i
Sterically Hindered Amine based Absorbents and Application for CO2 Capture in Membrane Contactors
Thèse
Francis Bougie
Doctorat en génie chimique Philosophiae Doctor (Ph.D.)
Québec, Canada
© Francis Bougie, 2014
ii
iii
Résumé La séparation des gaz dans des contacteurs à membrane (MC) est une technologie de
pointe qui offre plusieurs avantages par rapport aux contacteurs traditionnels (colonnes
garnies), mais très peu d'efforts ont été consacrés pour développer de nouvelles solutions
absorbantes spécialement optimisées pour les applications dans les MC. Actuellement,
aucun absorbant disponible ne répond complètement aux exigences pour la mise en œuvre
de la séparation industrielle des gaz acides, le CO2 en particulier, dans les contacteurs à
membranes. L'objectif principal de ce travail a été de développer un absorbant à base
d’alcanolamine à encombrement stérique (SHA), présentant les caractéristiques spécifiques
exigées pour application dans les MC (bonnes capacité et cinétique d’absorption,
régénération facile et plus économique, résistance à la dégradation, compatibilité avec les
membranes et haute tension superficielle) et d’étudier son efficacité pour la capture du CO2
dans différentes configurations de contacteurs à membrane et conditions opératoires.
Bien que les alcanolamine fortement encombrées stériquement sont caractérisées par
une faible cinétique d’absorption du CO2, le fait qu’elles possèdent un grand potentiel pour
réduire la consommation d'énergie lors de la régénération des solutions riches en CO2 a été
l’un des paramètres clés dans le choix de l’AHPD (2-amino-2-hydroxyméthyle-1,3-
propanediol). Pour améliorer le taux d'absorption, la pipérazine (Pz) s'est avérée un
activateur très efficace; l'addition de petites quantités de Pz aux solutions aqueuses
d’AHPD améliore significativement la cinétique d'absorption du CO2. Il a été aussi trouvé
que le mélange AHPD-Pz a également une très bonne capacité d’absorption. L'étude de la
régénération des solutions d’amines usées (contenant du CO2) a révélé que des solutions à
base d’alcanolamines fortement encombrées stériquement (AHPD en particulier), sont
beaucoup plus facilement régénérables par rapport à la MEA, l'amine de référence utilisée
industriellement dans la séparation des gaz acides. De plus, l'ajout d'une petite quantité de
Pz dans une solution aqueuse d’AHPD permet d’obtenir presque la même capacité cyclique
et efficacité de régénération que les solutions non-activées par la Pz, mais pour la moitié de
la durée du processus d'absorption.
Outre les propriétés absorbantes des liquides, les performances des MC pour la
séparation du CO2 dépendent fortement de la compatibilité entre la membrane et
iv
l’absorbant. Sur la base des propriétés liées au mouillage des membranes, comme la tension
superficielle du liquide, l’angle de contact, la pression de percée et la stabilité chimique,
une nouvelle méthode graphique d’estimation de la tension superficielle des solutions
aqueuses d'amines, d'alcools ou d’alcanolamines a été développée pour permettre la
sélection des meilleures conditions pour éviter le mouillage des membranes. Il a été trouvé
que les solutions à base d’AHPD (comme AHPD + Pz) ont un fort potentiel d'utilisation
dans les MC en raison de leur tension superficielle élevée. La méthode développée a aussi
permis d'identifier de nouvelles amines potentielles pouvant être utilisées dans les MC.
Une bonne stabilité et résistance à la dégradation est une autre caractéristique
importante des solutions absorbantes. L'étude de la stabilité de différentes solutions
aqueuses d’amines à la dégradation thermique et oxydative, en absence et en présence de
CO2, a révélé que les SHA sont plus résistantes à la dégradation thermique que les amines
conventionnelles, mais que la présence d'oxygène les dégrade plus significativement en
absence de CO2. Toutefois, la présence de CO2 dans les solutions à base de SHA est
bénéfique, car la formation préférentielle du bicarbonate conduit à une réduction
significative du taux de dégradation oxydative. Le faible degré de dégradation de la
solution aqueuse AHPD + Pz confirme son potentiel comme absorbant pour le CO2.
Finalement, la performance des solutions aqueuses AHPD + Pz pour la capture du CO2
dans des MC a été étudiée dans différentes conditions opératoires et configurations des
modules (fibres creuses et membranes plates, membranes en PTFE, PP et laminées
PTFE/PP, différents débits du liquide, compositions de gaz et orientations des flux gazeux
et liquide (co- et contre-courant)). Les solutions AHPD + Pz ont montré une excellente
performance. Sur la base des données expérimentales, une étude de modélisation de la
capture du CO2 dans des MC à fibres creuses PTFE a démontré l'effet positif des solutions
présentant une tension superficielle élevée sur la réduction du mouillage de la membrane.
En conclusion, les résultats de cette thèse ont montré que les solutions aqueuses AHPD
+ Pz possèdent une bonne capacité et cinétique d’absorption, régénération plus facile et
moins énergivore, résistance à la dégradation, haute tension superficielle et démontre
d'excellentes performances pour la capture du CO2 dans les MC, en représentant une
alternative intéressante à la MEA.
v
Abstract Gas separation in membrane contactors (MC) is a forefront technology offering several
advantages over traditional packed columns, but very few efforts have been made to
develop new absorbent solutions optimized specifically for application in MC. Currently,
no available absorbent meets all required characteristics for the implementation of
membrane contactors for acid gas separation (CO2 in particular) in industrial units. The
main objective of this work was to develop a dedicated sterically hindered alkanolamine
(SHA) based absorbent with improved characteristics for application in MC (good
absorption capacity and reaction kinetics, regeneration facility, resistance to degradation,
compatibility with membranes and high surface tension) and to investigate its efficiency for
CO2 capture in different membrane contactor configurations and operation conditions.
Although low kinetics characterizes highly sterically hindered alkanolamines, their
potential to reduce the energy consumption during the regeneration step brings us to focus
on AHPD (2-amino-2-hydroxymethyl-1,3-propanediol). To improve the absorption rate,
piperazine (Pz) was found to be a very effective activator; the addition of small amounts of
Pz to aqueous AHPD solutions has significant effect on the enhancement of the CO2
absorption rate. The blend AHPD-Pz was also found to present very good absorption
capacity. The investigation of the regeneration of loaded (CO2 containing) amine solutions
revealed that highly hindered SHA based solutions (AHPD in particular) are much easier to
regenerate compared to MEA, the benchmark amine industrially used in acid gas
separations. Moreover, the addition of small amount of Pz into AHPD aqueous solution
allowed to obtain almost the same cyclic capacity and regeneration efficiency as non-
activated solutions, but for half of the absorption time.
Besides the liquid absorbent properties, the performances of MC for CO2 separation
strongly depend on the compatibility between absorbent and membrane. Based on wetting-
related properties like liquid surface tension, contact angle, membrane breakthrough
pressure and chemical stability, a new graphical surface tension estimation method for
aqueous amine, alcohol or alkanolamine solutions was developed to select the best
conditions to elude the unwanted membrane wetting phenomenon. AHPD-based solutions
(like the AHPD + Pz solution) were found to have a strong potential for use in MC because
vi
of their very high surface tension. In addition, the developed method allowed to identify
new potential amines for use in MC.
A good stability and resistance to degradation is another important feature of CO2
absorbents. The investigation of the stability of different aqueous amine solutions to
thermal and oxidative degradation, in the absence and the presence of CO2, revealed that
SHA are more resistant to thermal degradation than conventional amines, but the presence
of oxygen degraded them more significantly in the absence of CO2. However, the presence
of CO2 is beneficial to SHA as the preferential bicarbonate formation in solutions reduces
by a large extent the oxidative degradation rate. The low degradation degree of the AHPD
+ Pz aqueous solution reaffirms its potential as CO2 absorbent.
Finally, the performance of the AHPD + Pz aqueous solution for CO2 capture in MC
was investigated in different operational conditions and module configurations (hollow
fibers and flat sheets membranes, PTFE, PP and laminated PTFE/PP membranes, various
liquid flow rates, gas compositions and flow orientation (co- and counter-current)).
Excellent performance was found for AHPD + Pz solutions. Based on experimental data, a
modeling study of CO2 capture in PTFE hollow fiber MC revealed the positive effect of
solutions presenting high surface tension on the reduction of membrane wetting.
In summary, the results of this thesis showed that AHPD + Pz aqueous solution possess
good absorption capacity, reaction kinetics, regenerative potential, and degradation
resistance, as well as high surface tension and showed excellent performance for CO2
capture in MC, representing an interesting alternative to MEA.
vii
Table of Contents Résumé .................................................................................................................................. iii Abstract .................................................................................................................................. v Table of Contents ................................................................................................................. vii Index of Tables ................................................................................................................... xiii Index of Figures ................................................................................................................. xvii Nomenclature ...................................................................................................................... xxi Acknowledgement ........................................................................................................... xxvii Preface .............................................................................................................................. xxix Chapter 1. Introduction .......................................................................................................... 1
1.1. Background .................................................................................................................. 1 1.2. Sterically hindered amines based absorbents for the removal of CO2 from gas
streams ........................................................................................................................ 6 1.2.1. Introduction ........................................................................................................... 7 1.2.2. Structure and properties of SHA ........................................................................... 8
1.2.2.1. Structure of SHA ........................................................................................... 8 1.2.2.2. Physical properties of single and mixed SHA aqueous mixtures .................. 8
1.2.3. Mechanism of reaction between CO2 and SHA. Influence of steric hindrance on carbamate stability ............................................................................................... 36
1.2.4. Absorption capacity ............................................................................................. 38 1.2.4.1. CO2 chemical solubility in single amine aqueous solutions ........................ 38 1.2.4.2. CO2 chemical solubility in SHA based mixed solvents ............................... 43 1.2.4.3. CO2 physical solubility in single and mixed solvents .................................. 47
1.2.5. Absorption kinetics ............................................................................................. 50 1.2.5.1. Single AMP systems .................................................................................... 51 1.2.5.2. Blended AMP systems ................................................................................. 55 1.2.5.3. Other SHA systems ...................................................................................... 58
1.2.6. Regeneration capability ....................................................................................... 63 1.2.7. Conclusions and recommendations for future research ...................................... 69
1.3. CO2 capture in amine solution absorbents using membrane contactors .................... 70 1.3.1. Principle of gas absorption in MC ....................................................................... 73 1.3.2. Membrane module configurations ...................................................................... 74 1.3.3. Absorbent screening for MC and liquid/membrane compatibility with polymeric
membranes ........................................................................................................... 77 1.3.4. CO2 absorption in membrane contactors using SHA .......................................... 82
1.4. Conclusions ................................................................................................................ 88 1.5. Objective of the work ................................................................................................ 90
Chapter 2. Kinetics of absorption of carbon dioxide into aqueous solutions of 2-amino-2-hydroxymethyl-1,3-propanediol...................................................................................... 93 2.1. Introduction ................................................................................................................ 95 2.2. Theory ........................................................................................................................ 97
2.2.1. Physical absorption ............................................................................................. 97 2.2.2. Chemical absorption ............................................................................................ 98
2.3. Experimental ............................................................................................................ 100
viii
2.3.1. Reagents ............................................................................................................ 100 2.3.2. Experimental setup ........................................................................................... 100 2.3.3. Experimental procedure .................................................................................... 101
2.4. Results and Discussions .......................................................................................... 103 2.4.1. Physicochemical properties of aqueous AHPD solutions ................................ 103 2.4.2. Physical absorption ........................................................................................... 103 2.4.3. Chemical absorption ......................................................................................... 107 2.4.4. Hindrance effect on the SHA properties ........................................................... 114
2.5. Conclusion ............................................................................................................... 117 Chapter 3. Acceleration of the reaction of carbon dioxide into aqueous 2-amino-2-
hydroxymethyl-1,3-propanediol solutions by piperazine addition ................................ 119 3.1. Introduction ............................................................................................................. 121 3.2. Theory ..................................................................................................................... 123
3.2.1. Physical absorption ........................................................................................... 123 3.2.2. Chemical absorption ......................................................................................... 124
3.3. Experimental ........................................................................................................... 126 3.3.1. Reagents ............................................................................................................ 126 3.3.2. Experimental setup and procedure .................................................................... 126
3.3.2.1 Density and viscosity measurements.......................................................... 126 3.3.2.2 Physical absorption and CO2 absorption rate measurements ..................... 126
3.4. Results and discussion ............................................................................................. 129 3.4.1. Physicochemical properties of solutions .......................................................... 129 3.4.2. Physical absorption ........................................................................................... 130 3.4.3. Chemical absorption ......................................................................................... 132
3.4.3.1 Data analysis and kinetic reaction rate constants ....................................... 132 3.4.3.2 Fast pseudo-first-order regime verification ................................................ 136 3.4.3.3 Enhancement effect of PZ additions in SHA solutions .............................. 139
3.4.4. Prospective and future studies .......................................................................... 140 3.5. Conclusion ............................................................................................................... 140
Chapter 4. CO2 absorption into mixed aqueous solutions of 2-amino-2-hydroxymethyl-1,3-propanediol and piperazine ............................................................................................ 143 4.1. Introduction ............................................................................................................. 145 4.2. Experimental ........................................................................................................... 147
4.2.1 Reagents ............................................................................................................. 147 4.2.2 Apparatus and procedures .................................................................................. 147
4.3. Thermodynamic modeling of the vapour-liquid equilibrium .................................. 149 4.3.1. Chemical equilibrium in the liquid phase ......................................................... 149 4.3.2. Vapour-liquid equilibrium ................................................................................ 151 4.3.3. Thermodynamic properties ............................................................................... 151 4.3.4. Pitzer’s GE model for activity coefficients and interaction parameters ............ 152
4.3.4.1. The system AHPD-CO2-H2O .................................................................... 156 4.3.4.2. The system AHPD-Pz-CO2-H2O ............................................................... 156
4.4. Results and discussions ........................................................................................... 157 4.4.1 Experimental setup verification ......................................................................... 157 4.4.2 Solubility measurements .................................................................................... 157
ix
4.5. Conclusion ............................................................................................................... 166 Chapter 5. CO2 Absorption in Aqueous Piperazine Solutions: Experimental Study and
Modeling ....................................................................................................................... 169 5.1. Introduction .............................................................................................................. 171 5.2. Experimental section ................................................................................................ 172
5.2.1 Reagents ............................................................................................................. 172 5.2.2 Apparatus and procedures .................................................................................. 172
5.3. Thermodynamic modeling of the vapour-liquid equilibrium .................................. 174 5.3.1. Chemical equilibrium in the liquid phase ......................................................... 174 5.3.2. Vapour-liquid equilibrium ................................................................................. 175 5.3.3. Pitzer’s GE model for activity coefficients ........................................................ 176
5.3.3.1 Interaction parameters for the system CO2-Pz-H2O ................................... 177 5.4. Results and discussions ............................................................................................ 178
5.4.1 CO2-Pz-H2O solubility database ........................................................................ 178 5.4.2 Solubility measurements .................................................................................... 179 5.4.3 Modeling results ................................................................................................. 183
5.5. Conclusions .............................................................................................................. 186 Chapter 6. Analysis of regeneration of sterically hindered alkanolamines aqueous solutions
with and without activator. ............................................................................................ 189 6.1 Introduction ............................................................................................................... 191 6.2. Material and methods ............................................................................................... 192
6.2.1 Reagents ............................................................................................................. 192 6.2.2 Apparatus and procedures .................................................................................. 193
6.3. Results and discussion ............................................................................................. 194 6.3.1 Analysis of the regeneration time and temperature ............................................ 194 6.3.2 Amine influence on regeneration efficiency ...................................................... 197 6.3.3 Effect of activator addition on regeneration efficiency ...................................... 199
6.4. Conclusions .............................................................................................................. 201 Chapter 7. Analysis of Laplace-Young equation parameters and their influence on efficient
CO2 capture in membrane contactors ............................................................................ 203 7.1. Introduction .............................................................................................................. 205 7.2. Experimental ............................................................................................................ 206
7.2.1 Reagents ............................................................................................................. 206 7.2.2 Apparatus and Procedures .................................................................................. 208
7.2.2.1 Surface tension ............................................................................................ 208 7.2.2.2 Density and viscosity of solutions .............................................................. 208 7.2.2.3 Contact angle .............................................................................................. 209 7.2.2.4 Breakthrough pressure ................................................................................ 209
7.3. Results and Discussion ............................................................................................ 210 7.3.1 Absorbent density and viscosity ......................................................................... 210 7.3.2 Absorbent surface tension .................................................................................. 211 7.3.3 Membrane/absorbent contact angle .................................................................... 218 7.3.4 Breakthrough pressure ........................................................................................ 220
7.3.4.1 Relationship between membrane long-term stability and breakthrough pressure ................................................................................................................... 222
x
7.3.4.2 Viscosity influence on breakthrough pressure ........................................... 226 7.4. Conclusions ............................................................................................................. 226
Chapter 8. Solubility of CO2 in and Density, Viscosity and Surface Tension of Aqueous 2-Amino-1,3-propanediol (Serinol) Solutions .................................................................. 229 8.1. Introduction ............................................................................................................. 231 8.2. Experimental section ............................................................................................... 233
8.2.1 Reagents ............................................................................................................. 233 8.2.2 Apparatus and Procedures ................................................................................. 234
8.2.2.1 Density and viscosity of solutions .............................................................. 234 8.2.2.2 Surface tension of solutions ....................................................................... 234 8.2.2.3 CO2 Solubility measurements .................................................................... 234
8.3. Results and Discussion ............................................................................................ 235 8.3.1. Density and viscosity of solutions .................................................................... 235 8.3.2. Surface tension of solutions .............................................................................. 238 8.3.3. CO2 Solubility ................................................................................................... 240
8.3.3.1 Solution concentration effect on solubility ................................................ 240 8.3.3.2 Temperature effect on solubility ................................................................ 242
8.4. Conclusions ............................................................................................................. 245 Chapter 9. Thermal and oxidative degradation of aqueous amine solutions used for CO2
capture ........................................................................................................................... 249 9.1. Introduction ............................................................................................................. 251 9.2. Material and methods .............................................................................................. 252
9.2.1. Chemicals ......................................................................................................... 252 9.2.2. Thermal degradation: typical experimental run ................................................ 254 9.2.3. Combined thermal and oxidative: typical experimental degradation run ......... 254 9.2.4. Degradation in the presence of CO2 ................................................................. 255 9.2.5. HPLC analysis .................................................................................................. 255
9.3. Results ..................................................................................................................... 256 9.3.1. Percentage of amine loss .................................................................................. 256 9.3.2. Amine degradation first-order rate constant ..................................................... 256 9.3.3. Qualitative observations ................................................................................... 259
9.4. Discussions .............................................................................................................. 260 9.4.1. Effect of process conditions on amine degradation .......................................... 260
9.4.1.1 Pure thermal degradation ........................................................................... 260 9.4.1.2. Oxygen effect on amine degradation ........................................................ 261 9.4.1.3. CO2 effect on amine degradation .............................................................. 261
9.4.2. Degradation analysis of the AHPD + Pz blend ................................................ 262 9.5. Conclusions ............................................................................................................. 263
Chapter 10. Absorption of CO2 into Pz-activated AHPD aqueous solutions in PTFE hollow fiber membrane contactors: Experimental and modeling study. ................................... 265 10.1. Introduction ........................................................................................................... 267 10.2. Membrane contactor model ................................................................................... 269
10.2.1. Porous membrane scale model ....................................................................... 269 10.2.2. Liquid boundary layer (liquid film) scale model ............................................ 271 10.2.3. Gas–liquid membrane contactor scale model ................................................. 272
xi
10.2.4. Model parameters ............................................................................................ 273 10.2.5. Numerical implementation .............................................................................. 274
10.3. Experimental .......................................................................................................... 274 10.3.1. Chemicals ........................................................................................................ 274 10.3.2. Membrane module ........................................................................................... 274 10.3.3. Absorption setup and procedure ...................................................................... 275
10.4. Results and Discussion .......................................................................................... 277 10.4.1. Effect of liquid flow rate on CO2 absorption .................................................. 277 10.4.2. Effect of gas phase composition on CO2 absorption ....................................... 278 10.4.3. Flow configuration and CO2 removal efficiency ............................................ 279 10.4.4. Model analysis – effect of membrane wetting ................................................ 280
10.5. Conclusion ............................................................................................................. 282 Chapter 11. Flat sheet membrane contactors (FSMC) for CO2 separation in aqueous amine
solutions ........................................................................................................................ 285 11.1. Introduction ............................................................................................................ 287 11.2. Experimental .......................................................................................................... 289
11.2.1. Chemicals ........................................................................................................ 289 11.2.2. Flat sheet membrane contactor ........................................................................ 290 11.2.3. Absorption setup and procedure ...................................................................... 290
11.3. Results and Discussion .......................................................................................... 292 11.3.1. Effect of liquid flow rate on CO2 absorption flux ........................................... 292 11.3.2. Membrane quantity effect on CO2 absorption rate .......................................... 293 11.3.3. Effect of gas phase composition and flow configuration on CO2 absorption flux
........................................................................................................................... 294 11.3.4. CO2 removal percentage .................................................................................. 295 11.3.5. Influence of membrane properties ................................................................... 296
11.4. Conclusion ............................................................................................................. 297 Chapter 12. General Conclusions and Suggestions for Future work ................................. 299 References .......................................................................................................................... 305 Appendix A ........................................................................................................................ 327
xii
xiii
Index of Tables Table 1.1. Structure of several sterically hindered amines ................................................. 10
Table 1.2. Current research for CO2 capture in MC ........................................................... 79
Table 2.1. Regressed coefficients for density, viscosity and ( )2 2
1/ 2N O N O AHPD
/D H correlations
........................................................................................................................................... 104
Table 2.2. Kinetic data for absorption of CO2 in AHPD aqueous solutions at 303.15 K ......................................................................................................................... 106
Table 2.3. Kinetic data for absorption of CO2 in AHPD aqueous solutions at 313.15 K ......................................................................................................................... 107
Table 2.4. Kinetic data for absorption of CO2 in AHPD aqueous solutions at 323.15 K ......................................................................................................................... 107
Table 2.5. Reaction rate parameters for CO2 absorption in aqueous AHPD solutions ..... 112
Table 3.1. Densities and viscosities of PZ-AHPD solutions ............................................. 129 Table 3.2. Regressed coefficients for density, viscosity and ( )2 2
1/ 2N O N O Amines
/D H correlations
........................................................................................................................................... 130
Table 3.3. Kinetic data for absorption of CO2 in PZ-AHPD aqueous solutions ............... 133
Table 3.4. Parameters for pseudo-first order regime verification of PZ-AHPD-H2O systems ........................................................................................................................................... 134
Table 4.1. Henry’s constant for the solubility of carbon dioxide in pure water ............... 152
Table 4.2. Equilibrium constants for chemical reactions (4.1)-(4.10). ............................. 153
Table 4.3. Interaction parameters in Pitzer’s GE equation for the system AHPD-PZ-CO2-H2O .................................................................................................................................... 154
Table 4.4. Henry's law constants for N2O in Pz (1)-AHPD (2) solutions ......................... 158
Table 4.5. CO2 solubility in AHPD aqueous solutions ..................................................... 160
Table 5.1. Chemical Equilibrium Constant (on the molality scale) for the Chemical Reaction R, Expressed on the Molality Scale, and Temperature Range of Validity. .............. ........................................................................................................................................... 175
Table 5.2. Number of Reliable Data of CO2 (1) Solubility in Aqueous Solution of Piperazine (2) and their Source .......................................................................................... 179
xiv
Table 5.3. Solubility of CO2 (1) in Aqueous Solution of Piperazine (2) at T = 287.1 K (∆T = ± 0.1 K) ........................................................................................................................... 180
Table 5.4. Solubility of CO2 (1) in Aqueous Solution of Piperazine (2) at T = 293.1 K (∆T = ± 0.1 K) ........................................................................................................................... 181
Table 5.5. Solubility of CO2 (1) in Aqueous Solution of Piperazine (2) at T = 298.1 K (∆T = ± 0.1 K) ........................................................................................................................... 181
Table 5.6. Solubility of CO2 (1) in Aqueous Solution of Piperazine (2) at T = 303.1 K (∆T = ± 0.1 K) ........................................................................................................................... 182
Table 5.7. Solubility of CO2 (1) in Aqueous Solution of Piperazine (2) at T = 313.1 K (∆T = ± 0.1 K) ........................................................................................................................... 182
Table 5.8. Interaction Parameters in Pitzer's GE Equation for the Ternary CO2-Pz-H2O System as in Eq. (5.17) for a Temperature range of 287.1 K to 395.1 K........................... 184
Table 6.1. Regeneration efficiency of various amines. ..................................................... 197
Table 6.2. Regeneration of AHPD with or without Pz. ..................................................... 199
Table 7.1. Characteristics of membranes used in this work. ............................................. 207
Table 7.2. Density and viscosity of aqueous amine solutions. .......................................... 210
Table 7.3. Surface tension of aqueous amine solutions. .................................................... 211
Table 7.4. Surface tension around 298 K and 30 wt.% of various aqueous amine solutions and their carbon and hydrophilic numbers. ........................................................................ 214
Table 7.5. Contact angles for several absorbent/membrane combinations. ...................... 218
Table 7.6. Alkalinity of tested amine solutions ................................................................. 219
Table 7.7. Experimental breakthrough pressure (∆PB.P.exp) for maximum pore size determination using water at 298.2 K. ............................................................................... 221
Table 7.8. Breakthrough pressure using water and aqueous amine solutions with PTFE 2. ............................................................................................................................................ 223
Table 8.1. Chemicals information. .................................................................................... 233
Table 8.2. Experimental Values of Density ρ and Viscosity µ of Aqueous Serinol Solutions determined at Temperature T, Amine-Molality m and Atmospheric Pressure (P = 101.3 kPa) ..................................................................................................................................... 236
Table 8.3. Values of the Regressed Coefficients for Eqs (8.1) to (8.3). ............................ 238
xv
Table 8.4. Experimental Values of Surface Tension σ of Aqueous Serinol Solutions determined at Temperature T, Amine-Molality m and Atmospheric Pressure (P = 101.3 kPa) .................................................................................................................................... 239
Table 8.5. Experimental Values of CO2 Solubility mCO2 at Temperature T = 313.15 K in Aqueous Serinol Solutions of Amine-Molality m ............................................................. 240
Table 8.6. Experimental Values of CO2 Solubility mCO2 at Temperature T in Aqueous Serinol Solutions of Amine-Molality m = 4.704 mol·kg-1 ................................................ 243
Table 8.7. Experimental Values of CO2 Solubility mCO2 at a Temperature T in Aqueous AHPD + Pz Solutions of Amine-Molality m = (2.712 + 1.161) mol·kg-1 ......................... 243
Table 9.1. Amines studied in this work............................................................................. 253
Table 9.2. Degradation first-order rate constants. ............................................................. 258
Table 10.1. Membrane and module specifications. ........................................................... 275
Table 11.1. Flat membrane and module specifications. .................................................... 290
xvi
xvii
Index of Figures Figure 1.1. CO2 capture technology (IPCC, 2005). .............................................................. 2
Figure 1.2. Typical CO2 absorption process (Tobiesen and Svendsen, 2006) ...................... 3
Figure 1.3. Literature density values of 2-PE + H2O solutions and results calculated with Eq. (1.1). .............................................................................................................................. 20
Figure 1.4. Literature viscosity values of 2-PE + MEA + H2O solutions with a total amine content of 30 wt% and results calculated with Eq. (1.4). .................................................... 22
Figure 1.5. Henry’s law constant of CO2 in aqueous AMP + MEA mixtures for a total amine content of 30 wt%. .................................................................................................... 48
Figure 1.6. Henry’s law constant of CO2 in aqueous AMP + DEA mixtures for a total amine content of 30 wt%. .................................................................................................... 49
Figure 1.7. Gas diffusion in membrane contactor (Hoff et al., 2004) ................................. 73
Figure 1.8. Mass transfer in membrane contactor ............................................................... 74
Figure 1.9. Hollow fiber membrane contactors: a) parallel flow; b) cross flow provided by TNO-MEP; c) MC module commercialized by Membrana Co .......................................... 75
Figure 2.1. Schematic overall experimental flowsheet ..................................................... 101
Figure 2.2. 2 2
1/ 2CO CO/D H ratio for the absorption of CO2 in water as a function of
temperature. Dotted lines are for trend only. ..................................................................... 104
Figure 2.3. 2 2
1/ 2N O N O/D H ratio for N2O in aqueous AHPD solutions ................................... 105
Figure 2.4. Specific absorption rate as a function of amine concentration for 2COy = 0.8.108
Figure 2.5. Specific absorption rate as a function of CO2 partial pressure for an aqueous AHPD solution of 1.5 kmol m-3 ......................................................................................... 108
Figure 2.6. Concentration profile of amine in the liquid film: a) exit of the liquid; b) entry of the liquid. Conditions: T = 313.15 K,
2COy = 0.41. ........................................................ 113
Figure 2.7. Concentration profile of dissolved CO2 in the liquid film: a) exit of the liquid; b) entry of the liquid. Conditions: T = 313.15 K,
2COy = 0.41 ........................................... 114
Figure 2.8. Variation of kov with the amine concentration for AHPD, AEPD, AMPD and AMP at 303.15 K. .............................................................................................................. 116
Figure 2.9. Modified Brønsted plot for AHPD, AEPD, AMPD and AMP at 303.15 K. .. 117
xviii
Figure 3.1. Structure of PZ ................................................................................................ 125
Figure 3.2. 2 2
1/ 2N O N O/D H ratio for N2O absorption in aqueous PZ-AHPD solutions ........... 131
Figure 3.3. Specific absorption rate as a function of amines concentrations for yCO2 = 0.02 ............................................................................................................................................ 132
Figure 3.4. Arrhenius plot of the second-order rate constant k2,PZ as a function of temperature. ........................................................................................................................ 136
Figure 3.5. The overall pseudo-first-order rate constant as a function of PZ concentration. ............................................................................................................................................ 137
Figure 3.6. Enhancement effect of PZ in 1 kmol m-3 AHPD solutions ............................. 140
Figure 4.1. Schematic diagram of the solubility apparatus ............................................... 148
Figure 4.2. CO2 solubility in water: comparison with literature values. ........................... 157
Figure 4.3. CO2 solubility in Pz aqueous solution: comparison with literature values ( Pzm = 2.0 mol.kg-1). ...................................................................................................................... 159
Figure 4.4a. CO2 solubility in AHPD aqueous solution at 298.15 K ( AHPDm = 0.9172 mol.kg-1). ............................................................................................................................ 159
Figure 4.4b. CO2 solubility in AHPD aqueous solution at 323.15 K, comparison with Park et al. (2002a) ( AHPDm = 0.9172 mol.kg-1). ........................................................................... 160
Figure 4.5. CO2 solubility in Pz-AHPD aqueous solutions at 288.15 and 333.15 K. ....... 163
Figure 4.6. Predicted species distribution in the AHPD+CO2+H2O system at 298.15 K (AHPDm = 0.9172 mol.kg-1). ................................................................................................... 163
Figure 4.7. Predicted species distribution in the Pz-AHPD+CO2+H2O system at 298.15 K (AHPD = 1.0 kmol.m-3 and Pz = 0.3 kmol.m-3). ................................................................ 164
Figure 4.8a. CO2 solubility in aqueous solution of AHPD. Experimental results of this work, AHPDm = 4.0 mol.kg-1. ................................................................................................ 165
Figure 4.8b. CO2 solubility in aqueous solution of AHPD. Experimental results by Le Tourneux et al. (2008), different AHPD molalities ........................................................... 165
Figure 5.1. Comparison of various solubility data of CO2 (1) in piperazine (2) aqueous solutions of concentration m2/mol·kg-1 at temperature T/K ............................................... 180
Figure 5.2. Equilibrium pressure above aqueous solutions of CO2 (1) - piperazine (2) at concentration m2/mol·kg-1 and temperature T/K as a function of solution CO2 loading (α) ............................................................................................................................................ 183
xix
Figure 5.3. Species distribution in the aqueous CO2 (1) – Pz (2) system at 298.1 K (m2/mol·kg-1 = 1.00) as a function of solution CO2 loading .............................................. 185
Figure 5.4. Calculated activity coefficients in the aqueous CO2 (1) – Pz (2) system at 298.1 K (m2/mol·kg-1 = 1.00) as a function of solution CO2 loading .......................................... 185
Figure 6.1. a) Schematic diagram of the absorption flask and b) schematic diagram of the vapor-liquid equilibrium cell used for the regeneration. ................................................... 194
Figure 6.2. Optimal regeneration temperature determination (the curve shows the trend) ........................................................................................................................................... 196
Figure 6.3. Standard desorption curve for a 1 kmol.m-3 aqueous AHPD solution at 383.15 K ......................................................................................................................... 196
Figure 6.4. Comparison of desorption curves of MEA and Pz. ........................................ 199
Figure 6.5. Effect of Pz on desorption of AHPD aqueous solutions. ............................... 200
Figure 7.1. Breakthrough pressure apparatus ................................................................... 209
Figure 7.2. Influence of the carbon and hydrophilic numbers on surface tension of various aqueous solutions. .............................................................................................................. 216
Figure 7.3. SEM pictures of some tested membranes. ...................................................... 224
Figure 8.1. Structure of monoethanolamine (MEA), 2-amino-1,3-propanediol (Serinol) and 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD). ....................................................... 232
Figure 8.2. Densities of aqueous Serinol solutions as a function of amine-molality m and temperature T ..................................................................................................................... 237
Figure 8.3. Viscosities of aqueous Serinol solutions as a function of amine-molality m and temperature T ..................................................................................................................... 237
Figure 8.4. Surface tensions of aqueous Serinol solutions as a function of amine-molality m and temperature T .............................................................................................................. 239
Figure 8.5a. CO2 molality-based solubility in aqueous Serinol solutions at T = 313.15 K as a function of Serinol molality m ........................................................................................ 241
Figure 8.5b. CO2 loading-based solubility in aqueous Serinol solutions at T = 313.15 K as a function of Serinol molality m ........................................................................................ 242
Figure 8.6. CO2 solubility in an aqueous Serinol solution of m = 4.704 mol·kg-1 as a function of temperature T .................................................................................................. 244
Figure 8.7. Comparison of CO2 solubility data in Serinol (black symbols), AHPD + Pz (grey symbols) and MEA (white symbols, (Shen and Li, 1992)) solutions at temperature T = 313.15 K (circular symbols) and 373.15 K (square symbols). ....................................... 245
xx
Figure 9.1. Amine structures. ............................................................................................ 253
Figure 9.2. Experimental setup for degradation involving gas introduction. .................... 254
Figure 9.3. Amine degradation loss after 14 days (except for AMP). .............................. 257
Figure 9.4. Effect of process conditions on Pz degradation. Solid lines are calculated using Eq. (9.1) and constants from Table 9.2. ............................................................................. 258
Figure 9.5. Effect of process conditions on AMP degradation. Solid line is calculated using Eq. (9.1) and constant in Table 2, whereas dashed lines are for trend only. ...................... 259
Figure 10.1. Schematic diagram of CO2 (A) and amine (B) concentration profiles in membrane contactor. .......................................................................................................... 270
Figure 10.2. Experimental setup for CO2 absorption using the membrane contactor in counter-current flow circulation (the co-current flow is performed by switching the gas connexions in the contactor module). ................................................................................ 276
Figure 10.3. CO2 absorption flux as a function of liquid flow rate with a pure CO2 gas flow rate of 100 ml/min in counter-current mode. ..................................................................... 277
Figure 10.4. CO2 absorption flux as a function of the inlet CO2 volumetric percentage with a total gas flow of 100 ml/min and liquid flow rates of 30 ml/min for AHPD, AHPD + Pz and MEA solutions. ............................................................................................................ 278
Figure 10.5. CO2 removal efficiency for the aqueous AHPD + Pz solution (counter-current, total gas flow of 100 ml/min and liquid flow rate of 30 ml/min)....................................... 280
Figure 10.6. Variation of the membrane wetted pore fraction for data of Figure 10.3. .... 281
Figure 10.7. Variation of the wetted pore fraction for data of Figure 10.4. .................... 281
Figure 11.1. Experimental setup for CO2 absorption using the FSMC. ............................ 291
Figure 11.2. CO2 absorption flux in 3-FSMC (PTFE) as a function of liquid flow rate (pure CO2 gas flow rate of 100 ml/min in counter-current mode). ............................................. 293
Figure 11.3. CO2 absorption rate in a PTFE membrane FSMC as a function of AHPD + Pz solution flow rate with a pure CO2 gas flow rate of 100 ml/min in counter-current mode. ............................................................................................................................................ 294
Figure 11.4. CO2 absorption flux as a function of the gas inlet CO2 volumetric percentage for an AHPD + Pz absorbent flow rate of 20 ml/min using 3-FSMC (PTFE). .................. 295
Figure 11.5. CO2 removal percentage for Figure 11.4 counter-current data ..................... 296
Figure 11.6. Effect of membrane properties on CO2 flux in 2-FSMC as a function of liquid flow rate (pure CO2 gas flow rate of 100 ml/min in counter-current mode). ..................... 297
xxi
Nomenclature ai activity of species i (Chapters 4 and 5) ai, bi, ci, di, ei correlation regressed coefficients av gas-liquid contact area, m2/m3 Aφ
Debye-Hückel parameter for the osmotic coefficient Bi,j second virial coefficients between species i and j
ALC gas concentration in the liquid phase, kmol/m3: ( )A ,C z x represents the variation concentration in the film for all z
A_LC gas concentration in the liquid phase, kmol/m3: ( )A_LC z represents the variation concentration on the length of the contactor
A_GC gas concentration in the gas phase, kmol/m3
A_G,0C initial CO2 concentration in the gas phase, kmol/m3
A, iC gas concentration at the G/L interface, kmol/m3
BaseC concentration of one of possible bases in the liquid phase, kmol/m3
BLC amine concentration in the liquid phase, kmol/m3: ( )B ,C z x represents the variation concentration in the film for all z
B_LC amine concentration in the liquid phase, kmol/m3: ( )B_LC z represents the variation concentration on the length of the column
B,0 D,0,C C initial amine concentration in the solution, kmol/m3
B,exit D,exit,C C amine concentration in the solution at the liquid exit, kmol/m3
DLC second amine concentration in the liquid phase, kmol/m3: ( ),DC z x represents the variation concentration in the film for all z
D_LC second amine concentration in the liquid phase, kmol/m3: ( )D_LC z represents the variation concentration on the length of the column
Ci concentration of specie i, kmol/m3 Cj concentration of species j in solution, kmol/m3 Cp heat capacity, J/mol.K d, ds respectively, the diameter and the diameter including film thickness of
the wetted wall column, m D dielectric constant or relative permittivity of pure water dm diameter of hollow fiber membrane, m dH hydraulic diameter of the annulus in the wetted wall column contactor,
m jD diffusion coefficient of species j in solution, m2/s: j can represent gas or
amine Dj,α molecular diffusivity coefficient of species j in α phase ( ,gα = ), m2/s dp,max membrane maximale pore diameter, m
wD amine diffusion coefficient at infinite dilution state in solution, m2/s
xxii
e electronic charge E enhancement factor Ea activation energy, kJ/mol Einf, E∞ infinite enhancement factor Einf,j infinite enhancement factor for the amine j f function, as defined by Eq. (4.19) g gravitational acceleration, m/s2 h effective height of the wetted wall column, m H membrane length, m Ha Hatta number Ha,j Hatta number for the amine j
( )2 2, ,m sat
CO H O wH T P Henry’s law constant for the solubility of carbon dioxide in pure water on the molality scale
Hj Henry’s law constant of species j in solution; j can represent N2O or CO2 (component A), kPa.m3/kmol
Hsol enthalpy of solution, J/mol I ionic strength of solutions, kmol/m3 k Boltzmann’s constant, J/K k-1 reverse first order reaction rate constant, 1/s
2k second order forward reaction rate constant, m3/kmol.s
2,PZk second order forward reaction rate constant for Pz, m3/kmol.s
AMk reaction rate constant as defined in Eq. (2.8), m3/kmol.s kapp pseudo-first-order apparent rate constant, 1/s kb second order reaction rate constant for base b, m3/kmol.s
2
*H Ok rate constant for CO2-H2O reaction, m3/kmol.s
2H Ok reaction rate constant as defined in Eq. (2.10), m3/kmol.s k
, kL liquid-phase mass transfer coefficient, m/s kg gas-phase mass transfer coefficient, kmol/s.m.kPa Kg overall mass transfert coefficient based on the gas phase, kmol/s.m.kPa
*OH
k − rate constant for CO2-OH- reaction, m3/kmol.s
-OHk reaction rate constant as defined in Eq. (2.9), m3/kmol.s kov pseudo-first-order overall rate constant, 1/s Kp protonation constant for AHPD, kmol/m3 KR equilibrium constant for the chemical reaction R, expressed on the
molality scale Kw dissociation constant for water, kmol2/m6 L liquid flow rate, m3/s m distribution coefficient mi true molality of species i in solution, mol/kg
im~ stoichiometric molality of component i, mol/kg M molarity, kmol/m3
xxiii
Mw molar mass of water, kg/mol n number of mole, mol NA Avogadro’s number, 1/mol Nj specific absorption rate of gas j, kmol/m2.s
TjN total absorption rate of gas j, kmol/s
O.A.D.% overall average deviation percentage P pressure, kPa Pj, Pj, i respectively partial pressure of gas j in bulk phase and at interface, kPa
satwP saturated vapour pressure of water, kPa
q0, q1 coefficients, as defined in Eqs. (4.23) or (5.17) Qg gas flow rate, m3/s r radial position within porous membrane and liquid film, m rA reaction rate of CO2 in the liquid phase, kmol/m3.s
2CO jr − reaction rate of CO2 with amine j, kmol/m3.s
ir reaction rate, kmol/m3.s R universal gas constant, J/mol.K R2 determination coefficient Rj reaction rate of the component j, kmol/m3.s
fR radius of liquid film in hollow fiber membrane, m gmR radius of gas-liquid interface in hollow fiber membrane, m inmR inner radius of hollow fiber membrane, m outmR outer radius of hollow fiber membrane, m
Re Reynolds number S transverse section of the column apparatus, m2 Sc Schmidt number Sh Sherwood number t regeneration time, min tc contact time, s T absolute temperature, K ug superficial gas velocity, m/s uℓ superficial liquid velocity, m/s V (partial) molar volume, m3/kmol w mass fraction x radial coordinate, m y vapour phase mole fraction z axial coordinate, m zi charge of ion i
Greek letters αi CO2 loading in solution, i, if present, can represent “R”, the rich solution
and “L” the lean solution. β amine bulkiness, m3/kmol (chapter 2) or exponent in Stokes-Einstein
relation
xxiv
( ) ( )0 1,ij ijβ β binary interaction parameters between species i and j in Pitzer’s equation
∆ uncertainty of specified value Lδ liquid film, m
0ε permittivity of free space, F/m η regeneration efficiency, as defined by Eq. (6.1)
,miγ ∗
activity coefficient of component i normalized to infinite dilution, on molality scale
( )ij Iλ second virial coefficient in Pitzer’s equation µ liquid viscosity, kg/m.s ν ,ν j stoichiometric coefficient
,i Rν stoichiometric coefficient of component i in the reaction R
iϕ fugacity coefficient of component i, kPa ρ density, kg/m3 σ surface tension, mN/m
ijkτ ternary interaction parameter in Pitzer’s equation θ contact angle, ° Subscripts and superscripts
A gas Am amine B amine eff effective in inside, inlet g gas phase liquid phase f liquid film
m membrane out outer, outside R reaction R sat saturation m molality w water ∞ infinite dilution in pure water Chemical name
2-PE 2-piperidineethanol
AEPD 2-amino-2-ethyl-1,3-propanediol
xxv
AHPD 2-amino-2-hydroxymethyl-1,3-propanediol
AM ammonia
AMP 2-amino-2-methyl-1-propanol
AMPD 2-amino-2-methyl-1,3-propanediol
AP 3-amino-1-propanol
CO carbon monoxide
CO2 carbon dioxide
DEA diethanolamine
DETA diethylenetriamine
DGA diglycolamine
DIBA diisobutylamine
DIPA diisopropanolamine
EAE 2-(ethylamino)ethanol
EDA ethylenediamine
EMEA 2-(ethylamino)ethanol
H2 hydrogen
HMDA hexamethylenediamine
MAE 2-(methylamino)ethanol
MDEA N-methyldiethanolamine
MEA monoethanolamine
MIPA 1-amino-2-propanol
MMEA 2-(methylamino)ethanol
N2 nitrogen
N2O nitrous oxide
NMP N-methylpyrrolidone
PP polypropylene
Pz or PZ piperazine
Serinol 2-Amino-1,3-propanediol
TBA tert-butylamine
TBAE 2-(tert-butylamino)ethanol
xxvi
THAM 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD)
TMS sulfolane (tetramethylene sulfone)
Technical acronym
CCS carbon capture and storage
FSMC flat sheet membrane contactor
HFMC hollow fiber membrane contactor
MC membrane contactor
SHA Sterically hindered alkanolamine
xxvii
Acknowledgement
I wish to express my sincere gratitude and admiration to my thesis director, Pr. Maria-
Cornelia Iliuta, for her guidance, confidence and encouragement over the entire course of
this Ph.D project. Without her help and support, this work would not have been possible.
I would like to thanks Dr. Ion Iliuta for his advices and modeling contribution to some
articles.
The financial support provided by the Natural Sciences and Engineering Research
Council of Canada (NSERC), FQRNT Centre in Green Chemistry and Catalysis (CGCC),
Rio Tinto Alcan (Canada) and Centre de Recherche en Catalyse et Chimie Verte (C3V,
Laval University) is gratefully acknowledged.
I also want to acknowledge the kind contribution of the following companies in
supplying membranes: Markel Corporation, Donaldson, AY Tech LLC, Membrana and
Celgard.
xxviii
xxix
Preface
This dissertation contains 12 chapters.
Chapter 1 (Introduction) starts with the importance of CO2 removal from gaseous
emissions and methods for its mitigation. A review on different aspects concerning several
binary and multi component systems CO2 - sterically hindered amines based absorbents and
CO2 capture in amine based absorbents using membrane contactors is then performed. The
Introduction chapter ends with Conclusions and Objectives of this work. Chapter 2 contains
the kinetic study of the reaction between CO2 and a sterically hindered alkanolamine, 2-
amino-2-hydroxymethyl-1,3-propanediol (AHPD). In Chapter 3, the influence of Pz
(piperazine) addition into AHPD solutions, as reaction accelerator, is discussed. Chapters 4
and 5 studied the CO2 absorption capacity of aqueous AHPD +Pz and Pz solutions,
respectively. Chapter 6 presents a comparison of the regeneration capability of different
single sterically hindered alkanolamines (SHA: AMP (2-amino-2-methyl-1-propanol),
AEPD (2-amino-2-ethyl-1,3-propanediol), AMPD (2-amino-2-methyl-1,3-propanediol),
AHPD) and Pz-activated aqueous solutions, with that of single monoethanolamine (MEA)
aqueous solution. Based on wetting-related properties like liquid surface tension, contact
angle, membrane breakthrough pressure and chemical stability, a thorough analysis of these
properties is performed in Chapter 7 on different potential membrane/liquid combinations
in order to develop an appropriate way to select the best conditions to elude the unwanted
wetting phenomenon in membrane contactors (MC). Following a new classification method
for the estimation of surface tension of aqueous amine solutions proposed in Chapter 7, the
aim of the work presented in Chapter 8 is to investigate the potential of Serinol (2-Amino-
1,3-propanediol) solutions as an efficient CO2 absorbent to be used in MC. In Chapter 9,
stability to thermal and oxidative degradation of aqueous AHPD, MEA, AMP, Pz and
Serinol solutions is investigated under various experimental conditions. Chapters 10 and 11
include the application of aqueous AHPD + Pz solution for CO2 removal in hollow fiber
and flat sheet MC, respectively. The general conclusions and suggestions for future work
are given in chapter 12.
xxx
This thesis was prepared based on the following published or submitted papers in/to
scientific journals:
1 - Bougie, F., Iliuta, M.C., Sterically hindered alkanolamines based absorbents for removal of CO2 from gas streams. Invited Review. J. Chem. Eng. Data 2012, 57, 635–669 (Chapter 1/1.2).
2 - Bougie, F., Iliuta, M.C., Kinetics of absorption of carbon dioxide into aqueous solutions of 2-amino-2-hydroxymethyl-1,3-propanediol. Chem. Eng. Sci. 2009, 64, 153-162 (Chapter 2).
3 - Bougie, F., Lauzon-Gauthier, J.1, Iliuta, M.C., Acceleration of the reaction of carbon dioxide into aqueous 2-amino-2-hydroxymethyl-1,3-propanediol solutions by piperazine addition. Chem. Eng. Sci. 2009, 64, 2011-2019 (Chapter 3).
4 - Bougie, F., Iliuta, M.C., CO2 absorption into mixed aqueous solutions of 2-amino-2-hydroxymethyl-1,3-propanediol and piperazine. I&EC Res. 2010, 49, 1150–1159 (Chapter 4).
5 - Bougie, F., Iliuta, M.C., CO2 Absorption in Aqueous Piperazine Solutions: Experimental Study and Modeling. J. Chem. Eng. Data 2011, 56, 1547-1554 (Chapter 5).
6 - Bougie, F., Iliuta, M.C., Analysis of regeneration of sterically hindered alkanolamines aqueous solutions with and without activator. Chem. Eng. Sci. 2010, 65, 4746–4750 (Chapter 6).
7 - Bougie, F., Iliuta, M.C., Analysis of Laplace-Young equation parameters and their influence on efficient CO2 capture in membrane contactors. Sep. Purif. Technol. 2013, 118, 806–815 (Chapter 7).
8 - Bougie, F., Iliuta, M.C., Solubility of CO2 in and Density, Viscosity and Surface Tension of Aqueous 2-Amino-1,3-propanediol (Serinol) Solutions. J. Chem. Eng. Data 2014, 59, 355–361 (Chapter 8).
9 - Bougie, F., Iliuta, M.C., Stability of aqueous amine solutions to thermal and oxidative degradation in the absence and the presence of CO2. Submitted (Chapter 9).
1 Undergraduate student
xxxi
10 - Bougie, F., Iliuta, I., Iliuta, M.C., Absorption of CO2 into Pz-activated AHPD aqueous solutions in PTFE hollow fiber membrane contactors: Experimental and modeling study. Submitted (Chapter 10).
11 - Bougie, F., Iliuta, M.C., Flat sheet membrane contactors (FSMC) for CO2 separation in aqueous amine solutions. Submitted (Chapter 11).
The author has main contribution in all stages of the work presented in papers 1 to 11,
including planning and performing experiments, as well as writing the papers by taking into
account the supervisor’s comments, except for the modeling part of the paper being the
object of Chapter 10.
The results of the present thesis were also presented in the following academic national
and international conferences:
1 - Bougie, F., Iliuta, M.C., Thermodynamic study of absorption capacity of carbon dioxide into mixed aqueous solutions based on AHPD. 8th World Congress of Chem. Eng. (WCCE8), Montréal, August 23-27, 2009 (oral presentation).
2 - Bougie, F., Iliuta, M.C., Aqueous solutions of sterically hindered amines for the removal of CO2 from gas streams using membrane contactors. 8th World Congress of Chem. Eng. (WCCE8), GLS 9, Montréal, August 23-27, 2009 (poster).
3 - Bougie, F., Iliuta, M.C., Energy friendly absorbents for CO2 capture. 3th IUPAC Int. Conf. Green Chem., (ICGC 2010), Ottawa, August 15-20, 2010. (poster).
4 - Bougie, F., Lalonde, J.1, Iliuta, M.C., Characterisation of polymeric flat membranes and compatibility with various aqueous amine solutions used for CO2 capture. 61st Canadian Chem. Eng. Conf., London, October 23-26, 2011 (oral presentation).
5 - Bougie, F., Iliuta, M.C., Activator effect on CO2 capture by AHPD aqueous solutions in PTFE hollow fiber membrane contactor - experimental and modeling. 62nd Canadian Chem. Eng. Conf., Vancouver, October 14-17, 2012 (poster).
6 - Bougie, F., Iliuta, M.C., CO2 capture by aqueous amine solutions in membrane contactors – Analysis of amine solutions and membrane contactor modifications on absorption performance, 63th Canadian Chem. Eng. Conf., Fredericton, October 20-23, 2013 (oral presentation).
1 Undergraduate student
xxxii
7 - Bougie, F., Moreau, V.1, Iliuta, M.C., Thermal and oxidative degradation of AHPD solutions for CO2 capture, 63th Canadian Chem. Eng. Conf., Fredericton, October 20-23, 2013 (poster).
The results of the present thesis have also been presented in annual meetings of several
Research Centers: CCVC (FRQNT Centre en chimie verte et catalyse), CQMF (FRQNT
Centre québécois sur les matériaux fonctionnels) and CERPIC (Centre en catalyse et
chimie verte, Université Laval)
1 - Bougie, F., Lalonde, J.1, Iliuta, M.C. Study of the breakthrough pressure for amine solutions in several polymeric porous flat membranes. 2ème conférence annuelle CCVC, Montréal, décembre 2010. (poster).
2 - Bougie, F., Lalonde, J.1, Iliuta, M.C. Study of the breakthrough pressure for amine solutions in several polymeric porous flat membranes. 3ème colloque annuel CQMF, Sherbrooke, octobre 2010. (poster).
3 - Bougie, F., Iliuta, M.C. Capture et valorisation du CO2. Présentation du groupe de recherche de Maria Iliuta. 4ème colloque annuel CQMF, Québec, octobre 2011. (poster).
4 - Bougie, F., Iliuta, M.C. Absorption du CO2 dans les contacteurs à membranes. Rencontre annuelle du CERPIC, Québec, mars 2011. (Oral presentation).
5 - Bougie, F., Iliuta, M.C. Capture du CO2. Rencontre annuelle du CERPIC, Québec, mai 2012. (oral presentation).
6 - Bougie, F., Lalonde, J.1, Iliuta, M.C. Study of the breakthrough pressure for liquid absorbents in several polymeric porous flat membranes. 4ème Conférence annuelle CCVC, Montréal, 10 Mai 2012. (poster).
1 Undergraduate student
1
Chapter 1. Introduction
1.1. Background
Carbon dioxide (CO2) is considered as one of the principal greenhouse gases. Due to
the dependence of world economy on fossil fuels used for generating energy,
approximately one third of all anthropogenic CO2 emissions come from fossil fuels such as
coal, oil and natural gas (23 Gton CO2/year (IPCC, 2005)). A variety of industrial processes
also emit large amounts of CO2 from each plant, for example oil refineries, cement works,
and iron production. There is growing political and public concern supported by consensus
among the scientific community that global emissions growth will soon drive atmospheric
CO2 concentrations to levels never seen, bringing a growing risk of fast climate change.
The Canadian Environmental Protection Act (CEPA, 2005) is the legislative authority in
Canada that pushes Canadian companies to reduce their greenhouse gas production. These
atmospheric emissions could be reduced substantially by using non-carbon energy
ressources like renewable ones (wind, water and solar energy). However, until the complete
change to new sources of energy which can take several years, the necessary energy will
still be obtained from fossil fuels. In this context, capturing and storing (CCS) the CO2
from different flow emisisons will give the opportunity to use the existing fossil fuels while
stabilizing the CO2 concentrations in the atmosphere. In the global CCS process, the CO2
capture represents the major cost (Tobiesen and Svendsen, 2006). The aim of CO2 capture
is its separation from different gaseous emission sources to get it as a concentrated stream,
ready for sequestration or further use (like conversion into valuable products). The CO2
capture can be performed by three technological concepts: post-combustion, oxy-
combusion and pre-combustion (Figure 1.1) (IPCC, 2005). Depending on the configuration,
it can be implemented in new plants or it may be retrofitted to existing plants.
The post-combustion capture is a well-known technology which involves the capture
of CO2 from flue gases after a fossil fuel has been burned (carbon removal after fuel
combustion). Due to the low CO2 content of the exhaust gas (3-15 vol% at near
atmospheric pressure and N2 the main constituent), the most effective method of CO2
2
capture is by chemical reaction with highly reactive components (amines), which is also
widely used for separating CO2 and other acid gases from natural gas (natural gas
sweetening). For oxy-combustion capture, the fossil fuel is burned in pure oxygen instead
of air, so that the resulting exhaust contains mainly CO2 and water vapor, being easily
separated by condensation. High CO2 concentrations can then be obtained in the exhaust
gas (greater than 80% by volume). The pre-combustion capture involves the fossil fuel
gasification, instead of direct combustion (carbon removal before fuel combustion). In the
presence of steam and air (oxygen), the fuel is transformed into synthesis gas (mainly
consisting in CO and H2). If H2 production is the main objective, additional hydrogen can
be obtained toghether with CO2 in the presence of an exces of steam. The CO2 separation
can lead to a concentrated flow (up to 60 vol% on a dry basis) and at high pressure.
Figure 1.1. CO2 capture technologies (IPCC, 2005).
CO2 separation can be performed by applying several methods: absorption in solutions
(use of selective liquids to separate gases), adsorption on solids (use of selective solid
materials to separate gases), membranes (use of selective barriers, porous or nonporous
3
materials, to separate gases) and cryogenic distillation (use of the difference in boiling
points to separate gases) (Abu-Khader, 2006). The choice of a specific method highly
depends on different parameters like gas concentration and composition (presence of other
components), temperature and pressure.
The present thesis concerns hybrid systems (membrane contactors) that combine
chemical absorption with membrane technology, as alternatives to traditional
absorption columns. The literature review will therefore be limited to this research
area.
The absorption is a commun process in chemical engineering and it is largely applied
in the industrial acid gas treatment (Kohl and Nielsen, 1997). In the absorption process
(Figure 1.2), the gas mixture is put in contact with the absorption solution in an absorber
(gas-liquid contactor) where about 85-90% of CO2 is removed, thus leading to a rich-
loaded CO2 solution (high CO2 content). This solution is then regenerated by heating in a
stripper to release the CO2 (concentrated stream ready to be compressed for further use) and
to produce a lean-loaded CO2 solution (low CO2 content) that is recycled back to the
absorber.
Figure 1.2. Typical CO2 absorption process (Tobiesen and Svendsen, 2006)
4
The choice of the absorbent is based on several important parameters, such as
absorption capacity, absorption kinetics, regeneration facility, and corrosiveness. Aqueous
solutions of a wide variety of amines can be used, such as monoethanolamine (MEA, a
primary amine), diethanolamine (DEA, a secondary amine), diisopropanolamine (DIPA, a
secondary amine), N-methyldiethanolamine (MDEA, a tertiary amine) and 2-amino-2-
methyl-1-propanol (AMP, a sterically hindered amine (SHA)) (Kohl and Nielsen, 1997).
The use of blended alkanolamine solutions has recently become very attractive because of
the combination of each amine advantages: a fast reactivity from a primary or secondary
alkanolamine (e.g. MEA, DEA) coupled with the high absorption capacity and low solvent
regeneration cost from a tertiary or sterically hindered alkanolamine (e.g. MDEA, AMP).
Other potential solutions contain piperazine (Pz) which is not an alkanolamine but has
proven to have a higher absorption rate than MEA (Derks et al., 2006). Pz is usually used in
a mixture with other amines presenting lower kinetics.
This thesis concerns the application of sterically hindered alkanolamine based
solutions for CO2 separation. The literature highlighting different aspects concerning
several binary and multi component sterically hindered alkanolamine based absorbents is
reviewed in the section 1.2.
Membrane technology is a powerful tool in developing new industrial processes, with
the advantage of reduced equipment size, energy use and waste generation (Bernardo et al.,
2009). Membrane gas separation is a pressure-driven process where the membrane allows
the selective permeation of a gas component through it by diffusion. The efficiency of gas
separation by membranes is dependent on the membrane material (permeability and
selectivity), membrane structure and thickness, membrane configuration (hollow fiber,
spiral, tubular, and flat) and the membrane module and process design. Several types of
membrane materials (polymeric, metallic, and ceramic) can be used in gas separations. In
the case of CO2 capture from a diluted flue gas and low pressure, the low CO2 partial
pressure provides low driving force for the separation, which results in lower efficiency
compared to traditional chemical absorption (lower percentage of CO2 removed and higher
energy penalty) (Feron et al., 1992). An increase in efficiency could be obtained for more
5
selective membranes, but the increase in selectivity will lead to a decrease in permeability.
An interesting option is a combination of the absorption process with the membrane
technology in a hybrid system, called membrane contactor (membrane gas absorption
system). The absorbent (chemical) assures a very good selectivity, while a highly porous
membrane assures a good permability (contact between gas and liquid).
This thesis concerns the application of sterically hindered alkanolamine based
absorbents for CO2 separation in membrane contactors; the literature highlighting different
aspects concerning CO2 capture in amine based absorbents using membrane contactors is
reviewed in the section 1.3.
6
1.2. Sterically hindered amines based absorbents for the removal of CO2
from gas streams
Résumé
La séparation du CO2 de mélanges gazeux par un procédé d’absorption possède de nombreuses applications, particulièrement dans l’industrie chimique et pétrolière mais aussi pour la protection de l’environnement. Le choix d’une amine (seule ou combinée avec d’autres amines en solution) pour la capture du CO2 est principalement basé sur sa capacité d’absorption, sa vitesse de réaction, sa capacité de régénération et sa résistance à la dégradation. Beaucoup de ces propriétés sont supérieures pour les amines à encombrement stérique comparativement aux amines conventionnelles. L’objectif de cette revue littéraire est donc de mettre à jour et d’analyser les données disponibles de différentes solutions à base d’amines à encombrement stérique utilisées pour la capture du CO2. Ces données sont essentielles pour le design et l’opération des équipements liés à l’absorption et concernent principalement : la densité, la viscosité, la pression de vapeur, la capacité calorifique, la chaleur d’absorption, les coefficients de diffusion du CO2 et de l’amine en solution, la capacité d’absorption, les constantes cinétiques et la capacité de régénération.
Abstract
Gas absorption process for CO2 separation from gas streams is of high interest in various applications in chemical, oil and gas industries, as well as in environmental protection. The choice of a certain amine (single or blended amine) for CO2 capture is mainly based on the absorption capacity, reaction kinetics and regenerative potential and facility. The application of sterically hindered amines in gas-treating technology offers absorption capacity, absorption rate, and degradation resistance advantages over conventional amines for CO2 removal from gases. The aim of this review is to bring an update of different aspects concerning several CO2 - SHA based binary and multi component absorbents, essential for the design and operation of absorption equipments (physical properties like density, viscosity, vapour pressure, heat capacity and heat of absorption, CO2 and amine diffusivity, CO2 absorption capacity and kinetics, regeneration capability).
7
1.2.1. Introduction
It is well known that approximately one third of all anthropogenic CO2 emissions come
from fossil fuels such as coal and oil used for generating energy. In addition, different
industrial processes emit large amounts of CO2 from each plant, as oil refineries, cement
works, and iron production (IPCC, 2005). A typical CO2 generation rate from power plant
is 400 × 103 kg·h-1 with stack gas flow rates of 484 m3·s-1 and approximately 13% CO2
(Rangwala, 1996). There is growing political and public concern supported by consensus
among the scientific community that global emissions growth will soon drive atmospheric
CO2 concentrations to very high levels, bringing a growing risk of fast climate change. In
Canada, the Canadian Environmental Protection Act (CEPA, 2005) is the legislative
authority that pushes the companies to reduce their greenhouse gas production. The CO2
emissions could be reduced substantially by capturing and storing the CO2.
Industrially often used alkanolamines are MEA, DEA, MDEA, AMP (Kohl and
Nielsen, 1997). The choice of a certain amine (single or blended amine) is mainly based on
the absorption capacity, reaction kinetics and regenerative potential and facility. The key
advantage of the primary and secondary alkanolamines such as MEA and DEA is their fast
reactivity due to the formation of stable carbamates. Conversely, this will lead to very high
solvent regeneration cost. On the absorption capacity side, they have the drawback of a
relatively low CO2 loading (limited to 0.5 mol CO2·mol amine-1). Tertiary alkanolamines,
like MDEA, have a very low reactivity with respect to CO2, due to the exclusive formation
of bicarbonates by CO2 hydrolysis. However, this will lead to a very low solvent
regeneration cost. Another advantage of these amines is the high CO2 theoretical loading
capacity of 1 mol of CO2·mol of amine-1. The application of SHA, e.g., AMP in gas-
treating technology offers absorption capacity, absorption rate, selectivity and degradation
resistance advantages over conventional amines for CO2 removal from gases (Goldstein et
al., 1984; Sartori and Savage, 1983; Say et al., 1984). Due to the hindrance of the bulky
group adjacent to the amino group, SHA form unstable carbamates. Hydrolysis of the
voluminous carbamates leads to a preferential bicarbonate formation process, resulting in
the theoretical loading capacity up to 1.0. Reaction kinetics significantly higher than those
related to tertiary amines, coupled with a low solvent regeneration cost offer to SHA
8
important industrial advantages. The use of blended alkanolamines solutions has also
become very attractive because of the combination of each amine advantages: a fast
reactivity from a primary or secondary alkanolamine (e.g. MEA, DEA) coupled with the
high absorption capacity and low solvent regeneration cost from a tertiary or sterically
hindered alkanolamine (e.g. MDEA, AMP).
The aim of this review is to bring an update of different aspects concerning several
CO2 - SHA based binary and multi component absorbents, essential for the design and
operation of absorption equipments (physical properties like density, viscosity, vapour
pressure, heat capacity and heat of absorption, CO2 and amine diffusivity, CO2 absorption
capacity and kinetics, regeneration capability).
1.2.2. Structure and properties of SHA
1.2.2.1. Structure of SHA
A hindered amine was originally defined by Sartori and Savage (1983) as an amine
belonging to one of the following categories: (i) a primary amine in which the amino group
is attached to a tertiary carbon; (ii) a secondary amine in which the amino group is attached
to at least one secondary or tertiary carbon.
An example of SHA, the well-known AMP is the hindered form of MEA obtained by
substituting two hydrogen atoms attached to the alpha carbon atom to the amino group in
MEA by two methyl groups. These substitutions influence significantly amine properties
and absorption capacity (Yoon and Lee, 2003). All sterically hindered amines found in the
literature that were linked to CO2 absorption (solubility, kinetics) or for which any other
properties necessary to operate a gas-liquid contactor are important (density, viscosity,
superficial tension, vapour pressure) are given in Table 1.1.
1.2.2.2. Physical properties of single and mixed SHA aqueous mixtures
Physical properties of amine solutions, as it will be explained in the next sections, are
necessary to design properly CO2 absorption and regeneration processes. It should be
mentioned here that without indication, all data presented in the next sections are for fresh
(unloaded) solutions. It seems however that CO2 loading could have a significant effect on
parameter’s values for conventional amines (MEA, DEA and MDEA), as it can be
9
demonstrated in Weiland et al. (1998). Unfortunately, except for some studies concerning
the heat of absorption and the vapour pressure, information concerning the loading effect
on SHA solution properties is extremely scarce and future research on the topic would be
very welcomed.
1.2.2.2.1. Density and Viscosity
Knowledge of physical properties like density and viscosity of solutions is
necessary for the operation of process equipments such as pumps and heat exchangers as
well as for the design of gas-liquid contactors. In addition, these data are useful for
estimating the liquid diffusivity and reaction rate constant, for example when a wetted-wall
column is used for kinetic studies. Solution density and viscosity are also important in the
mass transfer rate modeling of absorbers and regenerators because these properties affect
the liquid film coefficient for mass transfer, kL. Viscosity was also found to significantly
affect membrane contactor performance as mentioned by Lin et al. (2008).
Tables A.1 and A.2 report, respectively, all density information found in the
literature concerning AMP and the various SHA (other than AMP). In the same way,
Tables A.3 and A.4 concern viscosity data. AMP is the most studied SHA and this is
reflected by the large amount of reported density and viscosity values. More than 30
articles giving densities and/or viscosities were found in the open literature concerning this
alkanolamine. Therefore, AMP based systems will be discussed in a separate section. For
SHA solutions under a temperature range related to CO2 capture and regeneration, it was
found that the values of density and viscosity data are almost always in the range of 0.85-
1.11 g·cm-3 and 0.40-8.0 mPa·s (total amine concentration less than 40 wt%) respectively.
10
Table 1.1. Structure of several sterically hindered amines
Acronym Name CAS number Structure M /g·mol-1
2-PE 2-piperidineethanol 1484-84-0 129.20
2-PM 2-piperidinemethanol 3433-37-2 115.17
AEPD 2-amino-2-ethyl-1,3-propanediol 115-70-8 119.16
AHPD 2-amino-2-hydroxymethyl-1,3-propanediol 77-86-1 121.14
AMP 2-amino-2-methyl-1-propanol 124-68-5 89.14
AMPA 2-amino-2-methylpropionic acid 62-57-7 103.12
AMPD 2-amino-2-methyl-1,3-propanediol 115-69-5 105.14
APPA 2-amino-2-phenylpropionic acid 565-07-1 165.19
DIPA diisopropylamine 108-18-9 101.19
MDA 1,8-p-menthanediamine 80-52-4 170.30
PA pipecolinic acid 4043-87-2 129.16
TBA tert-butylamine 75-64-9 73.14
TBAE 2-(tert-butylamino)ethanol 4620-70-6 117.19
11
1.2.2.2.1.1. AMP systems
1.2.2.2.1.1.1. Pure and binary systems: AMP and AMP + H2O
Li and Lie (1994) reported densities and viscosities of pure AMP from 303 to 353 K in
order to correlate tertiary systems containing AMP by a Redlich-Kister equation for density
and a Grunberg and Nissan equation for viscosity. Kundu et al. (2003) does not report
experimental data but derived an empirical expression to calculate pure AMP density at
293-353 K. Álvarez et al. (2006) measured densities as well as kinematic viscosities of pure
AMP at temperature from 298.15 to 323.15 K. Pure AMP data were also reported along
with the aqueous binary data, as it will be mentioned further.
Yih and Shen (1988) were among the first to report some density and viscosity data for
the aqueous binary system, being necessary for kinetic studies using a wetted-wall column.
Amine concentration was varied between 0.258 and 3.0 kmol·m-1 (2 to 27 wt%) and the
temperature was kept at 313 K. Bosch et al. (1990) reported later viscosity of AMP aqueous
solutions of concentrations between 0.258 and 2.484 kmol·m-3 (2 to 22 wt%) at 298 K. Xu
et al. (1991) measured densities and viscosities over a large temperature and concentration
range (293-363 K and 9.05-100 wt%). Data were found in good agreement with those by
Yih and Shen (1988). However, at around 298 K and 18 wt%, viscosities differed from
those of Bosch et al. (1990). Littel et al. (1992) presented polynomial equations to calculate
density and viscosity values at 303 K and for concentrations up to 5.009 kmol·m-3 (45
wt%) and 3.979 kmol·m-3 (35.5 wt%), respectively. However, these two correlations are
not very useful as they are limited to one temperature only and that they require
concentration expressed in molarity instead of molality or mass fraction. Saha et al. (1993)
measured viscosity values at temperatures between 294 and 318 K and for AMP
concentration of 0.5-2.0 kmol·m-3 (4.5 to 18 wt%). Density values were unfortunately only
graphically represented over the same concentration range and for temperatures between
288 and 313 K. Zhang et al. (2002) measured densities for aqueous AMP solutions (293.15-
353.15 K) and pure AMP (303.15-353.15 K). All reported densities for the aqueous
solutions are relative to the density of pure water at the same temperature. The work by
Chan et al. (2002) represents one of those reporting the most density values for the aqueous
binary system over a large temperature and concentration range (298-353 K and 4-100
12
wt%). Data of that work were found to be excellent agreement with those of Zhang et al.
(2002) and Aguila-Hernandez et al. (2001). Henni et al. (2003) reported density and
viscosity of aqueous solutions at six temperatures in the range 298 to 343 K and over a
wide concentration range (21-100 wt%). Pure AMP densities were found to be in excellent
agreement, but consistently higher, than those by Li and Lie (1994), Zhang et al. (2002),
and Aguila-Hernandez et al. (2001). On average, the reported experimental values were
0.17% higher than those of Li and Lie (1994) , so well below their reported accuracy of
0.5% and 0.24% higher than those of Aguila-Hernandez et al. (2001). The only data
available at high pressure were given for AMP densities at 298.32 K and concentrations of
15 and 30 wt% (Arcis et al., 2007).
1.2.2.2.1.1.2. Tertiary and other systems AMP + Amine(s) + H2O
AMP + MEA + H2O
The aqueous system AMP + MEA has been widely studied in the literature. Density
and viscosity data for this system were reported mainly by Lie and Lie (1994), Chenlo et al.
(2001) and Mandal et al. (2003b), covering a wide range of temperatures and
concentrations. Data reported by Li and Lie (1994) for density and viscosity from 303 to
353 K and concentrations between 20 and 30 wt% were correlated by a Redlich-Kister
equation for the density and a Grunberg and Nissan equation for the viscosity. Chenlo et al.
(2001) measured kinematic viscosities at various concentrations from 0.25 to 2.0 mol·kg-1
and temperatures from 293.1 K to 323.1 K but dynamic viscosity values are not available as
no density data are given for the studied concentrations and temperature. Densities and
viscosities measured by Mandal et al. (2003b) at 293-323 K for total amine concentration
of 30 wt% were found in good agreement with previous data. For 30 wt% AMP and 24.0
wt% AMP + 6.0 wt% MEA blend, over the temperature range 303 to 323 K, densities
showed 0.04% and 0.05% deviations, respectively, while viscosities showed 3.02% and
3.08% deviations, respectively, from the experimental data of Li and Lie (1994). In
addition to these three works, some other publications were found reporting density and
viscosity values over limited range of temperatures and concentrations. Hsu and Li (1997a,
b) reported densities and viscosities of aqueous mixtures of AMP + MEA over a
temperature range of 303-353 K and for a 10/10 wt% amine blend. Data were correlated
13
together with those by Li and Lie (1994) using a Redlich-Kister equation for the excess
volume and viscosity deviation. Xiao et al. (2000) measured density and viscosity at 303
and 313 K for solutions containing 1.5 or 1.7 kmol·m-3 (13.5 or 15.3 wt%) AMP with small
additions of MEA (0.1-0.4 kmol·m-3; 0.6-2.5 wt%). Mandal and Bandyopadhyay (2006)
gave density and viscosity at 313 K for various solutions of 30 wt% AMP, 28.5 wt% AMP
+ 1.5wt% MEA, 27 wt% AMP + 3wt% MEA and 25.5 wt% AMP + 4.5 wt% MEA. Values
were found to be in good agreement with those of Li and Lie (1994) and of Mandal et al.
(2003b).
AMP + DEA + H2O
The system AMP + DEA + H2O has also been widely studied in the literature. Density
and viscosity data for this system were mainly reported by Hsu and Li (1997a, b), Aguila-
Hernández et al. (2001) and Mandal et al. (2003b), covering a wide range of temperatures
and concentrations. Hsu and Li (1997a, b) reported densities and viscosities at 303-353 K
and total amine concentration of 30 wt% (6/24, 12/18, 18/12 and 24/6 AMP/DEA wt%) and
20 wt% (5/15, 10/10 and 15/5 AMP/DEA wt%). At constant temperature, the increase of
AMP concentration leads to the decrease in density and the increase in viscosity. Aguila-
Hernández et al. (2001) measured density at 313.15, 323.15, and 333.15 K and the total
amine concentration was in the range of 30-95 wt%. The correlation made by Hsu and Li
(1997a) applied to data of Aguila-Hernández et al. (2001) were found to represent them
with good agreement. Data by Mandal et al. (2003b) given at 293-323 K and total amine
concentration of 30 wt% were in good agreement with previous data: 0.19% and 3.12%
deviations, respectively, from experimental density and viscosity data of Hsu and Li
(1997a, b) for the system 24.0 wt% AMP + 6.0 wt% DEA, over the temperature range 303
to 323 K. In addition to these four works, some other publications were found reporting
density and viscosity values over limited range of temperatures and concentrations. Chenlo
et al. (2001) reported kinematic viscosities at 0.25-2.0 mol·kg-1 and between 293.1 and
323.1 K. Mandal et al. (2003a) reported densities and viscosities at 313 K for four aqueous
blends of total amine concentration of 30 wt%. Wang and Li (2004) reported density and
viscosity of 1.0 and 1.5 kmol·m-3 AMP (9 and 13.5 wt%) aqueous solution containing small
additions of DEA (0.1 to 0.4 kmol·m-3; 1.1 to 4.2 wt%). Mandal and Bandyopadhyay
14
(2005) studied the absorption of CO2 and H2S in AMP + DEA aqueous solutions in a
wetted-wall column. For complete system characterisation, the authors measured density
and viscosity for total amine concentration of 3.0 kmol·m-3 and temperatures between 293
and 313 K. Density data showed excellent correspondence with those of Hsu and Li
(1997a), Mandal et al. (2003a), and Aguila-Hernández et al. (2001) while it was possible to
observe a good agreement between their viscosity data and those of Hsu and Li (1997b).
Other AMP based systems
Densities and kinematic viscosities of aqueous blends of AMP + MDEA have been
reported by Welsh and Davis (1995) within the temperature range of 283-353 K for
densities and 283-333 K for viscosities for a total amine concentration of 50 wt% for
density (10-50 wt% AMP), and 5-50 wt% for viscosity. By extending the range of
compositions, the same research group published in Davis and Pogainis (1995) densities for
aqueous amine solutions of 25 wt% AMP + (5 to 20 wt%) MDEA over the temperature
range 283-333 K. Aguila-Hernández et al. (2001) published density at 313.15, 323.15, and
333.15 K and for solutions of total amine concentration of 30, 40 and 50 wt%. The same
paper also reported the only density data available for the aqueous AMP + NMP system.
Density and viscosity for the aqueous AMP + Pz system were reported by Sun et al.
(2005), Paul and Mandal (2006c) and Samanta and Bandyopadhyay (2006), covering the
temperature range of 288-333 K and total amine concentrations between 9 and 30 wt% Pz.
Densities and viscosities decreased with increasing temperature and decreasing mass
fraction of PZ in the mixture. For 30 wt% AMP, 0.04 % deviation was found between
density data of Li and Lie (1994) and those by Samanta and Bandyopadhyay (2006).
Viscosity values of Paul and Mandal (2006c) and Samanta and Bandyopadhyay (2006)
showed excellent agreements.
Density and viscosity for aqueous ternary solutions of 2-(methylamino)ethanol (MAE;
MMEA) and 2-(ethylamino)ethanol (EAE; EMEA) with AMP are given by Álvarez et al.
(2006) at 298.15-323.15 K and total amine concentration of 50 wt% (AMP/(MMEA or
EMEA) wt% ratio was varied from 10/40 to 50/0, with 10 wt% increments). Similar data
for density were reported by Venkat et al. (2010a) for 30 wt% total amine concentration. It
15
was observed that the density of the ternary mixture decreased with increasing temperature
and with decreasing mass fraction of MMEA in the mixtures. No similar data are available
for viscosity.
Only one quaternary system was studied in the literature. Density and viscosity were
reported between 303.15 and 343.15 K for aqueous solutions of three alkanolamines
composed by 32.5 wt% MDEA + 12.5 wt% DEA + (2, 4, 6, 8, or 10 wt%) AMP
(Rebolledo-Libreros and Trejo, 2006). Since the pure AMP density was always lower than
that of DEA or MDEA in the range of temperature considered, the density values of the
studied solutions decreased as the AMP concentration increased. It was also found that the
viscosity values increased as the AMP concentration increased. Equations were developed
to allow the calculation of density and viscosity for aqueous solutions of MDEA and DEA
as a function of AMP concentration and temperature.
1.2.2.2.1.2. Other SHA systems
2-PE systems
Data for binary aqueous 2-PE systems were given by Shen et al. (1991), Xu et al.
(1992b), Aguila-Hernández et al. (2001) and Paul and Mandal (2006a). Densities for all
concentrations and temperatures were found to be in good agreement when coming from
Shen et al. (1991), Xu et al. (1992b) and Paul and Mandal (2006a). For aqueous solutions
of 10 wt% and 30 wt% 2-PE over the temperatures of 298 and 323 K, densities reported by
Paul and Mandal (2006a) are different respectively only by 0.09 and 0.08 % from those of
Xu et al. (1992b). Data from Aguila-Hernández et al. (2001) agreed well the others at 313
K but were significantly lower at the temperature of 323.15 and 333.15 K. Xu et al. (1992b)
stated that viscosity of aqueous 2-PE solutions is difficult to correlate or estimate, since in
solution 2-PE has not only polarity but also molecule association effects. A comparison
between viscosity data by Xu et al. (1992b) and Paul and Mandal (2006a) reported for 10
wt% and 30 wt% 2-PE over a temperature range of 298-313 K showed, respectively, 0.60%
and 3.27% deviation. A comparison between viscosity data of Shen et al. (1991) and Paul
and Mandal (2006a) was possible at 313 K and showed only a mean deviation of 0.90%
indicating good correlation between these data.
16
Mixtures between 2-PE and commonly used CO2 absorbents like MEA, DEA, MDEA
and Pz have also been of interest in the literature. The system 2-PE + MEA was first
considered by Hsu and Li (1997a, b) who reported densities and viscosities between 303
and 353 K for systems containing 30 wt% total amine (6/24, 12/18, 18/12 and 24/6 wt% of
2-PE and MEA respectively) and 20 wt% total amine (5/15, 10/10 and 15/5 wt% of 2-PE
and MEA respectively). It was found that for all temperatures, the increase in MEA
concentration in the blend leads to an increase of the density and a decrease of the
viscosity. Since 1997, the only data for this system were given by Paul and Mandal
(2006a). The authors measured densities and viscosities between 288 and 333 K for 30 wt%
total amine concentration. At 303, 313, 323, and 333 K, density data showed 0.03%,
0.06%, 0.10%, and 0.17% deviations, respectively, from those reported by Hsu and Li
(1997a), while viscosity data showed 0.68%, 0.67%, 0.77%, and 0.85% deviation,
respectively, from those reported by Hsu and Li (1997b), which is quite satisfying.
Density for the system 2-PE + DEA was measured by Aguila-Hernández et al. (2001)
at 313 K (total amine concentration varying between 30 and 50 wt%) and by Paul and
Mandal (2006a) between 288 and 333 K (total amine concentration kept at 30 wt%). For all
temperatures, densities increased with the increase of DEA concentration in the blend. At
313 K and for a total amine content of 30 wt%, experimental data by Paul and Mandal
(2006a) diverged at low 2-PE wt% ratio from data of Aguila-Hernández et al. (2001) but
became similar for concentrations above 20 wt% of 2-PE. For this mixture, the only data
available for viscosity are reported by Paul and Mandal (2006a) between 288 and 333 K
and for a total amine concentration of 30 wt%. No similar data are then available in the
open literature for comparison.
For the system 2-PE + MDEA, the only data available concern density at 313.15,
323.15, and 333.15 K, for total amine concentration of 30-60 wt% (Aguila-Hernandez et
al., 2001). It was found that for all temperatures, densities increase as MDEA concentration
increases. No similar data are available for comparison.
Densities and viscosities for the aqueous system 2-PE + Pz were reported by Paul and
Mandal (2006b) between 288 and 333 K and for total amine mass fraction of 30%. At
17
constant temperature, the increase of Pz concentration in the blend leads to an increase in
density and a decrease in viscosity. No similar data are available for comparison.
Mixed chemical/physical solvents can also been used to remove acid gases from gas
streams. They combine the advantages of chemical (usually, aqueous solutions of
alkanolamines) and physical solvents (usually, organic compounds with high boiling
points). Xu et al. (1993b) reported and correlated densities and viscosities of aqueous
blends of 2-PE and sulfolane (TMS), a physical solvent. At 298 K, densities and viscosities
are given for various aqueous solution of 2-PE (10-65 wt%) + TMS (1.82-44.44 wt%). For
blends of 45 wt% 2-PE + 40 wt% TMS and 55 wt% 2-PE + 10 wt% TMS, data were
measured between 293 and 358 K for densities and over 293-364 K for viscosities. No
similar data are available in the open literature for comparison.
AEPD systems
Density and viscosity data for aqueous AEPD systems are very scarce in the open
literature. Only two publications from the same research group (Yoon et al., 2002a; Yoon et
al., 2002b) were found to report useful information. Yoon et al. (2002a) reported density
and viscosity for AEPD for solution of 5 to 25 wt% by 5 wt% increment and from 303.15
to 318.15 K. The second publication (Yoon et al., 2002b) provided additional data by
extending the range of concentration (20-100 and 20-80 wt% AEPD for density and
viscosity measurements, respectively) and temperature (up to 343.15 K).
AMPD systems
Density and viscosity data for aqueous AMPD system are even scarcer than those
concerning the AEPD system. Only one publication was found giving useful information.
Baek et al. (2000) published density and viscosity data of AMPD binary system of 10, 20
and 30 wt% and over a temperature range of 303 to 343 K. Data were correlated with a
polynomial equation for densities and an exponential one for viscosities. The maximum
deviations between the measured and calculated data were less than 0.005% for densities
and 0.3% for viscosities. Data by Baek et al. (2000) were taken by Yoon et al. (2003) in
their kinetic study using a wetted-wall column absorber.
18
AHPD systems
The system containing AHPD was quite well covered in the literature. Park et al.
(2002a) measured densities and viscosities of aqueous AHPD solutions between 303.15 and
343.15 K and for AHPD concentrations ranging from 5 to 25 wt%. Le Tourneux et al.
(2008) brought new experimental data for solutions of concentrations between 0.15 and 10
wt% AHPD and temperatures of 283.15-313.15 K. The low concentration range was
compatible with aqueous solutions required for developing an enzymatic CO2 capture
process. Density and viscosity values for AHPD aqueous solution of 10 wt% at 303.15 and
313.15 K are in excellent agreement with the results reported by Park et al. (2002a)
(average absolute deviation of 0.025% for density and 1.3% for viscosity). Paul et al.
(2009c) reported polynomial equations (no tabulated results) of density and viscosity of
aqueous AHPD solutions under a temperature range from 298 to 323 K. The concentration
of AHPD in the solution was varied between 2.17 and 21.7 w%. For 10 wt% AHPD
solution and over the temperatures of 298 to 313 K, density and viscosity data showed good
agreement with respectively 0.18% and 2.65 % deviations from data of Le Tourneux et al.
(2008). The only ternary system involving AHPD was considered by Bougie et al. (2009)
who measured density and viscosity of aqueous AHPD + Pz solutions containing 1 kmol.m-
3 AHPD (11.8 wt%) and small amounts of Pz (0.1 to 0.4 kmol·m-3; 0.8 to 3.4 wt%) at
temperatures between 303.15 and 323.15 K. No similar data are available for comparison.
1.2.2.2.1.3. Density and viscosity correlations
Only reliable data from the available references were correlated using simple
polynomial linear equations. They can be very useful to calculate data at given
temperatures and concentrations in the ranges corresponding to data given in Tables A1-A4
(information about data used for correlations are given in §1.2.2.2.1.3.1 and §1.2.2.2.1.3.2).
1.2.2.2.1.3.1. Density correlations
For pure and binary systems (SHA + H2O), all the references indicated in Tables A.1
and A.2 were used in our database at the exception of (i) for AMP: Littel et al. (1992),
Kundu et al. (2003), Saha et al. (1993) and Arcis et al. (2007), (ii) for 2-PE: Aguila-
19
Hernández et al. (2001) at 323.15 and 333.15 K, and (iii) for AHPD: Paul et al. (2009c) for
reasons mentioned in §1.2.2.2.1.1 and §1.2.2.2.1.2.
The equation correlating the selected pure and binary density data for these sterically
hindered amines, where w is the amine mass percentage and T the absolute temperature, is
the following:
[ ] i
iiiii Twdwcwba 2
1
0
323 )K/(wt%)/()wt%/()wt%/( cmg/ ⋅⋅+⋅+⋅+=⋅ ∑=
−ρ (1.1)
Table A.5 give the coefficients of Eq. (1.1) along with the determination coefficient (R2)
and the overall average deviation percentage (O.A.D.%) of the calculated data relatively to
the literature data. It should be mention that only the statistically significant coefficients
were found, the others equal zero. This will apply also for the other presented correlations.
For ternary systems (SHA + other amine + H2O), all the references indicated in Tables
A.1 and A.2 were used in our database at the exception of (i) for AMP + MDEA: first data
of Davis and Pogainis (1995) for 25 wt% AMP + 5 wt% MDEA at 333.15 K which seems
odd. The equation correlating the selected ternary density data, where w1 is the mass
percentage of the SHA, w2 is the mass percentage of the other species and T the absolute
temperature, is the following:
( )
1 111 23 2
1 01 2
( / wt%) ( / wt%)/ K/ g cm ( / K)
/ wt% / wt%
i iii i i i
iii
ba c w d wT T
e w wρ
+ +
−
+=
+ + ⋅ + ⋅ ⋅ = ⋅ + ⋅ ⋅
∑
(1.2)
Tables A.6 and A.7 give the correlation coefficients of Eq. (1.2) for the ternary systems
without and with AMP, respectively. It should be mentioned that the coefficients found for
the systems 2-PE + Pz and AMP + EMEA are only specific for the data considered here:
total amine concentration of 30 wt% and 50 wt%, respectively, as only these data were
available for correlations. In general, Eqs. (1.1) and (1.2) applied to correlate density of
pure, binary and tertiary systems give excellent agreement as it can be seen in Figure 1.3.
Quaternary data presented by Rebolledo-Libreros and Trejo (2006) were not used because
the authors presented their own correlation and no similar data were available in the
literature.
20
Figure 1.3. Literature density values of 2-PE + H2O solutions and results calculated with Eq. (1.1).
1.2.2.2.1.3.2. Viscosity correlations
In comparison with the density correlations (§1.2.2.2.1.3.1), it was much more difficult
to find an accurate and simple linear correlation for viscosity data with a limited number of
correlation coefficients. Another difficulty arose from the fact that several studies reported
kinematic viscosity data without the respective density value to calculate the dynamic
viscosity what limited our database. Therefore, not all systems indicated in Tables A.3 and
A.4 have been correlated here.
For pure and binary systems, only AHPD and AMPD viscosity data were successfully
correlated using this equation:
i
iii
ii Twdwc
Tb
asmPa 21
0
2 )K/(wt%)/()wt%/(K/
)/ln( ⋅
⋅+⋅++=⋅ ∑
=
µ
(1.3)
For AHPD, the paper of Paul et al. (2009c) was not considered as no tabulated data
were available. Table A.8 give the information about the correlation coefficients of Eq.
(1.3) for these two systems. It should be mentioned that in Table A.8, R2 is linked to
21
ln(µ/mPa·s) whereas the stated O.A.D.% is associated directly to µ/mPa·s. For comparison
seek, Eq. (1.3) applied to pure and binary 2-PE, AEPD and AMP viscosity data of Tables
A.3 and A.4 gave respectively overall average deviations of 5.6, 3.6 and 8.4% what seemed
too high to be of interest.
Concerning the viscosity data of ternary systems, an equation similar to Eq. (1.2) was
chosen (i.e. Eq. (1.4)) to correlate them. Tables A.9 and A.10 display the regression
coefficients found for the selected systems, only the AMP + MDEA system was discarded
as the O.A.D.% was too high. Our database was composed of the articles indicated in
Tables A.3 and A.4 at the exception of Chenlo et al. (2001) for AMP + DEA and AMP +
MEA. The system AMP + MMEA was correlated with the kinematic viscosity instead of
the dynamic one for more accuracy. It should be mentioned that the coefficients found for
the system 2-PE + TMS are only specific for 45 wt% 2-PE + 40 wt% TMS and 55 wt% 2-
PE + 10 wt% TMS at temperatures between 293 and 364 K. Also, the coefficients found for
the system AHPD + Pz are only specific for the system containing 11.8 wt% AHPD, as
only these data were available for correlations. Figure 1.4 shows some data of the literature
along with values calculated with the Eq. (1.4).
( )
1 111 2 2
1 01 2
( / wt%) ( / wt%)/ Kln( / mPa s) ( / K)
/ wt% / wt%
i iii i i i
iii
ba c w d wT T
e w wµ
+ +
+=
+ + ⋅ + ⋅ ⋅ = ⋅ + ⋅ ⋅
∑
(1.4)
1.2.2.2.2. Surface tension
Surface tension of mixtures is an important property for the design of contacting
equipment like packed columns and membrane contactors used in gas absorption. Surface
tension affects the hydrodynamics and transfer rates of such systems where a gas-liquid
interface exits. In packed columns, surface tension was found to be one of the most
sensitive parameter in CO2 absorption by influencing the effective mass transfer area
(Gabrielsen et al., 2006). In membrane contactors, surface tension of solutions and the
hydrophobicity of the membrane strongly influence membrane wettability. In addition,
values of surface tension are also necessary to estimate the breakthrough pressure of the
solution through the pore of the membrane by using the Laplace-Young equation. Table
22
A.11 reports the aqueous amine systems for which data of surface tension were found in the
literature. For conventional amine solution concentrations (less than 40 wt%) and under a
temperature range of 293-393K, surface tension values of SHA solutions were found to
usually be between 38 and 72 mN·m-1.
Figure 1.4. Literature viscosity values of 2-PE + MEA + H2O solutions with a total amine content of 30 wt% and results calculated with Eq. (1.4).
In a study concerning membrane wetting, Rongwong et al. (2009) reported punctual
values of surface tension of 1 kmol·m-3 AMP aqueous solution and of 0.25 kmol·m-3 AMP
+ 0.25 kmol·m-3 (DEA or MEA) at 303 K. Authors mentioned that important measures to
prevent the wetting problems include the selection of liquids with suitable surface tension.
It was reported that when the liquid surface tension decreased from about 33 mN·m to 30
mN·m, the transmembrane pressure difference in polypropylene (PP) membranes was
decreased from about 0.9 bar to 0.1 bar, leading to the rapid increase of membrane wetting.
Another study reporting surface tension of AMP and AMP + MEA aqueous solution was
made by Vázquez et al. (1997). They measured surface tension at temperatures from 298 to
323 K and total amine concentration varying between 5 and 100 wt% for the binary AMP
system or kept at 50 wt% for tertiary mixtures. The experimental binary values were
23
correlated with temperature and mole fractions. For all studied systems, surface tension
decreased with increasing temperature for any given concentration and decreased when
wt% ratio of AMP increased in ternary system for a given temperature. Álvarez et al.
(2003; 1998) measured the surface tension of aqueous solutions of AMP + MDEA, AMP +
3-amino-1-propanol (AP) and AMP + 1-amino-2-propanol (MIPA) at 298-323 K. For these
tertiary mixtures, the concentration range for each amine was 0-50 wt% by 10 wt%
increments. Yoon et al. (2002b) reported surface tension of aqueous AEPD for temperature
ranging from 303.15 to 343.15 K and AEPD concentration of 20-80 wt%. The experimental
data were correlated as a function of temperature and AEPD concentration with an average
absolute deviation of 0.4%. Paul and Mandal (2006b) measured the surface tension of
aqueous blends of Pz as activator with 2-PE or AMP between 293 and 323 K and total
amine mass fraction of 30 %. Surface tension of the ternary mixtures decreased with
increasing temperature and decreasing mass fraction of Pz in the mixture. Ventak et al.
(2010a) reported experimental surface tension data of aqueous blends of AMP + MMEA at
298-323 K and total amine mass fraction of 30 %, as well as correlations with temperature
and amine concentration. The surface tension increased with decreasing temperature and
increasing mass fraction of MMEA in the mixture. One study has been found in the
literature concerning the surface tension of mixture of three alkanolamines. Águila-
Hernández et al. (2007) determined the equilibrium surface tension for aqueous solutions
composed of 32.5 wt% MDEA + 12.5 wt% DEA + (2, 4, 6, 8, or 10 wt%) AMP between
303.15 and 343.15 K. In the temperature range studied, the experimental surface tension
values of the aqueous blends of three alkanolamines decreased linearly with the increase of
AMP concentration and temperature. The authors mentioned that this behaviour was highly
consistent with the fact that the surface tension of pure AMP was lower than that of pure
MDEA and pure DEA, and consequently, the surface tension of aqueous solutions at a
given AMP concentration was lower than that of aqueous solutions of MDEA and DEA,
individually, under the same conditions of concentration and temperature. This behaviour
led to the statement that the lower the solution surface tension, the larger its absorption
capacity towards acid gases in conventional gas-liquid contactor. Furthermore, an analysis
24
of the excess surface adsorption clearly indicated the existence of an excess of amine
molecules at the liquid–vapour interface with respect to those of solvent.
From all these works, the only possible comparison can be made for AMP surface
tension at 303.15 K and 10 wt%: 52.87 mN·m-1 from Vázquez et al. (1997) versus 58.81
mN·m-1 from Rongwong et al. (2009).
1.2.2.2.3. Vapour pressure
For solubility measurements or modeling CO2 absorption in aqueous amine solutions,
the vapour phase needs to be analysed to determine the exact CO2 content. Most studies
consider that only water and CO2 are volatile compounds and therefore, amine volatility
can be neglected. However, it is often stated that MEA, the most used conventional
alkanolamine have a high vapour pressure and high amine loses occur industrially. Studies
on SHA volatility should then be made to explore the potential use of these amines.
Nguyen et al. (2010) mentioned that an excessive volatility may result in significant
economic losses and environmental impact. According to the authors, volatility is of
greatest interest at the top of the absorber at 313 K and at nominal lean loading because
aqueous amine absorbers are designed to operate near this temperature and that cleaned flue
gas leaving the absorber will tend to be in equilibrium with lean amine solution. Their study
reported amine volatility in 7 mol·kg-1 MEA, 8 mol·kg-1 Pz, 7 mol·kg-1 MDEA + 2 mol·kg-
1 Pz, 12 mol·kg-1 EDA (ethylenediamine), and 5 mol·kg-1 AMP at 313-333 K with lean and
rich loadings giving CO2 partial pressures of 0.5 and 5 kPa at 313 K. Data were obtained
from FTIR spectroscopy for both unloaded and nominal lean and rich CO2 systems. The
results showed that amine solutions were ranked in order of increasing amine volatility as
follows: 7 mol·kg-1 MDEA + 2 mol·kg-1 Pz (6/2 ppm), 8 mol·kg-1 Pz (8 ppm), 12 mol·kg-1
EDA (9 ppm), 7 mol·kg-1 MEA (31 ppm), and 5 mol·kg-1 AMP (112 ppm). 5 mol·kg-1
AMP was found the most volatile amine at the CO2 partial pressure of interest, 0.1-0.5 kPa
at 313 K. This behaviour may come from the fact that SHA, as they do not form stable
carbamates, existed in their free form and not in reacted, non-volatile species in solution,
increasing therefore their volatility.
25
AMP volatility was also studied by Pappa et al. (2006). In that work, AMP vapour
pressures were measured in the temperature range of 373.3 to 436.9 K. Data were
correlated with an Antoine expression with an mean deviation of 0.5%:
−=
32.107K/6.3472 - 15.155exp /kPasat
AMP TP (1.5)
1.2.2.2.4. Heat capacity
In a conventional industrial CO2 absorption process, a lean aqueous solution first
absorbs CO2 and is then sent to a stripper where CO2 is recovered and compressed. The
absorption takes usually place at room temperature or slightly above (298-323 K), whereas
solution regeneration is around 383 K. Heat capacity data for alkanolamine solutions are
required for the design of heat-exchangers included in the absorption/desorption
installation. Table A.12 reports the works where heat capacity data for various SHA were
found in the open literature. It was found that usually, SHA solution heat capacity values
can fluctuate between around 90 and 300 J·mol-1·K-1.
Some estimation methods to predict molar heat capacity can be found in the literature,
like for example those of Missenard (1965), Chueh and Swanson (1973a, b) and Nagvekar
and Daubert (1987). However, these estimations cannot always be considered as reliable as
true experimental data. It’s worth mentioning that except for the aqueous AMP were
several works have been published, no comparable experimental data are available for
comparison for the other systems.
Since 1999, the research group of Meng-Hui Li published several studies concerning
heat capacity of pure or aqueous alkanolamine solutions used for CO2 absorption. Chiu and
Li (1999) and Chiu et al. (1999) reported heat capacities of pure and aqueous solutions of
2-PE, AMP and several other conventional amines from 303 to 353 K. A comparison
showed that at 323 K, good agreement was found between the reported AMP Cp value
(2.80 kJ·kg-1·K-1) and the one estimated from Missenard (1965) (2.734 kJ·kg-1·K-1) with a
deviation of 2.4%. However, at temperatures of 293 and 298 K, both Cp estimation methods
of Missenard (1965) and Chueh and Swanson (1973a, b) yielded poor results compared to
26
the measured AMP Cp values, but good results compared to the 2-PE ones. It was observed
that the order of Cp for alkanolamine aqueous solutions generally follows the order of Cp
for pure alkanolamines. Among the eight studied alkanolamine aqueous solutions (MEA,
DEA, DGA, DIPA, TEA, MDEA, AMP, and 2-PE), the AMP system showed the strongest
non-ideality behaviour. Shih et al. (2002) determined the heat capacity of aqueous and non-
aqueous mixture of 2-PE + MEA from 303 to 353 K. The Redlich-Kister equation
correlated the ternary system with an overall average absolute deviation of 0.2% for 176
data points. Shih and Li (2002) measured heat capacities of non-aqueous AMP + DEA (0.1
to 0.9 AMP mole fractions) and of 16 aqueous ternary solutions. It was observed that at
constant temperature, the heat capacity of AMP + DEA increased as the mole fraction of
DEA increased. Heat capacities of aqueous AMP and aqueous and non-aqueous AMP +
MEA solutions from 303 to 353 K (eight binary and sixteen ternary systems) were given by
Chen and Li (2001). Probably due to the use of higher AMP purity, the values of Cp
obtained in this study were slightly higher than those of Chiu et al. (1999). However,
excellent agreement with Maham et al. (1997) was found.
Ho et al. (2007) reported heat capacities of aqueous solutions of AMP with TMS over
a temperature range from 303.15 to 353.15 K. Since the mole fraction of water in aqueous
alkanolamine solution is normally greater than 0.5 (Kohl and Nielsen, 1997), twelve
solutions of AMP + TMS + water that covered the mole fractions of water from 0.6 to 0.8
were studied. Heat capacities of AMP + sulfolane were also determined. For 132 data
points of AMP + sulfolane + water, the fitted results of heat capacity calculations using a
Redlich-Kister equation (average absolute percentage deviation (AAD%)) were 0.3 and
7.7% for the molar heat capacity and the excess molar heat capacity, respectively.
In addition to studies from Li’s research group, some works concerning AMP over
different temperature ranges are worth to be mentioned. Maham et al. (1997) measured
molar heat capacities of 14 pure alkanolamines (including AMP) at various temperature
from 299.1 (323 for AMP) to 397.8 K. The molar heat capacity was represented through a
structural dependence model, where the molar heat capacity of one molecule was
considered as the sum of various group (CH2, OH, NH and N) contributions. An analysis of
27
their model indicated that the molar heat capacities of alkanolamines were dominated by
CH2 and OH group contributions and that these contributions increased with increasing
temperature. In the work by Zhang et al. (2002), heat capacities of pure and aqueous
solutions of AMP were measured, respectively, at temperatures from 303.15 to 368.15 K
and from 278.15 to 368.15 K. Experimental Cp data for pure AMP were compared with
literature values. While the values from Zhang et al. (2002) and those from Chiu et al.
(1999) , Chen and Li (2001) and Maham et al. (1997) were in good agreement considering
the uncertainties, some deviations appeared with the data estimated from the works of
Chueh and Swanson (1973a, b) and Missenard (1965).
Based on the studies reporting pure and binary AMP Cp values indicated in Table
A.12, our own correlation was elaborated as:
[ ] i
iiiiiP TwdwcwbaC 2
1
0
321-1 )K/(wt%)/()wt%/()wt%/( KmolJ/ ⋅⋅+⋅+⋅+=⋅⋅ ∑=
−
(1.6)
Table A.13 reports the correlation coefficient of Eq. (1.6). Based on the five studies
(Chen and Li, 2001; Chiu and Li, 1999; Chiu et al., 1999; Maham et al., 1997; Zhang et al.,
2002), 328 data were correlated with an overall mean deviation of only 1.1%.
1.2.2.2.5. Heat of absorption
When designing absorption with chemical reaction there are several factors to account
for. One of the most important considerations is the temperature variation within the
absorber arising from the heat of absorption of the acid gas. The temperature influences not
only the equilibrium line, but also the rate of the chemical reactions involved and the
physical properties of the liquid and the gas (Gabrielsen et al., 2006). The use of constant
heat of absorption values in the calculations often leads to inaccurate results since the
magnitude of this phenomenon varies with temperature and CO2 content in the
alkanolamine solutions (CO2 loading). Experimental data are then necessary and they can
be derived either from solubility data or by direct calorimetric measurements. The
exothermic effect of the CO2 absorption cause an increase of the enthalpy of solution
(referred as the differential enthalpy of solution, ∆Hsol) and this can be calculated from
28
solubility data using Eq. (1.7). The differential enthalpy is then integrated following Eq.
(1.8) in order to get the enthalpy of solution Hsol.
R
)d(1/
ln d2CO solH
TP ∆
=
α
(1.7)
∫ ∆=α
αα 0
sol d1 solHH (1.8)
There are few direct measurements of the enthalpy of solution and the available
measurements showed considerable scatter with respect to both temperature and
concentration of alkanolamine (Rebolledo-Libreros and Trejo, 2004). Based on solubility
data of CO2 in aqueous alkanolamine solutions, Murrieta-Guevara et al. (1998) derived the
differential enthalpy of solution (∆Hsol) at 343.15 K for systems of 10 wt% AMP + 20 wt%
DEA and 5 wt% AMP + 25 wt% DEA. Values were obtained for CO2 loadings of 0.5, 0.6
and 0.7. It was seen that within ±10%, ∆Hsol was a linear function of α for all systems
considered and it changed slightly with the concentration of each amine of the blend. From
the same research group, Rebolledo-Libreros and Trejo (2004) obtained experimental gas
solubility data for CO2 in aqueous solutions of 32.5 wt% MDEA + 12.5 wt% DEA + (4, 6,
or 10) wt% AMP at 313.15, 343.15, and 393.15 K. They showed that the plots of 2COln P
versus 1/T were linear with a correlation coefficient of 0.99, indicating that ∆Hsol was
independent of temperature over the range of temperature studied. For each temperature,
pressure values were smoothed with a polynomial function to carry out interpolations of
∆Hsol at constant values of α. Differential enthalpies of solution were extracted at a mean
temperature of 350 K for loadings from 0.1 to 0.7. They found that within ±20%, the
calculated values of ∆Hsol were not influenced by the change of AMP concentration. An
explicit model for CO2 solubility in an aqueous solution of AMP has been proposed by
Gabrielsen et al. (2006) and an expression for the heat of absorption of CO2 has been
developed as a function of loading and temperature.
+=⋅∆
K/47652 8161-R mol/J 1-
TH sol
α
(1.9)
A rate-based steady-state model for CO2 absorption into an AMP solution has also
been developed (Gabrielsen et al., 2006), using both the proposed expression for the CO2
29
solubility and the expression for the heat of absorption along with an expression for the
enhancement factor and physicochemical data from literature. The proposed model was
successfully applied to absorption of CO2 into an AMP solution in a packed tower and
validated against pilot-plant data from literature. Arcis et al. (2007) measured the enthalpies
of solution of CO2 in 15 and 30 wt% AMP aqueous solutions at 322.5 K and for total
pressures from 0.2 to 5 MPa. The experimental enthalpies of solution were compared to the
values derived from vapour-liquid equilibrium data available in the literature. The
enthalpies estimated from Park et al. (2002c) for a 30 wt% AMP aqueous solution were
found to be in good agreement with their experimental enthalpies, but only for CO2 loading
over 0.4. The calorimetric data also allowed the determination of gas solubility in the liquid
phase.
1.2.2.2.6. Corrosion and amine degradation
According to Kohl and Nielsen (1997), the most serious operating problem
encountered in acid gas separation plants is corrosion. The corrosion problem leads to
direct impacts on a plant’s economy since it causes unplanned downtime, production
losses, reduced equipment life, and even injury or death. Veawab et al. (1996, 1997; 1999)
studied corrosion and corrosion inhibition in AMP aqueous solutions (1, 2.5, 5, and 7
kmol·m-3; 9 to 63.3 wt%) by static weight loss tests. The corrosion data were obtained
under boiling conditions in order to simulate the service environment in reboilers and
regenerators and were compared with those of MEA, tested under the same conditions. The
results indicated that AMP solutions were less corrosive to carbon steel than MEA ones in
environments of both pure CO2 and a mixture of CO2 and air (10% O2).
It is also often related that amine degradation products can influence significantly
corrosion caused by amine solutions. Thermal degradation occur as the amine solution is
circulating from the absorber (temperature up to 330 K) to the regeneration column where
temperature can go as far 413 K. Oxidative degradation is mainly caused by the presence of
oxygen in the flue gas and is then occurring in the absorber where oxygen concentration is
higher.
30
Detailed studies (Lepaumier et al., 2009a, b; Supap et al., 2006) were found in the
literature describing existing degradation mechanisms, giving degradation rate and
indicating plenty of possible degradation products found in solutions for conventional
alkanolamines (e.g. MEA, DEA, MDEA). Unfortunately, details concerning SHA are very
scarce and concern mainly AMP solutions where AMP was found more stable than usual
alkanolamines (Freeman et al., 2010; Reza and Trejo, 2006). It appeared that corrosion and
amine degradation are linked and that this field of research is very complex as many
degradation mechanisms exist, degradation products are abundant (e.g. more than 15 for
MMEA (Lepaumier et al., 2011)) and because many other parameters can be taken into
account: presence of metal ion, of dissolved CO2, of other reactive sour gas, of oxygen, etc.
Therefore, this section will not be discussed in depth in this review and readers are
encouraged to read selected literature on the subject for more information.
1.2.2.2.7. CO2 diffusivity in SHA solutions
Gas diffusivity in solutions is one essential parameter for the design of gas/liquid
contactors. It is also needed for the operation of certain types of contactors, in particular the
wetted-wall column, often used for kinetic studies. 2COD is used to calculate the
enhancement factor (E) and the liquid-side mass-transfer coefficient. However, the
diffusion coefficient of CO2 in amine solution cannot be measured directly as the acid gas
reacts with the amine. Therefore, some methods are usually adopted to estimate it, namely
the N2O analogy and the Stokes-Einstein relation. In the N2O analogy, CO2 diffusion
coefficient in amine aqueous solution can be estimated from N2O diffusion coefficient in
the same solution and the diffusivity ratio of these two gases in water at the same
temperature, according to the following equation:
( )waterON
CO
amineONamineCO2
2
22 )(
=
DD
DD
(1.10)
It is commonly accepted that values of nitrous oxide and carbon dioxide diffusion
coefficients in water can be calculated from equations proposed by Versteeg and van
Swaaij (1988).
31
( )
×=⋅ −
/K2119-exp 10 2.35 sm/ 6-12
CO2 TD
water
(1.11)
( )
×=⋅ −
/K2371-exp 10 5.07 sm/ 6-12
waterON2 TD
(1.12)
The use of the Stokes-Einstein relation allows the reduction of the number of
experiments. In Eq. (1.13), the N2O diffusion coefficient is estimated based on the
viscosities of amine solution and water and on the diffusion coefficient of N2O in water; the
last parameter can be calculated by Eq. (1.12) at a given temperature. However, many
uncertainties concern the exponent value (β) related to viscosities.
( ) ( )waterONAmineON 22
constant ββ µµ ⋅==⋅ DD
(1.13)
Table A.14 presents experimental values of N2O diffusion coefficient in various SHA
aqueous solutions. It is traditional to get values between 0.6 - 2.0 ×10-9 m2·s-1 but higher
diffusivity value can be found at higher temperature. As mentioned earlier, these values can
be used to estimate the CO2 diffusion coefficient in the same solution with the N2O
analogy. When the ratio 22 CO
1/2CO / HD is obtained experimentally, the diffusivity can be
calculated on the basis of Henry’s constant obtained from solubility measurements.
1.2.2.2.7.1. AMP systems
AMP + H2O
Yih and Shen (1988) measured the ratio 22 CO
1/2CO / HD in aqueous AMP solutions using
the N2O analogy. Nitrous oxide absorptions were performed at 313 K for AMP solutions of
0.258-3.0 kmol·m-3 (2.3-27 wt%). Data by Xu et al. (1991) obtained at 294-348.5 K for
AMP solutions of 2 and 3 kmol·m-3 (18 and 27 wt%) differ by 15% for the 3 kmol·m-3
AMP solution in respect to those given by Yih and Shen (1988). The authors recommended
the use of the N2O analogy method to estimate the diffusivity of CO2 in AMP solutions,
instead of the Stokes-Einstein relation. Even a value of 0.80 for β in the Stokes-Einstein
relation, as recommended by Versteeg and van Swaaij (1988), did not result in satisfactory
estimations. Saha et al. (1993) measured N2O solubility and diffusivity between 294-318 K
32
in 0.5, 1.0, 1.5, and 2.0 kmol·m-3 (4.5 to 18 wt%) AMP aqueous solutions. It was observed
that the diffusivity results did not follow the Stokes-Einstein relation strictly. Authors also
recommended the use of the N2O analogy. Messaoudi and Sada (1996) reported the ratio
ON1/2
ON 22 / HD in AMP solutions of 0.4 to 2.0 kmol·m-3 (3.6 to 18 wt%) and for temperatures
of 293, 303, and 313 K. The results were correlated using the Eq. (1.14) and linear
relationships were obtained at constant temperatures.
b3AMP
amineON
1/2ON
ON
1/2ON )mkmol/a( log
2
2
2
2 −⋅=
C
HD
HD
water (1.14)
where a and b are regressed parameters.
Eq. (1.14) was also applied to represent data reported by Yih and Shen (1988), Xu et
al. (1991) and Saha et al. (1993) and a comparison was made between all these sources.
Data of Messaoudi and Sada (1996), Yih and Shen (1988) and Xu et al. (1991) exhibited a
moderate dependence on amine concentration, but slightly diverge from each other. Data of
Saha et al. (1993) showed a strong amine concentration dependence and it was hard to
distinguish data at 293 K from those at 303 K.
In Ko et al. (2001), diffusivities of N2O were measured in several aqueous
alkanolamine solutions (MEA, DEA, DIPA, TEA, and AMP) at 303, 308, and 313 K and
for AMP concentration from 0.5 to 2.5 kmol·m-3 (4.5 to 22.4 wt%).
Taken all available N2O diffusion coefficient data in AMP-H2O solutions from the
literature, it appeared that the values from Xu et al. (1991), Saha et al. (1993), Ko et al.
(2001), and Xiao et al. (2000) deviate respectively by 5.9%, 2.8%, 3.5% and 1.3% from our
correlation based on all data reported in these four papers (Eq. 1.15). Data of Xu et al.
(1991) were found to deviate more significantly, reaching for some data a maximum
deviation of 22.8%. O.A.D.% indicated in Table A.15, without considering data of Xu et al.
(1991), reduce to 2.8%. Ko et al. (2001) also indicated that some data from that study (Xu
et al., 1991) deviated considerably. N2O diffusion coefficient data reported in Bosch et al.
(1990) were not considered in this correlation because they were based on estimation only.
33
N2O diffusivity data in 30 wt% AMP solutions reported by Mandal et al. (2004) and Li and
Lai (1995) were also not included in the correlation because for all temperatures, these data
were clearly above the trend created by all the other considered data.
i
iii
ii Twdwc
Tb
aD 21
0
2129ON )K/(wt%)/()wt%/(
K/ sm/10
2⋅
⋅+⋅++=⋅⋅ ∑
=
− (1.15)
AMP + MEA + H2O
Three works are available for this system, covering the temperature range of 293-313
K (Li and Lai, 1995; Mandal et al., 2005; Xiao et al., 2000). Li and Lai (1995) measured
the solubility and diffusivity of N2O in several AMP + MEA aqueous systems of total
amine concentration of 30 wt% at 303, 308 and 313 K. In their study, Xiao et al. (2000)
measured the diffusivity of N2O in AMP + MEA aqueous systems of 1.5 and 1.7 kmol·m-3
AMP + (0.1 to 0.4) kmol·m-3 MEA. Mandal et al. (2005) measured N2O diffusivity
between 293 and 313 K for total amine concentration of 30 wt%. Good agreement was
found between these data and those by Li and Lai (1995) for 24 wt% AMP + 6 wt% MEA,
over the temperature range of 303 to 313 K.
AMP + DEA + H2O
As in the case of the aqueous AMP + MEA, three works are also available for this
system, covering the temperature range of 293-313 K (Li and Lee, 1996; Mandal et al.,
2004; Wang and Li, 2004). Li and Lee (1996) measured N2O solubility and diffusivity in
solutions of 30 total amine mass percent. It was observed that the experimental diffusivities
at 303 and 313 K did not follow the Stokes-Einstein relation strictly. In their kinetics study,
Wang and Li (2004) measured N2O diffusivity in aqueous solutions of (1.0 and 1.5)
kmol·m-3 (9 and 13.4 wt%) AMP + (0.1 to 0.4) kmol·m-3 (1.1 to 4.2 wt%) DEA at 303,
308, and 313 K. It was found that diffusivities in 1.5 kmol·m-3 AMP + DEA + H2O are
smaller than in 1.0 kmol·m-3 AMP + DEA + H2O, due to the higher viscosity values of the
former system. Also, the diffusivity of N2O was found to decrease as the concentration of
DEA increased at a given temperature and increased as the temperature increased at a given
concentration. Mandal et al. (2004) reported the diffusivity of N2O in aqueous solutions of
total amine concentration of 30 wt% between 293 and 313 K. For 24 wt% AMP + 6 wt%
DEA, over the temperature range of 303 to 313 K, the deviation of the experimental data
34
was within 2.5% in respect to those by Li and Lee (1996). As in Li and Lee (1996), it was
observed that the experimental diffusivities of N2O in AMP + DEA + H2O did not follow
the Stokes-Einstein relation strictly.
AMP + Pz + H2O
Sun et al. (2005) measured solubility and diffusivity of N2O in aqueous mixtures of
AMP and Pz using a wetted-wall column with an estimated error of ±2%. Diffusivity was
measured between 303 and 313 K for solutions containing 1.0 and 1.5 kmol·m-3 AMP (9
and 13.5 wt%) and small addition of Pz (0.1 to 0.4 kmol·m-3; 0.9 to 3.5 wt%). Data were
necessary to interpret kinetic results.
Samanta and Bandyopadhyay (2009) also reported N2O diffusivity in aqueous
solutions of AMP + Pz, over a temperature range of 298-313 K and for solutions with a
total amine content of 30 wt%. By a parametric sensitivity analysis, this study showed that
Henry’s law constant and the estimated CO2 diffusivity in aqueous amine solutions were
among the most influential parameters for the prediction of the absorption rate. Importance
of reliable diffusivity data was also reported by Mandal and Bandyopadhyay (2006).
1.2.2.2.7.2. Other SHA systems
The aqueous 2-PE system was studied by Shen et al. (1991) and Xu et al. (1993a). At
313 K, the ratio 22 CO
1/2CO / HD was found to decrease with the increase of the amine
concentration. Xu et al. (1993a) mentioned that N2O diffusivity decreased with the increase
of amine concentration at a given temperature (at 293 and 313 K) and increased with an
increase of the temperature for a given concentration (between 5 and 40 wt%). Only one
publication was found reporting N2O diffusivity values in aqueous AEPD solutions (Yoon
et al., 2002a). As mentioned by Shen et al. (1991), it was found that the ratio 22 CO
1/2CO / HD
decreased with the increase of amine concentration at a given temperature (between 303.15
to 318.15 K) and decreased when temperature increased at a given concentration (between
5 and 25 wt%). One work was also found for aqueous AMPD system (Yoon et al., 2003).
2COD was determined by the N2O analogy; solubility were taken from Baek et al. (2000).
Bougie and Iliuta (2009) measured the ratio 22 CO
1/2CO / HD by the absorption of N2O in
35
AHPD solutions of 0.5 to 2.4 kmol·m-3 (6 to 27 wt%) between 303.15 and 323.15 K for
AHPD solution and the values were found to follow the same trend as those obtained by
Yoon et al. (2002a) for AEPD systems. For the same system, Paul et al. (2009c) reported
distinctly Henry’s law constants and N2O diffusivity in aqueous solutions of 2.17-21.7
wt%, over the temperature range from 298 to 323 K. It was found that N2O diffusivity in
the aqueous AHPD does not follow the Stokes-Einstein relation strictly.
1.2.2.2.8. Amine diffusivity in SHA solutions
In addition to the CO2 diffusivity, amine diffusivity in aqueous solutions is an
important physical parameter necessary for reaction kinetics study. According to Snijder et
al. (1993), the alkanolamine diffusivity can also be estimated with a modified Stokes-
Einstein relation (β = 0.60):
( ) ( )waterAminesolutionAmine constant ββ µµ ⋅==⋅ DD (1.16)
However, as it was observed in the case of CO2 (N2O) diffusivity, the reliability of this
relation is questionable. The calculations also require the values of amine diffusivity in
water at infinite dilution. Several correlations were found in the literature to estimate amine
diffusion in water at infinite dilution: Othmer and Thakar (1953), Scheibel (1954), Hayduk
and Laudie (1974), and the modified Wilke-Chang relation (Hayduk and Laudie, 1974).
Recently, Mandal et al. (2003a) used Glasscock’s correlation (Glasscock, 1990) to estimate
AMP diffusivity in water. However, works reporting values of SHA diffusivities in their
aqueous solution are quite scarce in the literature.
Chang et al. (2005) measured the diffusion coefficients of AMP, 2-PE and other
conventional alkanolamines in water at infinite dilution, as well as in concentrated solutions
(up to 4 kmol·m-3 (35.8 wt%) for AMP and 3 kmol·m-3 (38.1 wt%) for 2-PE), from 303 to
343 K and at atmospheric pressure. It was found that infinite dilution diffusivity
coefficients of alkanolamines in water depended on the characteristics of the solutions, such
as the sizes of solute and solvent and the intermolecular interactions between solute and
solvent. The following order was given for diffusivity coefficients of alkanolamines in
water: AMP (molar mass 89.14) > DGA (105.14) > 2-PE (129.2) > TEA (149.19). This
36
indicated that a lighter solute (alkanolamine) diffuses faster in water. An equation
representing the diffusion coefficient as a function of temperature and solution
concentration was applied to correlate all experimental data. Deviations between calculated
and experimental data for AMP and 2-PE solutions were 2.4 and 4.5%, respectively.
1.2.3. Mechanism of reaction between CO2 and SHA. Influence of steric hindrance on
carbamate stability
In general, only aliphatic and cycloaliphatic amines are suitable for gas treating
(Sartori et al., 1987). Due to their lower basicity, aromatic amines have low absorption
capacity and rate. When CO2 is absorbed in an amine aqueous solution, the following
reactions can occur (reaction mechanisms are presented for primary, secondary, tertiary and
sterically hindered amine for comparison).
Primary (RNH2) and secondary (R2NH) amines
An example is given for a primary amine:
- zwitterion (RNH2+COO-) formation
CO2 + RNH2 ↔ RNH2+COO- (1.17)
- carbamate (RNHCOO-) and protonated amine (RNH3+) formation
RNH2+COO- + RNH2 ↔ RNHCOO- + RNH3
+ (1.18)
Global reaction:
CO2 + 2 RNH2 ↔ RNHCOO- + RNH3+ (1.19)
The key advantage of the primary and secondary alkanolamines such as MEA and
DEA is their fast reactivity due to the formation of stable carbamates. Conversely, this will
lead to high solvent regeneration cost. On the absorption capacity side, they have the
drawback of a relatively low CO2 loading (stoichiometric loading limited to 0.5 mol
CO2·mol amine-1). Loadings greater than 0.5 mol CO2·mol amine-1 can be achieved only at
high CO2 partial pressures.
Tertiary (R3N) amines
- bicarbonate (HCO3-) formation
CO2 + H2O ↔ HCO3- + H+ (1.20)
37
- amine protonation
R3N + H+ ↔ R3NH+ (1.21)
Global reaction:
R3N + CO2 + H2O ↔ R3NH+ + HCO3- (1.22)
Tertiary alkanolamines, like MDEA and TEA, have a low reactivity with respect to
CO2, due to the formation of bicarbonates by CO2 hydrolysis. However, this will lead to a
very low solvent regeneration cost. Another advantage of these amines is the high CO2
theoretical loading capacity of 1 mol of CO2·mol of amine-1.
Sterically hindered amines
The reaction between a sterically hindered amine and CO2 be can described through
three simultaneous mechanisms:
a) Bicarbonate formation following the same mechanism as tertiary amines (Eq.
(1.22)).
b) Bicarbonate formation by zwitterion hydrolysis:
RNH2 + CO2 ↔ RNH2+COO- (1.17)
RNH2+COO- + H2O ↔ RNH3
+ + HCO3- (1.23)
Global reaction:
RNH2 + CO2 + H2O ↔ RNH3+ + HCO3
- (1.24)
c) Bicarbonate formation by carbamate hydrolysis
RNH2 + CO2 ↔ RNH2+COO- (1.17)
RNH2+COO- + RNH2 ↔ RNHCOO- + RNH3
+ (1.25)
RNHCOO- + H2O + (RNH3+) ↔ RNH2 + (RNH3
+) + HCO3- (1.26)
Global reaction:
RNH2 + CO2 + H2O ↔ RNH3+ + HCO3
- (1.24)
Due to the hindrance of the bulky group adjacent to the amino group, sterically
hindered amines form unstable carbamates whose hydrolysis leads to the formation of
bicarbonate, resulting in the theoretical loading capacity up to 1.0, like the tertiary amines.
Due to the very low kinetics of the physical CO2 absorption, bicarbonate formation through
mechanism a) is much less probable than b) and c).
38
1.2.4. Absorption capacity
CO2 solubility data are of great interest because they are essential for the design and
operation of absorption scrubbing equipment in many technical applications like chemical
industry, oil and gas industry and in environmental protection as well. Tables A.16 to A.19
present all experimental data published in the open literature concerning the solubility of
CO2 in single SHA aqueous solutions (Tables A.16 and A.18) and in SHA based mixed
aqueous solutions (Tables A.17 and A.19) up to date.
1.2.4.1. CO2 chemical solubility in single amine aqueous solutions
1.2.4.1.1. CO2 absorption in AMP aqueous solutions
As the solubility measurements for the CO2 - AMP system attracted many researchers
and the available data are abundant in respect to data for other SHA, this system is
discussed in its own section.
Sartori and Savage (1983) measured CO2 solubility in unhindered MEA and hindered
AMP aqueous solutions (3.0 kmol·m-3; 26.8 wt%) at 313 and 393 K and studied the steric
hindrance and basicity on CO2 – amine reactions. The higher CO2 loadings observed at 313
K for the hindered amine, AMP, confirmed the formation of unstable carbamates. At 393
K, a temperature close to that of regeneration, CO2 loadings in AMP were lower relatively
to MEA, which was in agreement with the thermodynamic model predictions. Chakraborty
et al. (1986) investigated the behaviour of CO2 in AMP aqueous solutions using different
experimental setups. Data are not tabulated but graphically discussed. An acidic species
such as CO2 could, in principle, react with both the amino and alcohol groups present in the
AMP molecule. However, the possible formation of an alkyl carbonic ion from reaction
between CO2 and the alcohol group of AMP can be neglected because the solution pH
never exceeds 12.0 (it is usually between 7.5 and 9). CO2 reaction with the amino group
leads to the formation of chemically combined CO2 forms: carbamate, bicarbonate and
carbonate ions. Since the basicity of AMP is low enough to guarantee that the carbonate-
bicarbonate equilibrium is shifted towards the bicarbonate, the carbonate formation can
entirely neglected. From Cl3 NMR spectra of liquid samples after reaction, the authors
39
could not identify the carbamate peak and they concluded that the concentration of
carbamate was lower than the instrument sensitivity.
Roberts and Mather (1988a) studied the CO2 absorption in aqueous AMP solutions at
313 K (2.0 and 3.0 kmol·m-3 AMP; 17.9 and 26.8 wt%) and 373 K (2.0 kmol·m-3 AMP)
over a wide range of CO2 partial pressures (generally from 1.25 to 5870 kPa). Excellent
agreement was found between their data and those reported by Sartori and Savage (1983).
Experimental data were also compared with previously reported solubility in aqueous MEA
solutions. It was shown that CO2 solubility was much greater in aqueous AMP solutions
than in MEA solutions at loadings between 0.5 and 1.0, which was in agreement with the
behaviour of SHA in respect to primary amines (see §1.2.3). The lower solubility in the
aqueous MEA solution was due to the stable carbamate formation which limited the
stoichiometric loading to 0.5. The formation of the unstable carbamate ion by reaction of
AMP with CO2 was followed by its hydrolysis and thus a solution loading of 1.0 may be
more easily attained. The experimental data reported in that work were used later by Hu
and Chakma (1990) for comparison with the predictions obtained using a mathematical
model developed for the determination of the equilibrium solubility of CO2 in aqueous
AMP solutions. The same experimental method (Jou et al., 1982) was used by Teng and
Mather (1989) for measuring CO2 solubility in 3.43 kmol·m-3 (30.7 wt%) aqueous AMP
solutions at 323 K and CO2 partial pressures varying between 4.32 and 5645 kPa. Solubility
data were correlated using the Deshmukh and Mather model (Deshmukh and Mather,
1981). Based on the work of Sartori et al. (1987), the authors neglected the carbamate
formation. It was shown that the model reproduced the experimental data within the
experimental uncertainty. Tontiwachwuthikul et al. (1991) measured CO2 solubility in 2.0
and 3.0 kmol·m-3 (17.9 and 26.8 wt%) aqueous AMP solutions at 293, 313, 333 and 353 K
and CO2 partial pressures varying between 1.59 and 98.93 kPa using a thermostated
gas/liquid contactor (Muhlbauer and Monaghan, 1957). The authors found a very good
agreement between their data at 313 K and those found in the literature (Roberts and
Mather, 1988a; Sartori and Savage, 1983). A modified Kent-Eisenberg model (Kent and
Eisenberg, 1976) was found to represent experimental data accurately. Haji-Sulaiman and
Aroua (1996) measured CO2 solubility in aqueous 2.0 kmol·m-3 (17.9 wt%) AMP solutions
40
at 303, 313, 323, 333 (1 data point) and 353 K (1 data point) and over CO2 partial pressures
of 0.5 to 100 kPa, by using a thermostated stirred cell reactor. Data were correlated using
the Deshmukh and Mather model. Using a similar experimental setup (Haji-Sulaiman and
Aroua, 1996; Haji-Sulaiman et al., 1998), additional measurements were provided later by
the same research group at 303, 313 and 323 K, but data were not tabulated but graphically
represented (Aroua et al., 2002). Experimental data were graphically compared with
predictions obtained by applying the electrolyte NRTL model (Austgen et al., 1989), using
the AspenPlus software. Jane and Li (1997) measured CO2 solubility in 2.0 kmol·m-3 (17.9
wt%) aqueous AMP solutions at 313 K and a good agreement was found with data reported
by Roberts and Mather (1988a) (5% deviation). Park et al. (2002c) measured CO2 solubility
in 30 wt% aqueous AMP solutions at 313, 333 and 353 K, but experimental data were not
tabulated, only graphically represented. Based on the experimental data, a modified Kent-
Eisenberg model was used to determine the equilibrium constants corresponding to amine
protonation and carbamate hydrolysis. CO2 solubility in aqueous 30 wt% AMP solution
was also measured by Seo and Hong (1996) at 313, 333 and 353 K and data were in good
agreement with those reported by Li and Chang (1994) (see also §1.2.4.2). Kundu et al.
(2003) measured the solubility of CO2 in 18, 25 and 30 wt% aqueous AMP solutions over a
temperature range of 303 to 323 K and over CO2 partial pressures ranging between 3.2 and
94 kPa. The modified Clegg-Pitzer equation was used to correlate and predict equilibria for
this system. Generally, predicted results were found in good agreement with previous
published data (Jane and Li, 1997; Li and Chang, 1994; Seo and Hong, 1996; Teng and
Mather, 1989). Teng and Mather (1990) measured CO2 solubility in 2.0 kmol·m-3 (17.9
wt%) aqueous AMP solutions at 313 and 343 K and a wide range of pressures between
0.162 and 5279 kPa. Data at 313 K agreed well with those reported by Roberts and Mather
(1988a). The authors observed that the solubility of CO2 in AMP solutions was higher than
that in comparable DEA or TEA solutions. Moreover, at CO2 loading higher than unity, the
temperature had little effect on gas solubility, as noted for MDEA solutions (Jou et al.,
1982). Silkenbäumer et al. (1998) measured CO2 solubility in aqueous AMP solutions at
different molal concentrations between 2.43 and 6.242 mol·kg-1 and at temperatures of 313,
333 and 353 K. A model taking into account CO2 absorption coupled with the chemical
41
reaction in the liquid phase was used to correlate experimental data. Activity coefficients
for both molecular and ionic species were calculated from the Pitzer equation (Pitzer,
1973). At low pressures, the authors found a good agreement between the correlation
results and experimental data by Roberts and Mather (1988a) obtained at 313 k and AMP
concentrations of 2.0 and 3.0 kmol·m-3. However, at high CO2 partial pressures, previous
data were systematically higher than the correlation results. A good agreement was also
observed between the correlation results and solubility data by Tontiwachwuthikul et al.
(1991) measured from 293 to 353 K and for AMP concentrations of 2.0 and 3.0 kmol·m-3
(17.9 and 26.8 wt%), excepted for data at 313 K and 3.0 kmol·m-3 where the experimental
partial pressures were higher that the calculated ones. Finally, solubility data by Teng and
Mather (1990) measured for 2.0 kmol·m-3 aqueous AMP solutions and at 343 K were found
to agree well with the correlation results, at low and high pressures. CO2 solubility given by
Yang et al. (2010) at 313 K and pressures between 0.89 and 151.9 kPa were found in good
agreement with those reported by Roberts and Mather (1988a) and Tontiwachwuthikul et
al. (1991). Xu et al. (1992c) developed a mathematical model based on the extended
Debye-Hückel equation for representing CO2 solubility in aqueous AMP solutions. Model
parameters were obtained on the basis of selected experimental data reported by Roberts
and Mather (1988a, b) and Teng and Mather (1989, 1990), because they were measured
using the same method and covered a wide range of amine concentration, temperature and
pressure. The stability constant of carbamate in solution was estimated. According to the
correlation results, the authors concluded that the formation of protonated amine and
bicarbonate ions is the dominant reaction. Carbamate ion concentration was found between
the order of 10-5 and 10-2.
1.2.4.1.2. CO2 absorption in other SHA aqueous solutions
Two works are available for the aqueous CO2-AMPD system. CO2 solubility in 10 and
30 wt% aqueous solutions at 303, 313, and 333 K and over CO2 partial pressures ranging
between 0.6 and 3064 kPa was determined by Baek and Yoon (1998). A comparison
between CO2 solubility in aqueous AMPD solutions and that in aqueous MEA, MDEA and
AMP solutions (Jou et al., 1994; Seo and Hong, 1996) showed that the tendency of the
solubility in AMPD solutions was similar to that in MDEA solutions. At low partial
42
pressures, CO2 solubility was lower in MDEA solutions and became higher at high
pressures. Puxty et al. (2009a) measured CO2 solubility in 1 kmol·m-3 (10 wt%) AMPD
aqueous solutions at 313 K based on a synthetic method and by using a thermostated glass
reactor. Data by Puxty et al. (2009a) were higher than those reported by Baek and Yoon
(1998) at low partial pressures and became lower at higher pressures.
CO2 solubility in the aqueous AEPD system was only measured by Park et al. (2002b)
at 313, 323, and 333 K and over CO2 partial pressures ranging between 1.8 and 2849 kPa.
A comparison with other amines such as MEA (Jou et al., 1995; Park et al., 2002b), AMPD
(Baek and Yoon, 1998), AMP (Seo and Hong, 1996) and MDEA (Jou et al., 1994) showed
that the tendency of CO2 solubility in aqueous AEPD solutions was similar to those in
MDEA and AMPD solutions. At low partial pressures, CO2 solubility in aqueous MEA
solutions was higher than in AMP, MDEA, AMPD or AEPD solutions, but became lower
at higher pressures (more than about 10-90 kPa, depending on the amine type).
The aqueous CO2-AHPD system attracted more attention. CO2 solubility in 10 and 20
wt% aqueous solutions at 313, 323, and 333 K and over CO2 partial pressures ranging
between 21.7 and 1839.8 kPa was first determined by Park et al. (2002a). Solubilities in 10
wt% aqueous AHPD were compared with those in aqueous solutions of MEA (Park et al.,
2002b) and other hindered amines such as AMPD (Baek and Yoon, 1998) and AEPD (Park
et al., 2002b). At partial pressures higher than about 40 kPa, CO2 loading capacity of
aqueous AHPD solutions was higher than that in MEA solutions. Moreover, the loading
capacity of all sterically hindered amines analyzed (AMPD, AEPD and AHPD) was found
to be higher than that in MEA, following the order AHPD > AEPD > AMPD. At lower
partial pressures, CO2 loading capacity in aqueous MEA solutions became higher than in
AHPD. New data for this system at 298 K and aqueous AHPD solution concentration of 10
wt% were reported later by the same research group (Park et al., 2003). Le Tourneux et al.
(2008) measured CO2 solubility in aqueous AHPD solutions of concentrations between
0.15 and 2.5 wt%, at 283, 298 and 313 K and over CO2 partial pressures ranging between
1.91 and 74.8 kPa. The low concentration range was compatible with aqueous solutions in
use in an enzymatic CO2 capture process. It was shown that the enzyme did not influence
43
the CO2 solubility, but only accelerated reaching the equilibrium. Data were correlated
using the modified Kent-Eisenberg model. Additional solubility data for CO2 in 10 wt%
aqueous AHPD solutions were compared with those reported by Park et al. (2003) and it
was shown that data by Park et al. (2003) were lower than those reported by Le Tourneux
et al. (2008). New CO2 solubility data were recently measured by Bougie and Iliuta (2010b)
for concentrations of 0.917, 2, 3 and 4 mol·kg-1, temperatures between 285 and 333 K and
over CO2 partial pressures ranging between 0.314 and 2637.6 kPa. When comparison was
possible, it was shown that data by Bougie and Iliuta (2010b) agreed well with those given
by Le Tourneux et al. (2008) which were obtained using a different experimental setup.
However, several data by Bougie and Iliuta (2010b) disagreed from those reported by Park
et al. (2003; 2002a).
1.2.4.2. CO2 chemical solubility in SHA based mixed solvents
1.2.4.2.1. CO2 absorption in AMP based mixed solvents
Mixed solvents represent a combination of chemical and physical pure solvents. The
use of blended alkanolamines for the removal of acid gases from gas streams has become
very attractive because of their advantages over traditional treating solvents (single aqueous
amine solutions). The mixed solvents combine the advantages of each amine present in the
mixture: the fast reactivity of primary or secondary alkanolamine (e.g. MEA, DEA) is
coupled with the high absorption capacity and low solvent regeneration cost of tertiary (e.g.
MDEA) or SHA (e.g. AMP) amines.
Roberts and Mather (1988b) measured the CO2 solubility in a mixed solvent consisting
of AMP (16.5 wt%), TMS (32.2 wt%), and water at 313 and 373 K and at CO2 partial
pressures between 2.63 and 6050 kPa. The solubility in the mixed solvent was compared
with the solubility in an aqueous solution of equivalent amine concentration. It was shown
that the solubility of CO2 was significantly lower in the mixed solvent than in the aqueous
AMP solvent at low acid gas partial pressures. With the increase of the CO2 partial pressure
this difference in the solubility decreased and at high partial pressures (much larger at 313
K (around 1400 kPa) than at 373 K (around 120 kPa)) the solubility of CO2 became larger
in the mixed solvent. Li and Chang (1994) measured CO2 solubility in aqueous AMP +
44
MEA solutions at 313, 333, 353 and 373 K for various ratios AMP/MEA for a total amine
concentration of 30 wt%. Based on the experimental data, a modified Kent-Eisenberg
model was used to determine the equilibrium constants corresponding to AMP and MEA
protonation and MEA carbamate hydrolysis. Park et al. (2002c) measured CO2 solubility in
aqueous AMP + MEA and AMP + DEA solutions at 313, 333 and 353 K, keeping the total
amine concentration at 30 wt%. Experimental data were not tabulated, only graphically
represented. As also observed by Li and Chang (1994), the equilibrium curve 2
/COP α for
the system MEA + CO2 crossed the one corresponding to AMP + CO2 system. At low CO2
partial pressures and up to a CO2 loading of about 0.5, the addition of AMP to an aqueous
MEA solution led to a decrease of CO2 solubility. AMP addition favoured CO2 solubility at
higher pressures. These observations agreed to the behaviour of SHA which can reach CO2
loadings up to 1 due to carbamate formation followed by its hydrolysis and conversion to
bicarbonate, coupled with the high reactivity of MEA which formed stable carbamate.
Being less reactive than MEA, DEA did not have a similar influence on the solubility of
CO2 in AMP solutions. While MEA addition to an aqueous AMP solution resulted in the
increase of CO2 solubility, the addition of DEA did not. It was observed that at low
loadings, DEA nearly had the same tendency to absorb CO2 like AMP. However, at higher
loadings DEA behaved in the same way like MEA due to the formation of stable
carbamates. CO2 solubility in aqueous AMP + DEA solutions at 313, 333 and 353 K was
also measured by Seo and Hong (1996) for the same total amine concentration of 30 wt%
but different ratios between AMP and DEA than those considered by Park et al. (2002c).
However, Park et al. (2002c) did not compare their data with those reported by Seo and
Hong (1996), even if they mentioned this reference in their work. Even if Park et al.
(2002c) did not report any tabulated data, the results of these two studies were found to
agree well. In order to test the predictive capability of the model used to correlated
experimental data of CO2 solubility in aqueous AMP solutions (described previously in
§1.2.4.1.1), Silkenbäumer et al. (1998) measured CO2 solubility in aqueous mixtures of
AMP (1.266 mol·kg-1) and MDEA (1.278 mol·kg-1) at 313 K and for total pressures
between 12.5 and 4020 kPa. It was found that at constant total pressure, the addition of
AMP, a stronger base than MDEA, to an aqueous MDEA solution increased CO2 loadings.
45
The model based only on data for the aqueous systems CO2 + MDEA and CO2 + AMP
predicted well the CO2 solubility in the aqueous mixed solvent. Murietta-Guevara et al.
(1998) reported the solubility of CO2 in aqueous mixtures of AMP + DEA at 313 and 373
K for a total amine concentration of 30 wt%, with different compositions of the individual
alkanolamines. Data analysis revealed a general trend: CO2 solubility increased with the
increase of AMP concentration. Using the same apparatus and methodology (Murrieta-
Guevara et al., 1992, 1994; Murrieta-Guevara et al., 1998), Rebolledo-Libreros and Trejo
(2004) measured CO2 solubility in aqueous solutions containing three amines: MDEA (32.5
wt%), DEA (12.5 wt%) and AMP (4, 6, and 10 wt%). The authors found that the increase
of AMP concentration in a mixture DEA + MDEA led to the increase of CO2 solubility.
Aroua et al. (2002) measured CO2 solubility in aqueous AMP and MDEA mixtures (2.0
kmol·m-3 total amine concentration in all measurements) at 303, 313 and 323 K and over
CO2 partial pressures of 0.1 to 100 kPa. Data were not tabulated; an example was given
graphically for 303 K and compared with predictions obtained by applying the electrolyte
NRTL model (Austgen et al., 1989) using the AspenPlus software. You et al. (2008)
studied the effect of AMP addition on CO2 absorption in aqueous ammonia at 298 K. The
mixed solvent contained 10 wt% ammonia and 1 wt% AMP. Data were not tabulated and
they were expressed graphically as CO2 removal efficiency of the absorbent from a feed gas
containing 15 vol% CO2 and 85 vol% N2. It was shown that AMP addition led to the
reduction of ammonia vaporisation and slightly increased CO2 absorption capacity. Yang et
al. (2010) measured CO2 solubility in aqueous mixtures containing AMP and Pz (as
activator) at 313, 333 and 353 K, and pressures up to 139.9 kPa. AMP concentrations in the
mixed solvent were 2 and 3 kmol·m-3 (17.9 and 26.8 wt%), while Pz concentrations were
0.5, 1 and 1.5 kmol·m-3 (4.3 to 12.9 wt%). It was observed that at constant temperature and
total amine concentration, CO2 solubility increased with increasing partial pressure. At
constant temperature and AMP concentration, Pz addition led to an increase in CO2
solubility.
1.2.4.2.2. CO2 absorption in other SHA based mixed solvents
You et al. (2008) studied the effect of AMPD, AEPD and AHPD (THAM) addition on
CO2 absorption in aqueous ammonia (AM) at 298 K. The mixed solvent contained 10 wt%
46
ammonia and 1 wt% AMPD, AEPD or AHPD. Data were not tabulated and they were
expressed graphically as CO2 removal efficiency of the absorbent from a feed gas
containing 15 vol% CO2 and 85 vol% N2. It was shown that the addition of all SHA tested
led to the reduction of ammonia vaporisation and maintained or slightly increased CO2
absorption capacity. CO2 removal capacity had the following trend (this includes AMP
effect described in the previous section): AM < (AM + AMPD) < (AM + AEPD) < (AM +
AMP) < (AM + AHPD). The positive effect of SHA addition was attributed to
intermolecular interactions between the alkanolamines and CO2. The loss of ammonia
decreased as following: AM > (AM + AMPD) > (AM + AEPD) > (AM + AMP) > (AM +
AHPD). The effect of SHA addition was attributed to the interactions between the hydroxyl
groups of SHA and ammonia via hydrogen bonding. Lal et al. (1998) measured CO2
solubility in an aqueous mixed solvent containing 55 wt% 2-PE and 10 wt% sulfolane at
313 and 373 K and over CO2 partial pressures ranging between 0.274 and 5548 kPa. The
same research group (Jou et al., 1998) also reported CO2 solubility in the same mixed
solvent but at a different concentration, namely 45 wt% 2-PE and 40 wt% sulfolane, at 298,
313, 343, 373 and 403 K and over a very large CO2 partial pressures range between
0.00156 and 18900 kPa. The authors (Jou et al., 1998) mentioned that 50% of their reported
data “were determined in 1981 using a wet chemical analysis and the other values were
determined in 1993 mainly using chromatographic analysis”. However, it was not clear if
these data have already been published elsewhere because the corresponding references
were not given. The formation of a second liquid phase consisting in almost pure sulfolane
was noted at certain conditions. The presence of the physical solvent (sulfolane) led to
loadings larger than unity. Li and Mather (1998) used simplified Clegg-Pitzer equations
(Clegg and Pitzer, 1992) to correlate solubility data of CO2 in this aqueous mixed solvent
containing 45 wt% 2-PE and 40 wt% sulfolane. Bougie and Iliuta (2010b) recently studied
the effect of Pz addition (as activator) on CO2 absorption in AHPD aqueous solutions
between 288 and 333 K. AHPD concentration in the mixed solvent was varied from 1.1 to
4.2 mol·kg-1, while Pz concentration was varied from 0.01 to 0.66 mol·kg-1. It was shown
that at constant total amine concentration and CO2 partial pressure, an increase in
temperature led to a decrease of CO2 loading. At constant temperature, an increase in the
47
total amine concentration led to a decrease of CO2 solubility. As expected, at constant
temperature the Pz addition in an aqueous AHPD solution increased the CO2 loading
capacity.
1.2.4.3. CO2 physical solubility in single and mixed solvents
Physical solubility data of acid gases (like CO2 and H2S) in single and mixed amine
solutions, usually expressed in term of Henry’s law constants, gasH , represent key
parameters needed for the design of absorption scrubbing equipments. Henry’s law
constants are particularly useful to calculate the CO2 diffusion coefficient, gasD in solution
from experimental values of the ratio 1/2 /gas gasD H . However, because of the gas reaction
within the amines, the genuine gas physical solubility cannot be measured directly. Henry’s
law constants in these solutions can be determined by the application of the N2O analogy
method (Li and Lai, 1995; Tsai et al., 2000; Versteeg and Vanswaaij, 1988; Wang et al.,
1992; Xu et al., 1991), by using N2O and CO2 solubility in water and N2O solubility in the
single or the mixed solvent. Relying on artificial neural networks, Bensetiti et al. (1999)
used an exhaustive N2O solubility database for developing correlation for N2O solubility in
water, AMP, DEA, MDEA, MEA and their mixtures. Combined with the N2O analogy
method, this correlation allowed the calculation of CO2 solubility in single or blend
solutions over wide ranges of amine concentrations and temperatures.
Saha et al. (1993) reported CO2 physical solubility data in aqueous AMP solutions of
concentrations between 0.5 and 2.0 kmol·m-3 (4.5 and 17.9 wt%) at 288.5, 293, 298, and
303 K. It was observed that CO2 solubility decreased with the increase of temperature. At
constant temperature, the solubility decreased when the amine concentration increased. The
same system was also studied by Mandal et al. (2005; 2004) who measured N2O solubility
at 293, 298, 303, 308, and 313 K and amine concentration between 2.0 and 3.0 kmol·m-3
(both papers contain the same estimated CO2 solubility data in aqueous AMP solutions).
For 2.0 kmol·m-3 AMP aqueous solutions, data by Mandal et al. (2005; 2004) agreed well
with those reported by Saha et al. (1993) (mean deviation of 2.8%).
Li and Lai (1995) used a similar apparatus as Saha et al. (1993) in order to determine
physical CO2 solubility in aqueous mixed AMP + MEA solution at 303, 308 and 313 K.
48
Mandal et al. (2005) estimated CO2 solubility in the same aqueous system, AMP + MEA, at
293, 298, 303, 308, and 313 K. In both works the amine concentration was kept at 30 wt%
in the mixed solvent, but the ratios between AMP and MEA were different. A comparison
of solubility data is given in Figure 1.5. Data by Mandal et al. (2005) were constantly lower
than those given by Li and Lai (1995). The highest deviations of data by Mandal et al.
(2005) (from those reported by Li and Lai (1995)) were observed at 313 K (e.g. 11.9% at
30 wt% AMP). However, it was observed that for constant total amine concentration, CO2
solubility decreased with the increase of temperature. At constant temperature, CO2
solubility decreased with the increase of AMP concentration.
Figure 1.5. Henry’s law constant of CO2 in aqueous AMP + MEA mixtures for a total amine content of 30 wt%.
Physical CO2 solubility in aqueous mixed AMP + DEA solutions was studied by Li
and Lee (1996) at 303, 308 and 313 K and by Mandal et al. (2004) at 293, 298, 303, 308,
and 313 K. In both works, the total amine concentration in the mixed solvent was kept at 30
wt%. A comparison of solubility data is given in Figure 1.6. Data by Mandal et al. (2004)
were constantly lower than those given by Li and Lee (1996). For a solution of 24 wt%
AMP, the absolute deviation of data by Mandal et al. (2004) (from those reported by Li and
Lee (1996) was 9.6%. As a general trend, for constant total amine concentration, CO2
49
solubility decreased with the increase of temperature. At constant temperature, CO2
solubility increased with the increase of AMP concentration.
Figure 1.6. Henry’s law constant of CO2 in aqueous AMP + DEA mixtures for a total
amine content of 30 wt%.
Baek et al. (2000) measured N2O solubility in 10, 20, and 30 wt% aqueous AMPD
solutions at 303, 313, and 323 K. Data can be used to determine CO2 physical solubility in
these amine solutions. Le Tourneux et al. (2008) measured N2O solubility in aqueous
AHPD solutions of concentrations between 0.15 and 10 wt%, at 283.15, 298.15 and 313.15
K. Data were used to estimate Henry’s law constant for CO2 in the corresponding AHPD
aqueous solutions. Paul et al. (2009c) estimated physical CO2 solubility in aqueous AHPD
solutions of concentrations between 2.17 and 21.7 wt%, at 298, 303, 313 and 323 K and
atmospheric pressure. Data were correlated as a function of temperature and amine
concentration. Bougie and Iliuta (2010b) measured N2O solubility in AHPD + Pz mixed
solvent at 288, 298, 313 and 333 K. AHPD concentration in the mixed solvent was varied
from 1.1 to 4.2 mol·kg-1, while Pz concentration was varied from 0.1 to 0.6 mol·kg-1. Data
can be used to determine CO2 physical solubility in the mixed solvent.
50
1.2.5. Absorption kinetics
Kinetics data represent essential information in CO2 absorption. In order to improve
CO2 capture, aqueous amine solutions not only require high absorption capacity but also an
important absorption rate. For SHA applications in CO2 separations, knowledge about the
reaction mechanism and kinetic constants for various SHA is of major importance. Even
though the available kinetic reviews (Mahajani and Joshi, 1988; Vaidya and Kenig, 2007;
Versteeg et al., 1996) offer detailed description on possible kinetic mechanisms between
CO2 and primary, secondary as well as tertiary amine solutions, only very limited
information on SHA are reported. We consider therefore that bringing together all kinetic
available information related to SHA is highly needed.
Absorption rate of CO2 in aqueous amine solution is usually described by a simple
second-order reaction or by the zwitterion mechanism. The expression for the second-order
reaction is given by:
BA2amineCO 2
CCkr =− (1.27)
while with the zwitterion mechanism:
...
1 1
OH1-
OH2OH
1-
OH2B
1
AM22
BAamineCO
2
2
2
++++
=
−
−
−
−
Ckkk
Ckkk
Ckkkk
CCr
(1.28)
It should be noted that the second term at the denominator contain kinetic parameters
involved in the deprotonation of the zwitterion by bases in solution. The contribution of
each base depends on its concentration as well as how strong the base is. Additional terms
can therefore be present if mixtures of more than one amine are used. This mechanism also
explains the shift in the order with respect to the amine often observed in kinetic
experiments. For the same amine aqueous system and temperature, it should be expected
that the values of k2 determined from each of the Eqs. (1.27) and (1.28) are not exactly the
same, because other kinetic constants are determined simultaneously in the zwitterion
mechanism. However, these values should be of the same magnitude, as demonstrated by
Shen et al. (1991). Values of the kinetic constants for various SHA (except for AMP) and
51
AMP, together with the corresponding temperature and amine concentration ranges, are
indicated in Tables A.20 and A.21 respectively.
1.2.5.1. Single AMP systems
AMP is the most popular SHA; it is the reason why it will be discussed in the
following two sections, separately from the other SHA. Its high CO2 loading capacity was
first pointed out by Sartori and Savage (1983). Since, a high amount of works was found in
the literature concerning kinetics of AMP. More than 15 papers were found giving details
on the reaction mechanism and/or kinetic constants on single and blended aqueous amine
solutions.
Chakraborty et al. (1986) studied the kinetics between pure CO2 and aqueous AMP
solutions at 315 K. The authors assumed that the forward reaction rate would be first order
in respect to both CO2 and AMP. A value as low as 100 m3·kmol-1·s-1 was found for k2.
However, the concentrations of the solutions used were not given and only one temperature
was considered, which is not quite sufficient to obtain reliable kinetic constants. Yih and
Shen (1988) mentioned that although Sartori and Savage (1983) have noted that steric
hindrance generally has an adverse effect on the CO2-amine reaction rate constants, as
indicated from data by Sharma (1965), the above value of k2 obtained by Chakraborty et al.
(1986) seemed too low in comparison with conventional amines. Therefore, the research by
Yih and Shen (1988) was undertaken to investigate the kinetic order with respect to both
CO2 and AMP and to obtain the second-order forward rate constant at 313 K.
Concentrations of 0.258-3.0 kmol·m-3 were considered. The authors found that the reaction
was first order in respect to both CO2 and AMP, as it was also mentioned in Chakraborty et
al. (1986). The new k2 value of 1270 m3·kmol-1·s-1 obtained in their study was about 6
times lower than the value of k2 for CO2-MEA, which confirmed Sartori and Savage (1983)
statement that steric hindrance has an adverse effect on the CO2-amine rate constants. Alper
(1990) investigated the mechanism and kinetics of the reaction between aqueous solutions
of CO2 and AMP (0.013-1.5 kmol·m-3) at 278-298 K. Experiments were also carried out
with MEA solution. They found that the reaction was first order in respect to CO2 but 1.14-
1.15 in respect to AMP. A fractional order between 1 and 2 would be expected if the
52
deprotonation of the zwitterion was not instantaneous. However, kinetic constants were
extracted as if the order with respect to AMP was unity. The corresponding second-order
rate constants at 298 K were found to be 520 and 5545 m3·kmol-1·s-1 for AMP and MEA,
respectively, with the corresponding activation energies of 41.7 and 46.7 kJ·mol-1. The
predicted rate constant at 313 K was 1165 m3·kmol-1·s-1, which agreed well with the value
of 1270 m3·kmol-1·s-1 reported by Yih and Shen (1988). Bosch et al. (1990) mentioned that,
following their analysis of the paper of Chakraborty et al. (1986) carried out in Bosch et al.
(1989), the CO2 absorption rates observed in sterically hindered amine solutions could
probably be explained satisfactorily with the zwitterion mechanism. In order to verify this
hypothesis, new CO2 absorption data for aqueous AMP solutions have been collected and
were presented in their paper (Bosch et al., 1990). Experimental work was conducted at 298
K for AMP solutions of 0.202 to 2.373 kmol·m-3. Unfortunately, from the observed
decrease of CO2 pressure with time, it was concluded that for none of the absorption
experiments the simple pseudo-first-order conditions prevailed. The reaction rate constant
for the zwitterion formation, k2, could not be calculated accurately (estimated inaccuracy of
100%); however, a value of 10000 m3·kmol-1·s-1 was reported at 298 K. This value seemed
quite high since steric considerations should have given a value of k2 for AMP smaller than
that for MEA, as reported in Alper (1990). In the paper of Saha et al. (1995), the
mechanism and kinetics of the reaction between CO2 and AMP aqueous solution were
investigated at 294-318 K. The reaction was found to be first order with respect to both
CO2 and AMP. Values of the second order rate constant were found to be 439, 687, 1179
and 1650 m3·kmol-1·s-1 at 294, 301.5, 311.5 and 318 K, respectively, in the amine
concentration range 0.5-2.0 kmol·m-3. These results were in close agreement with those
reported by Yih and Shen (1988) and Alper (1990), even though the latter adopted a
completely different methodology. The corresponding value of the activation energy was
found to be 43 kJ·mol-1. The study by Xu et al. (1996) was among the first to treat
absorption data over large concentration and temperature ranges in AMP solutions using
the zwitterion mechanism. Reaction rate constants for the reaction between CO2 and AMP
were determined from measurements of the absorption rate of CO2 into aqueous AMP and
non-aqueous (1-propanol + AMP) solutions. The kinetic parameters for aqueous AMP
53
solutions were obtained for temperatures from 288 to 318 K over an AMP concentration
range of 0.17-3.5 kmol·m-3, and at 298 K over a concentration range of 0.40-3.55 kmol·m-3
of AMP in 1-propanol solutions. The absorption of CO2 in AMP + l-propanol was studied
to help confirming the validity of using the zwitterion mechanism to interpret the kinetics
between CO2 and AMP. The authors found that the partial order in respect to AMP was
larger than unity in both solutions. In aqueous solutions, the reaction orders for AMP varied
from 1.15 at 288 K to 1.32 at 318 K, while it was 1.28 in 1-propanol solutions at 298 K.
The second-order rate constant, k2, and the kinetic constants 1OH2 2 −kkk and 1AM2 −kkk
were correlated as a function of temperature using Arrhenius type equations. The authors
compared their results with data from literature using the overall pseudo-first order reaction
rate constant. Their values of kov obtained at 298 K were in good agreement with those of
Bosch et al. (1990) and Alper (1990) at lower concentrations of AMP, but were slightly
higher when the concentration of AMP was greater than about 0.7 kmol·m-3. Also, the kov’s
measured at 288 K (Xu et al., 1996) were slightly higher than those determined from
Alper's results; at 313 K, the values were somewhat lower than those of Yih and Shen
(1988). The use of kov’s as a basis of comparison assumed that all experiments were carried
out in the pseudo-first order reaction regime, what may have not been the case in Bosch et
al. (1990). Messaoudi and Sada (1996) investigated the absorption of CO2 into aqueous
AMP solutions (0.5 to 2.0 kmol·m-3). The reaction was found to be first order with respect
to both CO2 and AMP. The second-order reaction rate constants at 293, 303 and 313 K
were found to be 190, 369 and 740 m3·kmol-1·s-1, respectively. These values were
constantly lower than those of Saha et al. (1995), although almost the same concentration
and temperature ranges were considered. Mandal and Bandyopadhyay (2005) performed
experimental and theoretical investigation of the simultaneous absorption of CO2 and H2S
in aqueous solutions of AMP + DEA. Kinetic information concerning AMP was taken from
Mandal’s thesis who reported an equation for the second order rate constant, k2. It was
assumed that the temperature and concentration ranges considered were adequately covered
by this equation. Kinetic constants calculated with that equation were in good agreement
with the values reported by Saha et al. (1995). Ali (2005) studied the effect of mixing AMP
with a primary amine (MEA) and a secondary amine (DEA) on the kinetics of the reaction
54
with carbon dioxide in aqueous media. Experimental work was conducted at 298, 303, 308,
and 313 K using aqueous AMP solutions of concentrations varying between 0.05 and 0.35
kmol·m-3. For blended aqueous solutions, AMP + MEA and AMP + DEA, various amine
concentrations were used and MEA/AMP and DEA/AMP molar ratios of (0.05, 0.09, 0.15,
0.22 and 1.08) and (0.06, 1.01 and 19) were respectively selected. A model based on the
zwitterion mechanism for all the amines involved (AMP, MEA, and DEA) was applied.
Blending AMP with either MEA or DEA resulted in overall pseudo-first-order reaction rate
constant values (kov) larger than the sum of the kov values corresponding to the respective
pure amines. This should be due to the role played by one amine in the deprotonation of the
zwitterion of another one. The kov values of Ali (2005) at a given temperature were found
comparable with those reported by Alper (1990) (using the stopped-flow technique), Xu et
al. (1996) (derived from absorption experiments using a stirred cell reactor) and Bosch et
al.(1990). The activation energy for the zwitterion formation step for AMP (a primary
amine) was found closer to that for MEA (a primary amine) than that for DEA (a secondary
amine). This appeared to suggest that the nature of the amine (i.e., whether it is primary or
secondary) had a great bearing on the energy barrier that had to be overcome to form the
zwitterion intermediate in the first step. For the aqueous AMP system, the activation energy
value for the zwitterion formation step obtained in Ali (2005) (41.9 kJ·mol-1) was found to
be very close to that obtained by Alper (1990) (41.7 kJ·mol-1) and comparable to that
obtained by Saha et al. (1995) (43.0 kJ·mol-1), despites the fact that these two last studies
treated their data using an overall second order reaction. Also, the Ea value obtained by Xu
et al. (1996) (24.3 kJ·mol-1) was found to be lower, while the data obtained by Messaoudi
and Sada (1996) (51.5 kJ·mol-1) was found to be higher, as compared to the value obtained
by Ali (2005). An analysis of the kinetic parameter involved in the zwitterion mechanism
showed that MEA had higher deprotonating ability than AMP but the AMP-DEA analysis
was quite ambiguous. The authors (Ali, 2005) succeeded to obtain almost the same kinetic
parameters for all three systems involving AMP (single AMP, AMP + DEA, AMP +
MEA). Reported k2 values for AMP were found to be very close to those of Saha et al.
(1995), while values of 1OH2 2 −kkk , 1AM2 −kkk , and 1MEA2 −kkk were, respectively,
55
higher, lower, and higher than those reported in the literature (Seo and Hong, 2000; Xiao et
al., 2000; Xu et al., 1996).
In Choi et al. (2007), experiments were carried out to investigate the characteristics of
CO2 absorption rate in AMP solution with small additions of hexamethylenediamine
(HMDA), MDEA or piperazine. Additive concentrations of 1, 3, and 5 wt% were added for
each 30 wt% AMP solution. To check the validity of the method, the authors studied the
CO2-AMP reaction and found a first order dependence with respect to CO2 and AMP. A
value of 731 m3·kmol-1·s-1 for the second-order reaction rate constant (k2) at 313 K was
obtained, which was in good agreement with that reported by Messaoudi and Sada (1996)
(740 m3·kmol-1·s-1). It should be noted that the values of Messaoudi and Sada (1996) were
well below any other reported k2 values in the literature. Choi’s experiments showed that
the addition of HMDA, MDEA or piperazine into AMP solutions increased the absorption
rate as compared to AMP alone. Surprisingly, authors found that MDEA addition in AMP
solution produced a larger or somewhat equivalent increase in the absorption rate than Pz
addition. No explanations of these results were given. The same research group also
published a study concerning CO2 absorption into aqueous AMP + MEA solutions at 293,
303 and 313 K (Choi et al., 2009). The reported kinetic constants concerned the blended
solutions and not AMP alone. However, they found that MEA was more reactive than
AMP.
1.2.5.2. Blended AMP systems
The presence of a second amine in solution can enhance the deprotonation mechanism
of the zwitterion. A new kinetic constant should be added: 1#2 Am2 −kkk which represent the
contribution to the deprotonation of the zwitterion by this new base in solution.
Kinetics of CO2 in aqueous AMP + DEA solutions at 303, 308 and 313 K was
studied in a wetted-wall column by Wang and Li (2004). AMP concentration were 1.0 and
1.5 kmol·m-3, with DEA addition of 0.1, 0.2, 0.3 or 0.4 kmol·m-3. A hybrid rate model was
applied: second-order reaction for AMP and zwitterion mechanism for DEA. This model
succeeded to represent experimental data with 7.2% deviation. Results of k2 for AMP were
56
reported by an equation. Comparison of calculated k2 values indicated a good agreement
with the values given by Saha et al. (1995) and Ali (2005).
CO2 absorption rate into aqueous solution of AMP + MEA was investigated by Xiao et
al. (2000) at 303, 308, and 313 K, using a wetted-wall column. Ten systems where 1.5 and
1.7 kmol·m-3AMP was mixed with various MEA concentrations (0, 0.1, 0.2, 0.3, and 0.4
kmol·m-3) were studied. CO2 absorption into 0.9 kmol·m-3 aqueous AMP at 313 K has been
carried out to check the validity of the method; kov obtained was found to be 728 s-1, which
was in a good agreement with data reported by Xu et al. (1996). kov values at 303 and 313
K for 1.5 and 1.7 kmol·m-3AMP were also found to be in good agreement with those of
Saha et al. (1995) and Xu et al. (1996), respectively. In order to represent the kinetic data,
authors suggested a reaction model consisting of a first order reaction mechanism for MEA
and a zwitterion mechanism for AMP. Comparing the kov calculated using the zwitterion
regression with the experimental kov, large deviations were found at 1.7 kmol·m-3 AMP +
MEA at 308 and 313 K and these deviations seemed to increase as MEA concentration
increased. Calculated kinetic constants for MEA and AMP were expressed as a function of
temperature. The comparison between the kinetic constants for AMP and those obtained by
Xu et al. (1996) showed a good agreement for k2 values only; the other kinetic constants
were quite different. This may come from the fact that in AMP + MEA systems, a new
parameter ( 1MEA2 −kkk ) modified the value of the other kinetic parameters obtained by a
non-linear regression. It should be noted that the values of this new kinetic parameter
involving MEA in the deprotonation of AMP zwitterion are lower at 303 and 308 K than
1AM2 −kkk , which seems inconsistent with the fact that MEA kinetics was well described
by a second order overall reaction in the literature, indicating that MEA usually
deprotonated its zwitterion almost instantaneously.
Seo and Hong (2000) investigated the absorption of CO2 into AMP + Pz solutions at
303 and 313 K using a wetted-sphere absorption apparatus. The concentration of AMP was
in the range of 0.55-3.35 kmol·m-3 and Pz additions of 0.058, 0.115, and 0.233 kmol·m-3
were made for each AMP solution. To validate the apparatus, kinetics of aqueous single
solutions of AMP was investigated under the same concentration and temperature ranges.
57
The reaction orders in respect to AMP were determined and they varied from 1.29 at 303 K
to 1.32 at 313 K, which could be explained by the zwitterion mechanism. Kinetic constants
were reported for single AMP aqueous systems, as well as for the AMP + Pz aqueous
systems. Concerning the system AMP + H2O, the second order rate constant for AMP, k2,
at 313 K was found to be in good agreement with the results of Yih and Shen (1988) and
Xu et al. (1996). Concerning the blended AMP + Pz + H2O system, kinetic constants
involving AMP were quite different from what have been found for the single AMP + H2O
system in the same work (Seo and Hong, 2000), but also from the work of Xu et al. (1996)
and Xiao et al. (2000). Relatively high CO2 partial pressures were used in Seo and Hong
(2000), resulting, according to Bishnoi and Rochelle (2000), in substantial depletion of Pz
at the gas-liquid interface that could have altered kinetic results. It could be seen however
that the kinetic constants 1Pz2 −kkk were very high, indicating that Pz facilitated AMP
zwitterion deprotonation that may have promoted the overall CO2 absorption rate. Pz
promoting effect in AMP solutions was also later reported by Samanta and Bandyopadhyay
(2009).
In Sun et al. (2005), the reaction kinetics of the absorption of CO2 into mixed aqueous
solutions of AMP and PZ were investigated using a wetted-wall column at 303, 308 and
313 K. The aqueous blends chosen for this kinetic study were 1.0 and 1.5 kmol·m-3 AMP
with various Pz concentrations (0.1, 0.2, 0.3, and 0.4 kmol·m-3). A second-order reaction
for the reaction of CO2 with Pz and a zwitterion mechanism for the reaction of CO2 with
AMP were considered to model the kinetic data. Arrhenius type equations were given for
each calculated kinetic parameter. Reported k2 values were higher than literature values
(Ali, 2005; Saha et al., 1995; Xiao et al., 2000) but similar to k2 values obtained by Seo and
Hong (2000) for the blended system AMP + Pz. All the other kinetic parameters related to
the deprotonation of AMP zwitterion given by Sun et al. (2005) were in disagreement with
what have been presented so far. The equation for the kinetic parameter 1Pz2 −kkk even
seems to be misprinted because the calculations give odd values.
Following the analysis of all these works concerning AMP, it seems that no clear
consensus was found concerning reliable kinetic constants. Selecting the right kinetic
constant and mechanism becomes even more ambiguous because two different sets of
58
kinetic parameters, taken either from Saha et al. (1995) (second-order reaction) or Xu et al.
(1996) (zwitterion mechanism), have successfully been applied in simulation/modeling
(Mandal et al., 2003a; Saha et al., 1999; Zhang et al., 2007).
Zwitterion mechanism could explain the order deviation for AMP found in several
works, as well as an apparent first order, but kinetic constants can take various values as
they are obtained simultaneously (Bosch et al., 1990; Seo and Hong, 2000; Xiao et al.,
2000). Utilisation of the same kinetic parameters for various systems (single and blended
aqueous AMP solutions) was successfully made by Ali (2005) but it would be interesting to
extend that study using higher AMP concentrations.
Another parameter that can influence the scattering of k2 values found for AMP or any
other amine may be the thermal effect associated with CO2 absorption (see §1.2.2.2.5).
Camacho et al. (2005) studied the kinetics of CO2 absorption in AMP solutions by
considering this thermal effect at the gas-liquid interface. All experiments were performed
using a stirred gas-liquid contactor. The variables considered were the AMP concentration
(0.1-3.0 kmol·m-3) and temperature (288-313 K). An iterative process has been used to
determine the interface temperature that was found significantly higher than the bulk
temperature. At 313 K, they obtained a kinetic constant k2 of 161.0 m3·kmol-1·s-1. This
value was of the same order as that reported by Chakraborty et al. (1986), but lower than
what have been presented elsewhere in the literature. The authors mentioned that these
different research groups that have worked in CO2 absorption in AMP solutions did not
consider thermal effects what caused these deviations. In the future, it should then be
interesting to see more kinetic publications taking into account or addressing this thermal
effect.
1.2.5.3. Other SHA systems
1.2.5.3.1. 2-PE systems
Shen et al. (1991), Xu et al. (1993a) and Paul et al. (2009a) studied the kinetics
between CO2 and aqueous 2-PE solutions at 313 K, 283-313 K and 303-323 K,
59
respectively. Xu et al. (1993a) performed experiments in a stirred-cell, while Shen et al.
(1991) and Paul et al. (2009a) used a wetted-wall column.
Shen et al. (1991) found the reaction to be first-order with respect to both CO2 and 2-
PE. The second-order forward rate constant at 313 K had a value of 195 m3·kmol-1·s-1 and
was extracted for amine concentration range of 0.218-1.0 kmol·m-3. Such a low
concentration range may not be sufficient for a reliable industrial-applicable kinetics study.
The result was much lower than that of Xu et al. (1993a) (k2 of 1468 m3·kmol-1·s-1 at 313
K). These values, however, do not have the same meaning, although their units are the
same, since Xu et al. (1993a) applied the zwitterion mechanism to treat their data. If the
second-order rate constant of Xu et al. (1993a) at 313 K was correlated using the method of
Shen et al. (1991), its value would become 1207 m3·kmol-1·s-1 with an absolute error as
high as ±13%, which would be still larger than the value of Shen et al. (1991).
In the study of Xu et al. (1993a), the authors made a comparison of the kinetics of 2-PE
versus AMP at 293 and 313 K. They showed that the apparent kinetic rate constants of 2-
PE were dramatically lower than those of AMP. This signifies that the reaction of CO2 with
2-PE was not as fast as that with AMP. A similar observation was revealed under other
experimental conditions (Sartori and Savage, 1983). However, the k2 value of 1468
m3·kmol-1·s-1 at 313 K reported by Xu et al. (1993a) was above the second-order rate
constant for AMP at the same temperature reported in the literature (Mandal and
Bandyopadhyay, 2005; Seo and Hong, 2000; Xu et al., 1996).
Paul et al. (2009a) studied the kinetics of CO2 absorption in 2-PE solutions of 0.14-
1.13 kmol·m-3. The reaction order was found to be between 1.10 and 1.12 with respect to
amine, which could be explained by the zwitterion mechanism, but the authors treated their
results by considering a second-order reaction. The second-order rate constants, k2, were
696, 1147, and 2047 m3·kmol-1·s-1 at 303, 313, and 323 K, respectively, with an activation
energy of 45.2 kJ·mol-1. The results at 303 and 313 K were lower to those reported by Xu
et al. (1993a) and may therefore reconcile the fact that 2-PE reacts slower than AMP.
However, the results reported by Paul et al. (2009a) should be considered with care as
almost all their Hatta numbers were higher than the calculated instantaneous enhancement
60
factor (E∞). An intermediate regime should have been presented instead of the desired fast
pseudo-first-order regime, even if the extraction of reliable kinetics results would have been
much more difficult (Derks et al., 2006).
Considering these three studies, zwitterion mechanism seems to describe well the
absorption of CO2 in 2-PE solutions, but more studies would be necessary to obtain reliable
kinetic constants (k2 and zwitterion deprotonation kinetic constants).
1.2.5.3.2. AEPD systems
Only the publication of Yoon et al. (2002a) was found in the open literature concerning
the kinetics of reaction between aqueous AEPD and CO2. The study was performed at
305.15, 313.15, and 318.15 K for aqueous solutions from 5 to 25 wt% AEPD, using a
wetted-wall column absorber. As commonly observed in kinetic studies between CO2 and
alkanolamines (Gianetto et al., 1986), a first order rate dependence in respect to CO2 was
found. Zwitterion mechanism was used to treat the experimental data. Three reaction rate
parameters, k2, 1OH2 2 −kkk , and 1AM2 −kkk , were determined simultaneously by a nonlinear
regression method and values were reported at each temperature. The parameter
1OH2 - −kkk was neglected because the contribution of the hydroxyl ion was considered
negligible. Arrhenius type equations have been used here to correlate the kinetic
parameters:
−=⋅⋅ −−
K/7820730.31exp skmolm/ 113
2 Tk
(1.29)
−=⋅⋅ −−
− K/22843316.72exp skmolm/ 126
1
OH2 2
Tkkk
(1.30)
−=⋅⋅ −−
− K/4809902.21exp skmolm/ 126
1
AM2
Tkkk
(1.31)
The activation energy (based on k2) was found to be 65.0 kJ·mol-1 with an absolute
error of 2%. It was observed that the overall absorption rate constant (kov) indicated in table
1 given in Yoon et al. (2002a) differed from those reported in tables 2-4 of the same paper.
61
Because this is the single work found in the literature concerning AEPD kinetics, more
studies would be compulsory to shed a light upon those discrepancies.
1.2.5.3.3. AHPD systems
Two kinetic studies were found in the literature concerning CO2 absorption in AHPD
solutions. Both works by Bougie and Iliuta (2009) and Paul et al. (2009b) used a wetted-
wall column absorber and studied the reaction kinetics at 303.15, 313.15 and 323.15 K.
In Bougie and Iliuta (2009), AHPD concentration was varied between 0.5 and 2.4
kmol·m−3 and the chemical absorption was described using the zwitterion mechanism. The
fast pseudo-first-order regime was verified by analysing gas and amine concentration
profiles in the liquid film. Three reaction rate parameters, k2, 1OH2 2 −kkk , and 1AM2 −kkk ,
were determined using a non-linear regression method for each studied temperature and
correlated using Arrhenius type equations. The calculated activation energy for 2k was
found to be 53.7 kJ·mol-1. Authors analysed the overall absorption rate constants of various
SHA and observed that the amines reactivity varied in the following ascending order
AEPD, AHPD, AMPD, and AMP, which represents the opposite order of the amines
bulkiness (steric hindrance). This seemed to confirm the assumption that a reduced steric
hindrance leads to a more pronounced reaction rate constant (more reactivity).
Paul et al. (2009b) used AHPD concentration of 0.179 to 1.789 kmol·m-3. The reaction
order was found to be in between 1.0 and 1.1 with respect to amine for the above-
mentioned concentration range. Kinetic rate parameters were calculated and presented at
each experimental condition assuming an overall second-order reaction. Second-order rate
constants, k2, were found to be 532.7, 1096, and 2380 m3·kmol-1·s-1 at 303, 313, and 323 K,
respectively, with an activation energy of 65.2 kJ·mol-1. These results were significantly
higher than those of Bougie and Iliuta (2009), but it should be recalled that these values
were not obtained on the basis of the same reaction mechanism. Paul et al. (2009b)
performed a parametric sensitivity analysis and found that Henry’s law constant values for
CO2 in solution had a huge impact on the calculated CO2 absorption rates.
62
1.2.5.3.4. AMPD systems
As for AHPD, only two kinetic studies were found in the literature concerning CO2
absorption in AMPD solutions.
Bouhamra et al. (1999) studied the mechanism and the kinetics of CO2 absorption in
AMPD solutions by a stopped-flow technique between 278 and 303 K. Concentrations
were varied between 0.025 and 1.600 kmol·m-3 AMPD. They found that the partial order
related to the amine varied between 1.26 and 1.33 what could be explained by the
zwitterion mechanism. Based on this mechanism, they extracted corresponding kinetic
constants for each temperature and correlated them following an Arrhenius law. The
activation energies for k2, 1AM2 −kkk and 1OH2 2 −kkk are respectively, 33.7, 44.7 and 62.05
kJ·mol-1. Comparison were made by the authors with AMP values from the literature
(Alper, 1990; Xu et al., 1996) and, as expected, the observed reaction rate for AMPD were
smaller than that of AMP which was caused by added hindrance and charge effect of an
hydroxyl group which replaced one hydrogen in AMP.
Concerning the second study, Yoon et al. (2003) with a wetted-wall column obtained
the kinetics constant for AMPD solutions of concentration between 0.236 to 2.963 kmol·m-
3 (2.5 to 30 wt%) and for temperature ranging from 303-323 K. As in Bouhamra et al.
(1999), they used the zwitterion mechanism to interpret their data and found that the partial
order for the amine was varying from 1.36 to 1.41. The activation energy for k2 was
calculated to be 38.3 kJ·mol-1 with an absolute error of 3%. Kinetic constant values of each
study were analysed and it was found that k2 values of Yoon et al. (2003) followed almost
the same trend as values of Bouhamra et al. (1999). Values of the kinetic parameter
1OH2 2 −kkk , were also found to follow the same trend if the value at 303 K from Bouhamra
et al. (1999) was not taken into account. 1AM2 −kkk values from both study were in
disagreement. A set of kinetic parameter coming from the combination of the absorption
data of both studies may correct these discrepancies but data in Bouhamra et al. (1999)
were not tabulated what limited this opportunity.
63
1.2.5.3.5. Other SHA systems
Ali et al. (2002) investigated the kinetics of the reaction between aqueous solutions
of carbon dioxide and TBAE over a temperature range of 283-308 K by using a direct
stopped-flow technique. Steric factors caused TBAE to react slower than its unhindered
constitutional isomer (2-(n-butylamino)ethanol), but with the increase in temperature, the
detrimental effect of these steric factors on the reaction rates was found to decrease.
Authors mentioned that the reaction mechanism of TBAE was similar to that for tertiary
amines, while the obtained k2 values of TBAE are significantly higher than those
corresponding to MDEA and TEA at 298 K. Sharma (1965) reported values of the second-
order rate constant (k2) for the reaction of CO2 with various SHA (AHPD, AMP, AMPD,
DIBA, DIPA, TBA) for 1 kmol·m-3 aqueous solutions at 291 and/or 298 K. However, the
errors in the reported values were estimated to be higher than 25%. A comparison with
other works (Alper, 1990; Bougie and Iliuta, 2009; Bouhamra et al., 1999) also revealed
major deviations of k2 values for AMP, AHPD and AMPD solutions.
1.2.6. Regeneration capability
Compared to the extensive number of studies on CO2 absorption in the open literature,
there are relatively few information related to CO2 thermal desorption processes, despite
the fact that the stripping unit is usually highly energy-consuming and it is responsible for
the main operational cost of the process (Tobiesen and Svendsen, 2006). For that reason,
amine solutions with low regeneration cost are essential for economic viability of the
absorption/desorption processes.
In comparison to conventional primary and secondary alkanolamines like MEA and
DEA, SHA (e.g. AMP) form unstable carbamates due to the hindrance of the bulky group
adjacent to the amino group (Sartori and Savage, 1983). The presence of carbamates
influences the regeneration efficiency of alkanolamine solutions. Stable carbamates are
difficult to revert to fresh amines, leading therefore to longer regeneration time and more
energy consuming (Barzagli et al., 2010; Sakwattanapong et al., 2005). Hydrolysis of the
voluminous carbamates leads to a preferential bicarbonate formation process and it is
expected that a solution containing a larger proportion of bicarbonate undergoes desorption
64
at a higher rate (requiring less energy) and produces a lean solution containing less
physically and chemically absorbed CO2 (Hook, 1997; Sartori and Savage, 1983;
Tontiwachwuthikul et al., 1991).
In a large scale continuous process, the solvent is continuously circulating between the
absorber and desorber, so that neither the regenerated amine is saturated by CO2 nor the
loaded amine solution needs to be fully regenerated. There is then place for high quantity of
possible configurations for an optimal absorption-regeneration process depending on
solution flow rate, amine concentration, lean and rich loading, and absorption and
regeneration temperatures. To improve the efficiency of the carbon dioxide cycling process
and to reduce the regeneration energy consumption, SHA regenerative behaviour over
conventional alkanolamines was investigated in some studies.
Hook (1997) studied the CO2 absorption/desorption capacity of solutions of eight
different amine compounds including MEA, AMP and six potassium amino salts. The aim
of that work was to identify the absorbent which minimizes the power consumption of the
regeneration step compared to MEA solutions in non-nuclear submarines where power
conservation is crucial. Carbon dioxide absorption experiments were performed using
100% CO2 and mixtures of 4.7 vol% and 1.1 vol% CO2 in air. Carbon dioxide absorption
was measured by following the volume changes of a CO2 gas “reservoir” which provided
the atmosphere over 10 mL of a stirred 2.5 kmol·m-3 aqueous amine solution for 5 h
(equilibrium 20 h) at 295 ± 0.5 K. For desorption experiments, the volume of gas generated
by the equilibrated solutions when stirred in a 393 K oil bath was measured. Desorption
experiments were conducted for 1 h, well in excess of the equilibrium time. Solutions
reached 363 K in 2.5 min and 372 K in 8 min and then remained at 372-373 K. Carbon
dioxide cycling experiments were performed by incorporating at least three absorptions and
two desorptions sequentially. From their results, some interesting observations appeared. It
was found, as expected, that the position and the nature of the substitution around the
amino group influenced the absorption rate and the absorption capacity of the studied
solutions. The slow absorption of the N-substituted, R-dimethylated (secondary) amines
relatively to sterically hindered primary amines indicated that the presence of three bulky
65
groups around the reaction site caused important sterically restriction, thus significantly
impeding the reaction. If only desorption kinetics was considered, the calculated CO2
released during the first 5 min of desorption led to the following order: AMP (0.69 mol of
CO2 released/mol of amine) > MEA (0.38). AMP was desorbed to a level of 0.1 mol/mol,
while MEA reached only 0.2 mol/mol. Tested potassium amino salt failed to desorb to the
levels reached by the alkanolamines. No polyalcohols were tested to verify if adding more
hydroxyl group increased the regeneration performances. Globally, the authors observed
that potassium amino salts exhibited precipitation problems which limited their application.
As a general trend, it was observed that the amines which allowed the higher CO2
absorption, by generating the most bicarbonate, produced the fastest CO2 stripping upon
heating. Although AMP exhibited encouraging desorption characteristics, the rate of CO2
absorption at low partial pressures versus MEA was likely to restrict its use. However, at
higher CO2 concentrations, as encountered in several industrial processes, AMP may be
potentially superior to MEA.
In Sakwattanapong et al. (2005), the reboiler heat duty for regeneration of loaded
aqueous single and blended alkanolamines was experimentally evaluated in a bench-scale
regeneration column under atmospheric pressure. Various alkanolamines, including MEA,
DEA, MDEA, and the mixtures of MEA + MDEA, DEA + MDEA, and AMP + MEA were
included in this study. The results indicated that the reboiler heat duty was dependent on
the CO2 loading of lean and rich solutions, alkanolamine type and concentration, as well as
on the composition of blended alkanolamines. MEA required the highest reboiler heat duty,
followed by DEA and MDEA. Unfortunately, single AMP aqueous solutions were not
evaluated since it was reported that these solutions underwent crystallization under the
tested conditions (solutions of 4, 5 and 7 kmol·m-3). In general, the use of more
concentrated solutions led to the reduction of the reboiler heat duties. Similar conclusions
were reported by Mejdell et al. (2010a) who studied different combinations of AMP +
MEA and found that aqueous mixtures of 20 wt% AMP + 30 wt% MEA and 25 wt% AMP
+ 25 wt% MEA offered net cyclic capacity advantage over 30 wt% MEA aqueous
solutions. For aqueous blended amine solutions, the heat duties were found to be between
the heat duties of their parent alkanolamines. Concerning the loading influence, the results
66
indicated that the reboiler heat duty was in inverse relationship with the achieved lean CO2
loading; i.e., it decreased with increasing lean CO2 loading. It was shown that the reboiler
heat duty did not have a linear correlation with lean CO2 loading; two distinct regions
seemed to be present. In the first region where the lean CO2 loading was below around 0.1,
a significant amount of additional heat duty was required for a small reduction in lean CO2
loading. In some cases, the lean CO2 loading remained virtually unchanged regardless of
the amount of energy supplied. This presented an unfavourable operating region that
consumed excessive energy during solvent regeneration. In the second region, where the
lean CO2 loading was above about 0.1, only a small amount of additional heat duty was
required to achieve a substantial reduction in lean CO2 loading, thus presenting a
favourable operating region. In addition, it was apparent that, at a given lean CO2 loading, a
reduction in rich CO2 loading (from 0.5 to 0.3) caused the reboiler heat duty to increase
substantially. Lowering the rich loading caused the CO2 partial pressure in equilibrium to
be reduced accordingly, increasing therefore the need of heating for producing more water
vapour, which required much more energy at the reboiler.
Zhang et al. (2008) studied the regeneration of loaded aqueous AMP solutions. All
absorption experiments were conducted in a double stirred-cell contactor at a temperature
of 303 K and with a gas mixture containing 15% CO2 and 85% N2. AMP concentration was
keep at 1.0 kmol·m-3. Regeneration experiments were run at 358, 368, 378, 383, 393, and
403 K. Each regeneration run lasted for 2 to 3 hours. An analysis of the optimum
regeneration temperature indicated that the regeneration efficiency increased from 86.2% to
98.3% when temperature increased from 358 to 403 K. The most suitable regeneration
temperature for AMP was found to be 383 K. After six absorption/regeneration cycles, the
regeneration efficiency for AMP solution sloped only from 98.3% to 94.0%, possibly
because of the formation of heat-stable and non-regenerable salts. For similar experimental
conditions (383 K and regeneration runs of 1.5 hour), a comparison of the regeneration
efficiency of different amine solutions was performed after three cycles of
absorption/regeneration. The results indicated that the aqueous AMP solution was easier to
regenerate, with less loss in the absorption capacity than the other amines. The regeneration
performance were ranked in the following order: AMP > MDEA > DETA
67
(diethylenetriamine ) > DEA > MEA. However, an analysis of the absorption rate led to the
following ranking: DETA > MEA > DEA > AMP > MDEA at the beginning of the
reaction. All these results led the authors to the conclusion that AMP may be more suited
for application in industrial processes where CO2 partial pressures are higher. AMP
solutions could then take advantage of its higher absorption capacity and appreciable
absorption rate.
Another work concerning the regeneration of SHA was recently published by Bougie
and Iliuta (2010a). The aim of this study was to compare the regeneration capability of
different single sterically hindered alkanolamines (AMP, AEPD, AMPD, AHPD) or Pz-
activated aqueous solutions with that of single MEA or Pz aqueous solutions. The
absorption/regeneration cycles were performed in the following conditions of solution
concentrations and regeneration temperatures: (i) 1.00 kmol·m-3 AHPD for a regeneration
temperature between 353.2 and 393.2 K and (ii) 1.00 kmol·m-3 (AEPD, AMPD, AMP,
AHPD or Pz), 2.00 kmol·m-3 MEA and 0.90 kmol·m-3 AHPD + 0.10 kmol·m-3 Pz for a
regeneration temperature of 383.2 K. The desorption rate was calculated on the basis of the
CO2 released, which was measured on-line using a microGC. Taken together, the results of
that work revealed that the regeneration efficiency can be classified in the following order:
AHPD (76.0) >> AMPD (62.6) ≥ AEPD (60.2) > MEA (43.9) ≥ Pz (42.3) > AMP (34.8).
These results demonstrated that solutions of the three most hindered alkanolamine (AHPD,
AMPD and AEPD), and in particular AHPD, were easier to regenerate because they
possibly did not form (or very few) stable carbamates in solution. However, the results
obtained for AMP solutions showed that the calculated cyclic capacity and the regeneration
efficiency, under the mentioned experimental conditions, were the lowest of all tested
amines. MEA and Pz showed almost the same cyclic capacity and regeneration efficiency.
However, Pz, with its higher kinetic constants over MEA seemed to be the best activator.
Finally, it was found that the addition of a small amount of Pz to AHPD aqueous solution
allowed obtaining almost the same cyclic capacity and regeneration efficiency as non-
activated solutions but for half of the absorption time. Furthermore, based on the results
and economic considerations (the prices for the three best SHA were 0.06, 0.22 and 0.57
68
US$/g, respectively, for AHPD, AEPD and AMPD) and amine availability, the aqueous
mixture AHPD + Pz seemed to be a potential new solvent for CO2 capture.
Choi et al. (2009) studied absorption and regeneration performance of loaded aqueous
blends of AMP + MEA (wt% AMP / wt% MEA: 30/0, 24/6, 18/12, 12/18, 6/24, 0/30). The
absorption was performed at 313 K while the effect of the regenerator temperature on the
stripping efficiency was investigated at 363, 373, and 383 K. The authors found that a
regeneration temperature of 383 K gave the highest stripping efficiency, so they kept this
temperature in the following experiments. The results showed that the CO2 removal
efficiency was optimal at 30 wt%. Further amine additions in the solution did not lead to
significant amelioration of the removal efficiency. They mentioned that the amine
degradation might have caused this behaviour. In single amine solutions, AMP had a better
stripping efficiency than MEA. In blended amine solutions, the stripping efficiency was
influenced by the ratio between AMP and MEA. According to the reactivity and the
regeneration efficiency, the optimum blend AMP + MEA was found at a concentration
ratio of 18/12 wt%.
Recently, Barzagli et al. (2010) studied experimentally the performances of CO2
capture by aqueous solutions of single alkanolamines DEA, MDEA and AMP (0.667, 1.33
and 2.00 kmol·m-3), as well as some alkanolamine blends (total amine content of 2.00
kmol·m-3). CO2-loaded and regenerated amine solutions were continuously circulated at the
same rate of 0.60 dm3·h-1 in a closed system between the absorber (set at 293 K) and the
desorber (set at 363, 373 and 363-388 K). The gas mixture of 12 vol% CO2 in air,
simulating the flue gas, continuously flowed at the bottom of the absorber through a
sintered-glass diffuser. CO2-amine reaction equilibria have been investigated by 13C NMR
spectroscopy, for establishing the regeneration efficiency and the loading capacity for each
single amine. It was found that AMP displayed the highest absorption efficiency, and
MDEA the highest regeneration efficiency, at every given amine concentration and
desorber temperature. Under the same operating conditions, blended AMP + MDEA and
AMP + DEA aqueous systems (1/2 and 2/1 molar ratios for a total of 2 kmol·m-3)
significantly enhanced the absorption efficiency (in the range 7-14%) with respect to single
69
amines. AMP + MDEA blends displayed better performances than AMP + DEA due to the
lower efficiency of DEA carbamate in both CO2 absorption and amine regeneration. Owing
to a higher thermal stability, AMP and MDEA solutions surpassed DEA, as no degradation
product were detected by 13C NMR analysis after heating AMP and MDEA solutions at
403 K up to fourteen days, whereas a degradation rate of about 0.4%/day for DEA solution
was identified.
1.2.7. Conclusions and recommendations for future research
An update of different aspects which are essential for the design and operation of CO2
absorption apparatus using solutions containing sterically hindered amines, such as physical
properties (density, viscosity, vapour pressure, heat capacity and heat of absorption, CO2
and amine diffusivity), CO2 absorption capacity and kinetics, regeneration capability, has
been presented here. It was observed that AMP was by far the most studied SHA in the
literature. Very limited information was found concerning other SHA; new works reporting
data on different aspects covered here would be saluted.
Several conclusions were made for each particular section. As it can be shown in the
tables and also mentioned in the analysis of existing data, new experimental work for
various systems would be useful for the elucidation of contradictory behaviors or for
completing the existing data base, as for example: (1) surface tension for aqueous AMP
solutions, as well as for various other SHA, in order to be able to compare and analyse data;
(2) vapor pressure and heat capacity for aqueous solutions of various SHA (except AMP)
where data are very limited or even unavailable; (3) amine diffusivity for all SHA; (4) CO2
solubility in aqueous AMPD and AEPD solutions; (5) physical solubility (Henry’s
constants) for AMP + MEA or DEA where data are quite contradictory (cf. to Figures 1.5
and 1.6); (6) new kinetic studies for all SHA, even for AMP, where the values for kinetic
parameters are quite spread, would be much welcome. Kinetic studies for single amine
solutions, using the zwitterion mechanism to treat CO2 reaction rate in a well-defined
reaction regime over large temperature and concentration ranges and taking into account
the thermal effect that happened at the gas-liquid interface, may help to get reliable sets of
kinetic rate constants.
70
1.3. CO2 capture in amine solution absorbents using membrane
contactors
Absorption (especially using amine-based absorbents) is the most commonly used
method for CO2 removal mainly due to its high CO2 removal efficiency, particularly at low
CO2 partial pressure. The gas absorption process for CO2 absorption can be carried out in
different reactors, such as bubble columns, sieve trays, packed towers, and venture
scrubbers. Although the traditional packed columns have attained considerable success in
industrial applications, they suffer from various operational problems like foaming,
flooding, channeling, and liquid entrainment (Gabelman and Hwang, 1999). As promising
alternative, the membrane contactor (MC) process has become one of the research focuses
because of various advantages over the traditional gas absorption processes (Bernardo et
al., 2009). The idea of MC was first introduced in the literature by Zhang and Cussler
(Zhang and Cussler, 1985a, b) in the context of CO2 absorption in aqueous NaOH solutions
using PP hollow fiber membranes. The gas absorption based on this hybrid process
combines the benefit of the absorption (high selectivity) and those of the membranes
(operational flexibility and easy linear scale up, low capital and operation costs, high mass
transfer rate). The overall absorption process is the same as in conventional absorption-
desorption cycle and the energy requirement depends on the solvent performance and
process optimization.
Bernardo et al. (2009) discussed the most promising areas of research in membrane gas
separations, their industrial applications and the opportunities for the integration of
membrane gas separation units in hybrid systems for process intensification. In the short
section dedicated to hybrid systems, the authors mentioned the application of hybrid
membrane/amine solutions for CO2 separation. Four reviews dedicated to hollow fiber
membrane contactors are available in the literature: Gabelman and Hwang (1999), Drioli et
al. (2005), Li and Chen (2005) and Mansourizadeh and Ismail (2009). The last two ones are
especially directed on the MC application for acid gas capture, most researches focussing
on CO2 removal. Recently, Cui and deMontigny (2013) shortly reviewed the recent
progress of CO2 capture using hollow fiber MC.
71
Compared to conventional gas-liquid contactors, membrane contactors have several
main advantages:
(i) Large contact area for promoting an efficient gas-liquid mass transfer.
Membrane contactors offer typically more surface area to volume ratio than traditional
gas absorbtion contactors (Gabelman and Hwang, 1999), thus reducing considerably the
contactor size (capital cost). As example, Hoff et al. (2004) and Kumar et al. (2002)
reported that contact areas in MC can reach 500 to 2000 m2/m3 (usually >1000 m2/m3) in
comparison with 100-300 m2/m3 for packed columns. In a recent work by Hoff and
Svedsen (2013) concerning a comparison between MC and absorption towers for post-
combustion CO2 capture and for natural sweetening, the results showed that the size of the
contactor may potentially be reduced by 75% using MC with the liquid flowing on the shell
side of the membrane unit.
(ii) High modularity (operation over a wide capacity range) and compatibility for an easy scale-up.
A predictable increase in capacity can be reached by simply adding membranes to the
modules or using more modules.
(iii) The possibility of varying stream flow rates independently and without the occurrence of flooding, entrainement, channeling and foaming.
This is due to the presence of the membrane between gas and liquid (Tesser et al., 2005).
(iv) Easier performance prediction.
This is due to the fact that the interfacial area (equal to the effective membrane surface
area) is known and constant (Li and Chen, 2005).
Despite their important advantages, the main MC disadvantages are the following:
(i) Additional resistance to mass transfer introduced by the membrane itself.
However, this resistance is strongly dependent on membrane porosity, permeability,
thickness and wettability and therefore, it is not always very important. The large interfacial
72
mass transfer area that can lead to a sufficiently high mass transfer rate can, in real
conditions, make the MC a much more efficient absorber compared to packed columns (Li
and Chen, 2005). For example, deMontigny et al. (2005) obtained Kgav values up to 4 times
larger than those obtained in a packed column containing Sulzer DX structured packing
(CO2 absorption using AMP solution in PTFE MC).
(ii) Membrane wetting.
This can become important when the membrane is wetted by the absorbent (pores
partially or totally filled by liquid) (Mansourizadeh and Ismail, 2009). In the wetted (even
partially) condition, the mass transfer is reduced due to the presence of a stagnant liquid
film in the membranes pores and consequently, MC performance can be dramatically
reduced (Dindore et al., 2004). As the contact between the gas and the liquid is made after
the gas diffuses through the membrane pores, in non-wetted conditions (pores filled with
gas), the gas-liquid interface is formed at the pores opening adjacent to the liquid. The non-
wetted operation mode is therefore favorable since gas phase coefficients for CO2 are
higher than those in the liquid phase. However, in real conditions, it is not be possible to
maintain a non-wetted mode over time and the partially-wetted mode is usually
encountered. The membrane wetting phenomenon and the research for finding appropriate
actions to be taken for limiting this unwanted issue before implementation of MC
technology in industrial units has attracted significant attention, but the subject continues to
offer significant research challenge (Al-Marzouqi et al., 2008; Rongwong et al., 2009;
Wang et al., 2005).
The performances of MC for CO2 separation from different industrial flue gases
strongly depend on the properties of both absorption liquid and membrane, the
compatibility between them and the constructive characteristics of MC modules (flow
configuration, module geometry, and operating parameters). In order to avoid wetting
phenomena and mixing between contacting phases, highly hydrophobic membranes are
required. The proper choice of the membrane/absorption liquid combination is a crucial
step in developing the CO2 absorption process in membrane contactors.
73
1.3.1. Principle of gas absorption in MC
In MC, the mass transfer between the gas and the liquid takes place without
dispersing one phase into another (Figure 1.7). The microporous membrane acts as a barrier
between the gas and liquid. The gas diffuses through the membrane pores, but the
membrane is not selective (as it is in the gas separation by selective membranes). It is the
liquid (absorbent) that assures the selectivity. In the ideal case, the hydrophobic membrane
pores are filled with the gas and the absorption takes place at the liquid side of the
membrane.
Figure 1.7. Gas diffusion in membrane contactor (Hoff et al., 2004).
Based on the film theory, the mass transfer of CO2 in an absorption liquid (e.g., amine
solution) using a gas-liquid MC involves 3 consecutive processes: (i) diffusion of CO2 from
the bulk gas phase towards the membrane surface, (ii) diffusion of CO2 through the
membrane pores towards the liquid interface, and (iii) diffusion of CO2 into the liquid
solution with chemical reaction (Figure 1.8). The overall rate of mass transfer (CO2 flux,
NA) is given by:
( ) ( ) ( )ALiALliAGMAGmMAGAGgA CCEkCCkCCkN −=−=−= ,,,, (1.32)
where m, , andg lk k k are, respectively, the gas phase, the membrane and the liquid phase
transfer coefficients. E is the enhancement factor due to the chemical reaction.
74
Figure 1.8. Mass transfer in membrane contactor.
The overall mass transfer resistance based on the gas phase (1/ GK ) will then consist of
three resistances in series: the resistance of the gaseous phase boundary layer (1/ gk ), the
membrane resistance (1/ mk ) and the resistance of the liquid phase boundary layer (1/ lk ):
1 1 1 1
G g m lK k k mk= + + (1.33)
where m is the distribution coefficient between gas and liquid phases. Various correlations
are available in the literature for calculating the individual mass transfer coefficients
(Gabelman and Hwang, 1999; Li and Chen, 2005; Mansourizadeh and Ismail, 2009).
1.3.2. Membrane module configurations
The essential element in the MC module is the microporous hydrophobic membrane
(hydrophobicity, structure, thermal and chemical resistance). However, the efficiency of the
MC also strongly depends on absorbent properties, flow configuration and module
geometry.
75
Although several membrane module geometries are possible, with very few exceptions,
most works in the literature concern hollow fiber membrane contactors (HFMC) (Figure
1.9). Typically, a HFMC module consists in a bundle of hollow fibers packed in parallel
alignement in a shell, similar to a tubular heat exchanger (Figure 1.9a). HFMC are
commercialised by different companies (Figure 1.9c) like Celgard LLC, Membrana Co,
Mitsubishi Rayon, Sumitomo, NeoMecs, Dic Corporation, Hoechst Celanese Corporation,
W.L.Gore & Associates, etc.
a)
b) c)
Figure 1.9. Hollow fiber membrane contactors: a) parallel flow; b) cross flow provided by TNO-MEP; c) MC module commercialized by Membrana Co.
To improve the mass transfer and avoid fluid channeling and bypassing on the shell
side due to possible non-uniform fiber distribution, many works have been focussed on
fibers regularity, packing density, and relative flow directions of gas and liquid phases as
parallel (co-current or counter-current) (Figure 1.9a,c) and cross-flows (Figure 1.9b)
(Dindore et al., 2005; Liu et al., 2005). Modules with parallel flow circulation are generally
adopted and applied in most investigations for gas separation due to the simplicity in
manufacturing and suitability for predicting mass transfer rates. The gas phase flows
parallel to the liquid phase on the opposite side of the membrane fibers (Wang and Cussler,
1993). Several works have shown that the counter-current flow might offer higher mass
transfer coefficients than the co-current configuration (deMontigny et al., 2006), but
76
depending on the process conditions, the difference can also be insignificant (Kreulen et al.,
1993). For industrial applications, the increase of gas-liquid contact area can be obtained by
arrangements of membrane modules in multistage cascade (Faiz et al., 2011), with the main
advantage of improving the system performance. As the present thesis concerns the
application of MC for CO2 removal using SHA based absorbents, the literature review is
limited to this topic. A complete description of the available works related to CO2
absorption in SHA solutions using HFMC is given in the section 1.3.4.
Compared to HFMC, information on CO2 absorption in flat sheet MC (FSMC) are
extremely scarce (Ahmad et al., 2010; Dindore et al., 2004; Lin et al., 2009b; Paul et al.,
2008; Zhang et al., 2006) and they are, therefore, not included in any review paper
concerning MC. However, this type of contactors has some advantages compared to
HFMC, like easiness in membrane fabrication and module assembly, and higher flux for
the same gas-liquid contact area (Baker, 2004). All investigations available in the open
literature are based on the use of just one membrane in the module. A short overview
concerning FSMC is given here and a complete description is available in the section 1.3.4.
and the Chapter 11.
A FSMC was used by Zhang et al. (2006) to study the effect of membrane porosity and
pore size on pure CO2 absorption in water and NaOH aqueous solutions. Dindore et al.
(2004) measured the critical entry pressure, a very useful parameter in membrane operation,
and determined the mass transfer coefficient for CO2 absorption in different plysical
solvents. The first work related to the application of FSMC for CO2 absorption in amine
solutions was given by Paul et al. (2008) who performed a theoretical study of CO2
absorption (pure CO2 and CO2/N2 mixture) by different single and blended alkanolamines
(MEA, DEA, MDEA, AMP, MEA + AMP) considering one hypothetical membrane.
FSMC with one PVDF or plasma-treated PVDF or PTFE membrane was used by Lin et al.
(2009b) to study the CO2 absorption from CO2/N2 mixtures in MDEA, AMP and AMP+Pz
aqueous solutions (influence of liquid and gas fow rates and absorbent concentration).
Finally, Ahmad et al. (2010) investigated the absorption of CO2 from CO2/N2 mixture in
aqueous AMP solutions using a PVDF flat sheet membrane.
77
1.3.3. Absorbent screening for MC and liquid/membrane compatibility with
polymeric membranes
Along with the constructive characteristics of MC modules (flow configuration,
module geometry, operating parameters), the performance of MC for CO2 separation
strongly depend on the properties of the absorbent, the membrane and on the compatibility
between them. In order to avoid the mixing between contacting phases and the unwanted
wetting phenomenon (one of the main drawbacks of this kind of contactors, and which is a
major obstacle to their implementation in industrial separation processes), highly
hydrophobic membranes that should be compatible with amine solutions (the most used
absorbent in acid gas separations) are required. The proper choice of the
membrane/absorption liquid combination is therefore a crucial step in developing CO2
absorption in MC.
Various liquid absorbents have been considered for CO2 separation in MC,
including water, aqueous solutions of different bases (NaOH, KOH, Na2CO3, K2CO3,
NaHCO3, Na2SO3, NH3, amines (alkanolamines)) and amino acid salts (Mansourizadeh and
Ismail, 2009). As this thesis concerns amine solutions and more specifically, SHA based
solutions, this review will be limited to this kind of compounds.
Industrially, the most used amines for CO2 removal from different gas mixtures are
MEA, DEA, DIPA, MDEA, Pz and AMP (Kohl and Nielsen, 1997). The choice of a certain
amine (single or blended) is mainly based on the absorption capacity, reaction kinetics,
regenerative potential, corrosiveness, price and availability. As largely discussed in the
section 1.2, the key advantage of primary and secondary alkanolamines (like MEA, DEA)
is their fast reactivity due to the formation of stable carbamates. However, this will lead to
very high solvent regeneration cost (mainly, energy penalty). They also have the drawback
of a relatively low CO2 loading (theoretically, 0.5 mol CO2/mole amine). Tertiary
alkanolamines (like MDEA) have a low reactivity in respect to CO2, due to the exclusive
formation of bicarbonates, but this will lead to a very low solvent regeneration cost, which
is a positive feature. Another advantage of tertiary amines is the high CO2 loading capacity
(theoretical, 1 mol of CO2/mol of amine). More recenty, the sterically hindered amines
78
(AMP, the simpler hindrance form of MEA, being the first SHA introduced in the
literature) attracted the attention due to the formation of unstable carbamates whose
hydrolysis leads preferentially to bicarbonate formation and in consequence, to a theoretical
loading capacity of 1 mol of CO2/mol of amine (Sartori and Savage, 1983). The reaction
kinetics is significantly higher compared to tertiary amines but lower than primary and
secondary amines. It is the reason why they are usually used in mixture with high reactive
amines (like MEA, DEA, Pz).
As it can be seen in Table 1.2 which reviews the available works concerning the
application of MC for CO2 absorption, among several possible aqueous amine solutions,
MEA is the most investigated absorbent for CO2 capture and this is obvious, taking into
account the fact that MEA is the benchmark amine industrially used in acid gas (CO2 in
particular) separation. Moreover, many works investigating the absorption efficiency of
other amine solutions also include in their analysis a comparison with the MEA
performance.
AMP, alone or mixed with other amines (usually, accelerators for improving the
absorption kinetics) is the only investigated SHA for CO2 separation in MC.
From the membrane point of view, PP, PTFE and PVDF are usually employed in the
MC modules fabrication for CO2 capture using amine solutions due to their hydrophobicity
and low surface energy. Membrane wetting by the absorbent liquid is generally favored by
a high polymer surface energy. PTFE presents a significantly lower surface energy
compared to other typical polymers, 17-22 mN/m (Fu et al., 2004) and as a result, is the
most resistant material to wetting by different aqueous amine solutions. Moreover, PTFE is
chemically and thermally stable and inert, thus preventing the polymer from changing its
properties over time (deMontigny et al., 2006; Hoff et al., 2004; Nishikawa et al., 1995; Sea
et al., 2002). Currently, PTFE is suggested to be the only suitable membrane material for
use in the presence of alkanolamines (Falk-Pedersen and Dannström, 1997). It was reported
that PTFE membranes can preserve their efficiency even after several months of use
(Dindore et al., 2004).
79
Table 1.2. Current research for CO2 capture in MC
Absorbent Membrane type (aqueous solutions)
PTFE PP PVDF
MEA (deMontigny et al., 2005) (Nishikawa et al., 1995) (Yeon et al., 2003) (Kim and Yang, 2000) (Sea et al., 2002) (Rajabzadeh et al., 2009) (Hoff et al., 2004) (deMontigny et al., 2005) (Sea et al., 2002) (deMontigny et al., 2006) (deMontigny et al., 2006) (Yeon et al., 2005) (Yeon et al., 2003) (Lv et al., 2010) (Nishikawa et al., 1995) (Wang et al., 2013) (Falk-Pedersen and
Dannström, 1997) (Falk-Pedersen and Dannström, 1997)
(Marzouk et al., 2012) (Yan et al., 2007) (Nii and Takeuchi, 1994) (Vogt et al., 2011) (Constantinou et al., 2014) (Rajabzadeh et al., 2009) (Sea et al., 2002)
DEA (Marzouk et al., 2012) (Rangwala, 1996) (Constantinou et al., 2014) (Wang et al., 2013)
MDEA (Kim and Yang, 2000) (Lu et al., 2005) (Lin et al., 2009b)* (Hoff et al., 2004) (Lin et al., 2009c) (Lin et al., 2009c) (Lin et al., 2009b)* (Lin et al., 2009a) (Lin et al., 2009a) (Lv et al., 2010) (Wang et al., 2013) (Yan et al., 2007)
AMP (deMontigny et al., 2005) (deMontigny et al., 2005) (Lin et al., 2009b)* (Kim and Yang, 2000) (deMontigny et al., 2006) (Ahmad et al., 2010)* (deMontigny et al., 2006) (Lu et al., 2007) (Lin et al., 2008) (Lin et al., 2009b)* (Lin et al., 2009c) (Lin et al., 2009c) (Nii and Takeuchi, 1994) (Lin et al., 2009a) (Rongwong et al., 2009) (Kumazawa, 2000)
Pz (Lin et al., 2009c) (Lin et al., 2009c) (Lin et al., 2009a)
AMP/DEA (Wang et al., 2013) AMP/Pz (Lin et al., 2009b)* (Lin et al., 2009a) (Lin et al., 2009b)*
(Lin et al., 2008) (Lin et al., 2009a)
MDEA/AMP (Lu et al., 2007) MDEA/MEA (Wang et al., 2013) MDEA/Pz (Lu et al., 2005) (Lin et al., 2009a)
(Lu et al., 2007) (Wang et al., 2013)
MEA/AMP/PZ (Chen et al., 2011)* *Flat sheet membranes
However, the high price and the unavailability in various structures (e.g., internal
diameter of fibers lower than 0.8 mm and large range of pore size and porosity) are the
main drawbacks of PTFE membranes. Their small specific interfacial area does not allow
the fabrication of modules with a gas-liquid contact surface as high as for PP modules.
80
PP membranes have been largely used in CO2 separation in MC because of their low
price (around 1100 times less expesive than PTFE (deMontigny et al., 2006)), availability
in a large range of fiber diameter, membrane thickness and porosity (deMontigny et al.,
2005; Sea et al., 2002), as well as the easiness of potting for module fabrication. However,
due to a higher surface energy (around 33 mN/m (Mittal, 2003)) compared to PTFE, most
works reported that PP membranes were wetted (even for short-term applications) by all
aqueous amine solutions (deMontigny et al., 2005; Kreulen et al., 1993; Nishikawa et al.,
1995; Rangwala, 1996; Wang et al., 2004), thus loosing their separation efficiency.
Membrane degradation and penetration of liquid into the pores seem to explain this
behaviour. For example, Yan et al. (2008) performed SEM analysis of PP membranes kept
in contact with MEA aqueous solutions and reported modifications and enlargement of
membrane pores.
The chemical stability of the membrane material has significant influence on its long-
term stability and consequently, on the absorption efficiency. PVDF membranes are
hydrophobic (around 30 mN/m, (Mittal, 2003)). However, although PVDF is known to be
very stable in most corrosive media (acids, oxidants and halogens), the use of PVDF
membranes is conditionally suitable for alkaline solutions (they can be attacked by medium
concentrated solutions) (Mansourizadeh and Ismail, 2009). Several works reported that
PVDF interacted with aqueous amine solutions (Atchariyawut et al., 2006; Sea et al., 2002;
Yeon et al., 2005).
An important aspect for the absorption liquid selection is related to the liquid
surface tension and the membrane/liquid contact angle. Even though the polymeric
microporous membranes used for the gas/liquid separation are hydrophobic, the absorption
solution with low surface tension can penetrate inside the membrane pores causing
membrane wetting. Membrane wettability is one of the main problems affecting the
performances of the membrane contactors in long-time cyclic operation. As the addition of
an organic component reduces the surface tension of water, most conventional
alkanolamine aqueous solutions gradually wet the membranes with time, leading to the
increase of the mass transfer resistance.
81
For a specific membrane material, the degree of pore wetting mainly depends on the
absorbent surface tension and its contact angle with the membrane (Gabelman and Hwang,
1999). The maximum pressure (breakthrough pressure, ΔPc) which can be applied on the
liquid to enter the membrane pores is determined by the Laplace-Young equation:
L
p,max
-4 cos cPdσ θ
∆ = (1.34)
where Lσ , θ and p,maxd represent, respectively, the liquid surface tension, the contact angle
between the liquid and the membrane, and the maximum membrane pore radius. In order to
prevent membrane wetting, the MC should be operated at a liquid pressure lower than the
breakthrough pressure (Li and Chen, 2005).
Dindore et al. (2004) evaluated several important criteria for the selection of
combinations membrane-solvent, like the critical entry pressure, the contact angle and the
critical solvent surface tension. Based on compatibility tests on several membrane
combinations (PTFE, PP, PVDF, PES, PS)/physical solvents (water, propylene carbonate,
selexol, N-methylpyrrolidone, dimethylformamide, tributylphosphate, glycerol triacetate, n-
formylmorpholine), the authors selected PTFE for determining the critical entry pressure
and contact angle. The other membranes showed incompatibility with the selected organic
solvents (morphological damage, swelling, shrinkages, color change, and dissolution).
Measurements were performed at room temperature using both flat and hollow fiber
membranes. CO2 absorption using selected membrane-solvent combinations was also
studied in order to determine the effect of membrane resistance on the overall performance
of the process and the influence of the membrane wetting behaviour. The results indicated
that the critical surface tension was independent of the porous or non-porous structure of
the material and that the contact angles decreased with the decrease of the liquid surface
tension. The membrane mass transfer resistance was found negligible in the non-wetted
operation mode and the Leveque equation (Kreulen et al., 1993) could be applied. In the
case of partial wetting, the overall mass transfer coefficients were lower than those
predicted by the Leveque equation and this difference increased with the increase of the
liquid velocity.
82
1.3.4. CO2 absorption in membrane contactors using SHA
Nii et Takeuchi (1994) experimentally studied CO2 absorption from a CO2/N2 gas
mixture in PTFE hollow fiber MC using various absorbents like AMP, MEA, DEA, DIPA,
MDEA, NaOH, K2CO3 and Na2SO3. Compared to carbonates, alkanolamines have the
advantage of high absorption capacity and rate. However, the use of aqueous alkanolamine
solutions had the drawback of higher regeneration energy requirements. The presence of
alkanolamines in aqueous carbonate solutions can lead to the enhancement of the CO2
absorption rate by improving simultaneously the regeneration efficiency. The work also
investigated the influence of the addition of small amounts of various alkanolamines to
aqueous carbonate solutions on the CO2 absorption flux, as well as the applicability of MC
for CO2 separation in the absence and the presence of SO2.
Kim and Yang (2000) used MC with microporous PTFE membranes to separate CO2-
N2 mixtures using AMP aqueous solutions. Mass transfer coefficients were determined 275
and 333 K. AMP solution concentrations ranged from 4 to 12 wt%. The separation
efficiency of AMP was compared with that MEA and MDEA, as well as with water. It was
observed that at high temperatures, the evaporated water filled the membrane pores and the
shell side, leading to a loss in the separation efficiency. Among the absorbents considered,
AMP exhibited higher absorption capacity and moderate absorption rate. CO2 removal
efficiency was found to increase with the increase of the liquid flow rate.
Wang et al. (2004) performed a theoretical simulation of pure CO2 capture in a non-
wetted hollow fiber MC using aqueous solutions of AMP, DEA and MDEA. The authors
investigated the influence of several parameters on CO2 removal efficiency, such as the
absorption liquid (physical properties, reaction kinetics), the operation conditions (liquid
flow velocity and concentration), and membrane characteristics (fiber length and radius).
AMP solution showed the best CO2 absorption capacity, followed by DEA and MDEA.
AMP and DEA absorption fluxes were found to be much higher compared to MDEA;
however, AMP and DEA concentrations dropped considerably due to depletion. Because of
faster kinetics, the liquid velocity and concentration, fiber length and radius showed a
significant influence on the CO2 absorption by AMP and DEA solutions.
83
deMontigny et al. (2005) compared the CO2 absorption performance of packed
columns and MC on the basis of the overall mass transfer coefficient used for evaluating
the performances of these absorption systems. Experimental data were obtained using
PTFE and PP hollow fiber membranes and MEA and AMP aqueous solutions as
absorbents. The effect of several operating parameters was studied (gas and liquid flow
rates and solution concentration). The results showed that at similar experimental
conditions, the MC system consisting in one, two and three modules in series performed
better than the columns containing Sulzer DX structured packing (Sulzer Chemtech. Ltd.).
However, the degree of improvement depended on the system configuration and membrane
type. PTFE membranes performed better than PP ones for both absorbents. On average, PP
and PTFE membranes gave mass transfer coefficients in the AMP/MC system that were
respectively, 18% and 430% better than the packed column. Similarly, PP and PTFE
membranes gave mass transfer coefficients in the MEA/MC system of 81% and 167% in
respect to the packed column. The reduced performance of PP membranes compared to
PTFE was attributed to membrane wetting and lower porosity of PP (35% for PP and 50%
for PTFE), as well as to the liquid channelling through the PP based module. MEA
performed better than AMP due to the faster reaction rate with CO2 compared to AMP.
deMontigny et al. (2006) extended the previous investigation concerning CO2
absorption in MEA and AMP aqueous solutions using PTFE and PP hollow fiber based
MC. The new experiments aimed to test the effect of module configuration and operation
conditions (gas phase circulating through the fiber lumen and liquid phase circulating
through the shell and vice-versa, co-current and counter-current flow orientation, and the
effect of using one, two or three modules in series). Compared to the previous publication
(deMontigny et al., 2005), the new results revealed that on average, the counter-current
operation mode performed 20% better than the co-current mode. Also, the circulation of the
liquid through the fiber lumen was shown to offer a significant improvement in the
performance compared to the liquid circulating through the shell side, due to the better
contact between the two phases.
84
Lu et al. (2007) investigated the CO2 capture from a CO2/N2 gas mixture in a hollow
fiber MC using aqueous solutions of MDEA in the absence and the presence of AMP or Pz
as activators. Mathematical simulations were validated with experimental data obtained at
room temperature (292-299 K) and atmospheric pressure for aqueous amine solutions of
total concentration of 2.5 kmol·m-3 (2.5 kmol·m-3 MDEA; 2.0 kmol·m-3 MDEA + 0.5
kmol·m-3 AMP; 2.0 kmol·m-3 MDEA + 0.5 kmol·m-3 Pz). Surprisingly, even though the
characteristics of the hollow fiber membrane module were given (fiber diameter and length,
thickness, average pore size, porosity), the membrane type (material) was not specified.
Experimental data and simulations results showed that the mass transfer in the MC can be
effectively enhanced by the addition of small amounts of activators in MDEA aqueous
solutions. However, as expected, Pz was found to be more efficient than AMP.
Paul et al. (2007) analysed theoretically the CO2 capture in a hollow fiber MC using
different aqueous single and blended alkanolamine solutions (MEA, DEA, MDEA, AMP,
MEA + MDEA, DEA + MDEA, MEA + AMP and DEA + AMP) of total amine
concentration of 10 wt%. Simulations were performed for the case of pure CO2 and a
CO2/N2 mixture containing 20 vol% CO2. It was concluded that the absorption fluxes of
CO2 in MEA + MDEA and MEA + AMP were higher than those in other blends.
Concerning the single amine solvents, the CO2 absorption capacity followed the sequence
MEA > AMP > DEA > MDEA, that was justified by the corresponding reaction kinetics.
Boucif et al. (2008) performed a numerical analysis of CO2 capture in PP HFMC using
aqueous solutions of AMP, DEA and DIPA. The analysis included the effect of various
parameters (liquid velocity, amine concentration, diameter and length of the fibers, and
external mass transfer coefficient) on the outlet CO2 concentration. The simulation results
showed that the use of aqueous AMP solutions leaded to a much higher absorption capacity
in comparison with the other two amine solutions, which was thought to be exclusively
determined by their difference in the reaction kinetics.
Lin et al. (2008) experimentally studied the performance of PVDF hollow fiber for
CO2 capture from CO2/N2 mixtures (1-15 vol% CO2) in a MC using as absorbents blended
aqueous solutions containing Pz (0.1-0.4 kmol·m-3) and AMP (1 kmol·m-3). As expected,
85
Pz addition to aqueous AMP solutions enhanced the CO2 absorption rate. In addition, the
increase of the Pz concentration in a blended AMP + Pz aqueous solution leaded to the
increase of the solution viscosity, which was believed to influence the membrane
wettability. That work also investigated the influence of gas and liquid flow rates on CO2
absorption rate and interfacial area, as well as the effect of absorbents on membrane
wetting phenomena and the resistance of mass transfer. An additional analysis concerning
CO2 absorption from CO2/N2 mixtures (1-15 vol% CO2) in PP and PVDF hollow fiber MC
using blended aqueous solutions containing Pz (0.1-0.4 kmol·m-3) + AMP (1 kmol·m-3) and
Pz (0.1-0.4 kmol·m-3) + MDEA (1 kmol·m-3) was published elsewhere by the same
research group (Lin et al., 2009a).
Lin et al. (2009c) aimed to study CO2 absorption from CO2/N2 mixtures (1-15 vol%
CO2) in a plasma-treated PP hollow fiber MC using various aqueous solutions containing 1
kmol·m-3 AMP, 1 kmol·m-3 MDEA and 0.1-0.4 kmol·m-3 Pz (single or blended amines).
The work investigated the effect of membrane material (non-treated PP, plasma-treated PP,
and PVDF), as well as the influence of gas and liquid flow rates and absorbent
concentration on CO2 absorption fluxes. It was observed that the durability of the PP-
plasma treated membranes was greatly improved in comparison with the non-treated PP
membranes. It was also found that PP durability was better than that of PVDF and
comparable to that of PTFE (deMontigny et al., 2006). However, the improvement was
more obvious for absorbents presenting a lower viscosity (single AMP aqueous solutions)
because of their lower tendency to penetrate the membrane pores, which therefore
diminished for blended AMP + Pz with the increase of Pz concentration (Lin et al., 2009c).
A poorer performance of PP membranes in contact with AMP aqueous solutions in
comparison with aqueous MEA solutions, related perhaps to the difference in the solution
viscosity and implicitly to the wetting behaviour was also observed by deMontigny et al.
(2005, 2006). Membrane mass transfer coefficients of the plasma-treated PP were shown to
be comparable to those of the PTFE hollow fibers reported by Yeon et al. (2003). The same
research group (Lin et al. (2009b)) also investigated CO2 absorption in plasma-treated flat
PVDF and PTFE MC using as absorbents aqueous solutions containing AMP, MDEA and
Pz (single or blended amines). Using aqueous AMP solutions, an increase of the CO2
86
absorption flux was observed for plasma-treated PVDF membranes in comparison with the
non-treated PVDF and PTFE ones.
CO2 absorption from CO2/N2 mixtures (1-15% CO2) in MDEA (1 M), AMP (1 M) and
AMP+Pz (1 M AMP+0.2 M Pz) aqueous solutions was also investigated by Lin et al.
(2009b) using flat PVDF, plasma-treated PVDF and PTFE membranes (one flat membrane
in the MC). The CO2 flux increased with the increase of gas flow rate and absorbent
concentration, the absorption process being dominantly governed by gas film diffusion and
membrane diffusion. The diffusion resistance in the membrane was not significant for SHA
based solutions. The treated PVDF membranes presented a higher water contact angle
(higher hydrophobicity) compared to pristine PVDF membranes. In consequence, both CO2
absorption flux and membrane durability were improved for the treated PVDV membranes
compared to pristine ones. Because of the smaller thickness of PVDF membranes, the CO2
absorption flux of the non-treated PVDV membranes was larger than that of PTFE.
However, the performance of non-treated PVDV membranes dropped after 12 days
compared to PTFE (still effective after 30 days) due to membrane deterioration by the
penetration of solution into the pores, behaviour that has already been reported in the
literature for PP (deMontigny et al., 2006) and PVDF (Yeon et al., 2005; Yeon et al., 2004)
membranes.
Rongwong et al. (2009) experimentally studied the performance of single and mixed
aqueous alkanolamine solutions containing AMP, MEA and DEA on CO2 absorption in
PVDF MC. The authors investigated CO2 absorption capacity and membrane wetting, as
well as the effect of the addition into the MEA solution of inorganic or organic salts (NaCl
and sodium glycinate) on the CO2 absorption flux and membrane wetting. Experiments
were performed at 303 K using a feed gas (CO2/N2) containing 20 vol% CO2. The results
showed that the absorption performance of single alkanolamine solutions followed the
order MEA > AMP > DEA. The mixed absorbents containing MEA provided higher
absorption flux, following the sequence MEA/AMP > MEA/DEA > AMP/DEA. The use of
mixed solutions did not elude the membrane wetting which leaded to the reduction of the
CO2 flux in the order AMP/DEA > AMP/MEA >MEA/DEA.
87
Ahmad et al. (2010) briefly investigated experimentally CO2 removal from a CO2/N2
feed gas containing 10-100 vol% CO2 in a PVDF flat-sheet MC using aqueous solutions of
AMP (1-5 kmol·m-3). The effect of CO2 concentration in the gaseous feed, alkanolamine
concentration and the membrane wetting behaviour were analysed.
Sohrabi et al. (2011) developed a 2D mathematical model to study CO2 transport
through hollow fiber MC, considering as absorbents aqueous solutions of AMP, MEA,
DEA, MDEA and K2CO3. Modeling results were validated with experimental data for CO2
absorption in MC using aqueous solutions of MEA taken from data already available in the
literature. The simulations indicated that CO2 removal increased with the increase of liquid
velocity and that the use of alkanolamines was more efficient than that of aqueous K2CO3.
CO2 absorption in a ternary mixture AMP+MEA+Pz was studied by Chen et al. (2011)
by using symmetric (traditional) and asymmetric (obtained by heating) PTFE hollow fiber
membranes. For all modules, CO2 recovery increased by increasing the liquid flow rates.
The authors found that the asymmetric membranes only brougt little improvement in CO2
recovery than the symmetric one, but the operational stability and durability of asymmetric
membranes was in long-term cyclic opearation.
Wang et al. (2013) investigated experimentally and theoretically CO2 absorption using
MEA, DEA and MDEA aqueous solutions, as well as the aqueous mixtures MEA/MDEA,
MDEA/Pz and DEA/AMP, in PP hollow fiber MC. For all amines solutions investigated,
the increase in the liquid velocity was found to slightly improve the mass transfer
coefficient due to the reduction of liquid side mass transfer resistance. For the AMP
containing mixed amine solution, the optimum composition to achieve the highest mass
transfer coefficient was found to be 15 wt % DEA with AMP in proportions from 0.5 to
0.8. For DEA/AMP blend, CO2 absorption was found to be controlled by the liquid-phase
mass transfer, compared to MEA/MDEA and MDEA/Pz blends where CO2 absorption was
controlled, respectivey, by combined liquid−gas phases and by a gradual transition from
liquid-side controlled to liquid−gas combined controlled as the concentration of Pz
increased.
88
1.4. Conclusions
The gas absorption using membrane contactors, combining the benefit of the
absorption (high selectivity) and those of the membranes (operational flexibility and easy
linear scale up, low capital and operation costs, high mass transfer rate), is a promising
alternative to traditional gas-liquid contactors. Despite several important advantages like (i)
large contact area for promoting an efficient gas-liquid mass transfer, (ii) high modularity
and compatibility for an easy scale-up, (iii) the possibility of varying stream flow rates
independently and without the occurrence of loading or flooding, (iv) easier performance
prediction due to the fact that the interfacial area is known and constant, the main drawback
is the additional mass transfer resistance taking place in the membrane which can become
important when the membrane is wetted by the absorbent (pores partially or totally filled by
liquid).
The performances of MC for CO2 separation from different industrial flue gases
strongly depend on the properties of both absorption liquid and membranes, the
compatibility between them and the constructive characteristics of MC modules (flow
configuration, module geometry, operating parameters). The prevention of the unwanted
wetting phenomenon, a major obstacle in their implementation in industrial separation
processes, requires membranes which are highly hydrophobic, liquid absorbents with
dedicated properties and good compatibility between membranes and liquid. The proper
choice of the membrane/absorption liquid combination is therefore a crucial step in
developing CO2 absorption in MC.
Among the available polymeric membranes (usually PP, PVDF and PTFE), only PTFE
is suggested as a suitable membrane for use in the presence of alkanolamines. The use of
the two others showed the occurrence of pore wetting and/or incompatibility with the amine
solutions (morphological damage, swelling, shrinkage, and color change).
The amines are the most used absorbents for CO2 removal from different gas mixtures.
Among them, the sterically hindered amines (AMP, a hindrance form of MEA, being the
first SHA introduced in the literature) attracted recently much attention. Their application
89
in gas-treating technology offers absorption capacity, absorption rate, and degradation
resistance advantages over conventional amines. However, except for AMP, very limited
information concerning the properties of other potential SHA is available, including their
capacity for CO2 separation.
MEA is the most investigated absorbent for CO2 capture in MC, certainly because
MEA is the benchmark amine industrially used in acid gas separations. AMP, alone or
mixed with other amines (usually, accelerators for improving the absorption kinetics) is the
single SHA investigated. However, its low surface tension does not make it very
appropriate for use in MC. Being the less hindered amine, its low sterically hindered
character leads to the formation of carbamate, along with bicarbonate. More hindered
amines are less favourable to carbamate formation. Especially based on their high
absorption capacities and low regeneration cost, SHA seem to be very appropriate to be
used in the blend solutions. More advanced investigations of other SHA are therefore
needed.
Most CO2 absorption studies concern hollow fiber membrane contactors. Despite the
advantages of flat sheet membrane contactors, information on CO2 absorption in this type
of contactor is extremely scarce and the very few available studies are limited to a single
membrane in the MC module.
Very few efforts have been made to investigate new CO2 absorbents especially
optimized for application in MC. Besides appropriate performances in CO2 separation
(absorption capacity, absorption kinetics, degradation resistance and regeneration facility),
it is crucial for the absorption solutions intended to be used in MC to have a high surface
tension in order to reduce the membrane wetting tendency. Thorough studies need to be
done in the development of dedicated absorbents and the evaluation of all their properties
related to their absorption/regeneration efficiency, stability and resistance to degradation,
compatibility with the membrane (evaluation of surface tension and wetting behaviour) and
application for CO2 absorption in MC.
90
1.5. Objective of the work
For successful gas separation applications in membrane contactors, the absorbent requires
several important characteristics:
(i) Good absorption properties for assuring a high efficiency of the process (high absorption capacity and kinetics)
(ii) Facility in solvent regeneration, low degradation degree and low corrosive character
(iii) Good compatibility with the membrane (high surface tension, no chemical degradation on membranes)
(iv) Availability and low cost
Very few efforts have been made to investigate new absorbent solutions especially
optimized for application in MC. Currently, no available absorbent meets all required
characteristics for implementation the membrane contactors for CO2 capture in industrial
units.
In this context, the following objectives were defined:
General objectives
The main objectives of this thesis are (i) to develop a dedicated sterically hindered
alkanolamine based CO2 absorbent with improved characteristics for application in MC and
(ii) to investigate its efficiency for CO2 separation in both hollow fiber and flat sheet MC.
Specific objectives
• study of the hindrance effect on SHA absorption/regeneration properties
• study of the compatibility between absorbent and membrane (PTFE membrane was
chosen for this work)
• investigation of the absorbent performance for CO2 absorption in PTFE hollow fiber
and flat sheet membrane contactors, and optimization of operating conditions (effect of
operation parameters)
91
92
First, the choice of the absorbent to be used in MC has to be based on properties related to
its behavior in reaction with CO2 (absorption capacity, absorption kinetics, regeneration
facility, and degradation resistance). As we chose sterically hindered alkanolamine (SHA)
based solutions as potential absorbents, the first chapter concerns the study of the
hindrance effect on the kinetics of different SHA: AMP (a simple hindrance form of MEA)
and three SHA derived from AMP (AEPD, AMPD, and AHPD). For this study, the kinetics
of the reaction between CO2 and AHPD was performed experimentally in a wetted wall
contactor and discussed together with data available in the literature for the other systems.
93
Chapter 2. Kinetics of absorption of carbon dioxide into aqueous solutions of 2-amino-2-hydroxymethyl-1,3-propanediol
Résumé
Dans cette étude, la cinétique de la réaction entre le CO2 et 2-amino-2-hydroxyméthyle-1,3-propanediol (AHPD), une amine à encombrement stérique, a été déterminée à 303.15, 313.15 et 323.15 K dans une colonne à parois mouillée. La plage de concentrations de la solution aqueuse d’AHPD a été 0.5 - 2.4 kmol m-3. Sur la base des données d’absorption physique de CO2 et N2O dans l’eau et du N2O dans les solutions d’amines, le rapport entre le coefficient de diffusion et la constante d’Henry pour le CO2 dans les solutions a été estimé par l’analogie avec le N2O. En considérant le pseudo-ordre 1 pour l’absorption du CO2, les constantes de vitesse ont été aussi déterminées. Les constantes de déprotonation du zwitterion et la constante de vitesse d’ordre 2 ont été calculées sur la base du mécanisme de zwitterion pour la réaction entre le CO2 et l’AHPD. Pour les trois températures, 303.15, 313.15 et 323.15 K, les valeurs suivantes ont été obtenues pour la constante de vitesse d’ordre 2 (k2): 285, 524 et 1067 m3 kmol−1 s−1, respectivement.
94
Abstract
In this work the kinetics of the reaction between CO2 and a sterically hindered alkanolamine, 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD) were determined at temperatures of 303.15, 313.15 and 323.15 K in a wetted wall column contactor. The AHPD concentration in the aqueous solutions was varied in the range 0.5 - 2.4 kmol m-3. The ratio of the diffusivity and Henry’s law constant for CO2 in solutions was estimated by applying the N2O analogy, using the physical absorption data of CO2 and N2O in water and of N2O in amine solutions. Based on the pseudo-first-order for the absorption of CO2, the overall pseudo-first-order rate constants were determined from the kinetics measurements. By considering the zwitterion mechanism for the reaction of CO2 with AHPD, the zwitterion deprotonation and second-order rate constants were calculated. The second-order rate constant, 2k , was found to be 285, 524, and 1067 m3 kmol−1 s−1 at 303.15, 313.15, and 323.15 K, respectively.
95
2.1. Introduction Globally, approximately one third of all anthropogenic CO2 emissions come from
fossil fuels such as coal and oil used for generating energy. A variety of industrial
processes also emit large amounts of CO2 from each plant, for example oil refineries,
cement works, and iron production (IPCC, 2005). There is growing political and public
concern supported by consensus among the scientific community that global emissions
growth will soon drive atmospheric CO2 concentrations to levels never seen, bringing a
growing risk of fast climate change. The Canadian Environmental Protection Act (CEPA,
2005) is the legislative authority in Canada that pushes Canadian companies to reduce their
greenhouse gas production. These emissions could be reduced substantially by capturing
and storing the CO2.
Actual industrial absorption processes use aqueous solutions of alkanolamines. For
technical, economical and environmental concerns, this technique is widely applied for (i)
acid gases (CO2, H2S) removal during natural gas sweetening and (ii) CO2 capture from
fossil-fuel-fired power plants, as well as some other important industries such as chemical
and petrochemical, steel, and cement production. Industrially more often used
alkanolamines are monoethanolamine (MEA), diethanolamine (DEA), diisopropanolamine
(DIPA), N-methyldiethanolamine (MDEA), 2-amino-2-methyl-1-propanol (AMP) (Kohl
and Nielsen, 1997). The choice of a certain amine (single or blended amine) is mainly
based on the absorption capacity, reaction kinetics and regenerative potential and facility.
The key advantage of the primary and secondary alkanolamines such as MEA and DEA is
their fast reactivity due to the formation of stable carbamates. Conversely, this will lead to
very high solvent regeneration cost. On the absorption capacity side, they have the
drawback of a relatively low CO2 loading (limited to 0.5 mol CO2/mole amine). Tertiary
alkanolamines, like MDEA, have a low reactivity with respect to CO2, due to the exclusive
formation of bicarbonates by CO2 hydrolysis. However, this will lead to a very low solvent
regeneration cost. Another advantage of these amines is the high CO2 theoretical loading
capacity of 1 mol of CO2/mol of amine. The application of sterically hindered amines, e.g.,
2-amino-2-methyl-1-propanol (AMP) in gas-treating technology offers absorption capacity,
absorption rate, selectivity and degradation resistance advantages over conventional amines
96
for CO2 removal from gases (Sartori and Savage, 1983). Sterically hindered amines (SHA)
form unstable carbamates due to the hindrance of the bulky group adjacent to the amino
group. Hydrolysis of the voluminous carbamates leads to a preferential bicarbonate
formation process, resulting in the theoretical loading capacity up to 1.0. Reaction kinetics
significantly higher than those related to tertiary amines, coupled with a low solvent
regeneration cost offer to SHA important industrial advantages. However, except for AMP,
data concerning the other potential SHA are quite scarce. Moreover, except for one work on
the substituent effect in amine-CO2 interaction investigated by NMR and IR spectroscopies
(Yoon and Lee, 2003), systematic studies on the relation structure-properties in close
connection to the CO2 absorption process are practically inexistent.
In our laboratory, extensive studies of CO2 capture in membrane contactors using SHA
based alkanolamine mixtures are in progress. In this context, in order to study the hindrance
effect on the absorption capacity and kinetics of SHA, a set of four SHA was chosen (Table
1.1. (Chapter 1)). It concerns AMP, a simple hindrance form of MEA, and three SHA
derived from AMP: 2-amino-2-methyl-1,3-propanediol (AMPD), 2-amino-2-ethyl-1,3-
propanediol (AEPD) and 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD). Few kinetic
studies on the systems CO2-AMP (or AMPD or AEPD) are available in the open literature:
AMP (Alper, 1990; Saha et al., 1995; Xu et al., 1996; Yih and Shen, 1988), AMPD
(Bouhamra et al., 1999; Yoon et al., 2003); AEPD (Yoon et al., 2002a). On our knowledge,
kinetic studies involving AHPD are not available in the open literature.
In this work, kinetics study of AHPD has been performed using a wetted wall
contactor. The ratio between the diffusion coefficient and Henry’s law constant, given by
the function 1/ 2A A/D H (Danckwerts, 1970), was estimated by applying the N2O analogy and
the Higbie penetration theory, using the physical absorption data of CO2 and N2O in water
and of N2O in amine solutions. Based on the pseudo-first-order for the absorption of CO2,
the overall pseudo-first-order rate constants were determined from the kinetics
measurements. By considering the zwitterion mechanism for the reaction of CO2 with
AHPD, the zwitterion deprotonation and second-order rate constants were calculated at
303.15, 313.15 and 323.15 K.
97
2.2. Theory 2.2.1. Physical absorption
Physicochemical properties of CO2 in aqueous alkanolamine solutions such as the
diffusion coefficient and Henry’s law constant cannot be found directly as CO2 react in
solutions. Hence, the N2O analogy is a useful method widely used in similar works (Dang
and Rochelle, 2003; Yih and Shen, 1988; Yoon et al., 2002a). For the analogy to apply, the
parameters characterising the physical absorption of CO2 and N2O in water and of N2O in
amine solutions need to be known.
With initial gas-free liquids and for short contact time between the gas j and the liquids
in the wetted wall contactor, the Higbie penetration theory (Higbie, 1935) is commonly
used (Alvarez-Fuster et al., 1980; Danckwerts, 1970) and gives the specific absorption rate
as:
1/2
2 .j jj
c j
D PN
t Hπ
= (2.1)
The contact time (tc) can be derived from the wetted wall column hydrodynamics (Roberts
and Danckwerts, 1962):
1/32 /32 3 .
3ch dt
L gπ µ
ρ =
(2.2)
The combination of Eqs. (2.1) and (2.2) gives:
1/61/2 1/31/2 2 3 .
2 3j j
j j
D N h dH P L g
π π µρ
= (2.3)
From this last equation, the ratio of the diffusivity and Henry’s law constant can be
calculated by the specific gas absorption rate at several flow rates, L, and for different
heights of effective wetted surface, h, (Nysing and Kramers, 1958) for a constant
temperature and liquid concentration. Here, Nj can be calculated from the total absorption
rate divided by the effective absorption area:
,Tj
js
NN
d hπ= (2.4)
and ds is the diameter of the wetted wall column including the thickness of the laminar film:
98
1/3
3 .sLd d
gdµ
πρ
= +
(2.5)
2.2.2. Chemical absorption
The kinetics of primary and secondary alkanolamines with CO2 can be described using
the zwitterion mechanism proposed first by Caplow (1968) and reintroduced later by
Danckwerts (1979). This mechanism has been used successfully with conventional and
sterically hindered alkanolamines such as DEA, DIPA, AMP, AEPD, AMPD and 2-
Piperidineethanol (2-PE) (Blauwhoff et al., 1984; Shen et al., 1991; Sun et al., 2005; Yoon
et al., 2003; Yoon et al., 2002a). The first step of this mechanism in the reaction of CO2
with AHPD is the formation of a zwitterion:
2 + -2 2 2
-1CO + RNH RNH COO .
k
k→← (2.6)
The second reaction is the removal of the proton of the zwitterion by any base existing in
the solution.
+ - - +b2RNH COO + Base RNHCOO + BaseH .k→ (2.7)
In our solutions, AHPD, OH- or H2O can contribute to this step:
AM+ - - +2 2 3RNH COO + RNH RNHCOO + RNH ,k→ (2.8)
-OHk+ - - -
2 2RNH COO + OH RNHCOO + H O ,→ (2.9)
H O2k+ - - +2 2 3RNH COO + H O RNHCOO + H O .→ (2.10)
Assuming a quasi-steady-state condition for the zwitterion concentration and an irreversible
deprotonation step by bases, the kinetic rate equation for CO2-AHPD is given by:
2CO -AM app A = ,r k C (2.11)
where the pseudo-first-order apparent reaction rate constant, kapp, is defined as:
99
- 2- 2
Bapp
2 H O22 OH2 AMB H OOH
-1 -1 -1
= .1 1 + + +
Ck
k kk kk k k C C Ck k k
(2.12)
In an aqueous system other reactions can also occur:
*H O2 - +
2 2 3CO +H O HCO + H ,k←→ (2.13)
*
-OH- -2 3CO + OH HCO .k
←→ (2.14)
Eq. (2.13) may usually be neglected because it proceeds very slowly: 2
*H Ok = 0.026 s-1 at
298.15 K (Pinsent et al., 1956). The second reaction is the bicarbonate formation and it can
enhance mass transfer even when the concentration of hydroxyl ion is low (Pinsent et al.,
1956). Therefore, the expression of the kinetic rate equation for CO2-OH- can be expressed
as:
- - -2
*ACO -OH OH OH
= ,r k C C (2.15)
where
-*OH
2895log = 13.635 - +0.08 .k IT
(2.16)
In Eq. (2.16), I is the ionic strength, defined as ½ the sum of the molarities of the different
ions ( )- - 2- + +3 3 3OH , HCO , CO , RNH , H multiplied by the square of the corresponding
electric charge. In sterically hindered alkanolamine solutions (Astarita et al., 1983) the
hydroxyl ion concentration was estimated by the relation (2.17):
-
-3wOH
p
-3wB
p
1- = , 10 ,
= , 10 .
KCK
K CK
α αα
α
≥
< (2.17)
The values of the dissociation constant for water, Kw, and the protonation constant for
AHPD, Kp, were taken from Covington et al. (1977) and Perrin (1965), respectively. The
H+ concentration was calculated from the water dissociation constant. The concentration of
bicarbonate and carbonate ions was calculated using the pH value, the second dissociation
constant of carbonic acid (Edwards et al., 1978) and the assumption that all absorbed CO2
100
is converted into these two species. Protonated amine concentration was calculated with the
AHPD protonation constant and the pH value.
Based on the Eqs. (2.11) and (2.15), the overall pseudo-first-order reaction rate
constant can therefore be expressed as:
- -*
ov app OH OH = + k k k C (2.18)
2.3. Experimental 2.3.1. Reagents
Aqueous AHPD solutions were prepared with degassed distilled water and 2-amino-
2-hydroxymethyl-1,3-propanediol with a minimum purity of 99.9 %. Tween 80 was used as
a surface active agent and was added at 0.04 vol% in AHPD solutions to avoid ripple
formation. All chemicals (Laboratoire MAT, Quebec, Canada) were used without further
purification. Gases (CO2, N2O and N2) were of commercial grade with a minimum purity of
99.9 % (Praxair).
2.3.2. Experimental setup
A wetted wall column similar to the apparatus described by Robert and Danckwerts
(1962) was build and used in this study. The column, made of stainless steel, has an outside
diameter of 1.905×10-2 m and the length of the absorption surface could be varied between
0.03 and 0.11 m. An overall flow diagram of the experimental setup is shown in Figure 2.1.
The column assembly was kept in an air bath controlled within ± 0.1 K with a temperature
controller (OMEGA, CN76000). The aqueous alkanolamine solutions were also kept in a
thermostated reservoir controlled by the same type of controller. The flow rate of input
gases was adjusted with mass flow controllers (OMEGA, FMA-100 series) and each
controller was calibrated for a specific gas using a bubble flowmeter. The accuracy of the
flow was estimated to be ± 0.5%. Gas chromatography (Perkin Elmer, AutoSystem) was
used to determine the inlet and outlet gas composition (for chemical absorption) and to
confirm complete removal of the air in the wetted wall contactor before each run. Aqueous
alkanolamine solutions were supplied to the column from 2×10-6 m3 s-1 to 5×10-6 m3 s-1
101
using a digital gear pump (Cole-Parmer, K-74014-40) and a digital volumetric flowmeter
(Cole-Parmer, K-32718-24) with an accuracy of ± 1%.
Figure 2.1. Schematic overall experimental flowsheet.
2.3.3. Experimental procedure
In a typical run, the aqueous alkanolamine solutions and the air bath were first
brought to the desired temperature. All experiments were done at 303.15, 313.15 and
323.15 K and for solutions concentration varying between 0.5 and 2.4 kmol m-3. The
concentration of the amine solutions (prepared gravimetrically) was checked with HCl
solutions and a methyl red-bromocresol green pH indicator mix. For physical absorption of
CO2 in water or of N2O in amine solutions, pure gases were used and absorption rates of
CO2 and N2O were measured by a bubble flowmeter. For chemical absorption, CO2 was
mixed with nitrogen to give a range of CO2 partial pressures from 10 to 82 kPa. The
absorption rate was measured as a function of the inlet gas flow rate and the difference
between the inlet and the outlet CO2 composition in the gas determined by gas
chromatography. The gas chromatograph was equipped with a thermal conductivity
detector and a Carboxen 1010 plot capillary column (30m×0.53mm). A carrier gas flow
102
rate of 3.08×10-7 m3s-1 was used and the temperatures of the detector and the column were
of 503.15 K and 398.15 K, respectively. The measured flow rate was always corrected for
the vapour pressure of water as a function of temperature and the value of the CO2 partial
pressure used for the calculation purpose was taken as the logarithmic mean between the
inlet and the outlet CO2 partial pressure (Dang and Rochelle, 2003). The liquid flow rate
and the height of the column were selected in such a way that the chemical absorptions
occurred in the fast pseudo-first-order reaction regime (Danckwerts, 1970). In this regime,
the Hatta number almost equals the enhancement factor when
inf2 < ,aH E (2.19)
where the Hatta number, Ha, is defined, after the assumption that the partial order for CO2
is one (this hypothesis will be verified later), as
A ov
L
= ,a
D kH
k (2.20)
and the infinite enhancement factor based on penetration theory, Einf, is
11/ 2
A A Binf B
B A A
= .D D CE DD H v P
−
+ ⋅ (2.21)
The liquid phase mass-transfer coefficient for physical absorption is calculated with the
definition given by the Higbie penetration theory as:
AL = 2
c
Dktπ
(2.22)
In this regime, the specific absorption rate is then
1/ 2A
A A ovA
.DN P kH
=
(2.23)
By experimental data regression, the Eq. (2.23) allowed us to determine the partial order of
the reaction with respect to CO2 and the kinetic reaction rate constants. It must be noted
that for physical and chemical absorptions the liquid flow on the wetted wall column was
103
always laminar. The highest Reynolds number obtained was around 150, which is much
less than the criterion of 250 proposed by Danckwerts (1970).
2.4. Results and Discussions 2.4.1. Physicochemical properties of aqueous AHPD solutions
Based on the experimental data of Park et al. (2002a), available for the temperature
range 303.15-343.15 K and AHPD concentration range 5-25 mass %, we obtained the
following correlations for the density and viscosity of aqueous AHPD solutions:
( )2
-3 2B
i = 0/(kg m ) = 1000 ( + + ) ,i
i i ia b T c T Cρ ⋅ ⋅ ⋅∑ (2.24)
( )
22
B-1 -1 i = 0
( + + )/ (kg m s ) = ,
1000
ii i ia b T c T C
µ⋅ ⋅∑
(2.25)
where ai, bi and ci are the regressed coefficients presented in Table 2.1, T is the absolute
temperature and the amine concentration is expressed in kmol m-3. The “Stepwise”
regression method was used (Montgomery and Runger, 1999); only the statistically
significant coefficients are therefore found in the regression, the others equal 0. R2 for the
Eqs. (2.24) and (2.25) are 0.999 and 0.998, respectively.
2.4.2. Physical absorption
In order to validate the wetted wall column contactor and the experimental
procedure, physical absorption of CO2 in water was performed at 305.65 K in order to
obtain the ratio2 2
1/ 2CO CO/D H , as explained in the section 2.2.1. The measured value, 1.35×10-8
kmol kPa-1m-2s-1/2 obtained with an estimated experimental uncertainty of ±2%, is in good
agreement with the literature values (Al-Ghawas et al., 1989; Li and Lai, 1995; Li and Lee,
1996; Mandal et al., 2004; Saha et al., 1993) as shown in Figure 2.2.
104
Figure 2.2.
2 2
1/ 2CO CO/D H ratio for the absorption of CO2 in water as a function of
temperature. Dotted lines are for trend only.
Table 2.1. Regressed coefficients for density, viscosity and ( )2 2
1/ 2N O N O AHPD
/D H
correlations
Density ai bi ci
i = 0 1.0666 0 -7.5553 × 10-7
i = 1 2.9602 × 10-2 0 0 i = 2 0 0 0
Viscosity ai bi ci
i = 0 2.2079 × 101 -1.2247 × 10-1 1.7347 × 10-4
i = 1 0 0 0 i = 2 5.6873 -3.1314 × 10-2 4.3489 × 10-5
( )2 2
1/ 2N O N O AHPD
/D H
ai bi ci
i = 0 3.8551 × 10-8 -9.5593 × 10-11 0 i = 1 0 0 -1.7436 × 10-14
i = 2 0 0 1.6068 × 10-15
105
The absorption of N2O in AHPD aqueous solutions ranging from 0.5 to 2.4 kmol m-
3 was performed at 303.15, 313.15 and 323.15 K. Figure 2.3 shows the measured values of
the ratio 2 2
1/ 2N O N O/D H along with the curves obtained using the following correlation:
( )2
2
1/ 2 2N O -1 -2 -1/2 2
Bi = 0N O AHPD
/(kmol kPa m s ) = ( + + ) ,ii i i
Da b T c T C
H
⋅ ⋅
∑ (2.26)
where the regressed coefficients obtained using the “Stepwise” regression method (as
explained in the section 2.4.1.) are found in Table 2.1. Eq. (2.26) agrees to our
experimental data within a mean absolute deviation of 1.6%. In Figure 2.4, 2 2
1/ 2N O N O/D H
values for the system N2O-water were taken from the literature (Horng and Li, 2002; Sun et
al., 2005; Versteeg and Vanswaaij, 1988). Experimental data show a decrease in the ratio
value with an increase in amine concentration or an increase in temperature. This trend is in
agreement with the data of Yih and Shen (1988) and Yoon et al. (2002).
Figure 2.3. 2 2
1/ 2N O N O/D H ratio for N2O in aqueous AHPD solutions.
2 2
1/ 2CO CO/D H in the amine solutions was calculated using the N2O analogy:
106
2 2
2 2 2 2 2
2 2 2
1/ 2N O N O AHPD1/ 2 1/ 2
CO CO AHPD CO CO H O1/ 2N O N O H O
( / )( / ) = ( / ) ,
( / )D H
D H D HD H
⋅ (2.27)
where the diffusivity and the Henry’s law constant of CO2 and N2O in water are given by
the regressed equations of Versteeg and van Swaaij (1988):
( )22
2 -1 -6CO H O
-2119/(m s ) = 2.35 10 exp ,DT
×
(2.28)
( )22
2 -1 -6N O H O
-2371/ (m s ) = 5.07 10 exp ,DT
×
(2.29)
( )22
3 -1 6CO H O
-2040/(kPa m kmol ) = 2.825 10 exp ,HT
×
(2.30)
( )22
3 -1 6N O H O
-2284/(kPa m kmol ) = 8.547 10 exp ,HT
×
(2.31)
where T is the absolute temperature.
Table 2.2. Kinetic data for absorption of CO2 in AHPD aqueous solutions
at 303.15 K
CAHPD 2COP
2CON × 106 kL × 104 kov kapp Ha E Einf
(kmol m-3) (kPa) (kmol m-2 s-1) (m s-1) (s-1) (s-1) 0.50 11.86 1.224 1.705 68.7 56.1 2.08 2.25 101.91 0.50 38.88 4.108 1.705 72.1 68.3 2.13 2.27 32.10 0.50 81.77 8.556 1.760 70.7 68.7 2.04 2.16 16.02 1.00 65.66 9.909 1.575 172.8 169.9 3.42 3.43 38.89 1.51 12.15 2.222 1.588 298.0 281.7 4.29 4.36 312.92 1.51 38.15 7.228 1.631 320.0 312.0 4.33 4.34 100.70 1.51 80.32 14.416 1.672 287.2 282.8 4.00 3.97 48.61 2.40 11.50 2.384 1.318 496.4 476.1 6.26 6.30 528.90 2.40 38.72 8.462 1.272 552.2 544.8 6.84 6.77 158.25 2.40 80.81 16.862 1.318 503.4 499.2 6.31 6.15 76.64
107
2.4.3. Chemical absorption
Following the procedure and equations described in the sections 2.2.2 and 2.3.3, the
chemical absorption of CO2 in AHPD solutions was studied in order to determine the order
of carbon dioxide in the CO2-AHPD reaction and the kinetic reaction rate constants. The
experimental results are presented in Tables 2.2-2.4. Figure 2.4 presents the specific
absorption rates of CO2 in the amine at different temperatures and for a CO2 partial
pressure of around 80 kPa. It can be seen that the trends are in agreement with other CO2-
alkanolamine systems studied in the literature (Gianetto et al., 1986).
Table 2.3. Kinetic data for absorption of CO2 in AHPD aqueous solutions at 313.15 K
Table 2.4. Kinetic data for absorption of CO2 in AHPD aqueous solutions
at 323.15 K
CAHPD 2COP
2CON × 106 kL × 104 kov kapp Ha E Einf
(kmol m-3) (kPa) (kmol m-2 s-1) (m s-1) (s-1) (s-1) 0.50 12.41 1.522 1.812 120.6 102.6 2.87 2.98 120.59 0.50 38.38 4.816 1.773 126.3 121.0 3.00 3.05 39.97 0.50 80.11 10.264 1.773 131.7 129.2 3.06 3.03 19.90 1.01 64.69 10.906 1.756 279.0 273.0 4.21 4.17 50.38 1.50 11.36 2.259 1.682 470.6 436.2 5.35 5.40 433.63 1.50 37.05 7.387 1.682 472.9 458.0 5.36 5.35 133.89 1.50 78.56 15.096 1.708 439.3 431.7 5.09 5.01 63.88 2.40 11.38 2.574 1.371 849.8 806.8 7.89 7.90 726.41 2.40 36.57 8.322 1.371 860.8 842.1 7.94 7.87 227.12 2.40 78.80 17.935 1.323 860.8 853.0 8.22 8.00 106.14
CAHPD 2COP
2CON × 106 kL × 104 kov kapp Ha E Einf
(kmol m-3) (kPa) (kmol m-2 s-1) (m s-1) (s-1) (s-1) 0.49 10.16 1.431 2.063 204.7 158.4 3.43 3.52 182.67 0.49 34.58 5.040 2.026 219.1 206.7 3.61 3.63 54.60 0.49 74.48 10.147 2.026 191.4 185.3 3.37 3.33 26.07 1.00 61.03 11.510 1.897 470.5 457.7 5.13 5.05 68.07 1.50 10.86 2.459 1.654 868.5 799.8 7.25 7.27 606.32 1.50 35.05 7.704 1.707 819.2 789.8 6.82 6.77 188.78 1.50 73.50 16.126 1.758 816.0 800.7 6.61 6.47 90.65 2.40 10.99 2.810 1.267 1721.8 1636.1 11.19 11.18 1069.39 2.40 34.57 8.795 1.312 1705.4 1669.0 10.76 10.64 340.82 2.40 74.79 18.641 1.312 1636.5 1619.3 10.54 10.26 158.13
108
-13.5
-13.0
-12.5
-12.0
-11.5
-11.0
-10.5
2.0 2.5 3.0 3.5 4.0 4.5 5.0
ln [N
A/(k
mol
m-2
s-1)]
ln [PA/(kPa)]
, 303.15 K
, 313.15 K
, 323.15 K
Figure 2.4. Specific absorption rate as a function of amine concentration for 2COy = 0.8.
Figure 2.5. Specific absorption rate as a function of CO2 partial pressure for an aqueous AHPD solution of 1.5 kmol m-3.
4
6
8
10
12
14
16
18
20
0 0.5 1 1.5 2 2.5 3C B/(kmol m-3)
, 303.15 K
, 313.15 K
, 323.15 K
109
As shown by Yih and Shen (1988), the value of the slope ln NCO2 versus ln PCO2 at a
constant temperature and at constant amine concentration equals one if the partial order of
CO2 in the CO2-AHPD reaction is one. In this study, the slopes of the lines of similar plots
were ranging from 0.983 to 1.023 for all studied temperatures and concentrations. It is then
possible to confirm that the reaction between CO2 and AHPD is first order with respect to
CO2. Figure 2.5 is an example of theses plots; the lines represent the data for the CO2
absorption in 1.5 kmol m-3 AHPD solutions. Similar conclusions were found for others
alkanolamines like, for example, AMP (Yih and Shen, 1988) or AEPD (Yoon et al., 2002).
It has been shown in the literature (Gianetto et al., 1986) that the partial order with
respect to CO2 is always 1, as we obtained in this work based on the experimental data, but
the partial order with respect to the amine can vary between 1 and 2 depending on the
chosen amine. In the case of the system CO2-aqueous AHPD, a zwitterion mechanism was
considered (Eqs. (2.6)-(2.10)). By combining Eqs. (2.11) and (2.12), we obtain the
following general kinetic rate equation applied to the zwitterion mechanism:
2
1 12CO -AM A B
-1
Base
= , 1 +
b
kr C Ckk C
∑
(2.32)
where the “base” partial orders applied to the CO2 and the amine are emphasised
(corresponding exponents equal to 1). Taking into account that the term in the parenthesis
of Eq. (2.32) is a function of the amine concentration and it does not depend on the gas
concentration, the partial order with respect to CO2 cannot vary from 1, which is not the
case for the amine. Depending on the expression in the parenthesis of Eq. (2.32), the
“apparent” partial order with respect to the amine may shift from 1 to 2 following the two
extreme cases (Alvarez-Fuster et al., 1981; Danckwerts, 1979; Derks et al., 2006):
i) if -1
B Base
1kk C∑
, the “apparent” amine partial order equals 1, (2.33)
ii) if -1
B Base
1kk C∑
, the “apparent” amine partial order equals 2. (2.34)
110
Between these two extreme cases, the “apparent” partial order with respect to the amine can
vary from 1 to 2, as shown, for example, by Yoon et al. (2003) for the CO2-aqueous AMPD
system. The same behaviour can be observed for the system studied in this work, CO2-
aqueous AHPD, where the ratio -1 B Base / k k C∑ varies between 0.35 and 2.10. However, in
order to determine the reaction rate parameters ( 2k , 2 AM 1/k k k− and 22 H O 1/k k k− ) by
regression, it is not necessary to know the exact value of “apparent” partial order with
respect to the amine, as it will be described subsequently .
Based on the Eqs. (2.18) and (2.23), kov and kapp were calculated and listed in Tables
2.2-2.4. The average absolute deviation between kov and kapp was found to be 4.8% when all
experimental points are considered, and 2.5% when the data obtained for the smallest CO2
partial pressure (around 10 kPa) are neglected. The explanation for this phenomenon is that
higher CO2 partial pressures lead to higher specific absorption rates and therefore, for a
constant liquid flow, higher loadings are obtained. The contribution of hydroxyl ions to the
overall absorption rate is then lowered. Nevertheless, the contribution of hydroxyl ions is
still low and the presence of related terms in the kapp and kov expressions may then be
neglected without significant loss of accuracy. The same behaviour was also observed by
others authors (Xu et al., 1996; Horng and Li, 2002; Sun et al., 2005).
The Hatta number, Ha and the infinite enhancement factor, Einf are calculated from
Eqs. (2.20) and (2.21), respectively and are given in Tables 2.2-2.4. The diffusion
coefficient of CO2 in aqueous AHPD solutions, ( )2CO AHPDD was estimated using the
( )2 2
1/2CO CO AHPD
/D H ratio obtained in the present experimental work and the extrapolated
values of ( )2CO AHPDH taken from Le Tourneux et al. (2008). The diffusion coefficient of
AHPD in aqueous AHPD solutions, DB, was calculated using a modified Stokes-Einstein
relation (Versteeg and van Swaaij, 1988):
-0.6B w w = ( / ) ,D D µ µ⋅ (2.35)
where the subscript « w » refers to infinite dilution state (pure water). Dw was obtained
using an equation based on molecular volumes of the solute and the solvent, as described in
Scheibel (1954). This correlation was chosen because it gives the lowest average error
111
between the predicted and experimental data for large molecules like AHPD (Hikita et al.,
1979; Snijder et al., 1993). The calculated values of Ha and Einf proved that Eq. (2.19) is
respected.
The fast pseudo-first-order regime was also verified on the entire length of the
column, by analysing the gas and amine concentration profiles in the liquid film. They were
obtained by solving the equations set (2.36)-(2.40) based on the following assumptions: (i)
steady-state and isothermal conditions; (ii) plug flow for both liquid and gas,
L
A_L ALA 0 ,L V
x
dC dCu D adz dx δ=
+ = (2.36)
L
B_L BL 0 ,L B Vx
dC dCu D adz dx δ=
+ = (2.37)
A_G ALG
0
0 ,A Vx
dC dCu D adz dx =
+ = (2.38)
2
ALA A2 0 , CD r
x∂
− =∂
(2.39)
2
BLB B A2 0 ,CD r
xν∂
− =∂
(2.40)
with the following boundary conditions in the axial and radial directions (counter-current
flow):
at 0z = : AG,0A_L
A
C RTC
H= , B_L B,exitC C= , A
A_G A_G,0PC CRT
= = , (2.41)
at 0x = , for all z: A_GAL A,i A_G
A
C RTC C mC
H= = = ; BL 0dC
dx= , (2.42)
at Lx δ= , for all z: AL A_LC C= , BL B_LC C= , (2.43)
The calculation results showed that: (i) the amine concentration does not vary
significantly in the liquid film and (ii) the CO2 is completely consumed in the liquid film,
and proved, therefore, the maintain of the fast pseudo-first-order regime on the entire area
of the experimental conditions. Figures 2.6 and 2.7 represent an example of the radial
112
concentration profiles of AHPD and CO2 in the liquid film, respectively, at 313.15 K and at
a CO2 mole fraction in the gas phase of 0.41.
By setting the “base” amine partial order to 1 and using the kapp data obtained, three
reaction rate parameters, 2k , 2 AM 1/k k k− and 22 H O 1/k k k− (Eq. (2.12)) are determined using a
non-linear regression method for each studied temperature and are listed in Table 2.5. The
parameter -2 1OH/k k k− was neglected because hydroxyl ions contribution was found
negligible. The average absolute deviation for the calculation of kapp is 1.3%. The
temperature dependence of the reaction rate constants was obtained using Arrhenius type
equations (T is the absolute temperature):
3 -1 -1 112
-6465/(m kmol s ) = 5.08 10 exp ,kT
×
(2.44)
6 -2 -1 62 AM
-1
-3124/(m kmol s ) = 8.88 10 exp ,k kk T
×
(2.45)
22 H O 6 -2 -1 5
-1
-3315/(m kmol s ) = 1.20 10 exp .k k
k T ×
(2.46)
The calculated activation energy for 2k is 53.7 kJ mol-1. This value is comparable to other
sterically hindered alkanolamines 2k activation energy found with the zwitterion
mechanism such as 38.3 kJ mol-1 and 65 kJ mol-1 for AMPD and AEPD, respectively
(Yoon et al., 2003; Yoon et al., 2002).
Table 2.5. Reaction rate parameters for CO2 absorption in aqueous AHPD solutions
T 2k 2 AM 1/k k k− 22 H O 1/k k k−
(K) (m3 kmol-1 s-1) (m6 kmol-2 s-1) (m6 kmol-2 s-1) 303.15 285 302 2.14 313.15 524 398 3.00 323.15 1067 572 4.21
113
a)
0.5
0.6
0.7
0.8
0.9
1.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0x/δ L
CBL
/CB_
L
b)
0.5
0.6
0.7
0.8
0.9
1.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0x/δ L
CBL
/CB_
L
Figure 2.6. Concentration profile of amine in the liquid film: a) exit of the liquid; b) entry
of the liquid. Conditions: T = 313.15 K, 2COy = 0.41.
114
b)
0.00.1
0.20.3
0.40.5
0.60.7
0.80.9
1.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0x/δ L
CA
L/m
C A_G
a)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0x/δ L
CA
L/m
C A_G
Figure 2.7. Concentration profile of dissolved CO2 in the liquid film: a) exit of the liquid; b) entry of the liquid. Conditions: T = 313.15 K,
2COy = 0.41.
2.4.4. Hindrance effect on the SHA properties
The number of studies about sterically hindered alkanolamines quickly increased
since the 80’s especially because of the popularity of AMP, a hindrance form of MEA.
115
Nevertheless, very few studies have been made on AMP’s derivative like AMPD, AHPD
and AEPD. These alkanolamines have a higher steric factor than AMP, as shown in Table
1.1. The addition of a hydroxyl and/or a methyl group leads the molecule structure more
congested. With the increase of the hindrance effect, the probability for a stable carbamate
formation will decrease, by increasing the probability of bicarbonate formation by
enhanced hydrolysis (Chakraborty et al., 1986). Hence, the increase of the hindrance effect
should determine (i) a decrease of the reaction rate value (as it will be discussed later in this
section, based on the existent data) and (ii) an increase of the regeneration facility of the
depleted solutions.
To help us differentiate the hindrance factor of each studied amines, which is
defined as the size of the groups attached to the alpha carbon, we used a method of
estimation of the molecular volumes, as applied in the calculation of the amines diffusion
coefficient in solutions (Othmer and Thakar, 1953; Scheibel, 1954). We found that the
ascending order of the amines bulkiness, β (given at the normal amine boiling point and
expressed in m3kmol-1) is the following: AMP (0.1067), AMPD (0.1252), AHPD (0.1326),
and AEPD (0.1474).
Based on the actual experimental data and those reported in Yoon et al. (2002)
(AEPD), Yoon et al. (2003) (AMPD) and Xu et al. (1996) (AMP), the pseudo-first-order
overall rate constant kov data for AHPD, AEPD, AMPD and AMP as a function of the
amine concentration, at a constant temperature of 303.15 K, are shown in Figure 2.8. It’s
important to mention that because of some discrepancies that exist between kinetic data for
AEPD (in the paper of Yoon et al. (2002), kov values given in Table 1 are not similar to
those given in Tables 2-4 at the same temperature and amine concentration), we chose the
data from Table 1 (from Yoon et al. (2002)) because they are replicated and they could then
be considered more reliable. In Figure 2.8, the slope of the drawn trend lines is an indicator
of the reactivity of theses amines: a higher slope value indicates a higher reaction rate. It is
then possible to see that the obtained slopes values vary in the following ascending order
AEPD, AHPD, AMPD, and AMP, which is the opposite of the amines bulkiness (steric
hindrance) order shown above. This seems to confirm the previously stated assumption that
116
a reduced steric hindrance leads to a more pronounced reaction rate constant (more
reactivity).
Figure 2.8. Variation of kov with the amine concentration for AHPD, AEPD, AMPD and AMP at 303.15 K.
In addition to steric hindrance, other factors seem to influence the amine reactivity
with carbon dioxide in solution (Sartori and Savage, 1983), like the amine category
(primary or secondary) and basicity (related to the pKa). Between AMP, AMPD, AHPD
and AEPD, the effect of amine category can be neglected as they are all primary amines.
Hence, the two main factors influencing the global reaction rate would be the steric
hindrance and the pKa of the amines. To illustrate this, a Brønsted type plot (Derks et al.,
2006) modified to take into account the amines bulkiness, β, is shown in Figure 2.9 and it is
possible to observe almost a perfect linear relation between these four alkanolamines.
Moreover, it seems that the amine reactivity increases with the increase of the function
pKa/β. It would be interesting to study the behaviour of different other primary sterically
hindered alkanolamines to see if the tendency is respected.
0
200
400
600
800
1000
1200
1400
1600
1800
0 0.5 1 1.5 2 2.5 3 3.5
k ov
/(s-1
)
CB /(kmol m-3)
AEPD
AHPD
AMPD
AMP
117
Figure 2.9. Modified Brønsted plot for AHPD, AEPD, AMPD and AMP at 303.15 K.
In order to determine the most effective absorbents for the CO2 absorption in
membrane contactors, other parameters will be essential to investigate for these amines.
Among them, the cyclic absorption capacity, the effect of an activator, the surface tension
and thermal degradation resistance may be the most important ones. This work is currently
in progress in our laboratory.
2.5. Conclusion In this work, the kinetics of the reaction between CO2 and AHPD has been
investigated at different temperatures and solution concentrations. A wetted wall column
apparatus was used in this study and all experimental conditions were selected to be in the
fast pseudo-first-order regime. The zwitterion mechanism was found to fit the experimental
data very well. Based on this mechanism, the reaction rate parameters were calculated with
a non-linear regression from the apparent reaction rate constant. The activation energy for
2k in the CO2-AHPD reaction is found to be 53.7 kJ mol-1.
AHPD
AMP
AEPD
AMPD
10
100
1000
50 60 70 80 90 100
k ov
/ (s-1
)
(pKa/β) / (kmol m-3)
CB = 1 kmol m-3
118
In the previous chapter, we concluded that the kinetics of AHPD is quite low, due to the
important hindrance effect. For improving it, blended solutions can be used. Piperazine, an
amine presenting higher absorption rate than MEA was chosen. The kinetics of the reaction
between CO2 and piperazine-activated aqueous solutions of AHPD was therefore
performed experimentally using a wetted wall contactor.
119
Chapter 3. Acceleration of the reaction of carbon dioxide into aqueous 2-amino-2-hydroxymethyl-1,3-propanediol solutions by piperazine addition
Résumé
Dans ce travail, la cinétique de la réaction entre le CO2 et des solutions à base d’une amine à encombrement stérique, le 2-amino-2-hydroxyméthyle-1,3-propanediol (AHPD), activées par la pipérazine (Pz), a été étudiée dans une colonne à parois mouillée à 303.15, 313.15 et 323.15 K. La concentration d’AHPD a été maintenue constante à 1 kmol m-3 et la concentration en Pz a été variée dans le domaine 0.1 - 0.4 kmol m-3. Les constantes globales de vitesse et les paramètres cinétiques ont été déterminés en considérant le pseudo-ordre 1 pour l’absorption du CO2. Le rapport entre le coefficient de diffusion et la constante d’Henry pour le CO2 dans les solutions d’amine ont été estimés par l’analogie avec le N2O en utilisant les données d’absorption physique de CO2 et N2O dans l’eau et du N2O dans les solutions d’amines. Les résultats ont démontré l’efficacité de la pipérazine comme activateur pour l’AHPD. Pour toutes les températures étudiées, l’addition de petites quantités de Pz a un effet significatif sur la cinétique de l’absorption du CO2.
120
Abstract
In this work, the kinetics of the reaction between CO2 and piperazine-activated aqueous solutions of a sterically hindered alkanolamine, 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD) was studied in a wetted wall column contactor at 303.15, 313.15 and 323.15 K. The AHPD concentration in the aqueous solutions was kept at 1 kmol m-3 while the piperazine (PZ) concentration varied in the range 0.1 - 0.4 kmol m-3. Under pseudo-first-order CO2 absorption conditions, the overall pseudo-first-order rate constants were determined and reaction rate parameters were calculated with a non-linear regression from the overall reaction rate constant. The ratio of the diffusivity and Henry’s law constant for CO2 in solutions was estimated by applying the N2O analogy using the physical absorption data of CO2 and N2O in water and of N2O in amine solutions. Piperazine was found to be an effective activator in the aqueous AHPD solutions, as the addition of small amounts of PZ to these solutions has a significant effect on the enhancement of the CO2 absorption rate for all studied temperatures.
121
3.1. Introduction The global climate change, where carbon dioxide (CO2) is found to be a major
contributor with the increasing industrial development, is one of the most important and
challenging environmental issues facing the world community. A variety of industrial
processes emit large amounts of CO2 from each plant, for example oil refineries, cement
works, and iron production (IPCC, 2005). The Canadian Environmental Protection Act
(CEPA, 2005) is the legislative authority in Canada that pushes Canadian companies to
reduce their greenhouse gas production. These emissions could be reduced substantially by
capturing and storing the CO2.
For technical and economical concerns, the majority of the actual industrial
absorption processes use aqueous solutions of alkanolamines. A wide variety of solvents
can be used such as solutions of monoethanolamine (MEA), diethanolamine (DEA),
diisopropanolamine (DIPA), N-methyldiethanolamine (MDEA) and 2-amino-2-methyl-1-
propanol (AMP) (Kohl and Nielsen, 1997). The use of blended alkanolamines solutions has
also recently become very attractive because of the combination of each amine advantages:
a fast reactivity from a primary or secondary alkanolamine (e.g. MEA, DEA) coupled with
the high absorption capacity and low solvent regeneration cost from a tertiary or sterically
hindered alkanolamine (e.g. MDEA, AMP). Other potential blended solutions can use
piperazine (PZ) as an activator, which is not an alkanolamine but has proven to have a
higher absorption rate than MEA. Some studies of the reaction of CO2 with PZ-activated
solutions have been performed in the literature: PZ/MDEA (Xu et al., 1992a; Zhang et al.,
2001), PZ/ N,N-diethylethanolamine (DEEA) (Vaidya and Kenig, 2008), PZ/AMP (Seo and
Hong, 2000; Sun et al., 2005), PZ/triethanolamine (TEA) (Yeon et al., 2004). In these
studies, reaction rate constants of CO2 with PZ were either determined from the absorption
experiments or taken from other sources and used to obtain other amine or gas-liquid
contactor properties. In both cases, very accurate values of the kinetic parameters are
necessary in order to get reliable results. However, quite different values of the second
order rate constant, k2,PZ, were found in the above mentioned studies concerning the CO2-
piperazine based mixed solvents systems, as well as in the studies concerning the
122
absorption of CO2 in pure piperazine solutions (Bishnoi and Rochelle, 2000; Derks et al.,
2006; Samanta and Bandyopadhyay, 2007; Sun et al., 2005).
In our laboratory, extensive studies of CO2 capture in membrane contactors using
sterically hindered alkanolamines (SHA) based alkanolamine mixtures are in progress. In
this context, in order to study the hindrance effect on the absorption capacity and kinetics
of SHA, a set of four SHA was chosen (Table 1.1, Chapter 1). It concerns AMP, a simple
hindrance form of MEA, and three SHA derived from AMP: 2-amino-2-methyl-1,3-
propanediol (AMPD), 2-amino-2-ethyl-1,3-propanediol (AEPD) and 2-amino-2-
hydroxymethyl-1,3-propanediol (AHPD). Few kinetic studies involving single-amine
aqueous solutions of these four SHA with CO2 are available in the literature (Bougie and
Iliuta, 2009; Yih and Shen, 1988; Yoon et al., 2003; Yoon et al., 2002a). However, except
for AMP, no studies are available in the open literature concerning the characterization of
blended solutions of other potential SHA (like AEPD, AMPD, AHPD) with an activator,
like PZ that was chosen in this work.
The global aim of this work is to study the kinetics of the reaction between CO2 and
piperazine-activated aqueous solutions of AHPD in a wetted wall column contactor at
303.15, 313.15 and 323.15 K. The AHPD concentration in the aqueous solutions was kept
at 1 kmol m-3 while the piperazine (PZ) concentration varied in the range 0.1 - 0.4 kmol m-
3. The work concerns particularly i) the determination of the second order rate constants of
CO2 with PZ from the absorption data of CO2 in blended amine solutions containing AHPD
and ii) the investigation of the enhancement effect of PZ addition on the absorption rate of
CO2 into aqueous AHPD solutions. In order to interpret the experimental data, a second
order reaction for CO2 with PZ and a zwitterion reaction mechanism for CO2 with AHPD
were used. The ratio between the diffusion coefficient and Henry’s law constant, given by
the function 1/ 2A A/D H (Danckwerts, 1970), was estimated by applying the N2O analogy and
the Higbie penetration theory, using the physical absorption data of CO2 and N2O in water
and of N2O in amine solutions. New physicochemical property data (density, viscosity) of
the mixed solvent, needed for calculations related to the wetted wall column, were also
obtained in this work.
123
3.2. Theory 3.2.1. Physical absorption
Physicochemical properties of CO2 in aqueous amine solutions such as the diffusion
coefficient and Henry’s law constant cannot be found directly as CO2 react in solutions.
Hence, the N2O analogy is a useful method widely used in similar works (Dang and
Rochelle, 2003; Yih and Shen, 1988; Yoon et al., 2002a). For the analogy to apply, the
parameters characterising the physical absorption of CO2 and N2O in water and of N2O in
amine solutions need to be known.
With initial gas-free liquids and for short contact time between the gas j and the
liquids in the wetted wall contactor, the Higbie (1935) penetration theory is commonly used
(Alvarez-Fuster et al., 1980; Danckwerts, 1970) and gives the specific absorption rate as: 1/2
, i 2 .π
j jj
c j
D PN
t H
= (3.1)
The contact time (tc) can be derived from the wetted wall column hydrodynamics (Roberts
and Danckwerts, 1962): 1/32 /32 3 .
3ch dt
L gπ µ
ρ =
(3.2)
The combination of Eqs. (3.1) and (3.2) gives: 1/61/2 1/31/2
, i
2π π 3 .2 3 g
j j
j j
D N h dH P L
µρ
= (3.3)
From this last equation, the ratio of the diffusivity and Henry’s law constant can be
calculated by the specific gas absorption rate at several flow rates, L, and for different
heights of effective wetted surface, h, (Nysing and Kramers, 1958) for a constant
temperature and liquid concentration. Here, Nj can be calculated from the total absorption
rate divided by the effective absorption area:
,π
Tj
js
NN
d h=
(3.4)
and ds is the diameter of the wetted wall column including the thickness of the laminar film:
124
1/33 d .π gds
Ld µρ
= +
(3.5)
3.2.2. Chemical absorption
The kinetics of primary and secondary alkanolamines with CO2 can be described
using the zwitterion mechanism proposed first by (Caplow, 1968) and reintroduced later by
Danckwerts (1979). This mechanism has been used successfully with conventional and
sterically hindered alkanolamines such as DEA, DIPA, AEPD and AMPD (Blauwhoff et
al., 1984; Yoon et al., 2003; Yoon et al., 2002a). The first step of this mechanism is the
formation of a zwitterion
2 + -2 2 2
-1CO + RNH RNH COO ,
k
k→←
(3.6)
which can then be deprotonated by bases existing in solution: + - - +b2RNH COO + Base RNHCOO + BaseH .k→ (3.7)
Assuming a quasi-steady-state condition for the zwitterion concentration and an
irreversible deprotonation step by bases, the kinetic rate equation for CO2-RNH2 is given
by:
2 2 2
2CO -RNH A RNH
-1
Base
= , 1 +
b
kr C Ckk C
∑
(3.8)
Eq. (3.8) applies for the reaction between CO2 and AHPD (Bougie and Iliuta, 2009).
For the reaction between CO2 and PZ, a second-order reaction rate is chosen as used
by Bishnoi and Rochelle (2000) and Sun et al. (2005):
2CO -PZ 2,PZ PZ A = . r k C C (3.9)
In an AHPD-PZ-H2O system other reactions can also occur: *H O2 - +
2 2 3CO + H O HCO + H ,k←→ (3.10) *
-OH- -2 3CO + OH HCO ,k
←→ (3.11)
125
*-PZCOO- - +
2 2CO + PZCOO PZ(COO ) + H .k←→ (3.12)
The reaction (3.10) may usually be neglected because it proceeds very slowly: 2
*H Ok
= 0.026 s-1 at 298.15 K (Pinsent et al., 1956). The second reaction (3.11) is the bicarbonate
formation and it was found that it is negligible in AHPD solutions (Bougie and Iliuta,
2009). Therefore, in the presence of an activator, this reaction will become insignificant and
can be ignored. As piperazine is a diamine (Figure 3.1), the reaction (3.12) can occur and it
represents the reaction of CO2 with piperazine carbamate, which contains a free amine
group, to form piperazine dicarbamate. However, this reaction can be neglected in specific
situations where the ratio of the concentration of piperazine carbamate to PZ is low. This
can happen if the experimental conditions are favourable for a fast pseudo-first-order
reaction regime when the amine concentrations remain almost constant in the liquid film
and the products concentration is relatively low. If this regime is correctly set (as it will be
verified later), the kinetic rate equation for the absorption of CO2 in the mixed amine
solutions is given by
2CO -Amines ov A = ,r k C (3.13)
where the overall pseudo-first-order reaction rate constant, kov, can be expressed as
2 AHPDov 2,PZ PZ
-1
Base
= + . 1 +
b
k Ck k Ckk C∑
(3.14)
In Eq. (3.14), the possible bases can be AHPD, PZ or H2O (with the hypothesis of low
concentrations of OH- and PZCOO-).
Figure 3.1. Structure of PZ
126
3.3. Experimental 3.3.1. Reagents
Aqueous AHPD-PZ solutions were prepared with degassed distilled water, 2-amino-
2-hydroxymethyl-1,3-propanediol with a minimum purity of 99.9 % and piperazine with a
minimum purity of 99%. Tween 80 was used as a surface active agent and was added at
0.04 vol% in solutions to avoid ripple formation. All chemicals (Laboratoire MAT,
Quebec, Canada) were used without further purification. Gases (CO2, N2O and N2) were of
commercial grade with a minimum purity of 99.9 % (Praxair).
3.3.2. Experimental setup and procedure
3.3.2.1 Density and viscosity measurements
Densities of aqueous PZ-AHPD solutions were measured by using a calibrated
pycnometer having a bulb volume of 1×10-5 m3 and a Mettler AE240 balance with a
precision of ±1×10-4 g. Temperature of the pycnometer was within ±0.1 K and was
measured with a precision mercury-filled thermometer. The reproducibility of the measured
density was within ±0.3 kg m−3. The kinematic viscosities of solutions were measured by
means of a Cannon-Fenske routine viscometer, size 25. Measurements were made in a
water bath whose temperature was kept constant within ±0.1 K. Kinematic viscosities were
calculated from the efflux times measured with an electronic stopwatch with a resolution of
0.01 s. The experimental errors were estimated to be within ±2.0%. The dynamic
viscosities were calculated by multiplying the kinematic viscosities with the corresponding
densities of the solutions.
3.3.2.2 Physical absorption and CO2 absorption rate measurements
A wetted wall column was used as contactor for physical N2O absorption and for
CO2 absorption rate measurements in amine solutions. A schematic diagram of the
experimental setup is shown in Figure 2.1 (Chapter 2). The column, made of stainless steel,
has an outside diameter of 1.905×10-2 m and the length of the absorption surface could be
varied between 0.03 and 0.11 m. Aqueous amine solutions were supplied to the column
from 3×10-6 m3 s-1 to 4×10-6 m3 s-1 using a digital gear pump (Cole-Parmer, K-74014-40)
and a digital volumetric flowmeter (Cole-Parmer, K-32718-24) with an accuracy of ± 1%.
127
The gas and the solution were circulating in the contactor countercurrently. Complete
information about the column assembly can be found in our previous work (Bougie and
Iliuta, 2010b).
In a typical experimental run, the apparatus and solutions were first brought to the
desired temperature. All experiments were done at 303.15, 313.15 and 323.15 K and for
solutions concentration of 1 kmol m-3 of AHPD with PZ concentration between 0.1 and 0.4
kmol m-3. The concentration of the amine solutions (prepared gravimetrically) was checked
with HCl solutions and a methyl red-bromocresol green pH indicator mix. For physical
absorption of N2O in amine solutions, pure gas was used and absorption rates were
measured by a bubble flowmeter. As pure nitrous oxide was used, no gas phase resistance
was considered in the calculation. For chemical absorption, CO2 was mixed with nitrogen
to give low CO2 partial pressures and to avoid therefore amine depletion at the interface.
The absorption rate was measured as a function of the inlet gas flow rate and the difference
between the inlet and the outlet CO2 composition in the gas determined by gas
chromatography. The gas chromatograph was equipped with a thermal conductivity
detector and a Carboxen 1010 plot capillary column (30m×0.53mm). A carrier gas flow
rate of 3.08×10-7 m3s-1 was used and the temperatures of the detector and the column were
of 423.15 K and 398.15 K, respectively. The measured flow rate was always corrected for
the vapour pressure of water as a function of temperature and the value of the bulk CO2
partial pressure used for the calculation purpose was taken as the logarithmic mean between
the inlet and the outlet CO2 partial pressure (Dang and Rochelle, 2003). As dilute CO2
mixtures were used, the gas phase resistance was taken into consideration to calculate the
CO2 partial pressure at the gas-liquid interface from the bulk CO2 partial pressure:
2 2
ACO , i CO
g
= - , NP Pk
(3.15)
where the gas mass transfer coefficient was taken as an average from two values calculated
with the correlation of Hobler (1966) and Pacheco et al. (2000) (Eqs. (3.16) and (3.17),
respectively):
H 0.5 Re ,dSh Sch
= (3.16)
128
0.85H = 1.075 Re .dSh Sc
h
(3.17)
The liquid flow rate and the height of the column were selected in such a way that
the chemical absorptions occurred in the fast pseudo-first-order reaction regime for both of
the amines (Iliuta, 2002). In this regime, the Hatta number equals the enhancement factor
when
, inf,2 < ,a j jH E (3.18)
where the Hatta number for an amine j, Ha,j, is defined as
A 2 AHPD
-1
Base,AHPD
L
1 + = ,b
a
D k Ckk C
Hk∑ (3.19)
A 2,PZ PZ,PZ
L
= ,a
D k CH
k (3.20)
A ov
L
= ,a
D kH
k (3.21)
and the infinite enhancement factor based on penetration theory, Einf,j, is 11/2
jA Ainf, j
j A j A, i
= .j
CD DE DD H v P
−
+ ⋅ (3.22)
The liquid phase mass-transfer coefficient for physical absorption is calculated with the
definition given by the Higbie penetration theory as:
AL = 2 .
π c
Dkt
(3.23)
In this regime, the specific absorption rate is then 1/2A
A A, i ovA
.DN P kH
=
(3.24)
By experimental data regression, Eq. (3.24) allowed us to determine the kinetic reaction
rate constants present in Eq. (3.14). It must be noted that for physical and chemical
absorptions the liquid flow on the wetted wall column was always laminar. The highest
129
Reynolds number obtained was around 90, which is much less than the criterion of 250
proposed by Danckwerts (1970).
3.4. Results and discussion 3.4.1. Physicochemical properties of solutions
The measured density and viscosity of aqueous PZ-AHPD solutions are presented in
Table 3.1. We observed as expected that these two properties increase when PZ
concentration increases for a constant temperature and at constant concentration measured
values decrease with temperature increase. The “Stepwise” regression method
(Montgomery and Runger, 1999) was used to correlate the data to a general equation:
( )-3 2
2PZ-1 -1
i = 0
/(kg m ) = ( + + ) ,
/ (kg m s )i
i i ia b T c T Cρ
µ
⋅ ⋅
∑ (3.25)
where ai, bi and ci are the regressed coefficients presented in Table 3.2, T is the absolute
temperature and the PZ concentration is expressed in kmol m-3. As the “Stepwise”
regression method was used, only the statistically significant coefficients are therefore
found in the regression, the others equal 0. Eq. (3.25) agrees to our experimental data
within an average relative deviation of 0.01% and 0.68% for densities and viscosities
respectively.
Table 3.1. Densities and viscosities of PZ-AHPD solutions
T (K) kmol m-3 PZ + kmol m-3 AHPD
Density ρ (kg m-3)
Viscosity µ × 103 (kg m-1 s-1)
303.15 0.1 + 1.0 1026.0 1.150 0.2 + 1.0 1026.4 1.198 0.3 + 1.0 1026.8 1.245 0.4 + 1.0 1027.1 1.298
313.15 0.1 + 1.0 1022.4 0.931 0.2 + 1.0 1022.8 0.964 0.3 + 1.0 1023.1 0.996 0.4 + 1.0 1023.5 1.035
323.15 0.1 + 1.0 1018.7 0.762 0.2 + 1.0 1019.1 0.793 0.3 + 1.0 1019.5 0.815 0.4 + 1.0 1019.8 0.845
130
Table 3.2. Regressed coefficients for density, viscosity and ( )2 2
1/ 2N O N O Amines
/D H
correlations
Density ai bi ci
i = 0 1.0788 × 103 0 -5.7829 × 10-4
i = 1 3.6094 0 0 i = 2 0 0 0
Viscosity ai bi ci
i = 0 4.2851 × 10-2 -2.4966 × 10-4 3.6929 × 10-7
i = 1 2.3546 × 10-3 0 -2.0111 × 10-8
i = 2 0 0 0 ( )2 2
1/ 2N O N O Amines
/D H
ai bi ci
i = 0 2.5560 × 10-8 0 -1.8900 × 10-13
i = 1 0 -1.9443 × 10-11 0 i = 2 5.5354 × 10-9 0 0
3.4.2. Physical absorption
To use the N2O analogy, the absorption of N2O in aqueous solutions of 1 kmol m-3
of AHPD with PZ concentration between 0.1 and 0.4 kmol m-3 was performed at 303.15,
313.15 and 323.15 K. Fig. 3.2 shows the calculated values of the ratio 2 2
1/ 2N O N O/D H along
with the curves obtained by a correlation of the same type as for densities and viscosities:
( )2
2
1/ 2 2N O -1 -2 -1/2 2
Bi = 0N O Amines
/(kmol kPa m s ) = ( + + ) .ii i i
Da b T c T C
H
⋅ ⋅
∑ (3.26)
The regressed coefficients for Eq. (3.26) are found in Table 3.2. This last equation agrees to
our experimental data within a mean relative deviation of 2.1%. In Fig. 3.4, 2 2
1/ 2N O N O/D H
values for the system N2O-AHPD-H2O were taken from our last work (Bougie and Iliuta,
2009). Experimental data show a decrease in the ratio value with an increase in amine
concentration or an increase in temperature. This trend is in agreement with the data of
131
0
1
2
3
4
5
6
7
8
9
0 0.1 0.2 0.3 0.4CPz / (kmol m-3)
CAHPD = 1 kmol m-3
, 303.15 K, 313.15 K, 323.15 K, Eq. (3.26), 303.15 K, 1 kmol/m³ AMP, Sun et al. (2005)
Bougie and Iliuta (2009), Yih and Shen (1988) and Yoon et al. (2002a). Data for the
aqueous system PZ-AMP at 303.15 K of Sun et al. (2005) were also added in Fig. 3.2 to
observe that the PZ addition causes a similar rate of decrease of the ratio2 2
1/ 2N O N O/D H
regardless of the sterically hindered alkanolamine used in solutions.
Figure 3.2. 2 2
1/ 2N O N O/D H ratio for N2O absorption in aqueous PZ-AHPD solutions.
The ratio 2 2
1/ 2CO CO/D H in amine solutions was calculated using the N2O analogy and
is shown in Table 3.3:
2 2
2 2 2 2 2
2 2 2
1/ 2N O N O AHPD1/ 2 1/ 2
CO CO AHPD CO CO H O1/ 2N O N O H O
( / )( / ) = ( / ) ,
( / )D H
D H D HD H
⋅ (3.27)
where the diffusivity and the Henry’s law constant of CO2 and N2O in water are given by
these regressed equations (Versteeg and Vanswaaij, 1988):
( )22
2 -1 -6CO H O
-2119/(m s ) = 2.35 10 exp ,DT
×
(3.28)
( )22
2 -1 -6N O H O
-2371/ (m s ) = 5.07 10 exp ,DT
×
(3.29)
132
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
0 0.1 0.2 0.3 0.4 0.5
NA
x10
6/ (
kmol
m-2
s-1)
CPZ / (kmol m-3)
CAHPD = 1 kmol m-3
, 303.15 K, 313.15 K, 323.15 K
( )22
3 -1 6CO H O
-2040/(kPa m kmol ) = 2.825 10 exp ,HT
×
(3.30)
( )22
3 -1 6N O H O
-2284/(kPa m kmol ) = 8.547 10 exp .HT
×
(3.31)
3.4.3. Chemical absorption
3.4.3.1 Data analysis and kinetic reaction rate constants
Following the procedure and equations described in the sections 3.2.2 and 3.3.2.2,
the chemical absorption of CO2 in PZ-AHPD solutions was studied in order to determine
first the kinetic reaction rate constants. The experimental results are presented in Table 3.3.
Fig. 3.3 presents the specific absorption rates of CO2 in the amine solutions at different
temperatures and for a CO2 partial pressure of around 2 kPa. It can be seen that the trends
are in agreement with other CO2-alkanolamine systems studied in the literature (Gianetto et
al., 1986).
Figure 3.3. Specific absorption rate as a function of amines concentrations for
2COy = 0.02.
133
Table 3.3. Kinetic data for absorption of CO2 in PZ-AHPD aqueous solutions
T kmol m-3 PZ ( )2 2
1/ 2 9CO CO Amines
/ 10D H × PCO2,i NCO2 × 106 kL × 104 tc kov
(K) + kmol m-3 AHPD (kmol kPa-1m-2s-1/2) (kPa) (kmol m-2 s-1) (m s-1) (s) (s-1) 303.15 0.1 + 1.0 10.83 2.23 1.916 1.23 0.128 6275
0.2 + 1.0 10.23 1.81 2.188 1.13 0.135 14039 0.3 + 1.0 9.79 1.64 2.302 1.08 0.137 20467 0.4 + 1.0 9.50 1.67 2.550 1.07 0.133 25931
313.15 0.1 + 1.0 9.26 2.23 1.754 1.33 0.120 7228 0.2 + 1.0 8.63 1.81 2.146 1.26 0.118 18940 0.3 + 1.0 8.16 1.59 2.240 1.15 0.126 29943 0.4 + 1.0 7.84 1.54 2.398 1.11 0.127 39210
323.15 0.1 + 1.0 7.60 2.11 1.663 1.41 0.112 10769 0.2 + 1.0 6.93 1.73 1.915 1.30 0.111 27294
0.3 + 1.0 6.42 1.47 1.979 1.17 0.121 44075 0.4 + 1.0 6.07 1.54 2.230 1.15 0.114 56552
134
Table 3.4. Parameters for pseudo-first order regime verification of PZ-AHPD-H2O systems
T kmol m-3 PZ DAHPD ×109 DPZ ×109 Ha,AHPD Ha,PZ Einf,AHPD Einf,PZ Einf,PZ/(Ha,PZ) (K) + kmol m-3 AHPD (m2 s-1) (m2 s-1)
303.15 0.1 + 1.0 0.77 0.89 4.45 26.13 1152 63.0 2.4 0.2 + 1.0 0.75 0.86 4.70 37.69 1442 155.7 4.1 0.3 + 1.0 0.73 0.84 4.87 46.46 1681 271.3 5.8 0.4 + 1.0 0.71 0.82 4.90 52.92 1689 363.1 6.9
313.15 0.1 + 1.0 0.97 1.11 5.60 30.30 1511 82.1 2.7 0.2 + 1.0 0.95 1.09 5.91 42.75 1978 213.0 5.0 0.3 + 1.0 0.93 1.07 6.34 54.26 2356 379.7 7.0 0.4 + 1.0 0.91 1.04 6.52 62.92 2488 534.3 8.5
323.15 0.1 + 1.0 1.19 1.36 7.00 35.19 2152 116.4 3.3 0.2 + 1.0 1.17 1.34 7.48 49.70 2846 306.0 6.2
0.3 + 1.0 1.15 1.32 8.17 63.71 3593 578.5 9.1 0.4 + 1.0 1.12 1.29 8.15 71.29 3576 767.4 10.8
135
By using the kov data indicated in Table 3.3 (except for those at 0.1 kmol m-3 of PZ,
as explained in the section 3.4.3.2), the second-order rate constant for the reaction of CO2
with PZ, 2, Pzk , and the PZ group of constant implicated in the deprotonation of the
zwitterion, 2 PZ 1/k k k− , are determined from Eq. (3.14) using a non-linear regression method
for each studied temperature. The other kinetic rate constant for AHPD were taken from
Bougie and Iliuta (2009). The obtained values for 2, Pzk are 66 450, 97 984 and 141 613 m3
kmol-1 s-1 at 303.15 K, 313.15 K and 323.15 K, respectively. The temperature dependence
of these reaction rate constants follows an Arrhenius type equation with an R2 of 0.999:
3 -1 -1 102,PZ
-3706/(m kmol s ) = 1.353 10 exp .kT
×
(3.32)
Figure 3.4. shows a comparison of the results of this study with literature values
(Bishnoi and Rochelle, 2000; Seo and Hong, 2000; Sun et al., 2005; Zhang et al., 2001). It
can be seen that our results are in good agreement with those reported in other works
involving PZ (Bishnoi and Rochelle, 2000; Zhang et al., 2001). However, Seo and Hong
(2000) and Sun et al. (2005) obtained lower values. The reason of these discrepancies may
be that the authors did not respect the pseudo-first-order reaction regime and PZ depletion
occurred at gas-liquid interface. This can happen if very low PZ concentrations are used
(Seo and Hong, 2000) and if the ratio of the Hatta number to the infinite enhancement
factor is not high enough (Sun et al., 2005). Similar comments on Sun et al. (2005) work’s
were made by Derks et al. (2006).
Based on a parameter sensitivity analysis, the parameter 2 PZ 1/k k k− obtained in for
this system was found to be not statistically significant and therefore no realistic values
were established for it from the non-linear regression. Some authors (Derks et al., 2006;
Samanta and Bandyopadhyay, 2007) suggest that the zwitterion mechanism applies to PZ.
However, as PZ has very high second-order rate constant and pKa (9.68 at 303.15 K;
(Pagano et al., 1961)), the deprotonation of the zwitterion would be very fast and the term
in parenthesis of Eq. (3.8) would tend to the k2,PZ value, which would lead to a second-order
reaction as set in this study. From this point of view, it is normal that we cannot find
consistent values for the parameter k2kPZ/k-1. Regressed values for this parameter were
136
however given in several works (e.g. Sun et al. (2005), but the corresponding Arrhenius
equation (Eq. (23) given in Sun et al. (2005) make no sense because the calculated value
are extremely low; this is in contradiction to the expected high values for a very fast
deprotonation process.
Figure 3.4. Arrhenius plot of the second-order rate constant k2,PZ as a function of temperature.
3.4.3.2 Fast pseudo-first-order regime verification
As discussed previously in sections 3.2.2 and 3.4.3.1, the fast pseudo-first order
regime is important to set correctly in order to get reliable results. Therefore, a verification
of Eq. (3.18) is necessary. The Hatta number, Ha,j and the infinite enhancement factor, Einf,j
are calculated from Eqs. (3.19)-(3.22) and are given in Table 3.4. The diffusion coefficient
of CO2 in aqueous PZ-AHPD solutions, ( )2CO AminesD was estimated using the
( )2 2
1/ 2CO CO Amines
/D H ratio obtained in this work and the values of ( )2CO AHPDH taken from Paul
et al. (2009b). The diffusion coefficients of AHPD and of PZ in aqueous PZ-AHPD
137
0
20000
40000
60000
80000
0 0.1 0.2 0.3 0.4 0.5
k ov
/ (s-1
)
CPZ / (kmol m-3)
CAHPD = 1 kmol m-3
, 303.15 K, 313.15 K, 323.15 K, calculated by Eq. (3.14)
solutions, Dj, were calculated using a modified Stokes-Einstein relation (Versteeg and
Vanswaaij, 1988): -0.6
j w w = ( / ) ,D D µ µ⋅ (3.33)
where the subscript « w » refers to infinite dilution state (pure water). Dw was obtained
using an equation based on molecular volumes of the solutes and the solvent, as described
elsewhere (Othmer and Thakar, 1953). This correlation was chosen because it was the only
correlation able to predict the PZ diffusion with an acceptable error over the complete
temperature range (Derks et al., 2008).
Figure 3.5. The overall pseudo-first-order rate constant as a function of PZ concentration.
In Table 3.4, it is possible to observe that for AHPD, the Hatta numbers are all
greater than 2 and considerably lower than the corresponding infinite enhancement factor.
For PZ, the Hatta numbers are all greater than 2 and are lower than the corresponding
infinite enhancement factor. However, for a concentration of 0.1 kmol m-3 PZ, we note that
the ratio between Einf,PZ and Ha,PZ is below 4 at all temperatures, which seems to represent
the minimum conditions to ensure the pseudo-first order regime for this system. As it can
138
be seen in Figure 3.5, this low ratio between Einf,PZ and Ha,PZ for PZ concentration of 0.1
kmol m-3 could explain the larger deviation obtained by the model (Eq. 3.14) compared to
experimental data. An average relative deviation of 9.2% is obtained when all results are
taken into consideration and 2.8% is obtained when values at 0.1 kmol m-3 PZ are
neglected.
The fast pseudo-first-order regime was also verified on the entire length of the
column, by analysing the gas and amine concentration profiles in the liquid film. They were
obtained by solving the equations set (3.34)-(3.43) based on the following assumptions: (i)
steady-state and isothermal conditions; (ii) plug flow for both liquid and gas,
L
A_L ALA 0 ,L V
x
dC dCu D adz dx δ=
+ = (3.34)
L
B_L BL 0 ,L B Vx
dC dCu D adz dx δ=
+ = (3.35)
L
D_L DL 0 ,L D Vx
dC dCu D adz dx δ=
+ = (3.36)
A_G ALG
0
0 ,A Vx
dC dCu D adz dx =
+ = (3.37)
2AL
A A2 0 , CD rx
∂− =
∂ (3.38)
2BL
B B A2 0 ,CD rx
ν∂− =
∂ (3.39)
2DL
A2 0 ,D DCD rx
ν∂− =
∂ (3.40)
with the following boundary conditions in the axial and radial directions (counter-current
flow):
at 0z = : AG,0A_L
A
C RTC
H= , B_L B,exitC C= , D_L D,exitC C= , A
A_G A_G,0PC CRT
= = , (3.41)
at 0x = , for all z: A_GAL A,i A_G
A
C RTC C mC
H= = = ; BL 0dC
dx= , DL 0dC
dx= , (3.42)
at Lx δ= , for all z: AL A_LC C= , BL B_LC C= , DL D_LC C= , (3.43)
139
The calculation results showed that, except for the lowest PZ concentration (0.1
kmol m-3): (i) the amine concentration does not vary significantly in the liquid film and (ii)
the CO2 is completely consumed in the liquid film. That therefore proves the maintain of
the fast pseudo-first-order regime on the considered range of experimental conditions taken
into account for data regression, as previously discussed.
3.4.3.3 Enhancement effect of PZ additions in SHA solutions
To analyse the enhancement effect of PZ in PZ-AHPD aqueous solutions, ratio
between kov indicated in Table 3.3 of this study and kov obtained in AHPD solutions only:
172.8 s-1, 279 s-1 and 470.5 s-1 at 303.15 K, 313.15 K and 323.15 K respectively (Bougie
and Iliuta, 2009), are calculated and shown in Fig. 3.6. We can see that the addition of
small concentration of PZ to AHPD aqueous solutions improve considerably the absorption
of CO2. An increase in PZ concentration increases the enhancement effect as expected,
while increasing temperature decreases this enhancement effect. The reason of this latter
tendency is because the activation energy of the second-order rate constant of AHPD, k2,
(53.7 kJ mol-1; (Bougie and Iliuta, 2009)) is higher than the activation energy of the second-
order rate constant of PZ, k2,PZ (calculated in this study as 30.8 kJ mol-1).
Xu et al. (1996) obtained a kov of 710.1 s-1 for the system AMP (0.977 kmol m-3) +
H2O at 305 K. Sun et al. (2005) calculated a kov of 14 820 s-1 at 303.15 K for the system PZ
(0.4 kmol m-3) + AMP (1 kmol m-3) + H2O. This value represents therefore an enhancement
factor of 20.9 caused by the PZ addition to AMP solutions. Comparatively, as seen in Fig.
3.6, an addition of 0.4 kmol m-3 of PZ to AHPD solutions at 303.15 K results in an
enhancement factor of 150.1. The enhancement effect of PZ addition is then more
pronounced in AHPD solutions than in AMP solutions.
As it was also shown in the literature (Bishnoi and Rochelle, 2000), PZ is an
effective promoter for carbon dioxide removal from gas streams. The fact that the rate
constant of PZ was found to be an order of magnitude higher than primary amines such as
MEA or DGA, justifies the choice of PZ as activator in AHPD solutions used in this work.
The high reactivity of piperazine compared to other amines with similar pKa values can be
due to its cyclic and diamine nature
140
0
20
40
60
80
100
120
140
160
0 0.1 0.2 0.3 0.4
k ov,
PZ-A
HPD
/ kov
,AH
PD
CPZ / (kmol m-3)
CAHPD = 1 kmol m-3
, 303.15 K
, 313.15 K
, 323.15 K
Figure 3.6. Enhancement effect of PZ in 1 kmol m-3 AHPD solutions.
3.4.4. Prospective and future studies
In order to determine the most effective absorbents for the CO2 absorption in
membrane contactors, other parameters will be essential to investigate for the potential
SHA and activated-SHA solutions. Among them, the cyclic absorption capacity, the surface
tension, the liquid-membrane compatibility and the thermal degradation resistance may be
the most important ones. This work is currently in progress in our laboratory.
3.5. Conclusion In this work, the kinetics of the reaction between CO2 and PZ-AHPD aqueous
solutions has been investigated at different temperatures and solution concentrations. A
wetted wall column apparatus was used and the experimental conditions were selected to be
in the fast pseudo-first-order regime. Reaction rate parameters were calculated with a non-
linear regression from the overall reaction rate constant. The second-order rate constant for
the reaction of CO2 with PZ, 2, Pzk , was found to be 66 450 m3 kmol−1 s−1 at 303.15 K with
an activation energy of 30 812 kJ kmol-1. Piperazine, which has a very high reaction rate
constant due to its cyclic diamine structure, was found to be an effective activator in these
solutions as the addition of small amounts of PZ to aqueous AHPD solutions has significant
effect on the enhancement of the CO2 absorption rate.
141
142
Along with good kinetics, the CO2 absorbent needs to present a good absorption capacity.
In the following chapter, the thermodynamics of the aqueous CO2 + AHPD + Pz system
was investigated experimentally using a vapor-liquid equilibrium apparatus based on a
static-synthetic method and data were modelled with a modified Pitzer’s thermodynamic
model for the activity coefficients.
143
Chapter 4. CO2 absorption into mixed aqueous solutions of 2-amino-2-hydroxymethyl-1,3-propanediol and piperazine
Résumé
La solubilité du CO2 dans des mélanges aqueux de 2-amino-2-hydroxyméthyl-1,3-propanediol (AHPD) et pipérazine (Pz) a été mesurée sur une plage de températures de 288.15 à 333.15 K et pour une concentration totale d'amine variant jusqu'à 3.1 kmol.m-3, en utilisant un appareil d’équilibre liquide-vapeur basé sur la méthode statique-synthétique. La pression partielle du CO2 a été variée dans le domaine 0.21 – 2 637 kPa. La solubilité du N2O dans les solutions aqueuses Pz-AHPD a également été mesurée afin de déterminer la constante d’Henry du CO2 dans ces solutions, par l'analogie avec le N2O. Les données expérimentales pour le système ternaire AHPD-CO2-H2O ont été corrélées en utilisant un modèle thermodynamique modifié de Pitzer pour les coefficients d'activité, associé à l'équation du viriel pour les coefficients de fugacité. La solubilité du dioxyde de carbone dans les solutions aqueuses d'amine mixte (Pz + AHPD) a été prédite en considérant que les paramètres des systèmes ternaires sont essentiels pour décrire le comportement du système quaternaire.
144
Abstract
Solubility data of CO2 in aqueous mixtures of 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD) and piperazine (Pz) were measured over a range of temperature from 288.15 to 333.15 K and for total amine concentrations up to 3.1 kmol.m-3. The CO2 partial pressure was kept within 0.21 – 2 637 kPa using a VLE apparatus based on a static-synthetic method. The solubility of N2O in the Pz-AHPD aqueous solutions was also performed in order to determine, with the N2O analogy, the Henry’s law constant of CO2 in these solutions. The experimental data for the ternary system AHPD-CO2-H2O were correlated using a modified Pitzer’s thermodynamic model for the activity coefficients combined with the virial equation of state for representing the fugacity coefficients. The solubility of carbon dioxide in aqueous solutions of mixed amine (Pz+AHPD) was predicted by supposing that the parameters characterising the single amine systems are essential for describing the quaternary system behaviour.
145
4.1. Introduction Since a few decades, removal of CO2 has become one of the most important
environmental issues facing the word community. This has motivated intensive research on
CO2 capture where new and more energy-efficient absorbents are essential. Actual
industrial CO2 absorption processes use aqueous solutions of alkanolamines. For technical,
economical and environmental concerns, this technique is widely applied for (i) acid gases
(CO2, H2S) removal during natural gas sweetening and (ii) CO2 capture from fossil-fuel-
fired power plants, as well as some other important industries such as chemical and
petrochemical, steel, aluminium and cement production.
Industrially more often used alkanolamines are monoethanolamine (MEA),
diethanolamine (DEA), diisopropanolamine (DIPA), N-methyldiethanolamine (MDEA),
and 2-amino-2-methyl-1-propanol (AMP) (Kohl and Nielsen, 1997). The choice of a
certain amine is mainly based on the absorption capacity, reaction kinetics and regenerative
potential and facility. The key advantage of the primary and secondary alkanolamines such
as MEA and DEA is their fast reactivity due to the formation of stable carbamates.
Conversely, this will lead to very high solvent regeneration cost. On the absorption capacity
side, they have the drawback of a relatively low CO2 loading (limited to 0.5 mol CO2/mole
amine). Tertiary alkanolamines, like MDEA, have a low reactivity with respect to CO2, due
to the exclusive formation of bicarbonates by CO2 hydrolysis. However, this will lead to a
very low solvent regeneration cost. Another advantage of these amines is the high CO2
theoretical loading capacity of 1 mol of CO2/mol of amine. The application of sterically
hindered alkanolamines (SHA) e.g., AMP in gas-treating technology offers absorption
capacity, absorption rate, selectivity and degradation resistance advantages over
conventional amines for CO2 removal from gases (Sartori and Savage, 1983). SHA form
unstable carbamates due to the hindrance of the bulky group adjacent to the amino group.
Hydrolysis of the voluminous carbamates leads to a preferential bicarbonate formation
process, resulting in the theoretical loading capacity up to 1.0. Reaction kinetics
significantly higher than those related to tertiary amines, coupled with a low solvent
regeneration cost offer to SHA important industrial advantages. The use of blended
alkanolamines solutions has also recently become very attractive because of the
146
combination of each amine advantages: a fast reactivity from a primary or secondary
alkanolamine coupled with the high absorption capacity and low solvent regeneration cost
from a tertiary or sterically hindered alkanolamine.
In our laboratory, extensive studies of CO2 capture in membrane contactors using
activated (piperazine) aqueous SHA solutions are in progress. A set of four SHA was
chosen (Bougie and Iliuta, 2009; Bougie et al., 2009). It concerns AMP, a simple hindrance
form of MEA, and three SHA derived from AMP: 2-amino-2-methyl-1,3-propanediol
(AMPD), 2-amino-2-ethyl-1,3-propanediol (AEPD) and 2-amino-2-hydroxymethyl-1,3-
propanediol (AHPD). The kinetics of these SHA has been discussed previously (Bougie
and Iliuta, 2009) as well as the influence of the addition of an activator (Pz) in AHPD
solutions (Bougie et al., 2009). In order to have a more accurate insight of the properties of
the studied solutions, data concerning the solubility of CO2 and N2O in aqueous amine
solutions are needed respectively to i) determine the equilibrium loading of CO2 in these
solutions for a wide range of temperature, solutions concentration, CO2 partial pressure and
ii) determine the Henry’s law constant of CO2 in these solutions by the application of the
widely known N2O analogy. Henry’s law constants are particularly useful to calculate the
CO2 diffusion coefficient in solution from values of the ratio 2 2
1/ 2CO CO/D H . This ratio is found
by the use of the wetted wall column contactor as explain in our previous works (Bougie
and Iliuta, 2009; Bougie et al., 2009). The number of studies about CO2 solubility and
Henry’s law constant in aqueous solutions of AMPD, AEPD or AHPD is quite low (Baek
and Yoon, 1998; Baek et al., 2000; Le Tourneux et al., 2008; Park et al., 2003; Park et al.,
2002a; Park et al., 2002b; Paul et al., 2009c) and disagreements were found between the
reported equilibrium solubility of CO2 in AHPD solution between the study of Park et al.
(2003) and Le Tourneux et al. (2008). Furthermore, except for AMP, no study was found
concerning the equilibrium solubility of CO2 and of N2O in Pz-activated aqueous solutions
of these SHA.
The main objective of this work is the experimental characterization and the
thermodynamic modeling of the CO2 solubility in aqueous Pz-activated AHPD solutions.
The solubility measurements were performed in a static vapor-liquid equilibrium apparatus
147
for a large range of temperature, solution concentrations and CO2 partial pressures. A
thermodynamic model based on the Pitzer’s equations for the activity coefficients coupled
with the truncated virial equation of state for representing the non ideality of the vapour
phase was used to correlate the experimental data for the ternary AHPD-CO2-H2O system.
The solubility of carbon dioxide in aqueous solutions of mixed amine (Pz+AHPD) was
predicted by supposing that the parameters characterising the single amines systems are
essential for describing the quaternary system behaviour. The solubility of N2O in the Pz-
AHPD aqueous solutions was also performed in order to determine, with the N2O analogy,
the Henry’s law constant of CO2 in these solutions. At our knowledge, similar data are not
available in the open literature.
4.2. Experimental 4.2.1 Reagents
Aqueous Pz-AHPD solutions were prepared with degassed distilled water, 2-amino-
2-hydroxymethyl-1,3-propanediol and piperazine. The amines (from Laboratoire MAT,
Quebec, Canada) had a minimum purity of 99.9 % and were used without further
purification. CO2 and N2O gases were of commercial grade with a minimum purity of 99.5
% and were supplied by Praxair.
4.2.2 Apparatus and procedures
The experimental setup for the solubility measurements (Armines, France) used in
this work is shown in Figure 4.1. It consists of an equilibrium cell made of TA6V titanium
with an internal volume of about 1.15×10-4 m3. The equilibrium cell is equipped with a
magnetic rod covered with titanium and the cell is located in a modified XU027 laboratory
oven from France Etuves. This oven came with a C3000 temperature controller (by France
Etuves) which allows temperature control of ±0.1 K. A special feature of this apparatus is
the addition in the oven of a coil refrigerated with a thermostated bath (K-12108-10 from
Cole-Palmer). This coil allowed us to made solubility measurement under room
temperature (273.15 to 303.15 K) with the same temperature precision. Pressure in the cell
was measured by means of one or two of the two installed absolute pressure transducers
(Druck PTX-611, 0-100 kPa and 0-16000 kPa) according to the pressure range. Two 100
148
ohms platinum resistance thermometer were used for temperature measurements of the
equilibrium cell. Liquid introduction inside the equilibrium cell was made with a variable
volume press (stainless steel 316, internal diameter of 3.002×10-2 m). This press has been
equipped with a linear encoder (Heidenhain, LS487C) which allowed knowing the exact
longitudinal position of the piston in the press with an accuracy of ± 2×10-6 m. Gas
introduction in the cell was made by a thermostated small gas cylinder with an internal
volume of about 7×10-5 m3. This small gas cylinder was equipped with a Druck PTX-611
0-16000 kPa absolute pressure transducer.
Figure 4.1. Schematic diagram of the solubility apparatus: A, Equilibrium Cell; B, Magnetic Rod; C, Platinum Resistance thermometer; Di, Gears; E, Coil; F, Pressure
Transducer (F1, Low pressure values; F2, High pressure Values); G, Valve; H, Stirrer; I, Temperature controller; J, Computer; K, Circulating bath; L, Variable volume press for
liquid introduction; M, Small gas cylinder; N, Gas cylinder; Oi, Needle valve; Pi, Valves; Q, Laboratory oven.
A standard experimental run consisted of a sequence of successive step. First, the
amines aqueous solution (total amines molalities from 0.91 to 4.36 mol.kg-1) was prepared
to its specific concentration by gravimetric method using a Mettler Toledo AE204 balance
with a precision of ±0.0001 g. Then the solution was degassed under vacuum and the amine
N
0 m carré0 m carré0 m carré
P-11
P-15
EC
K
L
E
J
F2F1
D4
H
M
A
P-12
A
D3D2D1
I
L
O1
O2
O3
P2
P1
Q
F
B
C
G
149
concentration of the resulting solution was checked with HCl and a methyl red-bromocresol
green pH indicator mix to verify the possible change in concentration due to solvent or
solute lost. The degassed solution was then transferred under vacuum inside the variable
volume press and subsequently, with the piston, in the equilibrium cell previously brought
to vacuum. The equilibrium cell was heated to the desired temperature and the solution was
agitated. At this stage, the vapour pressure of the solution was measured by the low
pressure transducer. This was followed by the introduction of the gas to be absorbed (CO2
or N2O) in the equilibrium cell via the small gas cylinder. Introduced gas mole number was
calculated by using the cylinder volume, its temperature as well as the observed pressure
drop in the cylinder after the gas introduction. System equilibrium was reached when the
pressure inside the equilibrium cell was varying less than 0.5% for at least 30 minutes. It
took about two hours after the gas introduction for chemical absorption of CO2 and 30
minutes for physical absorption of N2O. The difference between the introduced and the
remaining gas mole number in the head space of the equilibrium cell was then calculated
which lead to the concentration of absorbed gas in the solution.
4.3. Thermodynamic modeling of the vapour-liquid equilibrium 4.3.1. Chemical equilibrium in the liquid phase
Due to chemical reactions in the liquid phase, carbon dioxide can be found in the
liquid phase in both neutral and non-volatile ionic form. The model applied to
correlate/predict the solubility of carbon dioxide in aqueous solutions of AHPD and
Pz+AHPD considers the following equilibriums for the chemical species in the liquid
phase: the formation and dissociation of bicarbonate (reactions 4.1 and 4.2), the
autoprotolysis of water (reaction 4.3), the protonation of AHPD (reaction 4.4), the
formation of AHPD carbamate (reaction 4.5), the protonation and diprotonation of
piperazine (reactions 4.6 and 4.7), and the formation of piperazine carbamate, piperazine
dicarbamate and protonated piperazine carbamate (reactions 4.8-4.10). In the reactions 4.4
and 4.5, “R” denotes the (HO-CH2)2-C group in AHPD.
2 2 3CO + H O HCO +H ,IK − +
(4.1) 2 +
3 3HCO CO + H ,IIK− −
(4.2)
150
2H O H + HO ,IIIK + −
(4.3)
2 3RNH + H RNHIVK+ +
(4.4)
2 3 2RNH + HCO RNHCOO + H O,VK− −
(4.5) + +Pz + H PzH ,VIK
(4.6) + + 2
2PzH + H PzH ,VIIK +
(4.7) -
3 2Pz + HCO PzCOO + H OVIIIK−
(4.8)
3 2 2PzCOO + HCO Pz(COO ) H O ,IXK− − − +
(4.9)
PzCOO + H PzH COO ,XK− + + −
(4.10) The condition for chemical equilibrium for a chemical reaction R is:
( ) ( ), 1,...,10i RR i
i
K T a Rν= =∏ (4.11)
where ai is the activity of species i.
In the addition of the above equilibrium equations, overall species mole and charge
balance must be satisfied. In the balance equations for carbon dioxide, AHPD and Pz in the
liquid phase (Eqs. 4.12-4.14) AHPDm and Pzm denote the stoichiometric molalities of AHPD
and Pz, respectively and ∝ denotes the CO2 loading in the solutions expressed as total
moles of CO2 absorbed both chemically and physically per mole of amine.
2 3RNH RNHCOO RNHAHPDm m m m− += + + (4.12)
( )+ 22
2Pz PzH PzH PzCOO PzH COOPz COO
Pzm m m m m m m+ − + −−= + + + + + (4.13)
( )22 3 32
CO HCO CO RNHCOO PzCOO PzH COOPz COO( ) 2AHPD Pzm m m m m m m m mα − − − − + −−+ = + + + + + + (4.14)
( )
23 2
23 3
2
H RNH PzH PzH
OH HCO CO RNHCOO PzCOO Pz COO
2
2 2
m m m m
m m m m m m
+ + + +
− − − − − −
+ + + =
+ + + + + (4.15)
Solving this set of fourteen independent equations (Eqs. 4.11-4.15) for a given
temperature and solution overall molality results in the true (equilibrium) composition of
the liquid phase, expressed as the molality of each species, needed for solving the vapour-
liquid equilibrium equations.
151
4.3.2. Vapour-liquid equilibrium
Only water is treated as a solvent species. Carbon dioxide, AHPD, Pz and the ions
are treated as solute species. The reference state for the chemical potential of water is the
pure liquid at the system temperature and pressure. The chemical potential of a solute
species is a 1 molal solution in pure water at the system temperature and pressure.
The condition of vapour-liquid equilibrium (VLE) is applied in order to calculate
the total pressure and the composition of the gas phase. The extended Raoult’s law is used
to express the VLE for water (Eq. 4.16) and the extended Henry’s law is used to express the
equilibrium for carbon dioxide (Eq. 4.17):
( )exp
satw wsat sat
w w w w w
V P PP a Py
RTϕ ϕ
− =
(4.16)
( ) ( )2 2
2 2 2 2 22
,,, , exp
CO
satCO H O wm m sat
CO CO H O w CO CO
V P Pm H T P Py
RTγ ϕ
∞∗
− =
(4.17)
Because the vapour pressures of both amines used in this work are very low in the
temperature range considered here, the presence of AHPD and piperazine in the vapour
phase was neglected.
4.3.3. Thermodynamic properties
The VLE calculation requires the knowledge of the following properties:
(i) Henry’s constants for the solubility of carbon dioxide in pure water on the molality
scale, ( )2 2, ,m sat
CO H O wH T P , were taken from Rumpf and Maurer (1993) (Table 4.1).
(ii) The temperature dependent equilibrium constants for the reactions (4.1)-(4.10) are
given in Table 4.2. Except for the equilibrium constant K5 which was calculated based on
the experimental data for the system AHPD-water-CO2, all other constants were taken from
Edwards et al.(1978) , Perrin (1965) , Hetzer et al. (1968), and Ermatchkov et al. (2003).
(iii) The vapour pressure satwP and the molar volume wV of pure water were taken from
Saul and Wagner (1987).
(iv) The fugacity coefficients iϕ were calculated using a truncated virial equation of
state. Pure component second virial coefficients 2 2H O,H OB and
2 2CO ,COB for water and carbon
152
dioxide, respectively, were calculated on the basis of the data given by Dymond and Smith
(1980). The mixed second virial coefficients 2 2CO ,H OB were taken from Hayden and
O’Connell (1975) and correlated as a function of temperature.
(v) The partial molar volumes 2 2,CO H OV ∞ of carbon dioxide dissolved at infinite dilution
in water were calculates as recommended by Brelvi and O’Connell (1972) and correlated as
a function of temperature.
Table 4.1. Henry’s constant for the solubility of carbon dioxide in pure water
( 273 / 473T K≤ ≤ )12.
( )2 2 2 2 2 2 2 2 2 2
1, , , , ,ln , / (MPa kg mol ) / ( / K) ( / K) ln( / K)m sat
CO H O w CO H O CO H O CO H O CO H OH T P A B T C T D T−⋅ ⋅ = + + +
2 2,CO H OA 2 2,CO H OB
2 2,CO H OC 2 2,CO H OD
192.876 -9624.4 0.01441 -28.749
4.3.4. Pitzer’s GE model for activity coefficients and interaction parameters
In the literature, several models are used to characterize VLE of CO2 in aqueous
amines solutions. Among them, the Kent and Eisenberg (1976) and the Deshmukh and
Mather (1981) models are frequently used. However the former one doesn’t take into
account the activity coefficients in solution and the latter is limited to low concentration
because the activity coefficients are calculated with the Guggenheim’s equation (Pitzer,
1973). In this research, a more rigorous model is then used to cover the wide range of
amines concentration.
Activity coefficients of both neutral and ionic species were calculated using a
modified Pitzer model for the excess Gibbs energy of aqueous electrolyte solutions (Pitzer,
1973), :
(4.18)
is a modified Debye-Hückel term depending on ionic strength, temperature and
solvent (water) properties:
( ) ( )1
E
i j ij i j k ijki w j w i w j w k ww w
G f I m m I m m mRTn M
λ τ≠ ≠ ≠ ≠ ≠
= + +∑∑ ∑∑∑
( )1f I
153
(4.19)
where I is the ionic strength and is the Debye-Hückel parameter for the osmotic coefficient:
(4.20)
. (4.21)
Table 4.2. Equilibrium constants for chemical reactions (4.1)-(4.10).
2 3ln / ( / K) ln( / K) ( / K) ( / K) ( / K)R R R R R R RK A T B T C D T E T F T= + + + + +
R AR BR CR DR ER FR Ref. T/K
1 -12091.1 -36.7816 235.482 0 0 0 Edwards et al. (1978)
273-498
2 -12431.7 -35.4819 220.067 0 0 0 Edwards et al. (1978)
273-498
3 -13445.9 -22.4773 140.932 0 0 0 Edwards et al. (1978)
273-498
4 0 0 22.61853 0.591854 -2.360429·10-
3 2.814271·10-6 Perrin
(1966) 273-323
5 0 0 213.8527 -2.123369 7.033246·10-3 -7.884854·10-6 this work 288-333
6 3814.4 0 14.119 -1.51·10-2 0 0 Hetzer et al. (1968)
278-328
7 2192.3 0 10.113 -1.74·10-2 0 0 Hetzer et al. (1968)
273-323
8 1570.4 0 -3.75 0 0 0 Ermatchkov et al. (2003)
273-323
9 574.2 0 -1.587 0 0 0 Ermatchkov et al. (2003)
273-323
10 1517 0 4.354 0 0 0 Ermatchkov et al. (2003)
273-323
The dielectric constant of pure water, D was taken from Bradley and Pitzer (1979).
( ) ( ) ( )1 4 /1.2 ln 1 1.2f I A I Iφ= − +
Aφ
212 i i
iI m z= ∑
( )3/22
1/2
0
1 23 4A w
eA NDkTφ π ρ
πε
=
154
( )ij Iλ is the ionic strength dependent second virial coefficient:
( ) ( ) ( ) ( ) ( )( )0 1 22 / 1 1 xij ij ijI x x eλ β β − = + − +
(4.22)
where 2x I= .
The influence of temperature on the binary interaction parameters ( )0ijβ and ( )1
ijβ is
approximated by the relation:
10( )
/qf T q
T K= +
(4.23)
The ternary interaction parameters ijkτ are considered independent of temperature.
The equation for the activity coefficients of dissolved species follows from the
appropriate derivative of GE and water activity is calculated from the Gibbs-Duhem
equation:
( ) ( )
( )
, 2
1 22
2
2ln ln 1 1.2 21.21 1.2
1 1 32
mi i j ij
j w
jk xi j k j k ijk
j w k w j w k w
IA z I m II
xz m m x e m mIx
φγ λ
βτ
∗
≠
−
≠ ≠ ≠ ≠
= − + + + − +
− + + +
∑
∑∑ ∑∑
(4.24)
( ) ( )( )1.5
0 1ln 21 1.2
2
xw w i j ij ij
i w j w
w i j k ijk ii w j w k w i w
Ia M A m m eI
M m m m m
φ β β
τ
−
≠ ≠
≠ ≠ ≠ ≠
= − + −
+
+
∑∑
∑∑∑ ∑
(4.25)
All interaction parameters used in this work are given in Table 4.3.
Table 4.3. Interaction parameters in Pitzer’s GE equation for the system AHPD-PZ-CO2-H2O
10( )
/qf T q
T K= +
Parameter 0q 1q Subsystem Reference
( )2 3
0CO ,HCO
β − 2.256 -379.5 AHPD+CO2+H2O this work
155
( )+
2 3
0CO ,RNH
β -4.547 917.7
( )+
3 3
0HCO ,RNH
β − 0.700 -400.2
( )+
3 3
1HCO ,RNH
β − 1.017 -1050
( )2
0CO ,RNHCOO
β − -6.600 995.2
( )2 +3 3
0CO ,RNH
β − 3.400 -1006
( )+
2 3
0RNH ,RNH
β 0.300 -100.0
- +2 3 3CO ,HCO ,RNH
τ 0.0707 0
- - +3 3 3HCO ,HCO ,RNH
τ 0.0480 0
( )2
0CO ,PzH
β + 0.14624 -187.24 PZ+CO2+H2O Kamps et al. (2003)
( )+
3
0HCO ,PzH
β − 0.55489 2.0459
( )+
3
1HCO ,PzH
β − 1.8949 776.48
( )2
0CO ,PzH COO
β + − 0.55705 -196.84
( )0PzH COO ,PzH COO
β + − + − 0.096213 -72.2
( )1PzH COO ,PzH COO
β + − + − -0.83929 324.79
( )0PzH ,PzCOO
β + − -2.0678 776.43
( )( )
2
0PzH ,Pz COO
β + − -1.3044 440.98
( )0Pz,PzCOO
β − 0.34964 -83.169 Ermatchkov et al. (2006)
156
4.3.4.1. The system AHPD-CO2-H2O
Interaction parameters for the ternary system AHPD-CO2-H2O were determined on
the basis of experimental data taken from the literature (Le Tourneux et al., 2008; Park et
al., 2002a) and from the present work. In this system, eight species are present in the liquid
phase: 2CO , 3HCO− , 23CO − , 2RNH , 3RNH+ , RNHCOO− , H+ and OH− . Due to the very
low concentration of H+ and OH− with respect to the other species, their interactions with
all other species were ignored and therefore, the corresponding interaction parameters were
set to zero. Binary and ternary interaction parameters between neutral species, 2CO and
2RNH were considered negligible and were set to zero. Except for the binary interaction
parameter between 2RNH and 3RNH+ , all binary and ternary interaction parameters
between 2RNH and any other species were also set to zero. In addition, the ionic strength
dependence of the second virial coefficient (Eq. 4.22) was neglected for the all interactions
except for 3RNH+ - 3HCO− . In order to reduce the number of parameters, all binary and
ternary interaction parameters involving species with the same sign of charge were
neglected. Only the parameters which were found to have a significant influence on the
liquid phase species distribution were optimized based on the experimental data: ( )2 3
0CO ,HCO
β − ,
( )+
2 3
0CO ,RNH
β , ( )+
3 3
0HCO ,RNH
β − , ( )+
3 3
1HCO ,RNH
β − , ( )2
0CO ,RNHCOO
β − , ( )2 +3 3
0CO ,RNH
β − and ( )+
2 3
0RNH ,RNH
β . A sensitivity
study revealed that all other possible interaction parameters that appear in the expressions
for the activity coefficients (Eqs. 4.24 and 4.25) can be neglected without reducing the
accuracy of VLE representation of this system. Parameters 0q , 1q and the ternary ones ijkτ
were fitted simultaneously to the selected experimental data chosen as it will be described
in the section 4.4.2.
4.3.4.2. The system AHPD-Pz-CO2-H2O
Based on the thermodynamic description of the solubility of carbon dioxide in a
single amine system: AHPD-CO2-H2O and Pz-CO2-H2O, the carbon dioxide solubility in
the mixed AHPD+Pz aqueous system was predicted using the available interaction
parameters (Table 4.3). Interaction parameters for the ternary system AHPD-CO2-H2O
157
were determined from the experimental data of the present work and from literature (Le
Tourneux et al., 2008; Park et al., 2002a), as described in the previous section. No other
parameters were found in the literature concerning this ternary system. Interaction
parameters for the ternary system Pz-CO2-H2O were taken from Kamps et al. (2003) and
Ermatchkov et al. (2006).
4.4. Results and discussions 4.4.1 Experimental setup verification
To check the validity of the experimental setup and procedures, physical absorption
of CO2 in water was made at 293.15 K and 313.15 K and for several CO2 partial pressures.
The experimental data were compared with literature values in Figure 4.2. It is possible to
see that our results are in excellent agreement with literature values over the entire pressure
range. As expected, the CO2 concentration in water decrease when temperature increase at
constant CO2 partial pressure.
Figure 4.2. CO2 solubility in water: comparison with literature values.
4.4.2 Solubility measurements
N2O solubility
The absorption of N2O in Pz-AHPD aqueous solutions ranging from 0.10 to 0.50
kmol.m-3 Pz and 1.0 to 3.0 kmol.m-3 AHPD was performed between 288.15 to 333.15 K.
The experimental results, expressed in term of Henry’s law constant are indicated in Table
4.4. The uncertainties of indicated values are calculated to be within 2%. As expected,
158
Henry’s law constant values increase with increasing either temperature for a given solution
concentration or amine concentration in aqueous solutions at constant temperature.
Table 4.4. Henry's law constants for N2O in Pz (1)-AHPD (2) solutions T m1 m2 HN2O
(K) (mol.kg-1) (mol.kg-1) (kPa.m3.kmol-1) 288.15 1.1141 0.1114 3111.0 288.15 2.5881 0.6470 3687.2 288.15 4.2016 0.1401 4673.5 298.15 1.1351 0.3405 3865.4 298.15 3.3634 0.4036 5131.0 313.15 1.1570 0.5785 6518.2 313.15 3.3629 0.4035 6960.0 333.15 1.1149 0.1115 9265.3 333.15 2.5912 0.6478 12817.5 333.15 4.2046 0.1402 13158.4
CO2 solubility
CO2 solubility measurements were made in three different reactive systems: CO2-
Pz-H2O, CO2-AHPD-H2O, and CO2-Pz-AHPD-H2O in order to respectively: i) validate the
apparatus and the procedures for chemical absorption at high pressure, ii) obtain more CO2
solubility data in AHPD solution, needed to check the validity of literature sources between
those available; this is necessary to obtain good interaction parameters in the
thermodynamic model and, iii) obtain the CO2 solubility in the mixed Pz-AHPD aqueous
solutions to determine the effect of Pz on the equilibrium solubility of AHPD and to test the
prediction capacity of the developed VLE model in representing the experimental data for
the quaternary system based on the interaction parameters for the corresponding ternary
systems.
For the system CO2-Pz-H2O, CO2 chemical absorption was made at 313.15 K in a
solution containing 2 kmol.kg-1 of Pz and up to a total pressure of 2 900 kPa. This pressure
is above the highest pressure reached for the CO2 absorption in the mixed solvent and it is
possible to see in Figure 4.3 that even at this high pressure, the correlation between our data
and those of Kamps et al. (2003) is particularly good.
159
0.1
1
10
100
1000
10000
0.0 0.5 1.0 1.5 2.0
CO
2pa
rtia
l pre
ssur
e / (
kPa)
CO2 loading / (kmol CO2.kmol-1 AHPD)
T : 298.15 K
Park et al., 2003
Le Tourneux et al., 2008
This work
Figure 4.3. CO2 solubility in Pz aqueous solution: comparison with literature values ( Pzm = 2.0 mol.kg-1).
Figure 4.4a. CO2 solubility in AHPD aqueous solution at 298.15 K ( AHPDm = 0.9172
mol.kg-1).
For the system CO2-AHPD-H2O, some disagreements were found between the
reported equilibrium solubility of CO2 by the research group of Park et al. (2003) when
compared to the data of Le Tourneux et al. (2008), as it can be seen in Figure 4.4a. Our
results agree very well with those of Le Tourneux et al. (2008), which were performed
using a different experimental setup. We therefore consider the data from this source
reliable to be used in the interaction parameter determination. A verification of some of the
160
data reported in another article by Park et al. (2002a) was also made and shown in Figure
4.4b. It is possible to notice that Park’s data disagree from our result at pressure larger than
around 500 kPa. We therefore decided not to include these data (P > 500 kPa) in the
database used for the parameters estimation. Consequently, the number of reliable data for
the system CO2-AHPD-H2O is 177: 84 from Le Tourneux et al. (2008), 17 from Park et al.
(2002a) and 76 from this work (Table 4.5).
Fig. 4.4b. CO2 solubility in AHPD aqueous solution at 323.15 K, comparison with Park et al. (2002a)( AHPDm = 0.9172 mol.kg-1).
Table 4.5. CO2 solubility in AHPD aqueous solutions
T mAHPD PCO2 CO2 loading (K) (mol.kg-1) (kPa) (mol CO2.mol-1 AHPD)
298.15 0.917 0.31448 0.0745 298.15 0.917 1.1729 0.1817 298.15 0.917 4.4127 0.3652 298.15 0.917 27.331 0.7010 298.15 0.917 232.00 1.0208 298.15 0.917 533.60 1.1497 298.15 0.917 1237.6 1.3953 298.15 0.917 1938.4 1.6171 298.15 0.917 2637.6 1.8545 323.15 0.917 5.8978 0.1967 323.15 0.917 26.232 0.4010 323.15 0.917 68.496 0.5830 323.15 0.917 130.17 0.7272 323.15 0.917 236.23 0.8540
161
323.15 0.917 526.82 1.0287 323.15 0.917 1303.9 1.3921 333.15 0.917 2106.0 1.6782 284.84 2.000 0.90620 0.1602 284.82 2.000 2.3588 0.3283 284.76 2.000 6.7321 0.5307 284.53 2.000 23.350 0.7551 284.70 2.000 100.88 0.9452 284.65 2.000 241.76 1.0257 284.71 2.000 450.35 1.0825 284.78 2.000 927.36 1.1881 284.82 2.000 1249.1 1.2345 303.16 2.000 0.92003 0.0940 303.18 2.000 2.8518 0.1914 303.20 2.000 6.8983 0.3063 303.18 2.000 15.729 0.4396 303.16 2.000 35.435 0.5824 303.14 2.000 83.534 0.7289 303.11 2.000 259.50 0.8948 303.12 2.000 619.37 1.0011 303.11 2.000 1034.1 1.0715 333.14 2.000 7.6416 0.1209 333.13 2.000 25.197 0.2284 333.12 2.000 55.973 0.3357 333.11 2.000 94.976 0.4310 333.11 2.000 152.18 0.5138 333.11 2.000 216.84 0.5815 333.11 2.000 377.32 0.6940 333.13 2.000 683.61 0.8172 284.29 3.000 0.51300 0.1422 284.43 3.000 1.5500 0.2953 303.26 3.000 0.90389 0.1057 303.29 3.000 3.3283 0.2161 303.31 3.000 8.9937 0.3399 303.29 3.000 24.050 0.4897 303.31 3.000 59.078 0.6338 303.32 3.000 184.42 0.8054 303.27 3.000 456.96 0.9163 303.28 3.000 775.93 0.9741 333.20 3.000 22.759 0.0796 333.11 3.000 35.534 0.1623 333.12 3.000 59.999 0.2524 333.13 3.000 98.446 0.3399
162
333.12 3.000 198.23 0.4743 333.12 3.000 416.19 0.6144 333.14 3.000 676.89 0.7063 333.22 3.000 914.81 0.7642 293.26 4.000 0.60554 0.1303 293.20 4.000 2.2092 0.2569 303.16 4.000 1.1836 0.1211 303.17 4.000 4.6753 0.2462 303.17 4.000 11.897 0.3651 303.17 4.000 40.680 0.5466 303.15 4.000 77.589 0.6436 303.13 4.000 258.72 0.8125 333.17 4.000 24.854 0.1933 333.20 4.000 95.209 0.3628 333.20 4.000 176.18 0.4585 333.08 4.000 373.09 0.5864 333.09 4.000 601.44 0.6704 333.11 4.000 849.99 0.7331 333.22 4.000 1079.1 0.7765
For the system CO2-Pz-AHPD-H2O, all experimental results are listed in Table
A.22. The CO2 partial pressure and CO2 loading reported in Table 4.5 and Table A.22 have
respectively a maximum calculated uncertainty of 0.6% and 1.5%. From Figure 4.5, it is
possible to observe that at constant amine concentrations and CO2 partial pressure an
increase in temperature leads to a decrease of the CO2 loading capacity. Furthermore, as
expected, at constant temperature, an increase in the total amine concentration leads to a
decrease of the gas absorption capacity.
The absorption of CO2 in the AHPD aqueous solution was shown to be similar to
that in the AMP aqueous solution (Silkenbaumer et al., 1998). The system pressure
increases very slowly during the chemical absorption when the gas is mostly dissolved in
non-volatile ionic form then the pressure show a sharp increase above the stoichiometric
gas-amine ratio in the physical absorption section.
An example of the speciation in an aqueous AHPD by carbon dioxide addition,
based on the equilibrium model, is shown in Figure 4.6. The concentration profile for
several species is represented as a function of the CO2 loading for an AHPD aqueous
solution with a molality of 0.9172 mol.kg-1 at 298 K. Because H+ and OH− concentrations
163
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 0.5 1 1.5 2
mi/
(mol
.kg-1
)
CO2 loading / (kmol CO2.kmol-1 amines)
CO2
RNH2
RNHCOO-
HCO3-
RNH3+
CO32- T : 298.15 K
0
500
1000
1500
2000
2500
0.0 0.5 1.0 1.5 2.0
CO
2Pa
rtia
l pre
ssur
e / (
kPa)
CO2 loading / (kmol CO2. kmol-1 amines)
2 M AHPD 0.5 M Pz, 333.15 K1 M AHPD 0.1 M Pz, 333.15 K2 M AHPD 0.5 M Pz, 288.15 K1 M AHPD 0.1 M Pz, 288.15 K
are much lower than the concentrations of all other species, the corresponding curves were
not represented. The amine concentration decreases rapidly at CO2 loadings less than about
1 mol/mol of amine, while the bicarbonate and the protonated amine sharply increase. It
can be noted that at a CO2 loading of about 1 mol/mol of amine, practically all amine is
converted preferentially into the protonated amine and bicarbonate. Moreover, the
carbamate (RNHCOO-) concentration is very low, which is consistent with the behaviour of
the sterically hindered amines, especially when the hindered character is very important,
like in the case of AHPD (Bougie and Iliuta, 2009; Park et al., 2003).
Figure 4.5. CO2 solubility in Pz-AHPD aqueous solutions at 288.15 and 333.15 K.
Figure 4.6. Predicted species distribution in the AHPD+CO2+H2O system at 298.15 K (AHPDm = 0.9172 mol.kg-1).
164
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0.0 0.5 1.0 1.5
mi/
(mol
.kg-1
)
CO2 loading / (kmol CO2.kmol-1 amines)
RNH3+ HCO3
-
Pz
PzHCOO- PzH+
T : 298.15 KRNH2
CO2
PzH+COO-
Figure 4.7. Predicted species distribution in the Pz-AHPD+CO2+H2O system at 298.15 K (AHPD = 1.0 kmol.m-3 and Pz = 0.3 kmol.m-3).
Same type of concentration profiles are represented in Figure 4.7 for the quaternary
system AHPD-Pz-CO2-H2O for an aqueous solution containing 1 kmol·m-3 AHPD and 0.3
kmol·m-3 Pz at 298.15 K. For the same reasons mentioned for the ternary system AHPD-
CO2-H2O, the curves corresponding to H+ and OH− concentrations were not represented. It
can be seen that the AHPD behaviour in the presence of Pz is similar to that observed in the
single aqueous amine (AHPD) system. The AHPD concentration decreases rapidly at CO2
loadings less than about 1 mol/mol of amine, while the bicarbonate and the protonated
amine ( 3RNH+ ) increase sharply. At a CO2 loading of about 1 mol/mol of amine, practically
all amine is converted preferentially into the protonated amine and bicarbonate and the
AHPD carbamate ( RNHCOO− ) concentration is very low. On the contrary, Pz reacts very
rapidly at very low CO2 loadings (up to about 0.2 mol/mol of total amine) and it is
preferentially converted into Pz carbamate, PzCOO− (Bougie et al., 2009; Ermatchkov et
al., 2003). Pz dicarbamate, 2Pz(COO )− and diprotonated Pz ( 22PzH + ) concentrations are
very low. Moreover, with the increase of CO2 loading, Pz carbamate, PzCOO− and
protonated Pz, PzH+ are converting into protonated Pz carbamate, PzH COO+ − .
165
1
10
100
1000
10000
0.0 0.2 0.4 0.6 0.8 1.0
Pres
sure
/ (k
Pa)
CO2 loading / (kmol CO2.kmol-1 AHPD)
293.15 K303.15 K333.15 KCorrelation
1
10
100
0.0 0.5 1.0 1.5 2.0 2.5
Pres
sure
/ (k
Pa)
CO2 loading / (kmol CO2.kmol-1 AHPD)
m = 0.0123 mol.kg-1m = 0.0410 mol.kg-1m = 0.0821 mol.kg-1m = 0.2061 mol.kg-1CorrelationT = 313.15 K
Figure 4.8a. CO2 solubility in aqueous solution of AHPD. Experimental results of this work, AHPDm = 4.0 mol.kg-1.
Figure 4.8b. CO2 solubility in aqueous solution of AHPD. Experimental results by
Le Tourneux et al. (2008), different AHPD molalities.
166
All 177 selected experimental data for the system AHPD-CO2-H2O, covering a
large range of amine concentrations (between 0.0125 and 4 mol.kg-1), temperature (between
283.15 and 333.15 K) and total pressure (between 1.85 and 2640.8 kPa) were correlated
together with an average relative deviation of 22.7%. The interaction parameters for this
system are valid for the entire range of temperature, pressure and amine concentration
(Table 4.3). Generally, higher deviations were obtained at very large amine concentration
and very high pressures. Figures 4.8a and 4.8b show some comparisons between
experimental and calculated total pressure at low and large amine concentrations.
The solubility of carbon dioxide in aqueous solutions of mixed amine (AHPD+Pz)
was predicted by supposing that the parameters characterising the single amines systems
are essential for describing the quaternary system behaviour. Predictions of the CO2 partial
pressure correspond to an average relative deviation of 37%. This is believed to mainly due
to the fact that 49% of our quaternary experimental data are obtained at temperatures lower
than 313 K, the lowest valid temperature of the interaction parameters available in the
literature for the system Pz-CO2-H2O. For the same reason, no attempt was made to
correlate the experimental data of the quaternary system. New experimental work is
presently in progress in our laboratory in order to enlarge the experimental data base for the
Pz-CO2-H2O system at lower temperatures, which will allow us to revise and extend the
available interaction parameters using the new data.
4.5. Conclusion In the present work, new data concerning the solubility of CO2 and N2O in aqueous
mixtures of 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD) and piperazine (Pz) were
obtained over a large range of temperature (283.15-333.15 K) and amines concentrations
(0.91-4.36 mol.kg-1). Based on the experimental data, Henry’s law constant for CO2 in
these solutions were calculated using the N2O analogy. The experimental data for the
ternary system AHPD-CO2-H2O were satisfactorily correlated using a modified Pitzer’s
thermodynamic model for the activity coefficients combined with the virial equation of
state for representing the fugacity coefficients. The solubility of carbon dioxide in aqueous
solutions of mixed amine (AHPD+Pz) was predicted by supposing that the available
parameters characterising the single amines systems are appropriate for describing the new
167
data for the quaternary system behaviour. However, the quite large deviations obtained
between experimental and calculated equilibrium pressure led to the conclusion that more
experimental data for the Pz-CO2-H2O system at lower temperatures are necessary in order
to allow the revision of the interaction parameters for this system.
168
Because Pz was chosen as accelerator for CO2 absorption in AHPD aqueous solutions, the
precedent chapter concerned the thermodynamic study of the aqueous CO2 + AHPD + Pz
system. For modeling purpose, all data available in the literature for the aqueous Pz
system were considered. However, half of these data were obtained in experimental
conditions different from the aqueous CO2 + AHPD + Pz system. Therefore, in Chapter 5,
new solubility data of CO2 in aqueous piperazine solutions were obtained experimentally
using a vapor-liquid equilibrium apparatus based on a static-synthetic method, and data
were modelled with a modified Pitzer’s thermodynamic model for the activity coefficients.
169
Chapter 5. CO2 absorption in aqueous piperazine solutions: Experimental study and modeling
Résumé
Dans cette étude, de nouvelles données de solubilité du CO2 dans des solutions aqueuses de pipérazine (Pz) ont été mesurées dans le domaine de température 287.1 - 313.1 K et concentration d’amines m variant de 0.10 à 2.00 mol.kg-1. Les mesures ont été réalisées à des pressions partielles de CO2 entre 0.11 et 525.17 kPa, en utilisant un appareil d’équilibre liquide-vapeur basé sur la méthode statique-synthétique. Ces données expérimentales couplées avec celles disponibles dans la littérature pour le système ternaire Pz-CO2-H2O ont été corrélées en utilisant un modèle qui combine l’équation de viriel pour le calcul du coefficient de fugacité avec un modèle thermodynamique modifié de Pitzer pour les
coefficients d’activité. Sur la base de nouveaux coefficients d’interaction 0,i jβ et
1,i jβ
couvrant un large domaine de températures, pressions partielles de CO2 et concentrations d’amine, le modèle a montré une capacité satisfaisante de corrélation des données expérimentales de solubilité.
170
Abstract
In this work, new solubility data of CO2 in aqueous piperazine (Pz) solutions were measured over a temperature range from T = (287.1 to 313.1) K and for amine concentrations from m = (0.10 to 2.00) mol.kg-1. The CO2 partial pressure was kept within
2COP = (0.11 to 525.17) kPa using a VLE apparatus based on a static-synthetic method. These experimental data and those found in the literature for the ternary system Pz-CO2-H2O were correlated using a model combining the virial equation of state to calculate the fugacity coefficients with a modified Pitzer’s thermodynamic model for the activity coefficients. With the new extended interaction parameters 0
,i jβ and 1,i jβ that cover a wide
range of temperature, CO2 partial pressure and amine concentration, the model is able to correlate satisfactorily the available reliable experimental solubility data.
171
5.1. Introduction In the last few years, large human emission of greenhouse gases has become one of
the most discussed environmental issues around the word. This has motivated intensive
research on CO2 capture where new and more energy-efficient absorbents are essential. For
technical, economical and environmental concerns, actual industrial CO2 absorption
processes use aqueous solutions of alkanolamines. This technique is widely applied for acid
gases (CO2, H2S) removal during natural gas sweetening as well as for CO2 capture from
fossil-fuel-fired power plants or from other important industries such as chemical and
petrochemical, steel, aluminium and cement production.
Industrially more often used alkanolamines are monoethanolamine (MEA),
diethanolamine (DEA), N-methyldiethanolamine (MDEA), and 2-amino-2-methyl-1-
propanol (AMP) (Kohl and Nielsen, 1997). The choice of a certain amine is mainly based
on the absorption capacity, reaction kinetics and regenerative potential and facility. The use
of blended alkanolamines solutions has also recently become very attractive because of the
combination of each amine advantages: a fast reactivity from a primary or secondary amine
coupled with the high absorption capacity and low solvent regeneration cost from a tertiary
or sterically hindered alkanolamine (SHA).
In our laboratory, extensive studies of CO2 capture in membrane contactors using
Piperazine-activated aqueous SHA solutions are in progress. A set of four SHA was chosen
in order to study the hindered effect on the absorbent properties (Bougie and Iliuta, 2009).
It concerns AMP, a simple hindrance form of MEA, and three SHA derived from AMP: 2-
amino-2-methyl-1,3-propanediol (AMPD), 2-amino-2-ethyl-1,3-propanediol (AEPD) and
2-amino-2-hydroxymethyl-1,3-propanediol (AHPD). Based on these solutions kinetics
(Bougie and Iliuta, 2009; Yih and Shen, 1988; Yoon et al., 2003; Yoon et al., 2002a),
equilibrium data (Baek et al., 2000; Park et al., 2002b; Paul et al., 2009c; Xu et al., 1992c)
and their regenerative capacity (Bougie and Iliuta, 2010a), it appeared that the Pz-AHPD
mixture may be an interesting alternative to conventional amine solutions. Development of
a model describing this solution thermodynamic equilibrium would be of great interest as
deviation from equilibrium provides the driving force in kinetically controlled absorption.
172
Such work was reported in a previous paper (Bougie and Iliuta, 2010b): a thermodynamic
model based on the Pitzer’s equations for the activity coefficients coupled with the
truncated virial equation of state for representing the non ideality of the vapour phase was
used to predict the CO2 solubility in the CO2-Pz-AHPD-H2O system with the assumption
that the interaction parameter describing the ternary subsystem (CO2-Pz-H2O and CO2-
AHPD-H2O) are necessary to describe the quaternary system CO2-Pz-AHPD-H2O. The
resulting model prediction showed large deviation (average relative deviation of 37%) with
our experimental data. This was believed to come from the fact that 49% of our quaternary
experimental data were obtained at temperatures lower than 313.1 K, the lowest reported
temperature of the interaction parameters available in the literature12 for the system CO2-
Pz-H2O.
In this work, new solubility data of CO2 in aqueous piperazine solutions were
obtained over a temperature range from T = (287.1 to 313.1) K and for amine
concentrations from m = (0.10 to 2.00) mol.kg-1 using a VLE apparatus based on a static-
synthetic method. These data will be used i) to increase the very scarce reliable database of
CO2 solubility in aqueous Pz solutions below 313 K, and ii) along with all reliable data
found in the literature for CO2-Pz-H2O at all temperatures, to revise and extend the
available interaction parameters for this subsystem.
5.2. Experimental section 5.2.1 Reagents
All aqueous piperazine solutions used in this work were prepared with degassed
distilled water and piperazine (CAS # 110-85-0). The amine (from Laboratoire MAT,
Quebec, Canada) was supplied with a mass fraction of 0.999 and was used without further
purification. CO2 gas bottle was of commercial grade with a minimum purity of 99.9 % and
was supplied by Praxair.
5.2.2 Apparatus and procedures
The experimental setup for the CO2 solubility measurements used in this work is
shown in Figure 4.1 (Chapter 4). As the same setup and procedures were used in our
previous work (Bougie and Iliuta, 2010b), only the main details will be mentioned here.
173
The vapour-liquid equilibrium cell (from Armines, France) is made of TA6V titanium and
has an internal volume of about 1.15×10-4 m3. The cell is agitated with a magnetic rod and
is located in a modified XU027 laboratory oven from France Etuves, which allows a
temperature control of ± 0.1 K. A special feature of this apparatus, compared to similar
ones, is the addition in the oven of a coil refrigerated with a thermostated bath (K-12108-10
from Cole-Palmer). This coil allowed us to made solubility measurement under room
temperature (273.15 to 303.15 K) with the same temperature precision. Pressure in the cell
was measured by one of the two installed absolute pressure transducers (Druck PTX-611,
0-100 kPa and 0-16000 kPa) according to the pressure range with a precision of 0.08%.
Liquid introduction inside the equilibrium cell was made with a variable volume press
(stainless steel 316, internal diameter of 3.002×10-2 m) equipped with a linear encoder
(Heidenhain, LS487C) which allowed knowing the exact longitudinal position of the piston
in the press with an accuracy of ± 2×10-6 m. Gas introduction inside the equilibrium cell
was made by a thermostated small gas cylinder with an internal volume of about 7×10-5 m3.
This small gas cylinder was equipped with a Druck PTX-611 0-16000 kPa absolute
pressure transducer.
A standard CO2 solubility experiment consisted of a sequence of successive step.
First, the piperazine aqueous solution was prepared to its specific concentration, m = (0.10
to 2.00) mol.kg-1, by a gravimetric method. A Mettler Toledo AE204 balance with a
precision of ±0.001 g was used. The solution was then degassed, put inside the variable
volume press and subsequently, transferred with the piston in the equilibrium cell. All these
steps were made under vacuum. The equilibrium cell was next heated to the desired
temperature and the solution was agitated. At this stage, the vapour pressure of the solution
was measured by the low pressure transducer. This was followed by the introduction of the
CO2 in the equilibrium cell via the small gas cylinder. Introduced CO2 mole number was
calculated by using the cylinder volume, its temperature as well as the observed pressure
drop in the cylinder after the gas introduction. System equilibrium was reached when the
pressure inside the equilibrium cell was varying less than 0.5% for at least 30 minutes. The
remaining CO2 mole number in the cell was calculated based on the temperature, the
equilibrium pressure and the head space volume, and corrected by the compressibility
174
factor. The difference between the introduced and the remaining gas mole number in the
head space of the equilibrium cell was then calculated which lead to the concentration of
absorbed gas in the solution.
5.3. Thermodynamic modeling of the vapour-liquid equilibrium 5.3.1. Chemical equilibrium in the liquid phase
When CO2 is absorbed in piperazine solutions, many chemical reactions happen in
the liquid phase. The model applied to correlate/predict the solubility of carbon dioxide in
this solution considers the following equilibriums for the chemical species in the liquid
phase: the formation and dissociation of bicarbonate (reactions 5.1 and 5.2), the
autoprotolysis of water (reaction 5.3), the protonation and diprotonation of piperazine
(reactions 5.4 and 5.5), and the formation of piperazine carbamate, piperazine dicarbamate
and protonated piperazine carbamate (reactions 5.6 to 5.8).
, H HCO OH CO -322
1 ++→←+ K (5.1)
, H CO HCO -23
-3
2 ++→←K (5.2)
,OH H OH -2
3 +→← +K (5.3)
,PzH H Pz 4 ++ →←+ K (5.4)
,PzH H PzH 22
5 +++ →←+ K (5.5)
O,H PzCOO HCO Pz 2--
36 +→←+ K (5.6)
O,H )Pz(COO HCO PzCOO 22--
3- 7 +→←+ K (5.7)
. COOPzH H PzCOO -- 8 ++ →←+ K (5.8)
The condition for chemical equilibrium for a chemical reaction R is: . 8) to1( where)( ,∏ ==
iiR RaTK Riν (5.9)
Constants for the calculation of the various KR as a function of temperature as well as their
sources are given in Table 5.1.
In addition to the above equilibrium equations, overall Pz and CO2 concentrations
(mol.kg-1) as well as charge balance must be satisfied. In the balance equations for Pz and
carbon dioxide in the liquid phase (eqs. 5.10 and 5.11) Pz~m denote the stoichiometric
175
concentration of Pz (mol.kg-1) and ∝ denotes the CO2 loading in the solutions, expressed as
total moles of CO2 absorbed both chemically and physically per mole of amine.
-2
--22 COOPzH)Pz(COOPzCOOPzHPzH ~
+++ +++++= mmmmmmm PzPz (5.10)
-2
---23
-32 COOPzH)Pz(COOPzCOOCOHCOCO 2 ~
++⋅++++=⋅ mmmmmmmPzα (5.11)
2---2
3-3
-22 )Pz(COOPzCOOCOHCOOHPzHPzHH 2 2 2 mmmmmmmm ⋅++⋅++=⋅++ +++ (5.12)
Solving this set of eleven independent equations (eqs. 5.9 to 5.12) for a given
temperature, Pz overall concentration and CO2 loading results in the true (equilibrium)
composition of the liquid phase, expressed as the molality of each species (mol.kg-1),
needed for solving the vapour-liquid equilibrium equations.
Table 5.1. Chemical Equilibrium Constant (on the molality scale) for the Chemical
Reaction R, Expressed on the Molality Scale, and Temperature Range of Validity.
2R /K)(E /K)(D /K)ln(C
/K)(B A ln
TTT
TK +⋅+⋅++=
KR A B C 102 · D 10-5 · E T/K References
K1 -1203.01 68359.6 188.444 -20.6424 -47.1291 273.1 to 673.1 Patterson et al. (1982)
K2 175.360 -7230.6 -30.6509 1.31478 -3.72805 273.1 to 523.1 Patterson et al. (1984)
K3 140.932 -13445.9 -22.4773 - - 273.1 to 498.1 Edwards et al. (1978)
K4 14.119 3814.44 - -1.5096 - 273.1 to 323.1 Hetzer et al. (1968)
K5 10.113 2192.3 - -1.7396 - 273.1 to 323.1 Hetzer et al. (1968)
K6 -8.635 3616.0 - - - 283.1 to 333.1 Ermatchkov et al. (2003)
K7 -3.654 1322.1 - - - 283.1 to 333.1 Ermatchkov et al. (2003)
K8 10.025 3493.0 - - - 283.1 to 333.1 Ermatchkov et al. (2003)
5.3.2. Vapour-liquid equilibrium
In this study, only water is treated as a solvent species. Carbon dioxide, piperazine
and the several ions are treated as solute species. The reference state for the chemical
potential of water is the pure liquid and defined as a 1 molal solution in pure water for the
solute species, both at the system temperature and pressure.
The condition of vapour-liquid equilibrium (VLE) is applied in order to calculate
the total pressure and the composition of the gas phase. The extended Raoult’s law is used
176
to express the VLE for water (Eq. (5.13)) and the extended Henry’s law is used to express
the equilibrium for carbon dioxide (Eq. (5.14)). It was assumed that the presence of
piperazine in the gas phase could be neglected.
ww
satwwsat
wsat
w PyaRT
PPVP ϕϕ w
)(exp =
− (5.13)
22
22
2222 COCO
satwOH,COsat
wOH,CO,
COCO )-(
exp ),( ϕγ PyRT
PPVPTHm mm =
∞∗ (5.14)
The VLE calculation requires the knowledge of the following properties:
(vi) Henry’s constants for the solubility of carbon dioxide in pure water on the molality
scale, ),( satwOH,CO 22
PTH m , were taken from Rumpf and Maurer (1993).
(vii) The vapour pressure satwP and the molar volume wV of pure water were taken from
Saul and Wagner (1987).
(viii) The fugacity coefficients iϕ were calculated using a truncated virial equation of
state. Pure component second virial coefficients 2 2H O,H OB and
2 2CO ,COB for water and carbon
dioxide, respectively, were calculated on the basis of the data given by Dymond and Smith
(1980). The mixed second virial coefficients 2 2CO ,H OB were calculated according to the
correlations of Hayden and O’Connell (1975).
(ix) The partial molar volumes 2 2,CO H OV ∞ of carbon dioxide dissolved at infinite dilution
in water were calculates as recommended by Brelvi and O’Connell (1972) and correlated as
a function of temperature.
5.3.3. Pitzer’s GE model for activity coefficients
In this paper, activity coefficients of both neutral and ionic species were calculated
using a modified Pitzer model for the excess Gibbs energy of aqueous electrolyte solutions
(Pitzer, 1973) (Eq. 5.15). Only the main equations of this model are recalled here. More
details can be found in our previous publication (Bougie and Iliuta, 2010b).
∑∑ ∑∑∑≠ ≠ ≠ ≠ ≠
++=w w
1 )( )( i wj i wj wk
ijkkjiijjiww
E
mmmImmIfMRTn
G τλ (5.15)
177
( )1f I is a modified Debye-Hückel term depending on ionic strength (I), temperature and
solvent (water) properties. ( )ij Iλ is the ionic strength dependent second virial coefficient:
[ ], ))1(1)(/2( )( 2)1()0( xijijij exxI −+−+= ββλ
(5.16)
where 2x I= .
The influence of temperature on the binary interaction parameters ( )0ijβ and ( )1
ijβ is
approximated by the relation:
/K or 1
0(1))0(
Tqqijij +=ββ
(5.17)
The ternary interaction parameters ijkτ are considered independent of temperature.
The equation for the activity coefficients of dissolved species follows from the
appropriate derivative of GE and water activity is calculated from the Gibbs-Duhem
equation:
( ) ( )
( )
, 2
1 22
2
2ln ln 1 1.2 21.21 1.2
1 1 32
mi i j ij
j w
jk xi j k j k ijk
j w k w j w k w
IA z I m II
xz m m x e m mIx
φγ λ
βτ
∗
≠
−
≠ ≠ ≠ ≠
= − + + + − +
− + + +
∑
∑∑ ∑∑
(5.18)
( ) ( )( )1.5
0 1ln 21 1.2
2
xw w i j ij ij
i w j w
w i j k ijk ii w j w k w i w
Ia M A m m eI
M m m m m
φ β β
τ
−
≠ ≠
≠ ≠ ≠ ≠
= − + −
+
+
∑∑
∑∑∑ ∑
(5.19)
5.3.3.1 Interaction parameters for the system CO2-Pz-H2O
Interaction parameters for the ternary system CO2-Pz-H2O were determined on the
basis of experimental data taken from the literature and from the present work, as it will be
explain in the section 5.4.1. In this system, eleven species are present in the liquid phase:
2CO , 3HCO− , 23CO − , Pz , +PzH , +2
2PzH , -PzCOO , 2- )Pz(COO , -COOPzH + , H+ and
OH− . Due to the very low concentration of H+ and OH− with respect to the other species,
their interactions with all other species were ignored and therefore, the corresponding
178
interaction parameters were set to zero. Based on the results of Derks et al. (2005b), all the
interaction parameters associated with +22PzH were also neglected. The second pKa of
piperazine, which is 5.3 at 298 K, is too low considering the pH range of interest for the
CO2 absorption. Therefore, +22PzH concentration is supposed to be very small and
interactions with this ion can be neglected. Another simplification can be made concerning
the 23CO −
interactions considering that the CO2 absorption decreases the pH, lowering
considerably the carbonate concentration (Derks et al, 2005b). In order to additionally
reduce the number of parameters, all binary and ternary interaction parameters involving
species with the same sign of charge were neglected. Only the parameters which were
found to have a significant influence on the liquid phase species distribution were
optimized based on the experimental data: )0(HCO,CO 32
−β , )0(PzH,CO2
+β , -2
(0)CO ,PzH COO
β + , )0(PzH,HCO-
3+β ,
)0(PzCOOPz, −β , )0(
COOPzH,PzH −++β ,
)0(PzCOO,PzH −+β ,
2
(0)PzH ,Pz(COO )
β + − , (0)PzH COO ,PzH COO
β + − + − , -3
(1)HCO ,PzH
β + and
(1)PzH COO ,PzH COO
β + − + − . Parameters 0q , 1q were fitted simultaneously to the selected
experimental data.
5.4. Results and discussions 5.4.1 CO2-Pz-H2O solubility database
In addition to obtain CO2 solubility data in piperazine aqueous solution at
temperature lower than 313 K, it was imperative to gather from literature all other reliable
solubility data for this system in order to get an interaction coefficient parameters set for
the model, able to cover large temperature, pressure and amine composition ranges. A
survey of the literature shown that five others independent research groups (Aroua and
Salleh, 2004; Bishnoi and Rochelle, 2000; Derks et al., 2005b; Kadiwala et al., 2010;
Kamps et al., 2003; Nguyen et al., 2010) reported solubility data for the CO2-Pz-H2O
system. A comparison of our data with some of these sources is made in Figure 5.1. In this
figure, good agreements between our data and those of two of these five independent
groups were found: Derks et al. (2005b) and Kamps et al. (2003), respectively for
piperazine molalities of 0.60 and 2.00 mol·kg-1 and at temperatures of 298.1 and 313.1 K.
179
However, quite large deviations appear between our data and those of Aroua and
Salleh (2004), as for example, for a piperazine molality of 0.60 mol·kg-1 and 303.1 K
(Figure 5.1). In general, data reported in that work are constantly right-shifted
comparatively to ours: for a given CO2 partial pressure, equilibrium loading given by
Aroua and Salleh (2004) is much higher. Similar disagreements were also reported by
Ermatchkov et al. (2006).
Concerning the two remaining independent sources, Rochelle’s research group
(Bishnoi and Rochelle, 2000; Nguyen et al., 2010) and Kadiwala et al. (2010), data of the
former were verified in Ermatchkov’s (2006) work and Kadiwala compared their data
against those of Kamps et al. (2003) and found good agreements. Therefore, all data from
these two sources were considered reliable and were added to the databank used in the
parameter regression. 354 data points were finally included in the databank. Table 5.2
summarizes the origin and the number of data used in this work the parameter estimation.
Table 5.2. Number of Reliable Data of CO2 (1) Solubility in Aqueous Solution of Piperazine (2) and their Source
5.4.2 Solubility measurements
CO2 solubility measurements were made, following the procedure described in
section 2.2, in aqueous piperazine solutions over a temperature range from T = (287.1 to
313.1) K and for amine concentrations from m = (0.10 to 2.00) mol.kg-1. The CO2 partial
pressure was kept within 2COP = (0.11 to 525.17) kPa using a VLE apparatus based on a
static-synthetic method. The validity of the apparatus and procedure were verified in our
Source N T m2 α - - K mol·kg-1 -
This work 64 287.1 to 313.1 0.10 to 2.00 0.10 to 2.68 Derks et al. (2005b) 58 298.1 to 343.1 0.2 to 0.64 0.36 to 1.23
Ermatchkov et al. (2006) 52 313.1 to 393.1 1.0 to 4.4 0.05 to 0.95 Kadiwala et al. (2010) 42 313.1 to 343.1 0.3 to 1.4 0.92 to 2.77
Bishnoi and Rochelle (2000) 17 313.1 to 343.1 0.64 0.16 to 0.96 Kamps et al. (2003) 92 314.1 to 395.1 2.00 to 3.96 0.50 to 1.64 Nguyen et al. (2010) 29 313.1 to 333.1 2.0 to 8.0 0.26 to 0.86
180
1.0
10.0
100.0
1000.0
10000.0
0.0 0.5 1.0 1.5 2.0
P/kP
a
α
previous work (Bougie and Iliuta, 2010b). All the results are shown in Tables 5.3 to 5.7
along with their experimental uncertainties. In these tables, y1 is the mole fraction of CO2 in
the gas phase.
Figure 5.1. Comparison of various solubility data of CO2 (1) in piperazine (2) aqueous solutions of concentration m2/mol·kg-1 at temperature T/K: ■, m2 = 0.60 and T = 298.1; □,
m2 = 0.60 and T = 298.1, (Derks et al., 2005b); ♦, m2 = 0.60 and T = 303.1; ◊, m2 = 0.60 and T = 303.1, (Aroua and Salleh, 2004); ▲, m2 = 2.00 and T = 313.1; ∆, m2 = 2.00 and T =
313.1, (Kamps et al. 2003); lines, model correlation.
Table 5.3. Solubility of CO2 (1) in Aqueous Solution of Piperazine (2) at T = 287.1 K (∆T = ± 0.1 K)
m2 ∆m2 α ∆α P ∆P y1 ∆y1
mol·kg-1 mol·kg-1 - - kPa kPa - - 0.10 0.002 1.37 0.12 68.29 0.05 0.978 0.002 0.10 0.002 1.89 0.31 181.1 0.1 0.992 0.002 0.10 0.002 2.68 0.57 338.7 0.3 0.995 0.002 0.50 0.0005 0.316 0.004 2.130 0.002 0.257 0.002 0.50 0.0005 0.65 0.01 2.822 0.002 0.441 0.002 0.50 0.0005 0.92 0.02 18.61 0.01 0.915 0.002 0.50 0.0005 1.05 0.06 93.90 0.08 0.983 0.002 1.00 0.0003 0.248 0.002 1.931 0.002 0.218 0.002 1.00 0.0003 0.563 0.003 2.581 0.002 0.418 0.002 1.00 0.0003 0.89 0.01 6.327 0.005 0.764 0.002 1.00 0.0003 1.05 0.03 86.55 0.07 0.983 0.002 1.00 0.0003 1.10 0.04 157.2 0.1 0.991 0.002
181
Table 5.4. Solubility of CO2 (1) in Aqueous Solution of Piperazine (2) at T = 293.1 K (∆T = ± 0.1 K)
m2 ∆m2 α ∆α P ∆P y1 ∆y1
mol·kg-1 mol·kg-1 - - kPa kPa - - 0.100 0.002 1.15 0.13 44.52 0.04 0.948 0.002 0.100 0.002 1.32 0.25 81.34 0.07 0.972 0.002 0.100 0.002 1.49 0.42 135.4 0.1 0.983 0.002 0.50 0.0005 0.421 0.004 2.504 0.002 0.079 0.002 0.50 0.0005 0.79 0.01 3.201 0.003 0.282 0.002 0.50 0.0005 1.02 0.03 33.66 0.03 0.932 0.002 0.50 0.0005 1.07 0.05 70.52 0.06 0.967 0.002 0.50 0.0005 1.17 0.10 161.7 0.1 0.986 0.002 1.09 0.0002 0.330 0.002 2.716 0.002 0.173 0.002 1.09 0.0002 0.619 0.004 3.261 0.003 0.315 0.002 1.09 0.0002 0.82 0.01 6.148 0.005 0.638 0.002 1.09 0.0002 0.93 0.01 33.28 0.03 0.933 0.002 1.09 0.0002 0.98 0.03 94.66 0.08 0.977 0.002 1.09 0.0002 1.04 0.06 195.0 0.2 0.989 0.002
Table 5.5. Solubility of CO2 (1) in Aqueous Solution of Piperazine (2) at T = 298.1 K (∆T = ± 0.1 K)
m2 ∆m2 α ∆α P ∆P y1 ∆y1
mol·kg-1 mol·kg-1 - - kPa kPa - - 0.10 0.002 0.45 0.01 3.256 0.003 0.034 0.002 0.10 0.002 1.00 0.05 13.93 0.01 0.775 0.002 0.10 0.002 1.12 0.11 41.15 0.03 0.924 0.002 0.10 0.002 1.24 0.19 73.40 0.06 0.957 0.002 0.63 0.0004 0.357 0.002 3.519 0.003 0.118 0.002 0.63 0.0004 0.665 0.005 4.049 0.003 0.236 0.002 0.63 0.0004 0.92 0.01 10.204 0.008 0.698 0.002 0.63 0.0004 1.01 0.02 42.81 0.03 0.928 0.002 0.63 0.0004 1.06 0.04 86.96 0.07 0.965 0.002 0.63 0.0004 1.10 0.07 150.8 0.1 0.980 0.002 1.00 0.0003 0.236 0.002 3.410 0.003 0.089 0.002 1.00 0.0003 0.494 0.003 3.793 0.003 0.184 0.002 1.00 0.0003 0.713 0.005 4.983 0.004 0.381 0.002 1.00 0.0003 0.87 0.01 22.00 0.02 0.860 0.002 1.00 0.0003 0.92 0.02 61.73 0.05 0.950 0.002 1.00 0.0003 0.95 0.03 102.30 0.08 0.970 0.002
182
Table 5.6. Solubility of CO2 (1) in Aqueous Solution of Piperazine (2) at T = 303.1 K (∆T = ± 0.1 K)
m2 ∆m2 α ∆α P ∆P y1 ∆y1
mol·kg-1 mol·kg-1 - - kPa kPa - - 0.10 0.002 0.66 0.02 4.485 0.004 0.058 0.002 0.10 0.002 1.00 0.05 14.49 0.01 0.708 0.002 0.10 0.002 1.09 0.12 37.46 0.03 0.887 0.002 0.10 0.002 1.18 0.19 63.87 0.05 0.934 0.002 0.10 0.002 1.33 0.29 104.67 0.08 0.960 0.002 0.63 0.0004 0.298 0.002 4.540 0.004 0.080 0.002 0.63 0.0004 0.596 0.005 5.202 0.004 0.200 0.002 0.63 0.0004 0.85 0.01 21.75 0.02 0.809 0.002 0.63 0.0004 0.91 0.03 73.04 0.06 0.943 0.002 0.63 0.0004 0.93 0.04 104.74 0.08 0.960 0.002 1.00 0.0003 0.193 0.002 4.769 0.004 0.128 0.002 1.00 0.0003 0.428 0.003 5.516 0.004 0.249 0.002 1.00 0.0003 0.65 0.01 6.561 0.005 0.371 0.002 1.00 0.0003 0.86 0.01 12.95 0.01 0.683 0.002 1.00 0.0003 0.97 0.03 74.28 0.06 0.945 0.002
Table 5.7. Solubility of CO2 (1) in Aqueous Solution of Piperazine (2) at T = 313.1 K (∆T = ± 0.1 K)
m2 ∆m2 α ∆α P ∆P y1 ∆y1
mol·kg-1 mol·kg-1 - - kPa kPa - - 2.00 0.0001 0.097 0.001 9.34 0.01 0.238 0.002 2.00 0.0001 0.247 0.002 12.81 0.01 0.447 0.002 2.00 0.0001 0.436 0.002 17.26 0.01 0.592 0.002 2.00 0.0001 0.671 0.003 20.79 0.02 0.664 0.002 2.00 0.0001 0.907 0.005 33.17 0.03 0.791 0.002 2.00 0.0001 1.08 0.02 229.0 0.2 0.970 0.002 2.00 0.0001 1.16 0.04 532.0 0.4 0.987 0.002
In Figure 5.2 as well from the Tables 5.3 to 5.7 and considering the uncertainties, it
can be shown that at a constant amine concentrations and CO2 partial pressure an increase
in temperature leads to a decrease of the CO2 loading capacity. Furthermore, as expected
and also observed in other works (Speyer et al., 2010; Vahidi et al., 2009; Yang et al.,
2010), at a fixed temperature, an increase in piperazine concentration leads to a decrease of
the solution CO2 loading.
183
1.0
10.0
100.0
1 000.0
0.0 0.5 1.0 1.5
P/kP
a
α
The equilibrium CO2 partial pressure increases at first very slowly with respect to
the loading during the chemical absorption when the gas is mostly dissolved in non-volatile
ionic form. Then, at a loading near the unity, all further CO2 absorption in the equilibrium
cell can be related to physical absorption: the pressure increases sharply as the loading
increases.
Figure 5.2. Equilibrium pressure above aqueous solutions of CO2 (1) - piperazine (2) at concentration m2/mol·kg-1 and temperature T/K as a function of solution CO2 loading (α): ■, m2 = 0.10 and T = 293.1; ♦, m2 = 0.50 and T = 293.1; ▲, m2 = 1.09 and T = 293.1; ×, m2 = 1.00 and T = 298.1; lines, model correlation.
5.4.3 Modeling results
All 354 selected experimental data for the system CO2-Pz-H2O, covering a large
range of amine concentrations, temperature and solution loading were correlated together
with our regressed set of interaction parameter (Table 5.8) with a pressure average relative
deviation of 26.1%. This percentage is quite satisfying, taking into account the wide range
of amine concentrations, temperature and solution loading in the solubility database
considered for parameters estimation. Generally, higher deviations were obtained at very
large amine concentration and very high ionic strengths or at very low solution loading
(very low CO2 true molality in the liquid phase). In these regions, the addition of more
ionic strength dependent parameters and some amine-water interaction parameters (Chang
et al., 1993; Ermatchkov et al., 2003) might lead to the increase of the modeling accuracy.
184
Figures 5.1 and 5.2 show a comparison between some of our data and the model
correlation.
Table 5.8. Interaction Parameters in Pitzer's GE Equation for the Ternary CO2-Pz-H2O System as in Eq. (5.17) for a Temperature range of 287.1 K to 395.1 K.
parameters q0 q1
)0(HCO,CO 32
−β 5.8194 -2201.7 )0(
PzH,CO2+β
-5.2153 1911.2
)0(COOPzH,CO2
−+β
-0.3542 184.13
)0(HCO,PzH 3
−+β
0.4900 -179.00
)1(HCO,PzH 3
−+β
2.5000 -870.92 (0)Pz,PzCOO
β − 0.1500 -21.00
)0(PzCOO,PzH −+β
0.0200 55.344
)0()Pz(COO,PzH 2
−+β
5.0011 -1480.00
)0(COOPzH ,PzH −++β
-1.7999 580.00
)0(COOPzH ,COOPzH - −++β
0.4001 -221.99
)1(COOPzH ,COOPzH - −++β
2.2000 -580.00
Based on the equilibrium model, an example of the speciation of several ions and of
their activity coefficient in an aqueous Pz solution is shown respectively in Figures 5.3 and
5.4. The concentration profiles are represented as a function of the CO2 loading for a Pz
aqueous solution with a concentration of 1.00 mol·kg-1 at 298.1 K. Because H+ , OH− , and 22PzH +
concentrations are much lower than the concentrations of all other species, the
corresponding curves were not represented. Piperazine concentration decreases rapidly at
CO2 loadings up to about 1 mol/mol of amine, while the protonated amine PzH+ and the
amine carbamate PzCOO− concentrations show a fast increase. When the CO2
stoichiometric concentration is less than that of piperazine, CO2 is practically completely
chemically dissolved (this is observed on the entire CO2 loading range shown) and mainly
converted into piperazine carbamate, piperazine dicarbamate and protonated piperazine
185
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0.0 0.2 0.4 0.6 0.8 1.0
mi /m
ol·k
g-1
α
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0.2 0.4 0.6 0.8 1.0
γ i
α
species. The same behaviour was observed for other amine concentrations (Derks et al.,
2005b; Kamps et al. 2003).
Figure 5.3. Species distribution in the aqueous CO2 (1) – Pz (2) system at 298.1 K (m2/mol·kg-1 = 1.00) as a function of solution CO2 loading: ____ Pz; __ .. __ +PzH ; - - -
-PzCOO ; ..... -COOPzH + ; __ . __ 2
- )Pz(COO ; - . - 3HCO− ; __ __ CO2.
Figure 5.4. Calculated activity coefficients in the aqueous CO2 (1) – Pz (2) system at 298.1 K (m2/mol·kg-1 = 1.00) as a function of solution CO2 loading: ____ Pz; __ .. __ +PzH ; - - -
-PzCOO ; ..... -COOPzH + ; __ . __ 2
- )Pz(COO ; - . - 3HCO− ; __ __ CO2.
186
5.5. Conclusions In the present work, new data concerning the solubility of CO2 in aqueous
piperazine (Pz) solutions were obtained for a temperature range of T = (287.1 to 313.1) K
and for amine concentrations from m = (0.10 to 2.00) mol.kg-1. The CO2 partial pressure
was kept within 2COP = (0.11 to 525.17) kPa using a VLE apparatus based on a static-
synthetic method). Based on these experimental data and from selected data from literature,
354 data for ternary system CO2-Pz-H2O were satisfactorily correlated with an mean
average deviation of 26.1 % using a modified Pitzer’s thermodynamic model for the
activity coefficients combined with the virial equation of state for representing the fugacity
coefficients. A new set of interaction parameters for this system was found in this work in
order to cover a wider range of temperature, pressure and amine concentration.
187
188
Knowledge about the regeneration of loaded (CO2 containing) amine solutions are
essential for economic viability of the absorption/desorption processes. To represent an
interesting absorbent for CO2 separation, the aqueous AHPD +Pz solution studied in the
previous chapters from the point of view of solubility and kinetics of CO2 absorption,
should also have appropriate facility in the regeneration step. In this chapter, we therefore
compared the regeneration capability of different single SHA or Pz-activated aqueous
solutions with that of single MEA aqueous solution (the most used amine in industrial
applications).
189
Chapter 6. Analysis of regeneration of sterically hindered alkanolamines aqueous solutions with and without activator
Résumé
Dans cette étude, la capacité de régénération de différentes solutions aqueouses d’amines à encombrement stériques, seules (SHA: AMP, AEPD, AMPD, AHPD) ou activées par l’ajout de la Pz, a été comparée avec celle d’une solution aqueuse de MEA. Les résultats ont montré une meilleure capacité de régénération des amines AEPD, AMPD and AHPD, par rapport aux alcanolamines conventionnelles (MEA). L’ajout de petites quantités de Pz aux solutions aqueuses d’AHPD a une influence positive sur les performances de la solution.
190
Abstract
This work concerns the comparison of the regeneration capability of different single sterically hindered alkanolamines (SHA: AMP, AEPD, AMPD, AHPD) or Pz-activated aqueous solutions with that of single MEA aqueous solution. It was found that AEPD, AMPD and AHPD offer an easier and faster regeneration than conventional alkanolamines (MEA). Small additions of Pz to single AHPD aqueous solutions were found to have a beneficial influence on the solution performances.
191
6.1 Introduction Many industrial processes (e.g. chemical and petrochemical, steel, aluminum and
cement production) annually release a large amount of CO2 into the atmosphere. In almost
all cases, methods used by industries to reduce or eliminate emanations of this greenhouse
gas consist of its removal by chemical absorption/desorption processes with alkanolamine-
based aqueous solutions in which the amines are regenerated to be reused (Kohl and
Nielsen, 1997). Compared to the extensive number of studies on CO2 absorption in the
open literature, there are relatively few information related to CO2 thermal desorption
processes despite the fact that the stripping unit is usually highly energy-consuming and it
is responsible for the main operational cost of the process (Tobiesen and Svendsen, 2006).
For that reason, amine solutions with low regeneration cost are essential for economic
viability of the absorption/desorption processes.
It is recognized that the presence of carbamates influences the regeneration
efficiency of alkanolamine solutions. The stable carbamates are difficult to revert to fresh
amine, leading therefore to longer regeneration time and more energy consuming
(Sakwattanapong et al., 2005). In comparison to conventional primary and secondary
alkanolamines like monoethanolamine (MEA) and diethanolamine (DEA), sterically
hindered alkanolamines (SHA) (e.g. 2-amino-2-methylpropanol - AMP) form unstable
carbamates due to the hindrance of the bulky group adjacent to the amino group (Sartori
and Savage, 1983). Hydrolysis of the voluminous carbamates leads to a preferential
bicarbonate formation process and it is expected that a solution containing a greater
proportion of bicarbonate undergoes desorption at a greater rate (requiring less energy) and
produce a lean solution containing less physically and chemically absorbed CO2 (Hook,
1997; Sartori and Savage, 1983; Tontiwachwuthikul et al., 1991).
In our laboratory, extensive studies concerning the CO2 capture in membrane
contactors using SHA based alkanolamines mixtures are in progress. In order to study the
hindrance effect on the absorption ability of SHA, a set of four SHA was chosen. It
concerns AMP, the simple hindrance form of MEA, and three SHA derived from AMP: 2-
amino-2-methyl-1,3-propanediol (AMPD), 2-amino-2-ethyl-1,3-propanediol (AEPD) and
192
2-amino-2-hydroxymethyl-1,3-propanediol (AHPD). The kinetic and the thermodynamic
characterisation of the CO2 absorption into aqueous solutions of these SHA has been
discussed previously in the literature (Baek and Yoon, 1998; Bougie and Iliuta, 2009)
(Bougie and Iliuta, 2010b; Park et al., 2002b; Teng and Mather, 1989; Yih and Shen, 1988;
Yoon et al., 2003; Yoon et al., 2002a), as well as the influence of the addition of Pz
(piperazine) as activator in AMP and AHPD solutions (Bougie and Iliuta, 2010b; Bougie et
al., 2009; Choi et al., 2007). Nevertheless, as mentioned earlier, very few information are
available about the CO2 stripping efficiencies of these alkanolamine aqueous solutions. To
our knowledge, except for single AMP aqueous solutions (e.g. (Hook, 1997; Zhang et al.,
2008)), no information were found in the open literature concerning the regeneration of
aqueous solutions of the other investigated SHA with or without activator.
The main objective of this work is to compare the regeneration capability of
different single SHA or Pz-activated aqueous solutions with that of single MEA aqueous
solution (the most used amine in industrial applications). This research was then intended to
verify the assumptions that i) SHA offer an easier and faster regeneration than conventional
alkanolamines (e.g. MEA – monoethanolamine) and ii) small additions of activator to
single SHA aqueous solutions do not impede the desorption performance.
6.2. Material and methods 6.2.1 Reagents
Aqueous amines solutions were prepared with degassed distilled water and either
one or two of the following amines: 2-amino-2-methyl-1-propanol (AMP), 2-amino-2-
methyl-1,3-propanediol (AMPD), 2-amino-2-ethyl-1,3-propanediol (AEPD), 2-amino-2-
hydroxymethyl-1,3-propanediol (AHPD), piperazine (Pz) or monoethanolamine (MEA).
The amines (from Laboratoire MAT, Quebec, Canada, except for MEA from Aldrich) had
respectively a minimum purity of 95, 99, 97, 99.9, 99 and 99% and were used without
further purification. CO2 and N2 gases were of commercial grade with a minimum purity of
99.5 % and were supplied by Praxair.
193
6.2.2 Apparatus and procedures
In order to study the regenerative capacity of amines solutions, two different
experimental setups were used in this work: an absorption flask for the CO2 absorption and
a liquid-vapor equilibrium cell where the regeneration took place. A schematic diagram of
the absorption flask is shown in Figure 6.1a. It consists mainly of a thermostated and
magnetically agitated 500.0 × 10-6 m3 Pyrex flask where 250.0 × 10-6 m3 of amine solution
was put into contact with pure CO2 at a constant pressure of 120 kPa (uncertainty of ± 5
kPa) and at a saturation temperature of 303.2 K (uncertainty of ± 0.2 K). After a fixed
absorption time, 10.00 × 10-6 m3 sample was analysed with the barium chloride
precipitation method (Ma'mun et al., 2006) to determine the total CO2 content of the
solution. The rich solution was then transferred into a vapor-liquid equilibrium cell to
perform the regeneration. This equilibrium cell (Iliuta and Thyrion, 1995) shown in Figure.
6.1b, was magnetically agitated and heated by a 200 watts cartridge electric heater
(Chromalox CIR-2051, Omega). The cell was kept at atmospheric pressure and in order to
avoid water losses, two condensers in series were installed. When heat was supplied to the
saturated solution, CO2 was released and it was possible to calculate its desorption rate by
means of the gas chromatographic technique. A small and well known nitrogen reference
flow was sent to the equilibrium cell and the exit gaseous mixture (N2 + released CO2) was
analysed on-line every two minutes by a gas chromatograph (Micro GC 3000A, Agilent
Technologies) to measure the N2 volumetric percentage. Based on the nitrogen flow rate, its
percentage in the total flow, the temperature and the pressure, the instantaneous CO2 flow
rate and the CO2 desorbed mole number were then calculated. After the regeneration time, a
sample of the lean solution was analysed by the barium chloride precipitation method to
determine the total residual CO2 concentration. To ensure reliable results, all analysis were
accompanied by a blank duplicate to take into account the presence of dissolved CO2 in the
reagents.
Two independent methods were used to calculate the cyclic absorption capacity of
solutions: i) subtract the lean CO2 concentration from the rich CO2 concentration, on the
basis of the precipitation method and, ii) calculate the total CO2 mole number desorbed
from the solution by Micro GC analysis and link it with the solution volume and amine
194
concentration. In this work, the difference between the results obtained by these two
methods was never more than 4%; a mean value was therefore considered.
The absorption/regeneration cycles were performed in the following aqueous
solution concentration (uncertainty of ± 0.01 kmol.m-3) and regeneration temperature
(uncertainty of ± 0.1 K) conditions: (i) 1.00 kmol.m-3 AHPD for a regeneration temperature
between 353.2 and 393.2 K and (ii) 1.00 kmol.m-3 amine (AEPD, AMPD, AMP, AHPD or
Pz), 2.00 kmol.m-3 MEA and 0.90 kmol.m-3 AHPD + 0.10 kmol.m-3 Pz for a regeneration
temperature of 383.2 K.
a) b)
Figure 6.1. a) Schematic diagram of the absorption flask. A: thermostated tank; B: agitated saturation flask; b) Schematic diagram of the vapor-liquid equilibrium cell used for the regeneration. A: saturator; B: vapor-liquid equilibrium cell; C: bubble flowmeter; D: microGC.
6.3. Results and discussion 6.3.1 Analysis of the regeneration time and temperature
To find the optimal regeneration time and temperature, a first set of experiments
was performed for an aqueous solution of 1.00 kmol.m-3 AHPD and for a regeneration
temperature between 353.2 and 393.2 K.
195
To verify the accuracy of the barium chloride precipitation method, the amine
solutions were saturated at 120 kPa and 303.2 K and the CO2 concentrations in the rich
solutions determined experimentally with the precipitation method were compared to the
values obtained using a different method (static vapor-liquid equilibria) (Bougie and Iliuta,
2010b). For the five experiments performed (step of 10 K between 353.2 and 393.2 K), a
mean value of 0.88 kmol CO2 / kmol AHPD was obtained (loading uncertainty of ± 0.02
kmol CO2 / kmol AHPD). This value was found to be in the expected range (Bougie and
Iliuta, 2010b): between 0.785 kmol CO2 / kmol AHPD (303.15 K, 1.69 kmol.m-3 AHPD)
and 0.945 kmol CO2 / kmol AHPD (298.15 K and 0.85 kmol.m-3 AHPD). As expected, the
loading increases when the amine concentration decreases at the same temperature and the
loading decreases when the temperature and/or the amine concentration increase. In
addition, the method was also verified by analysing different aqueous K2CO3 solutions of
precise concentrations.
The saturated solutions were transferred into the vapor-liquid equilibrium cell for
regeneration and heated by the electric heater whose surface temperature was fixed and
designated as the regeneration temperature. To determine the optimum temperature, the
regeneration efficiency is used as a basis of comparison and its definition is given by Eq.
(6.1):
- = 100% , R L
R
α αηα
⋅ (6.1)
where αR and αL are respectively the rich and the lean CO2 loading in solutions. The results
are shown in Figure 6.2 where the efficiency divided by the regeneration time was plotted
against the regeneration temperature. It is possible to see that around 383.2 K, the curve
flattens indicating that increasing the temperature additionally does not create a significant
increase of the efficiency for a constant regeneration time. 383.2 K was then considered as
the optimum regeneration temperature. This result is similar to the temperature found for
AMP by Zhang et al. (2008).
Using this optimum regeneration temperature, the regeneration time used in the
subsequent experiments was selected as the time necessary to reach on average 80% of the
196
0
10
20
30
40
50
60
70
80
90
100
0 100 200 300
100
-η
Regeneration time (min)
Desorption curveAHPD, 1.00 kmol.m-3
T = 383.2 K
0.20
0.25
0.30
0.35
0.40
0.45
0.50
350 360 370 380 390 400
Eff
icie
ncy
/ reg
. tim
e (η
/t) /
(min
-1)
Regeneration temperature (K)
optimum efficiency, considering that it is not advantageous to complete the regenerations
until full CO2 desorption. As it can be seen in Figure 6.3, which represents a standard
desorption curve obtained for AHPD at 383.2 K, the second half time of the desorption
process do not increase significantly the regeneration efficiency (it is equivalent to about
20% of the optimum efficiency), although heat is still supplied to the solution. The
regeneration time (155.0 minutes) and temperature (383.2 K) have then been used for the
following regeneration experiments.
Figure 6.2. Optimal regeneration temperature determination (the curve shows the trend).
Figure 6.3. Standard desorption curve for a 1 kmol.m-3 aqueous AHPD solution at 383.15 K.
197
6.3.2 Amine influence on regeneration efficiency
A second set of experiments was performed using the regeneration time and
temperature established in the section 6.3.1, for the following aqueous solutions: 1.00
kmol.m-3 amine (AEPD, AMPD, AMP, AHPD or Pz), and 2.00 kmol.m-3 MEA. These
concentrations were selected in order to compare the regeneration of solutions having
almost the same chemical loading of 1 kmol CO2 / m3 of solution given by the
stoichiometry of the absorption reaction: theoretically ½ kmol of CO2/ kmol of MEA or Pz
(diamine) and 1 kmol of CO2/ kmol of each SHA. Furthermore, a fixed absorption time of
360.0 minutes (uncertainty of ± 0.5 minute) was preset for all amines before regeneration.
Therefore, all the experimental conditions for the absorption and regeneration were kept the
same for all amine solutions. The results obtained after three absorption/regeneration cycles
for each amine solution, are indicated in Table 6.1 and interesting observations can be
made.
Table 6.1. Regeneration efficiency of various amines
Amines αR αR - αL η (kmol/m3) (kmol/m3) (-)
AEPD 0.93 0.56 60.2 AHPD 0.77 0.58 76.0 AMP 0.98 0.34 34.8
AMPD 0.97 0.61 62.6 MEA 1.04 0.46 43.9
Pz 1.01 0.43 42.3
The loading of the rich amine solutions reached a value near the theoretical loading
of 1 kmol CO2 /m3 solution, except for the AHPD solutions. This indicates that for AHPD,
360.0 minutes of absorption are not enough to achieve full saturation of the 250.0 × 10-6 m3
solution used in this work. For the tested SHA, the rich loading concentration might be
classified as follows: 0.98 (AMP) ≥ 0.97 (AMPD) > 0.93 (AEPD) > 0.77 (AHPD) kmol
CO2 / kmol amine. This ranking is due to a combination of kinetics, thermodynamics and
steric hindrance of each amine.
Concerning the cyclic capacity (αR – αL) depicting the true net CO2 mol removal
per m3 of solution after each absorption/regeneration cycle, the results can be classified as
198
follows: AMPD (0.61) ≥ AHPD (0.58) ≥ AEPD (0.56) > MEA (0.46) ≥ Pz (0.43) > AMP
(0.34). In addition to the regeneration efficiency ranking: AHPD (76.0) >> AMPD (62.6) ≥
AEPD (60.2) > MEA (43.9) ≥ Pz (42.3) > AMP (34.8), these results demonstrate clearly
that the three most hindered amine solutions (AHPD, AMPD and AEPD), and in particular
AHPD, are more easy to regenerate because they do not form (or form very few) stable
carbamates in solution. Similar conclusions were found in the literature for AHPD systems
(Bougie and Iliuta, 2010b; Park et al., 2003). The MEA and Pz solutions gave comparable
results, as expected from amines that form high proportions of stable carbamates (Derks et
al., 2005b; Park et al., 2003). However, the results obtained for AMP solutions shows that
the calculated cyclic capacity and the regeneration efficiency are the lowest of all tested
amines. It is possible that part of the CO2 released from the bicarbonate decomposition
during the beginning of the regeneration could react again with the free amine molecules
formed in the solution due to the increase of the solution pH. This behaviour might be more
important for amines that form an important bicarbonate amount in solution and possess
higher kinetics, like AMP. However, this is not the case for the other SHA amines, even if
they form preferentially bicarbonate in solution, because of the relatively low kinetics in
comparison to AMP. On the contrary, this is less evident for MEA and Pz because most of
the amine molecules exist as stable carbamates in the rich solutions and the amount of
molecules that might react again is low.
Additional experimental data were performed for AMP solution in different
absorption & regeneration conditions than those established for the second set of
experiments, than means: absorption for 180.0 minutes and regeneration at 383.2 K for
155.0 minutes. As expected, the decrease of the absorption time for the AMP solutions
leads to the decrease of the values of the loading of the rich solutions (0.93). It was
observed that AMP solutions gave almost the same, and the lowest, cyclic capacity (0.37)
and regeneration efficiency (39.4); that validates the results given in Table 6.1. As
expected, the decrease of the absorption time for the AHPD and AMP solutions leads to the
decrease of the values of the loading of the rich solutions (Table 6.2).
199
40
50
60
70
80
90
100
0 50 100 150
100
-η
Regeneration time (min)
Desorption curveT = 383.2 K
Pz, 1.00 kmol/m³MEA, 2.00 kmol/m³
Table 6.2. Regeneration of AHPD with or without Pz
System Concentration αR αR - αL η Absorption time (kmol.m-3) (kmol/kmol) (kmol/kmol) (-) (min)
AHPD 1.00 0.77 0.58 76.0 360.0 AHPD 1.00 0.65 0.48 73.9 180.0
AHPD + Pz 0.90 + 0.10 0.74 0.56 76.4 180.0
6.3.3 Effect of activator addition on regeneration efficiency
To use the advantage of different types of amines (primary, secondary, tertiary and
SHA), blended amines are usually proposed in the literature (Dang and Rochelle, 2003; Xu
et al., 1992a). Activated alkanolamine solutions combine the advantage of the fast
reactivity of the activator molecules and that of the easiness of the regeneration of tertiary
or SHA amines. The choice of the activator is crucial for industrial applications and the
most common used amines are MEA and Pz. The results given in the section 6.3.2 show
that these two amines possess almost the same cyclic capacity and regeneration efficiency.
However, it is shown in Figure 6.4 that Pz regeneration is faster than that of MEA; the Pz
curve lies under the MEA’s one. Furthermore, on the kinetic side of view, it is well
established that Pz have kinetic constants an order of magnitude higher than MEA. Pz
seems therefore to be a better activator than MEA.
Figure 6.4. Comparison of desorption curves of MEA and Pz.
200
0
10
20
30
40
50
60
70
80
90
100
0 50 100 150
100
-η
Regeneration time (min)
Desorption curveT = 383.2 K
AHPD, 1.00 kmol/m³
AHPD, 0.90 kmol/m³ + Pz, 0.10 kmol/m³
In order to study the effect of small additions of activator to single SHA aqueous
solutions, a third set of experiments combining the SHA possessing the best cyclic capacity
and efficiency (AHPD) with the best activator (Pz) were performed at the same
regeneration temperature (383.2 K) and time (155.0 minutes), but for an absorption time of
180.0 minutes. Reducing the absorption time allowed us to observe better the effect of the
activator and to compare the new results with the previous experiments (section 6.3.2). The
results are indicated in Table 6.2 and Figure 6.5. As expected, the decrease of the
absorption time for the AHPD solutions leads to the decrease of the values of the loading of
the rich solutions (Table 6.2).
Keeping the same total amine concentration but replacing 0.10 kmol.m-3 of AHPD
by Pz leads to an increase of the loading of the rich solutions related to an enhancement of
the kinetics. Moreover, it is very interesting to note an increase of the cyclic capacity. For
all experiments performed with single or activated AHPD solutions, the regeneration
efficiency was found to be practically stable and Figure 6.5 illustrates the same desorption
pattern between those two aqueous solution. In summary, the addition of a small amount of
Pz into AHPD aqueous solution allowed to obtain almost the same cyclic capacity and
regeneration efficiency as non-activated solutions but for half of the absorption time.
Figure 6.5. Effect of Pz on desorption of AHPD aqueous solutions.
201
6.4. Conclusions Taken together, the results of this work have revealed that the regeneration
efficiency can be classified as following: AHPD (76.0) >> AMPD (62.6) ≥ AEPD (60.2) >
MEA (43.9) ≥ Pz (42.3) > AMP (34.8). These results demonstrate clearly that the three
most hindered amine solutions (AHPD, AMPD and AEPD), and in particular AHPD, are
more easy to regenerate because they do not form (or form very few) stable carbamates in
solution. However, the results obtained for AMP solutions show that the calculated cyclic
capacity and the regeneration efficiency are the lowest of all tested amines. The use of Pz
as activator seems to offer advantages over MEA. Finally, it was found that the addition of
a small amount of Pz into AHPD aqueous solution allowed to obtain almost the same cyclic
capacity and regeneration efficiency as non-activated solutions but for half of the
absorption time. In conclusion, based on the present study and for economic considerations
(the prices for the three best SHA are 0.06, 0.22 and 0.57 US$/g, respectively for AHPD,
AEPD and AMPD) and amine availability, the mixture AHPD-Pz seems to be the most
appropriate solvent for CO2 capture.
202
Besides the liquid absorbent properties, the performances of MC for CO2 separation
strongly depend on the compatibility between liquid and membrane. In the following
chapter, based on wetting-related properties like liquid surface tension, contact angle,
membrane breakthrough pressure and chemical stability, a thorough analysis of these
properties on different potential membrane/liquid combinations is performed in order to
develop an appropriate way to select the best conditions to elude the unwanted wetting
phenomenon.
203
Chapter 7. Analysis of Laplace-Young equation parameters and their influence on efficient CO2 capture in membrane contactors
Résumé
Sur la base des propriétés liées au mouillage des membranes, comme la tension superficielle du liquide, l’angle de contact, la pression de percée et la stabilité chimique, une analyse approfondie de l’effet de ces propriétés sur différentes combinaisons membrane/liquide a été réalisée afin de développer un moyen approprié pour sélectionner les meilleures conditions qui permettraient d’éviter le phénomène de mouillage dans les contacteurs à membrane (MC). Une étude systématique de la littérature combinée à de nouvelles données expérimentales ont permis d’obtenir des résultats intéressants. Tout d'abord, une nouvelle méthode très simple a été développée pour estimer la tension superficielle des solutions aqueuses d'amines, d'alcools ou d’alcanolamines. Deuxièmement, en plus de polytétrafluoroéthylène (PTFE) et polypropylène (PP) (dans une moindre mesure), les membranes PTFE/PP laminées se sont avérées une alternative intéressante à considérer dans la contacteurs à plaques en raison des valeurs plus élevées de l'angle de contact. Finalement, un nouveau critère de performance à long terme de l'absorption des gaz dans les MC a été proposé. On estime qu'un rapport entre la surpression du liquide dans le MC et la pression de percée nominale d’au moins 1,5% pourrait être un critère utile pour éviter le mouillage de la membrane. Dans ce contexte, plusieurs actions spécifiques à respecter lors de l’opération des MC ont été proposées.
204
Abstract
Based on wetting-related properties like liquid surface tension, contact angle, membrane breakthrough pressure and chemical stability, this work aims to perform a thorough analysis of these properties on different potential membrane/liquid combinations in order to develop an appropriate way to select the best conditions to elude the unwanted wetting phenomenon in membrane contactors (MC). From new experimental data and a systematic review from literature, several significant results were obtained. First, a new and very simple classification method for the estimation of surface tension of aqueous amine, alcohol or alkanolamine solutions was developed. Second, in addition to polytetrafluoroethylene (PTFE) and polypropylene (PP) (to a lesser extend), laminated PTFE/PP membranes were found to be an interesting alternative to be considered in plate MC because the lamination process leads to higher contact angle values. Finally, a new criterion for long-term performance of gas absorption in MC was proposed. It was estimated that a ratio between the liquid overpressure and the nominal breakthrough pressure less than 1.5% could be a useful criterion to prevent membrane wetting and several actions were suggested to respect it.
205
7.1. Introduction The removal of acid gases, such as CO2, from industrial gases is frequently carried
out by an absorption-desorption process using alkanolamine aqueous solutions as liquid
absorbents. Blends of an activator (usually a primary or secondary amine) with a tertiary or
sterically hindered amine combine the higher rate of reaction with CO2 of the former with
the lower reaction heat of the latter, thereby achieving higher rates of absorption in the
absorption unit while requiring less regeneration energy in the stripper unit (Bougie and
Iliuta, 2012).
The gas absorption process for CO2 absorption can be carried out in different
reactors, such as bubble columns, sieve trays, packed towers, and venture scrubbers.
Among various techniques for CO2 capture, the membrane contactor (MC) process has
become one of the research focuses because of various advantages over the traditional gas
absorption processes: (i) large contact area for promoting an efficient gas-liquid mass
transfer, (ii) high modularity and compatibility for an easy scale-up, (iii) the possibility of
varying fluid flow rates independently and without the occurrence of loading or flooding,
and (iv) less interaction between the absorbent solution and the oxygen contained in the
flue gas that cause amine oxidative degradation (Gabelman and Hwang, 1999). However,
as the contact between the gas and the liquid is made after the gas diffuses through the
membrane pores, an additional membrane mass transfer resistance is added. Pores can be
gas filled (non-wetting conditions) or be partially or fully liquid filled (wetting conditions).
As the CO2 diffusion coefficient in the gas phase is much higher than in the aqueous phase,
the non-wetted mode is preferred to get the highest absorption flux. For example,
simulation results by Wang et al. (2005) showed that the CO2 absorption rate was six times
higher in the non-wetted mode than in the wetted one. Wetting phenomena of porous
membranes by liquid absorbents is therefore considered as the major problem in MC,
reducing their performances and restricting their industrial applications.
Membrane wettability strongly depends on the properties of both absorption liquid
and membrane and on the compatibility between them (El-Naas et al., 2010). The Laplace-
Young equation (Eq.7.1), used to calculate the minimum liquid overpressure against the gas
206
phase (breakthrough pressure, ∆PB.P.) required for the liquid to instantly penetrate into the
membrane pores, links some important properties like the surface tension of the solution
(σ), the solution/membrane contact angle (θ), and the maximum membrane pore size
diameter (dmax).
maxB.P.
cos 4- d
P θσ=∆ (7.1)
In the literature, this equation is mainly used to calculate the breakthrough pressure,
for ensuring the operation of MC at a lower liquid pressure (Atchariyawut et al., 2007).
This should theoretically guarantee operations in the non-wetted mode. However and
unfortunately, the opposite is often observed. The existence of membrane wetting is either
proved by modeling, by calculating the membrane mass transfer coefficient on the basis of
experimental data (Keshavarz et al., 2008) or assumed on the basis of the decrease of
absorption performance in time (Lin et al., 2009c).
Based on the Laplace-Young equation and wetting-related properties like liquid
surface tension, contact angle, membrane breakthrough pressure and chemical stability for
several aqueous amine solutions (MEA (monoethanolamine), AMP (2-amino-2-methyl-1-
propanol), AHPD (2-amino-2-hydroxymethyl-1,3-propanediol), Pz (piperazine) and AHPD
+ Pz) and polymeric flat membranes (PTFE (polytetrafluoroethylene), PVDF
(polyvinylidene fluoride), PP (polypropylene), and laminated PTFE/PP and PP/PP), this
work aims to perform a thorough analysis of these properties on different potential
membrane/liquid combinations in order to develop an appropriate way to select the best
conditions to elude the unwanted wetting phenomenon in membrane contactors. New
experimental data were therefore combined with a very systematic review concerning
surface tension of aqueous alcohol, amines and alkanolamine solutions, as well as
membrane contactor operation.
7.2. Experimental 7.2.1 Reagents
Aqueous amines solutions were prepared by gravimetric method using distilled
water and either one or two of the following amines: 2-amino-2-methyl-1-propanol (AMP,
207
CAS No. 124-68-5), 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD, CAS No. 77-86-
1), piperazine (Pz, CAS No. 110-85-0) and monoethanolamine (MEA, CAS No. 141-43-5).
The amines (from Laboratoire MAT, Quebec, Canada, except for MEA from Aldrich) had
respectively a minimum purity of (95, 99.9, 99 and 99)% and were used without further
purification. A Mettler AE240 balance with a precision of ±1×10-4 g was used to prepare
the solutions and it was calculated that the uncertainties of the reported concentrations were
less than 0.1 wt.%.
Several PP, PVDF or PTFE based commercial membranes were used in this study.
In addition to porous hydrophobic membranes fabricated of single polymeric material (PP
or PTFE), four laminated membranes were also tested (PTFE/PP and PP/PP). Laminated
membranes combine two layers: one represents a typical porous membrane, while the other
is made of microfibers used as support to stiffen the whole assembly, offering a better
mechanical resistance. For example, in the PTFE/PP membrane, PTFE represents the
membrane which is used in contact with the absorption solution and PP is the supporting
layer. Supporting layers are constituted by microfibers that could be woven (structurally
well-arranged like a wire mesh) or non-woven (randomly assembled). A summary of
membranes characteristics is given in Table 7.1.
Table 7.1. Characteristics of membranes used in this work.
Company Membrane Thickness (µm)
Nominal pore diameter (µm)
Designation
Donaldson PTFE 127 0.1 PTFE 1 PTFE 203 0.1 PTFE 2
AY Tech LLC
PTFE 25 0.25 PTFE 3
GE PVDF 140 - 250 0.45 PVDF PP 75 - 111 0.1 PP 1
Pall PTFE/PP woven 178 - 246 0.2 PTFE 4 PTFE/PP non-woven 178 - 279 0.2 PTFE 5
Membrana PP 100 ± 15 0.1 PP 2 PP/PP non-woven 170 ± 15 0.2 PP 3
Celgard PP/PP non-woven 25/110 0.064 PP 4
208
7.2.2 Apparatus and Procedures
The experiments were performed using aqueous solutions of MEA (30.0 wt.%),
AMP (30.0 wt.%), AHPD (23.0 wt.%), Pz (7.0 wt.%), and AHPD (23.0 wt.%) + Pz (7.0
wt.%).
7.2.2.1 Surface tension
Surface tension data were measured at 298.2 K and 313.2 K using an optical contact
angle analyzer (OCA 15 Plus, Future Digital Scientific Corp, USA) based on the pendant
drop method. The apparatus was equipped with a thermostated chamber controlled with a
precision of 0.1 K using a refrigerated/heating circulator with high precision external
temperature control (Julabo F25-ME). Droplet geometry was analysed by digitizing the
image from a camera and the device’s software calculated the surface tension based on the
difference between the ambient phase and solution densities, drop maximum diameter and
form factor (Aguila-Hernandez et al., 2007). Ambient phase (air) density was corrected
considering 90% humidity. Increasing humidity in the measurement chamber was essential
to avoid evaporation of the drop that can affect the surface tension values. All
measurements were made at least in triplicate and average values are reported.
7.2.2.2 Density and viscosity of solutions
Densities of aqueous solutions are necessary to determine the surface tension
values, as mentioned in §7.2.2.1. They were measured by using a calibrated pycnometer
having a bulb volume of 1×10-5 m3 and a Mettler AE240 balance with a precision of ±1×10-
4 g. Temperature of the pycnometer was kept within ±0.1 K using a precision thermometer.
The calculated uncertainties of the measured density were within ±0.06 kg/m3. As
mentioned by Lin et al. (2008), higher liquid viscosity may lead to a lower penetration of
liquid into membrane pores, thus reducing membrane wettability. The kinematic viscosities
of solutions were therefore measured by means of a Cannon-Fenske viscometer.
Measurements were made in a water bath whose temperature was kept constant within ±0.1
K. Kinematic viscosities were calculated from the efflux times measured with an electronic
stopwatch with a resolution of 0.01 s. The experimental uncertainties were calculated to be
209
within ±0.3%. The dynamic solution viscosities were calculated by multiplying the
kinematic viscosities with the corresponding densities.
7.2.2.3 Contact angle
Contact angle measurements were performed using an optical contact angle
analyzer (OCA 15 Plus) based on the sessile drop method. A small droplet was deposited
on the surface of a membrane and the contact angles (that could be associated to advancing
contact angles) were determined from images acquired by camera. At least three droplets
were dispensed on each tested membrane and a mean value was recorded. Prior to each test,
membranes were thoroughly cleaned with alcohol and warm water and then dried overnight
at 333.2 K to remove the liquid remaining in the pores. Data for each specific membrane
were measured with an average uncertainty of ±3°.
Figure 7.1. Breakthrough pressure apparatus
7.2.2.4 Breakthrough pressure The breakthrough pressure was measured based on the Laplace-Young equation
(Eq. 7.1), using the setup shown in Figure 7.1. Water or amine solutions were pressurized
by nitrogen and the liquid pressure was measured using a pressure transducer (PX319,
Omega) with a precision of ±0.7 kPa. Experiments were performed at constant temperature
by keeping the membrane setup in a thermostated air bath (Julabo F12-ED). Liquid
temperature was measured using a thermocouple with a precision of ±0.1 K. As high
pressures can be reached in the experiments, a wire mesh support screen was installed
above the membrane to avoid as much as possible membrane deformation. Liquid pressure
was gradually increased (around 14 kPa/min) until small droplets were visually observed
on the membrane surface. Used membranes were removed, washed and dried and reused
210
when specific tests were necessary. The membrane area exposed to the liquid in all
experiments was 12.57 ± 0.06 cm2.
7.3. Results and Discussion 7.3.1 Absorbent density and viscosity
Density and viscosity data are given in Table 7.2. For 30.0 wt.% MEA aqueous
solutions, data obtained at 298.2 and 313.2 K were compared to literature values. For
density, excellent agreement is observed for both temperatures; a maximum deviation of
0.11% was found from data of Amundsen et al. (2009) or Han et al. (2012). Concerning
viscosity, excellent agreement was also observed; our result at 298.2 K (2.40 mPa·s) was
found to be between 2.32 mPa·s (Islam et al., 2004) and 2.48 mPa·s (Amundsen et al.,
2009) while at 313.2 K, our value (1.55 mPa·s) is very close to the value of Islam et al.
(2004) (1.536 mPa·s).
Density and viscosity values decrease with the increase of temperature, as expected.
Furthermore, an increase in total amine concentration (AHPD + Pz solution versus Pz
solution or AHPD solution) leads to an increase of both densities and viscosities. It is worth
mentioning that the addition of Pz in AHPD aqueous solution causes a very small increase
in density (0.4%), while the viscosity strongly increased by 60% and 52% at 298.2 K and
313.2 K, respectively. The viscosity influence on membrane breakthrough pressure and
wettability will be discussed in §7.3.4.2.
Table 7.2 Density and viscosity of aqueous amine solutions. Solution Concentration T ρ µ
(wt.%) (K) (kg/m3) (mPa·s)
MEA 30.0 298.2 1011.4 2.40 313.2 1004.5 1.55
AHPD 23.0 298.2 1056.9 1.84 313.2 1051.0 1.23
Pz 7.0 298.2 1000.0 1.19 313.2 995.2 0.84
AHPD + Pz 23.0 + 7.0 298.2 1061.4 2.96 313.2 1054.7 1.87
AMP 30.0 298.2 996.9 3.61 313.2 988.3 2.07
211
Table 7.3. Surface tension of aqueous amine solutions.
Solution Concentration T σ Literature values (wt.%) (K) (mN/m) σ (mN/m) Reference
Water - 298.2 72.1 ± 0.2 72.0 Perry (1997) 313.2 70.1 ± 0.5 69.6 Perry (1997)
MEA 30.0 298.2 63.9 ± 0.3 60.4 / 64.0* (Vazquez et al., 1997) / (Han et al., 2012)
313.2 61.5 ± 0.4 57.9 / 62.6 (Vazquez et al., 1997) / (Han et al., 2012)
AHPD 23.0 298.2 71.2 ± 0.3 57.6* (Murshid et al., 2011c)
313.2 69.4 ± 0.5 55.5* (Murshid et al., 2011c)
Pz 7.0 298.2 70.1 ± 0.2 70.7*/ 70.4* / 69.8* (Muhammad et al., 2009) / (Murshid et
al., 2011b) / (Murshid et al., 2011a)
313.2 67.6 ± 0.6 68.2* / 68.3* / 68.1* (Muhammad et al., 2009) / (Murshid et
al., 2011b) / (Murshid et al., 2011a)
AHPD + Pz
23.0 + 7.0 298.2 70.2 ± 0.2 48.8* (Murshid et al., 2012)
313.2 67.1 ± 0.7 46.0* (Murshid et al., 2012) AMP 30.0 298.2 46.1 ± 0.3 43.4 / 46.8* / 47.0* /
45.0* (Vazquez et al., 1997)
/ (Venkat et al., 2010b) / (Paul and Mandal, 2006b) / (Murshid et al.,
2011a) 313.2 44.0 ± 0.3 41.7 / 44.7* / 44.3* /
42.5* (Vazquez et al., 1997)
/ (Venkat et al., 2010b) / (Paul and Mandal, 2006b) / (Murshid et al.,
2011a) * Extrapolated or interpolated values
7.3.2 Absorbent surface tension
Surface tension values for amine solutions were determined at 298.2 K and 313.2 K
because data available in the literature either had to be estimated by
extrapolation/interpolation or were found to be contradictory. All measured data are given
212
in Table 7.3, together with literatures values. As expected, an increase in temperature
reduces the surface tension of all tested liquids.
For water, our data are within the experimental error of the values reported in the
literature (Perry, 1997). For MEA or AMP solutions, the present experimental data are in
average 3.0 mN·m-1 higher than the values reported by Vázquez et al. (1997), but they are
in good agreement with all other literature data (Han et al., 2012; Murshid et al., 2011a;
Paul and Mandal, 2006b; Venkat et al., 2010b). For Pz solutions, the present data for
aqueous 7.0 wt.% solution agree well with the estimated values from literature
(Muhammad et al., 2009; Murshid et al., 2011a; Murshid et al., 2011b). As also mentioned
by Derks et al. (2005a), it was found that the addition of small amounts (up to 12.9 wt.%)
of piperazine in water or in an aqueous amine solution does not have an important effect on
the surface tension values. Concerning AHPD containing solutions, a significant deviation
of around 20% and 30% for AHPD and AHPD + Pz solutions, respectively, were found in
comparison with data reported by Murshid et al. (Murshid et al., 2011c, 2012). In order to
elucidate this significant deviation, we attempted to associate the values of the surface
tension of aqueous amine solutions to the solute molecular structure.
The surface tension value of a liquid in contact with a gaseous phase is known to be
related to the difference in intermolecular interactions (e.g. Van der Waals forces, hydrogen
bonding) of surface molecules (attracted into the liquid by their neighbours, as there is
practically no force attracting the surface molecules away from the liquid) compared to the
bulk ones (attracted equally in all directions) (Adamson and Gast, 1997). Water molecules,
for example, are known to develop several intermolecular interactions (especially hydrogen
bonding) according to their composition, molecular structure and polarity. Consequently,
surface molecules are strongly attracted within the liquid by their neighbours, thus giving to
water a high surface tension value. The following hypotheses can therefore be considered:
i) the addition in pure water of larger molecules (alkanolamines, as considered in this work)
developing less intermolecular interactions than those present in water, will reduce water
surface tension value and ii) molecules having more non-polar (hydrophobic) groups than
213
polar (hydrophilic) ones will develop less intermolecular interactions with other molecules
(amines and especially water), thus reducing the surface tension more significantly.
The values of aqueous alkanolamine solutions given in Table 7.3 appear to confirm
theses hypotheses. A comparison of experimental data for water and aqueous solutions of
MEA, AMP and AHPD leads to the following order of the surface tension at both
temperatures: water > AHPD > MEA >> AMP. Following the hypothesis mentioned
before, this order can be explained by the fact that AHPD molecule has four hydrophilic
groups (three hydroxyls and one amino), compared to two in the MEA molecule (one
hydroxyl and one amino). The AMP molecule has also two hydrophilic groups (one
hydroxyl and one amino), but possesses two more hydrophobic (methyl) groups than MEA,
leading therefore to a more important reduction of the surface tension. Águila-Hernández et
al. (2007) mentioned that AMP molecular configuration (presence of two methyl and one
hydroxyl groups) is responsible for its pseudosurfactant behaviour that leads to a significant
reduction of water surface tension. Asprion (2005) found a strong relationship between the
molecular configuration of different alkanolamines and a surface tension model parameter
called the binding constant, obtained by fitting experimental data for aqueous alkanolamine
solutions. The author mentioned that hydrophilic (-NH2, -OH) or hydrophobic (methyl)
groups have a significant impact on the binding constant and therefore, on the calculated
surface tension.
214
Table 7.4. Surface tension around 298 K and 30 wt.% of various aqueous amine solutions and their carbon and hydrophilic numbers.
Carbon # Hydrophilic # Molecule σ(mN/m) Reference
1 1 Methanol 40.4 (Vazquez et al., 1995) 2 1 Ethanol 32.4 (Vazquez et al., 1995) 2 Ethylene Glycol 62.1 (Hoke and Chen, 1991) 2 MEA 63.9 This work
3 1 1-Propanol 26.0 (Vazquez et al., 1995) 1 2-Propanol 26.8 (Vazquez et al., 1995) 2 1,2-Diaminopropane 49.7 (Blanco et al., 2012) 2 1,2-Propanediol 50.6 (Nakanishi et al., 1971) 2 2-(Methylamino)ethanol 52.5 (Venkat et al., 2010b) 2 1-Amino-2-Propanol 53.6 (Alvarez et al., 2003) 2 1,3-Propanediol 57.1 (Nakanishi et al., 1971) 2 3-Amino-1-Propanol 60.5 (Alvarez et al., 2003) 3 Glycerol 68.5 (Ernst et al., 1936)
4 1 2-Methyl-2-Propanol 23.8 (Cheong and Carr, 1987)
2 Dimethylethanolamine 44.7 (Maham and Mather, 2001)
2 2-(Ethylamino)ethanol 44.9 (Alvarez et al., 2008) 2 AMP 46.1 This work 2 1,3-Butanediol 49.8 (Nakanishi et al., 1971) 2 1,4-Butanediol 55.1 (Nakanishi et al., 1971) 3 Diethanolamine 60.8 (Vazquez et al., 1996) 4 AHPD 70.9 This work
5 3 N-Methyldiethanolamine 53.5 (Alvarez et al., 1998) 3 2-Amino-2-ethyl-1,3-
propanediol 55.2 (Yoon et al., 2002b)
6 1 Triethylamine 22.0 (Livingston et al., 1916)
3 Diisopropanolamine 47.8 (Kelayeh et al., 2011)
4 Triethanolamine 57.8 (Vazquez et al., 1996)
9 4 Triisopropanolamine 38.4 (Chauhan et al., 2003)
215
Based on the above mentioned hypotheses, we propose a new and very simple way
to estimate the surface tension of aqueous solutions on the basis of the number of carbon
atoms (called here the carbon number) and that of hydrophilic groups (called here the
hydrophilic number) of the solute. According to this approach, solutes like alkanolamines
can be characterized by the number of carbon atoms (and not only methyl groups)
representing hydrophobic groups and the number of hydrophilic groups (here, hydroxyl and
amino) present in its structure. All the surface tension data found in the literature for
aqueous solutions of alcohols, amines and alkanolamines are given in Table 7.4 (at around
298 K and 30 wt.%). Table 7.4 contains several representative classes of compounds, such
as primary and tertiary amines, primary, secondary, tertiary and sterically hindered
alkanolamines, as well as primary, secondary, and tertiary alcohols.
Strong tendencies can be observed between the value of surface tension and both
carbon and hydrophilic numbers (Figure 7.2), confirming that the molecular configuration
of the solute (amine or alcohol) and the number and nature of its constitutional groups
influence its aqueous surface tension.
Several trends can be distinguished:
- increasing the carbon number for the same hydrophilic number reduces the aqueous
surface tension (this is in agreement with Traube’s rule (Traube, 1891) describing
approximately the decrease of surface tension in homologous series with the addition of
CH2 groups);
- increasing the hydrophilic number for the same carbon number increases the aqueous
surface tension;
- replacing one hydroxyl in polyols by an amino group increases the aqueous surface
tension (e.g., MEA vs. ethylene glycol; 3-amino-1-propanol vs. 1,3-propanediol; 1-
amino-2-propanol vs. 1,2-propanediol);
- linear molecules having two hydrophilic terminal groups have higher aqueous surface
tension than branched molecules having the same carbon and hydrophilic numbers. This
can be easily seen for a hydrophilic number equals 2, where several examples are
available (e.g., propylene glycol vs. 1,2-propanediol; 1,4-butanediol vs. 1,3-butanediol).
216
Figure 7.2. Influence of the carbon and hydrophilic numbers on surface tension of various
aqueous solutions.
This new approach to estimate the surface tension of aqueous amine, alcohol and
alkanolamine solutions offers several advantages in comparison with what has been
published in the literature concerning surface tension estimation.
First, this method is very simple, does not require any experimental data concerning
the solute of interest in order to estimate the surface tension of its aqueous solutions and
there is no fitted parameters in comparison with available models (Asprion, 2005; Li and
Lu, 2001; Mejia et al., 2005).
Second, various surface tension models are valid only for pure substances or
aqueous solutions with weight percentage usually larger than about 30% (equivalent to
about 0.1 mole fraction). Because in the very low mole fraction region (lower than about
0.1), the surface tension of solutions decreases extremely sharply with the solution
217
concentration, the corresponding models are not always appropriate for solutions usually
used in the CO2 capture process (weight percentage lower than about 30%) (Li and Lu,
2001; Luck, 2001). Molecular interactions in dilute/concentrated mixtures are different and
this has an impact on the surface tension trends of pure substances compared to those
corresponding to aqueous solutions. For example, surface tension of mono and poly-
alcohols presented in Figure 7.2 are well explained by our approach for aqueous solutions,
but the values corresponding to pure compounds (methanol 22.51 > ethanol 21.82 <
propanol 23.28 (Vazquez et al., 1995) or ethylene glycol 46.24 < 1,3-propanediol 46.95 >
1,4-butanediol 43.79 (Nakanishi et al., 1971)) do not follow the same trend. Besides, pure
substances are not always liquids and therefore, it is not possible to consider the solid
compounds (e.g. AHPD and di- or tri- isopropanolamine) in the models based on pure
liquid compounds.
Third, this method allows to (approximately) delimitate graphically two regions
were no surface tension data were found in the open literature to perform this analysis
(except for triethylamine). These two symbolic areas were shaded in Figure 7.2. It is
important to mention that the present analysis was performed at around 298 K and 30 wt.%
aqueous solutions, because most available experimental data were obtained in these
conditions; similar tendencies should be valid for other conditions. The bottom left area
includes molecules having possibly too many hydrophilic groups in their structures for the
small number of carbon they contain (chemically unstable). On the other hand, the upper
right area contains molecules that are usually very little soluble in water and therefore,
surface tension data for 30 wt% aqueous solution were not available.
Fourth, one can use the observed trends for the selection of new absorbents for acid
gas separation (e.g., CO2 capture) in membrane contactors where a high surface tension is
required, or in packed columns which need low surface tension absorbents. For example,
using this method and the third trend mentioned above, we predicted that one derivative of
glycerol (3 hydrophilic groups for 3 carbon groups), the 2-amino-1,3-propanediol, obtained
by replacing one hydroxyl by an amino group, should give a 30 wt.% aqueous solution with
a very high surface tension (around 70.5 compared to 68.5 mN·m-1 for glycerol) at 298.15
218
K. Our very recent work, actually initiated from the analysis of Figure 7.2, confirmed this
prediction (Bougie and Iliuta, 2013b).
Finally, this method allows to elucidate the significant deviation of the values of the
surface tension of AHPD aqueous solutions obtained in the present work, compared to the
literature ones, as mentioned at the beginning of this section. As shown in Figure 7.2, the
position of AHPD (4 hydrophilic groups for 4 carbon groups) predicts a high surface
tension of its solutions. This agrees very well with our experimental data and confirms
therefore their reliability in comparison with literature data (Table 7.3). In addition to a
high surface tension, it was already shown that AHPD based mixtures present good
absorption and regeneration performances, hence showing its great potential for use in MC
(Bougie and Iliuta, 2012).
7.3.3 Membrane/absorbent contact angle
Contact angle data are shown in Table 7.5. Data assigned to PTFE are mean values
of tests made on PTFE membranes (PTFE 1 to 5, as indicated in Table 7.1). Similarly, PP
concerns PP 1 to 4. The highest contact angle was obtained for water, in agreement to its
highest surface tension. However, AMP, presenting the lowest surface tension, leads to the
lowest contact angle only in contact with PTFE, but not with PP. In the case of PP, the
lowest contact angle was obtained for MEA solution. From 298.2 to 313.2 K, contact
angles slightly decrease, following the similar surface tension tendency.
Table 7.5. Contact angles for several absorbent/membrane combinations.
T (K) Material Water AMP MEA AHPD Pz AHPD + Pz
298.2 PTFE 135 130 133 133 133 133 PVDF 141 - - 134 143 144 PP 117 113 107 111 109 112
313.2 PTFE 138 126 130 135 134 132 PVDF 140 - - 133 142 142 PP 114 107 104 107 109 111
It can be seen that the highest contact angles were obtained on PVDF for all tested
solutions and temperatures. Data gives the following general trend for the contact angle:
219
PVDF > PTFE > PP. It is interesting to note however that membranes from other
manufacturers can result in a different ranking due to a possible different membrane surface
morphology. Even if the contact angle data for PVDF usually gives higher values, which
could be advantageous for the use in membrane contactors where a high hydrophobicity is
needed, PVDF presented stability problems. As indicated in Table 7.5, no contact angle
data were given for AMP and MEA 30 wt% solutions, as an almost instant chemical
degradation of the membrane in contact with the amine solutions was observed. This
confirmed that the PVDF chemical resistance in respect to highly alkaline alkanolamine
solutions is not satisfactory. This behaviour agrees with literature data who clearly showed
that PVDF membranes have been generally used for CO2 separation using water or diluted
absorbent solutions (Atchariyawut et al., 2007; Khaisri et al., 2009; Lin et al., 2008; Lin et
al., 2009a; Mansourizadeh and Ismail, 2010; Naim et al., 2012; Rajabzadeh et al., 2009;
Yeon et al., 2003). An analysis of the absorbent solutions pH (Table 7.6) confirmed higher
alkalinity for AMP and MEA solutions. Even in contact with AHPD solutions that present
the lowest pH value, the PVDF membranes showed chemical degradation after 3 days. The
use of PVDF membranes is therefore unsuitable for CO2 capture in membrane contactors
since highly concentrated absorbents are needed for better absorption/stripping energetic
efficiency (Mejdell et al., 2010b; Sakwattanapong et al., 2005).
Table 7.6. Alkalinity of tested amine solutions.
Solutions pH AHPD (23.0 wt.%) 10.92 AHPD (23.0 wt.%) + Pz (7.0 wt.%) 11.69 MEA (30.0 wt.%) 12.10 AMP (30.0 wt.%) 12.16 Pz (7.0 wt.%) 11.78
Concerning the tested laminated membranes, contact angle experiments revealed an
interesting detail. While contact angles for single polymer membranes (PTFE or PP)
remained almost constant, even for membranes from different sources, the lamination
process seemed to influence the contact angle values obtained for laminated membranes.
An average increase of 5° was observed for laminated PTFE/PP and PP/PP in comparison
220
with the value corresponding to standard membranes (PTFE and PP). The lamination
process seems then to create some roughness at the polymer surface, which can therefore
lead to an increase in the contact angle values (Mosadegh-Sedghi et al., 2013). Considering
that the addition of a PP supporting layer to a PTFE thin membrane can offer the membrane
additional durability, stability, and stiffness at a low price and, as indicated, additional
hydrophobicity, laminated PTFE/PP membranes could represent an interesting alternative
to standard membranes in membrane contactors.
7.3.4 Breakthrough pressure
First, experimental data of breakthrough pressure (∆PB.P.exp) were performed with
water at 298.2 K in order to determine the maximum pore size of each membrane, based on
Eq. (7.1). Water was chosen to avoid chemical reaction with all membranes and also
because its surface tension is well known. The results, together with the theoretical
breakthrough pressure calculated with the nominal pore size of each membrane (∆PB.P.nom)
(given for comparison), are presented in Table 7.7.
nomB.P.nom
cos 4- d
P θσ=∆ (7.2)
Maximum pore sizes were determined using fresh (unused) membranes only. As
expected, the measured maximum pore sizes were found to be higher than the nominal
(mean) values given by the manufacturers. No values were measured for the PP membranes
1, 2 and 4 as they all were broken around 350-400 kPag before any droplets were detected
on the membrane surface. The thickness of PP 1, 2 and 4 is around 100 µm, which is
comparable to PTFE 1. However, PTFE 1 membrane can tolerate a much higher pressure
without rupture, just as PTFE 3 which presents a very low thickness (25 µm). Nevertheless,
handling the very thin PTFE 3 membrane was very difficult as it folds up on itself very
quickly due to electrostatic attraction. In order to avoid it, a possible alternative is the use
of laminated PTFE membranes.
PVDF membranes present the lowest breakthrough pressure among all tested
membranes. A large nominal pore size of 0.45 micrometer could explain this result (Table
7.1). All tests with laminated membranes were performed with the solution in contact with
the membrane side, because the contact between the liquid and the supporting layer caused
221
a delamination of the membrane assembly. A comparison between the theoretical water
breakthrough pressures calculated with the nominal pore size and the experimental values
show a reduction of 30 to 84% (Table 7.7). These results lead to the conclusion that it is
highly important to consider the maximum pore size in the membrane contactor design
where the liquid pressure has to represent a small percentage of the real breakthrough
pressure, in order to maintain long-term absorption performance and to avoid membrane
wetting.
Table 7.7. Experimental breakthrough pressure (∆PB.P.exp) for maximum pore size determination using water at 298.2 K.
Membrane ∆PB.P.nom (kPag) ∆PB.P.exp (kPag) ∆PB.P. Reduction (%) dmax (µm) PTFE 1 1945 1269 35 0.16 PTFE 2 1945 1225 37 0.16 PTFE 3 778 544 30 0.36 PTFE 4 1068 573 46 0.37 PTFE 5 1068 444 58 0.48
PP 1 1116 354* 68 - PP 2 1116 343* 69 - PP 3 768 270 65 0.57 PP 4 2399 393* 84 - PVDF 498 154 69 1.45 *Membrane rupture
A second experimental set was performed using both water and aqueous amine
solutions to perform successive measurements of the breakthrough pressure where the same
membrane was used for each tested solution. Membranes were washed and dried overnight
between each test. As example, the results for PTFE 2 at 298.2 K are given in Table 7.8. It
can be seen that successive use of a membrane reduced gradually the breakthrough
pressure. As in the breakthrough pressure measurements the first droplets observed at the
surface of the membrane correspond to the largest pores (Eq. 7.1), it seems that successive
use of membranes leads to the enlargement of the biggest pores, due to the pressure applied
to the polymer by the solution. It is therefore important in membrane contactors to avoid
exposing the membrane to high liquid pressure, as pore size modification can lead to
important decrease in membrane performance in long-time operation.
222
For all tested solutions, the Laplace-Young equation applied considering the
maximal pore diameter was found to overestimate the experimental breakthrough pressure
(second column in Table 7.8). Experimental data combining all tested membranes and
solutions are, in average, 15% lower that the predicted values calculated with Laplace-
Young equation (third column in Table 7.8). One possible explanation can be associated to
the pore structure of each kind of membrane tested. As mentioned by Dindore et al. (2004),
some membranes have a fibrous structure and the pores represent irregular spaces that
remain between adjacent fibers, while other membranes have a spongy structure. In order to
take into account these irregular pore structures comparatively to a perfect cylindrical pore
structure, Franken et al. (1987) introduced a geometric coefficient of the pore, Β, at the
right-hand side of Laplace-Young equation:
maxB.P.
cos 4- d
BP θσ=∆ (7.3)
where Β = 1 for cylindrical pores and 0 < Β < 1 for non-cylindrical pores. From SEM
pictures of tested membranes (Figure 7.3), it can be seen that the pore structures are not
cylindrical. This membrane-related factor would reduce the predicted values of the
breakthrough pressure but it cannot explain that the highest deviations were obtained for
solution having the lowest surface tension (Table 7.8). Here again, this trend reveals the
importance of the surface tension of absorbents; a solution with high surface tension will
present a high breakthrough pressure, which will also be closer to the value calculated
using the Laplace-Young equation. This value is necessary to set the correct fluid pressures
in MC.
7.3.4.1 Relationship between membrane long-term stability and breakthrough pressure
For this analysis, a review of more than 135 literature works reporting data
concerning the use of hydrophobic microporous membranes in MC for CO2 absorption with
water or several kinds of absorbent solutions was performed. However, only those which
contain enough information to calculate or estimate the nominal breakthrough pressure
were selected (Kosaraju et al., 2005; Lin et al., 2008; Lin et al., 2009a; Lin et al., 2009b;
Lin et al., 2009c; Mavroudi et al., 2003, 2006; Rongwong et al., 2009; Sea et al., 2002; Yan
223
et al., 2007). As the membrane nominal pore size is usually available in the literature
instead of the actual maximal pore size, a nominal breakthrough pressure, B.P.nomP∆ (Eq.
7.2) was calculated instead of B.P.P∆ (Eq. 7.1). The required contact angles were either
directly reported or could be estimated to be 110° for PP or 130° for PTFE. Concerning
PVDF, when contact angle data were not given, the values could not be estimated because
of the large variability of reported values for this kind of membrane.
Table 7.8. Breakthrough pressure using water and aqueous amine solutions with PTFE 2.
Solution ∆PB.P. (kPag) ∆PB.P.exp (kPag) ∆PB.P. deviation (%)
water 1225* 1225 0 1206 -2 1199 -2
MEA 1044* 855 -18 830 -20 798 -24
AHPD 1164* 1108 -5 1050 -10
961 -17 Pz 1140* 1162 2
1108 -3 1078 -5
AHPD + Pz 1155* 1093 -5 1030 -11 998 -14
AMP 759* 483 -36 476 -37 470 -38
*Calculated using dmax indicated in Table 7.7
Data analysis revealed that where the liquid overpressure (∆Pliq-gas) was below 1.5%
of the nominal breakthrough pressure, a stable long-term performance of gas absorption
was reported, while the other works reported a decrease of performance and/or membrane
wetting.
224
. .
1.5%liq gas
B P nom
PP
−∆<
∆ (7.4)
Such a low ratio can be explained by the fact that the use of the nominal breakthrough
pressure overestimates the real breakthrough pressure of a membrane, which should be
higher than the operational, ∆Pliq-gas, for several reasons: (i) the maximal pore size rather
than the nominal one determines the real breakthrough pressure, as seen in Table 7.7, (ii)
the non-ideality of the pore structure, (iii) the deviation caused by low surface tension
absorbents, as seen in Table 7.8 and (iv) the security factor taken into consideration to
avoid instantaneous liquid intrusion into the pores, as defined by the concept of
breakthrough pressure.
Figure 7.3. SEM pictures of some tested membranes.
To respect this criterion, high values for nominal breakthrough pressure, ∆PB.P.nom
have to be assured. For this, the following measures could be highlighted:
• Use of CO2 absorbent solutions having high surface tension values even at high
concentrations. As shown in the section 7.3.2, it is possible to select amine aqueous
solutions that possess high surface tension values. Among the usual alkanolamines
considered for acid gas separation, AMP (a well-known sterically hindered
225
alkanolamine) should be avoided in membrane contactors due to its low surface tension
value; e.g., wetting problems were reported in the case the use of AMP in MC
(deMontigny et al., 2006; Lin et al., 2009a). Moreover, the operation temperature is
another important parameter that has to be considered as an increase in temperature
leads to the reduction of surface tension and consequently, a reduction of contact angle.
• Use of membranes showing high hydrophobicity and resistance to wetting. Currently,
PTFE seems to represent the best choice over PVDF and PP, showing high contact
angles and excellent chemical stability. Combining absorbents with high surface
tension with very hydrophobic membranes will lead to the highest contact angles.
• Use of membranes with low nominal pore size. However, the CO2 diffusivity through
the membrane can be restricted if too small pore sizes (e.g., 0.02 µm) are used, which
will increase significantly the membrane resistance (Lu et al., 2008).
In addition to all these measures to obtain high ∆PB.P.nom, it also appears important
to keep in the MC the lowest possible liquid overpressure (∆Pliq-gas). This aspect is often
neglected in the literature, despite the fact that it is of major importance at the contactor
liquid inlet position where the maximum value is obtained (highest liquid pressure and
lowest gas pressure in a countercurrent fluid configuration). Several actions can be
envisaged to lower the value of the liquid overpressure:
• Maintain low liquid flow rate. This will reduce the liquid pressure, as well as the
pumping energy and regeneration cost (Yan et al., 2007);
• Keep low liquid pressure drop in MC. Feeding the absorbent solution inside very small
hollow-fiber membranes (inside diameter < 250 µm) may lead to high pressure drops,
as it is often observed with bundle of thousand PP membranes. Considering
membranes with larger inside diameters can reduce the pressure drop, allowing
therefore the use of the liquid flow in the fiber lumen, which is a better choice than
having the liquid flowing on the shell side (increases the absorption performance
(Mavroudi et al., 2003));
• Increase the gas pressure. This will reduce the liquid overpressure and it can increase
the absorption performance due to the increase of the CO2 partial pressure.
226
7.3.4.2 Viscosity influence on breakthrough pressure An increase of solution viscosity was mentioned in the literature to reduce
membrane wetting (Lin et al., 2008; Lin et al., 2009a). It was reported that, being more
viscous, the absorbent solution might have more difficulty to enter the membrane pores.
However, experimental breakthrough pressure for aqueous AMP solutions having the
highest viscosity among the tested solutions (Table 7.2) was not found to be closer to the
calculated value based on the Eq. (7.1), compared to solutions presenting lower viscosity.
Data presented in Tables 7.2 and 7.8 show that the breakthrough pressure does not seem to
be influenced by viscosity. Additional thorough studies are necessary to investigate this
behavior more in depth.
7.4. Conclusions In this study, several parameters related to membrane wetting and linked to Laplace-
Young equation were investigated. A high surface tension is one of the key parameters to
be considered in the choice of absorbents to be used in MC. A new classification method
that could be very useful for the estimation of surface tension of aqueous amine or alcohol
solutions (aqueous binary systems) was developed here. Molecular structure of a solute has
shown to have a strong influence on the surface tension of its corresponding aqueous
solution. As example, AHPD, a sterically hindered alkanolamine with 4 hydrophilic groups
and a carbon number of 4, seems to be very appropriate to be considered for use in MC
because of it very high surface tension.
PVDF membranes were found to degrade over highly concentrated absorbents, their
use being therefore restricted to very dilute absorbent solutions. In addition to PTFE (high
hydrophobicity and superior chemical and mechanical resistance) and PP (much less
expensive, but wetted by aqueous absorbents), laminated PTFE/PP membranes who
combined the advantages of these two materials were found to be an interesting alternative
to be considered in plate MC because the lamination process seems to increase the surface
roughness that leads to higher contact angle values. Moreover, the fabrication of this kind
of membranes requires less amount of PTFE (an expensive material).
227
Membrane pore size and hydrophobicity, liquid surface tension and liquid pressure
are the most important parameters influencing the long-term absorption capacity in MC. In
this context, a new criterion for long-term performance of gas absorption in MC was
proposed here. It was estimated that a ratio between the liquid overpressure and the
nominal breakthrough pressure less than 1.5% seems to ensure long-term absorption
performance by preventing membrane wetting and several actions were suggested to
respect this criterion.
Finally, it is worth mentioning that only fresh (not degraded) solutions were
considered in this study because a very large variety of degradation products can be found
as impurities in the used solutions; each amine has its own degradation products which
even change with experimental conditions. A new study dedicated to the analysis of the
effect of degradation products on polymeric membrane wettability, for specific
absorbent/membrane systems, would be interesting for future works.
228
In the precedent chapter, we introduced an easy and very simple graphical molecular
classification method that can be used to identify potential amines whose aqueous solutions
present surface tensions appropriate for special gas separation applications (e.g., high
surface tension required for use in membrane contactors). Following this method, Serinol
(2-amino-1,3-propanediol) seemed to be an amine whose aqueous solution surface tension
should be higher than that of typical amine solutions used for acid gas separation. We
found therefore interesting to investigate the potential of this amine as an efficient CO2
absorbent to be used in MC and this will be the object of Chapter 8. Serinol is not
necessarily a SHA, in the light of the usual definition of these compounds, but it is
nevertheless more hindered than MEA and, in the same time, could have the advantage of a
much better kinetics toward CO2 compared to SHA.
229
Chapter 8. Solubility of CO2 in and density, viscosity and surface tension of aqueous 2-Amino-1,3-propanediol (Serinol) solutions
Résumé
Dans cette étude, les solutions aqueuses de 2-amino-1,3-propanediol (Sérinol) ont été caractérisées par la densité, la viscosité, la tension superficielle et la solubilité du CO2, afin d'évaluer l'utilisation potentielle de cette alcanolamine pour la capture du CO2 des mélanges gazeux. La densité et la viscosité ont été mesurées à des températures entre 293.2 et 313.2 K et pour des concentrations (molalités) d'amine entre 0.953 et 4.693 mol·kg-1. La tension superficielle a été mesurée pour les mêmes concentrations d’amine, mais à 298.2 et 313.2 K. La solubilité du CO2 dans des solutions de Sérinol de molalités entre 0.953 et 4.704 mol·kg-1 a été déterminée à 313.15 K et pour des concentrations de 4.704 mol·kg-1 à 343.15 et 373.15 K. Les capacités d’absorption du CO2 ont été comparées aux données de la littérature pour la monoéthanolamine (MEA), ainsi qu’à celles obtenues pour le système 2-amino-2-hydroxyméthyl-1,3-propanediol (AHPD) + pipérazine (Pz) (concentration de 2.712 + 1.161 mol·kg-1) à 313.15 et 373.15 K. Les résultats ont montré que les solutions de Sérinol ont des tensions superficielles plus élevées par rapport aux absorbants classiques, ce qui les rend très appropriées pour la séparation du CO2 dans des contacteurs à membrane. Les mesures de solubilité ont montré aussi que les carbamates formés par la réaction du CO2 avec le Sérinol peuvent être régénérés plus facilement par rapport à ceux produits dans des solutions de MEA. La capacité cyclique du processus d’absorption du CO2 dans le Sérinol de 58% est plus élevée que la valeur obtenue pour MEA, mais proche de celle correspondante au système AHPD + Pz.
230
Abstract
In this work, 2-amino-1,3-propanediol (Serinol) aqueous solutions were characterized through density, viscosity, surface tension and CO2 solubility measurements in order to evaluate the potential use of this alkanolamine for CO2 removal from different gas mixtures. Density and viscosity were measured from temperature T = (293.2 to 313.2) K and for amine concentrations from molality m = (0.953 to 4.693) mol·kg-1. Surface tension data were measured for the same solution concentrations but at T = (298.2 and 313.2) K. CO2 solubility in Serinol solutions from m = (0.953 to 4.704) mol·kg-1 was determined at T = 313.15 K and at T = (343.15 and 373.15) K for the m = 4.704 mol·kg-1 solution. CO2 loading capacities were compared to literature data for monoethanolamine (MEA) and those obtained for (2.712 + 1.161) mol·kg-1 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD) + piperazine (Pz) solution at T = (313.15 and 373.15) K. It was found that Serinol solutions have higher surface tensions compared to conventional absorbents, making them very suitable for CO2 removal using membrane contactors. Solubility measurements showed that Serinol carbamates formed by the reaction with CO2 can be more easily regenerated in comparison with those produced in contact with MEA. The CO2 cyclic capacity of Serinol was found to be 58% higher than that of MEA and close to the value obtained for the AHPD + Pz system.
231
8.1. Introduction The removal of acid gas such as CO2 and H2S in natural gas sweetening and CO2
capture from fossil-fuel-fired power plants or petrochemical, steel, and cement production
is of high interest for technical, economic and environmental concerns (Kohl and Nielsen,
1997). Among various possible techniques, the chemical absorption by aqueous
alkanolamines solutions is today’s best available technology (Bernardo et al., 2009) and
monoethanolamine (MEA) is considered since many decades as the benchmark amine for
this process. Numerous investigations have been performed to find better absorbents with
better kinetics (Ma'mun et al., 2007), higher CO2 solubility (Puxty et al., 2009a), less
corrosiveness (Veawab et al., 1999) and lower degradation rate (Lepaumier et al., 2009a),
as well as to avoid excessive energy requirement at the stripper (Rochelle, 2012). In
addition to all these important parameters related to the absorption liquid, the choice of the
gas-liquid contactor is another key factor to be considered in the choice of appropriate
absorbents for industrial applications. In this context, the use of membrane contactors (MC)
as highly efficient alternatives to packed columns (deMontigny et al., 2005) requires
absorbents with very high surface tension, in order to avoid the unfavourable wetting
phenomenon (Rongwong et al., 2009).
We recently proposed the aqueous mixture 2-amino-2-hydroxymethyl-1,3-
propanediol + piperazine (AHPD + Pz) as a potential alternative absorbent to MEA
solution, for its excellent kinetics (Bougie and Iliuta, 2009; Bougie et al., 2009), CO2
solubility (Bougie and Iliuta, 2010b), and easier regeneration (Bougie and Iliuta, 2010a).
Moreover, the high surface tension (Bougie and Iliuta, 2013a) of this absorbent makes it an
ideal candidate to be used in MC. In a very recent study (Bougie and Iliuta, 2013a)
concerning the surface tension tendency of alkanolamine solutions, we introduced an easy
and very simple graphical molecular classification method that could be used to identify
potential amines whose aqueous solutions present surface tensions appropriate for special
gas separation applications (high surface tension required for membrane contactors or low
surface tension required for packed columns). It was shown that the surface tension of
aqueous solutions increases when the solute molecule presents a low number of carbon
atoms (smaller molecules) and a higher number of hydrophilic groups (like hydroxyl or
232
amino). Following this method, aqueous solutions of derivatives of glycerol (1-amino-2,3-
propanediol and 2-amino-1,3-propanediol) obtained by the substitution of one hydroxyl by
an amino group are possibly the smallest alkanolamines (3 carbons) having 3 hydrophilic
groups and should therefore have very high surface tensions. Among these two potential
compounds for CO2 absorption using membrane contactors, 2-amino-1,3-propanediol
(Figure 8.1), called Serinol, seemed to be the most interesting amine because the amino
group is located between two hydroxyl groups, thus creating some steric hindrance around
it. According to Sartori’s definition (Sartori and Savage, 1983), Serinol cannot be defined
as a primary sterically hindered alkanolamine (SHA) like AHPD, which can offer it the
advantage of a much better kinetics toward CO2 compared to SHA, but it is nevertheless
more hindered than MEA. This should reduce the carbamate formation and be beneficial
for CO2 absorption capacity and regeneration performance.
Figure 8.1. Structure of monoethanolamine (MEA), 2-amino-1,3-propanediol (Serinol) and 2-amino-2-
hydroxymethyl-1,3-propanediol (AHPD).
A literature survey revealed that there is very limited information concerning the
properties of Serinol aqueous solutions. Fernandes et al. (2012) reported the protonation
constant of various amines or alkanolamines (including Serinol) from temperatures T =
(288 to 318) K. Puxty et al. (2009b) made a screening study and reported the approximate
CO2 absorption capacity and initial absorption rate of 76 amine moieties including Serinol.
No carbamate formation was found for Serinol in the kinetic experiments of Conway et al.
(2013). However, the authors performed their experiments at very low absorbent
concentration and only mentioned that data at higher concentrations could be different.
In this context, the aim of this work is to investigate the potential of Serinol as an
efficient CO2 absorbent. Primarily, CO2 solubility measurements were performed from T =
233
(313.15 and 373.15) K. Physical properties of aqueous Serinol solutions like density (ρ),
viscosity (µ), and surface tension (σ), also necessary to evaluate its potential use in MC and
to calculate other properties such as liquid diffusivities and reaction rate constants, were
measured over the temperature range of T = (293.15 to 313.15) K. The solution
concentration range considered in this work was from a molality m = (0.953 to 4.693)
mol·kg-1, which corresponds to a solution of (8 to 30) mass %. To our best knowledge,
similar data are not available in the open literature.
8.2. Experimental section 8.2.1 Reagents
Aqueous amines solutions were prepared by gravimetric method using distilled
water and the following amines: 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD, CAS
No. 77-86-1), piperazine (Pz, CAS No. 110-85-0) and 2-amino-1,3-propanediol (Serinol,
CAS No. 534-03-2). The purity and the source of these amines are given in Table 8.1. The
amines were used without further purification. A Mettler AE240 balance with a precision
of ± 1·10-4 g was used to prepare the solutions and the uncertainties of the reported
concentrations were calculated to be less than m = 0.001 mol·kg-1.
Table 8.1. Chemicals information.
Chemical name Source Minimal Mass Fraction Purity
2-amino-2-hydroxymethyl-1,3-propanediol Laboratoire MAT
0.999
piperazine Laboratoire MAT
0.99
2-amino-1,3-propanediol Canchemia 0.99 water VWR 0.9999999 carbon dioxide Praxair 0.999a
a Minimal mole fraction purity
234
8.2.2 Apparatus and Procedures
8.2.2.1 Density and viscosity of solutions
Density and viscosity were measured following the procedures described in a
previous work (Bougie et al., 2009). Densities of aqueous Serinol solutions were measured
with a calibrated pycnometer having a bulb volume of 1·10-5 m3 and a Mettler AE240
balance with a precision of ± 1·10-4 g. Kinematic viscosities of solutions were measured
with a Cannon-Fenske viscometer. Measurements were performed in a water bath whose
temperature was kept constant within ± 0.1 K. Kinematic viscosities were calculated from
the efflux times measured with an electronic stopwatch with a precision of 0.01 s. Dynamic
viscosities were calculated by multiplying the kinematic viscosities by the corresponding
densities of solutions. Data were obtained at temperatures of T = (293.2, 303.2, 307.2,
313.2) K and for solutions with concentrations of m = (0.953, 2.052, 3.464, and 4.693)
mol·kg-1. The uncertainties of the measured densities and viscosities were calculated to be
within ± 0.06 kg·m-3 and ± 0.008 mPa·s, respectively.
8.2.2.2 Surface tension of solutions
Surface tension data were measured at T = (298.2 and 313.2) K using an optical
contact angle analyzer (OCA 15 Plus, Future Digital Scientific Corp, USA) based on the
pendant drop method. Droplet geometry was analysed by digitizing the image from a
camera and the surface tension was calculated by the device’s software. More details and
the complete method description can be found elsewhere (Bougie and Iliuta, 2013a).
Solutions of concentration of m = (0.959, 2.151, 3.459, and 4.745) mol·kg-1 were used and
the uncertainties of the measured values were found to be ± 0.2 mN·m-1 at T = 298.2 K and
± 0.3 mN·m-1 at T = 313.2 K.
8.2.2.3 CO2 Solubility measurements
As a full description of the apparatus and the procedure to determine the CO2
solubility in aqueous alkanolamine solutions can be found in Bougie and Iliuta (2010b),
only a summary is given here. The experimental setup for the solubility measurements used
in this work consists of a titanium equilibrium cell (Armines, France) equipped with a
stirring magnetic rod, two absolute pressure transducers (Druck PTX-611, 0-100 kPa and 0-
235
16000 kPa) and two 100 ohms platinum resistance thermometers. The cell is placed in a
modified XU027 laboratory oven (France Etuves) for temperature control. Liquid insertion
into the equilibrium cell was made with a variable volume press (internal piston diameter of
3.002·10-2 m). Gas addition into the cell was made by a thermostated small gas cylinder
with an internal volume of about 7·10-5 m3, equipped with an absolute pressure transducer
(Druck PTX-611, 0-16000 kPa).
A standard experimental run consisted of a sequence of successive steps. First, the
amines aqueous solution was degassed under vacuum, transferred inside the variable
volume liquid press, and subsequently, transferred in the equilibrium cell previously
brought to vacuum. The equilibrium cell was heated to the desired temperature and the
vapour pressure of the solution was measured by the low pressure transducer. This was
followed by the introduction of CO2 (purity and source given in Table 8.1) in the
equilibrium cell from the small gas cylinder. The mole number of gas introduced in the cell
was calculated by using the cylinder volume and temperature, as well as the observed
pressure drop in the cylinder after the gas transfer. Equilibrium was reached when the
pressure inside the equilibrium cell was varying less than 0.5 % for at least 30 minutes. The
difference between the introduced and the remaining CO2 mole number in the head space of
the equilibrium cell was then calculated and used to determine the concentration of
absorbed gas in the solution.
In the present work, CO2 solubility in Serinol solutions was determined at T =
313.15 K for solution concentrations of m = (0.953, 2.097, 3.464, and 4.704) mol·kg-1 and
at T = (343.15 and 373.15) K for concentration of m = 4.704 mol·kg-1. For comparison
purpose, the CO2 solubility was also evaluated in m = (2.712 + 1.161) mol·kg-1 AHPD + Pz
solution (total amine content of 30 mass %) at T = (313.15 and 373.15) K.
8.3. Results and Discussion 8.3.1. Density and viscosity of solutions
Density and viscosity data of aqueous Serinol solutions are presented in Table 8.2.
As can be seen in Figures 8.2 and 8.3, the values of both properties increase, as expected,
with the increase of Serinol concentration or the decrease in temperature.
236
Table 8.2. Experimental Values of Density ρ and Viscosity µ of Aqueous Serinol Solutions determined at Temperature T, Amine-Molality m and Atmospheric
Pressure (P = 101.3 kPa)a
T/K m/mol·kg-1 ρ/kg·m-3 µ/mPa·s 293.2 0.953 1012.63 1.244 293.2 2.052 1026.97 1.606 293.2 3.464 1041.67 2.157 293.2 4.693 1052.77 2.784 300.2 0.953 1010.28 1.056 300.2 2.052 1024.18 1.335 300.2 3.464 1038.80 1.769 300.2 4.693 1050.01 2.261 307.2 0.953 1008.07 0.896 307.2 2.052 1021.61 1.118 307.2 3.464 1036.80 1.458 307.2 4.693 1047.38 1.845 313.2 0.953 1005.49 0.800 313.2 2.052 1019.22 0.983 313.2 3.464 1033.49 1.275 313.2 4.693 1044.53 1.588
a Standard uncertainties u are u(T) = 0.1 K, u(m) = 0.001 mol·kg-1, and the combined expanded uncertainties Uc are Uc(ρ) = 0.06 kg∙m‐3, and Uc(µ) = 0.008 mPa∙s (level of confidence = 0.95).
Eqs. (8.1) and (8.2), where m/mol·kg-1 is the Serinol molality defined as mol of
Serinol per kilogram of water and T/K the absolute temperature, were found to correlate our
data very satisfactorily with average relative deviations (A.R.D.) of 0.03% and 0.33% for
density and viscosity, respectively. The regressed coefficients for these equations are given
in Table 8.3.
( )1
3 2 2
0/ kg m i
i i ii
a b m c m Tρ −
=
⋅ = + ⋅ + ⋅ ⋅∑ (8.1)
12 2
0ln( / mPa s) ii
i i ii
da b m c m TT
µ=
⋅ = + ⋅ + ⋅ + ⋅
∑ (8.2)
237
980
1000
1020
1040
1060
0 1 2 3 4 5
ρ / k
g·m
-3
m / mol·kg-1
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0 1 2 3 4 5
µ/ m
Pa·s
m / mol·kg-1
Figure 8.2. Densities of aqueous Serinol solutions as a function of amine-molality m and temperature T: , 293.2 K; , 300.2 K; , 307.2 K; , 313.2 K. Dotted lines correspond
to calculated values using Eq. (8.1) and parameters given in Table 8.3.
Figure 8.3. Viscosities of aqueous Serinol solutions as a function of amine-molality m and temperature T: , 293.2 K; , 300.2 K; , 307.2 K; , 313.2 K. Dotted lines correspond
to calculated values using Eq. (8.2) and parameters given in Table 8.3.
238
Table 8.3. Values of the Regressed Coefficients for Eqs (8.1) to (8.3).
coefficient Eq. (8.1) Eq. (8.2) Eq. (8.3) a0 1045.3 -16.3524 94.9 b0 14.9 0.4296 -0.38 c0 0 0 -0.11 d0 - 3910.5 - a1 -5.4·10-4 3.5106·10-5 -2.57·10-4
b1 0 -2.286·10-6 0 c1 -8.4·10-6 -3.9·10-8 1.3·10-6
d1 - 0 - A.R.D 0.03% 0.33%* 0.04%
*Measured on µ
8.3.2. Surface tension of solutions
Surface tension data of aqueous Serinol solutions are presented in Table 8.4. As
expected, the surface tension decreases with the increase of concentration and temperature
(Figure 8.4). Eq. (8.3), in which m/mol·kg-1 is the amine molality and T/K the absolute
temperature, was found to correlate our experimental data very adequately as an average
relative deviation (A.R.D.) as low as 0.04% was obtained. The regressed coefficients are
indicated in Table 8.3. A comparison between the Serinol surface tension value at T =
298.2 K and m = 4.745 mol·kg-1 (70.4 mN·m-1) with those of common alkanolamines
solutions like MEA (63.9 mN·m-1) (Bougie and Iliuta, 2013a), diethanolamine (DEA, 60.8
mN·m-1) (Vazquez et al., 1996), and N-methyldiethanolamine (MDEA, 53.5 mN·m-1)
(Alvarez et al., 1998) at the same conditions of mass % concentration and temperature
demonstrates that this absorbent can be an interesting potential candidate for MC
applications where a high surface tension is of prime importance. This high surface tension
value for Serinol validates the predictive capacity of the recently developed surface tension
estimation method (Bougie and Iliuta, 2013a).
( )1
1 2 2
0/ mN m i
i i ii
a b m c m Tσ −
=
⋅ = + ⋅ + ⋅ ⋅∑ (8.3)
239
66
67
68
69
70
71
72
73
0.0 1.0 2.0 3.0 4.0 5.0
σ/ m
N·m
-1
m / mol·kg-1
Table 8.4. Experimental Values of Surface Tension σ of Aqueous Serinol Solutions determined at Temperature T, Amine-Molality m and Atmospheric Pressure (P =
101.3 kPa)a
T/K m/mol·kg-1 σ/mN·m-1
298.2 0.959 71.7 298.2 2.151 71.3 298.2 3.459 70.8 298.2 4.745 70.4 313.2 0.959 69.4 313.2 2.151 69.0 313.2 3.459 68.6 313.2 4.745 68.3
a Standard uncertainties u are u(T) = 0.1 K, u(m) = 0.001 mol·kg-1, and the combined expanded uncertainties Uc are Uc(σ) = 0.2 mN∙m‐1 at 298.15 K, and Uc(σ) = 0.3 mN∙m‐1 at 313.15 K (level of confidence = 0.95).
Figure 8.4. Surface tensions of aqueous Serinol solutions as a function of amine-molality m and temperature T: , 298.2 K; , 313.2 K. Dotted lines correspond to calculated values
using Eq. (8.3) and parameters given in Table 8.3.
240
8.3.3. CO2 Solubility
8.3.3.1 Solution concentration effect on solubility
CO2 solubility in a solution is among the most important properties to be considered
in the evaluation of its potential to separate CO2 in industrial applications. CO2 solubility
was first determined in Serinol solutions from amine molality m = (0.953 to 4.704) mol·kg-
1 at T = 313.15 K to evaluate the effect of the amine concentration on gas solubility. The
results are indicated in Table 8.5.
Table 8.5. Experimental Values of CO2 Solubility mCO2 at Temperature T = 313.15 K in Aqueous Serinol Solutions of Amine-Molality m a
m/mol·kg-1 = 0.953 m/mol·kg-1 = 2.097 PCO2/kPa u(PCO2)/kPa mCO2/mol·kg-1 Uc(mCO2) PCO2/kPa u(PCO2)/kPa mCO2/mol·kg-1 Uc(mCO2)
1.346 0.001 0.207 0.001 1.500 0.001 0.457 0.003 5.063 0.004 0.375 0.003 7.628 0.006 0.873 0.004 34.47 0.03 0.519 0.004 26.32 0.02 1.123 0.005 77.24 0.06 0.614 0.005 67.95 0.05 1.300 0.007 174.2 0.1 0.717 0.007 144.6 0.1 1.447 0.008 326.4 0.3 0.805 0.008 346.1 0.3 1.65 0.01 590.9 0.5 0.90 0.01
m/mol·kg-1 = 3.464 m/mol·kg-1 = 4.704
PCO2/kPa u(PCO2)/kPa mCO2/mol·kg-1 Uc(mCO2) PCO2/kPa u(PCO2)/kPa mCO2/mol·kg-1 Uc(mCO2)
0.5239 0.0004 0.356 0.001 0.8511 0.0007 0.647 0.002 1.276 0.001 0.745 0.003 2.254 0.002 1.315 0.003 3.060 0.002 1.176 0.004 7.475 0.006 1.994 0.005 10.675 0.009 1.578 0.006 68.42 0.05 2.678 0.007 47.42 0.04 1.928 0.007 186.3 0.1 3.006 0.008 105.59 0.08 2.131 0.009 413.6 0.3 3.33 0.01 227.2 0.2 2.35 0.01 464.8 0.4 2.60 0.01
a Standard uncertainties u are u(T) = 0.01 K, u(m) = 0.001 mol·kg-1.
From Figure 8.5a, it can be seen that the relationship between CO2 partial pressure
and CO2 solubility in the aqueous phase expressed as a semi-log plot, displays a linear
trend, similar to that observed in the literature for MEA (Shen and Li, 1992). For a given
CO2 partial pressure, it can be seen that a more amine-concentrated solution can be
industrially a more attractive option as the amount of CO2 absorbed per kg of solvent is
241
more important, thus reducing the need of a large liquid capacity, as well as the size of
liquid-related equipment (pumps).
Figure 8.5a. CO2 molality-based solubility in aqueous Serinol solutions at T = 313.15 K as a function of Serinol molality m: , 0.953 mol·kg-1; , 2.097 mol·kg-1; , 3.464 mol·kg-
1; , 4.704 mol·kg-1. Dotted lines represent the trends only.
Moreover, the loading-based solubilities (Figure 8.5b) where α/molCO2·molamine-1 is
the CO2 loading in solution (Eq. 8.4) show that, for a given CO2 partial pressure, all
solutions present almost the same loading capacity only up to a maximal value of around
0.55. This loading value of 0.55 seems to confirm that the CO2 absorption in Serinol leads
to carbamate formation at low CO2 partial pressures, similar to unhindered primary
alkanolamines like MEA, characterized by a theoretical chemical loading of around 0.5
when carbamate is the main product of reaction in solution (Bougie and Iliuta, 2012). For
loading above 0.5-0.55, it is expected that the hydrolysis of the Serinol carbamates and the
physical CO2-bicarbonate equilibrium contribute to bicarbonate formation. The salting-out
effect (Figure 8.5b) could then explain the solubility data above 0.55: a less amine-
concentrated solution will have a better solubility than a more amine-concentrated one.
2 2
2
-1CO CO solvent1
CO amine -1amine solvent
/ mol kg/ mol mol =
/ mol kgmm
α − ⋅⋅
⋅ (8.4)
242
Figure 8.5b. CO2 loading-based solubility in aqueous Serinol solutions at T = 313.15 K as a function of Serinol molality m: , 0.953 mol·kg-1; , 2.097 mol·kg-1; , 3.464 mol·kg-
1; , 4.704 mol·kg-1.
8.3.3.2 Temperature effect on solubility
A second set of CO2 solubility experiments was performed with 30 mass % total
amine solutions of Serinol at T = (343.15 and 373.15) K and AHPD + Pz at T = (313.15 and
373.15) K. Experimental data are indicated, respectively, in Tables 8.6 and 8.7. First, the
temperature effect on CO2 solubility in Serinol solutions can be seen in Figure 8.6. For a
given CO2 partial pressure, the solubility decreases with the increase of temperature. Based
on CO2 solubility in aqueous 30 mass % MEA solution (Shen and Li, 1992) and the present
experimental data, an analysis of the potential CO2 cyclic capacity of Serinol, AHPD + Pz
and MEA solutions between T = (313.15 and 373.15) K can be made. To estimate the CO2
cyclic capacity, defined as the difference between rich loading (CO2 loading after
absorption) and lean loading (CO2 loading after regeneration), α at T = 373.15 K at PCO2 =
5 kPa was subtracted from α at T = 313.15 K and PCO2 = 20 kPa. It was considered that
data for CO2 solubility at T = 373.15 K can simulate the conditions corresponding to lean
loaded (regenerated) solutions. From Figure 8.7, it can be seen that the cyclic capacity
(represented by the difference in α between the ends of the lines) of MEA solution is 0.263.
0.1
1.0
10.0
100.0
1000.0
0.0 0.2 0.4 0.6 0.8 1.0
P CO
2/ k
Pa
α
243
For comparison, the cyclic capacities of Serinol and AHPD + Pz solutions are, respectively,
0.416 and 0.449. An increase of 58 % (Serinol) and 71 % (AHPD + Pz) in respect to MEA
indicates that these two solutions show great potential to replace MEA in industrial
applications.
Table 8.6. Experimental Values of CO2 Solubility mCO2 at Temperature T in Aqueous Serinol Solutions of Amine-Molality m = 4.704 mol·kg-1 a
T = 343.15 K T = 373.15 K PCO2
/kPa u(PCO2)/kPa mCO2
/mol·kg-1 Uc(mCO2) PCO2
/kPa u(PCO2)/kPa mCO2
/mol·kg-1 Uc(mCO2)
2.439 0.002 0.680 0.002 4.038 0.003 0.289 0.001 13.17 0.01 1.409 0.003 14.95 0.01 0.628 0.003 51.99 0.04 1.982 0.005 32.99 0.03 0.923 0.004 95.95 0.08 2.200 0.006 59.16 0.05 1.174 0.006 202.4 0.2 2.786 0.008 111.39 0.09 1.460 0.007
202.7 0.2 1.738 0.009 a Standard uncertainties u are u(T) = 0.01 K, u(m) = 0.001 mol·kg-1.
Table 8.7. Experimental Values of CO2 Solubility mCO2 at a Temperature T in
Aqueous AHPD + Pz Solutions of Amine-Molality m = (2.712 + 1.161) mol·kg-1 a T = 313.15 K T = 373.15 K
PCO2/kPa u(PCO2
)/kPa mCO2/mol·kg-1 Uc(mCO2
) PCO2/kPa u(PCO2
)/kPa mCO2/mol·kg-1 Uc(mCO2
)
0.8750 0.0007 0.566 0.001 4.758 0.003 0.265 0.001 2.960 0.002 1.107 0.003 11.461 0.009 0.532 0.002 10.014 0.009 1.669 0.004 27.99 0.02 0.752 0.004 28.16 0.02 2.16 0.01 54.27 0.04 0.932 0.005 66.64 0.05 2.59 0.02 94.32 0.08 1.098 0.006 132.02 0.09 2.92 0.02 151.9 0.1 1.260 0.006 229.4 0.2 3.17 0.03
a Standard uncertainties u are u(T) = 0.01 K, u(m) = 0.001 mol·kg-1.
244
0.1
1.0
10.0
100.0
1000.0
0.0 1.0 2.0 3.0 4.0
P CO
2/ k
Pa
mCO2/ mol.kg-1
Figure 8.6. CO2 solubility in an aqueous Serinol solution of m = 4.704 mol·kg-1 as a function of temperature T: , 313.15 K; , 343.15 K; , 373.15 K.
In addition to the discussion given in section 8.3.3.1 (Figure 8.5b) concerning
Serinol carbamate formation, the analysis of Figure 8.7 provides some information about
the Serinol carbamate stability. The CO2 solubility data of Serinol at T = 373.15 K are
much more closer to those corresponding to the sterically hindered based AHPD + Pz
solution than to MEA ones, showing that Serinol carbamates in solution can be easily
regenerated compared to those formed in MEA solution. It can be concluded that the
addition of an extra hydroxymethyl group to the nitrogen alpha carbon of MEA to form the
Serinol molecule decreases significantly the Serinol carbamate stability compared to MEA
carbamate.
245
0.1
1.0
10.0
100.0
1000.0
0.0 0.2 0.4 0.6 0.8
P CO
2/ k
Pa
α
Figure 8.7. Comparison of CO2 solubility data in Serinol (black symbols), AHPD + Pz (grey symbols) and MEA (white symbols, (Shen and Li, 1992)) solutions at temperature T
= 313.15 K (circular symbols) and 373.15 K (square symbols). Cyclic capacities are represented by the difference in α between the ends of the lines: Serinol (0.416, large
dotted line), MEA (0.263, small dotted line), and AHPD + Pz (0.449, solid line).
8.4. Conclusions In this work, Serinol aqueous solutions were characterized through density,
viscosity, surface tension, and CO2 solubility measurements in order to evaluate the
potential of aqueous solutions of this amine for CO2 removal from gas mixtures. Density,
viscosity and surface tension data were correlated with mean relative deviations of (0.03,
0.33 and 0.04) %, respectively. The higher surface tension data of Serinol solutions,
compared to conventional alkanolamines, make them very suitable for CO2 absorption
using membrane contactors. Moreover, surface tension data validated the predictive
capacity of a previously developed method for estimating the surface tension of amines,
alcohols and alkanolamines aqueous solutions. CO2 solubility measurements showed that
the formation of Serinol carbamates is similar to other primary unhindered alkanolamines
like MEA. However, Serinol carbamates can be easily regenerated. The CO2 cyclic
246
capacity of Serinol was found to be 58 % higher than that of MEA. On the whole, the
experimental results confirmed the potential capacity of this alkanolamine to be used for
CO2 removal especially in membrane contactors.
247
248
A good stability and resistance to degradation is another important feature absorbents
should have for being used in the CO2 absorption process. In this context, the following
chapter evaluates the stability of aqueous AHPD + Pz solution to thermal and oxidative
degradation, in the absence and the presence of CO2, and compares the results with those
obtained for AMP (the most studied sterically hindered alkanolamine), MEA (the
benchmark amine used in CO2 capture) and Serinol (a potential alkanolamine for CO2
capture in MC investigated in Chapter 8)
249
Chapter 9. Stability of aqueous amine solutions to thermal and oxidative degradation in the absence and the presence of CO2
Résumé
La stabilité à la dégradation thermique et oxydative de cinq solutions aqueuses d’amine simple (2-amino-2-hydroxyméthyl-1,3-propanediol (AHPD), pipérazine (Pz) 2-amino-1,3-propanediol (Sérinol), 2-amino-2-méthyl-1-propanol (AMP) et monoéthanolamine (MEA)) et une solution mixte (AHPD + Pz), en présence ou non de CO2, a été étudiée par chromatographie en phase liquide. Il a été observé que la présence d’O2 et de CO2 a influencé significativement la dégradation thermique. Les amines à encombrement stérique étudiées (AMP et AHPD) ont démontré être plus résistante à la dégradation thermique que les amines (non encombrées) conventionelles. Par contre, en absence de CO2, l’oxygène les dégrade plus significativement. L’addition de Pz à la solution d’AHPD a démontré réduire la dégradation oxydative de cette dernière. En conclusion, la solution AHPD + Pz s’est révélée être un absorbant potentiellement intéressant pour remplacer la MEA dans les procédés de capture du CO2 industriels, surtout pour une application dans des contacteurs à membrane.
250
Abstract
The stability to thermal and oxidative degradation of five single amine aqueous solutions (2-amino-2-hydroxymethyl-1,3-propanediol (AHPD), piperazine (Pz), 2-amino-1,3-propanediol (Serinol), 2-amino-2-methyl-1-propanol (AMP) and monoethanolamine (MEA)) and one mixed aqueous solution (AHPD + Pz), in the absence and the presence of CO2, was investigated by high-performance liquid chromatography. The results showed that the presence of O2 and CO2 influenced significantly the degree of thermal degradation. The sterically hindered alkanolamines investigated (AMP and AHPD) were found more resistant to thermal degradation than conventional (unhindered) amines. However, in the absence of CO2, the oxygen degraded them more significantly. The addition of Pz to AHPD solution reduces the AHPD oxidative degradation. It was concluded that AHPD + Pz amine aqueous blend could be a potentially interesting absorbent to replace MEA for industrial CO2 capture applications, especially using membrane contactors.
251
9.1. Introduction For many decades now, the removal of acid gases such as CO2 and H2S in natural
gas sweetening and CO2 capture from fossil-fuel-fired power plants or industrial
applications is of high interest for economic, technical, and environmental concerns (Kohl
and Nielsen, 1997). Among possible techniques to capture or separate CO2 from other
gases, the chemical absorption by aqueous amines solutions is today’s best available
technology (Bernardo et al., 2009) and monoethanolamine (MEA) is considered as the
benchmark amine for this process. More recently, great interest has been given to aqueous
amine mixtures by combining the fast reactivity of primary and secondary amines to the
high absorption capacity and low regeneration cost of ternary and sterically hindered
amines (SHA). In this context, we recently proposed the aqueous mixture of 2-amino-2-
hydroxymethyl-1,3-propanediol (AHPD), a sterically hindered alkanolamine, and
piperazine (Pz), a secondary diamine activator, as a potential alternative absorbent to MEA
solution, for its excellent kinetics (Bougie et al., 2009), CO2 solubility (Bougie and Iliuta,
2010b), and easier regeneration (Bougie and Iliuta, 2010a). Moreover, the high surface
tension of this absorbent makes it an ideal candidate to be used in membrane contactors
(Bougie and Iliuta, 2013a) which were found to be more effective than traditional packed
columns to perform the CO2 absorption (deMontigny et al., 2005).
Another important feature absorbents should have in the CO2 absorption process is a
good stability and resistance to degradation. MEA is known to degrade more significantly
than other conventional alkanolamines and its corrosiveness, which could be increased by
the presence of degradation products, may cause severe damage to process facilities
(Veawab et al., 1997). In the cyclic absorption-regeneration process, the presence of
oxygen in the flue gas and an elevated temperature in the stripper cause, respectively,
oxidative and thermal degradation. The amine degradation may produce several negative
effects in the operation of a gas treating unit such as amine losses, reduction of capture
capacity, foaming and increase of the solution corrosiveness (Freeman et al., 2010; Supap
et al., 2006). Several exhaustive studies and review papers have been published so far on
amine degradation especially concerning 2-amino-2-methyl-1-propanol (AMP),
diethanolamine (DEA), N-methyldiethanolamine (MDEA), MEA and piperazine (Pz)
252
(Gouedard et al., 2012; Islam et al., 2011; Rochelle, 2012). These works, detailing thermal
and oxidative degradation mechanisms and products formed in the presence and absence of
CO2, revealed that the degradation process is a complex phenomenon; each amine can
produce a large quantity of degradation products and by several possible degradation
pathways. From studies evaluating the degradation rate or degradation percentage of
various amines moieties (Freeman et al., 2010; Freeman and Rochelle, 2012a; Lepaumier et
al., 2009a, b), it was observed that Pz and other cyclic amines showed an improved
resistance to degradation compared to aliphatic amines. These studies also demonstrated
that AMP, a sterically hindered alkanolamine, was among the most resistant amines to
degradation.
The available results seem to indicate that the use of aqueous mixtures containing
Pz and sterically hindered alkanolamines like AHPD could be beneficiary to minimise
amine degradation in industrial CO2 capture processes. However, the oxidative and thermal
degradation of AHPD or AHPD + Pz aqueous solutions have never been studied in the
literature to confirm this assumption. As already mentioned before, the aqueous blend
AHPD + Pz was found to possess good absorption capacity, reaction kinetics, regenerative
potential, as well as high surface tension required for use in membrane contactors (Bougie
and Iliuta, 2010a, b, 2013a; Bougie et al., 2009). Therefore, the main objective of this work
is to evaluate the stability of aqueous 23 wt% AHPD and 23 wt% AHPD + 7 wt% Pz
solutions to thermal and oxidative degradation in the absence and the presence of CO2. For
comparison purposes, the stability to degradation of 30 wt% aqueous solution of AMP (the
most studied sterically hindered alkanolamine), of MEA (the benchmark amine used in CO2
capture) and of 2-amino-1,3-propanediol (Serinol; a potential alkanolamine for CO2 capture
assessed in a previous work) (Bougie and Iliuta, 2014a) was also investigated under the
same experimental conditions. The degradation process was monitored by the change of
amine concentration measured by high-performance liquid chromatography.
9.2. Material and methods 9.2.1. Chemicals
Aqueous amines solutions used in this work were prepared by gravimetric method
using distilled water and the following amines: 2-amino-2-hydroxymethyl-1,3-propanediol
253
(AHPD, CAS No. 77-86-1), piperazine (Pz, CAS No. 110-85-0), 2-amino-1,3-propanediol
(Serinol, CAS No. 534-03-2), monoethanolamine (MEA, CAS No. 141-43-5) and 2-amino-
2-methyl-1-propanol (AMP, CAS No. 124-68-5). The purity and the source of these amines
are given in Table 9.1 (the structures of these amines are represented in Figure 9.1). All
amines were used without further purification. A Mettler AE240 balance with a precision
of ± 1×10-4 g was used to prepare the solutions and the uncertainties of the reported
concentrations were calculated to be less than 0.01 wt%.
Table 9.1. Amines studied in this work. Chemical Source Purity 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD) Laboratoire MAT 0.999 piperazine (Pz) Laboratoire MAT 0.99 2-amino-1,3-propanediol (Serinol) Canchemia 0.99 monoethanolamine (MEA) Sigma-Aldrich 0.99 2-amino-2-methyl-1-propanol (AMP) Laboratoire MAT 0.95
Figure 9.1. Amine structures.
254
9.2.2. Thermal degradation: typical experimental run
In thermal degradation experiments, 0.635 cm outside diameter stainless steel tubes
having an internal volume of 2.5 ml were filled with amine solutions and closed with two
end caps. Several tubes were prepared for each tested solution and placed in an oven at 403
K. This temperature was chosen to be slightly higher than the conventional stripper
temperature range (373-393 K) (Islam et al., 2011) in order to accelerate the amine
degradation rate. The tubes were individually removed from the oven (during 14 days),
cooled to avoid further degradation (Wang and Jens, 2012) and the content was transferred
to screw-cap glass vials stored at 277 K until the analysis of the amine content.
Figure 9.2. Experimental setup for degradation involving gas introduction.
9.2.3. Combined thermal and oxidative: typical experimental degradation run
All degradation experiments involving gas introduction were done in 194 ml
insulated stainless steel reactors as shown in Figure 9.2. Each reactor was equipped with E-
type thermocouple, inlet and outlet gas valves, a liquid sampling valve and a 50 psig relief
valve. A PTFE coated magnetic stirrer inside the reactor allowed solution concentration
homogeneity and a temperature stability of ±2 K. At the beginning of a degradation run,
150 g of aqueous amine solution was introduced in the reactor at room temperature. The
reactor was sealed and oxygen was injected to purge the air through the gas outlet valve.
This valve was then closed to pressurise the setup at 446 kPa and the reactor was heated at
255
403 K under stirring. The combined effect of oxygen and high temperature on amine
degradation could then be studied in these experiments since, as pointed out by Wang and
Jens (2012), any absorbed oxygen in the absorber is carried over to the stripper and has the
potential to cause oxidation at high temperature. A sample of 2 ml of the partially degraded
amine solution was withdrawn each day through the liquid sampling valve after allowing 1
ml liquid discharge to clean the inside of the sampling tube. The samples were stored in
small screw-cap glass vials at 277 K until the analysis for the amine content. To keep a
constant pressure after each sampling, the reactor was pressurized again with O2 up to the
check valve limit. Taking into consideration the water vapor pressure, the oxygen partial
pressure in the reactor was calculated to be close to 189 kPa, which should favor an
accelerated oxidative degradation. The experiments were performed for 14 days.
9.2.4. Degradation in the presence of CO2
The degradation experiments involving CO2 were performed in the 194 ml stainless
steel reactors (Figure 9.2), in a similar way to that described in the section 9.2.3. First, 150
g of aqueous amine solution was introduced in the reactor and allowed to saturate under a
CO2 partial pressure of 20 kPa at 298.15 K. Two kinds of degradation experiments were
performed: thermal + oxidative + CO2 and thermal + CO2. In the thermal + oxidative + CO2
degradation experiment, oxygen was introduced into the reactor containing the CO2
saturated solutions to reach a total pressure of 446 kPa and the reactor was then heated to
403 K. Samples were withdrawn every day for 14 days and the reactor was pressurised with
oxygen after each sampling. In the case of thermal + CO2 experiments, the CO2 saturated
solutions were directly heated to 403 K where CO2 partial pressure reached 189 kPa. No
CO2 was added after each sampling to avoid modifications of the gas phase composition.
All samples were kept in closed glass vials at 277 K until analysis.
9.2.5. HPLC analysis
All samples analyses of this work were performed by liquid chromatography. The
HPLC, from Mandel Scientific Company Inc., was equipped with an inline mobile phase
vacuum degasser, an autosampler, a column oven and a refractive index detector (RID). To
analyse the remaining amine concentration of the samples, a 4.6 × 100 mm universal cation
256
HR column was selected and an aqueous mobile phase of 20 mM methanesulfonic acid was
used. For typical analysis, samples kept at 277 K were brought at 298 K and diluted with
distilled water by a factor of 25 for AHPD or Pz solutions, of 35 for AMP solutions, of 40
for Serinol solutions and of 50 for MEA solutions. 5 µl of these diluted samples was
injected. All analyses were done using a simple isocratic mode in which 100% of the
mobile phase was flowing at a rate of 1 mL/min. The RID optical unit was set at a
temperature of 313 K and operated under positive mode. Samples were analysed at least 3
times to check the reproducibility which was found to be ±1%.
9.3. Results 9.3.1. Percentage of amine loss
The stability to degradation of six amine systems (five single aqueous amine
solutions (MEA, AMP, Serinol, AHPD and Pz) and one mixed AHPD + Pz aqueous
solution) was investigated at 403 K mainly under three degradation conditions: thermal,
thermal + oxidative and thermal + oxidative + CO2. The initial concentrations of the
solutions at 298 K were 30 wt% for AMP (3.35 M), MEA (4.95 M) and serinol (3.46 M),
23 wt% for AHPD (1.98 M), 7 wt% for Pz (0.76 M) and 23 + 7 wt% for the AHPD + Pz
(2.03 + 0.84 M) mixed solution. The remaining amine concentration for each studied
system and degradation condition was tracked by HPLC analysis during the experiments
(up to 14 days) and the percentages of amines loss at the end of this period are shown in
Figure 9.3. There is one exception: the thermal + oxidative degradation of AMP was
stopped after 7 days due to high amine loss. The columns identified as “AHPD mixt.” and
“Pz mixt.” represent, respectively, the amine loss for AHPD and Pz from the blend
solution.
9.3.2. Amine degradation first-order rate constant
As mentioned by Freeman and Rochelle (2012a), the loss of amine during
degradation is often well represented by a first-order dependence on the amine
concentration.
257
Figure 9.3. Amine degradation loss after 14 days (except for AMP).
The experimental amine concentration profiles were therefore correlated by an
exponential equation where the amine concentration (CAmine) is a function of the initial
amine concentration (C0,Amine), a first-order rate constant (k1) and time (t):
t-k e C C 1,0 AmineAmine = (9.1)
All determined first-order rate constants were brought together in Table 9.2 and an example
of amine concentration profiles is shown in Figure 9.4 for Pz. The first-order rate constants
allow the calculation of the amine concentrations with time and, as these constants are
occasionally available in literature for some amines, they will allow the comparison of data
obtained in this work with other degradation studies.
The first-order constants for all amine systems that could be well correlated with the
exponential equation (Eq. (9.1)) are indicated in Table 9.2. This was however not the case
for AMP concentrations determined under the thermal + oxidative and thermal + oxidative
+ CO2 conditions, as it can be seen in Figure 9.5, where the presence of oxygen seemed to
accelerate the amine degradation with time and the exponential regression could not be
applied. For this reason, no k1 value was given for these two conditions.
258
Table 9.2. Degradation first-order rate constants. Amine system k1 / 10-3 day-1
Thermal Thermal + oxidative
Thermal + oxidative + CO2
MEA 5.27 12.12 4.12 AMP 1.13 (1.61)* - - Serinol 9.75 25.33 27.87 AHPD 1.79 (1.86)* 59.11 2.21 Pz 2.87 12.32 9.00 AHPD mixt. 3.49 12.41 1.94 Pz mixt. 4.34 17.49 11.17
Figure 9.4. Effect of process conditions on Pz degradation. Solid lines are calculated using
Eq. (9.1) and constants from Table 9.2.
In addition to the results presented so far, two supplementary degradation
experiments were performed to analyse the thermal + CO2 degradation of both SHA (AMP
and AHPD) aqueous solutions. The aim of these tests was to evaluate if CO2 alone has the
same significant detrimental effect on thermal SHA degradation as oxygen (Figure 9.3). An
amine loss of only 2.2% and 2.6% was obtained after 14 days for AMP and AHPD,
respectively (Figure 9.3), compared to 63% and 56% for thermal + oxidative degradation
259
(as mentioned before, the percentage for AMP corresponds to a 7 days experiment). The
corresponding first-order rate constants are given in Table 9.2.
Figure 9.5. Effect of process conditions on AMP degradation. Solid line is calculated using
Eq. (9.1) and constant in Table 9.2, whereas dashed lines are for trend only.
9.3.3. Qualitative observations
In addition to data presented in Figure 9.3 and Table 9.2, some qualitative
observations were made during the experiments and they appear to confirm the
experimental results. It was observed that the color of the samples followed the degree of
amine degradation. Except for Serinol, the solutions turned progressively from clear to
yellow, orange, brown and finally black, depending of their degradation degree. Serinol
solutions samples took a purple shade before turning black. Similar observations were
made by Reza and Trejo (2006) from degradation experiments involving AMP, DEA and
MDEA. MEA degradation, as it can be seen in Figure 9.3, remained relatively at a low
level for all degradation experiments, what was also confirmed by the final yellow or
slightly orange color of the samples at the end of the period of 14 days. However, only this
amine solution corroded extensively the inside of the stainless steel reactors, thus validating
the aqueous MEA solution reputation of being a very corrosive media.
260
9.4. Discussions 9.4.1. Effect of process conditions on degradation of single amine aqueous systems
The degree of degradation of several amines was studied at temperature and oxygen
partial pressure higher than usually used in industrial applications. These conditions were
selected to accelerate amine degradation rates and to reduce the length of the experiments
because industrial amine degradation is usually a slow phenomenon (Lepaumier et al.,
2009b).
9.4.1.1 Pure thermal degradation
As a temperature of 403 K was used in all experiments, the thermal induced
degradation can serve as a base case to analyse the effect of other experimental parameters
on amine degradation (presence of oxygen and/or of CO2). It can be seen in Figure 9.3 that
the pure thermal degradation is the highest for Serinol, followed by MEA (both
representing primary “conventional” alkanolamines), Pz (a cyclic diamine) and AMP and
AHPD (two primary sterically hindered alkanolamines). This ranking is in agreement with
the results of Lepaumier et al. (2009b), one of the rare studies reporting pure thermal amine
degradation percentage data. As mentioned by the authors, a possible radical mechanism is
assumed to occur in this type of degradation, causing dealkylation, dimerization and
cyclisation of primary amines. In the present case, it seems that dealkylation was the main
pathway for the thermal degradation and this could explain the higher degradation rate of
MEA and Serinol over sterically hindered alkanolamines. The steric hindrance prevents the
dealkylation process, as the formation of one radical on the alpha-carbon of the nitrogen
atom is essentially impossible due to the absence of hydrogen which inhibits the C-N bond
cleavage (Lepaumier et al., 2009b). The secondary diamine cyclic structure of Pz (Figure
9.1) limited the thermal degradation to a lower degree in comparison with MEA, as
previously mentioned in the literature (Freeman et al., 2010). Based on the results of
Lepaumier et al. (2009b) and Wang and Jens (2012), k1 values of 4 × 10-3 and 0.4 × 10-3
day-1 were estimated for MEA and AMP, respectively, assuming a first-order amine
degradation behaviour and were found to be close to the values obtained in this work
(Table 9.2).
261
9.4.1.2. Oxygen effect on amine degradation
In addition to thermal degradation experiments, degradation tests were performed in
the presence of oxygen (partial pressure of around 189 kPa). As expected, the presence of
oxygen increased all amine thermal degradation percentages, as seen in Figure 9.3. This
source of amine degradation was found to be more significant compared to pure thermal
degradation, all amine degradation percentages increasing more than twice. A similar trend
can be observed from the results by Lepaumier et al. (2009b) for all 12 amine compounds
studied in their work. As it can be seen in Figure 9.3, AMP is the amine that unexpectedly
degraded the most in the presence of oxygen, contrary to the general assumption that
sterically hindered alkanolamines were more stable to oxidative degradation than other
conventional amines (Islam et al., 2011). In addition, AHPD, the other tested SHA which
presents a much higher degree of sterically hindrance compared to AMP, also degraded to a
large extent. The general oxidative degradation ranking for single amine is AMP >> AHPD
> Serinol > MEA ≈ Pz. Based on these results, it seems that the presence of oxygen could
promote demethylation type reactions, as AMP is the only tested amine having free methyl
groups (2) in its structure (Figure 9.1). In the same way, Lepaumier et al. (2009b) found
that a higher amount of methylated compounds was produced under an oxidative
environment by AMP in comparison to MEA. In the oxidative degradation ranking given
above, AMP is followed by AHPD, Serinol and MEA. Here it seems that the number of
hydroxyl groups, respectively 3, 2 and 1, increases the degradation rate of these amines.
This could be explained by a higher frequency of alcohol-carboxylic acids reactions
(Lepaumier et al., 2009b) as the number of hydroxyl group increases. Several organic acids
produced during amine oxidative degradation can participate to this reaction (Bedell, 2009;
Supap et al., 2006). Finally, Pz, without hydroxyl or methyl groups in its structure degraded
at a lower rate, possibly following other degradation reactions like ring opening reactions
that could also be influenced by the presence of acidic compounds in solution (Rochelle,
2012).
9.4.1.3. CO2 effect on amine degradation
In addition to thermal + oxidative degradation experiments, the influence of the
presence of CO2 on the degree of degradation was investigated using CO2 saturated amine
262
solutions. Except for Serinol where the presence of CO2 increased the oxidative
degradation percentage, the degradation percentage of all other studied amines, particularly
SHA, was found to be lower than that in the systems with O2 alone (thermal + oxidative
degradation) (Figure 9.3). A similar behaviour was observed by Supap et al. (2006) for
MEA degradation, where the presence of CO2 was found to induce more stable degradation
products than with O2 only. As mentioned by Freeman and Rochelle (2012a, b), the
degradation rate is a function of CO2 concentration which in turn depends on the speciation
in the solution. The presence of bicarbonate at high loading for conventional amines or
produced preferentially by SHA after CO2 absorption instead of amine carbamates (Bougie
and Iliuta, 2012) is assumed to have decreased the oxidative degradation rates.
The first-order degradation rate constants for the thermal + CO2 degradation
experiments performed for AMP and AHPD aqueous solutions are indicated in Table 9.2.
The presence of CO2 increases the thermal degradation of these two SHA, but considerably
less so than the presence of O2. This indicates the important effect of oxygen on these
amine solutions. In this context, the use of membrane contactors instead of packed columns
should be more advantageous in industrial applications as gas and liquid phases are
separated, limiting the interaction of oxygen with the amines in solution (Vogt et al., 2011).
9.4.2. Effect of process conditions on degradation of the aqueous AHPD + Pz blend
The degradation percentage of AHPD and Pz from their mixture can be compared to
the results corresponding to single amine systems (Figure 9.3). First, the trends mentioned
previously are similar: i) the thermal + oxidative degradation is more important compared
to thermal degradation or thermal + oxidative + CO2 degradation and ii) the presence of
CO2 decreased the thermal + oxidative degradation percentage. It is also possible to notice
that for AHPD, the thermal + oxidative degradation decreased considerably in its blend
with Pz (a reduction of 71%) in comparison with the single amine (AHPD) system, while
the degradation percentage corresponding to the other experimental conditions (thermal and
thermal + oxidative + CO2) remained very low. It can be concluded that the presence of Pz
keeps considerably AHPD from oxidative degradation. A similar beneficial behaviour has
been reported by Closmann et al. (2009) for Pz inhibiting MDEA oxidation in the
MDEA/Pz blend. However, although the Pz presence decreased AHPD oxidative
263
degradation, for all types of degradation of Pz in the mixture, its degradation degree was
shown to increase in comparison with the single Pz system (Figure 9.3). As mentioned in
the literature (Gouedard et al., 2012), this might be explained by a higher number of
degradation products caused by crossed reactions taking place in a two-amine system,
compared to single amine solutions.
9.5. Conclusions Thermal, thermal + CO2, thermal + oxidative and thermal + oxidative + CO2
degradation experiments were performed for several aqueous amine solutions including
conventional (MEA, Serinol), sterically hindered alkanolamines (AMP and AHPD) and
cyclic secondary diamine (Pz), in order to evaluate their stability to degradation. Single
amines and one blended system were investigated. It was found that the SHA are more
resistant to thermal degradation than the conventional amines investigated in this work, but
the presence of oxygen degraded them more significantly in the absence of CO2. The
presence of CO2 was beneficial to SHA as the preferential bicarbonate formation in
solutions reduces in a large extent the oxidative degradation rate observed in the absence of
CO2. The addition of Pz to AHPD solution also reduced the AHPD oxidative degradation
percentage; however, Pz degradation rate slightly increased, possibly due to crossed
reactions between the degradation products of each individual amine in solution.
In conclusion, the AHPD + Pz aqueous solution seems to be an interesting potential
absorbent to replace MEA solution in the industrial CO2 absorption process due to the low
degradation degree of the blend and also because this blend was found to be much less
corrosive than MEA solution. The use of the AHPD + Pz solution would be even more
beneficial in a membrane contactor compared to packed column because the oxidative
degradation could be minimised due to the reduced contact of the absorbent with the
oxygen contained in the flue gases. Future studies concerning the evaluation of degradation
products during CO2 absorption and degradation reaction mechanism of this specific blend
would therefore be helpful for industrial applications to optimize the operation conditions
and minimize the amine degradation process.
264
After the study of all CO2 absorption/regeneration properties and stability of AHPD + Pz
aqueous solution, as well as the solution/membrane compatibility, the performance of this
blend for CO2 absorption in PTFE hollow fiber membrane contactors is investigated in the
following chapter, under various liquid flow rates, gas compositions and flow orientation
(co- and counter-current). The results are compared to those obtained for aqueous AHPD
and MEA solutions in the same experimental conditions.
265
Chapter 10. Absorption of CO2 into Pz-activated AHPD aqueous solutions in PTFE hollow fiber membrane contactors: Experimental and modeling study
Résumé
Cette étude porte sur la séparation du CO2 de mélanges CO2/N2 par des solutions aqueuses de 2-amino-2-hydroxyméthyl-1,3-propanediol (AHPD), en présence et en absence de pipérazine (Pz) comme activateur, en utilisant des contacteurs à membrane à base de fibres creuses microporeuses en polytétrafluoroéthylène (PTFE). Les expériences ont été réalisées à différents débits de liquide, compositions de gaz et orientations des flux gazeux et liquide (co- et contre-courant). Les performances du procédé (efficacité de la capture et taux d'absorption) ont été comparées à celles correspondantes aux solutions aqueuses de MEA (l'amine de référence utilisée industriellement), dans les mêmes conditions expérimentales. Les taux d'absorption à travers les membranes augmentent avec l'augmentation de débit du liquide ou de la concentration du gaz en CO2. Les solutions d’AHPD activées par l’ajout de Pz ont montré des performances semblables ou meilleures que celles correspondantes aux solutions aqueuses de MEA. Un modèle mathématique représentant la diffusion du CO2 dans les pores de la membrane remplis entièrement par le gaz, la diffusion/réaction du CO2 et des amines dans les pores de la membrane dans le cas du mouillage et la diffusion/réaction du CO2 et des amines dans le film liquide, a été utilisé pour décrire le comportement des contacteurs à membrane. Les résultats de la modélisation montrent que le modèle peut très bien représenter les données expérimentales pour chacune des solutions aminées étudiées.
266
Abstract
This work investigates CO2 absorption from CO2/N2 mixtures in a microporous polytetrafluoroethylene (PTFE) hollow fiber membrane contactor using aqueous 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD) solutions in the presence and absence of piperazine (Pz). The absorption performance (absorption rate and capture efficiency) was compared to that of aqueous solutions of MEA (the benchmark amine used in CO2 removal) under the same experimental conditions. Experiments were conducted under various liquid flow rates, gas compositions and flow orientation (co- or counter-current). The absorption rates through the membranes increased with the increase of either liquid flow rate or CO2 gas concentration. Activated AHPD solution absorption performance was similar or better than that of conventional MEA aqueous solution. A two-scale model accounting for CO2 diffusion in the gas-filled membrane pores, CO2 and amines diffusion/reaction within the liquid-filled membrane pores and CO2 and amines diffusion/reaction in the liquid boundary layer was applied to describe the behaviour of the gas-liquid membrane contactor and agreed very well with the experimental results for each of the tested aqueous amine solutions.
267
10.1. Introduction The gas absorption process for CO2 separation is of high interest in various
applications in chemical, oil and gas industries, as well as in environmental protection
(Kohl and Nielsen, 1997). Among possible techniques to remove CO2 from different gas
mixtures, the chemical absorption by aqueous amines solutions is today’s best available
technology (Bernardo et al., 2009) and the process can traditionally be carried out in
different reactor types (bubble columns, sieve trays or packed towers). Membrane
contactors (MC) represent an interesting alternative that has been recently received lot of
attention due to several advantages like i) large and stable contact area promoting a more
efficient gas-liquid mass transfer than packed columns (deMontigny et al., 2005), ii) high
modularity and easy scale-up, and iii) the possibility of varying fluid flow rates
independently and without the occurrence of loading or flooding (Li and Chen, 2005). On
the negative side, the membrane itself adds an additional level of resistance to the mass
transfer process which can become important when the membrane pores are wetted by the
liquid absorbent (Li and Chen, 2005), leading to an important reduction of the absorption
process efficiency.
In membrane contactors, the gas and liquid phases flow on different sides of the
microporous membrane and the gas-liquid interface is formed, under non-wetting
conditions, at the membrane pores opening in the liquid phase. Under wetted conditions,
the pores are partially filled by liquid, depending on process conditions (liquid and
membrane type and characteristics, operation conditions, etc.) and the gas-liquid interface
is formed within the membrane. As the CO2 diffusion coefficient in the gas phase is much
higher than in the liquid phase, the non-wetted mode gives the highest absorption fluxes
(Rongwong et al., 2009). For large-scale CO2 absorption plants, membrane wettability is
one of the main obstacles facing this technology. The success of membrane contactors
implementation over conventional ones will then largely depend on the choice of the liquid
system and the type of membranes.
To avoid the wetting phenomena highly hydrophobic membranes are required to
repel the aqueous absorbent solutions. This lead mainly to the use of naturally low surface
energy membranes fabricated of polypropylene (PP), polyvinylidene fluoride (PVDF),
268
polytetrafluoroethylene (PTFE) or different membranes based on polymer modifications to
increase their hydrophobicity (such as asymmetric membranes and surface modified
membranes (Mosadegh-Sedghi et al., 2014). However, presently, only PTFE membranes
seem to be suitable for an industrial application principally because of their commercial
availability, higher hydrophobicity and chemical inertness (Falk-Pedersen and Dannström,
1997). PP membranes were often mentioned to be altered and wetted by amines solutions
(Barbe et al., 2000; deMontigny et al., 2006; Rangwala, 1996) and PVDF membrane
chemical stability in contact to amine solutions is questionable (Bougie and Iliuta, 2013a).
On the liquid side, apart from the frequently used amine solutions
(monoethanolamine (MEA), diethanolamine (DEA), N-methyldiethanolamine (MDEA) or
2-amino-2-methyl-1-propanol (AMP)) (Kim and Yang, 2000; Wang et al., 2004), very few
efforts have been made to investigate new absorbent solutions especially optimized for
application in MC and to compare their performance to those of MEA, the benchmark
amine used for CO2 capture, in the same experimental conditions. Besides their good
performance in CO2 separation (absorption capacity, absorption kinetics, degradation
resistance and regeneration facility), it is crucial for the absorption solutions intended to be
used in MC to have a high surface tension in order to reduce the membrane wetting
tendency. Kosaraju et al. (2005) studied CO2 absorption using a polyamidoamine
dendrimer aqueous solution, while amino acid salts aqueous solutions were introduced by
Feron and Jansen (2002) and Kumar et al. (2002), but no direct comparisons with MEA
solutions were performed in the same experimental conditions. Although these solutions
have high surface tensions, their price, elevated viscosities, and crystallisation problems
can limit their use.
Taking into consideration the requirements for a MC-optimized absorption solution
given above, we recently proposed the aqueous mixture of 23 wt% 2-amino-2-
hydroxymethyl-1,3-propanediol (AHPD, a sterically hindered alkanolamine) and 7 wt%
piperazine (Pz, a secondary diamine activator) as a potential alternative absorbent to MEA
solution. Previous works from our research group confirmed that this mixed solution can
provide good kinetics (Bougie et al., 2009), CO2 solubility (Bougie and Iliuta, 2010b,
2014a), regeneration capacity (Bougie and Iliuta, 2010a), resistance to degradation (Bougie
269
and Iliuta, 2014b) and higher surface tension in comparison with other conventional amines
(Bougie and Iliuta, 2013a).
In this context, the main objective of this research study is to evaluate the
performance of CO2 absorption process in a PTFE hollow fiber membrane contactor using
AHPD solutions in the presence and absence of piperazine. A comparison with CO2
absorption in MEA solution under the same experimental conditions was made.
Experiments were performed under various liquid flow rates, gas phase composition and
co- or counter-current flow orientations. The CO2 absorption rate and capture efficiency in
the membrane module were determined. The two-scale model developed by Iliuta et al.
(Iliuta et al., 2014) was applied to describe the behaviour of the gas-liquid membrane
contactor. On the basis of experimental results and numerical simulations, the fraction of
membrane pores wetted by absorbent was estimated.
10.2. Membrane contactor model The two-scale, isothermal, steady-state model developed by Iliuta et al. (2014)
accounting for CO2 diffusion in the gas-filled membrane pores, CO2 and amines
diffusion/reaction within the possible liquid-filled membrane pores and CO2 and amines
diffusion/reaction in the liquid boundary layer (Figure 10.1) was applied to describe the
comportment of the membrane contactor. As a complete description of the model is already
found in this article, only the main equations and concise explanations will be reminded
here.
10.2.1. Porous membrane scale model
With the absorbent solutions flowing inside the fiber lumen, steady-state mass
balance equations which describe CO2 (A) diffusion within the membrane gas-filled pores
and CO2 diffusion accompanied by chemical reaction within the membrane liquid-filled
pores are:
,,
1 0gA meff
A g
CD r
r r r r ∂∂ ∂
= ∂ ∂ ∂ (10.1)
270
( )2
,, , ,
1
1 0A meffA A i i j m
i
CD r r C
r r r rν
=
∂∂ ∂− = ∂ ∂ ∂ ∑
(10.2)
Figure 10.1. Schematic diagram of CO2 (A) and amine (B) concentration profiles in membrane contactor.
The corresponding boundary conditions are given as (gas-liquid interface is
positioned in membrane):
outmr R= ( ) ,
, , ,outm
outm
gA mg eff
g A g A m A gr Rr R
Ck C C D
r==
∂− = −
∂ (10.3)
gmr R= , ,
, ,g gm m
gA m A meff eff
A g A
r R r R
C CD D
r r= =
∂ ∂=
∂ ∂
(10.4)
, ,1
g gm m
gA m A mr R r R
C Cm= =
=
(10.5)
inmr R= ,,
, ,ininmm
A fA meffA A
r Rr R
CCD D
r r==
∂∂=
∂ ∂
(10.6)
271
Steady-state mass balance equation which describe amines (j=B,C) diffusion
accompanied by the chemical reaction within the liquid-filled portion of the pores is:
,,
1 0j meffj j
CD r R
r r r r
∂∂ ∂− = ∂ ∂ ∂
(10.7)
The corresponding boundary conditions are based on the following assumptions: the
amine is non-volatile and at the membrane-liquid interface the flux of component j in the
liquid film is equal to the flux in the wetted part of the membrane.
gmr R= ,
, 0gm
j meffj
r R
CD
r=
∂=
∂
(10.8)
inmr R= , ,
, ,ininmm
j m j feffj j
r Rr R
C CD D
r r==
∂ ∂=
∂ ∂
(10.9)
Under membrane all gas-filled pores conditions, the mathematical model describe only
the mass transfer of CO2 through the membrane pores and is reduced to the Eq. (10.1) with
the following boundary conditions:
outmr R= ( ) ,
, , ,outm
outm
gA mg eff
g A g A m A gr Rr R
Ck C C D
r==
∂− = −
∂ (10.10)
inmr R= ,,
, ,ininmm
gA fA meff
A g Ar Rr R
CCD D
r r==
∂∂=
∂ ∂
(10.11)
10.2.2. Liquid boundary layer (liquid film) scale model
The liquid film zone surrounding the inside membrane wall was described by the
nonlinear differential equations governing diffusion and reaction given by the film theory
(Lewis and Whitman, 1924).
( )2
,, , ,
1
1 0A fA A i i j f
i
CD r r C
r r r rν
=
∂ ∂ ∂− = ∂ ∂ ∂ ∑
(10.12)
( ),, ,
1 0j fj j j f
CD r R C
r r r r ∂ ∂ ∂
− = ∂ ∂ ∂
where j=B,C (10.13)
272
When the membrane pores are partially filled with liquid, the boundary conditions for
the liquid film concentrations are as follows:
inmr R= , ,
, ,ininmm
j m j feffj j
r Rr R
C CD D
r r==
∂ ∂=
∂ ∂
where j=A,B,C (10.14)
fr R= , ,fj f jr RC C
==
where j=A,B,C (10.15)
When the membrane pores are totally filled with gas, the boundary conditions for the
liquid film model are:
inmr R= , ,
1in inm m
gA f A mr R r R
C Cm= =
=
(10.16)
, 0inm
j f
r R
Cr
=
∂=
∂ where j=B,C (10.17)
fr R= , ,fj f jr RC C
==
where j=A,B,C (10.18)
10.2.3. Gas–liquid membrane contactor scale model
Due to the CO2-amine reaction in the membrane liquid-filled pores and in the liquid
film zone near the inside membrane wall, the depletion of amine as well as the saturation of
the bulk liquid with CO2 can be neglected in fully established region and the bulk liquid
flow within the hollow fiber can be modeled assuming “concentration plug flow” under
laminar flow conditions (Iliuta et al., 2013; Lee et al., 2000). The steady state mass balance
equations for CO2 and amines in the liquid phase are:
( )2
,,, , , ,
10
f
A fAA v in a i i j
ir R
CCu D a r C
z rν
==
∂∂+ − =
∂ ∂ ∑
(10.19)
( ), ,, , , 0
f
j j fj v in j j
r R
C Cu D a R C
z r=
∂ ∂+ − =
∂ ∂
where j=B,C (10.20)
Similarly, the steady state mass balance equation for CO2 in the gas phase within the
shell side is:
, ,, , 0
outm
gA g A meff
g A g v out
r R
C Cu D a
z r=
∂ ∂± − =
∂ ∂ (10.21)
273
where in Eq. (10.21), the sign “-” corresponds to the counter-current flow, and the sign
“+”corresponds to co-current flow.
The corresponding boundary conditions are given as:
0z = , ,0
inj jz
C C=
=
where j=B,C (10.22)
and
0z = , ,0
inA g A gz
C C=
= for co-current flow (10.23)
z H= , ,in
A g A gz HC C
== for counter-current flow (10.24)
10.2.4. Model parameters
The effective diffusion coefficients were evaluated using the correlation of Iversen
et al. (1997) for tortuosity factor. The diffusion coefficients for binary gas systems were
predicted with Chapman and Enskog equation (Reid et al., 1987). Knudsen diffusion
coefficient was evaluated using the correlation presented in (Treybal, 1967). The molecular
diffusion coefficients in the liquid phase was taken from Versteeg and van Swaaij (1988)
and Thomas and Furzer (1962) or calculated using the Wilke-Chang method (Reid et al.,
1987). The solubility of CO2 in the liquid phase was taken from Versteeg and van Swaaij
(1988) and Bougie and Iliuta (2009). For the liquid flow in the fiber lumen, the physical
liquid mass transfer coefficient was evaluated from the Graetz-Leveque correlation (Yang
and Cussler, 1986): 1/3
,
1.62 Rein inm m
j
k d dSh ScD H
= =
(10.25)
For the gas flow in the shell side, the mass transfer coefficient was evaluated with the
following correlation (Feron and Jansen, 2002):
0.5 0.33
,
0.9Reout
g mg g g
j g
k dSh Sc
D= = (10.26)
The kinetic constants and the rate expressions were taken from Bougie and Iliuta
(2009) and Bougie et al. (2009) for the aqueous AHPD or AHPD + Pz systems and from
Liao and Li (2002) for the aqueous MEA system.
274
10.2.5. Numerical implementation
Aspen Custom Modeler from Aspen Tech was used to generate the numerical
platform to solve the mixed ODE/algebraic system which models the gas-liquid hollow-
fiber membrane contactor. A 1st-order backward finite difference method was used for the
discretization in the axial direction and a 2nd -order central finite difference method in the
radial direction. A non-linear solver based on the Newton method was used to solve the set
of simultaneous model equations. The residual convergence determined by the difference
between the left and right hand sides of the equations was adopted.
10.3. Experimental 10.3.1. Chemicals
The aqueous amines solutions used in this work were prepared by gravimetric
method using distilled water and either one or two of the following amines: 2-amino-2-
hydroxymethyl-1,3-propanediol (AHPD, CAS No. 77-86-1), piperazine (Pz, CAS No. 110-
85-0) and monoethanolamine (MEA, CAS No. 141-43-5). The amines (from Laboratoire
MAT, Quebec, Canada, except for MEA from Sigma-Aldrich) had respectively a minimum
purity of (99.9, 99 and 99)% and were used without further purification. A Mettler AE240
balance with a precision of ±1×10-4 g was used to prepare the solutions and the
uncertainties of the reported concentrations were calculated to be less than 0.01 wt%. Gases
(CO2 and N2) were of commercial grade with a minimum purity of 99.9 % (Praxair).
10.3.2. Membrane module
The module used for CO2 absorption was fabricated from PTFE hollow fiber
membranes supplied by Markel Corporation (Pennsylvania, USA). The hollow fiber
membranes were potted with epoxy at both ends in stainless steel discs having small holes
positioned in a circular pattern. The length of the membrane inside the disc (0.03 m) on the
liquid entry side gave sufficient distance (>10din) for the laminar liquid flow inside the fiber
to be fully developed before it contacts the gas (Kumar et al., 2002). Additionally, the holes
in the discs were sufficiently distant one relative to each other to assure evenly spaced fiber
and no contact between them. This membrane assembly was put in a clear borosilicate
housing allowing visual inspections of the membranes to detect any possible liquid going to
275
the shell side through the membrane pores. Membrane and module specifications are
provided in Table 10.1. The gas-liquid contact area was calculated based on the membrane
inside diameter and the length of the membrane exposed to the gas flow.
Table 10.1. Membrane and module specifications.
Membrane Material PTFE Inside diameter (µm) 1830 Outside diameter (µm) 2440 Pore diameter (µm) 0.03-0.08 Porosity 0.2 Lenght (m) 0.178 Number 8
Module Inside diameter (m) 0.05 Length (m) 0.208 Gas-liquid contact area (m2) 0.0082
10.3.3. Absorption setup and procedure
The experimental setup for CO2 absorption using the membrane contactor is shown
in Figure 10.2. The gas circuit mainly consists of mass flow controllers (OMEGA, FMA-
2600A) to adjust the flow and composition of the inlet gas and of a bubble flowmeter and a
gas chromatograph (Micro GC 3000A, Inficon) to determine respectively the flow and
composition of the outlet gas. Aqueous amine solutions were supplied to the contactor
using a gear pump (Cole-Parmer, OF-75211) and a rotameter calibrated for each amine
solution was used to adjust the liquid flow. Inlet and outlet fluid pressures were measured
by four pressure transducers (Omega, PX481A) and a needle valve at the liquid exit of the
contactor was adjusted in all experiments to keep the liquid phase outlet pressure above the
gas phase pressure by at least 2-3 psig.
276
All experiments were performed at 298 K with the liquid flowing through the
membrane lumen and the gas supplied to the shell side. The fluids were circulating counter-
currently or co-currently by modifying the gas connexions in the contactor module. Three
aqueous solutions were tested. Aqueous 23 wt% AHPD solution was used as a base case
and the activation effect of Pz was studied for aqueous 23 wt% AHPD + 7 wt% Pz solution.
The 30 wt% MEA aqueous solution was also tested for comparison purpose.
Figure 10.2. Experimental setup for CO2 absorption using the membrane contactor in counter-current flow circulation (the co-current flow is performed by switching the gas
connexions in the contactor module).
In a typical run, liquid flow was first established through the contactor at a rate
between 1 and 120 ml/min and the liquid pressure was stabilized. A constant humidified
100 ml/min total gas flow rate with a volumetric fraction of CO2 ranging from 20 to 100%
in nitrogen (balance) was then supplied to the shell side of the contactor. Usually, around
15 minutes were necessary to reach steady-state conditions and the absorption rate was
measured based on the inlet gas flow rate and the difference between the inlet and the outlet
CO2 composition in the gas as determined with the bubble flowmeter and gas
chromatograph. The amine solutions were thermally regenerated and reused. As a
modification of the lean loading (mol of CO2 per mol of amine) can influence the
absorption fluxes, all solutions were subjected to several absorption-regeneration cycles
(Bougie and Iliuta, 2010a) until the lean loading become constant, as it would happen in an
277
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0 20 40 60 80 100 120 140
CO2
flux
(mol
/m2 .m
in)
Liquid flow rate (ml/min)
AHPD - experimentalMEAAHPD + PzAHPD - modelMEAAHPD + Pz
industrial absorption process. The solution lean loading values of 0.03, 0.05 and 0.16 were
obtained, respectively, for the AHPD, AHPD + Pz and MEA solutions.
10.4. Results and Discussion 10.4.1. Effect of liquid flow rate on CO2 absorption
The absorption performance, expressed as carbon dioxide flux through the
membrane, is shown in Figure 10.3 for the three studied aqueous amine solutions as a
function of the liquid flow rate. The data have been gathered from experiments using a gas
flow rate of 100 ml/min of pure CO2 in order to eliminate any gas phase resistance and
clearly see the effect of liquid flow rate and solution composition on CO2 absorption. This
range of liquid flow rate is frequent in the literature (Kim and Yang, 2000; Yeon et al.,
2003) and gives liquid velocities between 0.8 and 95 mm/s.
Figure 10.3. CO2 absorption flux as a function of liquid flow rate with a pure CO2 gas flow rate of 100 ml/min in counter-current mode.
It can be observed that for all solutions, the absorption flux increased at low liquid
flow rate, before becoming almost stationary and independent of the liquid flow (Figure
10.3). This could be explained by a reduced driving force at low liquid velocity, as the
solution loading increases more rapidly. At high liquid flow rate, the solution loading does
278
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0 20 40 60 80 100
CO2
flux
(mol
/m2 .m
in)
vol% CO2
AHPD - experimental - counter-currentMEAAHPD + PzAHPD - experimental - co-currentAHPD + PzAHPD - model - counter-currentMEAAHPD + Pz
not increase significantly, leading to an almost fixed driving force and consequently, to a
nearly stable absorption flux. A similar loading effect on the absorption flux can be
observed in Feron and Jansen (2002) for CO2 absorption in a dedicated absorption solution
(CORAL). AHPD + Pz solution outperforms both MEA and AHPD solutions (Figure 10.3).
This can be explained by the addition of Pz, an amine known to have a larger second order
reaction rate constant with CO2 compared to MEA (Derks et al., 2006), as activator for
AHPD.
10.4.2. Effect of gas phase composition on CO2 absorption
For practical considerations, it is more useful to obtain high CO2 absorption fluxes
at low absorbent flow rate to minimise liquid-related tank volumes and pump energy
consumption. Based on the results presented in Figure 10.3, the liquid flow rates were
therefore selected at 30 ml/min for the investigation of the effect of gas phase composition
on the absorption efficiency. CO2 absorption fluxes were measured under various gas phase
CO2 concentrations and data obtained for liquid flow rates of 30 ml/min respectively for
AHPD, AHPD + Pz and MEA solutions are shown in Figure 10.4.
Figure 10.4. CO2 absorption flux as a function of the inlet CO2 volumetric percentage with a total gas flow of 100 ml/min and liquid flow rates of 30 ml/min for AHPD, AHPD + Pz
and MEA solutions.
279
CO2 fluxes show a near linear increase below around 60% CO2, with a tendency to
level off for AHPD and MEA at higher CO2 percentage in the gas phase. However, for
AHPD + Pz a linear trend is kept up to 100% CO2. This could be explained by the fact that
because the membranes used in this study have a small pore size and a low porosity, the
AHPD + Pz system, being the more reactive, seems to be more affected by a higher gas
phase and membrane resistance in comparison to AHPD or MEA. Consequently, the gas
phase resistance and the diffusion limitation in the membrane pore become less significant
when the gas is more concentrated in CO2 and this increases the absorption rate of the Pz-
activated solution compared to AHPD or MEA. Similarly, Lin et al. (2008) mentioned that
the use of Pz as activator in AMP solution caused the decrease of the liquid phase
resistance by the increase of the enhancement factor, while the gas and membrane
resistances increased in respect to the global mass transfer resistance.
10.4.3. Flow configuration and CO2 removal efficiency
As shown in Figure 10.4, the difference between the absorption fluxes in co- and
counter-current flow circulation was found to be insignificant, similar to the results
reported by Kreulen et al. (1993) for CO2 absorption in aqueous NaOH solutions. This can
mainly be attributed to the relatively short length of the membranes and their low number
in the module.
Besides the CO2 absorption flux, the evaluation of CO2 removal efficiency is of
interest because a CO2 capture of 90% is usually targeted in industrial applications (Yan et
al., 2008). As example, based on data from Figure 10.4, the CO2 removal percentages were
calculated for the AHPD + Pz system obtained in counter-current flow condition and the
results are shown in Figure 10.5. It can be seen that CO2 removal percentages of 35 and
15% were obtained, respectively, when the total inlet gas flow rate (100 ml/min) contains
20 and 100% CO2. Although a removal percentage of 35% is higher than 15%, the absolute
amount of CO2 removed is higher for the pure gas (35% × 20% < 15% × 100%). The higher
absorption flux at higher CO2 content is due to the higher driving force.
280
0
5
10
15
20
25
30
35
40
0 20 40 60 80 100
CO2
rem
oval
%
vol% CO2
Figure 10.5. CO2 removal efficiency for the aqueous AHPD + Pz solution (counter-current, total gas flow of 100 ml/min and liquid flow rate of 30 ml/min).
The increase of CO2 removal efficiency could be obtained by the increase of the
number of membrane in the module or the increase of the number of modules, the use of
membranes with higher porosity or using an optimal absorption temperature. An increase in
temperature will increase the absorption kinetics and diffusion coefficients, but will be
detrimental to CO2 solubility, surface tension and wetting tendency of the membranes
(Feron and Jansen, 2002; Khaisri et al., 2010). Based on the results obtained in this work, a
parameter optimisation to increase the CO2 removal efficiency will therefore make the
object of a future publication.
10.4.4. Model analysis – effect of membrane wetting
In addition to experimental data, Figures 10.3 and 10.4 present the modeling results.
It can be seen an excellent agreement with the experimental data, thus confirming the
capacity of the model to describe the CO2 absorption performance in HFMC of various
amine systems (average relative deviation of 1.2%). Under the operation conditions of this
study, the average membrane pore wetted fraction was estimated at around 7.5% for
AHPD, 8% for AHPD+Pz and 10% for MEA. The values of membrane wetted pore
fractions issued from Figures 10.3 and 10.4 data are displayed respectively in Figures 10.6
and 10.7.
281
Figure 10.6. Variation of the membrane wetted pore fraction for data of Figure 10.3.
Figure 10.7. Variation of the wetted pore fraction for data of Figure 10.4.
282
It is known that the wetting of membrane pores depends on several parameters like
membrane configuration (e.g., pore size and distribution), physiochemical properties of the
liquid (e.g., surface tension), and operation parameters (Mosadegh-Sedghi et al., 2014). The
modeling results clearly show that a solution with higher surface tension is expected to wet
less the membrane pores. This predicted effect of the surface tension on membrane wetting
is in agreement with the experimental surface tensions of the three solutions (AHPD: 71.2
mN/m; AHPD+Pz: 70.2 mN/m; MEA: 63.9 mN/m) (Bougie and Iliuta, 2013a).
10.5. Conclusion CO2 removal by AHPD (23 wt%) + Pz (7 wt%) aqueous solution in a PTFE hollow
fiber membrane contactor was investigated. The results were compared to those obtained
with aqueous AHPD (23 wt%) and MEA (30 wt%) solutions under the same experimental
conditions. The experiments were conducted under various liquid flow rates, gas
compositions and flow orientation (co- or counter-current). For all tested solutions, the CO2
absorption rates increased with the increase of either liquid flow rate or CO2 gas
concentration. It was found that at higher liquid flow rates, AHPD + Pz solution
outperformed both MEA and AHPD solution due to the activator effect of Pz which has
very fast kinetics. At low flow rates, the performance of AHPD + Pz is similar to MEA, but
better compared to AHPD. At a constant liquid flow rate, the CO2 flux for AHPD + Pz
increased linearly with the CO2 concentration in the gas phase. The absorption performed
in co- and counter-current flow circulation showed no significant difference between the
absorption fluxes. An excellent agreement was found between the modeling results and
experimental data, thus confirming the capacity of the model to describe the CO2
absorption performance in membrane contactor of various amine systems. Moreover, the
modeling results clearly showed the effect of the surface tension on membrane wetting (the
wetting behaviour increases with the decrease of the surface tension).
283
284
Finally, the performance of AHPD + Pz aqueous solutions for CO2 absorption was also
investigated in different flat sheet membrane contactors (PTFE, PP and laminated
PTFE/PP membranes), under various liquid flow rates, gas compositions and flow
orientation (co- and counter-current). The results are compared to those obtained for
aqueous AHPD and MEA solutions in the same experimental conditions.
285
Chapter 11. Flat sheet membrane contactors (FSMC) for CO2 separation in aqueous amine solutions
Résumé
Un nouveau contacteur à membranes plates (FSMC) a été développé et utilisé pour étudier la capture du CO2 de mélanges CO2/N2 par une solution aqueuse de 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD) en présence et en absence de pipérazine (Pz) comme activateur. Le contacteur a été opéré dans différentes conditions expérimentales afin d'étudier l'effet du débit du liquide, la concentration en phase gazeuse et la configuration du contacteur (nombre de membranes, type de membrane (PTFE, PP et PTFE/PP laminées) et l'écoulement des fluides (co- et contre-courant). À des fins de comparaison, la solution aqueuse de MEA (l'amine de référence utilisée dans la capture du CO2) a également été testée dans les mêmes conditions expérimentales. Les taux d'absorption à travers les membranes augmentent avec l'augmentation du débit du liquide et la concentration du CO2 dans la phase gazeuse. La solution AHPD-Pz a montré de meilleures performances que la solution d’AHPD, mais semblables aux solutions aqueuses de MEA. Comme le taux d'absorption du CO2 augmente proportionnellement avec le nombre de membranes, plusieurs membranes plus peuvent être facilement ajoutées au module pour augmenter les performances.
286
Abstract
A new multi-flat-sheet membrane contactor was developed and used to investigate CO2 removal from CO2/N2 gas mixtures using aqueous 2-amino-2-hydroxymethyl-1,3-propanediol (AHPD) solution in the presence and the absence of piperazine (Pz) as activator. The FSMC was operated under various experimental conditions in order to study the effect of liquid flow rates, gas phase composition and contactor configuration (number of membranes, type of membrane (PTFE, PP and laminated PTFE/PP) and fluid flow orientation (co- and counter-current)). For comparison purpose, MEA aqueous solution (the benchmark amine used in the CO2 capture process) was also tested under the same experimental conditions. The absorption rates through the membranes were found to increase with the increase of liquid flow rate and CO2 concentration in the gas phase. Activated Pz-AHPD solution showed better performance than single AHPD solution, but similar absorption fluxes were obtained for AHPD + Pz and MEA solutions. As a proportional increase of the absorption rate with the number of membranes was observed, more membranes can be easily added to the module to increase the absorption performance.
287
11.1. Introduction The absorption is a common process in chemical engineering and it is largely
applied in the industrial acid gas treatment and environmental protection. Among possible
techniques, the chemical absorption by aqueous amines solutions is today’s best available
technology to remove CO2 from different gas mixtures (Bernardo et al., 2009). The
conventional technique is based on packed columns. However, at an industrial scale, these
gas-liquid contactors are very large, expensive to build, and suffer from a variety of
operational problems including liquid channeling, flooding, entrainment and foaming
(Wang et al., 2011). Membrane contactors (MC) represent an interesting alternative as they
are characterized by: i) large and stable gas-liquid contact area reducing the contactor size
and weight, ii) high modularity and easy scale-up, and iii) the possibility of varying the
membrane-separated fluid flow rates independently and without the occurrence of the
above-mentioned operational problems experienced in packed columns (Li and Chen,
2005). On the downside, the membranes in the absorption module add an additional level
of resistance to the mass transfer process and the pressure of both phases should be
controlled (Gabelman and Hwang, 1999). Actions can however be taken to minimize these
relatively few disadvantages which are then often outweighed by the numerous advantages
cited above. For this reason, membrane contactors have been received lot of attention in the
last decades.
Despite the fact that most common hydrophobic polymeric membranes used in CO2
capture application, i.e. those made in polypropylene (PP), polyvinylidene fluoride (PVDF)
and polytetrafluoroethylene (PTFE), are commercially available as hollow fiber or flat
membranes, it can be observed that these studies have been mainly focused on hollow fiber
membrane contactors (HFMC) (Paul et al., 2008). Compared to HFMC, information on
CO2 absorption in flat sheet MC (FSMC) are extremely scarce (Ahmad et al., 2010;
Dindore et al., 2004; Lin et al., 2009b; Zhang et al., 2006). However, this type of contactors
has some noteworthy advantages compared to HFMC, like an easiness in membrane
fabrication and characterization, facility of the module assembly (no membrane potting)
and higher flux for the same gas-liquid contact area (Baker, 2004).
288
A FSMC containing one flat sheet membrane of ePTFE (expanded teflon) was used
by Zhang et al. (2006) in order to investigate the effect of membrane porosity and pore size
on the absorption process. Several membranes having a surface area of 450 cm2 were used,
the mean pore size varying from 0.2 to 2 µm and porosity of 0.52-0.9. Tests were
performed using pure CO2 and water or NaOH aqueous solutions (0.1 M). It was concluded
that the porosity has a more significant effect on the absorption for a rapid mass transfer
process; for a slow mass transfer process, the porosity has almost no effect. Dindore et al.
(2004) used a simple FSMC in order to measure the critical entry pressure (a very useful
parameter in membrane operation) and to determine the mass transfer coefficient for CO2
absorption in different physical solvents. Only one flat sheet membrane of PP (thickness of
92.5 µm, maximum pore size of 0.36 µm) or PTFE (thickness of 158 µm, maximum pore
size of 0.45 µm) was used. The membrane mass transfer resistance was found negligible in
the not-wetted mode of operation.
The first work related to the application of FSMC for CO2 absorption in amine
solutions was given by Paul et al. (2008) who studied theoretically CO2 absorption (pure
CO2 and CO2/N2 mixture) by different single and blended alkanolamines (MEA,
diethanolamine (DEA), N-methyldiethanolamine (MDEA), 2-amino-2-methyl-1-pronanol
(AMP), MEA + AMP) considering a flat sheet membrane contactor (one hypothetical
membrane, length of 20 cm). The authors concluded that for all solutions considered in
their work the CO2 absorption flux in FSMC was higher than that in HFMC. FSMC
containing only one PVDF, plasma-treated PVDF or PTFE membrane was used by Lin et
al. (2009b) to study the CO2 absorption from CO2/N2 mixtures (1-15% CO2) in MDEA (1
M), AMP (1 M) and AMP+Pz (1 M AMP+0.2 M Pz) aqueous solutions. The effect of
several parameters on the CO2 absorption flux was investigated, like liquid and gas flow
rates and absorbent concentration. It was found that the CO2 flux increased with the
increase of gas flow rate and absorbent concentration and the absorption process being
dominantly governed by gas film and membrane resistances. The plasma treatment was
found to increase both the absorption flux and membrane durability compared to non-
treated PVDF membranes. Ahmad et al. (2010) investigated the absorption of CO2 from
CO2/N2 gas mixtures (10-100% CO2) using a FSMC containing one PVDF membrane
289
(porosity of 0.75 and pore size of 0.1 µm and 0.45 µm) and aqueous AMP solutions (1-5
M). However, no information about the module characteristics was given (membrane
thickness, gas-liquid contact area). Unexpectedly, the membranes with the biggest pore size
gave a lower mass transfer coefficient, which was attributed to membrane wetting.
All few studies related to the use of FSMC were limited to a single membrane and
one flow configuration type. Moreover, only three experimental works involved amine
solutions as CO2 absorbents. In this work, a new multi-flat-sheet membrane contactor was
developed and used to investigate CO2 removal from CO2/N2 gas mixtures using aqueous
2-amino-2-hydroxymethyl-1,3-propanediol (AHPD) solution in the presence and the
absence of piperazine (Pz) as activator. Aqueous 23 wt% AHPD solution was used as
reference and the activation effect of Pz was investigating for the aqueous 23 wt% AHPD +
7 wt% Pz system (30 wt% total amine). In our previous works, this blend combining AHPD
(a sterically hindered alkanolamine) and Pz (a secondary diamine activator with better
kinetic compared to MEA (Derks et al., 2006)) was found to represent a dedicated CO2
absorbent to be used in MC. Besides good absorption capacity (Bougie and Iliuta, 2010b,
2014a), kinetics (Bougie et al., 2009), regeneration capacity (Bougie and Iliuta, 2010a) and
resistance to degradation (Bougie and Iliuta, 2014b), it also presents a high surface tension
in comparison with other conventional amines, thus offering the potential to minimize the
membrane wetting tendency (Bougie and Iliuta, 2013a). The FSMC was operated under
various experimental conditions in order to study the effect of liquid flow rates, gas phase
composition and contactor configuration (number of membranes, type of membrane (PTFE,
PP and laminated PTFE/PP) and fluid flow orientation (co- and counter-current)). For
comparison purpose, a 30 wt% MEA aqueous solution (the benchmark amine used in the
CO2 capture process) was also tested under the same experimental conditions.
11.2. Experimental 11.2.1. Chemicals
The aqueous amines solutions used in this work were prepared by gravimetric
method using distilled water and either one or two of the following amines: 2-amino-2-
hydroxymethyl-1,3-propanediol (AHPD, CAS No. 77-86-1), piperazine (Pz, CAS No. 110-
290
85-0) and monoethanolamine (MEA, CAS No. 141-43-5). The amines (from Laboratoire
MAT, Quebec, Canada, except for MEA from Sigma-Aldrich) had respectively a minimum
purity of (99.9, 99 and 99)% and were used without further purification. A Mettler AE240
balance with a precision of ±1×10-4 g was used to prepare the solutions and the
uncertainties of the reported concentrations were calculated to be less than 0.01 wt%. Gases
(CO2 and N2) were of commercial grade with a minimum purity of 99.9 % (Praxair,
Canada).
11.2.2. Flat sheet membrane contactor
The membranes used in the FSMC were of different types. PTFE, PP and laminated
PTFE/PP flat membranes were supplied respectively by Donaldson Company (Minnesota,
USA), Membrana (North Carolina, USA) and Pall Canada Ltd (Quebec, Canada). All
membrane characteristics are reported in Table 11.1. Modules with 1 to 3 membranes were
tested (n-FSMC, with n=1,2,3). Before use, the membranes were washed with alcohol,
rinsed with distilled water and dried in a convection oven at 333 K overnight. The
membranes were then cut and mounted in the contactor assembly (gas-liquid contact area
per membrane of 0.0041 m2). In all experiments, the liquid solution was fed toward each
membrane from the bottom and some distance was given to the absorbent for the laminar
flow to be fully developed before it contacts the gas (Kumar et al., 2002). When more than
one membrane was used in the contactor module, the liquid flow at the outlet of the first
membrane was directed to the inlet of the second membrane and so on; the liquid flow
circulation in the contactor was in series in respect to all membranes.
Table 11.1. Flat membrane and module specifications.
Membrane PTFE PP Laminated PTFE/PP Thickness (µm) 203 100 178-246 Pore diameter (µm) 0.1 0.1 0.2 Porosity (µm) 0.8 0.8 0.8
11.2.3. Absorption setup and procedure
The experimental setup for CO2 absorption including the FSMC is shown in Figure
11.1. Mass flow controllers (OMEGA, FMA-2600A) regulated the inlet gas composition
291
and flow rate, while a bubble flowmeter and a gas chromatograph (Micro GC 3000A,
Inficon) were used to determine respectively the flow rate and composition of the leaving
gas. Aqueous amine solutions were supplied to the contactor using a gear pump (Cole-
Parmer, OF-75211) and a rotameter calibrated for each amine solution was used to adjust
the liquid flow rate. Inlet and outlet fluid pressures were measured by four pressure
transducers (Omega, PX481A) and a needle valve at the liquid exit of the contactor was
adjusted in all experiments to keep the liquid phase outlet pressure above the gas phase
pressure by at least 2-3 psig.
Figure 11.1. Experimental setup for CO2 absorption using the FSMC.
All experiments in this work were performed at 298 K with FSMC containing one,
two or three membranes as listed in Table 11.1. The fluids were circulating counter-
currently or co-currently by switching the gas connexions on the contactor module. Three
aqueous solutions were tested. Aqueous 23 wt% AHPD solution was used as reference and
the activation effect of Pz was studied for aqueous 23 wt% AHPD + 7 wt% Pz solution.
The 30 wt% MEA aqueous solution was also tested for comparison purpose. The liquid
flow rates were varied between 1 and 50 ml/min. A constant humidified 100 ml/min total
gas flow rate (unless otherwise specified) with a volumetric fraction of CO2 ranging from
20 to 100% in nitrogen (balance) was then supplied to the contactor. Usually, around 15
minutes were necessary to reach steady-state conditions and the absorption rate was
292
measured based on the inlet gas flow rate and the difference between the inlet and the outlet
CO2 composition in the gas as determined with the bubble flowmeter and gas
chromatograph. The amine solutions were thermally regenerated and reused (Bougie and
Iliuta, 2010a) as it would happen in an industrial absorption process and solution lean
loading values of 0.03, 0.05 and 0.16 were obtained before absorption experiments,
respectively, for AHPD, AHPD + Pz and MEA solutions.
11.3. Results and Discussion 11.3.1. Effect of liquid flow rate on CO2 absorption flux
The absorption performance expressed as CO2 absorption flux, obtained for the
three studied aqueous amine solutions using 3-FSMC in counter-current flow, as a function
of the liquid flow rate is shown in Figure 11.2. To eliminate the gas phase resistance and
clearly see the effect of liquid flow rate and solution composition on CO2 absorption, the
experiments were performed using pure CO2 (gas flow rate of 100 ml/min). For all
solutions, the absorption flux increases at low liquid flow rate, before becoming almost
stationary. This could be explained by a reduced driving force at low liquid flow rate
because the solution loading reaches a higher value compared to that obtained at high liquid
flow rate. Moreover, an increase of the liquid flow rate can also reduce the resistance of the
stagnant-layer close to the membrane obtained under laminar flow (Feron and Jansen,
2002).
As expected, due to Pz addition, the absorption fluxes for the activated AHPD
solution are higher than those corresponding to single AHPD solution. Compared to 30
wt% MEA (4.95 M), the blend AHPD + Pz containing 30 wt% total amine (23 wt% + 7
wt%; 2.859 M total amine) offers similar absorption fluxes. This confirms the potential of
the AHPD + Pz solution to replace MEA in industrial applications, and especially using
MC. As mentioned above, this blend offers good absorption capacity and kinetics,
regeneration capacity and resistance to degradation (Bougie and Iliuta, 2014b)(Bougie and
Iliuta, 2014b), as well as a high surface tension compared to conventional amines, for
minimizing the membrane wetting tendency.
293
Figure 11.2. CO2 absorption flux in 3-FSMC (PTFE) as a function of liquid flow rate (pure CO2 gas flow rate of 100 ml/min in counter-current mode).
11.3.2. Effect of the number of membranes on CO2 absorption rate
One interesting feature of FSMC compared to HFMC is the possibility to add new
membranes into a module, in order to increase the contactor performances. The
improvement of the contactor performance can be observed in Figure 11.3. CO2 absorption
rate in AHPD + Pz solution was measured under counter-current conditions with a pure
CO2 gas flow rate of 100 ml/min.
The results clearly demonstrate a proportional increase of the absorption rate with
the number of membranes: the maximum CO2 absorption rate values obtained using 2
membranes (0.0011 mol/min) and 3 membranes (0.0017 mol/min) are, respectively, two
and three times more than the value obtained with 1 membrane (0.00057 mol/min). This
indicates that the solution still keeps its absorption capacity at the exit of the third
membrane; more membranes could therefore easily be added to the module to increase the
absorption rate until the addition of a new membrane is no longer beneficiary. This one-
membrane-at-a-time optimisation procedure is another advantage of FSMC over HFMC
and the compact multi-membrane FSMC design would possibly be more attractive than
tubular HFMC modules connected in series (deMontigny et al., 2006).
294
0.0000
0.0002
0.0004
0.0006
0.0008
0.0010
0.0012
0.0014
0.0016
0.0018
0 10 20 30 40
CO2
abso
rptin
rate
(mol
/min
)
Liquid flow rate (ml/min)
1 membrane 2 membranes 3 membranes
Figure 11.3. CO2 absorption rate in a PTFE membrane FSMC as a function of AHPD + Pz solution flow rate with a pure CO2 gas flow rate of 100 ml/min in counter-current mode.
11.3.3. Effect of gas phase composition and flow configuration on CO2 absorption flux
To investigate the effect of the gas phase composition on the CO2 absorption flux in
FSMC, experiments were performed with a gas flow rate of 100 ml/min and at a liquid
(AHPD + Pz) flow rate of 20 ml/min. 3-FSMS module was used in both co- and counter-
current. To evaluate the effect of gas flow rate on CO2 absorption flux, additional
experiences were performed at constant gas composition (20% CO2) and different gas flow
rates (200 and 300 ml/min). The results are shown in Figure 11.4.
As expected, a near linear increase is observed, with a tendency to level off at
higher CO2 percentage in the gas phase. The increase of the absorption flux with the
increase of CO2 content is due to the increase of the driving force. For a total gas flow rate
of 100 ml/min, the difference between the absorption fluxes obtained in co- and counter-
current flow circulation is insignificant. This result is in agreement with many studies on
gas absorption in HFMC (Atchariyawut et al., 2007; Iliuta et al., 2014; Kreulen et al.,
1993).
From the two additional data obtained at 200 and 300 ml/min total gas flow rate and
20% CO2 content, it can be seen that the absorption flux slightly increases with the increase
of flow rate due to the reduction of the gas phase resistance. The significant difference
295
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0 20 40 60 80 100
CO2
flux
(mol
/m2 .m
in)
vol% CO2
300 ml/min total gas - Counter-current200 ml/min total gas - Counter-current100 ml/min total gas - Counter-current100 ml/min total gas - Co-current
between the 100 ml/min counter-current data (the symbol in Figure 11.4 is hidden behind
the co-current data) and the 200 ml/min data can be due to the fact that all CO2 present in
the gas phase at the lower gas flow rate of 100 ml/min was completely absorbed as it will
be seen in the next section (§11.3.4). Consequently, the flux at 20% CO2 and 100 ml/min
would have been higher and closer to those at 200 and 300 ml/min if more CO2 was
introduced into the contactor.
Figure 11.4. CO2 absorption flux as a function of the gas inlet CO2 volumetric percentage for an AHPD + Pz absorbent flow rate of 20 ml/min using 3-FSMC (PTFE).
11.3.4. CO2 removal percentage
Besides the CO2 absorption flux, the evaluation of CO2 removal percentage in the
gas phase is of interest because a CO2 capture of 90% is usually targeted in industrial
applications (Yan et al., 2008). Based on data of Figure 11.4, the CO2 removal percentages
were evaluated for the AHPD + Pz systems under counter-current flow conditions and the
results are shown in Figure 11.5. As mentioned in §11.3.3, the CO2 removal at 20% CO2
and 100 ml/min is complete. It can be seen that the CO2 removal percentage is a function of
the CO2 concentration in the gas flow and the flow rate. This kind of data can therefore be
useful to determine the absorption conditions to obtain a CO2 removal percentage in the
area between the two dashed lines.
296
0
20
40
60
80
100
0 20 40 60 80 100
CO2
rem
oval
%
vol% CO2
100 ml/min total gas200 ml/min total gas300 ml/min total gas
Figure 11.5. CO2 removal percentage for Figure 11.4 counter-current data.
11.3.5. Influence of membrane properties
The last parameter investigated in this study is the effect of membrane properties on
CO2 absorption flux. Data from AHPD + Pz counter-current experiments using pure CO2 at
a gas flow rate of 100 ml/min are displayed in Figure 11.6. It can be seen that the CO2
absorption flux obtained with PP membrane is slightly higher that PTFE absorption flux.
On the other side, lower fluxes were obtained with the PTFE/PP laminated membranes. As
the experiments were performed with pure CO2 (no gas phase resistance) and the same
solution (same liquid phase resistance), the influence of the membrane resistance should
explain these results.
Firstly, the difference between PP and PTFE membranes can be explained by the
lower thickness of PP membranes (100 µm) in comparison with PTFE (203 µm). CO2
diffusion limitations inside the membrane should be lower for PP membranes and therefore,
higher CO2 absorption fluxes are observed experimentally. Similar observation were given
by Lin et al. (2009b): higher CO2 flux was obtained with thinner PVDF membrane
compared to PTFE ones. Secondly, the lower absorption fluxes obtained with PTFE/PP
laminated membranes can be explained by the combination of their higher thickness and
bigger pore size compared to PP and PTFE. According to Laplace-Young relation, the
breakthrough pressure for membranes with bigger pore will be lower, thus indicating a
higher wetting tendency (Bougie and Iliuta, 2013a). Consequently, the membrane
297
resistance of laminated sheets should be higher compared to single PP and PTFE
membranes, thus reducing, as observed experimentally, the CO2 fluxes. A similar
behaviour was observed by Ahmad et al. (2010): higher mass transfer coefficients were
obtained using membranes with smaller pores. These results indicate therefore that
membranes with low pore size and low thickness should be preferred in FSMC to maximise
the CO2 absorption flux.
Figure 11.6. Effect of membrane properties on CO2 flux in 2-FSMC as a function of liquid flow rate (pure CO2 gas flow rate of 100 ml/min in counter-current mode).
11.4. Conclusion In this work, a new multi-flat-sheet membrane contactor was developed and used to
investigate CO2 removal from CO2/N2 gas mixture using aqueous 2-amino-2-
hydroxymethyl-1,3-propanediol (AHPD) solution in the presence and the absence of
piperazine (Pz) as activator. Aqueous 23 wt% AHPD solution was used as reference and
the activation effect of Pz was investigated for the aqueous 23 wt% AHPD + 7 wt% Pz
system (30 wt% total amine). The FSMC was operated under various experimental
conditions in order to study the effect of liquid and gas flow rates, gas phase composition
and contactor configuration (number of membranes, type of membrane (PTFE, PP and
laminated PTFE/PP) and fluid flow orientation (co- and counter-current)). For comparison
298
purpose, 30 wt% MEA aqueous solution (the benchmark amine used in the CO2 capture
process) was also tested under the same experimental conditions.
The absorption rates through the membranes were found to increase with the
increase of liquid flow rate or CO2 gas concentration. The absorption fluxes for activated
Pz-AHPD solution were higher than those corresponding to single AHPD solution and
similar to those obtained for the MEA solution. This confirms the potential of the AHPD +
Pz solution to replace MEA in industrial applications, especially using MC. Besides its
efficiency in CO2 removal, the AHPD + Pz blend was already found to offer good
regeneration capacity and resistance to degradation, as well as a high surface tension for
minimizing the membrane wetting tendency in comparison to conventional amines. As a
proportional increase of the absorption rate with the number of membranes was observed,
more membranes can easily be added to the module to increase the absorption rate. The
CO2 flux obtained with PP and PTFE membranes were close, but lower fluxes were
obtained with the laminated membranes which had higher thickness and bigger pore size
compared to PP and PTFE.
299
Chapter 12. General Conclusions and Suggestions for
Future work
For gas-liquid absorption processes, membrane contactor (MC) technology offers a
variety of advantages over traditional packed columns. Optimal operation conditions for
CO2 removal from gas mixtures using MC are based on the appropriate choice of
membranes and absorbent solutions. The main objective of this thesis was (i) to develop a
dedicated absorbent solution presenting specific important properties for efficient gas
separation, such as good absorption capacity and reaction kinetics, regenerative potential,
resistance to degradation and high surface tension, and (ii) to investigate its application for
CO2 capture in MC.
As the choice of the absorbent to be used in MC has to be based on properties
related to its behavior in reaction with CO2, sterically hindered alkanolamine (SHA) based
solutions were considered as they are known to form less carbamate in solution, thus
offering higher absorption capacity and easier regeneration over conventional amines. For
an appropriate selection of SHA, we first investigated the molecular steric hindrance effect
on CO2 absorption kinetics of a SHA series composed by AMP (a simple hindrance form of
MEA) and three SHA derived from AMP (AEPD, AMPD and AHPD). For this study, the
kinetics of the reaction between CO2 and AHPD was performed experimentally at different
temperatures and solution concentrations using a wetted wall contactor and the results were
discussed together with data available for the other systems. The steric hindrance was
found to be inversely proportional to the reaction rate of these amines with CO2.
Although the alkanolamines with high steric hindrance present low kinetics, their
potential to reduce the energy consumption during the regeneration process brought us to
focus on AHPD, one of the most hindered alkanolamine investigated. To improve the
absorption rate of AHPD solution, piperazine, an amine presenting higher absorption rate
than MEA, was chosen as activator. The kinetics of the reaction between CO2 and
piperazine-activated aqueous solutions of AHPD was therefore performed in a wetted wall
contactor. Piperazine was found to be an effective activator of aqueous AHPD solutions as
300
the addition of small amounts of Pz has a significant effect on the enhancement of the CO2
absorption rate.
Along with good kinetics, the CO2 absorbent needs to present a good absorption
capacity. The thermodynamics of the aqueous CO2 + AHPD + Pz system was therefore
experimentally investigated using a liquid-vapor equilibrium apparatus based on a static-
synthetic method, and data were modelled with a modified Pitzer’s thermodynamic model
for the activity coefficients. The solubility of carbon dioxide in AHPD + Pz aqueous
solutions was predicted by supposing that the parameters characterising the single amines
systems (AHPD-CO2-H2O and Pz-CO2-H2O) were appropriate for describing the
quaternary system behaviour (AHPD-Pz-CO2-H2O). The experimental data for the single
amine system AHPD-CO2-H2O were satisfactorily correlated. The larger deviation obtained
between experimental and predicted equilibrium pressure for the quaternary AHPD-Pz-
CO2-H2O system was due to the fact that half of data available in the literature for Pz-CO2-
H2O were obtained in experimental conditions different from the CO2 + AHPD + Pz
system. Additional experimental data for CO2 solubility in aqueous piperazine solutions
were therefore obtained using the same VLE apparatus.
In addition to the absorbent absorption capacity and reaction kinetics toward CO2,
knowledge about the regeneration of loaded (CO2 containing) amine solutions are essential
for the analysis of economic viability of the absorption/desorption process. For the blend
AHPD + Pz to be an interesting absorbent for CO2 separation, it should also have
appropriate facility to regenerate. We therefore compared the regeneration capability of
different single SHA and Pz-activated aqueous solutions with that of MEA aqueous
solution (the most used amine in industrial applications). The results revealed that the
regeneration efficiency was in the order AHPD >> AMPD ≥ AEPD > MEA ≥ Pz > AMP.
These results clearly demonstrate that the systems containing the most sterically hindered
amines (AHPD, AMPD and AEPD), and in particular AHPD, could more easily be
regenerated because they do not form (or form to a small extend) stable carbamates in
solution. Moreover, the use of Pz as activator can offer an advantage over MEA due to
similar regeneration facility but faster absorption rate. The addition of small amount of Pz
301
into AHPD aqueous solution allowed to obtain almost the same cyclic capacity and
regeneration efficiency as non-activated solutions, but for half of the absorption time.
Based on the regeneration results and economic considerations, the aqueous AHPD + Pz
solution is favoured over the other tested amine solutions.
Besides the liquid absorbent properties, the performances of MC for CO2 separation
strongly depend on the compatibility between the absorbent and the membrane. Based on
wetting-related properties like liquid surface tension, contact angle, membrane
breakthrough pressure and chemical stability, a thorough analysis of these properties on
different potential membrane/liquid combinations (including the aqueous AHPD + Pz
solution) were performed in order to develop an appropriate way to select the best
conditions to elude the unwanted membrane wetting phenomenon. From this study, a new
graphical surface tension estimation method was developed, showing that the molecular
structure of a solute has a strong influence on the surface tension of its corresponding
aqueous solution. AHPD-based solutions (like AHPD + Pz) were found to have a strong
potential for use in MC because of their very high surface tension. The PTFE membranes
(high hydrophobicity and superior chemical and mechanical resistance) proved to represent
best options over PVDF and PP.
The developed graphical surface tension estimation method was found to be an
interesting and easy way to identify potential amines whose aqueous solutions present high
surface tensions, being, in this way, appropriate for use in MC. Following this method,
Serinol (2-amino-1,3-propanediol) seemed to be an amine whose aqueous solution surface
tension should be higher than that of typical amine solutions used in acid gas separations.
Although, in the light of the usual definition of SHA, Serinol is not necessarily such kind of
compound, it is nevertheless more hindered than MEA and, in the same time, could have
the advantage of much better kinetics toward CO2 compared to SHA. We therefore found
interesting to investigate the potential of this amine as an efficient CO2 absorbent to be used
in MC. Serinol aqueous solutions were therefore characterized by density, viscosity,
surface tension, and CO2 solubility measurements. The higher surface tension data obtained
for aqueous Serinol solutions, compared to conventional alkanolamines, could make this
302
absorbent very suitable for CO2 absorption using MC. The results validated in the same
time the predictive capacity of our surface tension estimation method. In addition, the CO2
cyclic capacity of Serinol was found to be 58% higher than that of MEA. On the whole, the
experimental results confirmed the potential of this alkanolamine to be used for CO2
removal especially in MC.
A good stability and resistance to degradation is another important feature
absorbents should have for being used in the CO2 absorption process. In this context, the
evaluation of the stability of aqueous AHPD + Pz and Serinol solutions to thermal and
oxidative degradation, in the absence and the presence of CO2, were performed and
compared to the results obtained for AMP (the most studied sterically hindered
alkanolamine) and MEA (the benchmark amine used in CO2 capture). It was found that
SHA were more resistant to thermal degradation than conventional amines, but the
presence of oxygen degraded them more significantly in the absence of CO2. The presence
of CO2 was beneficial to SHA stability due to the preferential bicarbonate formation in
solutions, which reduced to a large extent the oxidative degradation rate observed in the
absence of CO2. The addition of Pz to AHPD solution also reduced the AHPD oxidative
degradation percentage. The low degradation degree of the aqueous AHPD + Pz solution
reaffirmed its potential for application in the gas separation process, while Serinol was
found to degrade significantly.
After the study of CO2 absorption/regeneration properties, solution/membrane
compatibility and stability to degradation, the performance of AHPD + Pz blend (23 wt%
AHPD + 7% Pz) for CO2 absorption in PTFE hollow fiber membrane contactors was
investigated experimentally and theoretically. The results were compared with those
obtained for aqueous AHPD (23 wt%) and MEA (30 wt%) solutions, under the same
experimental conditions. It was found that at higher liquid flow rates, AHPD + Pz solution
outperformed both MEA and AHPD solutions due to the activator effect of Pz which has
very fast kinetics. At low flow rates, the performance of AHPD + Pz was similar to MEA,
but better compared to AHPD. Moreover, the modeling results clearly showed the effect of
303
the surface tension on membrane wetting (the wetting behaviour increased with the
decrease of the solution surface tension).
Finally, the performance of the AHPD + Pz aqueous solution (23 wt% AHPD + 7%
Pz) for CO2 absorption was also investigated in different flat sheet membrane contactors
(PTFE, PP and laminated PTFE/PP membranes), under various liquid flow rates, gas
compositions and flow orientation (co- or counter-current). The results were compared to
those obtained for aqueous AHPD (23 wt%) and MEA (30 wt%) solutions, under the same
experimental conditions. The results showed again the excellent performance of the AHPD
+ Pz solution. It was found that the membranes with lower thickness and smaller pore size
allowed to obtain higher CO2 absorption fluxes.
It should be noted that, under the same experimental conditions (fluid flow rates and
compositions, gas-liquid contact area, and absorption temperature), the CO2 absorption
fluxes were higher in the FSMC compared to HFMC. However, although the membrane
material was the same (PTFE), hollow fiber membranes had different properties (porosity,
pore size, thickness) in comparison with flat membranes. Therefore, a reliable comparison
between the performances of the two types of contactor is difficult to perform.
In summary, the results of this thesis showed that the AHPD + Pz aqueous solution
possess good absorption capacity, reaction kinetics, regenerative potential, degradation
resistance and high surface tension. This absorbent represents therefore an interesting
alternative to MEA for CO2 absorption processes, especially to be performed in highly
efficient gas-liquid membrane contactors.
The results presented in this thesis can open several directions for future research
projects. A detailed parametric study and optimisation of CO2 absorption performance
using various possible configurations of FSMC, especially on a semi-pilot scale including
both absorption and regeneration steps in cyclic operation, would be very interesting for a
future work. Also, the development of MC with absorbent solutions able to work at high
temperatures would possibly allow energy saving and widen the practical range of
application of these systems. In this context, for example, it would be interesting to study
both absorption/regeneration using MC modules. A study of the performance of HFMC and
304
FSMC modules using membranes presenting the same properties (porosity, pore size,
thickness), coupled with a techno-economical evaluation would be very useful for eventual
industrial implementations. Finally, the development and application in MC of highly
hydrophobic and cheaper membranes (compared to PTFE), with low thickness and small
pore size would also enhance the absorption performance.
305
References Abu-Khader, M.M., 2006. Recent progress in CO2 capture/sequestration: A review. Energy
Sources Part A-Recovery Util. Environ. Eff. 28, 1261-1279. Adamson, A.W., Gast, A.P., 1997. Physical chemistry of surfaces, 6th ed. Wiley, New
York. Aguila-Hernandez, J., Gomez-Quintana, R., Murrieta-Guevara, F., Romero-Martinez, A.,
Trejo, A., 2001. Liquid Density of Aqueous Blended Alkanolamines and N-Methylpyrrolidone as a Function of Concentration and Temperature. J. Chem. Eng. Data 46, 861-867.
Aguila-Hernandez, J., Trejo, A., Garcia-Flores, B.E., 2007. Surface Tension and Foam Behaviour of Aqueous Solutions of Blends of Three Alkanolamines, as a Function of Temperature. Colloids Surf., A 308, 33-46.
Ahmad, A.L., Sunarti, A.R., Lee, K.T., Fernando, W.J.N., 2010. CO2 removal using membrane gas absorption. Int. J. Greenhouse Gas Control 4, 495-498.
Al-Ghawas, H.A., Hagewiesche, D.P., Ruizibanez, G., Sandall, O.C., 1989. Physicochemical Properties Important for Carbon Dioxide Absorption in Aqueous Methyldiethanolamine. J. Chem. Eng. Data 34, 385-391.
Al-Marzouqi, M., El-Naas, M., Marzouk, S., Abdullatiff, N., 2008. Modeling of chemical absorption of CO2 in membrane contactors. Sep. Purif. Technol. 62, 499-506.
Ali, S.H., 2005. Kinetics of the Reaction of Carbon Dioxide with Blends of Amines in Aqueous Media using the Stopped-flow Technique. Int. J. Chem. Kinet. 37, 391-405.
Ali, S.H., Merchant, S.Q., Fahim, M.A., 2002. Reaction Kinetics of Some Secondary Alkanolamines with Carbon Dioxide in Aqueous Solutions by Stopped-flow Technique. Sep. Purif. Technol. 27, 121-136.
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, 1725-1728.
Alvarez-Fuster, C., Midoux, N., Laurent, A., Charpentier, J.C., 1980. Chemical kinetics of the reaction of carbon dioxide with with amines in pseudo m-nth order conditions in aqueous and organic solutions. Chem. Eng. Sci. 35, 1717-1723.
Alvarez-Fuster, C., Midoux, N., Laurent, A., Charpentier, J.C., 1981. Chemical kinetics of the reaction of CO2 with amines in pseudo m-nth order conditions in polar and viscous organic solutions. Chem. Eng. Sci. 36, 1513-1518.
Alvarez, E., Cancela, A., Maceiras, R., Navaza, J.M., Taboas, R., 2003. Surface Tension of Aqueous Binary Mixtures of 1-Amino-2-propanol and 3-Amino-1-propanol, and Aqueous Ternary Mixtures of these Amines with Diethanolamine, Triethanolamine, and 2-Amino-2-methyl-1-propanol from (298.15 to 323.15) K. J. Chem. Eng. Data 48, 32-35.
Alvarez, E., Gomez-Diaz, D., La Rubia, M.D., Navaz, J.M., 2006. Densities and Viscosities of Aqueous Ternary Mixtures of 2-(Methylamino) ethanol and 2-(Ethylamino) ethanol with Diethanolamine, Triethanolamine, N-Methyldiethanolamine, or 2-Amino-1-methyl-1-propanol from 298.15 to 323.15 K. J. Chem. Eng. Data 51, 955-962.
Alvarez, E., Gomez-Diaz, D., La Rubia, M.D., Navazalt, J.M., 2008. Surface Tension of Aqueous Binary Mixtures of 2-(Methylamino)ethanol and 2-(Ethylamino)ethanol and
306
Aqueous Ternary Mixtures of These Amines with Triethanolamine or N-Methyldiethanolamine from (293.15 to 323.15) K. J Chem Eng Data 53, 318-321.
Alvarez, E., Rendo, R., Sanjurjo, B., Sanchez-Vilas, M., Navaza, J.M., 1998. Surface Tension of Binary Mixtures of Water + N-Methyldiethanolamine and Ternary Mixtures of This Amine and Water with Monoethanolamine, Diethanolamine, and 2-Amino-2-methyl-1-propanol from 25 to 50 °C. J. Chem. Eng. Data 43, 1027-1029.
Amundsen, T.G., Oi, L.E., Eimer, D.A., 2009. Density and Viscosity of Monoethanolamine + Water + Carbon Dioxide from (25 to 80) °C. J Chem Eng Data 54, 3096-3100.
Arcis, H., Rodier, L., Coxam, J.Y., 2007. Enthalpy of Solution of CO2 in Aqueous Solutions of 2-Amino-2-methyl-1-propanol. J. Chem. Thermodyn. 39, 878-887.
Aroua, M.K., Haji-Sulaiman, M.Z., Ramasamy, K., 2002. Modelling of Carbon Dioxide Absorption in Aqueous Solutions of AMP and MDEA and Their Blends Using Aspenplus. Sep. Purif. Technol. 29, 153-162.
Aroua, M.K., Salleh, R.M., 2004. Solubility of CO2 in aqueous piperazine and its modeling using the Kent-Eisenberg approach. Chem. Eng. Technol. 27, 65-70.
Asprion, N., 2005. Surface Tension Models for Aqueous Amine Blends. Ind Eng Chem Res 44, 7270-7278.
Astarita, G., Savage, D.W., Bisio, A., 1983. Gas treating with chemical solvents. John Wiley, New York.
Atchariyawut, S., Feng, C., Wang, R., Jiraratananon, R., Liang, D.T., 2006. Effect of membrane structure on mass-transfer in the membrane gas-liquid contacting process using microporous PVDF hollow fibers. J. Membr. Sci. 285, 272-281.
Atchariyawut, S., Jiraratananon, R., Wang, R., 2007. Separation of CO2 from CH4 by using gas-liquid membrane contacting process. J. Membr. Sci. 304, 163-172.
Austgen, D.M., Rochelle, G.T., Peng, X., Chen, C.C., 1989. Model of Vapor Liquid Equilibria for Aqueous Acid Gas Alkanolamine Systems Using the Electrolyte NRTL Equation. Ind. Eng. Chem. Res. 28, 1060-1073.
Baek, J.I., Yoon, J.H., 1998. Solubility of Carbon Dioxide in Aqueous Solutions of 2-Amino-2-methyl-1,3-propanediol. J. Chem. Eng. Data 43, 635-637.
Baek, J.I., Yoon, J.H., Eum, H.M., 2000. Physical and Thermodynamic Properties of Aqueous 2-Amino-2-methyl-1,3-propanediol Solutions. Int. J. Thermophys. 21, 1175-1184.
Baker, R.W., 2004. Membrane technology and applications, 2nd ed. J. Wiley, Chichester ; New York.
Barbe, A.M., Hogan, P.A., Johnson, R.A., 2000. Surface morphology changes during initial usage of hydrophobic, microporous polypropylene membranes. J. Membr. Sci. 172, 149-156.
Barzagli, F., Mani, F., Peruzzini, M., 2010. Continuous Cycles of CO2 Absorption and Amine Regeneration with Aqueous Alkanolamines: a Comparison of the Efficiency Between Pure and Blended DEA, MDEA and AMP Solutions by C13 NMR Spectroscopy. Energy Environ. Sci. 3, 772-779.
Bedell, S.A., 2009. Oxidative degradation mechanisms for amines in flue gas capture. Energy Procedia 1, 771-778.
Bensetiti, Z., Iliuta, I., Larachi, F., Grandjean, B.P.A., 1999. Solubility of Nitrous Oxide in Amine Solutions. Ind. Eng. Chem. Res. 38, 328-332.
307
Bernardo, P., Drioli, E., Golemme, G., 2009. Membrane Gas Separation: A Review/State of the Art. Ind. Eng. Chem. Res. 48, 4638-4663.
Bishnoi, S., Rochelle, G.T., 2000. Absorption of Carbon Dioxide into Aqueous Piperazine: Reaction Kinetics, Mass Transfer and Solubility. Chem. Eng. Sci. 55, 5531-5543.
Blanco, A., Garcia-Abuin, A., Gomez-Diaz, D., Navaza, J.M., 2012. Surface Tension and Refractive Index of Benzylamine and 1,2-Diaminopropane Aqueous Solutions from T = (283.15 to 323.15) K. J Chem Eng Data 57, 2437-2441.
Blauwhoff, P.M.M., Versteeg, G.F., Vanswaaij, W.P.M., 1984. A study on the reaction between CO2 and alkanolamines in aqueous solutions. Chem. Eng. Sci. 39, 207-225.
Bosch, H., Versteeg, G.F., Van swaaij, W.P.M., 1989. Gas-Liquid Mass Transfer with Parallel Reversible Reactions .1. Absorption of CO2 into Solutions of Sterically Hindered Amines. Chem. Eng. Sci. 44, 2723-2734.
Bosch, H., Versteeg, G.F., Van swaaij, W.P.M., 1990. Kinetics of the Reaction of CO2 with the Sterically Hindered Amine 2-Amino-2-methylpropanol at 298 K. Chem. Eng. Sci. 45, 1167-1173.
Boucif, N., Favre, E., Roizard, D., 2008. CO2 capture in HFMM contactor with typical amine solutions: A numerical analysis. Chem. Eng. Sci. 63, 5375-5385.
Bougie, F., Iliuta, M.C., 2009. Kinetics of absorption of carbon dioxide into aqueous solutions of 2-amino-2-hydroxymethyl-1,3-propanediol. Chem. Eng. Sci. 64, 153-162.
Bougie, F., Iliuta, M.C., 2010a. Analysis of regeneration of sterically hindered alkanolamines aqueous solutions with and without activator. Chem. Eng. Sci. 65, 4746-4750.
Bougie, F., Iliuta, M.C., 2010b. CO2 Absorption into Mixed Aqueous Solutions of 2-Amino-2-hydroxymethyl-1,3-propanediol and Piperazine. Ind. Eng. Chem. Res. 49, 1150-1159.
Bougie, F., Iliuta, M.C., 2012. Sterically Hindered Amine-Based Absorbents for the Removal of CO2 from Gas Streams. J. Chem. Eng. Data 57, 635-669.
Bougie, F., Iliuta, M.C., 2013a. Analysis of Laplace-Young equation parameters and their influence on efficient CO2 capture in membrane contactors. Sep. Purif. Technol. 118, 806-815.
Bougie, F., Iliuta, M.C., 2013b. Solubility of CO2 in and Density, Viscosity and Surface Tension of Aqueous 2-Amino-1,3-propanediol (Serinol) Solutions. J. Chem. Eng. Data, Submitted to publication.
Bougie, F., Iliuta, M.C., 2014a. Solubility of CO2 in and Density, Viscosity, and Surface Tension of Aqueous 2-Amino-1,3-propanediol (Serinol) Solutions. J. Chem. Eng. Data 59, 355-361.
Bougie, F., Iliuta, M.C., 2014b. Thermal and oxidative degradation of aqueous amine solutions used for CO2 capture. Submitted
Bougie, F., Lauzon-Gauthier, J., Iliuta, M.C., 2009. Acceleration of the reaction of carbon dioxide into aqueous 2-amino-2-hydroxymethyl-1,3-propanediol Solutions by piperazine addition. Chem. Eng. Sci. 64, 2011-2019.
Bouhamra, W., Bavbek, O., Alper, E., 1999. Reaction Mechanism and Kinetics of Aqueous Solutions of 2-Amino-2-methyl-1,3-propandiol and Carbon Dioxide. Chemical Engineering Journal 73, 67-70.
308
Bradley, D.J., Pitzer, K.S., 1979. Thermodynamics of electrolytes. 12. Dielectric properties of water and Debye-Huckel parameters to 350°C and 1 kBar. J. Phys. Chem. 83, 1599-1603.
Brelvi, S.W., Oconnell, J.P., 1972. Corresponding states correlations for liquid compressibility and partial volumes of gases at infinite dilution in liquids. A.I.Ch.E. J. 18, 1239-&.
Camacho, F., Sanchez, S., Pacheco, R., Sanchez, A., La Rubia, M.D., 2005. Thermal Effects of CO2 Absorption in Aqueous Solutions of 2-Amino-2-methyl-1-propanol. A.I.Ch.E. J. 51, 2769-2777.
Caplow, M., 1968. Kinetics of carbamate formation and breakdown. J. Am. Chem. Soc. 90, 6795-&.
CEPA, 2005. CEPA 1999 annual report, in: Canada. Environment Canada. (Ed.). Environment Canada, Ottawa, p. v.
Chakraborty, A.K., Astarita, G., Bischoff, K.B., 1986. CO2 Absorption in Aqueous Solutions of Hindered Amines. Chem. Eng. Sci. 41, 997-1003.
Chan, C., Maham, Y., Mather, A.E., Mathonat, C., 2002. Densities and Volumetric Properties of the Aqueous Solutions of 2-Amino-2-methyl-1-propanol, n-Butyldiethanolamine and n-Propylethanolamine at Temperatures from 298.15 to 353.15 K. Fluid Phase Equilibr. 198, 239-250.
Chang, H.T., Posey, M., Rochelle, G.T., 1993. Thermodynamics of alkanolamine water solutions from freezing point measurements. Ind. Eng. Chem. Res. 32, 2324-2335.
Chang, L.C., Lin, T., Li, M.H., 2005. Mutual Diffusion Coefficients of Some Aqueous Alkanolamines Solutions. J. Chem. Eng. Data 50, 77-84.
Chauhan, R.K., Yoon, S.J., Lee, H., Kang, M.C., Min, B.M., 2003. Physical and transport properties of aqueous triisopropanolamine solutions. J Chem Eng Data 48, 291-293.
Chen, S.-C., Lin, S.-H., Chien, R.-D., Wang, Y.-H., Hsiao, H.-C., 2011. Chemical absorption of carbon dioxide with asymmetrically heated polytetrafluoroethylene membranes. Journal of Environmental Management 92, 1083-1090.
Chen, Y.J., Li, M.H., 2001. Heat Capacity of Aqueous Mixtures of Monoethanolamine with 2-Amino-2-methyl-1-propanol. J. Chem. Eng. Data 46, 102-106.
Chenlo, F., Moreira, R., Pereira, G., Vazquez, M.J., Santiago, E., 2001. Viscosities of Single-Solute and Binary-Solute Aqueous Systems of Monoethanolamine, Diethanolamine, and 2-Amino-2-methyl-1-propanol. J. Chem. Eng. Data 46, 276-280.
Cheong, W.J., Carr, P.W., 1987. The Surface Tension of Mixtures of Methanol, Acetonitrile, Tetrahydrofuran, Isopropanol, Tertiary Butanol and Dimethylsulfoxide with Water at 25 °C. J Liq Chromatogr 10, 561-581.
Chiu, L.F., Li, M.H., 1999. Heat Capacity of Alkanolamine Aqueous Solutions. J. Chem. Eng. Data 44, 1396-1401.
Chiu, L.F., Liu, H.F., Li, M.H., 1999. Heat Capacity of Alkanolamines by Differential Scanning Calorimetry. J. Chem. Eng. Data 44, 631-636.
Choi, W.J., Cho, K.C., Lee, S.S., Shim, J.G., Hwang, H.R., Park, S.W., Oh, K.J., 2007. Removal of Carbon Dioxide by Absorption into Blended Amines: Kinetics of Absorption into Aqueous AMP/HMDA, AMP/MDEA, and AMP/Piperazine Solutions. Green Chem. 9, 594-598.
309
Choi, W.J., Seo, J.B., Jang, S.Y., Jung, J.H., Oh, K.J., 2009. Removal Characteristics of CO2 using Aqueous MEA/AMP Solutions in the Absorption and Regeneration Process. J. Environ. Sci. 21, 907-913.
Chueh, C.F., Swanson, A.C., 1973a. Estimating Liquid Heat-Capacity. Chem. Eng. Prog. 69, 83-85.
Chueh, C.F., Swanson, A.C., 1973b. Estimation of Liquid Heat-Capacity. Can. J. Chem. Eng. 51, 596-600.
Clegg, S.L., Pitzer, K.S., 1992. Thermodynamics of Multicomponent, Miscible, Ionic-Solutions - Generalized Equations for Symmetrical Electrolytes. J. Phys. Chem. 96, 3513-3520.
Closmann, F., Nguyen, T., Rochelle, G.T., 2009. MDEA/Piperazine as a solvent for CO2 capture. 1, 1351-1357.
Constantinou, A., Barrass, S., Gavriilidis, A., 2014. CO2 Absorption in Polytetrafluoroethylene Membrane Microstructured Contactor Using Aqueous Solutions of Amines. dx.doi.org/10.1021/ie403444t.
Conway, W., Wang, X.G., Fernandes, D., Burns, R., Lawrance, G., Puxty, G., Maeder, M., 2013. Toward the Understanding of Chemical Absorption Processes for Post-Combustion Capture of Carbon Dioxide: Electronic and Steric Considerations from the Kinetics of Reactions of CO2(aq) with Sterically hindered Amines. Environ. Sci. Technol. 47, 1163-1169.
Covington, A.K., Ferra, M.I.A., Robinson, R.A., 1977. Ionic product and enthalpy of ionization of water from electromotive force measurements. Journal of the Chemical Society-Faraday Transactions I 73, 1721-1730.
Cui, Z., deMontigny, D., 2013. Part 7: A review of CO2 capture using hollow fiber membrane contactors. Carbon Manag. 4, 69-89.
Danckwerts, P.V., 1970. Gas-liquid reactions. McGraw-Hill Book Co., New York,. Danckwerts, P.V., 1979. The reaction of CO2 with ethanolamines. 34, 443-446. Dang, H.Y., Rochelle, G.T., 2003. CO2 absorption rate and solubility in
monoethanolamine/piperazine/water. Sep. Sci. Technol. 38, 337-357. Davis, R.A., Pogainis, B.J., 1995. Solubility of Nitrous Oxide in Aqueous Blends of N-
Methyldiethanolamine and 2-Amino-2-methyl-1-propanol. J. Chem. Eng. Data 40, 1249-1251.
deMontigny, D., Tontiwachwuthikul, P., Chakma, A., 2005. Comparing the Absorption Performance of Packed Columns and Membrane Contactors. Ind. Eng. Chem. Res. 44, 5726-5732.
deMontigny, D., Tontiwachwuthikul, P., Chakma, A., 2006. Using polypropylene and polytetrafluoroethylene membranes in a membrane contactor for CO2 absorption. J. Membr. Sci. 277, 99-107.
Derks, P.W., Hogendoorn, K.J., Versteeg, G.F., 2005a. Solubility of N2O in and Density, Viscosity, and Surface Tension of Aqueous Piperazine Solutions. J. Chem. Eng. Data 50, 1947-1950.
Derks, P.W.J., Dijkstra, H.B.S., Hogendoorn, J.A., Versteeg, G.F., 2005b. Solubility of carbon dioxide in aqueous piperazine solutions. A.I.Ch.E. J. 51, 2311-2327.
310
Derks, P.W.J., Hamborg, E.S., Hogendoorn, J.A., Niederer, J.P.M., Versteeg, G.F., 2008. Densities, viscosities, and liquid diffusivities in aqueous piperazine and aqueous (piperazine+ N-methyldiethanolamine) solutions. J. Chem. Eng. Data 53, 1179-1185.
Derks, P.W.J., Kleingeld, T., van Aken, C., Hogendoom, J.A., Versteeg, G.F., 2006. Kinetics of Absorption of Carbon Dioxide in Aqueous Piperazine Solutions. Chem. Eng. Sci. 61, 6837-6854.
Deshmukh, R.D., Mather, A.E., 1981. A Mathematical Model for Equilibrium Solubility of Hydrogen Sulfide and Carbon Dioxide in Aqueous Alkanolamine Solutions. Chem. Eng. Sci. 36, 355-362.
Dindore, V.Y., Brilman, D.W.F., Geuzebroek, F.H., Versteeg, G.F., 2004. Membrane-solvent selection for CO2 removal using membrane gas-liquid contactors. Sep. Purif. Technol. 40, 133-145.
Dindore, V.Y., Brilman, D.W.F., Versteeg, G.E., 2005. Modelling of cross-flow membrane contactors: Mass transfer with chemical reactions. J. Membr. Sci. 255, 275-289.
Drioli, E., Curcio, E., Di Profio, G., 2005. State of the art and recent progresses in membrane contactors. Chem. Eng. Res. Des. 83, 223-233.
Dymond, J.H., Smith, E.B., 1980. The virial coefficients of pure gases and mixtures : a critical compilation. Clarendon Press; Oxford University Press, Oxford, New York.
Edwards, T.J., Maurer, G., Newman, J., Prausnitz, J.M., 1978. Vapor-liquid equilibria in multicomponent aqueous solutions of volatile weak electrolytes. A.I.Ch.E. J. 24, 966-976.
El-Naas, M.H., Al-Marzouqi, M., Marzouk, S.A., Abdullatif, N., 2010. Evaluation of the removal of CO2 using membrane contactors: Membrane wettability. J Membrane Sci 350, 410-416.
Ermatchkov, V., Kamps, A.P.S., Maurer, G., 2003. Chemical equilibrium constants for the formation of carbamates in (carbon dioxide + piperazine + water) from H-1-NMR-spectroscopy. J. Chem. Thermodyn. 35, 1277-1289.
Ermatchkov, V., Kamps, A.P.S., Speyer, D., Maurer, G., 2006. Solubility of carbon dioxide in aqueous solutions of piperazine in the low gas loading region. J. Chem. Eng. Data 51, 1788-1796.
Ernst, R.C., Watkins, C.H., Ruze, H.H., 1936. The physical properties of the ternary system ethyl alcohol-glycerin-water. J Phys Chem-Us 40, 627-635.
Faiz, R., El-Naas, M.H., Al-Marzouqi, M., 2011. Significance of gas velocity change during the transport of CO2 through hollow fiber membrane contactors. Chemical Engineering Journal 168, 593-603.
Falk-Pedersen, O., Dannström, H., 1997. Separation of carbon dioxide from offshore gas turbine exhaust. Energ Convers Manage 38, S81-S86.
Fernandes, D., Conway, W., Wang, X.G., Burns, R., Lawrance, G., Maeder, M., Puxty, G., 2012. Protonation constants and thermodynamic properties of amines for post combustion capture of CO2. J. Chem. Thermodyn. 51, 97-102.
Feron, P.H.M., Jansen, A.E., 2002. CO2 separation with polyolefin membrane contactors and dedicated absorption liquids: performances and prospects. Sep. Purif. Technol. 27, 231-242.
Feron, P.H.M., Jansen, A.E., Klaassen, R., 1992. Membrane technology in carbon dioxide removal. Energ Convers Manage 33, 421-428.
311
Franken, A.C.M., Nolten, J.A.M., Mulder, M.H.V., Bargeman, D., Smolders, C.A., 1987. Wetting Criteria for the Applicability of Membrane Distillation. J Membrane Sci 33, 315-328.
Freeman, S.A., Davis, J., Rochelle, G.T., 2010. Degradation of Aqueous Piperazine in Carbon Dioxide Capture. Int. J. Greenhouse Gas Control 4, 756-761.
Freeman, S.A., Rochelle, G.T., 2012a. Thermal Degradation of Aqueous Piperazine for CO2 Capture. 1. Effect of Process Conditions and Comparison of Thermal Stability of CO2 Capture Amines. Ind. Eng. Chem. Res. 51, 7719-7725.
Freeman, S.A., Rochelle, G.T., 2012b. Thermal Degradation of Aqueous Piperazine for CO2 Capture: 2. Product Types and Generation Rates. Ind. Eng. Chem. Res. 51, 7726-7735.
Fu, R.K.Y., Mei, Y.F., Wan, G.J., Siu, G.G., Chu, P.K., Huang, Y.X., Tian, X.B., Yang, S.Q., Chen, J.Y., 2004. Surface composition and surface energy of Teflon treated by metal plasma immersion ion implantation. Surface Science 573, 426-432.
Gabelman, A., Hwang, S.T., 1999. Hollow fiber membrane contactors. J. Membr. Sci. 159, 61-106.
Gabrielsen, J., Michelsen, M.L., Stenby, E.H., Kontogeorgis, G.M., 2006. Modeling of CO2 Absorber using an AMP Solution. A.I.Ch.E. J. 52, 3443-3451.
Gianetto, A., Silveston, P.L., Baldi, G., 1986. Multiphase chemical reactors : theory, design, scale up. Hemisphere Pub. Corp.; Distribution outside North America, Springer-Verlag, Washington, Berlin ; New York.
Glasscock, D.A., 1990. Modelling and experimental study of carbon dioxide absorption into aqueous alkanolamines. The University of Texas at Austin.
Goldstein, A.M., Edelman, A.M., Beisner, W.D., Ruziska, P.A., 1984. Hindered Amines Yield Improved Gas Treating. Oil Gas J. 82, 70-76.
Gouedard, C., Picq, D., Launay, F., Carrette, P.L., 2012. Amine degradation in CO2 capture. I. A review. Int. J. Greenhouse Gas Control 10, 244-270.
Haji-Sulaiman, M.Z., Aroua, M.K., 1996. Equilibrium of CO2 in Aqueous Diethanolamine (DEA) and Amino Methyl Propanol (AMP) Solutions. Chem. Eng. Commun. 140, 157-171.
Haji-Sulaiman, M.Z., Aroua, M.K., Benamor, A., 1998. Analysis of Equilibrium Data of CO2 in Aqueous Solutions of Diethanolamine (DEA), Methyldiethanolamine (MDEA) and Their Mixtures Using the Modified Kent Eisenberg Model. Chem. Eng. Res. Des. 76, 961-968.
Han, J.Y., Jin, J., Eimer, D.A., Melaaen, M.C., 2012. Density of Water (1) + Monoethanolamine (2) + CO2 (3) from (298.15 to 413.15) K and Surface Tension of Water (1) + Monoethanolamine (2) from (303.15 to 333.15) K. J Chem Eng Data 57, 1095-1103.
Hayden, J.G., O’Connell, J.P., 1975. A Generalized Method for Predicting Second Virial Coefficients. 14, 209-216.
Hayduk, W., Laudie, H., 1974. Prediction of Diffusion Coefficients for Nonelectrolytes in Dilute Aqueous Solutions. A.I.Ch.E. J. 20, 611-615.
Henni, A., Hromek, J.J., Tontiwachwuthikul, P., Chakma, A., 2003. Volumetric Properties and Viscosities for Aqueous AMP Solutions from 25 °C to 70 °C. J. Chem. Eng. Data 48, 551-556.
312
Hetzer, H.B., Robinson, R.A., Bates, R.G., 1968. Dissociation constants of piperazinium ion and related thermodynamic quantities from 0 to 50 degrees. J. Phys. Chem. 72, 2081-&.
Higbie, R., 1935. The rate of absorption of a pure gas into a still liquid during short periods of exposure. Transactions of the American Institute of Chemical Engineers 31, 365-389.
Hikita, H., Asai, S., Katsu, Y., Ikuno, S., 1979. Absorption of Carbon Dioxide into Aqueous Monoethanolamine Solutions. A.I.Ch.E. J. 25, 793-800.
Ho, S.C., Chen, J.M., Li, M.H., 2007. Liquid Heat Capacity of Aqueous Sulfolane + 2-Amino-2-methyl-1-propanol Solutions. J. Chin. Inst. Chem. Eng. 38, 349-354.
Hobler, T., 1966. Mass transfer and absorbers, [1st English ed. Pergamon Press, Oxford, New York,.
Hoff, K.A., Juliussen, O., Falk-Pedersen, O., Svendsen, H.F., 2004. Modeling and experimental study of carbon dioxide absorption in aqueous alkanolamine solutions using a membrane contactor. Ind. Eng. Chem. Res. 43, 4908-4921.
Hoff, K.A., Svendsen, H.F., 2013. CO2 absorption with membrane contactos vs. packed absorbers - Challenges and opportunities in post combustion capture and natural gas sweetening. 952-960.
Hoke, B.C., Chen, J.C., 1991. Binary Aqueous Organic-Surface Tension Temperature Dependence. J Chem Eng Data 36, 322-326.
Hook, R.J., 1997. An Investigation of Some Sterically Hindered Amines as Potential Carbon Dioxide Scrubbing Compounds. Ind. Eng. Chem. Res. 36, 1779-1790.
Horng, S.Y., Li, M.H., 2002. Kinetics of absorption of carbon dioxide into aqueous solutions of monoethanolamine plus triethanolamine. Ind. Eng. Chem. Res. 41, 257-266.
Hsu, C.H., Li, M.H., 1997a. Densities of Aqueous Blended Amines. J. Chem. Eng. Data 42, 502-507.
Hsu, C.H., Li, M.H., 1997b. Viscosities of Aqueous Blended Amines. J. Chem. Eng. Data 42, 714-720.
Hu, W., Chakma, A., 1990. Modeling of Equilibrium Solubility of CO2 and H2S in Aqueous Amino Methyl Propanol (AMP) Solutions. Chem. Eng. Commun. 94, 53-61.
Iliuta, I., 2002. Reactoare multifazice : gaz, lichid, solid. Editura Academiei Române, Bucuresti.
Iliuta, I., Iliuta, M.C., Larachi, F., 2013. Catalytic CO2 hydration by immobilized and free human carbonic anhydrase II in a laminar flow microreactor - Model and simulations. Sep. Purif. Technol. 107, 61-69.
Iliuta, M.C., Bougie, F., Iliuta, I., 2014. CO2 removal by single and mixed amines in a hollow-fiber membrane module - Investigation of contactor performance.
Iliuta, M.C., Thyrion, F.C., 1995. Vapor-liquid equlibrium for the acetone methanol inorganic salt system. Fluid Phase Equilibr. 103, 257-284.
IPCC, 2005. IPCC Special Report on Carbon Dioxide Capture and Storage, Metz, B. ed. Cambridge University Press, for the Intergovernmental Panel on Climate Change, Cambridge.
Islam, M.N., Islam, M.M., Yeasmin, M.N., 2004. Viscosity of aqueous solutions of 2-methoxyethanol, 2-ethoxyethanol, and ethanolamine. J Chem Thermodyn 36, 889-893.
313
Islam, M.S., Yussof, R., Ali, B.S., Islam, M.N., Chakrabarti, M.H., 2011. Degradation studies of amines and alkanolamines during sour gas treatment process. Int. J. Phys. Sci. 6, 5877-5890.
Iversen, S.B., Bhatia, V.K., DamJohansen, K., Jonsson, G., 1997. Characterization of microporous membranes for use in membrane contactors. J. Membr. Sci. 130, 205-217.
Jane, I.S., Li, M.H., 1997. Solubilities of Mixtures of Carbon Dioxide and Hydrogen Sulfide in Water + Diethanolamine + 2-Amino-2-methyl-1-propanol. J. Chem. Eng. Data 42, 98-105.
Jou, F.Y., Mather, A.E., Otto, F.D., 1982. Solubility of H2S and CO2 in Aqueous Methyldiethanolamine Solutions. Ind. Eng. Chem. Process Des. Dev. 21, 539-544.
Jou, F.Y., Mather, A.E., Otto, F.D., 1995. The Solubility of CO2 in a 30 Mass Percent Monoethanolamine Solution. Can. J. Chem. Eng. 73, 140-147.
Jou, F.Y., Otto, F.D., Mather, A.E., 1994. Vapor-Liquid Equilibrium of Carbon Dioxide in Aqueous Mixtures of Monoethanolamine and Methyldiethanolamine. Ind. Eng. Chem. Res. 33, 2002-2005.
Jou, F.Y., Otto, F.D., Mather, A.E., 1998. Solubility of H2S, CO2, and Their Mixtures in an Aqueous Solution of 2-Piperidineethanol and Sulfolane. J. Chem. Eng. Data 43, 409-412.
Kadiwala, S., Rayer, A.V., Henni, A., 2010. High pressure solubility of carbon dioxide (CO2) in aqueous piperazine solutions. Fluid Phase Equilibr. 292, 20-28.
Kamps, A.P.S., Xia, J.Z., Maurer, G., 2003. Solubility of CO2 in (H2O+piperazine) and in (H2O+MDEA+piperazine). A.I.Ch.E. J. 49, 2662-2670.
Kelayeh, S.A., Jalili, A.H., Ghotbi, C., Hosseini-Jenab, M., Taghikhani, V., 2011. Densities, Viscosities, and Surface Tensions of Aqueous Mixtures of Sulfolane + Triethanolamine and Sulfolane + Diisopropanolamine. J Chem Eng Data 56, 4317-4324.
Kent, R.L., Eisenberg, B., 1976. Better Data for Amine Treating. Hydrocarbon Process. 55, 87-90.
Keshavarz, P., Fathikalajahi, J., Ayatollahi, S., 2008. Mathematical modeling of the simultaneous absorption of carbon dioxide and hydrogen sulfide in a hollow fiber membrane contactor. Sep Purif Technol 63, 145-155.
Khaisri, S., Demontigny, D., Tontiwachwuthikul, P., Jiraratananon, R., 2009. Comparing membrane resistance and absorption performance of three different membranes in a gas absorption membrane contactor. Sep Purif Technol 65, 290-297.
Khaisri, S., deMontigny, D., Tontiwachwuthikul, P., Jiraratananon, R., 2010. A mathematical model for gas absorption membrane contactors that studies the effect of partially wetted membranes. J. Membr. Sci. 347, 228-239.
Kim, Y.S., Yang, S.M., 2000. Absorption of carbon dioxide through hollow fiber membranes using various aqueous absorbents. Sep. Purif. Technol. 21, 101-109.
Ko, J.J., Tsai, T.C., Lin, C.Y., Wang, H.M., Li, M.H., 2001. Diffusivity of Nitrous Oxide in Aqueous Alkanolamine Solutions. J. Chem. Eng. Data 46, 160-165.
Kohl, A.L., Nielsen, R., 1997. Gas Purification, 5th ed. Gulf Publishing Company, Houston, TX.
Kosaraju, P., Kovvali, A.S., Korikov, A., Sirkar, K.K., 2005. Hollow Fiber Membrane Contactor Based CO2 Absorption-Stripping Using Novel Solvents and Membranes. Ind. Eng. Chem. Res. 44, 1250-1258.
314
Kreulen, H., Smolders, C.A., Versteeg, G.F., Vanswaaij, W.P.M., 1993. Microporous hollow fibre membrane modules as gas-liquid contactors. Part 2. Mass transfer with chemical reaction. J. Membr. Sci. 78, 217-238.
Kumar, P.S., Hogendoorn, J.A., Feron, P.H.M., Versteeg, G.F., 2002. New absorption liquids for the removal of CO2 from dilute gas streams using membrane contactors. Chem. Eng. Sci. 57, 1639-1651.
Kumazawa, H., 2000. Absorption and desorption of CO2 by aqueous solutions of sterically hindered 2-amino-2-methyl-1-propanol in hydrophobic microporous hollow fiber contained contactors. Chem. Eng. Commun. 182, 163-179.
Kundu, M., Mandal, B.P., Bandyopadhyay, S.S., 2003. Vapor-Liquid Equilibrium of CO2 in Aqueous Solutions of 2-Amino-2-methyl-1-propanol. J. Chem. Eng. Data 48, 789-796.
Lal, D., Otto, F.D., Mather, A.E., 1998. Solubility of Acid Gases in a Mixed Solvent. Can. J. Chem. Eng. 76, 964-966.
Le Tourneux, D., Iliuta, I., Iliuta, M.C., Fradette, S., Larachi, F., 2008. Solubility of Carbon Dioxide in Aqueous Solutions of 2-Amino-2-hydroxymethyl-1,3-propanediol. Fluid Phase Equilibr. 268, 121-129.
Lee, J.C., Yetter, R.A., Dryer, F.L., Tomboulides, A.G., Orszag, S.A., 2000. Simulation and analysis of laminar flow reactors. Combust. Sci. Technol. 159, 199-212.
Lepaumier, H., Grimstvedt, A., Vernstad, K., Zahlsen, K., Svendsen, H.F., 2011. Degradation of MMEA at Absorber and Stripper Conditions. Chem. Eng. Sci. 66, 3491-3498.
Lepaumier, H., Picq, D., Carrette, P.L., 2009a. New Amines for CO2 Capture. I. Mechanisms of Amine Degradation in the Presence of CO2. Ind. Eng. Chem. Res. 48, 9061-9067.
Lepaumier, H., Picq, D., Carrette, P.L., 2009b. New Amines for CO2 Capture. II. Oxidative Degradation Mechanisms. Ind. Eng. Chem. Res. 48, 9068-9075.
Lewis, W.K., Whitman, W.G., 1924. Principles of gas absorption. Ind. Eng. Chem. 16, 1215-1220.
Li, J.L., Chen, B.H., 2005. Review of CO2 absorption using chemical solvents in hollow fiber membrane contactors. Sep. Purif. Technol. 41, 109-122.
Li, M.H., Chang, B.C., 1994. Solubilities of Carbon Dioxide in Water + Monoethanolamine + 2-Amino-2-methyl-1-propanol. J. Chem. Eng. Data 39, 448-452.
Li, M.H., Lai, M.D., 1995. Solubility and Diffusivity of N2O and CO2 in (Monoethanolamine + N-Methyldiethanolamine + Water) and in (Monoethanolamine + 2-Amino-2-Methyl-1-Propanol + Water). J. Chem. Eng. Data 40, 486-492.
Li, M.H., Lee, W.C., 1996. Solubility and Diffusivity of N2O and CO2 in (Diethanolamine + N-Methyldiethanolamine + Water) and in (Diethanolamine + 2-Amino-2-methyl-1-propanol + Water). J. Chem. Eng. Data 41, 551-556.
Li, M.H., 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, 444-447.
Li, Y.G., Mather, A.E., 1998. Correlation and Prediction of the Solubility of CO2 and H2S in an Aqueous Solution of 2-Piperidineethanol and Sulfolane. Ind. Eng. Chem. Res. 37, 3098-3104.
315
Li, Z.B., Lu, B.C.Y., 2001. On the Prediction of Surface Tension for Multicomponent Mixtures. Can J Chem Eng 79, 402-411.
Liao, C.H., Li, M.H., 2002. Kinetics of absorption of carbon dioxide into aqueous solutions of monoethanolamine + N-methyldiethanolamine. Chem. Eng. Sci. 57, 4569-4582.
Lide, D., 2008. CRC handbook of chemistry and physics, 89th ed. ed, Boca Raton, FL. Lin, S.H., Chiang, P.C., Hsieh, C.F., Li, M.H., Tung, K.L., 2008. Absorption of Carbon
Dioxide by the Absorbent Composed of Piperazine and 2-Amino-2-methyl-1-propanol in PVDF Membrane Contactor. J. Chin. Inst. Chem. Eng. 39, 13-21.
Lin, S.H., Hsieh, C.F., Li, M.H., Tung, K.L., 2009a. Determination of mass transfer resistance during absorption of carbon dioxide by mixed absorbents in PVDF and PP membrane contactor. Desalination 249, 647-653.
Lin, S.H., Tung, K.L., Chang, H.W., Lee, K.R., 2009b. Influence of fluorocarbon flat-membrane hydrophobicity on carbon dioxide recovery. Chemosphere 75, 1410-1416.
Lin, S.H., Tung, K.L., Chen, W.J., Chang, H.W., 2009c. Absorption of carbon dioxide by mixed piperazine-alkanolamine absorbent in a plasma-modified polypropylene hollow fiber contactor. J. Membr. Sci. 333, 30-37.
Littel, R.J., Versteeg, G.F., Van swaaij, W.P.M., 1992. Solubility and Diffusivity Data for the Absorption of COS, CO2, and N2O in Amine Solutions. J. Chem. Eng. Data 37, 49-55.
Liu, L.Y., Li, L.J., Ding, Z.W., Ma, R.Y., Yang, Z.R., 2005. Mass transfer enhancement in coiled hollow fiber membrane modules. J. Membr. Sci. 264, 113-121.
Livingston, J., Morgan, R., Egloff, G., 1916. The properties of mixed liquids II Phenol-water and triethylamine-water mixtures. J Am Chem Soc 38, 844-857.
Lu, J.G., Wang, L.J., Sun, X.Y., Li, J.S., Liu, X.D., 2005. Absorption of CO2 into aqueous solutions of methyldiethanolamine and activated methyldiethanolamine from a gas mixture in a hollow fiber contactor. Ind. Eng. Chem. Res. 44, 9230-9238.
Lu, J.G., Zheng, Y.F., Cheng, M.D., 2008. Wetting mechanism in mass transfer process of hydrophobic membrane gas absorption. J Membrane Sci 308, 180-190.
Lu, J.G., Zheng, Y.F., Cheng, M.D., Wang, L.J., 2007. Effects of activators on mass-transfer enhancement in a hollow fiber contactor using activated alkanolamine solutions. J. Membr. Sci. 289, 138-149.
Luck, W.A.P., 2001. Understanding of surface tension? Colloid Polym Sci 279, 554-561. Lv, Y., Yu, X., Tu, S.-T., Yan, J., Dahlquist, E., 2010. Wetting of polypropylene hollow
fiber membrane contactors. J. Membr. Sci. 362, 444-452. Ma'mun, S., Jakobsen, J.P., Svendsen, H.F., Juliussen, O., 2006. Experimental and
modeling study of the solubility of carbon dioxide in aqueous 30 mass % 2-((2-aminoethyl)amino)ethanol solution. Ind. Eng. Chem. Res. 45, 2505-2512.
Ma'mun, S., Svendsen, H.F., Hoff, K.A., Juliussen, O., 2007. Selection of new absorbents for carbon dioxide capture. Energ Convers Manage 48, 251-258.
Mahajani, V.V., Joshi, J.B., 1988. Kinetics of Reactions Between Carbon Dioxide and Alkanolamines. Gas Sep. Purif. 2, 50-64.
Maham, Y., Hepler, L.G., Mather, A.E., Hakin, A.W., Marriott, R.A., 1997. Molar Heat Capacities of Alkanolamines from 299.1 to 397.8 K - Group Additivity and Molecular Connectivity Analyses. J. Chem. Soc., Faraday Trans. 93, 1747-1750.
316
Maham, Y., Mather, A.E., 2001. Surface thermodynamics of aqueous solutions of alkylethanolamines. Fluid Phase Equilibr 182, 325-336.
Mandal, B.P., Bandyopadhyay, S.S., 2005. Simultaneous Absorption of Carbon Dioxide and Hydrogen Sulfide into Aqueous Blends of 2-Amino-2-methyl-1-propanol and Diethanolamine. Chem. Eng. Sci. 60, 6438-6451.
Mandal, B.P., Bandyopadhyay, S.S., 2006. Absorption of Carbon Dioxide into Aqueous Blends of 2-Amino-2-methyl-1-propanol and Monoethanolamine. Chem. Eng. Sci. 61, 5440-5447.
Mandal, B.P., Biswas, A.K., Bandyopadhyay, S.S., 2003a. Absorption of Carbon Dioxide into Aqueous Blends of 2-Amino-2-methyl-1-propanol and Diethanolamine. Chem. Eng. Sci. 58, 4137-4144.
Mandal, B.P., Kundu, M., Bandyopadhyay, S.S., 2003b. Density and Viscosity of Aqueous Solutions of (N-Methyldiethanolamine + Monoethanolamine), (N-Methyldiethanolamine + Diethanolamine), (2-Amino-2-methyl-1-propanol + Monoethanolamine), and (2-Amino-2-methyl-1-propanol + Diethanolamine). J. Chem. Eng. Data 48, 703-707.
Mandal, B.P., Kundu, M., Bandyopadhyay, S.S., 2005. Physical Solubility and Diffusivity of N2O and CO2 into Aqueous Solutions of (2-Amino-2-methyl-1-propanol + Monoethanolamine) and (N-Methyldiethanolamine + Monoethanolamine). J. Chem. Eng. Data 50, 352-358.
Mandal, B.P., Kundu, M., Padhiyar, N.U., Bandyopadhyay, S.S., 2004. Physical Solubility and Diffusivity of N2O and CO2 into Aqueous Solutions of (2-Amino-2-methyl-1-propanol + Diethanolamine) and (N-Methyldiethanolamine + Diethanolamine). J. Chem. Eng. Data 49, 264-270.
Mansourizadeh, A., Ismail, A.F., 2009. Hollow fiber gas-liquid membrane contactors for acid gas capture: A review. Journal of Hazardous Materials 171, 38-53.
Mansourizadeh, A., Ismail, A.F., 2010. Effect of LiCl concentration in the polymer dope on the structure and performance of hydrophobic PVDF hollow fiber membranes for CO2 absorption. Chem Eng J 165, 980-988.
Marzouk, S.A.M., Al-Marzouqi, M.H., Teramoto, M., Abdullatif, N., Ismail, Z.M., 2012. Simultaneous removal of CO2 and H2S from pressurized CO2-H2S-CH4 gas mixture using hollow fiber membrane contactors. Sep. Purif. Technol. 86, 88-97.
Mavroudi, M., Kaldis, S.P., Sakellaropoulos, G.P., 2003. Reduction of CO2 emissions by a membrane contacting process. Fuel 82, 2153-2159.
Mavroudi, M., Kaldis, S.P., Sakellaropoulos, G.P., 2006. A study of mass transfer resistance in membrane gas-liquid contacting processes. J Membrane Sci 272, 103-115.
Mejdell, T., Hoff, K.A., Juliussen, O., Svendsen, H.F., Tobiesen, A., Vassbotn, T., 2010a. Amines as Absorbent for CO2 Removal from Gases.
Mejdell, T., Hoff, K.A., Juliussen, O., Svendsen, H.F., Tobiesen, A., Vassbotn, T., 2010b. Amines. WO/2010/037825 A1.
Mejia, A., Segura, H., Wisniak, J., Polishuk, I., 2005. Correlation and prediction of interface tension for fluid mixtures: An approach based on cubic equations of state with the Wong-Sandier mixing rule. J Phase Equilib Diff 26, 215-224.
317
Messaoudi, B., Sada, E., 1996. Kinetics of Absorption of Carbon Dioxide into Aqueous Solutions of Sterically Hindered 2-Amino-2-methyl-1-propanol. J. Chem. Eng. Jpn. 29, 193-196.
Missenard, F.A., 1965. Methode Additive pour la Determination de la Chaleur Molaire des Liquides. C.R. Hebd. Seances Acad. Sci. 260, 5521-5523.
Mittal, K.L., 2003. Contact angle, wettability and adhesion. VSP, Utrecht ; Boston. Montgomery, D.C., Runger, G.C., 1999. Applied statistics and probability for engineers,
2nd ed. John Wiley Sons, New York. Mosadegh-Sedghi, S., Rodrigue, D., Brisson, J., Iliuta, M.C., 2013. Highly hydrophobic
microporous LDPE hollow fiber membranes by melt-extrusion coupled with salt-leaching technique. Polymers for Advanced Technologies 24, 584-592.
Mosadegh-Sedghi, S., Rodrigue, D., Brisson, J., Iliuta, M.C., 2014. Wetting phenomenon in membrane contactors - Causes and prevention. J. Membr. Sci. 452, 332-353.
Muhammad, A., Mutalib, M.I.A., Murugesan, T., Shafeeq, A., 2009. Thermophysical Properties of Aqueous Piperazine and Aqueous (N-Methyldiethanolamine + Piperazine) Solutions at Temperatures (298.15 to 338.15) K. J Chem Eng Data 54, 2317-2321.
Muhlbauer, H.G., Monaghan, P.R., 1957. Sweetening Natural Gas With Ethanolamine Solutions. The Oil and Gas Journal 55, 139-145.
Munjal, P., Stewart, P.B., 1971. Correlation equation for solubility of carbon dioxide in water, seawater, and seawater concentrates. J. Chem. Eng. Data 16, 170-&.
Murrieta-Guevara, F., Rebolledo-Libreros, E., Trejo, A., 1992. Solubility of Hydrogen Sulfide in Mixtures of N-Methylpyrrolidone with Alkanolamines. Fluid Phase Equilibr. 73, 167-174.
Murrieta-Guevara, F., Rebolledo-Libreros, E., Trejo, A., 1994. Gas Solubility of Hydrogen Sulfide and Carbon Dioxide in Mixtures of Sulfolane with Diethanolamine at Different Temperatures. Fluid Phase Equilibr. 95, 163-174.
Murrieta-Guevara, F., Rebolledo-Libreros, M.E., Romero-Martinez, A., Trejo, A., 1998. Solubility of CO2 in Aqueous Mixtures of Diethanolamine with Methyldiethanolamine and 2-Amino-2-methyl-1-propanol. Fluid Phase Equilibr. 151, 721-729.
Murshid, G., Shariff, A.M., Keong, L.K., Bustam, M.A., 2011a. Physical Properties of Aqueous Solutions of Piperazine and (2-Amino-2-methyl-1-propanol + Piperazine) from (298.15 to 333.15) K. J Chem Eng Data 56, 2660-2663.
Murshid, G., Shariff, A.M., Keong, L.K., Bustam, M.A., Ahmad, F., 2011b. Thermophysical Analysis of Aqueous Solutions of Piperazine (PZ) and 2-Amino-2-hydroxymethyl-1,3-propanediol + Piperazine (PZ + AHPD). Int. J. Chem. Environ. Eng. 2, 318-321.
Murshid, G., Shariff, A.M., Lau, K.K., Bustam, M.A., Ahmad, F., 2011c. Physical Properties and Thermal Decomposition of Aqueous Solutions of 2-Amino-2-hydroxymethyl-1, 3-propanediol (AHPD). Int J Thermophys 32, 2040-2049.
Murshid, G., Shariff, A.M., Lau, K.K., Bustam, M.A., Ahmad, F., 2012. Physical Properties of Piperazine (PZ) Activated Aqueous Solutions of 2-Amino-2-hydroxymethyl-1,3-propanediol (AHPD + PZ). J Chem Eng Data 57, 133-136.
Nagvekar, M., Daubert, T.E., 1987. A Group Contribution Method for Liquid Thermal-Conductivity. Ind. Eng. Chem. Res. 26, 1362-1365.
318
Naim, R., Ismail, A.F., Mansourizadeh, A., 2012. Preparation of microporous PVDF hollow fiber membrane contactors for CO2 stripping from diethanolamine solution. J Membrane Sci 392, 29-37.
Nakanishi, K., Matsumoto, T., Hayatsu, M., 1971. Surface Tension of Aqueous Solutions of Some Glycols. J Chem Eng Data 16, 44-45.
Nguyen, T., Hilliard, M., Rochelle, G.T., 2010. Amine Volatility in CO2 Capture. Int. J. Greenhouse Gas Control 4, 707-715.
Nii, S., Takeuchi, H., 1994. Removal of CO2 and or SO2 from Gas Streams by a Membrane Absorption Method. Gas Sep. Purif. 8, 107-114.
Nishikawa, N., Ishibashi, M., Ohta, H., Akutsu, N., Matsumoto, H., Kamata, T., Kitamura, H., 1995. CO2 removal by hollow fiber gas-liquid contactor. Energ Convers Manage 36, 415-418.
Novak, J., Fried, V., Pick, J., 1961. Loslichkeit des kohlendioxyds in wasser bei verschiedenen drucken und temperaturen. Collect. Czech. Chem. Commun. 26, 2266-2270.
Nysing, R., Kramers, H., 1958. Absorption of CO2 in carbonate bicarbonate buffer solutions in a wetted wall column. Chem. Eng. Sci. 8, 81-89.
Othmer, D.F., Thakar, M.S., 1953. Correlating Diffusion Coefficients in Liquids. Ind. Eng. Chem. 45, 589-593.
Pacheco, M.A., Kaganoi, S., Rochelle, G.T., 2000. CO2 absorption into aqueous mixtures of diglycolamine and methyldiethanolamine. Chem. Eng. Sci. 55, 5125-5140.
Pagano, J.M., Fernelius, W.C., Goldberg, D.E., 1961. Thermodynamic study of homopiperazine, piperazine and N-(2-aminoehtyl)-piperazine and their complexes with copper(II) ion. J. Phys. Chem. 65, 1062-&.
Pappa, G.D., Anastasi, C., Voutsas, E.C., 2006. Measurement and Thermodynamic Modeling of the Phase Equilibrium of Aqueous 2-Amino-2-methyl-1-propanol Solutions. Fluid Phase Equilibr. 243, 193-197.
Park, J.Y., Yoon, S.J., Lee, H., 2003. Effect of Steric Hindrance on Carbon Dioxide Absorption into New Amine Solutions: Thermodynamic and Spectroscopic Verification Through Solubility and NMR Analysis. Environ. Sci. Technol. 37, 1670-1675.
Park, J.Y., Yoon, S.J., Lee, H., Yoon, J.H., Shim, J.G., Lee, J.K., Min, B.Y., Eum, H.M., 2002a. Density, Viscosity, and Solubility of CO2 in Aqueous Solutions of 2-Amino-2-hydroxymethyl-1,3-propanediol. J. Chem. Eng. Data 47, 970-973.
Park, J.Y., Yoon, S.J., Lee, H., Yoon, J.H., Shim, J.G., Lee, J.K., Min, B.Y., Eum, H.M., Kang, M.C., 2002b. Solubility of Carbon Dioxide in Aqueous Solutions of 2-Amino-2-ethyl-1,3-propanediol. Fluid Phase Equilibr. 202, 359-366.
Park, S.H., Lee, K.B., Hyun, J.C., Kim, S.H., 2002c. Correlation and Prediction of the Solubility of Carbon Dioxide in Aqueous Alkanolamine and Mixed Alkanolamine Solutions. Ind. Eng. Chem. Res. 41, 1658-1665.
Patterson, C.S., Busey, R.H., Mesmer, R.E., 1984. Second Ionization of Carbonic Acid in NaCl Media to 250°C. 13, 647-661.
Patterson, C.S., Slocum, G.H., Busey, R.H., Mesmer, R.E., 1982. Carbonate equilibria in hydrothermal systems - 1st ionization of carbonic acid in NaCl media to 300°C. Geochim. Cosmochim. Acta 46, 1653-1663.
319
Paul, S., Ghoshal, A.K., Mandal, B., 2007. Removal of CO2 by single and blended aqueous alkanolamine solvents in hollow-fiber membrane contactor: Modeling and simulation. Ind. Eng. Chem. Res. 46, 2576-2588.
Paul, S., Ghoshal, A.K., Mandal, B., 2008. Theoretical studies on separation of CO2 by single and blended aqueous alkanolamine solvents in flat sheet membrane contactor (FSMC). Chemical Engineering Journal 144, 352-360.
Paul, S., Ghoshal, A.K., Mandal, B., 2009a. Absorption of Carbon Dioxide into Aqueous Solutions of 2-Piperidineethanol: Kinetics Analysis. Ind. Eng. Chem. Res. 48, 1414-1419.
Paul, S., Ghoshal, A.K., Mandal, B., 2009b. Kinetics of Absorption of Carbon Dioxide into Aqueous Solutions of 2-Amino-2-hydroxymethyl-1,3-propanediol. Sep. Purif. Technol. 68, 422-427.
Paul, S., Ghoshal, A.K., Mandal, B., 2009c. Physicochemical Properties of Aqueous Solutions of 2-Amino-2-hydroxymethyl-1,3-propanediol. J. Chem. Eng. Data 54, 444-447.
Paul, S., Mandal, B., 2006a. Density and Viscosity of Aqueous Solutions of 2-Piperidineethanol, (2-Piperidineethanol + Monoethanolamine), and (2-Piperidineethanol + Diethanolamine) from (288 to 333) K. J. Chem. Eng. Data 51, 1406-1410.
Paul, S., Mandal, B., 2006b. Density and Viscosity of Aqueous Solutions of (2-Piperidineethanol + Piperazine) from (288 to 333) K and Surface Tension of Aqueous Solutions of (N-Methyldiethanolamine + Piperazine), (2-Amino-2-methyl-1-propanol + Piperazine), and (2-Piperidineethanol + Piperazine) from (293 to 323) K. J. Chem. Eng. Data 51, 2242-2245.
Paul, S., Mandal, B., 2006c. Density and viscosity of aqueous solutions of (N-methyldiethanolamine + piperazine) and (2-amino-2-methyl-1-propanol + piperazine) from (288 to 333) K. J. Chem. Eng. Data 51, 1808-1810.
Perrin, D.D., 1965. Dissociation constants of organic bases in aqueous solution. Butterworths, London,.
Pinsent, B.R.W., Pearson, L., Roughton, F.J.W., 1956. The kinetics of combination of carbon dioxide with hydroxide ions. Trans. Faraday Soc. 52, 1512-1520.
Pitzer, K.S., 1973. Thermodynamics of Electrolytes .1. Theoretical Basis and General Equations. J. Phys. Chem. 77, 268-277.
Puxty, G., Allport, A., Attalla, M., 2009a. Vapour liquid equilibria data for a range of new carbon dioxide absorbents. Energy Procedia 1, 941-947.
Puxty, G., Rowland, R., Allport, A., Yang, Q., Bown, M., Burns, R., Maeder, M., Attalla, M., 2009b. Carbon Dioxide Postcombustion Capture: A Novel Screening Study of the Carbon Dioxide Absorption Performance of 76 Amines. Environ. Sci. Technol. 43, 6427-6433.
Rajabzadeh, S., Yoshimoto, S., Teramoto, M., Al-Marzouqi, M., Matsuyama, H., 2009. CO2 absorption by using PVDF hollow fiber membrane contactors with various membrane structures. Separation and Purification Technology 69, 210-220.
Rangwala, H.A., 1996. Absorption of Carbon Dioxide into Aqueous Solutions using Hollow Fiber Membrane Contactors. J. Membr. Sci. 112, 229-240.
320
Rebolledo-Libreros, M.E., Trejo, A., 2004. Gas Solubility of CO2 in Aqueous Solutions of N-Methyldiethanolamine and Diethanolamine with 2-Amino-2-methyl-1-propanol. Fluid Phase Equilibr. 218, 261-267.
Rebolledo-Libreros, M.E., Trejo, A., 2006. Density and Viscosity of Aqueous Blends of Three Alkanolamines: N-Methyldiethanolamine, Diethanolamine, and 2-Amino-2-methyl-1-propanol in the Range of (303 to 343) K. J. Chem. Eng. Data 51, 702-707.
Reid, R.C., Prausnitz, J.M., Poling, B.E., 1987. The properties of gases and liquids, 4th ed. McGraw-Hill, New York.
Reza, J., Trejo, A., 2006. Degradation of Aqueous Solutions of Alkanolamine Blends at High Temperature, under the Presence of CO2 and H2S. Chem. Eng. Commun. 193, 129-138.
Roberts, B.E., Mather, A.E., 1988a. Solubility of CO2 and H2S in a Hindered Amine Solution. Chem. Eng. Commun. 64, 105-111.
Roberts, B.E., Mather, A.E., 1988b. Solubility of CO2 and H2S in a Mixed Solvent. Chem. Eng. Commun. 72, 201-211.
Roberts, D., Danckwerts, P.V., 1962. Kinetics of CO2 absorption in alkaline solutions. 1. Transient absorption rates and catalysis by arsenite. Chem. Eng. Sci. 17, 961-969.
Rochelle, G.T., 2012. Thermal degradation of amines for CO2 capture. Curr. Opin. Chem. Eng. 1, 183-190.
Rongwong, W., Jiraratananon, R., Archariyawut, S., 2009. Experimental study on membrane wetting in gas-liquid membrane contacting process for CO2 absorption by single and mixed absorbents. Sep. Purif. Technol. 69, 118-125.
Rumpf, B., Maurer, G., 1993. An experimental and theoretical investigation on the solubility of carbon dioxide in aqueous solutions of strong electrolytes. Ber. Bunsen-Ges. Phys. Chem. Chem. Phys. 97, 85-97.
Saha, A.K., Bandyopadhyay, S.S., Biswas, A.K., 1993. Solubility and Diffusivity of N2O and CO2 in Aqueous Solutions of 2-Amino-2-methyl-1-propanol. J. Chem. Eng. Data 38, 78-82.
Saha, A.K., Bandyopadhyay, S.S., Biswas, A.K., 1995. Kinetics of Absorption of CO2 into Aqueous Solutions of 2-Amino-2-methyl-1-propanol. Chem. Eng. Sci. 50, 3587-3598.
Saha, A.K., Biswas, A.K., Bandyopadhyay, S.S., 1999. Absorption of CO2 in a Sterically Hindered Amine: Modeling Absorption in a Mechanically Agitated Contactor. Sep. Purif. Technol. 15, 101-112.
Sakwattanapong, R., Aroonwilas, A., Veawab, A., 2005. Behavior of Reboiler Heat Duty for CO2 Capture Plants using Regenerable Single and Blended Alkanolamines. Ind. Eng. Chem. Res. 44, 4465-4473.
Samanta, A., Bandyopadhyay, S.S., 2006. Density and Viscosity of Aqueous Solutions of Piperazine and (2-Amino-2-methyl-1-propanol + Piperazine) from 298 to 333 K. J. Chem. Eng. Data 51, 467-470.
Samanta, A., Bandyopadhyay, S.S., 2007. Kinetics and modeling of carbon dioxide absorption into aqueous solutions of piperazine. Chem. Eng. Sci. 62, 7312-7319.
Samanta, A., Bandyopadhyay, S.S., 2009. Absorption of Carbon Dioxide into Aqueous Solutions of Piperazine Activated 2-Amino-2-methyl-1-propanol. Chem. Eng. Sci. 64, 1185-1194.
321
Sartori, G., Ho, W.S., Savage, D.W., Chludzinski, G.R., Wiechert, S., 1987. Sterically-Hindered Amines for Acid-Gas Absorption. Sep. Purif. Methods 16, 171-200.
Sartori, G., Savage, D.W., 1983. Sterically Hindered Amines for CO2 Removal from Gases. Ind. Eng. Chem. Fund. 22, 239-249.
Saul, A., Wagner, W., 1987. International equations for the saturation properties of ordinary water substance. J. Phys. Chem. Ref. Data 16, 893-901.
Say, G.R., Heinzelmann, F.J., Iyengar, J.N., Savage, D.W., Bisio, A., Sartori, G., 1984. A New, Hindered Amine Concept for Simultaneous Removal of CO2 and H2S from Gases. Chem. Eng. Prog. 80, 72-77.
Scheibel, E.G., 1954. Liquid Diffusivities. Ind. Eng. Chem. 46, 2007-2008. Sea, B., Park, Y.I., Lee, K.H., 2002. Comparison of porous hollow fibers as a membrane
contactor for carbon dioxide absorption. Journal of Industrial and Engineering Chemistry 8, 290-296.
Seo, D.J., Hong, W.H., 1996. Solubilities of Carbon Dioxide in Aqueous Mixtures of Diethanolamine and 2-Amino-2-methyl-1-propanol. J. Chem. Eng. Data 41, 258-260.
Seo, D.J., Hong, W.H., 2000. Effect of piperazine on the kinetics of carbon dioxide with aqueous solutions of 2-amino-2-methyl-1-propanol. Ind. Eng. Chem. Res. 39, 2062-2067.
Sharma, M.M., 1965. Kinetics of Reactions of Carbonyl Sulphide and Carbon Dioxide with Amines and Catalysis by Bronsted Bases of Hydrolysis of COS. Trans. Faraday Soc. 61, 681-&.
Shen, K.P., Li, M.H., 1992. Solubility of Carbon Dioxide in Aqueous Mixtures of Monoethanolamine with Methyldiethanolamine. J. Chem. Eng. Data 37, 96-100.
Shen, K.P., Li, M.H., Yih, S.M., 1991. Kinetics of Carbon Dioxide Reaction with Sterically Hindered 2-Piperidineethanol Aqueous Solutions. Ind. Eng. Chem. Res. 30, 1811-1813.
Shih, T.W., Chen, Y.J., Li, M.H., 2002. Heat Capacity of Aqueous Mixtures of Monoethanolamine with 2-Piperidineethanol. Thermochim. Acta 389, 33-41.
Shih, T.W., Li, M.H., 2002. Heat Capacity of Aqueous Mixtures of Diethanolamine with 2-Amino-2-methyl-1-propanol. Fluid Phase Equilibr. 202, 233-237.
Silkenbaumer, D., Rumpf, B., Lichtenthaler, R.N., 1998. Solubility of Carbon Dioxide in Aqueous Solutions of 2-Amino-2-methyl-1-propanol and N-Methyldiethanolamine and Their Mixtures in the Temperature Range from 313 to 353 K and Pressures up to 2.7 MPa. Ind. Eng. Chem. Res. 37, 3133-3141.
Snijder, E.D., Teriele, M.J.M., Versteeg, G.F., Vanswaaij, W.P.M., 1993. Diffusion Coefficients of Several Aqueous Alkanolamine Solutions. J. Chem. Eng. Data 38, 475-480.
Sohrabi, M.R., Marjani, A., Moradi, S., Davallo, M., Shirazian, S., 2011. Mathematical modeling and numerical simulation of CO2 transport through hollow-fiber membranes. Applied Mathematical Modelling 35, 174-188.
Speyer, D., Ermatchkov, V., Maurer, G., 2010. Solubility of Carbon Dioxide in Aqueous Solutions of N-Methyldiethanolamine and Piperazine in the Low Gas Loading Region. J. Chem. Eng. Data 55, 283-290.
Sun, W.C., Yong, C.B., Li, M.H., 2005. Kinetics of the Absorption of Carbon Dioxide into Mixed Aqueous Solutions of 2-Amino-2-methyl-l-propanol and Piperazine. Chem. Eng. Sci. 60, 503-516.
322
Supap, T., Idem, R., Tontiwachwuthikul, P., Saiwan, C., 2006. Analysis of Monoethanolamine and its Oxidative Degradation Products during CO2 Absorption from Flue Gases: A Comparative Study of GC-MS, HPLC-RID, and CE-DAD Analytical Techniques and Possible Optimum Combinations. Ind. Eng. Chem. Res. 45, 2437-2451.
Teng, T.T., Mather, A.E., 1989. Solubility of H2S, CO2 and Their Mixtures in an AMP Solution. Can. J. Chem. Eng. 67, 846-850.
Teng, T.T., Mather, A.E., 1990. Solubility of CO2 in an AMP Solution. J. Chem. Eng. Data 35, 410-411.
Tesser, R., Bottino, A., Capannelli, G., Montagnaro, F., Vitolo, S., Di Serio, M., Santacesaria, E., 2005. Advantages in the use of membrane contactors for the study of gas-liquid and gas-liquid-solid reactions. Ind. Eng. Chem. Res. 44, 9451-9460.
Thomas, W.J., Furzer, I.A., 1962. Diffusion measurements in liquids by the Gouy method. Chem. Eng. Sci. 17, 115-120.
Tobiesen, F.A., Svendsen, H.F., 2006. Study of a Modified Amine-based Regeneration Unit. Ind. Eng. Chem. Res. 45, 2489-2496.
Tontiwachwuthikul, P., Meisen, A., Lim, C.J., 1991. Solubility of CO2 in 2-Amino-2-methyl-1-propanol Solutions. J. Chem. Eng. Data 36, 130-133.
Traube, I., 1891. Annalne Chemie 265, 27. Treybal, R.E., 1967. Mass-transfer operations, 2d ed. McGraw-Hill, New York,. Tsai, T.C., Ko, J.J., Wang, H.M., Lin, C.Y., Li, M.H., 2000. Solubility of Nitrous Oxide in
Alkanolamine Aqueous Solutions. J. Chem. Eng. Data 45, 341-347. Vahidi, M., Matin, N.S., Goharrokhi, M., Jenab, M.H., Abdi, M.A., Najibi, S.H., 2009.
Correlation of CO2 solubility in N-methyldiethanolamine + piperazine aqueous solutions using extended Debye-Huckel model. J. Chem. Thermodyn. 41, 1272-1278.
Vaidya, P.D., Kenig, E.Y., 2007. CO2-Alkanolamine Reaction Kinetics: A Review of Recent Studies. Chem. Eng. Technol. 30, 1467-1474.
Vaidya, P.D., Kenig, E.Y., 2008. Acceleration of CO2 reaction with N,N-diethylethanolamine in aqueous solutions by piperazine. Ind. Eng. Chem. Res. 47, 34-38.
Vazquez, G., Alvarez, E., Navaza, J.M., 1995. Surface Tension of Alcohol + Water from 20 °C to 50 °C. J Chem Eng Data 40, 611-614.
Vazquez, G., Alvarez, E., Navaza, J.M., Rendo, R., Romero, E., 1997. Surface Tension of Binary Mixtures of Water + Monoethanolamine and Water + 2-Amino-2-methyl-1-propanol and Tertiary Mixtures of these Amines with Water from 25 °C to 50 °C. J. Chem. Eng. Data 42, 57-59.
Vazquez, G., Alvarez, E., Rendo, R., Romero, E., Navaza, J.M., 1996. Surface Tension of Aqueous Solutions of Diethanolamine and Triethanolamine from 25 °C to 50 °C. J. Chem. Eng. Data 41, 806-808.
Veawab, A., Tontiwachwuthikul, P., Bhole, S.D., 1996. Corrosivity in 2-Amino-2-methyl-1-propanol (AMP)-CO2 System. Chem. Eng. Commun. 144, 65-71.
Veawab, A., Tontiwachwuthikul, P., Bhole, S.D., 1997. Studies of Corrosion and Corrosion Control in a CO2-2-Amino-2-methyl-1-propanol (AMP) Environment. Ind. Eng. Chem. Res. 36, 264-269.
323
Veawab, A., Tontiwachwuthikul, P., Chakma, A., 1999. Influence of Process Parameters on Corrosion Behavior in a Sterically Hindered Amine-CO2 System. Ind. Eng. Chem. Res. 38, 310-315.
Venkat, A., Kumar, G., Kundu, M., 2010a. Density and Surface Tension of Aqueous Solutions of (2-(Methylamino)-ethanol + 2-Amino-2-methyl-1-propanol) and (2-(Methylamino)-ethanol + N-Methyldiethanolamine) from (298.15 to 323.15) K. J. Chem. Eng. Data 55, 4580-4585.
Venkat, A., Kumar, G., Kundu, M., 2010b. Density and Surface Tension of Aqueous Solutions of (2-(Methylamino)-ethanol+2-Amino-2-methyl-1-propanol) and (2-(Methylamino)-ethanol + N-Methyl-diethanolamine) from (298.15 to 323.15) K. J Chem Eng Data 55, 4580-4585.
Versteeg, G.F., Van Dijck, L.A.J., Van Swaaij, W.P.M., 1996. On the Kinetics Between CO2 and Alkanolamines Both in Aqueous and Non-aqueous Solutions. An Overview. Chem. Eng. Commun. 144, 113-158.
Versteeg, G.F., Vanswaaij, W.P.M., 1988. Solubility and Diffusivity of Acid Gases (CO2, N2O) in Aqueous Alkanolamine Solutions. J. Chem. Eng. Data 33, 29-34.
Vogt, M., Goldschmidt, R., Bathen, D., Epp, B., Fahlenkamp, H., 2011. Comparison of membrane contactor and structured packings for CO2 absorption, in: Gale, J., Hendriks, C., Turkenberg, W. (Eds.), 10th International Conference on Greenhouse Gas Control Technologies, pp. 1471-1477.
Wang, H.M., Li, M.H., 2004. Kinetics of Absorption of Carbon Dioxide into Aqueous Solutions of 2-Amino-2-methyl-l-propanol + Diethanolamine. J. Chem. Eng. Jpn. 37, 267-278.
Wang, K.L., Cussler, E.L., 1993. Baffled membrane modules made with hollow-fiber fabric. J. Membr. Sci. 85, 265-278.
Wang, M., Lawal, A., Stephenson, P., Sidders, J., Ramshaw, C., 2011. Post-combustion CO2 capture with chemical absorption: A state-of-the-art review. Chem. Eng. Res. Des. 89, 1609-1624.
Wang, R., Li, D.F., Liang, D.T., 2004. Modeling of CO2 capture by three typical amine solutions in hollow fiber membrane contactors. Chem. Eng. Process. 43, 849-856.
Wang, R., Zhang, H.Y., Feron, P.H.M., Liang, D.T., 2005. Influence of membrane wetting on CO2 capture in microporous hollow fiber membrane contactors. Separation and Purification Technology 46, 33-40.
Wang, T., Jens, K.J., 2012. Oxidative Degradation of Aqueous 2-Amino-2-methyl-1-propanol Solvent for Postcombustion CO2 Capture. Industrial and Engineering Chemistry Research 51, 6529-6536.
Wang, Y.W., Xu, S., Otto, F.D., Mather, A.E., 1992. Solubility of N2O in Alkanolamines and in Mixed Solvents. Chem. Eng. J. 48, 31-40.
Wang, Z., Fang, M., Yan, S., Yu, H., Wei, C.-C., Luo, Z., 2013. Optimization of Blended Amines for CO2 Absorption in a Hollow-Fiber Membrane Contactor. Ind. Eng. Chem. Res. 52, 12170-12182.
Weiland, R.H., Dingman, J.C., Cronin, D.B., Browning, G.J., 1998. Density and Viscosity of Some Partially Carbonated Aqueous Alkanolamine Solutions and Their Blends. J. Chem. Eng. Data 43, 378-382.
324
Welsh, L.M., Davis, R.K., 1995. Density and Viscosity of Aqueous Blends of N-Methyldiethanolamine and 2-Amino-2-methyl-1-propanol. J. Chem. Eng. Data 40, 257-259.
Xiao, J., Li, C.W., 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, 161-175.
Xu, G.W., Zhang, C.F., Qin, S.J., Wang, Y.W., 1992a. Kinetics study on absorption of carbon dioxide into solutions of activated methyldiethanolamine. Ind. Eng. Chem. Res. 31, 921-927.
Xu, S., Otto, F.D., Mather, A.E., 1991. Physical Properties of Aqueous AMP Solutions. J. Chem. Eng. Data 36, 71-75.
Xu, S., Wang, Y.W., Otto, F.D., Mather, A.E., 1992b. Physicochemical Properties of 2-Piperidineethanol and Its Aqueous Solutions. J. Chem. Eng. Data 37, 407-411.
Xu, S., Wang, Y.W., Otto, F.D., Mather, A.E., 1992c. Representation of the Equilibrium Solubility Properties of CO2 with Aqueous Solutions of 2-Amino-2-methyl-1-propanol. Chem. Eng. Process. 31, 7-12.
Xu, S., Wang, Y.W., Otto, F.D., Mather, A.E., 1993a. Kinetics of the Reaction of Carbon Dioxide with Aqueous 2-Piperidineethanol Solutions. A.I.Ch.E. J. 39, 1721-1725.
Xu, S., Wang, Y.W., Otto, F.D., Mather, A.E., 1993b. The Physicochemical Properties of the Mixed Solvent of 2-Piperidineethanol, Sulfolane and Water. J. Chem. Technol. Biotechnol. 56, 309-316.
Xu, S., Wang, Y.W., Otto, F.D., Mather, A.E., 1996. Kinetics of the Reaction of Carbon Dioxide with 2-Amino-2-methyl-1-propanol Solutions. Chem. Eng. Sci. 51, 841-850.
Yan, S., Fang, M., Zhang, W., Zhong, W., Luo, Z., Cen, K., 2008. Comparative analysis of CO2 separation from flue gas by membrane gas absorption technology and chemical absorption technology in China. Energ Convers Manage 49, 3188-3197.
Yan, S.P., Fang, M.X., Zhang, W.F., Wang, S.Y., Xu, Z.K., Luo, Z.Y., Cen, K.F., 2007. Experimental study on the separation of CO2 from flue gas using hollow fiber membrane contactors without wetting. Fuel Processing Technology 88, 501-511.
Yang, M.C., Cussler, E.L., 1986. Designing Hollow-Fiber Contactors. A.I.Ch.E. J. 32, 1910-1916.
Yang, Z.Y., Soriano, A.N., Caparanga, A.R., Li, M.H., 2010. Equilibrium Solubility of Carbon Dioxide in (2-Amino-2-methyl-1-propanol + Piperazine + Water). J. Chem. Thermodyn. 42, 659-665.
Yeon, S.H., Lee, K.S., Sea, B., Park, Y.I., Lee, K.H., 2005. Application of pilot-scale membrane contactor hybrid system for removal of carbon dioxide from flue gas. J. Membr. Sci. 257, 156-160.
Yeon, S.H., Sea, B., Park, Y.I., Lee, K.H., 2003. Determination of mass transfer rates in PVDF and PTFE hollow fiber membranes for CO2 absorption. Sep. Sci. Technol. 38, 271-293.
Yeon, S.H., Sea, B., Park, Y.I., Lee, K.S., Lee, K.H., 2004. Absorption of carbon dioxide characterized by using the absorbent composed of piperazine and triethanolamine. Sep. Sci. Technol. 39, 3281-3300.
Yih, S.M., Shen, K.P., 1988. Kinetics of Carbon Dioxide Reaction with Sterically Hindered 2-Amino-2-methyl-1-propanol Aqueous Solutions. Ind. Eng. Chem. Res. 27, 2237-2241.
325
Yoon, J.H., Baek, J.I., Yamamoto, Y., Komai, T., Kawamura, T., 2003. Kinetics of Removal of Carbon Dioxide by Aqueous 2-Amino-2-methyl-1,3-propanediol. Chem. Eng. Sci. 58, 5229-5237.
Yoon, S.J., Lee, H., 2003. Substituent Effect in Amine-CO2 Interaction Investigated by NMR and IR Spectroscopies. Chemistry Letters 32, 344-345.
Yoon, S.J., Lee, H., Yoon, J.H., Shim, J.G., Lee, J.K., Min, B.Y., Eum, H.M., 2002a. Kinetics of Absorption of Carbon Dioxide into Aqueous 2-Amino-2-ethyl-1,3-propanediol Solutions. Ind. Eng. Chem. Res. 41, 3651-3656.
Yoon, S.J., Lee, H.S., Lee, H., Baek, J.I., Yoon, J.H., Eum, H.M., 2002b. Densities, Viscosities, and Surface Tensions of Aqueous 2-Amino-2-ethyl-1,3-propanediol Solutions. J. Chem. Eng. Data 47, 30-32.
You, J.K., Park, H., Yang, S.H., Hong, W.H., Shin, W., Kang, J.K., Yi, K.B., Kim, J.N., 2008. Influence of Additives Including Amine and Hydroxyl Groups on Aqueous Ammonia Absorbent for CO2 Capture. J. Phys. Chem. B 112, 4323-4328.
Zhang, K., Hawrylak, B., Palepu, R., Tremaine, P.R., 2002. Thermodynamics of Aqueous Amines: Excess Molar Heat Capacities, Volumes, and Expansibilities of {Water + Methyldiethanolamine (MDEA)} and {Water + 2-Amino-2-methyl-1-propanol (AMP)}. J. Chem. Thermodyn. 34, 679-710.
Zhang, P., Shi, Y., Wei, J.W., 2007. Kinetics Region and Model for Mass Transfer in Carbon Dioxide Absorption into Aqueous Solution of 2-Amino-2-methyl-1-propanol. Sep. Purif. Technol. 56, 340-347.
Zhang, P., Shi, Y., Wei, J.W., Zhao, W., Ye, Q., 2008. Regeneration of 2-Amino-2-methyl-1-propanol Used for Carbon Dioxide Absorption. J. Environ. Sci. 20, 39-44.
Zhang, Q., Cussler, E.L., 1985a. Microporous Hollow Fibers for Gas Absorption .1. Mass-Transfer in the Liquid. J. Membr. Sci. 23, 321-332.
Zhang, Q., Cussler, E.L., 1985b. Microporous Hollow Fibers for Gas Absorption .2. Mass-Transfer across the Membrane. J. Membr. Sci. 23, 333-345.
Zhang, X., Zhang, C.F., Qin, S.J., Zheng, Z.S., 2001. A kinetics study on the absorption of carbon dioxide into a mixed aqueous solution of methyldiethanolamine and piperazine. Ind. Eng. Chem. Res. 40, 3785-3791.
Zhang, Z.T., Gao, J., Zhang, W.D., Ren, Z.Q., 2006. Experimental Study of the Effect of Membrane Porosity on Membrane Absorption Process. Sep. Sci. Technol. 41, 3245-3263.
326
327
Appendix A
Table A.1. Density data of AMP systems System T
(K) ∆T (K)
[AMP] (wt%)
[Amine1] (wt%)
∆[AM2] (wt%)
∆ρ (g·cm-3)
Reference
AMP 313 - 2-27 - - - (Yih and Shen, 1988) AMP 293-363 0.05 9-100 - - 1×10-5 (Xu et al., 1991) AMP 303 - 1-45 - - 5×10-5 (Littel et al., 1992) AMP 288-313 - 4.5-18 - - - (Saha et al., 1993) AMP 303-353 0.05 100 - - 0.50% (Li and Lie, 1994) AMP 303-353 - 100 - - 0.002% (Zhang et al., 2002) AMP 293-353 - 40-99 - - 0.002% (Zhang et al., 2002) AMP 298-353 - 4-100 - - 8×10-5 (Chan et al., 2002) AMP 298-343 - 21-100 - 0.05 5×10-5 (Henni et al., 2003) AMP 313-333 0.002 100 - - 6×10-4 (Aguila-Hernandez et al.,
2001) AMP 293-353 - 100 - - - (Kundu et al., 2003) AMP 298-323 - 100 - 0.2 5×10-5 (Alvarez et al., 2006) AMP 298 - 15-30 - 0.01% - (Arcis et al., 2007)
AMP + DEA 303-353 0.05 5-24 5-24 - 0.05% (Hsu and Li, 1997a) AMP + DEA 313-333 0.002 5-95 5-95 - 6×10-4 (Aguila-Hernandez et al.,
2001) AMP + DEA 293-323 0.2 21-28.5 1.5-9 - 0.04% (Mandal et al., 2003b) AMP + DEA 313 0.2 25.5-30 1.5-4.5 - 0.04% (Mandal et al., 2003a) AMP + DEA 303-313 0.05 9-13 1-4 0.2 0.05% (Wang and Li, 2004) AMP + DEA 293-313 - 1.7-25 2-28 - - (Mandal and Bandyopadhyay,
2005) AMP + EMEA 298-323 - 10-50 10-40 0.2 5×10-5 (Alvarez et al., 2006) AMP + MDEA 283-353 0.05 10-50 10-50 0.05% 0.004 (Welsh and Davis, 1995) AMP + MDEA 283-333 0.05 25 5-20 0.05 0.001 (Davis and Pogainis, 1995) AMP + MDEA 313-333 0.002 5-50 5-50 - 6×10-4 (Aguila-Hernandez et al.,
2001) AMP + MEA 303-353 0.05 5-30 5-24 - 0.50% (Li and Lie, 1994) AMP + MEA 293-323 0.2 21-30 1.5-9 - 0.04% (Mandal et al., 2003b) AMP + MEA 302-353 0.05 10 10 - 0.05% (Hsu and Li, 1997a) AMP + MEA 303-313 0.05 13-15 1-4 0.2 0.05% (Xiao et al., 2000) AMP + MEA 313 0.2 25.5-30 1.5-4.5 - 0.04% (Mandal and Bandyopadhyay,
2006) AMP + MMEA 298-323 - 10-50 10-40 0.20 5×10-5 (Alvarez et al., 2006) AMP + MMEA 298-323 0.04 18-27 3-12 0.007% 7.7×10-4 (Venkat et al., 2010a) AMP + NMP 313-333 0.002 5 - 60 5-60 - 6×10-4 (Aguila-Hernandez et al.,
2001)
328
AMP + Pz 303-313 0.05 9-13 1-3.5 0.2 0.05% (Sun et al., 2005) AMP + Pz 288-333 0.1 18-27 3-12 - 4.8×10-4 (Paul and Mandal, 2006c) AMP + Pz 298-333 0.1 22-30 2-8 - 4.5×10-5 (Samanta and
Bandyopadhyay, 2006) AMP + MDEA
+ DEA 303-343 0.005 2-10 3 0.002 0.01% (Rebolledo-Libreros and Trejo,
2006) 1DEA or EMEA or MDEA or MEA or MMEA or NMP or Pz 2Concentration uncertainty of all amines in solutions 332.5 (MDEA) + 12.5 (DEA)
Table A.2. Density data of various SHA systems System T
(K) ∆T (K)
[SHA] (wt%)
[Amine1] (wt%)
∆[AM2] (wt%)
∆ρ (g·cm-3) Reference
2-PE 313 - 1-13 - - - (Shen et al., 1991) 2-PE 298-358 0.05 10-100 - - 1×10-5 (Xu et al., 1992b) 2-PE 313-333 0.002 30-100 - - 6×10-4 (Aguila-Hernandez et al., 2001) 2-PE 288-333 0.2 5-30 - - 0.06% (Paul and Mandal, 2006a)
2-PE + DEA 313 0.002 5-50 5-50 - 6×10-4 (Aguila-Hernandez et al., 2001) 2-PE + DEA 288-333 0.2 3-27 3-27 - 0.06% (Paul and Mandal, 2006a)
2-PE + MDEA 313-333 0.002 5-60 5-60 - 6×10-4 (Aguila-Hernandez et al., 2001) 2-PE + MEA 303-353 0.05 5-24 5-24 - 0.05% (Hsu and Li, 1997a) 2-PE + MEA 288-333 0.2 3-27 3-27 - 0.06% (Paul and Mandal, 2006a) 2-PE + Pz 288-333 0.1 18-27 3-12 0.007% 3.7×10-4 (Paul and Mandal, 2006b)
2-PE + TMS 293-358 0.05 10-65 2-44 - 1×10-4 (Xu et al., 1993b)
AEPD 303-318 - 5-25 - - - (Yoon et al., 2002a) AEPD 303-343 0.05 20-100 - - 2×10-4 (Yoon et al., 2002b)
AHPD 303-343 0.05 5-25 - - 3×10-4 (Park et al., 2002a) AHPD 283-313 0.1 0.2-10 - 0.02% 3×10-4 (Le Tourneux et al., 2008) AHPD 298-323 0.3 2.2-21.7 - - 3.5×10-5 (Paul et al., 2009c)
AHPD + Pz 303-323 0.1 11.8 1-3.5 - 3×10-4 (Bougie et al., 2009)
AMPD 303-343 0.01 10-30 - - 4×10-5 (Baek et al., 2000) 1DEA or MDEA or MEA or Pz or TMS 2Concentration uncertainty of all amines in solutions
329
Table A.3.Viscosity data of AMP systems System T
(K) ∆T (K)
[AMP] (wt%)
[Amine1] (wt%)
∆[AM2] (wt%)
∆µ (mPa·s)
Reference
AMP 313 - 2-27 - - - (Yih and Shen, 1988)
AMP 298 - 2-22 - - - (Bosch et al., 1990)
AMP 296-350 0.05 18-27 - - 0.001 (Xu et al., 1991)
AMP 303 - 1-35.5 - - 1×10-3 (Littel et al., 1992)
AMP 294-318 - 4.5-18 - - - (Saha et al., 1993)
AMP 303-353 0.05 100 - - 1.0% (Li and Lie, 1994)
AMP 298-343 0.01 21-100 - 0.05 0.50% (Henni et al., 2003)
AMP 298-323 0.05 100 - 0.20 5×10-4 * (Alvarez et al., 2006)
AMP + DEA 303-353 0.05 5-24 5-24 0.2 1.0% (Hsu and Li, 1997b)
AMP + DEA 293-323 0.2 21-28.5 1.5-9 - 0.03% (Mandal et al., 2003b)
AMP + DEA 293-323 0.05 2-14 2-17 0.02% 0.2% (Chenlo et al., 2001)
AMP + DEA 313 0.2 25-30 1.5-4.5 - 0.03% (Mandal et al., 2003a)
AMP + DEA 303-313 0.05 9-13 1-4 0.2 1.0% (Wang and Li, 2004)
AMP + DEA 293-313 - 1.7-25 2-29 - - (Mandal and Bandyopadhyay, 2005)
AMP + EMEA 298-323 0.05 10-50 10-40 0.2 5×10-4 * (Alvarez et al., 2006)
AMP + MDEA 283-333 0.05 5-50 5-50 0.05% 0.4% ** (Welsh and Davis, 1995)
AMP + MEA 303-353 0.05 5-30 5-24 - 1.0% (Li and Lie, 1994)
AMP + MEA 293-323 0.05 2-15 1-10 0.02% 0.2% (Chenlo et al., 2001)
AMP + MEA 293-323 0.2 21-30 1.5-9 - 0.03% (Mandal et al., 2003b)
AMP + MEA 303-353 0.05 10 10 0.2 1.0% (Hsu and Li, 1997b)
AMP + MEA 303-313 0.05 13-15 0.5-2.5 0.2 1.0% (Xiao et al., 2000)
AMP + MEA 313 0.2 25.5-30 1.5-4.5 - 0.03% (Mandal and Bandyopadhyay, 2006)
AMP + MMEA 298-323 0.05 10-50 10-40 0.2 5×10-4 * (Alvarez et al., 2006)
AMP + Pz 303-313 9-13 1-3.5 0.2 1.0% (Sun et al., 2005)
AMP + Pz 288-333 0.1 18-27 3-12 - 0.005 (Paul and Mandal, 2006c)
AMP + Pz 298-333 0.1 22-30 2-8 - 1.0% (Samanta and Bandyopadhyay, 2006)
AMP + MDEA + DEA
303-333 0.005 2-10 3 0.002 0.3% (Rebolledo-Libreros and Trejo, 2006)
* value in mm2·s-1; ** kinematic viscosity
1DEA or EMEA or MDEA or MEA or MMEA or Pz 2Concentration uncertainty of all amines in solutions 332.5 (MDEA) + 12.5 (DEA)
330
Table A.4. Viscosity data of various SHA systems System T
(K) ∆T (K)
[SHA] (wt%)
[Amine1] (wt%)
∆[AM2] (wt%)
∆µ (mPa·s)
Reference
2-PE 313 - 1-13 - - - (Shen et al., 1991)
2-PE 298-358 0.05 10-100 - - 0.001 (Xu et al., 1992b)
2-PE 288-333 0.2 5-30 - - 0.69% (Paul and Mandal, 2006a)
2-PE + DEA 288-333 0.2 3-27 3-27 - 0.69% (Paul and Mandal, 2006a)
2-PE + MEA 288-333 0.2 3-27 3-27 - 0.69% (Paul and Mandal, 2006a)
2-PE + MEA 303-353 0.05 5-24 5-24 0.2 1.0% (Hsu and Li, 1997b)
2-PE + Pz 288-333 0.1 18-27 3-12 0.007% 0.005 (Paul and Mandal, 2006b)
2-PE + TMS 293-364 0.05 45-55 10-40 - 0.001 (Xu et al., 1993b)
AEPD 303-318 - 5-25 - - - (Yoon et al., 2002a)
AEPD 303-343 0.05 20-80 - - 1% * (Yoon et al., 2002b)
AHPD 303-343 0.05 5-25 - - 1% * (Park et al., 2002a)
AHPD 283-313 0.1 0.2-10 - 0.02% 1.5% * (Le Tourneux et al., 2008)
AHPD 298-323 0.3 2.2-21.7 - - 1% (Paul et al., 2009c)
AHPD + Pz 303-323 0.1 11.8 1-3.5 - 2% (Bougie et al., 2009)
AMPD 303-343 0.05 10-30 - - 0.5% * (Baek et al., 2000)
* kinematic viscosity
1DEA or MEA or Pz or TMS 2Concentration uncertainty of all amines in solutions
331
Table A.5. Density correlation coefficients of Eq. (1.1) for binary aqueous amine systems
Binary aqueous SHA systems
Parameters 2-PE AEPD AHPD AMPD AMP
a0 1.04689E+00 1.02865E+00 1.05390E+00 1.05608E+00 1.06915E+00
b0 1.49927E-03 3.57266E-03 2.56901E-03 1.96949E-03 6.44832E-04
c0 8.34750E-06 -1.91885E-05 - - -
d0 -1.82209E-07 - - - -8.30901E-08
a1 -5.54067E-07 -4.85143E-07 -6.40281E-07 -6.65693E-07 -7.81603E-07
b1 -1.32719E-08 -1.75701E-08 - -5.50806E-09 -8.45613E-09
c1 - 1.08868E-10 - - 2.75376E-11
d1 7.24406E-13 - - - -
R2 0.9881 0.9992 0.9949 0.9992 0.9987
O.A.D.% 0.08 0.06 0.11 0.03 0.10
Table A.6. Density correlation coefficients of Eq. (1.2) for ternary aqueous amine systems without AMP
Ternary systems without AMP
Parameters 2-PE + DEA 2-PE + MDEA 2-PE + MEA 2-PE + Pz 2-PE + TMS AHPD + Pz
a0 1.41097E+00 1.08935E+00 1.20039E+00 1.10759E+00 1.09639E+00 1.02820E+00
b0 -6.74390E+01 - -2.83464E+01 - - 9.79061E+00
c0 - 6.62278E-04 1.02170E-03 -3.62000E-04 3.16279E-04 -
d0 4.79490E-04 1.18023E-03 8.94935E-04 - 2.04969E-03 4.03960E-04
e0 - -9.71944E-06 -2.48945E-05 - - -
a1 -1.98097E-06 -1.03664E-06 -1.27371E-06 -9.52907E-07 -1.07861E-06 -
b1 - - - - - -
c1 2.03541E-11 -6.10764E-11 -1.73320E-10 - - -2.71715E-09
d1 1.04730E-10 -5.54610E-11 -8.69695E-11 - - -
e1 2.72169E-13 -2.53118E-14 - - - -
R2 0.9889 0.9990 0.9963 0.9992 0.9807 0.9999
O.A.D.% 0.08 0.03 0.05 0.02 0.3 0.002
332
Table A.7. Density correlation coefficients of Eq (1.2) for ternary aqueous amine systems involving AMP Ternary systems involving AMP
Parameters AMP + DEA AMP + EMEA AMP + MDEA AMP + MEA AMP + MMEA AMP + NMP AMP + Pz
a0 1.22303E+00 1.09425E+00 1.08326E+00 1.20585E+00 1.04547E+00 7.40135E-01 1.34408E+00
b0 -3.27527E+01 - - -3.07163E+01 - 7.61748E+01 -5.49795E+01
c0 3.80116E-04 1.49372E-04 3.20405E-04 6.93539E-04 1.22377E-03 4.79469E-04 -
d0 1.49490E-03 - 1.19843E-03 9.97040E-04 1.69268E-03 1.11448E-03 6.25989E-04
e0 -1.42683E-05 - -1.15025E-05 -1.52999E-05 -4.14561E-05 -1.50504E-05 -
a1 -1.34911E-06 -1.22740E-06 -9.95555E-07 -1.27851E-06 -8.02988E-07 - -1.82348E-06
b1 - - - - - - -
c1 -1.00459E-10 - -8.67675E-11 -1.64584E-10 -1.90872E-10 -1.04511E-10 -1.45128E-11
d1 -4.60396E-11 - -5.11136E-11 -9.70532E-11 -2.62090E-10 -6.00757E-11 -2.33775E-10
e1 -1.12208E-14 - -4.36394E-14 -3.58123E-13 -6.30596E-14 -3.23054E-14 7.83703E-13
R2 0.9965 0.9994 0.9986 0.9956 0.9945 0.9985 0.9969
O.A.D.% 0.08 0.01 0.05 0.04 0.04 0.04 0.03
333
Table A.8. Viscosity correlation coefficients of Eq. (1.3) for pure and binary aqueous amine systems
Binary systems Parameters AHPD AMPD
a0 2.06480E+01 1.93980E+01
b0 - -
c0 3.96451E-02 2.62452E-02
d0 9.88914E-04 1.31608E-03
a1 1.55017E-04 1.34932E-04
b1 -1.15826E-01 -1.05473E-01
c1 -1.78664E-07 -
d1 -6.75681E-09 -9.85184E-09
R2 0.9995 0.9999
O.A.D.% 0.6 0.4
Table A.9. Viscosity correlation coefficients of Eq (1.4) for ternary aqueous amine systems without AMP
Ternary systems without AMP Parameters 2-PE + DEA 2-PE + MEA 2-PE + Pz 2-PE + TMS AHPD + Pz
a0 2.80407E+02 -1.67285E+01 -5.10680E+00 -1.34308E+02 -6.67761E+00
b0 -2.53385E+04 4.10418E+03 2.53784E+03 1.90170E+04 2.05722E+03
c0 2.17886E-02 7.63385E-02 -7.25504E-02 - -
d0 - 5.92013E-02 - - 4.31625E-02
e0 5.78432E-04 -6.10503E-04 - 1.25396E-04 -
a1 1.13136E-03 2.80800E-05 - -2.95581E-04 -
b1 -9.90823E-01 - - 3.34791E-01 -
c1 - -7.77203E-09 - - -
d1 6.04752E-09 -5.31004E-09 - - -
e1 - -2.88500E-11 -1.14183E-10 - -
R2 0.9956 0.9988 0.9970 0.9999 0.9995
O.A.D.% 1.9 1.4 1.9 0.9 0.3
334
Table A.10. Viscosity correlation coefficients of Eq. (1.4) for ternary aqueous amine systems involving AMP
Ternary systems involving AMP
Parameters AMP + DEA AMP + EMEA AMP + MEA AMP + MMEA* AMP + Pz
a0 2.32656E+02 -3.54098E+01 -2.38985E+01 -3.00208E+01 -1.05998E+01
b0 -2.11896E+04 8.96897E+03 5.65724E+03 7.78275E+03 3.11621E+03
c0 4.54955E-02 1.13858E-02 8.71964E-02 1.73432E-02 6.27013E-02
d0 1.52046E-02 - 6.82753E-02 - 3.28742E-02
e0 8.90657E-04 1.65873E-04 -1.23642E-03 - -
a1 8.91034E-04 8.03854E-05 4.92136E-05 6.48099E-05 -
b1 -8.07688E-01 - - - -
c1 - - -9.22850E-09 -1.43395E-09 -5.81697E-09
d1 8.51917E-09 1.75843E-09 -5.89416E-09 1.25089E-09 1.75049E-08
e1 - - -1.88802E-11 - -
R2 0.9958 0.9998 0.9960 0.9998 0.9954
O.A.D.% 2.6 0.4 2.4 0.4 2.6
*Correlation of the kinematic viscosity
Table A.11. Surface tension of various SHA System T
(K) ∆T (K)
[SHA] (wt%)
[Amine1] (wt%)
∆[AM2] (wt%)
∆σ (mN·m-1)
Reference
2-PE + Pz 293-323 0.1 18-27 3-12 0.007% 0.12 (Paul and Mandal, 2006b)
AEPD 303-343 0.1 20-80 - - 0.8% (Yoon et al., 2002b)
AMP 298-323 0.05 5-100 - 0.3% 0.02 (Vazquez et al., 1997) AMP 303 - 9 - - - (Rongwong et al., 2009)
AMP + AP 298-323 0.01 10-50 10-50 - 0.02 (Alvarez et al., 2003) AMP + DEA 303 - 2 2 - - (Rongwong et al., 2009)
AMP + MDEA 298-323 0.05 10-50 10-50 0.3% 0.02 (Alvarez et al., 1998) AMP + MEA 303 - 2 2 - - (Rongwong et al., 2009) AMP + MEA 298-323 0.05 10-50 10-50 0.3% 0.02 (Vazquez et al., 1997) AMP + MIPA 298-323 0.01 10-50 10-50 - 0.02 (Alvarez et al., 2003)
AMP + MMEA 298-323 0.2 18-27 3-12 0.007% 0.35 (Venkat et al., 2010a) AMP + Pz 293-323 0.1 18-27 3-12 0.007% 0.12 (Paul and Mandal, 2006b)
AMP + MDEA + DEA 303-343 0.005 2-10 3 0.002 0.21 (Aguila-Hernandez et al., 2007) 1Pz or AP or DEA or MDEA or MEA or MIPA or MMEA or Pz 2Concentration uncertainty of all amines in solutions 332.5 (MDEA) + 12.5 (DEA)
335
Table A.12. Heat capacity of various SHA solutions System T
(K) ∆T (K)
[SHA] (mole frac.)
[Amine1] (mole frac.)
∆[AM2] (mole frac.)
∆Cp (J·mol·K-1)
Reference
2-PE 303-353 0.1 0.2-0.8 - - 3% (Chiu and Li, 1999) 2-PE 303-353 0.1 1.0 - - 3% (Chiu et al., 1999)
2-PE + MEA 303-353 0.1 0.04-0.8 0.04-0.8 - 2% (Shih et al., 2002)
AMP 303-353 0.1 0.2-0.8 - - 3% (Chiu and Li, 1999) AMP 303-353 0.1 1.0 - - 3% (Chiu et al., 1999) AMP 303-368 - 1.0 - - 2% (Zhang et al., 2002) AMP 278-368 - 0.06-0.90 - - 2% (Zhang et al., 2002) AMP 303-353 0.1 1.0 - - 2% (Chen and Li, 2001) AMP 303-353 0.1 0.2-0.8 - - 2% (Chen and Li, 2001) AMP 323-398 0.08 1.0 - - 0.9% (Maham et al., 1997)
AMP + DEA 303-353 0.1 0.04-0.9 0.04-0.9 - 2% (Shih and Li, 2002) AMP + MEA 303-353 0.1 0.04-0.8 0.04-0.8 - 2% (Chen and Li, 2001) AMP + TMS 303-353 0.1 0.04-0.8 0.04-0.8 1.5×10-4 1% (Ho et al., 2007)
1MEA or DEA or TMS 2Concentration uncertainty of all amines in solutions
Table A.13. Heat capacity correlation coefficients of Eq. (1.6) for pure and binary AMP aqueous solutions
Parameters Pure and binary AMP systems a0 3.99560E+01
b0 3.80351E+00
c0 -8.38391E-02
d0 5.79857E-04
a1 -1.71029E-04
b1 6.52042E-06
c1 -
d1 3.44363E-10
R2 0.9983
O.A.D.% 1.1
336
Table A.14. N2O diffusion coefficient in various SHA solutions
System T
(K) ∆T (K)
[SHA] (wt%)
[Amine1] (wt%)
∆[AM2] (wt%)
∆DN2O (m2·s-1)
Reference
2-PE * 313 - 1-13 - - - (Shen et al., 1991) 2-PE 293-313 - 5-40 - - - (Xu et al., 1993a)
AEPD * 303-318 0.1 5-25 - - - (Yoon et al., 2002a)
AHPD * 303-323 0.1 6-27 - - 2% (Bougie and Iliuta, 2009) AHPD 298-323 0.2 2.17-21.7 - - 2% (Paul et al., 2009c)
AMP * 313 - 2.3-27 - - - (Yih and Shen, 1988) AMP 294-348.5 0.1 18-27 - - 5% (Xu et al., 1991) AMP 294-318 0.1 4.5-18 - - 5% (Saha et al., 1993)
AMP * 293-313 - 3.6-18 - - - (Messaoudi and Sada, 1996)
AMP 303-313 - 4.5-22.4 - - 2% (Ko et al., 2001) AMP 298 - 1.8-21.5 - - - (Bosch et al., 1990)
AMP + DEA 303-313 - 6-24 6-30 - 2% (Li and Lee, 1996) AMP + DEA 303-313 - 9-13.4 1-4 0.2 2% (Wang and Li, 2004) AMP + DEA 293-313 0.2 21-30 1.5-9 - 4% (Mandal et al., 2004) AMP + MEA 303-313 - 6-30 6-30 - 2% (Li and Lai, 1995) AMP + MEA 303-313 - 13.4-15.2 1-4 0.2 2% (Xiao et al., 2000) AMP + MEA 293-313 0.2 21-30 1.5-9 - 2% (Mandal et al., 2005) AMP + Pz 303-313 - 9-13.5 1-3.5 0.2 2% (Sun et al., 2005) AMP + Pz 298-313 0.1 22-30 2-8 - 4% (Samanta and
Bandyopadhyay, 2009)
AMPD 303-323 0.1 2.5-30 - - - (Yoon et al., 2003)
*Authors reported the ratio 22 CO
2/1CO / HD by using the N2O analogy
1DEA or MEA or Pz 2Concentration uncertainty of all amines in solutions
337
Table A.15. N2O diffusion correlation coefficients of Eq. (1.15) in AMP solutions Parameters Pure and binary AMP solutions
a0 -8.86770E+01
b0 1.68285E+04
c0 2.02569E-01
d0 -6.07155E-03
a1 3.83315E-04
b1 -
c1 -2.87647E-06
d1 7.18194E-08
R2 0.9933
O.A.D.% 3.7
Table A.16. CO2 solubility in single SHA aqueous solutions System T
(K) ∆T (K)
PCO2 (kPa)
∆PCO2 (kPa)
[SHA] (wt%)
∆α (mol·mol-1)
Reference
AMP* 313, 393 - 0.55-2068 - 26.8 - (Sartori and Savage, 1983)
AMP 313 0.5 1.25-216 0.1% 26.8 3% (Roberts and Mather, 1988a)
AMP 313 0.5 2.17-5740 0.1% 18 3% (Roberts and Mather, 1988a)
AMP 373 0.5 8.53-5870 0.1% 18 3% (Roberts and Mather, 1988a)
AMP 323 0.5 4.32-5645 - 30.7 3% (Teng and Mather, 1989)
AMP 293,313,333,353 0.5 1.59-98.93 - 18,26.8 - (Tontiwachwuthikul et al., 1991)
AMP* 303,313,323 - 0.1-100 - 18 12% (Aroua et al., 2002)
AMP 313 0.1 43.7-159 1.4 18 5% (Jane and Li, 1997)
AMP* 313, 333, 353 0.1 0.69-344 0.25% 30 - (Park et al., 2002c)
AMP** 288.5, 293, 298, 303
0.2 n.a. - 4.5-18 2% (Saha et al., 1993)
AMP** 293, 298, 303, 308, 313
0.1 n.a. 0.2 18-26.8 2% (Mandal et al., 2005)
AMP** 293, 298, 303, 308, 313
0.1 n.a. 0.2 18-26.8 2% (Mandal et al., 2004)
AMP 313, 333, 353 0.1 3.94-336.6 0.1% 30 3% (Seo and Hong, 1996)
AMP 303, 313, 323, 333
- 0.5-100 - 18 12% (Haji-Sulaiman and Aroua, 1996)
338
AMP 313, 333, 353, 373
0.1 1.05-197 1.4 30 3% (Li and Chang, 1994)
AMP 303 0.1 4.41-90.1 0.2 18 2% (Kundu et al., 2003)
AMP 303, 313, 323 0.1 3.20-94 0.2 25, 30.4 2% (Kundu et al., 2003)
AMP 313, 333, 353 0.04 7.3-2743 0.2% 17.6, 35.6 3% (Silkenbaumer et al., 1998)
AMP 313 0.5 0.162-283.7 - 18 3% (Teng and Mather, 1990)
AMP 343 0.5 0.586-5279 - 18 3% (Teng and Mather, 1990)
AMP 313 0.01 0.89-151.9 5/10 26.8 3% (Yang et al., 2010)
AMPD 313 - 0.34-881 0.5% 10.4 - (Puxty et al., 2009a)
AMPD 313 0.1 1.04-2991 0.1% 10 3% (Baek and Yoon, 1998)
AMPD 303, 313, 333 0.1 0.6-3064 0.1% 30 3% (Baek and Yoon, 1998)
AMPD** 303, 313, 325 0.1 n.a. - 10, 20, 30 3% (Baek et al., 2000)
AEPD 313, 323, 333 0.1 1.8-1927.4 0.1% 10 3% (Park et al., 2002b)
AEPD 333 0.1 7.7-2849 0.1% 30 3% (Park et al., 2002b)
AHPD 313, 323, 333 0.1 21.7-1839.8 0.1% 10 - (Park et al.,
2002a) AHPD 313 0.1 42.1-1451.5 0.1% 20 - (Park et al.,
2002a) AHPD 298 0.1 0.9-2427.3 0.1% 10 - (Park et al., 2003) AHPD 283, 298, 313 0.01 1.91-74.8 0.25% 0.15-10 1% (Le Tourneux et
al., 2008) AHPD** 283, 298, 313 0.01 n.a. 0.25% 0.15-10 1% (Le Tourneux et
al., 2008) AHPD 284, 293, 298,
303, 323, 333 0.1 0.31-2637.6 - 10-32.6 - (Bougie and Iliuta,
2010b) AHPD** 298, 303, 313,
323 0.3 n.a. 0.2 2.17-21.7 1.5% (Paul et al.,
2009c) *d.n.t: data not tabulated; ** p.s.: physical solubility; solubility uncertainties are on Henry’s constant
339
Table A.17. CO2 solubility in SHA based mixed solvents System T
(K) ∆T (K)
PCO2 (kPa)
PCO2 (kPa)
Concentration ∆α (mol·mol-1)
Reference
2-PE + TMS 313, 373 0.1 0.274-5548 0.1% 55 wt% 2-PE + 10 wt% sulfolane - (Lal et al., 1998) 2-PE + TMS 298, 313, 343,
373, 403 0.1 0.00156-
18900 0.1% 45 wt% 2-PE + 40 wt% sulfolane 4% (Jou et al., 1998)
AHPD + Pz* 288, 298, 313, 333
0.1 n.a. - (1.1-4.2) mol·kg-1 AHPD + (0.1-0.65) mol·kg-1 Pz 2% (Bougie and Iliuta, 2010b)
AHPD + Pz 288, 298, 313, 333
0.1 2.1-2310 - (1.1-4.2) mol·kg-1 AHPD + (0.01-0.66) mol·kg-1 Pz - (Bougie and Iliuta, 2010b)
AMP + DEA 313, 373 0.02 162-2908 3.5 5 wt% AMP + 25 wt% DEA 10% (Murrieta-
Guevara et al., 1998)
AMP + DEA 313, 373 0.02 22-2597 3.5 10 wt% AMP + 20 wt% DEA 10% (Murrieta-Guevara et al.,
1998) AMP + DEA* 303,308,313 0.5 n.a. - (6-24) wt% AMP + (6-24) wt% MEA 2% (Li and Lee,
1996) AMP + DEA* 293, 298, 303,
308, 313 0.1 n.a. 0.2 (21-30) wt% AMP + (0-9) wt% DEA 2% (Mandal et al.,
2004) AMP + DEA 313, 333, 353 0.1 0.69-344 0.25% (0-30) wt% AMP + (0-30) wt% DEA - (Park et al.,
2002c) AMP + DEA 313, 333, 353 0.1 1.61-357.3 0.1% (0-30) wt% AMP + (0-30) wt% DEA 3% (Seo and Hong,
1996) AMP + MDEA 313 0.04 12.5-4020** 0.2%/0.1% 1.266 mol·kg-1 AMP + 1.278 mol·kg-1 MDEA 3% (Silkenbaumer et
al., 1998) AMP + MDEA 303, 313, 323 - 0.1-100 - 2.0 kmol·m-3 total amine content 12% (Aroua et al.,
2002) AMP + MEA* 303,308,313 0.5 n.a. - (0-30) wt% AMP + (0-30) wt% MEA 2% (Li and Lai, 1995)
340
AMP + MEA* 293, 298, 303, 308, 313
0.1 n.a. 0.2 (21-30) wt% AMP + (0-9) wt% MEA 2% (Mandal et al., 2005)
AMP + MEA 313, 333, 353 0.1 0.69-344 0.25% (0-30) wt% AMP + (0-30) wt% MEA - (Park et al., 2002c)
AMP + MEA 313, 333, 353, 373
0.1 1-199 1.4 (0-30) wt% AMP + (0-30) wt% MEA 3% (Li and Chang, 1994)
AMP + Pz 313, 333, 353 0.01 0.97-139.9 5/10 (2.0, 3.0) kmol·m-3 AMP + (0.5, 1.0, 1.5) kmol·m-3 Pz 3% (Yang et al., 2010)
AMP + TMS 313, 373 0.5 2.63-6050 0.1% 16.5 wt% AMP + 32.2 wt% sulfolane 3% (Roberts and Mather, 1988b)
AMP + DEA + MDEA
313, 343, 393 0.02/0.5 10-1929 3.5 4 wt% AMP + 12.5 wt% DEA + 32.5 wt% MDEA - (Rebolledo-Libreros and Trejo, 2004)
AMP + DEA + MDEA
313, 343, 393 0.02/0.5 6.6-1999.1 3.5 6 wt% AMP + 12.5 wt% DEA + 32.5 wt% MDEA - (Rebolledo-Libreros and Trejo, 2004)
AMP + DEA + MDEA
313, 343, 393 0.02/0.5 3.1-1968.7 3.5 10 wt% AMP + 12.5 wt% DEA + 32.5 wt% MDEA - (Rebolledo-Libreros and Trejo, 2004)
*Physical solubility; solubility uncertainties are on Henry’s constant; **Total pressure
341
Table A.18. Estimated Henry’s law constants for CO2 in aqueous single SHA solutions using the N2O Analogy
System T (K)
[SHA] (kmol·m-3)
HCO2 (kPa·m3·kmol-1)
Reference
AMP 288.5 0.5 2463.2 (Saha et al., 1993) 1.0 2592.6 1.5 2696.2 2.0 2823.3 293.0 0.5 2801.7 1.0 2954.8 1.5 3072.4 2.0 3218.7 298.0 0.5 3062.5 1.0 3229.5 1.5 3351.4 2.0 3505.8 303.0 0.5 3466.8 1.0 3652.9 1.5 3779.7 2.0 3944.0
AMP 293 2.0 3157 (Mandal et al., 2005; Mandal et al., 2004)
2.5 3241 3.0 3320 298 2.0 3636 2.5 3721 3.0 3818 303 2.0 3846 2.5 3911 3.0 4004 308 2.0 4405 2.5 4485 3.0 4551 313 2.0 4530 2.5 4619 3.0 4693
AHPD 298 0.2 3170 (Paul et al., 2009c) 0.4 3202 0.9 3277 1.3 3390
342
1.9 3522 303 0.2 3535 0.4 3579 0.9 3670 1.3 3820 1.9 3982 313 0.2 4417 0.4 4492 0.9 4643 1.3 4810 1.9 4896 323 0.2 5527 0.4 5644 0.9 5767 1.3 5894 1.9 6027
AHPD 283 0.01 1934 (Le Tourneux et al., 2008) 0.04 1947 0.08 1952 0.2 1961 0.8 2087 298 0.01 3007 0.04 3017 0.08 3018 0.2 3041 0.8 3176 313 0.01 4242 0.04 4257 0.08 4262 0.2 4274 0.8 4463
343
Table A.19. Estimated Henry’s constants for CO2 in aqueous mixed SHA based solutions, using the N2O Analogy.
System T (K)
wt% (1) + wt% (2) HCO2 (kPa·m3·kmol-1)
Reference
AMP(1) + MEA(2) 303.0 0 + 30 3181.9 (Li and Lai, 1995) 6 + 24 3317.3 12 + 18 3582.3
18 + 12 3949.8 24 + 6 4083.0 30 + 0 4271.5 308.0 0 + 30 3382.2 6 + 24 3601.1 12 + 18 4073.6 18 + 12 4292.8 24 + 6 4539.7 30 + 0 4713.7 313.0 0 + 30 3646.6 6 + 24 3943.8 12 + 18 4360.7 18 + 12 4846.1 24 + 6 5081.7 30 + 0 5356.3
AMP(1) + MEA(2) 293.0 30 + 0 3328 (Mandal et al., 2005) 28.5 + 1.5 3306 27 + 3 3278 25.5 + 4.5 3247 14 + 6 3221 22.5 + 7.5 3182 21 + 9 3159 298.0 30 + 0 3829 28.5 + 1.5 3780 27 + 3 3731 25.5 + 4.5 3697 14 + 6 3667 22.5 + 7.5 3614 21 + 9 3560 303.0 30 + 0 4021 28.5 + 1.5 3970 27 + 3 3912 25.5 + 4.5 3856 24 + 6 3805
344
22.5 + 7.5 3750 21 + 9 3706
308.0 30 + 0 4569 28.5 + 1.5 4495 27 + 3 4434 25.5 + 4.5 4366 14 + 6 4307 22.5 + 7.5 4207 21 + 9 4141 313.0 30 + 0 4720 28.5 + 1.5 4622 27 + 3 4576 25.5 + 4.5 4521 14 + 6 4459 22.5 + 7.5 4405 21 + 9 4349
AMP(1) + DEA(2) 303.0 6 + 24 4799.0 (Li and Lee, 1996) 12 + 18 4590.3 18 + 12 4496.4
24 + 6 4404.0 308.0 6 + 24 6179.3 12 + 18 5740.0 18 + 12 5357.4 24 + 6 5046.0 313.0 6 + 24 8193.9 12 + 18 7236.5 18 + 12 6440.1 24 + 6 5828.1
AMP(1) + DEA(2) 293.0 30 3328 (Mandal et al., 2004) 28.5 3350 27 3381 25.5 3401 24 3405 22.5 3434 21 3447 298.0 30 3829 28.5 3862 27 3863 25.5 3886 24 3890 22.5 3916 21 3929 303.0 30 4021
345
28.5 4018 27 4024 25.5 4026 24 4034 22.5 4055 21 4071 308.0 30 4569 28.5 4576 27 4591 25.5 4604 24 4623 22.5 4632
21 4646 313.0 30 4720 28.5 4890 27 5109 25.5 5223 24 5267 22.5 5296 21 5308
346
Table A.20. Kinetic information of CO2 absorption by various SHA (other than AMP) solutions
System T [SHA] k2 at 298 K k2 k2kAm/k-1 k2kH2O/k-1 Reference
(K) (kmol·m-3) (m3·kmol-1·s-1) (m3·kmol-1·s-1) (m6·kmol-2·s-1) (m6·kmol-2·s-1)
2-PE 313 0.107 - 1.0 - 195 - - (Shen et al.,
1991) 2-PE 283-313 0.25 - 2.5 620 exp(24.439 - 44621/RT) exp(24.619 -41695/RT) exp(20.734 -
44206/RT) (Xu et al.,
1993a) 2-PE 303-323 0.14 - 1.13 495 exp(24.437 - 45171/RT) - - (Paul et al.,
2009a)
AEPD 303-318 0.417 - 2.154 242 exp(31.730 - 7820/T) exp(21.902 - 4809/T) exp(72.316 - 22843/T) (Yoon et al., 2002a)
AHPD 303-323 0.5 - 2.4 192 exp(26.953 - 6465/T) exp(15.999 -3124/T) exp(11.695 -3315/T) (Bougie and
Iliuta, 2009) AHPD 303-323 0.179 - 1.789 329 exp(32.093 - 65155/RT) - - (Paul et al.,
2009b)
AMPD 278-303 0.025 - 1.6 194 exp(19.058 - 4110.2/T) exp(25.157 - 5381.3/T) exp(24.201 - 7043.5/T) (Bouhamra et al., 1999)
AMPD 303-323 0.236 - 2.963 303* exp(21.158 - 4602.6/T) exp(17.190 - 3434.7/T) exp(11.860 - 3476.8/T) (Yoon et al., 2003)
TBAE 283-308 - 170 exp(31.330 - 7806/T) - - (Ali et al.,
2002) *Extrapolated value
347
Table A.21. Kinetic information for CO2 absorption by AMP solutions
System T [AMP] k2 at 298 K k2 k2kAm/k-1 k2kH2O/k-1 k2kAm#2/k-1 Reference
(K) (kmol·m-3) (m3·kmol-1·s-1) (m3·kmol-1·s-1) (m6·kmol-2·s-1) (m6·kmol-2·s-1) (m6·kmol-2·s-1)
AMP 315 - - 100 - - - (Chakraborty et al., 1986)
AMP 313 0.26 - 3.0 - 1270 - - - (Yih and Shen, 1988)
AMP 278-298 0.01 - 1.5 520 exp(23.079 - 5013.7/T) - - - (Alper, 1990)
AMP 298 0.202 - 2.373 10000 10000 127 8.36 - (Bosch et al., 1990)
AMP 294-318 0.5 - 2.0 555 exp(23.690 - 5176.49/T) - - - (Saha et al., 1995)
AMP 288-318 0.17 - 3.5 782 exp(16.454 - 24261/RT) exp(16.005 - 20678/RT) exp(19.311 - 45670/RT) - (Xu et al., 1996)
AMP 293-313 0.5 - 2.0 268 exp(26.500 - 6230.6/T) - - - (Messaoudi and Sada, 1996)
AMP 303 0.55 - 3.35 1105* 1150 1387 0.2611 - (Seo and Hong, 2000)
AMP 313 0.55 - 3.35 - 1241 2057 1.875 - (Seo and Hong, 2000)
AMP 293-313 0.2 - 2.8 570 exp(25.815 - 5801.7/T) - - - (Mandal and Bandyopadhyay, 2005)
AMP 298-313 0.05 - 0.35 578 exp(23.234 - 5028.5/T) exp(18.397 - 3522.1/T) exp(14.401-3413.9/T) - (Ali, 2005)
AMP 313 3.3 - 731 - - - (Choi et al., 2007)
AMP 288-313 0.1 - 3.0 27 exp(29.200 - 8186.9/T) - - - (Camacho et al., 2005)
AMP** 298 0.402 - 3.545 56 56.3 39 - - (Xu et al., 1996)
AMP + DEA 298-313 0.006 - 0.380 556 exp(22.829 - 4919.6/T) exp(13.996 - 2217.2/T) exp(14.424 - 3421/T) exp(23.799 - 4243.1/T) (Ali, 2005)
AMP + DEA 303-313 1.0 - 1.5 611 exp(19.509 - 3902/T) - - - (Wang and Li, 2004)
AMP + MEA 303-313 1.5 - 1.7 1098 10^(6.595 - 1059.2/T) 10^(13.23 - 3036.3/T) 10^(6.952 - 2392.9/T) 10^(19.607 - 5032.9/T) (Xiao et al., 2000)
AMP + MEA 298-313 0.073 - 0.256 559 exp(23.316 - 5063.2/T) exp(12.951 - 1872.1/T) exp(14.768 - 3532.7/T) exp(23.280 - 3547.6/T) (Ali, 2005)
AMP + Pz 303 0.55 - 3.35 1375* 1500 638.7 7.941 14693 (Seo and Hong, 2000)
AMP + Pz 313 0.55 - 3.35 - 1771 750.6 8.32 13767 (Seo and Hong, 2000)
AMP + Pz 303-313 1.0 - 1.5 1185* exp(17.259 -3034/T) exp(22.885 -4241/T) exp(27.708 - 5893/T) exp(13.248 - 45861/T) (Sun et al., 2005)
*Extrapolated values; ** In 1-propanol
348
Table A.22. CO2 solubility in Pz-AHPD aqueous solutions
T mAHPD mPz CO2 loading P yCO2
(K) (mol.kg-1) (mol.kg-1) (mol CO2.mol-1 amines) (kPa) -
288.10 1.0971 0.0110 0.1124 2.134 0.143 288.09 1.0971 0.0110 0.2898 3.233 0.436 288.08 1.0971 0.0110 0.5144 6.836 0.735 288.09 1.0971 0.0110 0.7939 24.65 0.927 288.08 1.0971 0.0110 0.9986 107.6 0.983 288.06 1.0971 0.0110 1.0978 256.5 0.993 288.08 1.0971 0.0110 1.1794 459.1 0.996 288.10 1.0971 0.0110 1.3355 928.5 0.998 288.09 1.0971 0.0110 1.5060 1533.5 0.999 333.13 1.1232 0.0112 0.0811 22.27 0.124 333.14 1.1232 0.0112 0.1939 33.38 0.417 333.16 1.1232 0.0112 0.2995 53.09 0.634 333.15 1.1232 0.0112 0.4256 94.72 0.795 333.15 1.1232 0.0112 0.5327 149.0 0.870 333.15 1.1232 0.0112 0.6468 248.3 0.922 333.14 1.1232 0.0112 0.7749 440.7 0.956 333.15 1.1232 0.0112 0.9932 922.7 0.979 333.14 1.1232 0.0112 1.1841 1442.1 0.987 333.15 1.1232 0.0112 1.4332 2110.2 0.991 288.19 1.1095 0.1109 0.1531 2.050 0.157 288.19 1.1095 0.1109 0.4282 4.186 0.590 288.19 1.1095 0.1109 0.7355 16.53 0.897 288.20 1.1095 0.1109 0.9614 83.98 0.980 288.19 1.1095 0.1109 1.0438 182.34 0.991 288.18 1.1095 0.1109 1.1292 366.44 0.995 288.18 1.1095 0.1109 1.2833 757.87 0.998 288.20 1.1095 0.1109 1.6100 1790.3 0.999 333.17 1.1297 0.1130 0.0857 20.60 0.057 333.19 1.1297 0.1130 0.1952 27.63 0.299 333.18 1.1297 0.1130 0.3178 46.31 0.583 333.19 1.1297 0.1130 0.4491 85.70 0.775 333.18 1.1297 0.1130 0.5891 158.7 0.879 333.21 1.1297 0.1130 0.7182 289.2 0.934 333.20 1.1297 0.1130 0.8322 491.5 0.961 333.20 1.1297 0.1130 1.0092 984.2 0.981 333.18 1.1297 0.1130 1.1379 1468.0 0.987 333.17 1.1297 0.1130 1.3798 2310.5 0.992 298.21 1.1345 0.3403 0.0954 3.350 0.071
349
298.21 1.1345 0.3403 0.2244 3.838 0.192 298.26 1.1345 0.3403 0.3749 5.427 0.431 298.14 1.1345 0.3403 0.5175 9.157 0.664 298.17 1.1345 0.3403 0.7754 34.60 0.912 298.15 1.1345 0.3403 0.9837 170.9 0.982 298.15 1.1345 0.3403 1.0869 436.3 0.993 298.16 1.1345 0.3403 1.2093 936.1 0.997 298.16 1.1345 0.3403 1.3060 1422.1 0.998 298.16 1.1345 0.3403 1.4718 2253.6 0.999 313.16 1.1633 0.5816 0.0932 7.578 0.042 313.15 1.1633 0.5816 0.1823 7.977 0.093 313.15 1.1633 0.5816 0.2932 9.192 0.215 313.15 1.1633 0.5816 0.4101 12.89 0.442 313.15 1.1633 0.5816 0.5735 27.81 0.743 313.16 1.1633 0.5816 0.7806 101.8 0.930 313.16 1.1633 0.5816 0.9424 358.6 0.980 313.16 1.1633 0.5816 1.0627 883.9 0.992 313.16 1.1633 0.5816 1.1588 1385.8 0.995 313.16 1.1633 0.5816 1.3069 2195.9 0.997 313.16 4.2294 0.1410 0.0670 7.394 0.074 313.16 4.2294 0.1410 0.1279 8.475 0.196 313.16 4.2294 0.1410 0.2054 10.89 0.378 313.16 4.2294 0.1410 0.3001 15.38 0.563 313.16 4.2294 0.1410 0.4399 28.00 0.763 313.15 4.2294 0.1410 0.6271 66.37 0.901 313.15 4.2294 0.1410 0.8617 216.0 0.970 313.15 4.2294 0.1410 1.0396 639.6 0.990 298.18 3.3604 0.4032 0.0518 3.300 0.080 298.22 3.3604 0.4032 0.1380 3.932 0.233 298.20 3.3604 0.4032 0.2198 4.909 0.389 298.23 3.3604 0.4032 0.3006 6.642 0.551 298.13 3.3604 0.4032 0.4577 13.79 0.786 298.20 3.3604 0.4032 0.6592 45.46 0.936 298.16 3.3604 0.4032 0.8742 229.6 0.988 298.18 3.3604 0.4032 0.9984 788.2 0.996 288.17 2.5792 0.6448 0.0637 1.876 0.113 288.17 2.5792 0.6448 0.1622 2.257 0.266 288.17 2.5792 0.6448 0.2569 2.797 0.411 288.19 2.5792 0.6448 0.4415 4.978 0.673 288.18 2.5792 0.6448 0.6146 11.85 0.864 288.18 2.5792 0.6448 0.8322 57.36 0.972 288.17 2.5792 0.6448 0.9919 317.9 0.995 288.17 2.5792 0.6448 1.0821 893.1 0.998
350
288.17 2.5792 0.6448 1.1367 1436.9 0.999 333.15 2.6430 0.6607 0.1025 19.52 0.050 333.15 2.6430 0.6607 0.2015 23.80 0.225 333.15 2.6430 0.6607 0.3121 38.73 0.527 333.13 2.6430 0.6607 0.4231 73.10 0.751 333.13 2.6430 0.6607 0.5152 122.8 0.853 333.13 2.6430 0.6607 0.5975 204.2 0.912 333.13 2.6430 0.6607 0.6917 366.8 0.951 333.13 2.6430 0.6607 0.8119 767.0 0.977
Top Related