Sterically hindered amine based absorbents and application for … · 2020. 8. 7. · 1.2.4.3. CO 2 physical solubility in single and mixed solvents.....47 1.2.5. Absorption kinetics

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


5.3. 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


6.3. 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 2.5. 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 5.8. 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

http://fr.wikipedia.org/wiki/Nomenclature

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).


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

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.


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


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



CAHPD 2COP



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)


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


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)


3.4.3.1 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.


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)


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)


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).


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.


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


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


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


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


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


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


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


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


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 triisopropanolamine) 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

==


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

==


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.


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

Documents

Sterically hindered amine based absorbents and application for … · 2020. 8. 7. · 1.2.4.3. CO 2 physical solubility in single and mixed solvents.....47 1.2.5. Absorption kinetics