from 0 = 0 (at the pole) to 0 = 7i/2 (at the equator) as given by NA =-ND =-2TCD1A0 J [G(0) + l]2(sin0)1/3 ^ d0 (4.97) v 3x j (o.e) Note that for absorption, the concentration gradient, v dx y(oe) , is a negative quantity and for desorption it is a positive value. The procedure to calculate NA and ND using equation (4.97) is described in Section (4.6). 4.3.5 Rate of Absorption or Desorption without Chemical Reaction When there is no reaction between the gas and liquid film, the general model reduces to a case of physical absorption or desorption. This situation is governed by a single partial differential equation, which under certain assumptions, can be solved analytically and an explicit expression for the mass transfer rate can be derived. The details of this derivation are given in Appendix C and the final results are reproduced below. 83 Rate of Physical Absorption or Desorption: N° = -N° =3.1774J^-^(c; -C°) (4.98) For absorption, the driving force (c\ -C°) is positive whereas for desorption it is negative. Physical Mass Transfer Coefficient: Sh = 1.26758Re05Sc05 (4.99) Contact Time: xc =1.5848TTR' (4.100) In equation (4.99), the Reynolds number (Re) and Schmidt number (Sc) are defined according to equations (4.75) and (4.76), respectively. The Sherwood number (Sh) is given by: Sh = -^—°- (4.101) The analytical solution for physical absorption (i.e., equation 4.98) is very important because it can be used to calculate the diffusivity of a gas in a liquid from experimental absorption rate measurements. A comparison of the mass fluxes for physical absorption predicted by numerical methods and analytical solution is given in Chapter 5. 84 4.3.6 Enhancement Factor Once the rate of absorption (or desorption) with and without chemical reaction is known, the enhancement factor can be determined from the following equation: N N EA = and ED==5- (4.102) A NA° D N° The enhancement factor is a measure of the effect of chemical reaction on the rate of mass transfer. 4.3.7 Overall Reaction Rate (rtotai) In order to solve the model equations (4.64) to (4.74) to calculate the rate of absorption or desorption, an expression for overall reaction rate (rtotai) is required (see equation 4.64). The term rtotai is the sum of all rates of reactions that either consume or produce 0O2 in the solution. The relationships for rtotai for the eight systems considered here are listed in Table 4.1. Note that the reactions rates presented in Section 4.2 have been derived for the general case including all possible reactions that may occur when C02 reacts with aqueous amine solutions. However, for amines like MEA, DEA and AMP, where the zwitterion mechanism applies, it has been shown previously (Danckwerts, 1979; Versteeg and van Swaaij, 1988; Little et al., 1992a,b; Hagewiesche, et al., 1995; Rinker et al., 1996) and has been confirmed by our own data regression and parametric sensitivity analysis presented in Chapter 5 that the contribution to mass transfer rates of reactions involving the 85 deprotonation of amine-zwitterion to water and hydroxyl ions is negligible. Therefore, for AMP, reactions (4.3) and (4.4), for MEA, reactions (4.12) and (4.13) and for DEA, reactions (4.18) and (4.19) can be neglected without any loss of accuracy. This means that the rate expressions for systems involving AMP, MEA and DEA can be simplified by setting the rate constants k3, k4, ki2, k13, ki8 and k19 equal to zero. Table 4.1: Overall reaction rate (rtotai) System Ttotal Equation No. C02+MEA+H20 r7 ri0-13 (4.32)+(4.35) CO2+DEA+H2O r7 + ri6-19 (4.32)+(4.38) CO2+MDEA+H2O r7+r22 (4.32)+(4.41) CO2+AMP+H2O r7+r,-4 (4.32)+(4.31) C02+MEA+MDEA+H20 r7 0-13,24 +r22 (4.32)+(4.42)+(4.41) C02+MEA+AMP+H20 r7 +l"l0-13,26 +l"l-4,25 (4.32)+(4.48)+(4.45) C02+DEA+MDEA+H20 r7 +f"l 6-19,27 +r22 (4.32)+(4.51)+(4.41) CO0+DEA+AMP+H2O r7 +l"l6-19,29 +l*1-4,28 (4.32)+(4.57)+(4.54) 4.4 Model Parameters The parameters needed to solve the mathematical model are listed in Table 4.2 and their determination is outlined below. 86 The density and viscosity of amine solutions were obtained using the correlations presented in Appendix E. The Henry's constant and the diffusion coefficient of C02 in aqueous amine solutions were obtained using the N20 analogy and they are presented in Appendix F and Appendix G, respectively. The gas-side mass transfer coefficient was estimated by absorbing a mixture of C02 and N2 in aqueous DEA solutions and its determination is presented in Appendix I. The equilibrium constants for fast reactions are presented in Appendix H. 4.5 Numerical Implementation The model equations (4.64) to (4.97) were solved using a commercial software package called Athena Visual Workbench licensed from Stewart and Associate Engineering Software Inc., Madison, Wisconsin. Athena Visual Workbench is a Windows based fully integrated environment for process modeling and nonlinear parameter estimation. It consists of three mathematical solvers: DDAPLUS, PDAPLUS and GREGPLUS. DDAPLUS is used for the integration and sensitivity analysis of initial-value problems with mixed algebraic and differential equations, PDAPLUS is used for the integration and sensitivity analysis of initial-boundary value problems with mixed algebraic and differential equations and GREGPLUS is used for non-linear parameter estimation. DDAPLUS and PDAPLUS employ the method of lines with finite differences, global orthogonal collocation and collocation on finite elements and GREGPLUS uses weighted least squares and Bayesian estimators with single as well as 87 Table 4.2: Parameters for absorption/desorption model Parameter Definition Source P Density of amine solutions Appendix E Viscosity of amine solutions Appendix E Hc02 Henry's constant for C02 in amine solutions Appendix F Di Diffusivity of C02 in amine solution Appendix G D2-D17 Diffusivities of amine and other ionic species in amine solutions Appendix G kg Gas-side mass transfer coefficient Appendix I K5 tO Kg, K14. K15, K20, K21, K23, Equilibrium constants for reaction (4.5) to (4.9), (4.14), (4.15), (4.20), (4.21) and (4.23) Appendix H k7 Forward rate constant for CO2 hydration reaction (Rxn. 4.7) Appendix H k-|,k10 ,k16,k22 Forward rate constants for reactions (4.1), (4.10), (4.16) and (4.22) Chapter 5 k2 k3 k4 k_! k^ k_., Combined rate constants for C02+AMP+H20 system Chapter 5 k-n k12 k13 k ' k ' k ^-IO -10 -10 Combined rate constants for CO2+MEA+H2O system Chapter 5 ^17 k18 k19 k ' k ' k •^--le ^-16 *-i9 Combined rate constants for CO2+DEA+H2O system Chapter 5 k 24 k 25 k 26 k ' k ' k ' Combined rate constants for mixed amine systems Chapter 5 k27 k28 k2g k-16 k-i6 k_16 88 multi-response data. Further information about this package can be obtained from Caracotsios and Stewart (1985, 1995), Stewart et al. (1996) and Bain and Stewart (1990). The present model is quite complex and cannot be solved directly using Athena. For a given set of operating conditions, the calculation of absorption and desorption rates involves: (a) calculation of physical-chemical properties, (b) solution of the equilibrium model (equations 4.84 to 4.95), (c) solution of the model equations (4.64) to (4.84) to calculate concentration gradients at different latitude 0, (d) calculation of the total absorption or desorption rate by integrating over the entire hemisphere. The Fortran code to perform tasks (a) and (d) were written as part of this study and that for tasks (b) and (c) were generated from Athena by implementing the model equations in the Athena Visual Workbench environment. All these subroutines were then combined according to the scheme shown in Figure 4.2 using the Compaq Visual-Fortran Developer Studio. For a typical run, an input data file consisting of identification flags and operating variables (see Figure 4.2) was generated in Excel and added to the Developer Studio's project files. The simulation was run using the fastest available desktop PC (Pentium 4, 1700 MHz, 256 RAM). The results were plotted using Microsoft Excel. 4.6 Parameter Estimation The unknown parameters such as ki, ki0, k-|6, k22, etc. can be determined by minimizing the following objective function: 89 S(k) = Z[yi-y(k)]Qi[yi-y(k)] (4.103) i=i where k = (ki, k10, ki6, k22, etc.)T is the unknown parameter vector, yt is the measured value of the state variable (in the present case, it is the absorption/desorption rate of C02) and y(k) is the calculated value of the state variable which is obtained by solving the model equations (4.64) to (4.97) for a set of assumed values of the unknown parameters. N denotes the number of experimental data and Q is the weighting matrix. For least squares estimation, Q is taken as the identity matrix. Note that the parameters for each case were estimated separately. However, the same methodology was used in all cases. The optimization was carried out using Athena's GREGPLUS solver according to the scheme shown in Figure 4.3. 90 (^START INPUT DATA IDENTIFICATION FLAGS MODE NAM IRXN I KG 1 = Absorption, 2 = Desorption 1 = MEA, 2 = DEA, 3 = MDEA, 4 = AMP 0 = without reaction, 1 = with reaction 0 = without k , 1 = with k OPERATING VARIABLES T(K), Ptotal(kPa), QL(mL/S), CAM1(wt%), CAM2(wt%) C02 Loading (mol/mol), FC02(std. mL/min), FN2(std. mL/min), FDN2 (std. mL/min), CALCULATE PHYSICAL-CHEMICAL PROPOERTIES CALCULATE INITIAL CONCENTRATIONS OF CHEMICAL SPECIES SOLVE MODEL EQUATIONS CALCULATE ABSORPTION OR DESORPTION RATE PRINT RESULTS STOP CALL SUBROUTINES . DENSITY . CONCENTRATION . VISCOSITY . DIFFUSIVITY .HENRY . GASFLOW . KGAS . EQCONSTANT . RATECONSTANT (Calculates physical-chemical properties from correlations given in Appendices E to H) CALL SUBROUTINE . EQUILIBRIUM (Solves eqs. (4.85 to 4.95) using Newton's Homotopy Continuation Method-Generated from Athena) CALL SUBROUTINE . PDAPLUS (Solves eqs. (4.64 to 4.84) using Method of Lines-Generated from Athena) CALL SUBROUTINE . INTEGRAL (Calculates absorption or desorption rate from eq. (4.97) using Simpson's Method) CALL SUBROUTINE . PRINT (Prints results for plotting) Figure 4.2: Numerical scheme for model solution 91 INPUT DATA IDENTIFICATION FLAGS MODE : -1 = Absorption, 2 = Desorption NAM : 1 = MEA, 2 = DEA, 3 = MDEA, 4 = AMP IRXN : 0 = without reaction, 1 = with reaction I KG : 0 = without kg , 1 = with kg OPERATING VARIABLES T(K), Ptotal(kPa), QL(mL/S), CAM1(wt%), CAM2(wt%), C02 Loading (mol/mol), FC02(std. mL/min), FN2(std. mL/min), FDN2 (std. mL/min), MEASURED RATES NA (kmol/s) or ND (kmol/s) INITIAL GUESS FOR UNKNOWN PARAMETERS GREG (PARAMETER ESTIMATION PACKAGE) •«-H PRINT PARAMETER ESTIMATES CALL SUBROUTINE . SOLVER (Calculates absorption or desorption rate for a given set of parameters using the numerical scheme shown in Figure 4.2) STOP Figure 4.3: Numerical scheme for parameter estimation 92 CHAPTER 5 RESULTS AND DISCUSSION This chapter presents the experimental and theoretical results and their interpretation for C02 absorption and desorption in aqueous amine solutions based on the novel hemispherical contactor and the comprehensive mathematical model discussed in Chapters 3 and 4 respectively. The results presented here pertain to the eight systems listed in Chapter 4. The bulk of the experimental work was focused on C02 desorption from aqueous amine systems. Some absorption experiments were also performed but the purpose of the absorption experiments was mainly to verify the experimental technique and the mathematical model because, for most amines, extensive absorption data are already available in the literature. 5.1 Model Verification 5.1.1 C02 Absorption/Desorption in Aqueous Amines Figure 5.1 shows a comparison of the predicted and experimentally observed absorption rates for the three most commonly used alkanolamines namely MEA, DEA and MDEA at 303 K. Note that, unless otherwise stated, predicted results throughout this thesis were obtained based on best fit parameter estimates from this work. In Figure 5.1, MEA is a primary amine and has the highest reactivity, DEA is a secondary amine with intermediate reactivity 93 and MDEA is a tertiary amine with very low reactivity with C02 in aqueous solutions. This behavior is clearly demonstrated in Figure 5.1, where absorption rates decrease from MEA to DEA to MDEA. 800 .C/3 O £ 600 • MEA A DEA • MDEA Model (Literature) Model (This work) i—i—i—i—|—i—i—i—i—|—i—i—r • • i i I i i i i I i i i i I i i i i I i i i i I i i i i_ 0.0 1.0 2.0 3.0 4.0 5.0 CAMINE (kmol/m3) 6.0 Figure 5.1: Predicted and experimental absorption rates of C02 in aqueous solutions of MEA, DEA and MDEA at 303 K (pC02= 97.0 kPa, QL = 2.0 mL/s) In Figure 5.1, the symbols represent the experimental data, the solid lines represent the model predictions based on the parameter estimates obtained in this work (described later in this chapter) and the dotted lines represents model predictions based on literature correlations (Hikita et al., 1977a; Rinker et al., 1996; Rinker et al., 1995) for the kinetic parameters. A slight deviation of the predicted trends using literature correlations is reasonable as different authors 94 have employed different experimental techniques (see Tables 2.1 to 2.3), different correlations to calculate physical properties and different ways to interpret their experimental data to estimate kinetic parameters. Furthermore, the presence of small amounts of impurities in the MEA and DEA solutions may cause significant variations in the measured absorption rates. Therefore, it can be safely said that the present model predicts absorption rates with fairly good accuracy. 100 40 o E E ° 60 x CD CD rz o CO CD Q 0 T i i T—rr—I—i—i—1—i—r • MEA —i—i—i—i—i—i—i—r T—'—i—i— • A DEA - • MDEA / / \ — Model / / ; / /m • • i i i i i i i i i i 0.0 Figure 5.2: 0.1 0.2 0.3 0.4 Loading (mol of C02 /mol of amine) 0.5 Predicted and experimental desorption rates of C02 from aqueous solutions of MEA, DEA and MDEA at 373 K (pco < 5.0 kPa, QL = 2.0 mL/s) Figure 5.2 presents desorption rates of C02 from aqueous solutions of MEA, DEA and MDEA at 373 K. In this figure, the symbols represent the experimental data and the solid lines represent the model predictions based on 95 the parameter estimates obtained in this work. The agreement between experimental and predicted desorption rates is excellent. Figure 5.2 also shows that MDEA, which does not directly bond with CO2, desorbs much faster compared to DEA and MEA, which react with C02 to form stable carbamate. 500 CO o 400 E " 300 x cu CO * 200 c o 8 100 CD Q 0 0.20 Figure 5.3: Experimental (this work) Predicted (literature) - Predicted (this work) -I I L 0.30 0.40 Loading (mole of C02/mole of MEA) 0.50 Predicted and experimental desorption rates for C02 desorption from aqueous MEA solution at 373 K (p^*1 5.0 kPa, QL = 2.0 mL/s) A comparison of the predicted and experimental desorption rates based on a literature correlation of Hikita et al. (1977) for the CO2-MEA-H2O system is shown in Figure 5.3. It can be seen from this figure that, although the predicted trend with respect to C02 loading is the same as that observed experimentally, the literature correlation over-predicts the desorption rates by a factor of 2 to 5. This is not surprising because most of these correlations were developed based 96 on absorption data only and (as discussed later in this chapter) cannot be reliably extrapolated to desorption conditions. 5.1.2 Detailed Profiles in the Hemispherical Film In order to further validate the model and the numerical scheme to solve the model equations, in this section we analyze detailed concentration profiles within the hemispherical film generated from the numerical simulation results obtained for both absorption and desorption at various operating conditions. The profiles are plotted in Figures 5.4 to 5.10. The plots for absorption were generated at 313 K with initially C02-free (unloaded) pure water and aqueous AMP solutions and the plots for desorption were generated at 383 K with partially loaded aqueous AMP solutions. These results were generated using the kinetic parameters obtained in this work as discussed later in this chapter. Other model parameters were calculated using the correlations given in Appendices E to H. The C02-AMP-H20 is chosen just as an example. Similar results were obtained for other amines. Figures 5.4 and 5.5 show the predicted concentration profiles at 0 = 30°, 60° and 90° from the pole. As expected, for both absorption and desorption, the C02 profiles flatten with increasing distance from the pole. This is because the concentration driving force decreases from the pole to the equator. 97 1 e = 30° 2 0 = 60° -3 6 = 90° '. 0.0 0.2 0.4 0.6 0.8 1.0 Dimensionless Distance from Interface Figure 5.4: Predicted C02 profiles at different latitudes for C02 absorption in pure water (T = 313 K, QL = 2.0 mL/s, pco = 92.7 kPa) 1 e = 30° 2 e = 60° 3 0 = 90° 0 Q I i i i i I i i • i I i i • i i • • • • I • i—i—i—I 0.0 0.2 0.4 0.6 0.8 1.0 Dimensionless Distance from Interface Figure 5.5: Predicted C02 profiles at different latitudes for C02 desorption from aqueous AMP solution (T = 383 K, QL = 2.0 mL/s, CAMP = 2.11 kmol/m3, Ptotai = 220 kPa, a = 0.2 mole of C02/mole of amine) 98 Figure 5.6 shows the effect of amine concentration on the CO2 profile in the film under absorption conditions. Here again the results are as expected. When the AMP concentration is high, more CO2 in the solution reacts with the amine and very little or no dissolved C02 is left in the solution thereby creating a higher driving force for mass transfer. 25.0 20.0 ^ 15.0 E E o 10.0 o 5.0 0.0 " 1 1 ii T—r T—1 1— _j , , , r, ^ | | ( | 1 1 1 j— 1 = 1 wt% AMP -2 = 5 wt% AMP -3 = 20 wt% AMP : -—'— 1 1 1 1 1 1 1 1 • 0.0 0.2 0.4 0.6 0.8 Dimensionless Distance from Interface 1.0 Figure 5.6: Effect of amine concentration on CO2 profile at 9 = 90° for CO2 absorption in aqueous AMP solution (T = 313 K, QL = 2.0 mL/s, pco = 92.7 kPa) Figures 5.7 to 5.10 show the predicted concentration profiles of all chemical species (ionic and non-ionic) that may be present when C02 is absorbed or desorbed in or from aqueous AMP solutions. Figures 5.7 and 5.9 are the profiles for absorption and desorption, respectively, at 45° from the pole and 99 o E E CO o o X b o o X z •<r LL. X- -D + " ZJ co o" i =• or -o o O CN -^ 1 CO ° o d O c o o Figure 100 80 60 40 20 0 —1—1—1—1— A3 i i i i i i i i 2 —i—r—i—i—i—i—i—i—i—i—i—i— - \ / 1 = C02 : 2 = R4NH2 - \ y 3 = R4NH3+ ; • X \ 4 = R4NHCOO" : - \ \ 5 = HCO3" \ \ 6 = coy2 : \ \ 7 = OH" \ \ -7 : 4 ^ • • 0.25 0.20 ~ 0.15 0.10 o CM X z a: o c o 1 0.05 0 0.00 0.0 1.0 0.2 0.4 0.6 0.8 Dimensionless Distance from Interface 5.7: Predicted concentration profiles for C02 absorption in aqueous AMP solution at 0 = 45° (T = 313 K, QL = 2.0 mL/s, CAMP = 0.223 kmol/m , pco = 92.7 kPa) CO o o X I o o o X z + CO x 1—1 1—1—r o E E •g cr j c 0d ~ O 0 O CN -14— CO o o 0 O c o o Figure 1 = C02 0.20 2 = R4NH2 3 = R4NH3+ ; 0.15 4 = R4NHCOO- : 5 = HC03-6 = CO32 ; 0.10 7 = OH" 0.05 7 : 0.25 o E X Z 01 o c o O 0.00 0.0 1.0 0.2 0.4 0.6 0.8 Dimensionless Distance from Interface 5.8: Predicted concentration profiles for C02 absorption in aqueous AMP solution at 0 = 90° (T = 313 K, QL = 2.0 mL/s, CAMP = 0.223 kmol/m3, pco, = 92.7 kPa) 100 CM i CO o o X o I o o o X z CN o o o c o O o E •g cr 10.0 8.0 6.0 4.0 2.0 0.0 —1—1—1—1—1—1—1—1—1—1—1—1—1—1—1—1—1—r—1 1 1 1 1 1 "^^^ R4NH2 -co2 R4NHCOO" ; Y R4NH3+ co32 : HC03" 2.0 o E 1.6 1 CO o o 1.2 f CO X 0.8 0.4 CN X z or o c o 0.0 O 0.0 1.0 Figure 5.9: 10.0 8.0 6.0 h 4.0 2.0 0.0 0.2 0.4 0.6 0.8 Dimensionless Distance from Interface Predicted concentration profiles for C02 desorption in aqueous AMP solution at 0 = 45° (T = 383 K, QL = 2.0 mL/s, CAMP = 2.11 kmol/m3, Ptotai = 220 kPa, a = 0.2 mole of C02/mole of amine) X 0 CJI ' CO 0 0 0 b E 0 E 0 ' " X E z 0? id Fil CN O inbr O _i 4— c O O c 0 0 1 1 1 1 I 1 1 1 1 1 1 1 T 1 1 1 1 1 1 1 1 1 1 1 RJslHo -in 1 ^\ — - R4NHCOO C02 -y~~— R4NH3+ co32 : HCO3" - -2.0 ^ 1 0 1.6 1 CO O O X 1.2 + CO X z a: 0.8 CN X z 0.4 0 0' c O 0.0 O 0.0 0.2 0.4 0.6 0.8 1.0 Dimensionless Distance from Interface Figure 5.10: Predicted concentration profiles for C02 desorption in aqueous AMP solution at 0 = 90° (T = 383 K, QL = 2.0 mL/s, CAMP = 2.11 kmol/m3, Ptotai = 220 kPa, a = 0.2 mole of C02/mole of amine) 101 Figures 5.8 and 5.10 are the corresponding profiles at 90° from the pole. It can be seen from these figures that the concentrations of various species in the film vary both in the 9 and x directions. As expected, for absorption, the concentrations of free amine and hydroxyl ions decreases from the bulk to the gas-liquid interface and from the pole to the equator, whereas the concentrations of protonated amine, carbamate, bicarbonate and carbonate ions at these locations increases with x. For desorption, exactly the opposite occurs. This is because during absorption the amine and hydroxyl ions are consumed and the other species (i.e., protonated amine, carbamate, bicarbonate and carbonate ions) are formed, whereas, during desorption, amine is regenerated and the concentrations of the other species are depleted. 5.1.3 Numerical versus Analytical Solutions The general model (equations 4.64 to 4.97) cannot be solved analytically. However, when there is no reaction between the absorbing gas and liquid, the general model reduces to a case of physical absorption or desorption (e.g., CO2 or N20 in pure water). This situation is governed by a single partial differential equation, which using certain critical assumption can be solved analytically and an explicit expression for mass transfer rate can be derived (see Appendix C). Figure 5.11 shows a comparison of the C02 absorption rates predicted from the analytical and numerical solutions of the model equations. The analytical result was calculated from equation (4.98) and the numerical solution was calculated from the general model by setting all the rate constants equal to 102 zero. Clearly, the agreement between the numerical and analytical solution is very good. The relative error is between -0.4 to +4.4 percent. Slight discrepancy may be due to analytical solution assumption that penetration depth is much smaller compared to the thickness of the liquid film. 40 "0 E E CO O X </> CD -4—» CO DZ c o o 00 XI < 25 20 h 15 10 h 5 h ~i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 r •Numerical •Analytical 0 1—1—1—1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 i_ 285 295 305 315 325 Temperature (k) Figure 5.11: Absorption rates of CO2 in pure water predicted from analytical and numerical solutions (QL = 2.0 mL/s, pco = 87-97 kPa) The analysis presented above verifies that the present model accurately describes the absorption and desorption of a gas in and from the hemispherical film. It further confirms that the experimental technique and the numerical scheme used to solve the model equations are accurate and reliable. 103 5.2 Parametric Sensitivity Analysis The present model consists of three types of parameters: (a) the operating parameters such as amine concentration, temperature, total pressure, CO2 partial pressure, CO2 loading and liquid and gas flow rates, (b) physical property parameters such as density, viscosity, gas-side mass transfer coefficient, liquid phase diffusivities and Henry's constants and (c) kinetic parameters such as rate constants and equilibrium constants. In this section we examine the effect of these parameters on the rate of absorption and desorption of C02 into and out of aqueous amine solutions. 5.2.1 Effect of Operating Parameters In order to study the effect of operating parameters on C02 absorption and desorption rates, a series of simulation runs were carried out with C02-AMP as the example system. The simulation results are plotted in Figures 5.12 to 5.17. These results were obtained for the base case conditions listed in Table 5.1. To demonstrate the effect of each operating variable separately, only one of the operating parameters in Table 5.1 was varied while the others were fixed at their base value. The physical properties and the equilibrium constants were calculated from the correlations given in Appendices E to I and the rate constants were calculated from the correlations developed in this work. 104 Table 5.1: Base case operating conditions for parametric sensitivity analysis Parameter Value Absorption Desorption T(K) 303 373 Ptotal(kPa) 101.3 202.6 pC02(kPa) 97.0 1.5 CAM (wt%) 20.0 20.0 ocinitiai (mole/mole) 0.0 0.2 QL (mL/s) Z0 2.0 The results presented in Figures 5.12 to 5.17 are plausible and self-explanatory with the exception of the effect of amine concentration given in Figure 5.12. This figure shows that both the absorption and desorption rates increase with increase in amine concentration. However, it should be noted that this increase in desorption rate is not because of the increase in amine concentration itself but because of the increase in total C02 present in the solution. As this plot was generated at a fixed C02 loading, which in our case is defined as moles of C02 (in all its forms) per mole of amine. Figure 5.12 also shows that the rate of increase in the absorption rate diminishes when the amine concentration exceeds 30-wt%. This is likely, because at higher amine concentrations, the kinetic effects are offset by the reduced water content, higher solution viscosity and lower physical solubility of C02 in the solution. The results presented in Figures 5.12 to 5.17 clearly demonstrate that the absorption rate is most sensitive to changes in amine concentration, C02 partial pressure and temperature and the desorption rate is most sensitive to 105 temperature, C02 loading and C02 partial pressure in the stripping gas. Similar trends were observed for other systems studied in this work and for the sake of brevity those results are not included here. As will be discussed later in this chapter, the same trends were observed experimentally. Based on these observations, it was decided to conduct absorption experiments by varying mainly the amine concentration and temperature, and desorption experiments by varying the C02 loading and temperature. Figure 5.12: Effect of amine concentration on C02 absorption and desorption rates in aqueous AMP system 106 0.0 0.1 0.2 0.3 0.4 C02 Loading (mol of C02/mol of amine) Figure 5.13: Effect of C02 loading on C02 absorption and desorption rates in aqueous AMP system 270 290 310 330 350 370 390 Temperature (K) Figure 5.14: Effect of temperature on C02 absorption and desorption rates in aqueous AMP system 107 250 CO 15 200 E E, o 150 x CD CO 100 c o o CO -O < 50 0 0 T 1 1 1 1 1 1 1 1 1 r Desorption Absorption -i 1 1 1 r _l I L. 20 40 60 C02 Partial Pressure (kPa) 20 CO o 15 | X 10 0) ro or c o "-+-» r CL 5 o CO CU Q 0 80 Figure 5.15: Effect of C02 partial pressure on C02 absorption and desorption rates in aqueous AMP system 300 CO "o E 250 E, CO O o 200 ro UL c g Qr 150 o CO -Q < 100 -i 1 1 1—i—i—i—i 1—r -i—i—i—r • Desorption Absorption _i i i I_I i i_ _l I L 100 150 200 250 Total Pressure (kPa) 20 CO X 10 OJ ro CU c o CO CU Q 0 300 Figure 5.16: Effect of total pressure on C02 absorption and desorption rates in aqueous AMP system 108 300 -i—i—i—i—i—i—|—i—i—i—|—i—i—i—i—i—i—i—i—i—i—r 40 JO o E £ CO O CD -*—< CO CC c o o < 250 200 • •Absorption • Desorption o 30 CO O X 20 a) CO CC A 10 o </) cu Q -1 I I I I I I I I I I I I I I I I I L I I |_ 0 .1.8 2.0 2.2 2.4 2.6 Liquid Flow Rate (mL/s) 2.8 " 3.0 Figure 5.17: Effect of liquid flow rate on C02 absorption and desorption rates in aqueous AMP system 5.2.2 Effect of Physical Property Parameters Table 5.2 shows the effect of errors in the values of physical property parameters on predicted absorption and desorption rates for MEA, DEA, MDEA and AMP systems. This table was produced by introducing ± 20 or ± 50% deviations in the physical properties calculated from the correlations used in this work. For all systems the operating variables were kept the same as given in Table 5.1. The results clearly show that the predicted rates are most sensitive to C02 diffusivity in the amine solution, Henry's constant of C02 and to a lesser extent density of the solution. The effect of the diffusivities of amines and other ionic species on the absorption/desorption rates is not very significant. 109 c o V-» CL i_ o CO CD TD TD C CD C o Q_ O CO -Q CD CN O a. o 2 c < o TD CO C CD imete DEA cu u. CD CL < >%LU ert Q CL o < £1 LU CL TO 4— O o CO CO hy E CL CD •4—' _C CO >» c CO g CO "CD o "> CD CD T3 4— aq o c o CO CD CD Eff 2 o CO CD Q o CO < < < LU Q < 111 Q < LU < < LU Q < LU Q < LU 5 CD 25 CD Q — Q_ E CD i_ CD CL CD i O •4—» co + o CM CO 1 + 2 O Oi co i LO CM • CN i CM • LO CM CM O 4—' O O O 4—» i o co +~ CM + CM + O +~ LO + CO CM i • CM T- LO i O O -*—> o 1 o o CD +~ LO + LO + + CM + •<- CM 1 + ^ o CD + CD CM i CO 1 CM O o O 4—• 4—» •4—« CO CM CM + + + CD + + o CD + i o LO + co + o 4—» CM CM O CM + I O •4—' + CM O + o o CD O • + •2 p LO CO + d o co co CM + co d o o co + co + o •4—» CM CM LO O •4—' co a o + CD d + o -c—» LO d CM 1 + H LO CM i O -4—« O CO CM d o o d i co Q co CD + CD CM CO + CO co N-CO + CM CM + o CO o o 1^. o 4—• co + CM LO CO CM i CM CM CM i O O O O o •4—» "<* T— CM T— + + CM X— + + + o o o o o o o CM CM LO LO LO LO LO +1 +1 +1 +1 +1 +1 +1 This table does not show a significant effect of the gas side mass transfer coefficient on the predicted desorption rates. This is because, in the case examined here, the CO2 partial pressure in the stripping gas was less than 1 mol%. However, it can easily be shown that under certain operating conditions this effect may also become significant. As is discussed later in this chapter, in all our desorption experiments, we kept the C02 partial pressure in the stripping gas well below 5 mol%. From the above analysis we find that, in order to predict absorption and desorption rates accurately, it is important that the values of the C02 diffusivities and Henry's constants in amine solutions are known fairly accurately. For this reason we developed new correlations for these parameters that are applicable for both single and blended amine systems and cover wider temperature ranges than the available literature data. These correlations are presented in Section 5.3 (see also Appendices F and G). 5.2.3 Effect of Kinetic Parameters Table 5.3 lists all possible reactions with their associated kinetic parameters for the eight systems considered in this work. Most of these parameters are important to calculate absorption and desorption rates using our model. However, from our sensitivity analysis, we observed that, for carbamate-forming amines such as MEA, DEA and AMP, the contribution of reactions involving the deprotonation of amine-zwitterion to water and hydroxyl ions ill CO L_ 0) OD E CD s , m ID-o * CD C Ik CO c o o E ZJ To Z3 -g > T3 C =5 : CT ILU io o £ JZ o co II - o T- CD CM T-o co CM T-CO - CO CM 00 - CO CD LO O t— T— o * * * - o o ^ CO o - CD T- CM . V CM T-* - m LO T-- CO IO v-£^ i- CD CM T-O CO CM T-CD c5 £ - CD CD O CM CO CD ' co CO co* CM LO CM - 0) r» CM X. CD CO 1-- di CO i-JS. CD CO c o o CD A—* CD L— T3 CD C IXI E , o O o CO o CM o j£ CD -Mi OJ JZ co CO CD CO CM CO CD CM CM CO V - 6 CD a T— CD CD CD T~ JH ^ f-- CM ~~co - CD T- CM ^ ^ ^ ^ LO n CM CM T-j*: T- CM oo cn N M T-J±L ^ J£ CO CO Zl c > CD CM jsc: CM - CM f~- CM o T JXL r~- CM CD JUL CO C o o CD or CO E CD 00 iff o TT + O CM I + < LU O O oc3 . CM 03 ^ ^ CN 05 ^ CD Tf' TT r-~ eg in CM ^ CM h- CM in" CD CM LO CM oc3 -^t ^ CM °? TT ^"-^ O ° ---CD 1J r- TI-°8 ca CD + + + < LU Q + CM o o o CM I + < LU Q + CM o o X + Q_ < + CM o o + + o CM X + < LU Q + < LU + CM o o TT Tl-+ Tf Tl-+ 0« + o CM X + D_ < + < LU + CM o o X + < UJ Q + < LU Q + CM o o . CM CM -TT' TT' CD 00 T- T- CM TT' TT' TT' + + o CM X + D_ < + < LU Q + CM o o (reactions 4.3 and 4.4 for AMP, 4.12 and 4.13 for MEA and 4.18 and 4.19 forDEA) to absorption and desorption rates is insignificant. As shown in Table 5.4, a change of four orders of magnitudes in the values of the rate constants associated with these reactions produce no significant change in the predicted rate of absorption. Therefore, these reactions can be ignored without significant loss of accuracy. Similar conclusions were arrived at previously by Rinker et al. (1996) in their work on the kinetics of C02 absorption in aqueous DEA solutions. This simplification greatly reduces the complexity of the kinetic rate expressions and cuts down the time required for parameter estimation by one third as the number of unknown parameters in each expression reduces from 5 to 3. With the present model, parameter estimation with 5 unknowns is extremely time consuming. It takes about 15 to 20 hours to obtain the estimates for a single data set even with the fastest PC (Pentium 4, 1700 Mhz, 256 RAM) available to us. This is because each function call to solve the model equations takes about 2 to 3 minutes and a typical parameter estimation run makes hundreds of such function calls. After this simplification is implemented, the reactions and kinetic parameters that remain are summarized in Table 5.5. Table 5.4: Effect of reactions involving zwitterion deprotonation to water and hydroxyl ions on the rate of absorption (see reactions 4.3, 4.4, 4.12, 4.13, 4.18, 4.19) Parameters Calculated Absorption Rates x 103 (mmol/s) Value MEA DEA AMP ki2/k_io & ki2/k_io kis/k.16 & ki9/k-i6 k3/k.i & Wk-i 0-0 525.99 309.02 263.37 1-1 555.19 309.09 274.76 10-10 556.88 309.31 276.40 100-100 557.06 309.44 276.58 1000-1000 557.08 309.46 276.59 113 CO £ "CD E co ICL o CO c I* ~o . CD c CO ~ c -Q o E " o E " .2 °^ .Q lo 1= zs '5 ^ C7 > UJ 'TJ c o in T— cn 00 co CN cn CO CN cn CO 1^ cn 00 CO in o cn co ,y CN - O . co * o c5* . CN co in T-co CN o CN cn oo . > CN - CD . cn •<- CN CN V * ^o CN cn* co* * o5 „ CN co in T-0 -I—» co i_ TJ *i CD CO o CO CN o I \— in cvi CN o CN T-CO CN CO t~-J2 CO _ I V —>. CN ob ^ CO CN T-CO CO 3 c •o o > o =5 a) 2 o CO o V - CN CN O CO CO c o » o co or Tt Tt' T£ o ® CN CD CM h- T-Tt o -Tt Tt CM CM T£ ST h~ CM LO Tt Tt Tt 5,-2 ^<CM" CM X LO Tt t^- T— CM Tt Tt Tt + LO LO CM Tt CO . i- CM 1 Tt Tt CO CD CM O CM ^ Tt ^T- LO T— X CM Tt Tt Tt + I T- CM Tt Tt Tt + o " S CN Tt^^^ ^ <cT CM ^1 ^ Tt T— CD CM CM Tt Tt Tt Tt + ro o + + + CO E 0 -4—> CO + < LU O o o CN X + < LU Q + CN o o o CN X + < LU Q + CN o o o CN X + < + CN o o o CN X + < LU Q + < UJ + CN o o + o CN X + < + < LU + CN o o + o CN X + < LU Q + < UJ Q + CN o o + o CN X + CL < + < UJ Q + CN o o 5.3 Henry's Constant and C02 Diffusivity in Amine Solutions With the help of Table 5.2, it was shown earlier that, in order to predict absorption and desorption rates accurately, fairly accurate values of C02 diffusivities and Henry's constants in amine solutions are required. In this section we examine existing literature data for these properties and present new data and new correlations that are applicable for both single and mixed amine systems and cover wider temperature ranges. 5.3.1 Correlations for Henry's Constants of C02 in Amine Solutions Since C02 reacts with aqueous amine solutions, it is not possible to determine its physical solubility in these solutions by direct measurements and an indirect method based on the N20 analogy is commonly used. According to this method, the ratio of the Henry's constants of C02 and N20 in an aqueous amine solution is the same as that in water at the same temperature: H H° C02 _ C02 (5 1) Ll 110 ' ' nN20 nN20 where Hcc.2 and HN20 denote the Henry's constants of C02 and N20 in the aqueous amine solution, and Hcc,2 and H£ 0denote the Henry's constants of C02 and N20 in water. A large amount of data is available in the literature for the Henry's constant of C02 in water and N20 in water and various amine solutions as a function of temperature. These data are summarized in Tables F.1 to F.10. It can be seen from these tables that most of the existing data are in the low 115 temperature range (293-323 K) typical of an absorber, and the data at stripper operating temperatures (373-393 K) simply do not exist. In this work new data for the Henry's constant of CO2 in water and N2O in water and N20 in aqueous amine solutions were obtained. These results are also listed in Tables F.1 to F.10. The details of the experimental apparatus and procedure are given in Appendix F. The data for H°C02, H^o and HN20 (from this work and from the literature) were correlated as a function of temperature. These correlations are given in Appendix F (see equations F.2 to F.4). The correlations represent the experimental data well. The absolute average percent deviation was within 7% (see Figures F.1 to F.4). 5.3.2 Correlations for C02 Diffusivities in Amine Solutions Like Henry's constant, the diffusivity of CO2 in amine solutions cannot be determined directly and an indirect method based on the N20 analogy was used again. According to this method, the ratio of the diffusivities of CO2 and N2O in an aqueous amine solution is the same as that in water at the same temperature: Dco Dco D D° where Dcc,2and DN20 denote the diffusivities of CO2 and N20 in the amine solution, and D°COz and D^0 denote the diffusivities of C02 and N20 in water. As shown in Appendix G, good quality data are available in the literature for Dcc,2, D^o and DNz0. Some correlations for these data as a function of 116 temperature and amine concentration also exist in the literature. However, most of these correlations, specifically for DN20, are based on the data collected by individual authors valid only for narrow concentration and temperature ranges. Furthermore, in most of the published literature on CGramine kinetics, DN20 at various temperatures and amine concentrations is calculated using the following modified Stokes-Einstein relation: DN>°.2 - D°20la°280 • (5.3) In order to check the validity of equation (5.3), we compiled all the available diffusivity data for pure water, and single and mixed amine systems from various sources (see Tables G.1 to G.10) and prepared Stokes-Einstein plots as shown in Figures G.3 and G.4. The data from our own diffusivity measurements using the hemispherical contactor were also included in these plots. It can be seen from these plots that the Stokes-Einstein type relationship does not represent the experimental data well. The average deviation (AAD%) for single amine systems is about 20% and that for mixed amine systems is about 30%. Therefore, in this work, new correlations were developed which are valid over a wider range of temperatures and amine concentrations. These correlations are discussed in detail in Appendix G (see equations G.5 and G.6). Figures G.5 and G.6 show a comparison of the measured and calculated diffusion coefficients using new correlations. In general, the agreement is very good. The overall AAD% for single amine systems is about 13% and that for mixed amine systems is about 9%. In light of the fact that different sources have used different absorption apparatus to measure diffusivities, this much deviation 117 is expected and the correlations can be safely used. Note that for CAM = 0, these correlations (equations G.5 and G.6) reduce to the correlation for N20-H20 system (equations G.3). This unique and very important feature does not exist in other similar correlations proposed in the literature (Li and Lai, 1995; Li and Lee, 1996). 5.4 C02 Absorption and Desorption in Aqueous Amine Solutions This section presents the experimental and theoretical results from the work on C02 absorption and desorption in aqueous amine solutions. The experimental data are given in Tables K.11 to K.27 and the experimental conditions for each system studied are summarized in Table 5.6. The bulk of the experiments were focused on desorption because the purpose of these experiments was to estimate kinetic parameters under desorption conditions. The absorption experiments were carried out mainly to verify experimental techniques, as for most amines, good quality absorption data already exist in the literature. The data were analyzed using the diffusion-reaction model presented in Chapter 4. Based on the sensitivity analysis results presented above, it was assumed that for carbamate-forming amines (i.e., MEA, DEA, AMP), the contribution of reactions involving zwitterion deprotonation to water and hydroxyl ions is negligible. In all cases, the physical properties and equilibrium constants were calculated from correlations given in Appendices E to I. In all cases, the values of 118 o CD 0_ CM O in o CO o co |Q LO CO no LO t7J CO £ o _ o IE o LL CL o ^ I— * CD Tf oo Tf o CO LO Is-Tf CO CD CO Is-CM Tf CO O o d d d d d q O CM 4 o CD q d CM CD CM CM LO T— d d d d d d d d o o o q o q o o Tl-" LO Tf LO LO LO Tf Tf LO CM co LO CO 06 Is-o d d d d LO d LO d LO d LO o CM LO CM o CM LO CM o CM CM +~ LO CM +~ LO CM +~ LO CM T LO CM CM CM CM T— 00 Is-co CM 00 op CO LO co 00 Is-cp co Is-cp CO Is-cp CO Is-00 oo Is-co co CO CO CO Tf co CO Tf co co CO CO co LO co co LO co co LO co co LO CO CO CL c ^ I-B CM (II O CO l-Q -Q_ O E "o E o CL IS o o d d o o d d o d o d Is-CD • CD CD Is-CD Is-CD Is-CD CO oo CD oo o op op cp op 4 oo 4 CN co co co CM CO CM CO 1 CO 1 o o co co CO CD o CD CO CM CO CO O o CO CO q d co o co CO E "55 >> o CM X + < o CM X + < UJ LU Q O CM X + < LU Q + CM + CM o o o o X + CL < + CM o o o o o CM X + < LU Q + < o CM X + CL < + < + CM LU LU + + CM CM o o o o o d Is- Is- Is- Is-CD CD CD CD LO LO LO LO CM CM CM CM CO O CO + < LU Q + < LU Q + CM CM o o o o o CM X + < + < UJ Q + the forward rate constant for the CO2 hydration reaction (k7) were calculated from the correlation reported by Pinsent et al. (1956) and corrected for ionic strength using an equation developed by Astarita et al. (1983) (see equation H.2). 5.4.1 C02 Absorption/Desorption in Aqueous Solutions of MEA, DEA and The data interpretation for C02 absorption and desorption in aqueous solutions of MEA, DEA and AMP is discussed together as each of these amines forms carbamate and their reactions with C02 can be discussed based on the zwitterion mechanism. The major difference between AMP and MEA or DEA is that AMP-carbamate is not very stable and quickly hydrolyzes to give bicarbonate and free amine. The experimental results for the CO2-MEA system are reported in Tables K.11 and K.20, for the CO2-DEA system in Tables K.12, K.13 and K.21 and for the CO2-AMP in Tables K.15 and K.23. The simplified rate expressions for each of these systems are: CO2-MEA System: AMP 10-11 _ (5.4) CO2-DEA System: k 16 VfXi6,xi7 y (5.5) 16-17 1 + V *17 J w13 120 CCVAMP System: ^1 ]C3C4 VK1K2 J C2 (5.6) 1-2 Equations (5.4), (5.5) and (5.6) are essentially the same and therefore the procedure for parameter estimation is identical. In order to see if the kinetic parameters obtained under absorption conditions could be used to predict desorption rates we first used literature correlations for the rate constants involved in equations (5.4) to (5.6). For the combined equilibrium constants (i.e., K1K2, K10K11 and K16K17), the correlations given in Appendix H were used. For AMP, ki and k_i/k2 were calculated from the correlations given by Xu et al. (1996), for DEA, k16 and k_16/k17were calculated from the correlations given by Rinker et al. (1996) and for MEA, k-io was calculated from correlation of Hikita et al. (1977a) and k.i0/kn was set equal to zero because no such correlation is available in the literature. The results are plotted in Figures 5.18 to 5.26. These plots clearly demonstrate that the predictions based on literature correlations are fairly good for absorption rates (see Figures 5.18 to 5.20), but those for desorption rates are consistently higher than the experimental values by a factor of 2 to 6 (see 5.4.1.1 Predicted Rates using Literature Correlations 121 Figures 5.21-5.26). This was expected because the literature correlations have been developed based on the absorption data only. Also, desorption rates are very sensitive to the rate constants for the reverse reactions and accurate values of these constants are not available in the literature. 800 -J" 700 o E E, 600 CO T< 500 | 400 c 300 o CO § 200 -i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r 100 0.0 Figure 5.18: Experimental Predicted (Literature) Predicted (This Work) _i i i i i i i_ 1.0 2.0 3.0 4.0 CMEA (kmol/m3) 5.0 6.0 Predicted and experimental absorption rates for C02 absorption in aqueous MEA solution at 303 K. 122 500 CO I 400 E, °° 300 x cu ro 200 c g o. 8 100 < 0 -i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 r Experimental Predicted (Literature) Predicted (This work) i i i i l \ i i i _l I I I I I I l_ 0.0 Figure 5.19: 1.0 2.0 CDEA (kmol/m3) 3.0 4.0 Predicted and experimental absorption rates for CO2 absorption in aqueous DEA solution at 303 K. 600 .CO o E E «T 400 o v— X 2 cr I 200 Q. \ O CO JO < 0 ~i—1—1—1—1—1—1—1—1—1—1—1—1—1—1—1—1—1—r Experimental Predicted (Literature) Predicted (This work) _i 1 1 1 1 1 1 1 I |_ 1 1 1 1 • 1 1 0.0 Figure 5.20: 1.0 2.0 3.0 4.0 5.0 CAMP (kmol/m ) Predicted and experimental absorption rates for CO2 absorption in aqueous AMP solution at 303 K. 123 500 CO I 400 E, ? 300 x ro * 200 c o co 100 CD Q 0 ~1 1 1 1 1 1 r-Experimental (this work) Predicted (literature) Predicted (this work) -i i t_ 0.20 Figure 5.21: 100 0.30 0.40 Loading (mole of C02/mole of MEA) 0.50 Predicted and experimental desorption rates for C02 desorption from aqueous MEA solution at 378 K. CO o 80 E E ~> o 60 X "co CU 40 c o CL O CO 20 CD Q 0 ~i 1 1—i—i—i—i 1 1 1 1—i—i—r i Experimenta (this work) - - Predicted (literature) -i 1 r / Predicted (this work) 0.00 Figure 5.22: 0.05 0.10 0.15 Loading (mole of C02/mole of DEA) 0.20 Predicted and experimental desorption rates for C02 desorption from aqueous DEA solution at 382 K. 124 400 CO o I 300 CO O J2 200 co or ! 100 CO CD Q 0 Experimental (this work) Predicted (literature) Predicted (this work) 0.20 0.30 0.40 Loading (mol of C02/mol of AMP) 0.50 Figure 5.23: Predicted and experimental desorption rates for C02 desorption from aqueous AMP solution at 378 K. 400 ~o E E x CD -I—' CO or c 200 o CL O CO CD Q 100 •a CD o CD 0 i i i T r •' 'i T •• A • i i i i i "i 1^ r —i 1 r— A -A -'_ A -: / -• V / + T = 333 K • T = 343 K -r i i i i i i i x T = 353 K • T = 363 K -A T = 373 K i i i i i i AT = I i i 378 K ; i i i 0 100 200 300 Experimental Desorption Rates x 103 (mmol/s) 400 Figure 5.24: Predicted and experimental desorption rates for C02 desorption in aqueous MEA solution at 333 to 378 K using literature correlations 125 .CO ~o E E CO O X CD CO CL. cz o o CO CD Q T3 CD -+—1 O TJ CD CL 300 200 h 100 h 0 • T = 343 K x T = 353 K • T = 363 K A T = 373 K A T = 382 K i i i i i i i_ 0 100 200 300 Figure Experimental Desorption Rates x 10 (mmol/s) 5.25: Predicted and experimental desorption rates for C02 desorption in aqueous DEA solution at 343 to 382 K using literature correlations 5f 300 o E E, o O T— X 200 £ ro or c o -1—» CL t O CO 100 CD Q •D £ O T3 CD L_ CL 0 + T = 333 K x T = 353 K AT = 373K • T = 343 K • T = 363 K A T = 378 K 0 100 200 300 Experimental Desorption Rates x 10 (mmol/s) Figure 5.26: Predicted and experimental desorption rates for C02 desorption in aqueous AMP solution at 343 to 378 K using literature correlations 126 Minor deviations in the predicted absorption rates are understandable, as different authors have used different experimental techniques to measure absorption rates (see Tables 2.1 to 2.5). Note that, for MEA, the predicted absorption rates based on literature correlations are about 5 to 10% higher than the measured rates, for DEA they are about 20 to 25% less than the measured rates and for AMP, they vary from -10 to +20% of the measured rates (see Figures 5.18 to 5.20). This strongly suggests that the variation in predicted rates is due to the differences in experimental apparatus and the methods used for data interpretation and not due to some consistent error in our measurements. For example, Hikita et al. (1977a) used a pseudo-first-order model, Rinker et al (1996) used a rigorous model and Xu et al. (1996) used the zwitterion mechanism to analyze their data. 5.4.1.2 Parameter Estimates for MEA, DEA and AMP The absorption and desorption rate data for MEA, DEA and AMP were regressed using our model and the kinetic parameters were estimated. These estimates are presented in Tables (5.7) to (5.9) and plotted as a function of temperature in Figures (5.27) to (5.35). Note that all three parameters (e.g., k-i, k1/K1K2 and k_1/k2) were easily estimated from the desorption data. However, we were unable to obtain good estimates of combined rate constants (i.e., k1/K1K2and k_1/k2; k10/K^K.,., and k_i0/ku; k16/K16K17 and k_16/k17), which represent the reverse reaction in the zwitterion mechanism, from absorption data. This is understandable, because for 127 initially unloaded solutions, the values of these parameters are so close to zero that it becomes impossible to obtain reliable estimates. Therefore, in this work, we first estimated all three parameters for each case using desorption data and then used the correlations for combined constants based on these estimates to obtain forward rate constants under absorption conditions (see Tables 5.7 to 5.9). Table 5.7: Estimates of rate constants in eq. (5.4) from absorption and desorption data T ^10 k_i0 /k^ Method (K) (m3/kmol s) (1/s) (kmol/m3) 303.4 6,828.6 - - Absorption 333.2 45,108.0 14.6 23.0 Desorption 343.2 65,248.0 36.2 44.3 Desorption 353.4 149,720.0 198.1 148.9 Desorption 363.3 276,760.0 523.7 333.1 Desorption 373.2 364,820.0 1,512.8 776.4 Desorption 378.3 540,670.0 3,732.9 1,297.6 Desorption Table 5.8 Estimates of rate constants in eq. (5.5) from absorption and desorption data T k-i6 /K16K17 k_16 /k17 Method (K) (m3/kmol s) (1/s) (kmol/m3) 303.2 3,055.8 - - Absorption 313.2 4,085.8 - - Absorption 323.2 7,290.7 - - Absorption 343.2 9,139.0 33.9 1.2 Desorption 353.4 12,271.0 122.4 8.7 , Desorption 363.3 15,119.0 531.3 43.2 Desorption 373.2 17,058.0 1,515.6 110.0 Desorption 382.3 21,898.0 3,288.7 257.9 Desorption 128 Table 5.9: Estimates of rate constants in eq. (5.6) from absorption and desorption data T ki k_1/k2 Method (K) (m3/kmol s) (1/s) (kmol/m3) 296.1 924.3 - - Absorption 303.5 1,288.9 - - Absorption 308.3 1,368.1 - - Absorption 313.2 1,868.4 - - Absorption 318.2 2,427.2 - - Absorption 322.9 3,234.9 - - Absorption 333.4 4,264.2 108.5 4.5 Absorption 343.4 6,838.5 374.6 10.4 Desorption 353.5 8,561.9 1,480.6 22.5 Desorption 363.4 10,039.0 3,578.9 47.8 Desorption 373.3 15,386.0 9,651.1 94.4 Desorption 378.4 20,678.0 14,728.0 184.9 Desorption I I I I I I I I I I I I I I I I I • I I I I I f • From desorption data -| £+3 i i i i i i i • i i i i i i i i i i i i i <—i—i—i—i 2.4 2.6 2.8 3.0 3.2 3.4 1000/T(1/K) Figure 5.27: Arrhenius plot of the estimates for k10 from absorption and desorption data 129 10000 00 o 1000 k 100 t 10 -i—i—i—l—i—i—i—i—I—i—i—i—i—i—i—i—i—i—I—i—i—i—r k<° =9.539x10"expf-16'061' K10K11 E = 133.53 kJ/mol FT = 0.9917 T J *l I i i i i i i i i i_ j i i_ J i i i i_ 2.6 2.7 2.8 2.9 1000/T(1/K) 3.0 3.1 Figure 5.28: Arrhenius plot of the estimates for k^/K^K^ from absorption and desorption data 10000 1000 o J 100 -2> 10 t "1—i—i—i—I—i—I—I—i—I—I—I—i—i—i—I—i—i—i—I—i—i—i—r ^ = 4.231x10"exPr-12'647" V T E = 105.15 kJ/mol R2 = 0.9943 -I I I I I I I I I I I I I I I I I I I I I I I l_ 2.6 2.7 2.8 2.9 1000/T(1/K) 3.0 3.1 Figure 5.29: Arrhenius plot of the estimates for k^o/k^ from absorption and desorption data 130 1E+5 F ^ 1E+4 o E CO CO * 1E+3 1E+2 i i i I i 1 r- I i —r —i 1 1 1 i 1 1 1 1—: • From absorption data -• From desorption data -Z - k16 = 3x107 exp 2759 ; \ E = 22.94 kJ/mol -; R2 = 0.979 • i i i i i i i i i • i • i i i i • i 2.5 2.7 2.9 3.1 1000/T(1/K) 3.3 3.5 Figure 5.30: Arrhenius plot of the estimates for k16 from absorption and desorption data CO CO 1E+5 1E+4 h 1E+3 k * 1E+2 \E 1E+1 h ~l 1 1 1 1 1- ~I I | I ! ! I'1 "|"""' f—-|——"| • | | | I I " ' I k.e o..,n2i_J-15,638^ K16K17 2x10^ exp v J E = 130.01 kJ/mol R^ = 0.9952 *| £+0 I i • i i I • i i i i i i i i i i i i i I 1 i i >-2.5 2.6 2.7 2.8 2.9 3.0 1000/T(1/K) Figure 5.31: Arrhenius plot of the estimates for k16/K16K17 from absorption and desorption data 131 10000 1000 k ^ 100 Lr E 10 h 1 t ni i i i i i i i | i i i i i i "i i i | i i i k_ie o ,^22 f-17,920^ 16 -8x10 exp E = 148.99 kJ/mol VX = 0.9805 Q I i I i I I i I I i I i I I I I I I i I I I [ i I I i I i i I i • i i 2.6 2.7 2.7 2.8 2.8 2.9 2.9 3.0 1000/T(1/K) Figure 5.32: Arrhenius plot of the estimates for k_16/k17 from absorption and desorption data 1E+5 « 1E+4 o E 1E+3 : i i i i —r i 1 1 1— ~ r i 1 1 1 1 1 1 1—: . *r,9 f-4117.2^ ! k1 =1x10 exp ; V t ) -E = 34.23 kJ/mol R2 = 0.9924 : • • From absorption data - • From desorption data • i i i • i i i i • i i i i i i i i i 1E+2 2.5 2.7 2.9 3.1 3.3 3.5 10007T(1/K) Figure 5.33: Arrhenius plot of the estimates for ki from absorption and desorption data 132 CM -I—i—i—i—|—i—i—r 1E+5 1E+4 1E+3 t 1E+2 •j I—'—i—i—" i < i • • i i i i i i i i i i i • • • • 2.6 2.7 2.8 2.9 10007T(1/K) 3.0 3.1 Figure 5.34: Arrhenius plot of the estimates for k1/K1K2 from absorption and desorption data 1E+4 1E+3 o § 1E+2 CM 1E+1 ~\—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—• k , _ ,rt13 f-10,041^ -^- = 5x1013exp k, E = 83.48 kJ/mol FT = 0.9936 •\ r£+fj '—1—1—1 1 1 • • 1 1 ' • 1 1 1 ' 1 1 1 1 ' 1 1 1 *-2.6 2.7 2.8 2.9 3.0 3.1 1000YT(1/K) Figure 5.35: Arrhenius plot of the estimates for k_1/k2 from absorption and desorption data 133 5.4.2 C02 Absorption/Desorption in Aqueous MDEA Solutions The experimental results for the C02-MDEA system are reported in Tables K.14 and K.22. As stated earlier in Chapter 4, MDEA does not react directly with C02 but catalyzes the C02 hydration reaction. The rate expression for this reaction is given by: k r22 =-l<22C1C16 + — C5C17 (5.7) K22 Equation (5.7) has two unknown parameters (i.e., k22, k_22 or k22/K22), which were determined from the absorption and desorption rate data given in Tables K.14 and K.22. Here again, first, the existing correlations for these parameters were used to compare the predicted absorption and desorption rates with our experimental data. The correlation to calculate K22 was taken from Appendix H and that for k22 was obtained from Rinker et al. (1995). The latter was developed based on absorption data only. These results are presented in Figures 5.36 to 5.38. In this case, the predictions based on the literature correlations are fairly good for both absorption and desorption rates. This is probably because the rate of the C02 reaction with MDEA is much slower compared to that MEA, DEA and AMP. In the next step, we used our model to estimate k22 and k.22 by using the parameter estimation technique implemented earlier for MEA, DEA and AMP. These estimates from absorption and desorption data are listed in Table (5.10) and plotted as a function of temperature in Figures 5.39 and 5.40. 134 100 In -(mmol 80 -CO O 60 -X -CD -ion Ral 40 -CL Absoi 20 -0 ~i i i i 1 I 1 I I 1 i i i 1 1 1 i 1 r 0.0 Predicted (Literature) Predicted (This work) Experimental -I I I I I I L_ 1.0 2.0 CAMP (kmol/m3) 3.0 4.0 Figure 5.36: Predicted and experimental absorption rates for C02 absorption in aqueous MDEA solution at 303 K. 200 CO o 160 E ^ 120 h x CD -f—' * 80 I-c g S 40 h <D Q 0 "i i i ii ™" i" i i i i ii i i | i ri i III i r~ • Experimental (this work) Predicted (literature) Predicted (this work) -i i i i i i i i i i i • • i ' • i i i i i i i i_ 0.05 Figure 5.37: 0.10 0.15 0.20 0.25 Loading (mol of C02/mol of MDEA) 0.30 Predicted and experimental desorption rates for desorption from aqueous MDEA solution at 378 K. 135 0 20 40 60 80 100 Experimental Desorption Rates x 103 (mmol/s) Figure 5.38: Predicted and experimental desorption rates for CO2 desorption in aqueous MDEA solution at 343 to 378 K using literature correlations Table 5.10: Estimates of rate constants in eq. (5.7) from absorption and desorption data T k22 k_22 Method (K) (m3/kmol s) (m3/kmol s) 303.2 9.9 - Absorption 343.4 72.2 2.3 Desorption 353.2 123.3 6.1 Desorption 363.3 194.3 14.9 Desorption 373.3 357.6 35.9 Desorption 378.3 461.0 51.0 Desorption 136 1000 2. 100 o E 10 i—i—i—i—I—r -i—I 1—i 1—i—i—|—I—i—i—i—|—i—i—I—r l9 r-5797.8^ k22 =2x10 exp • From absorption data • From desorption data _l I I I I I I • I » I L E = 48.20 kJ/mol R2 = 0.9922 J J i i i i i i i i i— 2.5 2.7 2.9 3.1 10007T(1/K) 3.3 3.5 Figure 5.39: Arrhenius plot of the estimates for k2  from absorption and desorption data 100 oo o 10 b -i—i—i—r 2.5 -| I I I I I I I I I I I l_ l_ _l I I I I I I I I L_ 2.6 2.7 2.8 1000/T(1/K) 2.9 3.0 Figure 5.40: Arrhenius plot of the estimates for k-22 from absorption and desorption data 137 5.4.3 C02 Absorption/Desorption in Aqueous Amine Blends The experimental data for C02 absorption and desorption in four amine blends (MEA+MDEA, MEA+AMP, DEA+MDEA and DEA+AMP) studied in this work are presented in Tables K.16 to K.19 and K.24 to K.27. These data were fitted to the overall rate expressions given in Table 5.5. It can be seen from this table that most of the unknown kinetic parameters have already been estimated using single amine systems. The only remaining unknowns in these expressions are: k24/k_10 for MEA+MDEA blend; k25/k_-i and k26/k_io for MEA+AMP blend; k27/k_16for DEA+MDEA blend and k2a/k-i and k29/k-i6 for DEA+AMP blend. These parameters were estimated by regressing desorption data of corresponding blend at 353 to 373 K. The results are listed in Table (5.11) and plotted as a function of temperature in Figures (5.41) to (5.46). Table 5.11: Estimates of the combined rate constants for amine blends from desorption data T (K) k-io/k24 (kmol/m3) k-i/k25 (kmol/m3) k-i0/k26 (kmol/m3) k.16/k27 (kmol/m3) k_i/k28 (kmol/m3) k-i6/k29 (kmol/m3) 353.0 427.2 2,700.0 788.43 63.8 1,006.9 93.09 363.0 738.9 9,360.0 1,305.60 178.3 3,270.3 184.82 373.0 1,473.8 23,400.0 3,048.40 897.2 15,366.0 394.80 138 10000 o J 1000 100 —i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r k 10 .... „ni3 f-8494.1A -^- = 1.111x1013exp 2.5 2.6 v24 V E = 70.62 kJ/mol J FT = 0.9939 —i i i i i i i i i i i i i_ 2.7 2.8 10007T(1/K) 2.9 3.0 Figure 5.41: Arrhenius plot of the estimates for k_-io/k24 from desorption data 100000 F 10000 F CM 7 1000 F 100 n—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r _J I I I I I l_ 2.5 2.6 ^ = 1x1021 exp -14,313 E = 119.0 kJ/mol FT = 0.9951 _i i i i i i i i i i i i i i_ 2.7 2.8 1000/T(1/K) 2.9 3.0 Figure 5.42: Arrhenius plot of the estimates for k_i/k25 from desorption data 139 10000 1000 b CO CN -i—i—i—i—i—i—i—r "T I l l I |" l 1 l l I I ll I k 10 -r „ni3 f-8926.6^ —^- = 7x101 exp[ k26 V T E = 74.22 kJ/mol FT = 0.9736 <| QQ I I i I i I I I I I 1 I I I I I I I i i L_i 2.5 2.6 2.7 2.8 2.9 3.0 1000/T(1/K) Figure 5.43: Arrhenius plot of the estimates for k.10/k26 from desorption data 10000 E 1000 o E •~cb 100 t 10 T—i—i—i—i—r 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 !--16 o A AC. A rv23 k_„ f-17572.4^ -2.445x10"J exp k27 V J E = 146.10 kJ/mol Rz = 0.9801 j i i i i i i i i i i i i i i_ _i i i i i i_ 2.5 2.6 2.7 2.8 100WTM/K) 2.9 3.0 Figure 5.44: Arrhenius plot of the estimates for k.i6/k27 from desorption data 140 100000 nE 10000 1000 h 100 "I—I—I—I—l—I—I—I—I—I—I—I—r—I—r—i—i—-i—I—i—I 1 1—r-k 1 o ,o25 f-i8,327^ = 3 x 1025 exp| ^28 E = 152.37 kJ/mol FT = 0.9896 -i i i i i i_ 2.5 2.6 2.7 2.8 1000/T(1/K) 2.9 3.0 Figure 5.45: Arrhenius plot of the estimates for k_i/k28 from desorption data 1000 o I 100 cn CM -co -i—i—i—i 1—i—i-n—i—i—i—i—i—i—i—i—r -| 0 I I I I I I L. 2.5 2.6 k ie o „rti3 f-9727.9N 16 =8x1013exp E = 80.88 kJ/mol R2 = 0.9971 _i i i i i • i ' i • i i 2.7 2.8 2.9 3.0 1000VTM/K) Figure 5.46: Arrhenius plot of the estimates for k.^k29 from desorption data 141 5.4.4 Comparison of the Parameter Estimates with Literature Data Figures 5.47 to 5.50 compare the forward rate constants (i.e., k10, k16, ki and k22) obtained in this work with those from the published literature. The estimates from this work are based on both absorption and desorption data whereas the literature values are from absorption data only. Clearly, our estimates from absorption data are in good agreement with those reported in the literature. These figures also show that in most cases the second-order rate constants obtained under absorption conditions cannot be extrapolated to desorption temperatures with good accuracy except for C02-MDEA-H20 system (see Figure 5.50). The reason for much better extrapolation of kinetic data for C02-MDEA-H20 system to desorption temperatures is because of very low reactivity of C02 in aqueous MDEA solution compared to MEA, DEA and AMP solutions. From these results, it can be concluded that to obtain good kinetic data, experiments must also be conducted under desorption conditions. In fact, the reverse rate constants can only be obtained accurately from desorption experiments. Table 5.12 shows a comparison of the activation energies of ki, ki0, ki6 and k22 determined in this work with those reported in the literature. It can be seen from this table that the activation energies determined in this work are slightly higher than the published values. This discrepancy could be because our correlations are based on both absorption and desorption measurements covering a much wider temperature range (from 303 to 382 K) as opposed to 142 293-323 K for the literature correlations that were developed based on absorption data only. The higher activation energies of the combined rate constants indicate that the reverse reaction is a strong function of temperature and to regenerate the solution the amine solutions must first be heated to sufficiently high temperatures. As expected, the activation energy for the reverse reaction is highest for MEA and lowest for MDEA, indicating thereby that MDEA is lot easier to regenerate than MEA. 1E+6 p 1E+5 CO "o E J1E+4 1E+3 + o X • A X A Astarita (1961) Danckwerts & Sharma (1966) Sada et al. (1976) Alvarez-Fuster et al. (1980) Donaldson & Nguyen (1980) Laddha & Danckwerts (1981) Littel et al. (1992b) This work (Desorption) This work (Absorption) •Correlation of Hikita et al. (1977a) • Correlation (This work) * + o I I I I I I I I 1 I I I I I I _1 I I I I I 2.4 2.6 2.8 3.0 3.2 1000/T(1/K) 3.4 3.6 Figure 5.47: Comparison of the second-order rate constant (k-|0) for CO2-MEA reaction determined in this work with those reported in the literature 143 1E+5 c 1E+4 b CO "o j=£ 1E+3 1E+2 1E+1 • Blanc & Demarais (1984) • x Versteeg & van Swaaij (1988) + x Littel etal. (1992b) A • This work (Desorption) • — Correlation (This work) Baluwhoffetal. (1984) Versteeg and Oyevaar (1989) Rinker et al. (1996) This work (Absorption) 2.4 Figure 5.48: 2.6 2.8 3.0 1000/T(1/K) 3.2 3.4 3.6 Comparison of the second-order rate constant (k-|6) for CCVDEA reaction determined in this work with those reported in the literature 144 1E+5 F 1E+4 00 "o I 1E+3 |r CO E 1E+2 1E+1 2.4 A Messaoudi & Sada (1996) + Saha etal. (1995) o Yih& Shen (1988) • Sharma (1965) • This work (Absorption) _l I I I I I I !_ A Xu etal. (1996) x Alper (1990) x Chakraborty et al. (1986) • This work (Desorption) — Correlation (This work) A x X 2.6 2.8 3.0 1000/T(1/K) 3.2 3.4 3.6 Figure 5.49: Comparison of the second-order rate constant (ki) for C02-AMP reaction determined in this work with those reported in the literature 145 1000 r 100 X • A Versteeg & van Swaaij (1988) Tomcej & Otto (1989) Rangwala et al. (1992) Rinker et al. (1995) This work (Desorption) This work (Absorption) •Correlation (This work) 10 h -J I I I I I I I L. J L _l I I I I I I 2.4 2.6 2.8 3.0 3.2 1000/T(1/K) 3.4 3.6 Figure 5.50: Comparison of the second-order rate constant (k22) for C02-MDEA reaction determined in this work with those reported in the literature Table 5.12: Comparison of the activation energies of k-i, k10 and ki6 Parameter Activation Energy (kJ/mol) This work Literature Values Source ki 34.23 24.26 Xu etal. (1996) km 57.07 41.20 Rinker etal. (1996) k« 22.94 14.14 Hikita etal. (1977a) k22 48.20 39.00 Rinker etal. (1995) 146 5.4.5 Predicted Absorption/Desorption Rates Based on Correlations Developed in this Work Figures 5.51 to 5.58 compare the predicted and experimental desorption rates based on the correlations developed in this work. In general the agreement between the predicted and measured rates is very good. The percent absolute average deviations for MEA, DEA, AMP and MDEA are 8.71, 10.34%, 6.98% and 6.39% respectively. Similar agreement exists for mixed amine systems. The average deviation for the latter was less than 6%. This further validates the reaction mechanisms assumed and the experimental techniques used to estimate the kinetic parameters, and confirms that the theory for gas absorption with reversible chemical reactions can be applied to predict desorption rates. A more rigorous error analysis based on 95% confidence interval of the parameters estimates could not be performed because of the limitation of the parameter estimation software used in this work. In the next step, the model with the correlations developed in the previous section was used to predict the C02 absorption rates in 25 wt% amine blends namely: MEA+MDEA, DEA+MDEA, MEA+AMP, DEA+AMP. The purpose of this exercise was to investigate the effect of an addition of MEA and DEA on the absorption rates of MDEA and AMP. The predicted and experimental results are plotted in Figures 5.59 to 5.62. The agreement was satisfactory. However, more work is required to improve the capability of the model to handle mixed amine solutions. It can be seen from Figures 5.59 to 5.60, that small addition of MEA or DEA (< 5 wt%) can substantially enhance the rate of absorption of C02. 147 00 co o X CD -*-» CO CU c o 160 120 h 80 h e- 40 b o 00 CD Q 0 0.2 0.3 0.4 Loading (mole of C02/mole of MEA) 0.5 Figure 5.51: Predicted and experimental desorption rates C02-MEA as a function of temperature and C02 loading 00 140 i—i—i—i—i—i—i—i—i—i—i—|—i—i—'—i—i—i—1—i—i—•~—1—i—'—'— 0 _l I I I I L + T = 333 K o T = 343 K x T = 353 K • T = 363 K A T = 373 K A T = 378 K • ' ' • • i i i i i i i i i i i i_ 20 40 60 80 100 120 140 Experimental Desorption Rates x 10 (mmol/s) Figure 5.52: Predicted versus experimental desorption rates C02-MEA calculated from the correlation developed in this work 148 -C/> o E E, CO O X CD CD or c o 100 75 h 50 h fr 25 CO CL) Q 0 0.0 Figure 5.53: 0.1 0.2 0.3 0.4 0.5 Loading (mole of C02/mole of DEA) Predicted and experimental desorption rates for CO2-DEA system as a function of temperature and C02 loading o E E 80 x CD 60 0T c o o. 40 o co CD Q -o o "0 •0 CD 20 0 • III , 11 ....]——,. , ^ , . 11 1 1 1 1 v -• - • T = 343 K x T = 353 K " • T = 363 K A T = 373 K -/*' 1 A T = 382 K 1 1 i 1 1 1 1 1 1 • 1 1 1 1 1 0 20 40 60 80 100 Experimental Desorption Rates x 10 (mmol/s) Figure 5.54: Predicted versus experimental desorption rates C02-DEA system calculated from the correlation developed in this work 149 1 00 I—i—i—i—i—|—i—i—i—i—i—i—i—i—i—|—i—i—i—i—i—i—i—i—i—I—i—i—i—r 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Loading (mol of C02/mol of AMP) Figure 5.55: Predicted and experimental desorption rates for C02-AMP system as a function of temperature and C02 loading «T 100 80 h E 60 Q. 40 20 0 0 -i 1 1 1 1 1 1 1 1 1 1 1 1 r + T = 333 K x T = 353 K A T = 373 K • T = 343 K • T = 363 K A T = 378 K 20 40 60 80 100 Experimental Desorption Rates x 10 (mmol/s) Figure 5.56: Predicted versus experimental desorption rates C02-AMP system calculated from the correlation developed in this work 150 T II I | 11 1 | | "|-..| • T • |' | | | i | | | J | I | I I I I J I I r~~T • T = 343K ; x T = 353K -• T = 363 K -A T = 373K ; A T = 378K " — Model • Q I I t I I 11 i • i i i i i i • i i i i i i i i • i i 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 Loading (mol of C02/mol of MDEA) Figure 5.57: Predicted and experimental desorption rates for C02-MDEA system as a function of temperature and C02 loading 0 20 40 60 80 100 Experimental Desorption Rates x 103 (mmol/s) Figure 5.58: Predicted versus experimental desorption rates C02-MDEA system calculated from the correlation developed in this work 151 When the MEA concentration was 20% of the 25 wt% MEA+MDEA blend, the absorption rate increased by about a factor of 5. Similarly, when the DEA concentration was 20% of the 25 wt% DEA+MDEA blend, the absorption rate went up by about a factor of 3. Addition of MEA to 25 wt% AMP blend does improve the absorption capability of AMP, however, this improvement is nominal. No improvement in the C02 absorption rate was observed when DEA was added to 25 wt% AMP blends. This is reasonable because the reactivities of these amines are comparable. The results for MEA+MDEA and DEA+MDEA mixtures are important. With the help of the model, an optimal mixture concentration could be obtained so that the desired amount of C02 slip could be achieved in the absorber column. This is significant in natural gas treating applications where one may not want to absorb too much C02 in the absorber to avoid unnecessary dilution of the acid gas going to the sulfur recovery plant. 152 800 100 •L 80 H 60 o cc c cu E CU 40 ° c co JZ. c UJ A 20 0 0 20 40 60 80 100 Cone, of MEA in 25 wt% MEA+MDEA Blend (% of total amine) Figure 5.59: Predicted versus experimental absorption rates for CO2-MEA+MDEA system at 303 K based on the correlation developed in this work (pco =97 kPa) 500 to o 400 h 0 Experimental Model 100 80 60 40 H 20 0 o -*—» o CO c cu E CD O c CO JZ c LU 0 20 40 60 80 100 Cone, of DEA in 25 wt% DEA+MDEA Blend (% of total amine) Figure 5.60: Predicted versus experimental absorption rates for CO2-DEA+MDEA system at 303 K based on the correlation developed in this work (pco =97 kPa) 153 800 -52 o £ 600 [• J2 400 h or c o e-200 h 00 < 0 "I 1 1 1 1 1 1 1— • Experimental -I 1 1 1 1 1 1 1 1 r-Model E "V" -I I l_ _l I I l_ 100 1 80 60 40 20 0 0 20 40 60 80 100 Cone, of MEA in 25 wt% AMP+MEA Blend (% of total amine) Figure 5.61: Predicted versus experimental absorption rates for C02-MEA+AMP system at 303 K based on the correlation developed in this work (pco =97 kPa) 400 £ 200 CO or c o o CO < 100 0 "T 1 1 1 1 1 1 1 1 1 1—-1 1 r—I 1 1 1— • Experimental Model E _i i i i • i i i_ _i i i i i i i_i i_ 100 80 60 40 20 0 20 40 60 80 100 Cone, of DEA in 25 wt% AMP+DEA Blend (% of total amine) Figure 5.62: Predicted versus experimental absorption rates for C02-DEA+AMP system at 303 K based on the correlation developed in this work (pco =97 kPa) 154 5.5 Conclusions The parametric sensitivity analyses indicate that accurate values of CO2 diffusivity, and the physical solubility of CO2 given by the Henry's law constant are essential for predicting accurate absorption and desorption rates. An extensive literature search was undertaken and all available property data were compiled. It was found that the values of these parameters at stripping temperatures are not available. A series of experiments involving IS^O-water and N2-alkanolamine systems were conducted to measure the G02 diffusivities and Henry's constants in water using the N20 analogy. Using data obtained in this work and those available in the literature, correlations were developed that can be used for pure water, as well as for single and binary amine systems covering a wide range of temperatures and amine concentrations. Such correlations are not available in the literature. The absorption and desorption rates of C02 in aqueous solutions of MEA, DEA, MDEA, AMP and their mixtures (MEA+MDEA, MEA+AMP, DEA+MDEA and DEA+AMP) were measured for amine concentrations in the range of 2 to 35 wt%. The absorption experiments were carried out at near atmospheric pressure using pure CO2 saturated with water at 293 to 323 K with initially unloaded solutions. The desorption experiments were performed at 333 to 383 K for CO2 loadings between 0.02 to 0.7 moles of C02 per mole of amine using humidified N2 gas as a stripping medium. The data were analyzed using the rigorous diffusion-reaction model developed in this work. The model predicts the experimental results well for all eight different amine systems discussed above. 155 The results indicate that the theory of absorption with reversible chemical reaction could be applied to predict desorption rates. The zwitterion mechanism adequately describes the reactions between CO2 and carbamate-forming amines such as MEA, DEA and AMP under both absorption and desorption conditions. The reactions between C02 and aqueous MDEA solutions are best described by a base-catalyzed hydration reaction mechanism. The kinetic data obtained show that desorption experiments can be used to determine both forward and backward rate constants accurately. The absorption experiments on the other hand could only be used to determine forward rate constants. For MEA, DEA and AMP, the kinetic data obtained under absorption conditions do not extrapolate well to desorption temperatures. Therefore, kinetic data at higher temperatures should be obtained from desorption experiments. The existing absorption data for the CO2-MDEA system can be extrapolated to desorption temperatures with reasonable accuracy. This is probably because MDEA is very slow reacting amine and does not form carbamate. For the blended amine systems, it was found that small additions of MEA or DEA (< 5 wt%) significantly enhance the absorption rates of CO2 in 25 wt% MDEA solutions. Addition of small quantity of MEA (< 5 wt%) to 25 wt% AMP blend, on the other hand, was found to have very nominal effect on the C02 absorption rates of AMP. No improvement in absorption rates was observed when DEA was added to 25 wt% AMP solutions. 156 CHAPTER 6 INTRODUCTION AND LITERATURE REVIEW 6.1 Background In Part 1 of this thesis, we focused mainly on determining the kinetic parameters for fast-reacting gas-liquid systems by measuring the absorption and desorption of C02 in aqueous amine solutions using a novel hemispherical contactor. In Part 2, we present a novel approach to determining kinetic coefficients for extremely slow reactions. The system under consideration is carbon monoxide absorption and its subsequent reactions with aqueous diethanolamine. The solubility of CO in pure water at 298 K and 1 bar is about 40 times less than that of C02 in water and its reactions with aqueous DEA at 298 K are a few orders of magnitude less than those of C02 with aqueous DEA. Why is it important then to study the kinetics of such slow reactions in gas treating applications? The answer to this question lies in the fact that, in gas treating units, the solvents are regenerated and reused over extended periods and are exposed to varying temperatures, about 313 K in the absorber and about 393 K in the stripper, reboiler and heat exchangers. Thus, feed gases containing appreciable quantities of CO (e.g. gas mixtures from hydrogen plants and fluid catalytic cracker units) may lead to significant absorption of CO in the amine solution at absorber temperatures that in turn may cause reactions of CO with 157 hydroxyl ions and DEA to form formate ions and formyl-diethanolamine (DEAF) respectively. These reactions may become particularly important at stripper and reboiler temperatures resulting in significant losses of DEA. The degradation of DEA is costly and it interferes with the proper running of gas treating units. Therefore, it is quite important to understand how CO reacts with aqueous DEA solutions and what major degradation products it forms, so that DEA losses can be quantified and preventive measures taken. Since the reactions between CO and aqueous DEA solutions are extremely slow, a short contact time apparatus like the hemispherical contactor cannot be used to study the kinetics of these reactions. The best way to monitor the progress of these reactions is to conduct experiments in a batch autoclave reactor of the type described in Chapter 7. Although the subject matter of Part 2 is slightly different from that of Part 1, the modeling and parameter estimation approach is similar. The knowledge acquired during the C02 absorption/desorption study was extremely useful in approaching this second problem. This work was funded by Equilon Enterprises LLC, Westhollow Technology Center, Houston Texas and Shell Global Solutions International, Amsterdam, The Netherlands. 6.2 Literature Review Aqueous solutions of diethanolamine (DEA) are widely used to remove acid gases, such as C02 and H2S, from sour gas mixtures including natural gas, synthesis gas, flue gas and various refinery gas streams. The efficiency of these 158 processes largely depends on the performance of the amine solution, which is continuously regenerated and used over extended periods. Although the principal acid gas-amine reactions are reversible, irreversible reactions may also occur resulting in products from which the amine cannot be regenerated under typical operating conditions. This phenomenon is called amine degradation. Numerous papers are available on the degradation of alkanolamines under the conditions typical of gas purification plants. However, most papers focus on alkanolamine degradation due to C02, carbon disulfide (CS2), carbonyl sulfide (COS) and H2S/C02 mixtures. Some sour gas mixtures (e.g. gas mixtures from hydrogen plants and catalytic crackers) contain significant quantities of carbon monoxide (CO) that may also cause amines degradation. A survey of the literature reveals that CO is a strong reducing agent which is only slightly soluble in water and other physical solvents (< 0.001 mole CO/mole of water at 298 K and CO partial pressure of 101.325 kPa, Kirk Othmer, 1993; Fogg and Gerrad, 1991). An industrially significant process is the reaction of CO with secondary amines to produce formamides. For example, the industrial solvent dimethylformamide (DMF) is manufactured by reacting CO with dimethylamine at 423-428 K and CO partial pressures of 7.5-10.3 MPa in the presence of a catalyst (Duranleau and Lambert, 1985): (CH3)2NH + CO metalcata|ysl )(CH3)2NCHO Similar reactions occur between alkanolamines and CO to produce formylalkanolamines. For instance, DEA is known to react with CO at high pressure (pco= 6.2 MPa) and temperature (423 K) over a 5 hour period to 159 produce (C2H4OH)2NCOH, i.e. formyldiethanolamine or DEAF (Lambert and Duranleau, 1985). These conditions are very severe and are not likely to be met in normal refinery and gas plant operations. However, DEAF has been found in DEA solutions used to treat gases from fluid catalytic crackers (FCC units). Koike et al. (1988) have suggested that DEAF is formed by the reaction between formic acid (or formate ions) and DEA according to: HCOOH + (HOC2H4)2NH-^(HOC2H4)2NCOH + H20 The formic acid (or formate ions) may be present in the amine solution because of the reaction between CO and OH- ions, and by the hydrolysis of HCN generally present in the FCC dry gas. In another study, Kim et al. (1988) examined the absorption rate of CO into aqueous solutions of potassium carbonate (K2C03), methyldiethanolamine (MDEA) and diethylethanolamine (DEAE) in a stirred tank reactor at 348-398 K and CO partial pressures of 0.75-3.1 MPa. There were no mass transfer limitations and the reaction between CO and hydroxyl ions (OH") to produce formate ions (HCOO") was found to control the rate of absorption. It was concluded that the rate of CO absorption into aqueous solutions of K2C03, DEAE and MDEA are much slower (by a factor 108) than the corresponding rates of C02 absorption. The activation energy for the CO-OH" reaction in aqueous K2C03 solutions was reported to be 122.9 kJ/mol, which is significantly higher then that of 55.4 kJ/mol for the liquid phase CO2-OH" reaction reported by Pinsent et al. (1956). This suggests that 160 temperature is one of the most important parameter for the CO-OH" reaction in basic solutions. Eickmeyer (1962) gave a brief account of the deactivation of DEA promoted potassium carbonate solutions in commercial gas treating plants. The formates were claimed to result from the very slow reaction between CO and OH" ions. Traces of oxygen were believed to catalyze the reaction. From the above information it is clear that little is known about the role of CO in amine degradation. Nonetheless, there is clear evidence from refineries using DEA, which suggests that CO is the leading cause for the buildup of DEAF and formate in the gas treating system. As a result, considerable amounts of valuable DEA are lost or rendered ineffective, and efficiency and cost effectiveness are compromised. The purpose of this study was therefore to investigate the mechanism by which CO reacts with aqueous DEA and to estimate the corresponding solubility and kinetic parameters over a range of temperatures and CO partial pressures, so that one could quantify DEA losses and take preventive measures. 6.3 Objectives The specific objectives of this part of the work were to: (a) identify the dominant DEAF formation route, (b) estimate the kinetic rate coefficients and (c) determine the physical solubility of CO in aqueous DEA solutions. To accomplish these objectives, a reaction mechanism was proposed and a mathematical model was developed. The model consists of a set of differential and algebraic 161 equations, which describe gas absorption with slow chemical reaction in a well-mixed batch reactor. The reason for using a batch reactor was due to the low CO solubility and its slow chemical reaction with aqueous DEA solutions observed in the series of exploratory experiments described in Chapter 7. 162 CHAPTER 7 EXPERIMENTAL APPARATUS AND EXPLORATORY EXPERIMENTS This chapter describes the experimental apparatus for studying the kinetics of CO-induced degradation of aqueous diethanolamine as well as some exploratory results obtained using this equipment. 7.1 Experimental Apparatus and Procedure All experiments presented in this study were carried out in a 660 mL stainless steel autoclave reactor (Model 4560, Parr Instrument Co. Moline, IL). The experimental setup, which mainly consists of a gas bomb and an autoclave, is shown in Figure 7.1. The autoclave was equipped with a variable speed magnetic stirrer with three impellers mounted at 4, 10 and 16 cm from the bottom of a common shaft (20 cm long), a heating jacket controlled by a PID temperature controller (Omega CN76000, Omega Eng. Co., Stamford, CT), gas and liquid sample lines, pressure transducer and a thermocouple inserted into the liquid phase. To ensure proper mixing of the gas and liquid phases, the liquid volume was selected such that the top impeller was always in the gas phase and the bottom two impellers were submerged in the liquid phase. The autoclave could be operated between 283-623 K and at pressures up to 13.8 MPa. The gas bomb was made of stainless steel and could be pressurized up to 20 MPa. 163 Pressure Sensor Gas Bomb Magnetic Stirrer Pressure Sensor j-{9<H Liquid Feed $3 •Vent/Vacuum —•Liquid Sample Temperature Controller Autoclave and Heater Figure 7.1: Schematic diagram of the experimental setup For data analysis it is essential to know the precise values of the internal working volumes of the gas bomb and autoclave together with their associated piping and fittings. The volume of the gas-bomb assembly was measured by difference in weights of the assemblage when filled with water and empty. A total of 5 measurements were performed and the average volume was found to be 511.28 mL. The internal working volume of the autoclave assembly was measured by admitting a known amount of nitrogen gas from the gas bomb to the autoclave. The pressure and temperature of the autoclave were recorded and the volume was calculated by dividing the amount of gas in the autoclave by its density at that temperature and pressure. A total of 15 measurements were made and the average volume was found to be 616.7 mL. 164 The pressures in the autoclave and gas bomb were measured by means of pressure transducers (Omega PX303). These transducers were re-calibrated by a dead weight meter. In all experiments, the pressure changes in the autoclave and the gas bomb were recorded by a computerized data logging system. The DEA and MDEA were obtained from Sigma-Aldrich with a lot purity greater than 99 wt%. The CO and N2 had a purity of 99.9% and were obtained from Praxair. Aqueous solutions of DEA and MDEA were prepared by weight using distilled-deionized water. In a typical absorption experiment, the CO was fed from a supply cylinder into the gas bomb and kept there to attain constant temperature and pressure. The autoclave was then filled with about 200-300 mL of aqueous amine solution, sealed and evacuated to a pressure of less then 7 kPa to ensure the removal of extraneous gases. While stirring the solution (at 1,380 rpm or 80% of the maximum speed), its temperature was raised and maintained at the desired value by means of the water bath or the heating jacket controlled by a PID type temperature controller. The solution pressure was noted and identified as its vapor pressure. CO was then admitted from the gas bomb and its pressure change recorded. The subsequent change in autoclave pressure was recorded continuously as a function of time. At the end of each experiment, liquid samples were taken and analyzed. In some experiments, liquid samples were also withdrawn periodically during the runs. The analytical techniques to determine 165 the concentration of the main reaction products (formate ions and DEAF) are given in Appendix I. Initially, exploratory experiments were made to identify the degradation products and evaluate the trends of CO absorption at different temperatures. The primary experiments, on the other hand, were used to establish the reaction kinetics for the CO-DEA-H20 system and estimate the kinetic and solubility parameters. The experimental procedure for both experiments was essentially the same and as described above. The main difference was that, in all primary experiments, the stirrer speed was fixed at 1,035 rpm (60% of the maximum speed) as opposed to 1,380 rpm speed used in the exploratory experiments. Also, in the primary experiments, the liquid volume was always greater than 250 mL to ensure a well-defined gas-liquid interface. It should be noted however that this is not a requirement as long as the volumetric mass transfers coefficient (kLa) can be determined accurately. 7.2 Exploratory Experiments 7.2.1 CO Absorption in Aqueous DEA Solutions To understand the effect of temperature on CO loading in aqueous DEA solutions, pure CO was absorbed in a 30 wt% DEA solutions for a period of 12 hours. The initial CO partial pressure (pco) was kept at 760-990 kPa and the temperature was varied from 313-403 K. The variation in CO partial pressure was due to the vapor pressure of water at different temperatures. Therefore, the lower value of p°c0 corresponded to a higher temperature and vice versa. The 166 results of these experiments are plotted in Figure 7.2. The CO loading represents, the total amount of CO in the solution, which includes CO in molecular form and CO that has reacted with water and DEA. The CO loading at a particular time was calculated as follow: CO loading (mol /mol) = ntotal _nt MCO "CO DEA where n™* represents the total number of moles of CO transferred from the gas bomb to the autoclave, n^o represents the number of moles of CO in the head space of the autoclave at any given time t and nDEA represents the total number of moles of DEA in the solution. The quantities n^' and n^o were calculated using the Redlich-Kwong-Soave equation of state. 70 LU 60 h Q 1 50 6 U 40 o E £ 30 ^ 20 CO o O 10 O • • • • •363 K • - • •• * • t s : •313K • 403 K • 393 K 373*K • 353 K •343 K 333 K 0 _1 I I I I I I I I I I I I I I I I—I—L. 0 6 8 Time (h) 10 12 14 Figure 7.2: CO loading as a function of time and temperature (DEA= 30 wt%, p°0= 760-990 kPa) 167 As can be seen from Figure 7.2, the CO loading increases with temperature and it does not reaches equilibrium even after 12 hours especially at temperatures higher then 333 K. This suggests that CO physically absorbs in and then chemically reacts with aqueous DEA solutions. The continuous increase in CO loading at higher temperatures indicates that the reactions may be irreversible. These results also show that at lower temperatures (< 313 K), the CO loading is mainly due to physical solubility whereas at high temperatures (> 333 K), chemical reactions become important and the loading is predominantly due to the reaction products. 7.2.2 Material Balance At the end of each experiment presented in Figure 7.2, liquid samples were withdrawn and the main reaction products were identified according to the procedure given in Appendix I. The analyses showed that, in all cases where reaction was significant, DEAF and formate ions were the only reaction products. The presence of DEAF was confirmed by analyzing the liquid sample using GC/MS and comparing the results with those reported by Koike et al. (1988). To ensure that DEAF and formate were the only principal reaction products, material balances were made for experiments in which CO was brought in contact with 30 wt% aqueous DEA solutions at 353, 393 and 413 K at initial CO partial pressures ranging from 650 to 850 kPa. The final solutions were analyzed for DEAF and formate according to the procedure given in Appendix I. Table 7.1 gives the corresponding CO material balances. The deviations 168 between the initial and final amounts of CO are less than 15% and fall well within the range of experimental error. No additional degradation product was observed from the analysis of liquid samples using GC and GC-MS. Table 7.1: Overall CO balance as a function of temperature for CO absorption in 30 wt% aqueous DEA solutions Temperature (K) 353 393 413 Initial Solution (g) 199.16 199.14 198.99 Conditions pco (kPa) after 5 min 852 654 711 Initial CO introduced (g) 3.368 2.394 2.642 Final Time (h) 48 5 23 Conditions Pco (kPa) 661 435 123 CO loaded (g) Formate (wt%) CO in formate (g) DEAF (wt%) CO in DEAF (g) Molecular CO in solution2 (g) Final CO in all forms (g) 0.827 0.224 0.270 1.061 0.445 0.132 0.847 0.842 0.267 0.324 0.811 0.340 0.091 0.757 2.204 1.106 1.340 1.292 0.541 0.026 1.907 Deviation (%)n -2.4 +10.1 +13.5 1 Defined as 100x(CO loaded-Final CO in all forms)/CO loaded 2 Based on Henry's constant values obtained by N2-analogy Table 7.2: Formate and DEAF concentrations of solutions resulting from exposing pure DEA to CO for 5 hours at p°Q = 690 kPa T Formate DEAF (K) (wt%) (wt%) 313 0.0 0.0 393 0.0 2.4 169 7.2.3 CO Absorption in Pure DEA To identify the pathways of the CO-DEA reactions, experiments were undertaken where CO was absorbed in 99.8% pure DEA solution at 313 and 393 K for 5 hours at an initial CO partial pressure of 690 kPa. At the end of each of these experiments, liquid samples were taken and analyzed by IC for DEAF and formate ions. The results of this analysis are presented in Table 7.2, which shows that at low temperature (313 K) neither DEAF nor formate ions are formed, whereas at higher temperatures (>393 K) no formate was detected although a significant amount of DEAF was formed. This suggests that CO may react directly with DEA to form DEAF similar to the industrial process for the manufacture of DMF (Duranleau and Lambert, 1985). The absence of formate ions in the sample solution also indicates that water plays an important role for the buildup of formate in the system. 7.3 Proposed Reaction Mechanism The exploratory experiments show that CO reacts with aqueous DEA solutions to form formate ions and DEAF. These experiments also indicate that two pathways may form the DEAF: (i) via direct insertion of CO into DEA and/or (ii) via formate formation. Based on these results, we propose that the CO reacts with aqueous DEA solution according to the following reaction mechanism: Formate Formation: CO + OH' k1 )HCOO" (7.1) 170 Direct Insertion: CO + R^NH^^R^NHCO (7.2) DEA-Formate Reaction: k K R^NH* + HCOO- <V R^NHCO + H20 (7.3) DEA Protonation: R1R2NH^+OH<iR1R2NH + H20 (7.4) Water Dissociation: 2H2O^OH-+H30+ (7.5) where R1R2NHand R1R2NHCO denote DEA and DEAF, respectively. Note that for DEA, Ri = R2 = -CH2CH2OH. The validation of the proposed reaction mechanism and the determination of the unknown kinetic coefficients are presented in Chapters 8 and 9. 171 CHAPTER 8 MATHEMATICAL MODEL This chapter presents the mathematical model for CO absorption in aqueous diethanolamine solution in a batch reactor. This model was used to validate the proposed reaction mechanism and to estimate the rate constants with the non-linear regression package GREG. 8.1 Reaction Mechanism As proposed in Chapter 7, when CO is absorbed into an aqueous DEA solution, the following reactions may occur: Formate Formation: CO + OH" *HCOO" (8.1) Direct Insertion: CO + R1R2NH—^—>R1R2NHCO (8.2) DEA-Formate Reaction: R1R2NH2 +HCOO" R^NHCO + HjO (8.3) DEA Protonation: R1R2NH2 +OH- ^R^NH + HaO (8.4) Water Dissociation: 2H2O^OH+H30+ (8.5) 172 where R^NHand R1R2NHCO denote DEA and DEAF, respectively. Also, for DEA, Ri = R2 = -CH2CH2OH. 8.2 Reaction Rates For convenience, the chemical species in reactions (8.1) to (8.5) are renamed as follows: C^CO,,,, C2=R1R2NH, C3=R1R2NH2-, C4=R^2NHCO (8.6) C5 = HCOCr C6 = OH", C7 = H30+, C8 = H20 Reactions (8.1) and (8.2) are considered to be irreversible and to proceed at finite rates. The latter are given by the following expressions: r^-ktCA (8.7) r2=-k2C,C2 (8.8Based on the experimental evidence presented below, reaction (8.3) was considered to be reversible with the overall rate given by: r3=-k3C3C5+k_3C4 (8.9) Reactions (8.4) and (8.5) are regarded as reversible equilibrium reactions since they involve only proton transfers. 8.3 Mathematical Model The mathematical model is based on the concept of gas absorption with slow chemical reaction in a well-mixed batch reactor as shown in Figure 8.1. The following equations govern the mass transfer with chemical reaction in a batch autoclave (liquid volume = VL, gas volume = VQ): 173 Vapor, VG Figure 8.1: Schematic diagram of chemical reaction autoclave 174 Gas Phase CO Balance: ^ = -^IvLEkLa dt V ' L Pco hi ^ CO-DEA j Liquid Phase CO Balance: dC, _ dt = EkLa Pco H CO-DEA + r1+r2 DEAF Balance: dC dt - = -r3-r2 Overall CO Balance: V^dpco + dC, dC±+ dC^ RT dt L dt L dt L dt Total DEA Balance: Electron Neutrality Balance: DEA Protonation: dC2 | dC3 [ dC4 = Q dt dt dt dC3 dC7 dC5 dC6 _ ^ dt dt dt dt X _ ^2 Water Dissociation: 175 The model equations (8.10)-(8.17) were derived based on the following assumptions: • The concentration of dissolved (molecular) CO at the gas-liquid interface corresponds to the physical solubility as determined by Henry's Law (i.e., Pco =HC,). • Both the gas and liquid phases are well mixed. • The gas phase resistance to mass transfer is negligible. The last assumption follows because pure CO was used. Also, this assumption was verified experimentally and by simulation. Consequently, there are eight ordinary differential and algebraic equations (equations 8.10-8.17) with eight unknown chemical species. The equations must be solved subject to the following initial conditions: pco=p°co and Ci=C° fori =1-7 (8.18) For the present experiments, the initial CO loading of the DEA solution is zero. Therefore, at t=0, C° = C° = C° =0. The values of C°, C°, C° and C° can be obtained by solving the equilibrium equations shown below. When the CO loading is zero, it follows that: Overall DEA Balance: C°+C°=[R^2NH]initial (8.19) Electron Neutrality Balance: C°+C°-C°=0 (8.20) 176 DEA Protonation: K4=-^- (8.21) 4 C°C° Water Dissociation: K5 = C°C° (8.22) For a given initial DEA concentration, the concentrations C°2, C°3, Cg and C° can be calculated by solving equations (8.19) to (8.22) simultaneously. 8.4 Model Parameters The parameters needed to solve the model equations (8.10) to (8.17) are listed in Table 8.1 and their determination is outlined below. Since the present reactions are very slow, one can safely assume that the enhancement factor (E) for the CO-DEA-H20 system is unity. Olofsson and Hepler (1975) reported the following correlation for the water dissociation constant (K5) in the temperature range of 293-573 K: log10 (K5) = 8909.483 -142^.136 - 4229.195 log10 (T) + 9.7384T - 0.0129638T2 + 1.15068x10"5T3-4.602x10"9T4 (8.23) Bower et al. (1962) correlated their data for (K4 K5) over the temperature range of 273-323 K according to the following equation: log10 (K4K5) = -4.0302 - 183°-15 + 0.0043261T (8.24) 177 Table 8.1: Model parameters Param. Definition Remarks E Enhancement factor For slow reactions, E = 1 K4 DEA protonation constant Bower et al. (1962) K5 Water dissociation constant Olofsson and Hepler(1975) kLa Mass transfer coefficient To be estimated HCO-DEA Henry's constant for CO To be estimated ki Rate constant for Rxn. (1) To be estimated k2 Rate constant for Rxn. (2) To be estimated k3 Rate constant for Rxn. (3) To be estimated k-3 Rate constant for reverse Rxn. (3) To be estimated The volumetric mass transfer coefficient (kLa), Henry's constant of CO in aqueous DEA solution (HCO-DEA) ar|d the reaction rate constants (ki, k2, k3 and k3) were obtained as part of this work. In addition to the parameters listed in Table 8.1, the densities and viscosities of the aqueous amine solutions must also be known as a function of temperature. These properties were obtained from the correlations given by Hsu and Li (1997a, b). 8.5 Parameter Estimation The model was solved using an algorithm called DDASAC (Caracotsios and Stewart, 1985). DDASAC is an extension of DASSL (Petzold, 1983; Brenan et al., 1989), which uses an implicit integrator for non-linear initial value problems containing ordinary differential equations with or without algebraic equations. The parameter estimation was carried out using the software package GREG (Stewart et al. 1992). 178 The six unknown parameters (kLa, HC0_DEA, ki, k2, k3, k.3) were determined by minimizing the following objective function: S(k) = £&l-y(k)]Q,|yl-y(k)] (8.25) i=i where k = (kLa,HC0_DEA, ki, k2, k3, k.3)T is the unknown parameter vector, yjs the measured value of the state variable (in the present case, it is the partial pressure of CO) and y(k) is the calculated value of the state variable which is obtained by solving the model equations (8.10) to (8.17) for some assumed values of kLa, HC0_DEA, ki, k2, k3, k_3. N denotes the number of experimental data and Q is the weighting matrix. For least squares estimation, Q is taken as the identity matrix. Note that not all six parameters were estimated simultaneously. First, kLa and k3 and k_3 were estimated from the data obtained in a separate series of experiments. Then using these known values kLa, k3and k.3, the rest of the unknown parameters (ki, k2, and HC0_DEA) were estimated. However, the same methodology was used in all three cases. The optimization was carried out using GREG (Stewart et al. 1992). 179 CHAPTER 9 RESULTS AND DISCUSSION This chapter presents the results and main conclusions from the experimental and theoretical studies on the kinetics of CO-induced degradation of aqueous diethanolamine. 9.1 Absorption and Degradation in CO-DEA System 9.1.1 Determination of k, a In order to obtain accurate CO partial pressure profiles in the batch absorption experiments using the reaction model given in Chapter 8, it is essential to know the values of kLa in each case. A series of experiments were conducted where pure N2 was absorbed into 30 wt% aqueous DEA solutions over the temperature range of 283-413 K in a batch autoclave. The experimental setup was exactly the same as that shown in Figure 6.1. In a typical experiment of this series, about 300 mL of a 30 wt% aqueous DEA solution were fed into the autoclave, a vacuum was drawn and the liquid solution along with the gas bomb containing pure N2 was brought to the desired temperature. Once this temperature was reached, a known amount of pure N2 was released from the gas bomb to the autoclave and the total pressure in the autoclave was recorded with time. Each experiment was run for about 105 minutes, which was found to be sufficient to completely saturate the liquid with nitrogen. To study the effect of 180 agitation on kLa, each experiment was performed at the two stirrer speeds of 1,035 and 1,550 rpm. Since N2 does not react with aqueous DEA, the data can be analyzed using a physical absorption model, which is obtained by setting the reaction rates in the reaction model to zero. Physical absorption in a batch autoclave (see Figure 8.3) can then be represented by the following equations: Gas Phase N2 Balance dpN? RT , R dt VG VLkLa PN2 N Liquid Phase N2 Balance dCN, = kLa ^Nj-DEA Pn'-- a (9.1) |_| ^ N2-DEA j (9.2) dt Initial Conditions at t = 0 pN2=p°2 and CNj =0 (9.3) In this model, there are two ordinary differential equations and two unknown parameters (kLa and HN2_DEA), which can be determined from the partial pressure (pN2) versus time data by applying the parameter estimation technique described earlier. Figures 9.1 and 9.2 show the experimental and predicted N2 partial pressures using the optimized kLa and HN2_DEA values at 1,035 rpm and 283-413 K. An excellent agreement between experimental and fitted values demonstrates the accuracy of the experimental procedure. The same set of experiments was 181 2000 Experimental Model 1900 ro Q_ CM z Q. 1800 h 1700 0 20 40 60 80 Time (min) 100 120 Figure 9.1: Nitrogen absorption in 30 wt% aqueous DEA solution at 1050 rpm and 284-333 K 2000 Experimental Model ; | }3g^;373K 393 K •413K 1200 '—1—1—1 1—1—1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 20 40 60 80 100 120 Time (min) Figure 9.2: Nitrogen absorption in 30 wt% aqueous DEA solution at 1050 rpm and 353-413 K 182 done at 1,550 rpm and similar agreement was found. The estimates of kLa so obtained were correlated by the following equation: DM kLa = 0.297-^Re°-77Sc0-5 (9.4) djmp where the Reynolds number (Re) and Schmidt number (Sc) are defined as follows: Re = ^E£L and Sc = (9.5) m. PLDN2 The impeller diameter (djmp) was 3.1 cm and the diffusivity of N2 in aqueous DEA solution (DN2) was calculated using Wilke and Chang's correlation given in Perry et al. (1963). Note that the exponent of Sc in equation (9.4) was assumed not fitted. This value is the same as reported previously (Critchfield and Rochelle, 1988; Haimour et al, 1985, Hikita et al., 1975; Danckwerts, 1970). A plot of the predicted and estimated kLa values is shown in Figure 9.3. The agreement between the measured and correlated kLa values at both stirrer speeds is within ± 25%. Consequently, equation (9.4) can be used to calculate kLa in the reaction model. 183 0 0.04 0.08 0.12 0.16 Measured kLa (1/min) Figure 9.3: Measured and predicted volumetric mass transfer coefficient 9.1.2 Determination of HCO-DEA by Nitrogen Analogy The second unknown parameter in the reaction model is the Henry's constant of CO in aqueous DEA solutions (HC0_DEA). Since CO reacts with aqueous DEA, it is not possible to measure HCO-DEA by conventional techniques used for physical absorption and a new technique called the "N2-Analogy" method was developed. This technique is based on the same principal as the "N20-Analogy" approach used previously for measuring the Henry's constants of C02 in alkanolamine solutions (Laddha et al., 1981). Like C02 and N20, the molecular masses of CO and N2 are nearly equal and the ratio of their Henry's constants in water is about 0.70 ± 0.02 over the temperature range of 273-413 K (Fogg and Gerrard, 1991; Perry et al., 1963). If it is assumed that this ratio 184 remains the same even in aqueous amine solutions, then H CO-DEA can be calculated from the following equation: H CO-H20 Hi H CO-DEA H (9.6) N2-H20 Nj-DEA The values of HC0_H20and HN2_H20were obtained from the correlations given in Fogg and Gerrald (1991) and the HN _DEA data were measured in this work using using the N2-Analogy (equations 9.6) are listed in Table 9.1. A comparison of the values of Henry's constant of nitrogen in water and in aqueous DEA solution indicates that nitrogen is several times more soluble in aqueous DEA solutions than in water and its solubility increases with temperature above 313 K. This trend is not surprising because the literature data on CO and N2 solubility (see Table 9.2) show that the solubility of these gases in pure organic solvents is 10 to 100 times higher than that in pure water, and in almost all cases, it increases with temperature above 298 K. The solubility measurements presented here were repeated at least three times and the reproducibility was within ± 3%. Due to our concern about the accuracy of the HC0_DEA values obtained from the N2-Anology, we did not use them in the reaction model to estimate the rate constant ki and k2. Instead, we considered HC0_DEA as one of the unknown parameters and found it directly from the reaction model (see section on determination of ki, k2 and HC0_DEA). As indicated later, the agreement between the HC0_DEA values obtained the procedure described in the previous section. The H CO-DEA values obtained by 185 using both methods is fairly good. This suggests that our measurements of HCO-DEA by the N2-Anology method are fairly accurate. Table 9.1: Henry's constant of CO in 30 wt% aqueous DEA solution from N2-Analogy T I_I + nN2-H20 I_I + nCO-H20 ^CO-HjO ^N2-H20 I_J ++ N2-DEA ^CO-DEA (K) (MPa) (MPa) - (MPa) (Mpa) 313 10,133 7,072 0.70 4,519 3,154 333 11,429 8,187 0.72 2,371 1,698 353 12,048 8,613 0.72 2,047 1,463 363 11,926 8,582 0.72 1,986 1,429 373 11,683 8,400 0.72 1,834 1,319 383 11,348 8,096 0.71 1,895 1,352 393 10,943 7,701 0.70 1,976 1,390 + Obtained from correlations given in Fogg and Gerrad (1991) Measured in this work 186 Table 9.2: Henry's constants of CO and N2 in water and organic solvents (Fogg and Gerrad, 1991) Solvent Temp. Hco HN2 Hco (K) (MPa) (Mpa) HN2 Water 298 5,857 8,552 0.68 313 7,072 10,133 0.70 333 8,187 11,429 0.72 353 8,613 12,048 0.71 363 8,582 11,926 0.72 373 8,400 11,683 0.72 383 8,096 11,348 0.71 393 7,701 10,943 0.70 413 6,708 9,960 0.67 Methyl acetate 298 119 173 0.69 303 118 170 0.70 313 116 163 0.71 2-propanone 298 130 187 0.69 303 129 184 0.70 313 126 177 0.71 1-1'-oxybisethane 293 60 82 0.74 298 60 81 0.74 303 60 80 0.75 313 59 78 0.76 Tetrachloromethane 298 117 159 0.74 303 116 156 0.74 313 114 151 0.75 323 111 146 0.76 333 109 141 0.77 Chlorobenzene 298 161 238 0.68 303 159 233 0.68 313 157 225 0.70 323 154 217 0.71 333 150 209 0.72 343 147 202 0.73 353 144 194 0.74 Benzene 298 157 228 0.69 333 135 191 0.71 Hexadecane 298 58 83 0.70 Methanol 298 308 371 0.83 Ethanol 298 220 284 0.77 1-Propanol 298 196 250 0.78 187 9.1.3 Determination of k3 and k.3 It was not possible to estimate k3 and k_3 together with ki, k2 and HC0_DEA by regressing the pressure versus time data obtained from CO absorption experiments. This was mainly because molecular CO is not directly involved in reaction (8.3). Consequently, the latter exerts no or little effect on the overall change in CO partial pressure due to reaction. Therefore, it was decided to estimate k3 and k_3 separately by following the liquid phase reaction between formic acid and aqueous DEA. In this series of experiments, 0.5-0.65 M formic acid was added to a 30 wt% aqueous DEA solution and the mixture was allowed to react in a stirred autoclave under a nitrogen blanket for about 40-170 hours in the temperature range of 303-413 K. Liquid samples were withdrawn at regular intervals from the autoclave and analyzed for formate and DEAF using IC. The initial formate concentration was set at about 5 times the maximum concentration predicted from the CO-DEA reaction model. 9.1.3.1 DEA-Formic Acid-Water Reaction Model To estimate k3 and k_3 from the data obtained by means of the above experiments, the following liquid phase reactions can be postulated: R^NH* + HCOO- K<i3 R^NHCO + H20 (9.7) R^NH^+OH1 ^R1R2NH + H20 (9.8) HCOOH + H2O^HCOO" +H30+ (9.9) 2H2O^OH"1+H30 + (9.10) 188 For convenience, we keep the same nomenclature as before except that species HCOOH is represented by C9. The reaction (9.7) is the same as reaction (8.3) and its rate expression is given by equation (8.9). The dissociation of formic acid represented by reaction (9.9) is an equilibrium reaction. The equilibrium constant, K6 was estimated from the following correlation: log™ (K6) = -39.06123 - 173t624 +17.88348 log10 (T) - 0.028039T (9.11) Equation (9.11) is based on the data of Harned and Embree (1934) obtained in the temperature range of 273-333 K. The equations governing the material and electronic balances in a batch reactor are: Formate Balance: dC 5=r3 (9.12) DEAF Balance: Total DEA Balance: dt dC„ — = -r3 (9.13) dt 3 K ' dC, dC, dC4 . ,nAAs —- + —-+—1 = 0 (9.14) dt dt dt v Electron Neutrality Balances: dC3 dC7 dC5 dC dt dt dt dt DEA Protonation: C = 0 (9.15) K4=—(9.16) 4 C,C« 189 Formic Acid Dissociation: K6=-^ (9.17) Water Dissociation: K5=C6C7 (9.18) There are seven ordinary differential and algebraic equations (equations 9.12 to 9.18) with seven unknown chemical species. The equations must be solved subject to following initial conditions: C{=C? fori =2-7,9 (9.19) At t=0, C° = 0. The values of C°, C°, C°, C°, C° and C° can be obtained by solving the equilibrium equations shown below. Since the DEAF concentration is initially zero, it follows that: Overall DEA Balance: Overall Formate Balance: Electron Neutrality Balance: DEA Protonation: C°+C°=[R^2NH]initial (9.20) C°+C°=[HCOOH]inrtjal (9.21) C°+C7-C°-C° =0 (9.22) K4=-^- (9.23) Formic Acid Dissociation: 190 K6=-!^ 0-24) Water Dissociation: K5=C°C° (9.25) For the present experiments, the initial DEA and formate concentrations are 2.85 and 0.5-0.65 mol/L respectively. Therefore, the equilibrium concentrations C2, C°, Cg, C°, C° and C°can be calculated by solving equations (9.20) to (9.25) simultaneously by means of Newton's Homotopy continuation method. Time (h) Figure 9.4: Measured and predicted formate ions and DEAF concentration 191 9.1.3.2 Parameter Estimation In the above model (equations 9.12 to 9.25), there are only two unknown parameters (k3 and k_3), which can be estimated from the formate versus time data by the non-linear least squares optimization technique described earlier. The formate ion and DEAF concentrations obtained from the experiments and the model using the optimized values k3 and k_3 are plotted in Figure 9.4. Good agreement between the measured and predicted formate and DEAF concentrations was obtained. Data for temperatures below 343 K are not shown in Figure 9.4 because no significant reaction was observed. In Figure 9.4, the concentration trends for Formate and DEAF clearly indicate that the reaction between formic acid and DEA is slow and reversible. The estimates for k3 and k_3 at various temperatures are listed in Table 9.3 and are plotted in Figures 9.5 and 9.6 respectively. Also, these estimates could be fitted by the following Arrhenius equations: k3 =3x1010exp(-10,177/T) (9.26) k_3 = 2x107 exp(-7,637.2/T) (9.27) where T is in Kelvin, k3 is in m3/kmol h and k.3 is in 1/h. The activation energies, as determined from equations (9.26) and (9.27), are 84.61 kJ/mol and 63.50 kJ/mol respectively. 192 Table 9.3: Estimates of the rate and equilibrium constants for reaction (8.3) T k3 k.3 K3 K3 (K) (m3/kmol h) (1/ h) (m3/kmol) (dimensionless) 343 4.666 xlO"3 5.260 x 10'3 0.887 34.91 353 6.194 x10"3 6.488 x 10"3 0.955 37.42 373 4.354 x10'2 2.735 x 10"2 1.592 61.93 393 1.125 x10'1 6.467 x10"2 1.740 67.15 413 6.920 x10"1 2.099 x10"1 3.297 126.27 1.E+00 F 1.E-03 t 1.E-04 '—1—1—'—1—1—1—1—1—1—1—1—1—1—1—1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2.5 3 3.5 1000/T(1/K) Figure 9.5: Arrhenius plot of the estimates for the forward rate constant of reaction (8.3) 193 1.E+00 F 1.E-03 b 1 .E-04 '—1—1—1—1—1—1—1—1—1 1 1 1 1 1 1 1 1 1 1 1 • • 1 • 2 2.5 3 3.5 1000/T(1/K) Figure 9.6: Arrhenius plot of the estimates for the reverse rate constant of reaction (8.3) From the estimates of k3 and k.3, the equilibrium constant, K3, was calculated and plotted in Figure 9.7. The values of K3 were fitted to the following Arrhenius equation: K3 = 1.355 x 103 exp(-2,540.1 /T) (9.28) where T is in Kelvin and K3 has units of m3/kmol. In the literature, the temperature dependence of amine hydrolysis constants is usually reported in terms of the expressions proposed by Harned and Robinson (1940). K3 was made dimensionless by dividing it by the concentration of water in the solution: log10(K3) = 10.339 + 11"r668 - 8.364 log10(T) + 0.025946T (9.29) 194 The constants of equation (9.29) were estimated by the method of least squares. The literature does not contain values and correlations for k3, k_3 and K3 and we are the first to report them. 10 o E £ 1 E K3 = 1.355x103 exp -2,540.1 R2 =0.9391 Q I i i t i i I I t i I i i i i i i i i i I i i i i i i i • • I 2.0 2.5 3.0 3.5 1000/T(1/K) Figure 9.7: Arrhenius plot of the estimates for the equilibrium constant of reaction (8.3) 9.1.4 Determination of ki, k2 and HCO-DEA The correlations for kLa, k3 and k_3 obtained above may be substituted into the reaction model (equations 8.10-8.18) to estimate the remaining unknown parameters (ki, k2 and HC0_DEA) from the pressure versus time data using the least squares minimization method discussed earlier. The experimental data were obtained from a series of batch absorption experiments, where pure CO was absorbed in 30 wt% aqueous DEA solutions in the temperature range of 313-413 K and CO partial pressure range of 800-1200 kPa. The change in total 195 reactor pressure due to absorption and reaction was recorded with time. The partial pressure of CO was calculated by subtracting the solution vapor pressure from the total reactor pressure at the corresponding temperature. At the end of each experiment, liquid samples were withdrawn and the concentrations of formate ions and DEAF were found. Initially, the parameters were estimated by considering that all reactions (8.1)-(8.5) were reversible. Since, we could not estimate the values of reverse rate constants for reactions (8.1) and (8.2) the reversibility for these reactions was not considered in subsequent data analysis. The estimates for ki, k2 and HCO-DEA are listed in Table 9.4 and a comparison of the measured and predicted CO partial pressures using these estimates is shown in Figure 9.8. Good agreement was found between the measured and predicted CO partial pressures at all temperatures. This is strong validation of the model and experimental procedure. Table 9.4: Estimates of ki, k2 and H CO-DEA k2 (m3/kmol h) H + CO-DEA CO-DEA (K} (m3/kmol h) (MPa) (MPa) 313 8.58 333 27.41 343 81.11 363 130.32 373 225.28 383 441.47 393 619.62 413 813.08 5.370 x 10"3 1.214 x 10"2 1.645 x 10"2 2.741 x 10'2 3.806 x 10-2 5.591 x 10"2 7.506 x 10"2 1.152 x IO-1 5,928 2,513 2,108 1,368 1,348 1,398 1,297 618 3,161 1,702 1,469 1,429 1,317 1,348 537 Calculated from N2-Analogy 196 1200 800 ro D_ s£ O o 0. 400 0 Measured Predicted 313K. 333 K 343 K 363 K 413K -i i i i i_ -i i i i i i i i i i i i i i i_ 0 400 800 1200 Time (min) 1600 Figure 9.8: Measured and predicted partial pressures for CO absorption in 30 wt% aqueous DEA solution We considered HC0_DEA as one of the unknown parameters and estimated it along with k-i and k2. The values of H CO-DEA obtained using this approach and the N2-Analogy method are listed in Table 9.4. The agreement between the two sets of H CO-DEA values and its trend with respect to temperature is good, particularly at temperatures above 333 K. This suggests that the N2-Anology is valid. The relatively large differences in the values HC0_DEA at temperatures lower then 333 K are probably due to the extremely slow reaction between CO and aqueous DEA solution. Consequently, the reaction model may not furnish good estimates of Hco_DEA at these temperatures. 197 1.E+04 1.E+03 h o I 1.E+02 != o 1.E+01 t 1.E+00 k, = 1x109 exp V •5,816.5 T E = 48.36 kJ/mol R2 =0.9608 -I I I I I L__l I I I I I I I I I I I I l_ 2.0 2.2 2.4 2.6 2.8 1000/T (1/K) 3.0 3.2 Figure 9.9: 1.E+00 c — 1.E-01 75 E 1.E-02 1.E-03 Arrhenius plot of the estimates for the second order rate constant of reaction (8.1) 2.0 2.2 k2 =1.6425 x103exp| E = 32.88 kJ/mol R2 =0.9933 (-3,955.3^ -i i i i i i i i i i i i i i i i i i i i i i_ 2.4 2.6 2.8 3.0 3.2 1000/T (1/K) Figure 9.10: Arrhenius plot of the estimates for the second order rate constant of reaction (8.2) 198 The estimates for k-i and k2 at various temperatures are plotted in Figures 9.9 and 9.10 and are fitted to the following Arrhenius equations: k, = 1x109 exp(-5,816.5/T) (9.30) k2 =1.6425x103exp(-3,955.3/T) (9.31) where T has units Kelvins, while ki and k2 are in units of m3/kmol h. The activation energies determined from equations (9.30) and (9.31) are 48.36 kJ/mol and 32.88 kJ/mol respectively. The data for both ki and k2 fall on straight lines thereby providing strong support for the validity of the present mathematical model and parameter estimation technique. McDonald (1963) studied the kinetics of the absorption of CO in basic solutions and reported a second-order rate constant (k-i) value of 93.96 m3/kmol h at 353 K. This value agrees well with the present estimates of 81.11 m3/kmol h at 343 K and 130.32 nrrVkmol h at 363 K (see Table 9.4). Comparison of the values of ki and k2 in Table 9.4 suggests that the formate formation reaction (reaction 8.1) is about three orders of magnitude faster then the DEAF formation reaction (reaction 8.2). Also, a comparison of rate constants for the DEAF formation reactions via direct insertion (k2 in Table 9.4) and formate-DEA reaction (k3 in Table 9.3) reveals that both constants are of the same order of magnitude. However, since reaction (8.3) is reversible and since the equilibrium favors the reverse reaction, its overall effect on the net formation of DEAF is negative, particularly at temperatures higher than 343 K. This phenomenon is better illustrated in Figure 9.11, which compares the model predictions for the formation of formate ions and DEAF in a batch reactor at 393 199 K with and without reaction (8.3). It is clear from this figure that, under conditions when reaction (8.3) is included, the net formation of DEAF is reduced and that of formate ions is increased. However, simulation results at temperatures lower than 343 K indicate that reaction (8.3) has no effect on the buildup of formate and DEAF in the reactor. From this analysis, we conclude that the dominant DEAF formation route is via direct insertion of CO into the DEA molecule (reaction 8.2). 80 1'60 O o o 0 •with reaction (5.3) •without reaction (5.3) -i—i i i i i i i i i • • • ' • 6 Time (h) 8 J I I I L. 10 12 Figure 9.11: Effect of reaction (8.3) on the formation of formate and DEAF at T = 393 K , pco = 1000 kPa and CDEA= 30 wt% 9.2 Absorption and Degradation in CO-MDEA System The reaction mechanism and the parameter estimation technique was further verified by conducting an experiment where pure CO was absorbed into 30 wt% aqueous MDEA solutions at 393 K. MDEA is a tertiary amine and is not 200 expected to react with CO to form diethanolacetamide (i.e. RiR2NCOR3). Consequently, reactions (8.2) and (8.3) do not exist and the CO-MDEA-H20 system can be represented by: CO + OH-^^HCOO- (9.32) R1R2R3NH++OH<iR1R2R3N + H20 (9.33) 2H20<ioi-r+H30+ (9.34) where R1R2R3N and R1R2R3NH+ denote MDEA and protonated MDEA respectively. Note that for MDEA, Ri = R2 = -CH2CH2OH and R3 = -CH3. The ionic strength of a 30 wt% aqueous MDEA solution is almost the same as that of a 30 wt% aqueous DEA solution. Therefore, if the reaction mechanism is right, the same value of ki should be found for both systems. We therefore estimated ki from the batch absorption data obtained in this experiment by using the same reaction model as before (equations 9.10-9.18) and by setting r2 and r3 equal to zero. The value of the MDEA protonation constant (K7) was calculated from the following correlation reported by Barth et al. (1981) over the temperature range of 298-333 K. It was extrapolated up to 393 K for this work: log10(K7K5) = -14.01 + 0.018T (9.35) The parameter estimates from the experiments on CO absorption in aqueous MDEA and DEA solution at 393 K are listed in Table 9.5. As expected, the estimated values of ki in both cases are in excellent agreement. This once 201 again validates the proposed reaction mechanism and the experimental technique for parameter estimation. Table 9.5: CO reaction with aqueous DEA and MDEA Amine T ki H C0-Am Formyl HCOO' (K) (m3/kmol h) (MPa) Alkanolamine DEA 393 620 1,297 Yes Yes MDEA 393 668 841 No Yes 9.3 Absorption and Degradation in CO-DEA+DEAF System In order to elucidate the reversibility of reaction (8.3), an experiment was conducted where CO was absorbed in the aqueous mixture of 30 wt% DEA and 18.4 wt% DEAF at 393 K. The stock solution of DEAF had a 85% purity and was provided by the Shell Technology Center, Houston, Texas. The impurities in the stock solution were mainly water and the acetates. Purer DEAF solutions were unavailable. The experiment was conducted in a batch mode and liquid samples were withdrawn at regular intervals. Using the parameters estimated in this work, a numerical simulation was performed to predict the rate of change of CO partial pressure and formate and DEAF concentrations in the liquid phase. These results are plotted in Figure 9.12. Clearly, the agreement between predicted and measured CO partial pressures and DEAF concentrations is excellent. However, the predicted formate ion concentrations were about 20-25% lower then those measured in our experiment. The reason for this discrepancy is not clear. One of the explanations is that the impurities in the DEAF stock solution might have converted to formate. Note that in Figure 9.12, the sudden drop in CO partial pressure at each 60-202 minute interval is because the numerical simulation was done by taking into account the changes in total pressure and the gas and liquid volume due to sample withdrawal. Time (min) Figure 9.12: Measured and predicted concentration and partial pressure profiles for CO absorption in aqueous solution of 30 wt% DEA and 18.4 wt% DEAF at 393 K 9.4 Model Predictions After establishing the reaction mechanism and the values of all the associated physico-chemical parameters, we are able to predict the DEAF and formate ion formation when aqueous DEA solutions are exposed to CO. We therefore measured DEAF and formate concentrations in the liquid samples withdrawn at the end of each experiment. The predicted and measured values of DEAF and formate ion concentrations in all experiments are listed in Tables 9.6 to 9.9. It can be seen from these tables that the agreement between predicted 203 and measured concentrations is good. In most cases, the relative error fell within ± 30%. It should however be noted that in a few cases the error was as high as ± 50%, which could not be due to experimental error as these measurements were rechecked and same values were found. In view of the errors associated with literature correlations in calculating various physico-chemical properties and the experimental errors associated with the IC technique, this agreement can still be regarded as good. Table 9.6: Measured and predicted Formate and DEAF concentrations for CO absorption in 30 wt% aqueous DEA solution T Time Formate Formate Error DEAF DEAF Error Measured Predicted Measured Predicted (K) (h) (mmol/l) (mmol/l) (%) (mmol/l) (mmol/l) (%) 313 62.2 5.2 5.2 0 11.0 8.2 25 333 24.0 15.4 17.2 -12 7.6 9.9 -31 343 24.0 25.4 26.0 -2 10.8 16.0 -47 363 23.1 48.0 55.1 -15 32.2 37.6 -17 373 7.5 31.6 36.2 -15 18.1 18.5 -2 383 7.5 38.7 42.3 -9 18.9 20.8 -10 393 6.5 41.9 46.7 -11 29.1 21.7 25 413 21.8 159.4 139.3 13 93.5 68.5 27 204 Table 9.7: Effect of amine concentration on CO reaction with aqueous DEA solution at 393 K DEA Time Formate Formate Error DEAF DEAF Error Measured Predicted Measured Predicted (wt%) (h) (mmol/l) (mmol/l) (%) (mmol/l) (mmol/l) (%) 5 7.0 30.4 32.6 -7 8.7 7.6 13 10 23.5 - 89.3 - - 27.9 -15 22.0 104.6 96.8 7 46.1 35.3 24 30 6.5 41.9 46.2 -10 29.1 21.8 25 50 28.0 104.2 121.8 -17 113.3 80.1 29 Table 9.8: CO absorption in aqueous solution of 30 wt% DEA and 18.4 wt% DEAF T Time Formate Formate Error DEAF DEAF Error Measured Predicted Measured Predicted (K) (h) (mmol/l) (mmol/l) (%) (mmol/l) (mmol/l) (%) 393 19.5 571.9 540.4 5 1031.5 956.3 7 Table 9.9: CO absorption in 30 wt% aqueous MDEA solution T Time Formate Formate Error DEAF DEAF Error Measured Predicted Measured Predicted (K) (h) (mmol/l) (mmol/l) (%) (mmol/l) (mmol/l) (%) 393 21.7 57.5 83.5 -45 0 0 0 205 9.5 Process Implications Figures 9.13 and 9.14 show the predicted concentrations of various species in the liquid phase for CO absorption in 30 wt% aqueous DEA solution in a batch autoclave at temperatures of 313 K and 413 K and at initial CO partial pressure of 1,000 kPa. These temperatures are typical for absorbers and strippers in commercial plants. These figures clearly show that under identical conditions the formation of DEAF and formate is about 10-15 times higher at the stripper temperature then that at the absorber temperature. This suggests that in the absorber, where the temperature is usually low (<323 K), the physical absorption of CO is facilitated. By contrast, in the stripper and its reboiler, where the temperatures are usually high (>363 K), amine degradation will occur. Although the rates of these degradation reactions are quite slow our simulation results show that, over time, one can expect substantial DEAF and formate ion buildup in the system. For the purpose of illustration, the data shown in Figures 9.13 and 9.14 were obtained using very high initial CO partial pressures. Using the parameters estimated in this work one can easily demonstrate the implication of CO induced degradation under real plant conditions. It should, however, be noted that translating these data to real plant conditions is not straightforward. The influence of process conditions and a totally different gas composition on the various parameters must be taken into account. 206 Figure 9.13: Predicted concentration profiles for CO absorption in 30 wt% aqueous DEA solution at T = 313 K and pc0 = 1000 kPa 80 60 h x o b o o X 5 140 LL < LU Q d 20 h 0 Figure 0 2 4 6 Time (h) 9.14: Predicted concentration profiles for CO absorption in 30 wt% aqueous DEA solution at T = 393 K and p°0 = 1,000 kPa 207 9.6 Conclusions The kinetics of CO-induced degradation of DEA were studied over the temperature range of 313-413 K and DEA concentrations of 5-50 wt%. A reaction mechanism was proposed and a mathematical model was developed to estimate previously unknown kinetic and solubility parameters from batch absorption experiments. The experimental data are best described by reactions (8.1) to (8.5). Numerical simulation results, based on the estimated parameters, indicate that the primary DEA degradation reaction is the direct reaction of DEA with molecular CO (reaction 8.2). The formate-DEA reaction (reaction 8.3), on the other hand, is relatively slow and reversible. Comparison of the simulation results with and without reaction (8.3) show that, at temperatures higher than 343 K, some of the DEAF formed by reaction (8.2) is hydrolyzed back to form formate ions and protonated DEA by reverse reaction (8.3). However, at temperatures lower than 343 K, reaction (8.3) seems to have no significant effect on the net formation of DEAF and formate ions in the system. The data from the experiment on CO absorption in aqueous solution containing 30 wt% DEA and 18.4 wt% DEAF, further confirms the reversibility of reaction (8.3). Parameter estimates of the rate constants suggest that the reaction rates are quite slow but, over time, nevertheless lead to substantial DEAF and formate ion buildup in the system. The Henry's constant of CO in aqueous DEA solution was determined by the N2-Anology and from the reaction model. The results are in fairly good agreement. Comparison of the values of HC0_DEA over the temperature range of 208 313-343 K shows that the value of HC0_DEA decreases with increasing temperature. This trend is similar to that of CO solubility in pure organic solvents reported in the literature (Fogg and Gerrard, 1991). An experiment conducted by absorbing CO in 30 wt% aqueous MDEA solution shows that, unlike DEA, MDEA does not react with CO. There was no experimental evidence suggesting the formation of diethanolacetamide (R1R2NCOCH3) even after 20 hours at 393 K. The rate constant for the formate ion formation reaction was found to be in excellent agreement with that obtained from CO-DEA-H20 system. This further validates the proposed reaction mechanisms and the experimental technique used to estimate the kinetic parameters. The DEAF and formic ion concentrations found at the end of each experiment agreed well with those predicted from the present mathematical model. In most cases the relative error was within ± 30%. 209 CHAPTER 10 OVERALL CONCLUSIONS AND RECOMMENDATIONS 10.1 Main Conclusions from C02 Absorption/Desorption Work A comprehensive mathematical model governing the diffusion-reaction process of C02 absorption and desorption in aqueous amine solutions falling as a laminar liquid film over a hemispherical surface has been developed. The model applies to a general case where the aqueous solution may contain a binary mixture of amines (e.g., MEA+AMP or MEA+MDEA) and where C02 undergoes a series of liquid phase reversible reactions. The model readily reduces to a single aqueous amine (e.g., MEA or AMP) or physical absorbent system by setting the initial concentration of one or both amines to zero. The model can also be reduced to the well-known pseudo-first-order kinetics when all reactions are lumped into a single, overall reaction by setting the rates of all but one reaction to zero. The predicted absorption and desorption rates from the model at various amine concentrations, C02 partial pressures, temperatures and C02 loadings are consistent with the experimentally observed rates. A numerical procedure was developed to solve the model equations and to estimate unknown model parameters using experimental data. To do this a computer code was written that performs four different tasks involving (a) calculation of physical-chemical properties for a given set of condition, (b) solution of equilibrium model for setting the initial conditions, (c) solution of 210 diffusion-reaction model to calculate concentration gradients at different latitudes 6, and (d) calculation of total absorption or desorption rate by integrating over the entire hemisphere. The Fortran code for tasks (a) and (d) were written as part of this study and that for tasks (b) and (c) were generated using Athena by implementing the model equations in the Athena Visual Workbench environment. These subroutines were compiled together to calculate absorption or desorption rates using the Compaq Visual-Fortran Developer Studio in Windows environment. For the purpose of parameter estimation using the present model and experimental data, the subroutines for tasks (a) to (d) were linked with an additional subroutine generated from Athena that uses weighted least squares and Bayesian estimators with single as well as multi-response data. The computer code has been successfully used to simulate a variety of different operating conditions and to estimate reaction parameters for a number of C02-amine-water systems. The model was solved analytically for the simple case of physical absorption of a gas over a hemispherical liquid film and a correlation was developed for the physical mass transfer coefficient as a function of dimensionless numbers. The correlation was later used to estimate the diffusivities of C02 in water and N20 in water and amine solutions. The calculated diffusivities are in good agreement with those reported in the literature. The analytical solution was also compared with the numerical solution and good agreement was found. 211 A novel hemispherical contactor was developed to measure absorption and desorption rates of C02 in aqueous amine solutions under a variety of different operating conditions encountered in actual absorption-stripping processes for gas treating. This contactor is, in principle, similar to the wetted sphere units used in many previous studies on C02-alkanolamine kinetics. However, the present design offers significant improvements over the conventional designs with respect to surface rippling commonly encountered in the lower half of the full sphere unit. The unit was built in-house with a computerized data logging system and was used successfully to measure absorption and desorption rates to calculate diffusivities and kinetic parameters for both physical and chemical solvents. The diffusion-reaction model developed in this work involves many physical chemical property parameters. A parametric sensitivity analyses indicated that accurate values of C02 diffusivity and C02 solubility given by a Henry's law constant are essential for predicting accurate absorption and desorption rates. An extensive literature search was undertaken and all available property data were compiled. It was found that the values of these parameters at stripping temperatures are not available. A series of experiments involving N20-water and N2-alkanolamine systems were conducted to measure diffusivities and Henry's constants of C02 in water using the N20 analogy. Using data obtained in this work and others available in the literature, correlations were developed that can be used for pure water as well as single and binary amine systems covering 212 a wide range of temperatures and amine concentrations. Such correlations are not available in the literature. The absorption and desorption rates of C02 in aqueous solutions of MEA, DEA, MDEA, AMP and their mixtures (MEA+MDEA, MEA+AMP, DEA+MDEA and DEA+AMP) were measured for amine concentrations in the range of 2 to 35 wt%. The absorption experiments were carried out at near atmospheric pressure using pure C02 saturated with water at 293 to 323 K with initially unloaded solutions. The desorption experiments were performed at 333 to 383 K for C02 loading between 0.02 to 0.7 moles of C02 per mole of amine using humidified N2 gas as a stripping medium. The data were analyzed using the rigorous diffusion-reaction model developed in this work. The model predicts the experimental results well for all eight different amine systems studied (see Chapter 4). The results indicate that the theory of absorption with reversible chemical reaction could be applied to predict desorption rates. A zwitterion mechanism adequately describes the reactions between C02 and carbamate forming amines such as MEA, DEA and AMP under both absorption and desorption conditions. The reactions between C02 and aqueous MDEA solutions are best described by a base-catalyzed hydration reaction mechanism. The kinetic data obtained show that desorption experiments can be used to determine both the forward and backward rate constants accurately. The absorption experiments on the other hand could only be used to determine forward rate constants. 213 For MEA, DEA and AMP, the kinetic data obtained under absorption conditions do not extrapolate well to desorption temperatures. Therefore, kinetic data at higher temperatures should be obtained from desorption experiments. The existing absorption data for the C02-MDEA system can be extrapolated to desorption temperatures with reasonable accuracy. This is probably because MDEA is very slow reacting amine and does not form carbamate. For the blended amine systems, it was found that small additions of MEA or DEA (< 5 wt%) significantly enhance the absorption rates of C02 in 25 wt% MDEA solutions. Addition of small quantity of MEA (< 5 wt%) to 25 wt% AMP blend, on the other hand, was found to have very nominal effect on the C02 absorption rates of AMP. No improvement in absorption rates was observed when DEA was added to 25 wt% AMP solutions. 10.2 Main Conclusions from the Work on CO-lnduced Degradation of DEA The kinetics of CO-induced degradation of DEA was studied over the temperature range of 313-413 K and DEA concentrations of 5-50 wt%. A reaction mechanism was proposed and a mathematical model was developed to estimate previously unknown kinetic and solubility parameters from batch absorption experiments. The experimental data are best described by reactions (8.1) to (8.5). Numerical simulation results, based on the estimated parameters, indicate that the primary DEA degradation reaction is the direct reaction of DEA with 214 molecular CO. The formate-DEA reaction, on the other hand, is relatively slow and reversible. Comparison of the simulation results with and without the DEA-formate reaction shows that, at temperatures higher than 343 K, some of the DEAF formed by the direct insertion reaction is hydrolyzed back to give formate ions and protonated DEA. However, at temperatures lower than 343 K, the DEA-formate reaction seems to have no significant effect on the net formation of DEAF and formate ions in the system. The data from the experiment on CO absorption in aqueous solution containing 30 wt% DEA and 18.4 wt% DEAF confirms that the DEA-Formate reaction reversible. Parameter estimates of the rate constants suggest that the reaction rates are quite slow but, over time, nevertheless lead to substantial DEAF and formate ion buildup in the system. The Henry's constant of CO in aqueous DEA solutions was determined by the N2-Anology and from the reaction model. The results are in fairly good agreement. Comparison of the values of HC0_DEA over the temperature range of 313-343 K shows that the value of HC0_DEA decreases with increasing temperature. This trend is similar to that of CO solubility in pure organic solvents reported in the literature. An experiment conducted by absorbing CO in 30 wt% aqueous MDEA solution shows that, unlike DEA, MDEA does not react with CO. There was no experimental evidence suggesting the formation of diethanolacetamide 215 (R1R2NCOCH3) even after 20 hours at 393 K. The rate constant for the formate ion formation reaction was found to be in excellent agreement with that obtained from CO-DEA-H2O system. This further validates the proposed reaction mechanisms and the experimental technique used to estimate the kinetic parameters. The DEAF and formic ions concentrations found at the end of each experiment agreed well with those predicted from the present mathematical model. In most cases the relative error was within + 30%. 10.3 Recommendations for Future Work This work provides important correlations for physical properties and kinetic parameters that cover both absorption and desorption conditions applicable to both single and blended amine systems. These correlations should be implemented in process simulators such as Aspen-Plus, TSWEET and Hysis to simulate industrial gas treating processes involving MEA, DEA, MDEA, AMP or their mixtures and the -model predictions should be compared with pilot plant or industrial plant data. Further work on C02 absorption in partially loaded amine solutions should be carried out to obtain accurate data representing the rich end of the absorber column. More experimental data on C02 absorption/desorption in mixed amine system are desirable. Special attention should be focused on the MEA+MDEA system. The presence of small amounts of MEA in aqueous MDEA solutions significantly improves the capacity of MDEA to absorb C02. To explore optimum 216 MEA concentrations, more experiments should be conducted with solutions containing less than 10 wt% MEA. Blends of MEA and MDEA could be used to economically capture CO2 from flue gases for the purpose of sequestration or enhanced oil recovery, as the regeneration cost of such solvent systems will be significantly lower. It will be useful to implement the kinetic mechanisms and the correlations of the kinetic parameters obtained from the work on CO-induced degradation of DEA in existing process simulators to estimate DEA losses for processes where significant amounts of CO is present in the feed gas. Innovative scheme could be developed to avoid or minimize DEA losses, which may run into millions of dollars in industry. The hemispherical contactor could easily be used to screen new physical or chemical solvents for C02 and/or H2S removal. 217 NOMENCLATURE PART I. a Interfacial area (m2) CAM Concentration of amine (kmol/m3) Cj Concentration of various ionic and non-ionic species as defined by eq. (4.30) Dj Diffusivity of ionic and non-ionic species in aqueous amine solution (m2/s) Dc0 Diffusivity of C02 in amine solution (m2/s) D£02 Diffusivity of C02 in pure water (m2/s) DN20 Diffusivity of N20 in amine solution (m2/s) D°20 Diffusivity of N20 in pure water (m2/s) E Enhancement factor EA Absorption enhancement factor ED Desorption enhancement factor Ec* Maximum possible enhancement factor Fco Flow rate of feed C02 (mmol/s) F^* Flow rate of feed N2 (mmol/s) F^1 Flow rate of dilution N2 (mmol/s) g Gravitational constant (m/s2) Ga Grashof number as defined by eq. (4.77) HC02 Henry's constant of C02 in amine solution (kPa-m3/kmol) H°02 Henry's constant of C02 in pure water (kPa-m3/kmol) HM n Henry's constant of N20 in amine solution (kPa-m3/kmol) 3/ H° Henry's constant of C02 in pure water (kPa-m3/kmol) N20 J kapp Apparent rate constant (1/s) k| Rate constant for reaction (i) k2nd Second order rate constant as defined by eq. (2.2) (m3/kmol s) kg Gas-side mass transfer coefficient (kmol/kPa m2 s) k° Physical mass transfer coefficient (m/s) M Hatta number as defined by eq. (2.2) NA Rate of absorption (kmol/s) No Rate of desorption (kmol/s) Pco2 Partial pressure of C02 in the gas bulk (kPa) Pco2 Equilibrium partial pressure of C02 corresponding to its concentration in the liquid bulk (kPa) Q Liquid flow rate (m3/s) r, Rate of reaction number (i), (kmol/m3 s) S Least square optimization objective function 218 R Radius of the hemisphere (m) Re Reynolds number as defined by eq. (4.75) Sc Schmidt number as defined by eq. (4.76) Sh Sherwood number as defined by eq. (4.101) T Temperature (K) V0 Velocity at the surface of the liquid film (m/s) V9 Velocity distribution in the liquid film (m/s) x Dimensionless distance from the gas-liquid interface y. Measured value of the state variable y(k) Calculated value of the sate variable z Zwitterion Greek Symbols a CO2 loading (mol of COVmol amine) Ae Liquid film thickness as a function of latitude 0 (m) Ao Liquid film thickness at the equator (m) u. Viscosity of aqueous amine solution, (kg/m s) v Kinematic viscosity (m2/s) VAM Stochiometric coefficient p Density of aqueous amine solution, (kg/m3) 0 Angle from the pole xc Contact time (s) PART II. a Gas-liquid interfacial area, (m"1) CDEA DEA concentration, (wt%) C| Liquid phase concentrations of various chemical species as defined in Eq. (6), (kmol/m3) DNz Diffusivity of nitrogen in aqueous DEA solution, (m2/min) dimp Impeller diameter, (m) E Enhancement factor in Eqs. (10) and (11) E Activation energy, (kcal/mol) H Henry's law constant, (kPa/m3 kmol) ki Rate constant for reaction (1), (m3/kmol h) k2 Rate constant for reaction (2), (m3/kmol h) k3 Rate constant of forward reaction (3), (m3/kmol h) L3 Rate constant of reverse reaction (3), (1/h) K3 DEAF hydrolysis constant of reaction (3), (in Eq. 53 m3/kmol, dimensionless in Eq. 54) K4 DEA protonation constant as defined in Eq. (16) K5 Water dissociation constant as defined in Eq. (17) K6 Formic acid dissociation constant as defined in Eq. (42) K7 MDEA protonation constant as defined in Eq. (60) kL Liquid side mass transfer coefficient as defined by Eq. (29), (m/min) 219 N Stirrer speed, (rpm) n'co' Total number of moles of CO transferred from gas bomb to autoclave (mol) n^o Number of moles of CO in the headspace of the autoclave (mol) nDEA Number of moles of DEA in the solution (mol) Pco Initial partial pressure of CO, (kPa) pc0 Partial pressure of CO, (kPa) PN2 Initial partial pressure of N2, (kPa) pN2 Partial pressure of N2, (kPa) r. Rate of reaction, (kmol/m3 h) S Least square optimization objective function R Gas constant Re Reynolds number as defined by Eq. (30) Sc Schmidt number as defined by Eq. (30) T Temperature, K VG Gas volume in the autoclave, (m3) VL Liquid volume in the autoclave, (m3) V Measured value of the state variable ji y(k) Calculated value of the sate variable Greek Symbols u.L Viscosity of aqueous DEA solution, (kg/m-min) pL Density of aqueous DEA solution, (kg/m3) Abbreviations (Part I and II) AMP 2-amino-2-methyl-1-propanol DEA Diethanolamine DEAE Diethylethanolamine DEAF Formyl-diethanolamine DMF Dimethylformamide FCC Fluid catalytic cracking HCN Hydrogen cyanide MDEA Methy-diethanolamine MEA Monoethanolamine RT -CH2CH2OH R2 -CH2CH2OH R3 -CH3 R4 -C(CH3)2CH2OH R-iNH2 MEA R^NH DEA R1R2R3N MDEA R4NH2 AMP 220 RINH; R^NH; R1R2R3NH+ R4NH3 RuNHCOO" R^NCOO R4NHCOO" RiRzNHCO Protonated MEA Protonated DEA Protanated MDEA Protanated AMP MEA carbamate DEA carbamate AMP carbamate DEAF REFERENCES Al-Ghawas, H.A., Ruiz-lbanez, G. and Sandall, O.C., 1989a, Absorption of carbonyl sulfide in aqueous methyldiethanolamine, Chem. Eng. Sci., 44(3), 631. Al-Ghawas, H.A., Hagewiesche, D.P., Ruiz-lbanez, G. and Sandall, O.C., 1989b, Physicochemical properties properties important for carbon dioxide absorption in aqueous methyldiethanolamine, J. Chem. Eng. Data, 34, 385. Alper, E., 1990, Reaction mechanism and kinetics of aqueous solutions of 2-amino-2-methyl-1- propanol and carbon dioxide, Ind. Eng. Chem. Res., 29, 1725. Alvarez-Fuster, C, Midoux, N., Laurent, A. and Charpentier, J.C., 1980, Chemical kinetics of the reaction of carbon dioxide with amines in pseudo m-nth order conditions in aqueous and organic solutions, Chem. Eng. Sci., 35,1717. Aroua, M. K., Ben Amor A., and Sulaiman M. Z., 1997, Temperature dependency of the equilibrium constant for the formation of carbamate from diethanolamine. J. Chem. Eng. Data., 42, 692. Astarita, G., 1961, Carbon dioxide absorption in aqueous monoethanolamine solutions, Chem. Eng. Sci., 16, 202. Astarita, G., Savage, D.W. and Bisio, A., 1983, Gas Treating with Chemical Solvents, John Wiley and Sons, N.Y. Austgen, D.M., Rochelle, G.T., Peng, X. and Chen, C.C., 1989, A model of vapor-liquid equilibrium for aqueous acid gas alkanolamine systems using electrolyte NRTL equation, Ind. Eng. Chem. Res., 28, 1060. Austgen, D.M., Rochelle, G.T. and Chen, C.C., 1991, Model of vapor-liquid equilibria for aqueous acid gas-alkanolamine systems 2. representation of H2S and C02 solubility in aqueous MDEA and C02 solubility in aqueous mixtures of MDEA with MEA or DEA, Ind. Eng. Chem. Res., 30, 1991. Bain, R.S. and Stewart, W.E., 1990, Application of robust nonmonotone descent criteria to the solution of nonlinear algebraic equations, Comp. and Chem. Eng., 15, 203. Barth, D., Tondre, C, Lappai, G. and Delpeuch, J.J., 1981, Kinetic study of carbon dioxide reaction with tertiary amines, J. phys. Chem., 85, 3660. Barth, D., Tondre C. and Delpuech, J.J., 1984, Kinetics and mechanisms of the reactions of carbon dioxide with alkanolamines: a discussion concerning the cases of MDEA and DEA, Chem. Eng. Sci., 39, 1753. 222 Barth, D., Tondre C. and Delpuech, J.J., 1986, Stopped-flow investigations of the reaction kinetics of carbon dioxide with some primary and secondary alkanolamines in aqueous solutions, Int. J. Chem. Kinetics, 18, 445. Blanc, C. and Demarais, G., 1984, The reaction rate of C02 with diethanolamine, Intl. Chem. Eng., 24, 43. Blauwhoff, P.M.M., Versteeg, G.F. and van Swaaij, W.P.M., 1984, A study on the reaction between C02and alkanolamines in aqueous solutions, Chem. Eng. Sci., 39, 207. Blauwhoff, P.M.M., Kamphuis, B.,' van Swaaij, W.P.M. and Westerterp, K.R., 1985, Absorber design in sour natural gas treatment plants: impact of process variables on operation and economics, Chem. Eng. Process., 19, 1. Bosch, H., Versteeg, G.F. and Van Swaaij, W.P.M., 1990a, Desorption of acid gases (C02 and H2S) from loaded alkanolamine solutions, In Gas Separation Technology, edited by Vansant, E.F. and Dewolfs, R., Elsvier Science Publishers B.V., Amsterdam, 505-512. Bosch, H., Versteeg, G.F. and van Swaaij, W.P.M., 1990b, Kinetics of the reaction of C02 with the sterically hindered amine 2-amino-2-methylpropanol at 298 K, Chem. Eng. Sci., 45, 1167. Bottoms, R.R., 1930, U.S. Patent 1,783, 901. Bower, V.E., Robinson, R.A. and Bates, R.G., 1962, Acidic dissociation constant and related thermodynamic quantities for diethanolammonium ion in water from 0 to 50 °C, J. Res. NBS-A. Physics and Chemistry, 66A, 71. Brenan, K.E., Campbell, S.L. and Petzold, L.R., 1989, Numerical solution of initial value problems in differential-algebraic equations, North Holland, Amsterdam. Browning, G.J. and Weiland, R.H., 1994, Physical solubility of carbon dioxide in aqueous alkanolamines via nitrous oxide analogy, J. Chem. Data, 39, 817. Caplow, M., 1968, Kinetics of carbamate formation and breakdown, J. Am. Chem. Soc, 90,6795. Caracotsios, M. and Stewart, W.E., 1985, Sensitivity analysis of initial value problems with mixed ODEs and algebraic equations, Comp. & Chem. Eng., 9, 359. Caracotsios, M. and Stewart, W.E., 1995, Sensitivity analysis of initial-boundary-value problems with mixed PDEs and algebraic equations; Applications to chemical and biochemical systems, Comp. and Chem. Eng., 19,1019. 223 Chakma, A, 1987, Studies on DEA and MDEA degradation, Ph.D. Thesis, University of British Columbia, Vancouver, B.C. Chakma, A, and Meisen, A., 1988, Identification of methyldiethanolamine degradation products by gas chromatography and gas chromatography-mass spectrometry, J. Chrom., 457, 287. Clarke, J.K.A., 1964, Kinetics of absorption of carbon dioxide in monoethanolamine solutions at short contact times, I & EC Fundamentals 3(3),239. Chakraborty, A.K., Astarita, G. and Bischoff, K.B., 1986, C02 absorption in aqueous solutions of hindered amines, Chem. Eng. Sci., 41(4), 997. Coldrey, P.W. and Harris, L., 1976, Kinetics of the liquid phase reaction between carbon dioxide and diethanolamine, Can. J. Chem. Eng., 51, 566. Critchfield, J. E. and Rochelle, G.T., 1987, C02 Absorption into aqueous MDEA and MDEA/MEA solutions, Presented at the AIChE National Meeting, Paper No. 43e, Houston, TX. Critchfield, J.E., 1988, C02 absorption desorption in methyldiethanolamine solutions promoted with monoethanolamine and diethanolamine: mass transfer and reaction kinetics, Ph.D. dissertation, The University of Texas, Austin. Critchfield, J.E. and Rochelle, G.T., 1988, C02 absorption from DEA and DEA-promoted MDEA solutions, Paper 68b, presented at AICHE spring national meeting, New Orleans, LA. Crooks, J.E. and Donnellan, J.P., 1989, Kinetics and mechanism of the reaction between carbon dioxide and amines in aqueous solutions, J .Chem. Soc. Perkin. Trans. II, 331. Danckwerts, P.V. and Sharma, M.M., 1966, The absorption of carbon dioxide into solutions of alkalis and amines (with some notes on hydrogen sulfide and carbonyl sulfide), Chem. Engr, Oct. CE244. Danckwerts, P.V. and McNeil, K.M., 1967, The absorption of carbon dioxide into aqueous amine solutions and the effects of catalysis, Trans. Inst. Chem. Eng., 45, T32. Danckwerts, P.V., 1970, Gas liquid reactions, McGraw Hill, NY. Danckwerts, P. V., 1970, The reaction of C02 with ethanolamines, Chem. Eng. Sci., 34, 443. 224 Danckwerts, P.V., 1979, The reaction of CO2 with ethanolamines, Chem. Eng. Sci., 34, 443. Davidson, J.F. and Cullen, E.J., 1957, The determination of diffusion coefficients for sparingly soluble gas in liquids, Trans. Inst. Chem. Engrs, 35, 51. Davis, R.A. and Sandall, O.C., 1993, Kinetics of the reaction of carbon dioxide with secondary amines in polyethylene glycol, Chem. Eng. Sci., 48(18), 3187. Dawodu, O.F. and Meisen, A., 1991, Identification of products resulting from carbonyl sulfide induced degradation of diethanolamine, J. Chrom., 587, 237. DeCoursey, W.J., 1982, Enhancement factors for gas absorption with reversible reaction, Chem. Eng. Sci., 37, 1483. DeCoursey, W.J., 1992, A simpler approximation for enhancement of mass transfer by second-order reversible reaction, IChemE Symp. Ser. No. 128, 2, B269. Dehouche, Z. and Lieto, J., 1995, Modelling and experimental study of key parameters of absorption on wetted sphere contactor, Chem. Eng. Sci., 50, 2899. Deshmukh, R.D. and Mather, A.E., 1981, A mathematical model for equilibrium solubility of hydrogen sulphide and carbon dixode in aqueous alkanolamine solutions, Chem. Engng. Sci., 36, 355. DiGuillo, R. M., Lee, R. J., Schaeffer, S. T., Brasher, L. L. and Teja, A. S., 1992, densities and viscosities of the ethanolamines, J. Chem. Eng. Data, 37, 239. Donaldson, T.L. and Y.N. Nguyen, 1980, Carbon dioxide reaction kinetics and transport in aqueous amine membranes, Ind. Eng. Chem. Fundam., 19, 260. Duranleau, R.G. and Lambert, C.L., 1985, Process for the preparation of formamide compunds, US patent 4529822, (TEXACO Inc). Duda, J.F., and Verntas, J.C., 1968, Laminar liquid jet diffusion studies, AIChE J., 14, 286. Edwards, T. J., Maurer, G., Newman, J., Prausnitz, J. M., 1978, Vapor-liquid equilibria in multicomponent aqueous solutions of volatile weak electrolytes. AIChE J., 24, 966. Eickmeyer, A.G., 1962, Catalytic removal of C02, Chem. Eng. Prog., 58(4), 89. 225 Ellis, A.J., and Golding, R.M., 1963, The solubility of carbon dioxide above 100 °C in water and in sodium chloride solutions, Am. J. Sci., 261, 47. Fogg, P.G.T. and Gerrard, W., 1991, Solubility of gases in Liquids, John Wiley & Sons, NY, 274. Glasscock, D.A., 1990, Modeling and experimental study of carbon dioxide absorption into aqueous alkanolamines, Ph.D. dissertation, The University of Texas, Austin. Glasscock, D.A., Critchfield, J.E., and Rochelle, G.T., 1991, C02 absorption/desorption in mixtures of methyldiethanolamine with monoethanolamine or diethanolamine, Chem. Eng. Sci., 46, 2829. Goldstein, A.M., Brown, E.C., Heinzelmann, F.J. and Say, G.R., 1986, New FLEXORB gas treating technology for acid gas removal. Energy. Prog., 7, 67. Groothius, H., 1966, de Ingenieur, 78, CH9. Hagewiesche, D.P., Ashour, S.S., Al-Ghawas, H.A. and Sandall, O.C., 1995a, Absorption of carbon dioxide into aqueous blends of monoethanolamine and n-methyldiethanolamine, Chem. Eng. Sci., 50, 1071. Hagewiesche, D. P., Ashour, S. S. and Sandall, O. C, 1995b, Solubility and diffusivity of nitrous oxide in ternary mixtures of water, monoethanolamine, and N methyldiethanolamine and solution densities and viscosities, J. Chem. Eng. Data, 40, 627. Haimour, N. and Sandall, O.C., 1984, Absorption of carbon dioxide into aqueous methyldiethanolamine, Chem. Eng. Sci., 39, 1791-1796. Haimour, N.M., 1990, Solubility of N20 in aqueous solutions of diethanolamines at different temperatures, J. Chem. Eng. Data., 35, 177. Haimour, N.A. Bidarian, A. and Sandall, O.C., 1985, Absorption of C02 into aqueous solutions of methyldiethanolamine, Paper presented at AICHE spring national meeting, Houston. Haimour, N., Bidarian, A. and Sandall, O.C., 1987, Kinetics of the reaction between carbon dioxide and methyldiethanolamine, Chem. Eng. Sci., 42(6), 1393. Harned, H.S. and Embree, N.D., 1934, The ionization constant of formic acid from 0 to 60 °C. J. Amer. Chem. Soc, 56, 1042-1044. Harned, H.S. and Robinson, R.A., 1940, Trans. Faraday Soc, 36, 973. 226 Hikita, H., Asai, S., Ishikawa, N. and Saito, Y., 1975, Kinetics of absorption of chlorine in aqueous acidic solutions of ferrous chloride, Chem. Eng. Sci. 30, 607. Hikita, H., S. Asai, H. Ishikawa, and M. Honda, 1977a, The kinetics of reactions of carbon dioxide with monoethanolamine, diethanolamine, and triethanolamine by a rapid mixing method, Chem. Eng. J., 13,7. Hikita, H., S. Asai, H. Ishikawa, and M. Honda, 1977b, The kinetics of reactions of carbon dioxide with monoisopropanolamine, diglycolamine, and ethylenediamine by a rapid mixing method, Chem. Eng. J., 14, 27. Hsu, C. and Li, M., 1997a, Densities of aqueous blended amines, J. Chem. Eng. Data. 42, 502. Hsu, C. and Li, M., 1997b, Viscosities of aqueous blended amines, J. Chem. Eng. Data. 42, 714. Jensen, A., Jorgensen, E., Faurholt, C, 1954a, Reactions between carbon dioxide and amino alcohols. I. monoethanolamine and diethanolamine, Acta Chemica Scandinavica, 8, 1137. Jensen, A., M.B. Jensen, and C. Faurholt, 1954b, Studies on carbamates: the carbamates of di-n-propylamine and di-iso-propylamine, Acta Chemica Scandinavica, 8,1129. Jorgensen, E., 1956, Reactions between carbon dioxide and amino alcohols III. diethanolamine, Acta Chemica Scandinavica, 10,747. Katti, S.S., 1995, Gas-liquid-solid system: an industrial perspective. Trans. IchemE. 73, 595. Katti, S.S. and McDougal, K.A., 1986, Rate approach in the design and simulation of chemical absorbers: theory and practice, Paper presented at the annual meeting, AICHE, Miami Beach Florida. Kennard, M.L. and Meisen, A., 1985, Mechanisms and kinetics of diethanolamine degradation, Ind. Eng. Chem. Fund., 24,129. Kell, G. S., 1975, Density, thermal expansivity, and compressibility of liquid water from 0 to 150 °C: correlations and tables for atmospheric pressure and saturation reviewed and expressed on 1968 temperature scale, J. Chem. Eng. Data, 20, 97. Kent, R.L. and Eisenberg, B., 1976, Better data for amine treating, hydrocarbon process., 55, 87. 227 Kim, C.J., 1988, Degradation of alkanolamines in gas treating solutions: kinetics of di-2-propanolamine degradation in aqueous solutions containing carbon dioxide, Ind. Eng. Chem. Res. 27, 1. Kim, C.J. and Sartori, G., 1984, Kinetics and mechanism of diethanolamine degradation in aqueous solutions containing carbon dioxide, Int. J. Chem. Kinetics, 16, 1257. Kim, C.J., Palmer, A.M. and Milliman, G.E., 1988, Absorption of carbon monoxide into aqueous solutions K2C03, methyldiethanolamine and diethyiethanolamine, Ind. Eng. Chem. Res., 27, 324. Kirk Othmer, 1993, Encyclopedia of Chemical Technology, 5, 4th edition. Kohl, A. L. and Riesenfeld, F. C. 1985, Gas Purification, 4th ed.; Gulf: Houston, TX. Ko, J.J. and Li, M.H., 2000, Kinetics of absorption of carbon dioxide into solutions of N-methyldiethanolamine + water, Chem. Eng. Sci., 55, 4139. Koike, L., Barone, J., Godinho, O.E.S., Aleixo, L.Z. and Reis, F., 1987, N-formyldiethanolamine: a new artifact in diethanolamine solutions, Chem. Ind., 17, 626. Krishnamurthy, R. and Taylor, R., 1985, A Non-equilibrium stage model of multicomponent separation processes, AIChE J., 31, 449-455, Ibid, 1985, 456. Laddha, S.S. and Danckwerts, P.V., 1981, Reaction of C02 with ethanolamines: kinetics from gas-absorption, Chem. Eng. Sci., 36, 479. Laddha, S.S., Diaz, J.M. and Danckwerts, P.V., 1981, The N20 analogy: the solubilities of C02 and N20 in aqueous solutions of organic compounds, Chem. Eng. Sci. 36, 228. Laddha, S.S., and Danckwerts, P.V., 1982, The absorption of C02 by amine-potash solutions, Chem. Eng. Sci., 37, 665. Lambert, CL. and Duranleau, R.G., 1985, Process for the preparation of formylalkanoalmines, US patent 4510326, (TEXACO Inc). Leder, F., 1971, The Absorption of C02 into chemically reactive solutions at high temperatures, Chem. Eng. Sci. 26, 1381. Levenspiel, O, 1972, Chemical reaction engineering, John Wiley & Sons Inc.,NY. 228 Lewis, W.K. and Whitman, W.G., 1924, Principles of gas absorption, Ind. Eng. Chem., 16 (12), 1215. Li, M. H. and Shen, K. P., 1992, Densities and solubilities of solutions of carbon dioxide in water + monoethanolamine + N-methyldiethanolamine, J. Chem. Eng. Data 1992, 37, 288. Li, M. H. and Lie, Y. C, 1994, Densities and viscosities of solutions of monoethanolamine + N-methyldiethanolamine + water and monoethanolamine + 2-amino-2-methyl-1 propanol + water, J. Chem. Eng. Data, 39, 444. Li, M.H. and Lai, M.D., 1995, Solubility and diffusivity of N20 and C02 in (monoethanolamine + N-methyldiethanolamine +water) and in (monoethanolamine + 2-amino-2-methyl-1-propanol + water), J. Chem. Eng. Data, 40, 486. Li, M.H. and Lee, W.C., 1996, Solubility and diffusivity of N20 and C02 in (diethanolamine + N-methyldiethanolamine +water) and in (diethanolamine + 2-amino-2-methyl-1 -propanol + water), J. Chem. Eng. Data, 41, 551. Littel, R.J., van Swaaij, W.P.M., and Versteeg, G.F., 1990a, The kinetics of carbon dioxide with tertiary amines in aqueous solution, Presented at the AIChE Spring National Meeting, Orlando, FL. Littel, R.J., Bos, M. Knoop, G.J., 1990b, Dissociation constants of some alkanolamines at 293, 303, 318, and 333 K, J. Chem. Eng. Data, 35, 276. Littel, R.J., Versteeg, G.F. and van Swaaij, W.P.M., 1992a, Kinetics of C02 with primary and secondary amines in aqueous solutions- I. zwitterion deprotonation kinetics for DEA and DIPA in aqueous blends of alkanolamines, Chem. Eng. Sci., 47, 2027. Littel, R.J., Versteeg, G.F. and van Swaaij, W.P.M., 1992b, Kinetics of C02 with primary and secondary amines in aqueous solutions- II. Influence of temperature on zwitterion formation and deprotonation rates, Chem. Eng. Sci., 47, 2037. Little, R.J., Versteeg, G.F. and van Swaaij, W.P.M., 1992c, Solubility and diffusivity data for the absorption of COS, C02 and N20 in amine solutions, J. Chem. Eng. Data, 37, 49. Lynn, S., Straatemeier, J.R. and Kramers, H., 1955, Absorption studies in the light of the penetration theory. III. absorption by wetted spheres, singly and in columns, Chem. Eng. Sci., 4, 63. Malinin, S.D., 1975, Thermodynamics of the H20-C02 systems, Geochem. Int., 11, 1060. 229 Markham, A.E. and Kobe, K.A.,1941, Ind. Eng. Chem., 63, 449. Messaoudi, B. and Sada, E., 1996, Absorption of carbon dioxide into loaded aqueous solutions of 2-amino-2-methyl-1 propanol, J. Chem. Eng. Japan, 29, 534. McDonald, R. D., 1963, The kinetics of carbon monoxide in basic solution at elevated temperatures, MSc Thesis, The University of British Columbia, Vancouver, BC. Olofsson, G. and Hepler, L.G., 1975, The thermodynamics of ionization of water over wide range of temperature and pressure, J. Soln. Chem., 4, 127. Pacheco, M.A., Kaganoi, S. and Rochelle, G.T., 2000, CO2 absorption into aqueous mixtures of diglycolamine and methyldiethanolamine, Chem. Eng. Sci., 55, 5125. Pani, F, Gaunand, A., Cadours, R, Bouallou, C. and Richon, D., 1997, Kinetics of absorption of C02 in concentrated aqueous methyldiethanolamine solutions in the range 296-343K, J.Chem. Eng. Data/42, 353-359. Penny, D.E. and Ritter, T.J., 1983, Kinetic study of the reaction between carbon dioxide and primary amines, J.Chem.Soc, Faraday Trans.79, 2103. Perry, J.H., Chilton, CH. and Kirkpatrick S.D., 1963, 1984, Chemical Engineers' Handbook, McGraw Hill, NY, 5th and 6th Edn, 14-4. Petzold, L.R., 1983, A description of DASSL: a differential/algebraic system solver in scientific computing, eds. Stepleman et al., North Holland, Amsterdam, 65-68. Prausnitz, J.M., Lichtenthaler, R.N. and de Azevedo, E.G., 1986, Molecular thermodynamics of fluid-phase equilibria, 2nd edn., Prentice-Hall Englewood Cliffs, NJ. Pinsent, B.R.W., Pearson, W.L., Roughton, F.J.W., 1956, The kinetics of the combination of carbon dioxide with hydroxide ions. Trans. Fraday Soc, 52,1512. Rangwala, R.A., Morrell, B.R., Mather, A.E. and Otto, F.D., 1992, Absorption of CO2 into aqueous tertiary amine solutions, Can. J. Chem. Eng., 70,482. Richard, A.D. and Sandall, O.C., 1993, Kinetics of the reaction of carbon dioxide with secondary amines in polyethylene glycol, Chem. Eng. Sci., 48, 3187. 230 Rinker, E.B., Ashour, S.S. and Sandall, O.C., 1995a, Kinetics and modeling of carbon dioxide absorption into aqueous solutions of N-methyldiethanolamine, Chem. Eng. Sci., 50(5),755. Rinker, E. B., Russell, J. W., Tamimi, A. and Sandall, 0. C, 1995b, Diffusivity of nitrous oxide in N-methyldiethanolamine + diethanolamine + water, J. Chem. Eng. Data, 40, 630. Rinker, E. B., Oelschlager, D. W., Colussi, A. T., Henry, K. R., and Sandall, O. C, 1994, Viscosity, density, and surface tension of binary mixtures of water and N-methyldiethanolamine and water and diethanolamine and tertiary mixtures of these amines with water over the temperature range 20-100 °C, J. Chem. Eng. Data, 39, 392. Rinker, E.B., Ashour, S.S. and Sandall, O.C., 1996, Kinetics and modeling of carbon dioxide absorption into aqueous solutions of diethanolamine, Ind. Eng. Chem. Res., 35, 1107. Read, A.J., 1975, The first ionization constant of carbonic acid from 25 to 250 °C and to 2000 bar, J. Soln. Chem., 4, 53. Reid, R.C, Prausnitz, J.M. and Poling, B.E., 1987, The properties of gases and liquids, 4th Edn, McGraw Hill, NY. Rowley, R.L., Adams, M.E., Marshall, T.L., Oscarson, J.L., Wilding, W.V. and Anderson, D.J., 1998, Measurement of the absorption rate of carbon dioxide into aqueous diethanolamine, J. Chem. Eng. Data, 43, 427. Sada, E., Kumazawa, H. and Butt, M.A., 1977, Solubilities of gases in aqueous solutions of amine, J. Chem. Eng. Data, 22, 277. Sada, E. and Kito, S., 1972, Kagaku Kogaku, 36, 218. Sada, E., Kumazawa, H. and Butt, M.A., 1976, Gas Absorption with consecutive chemical reaction: absorption of carbon dioxide into aqueous amine solutions, Can. J. Chem. Eng., 54, 421. Sada, E., Kumazawa, H. and Butt, M. A., 1978, Solubility and diffusivity of gases in aqueous solutions of amines. J. Chem. Eng. Data, 23, 161. Sada, E., Kumazawa, H., and Han, Z.Q., 1985, Kinetics of reaction between carbon dioxide and ethylenediamine in nonaqueous solvents, Chem. Eng. J., 31,109. 231 Saha, A.K., Bandyopadhyay, S.S. and Biswas, A.K., 1993, Solubility and diffusivity of N20 and C02 in aqueous solutions of 2-amino-2-methyl-1-propanol, J. Chem. Eng..Data, 38, 78. Saha, A.K., Bandyopadhyay, S.S. and Biswas, A.K., 1995, Kinetics of absorption of CO2 into aqueous solutions of 2-amino-2-methyl-1-propanol, Chem. Eng. Sci. 50(22), 3587. Sandall, O.C., Rinker, E.B. and Ashour, S., 1993, Acid gas treating by aqueous alkanolamines, Annual report to the Gas Research Institute. Sardar, H., 1985, Development of a non-equilibrium stage model for the design and simulation of gas processing units and verification with plant data, PhD Dissertation at Clarkson University, Pottsdam, N.Y. Sartori, G. and Savage, D.W., 1983, Sterically hindered amines for C02 removal from gases, Ind. Eng. Chem. Fundam., 22, 239. Sartori G., Ho, W. S. and Savage, D.W., 1987, Sterically hindered amines for acid gas absorption, Separation and Purification Methods, 16,171. Savage, D.W., Astarita, G. and Joshi, S., 1980, Chemical absorption and desorption of carbon dioxide from hot carbonate solutions, Chem. Eng. Sci., 35, 1513. Sharma, M.M., 1964, Thesis, Cambridge University. Sharma, M.M., 1965, Kinetics of reactions of carbonyl sulphide and carbon dioxide with amines and catalysis by bronsted bases of the hydrolysis of COS, Trans. Faraday Soc,61, 681. Song, J. H., Park, S. B., Yoon, J. H., Lee, H. and Lee, K. H., 1996, Densities and viscosities of monoethanolamine + ethylene glycol + water, J. Chem. Eng. Data, 41, 1152. Stewart, W.E., Caracotsios, M. and Sorenson, J.P., 1992, Parameter estimation from multi response data, AIChE J., 42, 1404. Stewart, W.E., Henson, T.L. and Box, G.E.P., 1996, Model discrimination and criticism with single-response data, AIChE J., 42, 3055. Supap, T., Idem, R., Veawab, A., Aroonwilas, A., Tontiwachwuthikul, P., Chakma, A. and Kybett, B. D., 2001, Kinetics of the oxidative degradation of aqueous monoethanolamine in a flue gas treating unit, Ind. Eng. Chem. Res., 40, 3445. 232 Takenouchi, S. and Kennedy, G.C., 1964, The binary system H2O-CO2 at high temperatures and pressures, Am. J. Sci., 262, 1055. Tamimi, A. Rinker, E.B. and Sandall, O.C., 1994a, Diffusion coefficients for H2S, C02 and N20 in water over the temperature range 293 - 368 K, J. Chem. Eng. Data, 39, 330. Tamimi, A. Rinker, E.B. and Sandall, O.C., Diffusivity of N2O in aqueous solutions of methyldiethanolamine and diethanolamine from 293 - 368 K, J. Chem. Eng. Data, 39, 396 (1994b). Toman, J.J. and Rochelle, G.T., 1989, Carbon dioxide absorption rates and physical solubility in 50% aqueous methyldiethanolamine partially neutralized with sulfuric acid, Presented at the AIChE Spring National Meeting, Houston, TX. Toman, J.J. and Rochelle, G.T., 1990, C02 Absorption rates into aqueous solutions of 2-(tert- butylamino)ethanol, presented at the Spring National AIChE Meeting, Orlando, FL. Tomcej, R.A. and F.D. Otto, 1989, Absorption of C02 and N20 into aqueous solutions of methyldiethanolamine, AIChE., 35(5), 861. Treybal, R.E., 1980, Mass transfer operations, McGraw-Hill, N.Y. Tseng, P.C., W.S. Ho, D.W. Savage, 1988, Carbon dioxide absorption into promoted carbonate solutions, AIChE J., 34, 922. Van Krevelen, D.W. and Hoftijzer, P.J., 1948, Kinetics and simultaneous absorption and chemical reaction, Chem. Eng. Prog., 44, 7, 529. Versteeg, G.F. and van Swaaij, W.P.M., 1988, On the kinetics between C02 and alkanolamines both in aqueous and non-aqueous solutions- II. tertiary amines, Chem. Eng. Sci., 43, 587. Versteeg, G.F. and van Swaaij, W.P.M., 1988, Solubility and diffusivity of acid gases (C02 and N20) in aqueous alkanolamine solutions, J. Chem. Eng. Data, 33, 29. Versteeg, G.F. and Oyevaar, M.H., 1989, The reaction between C02 and diethanolamine at 298 K, Chem. Eng. Sci., 44, 1264. Wang, Y. W., Xu, S., Otto, F. D. and Mather, A. E, 1992, Solubility of N20 in alkanolamines and in mixed solvents. Chem. Eng. J., 48, 31. 233 Weiland, R.H., Chakravarty, T. and Mather, A.E., 1993, Solubility of carbon dioxide and hydrogen sulfide in aqueous alkanolamines, Ind. Eng. Chem. Res., 32,1419. Wild, J.D., and Potter, O.E., 1968, Gas absorption with chemical reactions into liquid flowing on a sphere, Inst. Chem. Eng. Symp. Ser., 28, 30. Xiao, J., Li, C.W., and Li, M.H., 2000, Kinetics of absorption of carbon dioxide into aqueous solutions of 2-amino-2-methyl-1 -propanol + monoethanolamine, Chem. Eng. Sci., 55, 161. Xu, S., Wang, Y., Otto, F.D. and Mather, A.E., 1991a, Rate of absorption of C02 in a mixed solvent, Ind. Eng. Chem. Res., 30, 1213. Xu, S.; Otto, F. D.; Mather, A. E., 1991b, Physical properties of aqueous AMP solutions, J. Chem. Eng. Data, 36, 71. Xu, G.W, Zhang, C.F., Qin, S.I. and Wang, Y.W., 1992a, Kinetics study on absorption of carbon dioxide into solutions of activated methyldiethanolamine, Ind. Eng. Chem. Res., 31,921. Xu, S.; Wang, Y.; Otto, F. D.; Mather, A. E., 1992b, Physicochemical properties of 2-piperidineethanol and its aqueous solutions, J. Chem. Eng., 37, 407. Xu, G.W., Zhang, C.F., Qin, S.J. and Zhu, B.C., 1995, Desorption of C02 from MDEA and activated MDEA solutions, Ind. Eng. Chem. Res., 34, 874. Xu, S., Wang, Y.W., Otto, F.D. and Mather, A.E., 1996, Kinetics of the reaction of carbon dioxide with 2-amino-2-methyl-1 -propanol solutions, Chem. Eng. Sci., 51, 841. Yih, S. M. and Shen, K.P., 1988, Kinetics of carbon dioxide reaction with sterically hindered 2-amino-2-methyl-1-propanol aqueous solutions, Ind. Eng. Chem. Res., 41, 2237. Yaws, C. L, Miller, J. W., Shah, P. N., Schorr, G. R. and Patel, P. M., 1976, Chem. Eng., 83,153. Yu, W. C, Astarita, G. and Savage, D.W., 1985, Kinetics of carbon dioxide absorption in solutions of methyldiethanolamine, Chem. Eng. Sci., 40(8), 1585. Yu, W.C. and G. Astarita, 1987, Design of packed towers for selective chemical absorption, Chem. Eng. Sci., 42, 425 234 APPENDIX A DETERMINATION OF C02 LOADING IN AMINE SOLUTIONS A.1 Detection Principle The total CO2 content in the liquid phase or the CO2 loading was determined by a novel method developed by Shell Technology Center, Houston, Texas. In this method a small amount of liquid sample (0.1-1.0 gm) is injected into a solution of 30 wt% hydrochloric acid which instantly frees the CO2 chemically combined with the amine. The nitrogen gas carries the total C02 to a Gastec tube filled with an indicating layer of a pale white color substance that contains hydrazine. In the Gastec tube, CO2 reacts with hydrazine to form carbonic acid monohydrazide, which changes the color of redox indicator to a bluish-violet according to following reaction: C02 +N2H4 -> NH2NHCOOH The length of the stain formed is directly proportional to the C02 concentration in the sample. A. 2 Experimental Setup and Procedure Figure A.1 shows the experimental setup for determining CO2 in amine solution. The gas analysis vessel shown in Figure A.1 was specifically designed for this type of experiments and was supplied by Shell Technology Center, 235 Houston, Texas. The Gastec tubes used in this work were type 2HH and were obtained from Gastec Corporation, Fukaya, Japan. N2 In (120 mL/min @ 5 psig) Cone. NaOH Solution Vesse A Cone. HCL 15 mL •!v' M m Drying Tube Gastec Tube N2 Out Amine Sample (-0.1-1 gm) Gas Analysis Vessel Figure A.1 Experimental setup for CO2 measurement in Liquid Samples In a typical run, the gas analysis vessel was thoroughly rinsed with distilled water and 15 mL of 30-wt% HCL solution was dispensed into the vessel and the cap was tightened. A low-pressure nitrogen supply line was attached to the inlet of the analysis vessel. Before entering the analysis tube, the nitrogen stream was first bubbled through a tube containing concentrated NaOH solution and then to drying tube. This was done to remove any traces of CO2 or water vapor present in nitrogen gas from the supply cylinder. The N2 supply was turned on and the upstream pressure was adjusted to 5 psi. The entire system was sparged for 5 minutes and checked for leaks using snoop. After breaking the tips off both ends, the Gastec tube was inserted at the outlet of the analysis vessel and kept in horizontal position by a metal stand. The nitrogen gas was allowed to sparge through the tube for 5 minutes to make sure there was no C02 236 contamination. During this period If there was any stain, the tube was discarded and a new tube was used. Once the system was C02 free, small amount of sample (0.1-1.0 gm, depending on C02 loading) was withdrawn from the air-tight sample bottle using a Hamilton gas-tight syringe, weighed in an electronic balance and injected into the sample port through the septum. At this stage nitrogen flow was adjusted to 120 mL/min and was kept at this throughout the run. The sparging was continued for exactly 40 mins. The nitrogen flow rate and the sparge time were determined by conducting a number of trial runs during the development of the procedure. In the analysis, if the stain over ranged the tube, it was repeated using smaller sample size. If there was no stain, the analysis was repeated using larger sample size. For each run, the vessel was drained and rinsed with water. Also, each analysis was repeated at least twice using a new tube every time. After each run, the Gastec tube was quickly removed from the system and the stain length was measured using a Vernier scale. If the stain length was not uniform, the maximum and minimum lengths were measured and an average was taken. The amount of C02 in a liquid sample was calculated using the following equation: C02 Loading ( molsofCQ2 ^> ^mols of Amine Rf x Stain Length(cm)/44 Amine wt%xSample Size(mg)/Amine mol. wt. The response factor, Rf, was obtained from the calibration curve prepared by injecting a liquid sample with known amount of C02. To do this a standard solution that contained 10,000 ppm (weight) as C02 stock was prepared using 237 sodium carbonate (24.1 gm Na2C03 up to 1000 gm) in water. The relationship between the C02 content and the stain length was obtained by injecting standard samples ranging from 0 to 2.5 gm. Each injection was repeated at least twice and the responses were averaged. The raw data are reported in Table A.1 and the calibration curve is presented in Figure A.2. The value of the correlation coefficient (R2 = 0.9956) indicates that reproducibility of the measurements is excellent. 40 • Measurements Regression y = 6.3278x Gastec Tube No. = 2HH N2 Flow = 120 mL/min Sparge Time = 40 mins J L 0 2 3 4 5 6 Stain Length (cm) Figure A.2 Calibration Curve for Gastec Detector Tube No. 2HH for C02 Measurement in Liquid Samples 238 Table A.1: Calibration Data for Gastec Detector Tubes No. 2HH (Standard Solution: 10, 000 ppm (by weight) Na2C03, N2 Flow = 120 ml/min, Sparge Time = 40 min) Sample Volume Sample Weight C02in Sample Tube Reading (Avg. of 2 tubes) ml gm mg cm 0.0 0.00 0.00 0.00 0.5 0.51 5.10 1.03 1.0 1.02 10.20 1.63 1.5 1.53 15.30 2.54 2.0 2.04 20.40 3.14 2.5 2.55 25.50 3.95 3.0 3.06 30.60 4.84 239 APPENDIX B CALIBRATION OF MEASURING INSTRUMENTS B.1 Pressure Transducers The calibration of pressure transducers installed on the heating tank and absorption/desorption chamber was done by pressurizing the apparatus from ambient to 20 psig and then bringing it back to ambient pressure in a stepwise manner. The calibration curve was generated using the average values. The actual pressure was measured in mm of Hg using mercury manometer and the output signal from the transducer was measured in volts using the computer data acquisition system. The calibration curves and corresponding correlations are shown in Figure B.1 and B.2. B.2 Mass Flow Meters The mass flow meters were calibrated for CO2, N20 and N2 by means of a soap film meter. A soap film assembly as shown in Figure B.3 was set up and the ambient temperature and pressure was recorded. The time taken for a soap bubble to travel between two marks was noted using a stopwatch. For each flow at least three measurements were collected and average was used in generating the calibration curve. The output signal from the mass flow meter was measured using the computer data acquisition system. The same procedure was adopted 240 Figure B.1 Calibration Curve for Pressure Transducer on Absorption/Desorption Chamber(Omega Model PX202-030GV) 20 .9? 15 to lo Q. <D 3 10 co CD -y = 25.616x +0.2149 r R2 = 0.9997 • Measurements C_i i i 1 i_ — Regression i i i i i i—i—i—i— 0.0 0.2 0.4 0.6 0.8 Output Signal after Zeroing (volt) 1.0 Figure B.2 Calibration Curve for Pressure Transducer on Heating Tank (Omega Model PX202-030GV) 241 Im C02/N20/N2H^L-rfk|—K>!<]—J Data Acquisition Gas Out Mass Flowmeter Soap Film Meter Figure B.3 Setup for Mass Flow Meter Calibration using a Soap Film Meter 2000 I E 1000 500 0 i y = 402.26X - 60.821 : R2 = 0.9981 T = 22.5 °C P = 1 atm • Measurements • i i i l •• i i l i i Regression 0 1 2 3 4 5 Output Signal after Zeroing (volt) Figure B.4 Calibration Curve for Measuring Dilution N2 Flow Rate using Mass Flow Meter (Brooks Model 5700) 242 for all mass flow meters. The calibration curves for the three mass flow meters used in this apparatus are plotted in Figure B.4 to B.7. Figure B.5 Calibration Curve for Measuring Stripping N2 Flow Rate using Mass Flow Meter (Colepalmer Model GFM171) c E E o 1500 1000 „ 500 O O "y = 876.59x + 63.531 '. R2 = 0.9997 y T = 23.4 °C P = 1 atm JT • Measurements • • i i i i i i—i—i — Regression i i i i i i i i—i—i— 0.0 0.5 1.0 1.5 Output Signal after Zeroing (volt) 2.0 Figure B.6 Calibration Curve for Measuring C02 Flow Rate using Mass Flow Meter (Colepalmer Model GFM171) 243 1500 h y = 854.28X + 59.257 i i • • • 0.0 0.5 1.0 1.5 2.0 Output Signal after Zeroing (volt) Figure B.7 Calibration Curve for Measuring N20 Flow Rate Using Mass Flow Meter (Colepalmer Model GFM 171) B.3 C02 Analyzer The C02 absorption or desorption rate were determined by the difference of the CO2 flow rate in the gas stream into and out of the hemispherical contactor. The flow rate of C02 into the unit were determined using the mass flow meter as described in the previous section. To determine the flow rate out of the unit a calibration of the gas analyzer was required. NOVA 300 Infrared analyzer that had a range of 0-20% (volume basis) was used in this work. The calibration was carried out in the configuration used in the actual experiment. Two mass flow meters one calibrated for N2 and other for CO2 were used in the setup shown in Figure B.8. To calibrate the analyzer, N2 flow rate was fixed at a certain value and the CO2 flow rate was varied so that a gas mixture with C02 244 composition ranging from 0 to 20% (mole basis) was sent through the analyzer. The analyzer output signal was adjusted to zero by subtracting from it the output signal value at zero C02 concentration. At each setting the steady state reading of the analyzer output signal and the inlet gas mixture composition were recorded by the computer. The calibration curve obtained is plotted in Figure B.9 Throughout this work, before starting an experiment the calibration was checked by sending the gas through the by pass line to the C02 analyzer as shown in Figure 3.2. CO N2 Mass Flowmeter -4 CO, Analyzer 1 1 , •1 1 °°° 1 Mass ] W Nova 300 u lr-Data Acquisition Figure B.8 Setup for C02 Analyzer Calibration Gas Out 245 25 0 1 2 3 4 5 Output Signal after Zeroing (volt) Figure B.9 Calibration Curve for Gas-Phase C02 Measurement using Infrared Analyzer (Model NOVA-300) BA Gas Chromatograph In absorption experiments with N20, the exit gas composition was determined using a gas chromatograph (Shimadzu Model GC 8A) that was equipped with a thermal conductivity detector. The principal operating conditions are summarized in Table B.1. The GC was calibrated by injecting gas mixtures of known composition. The gas mixtures were prepared by mixing N20 and N2 gases in various proportions using mass flow meters. The standard gas samples were stored 1-liter Tedlar sampling bags. Each sample was injected at least three times and peak areas of N20, N2 and the total area were noted from the strip chart recorder. Average of the three values was used in generating the calibration curve. To compensate for error due to injection volume, the ratio of 246 N2O peak area to total peak area was plotted against the N20 mole fraction in the standard sample. The calibration data obtained are listed in Table B.2 and the calibration curve is shown in Figure B.10. The GC method described above was also used for measuring CO2 in the exit gas. This was done only for a few randomly selected runs to check the accuracy of the Infrared Analyzer. The column and operating conditions were identical to those used for N20 measurement (see Table B.1). The calibration data for CO2 are listed in Table B.3 and corresponding calibration curve is shown in Figure B.11. 0.20 0 0.18 CM z 1 0.16 o I 0.14 o ^ 0.12 0.10 -y y = 0.8932X - • R2 = 0.9988 - • Measurements • 1 1 1 1 1 1 1 1 — Regression 1 1 1 1 1 1 1 1 1 1 1 1 0.10 0.14 0.18 0.22 0.26 Peak Area Ratio (N20 Area/Total Area) Figure B.10 Calibration Curve for Gas Phase N20 measurement using GC (Shimadzu Model GC 8A, Column: Chromosorb 102) 247 0.5 r y = 0.8987x • Measurements 0.0 0.1 0.2 0.3 0.4 0.5 Peak Area Ratio (C02 Area/Total Area) Figure B.11 Calibration Curve for Gas-Phase C02 Measurement using GC (Shimadzu Model GC 8A, Column: Chromosorb 102) Table B.1 GC operating conditions for gas phase N20/C02 measurement Item Description Column Chromosorb 102, mesh size 80/100, 20'x1/8" SS packed column (supplied by Supelco Inc. Oakville, ON) Detector Thermal Conductivity Detector(TCD) Column temperature 40 °C Injector temperature 60 °C Carrier gas He (20 cm3/min) Injection volume 0.25 cm3 248 Table B.2 Calibration Data for N20 measurement using GC Ref. No. N20 in Std. Samples (mole fraction) N20 Peak Area Total Peak Area Peak Area Ratio (N20/Total) N2OCalib1 0.186 19715 94539 0.2085 19892 95505 0.2083 19580 94260 0.2077 N2OCalib2 0.177 18795 94158 0.1996 18946 95046 0.1993 18943 95251 0.1989 N2OCalib3 0.166 17346 93568 0.1854 17442 94098 0.1854 17411 94008 0.1852 N2OCalib4 0.155 16160 93307 0.1732 16140 93400 0.1728 16055 92836 0.1729 N20Calib5 0.141 14728 93316 0.1578 14758 93649 0.1576 14560 92791 0.1569 249 Table B.3 Calibration Data for C02 measurement using GC Ref. No. C02 in Std. Samples (mole fraction) C02 Peak Area Total Peak Area Peak Area Ratio (C02/Total) C02Calib1 0.401 45017 102002 0.4413 45859 103847 0.4416 45600 103212 0.4418 C02Calib2 0.350 39432 101738 0.3876 39647 102160 0.3881 39638 102043 0.3884 C02Calib3 0.300 34027 100098 0.3369 34241 101731 0.3366 34489 102486 0.3365 C02Calib4 0.2504 28213 100217 0.2815 28411 101089 0.2810 28431 101124 0.2811 C02Calib5 0.200 22343 99410 0.2248 22596 100552 0.2247 22566 100514 0.2245 C02Calib6 0.170 18889 98872 0.1910 18847 98828 0.1907 18776 98279 0.1910 C02Calib7 0.140 15792 98675 0.1600 15814 98786 0.1600 15763 98605 0.1598 C02Calib8 0.110 12151 97658 0.1244 12241 98647 0.1240 12201 98512 0.1239 250 Table B.3 Calibration Data for C02 measurement using GC (Contd.) C02Calib9 0.0801 8601 96902 0.0887 8670 98095 0.0884 8637 97649 0.0885 0.0651 6694 95439 0.0701 CO2Calib10 6785 97096 0.0698 6758 96683 0.0699 0.0447 4742 95268 0.04977 C02Calib11 4730 95063 0.04975 4745 95435 0.04972 C02Calib12 0.0324 3438 95411 0.03603 3453 96065 0.03594 3434 94917 0.03618 C02Calib13 0.0159 1698 95946 0.01769 1713 96050 0.01783 1692 95737 0.01767 251 APPENDIX C ANALYTICAL SOLUTION FOR PHYSICAL ABSORPTION/DESORPTION MODEL C. 1 Model for Physical Gas Absorption/Desorption The model equation for gas absorption/desorption over a hemispherical liquid film without chemical reaction is given by Initial Condition: at 9 = 0 forr>R C,=C° (C.2) Boundary Conditions: atr = Rfor0>O ^1 = 0 (C.3) atr = R + Ae for 0>O C^C1, (C.4) C.l.l Model Equations in Dimensionless form Equations (C.1) to (C.4) can be written in dimensionless form by substituting the following dimensionless variables (Wild and Potter, 1968): x = R + A°~r (C.5) A9 252 C = ^ and C°=-^ c; (C.6) dr) -V0A92 de = 47tR2D1 . . 5/3 . . -(sm5M 0)d0 3QA„ (C.7) and performing some algebraic manipulations (1"X >*T ^ (C.8) Initial Condition: at TI = 0 for x > 0 C = C° (C.9) Boundary Conditions: at x = 0 for T| > 0 C = 1 (C.10) atx = 1forr|>0 = 0 (C11) The differential equations (C.8) to (C.11) must be integrated from r\ = 0 (i.e. 6 = 0) to ri = r|2 (i.e. 6 = K/2). The upper limit of integration is defined by the following equations: "Ha 4TTR 2r\ it/2 A-rrR^n =± Jsin5/3ede= 1 (0.84133) (C.12) 3QA o 0 3QA 253 it/2 The fraction 0.84133 in the above equation is the value of Jsin5'3 0d9 and was 0 obtained by numerical integration. The total rate of absorption/desorption of a gas over a hemispherical film can then be computed from the following equation: 2 o v 3x y(x=ofl) dt| (C.13) Note that the concentration gradient 3C, is positive quantity for absorption and negative for desorption. The physical mass transfer coefficient can be obtained as follows: N° N° k°= ^ = =-2 (C.14) L 2TCR2(C;-C?) 2JIR2(C?-C1) C.2 Analytical solution In order to obtain an expression for physical mass transfer coefficient we need to solve the model equations (C.8) to (C.11) analytically. A quick analytical solution is possible if we assume that the penetration depth of the gas is much smaller compared to the thickness of the liquid film. Under this assumption, the liquid film can be treated as if it were infinitely deep and boundary condition at x = 1 ( Eq. C.11) can be replaced boundary conditions at x = ~ and the governing equations can be written as: fr=0 <c-15> 254 Initial Condition: atr| = 0 forx>0 C = C° (C.16) Boundary Conditions: atx = 0forr|>0 C = 1 (C.17) atx = oofom>0 C = C° (C.18) Equations (C.15) to (C.18) can be solved using Laplace transform. Taking the Laplace transform of Eq. (C.15) with respect to rj and denoting the transform of Cas C: sC-C(0,x) = -^ (C.19) From the initial condition Ci(0,x) = C°, which upon substitution in equation (C.19): dx Particular solution of Eq, (C.20): d2r -sC = -C° (C.20) C=-— (C.21) p s Homogenous solution of Eq. (C.20): C7(x,s) = A.e^ + A2e'rsx (C.22) 255 where Ai and A2 are the arbitrary constants. From boundary condition at x = 0, A^A2 = 1/s and from boundary condition at x = oo, Ai = 0. Therefore the general solution is: C(X,S)=* s s (C.23) or C(x,-n) = L -1 .-Vsx 91 s (C.24) From the inverse Laplace transform table: L-1 -Vsx = 1-erf < x ^ and L"1 v V . J 9L s = C° (C.25) Combining equations (C.24) and (C.25): C(x,T|) = 1-erf -C° (C.26) Recalling the definition of error function: erf(c;) = Aje^d<t) Vrc o (C.27) Applying the definition of error function from equation (C.27) to equation (C.26): C(X,TI) = 1—±= fe^-C0 V7C J0 (C.28) 256 c c° Substituting C = —1 and C° = into equation (C.28): c; c; 2C r 2 c1(x,1l) = c;—p- Je-* d«t>-C? Vrc o (C.29) Differentiating equation (C.29) with respect to x, obtaining y 3x y(x=o,n) and substituting it into equation (C.13): 2 o V^l (C.30) or -0_ frB_3Q(C'1-C?) ^ A _ _ 2 Vic (C.31) Substituting for r\2 from equation (C.12) into equation (C.31): N° =-ND° =3.1774, QR D., (c'i-c?) (C.32) Substituting forN° or from equation (C.32) into equation (C.14): k° =1.5887 D, ( A -D2 \ rt A0%R< Q v y (C.33) Equation (C.33) can be written in terms dimensionless numbers as follows: 0.5 o~0.5 Sh = 1.26758 Reu Sc (C.34) 257 where, ShJ^,Re = -5-.Sc^ (C.35) D, 2TIRV D, From Higbie's penetration theory the liquid phase mass transfer coefficient, k°, for physical absorption is given by k?=2 ^- (C.36) "c where, xc is the gas-liquid contact time and can be obtained by comparing equations (C.33) and (C.36): Tc =1.58487tR' (C.37) Equation (C.37) is similar to that given by Wild and Potter (1968) for complete sphere. 258 APPENDIX D DERIVATION OF RATE EXPRESSION FOR ZWITTERION MECHANISM This appendix gives the derivation for rate expression representing C02 reactions with a mixture of AMP and MEA in aqueous solution via zwitterion mechanism. The reactions involved are: AMP- Zwitterion Formation: k K C02+R4NH2<!4R4NH^COO- (D.1) AMP-Zwitterion Deprotonation: k K R4NH^COO" + R4NH2 R4NHCOO" + R4NH: (D.2) k ,K R4NH^COO" +H20^4R4NHCOO" +H30+ (D.3) k.3 k K R4NH^COO- + OH" ^R4NHCOCT + H20 (D.4) Deprotonation of AMP-Zwitterion to MEA: k26.K26 R^H^COO'+R^H, <-> R4NHCOO"+R1NH^ (D.5) k-26 Reactions (D.1) to (D.5) are same as reactions (4.1) to (4.4) and (4.25) respectively. The nomenclature used to denote chemical species and rate constants involved in these reactions is the same as in Chapter 4. The letter Z represents the zwitterion intermediate. 259 Rate of Accumulation of Z = Rate of formation of Z - Rate of consumption of Z — - k^Cj + k_2C4C3 + k_3C4C8 + k_4C4C9 + k_25C4C11 - ^ 10 dZ Assuming a steady state exist, then — = 0 and equation (D.6) can be solved for dt Z as follows: Z = k^C^C2 +k_2C4C3 +k_3C4C8 +k_4C4C9 +k_25C4C11 k_, -i-k2C2 -i-k3Cg H-k4Cy 4-k2gC (D.7) 10 Rate expression for reaction (D.1) is: *i (D.S) Substituting for Z from equation (D.7) into equation (D.8) gives: ri = -k^C^C2 + — k,C,C2 + k_2C4C3 + k_3C4Cg + k_4C4Cg +k_2gC4C^ k_, -t-k2C2 -i-k3Cg -f~k4Cy -i"k2gC 10 (D.9) or ri=-kiCA + k1C1C2 + k_2C4C3 + k_3C4Cg + k_4C4Cg +k_25C4C11 1 + P2 + ,x3 C9 + k-i + K25 vk_1y '10 (D.10) Let B = (k 2 k_"i P2 + 'k ^ A k-i C9 + K25 '10 (D.11) Combining equations (D.10) and (D.11) and rearranging gives: 260 -k, CjCgB — C^. k^ k^ 4 c9+- -25 '11 1 + B (D.12) or k2^ k.i v -1 j K,K; -C3 + K,K. -C8 + K1K4 c9 + *25 k 1 v "1 J '11 25 1 + B (D.13) fir \ Let A = fir ^ K1K: -C3 + Vk-V KiK--c8 + k-, -C9 + ^25 vk-1 y •C^ (D.14) 25 •k, C,C2 -C4 B 1 + 2 B (D.15) Since r, is derived from reactions (4.1) to (4.4) and (4.25), in Chapter 4 r, is denoted as r^^. When the reaction mixture has only AMP and water then k25 will be zero and r^^ will reduce to r,_4 261 APPENDIX E DENSITY AND VISCOSITY OF AQUEOUS AMINE SOLUTIONS E.1 Density The densities of binary and ternary aqueous amine mixtures were calculated using the correlation of Hsu and Li (1997a). According to this correlation the density of a liquid mixture, pm can be obtained from the following equation: where, Xj is the mole fraction of each component in the mixture, Mi is the molar mass of pure component i, and Vm is the molar volume of the liquid mixture. The molar volume of the liquid mixture is calculated from the following equation: where, v° is the molar volume of the pure fluid at the temperature of the system and VE is the excess volume of the liquid mixture which for a ternary system is assumed to be: (E.1) (E.2) VE = VE + VE + VE v — v12 -i- v13 T v23 (E.3) 262 The excess volume of the binary liquid mixture, v,| can be calculated using the following Redlich-Kister type equation: V1E2=x1x2^Ai(x1-x2)i (E.4) i=0 where, Aj are the binary interaction parameters which are assumed to have the following temperature dependence: A^a + bT + cT2 (E.5) The densities of pure fluids needed to calculate v° in eq. (E.2) were obtained from the correlation of the type: p = a, +a2T + a3T2 (E.6) The correlation coefficients a, b, c, a-\, a2 and a3 in eqs. (E.5) and (E.6) determined by Hsu and Li (1997a) are listed in Table E.1 and E.2. The parameter values are based on 686 data points measured in the temperature range of 303-353 K, including pure fluids, single-amine aqueous solutions, and ternary aqueous solutions of blended amines. The data include both their own density measurements and those reported in the literature. The overall average absolute percentage deviation for the density calculations was reported to be 0.041%. To check the accuracy of this correlation, we measured the density of various aqueous amine mixtures at 303 and 323 K. These values are listed in Table E.3 and a cross plot is presented in Figure E.1. The agreement between 263 our measurements and the estimates using eq. (E.1) is excellent. The relative error is within ± 1%. Table E.1 Binary Parameters of the Redlich-Kister Equation of the Excess Volume (eq. E.4 and E.5) Binary Pair Param. MEA+H2O DEA+H2O MDEA+H2O AMP+H2O A0 a -5.92024 x HT2 -3.31562 x 10 -2.88774x10 -6.51042 b -1.77290 x 10"4 8.11654 x 10"2 6.95810 x10"2 5.02584 x 10"3 c -1.10780 x 10'6 6.15156 x 10"6 -5.03040 x10'7 1.08578 x IO-6 Ai a 2.17490 -3.52516 x 10 -2.06623x10 5.55560 b 1.10385 x 10"5 9.75694 x 10"2 6.36707 x 10"2 -1.1325 x 10~2 c 0 0 0 0 Binary Pair MEA+MDEA MEA+AMP DEA+MDEA DEA+AMP Ac a -2.42756 x 10 5.53222 -1.24706x10 1.75649 x 103 b 1.89797 x 10"1 1.62914 x 10"1 1.00561 x 10"1 -1.06202x10 c -2.88250 x 10"4 -6.44380 xlO"4 -1.62790 xlO-4 1.59224 x 10"2 Ai a 0 0 0 0 b -1.05682 9.86571 x.10"1 9.63470 x 10"2 -1.28358 c 4.28233 x 10"3 -2.39399 x 10"3 1.02886 x IO-5 4.86136 x 10"3 A2 a 0 0 0 0 b -1.49472 x 10 -2.70341 x 10 -7.53195 x 10"3 -6.44203 x 10 c 1.52253 x 10"2 5.68765 x10"2 -3.65100 x10'3 2.35996 x 10"1 Based on data from: Al-Ghawas et al. (1989b), Xu et al. (1991), Xu et al. (1992), Li and Shen (1992), Rinker et al. (1994), Li and Lie (1994), Hagewiesche et al. (1995b), Hsu and Li (1997a). 264 Table E.2 Parameters of the Density Equation for Pure Fluids (eq. E.6) Pure Fluid ai a2 a3 H20 0.863559 1.21494 x IO'3 -2.57080 x 10"6 MEA 1.19093 -4.29990 x 10-4 -5.66040 x 10'7 DEA 1.20715 -1.51200 x 10-4 -7.66530 x 10'7 MDEA 1.22864 -5.44540 x 10-4 -3.35930 x10"7 AMP 1.15632 -6.76170 x 10^ -2.67580 x 10~7 Based on data from: Kell (1975), Perry and Chilton (1984), Al-Ghawas et al. (1989), Xu et al. (1991), Diguillo et al. (1992), Li and Shen (1992), Wang et al. (1994), Xu et al. (1992), Li and Lie (1994). 0.98 1.00 1.02 1.04 Calculated Density (g/cm3) Figure E.1 Measured and calculated densities at 303 and 323 K 265 Table E.3 Density of Aqueous Amine Blends at 303 K and 323 K (Total Amine = 25 wt%, Water = 75 wt%) Blend Density (g/crn^) T = 303 K T = 323K This From Error This From Error+ work eq. (E.1) (%) work eq. (E.1) (%) MEA + AMP 0 + 25 1.000 0.995 0.53 0.987 0.983 0.36 5 + 20 1.000 0.998 0.25 0.987 0.986 0.02 10 + 15 1.000 1.000 0.02 0.987 0.989 -0.22 15 + 10 1.000 1.002 -0.21 0.990 0.992 -0.18 20 + 5 1.005 1.005 0.05 0.990 0.994 -0.45 25 + 0 1.010 1.006 0.36 0.997 0.996 0.02 DEA + AMP 0 + 25 1.000 0.995 0.53 0.987 0.983 0.40 5 + 20 1.000 1.000 -0.05 1.000 0.990 1.03 10 + 15 1.010 1.006 0.35 1.000 0.996 0.36 15 + 10 1.015 1.012 0.25 1.000 1.003 -0.29 20 + 5 1.015 1.018 -0.34 1.010 1.009 0.08 25 + 0 1.020 1.025 -0.46 1.006 1.016 -0.97 MEA + MDEA 0 + 25 1.010 1.018 -0.80 1.006 1.008 -0.18 5 + 20 1.015 1.016 -0.07 0.997 1.005 -0.89 10 + 15 1.010 1.013 -0.32 0.997 1.003 -0.65 15 + 10 1.010 1.011 -0.09 0.997 1.001 -0.42 20 + 5 1.000 1.009 -0.88 0.997 0.999 -0.22 25 + 0 1.010 1.006 0.36 1.000 0.996 0.36 DEA + MDEA 0 + 25 1.010 1.018 -0.80 1.006 1.008 -0.18 5 + 20 1.020 1.019 0.08 1.010 1.009 0.08 10 + 15 1.016 1.020 -0.44 1.006 1.011 -0.47 15 + 10 1.020 1.022 -0.17 1.010 1.012 -0.23 20 + 5 1.020 1.023 -0.31 1.010 1.014 -0.39 25 + 0 1.020 1.025 -0.46 1.020 1.016 0.42 + Error = 100 x (measured-calculated)/measured 266 E.2 Viscosity The viscosities of binary and ternary aqueous amine mixtures were calculated using the correlation of Hsu and Li (1997b). According to this correlation the kinematic viscosity of a liquid mixture, vm, can be obtained from the following equation: n lnvm =5v + £xilnvi (E.7) i=1 where, Xj is the mole fraction of each component in the mixture, Vj is the kinematic viscosity (|Oj/pi) of pure fluid i, and 5v is the deviation in kinematic viscosity, which for a ternary system is assumed to be: 5v = 5vf2+5vf3+5vf3 (E.8) For a binary system, the 5vf2 is a function of temperature and mole fraction and can be calculated using the following Redlich-Kister type expression (Hsu and Li, 1997b): m 5v*2 =x1x2£Ai(x1-x2)i (E.8) i=0 where, Aj are the binary interaction parameters which are assumed to have the following temperature dependence: A:=a + —— (E.9) T + c 267 The viscosity of pure fluids required to calculate vi in eq. (E.7), is assumed to be the following expression: lnv = a1+—^— (E.10) T + a3 The correlation coefficients a, b, c, a^ a2 and a3 in eqs. (E.9) and (E.10) determined by Hsu and Li (1997b) are listed in Table E.4 and E.5. The parameter values are based on 499 data points measured in the temperature range of 303-353 K, including pure fluids, single-amine aqueous solutions, and ternary aqueous solutions of blended amines. The data include both their own viscosity measurements and those reported in the literature. The overall average absolute percentage deviation for the viscosity calculations was reported to be 1%. To check the accuracy of eq. (E.7), we measured the viscosity of aqueous amine mixtures at 303 K using a falling ball viscometer (Colepalmer Model P-08701-00). These measurements are listed in Table E.6 and a cross plot is presented in Figure E.2. The agreement between our measurements and the estimates from eq. (E.7) is within + 12%. 268 Table E.4 Binary Parameters of the Redlich-Kister Equation for the Viscosity Deviation (eqs. E.8 and E.9) Binary Pair Param. MEA+H20 DEA+H2O MDEA+H2O AMP+H2O Ao a 2.58323 x 10'1 2.76655 -6.26493 4.01239 b 5.05207 x 102 3.64795 x 103 1.59158 x103 2.49856 x 102 c -2.23155 x 102 6.78430x10 -1.79649 x 102 -2.65712 x102 Ai a -7.20106 1.71593 x 10 2.87926 -2.68462 b 2.30838 x 103 -4.75487 x 103 -4.03039 x103 0 c 0 0 0 0 Binary Pair MEA+MDEA MEA+AMP DEA+MDEA DEA+AMP Ao a 2.45414 x 10 -1.27691 x 102 -3.71143x10 -5.71403 b -7.79167 x 103 4.00392 x 104 7.70451 x 103 0 c 0 0 0 0 Ai a -1.56256 x 10 1.10232 x 102 -1.33407 x 10 -6.48408 x 10 b 0 0 0 0 c 0 0 0 0 Based on data from: Al-Ghawas et al. (1989), Xu et al. (1992), Rinker et al. (1994), Li and Lie (1994), Hagewiesche et al. (1995b), Song et al.(1996), Hsu and Li (1997b). 269 Table E.5 Parameters of the Viscosity Equation for Pure Fluids (eq. E.10) Pure Fluid ai a2 a3 H20 -3.28285 4.56029 x 102 -1.54576 x 102 MEA -3.51312 8.93173 x 102 -1.59612 x 102 DEA -4.99689 1.58400 x 103 -1.57449 x102 MDEA -4.06399 1.20196 x 103 -1.54419 x102 AMP -4.36785 9.96598 x 102 -1.92984 x 102 Based on data from: Yaws et al. (1976), Al-Ghawas et al. (1989), Diguillo et al. (1992), Xu et al. (1992), Li and Lie (1994), Song et al. (1996). 1.5 2.0 2.5 3.0 Calculated Viscosity (cp) Figure E.2 Measured and calculated viscosities at 303 K. 270 Table E.6 Viscosity of Aqueous Amine Blends at 303 K (Total Amine = 25 wt%, Water = 75 wt%) Blend Viscosity Viscosity Error+ (cp) (cp) (%) (This Work) (From eq. E.7) MEA + AMP 0 + 25 2.455 2.363 3.8 5 + 20 2.250 2.224 1.2 10 + 15 2.120 2.103 0.8 15 + 10 2.081 1.991 4.3 20 + 5 1.972 1.877 4.8 25 + 0 1.872 1.754 6.3 DEA + AMP 0 + 25 2.455 2.363 3.8 5 + 20 2.215 2.296 -3.7 10 + 15 2.372 2.228 6.1 15 + 10 2.155 2.162 -0.3 20 + 5 2.105 2.100 0.2 25 + 0 1.994 2.046 -2.6 MEA + MDEA 0 + 25 1.924 2.094 -8.9 5 + 20 1.927 2.014 -4.5 10 + 15 2.084 1.940 6.9 15 + 10 1.993 1.872 6.1 20 + 5 1.939 1.810 6.7 25 + 0 1.942 1.754 9.7 DEA + MDEA 0 + 25 1.924 2.094 -8.9 5 + 20 2.374 2.079 12.4 10 + 15 2.229 2.066 7.3 15 + 10 2.226 2.056 7.6 20 + 5 2.128 2.049 3.7 25 + 0 1.994 2.046 -2.6 + Error = 100 x (measured-calculated)/measured 271 APPENDIX F HENRY'S CONSTANT OF C02 AND N20 IN AQUEOUS AMINE F.1 N20 Analogy In order to predict the rate of absorption and desorption of C02 in aqueous amine solutions or analyze the rate measurements in terms of reaction kinetics, it is essential to estimate the physical solubility of C02 at various amine concentrations and temperatures. This physical solubility is calculated by multiplying the partial pressure of the gas above the solution with the inverse of the Henry's law constant. Since C02 reacts with aqueous amines, its physical solubility in these solutions cannot be determined by direct measurements and an indirect method based on the N20 analogy is commonly used. The analogy is based on the assumption that the ratio of the solubilities of N20 and C02 is the same in aqueous amine solutions as in water at the same temperature. In view of the similarities of N20 and C02 with regard to configuration, molecular volume and electronic structure, this assumption is considered reasonable and the physical solubility represented by Henry's constant can be estimated as follows: SOLUTIONS H H (F.1) 272 where Hcc,2 and HN20 denote the Henry's constants of C02 and N20 in aqueous amines, and H£02 and H^o denote Henry's constants of C02 and N20 in water. F.2 Henry's Constant of C02 and N20 in Water The data for Henry's constant of C02 in water and N20 in water from the present work and those reported in the literature are presented in Tables F.1 and F.2. These data were correlated as a function of temperature according to the following correlations: QCC7 VQO Log10(HCO2) = 69.39562 - ^ -22.29261log10(T)+0.003941096 T (F.2) 4373 35 Log10 (H°20) = 85.8485- ^- 27.716621 log10 (T) + 0.003397123 T (F.3) where T is in Kelvin and H°C02 and H°20 are in the units of kPa-m3/kmol. The correlation for C02 is valid over the temperature range from 273 to 523 K and the correlation for N20 is valid from 278 to 393 K. As shown in Figures F.1 and F.2, these correlations represent the solubility data reasonably well. The average absolute percent deviation (AAD%) for C02-H20 and N20-H20 is 2.32% and 4.1% respectively. Figures F.1 and F.2 also indicate that there is a good agreement between literature results and those of the present study particularly in the temperature range of 293-350 K where historically most of the data are reported. 273 F.3 Henry's Constant ofN20 and C02 in Aqueous Amine Solutions To estimate the Henry's constant of C02 in aqueous amine solutions at various amine concentrations and temperatures using N20 analogy (eq. F.1), corresponding data for N20 are needed. Therefore, an extensive literature survey was conducted and the data for the Henry's constant of N20 in aqueous solutions of MEA, DEA, MDEA, AMP and their mixtures (i.e., MEA+MDEA, MEA+AMP, DEA+MDEA and DEA+AMP) were complied. These data are listed in Tables F.3 to F.10. It can be seen from these tables that the published data are limited to absorber temperatures (i.e. 293-333 K) only. The data corresponding to stripper temperatures do not exist in the open literature and are usually obtained by extrapolating the low temperature data. However, in the present study it was found that these extrapolations might lead to inaccurate prediction of mass transfer rates as the literature correlations for H^0 and HNz0 proposed by Versteeg and van-Swaaij (1988), Li and Lai, (1995) and Li and Lee, (1996) over predict the Henry's constant value at stripper temperature by a factor of 3. Note that these correlations have been developed based on the solubility data up to 323 K. Consequently, in this work the solubility of N20 in aqueous solutions of MEA, DEA, MDEA, AMP, MEA+MDEA, MEA+AMP, DEA+MDEA and DEA+AMP were measured in the temperature range of 293-393 K and amine concentration of 10-30 wt%. These data are presented in Tables F.3 to F.10. The experimental apparatus and procedure used was identical to that described in Chapter 6 (see Figure 6.1). 274 To correlate the solubility of N2O in amine solutions, a semiempirical model proposed by Wang et al. (1992) was used. According to this method the Henry's constant of N2O in mixed solvent system can be obtained from the following equation: lnHN20=HE+XoilnH^0 (F.4) i=i where HN20 is the Henry's constant of N20 in the mixed solvent, HE is the excess Henry's constant, H'N20 is the Henry's constant of N20 in pure solvent i, and 4>j is the volume fraction of solvent i. The volume fraction is calculated as *.=^- (F-5) i=1 where Vj is the molar volume of pure solvent i, and Xj is the mole fraction of solvent i. Wang et al. (1992) measured the Henry's constants of N2O in pure amines such as MEA, DEA, MDEA and AMP in the temperature range of 293-355 K and correlated their data as HSne=aiexp (F.6) where ai and a2 are correlation coefficients and are listed in Table F.11. The densities of pure solvents were calculated according to the method described in Appendix E. The Henry's constant of N20 in pure water was calculated using eq. (F.3). The excess Henry's constant for binary system was correlated as: 275 (F.7) and that for ternary solvent systems was correlated as: Hyk =®&laii +0,OkaIk +OjCDkajk + 0>i<Dj<I>kaijk (F.9) where CCJJ, aik, oijk and 0% are the interaction parameters. For ternary systems (e.g. H20-MEA-MDEA), the parameters ay and ccik represent the water-amine interactions (e.g. H20-MEA and H20-MDEA), cqk represent the amine-amine interaction (e.g. MEA-MDEA) and ayk represent the amine-amine-water interaction (e.g. H20-MEA-MDEA). As in the work of Wang et al. (1992), ap was set to be constant and ay, aik, % were assumed to have the following temperature dependence: where is T in Kelvin and ai and a2 are constants. In this work, first the constants for binary systems (i.e. H20+MEA, H20+DEA, H20+MDEA and H20+AMP) were determined by regressing the data presented in Tables F.3 to F.6 and then using those values the constants for ternary systems (i.e. H20+MEA+MDEA, H20+MEA+AMP, H20+DEA+MDEA and H20+DEA+AMP) were estimated by regressing the data given in Tables F.6 to F.10. The regression was done using the software package GREG supplied by Stewart & Associate Engineering Software Inc. The estimates of the regression coefficients for both binary and ternary systems are presented in Table F.12. The comparisons of the calculated and experimental Henry's constants of N20 in binary and ternary aqueous amine a y =a1+a2/T (F.8) 276 solutions are shown in Figures F.3 and F.4 respectively. The estimates of the average absolute percent deviation (AAD%) for each system are reported in Table F.12. The results are satisfactory. The rather large deviation may be because the regression was performed using data from all sources. Also, the solubility data for mixed amine systems studied here are not plenty. The regression results indicate that the N20 solubility data for H20+DEA+MDEA and H20+DEA+AMP reported by Li and Lee (1996) do not correlate well with the experimental data obtained in this work and those reported by other sources as indicated by Figures 3 and 4. Overall, the N20 solubilities in the aqueous amine systems studied in this work are well correlated using the method of Wang et al. (1992). The correlations coefficients obtained in this study are valid over the temperature range of 293-393 K and the amine concentration of 10-30 wt%. With all the other quantities in eq. (F.1) estimated, the Henry's constant of C02 in unloaded aqueous amine solutions can be calculated. For the sake of brevity the results of these calculations are not presented here. 277 Table F.1 Henry's constant of C02 in water T (K) Hco2 (kPa m3/kmol) Reference T (K) Hco2 (kPa m3/kmol) Reference 293 2630 This work 318 4854 Versteeg & 304 3500 it 323 5155 van Swaaij (1988) 313 4234 if 333 6135 II 323 5179 ti 344 7143 II 333 6233 II 350 7576 II 343 7296 if 355 8333 II 353 8305 II 360 9259 II 363 9302 II 298 2889 Malinin (1974) 373 10056 II 308 3723 II 383 10828 II 323 5192 II 393 11225 II 348 7257 II 303 3420 Li & Lee (1996) 373 9829 II 308 3835 II 423 12485 II 313 4226 II 473 11913 II 303 3382 Li & Lai (1995) 523 9760 II 313 4227 it 383 9971 Takenouchi & 323 5136 II 423 12513 Kennedy (1964) 298 2993 Saha et al. (1993) 473 12652 II 293 2744 II 523 11984 n 313 4572 II 373 9895 Ellis & Golding 333 6476 II 423 13108 (1963) 353 8600 it 450 14362 it 303 3394 Al-Ghawas et al. 473 13496 it 313 4250 (1989) 523 12099 II 323 5167 II 273 1328 Perry et al. (1963) 291 2469 Versteeg & 278 1598 II 292 2410 van Swaaij (1988) 283 1897 II 292 2571 II 288 2225 II 293 2632 II 293 2590 II 298 2967 if 298 2991 II 298 3040 ii 303 3392 II 303 3571 II 308 3812 II 308 3937 ti 313 4249 it 311 4098 it 318 4687 it 313 4219 it 323 5161 II 313 4202 it 333 6219 II 278 Table F.2 Henry's constant of N20 in water T HN,O Reference T HN20 (kPa m3/kmol) Reference (K) (kPa m3/kmol) (K) 293 3701 This work 291 3344 Versteeg & 304 4903 II 292 3484 van Swaaij (1988) 313 5849 it 293 3333 II 323 7117 « 293 3425 n 333 8461 it 298 4132 II 343 9780 it 299 3774 ii 353 10911 it 303 4950 II 363 11845 II 308 5263 ii 373 12745 II 313 5917 II 383 13534 II 313 6061 it 393 14013 II 318 6993 II 303 4465 Li & Lee (1996) 323 7143 it 308 4813 it 323 7407 II 313 5822 II 340 10309 ti 303 4406 Li & Lai (1995) 353 12821 II 313 5725 it 359 14085 it 323 7264 II 298 4173 Haimour (1984) 298 4234 Browning (1994) 298 4132 Sada et al. (1977) 293 3694 Sandall et. al. 298 4154 Joosten (1972) 313 6339 (1993) 298 4171 Sada&Kito (1972) 333 9104 II 298 3906 Duda and Vrentas 353 11223 II 313 6211 (1968) 298 4176 Xu et al. 278 2134 Perry et al. (1963) 323 7254 (1991) 283 2572 II 348 12348 II 288 3029 II 293 3393 Al-Ghawas et al. 293 3617 it 298 3879 (1989) 298 4114 II 303 4321 it 303 4743 II 308 4703 it 308 5539 II 313 5009 ti 298 4212 Markham & Kobe (1941) 279 Table F.3 Henry's constant of N20 in aqueous MEA solutions T (K) CMEA (wt%) HN20 (kPa m3/kmol) Reference T (K) CMEA (wt%) HN20 (kPa m3/kmol) Reference 293 10.0 3944 This work 303 19.6 4920 Lttel et al. 303 10.0 5146 " 303 19.7 4901 (1992c) 313 10.0 6296 " 318 1.2 7223 II 323 10.0 7324 " 318 2.6 6976 II 333 10.0 8720 " 318 5.4 6957 II 353 10.0 11193 " 318 10.0 7013 II 373 10.0 12840 " 318 15.1 7088 u 393 10.0 13925 " 318 19.9 6921 u 293 20.0 4250 318 21.6 7184 u 303 20.0 4998 333 1.4 9107 u 313 20.0 6205 " 333 3.6 9137 u 323 20.0 7373 u 333 5.4 9018 It 333 20.0 8494 u 333 9.8 9077 a 353 20.0 10460 u 333 10.9 8817 u 373 20.0 12104 a 333 16.0 8817 u 393 20.0 12764 a 333 20.4 9167 u 293 30.0 4189 u 333 22.8 8264 u 303 30.0 5165 u 348 1.3 11001 u 313 30.0 6220 u 348 2.6 10445 u 323 30.0 7222 u 348 5.5 10755 u 333 30.0 8322 u 348 11.2 10676 a 353 30.0 10081 u 348 16.6 10715 a 373 30.0 11175 u 348 18.0 9774 u 393 30.0 12597 u 348 22.7 9807 u 303 30.0 4362 Li & Lai 348 24.3 9333 u 308 30.0 4696 (1995) 298 6.7 4132 Ladha et al. 313 30.0 5127 u 298 12.5 4202 (1981) 303 1.1 4564 Lttel et al. 303 2.4 4910 (1992c) 303 4.8 4780 it 303 9.5 4717 it 303 13.2 4863 II 280 Table F.4 Henry's constant of N20 in aqueous DEA solutions T (K) CDEA (wt%) HN20 (kPa m3/kmol) Reference T (K) CDEA (wt%) HN20 (kPa m3/kmol) Reference 293 10.0 3810 This work 303 2.6 5018 Littel et al. 303 10.0 4768 ii 303 3.2 4872 (1992c) 313 10.0 5772 II 303 6.3 5028 II 323 10.0 6982 II 303 6.3 5120 II 333 10.0 8272 II 303 8.3 4949 II 353 10.0 10971 II 303 9.0 5079 II 373 10.0 12352 II 303 12.6 5110 II 393 10.0 12984 II 303 17.9 5226 II 293 20.0 4108 n 303 26.6 5429 II 303 20.0 5111 ti 303 36.8 5699 n 313 20.0 6333 ti 318 2.8 7088 n 323 20.0 7267 u 318 4.4 7050 a 333 20.0 8291 a 318 7.1 7050 u 353 20.0 10743 u 318 8.6 7204 a 373 20.0 12006 u 318 14.3 7323 u 393 20.0 13681 a 318 15.9 7145 u 293 30.0 4064 u 318 25.2 7323 a 303 30.0 5128 u 318 34.6 7685 a 313 30.0 6824 u 333 4.3 8733 u 323 30.0 7975 u 333 8.9 8545 II 333 30.0 9169 u 333 17.8 8679 II 353 30.0 11156 u 333 25.4 8733 it 373 30.0 12875 u 333 33.2 8873 ii 393 30.0 14197 u 298 2.0 4235 Versteeg & 303 30.0 6890 Li & Lee 298 2.1 4206 Oyeaar 308 30.0 9613 (1996) 298 4.0 4286 (1989) 313 30.0 14377 u 298 5.5 4272 u 288 10.8 3293 Haimour 298 7.7 4362 u 288 21.3 3597 (1990) 298 11.2 4546 a 288 31.5 4205 u 298 15.4 4505 il 293 10.8 5016 a 298 29.0 4496 ii 293 21.3 3577 u 298 36.0 4764 il 293 31.6 3779 u 298 37.0 4848 n 298 10.8 4499 u 298 43.8 4985 u 298 21.4 5826 u 298 56.3 5151 a 298 31.6 3972 a 298 57.3 5140 u 303 10.8 4124 u 298 75.8 6132 u 303 21.4 5076 u 298 78.3 4719 a 303 31.7 7103 u 298 80.8 4424 u 298 80.8 4563 u 281 Table F.4 Henry's constant of N20 in aqueous DEA solutions Continued T (K) CDEA (wt%) ^N20 (kPa m3/kmol) Reference T (K) CDEA (wt%) HN20 (kPa m3/kmol) Reference 298 4.7 4164 Versteeg & 298 4.7 4164 Sada et al. 298 14.5 4393 van Swaaij 298 10.4 4213 (1977) 298 14.7 4432 (1988) 298 20.9 4385 it 298 16.1 4324 " 298 23.7 4456 it 298 20.9 4385 298 31.3 4631 it 298 22.9 4571 298 5.2 4149 Ladha et al. 298 23.7 4456 " 298 10.4 4202 (1981) 298 24.2 4666 " 298 20.6 4367 it 298 24.5 4480 298 31.3 4631 " 282 Table F.5 Henry's constant of N20 in aqueous MDEA solutions T (K) CMDEA (wt%) HN20 (kPa m3/kmol) Reference T (K) CMDEA (wt%) HN20 (kPa m3/kmol) Reference 293 10.0 3960 This work 298 30.0 4565 Al-Ghawas 303 10.0 5167 " 298 40.0 4840 et al. 313 10.0 6410 " 298 50.0 5229 (1989a) 323 10.0 7422 " 303 10.0 4483 " 333 10.0 8845 " 303 20.0 4892 " 353 10.0 11210 " 303 30.0 5101 •* 373 10.0 13025 " 303 40.0 5429 " 393 10.0 14039 " 303 50.0 5631 293 20.0 4110 " 308 10.0 4940 " 303 20.0 4965 " 308 20.0 5172 313 20.0 6410 " 308 30.0 5376 •* 323 20.0 7558 u 308 40.0 5668 333 20.0 8615 a 308 50.0 5970 " 353 20.0 10363 u 313 10.0 5205 " 373 20.0 11389 u 313 20.0 5630 a 393 20.0 12951 a 313 30.0 5898 u 293 30.0 4246 u 313 40.0 6206 u 303 30.0 5179 u 313 50.0 6426 u 313 30.0 6065 u 323 10.0 5563 u 323 30.0 6770 u 323 20.0 5841 u 333 30.0 8311 u 323 30.0 6074 u 353 30.0 10089 u 323 40.0 6220 u 373 30.0 11495 u 323 50.0 6370 u 393 30.0 12339 u 293 15.3 3663 Al-Ghawas 303 30.0 5169 Li & Lai 293 20.1 3800 et al. 308 30.0 5585 (1995) 293 30.2 4128 (1989b) 313 30.0 6188 u 298 15.3 4173 u 288 10.0 3103 Al-Ghawas 298 20.1 4308 u 288 20.0 3273 et al. 298 30.2 4623 u 288 30.0 3654 (1989a) 303 15.3 4638 u 288 40.0 4036 u 303 20.1 4769 u 288 50.0 4154 u 303 30.2 5070 it 293 10.0 3497 u 308 15.3 5039 a 293 20.0 3658 u 308 20.1 5167 a 293 30.0 4039 u 308 30.2 5453 a 293 40.0 4349 a 313 15.3 5363 u 293 50.0 4616 a 313 20.1 5490 it 298 10.0 3996 a 313 30.2 5762 u 298 20.0 4289 u 283 Table F.5 Henry's constant of N20 in aqueous MDEA solutions Continued T (K) CMDEA (wt%) (kPa m3/kmol) Reference T (K) CMDEA (wt%) ^N20 (kPa m3/kmol) Reference 293 4.1 3566 Versteeg & 318 6.0 6994 Versteeg & 293 4.9 3588 van Swaaij 318 11.3 6609 van Swaaij 293 8.5 3663 (1988) 318 11.9 7126 (1988) 293 9.4 3685 318 22.2 6994 " 293 11.8 3702 318 30.4 6976 " 293 13.0 3719 318 31.0 7364 " 293 16.3 3818 333 6.0 8873 " 293 17.7 3794 333 9.2 8624 293 21.9 3873 333 12.0 8931 " 293 30.6 4289 333 15.4 9167 298 7.3 4235 333 18.0 8545 " 298 10.0 4164 u 333 22.4 8733 " 298 14.6 4316 a 333 24.6 8571 298 14.8 4464 u 333 30.7 8733 n 298 19.3 4546 a ' 333 31.3 9077 " 298 19.5 4529 u 288 10.0 3207 Haimour 298 20.3 4480 M 288 15.0 3272 et al. 298 21.9 4596 a 288 20.0 3386 (1984) 298 22.2 4728 u 288 30.0 3773 u 298 28.2 4858 u 288 40.0 4162 u 298 29.3 4955 u 293 10.0 3662 tt 298 31.9 6057 a 293 15.0 3737 tt 308 7.4 5391 u 293 20.0 3830 tt 308 9.7 5414 u 293 30.0 4218 a 308 11.8 5495 u 293 40.0 4541 tt 308 14.8 5483 u 298 10.0 4273 u 308 15.6 5507 tt 298 15.0 4404 u 308 17.2 5542 tt 298 20.0 4561 u 308 17.8 5483 u 298 30.0 4846 u 308 19.7 5615 if 298 40.0 5126 u 308 21.0 5653 u 308 10.0 5530 u 308 21.8 5678 u 308 15.0 5602 tt 308 22.1 5628 u 308 20.0 5772 u 308 30.8 5941 u 308 30.0 5984 tt 318 5.8 6295 u 308 40.0 6285 u 284 Table F.6 Henry's constant of N20 in aqueous AMP solutions T (K) CAMP (wt%) HN2O (kPa m3/kmol) Reference T (K) CAMP (wt%) HN20 (kPa m3/kmol) Reference 293 10.0 3456 This work 288 4.5 3243 Saha et al. 302 10.0 4489 tt 293 4.5 3737 (1993) 313 10.0 5858 " 298 4.5 4216 tt 323 10.0 7293 n 303 4.5 4751 tt 334 10.0 8952 II 288 8.9 3413 ti 353 10.0 10796 II 293 8.9 3941 II 373 10.0 13391 298 8.9 4446 II 393 10.0 14886 it 303 9.0 5006 II 294 20.0 3857 II 288 13.3 3550 II 304 20.0 4931 » 293 13.4 4098 it 312 20.0 5936 it 298 13.4 4613 it 323 20.0 7230 u 303 13.4 5179 ti 333 20.0 8440 a 288 17.8 3717 II 353 20.0 10431 u 293 17.8 4293 II 373 20.0 12423 u 298 17.9 4826 II 393 20.0 13360 u 303 17.9 5405 it 294 30.0 5159 a 303 1.7 4863 it 303 30.0 6032 u 303 3.5 4910 u 313 30.0 6729 a 303 6.9 4988 u 323 30.0 7516 a 303 14.0 5038 u 333 30.0 8429 u 303 17.6 5237 a 343 30.0 9301 u 303 22.2 5383 u 353 30.0 10222 u 303 28.2 5686 u 363 30.0 10988 u 303 29.1 5712 u 373 30.0 11929 u 283 17.8 3213 Xu et al. 383 30.0 12973 u 288 17.8 3683 (1991b) 393 30.0 12861 u 298 17.9 4617 u 303 30.0 5856 Li and Lai 310 18.0 5986 u 308 30.0 6545 (1995) 283 26.6 3518 u 313 30.0 7531 u 288 26.7 4020 u 298 26.8 4842 u 310 27.0 6313 a 285 Table F.7 Henry's constant of N20 in MEA+MDEA+H20 T (K) CMEA (wt%) CMDEA (wt%) HN20 (kPa m3/kmol) Reference 293 12.5 12.5 3660 This work 303 12.5 12.5 4791 a 313 12.5 12.5 5764 a 323 12.5 12.5 6645 H 333 12.5 12.5 7899 ii 353 12.5 12.5 9606 ii 373 12.5 12.5 10802 li 393 12.5 12.5 11697 II 303 24.0 6.0 4481 Li & Lai (1995) 303 18.0 12.0 4580 II 303 12.0 18.0 4696 II 303 6.0 24.0 4868 II 308 24.0 6.0 4860 II 308 18.0 12.0 4957 it 308 12.0 18.0 5156 u 308 6.0 24.0 5392 a 313 24.0 6.0 5277 u 313 18.0 12.0 5396 tt 313 12.0 18.0 5624 tt 313 6.0 24.0 5935 tt 303 1.5 28.5 5140 Hagewiesche et al. 303 3.0 27.0 5270 (1995b) 303 4.5 25.5 5360 u 303 2.0 38.0 5580 a 303 4.0 36.0 5670 u 303 6.0 34.0 5850 a 313 1.5 28.5 6100 u 313 3.0 27.0 6290 u 313 4.5 25.5 6480 u 313 2.0 38.0 6270 u 313 4.0 36.0 6340 u 313 6.0 34.0 6440 u 323 1.5 28.5 6350 u 323 3.0 27.0 6550 u 323 4.5 25.5 6590 tt 323 2.0 38.0 6290 it 323 4.0 36.0 6320 It 323 6.0 34.0 6430 it 298 10.0 40.0 5680 Browning & 298 20.0 30.0 5465 Weiland (1994) Table F.8 Henry's constant of N20 in MEA+AMP+H20 T CMEA CAMP HN20 (kPa m3/kmol) Reference (K) (wt%) (wt%) 293 12.5 12.5 3919 This work 303 12.5 12.5 4920 a 313 12.5 12.5 5977 u 323 12.5 12.5 7054 if 333 12.5 12.5 8114 II 353 12.5 12.5 10049 II 373 12.5 12.5 11570 ll 393 12.5 12.5 12565 II 303 24.0 6.0 4548 Li & Lai (1995) 303 18.0 12.0 4911 II 303 12.0 18.0 5415 II 303 6.0 24.0 5597 II 308 24.0 6.0 5000 II 308 18.0 12.0 5656 it 308 12.0 18.0 5961 tt 308 6.0 24.0 6304 u 313 24.0 6.0 5545 u 313 18.0 12.0 6131 u 313 12.0 18.0 6813 u 313 6.0 24.0 7145 tt 287 Table F.9 Henry's constant of N20 in DEA+MDEA+H20 T (K) CDEA (wt%) CMDEA (wt%) HN20 (kPa m3/kmol) Reference 293 12.5 12.5 5106 This work 303 12.5 12.5 6313 a 313 12.5 12.5 7879 u 323 12.5 12.5 9140 II 333 12.5 12.5 10774 II 353 12.5 12.5 12927 it 373 12.5 12.5 14670 M 393 12.5 12.5 15264 II 303 24.0 6.0 6586 Li & Lee (1996) 303 18.0 12.0 6081 II 303 12.0 18.0 5938 it 303 6.0 24.0 5422 it 308 24.0 6.0 8480 it 308 18.0 12.0 7663 u 308 12.0 18.0 6797 tt 308 6.0 24.0 6090 it 313 24.0 6.0 11537 if 313 18.0 12.0 9632 tf 313 12.0 18.0 7979 u 313 6.0 24.0 6887 u 298 10.0 40.0 5680 Browning & 298 20.0 30.0 5465 Weiland (1994) 288 Table F.10 Henry's constant of N20 in DEA+AMP+H20 T CDEA CAMP HN2O (kPa m3/kmol) Reference (K) (wt%) (wt%) 293 12.5 12.5 6048 This work 303 12.5 12.5 7450 u 313 12.5 12.5 8654 a 323 12.5 12.5 10008 II 333 12.5 12.5 11334 11 353 12.5 12.5 13970 II 373 12.5 12.5 15917 it 393 12.5 12.5 16543 II 303 24.0 6.0 6579 Li & Lee (1996) 303 18.0 12.0 6293 II 303 12.0 18.0 6164 II 303 6.0 24.0 6037 II 308 24.0 6.0 8580 II 308 18.0 12.0 7970 u 308 12.0 18.0 7439 u 308 6.0 24.0 7007 u 313 24.0 6.0 11520 u 313 18.0 12.0 10174 u 313 12.0 18.0 9054 u 313 6.0 24.0 8194 u 289 Table F.11: Parameters in eq. (F.6) for Henry's constant of N20 in pure amines (Wang et al., 1992) System ai a2 No. of data AAD%+ MEA 1.207 x 105 -1136.5 6 1.39 DEA 1.638 x 105 -1174.6 5 2.43 MDEA 1.524 x 105 -1312.7 5 0.90 AMP 8.648 x 104 -1205.2 6 1.55 +AAD% = average absolute percent deviation Table F.12: Parameters in excess Henry's constant of N20 in binary and ternary solvent systems (this work) System ai a2 0C123 No. of data AAD%+ MEA+H20 -2.14841 1034.851 59 3.95 DEA+H20 -2.697683 1263.981 97 7.73 MDEA+H20 -2.72350 1342.437 147 5.64 AMP+H20 -1.367139 1089.998 62 3.88 M EA+M D EA+ H20 33.68282 -1436.122 -47.71418 40 5.18 MEA+AMP+H20 93.32045 -5199.785 -110.8438 20 4.53 DEA+MDEA+H20 -27.38076 2762.749 42.73698 22 7.63 DEA+AMP+H20 -50.12477 4131.152 72.02965 20 7.08 +AAD% = average absolute percent deviation 290 16000 — 12000 F o E CD D_ x • CM O O X 8000 F 4000 0 • This work o Literature data (Table F.1) — Correlation (Eq. F.2) _i i_ _J I L i I i i i i 250 300 350 400 450 500 550 Temperature (K) Figure F.1 Henry's constant of C02 in water 16000 — 12000 o E 2 8000 o CM X I i 4000 F o • This work o Literature data (Table F.2) — Correlation (Eq. F.3) -i i i_ j i_ i i 250 275 300 325 350 Temperature (K) 375 400 Figure F.2 Henry's constant of N20 in water 291 0 4000 8000 12000 16000 Measured HN2C-AM (kPa nrvVkmol) Figure F.3 Measured and calculated values of Henry's constant of N20 in aqueous amine solutions between 288 to 393 K. 0 4000 8000 12000 16000 Measured HN20.AM (kPa m3/kmol) Figure F.4 Measured and calculated values of Henry's constant of N20 in aqueous amine blends between 288 to 393 K. 292 APPENDIX G DIFFUSIVITY OF C02 AND N20 IN AQUEOUS AMINE SOLUTIONS G.3 N2Q Analogy Since CO2 reacts with aqueous amines, its diffusivity in these solutions cannot be measured by direct absorption experiments and an indirect method based on N20-Anology is frequently used. This analogy is based on the assumption that under identical conditions, the ratios of the diffusivities of C02 and N20 in water and aqueous amine solutions are equal: where DC02 and DN20 denote the diffusivity of C02 and N20 in amine solution, and Dco2 and D£20 denote the diffusivity of C02 and N20 in water. In order to calculate Dcc.2 using eq. (G.1), the values of D°C02, D£ 0 and DN20 are required which were calculated from the correlations given below. G.2 Diffusivity of C02 and N20 in Water Diffusivities of C02 and N20 in water were calculated using the following equations: D (G.1) D D° L-'M (G.2) 293 5.2457 x10"6 exp - 2388.9 (G.3) where T is in Kelvin and D°CCl2 and D^o are in the units of m2/s. These correlations were developed from the data acquired in this work and those reported in the literature. These data are tabulated in Tables G.1 and G.2; also, they are plotted on Figures G.1 and G.2. Evidently, the agreement between the measured and predicted values is good. The average absolute percent deviation (AAD%) for CO2-H2O system is 4.1% and for N20-H20 system is 10.9%. The rather large deviation for N20-H20 system is because the data reported by Al-Ghawas et al. (1989a, 1989b) are about 10-70% lower then those measured in this work and those reported in other sources (see Table G.2). In general, the diffusion coefficients measured in this work are in good agreement with the literature values. G.3 Diffusivity of N20 and Aqueous Amine Solutions Historically, the diffusivity of N2O in aqueous amine solutions at various temperatures and amine concentrations is calculated using the following modified Stokes-Einstein relation: In order to check the above relationship, N20 diffusivity data for MEA, DEA, MDEA, AMP and their mixtures (MEA+MDEA, MEA+AMP, DEA+MDEA & DEA+AMP) were compiled from various literature sources. These data along with those obtained in this work are tabulated in Tables G.3 to G.10. The NjOM-sol — '-'N2OM'H20 (G.4) 294 corresponding Stokes-Einstein plots for single and mixed amine systems are presented in Figures G.3 and G.4 respectively. It can be seen from these plots that the Stokes-Einstein type relationship does not represent the experimental data well. The average deviation (AAD%) for single amine systems is about 20% and that for mixed amine systems is about 30%. Therefore, in this work new correlations were developed. The data for single amine systems were correlated as: DNz0 =(5.2457x10"6 +Ai1CAM +Ai2CAM)exp| 2388.9 (G.5) and the data for mixed amine systems were correlated as: DN2o = '5.2457X10-6 + AL1CAM, + A12CAM, + A21CAM2 ^ + A22CAM2 +B12CAM'CAM2 J exp 2388.9 (G.6) where T is in Kelvin, CAM is in kmol/m3 and DN2Q are in the units of m2/s. The interaction parameters An, Ai2, and Bi2 were assumed to have the following temperature dependence: Aj^Ajj.B^ -a, (G.7) where T is in Kelvin and ai and a2 are constants. In this work, first the constants for single amine systems were determined by regressing the data given in Tables G.3 to G.6 and then using these results, the constants for mixed amine systems were estimated by regressing the data given in Tables G.7 to G.10. The regression was done using the software 295 package GREG supplied by Stewart & Associate Engineering Software Inc. The estimates of the regression coefficients for single and mixed amine systems are presented in Table G.11 and Table G.12. A comparison of the measured and calculated diffusion coefficients is shown in Figures G.5 and G.6. In general, the agreement is satisfactory. The overall AAD% for single amine systems is about 13% and that for mixed amine systems is about 9%. In light of the fact that different sources have used a different absorption apparatus to measure diffusivity, this much deviation is expected and the correlations can be safely used. Note that for CAM = 0, these correlations (eqs. G.5 and G.6) reduce to the correlation for N20-H20 system (eq. G.3). This unique and very important feature does not exist in other similar correlations proposed in the literature (Li and Lai, 1995 and Li and Lee, 1996). G.4 Diffusivity of C02 and Aqueous Amine Solutions Once the diffusivity of C02 in water, N20 in water and N20 in aqueous amines are known, the diffusivity C02 in aqueous amine solution was calculated using N20-analogy given by eq. (G.1). G.5 Diffusivity of the Alkanolamines in Aqueous Solutions The diffusion coefficients of various amines in their aqueous solutions were calculated using the correlation developed by Glasscock (1990): DAM=2.5x10-10 Mw 298 ] u^ln y (G.8) 296 The density of pure amine at 298 K (pA^ ) was calculated using eq. (E.6) and the viscosities of pure water (|iH20) and aqueous amine solutions (u.^,) were calculated from eqs. (E.10) and (E.7) respectively. The diffusion coefficients of all other ionic species were set equal to DAM-297 10.0 _C/3 CM E x O CM o o o Q 1.0 b 0.1 o Data from Table G.1 — Eq. (G.2) _l I L J I I I 1 I 1_ 1 2 3 4 I 1000/T (1/K) Figure G.1 Diffusivity of C02 in water as a function of temperature 10.0 .CO CM E x a> O O CM z o Q 1.0 0.1 o Data from Table G.2 — Eq. (G.3) -i i i i • • • i i i i_ 1 2 3 4 5 1000/T (1/K) Figure G.2 Diffusivity of N20 in water as a function of temperature. 298 7.0 6.0 5.0 J4.0 CN O Q z 3.0 2.0 1.0 0.0 o MEA+H20 •DEA+H20 x MDEA+H20 A AMP+H20 x X X X X X A X ^x xxxxx x x 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 (Viscosity of Amine Sol.A/iscosity of Water)0'8 Figure G.3 Stokes-Einstein plot for diffusivity of N20 in aqueous alkanolamine solutions. 8.0 6.0 o CM z Q ^5 4.0 CNJ z o Q 2.0 0.0 ; oMEA+MDEA+H20 : x MEA+AMP+H20 : • DEA+MDEA+H20 cP • ' a ; A DEA+AMP+H20 CO • n > • LJ / S< ' 1 1 ! : l 1 ' 1 1 ' ' 1 ' 'if \ 1 I 1 I ! 1 1 ...i i 1 ,1111 0.0 2.0 4.0 6.0 8.0 (Viscosity of Amine Sol.A/iscosity of Water)0'8 Figure G.4 Stokes-Einstein plot for diffusivity of N20 in aqueous alkanolamine blends. 299 Figure G.5 Measured and calculated diffusivities of N20 in aqueous amine solutions. 3.0 u ra O CO E x cn O O CN z a •a cu 2.0 IS 1.0 0.0 : o MEA+MDEA+H20 ; x MEA+AMP+H20 : • DEA+MDEA+H20 O >r [ A DEA+AMP+H20 DWK ° I I I ! I 1 ! 1 i ! 1 i 1 1 • 0.0 0.5 1.0 1.5 2.0 Measured DN2o (109xm2/s) 2.5 3.0 Figure G.6 Measured and calculated diffusivities of N20 in aqueous amine blends. 300 Table G.1 Diffusivity of C02 in water T 109xD°CO2 (m2/s) Reference T 109xD°CO2 (m2/s) Reference (K) (K) 298 1.75 This work 292 1.65 Versteeg & 308 2.12 ti 293 1.68 van Swaaij (1988)+ 318 3.11 ii 293 1.64 II 328 3.65 it 293 1.60 II 298 2.03 Rowley et al. (1998) 293 1.77 II 303 2.23 Li & Lee (1996) 298 1.98 tt 308 2.51 u 298 1.87 II 313 2.80 it 298 1.95 II 303 2.12 Li & Lai (1995) 298 2.05 II 308 2.46 it 298 1.85 ti 313 2.78 u 298 2.00 tt 293 1.76 Tamimi et al. (1994a) 298 1.94 tt 298 1.94 u 298 1.87 n 303 2.20 u 298 1.90 it 313 2.93 tt 298 1.74 tl 333 4.38 II 303 2.29 u 353 6.58 u 303 2.15 II 368 8.20 u 308 2.41 II 298 1.93 Xu et al. (1991b) 308 2.18 it 273 0.96 Versteeg & 313 2.80 tt 280 1.15 van Swaaij (1988)+ 318 3.03 fl 283 1.46 n 325 3.61 u 288 1.60 328 3.68 if 288 1.39 338 4.40 ft 289 1.57 II 338 4.30 It 291 1.71 •i 348 5.40 II +Versteeg & van Swaaij (1988) includes data from 30 different sources. 301 Table G.2 Diffusivity of N20 in water T (K) 109xD°2O (m2/s) Reference T (K) 109xD°2O (m2/s) Reference 297 1.59 This work 288 1.39 Versteeg & 303 2.19 » 290 1.70 van Swaaij (1988)+ 313 2.61 ft 291 1.47 II 323 3.00 u 292 1.56 II 303 2.11 Li & Lee (1996) 293 1.48 ii 308 2.34 n 293 1.52 ft 313 2.70 u 293 1.92 ft 303 2.01 Li & Lai (1995) 293 1.74 ft 308 2.30 u 293 1.45 " 313 2.65 « 293 1.65 293 1.84 Tamimi et al. (1994a) 298 2.09 " 298 1.88 tt 298 1.86 303 1.93 tt 298 1.69 313 2.61 tt 298 1.92 333 4.51 ti 298 1.78 353 6.50 tf 298 1.88 368 7.30 u 298 1.80 294 1.59 Saha etal. (1993) 303 2.27 302 1.95 tf 304 2.35 " 312 2.51 308 2.03 " 318 2.93 ti 308 2.34 u 288 1.29 Al-Ghawas et al. 313 2.35 u 293 1.44 (1989a) 313 2.55 u 298 1.57 tt 313 2.58 II 303 1.61 u 318 3.17 II 308 1.63 u 340 5.33 u 313 1.68 u 343 5.43 u 323 1.87 u 353 6.32 a 293 1.40 Al-Ghawas et al. 298 1.49 (1989b) 303 1.58 308 1.66 313 1.73 u "Versteeg & van Swaaij (1988) includes data from 30 different sources. 302 Table G.3 Diffusivity of N20 in aqueous MEA solutions T (K) CMEA (wt%) 109xDN2O (m2/s) Reference 303 25.0 1.70 This work 313 25.0 2.18 u 323 25.0 2.74 a 303 30.0 1.56 Li and Lai (1995) 308 30.0 1.73 313 30.0 1.88 u 298 4.5 1.74 Sada et al. (1978) 298 8.3 1.63 a 298 12.3 1.46 u 298 16.8 1.24 a 298 20.4 1.15 tt 303 Table G.4 Diffusivity of N2Q in aqueous DEA solutions T (K) CDEA (wt%) 109xDN2O (m2/s) Reference 303 25.0 1.32 This work 313 25.0 1.66 u 323 25.0 2.10 u 303 30.0 1.16 Li & Lee (1996) 308 30.0 1.32 » 313 30.0 1.50 u 293 10.0 1.23 Tamimi et al. (1994b) 298 10.0 1.31 M 303 10.0 1.51 a 313 10.0 1.71 u 333 10.0 2.91 u 353 10.0 3.95 a 368 10.0 4.23 u 293 20.0 0.96 u 298 20.0 0.98 u 303 20.0 1.19 u 313 20.0 1.57 u 333 20.0 2.40 u 353 20.0 3.48 u 368 20.0 3.73 u 293 30.0 0.57 u 298 30.0 0.69 u 303 30.0 0.78 a 313 30.0 1.17 u 333 30.0 1.88 u 353 30.0 2.88 u 368 30.0 3.44 u 298 3.1 1.66 Sada et al. (1978) 298 6.0 1.56 u 298 11.6 1.40 u 298 16.9 1.26 u 298 20.7 1.04 u 304 Table G.5 Diffusivity of N20 in aqueous MDEA solutions T CMDEA 1C 9 x DN2o Reference T CMDEA 109xDN2O (m2/s) Reference (K) (wt%) (m2/s) (K) (wt%) 303 25.0 1.23 This work 293 50.0 0.12 Tamimi et al. 313 25.0 1.54 it 298 50.0 0.19 (1994b) 323 25.0 1.95 u 303 50.0 0.23 ft 303 30.0 1.08 Li & Lai (1995) 313 50.0 0.41 it 308 30.0 1.18 333 50.0 0.97 it 313 30.0 1.29 « 353 50.0 1.23 ii 303 30.0 1.03 Hagewiesche 368 50.0 1.56 tf 303 40.0 0.82 et al. (1995b) 288 10.0 1.05 Al-Ghawas et al. 313 30.0 1.24 cc 293 1.0.0 1.16 (1989) 313 40.0 1.11 cc 298 10.0 1.33 it 323 30.0 1.42 11 303 10.0 1.43 tt 323 40.0 1.27 (1 308 10.0 1.48 Cf 293 10.0 1.36 Tamimi et al. 313 10.0 1.61 Cf 298 10.0 1.41 (1994b) 323 10.0 1.76 Cl 303 10.0 1.60 u 288 20.0 0.84 ft 313 10.0 1.93 u 293 20.0 0.99 Cl 333 10.0 3.20 it 298 20.0 1.15 it 353 10.0 4.13 u 303 20.0 1.29 tl 368 10.0 4.86 ti 308 20.0 1.41 CI 293 20.0 1.06 u 313 20.0 1.50 Cf 298 20.0 1.11 u 323 20.0 1.67 CC 303 20.0 1.25 u 288 30.0 0.71 Cl 313 20.0 1.67 Cl 293 30.0 0.82 ti 333 20.0 2.71 11 298 30.0 0.97 U 353 20.0 3.93 CI 303 30.0 1.03 it 368 20.0 4.29 11 308 30.0 1.16 tt 293 30.0 0.61 Cl 313 30.0 1.24 it 298 30.0 0.72 Cl 318 30.0 1.30 it 303 30.0 0.81 tl 323 30.0 1.42 tt 313 30.0 1.27 tl 288 40.0 0.59 it 333 30.0 2.02 It 293 40.0 0.64 CC 353 30.0 3.02 II 298 40.0 0.75 CC 368 30.0 3.71 ii 303 40.0 0.82 Cf 293 40.0 0.26 ft 308 40.0 0.99 CC 298 40.0 0.38 Cf 313 40.0 1.11 tf 303 40.0 0.49 u 318 40.0 1.17 Cf 313 40.0 0.80 if 323 40.0 1.27 ft 333 40.0 1.77 It 288 50.0 0.33 tt 353 40.0 2.58 tt 293 50.0 0.43 tf 368 40.0 2.97 tf 298 50.0 0.51 ft 305 Table G.5 Diffusivity of N20 in aqueous MDEA solutions Continued T CMDEA 1C 9 x DN2o Reference T CMDEA 109xDN2O (m2/s) Reference (K) (wt%) (m2/s) (K) (wt%) 303 50.0 0.56 Al-Ghawas et al. 298 11.7 1.64 Versteeg et al. 308 50.0 0.64 (1989) 298 12.5 0.95 (1988) 313 50.0 0.75 u 298 13.3 0.77 u 318 50.0 0.84 it 308 3.9 1.98 u 323 50.0 0.93 u 308 5.0 1.87 u 293 2.1 1.46 Versteeg et al. 308 7.6 1.64 a 293 2.6 1.32 (1988) 308 8.1 1.64 u 293 4.4 1.22 u 308 8.9 1.54 u 293 4.8 1.16 u 308 10.2 1.43 u 293 6.7 1.26 u 308 10.8 1.38 u 293 8.4 0.81 a 308 11.3 1.31 it 298 1.5 1.88 u 318 3.0 2.46 a 298 2.5 1.32 u 318 5.8 2.18 u 298 3.1 1.44 u 318 9.1 2.06 u 298 4.3 1.36 u 318 12.2 1.68 a 298 5.9 1.32 a 318 15.8 1.47 a 298 6.5 1.32 u 318 17.5 1.10 u 298 6.5 1.25 u 333 4.7 3.34 u 298 7.2 1.16 u 333 8.0 3.13 a 298 8.8 1.02 u 333 12.7 2.37 u 298 9.7 1.06 u 333 15.9 2.09 u 306 Table G.6 Diffusivity of N20 in aqueous AMP solutions T (K) CAMP (wt%) 109xDNaO (m2/s) Reference T (K) CAMP (wt%) 109xDN2O (m2/s) Reference 303 25.0 1.13 This work 298 12.2 0.81 Xu et al. 313 25.0 1.71 u 300 12.3 0.91 (1991b) 323 25.0 2.25 u 301 12.3 0.97 u 303 30.0 1.05 Li & Lai (1995) 304 12.3 1.02 it 308 30.0 1.16 u 313 12.3 1.47 it 313 30.0 1.30 a 323 12.4 2.09 u 294 3.1 1.57 Saha et al. 334 12.5 2.82 ft 294 6.1 1.30 (1993) 344 12.5 3.88 If 294 9.2 1.09 u 348 12.6 4.44 If 294 12.2 0.88 a 294 18.3 0.59 ff 302 3.1 1.83 u 296 18.4 0.62 If 302 6.1 1.55 u 308 18.5 1.06 ff 302 9.2 1.30 (( 319 18.6 1.57 ff 302 12.3 1.10 tl 328 18.7 2.27 ff 312 3.1 2.30 tl 338 18.8 3.00 f 312 6.2 1.98 a 348 18.9 4.07 ff 312 9.2 1.68 a 349 18.9 4.17 u 312 12.3 1.45 u 318 3.1 2.70 u 318 6.2 2.29 u 318 9.3 2.01 u 318 12.4 1.67 it 307 Table G.7 Diffusivity of N20 in MEA+MDEA+H20 T CMEA CMDEA 109xDN2O (m2/s) Reference (K) (wt%) (wt%) 303 12.5 12.5 1.31 This work 313 12.5 12.5 1.71 u 323 12.5 12.5 2.12 a 303 24.0 6.0 1.49 Li and Lai (1995) 303 18.0 12.0 1.40 u 303 12.0 18.0 1.34 u 303 6.0 24.0 1.24 u 308 24.0 6.0 1.67 u 308 18.0 12.0 1.55 u 308 12.0 18.0 1.46 u 308 6.0 24.0 1.35 u 313 24.0 6.0 1.81 u 313 18.0 12.0 1.68 u 313 12.0 18.0 1.57 u 313 6.0 24.0 1.47 u 303 1.5 28.5 1.14 Hagewiesche et al. 303 3.0 27.0 1.27 (1995) 303 4.5 25.5 1.39 u 303 2.0 38.0 0.84 u 303 4.0 36.0 0.91 tf 303 6.0 34.0 0.97 a 313 1.5 28.5 1.27 u 313 3.0 27.0 1.35 u 313 4.5 25.5 1.45 u 313 2.0 38.0 1.16 u 313 4.0 36.0 1.21 u 313 6.0 34.0 1.29 u 323 1.5 • 28.5 1.49 u 323 3.0 27.0 1.56 u 323 4.5 25.5 1.66 u 323 2.0 38.0 1.31 u 323 4.0 36.0 1.39 u 323 6.0 34.0 1.47 u 308 Table G.8 Diffusivity of N20 in MEA+AMP+H20 T CMEA CAMP 109xDNzO (m2/s) Reference (K) (wt%) (wt%) 303 12.5 12.5 1.27 This work 313 12.5 12.5 1.57 u 323 12.5 12.5 2.12 u 303 24.0 6.0 1.51 Li and Lai (1995) 303 18.0 12.0 1.42 u 303 12.0 18.0 1.32 a 303 6.0 24.0 1.21 u 308 24.0 6.0 1.63 tt 308 18.0 12.0 1.57 a 308 12.0 18.0 1.46 u 308 6.0 24.0 1.35 u 313 24.0 6.0 1.80 u 313 18.0 12.0 1.73 u 313 12.0 18.0 1.60 tt 313 6.0 24.0 1.49 u 309 Table G.9 Diffusivity of N20 in DEA+MDEA+H20 T CDEA CMDEA 109xDN2O (m2/s) Reference (K) (wt%) (wt%) 303 12.5 12.5 1.26 This work 313 12.5 12.5 1.66 u 323 12.5 12.5 1.93 u 303 24.0 6.0 1.15 Li and Lee (1996) 303 18.0 12.0 1.14 u 303 12.0 18.0 1.12 u 303 6.0 24.0 1.11 u 308 24.0 6.0 1.30 if 308 18.0 12.0 1.27 tf 308 12.0 18.0 1.25 u 308 6.0 24.0 1.22 tf 313 24.0 6.0 1.45 if 313 18.0 12.0 1.41 tf 313 12.0 18.0 1.38 u 313 6.0 24.0 1.34 tt 293 2.1 47.9 0.28 Rinker et al. (1995b) 313 2.1 47.9 0.42 u 333 2.1 47.9 0.82 ti 353 2.1 47.9 1.30 ti 293 9.0 41.0 0.28 ft 313 9.0 41.0 0.42 u 333 9.0 41.0 0.92 u 353 9.0 41.0 1.10 ti 293 15.3 34.7 0.30 ti 313 15.3 34.7 0.47 ft 333 15.3 34.7 0.77 ti 353 15.3 34.7 1.06 fi 293 18.5 31.5 0.24 u 313 18.5 31.5 0.45 ti 333 18.5 31.5 0.97 ti 353 18.5 31.5 1.03 tf 310 Table G.10 Diffusivity of N20 in DEA+AMP+H20 T CDEA CAMP 109xDNaO (m2/s) Reference (K) (wt%) (wt%) 303 12.5 12.5 1.11 This work 313 12.5 12.5 1.41 u 323 12.5 12.5 1.99 u 303 24.0 6.0 1.15 Li and Lee (1996) 303 18.0 12.0 1.12 u 303 12.0 18.0 1.11 u 303 6.0 24.0 1.08 u 308 24.0 6.0 1.30 u 308 18.0 12.0 1.27 u 308 12.0 18.0 1.24 u 308 6.0 24.0 1.20 u 313 24.0 6.0 1.46 u 313 18.0 12.0 1.43 u 313 12.0 18.0 1.39 u 313 6.0 24.0 1.54 u Table G.11: Constants in eq. (G.7) for interaction parameter Bi2 (this work) System Bi2 No. of data AAD%+ ai a2 M EA+M D EA+ H20 • -5.8800 x 10"6 1.8453 x 10"3 33 6.6 MEA+AMP+H20 -8.7465 x 10"6 2.7373 x 10"3 15 3.7 DEA+MDEA+H20 1.5371 x10"6 -4.3411 x 10"4 31 17.7 DEA+AMP+H20 1.4169 x10"6 -3.0807 xlO"4 15 4.2 +AAD% = average absolute percent deviation 311 + 0s Q 3 O CO ° 1 CM CO CN l< CO CN CO CO E 00 iff X x CD X X t-no in io m T- CO O) oo oq cn T-; CT) cd cvi p 8> S S OOOO uo co CM - CM W co CM CO |T -tf Tf oooo X X X o CD O ^ CM O X LO CO CD N- LO CO CD CD CO CO CO LO LO OOOO T— T— T— T— X X X X CD CD CO OO O LO LO O [•>-CM 00 CO co s •^t CM CD T— |to io co co b b b o T— T— T— T— X X X X CO LO CM CD CD CO LO CO LO O CO CO O ooTNo CN CJ J- CN x i i x < < LU 0-UJ LU Q ^ 2 Q S < APPENDIX H EQUILIBRIUM CONSTANTS The equilibrium constants required to calculates the absorption and desorption rates from the model presented in Chapter 4 were calculated as a function of temperature from the correlations given below. H.1 Water Dissociation Constant The water dissociation constant (K9) as defined by reaction (4.9) was calculated from the correlation reported by Olofsson and Hepler (1975) for the temperature range 293-573 K: log10(K9) = 8909.483 -142^13'6 -4229.195log10(T) + 9.7384T-0.0129638T2 + 1.15068x10-5T3-4.602x10"9T4 (H.1) H.2 Rate Constant for Bicarbonate Formation Reaction The forward rate constant for bicarbonate formation (k7) as defined by reaction (4.7), was calculated from the correlation reported by Pinsent et al. (1956) and corrected for ionic strength by Astarita et al. (1983) for the temperature range of 273-313 K: 313 Iog10(k7) = 13.635- 2895 T + 0.08 Ic (H.2) where, Ic is the ionic strength given by 1 n ^ i=1 H.3 Equilibrium Constant for Bicarbonate Formation Reaction The data for K7K9 were reported by read (1975) for the temperature range of 298-523 K and were correlated according to the following equation: 5652 1 log10 (K7K9) = 115.36 - ^ - 41.882 log10 (T) + 0.0029116T (H.3) H.4 Equilibrium Constant for COf Formation Reaction Equilibrium constant for reaction (4.8), K8, was obtained from the correlation given by Edward et al. (1978) for the temperature range of 273-498 K: H.5 MEA Protonation Constant MEA protonation constant (K15) as defined by reaction (4.15) was calculated from the correlation of Bates and Pinching (1951) for the temperature range of 273-323 K: log10 (K8Kg) = 95.5739- 5399.0187 T -35.4819 log10(T) (H.4) log10(K15K9) =-0.3869- 2677.91 T -0.0004277T (H.5) 314 H.6 DEA Protonation Constant DEA protonation constant (K2i) as defined by reaction (4.21) was calculated from the correlation of Bower et al. (1962) for the temperature range of 273-323 K: log™ (K2iK9) = -4-0302 _183°-15 +0.0043261T (H.6) H.7 MDEA Protonation Constant The MDEA protonation constant (K23) as defined by reaction (4.23) was calculated from the correlation developed using the data of Little et al. (1990) in the temperature range 293-333 K: 3684 5 log10(K23K9) = 37.956-^?^-13.833log10(T) (H.7) H.8 AMP Protonation Constant The AMP protonation constant (KeKg) as defined by equation (4.6) was obtained from the correlation developed using the data of Littel et al. (1990b) in the temperature range of 293-333 K: 2629 9 log10 (K6K9) = -0.39147- - 0.19958 log10 (T) (H.8) 315 H.9 MEA Carbamate Stability Constant MEA carbamate reversion constant (Ku) as defined by reaction (4.14) was calculated from the correlation reported by Kent and Eisenberg (1976) for the temperature range of 298-413 K: log10 (1 / K14) = -2.8451 +13368 - 0.01964log10(T) (H.9) where T is in Kelvin. Note that in the original correlation the temperature was in Rankin. To be consistent with other correlations we converted this into Kelvin. H.10 DEA Carbamate Stability Constant DEA carbamate reversion constant (K2o) as defined by reaction (4.20) was calculated from the correlation reported by Aroua et al. (1997) for the temperature range of 303-331 K: 1781 log10(1/K20) = -5.12 + -^ (H.10) H.11 AMP Carbamate Stability Constant AMP carbamate reversion constant (K5) as defined by reaction (4.5) was calculated using the correlation developed from the data of Xu et al. (1992) for the temperature range of 313-373 K: 6914 6 log10 (1/ K5) = -120.86 + + 38.991 log10 (T) (H.11) 316 H.11 Combined Equilibrium Constants The rate expressions for zwitterion mechanisms (e.g. eqs. 4.31-4.33) involve several combined equilibrium constants such as K<iK2, KK3, K1K4 etc. These equilibrium constants can be obtained by appropriately combining above independent equilibrium constants as follows: ^2=-^- (H.12) KlK3=^ (H.13) KiK4=£- (H.14) K5 ^5=77%- (H.15) K5K15 K1K28 =77^ (H.16) K5K21 ^u=J^7~ (H.17) K14K, 5 K10K12=^ (H.18) k14 K10K13=£- (H.19) K14 K10K24=-%- (H.20) K14K23 317 KloK26=K^R7 (H-21) Kl6Kl7=^cfc (H"22) K k" ^7^9 •Se^ia —j7— (H.23) Ki6Ki9=j^ (H.24) k,gK27=*5t <h-25) K^29=~~ (H.26318 APPENDIX I DETERMINATION OF GAS-SIDE MASS TRANSFER COEFFICIENT In absorption experiments, where mostly pure C02 was used, gas phase resistance to mass transfer can be considered negligible. However, in desorption experiments, where C02 from the liquid film is stripped into humidified N2 gas, the gas-side resistance to mass transfer may become significant and should be included in calculation for desorption rates. This appendix describes the procedure to develop a correlation for gas-side mass transfer coefficient that was used to interpret the desorption data. 1.1 Estimation ofkg To estimate kg, a number of absorption experiments were performed in which gas mixtures containing C02 and N2 were absorbed in 25 wt% aqueous DEA solutions at 303, 313 and 323 K. The mass flow rate of the gas per unit cross-section area was varied from 0.9 to 7 g-m"2-s"1 and the C02 partial pressure in the gas was varied from 5 to 60 mole%. In all these experiments, the liquid flow rate was fixed at 3.12 mL/s. The experimental conditions with measured absorption fluxes are given in Table K.13. The primary objective of these experiments was to conduct experiments in such a way that gas-side resistance mainly controls the mass transfer. The C02 reaction with 319 25 wt% aqueous DEA solution is very fast. So if the C02 partial pressure in the gas phase is low, the gas-side resistance to mass transfer becomes significant and kg can be estimated from the following equation: Equation (1.1) was derived assuming that the liquid phase reaction is in fast pseudo-first order regime with E = VM . The second order rate constant (k2nd) for C02-DEA reaction was calculated from the data of Rinker et al. (1996), the Henry's constant (Hcc,2) was calculated from the correlation given in Appendix F. The C02 partial pressure in the gas-phase (pcc.2) was taken as the log mean partial pressures of C02 at the inlet and out let of the absorption chamber. Table 1.1 shows the measured absorption fluxes and corresponding kg values calculated using equation (1.1). The kg values obtained in this work agree very well with similar measurements reported by Critchfield (1988) using a stirred cell reactor. For the experimental conditions studied in this work, the values of kg vary from 5.5 x 10"7 to 4.5 x 10"6 kmol/kPa m2 s, which constitutes about 20-30% of the total resistance. This implies that under desorption conditions, where C02 diffuses into mixture of N2 and water vapors, the gas-phase resistance to mass transfer should be included in the calculation. k9a aVk2-CAMDCo2 1 H — + —== (1.1) 320 1.2 Development of Correlation for kg The simplest method of representing data for kg is to relate the Sherwood number Sh _ k9RTd PBM to Reynolds number Re = Gd and the Schmidt number Sc = ^g PgDg as follows: Sh = a(Re)b (Sc)c (1.2) For wetted wall columns, Treybal (1980) has reported a = 0.023, b = 0.83 and c = 0.44. These values, however, may vary somewhat from one investigator to another (Treybal, 1980). The important point to note from these results is that kg °= G08, DG56 and PT/PBM-Therefore, we have correlated the kg values measured in this work according to the following equation: kg=aGb D° PT RTp BM (I.3) where a, b and c are constants which can be determined by regression. We were unable to determine the value of the constant c in equation (1.3) by regressing our data. Therefore, we set c = 0.5 as reported in the literature and estimated constants a and b by regression. These values were found to be 83.341 and 0.80531 respectively. The diffusivity of CO2 in gas mixture was calculated from the correlation given by Fuller et al. (1966). Figure 1.1 shows the comparison of the measured and calculated kg values. The agreement is satisfactory. The percent average absolute deviation (AAD%) is 28%. 321 Figure 1.1: Measured and calculated gas-side mass-transfer coefficients. 322 Table 1.1: Estimates of kg from experiments and correlation (CDEA = 25 wt%) T (K) ^av (g/m2s) p total (kPa) Im KC02 (kPa) Flux ( mmol ^ *° 2 (mmol/kPam s) K 2 (mmol/kPam s) U2s J measured calculated from eq. (5.3) 304.8 5.87 101.3 8.36 4.30 2.90 2.36 303.2 2.88 101.4 15.90 6.04 1.01 1.41 303.1 4.63 101.4 16.14 7.04 1.53 2.09 303.3 5.99 101.4 16.89 8.30 2.49 2.61 303.0 1.57 100.5 28.34 9.49 0.74 0.97 303.5 2.30 100.5 31.82 11.14 0.82 1.38 303.1 3.16 100.5 32.27 12.39 1.04 1.81 303.4 0.96 100.7 37.72 11.91 0.65 0.72 303.4 1.45 100.7 43.67 15.04 0.79 1.09 303.3 1.97 100.7 46.64 17.36 0.95 1.47 313.2 6.14 101.3 8.23 4.27 2.85 2.45 313.2 3.00 101.3 15.22 6.24 1.16 1.46 312.7 4.61 101.3 16.01 7.09 1.50 2.08 312.7 6.09 101.3 17.29 8.44 2.17 2.66 312.8 1.50 100.4 27.38 9.30 0.74 0.93 312.8 2.37 100.4 29.85 11.82 1.07 1.41 313.2 3.03 100.4 30.39 13.46 1.47 1.75 312.9 0.97 100.4 38.16 11.28 0.56 0.73 312.9 1.51 100.4 41,51 14.83 0.82 1.11 313.3 2.02 100.4 44.31 17.08 0.98 1.47 322.6 6.31 101.3 7.71 4.38 4.26 2.50 323.1 2.99 101.3 14.76 6.31 1.21 1.45 323.0 4.87 101.3 14.83 8.08 3.16 2.18 322.6 6.37 101.3 15.50 8.61 3.64 2.73 322.7 1.54 100.4 24.87 9.52 0.92 0.94 322.5 2.37 100.4 27.62 12.61 1.51 1.39 322.8 3.33 100.4 28.57 14.22 2.05 1.87 322.8 0.99 100.4 34.59 11.34 0.65 0.72 323.2 1.53 100.4 39.25 15.21 0.94 1.11 322.7 2.16 100.4 41.27 18.14 1.33 1.53 323 APPENDIX J DETERMINATION OF FORMATE AND DEAF J. 1 Formate Analysis The concentration of formate ions in the liquid sample was determined by using ion chromatography (IC). In this method the analyte solution is carried by an eluent through a resin-packed column containing numerous ionic exchange sites. The ion chromatograph used was a Dionex 2020i system (Dionex Corporation, Sunnyvale, CA). Detailed specifications with optimized operating conditions are summarized in Table J.1. The standard solution was prepared from analytical grade formic acid obtained from Sigma-Aldrich. Both the standard solution and sample dilution were made using deionized water. J.2 DEAF Analysis The DEAF content of the liquid samples was first determined using a Hewlett Packard gas chromatograph (model 5830A). The principal operating conditions are summarized in Table J.2. Using this method, we were able to identify a DEAF peak at a retention time of 14 minutes. However, it was not possible to quantify DEAF because of poor reproducibility, especially at low DEAF concentrations. Therefore, a new technique, which is based on the hydrolysis of DEAF to DEA and formate ions (reaction J.1), was developed. (HOC2H4)2NCOH + OH" @333K )(HOC2H4)2NH + HCOO" (J.1) 324 This method consists of three steps: (i) the liquid sample is analyzed for formate ion concentration using aforementioned IC technique, (ii) 2.0 mL of liquid sample is transferred to a 10 mL glass bottle and 0.2 mL of saturated NaOH solution are added. The bottle is closed and the mixture is kept for hydrolysis for 1 hour in a water bath at 333 K. During this time, all the DEAF is converted to DEA and formate ions, (iii) the hydrolyzed sample is analyzed for formate content by the IC technique. The DEAF content of the liquid sample is then determined from the difference in concentration of formate ions before and after hydrolysis. The accuracy and reproducibility of this method was confirmed by hydrolyzing standard DEAF solutions of various concentrations. The standard DEAF solution was prepared from 85% purity stock solution supplied by Shell Technology Center, Houston, Texas. The impurities in the stock solution were mainly water and the acetates. 325 Table J.1: IC operating conditions for formate analysis Item Description Column lonPac ICE-AS1 Detector Dionex Conductivity CDM-1 Regenerator Dionex AMMS-ICEII Eluent 1 mM Heptafluorobutyric acid Regenerant 5 mM Tetrabutyl ammonium hydroxide Sample injection Automatic Injection volume 10 p.L Flow 0.8 mL/min Pressure -6.2 MPa Pump Dionex gradient pump GPM-1 Peak Integrator Peak Simple Chromatography Data System, SRI Instruments, Torrance, CA Table J.2: GC operating conditions for DEAF analysis Item Description Column Tenax TA, 60/80 mesh packed in 6'x1/8" SS packed column (supplied by Supelco Inc. Oakville, ON) Detector H2 flame ionization (FID) Detector temperature 310 °C Injector temperature 280 °C Carrier gas N2 (24 mL/min) Injection volume 5uL Temperature program 180 °C hold for 2 min, ramp @ 8 °C/min to 320 °Cand hold for 5 min 326 APPENDIX K EXPERIMENTAL DATA K.1 C02 Absorption in water Table K.1: C02 absorption in pure water Series Date T QL .. ptotai PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (kPa) (kPa) (mmol/m2s) 1 5/19/98 298 1.25 100.9 97.7 1.897 5/19/98 298 1.85 100.9 97.5 2.177 5/19/98 298 2.50 100.9 97.5 2.484 5/19/98 298 3.10 100.9 97.5 2.540 5/19/98 298 3.61 100.9 97.5 2.627 2 5/20/98 308 1.25 100.9 95.1 1.532 5/20/98 308 1.85 100.9 95.1 1.845 5/20/98 308 2.50 100.9 95.0 2.026 5/20/98 308 3.10 100.9 95.0 2.198 5/20/98 308 3.61 100.9 95.0 2.551 3 5/22/98 318 1.25 100.9 91.1 1.232 5/22/98 318 1.85 100.9 90.9 1.986 5/22/98 318 2.50 100.9 90.9 2.201 5/22/98 318 3.10 100.9 90.9 2.145 5/22/98 318 3.61 101.3 91.3 2.233 4 5/21/98 328 1.25 101.3 85.2 1.107 5/21/98 328 1.85 101.3 85.0 1.552 5/21/98 328 2.50 101.3 85.0 1.725 5/21/98 328 3.10 101.3 84.9 1.886 5/21/98 328 3.61 101.3 84.8 2.134 327 K.2 N20 Absorption in water Table K.2: N20 absorption in pure water Series Date T QL ptotal PN20 Abs. Flux No. (m/d/y) (K) (mL/s) (kPa) (kPa) (mmol/m2s) 5 6/8/98 297 1.25 100.7 97.7 1.435 6/8/98 297 1.85 100.6 97.5 1.568 6/8/98 297 2.50 100.6 97.5 1.682 6/8/98 297 3.10 100.6 97.5 1.941 6/8/98 297 3.61 100.6 97.5 1.847 6 6/2/98 303 1.25 100.6 96.3 1.169 6/2/98 303 1.85 100.6 96.2 1.660 6/2/98 303 2.50 100.6 96.2 1.701 6/5/98 303 3.10 100.6 96.2 1.856 6/5/98 303 3.61 100.6 96.2 2.056 7 6/9/98 313 1.25 100.5 92.9 1.279 6/9/98 313 1.85 100.5 93.0 1.073 6/9/98 313 2.50 100.5 92.9 1.678 6/9/98 313 3.10 100.5 92.9 1.629 6/9/98 313 3.61 100.5 92.9 1.613 8 6/14/98 323 1.25 100.9 88.3 1.151 6/14/98 323 1.85 100.9 88.2 1.453 6/14/98 323 2.50 100.9 88.3 1.103 6/14/98 323 3.10 100.9 88.2 1.389 6/14/98 323 3.61 100.9 88.2 1.213 328 K.3 N20 Absorption in Aqueous Solutions of Single Amines Table K.3: N20 absorption in 25 wt% aqueous MEA solution Series Date T QL ptotai PN20 Abs. Flux No. (m/d/y) (K) (mL/s) (kPa) (kPa) (mmol/m2s) 9 6/16/98 303 1.65 100.9 96.5 1.084 6/16/98 303 2.11 100.9 96.5 1.161 6/16/98 303 2.85 100.9 96.5 1.220 6/16/98 303 3.53 100.9 96.5 1.341 6/16/98 303 4.12 100.9 96.5 1.418 10 6/17/98 313 1.65 100.5 93.0 1.048 6/17/98 313 2.11 100.5 93.0 1.083 6/17/98 313 2.85 100.5 93.0 1.139 6/17/98 313 3.53 100.5 92.9 1.304 6/17/98 313 4.12 100.5 92.9 1.286 11 6/18/98 323 1.65 100.7 88.2 1.037 6/18/98 323 2.11 100.7 88.2 1.079 6/18/98 323 2.85 100.7 88.2 1.084 6/18/98 323 3.53 100.7 88.1 1.137 6/18/98 323 4.12 100.7 88.1 1.185 329 Table K.4: N20 absorption in 25 wt% aqueous DEA solution Series Date T QL ptotai PN20 Abs. Flux No. (m/d/y) (K) (mL/s) (kPa) (kPa) (mmol/m2s) 12 6/20/98 303 1.68 100.5 96.2 0.867 6/20/98 303 2.15 100.5 96.2 0.928 6/20/98 303 2.91 100.5 96.1 0.975 6/20/98 303 3.60 100.5 96.1 1.072 6/20/98 303 4.20 100.5 96.1 1.133 13 6/21/98 313 1.68 100.6 93.1 0.838 6/21/98 313 2.15 100.6 93.1 0.866 6/21/98 313 2.91 100.6 93.1 0.910 6/21/98 313 3.60 100.6 93.1 1.042 6/21/98 313 4.20 100.6 93.1 1.028 14 6/23/98 323 1.68 100.5 87.9 0.829 6/23/98 323 2.15 100.5 87.9 0.863 6/23/98 323 2.91 100.5 87.9 0.867 6/23/98 323 3.60 100.5 87.9 0.909 6/23/98 323 4.20 100.5 87.9 0.947 330 Table K.5: N20 absorption in 25 wt% aqueous MDEA solution Series Date T QL ptotai PN20 Abs. Flux No. (m/d/y) (K) (mL/s) (kPa) (kPa) (mmol/m2s) 15 6/25/98 303 1.75 100.5 96.2 0.924 6/25/98 303 2.24 100.5 96.1 0.994 6/25/98 303 3.02 100.5 96.1 1.019 6/25/98 303 3.74 100.5 96.1 1.131 6/25/98 303 4.36 100.5 96.1 1.196 16 6/26/98 313 1.75 100.7 93.3 0.883 6/26/98 313 2.24 100.7 93.3 0.924 6/26/98 313 3.02 100.7 93.2 0.972 6/26/98 313 3.74 100.7 93.2 1.089 6/26/98 313 4.36 100.7 93.2 1.074 17 6/27/98 323 1.75 101.0 88.5 0.880 6/27/98 323 2.24 101.0 88.5 0.922 6/27/98 323 3.02 101.0 88.5 0.915 6/27/98 323 3.74 101.0 88.4 0.969 6/27/98 323 4.36 101.0 88.4 1.012 331 Table K.6: N20 absorption in 25 wt% aqueous AMP solution Series Date T QL ptotai PN20 Abs. Flux No. (m/d/y) (K) (mL/s) (kPa) (kPa) (mmol/m2s) 18 7/3/98 303 1.70 101.0 96.7 0.861 7/3/98 303 2.17 101.0 96.7 0.955 7/3/98 303 2.93 101.0 96.7 0.843 7/3/98 303 3.63 101.0 96.7 0.994 7/3/98 303 4.24 101.0 96.7 1.050 19 7/4/98 313 1.70 101.0 93.5 0.863 7/4/98 313 2.17 101.0 93.5 0.977 7/4/98 313 2.93 101.0 93.5 1.034 7/4/98 313 3.63 101.0 93.5 1.015 7/4/98 313 4.24 101.0 93.5 1.002 20 7/5/98 323 1.70 100.7 88.2 0.864 7/5/98 323 2.17 100.7 88.2 0.950 7/5/98 323 2.93 100.7 88.2 0.872 7/5/98 323 3.63 100.7 88.2 0.980 7/5/98 323 4.24 100.7 88.2 1.036 332 K.3 N2Q Absorption in Aqueous Solutions of Mixed Amines Table K.7: N20 absorption in aqueous solutions of 12.5 wt% MEA plus 12.5 wt% MDEA. Series Date T QL ptotai PN20 Abs. Flux No. (m/d/y) (K) (mL/s) (kPa) (kPa) (mmol/m2s) 21 7/25/98 303 1.70 100.6 96.3 1.047 7/25/98 303 2.01 100.6 96.3 1.136 7/25/98 303 2.72 100.6 96.3 1.150 7/25/98 303 3.55 100.6 96.3 1.289 7/25/98 303 4.15 100.6 96.2 1.365 22 7/26/98 313 1.70 100.6 93.1 1.016 7/26/98 313 2.01 100.6 93.1 1.091 7/26/98 313 2.72 100.6 93.1 1.150 7/26/98 313 3.55 100.6 93.1 1.233 7/26/98 313 4.15 100.6 93.1 1.217 23 7/27/98 323 1.70 101.0 88.4 1.016 7/27/98 323 2.01 101.0 88.4 1.076 7/27/98 323 2.72 101.0 88.4 1.015 7/27/98 323 3.55 101.0 88.4 1.097 7/27/98 323 4.15 101.0 88.4 1.146 333 Table K.8: N20 absorption in aqueous solutions of 12.5 wt% MEA plus 12.5 wt% AMP. Series Date T QL ptotal PN20 Abs. Flux No. (m/d/y) (K) (mL/s) (kPa) (kPa) (mmol/m2s) 24 7/29/98 303 1.87 101.3 96.9 1.029 7/29/98 303 2.22 101.3 96.9 1.125 7/29/98 303 2.99 101.3 96.9 1.124 7/29/98 303 3.91 101.3 96.9 1.260 7/29/98 303 4.56 101.3 96.9 1.337 25 8/2/98 313 1.87 101.1 93.6 0.972 8/2/98 313 2.22 101.1 93.6 1.058 8/2/98 313 2.99 101.1 93.6 1.116 8/2/98 313 3.91 101.1 93.6 1.159 8/2/98 313 4.56 101.1 93.6 1.143 26 8/3/98 323 1.87 101.0 88.4 0.956 8/3/98 323 2.22 101.0 88.4 1.006 8/3/98 323 2.99 101.0 88.4 1.078 8/3/98 323 3.91 101.0 88.4 1.083 8/3/98 323 4.56 101.0 88.4 1.131 334 Table K.9: N20 absorption in aqueous solutions of 12.5 wt% DEA plus 12.5 wt% MDEA. Series Date T QL ptotai PN20 Abs. Flux No. (m/d/y) (K) (mL/s) (kPa) (kPa) (mmol/m2s) 27 8/7/98 303 1.86 100.9 96.6 0.750 8/7/98 303 2.21 100.9 96.6 0.811 8/7/98 303 2.98 100.9 96.6 0.811 8/7/98 303 3.90 100.9 96.6 0.909 8/7/98 303 4.55 100.9 96.6 0.961 28 8/8/98 313 1.86 101.1 93.7 0.736 8/8/98 313 2.21 101.1 93.7 0.781 8/8/98 313 2.98 101.1 93.7 0.814 8/8/98 313 3.90 101.1 93.7 0.892 8/8/98 313 4.55 101.1 93.7 0.881 29 8/10/98 323 1.86 101.3 88.7 0.707 8/10/98 323 2.21 101.3 88.7 0.745 8/10/98 323 2.98 101.3 88.7 0.721 8/10/98 323 3.90 101.3 88.7 0.771 8/10/98 323 4.55 101.3 88.7 0.805 335 Table K.10: N20 absorption in aqueous solutions of 12.5 wt% DEA plus 12.5 wt% AMP. Series Date T QL ptotal PN20 Abs. Flux No. (m/d/y) (K) (mL/s) (kPa) (kPa) (mmol/m2s) 30 8/15/98 303 1.86 100.7 96.4 0.633 8/15/98 303 2.21 100.7 96.4 0.685 8/15/98 303 2.98 100.7 96.4 0.685 8/15/98 303 3.90 100.7 96.4 0.768 8/15/98 303 4.55 100.7 96.4 0.812 31 8/16/98 313 1.86 100,9 93.4 0.638 8/16/98 313 2.21 100.9 93.4 0.635 8/16/98 313 2.98 100.9 93.4 0.673 8/16/98 313 3.90 100.9 93.4 0.703 8/16/98 313 4.55 100.9 93.4 0.770 32 8/17/98 323 1.86 100.7 88.3 0.635 8/17/98 323 2.21 100.7 88.2 0.669 8/17/98 323 2.98 100.7 88.2 0.682 8/17/98 323 3.90 100.7 88.2 0.707 8/17/98 323 4.55 100.7 88.2 0.738 336 K.4 C02 Absorption in Aqueous Solutions of Single Amines Table K.11: CO2 absorption in aqueous MEA solutions Series Date T QL CMEA ptotal PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (wt%) (kPa) (kPa) (mmol/m2s) 33 11/11/98 303.4 2.06 4.74 101.2 96.91 21.11 11/11/98 303.2 2.05 12.02 101.2 96.96 39.08 11/11/98 303.2 2.10 14.26 101.2 96.96 49.24 11/11/98 303.2 2.04 16.32 101.2 96.96 54.55 11/11/98 303.2 2.05 19.60 101.2 96.96 59.62 11/11/98 303.8 2.04 23.03 101.2 96.81 66.39 11/11/98 303.8 2.09 31.41 101.2 96.81 70.79 Table K.12: CO2 absorption in aqueous DEA solutions Series Date T QL CDEA ptotal PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (wt%) (kPa) (kPa) (mmol/m2s) 34 8/30/98 303.2 1.70 2.95 100.3 96.09 9.99 9/7/98 303.5 1.70 7.05 100.3 96.02 18.29 9/7/98 303.2 1.67 12.40 100.3 96.09 23.51 9/7/98 303.2 1.74 20.00 101.1 96.89 30.26 10/16/99 303.4 1.83 25.00 101.5 97.24 33.87 9/7/98 303.2 1.65 30.01 100.3 96.09 35.86 9/7/98 303.2 1.67 36.11 100.3 96.09 36.55 337 Table K.13: C02 absorption in aqueous DEA solutions for estimation of kg Series Date T QL CDEA ptotai _ lm PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (wt%) (kPa) (kPa) (mmol/m2s) 35 7/29/99 303.4 3.12 25.0 101.4 4.50 3.10 7/30/99 304.8 3.12 25.0 101.3 8.36 4.30 8/3/99 303.2 3.12 25.0 101.4 15.90 6.04 8/3/99 303.1 3.12 25.0 101.4 16.14 7.04 8/3/99 303.3 3.12 25.0 101.4 16.89 8.30 8/5/99 303.0 3.12 25.0 100.5 28.34 9.49 8/5/99 303.5 3.12 25.0 100.5 31.82 11.14 8/5/99 303.1 3.12 25.0 100.5 32.27 12.39 8/9/99 303.4 3.12 25.0 100.7 37.72 11.91 8/9/99 303.4 3.12 25.0 100.7 43.67 15.04 8/9/99 303.3 3.12 25.0 100.7 46.64 17.36 7/29/99 303.7 3.12 25.0 101.3 42.43 30.51 7/29/99 303.5 3.12 25.0 101.3 44.60 22.49 7/29/99 303.5 3.12 25.0 101.3 52.39 15.35 7/30/99 313.3 3.12 25.0 101.3 4.32 2.99 8/4/99 313.2 3.12 25.0 101.3 8.23 4.27 8/4/99 313.2 3.12 25.0 101.3 15.22 6.24 8/4/99 312.7 3.12 25.0 101.3 16.01 7.09 8/6/99 312.7 3.12 25.0 101.3 17.29 8.44 8/6/99 312.8 3.12 25.0 100.4 27.38 9.30 8/6/99 312.8 3.12 25.0 100.4 29.85 11.82 8/10/99 313.2 3.12 25.0 100.4 30.39 13.46 8/10/99 312.9 3.12 25.0 100.4 38.16 11.28 8/10/99 312.9 3.12 25.0 100.4 41.51 14.83 7/30/99 313.3 3.12 25.0 100.4 44.31 17.08 7/30/99 313.2 3.12 25.0 101.3 46.73 16.29 7/30/99 313.2 3.12 25.0 101.3 39.17 23.00 7/30/99 313.2 3.12 25.0 101.3 40.15 31.07 Contd. 338 Table K.13: CO2 absorption in aqueous DEA solutions for estimation of k, (Contd.) Series Date T QL CDEA ptotal PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (wt%) (kPa) (kPa) (mmol/m2s) 35 7/30/99 322.6 3.12 25.0 101.3 4.14 2.66 8/4/99 322.6 3.12 25.0 101.3 7.71 4.38 8/4/99 323.1 3.12 25.0 101.3 14.76 6.31 8/4/99 323.0 3.12 25.0 101.3 14.83 8.08 8/6/99 322.6 3.12 25.0 101.3 15.50 8.61 8/6/99 322.7 3.12 25.0 100.4 24.87 9.52 8/6/99 322.5 3.12 25.0 100.4 27.62 12.61 8/10/99 322.8 3.12 25.0 100.4 28.57 14.22 8/10/99 322.8 3.12 25.0 100.4 34.59 11.34 8/10/99 323.2 3.12 25.0 100.4 39.25 15.21 7/30/99 322.7 3.12 25.0 100.4 41.27 18.14 7/30/99 322.5 3.12 25.0 100.9 36.20 17.30 7/30/99 323.1 3.12 25.0 100.9 31.64 22.67 7/30/99 323.0 3.12 25.0 100.9 34.36 30.52 Table K.14: C02 absorption in aqueous MDEA solutions Series Date T QL CMDEA ptotal PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (wt%) (kPa) (kPa) (mmol/m2s) 36 9/13/98 303.2 1.70 3.94 100.9 96.63 2.43 9/13/98 303.2 1.67 11.17 100.9 96.63 3.75 9/13/98 303.2 1.67 16.09 100.9 96.63 4.47 9/14/98 303.4 1.67 20.10 100.6 96.31 4.63 9/14/98 303.2 1.66 26.93 100.6 96.36 4.47 9/13/98 303.4 1.70 32.02 100.9 96.58 4.23 9/11/98 303.2 1.69 38.75 101.0 96.76 3.80 339 Table K.15: CO2 absorption in aqueous AMP solutions Series Date T QL CAMP ptotal PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (wt%) (kPa) (kPa) (mmol/m2s) 37 6/4/99 296.3 1.75 2.00 100.7 97.89 3.47 6/8/99 295.7 1.84 8.00 101.2 98.45 14.87 6/9/99 296.5 1.70 10.00 101.4 98.53 17.49 6/10/99 296.5 1.77 16.00 101.3 98.42 24.68 5/30/99 295.3 1.47 29.03 100.3 97.64 27.88 5/21/99 296.4 1.50 30.86 100.1 97.21 32.21 38 6/4/99 303.2 1.75 2.00 100.9 96.68 4.68 6/8/99 303.4 1.84 8.00 101.2 96.90 14.18 6/9/99 303.1 1.70 10.00 101.4 97.18 16.95 6/10/99 304.0 1.77 16.00 101.3 96.85 25.28 6/14/99 302.2 1.90 18.00 101.0 97.00 26.65 6/16/99 303.4 1.70 20.00 101.1 96.85 30.73 6/17/99 304.4 1.78 23.00 101.0 96.49 32.03 5/11/99 303.7 2.06 23.89 101.2 96.84 32.71 6/18/99 303.4 1.75 25.00 101.0 96.75 33.13 5/11/99 304.1 1.45 37.82 101.2 96.73 34.62 39 6/4/99 308.4 1.99 2.00 100.7 95.05 4.50 5/26/99 308.5 1.82 7.57 100.2 94.49 12.76 6/8/99 308.4 1.84 8.00 101.2 95.50 14.91 6/9/99 307.8 1.70 10.00 101.4 95.90 16.62 6/10/99 308.3 1.77 16.00 101.3 95.65 24.97 5/25/99 308.5 1.95 16.29 100.9 95.14 26.51 6/14/99 308.4 1.90 18.00 101.0 95.34 26.91 6/16/99 308.4 1.70 20.00 101.1 95.48 31.08 6/17/99 307.8 1.78 23.00 101.0 95.55 33.64 5/21/99 308.5 1.50 30.86 100.1 94.36 36.74 40 6/4/99 313.2 1.99 2.00 100.7 93.38 5.24 5/26/99 313.0 1.82 7.57 100.2 92.89 13.16 6/8/99 313.3 1.84 8.00 101.2 93.79 14.17 6/9/99 313.2 1.70 10.00 101.4 94.01 16.24 6/10/99 312.9 1.77 16.00 101.3 94.05 26.47 5/25/99 313.4 1.95 16.29 100.9 93.40 27.65 6/14/99 314.3 1.90 18.00 101.0 93.19 27.97 6/16/99 313.0 1.70 20.00 101.1 93.85 32.79 6/17/99 312.9 1.78 23.00 101.0 93.77 34.36 5/30/99 313.1 1.47 29.03 100.3 93.00 36.13 5/21/99 313.3 1.50 30.86 100.1 92.65 40.95 340 Table K.15 (contd.): C02 absorption in aqueous AMP solutions Series Date T QL CAMP ptotai PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (wt%) (kPa) (kPa) (mmol/m2s) 41 6/4/99 318.3 1.99 2.00 100.7 91.10 4.17 5/26/99 318.4 1.82 7.57 100.2 90.50 15.34 6/9/99 318.3 1.70 10.00 101.3 91.66 17.07 6/10/99 318.3 1.77 16.00 101.3 91.66 27.45 5/25/99 318.2 1.95 16.29 100.9 91.27 28.66 6/14/99 317.9 1.90 18.00 101.0 91.58 30.02 6/16/99 318.0 1.70 20.00 101.1 91.67 33.77 6/17/99 318.3 1.78 23.00 101.0 91.43 36.49 5/30/99 318.4 1.47 29.03 100.3 90.66 38.77 42 6/4/99 323.2 1.99 2.00 100.7 88.34 4.53 5/26/99 323.2 1.82 7.57 100.2 87.86 16.18 6/9/99 322.8 1.70 10.00 101.3 89.14 17.37 6/10/99 322.6 1.77 16.00 101.3 89.31 28.55 5/25/99 323.3 1.95 16.29 100.9 88.43 28.08 6/14/99 322.6 1.90 18.00 100.6 88.58 30.79 6/16/99 322.6 1.70 20.00 101.1 89.16 34.19 6/17/99 322.6 1.78 23.00 101.0 89.03 36.72 5/30/99 323.2 1.47 29.03 100.3 87.93 41.13 341 K.5 C02 Absorption in Aqueous Amine Blends Table K.16: C02 absorption in aqueous blends of MEA+MDEA Series Date T QL CMEA CMDEA ptotai PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (wt%) (wt%) (kPa) (kPa) (mmol/m2s) 43 8/17/99 303.3 2.56 0.0 25.0 101.4 97.17 4.42 8/19/99 303.4 3.04 5.0 20.0 101.3 96.98 22.78 8/23/99 304.5 3.10 10.0 15.0 100.6 96.09 43.35 8/24/99 303.5 2.94 15.0 10.0 100.7 96.39 59.02 8/24/99 303.5 2.72 20.0 5.0 100.7 96.36 67.80 6/25/99 303.5 1.95 25.0 0.0 101.2 96.87 68.58 Table K.17: CQ2 absorption in aqueous blends of MEA+AMP Series Date T QL CAMP CMEA ptotai PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (wt%) (wt%) (kPa) (kPa) (mmol/m2s) 44 6/18/99 303.4 1.75 0.0 25.0 101.0 96.75 30.94 6/22/99 303.5 1.74 5.0 20.0 101.2 96.86 35.10 6/23/99 303.5 1.68 10.0 15.0 100.8 96.47 42.31 6/24/99 303.2 1.74 15.0 10.0 100.2 95.99 51.52 6/25/99 303.4 1.74 20.0 5.0 100.9 96.56 57.11 6/25/99 303.5 1.95 25.0 0.0 101.2 96.87 68.58 342 Table K.18: C02 absorption in aqueous blends of DEA+MDEA Series Date T QL CMDEA CDEA ptotal PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (wt%) (wt%) (kPa) (kPa) (mmol/m2s) 45 8/17/99 303.3 2.56 0.0 25.0 101.4 97.17 4.42 8/20/99 303.2 3.01 5.0 20.0 101.1 96.87 11.50 8/20/99 303.5 3.08 10.0 15.0 101.1 96.78 19.76 8/20/99 303.5 3.12 15.0 10.0 101.0 96.68 24.02 8/23/99 303.5 2.99 20.0 5.0 100.7 96.39 28.90 10/16/99 303.4 1.83 25.0 0.0 101.5 97.24 33.87 Table K.19: CQ2 absorption in aqueous blends of DEA+AMP Series Date T QL CAMP CDEA ptotal PC02 Abs. Flux No. (m/d/y) (K) (mL/s) (wt%) (wt%) (kPa) (kPa) (mmol/m2s) 46 6/18/99 303.4 1.75 0.0 25.0 101.0 96.75 33.13 6/28/99 303.1 1.79 5.0 20.0 101.3 97.13 30.47 6/28/99 303.6 1.75 10.0 15.0 101.4 97.05 29.27 6/29/99 303.6 1.74 15.0 10.0 101.2 96.90 29.32 6/29/99 303.4 1.77 20.0 5.0 101.2 96.91 30.28 10/16/99 303.4 1.83 25.0 0.0 101.5 97.24 33.87 343 K.6 CO2 Desorption from Aqueous Solutions of Single Amines Table K.20: C02 desorption from aqueous MEA solutions Series Date T a QL CMEA ptotai _ lm PC02 Des. Flux No. (m/d/y) (K) (mol/mol) (mL/s) (wt%) (kPa) (kPa) (mmol/m2s) 47 10/27/98 333.2 0.335 2.20 20.0 100.6 0.50 0.31 10/31/98 333.2 0.384 2.24 20.0 100.7 0.93 0.74 10/23/98 333.2 0.422 2.14 20.0 100.5 0.99 1.49 11/2/98 333.2 0.487 2.14 20.0 101.0 2.47 2.92 48 10/27/98 343.2 0.335 2.20 20.0 100.6 0.74 0.51 10/31/98 343.2 0.384 2.22 20.0 100.7 1.49 1.62 10/23/98 343.2 0.422 2.14 20.0 100.5 1.50 2.65 10/21/98 343.2 0.439 2.10 20.0 100.9 1.70 3.02 10/21/98 343.2 0.443 2.22 20.0 101.3 1.73 3.05 10/21/98 343.2 0.480 2.27 20.0 100.9 2.62 4.73 49 10/27/98 353.2 0.335 2.20 20.0 100.6 1.12 1.40 11/2/98 353.5 0.362 2.22 20.0 101.0 1.18 2.94 10/31/98 353.2 0.384 2.24 20.0 100.9 1.46 3.72 11/2/98 353.5 0.421 2.27 20.0 101.0 1.60 5.23 11/2/98 353.6 0.435 2.25 20.0 101.0 1.55 6.05 50 10/30/98 363.2 0.226 2.23 20.0 202.2 1.00 0.42 10/30/98 363.3 0.269 2.23 20.0 201.2 1.15 0.80 10/28/98 363.3 0.299 2.22 20.0 201.0 1.46 1.37 10/25/98 363.4 0.370 2.27 20.0 200.8 2.99 3.56 10/31/98 363.4 0.398 2.24 20.0 202.3 3.98 5.07 51 10/28/98 373.2 0.205 2.38 20.0 202.7 1.35 0.95 10/24/98 373.3 0.265 2.30 20.0 201.3 1.51 2.27 10/31/98 373.3 0.279 2.23 20.0 202.2 1.85 2.59 10/25/98 373.2 0.360 2.27 20.0 200.8 3.42 6.20 10/30/98 373.4 0.398 2.24 20.0 202.3 4.26 9.71 52 10/28/98 378.3 0.205 2.38 20.0 202.4 1.54 1.43 10/31/98 378.3 0.218 2.35 20.0 202.3 0.66 2.19 10/30/98 378.4 0.279 2.23 20.0 202.3 2.10 4.76 10/31/98 378.5 0.319 2.29 20.0 202.3 1.84 6.88 10/31/98 378.1 0.377 2.24 20.0 202.3 3.03 12.95 344 Table K.21: C02 desorption from aqueous DEA solutions Series Date T a QL CDEA ptotal _. im PC02 Des. Flux No. (m/d/y) (K) (mol/mol) (mL/s) (wt%) (kPa) (kPa) (mmol/m2s) 53 10/17/98 343.2 0.232 2.16 20.0 101.3 0.42 0.73 10/17/98 343.2 0.297 2.28 20.0 101.3 0.88 1.53 10/17/98 343.2 0.369 2.29 20.0 101.3 1.22 2.14 10/17/98 343.2 0.402 2.34 20.0 101.3 1.84 3.25 10/17/98 343.2 0.476 2.25 20.0 101.3 2.43 4.35 54 7/19/99 353.7 0.138 2.63 25.0 206.1 0.48 0.42 7/14/99 353.5 0.205 2.48 25.0 196.3 1.12 0.99 7/12/99 353.6 0.298 2.61 25.0 204.3 2.74 2.49 10/18/98 353.5 0.342 2.52 25.0 204.3 3.85 3.13 7/3/99 353.3 0.380 ' 2.48 25.0 203.8 6.43 3.76 55 7/19/99 363.7 0.138 2.63 25.0 204.7 0.73 0.76 7/14/99 363.5 0.205 2.48 25.0 204.9 2.20 2.04 7/12/99 363.5 0.298 2.06 25.0 204.0 3.96 4.07 10/18/98 363.3 0.342 2.52 25.0 204.3 4.80 6.41 7/3/99 363.1 0.380 2.48 25.0 206.9 5.08 8.57 56 7/19/99 373.5 0.078 2.66 25.0 204.5 0.30 0.36 7/19/99 373.6 0.138 2.63 25.0 204.5 1.11 1.47 7/22/99 373.4 0.163 2.56 25.0 204.7 1.66 2.21 7/22/99 373.4 0.184 2.56 25.0 204.7 2.18 2.64 7/14/99 373.5 0.205 2.48 25.0 206.1 3.03 3.45 57 7/21/99 382.0 0.041 2.72 25.0 204.5 0.16 0.40 7/21/99 381.7 0.069 2.72 25.0 204.5 0.29 0.68 7/21/99 381.8 0.088 2.66 25.0 204.5 0.52 1.27 7/22/99 382.4 0.101 2.66 25.0 204.7 0.58 1.47 7/19/99 382.3 0.138 2.63 25.0 204.7 0.96 2.34 345 Table K.22: C02 desorption from aqueous MDEA solutions Series Date T a QL CMDEA ptotal im PC02 Des. Flux No. (m/d/y) (K) (mol/mol) ( mL/s) (wt%) (kPa) (kPa) (mmol/m2s) 58 10/2/99 343.4 0.132 1.87 25.0 100.6 0.59 0.43 10/4/99 343.5 0.161 1.88 25.0 100.7 0.90 0.95 10/5/99 343.3 0.188 1.89 25.0 100.5 0.92 1.63 10/7/99 343.4 0.223 1.88 25.0 101.3 1.16 2.06 10/9/99 343.4 0.304 1.90 25.0 100.9 1.75 3.54 59 10/2/99 353.3 0.132 1.88 25.0 100.6 0.97 1.19 10/4/99 353.2 0.161 1.88 25.0 101.0 1.24 1.95 10/5/99 353.3 0.188 1.86 25.0 100.9 1.43 2.80 10/7/99 353.3 0.223 1.89 25.0 101.0 1.60 3.54 10/9/99 353.3 0.304 1.89 25.0 101.0 2.72 6.57 60 10/10/99 363.2 0.087 1.87 25.0 202.2 1.42 0.94 10/2/99 363.3 0.132 1.87 25.0 201.2 2.70 1.98 10/4/99 363.3 0.161 1.88 25.0 201.0 3.76 3.39 10/5/99 363.4 0.188 1.89 25.0 200.8 4.12 4.33 10/7/99 363.4 0.223 1.88 25.0 202.3 4.38 5.61 61 10/10/99 373.3 0.087 1.90 25.0 202.7 1.65 1.71 10/2/99 373.2 0.132 1.87 25.0 201.3 2.48 3.35 10/4/99 373.4 0.161 1.88 25.0 202.2 3.66 5.07 10/5/99 373.5 0.188 1.87 25.0 200.8 4.26 6.94 10/7/99 373.4 0.223 1.87 25.0 202.3 4.88 8.69 62 10/10/99 378.3 0.087 1.88 25.0 202.4 1.84 2.03 10/2/99 378.3 0.132 1.91 25.0 202.3 2.51 4.12 10/4/99 378.4 0.161 1.92 25.0 202.3 3.07 6.36 10/5/99 378.4 0.188 1.92 25.0 202.3 2.81 7.43 10/7/99 378.5 0.223 1.89 25.0 202.3 4.24 10.44 346 Table K.23: C02 desorption from aqueous AMP solutions Series Date T a QL CAMP P P002 Des. Flux No. (m/d/y) (K) (mol/mol) (mL/s) (wt%) (kPa) (kPa) (mmol/m2s) ~~63 11/3/98 332.9 0.376 2.02 20.0 100.6 051 036 11/14/98 333.3 0.425 2.01 20.0 100.6 0.67 0.52 11/5/98 333.6 0.568 1.86 20.0 100.6 0.94 1.01 11/16/98 333.6 0.695 2.01 20.0 100.6 1.30 1.46 11/5/98 333.8 0.756 1.76 20.0 100.6 1.27 1J59 64 11/4/98 343.6 0.376 2.02 20.0 100.6 0.75 0.62 11/14/98 343.5 0.425 1.92 20.0 100.6 0.95 0.90 11/5/98 343.1 0.568 1.86 20.0 100.6 1.43 1.79 11/16/98 343.6 0.695 1.81 20.0 100.6 1.77 2.64 11/5/98 343.6 0.756 1.76 20.0 100.6 2.02 3A7_ 65 11/4/98 353.5 0.376 2.02 20.0 100.6 1.69 1.89 11/14/98 353.6 0.425 1.92 20.0 100.6 1.18 2.42 11/10/98 353.5 0.454 1.88 20.0 101.3 1.13 2.92 11/16/98 353.6 0.516 1.81 20.0 101.3 1.38 3.82 11/5/98 353.6 0.568 1.86 20.0 100.6 1.47 434 66 11/3/98 363.4 0.217 2.19 20.0 202.7 1.55 0.89 11/23/98 363.6 0.291 2.01 20.0 202.7 2.10 1.70 11/4/98 363.4 0.376 2.02 20.0 202.4 2.70 3.36 11/16/98 363.6 0.454 1.98 20.0 202.4 3.43 4.58 11/5/98 363.4 0.568 1.86 20.0 202.4 5.45 7,53 67 11/3/98 373.2 0.217 2.19 20.0 202.7 2.29 1.83 11/23/98 373.4 0.291 2.01 20.0 202.7 3.20 3.57 11/10/98 373.6 0.333 1.99 20.0 204.5 2.49 4.60 11/19/98 373.4 0.358 1.92 20.0 204.5 3.11 5.71 11/4/98 373.2 0.376 2.02 20.0 203.8 3.47 6J52 68 11/3/98 378.2 0.217 2.19 20.0 202.3 2.05 2.98 11/23/98 378.5 0.291 2.01 20.0 202.7 2.98 5.26 11/10/98 378.5 0.333 2.02 20.0 205.1 2.47 6.52 11/19/98 378.3 0.358 1.92 20.0 204.5 2.58 7.74 11/4/98 378.6 0.376 1.99 20.0 202.4 2.57 8.73 347 K. 7 CO2 Desorption from Aqueous Amine Blends Table K.24: C02 desorption from aqueous blends of MEA+MDEA Series Date T a QL CMEA CMDEA ptotai PC02 Des. Flux No. (m/d/y) (K) (mol/mol) (mL/s) (wt%) (mL/s) (kPa) (kPa) (mmol/m2s) 69 11/4/99 353.3 0.188 2.04 12.5 12.5 202.5 0.63 0.44 11/5/99 353.4 0.241 2.05 12.5 12.5 201.9 1.17 0.80 11/7/99 353.4 0.276 2.05 12.5 12.5 204.1 1.90 1.42 11/8/99 353.5 0.335 2.08 12.5 12.5 202.6 3.00 3.37 70 11/4/99 363.2 0.188 2.05 12.5 12.5 202.5 0.94 0.95 11/5/99 363.3 0.241 2.05 12.5 12.5 201.9 2.17 2.07 11/7/99 363.3 0.276 2.05 12.5 12.5 204.1 2.56 2.91 11/8/99 363.4 0.335 2.08 12.5 12.5 202.6 4.63 6.19 71 11/9/99 373.4 0.148 2.14 12.5 12.5 202.5 0.80 1.09 11/4/99 373.3 0.188 2.08 12.5 12.5 201.9 1.35 1.91 11/5/99 373.4 0.241 2.06 12.5 12.5 204.1 2.89 3.94 11/7/99 373.3 0.276 2.07 12.5 12.5 202.6 3.67 6.16 Table K.25: C02 desorption from aqueous blends of MEA+AMP Series Date T ot QL CMEA CAMP ptotai PC02 Des. Flux No. (m/d/y) (K) (mol/mol) (mL/s) (wt%) (mL/s) (kPa) (kPa) (mmol/m2s) 72 11/13/99 353.3 0.226 2.11 12.5 12.5 202.5 0.76 0.30 11/14/99 353.5 0.278 2.07 12.5 12.5 202.5 1.03 0.77 11/16/99 353.3 0.306 2.06 12.5 12.5 202.3 1.45 1.24 11/20/99 353.6 0.359 2.04 12.5 12.5 202.5 2.18 2.20 73 11/13/99 363.3 0.226 2.21 12.5 12.5 202.5 1.50 0.75 11/14/99 363.4 0.278 2.12 12.5 12.5 202.5 2.20 1.66 11/16/99 363.3 0.306 2.12 12.5 12.5 202.3 2.31 2.51 11/20/99 363.5 0.359 2.13 12.5 12.5 202.5 3.67 4.62 74 11/13/99 373.2 0.226 2.29 12.5 12.5 202.5 2.38 1.80 11/14/99 373.3 0.278 1.15 12.5 12.5 202.5 2.90 3.80 11/16/99 373.4 0.306 2.11 12.5 12.5 202.3 3.21 5.33 11/20/99 373.3 0.359 2.10 12.5 12.5 202.5 4.63 8.56 348 Table K.26: C02 desorption from aqueous blends of DEA+MDEA Series Date T a QL CDEA CMDEA ptotal _ im PC02 Des. Flux No. (m/d/y) (K) (mol/mol) (mL/s) (wt%) (mL/s) (kPa) (kPa) (mmol/m2s) 75 12/8/99 353.5 0.152 2.58 12.5 12.5 205.9 0.86 0.67 12/19/99 353.4 0.186 2.58 12.5 12.5 209.0 1.62 1.05 12/18/99 353.5 0.202 2.58 12.5 12.5 208.8 1.72 1.41 12/12/99 353.2 0.267 2.58 12.5 12.5 206.7 2.41 2.62 76 12/8/99 363.4 0.152 2.58 12.5 12.5 208.0 1.35 1.23 12/19/99 363.4 0.186 2.58 12.5 12.5 209.0 2.48 2.10 12/18/99 363.4 0.202 2.58 12.5 12.5 208.8 2.57 2.50 12/12/99 363.1 0.267 2.58 12.5 12.5 206.7 3.93 4.99 77 12/19/99 373.3 0.101 2.58 12.5 12.5 209.0 1.25 1.33 12/8/99 373.2 0.152 2.58 12.5 12.5 208.0 2.25 2.77 12/19/99 373.5 0.186 2.58 12.5 12.5 209.0 3.13 4.30 12/18/99 373.4 0.202 2.58 12.5 12.5 208.8 3.01 4.98 Table K.27: C02 desorption from aqueous blends of DEA+AMP Series Date T a QL CDEA CAMP ptotal PC02lm Des. Flux No. (m/d/y) (K) (mol/mol) (mL/s) (wt%) (mL/s) (kPa) (kPa) (mmol/m2s) 78 12/2/99 353.4 0.169 2.38 12.5 12.5 208.0 0.66 0.38 12/5/99 353.4 0.231 2.38 12.5 12.5 207.3 1.09 0.76 11/29/99 354.0 0.260 2.38 12.5 12.5 206.0 1.25 1.18 11/28/99 353.5 0.335 2.38 12.5 12.5 205.9 2.19 2.27 79 12/2/99 363.3 0.169 2.38 12.5 12.5 208.0 1.15 0.79 12/5/99 363.3 0.231 2.38 12.5 12.5 207.3 2.07 1.71 11/29/99 363.3 0.260 2.38 12.5 12.5 206.7 1.97 2.16 11/28/99 364.1 0.335 2.38 12.5 12.5 207.6 3.47 4.27 80 12/2/99 373.1 0.169 2.38 12.5 12.5 208.0 2.13 1.95 12/5/99 373.2 0.231 2.38 12.5 12.5 207.3 2.86 3.21 11/29/99 373.1 0.260 2.38 12.5 12.5 206.7 2.76 4.07 11/28/99 372.9 0.335 2.38 12.5 12.5 206.2 4.18 6.99 349 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Country Views Downloads
United States 96 21
United Kingdom 46 5
Canada 38 14
Germany 37 64
China 30 15
India 28 9
France 18 0
Malaysia 18 4
Algeria 16 1
Indonesia 15 5
Turkey 15 7
Japan 14 5
Norway 13 2
City Views Downloads
Unknown 195 102
Ashburn 19 3
Regina 18 4
Mountain View 14 3
Jakarta 11 5
Shenzhen 10 12
Beijing 8 2
Aberdeen 8 3
Seattle 7 0
Oshawa 6 0
Guwahati 6 5
London 6 0
Batu Caves 5 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058971/manifest

Comment

Related Items