HC E155q

CO2 capture using monoethanolamine solutions:
Development and validation of a process model based

on the SAFT-VR equation of state
Charles Victor Brand
A thesis submitted for the degree of

Doctor of Philosophy and the Diploma of
Imperial College London
Centre for Process Systems Engineering

Department of Chemical Engineering
Imperial College London
United Kingdom
May 2013
Declaration
I hereby declare that all the material in this dissertation is own and has been otherwise
appropriately referenced.
Charles Brand
May 2013
The copyright of this thesis rests with the author and is made available under a Creative
Commons Attribution Non-Commercial No Derivatives licence. Researchers are free to
copy, distribute or transmit the thesis on the condition that they attribute it, that they
do not use it for commercial purposes and that they do not alter, transform or build upon
it. For any reuse or redistribution, researchers must make clear to others the licence terms
of this work.
3
Abstract
The development of a predictive model for an absorber-desorber process for the separa-
tion of carbon dioxide (CO2 ) from a gas stream using an aqueous alkanolamine solution
as a solvent is presented. Post-combustion carbon dioxide capture by absorption with
aqueous amine solvents is likely to play an important role in climate change mitigation,
by helping to reduce a significant fraction of CO2 emissions from fossil fuel power plants.
There are, however, a number of concerns with the large scale deployment of this tech-
nology, including energy requirements, solvent degradation and the environmental and
health impact resulting from a potential loss of solvent and solvent degradation products.
Modelling studies can play an invaluable and complementary role in addressing some of
these issues, including the choice of solvent and operating conditions that yield optimal
performance. The model presented here incorporates state-of-the-art SAFT-VR thermo-
dynamics into a rate-based process model. A characteristic of the proposed approach is
that all the reactions are treated within a thermodynamic description, assuming chemical
equilibrium throughout. This greatly reduces the amount of experimental data required
to model the behaviour of the absorber. Furthermore, in contrast with many treatments
of reactive systems of this type, no enhancement factor is used in the process model. The
absorber-desorber process model is implemented in the gPROMS software platform and
validated using published pilot plant experimental data for the removal of CO2 from an air
and CO2 stream using monoethanolamine (MEA) solutions. A scaling of the diffusivity
in the liquid phase, that is found to be transferable to different operating conditions, is
proposed. Reliable predictions are obtained for the temperature and composition profiles
in the gas and liquid phases, including a good description of the temperature bulge which
sometimes appears along the height of the absorber column. The same transferable model
is used to describe both the absorber and the desorber columns. The influence of key pa-
rameters of the model for different operating conditions is assessed through a sensitivity
analysis. The model developed in this study is applied to simulate a complete amine-based
carbon capture absorber-desorber process. Given the relatively simple modelling of the
solvent/CO2 interactions, in which the reactions are treated implicitly through a physical
approach, the proposed model lends itself well to the investigation of other solvents.
5
Acknowledgements
The work presented in this thesis would not have been possible hadn’t it been for the
contribution of a number of people. First and foremost, I would like to express my sincer
gratitude to my supervisors Professors Claire Adjiman, Amparo Galindo and George
Jackson for their academic support and personal involvment during this PhD. Thank you
for sharing with me your knowledge in process modelling and thermodynamic.
Thanks to the members of the MSE group for the fun and the help during these years, in
particular Niall, Frances, Tom, Carlos, Javier, Alex, Vassilis, Jens, Andrew, Nina, Roochi,
Zara, Lara, Olga, Hendrick.
Thanks to my Father, Mother, Brother and Sisters for the regular support and the wise
guidance.
I am grateful to Nicolas, Quentin, Pierre, Philippe, Frank, Benjamin, and Paul for sharing
their advice and experience.
Thanks to you Sophie for your presence, you are of great importance to me.
6
Contents
1 Introduction 24
1.1 Motivation for CO2 capture . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.2 Objectives of this work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.3 Outline of thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2 Techniques for CO2 capture 29

2.1 CO2 capture technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.1.1 Calcium looping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.1.2 Oxy-fuel combustion . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.1.3 Integrated gasification combined cycle . . . . . . . . . . . . . . . . 32
2.1.4 Pressure swing adsorption . . . . . . . . . . . . . . . . . . . . . . . 32
2.1.5 Chemical absorption . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.2 Amine-based processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.2.1 Importance of the thermodynamic and the transport models . . . . 36
2.2.2 Column description . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.2.3 The concept of a stage . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.2.4 Rate-based vs Equilibrium models . . . . . . . . . . . . . . . . . . . 39
2.2.5 Two film theory and enhancement factor . . . . . . . . . . . . . . . 40
2.2.6 Mass transfer correlations . . . . . . . . . . . . . . . . . . . . . . . 42
7
2.2.7 Phase equilibrium and chemical reactions . . . . . . . . . . . . . . . 42
2.2.8 Integration of SAFT into process models . . . . . . . . . . . . . . . 44
2.2.9 Summary of existing process models . . . . . . . . . . . . . . . . . 45
2.2.10 Overview of available pilot plant . . . . . . . . . . . . . . . . . . . . 45
2.2.11 Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3 Thermodynamic theory and molecular model 49

3.1 Thermodynamic theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.1.1 Chemical approach . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.1.2 Quasi chemical approach . . . . . . . . . . . . . . . . . . . . . . . . 51
3.1.3 Molecular perturbation approach, SAFT . . . . . . . . . . . . . . . 52
3.1.4 SAFT-VR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.2 Molecular model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.3 Validation of the thermodynamic model . . . . . . . . . . . . . . . . . . . 62
3.4 Enthalpy of absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.4.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.4.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.5 Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4 Development of an absorber model using SAFT 69

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.2 Non-equilibrium absorber model . . . . . . . . . . . . . . . . . . . . . . . . 70
4.3 Model equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.4 Conluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
5 Absorber model results 87

5.1 Validation of the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.2 Model predictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.3 Sensitivity analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.3.1 Mass transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
8
5.3.2 Heat transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.3.3 Extent of water in the flue gas . . . . . . . . . . . . . . . . . . . . . 118
5.4 Conluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6 Model of the regeneration module 122

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
6.2 Model of the desorber column . . . . . . . . . . . . . . . . . . . . . . . . . 125
6.3 Reboiler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
6.4 Condenser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
7 Regeneration module results 133

7.1 Model predictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
7.2 Sensitivity analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
7.2.1 Mass transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
7.2.2 Heat transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
7.3 Energy requirements in the reboiler . . . . . . . . . . . . . . . . . . . . . . 149
7.4 Effect of the pressure in the reboiler . . . . . . . . . . . . . . . . . . . . . . 152
8 Application of the model 156

8.1 Closed flowsheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
8.1.1 Addition of MEA . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
8.2 Application to an other pilot plant . . . . . . . . . . . . . . . . . . . . . . 166
8.2.1 Absorber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
9 Concluding remarks 170

9.1 Summary of thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
9.2 Key contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
9.3 Directions for future work . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
9
9.3.1 Increase the model accuracy . . . . . . . . . . . . . . . . . . . . . . 173
9.3.2 Validate further . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
9.3.3 Process design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
9.3.4 Extending the model to account for impurities . . . . . . . . . . . . 174
9.3.5 Potential use of this model . . . . . . . . . . . . . . . . . . . . . . . 175
9.4 Publications and presentations . . . . . . . . . . . . . . . . . . . . . . . . . 175
9.4.1 Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
9.4.2 Conference presentations and posters . . . . . . . . . . . . . . . . . 176
10
List of Figures
2.1 Flowsheet of an amine-based carbon capture process . . . . . . . . . . . . 35

2.2 Diagram of a non-equilibrium stage. This stage represents a section of
packing in a packed column. . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.3 A schematic of the two-film model of a column stage. δ V and δ L represent
the thickness of the gas and liquid films, respectively. P V , P I , and P L are
the pressure in the bulk vapour phase, at the gas-liquid interface and in
the bulk liquid, respectively. As can be seen an isobaric profile is assumed.
The temperatures of the bulk vapour phase, at the gas-liquid interface, and
of the bulk liquid are denoted by T V , T I , and T L , respectively. Finally, yi
and xi are the mole fractions of component i in the bulk vapour and liquid
phases, respectively, and yiI and xIi are the mole fractions of component i at
the vapour-liquid interface in the vapour and the liquid phases, respectively. 41
3.1 The square-well potentials for a monomer, Φmono , and association between
the sites, Φassoc , employed in the SAFT-VR approach. (a) Φmono is defined
by a hard-core of diameter σ, range of attraction λσ and depth −ε. (b)
Φassoc is defined by an off-centre potential of depth −εab and of range rc;ab .
(c) The centre of the site is at a distance rd from the centre of the segment. 54
3.2 Schematic representation of the association scheme between MEA and CO2
showing two reaction products. . . . . . . . . . . . . . . . . . . . . . . . . 56
11
3.3 A schematic of the molecular model used in the SAFT-VR method. (a)
H2 O, (b) MEA, (c) CO2 and (d) N2 . The sites e represent the electron
lone pairs on the oxygen atom, the sites e∗ correspond to a lone pair on
the nitrogen atom, the sites H correspond to the hydroxyls, the sites H ∗
correspond to the amines hydrogens, and the sites α1 and 1 α2 are acceptor
sites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.4 Solubility of CO2 in a 30wt% solution of MEA. SAFT-VR model predic-
tions (continuous curves) from Rodrı́guez et al. (2012) and experimental
data (squares) from Jou et al. (1995) . . . . . . . . . . . . . . . . . . . . . 63
3.5 Predicted mole fraction of carbamate and bicarbonate in the ternary mix-
ture of MEA + H2 O + CO2 with a concentration of MEA in the liquid phase
of 30 wt%. (a) T = 313 K and (b) T = 353 K. The symbols (squares for
carbamate and diamonds for bicarbonate) correspond to the experimental
data (Böttinger et al., 2008) and the curves (continuous for carbamate and
dashed for bicarbonate) to the SAFT-VR calculations (Rodrı́guez et al.,
2012). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.6 Enthalpy of absorption of CO2 in a 30 wt% solution of MEA: SAFT-VR
model predictions with no correction at 313 K (dashed curve), SAFT-VR
model with a correction of δHCO2 = 50 kJ mol−1 at 313 K (continuous
curve), and experimental data at 313 K (squares) from Kim and Svendsen
(2007); SAFT-VR model predictions with no correction at 393 K (dotted
curve), SAFT-VR model with a correction of δHCO2 = 80 kJ mol−1 at 393
K (dot-dashed curve) and experimental data at 393 K (void squares) from
Kim and Svendsen (2007). . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.1 Diagram of a non-equilibrium stage. This stage represents a section of

packing in a packed column. . . . . . . . . . . . . . . . . . . . . . . . . . . 71
12
4.2 A schematic of the two-film model of a column stage. δ V and δ L represent
the thickness of the gas and liquid films, respectively, P V , P I , and P L are
the pressure in the bulk vapour phase, at the gas-liquid interface and in
the bulk liquid, respectively. As can be seen an isobaric profile is assumed.
The temperatures of the bulk vapour phase, at the gas-liquid interface and
of the bulk liquid are denoted by T V , T I , and T L , respectively. Finally, yi
and xi are the mole fractions of component i in the bulk vapour and liquid
phases, respectively, and yiI and xIi are the mole fractions of component i at
the vapour-liquid interface in the vapour and the liquid phases, respectively. 72
4.3 Variation of the diffusion coefficient of CO2 in the liquid phase with regard
to the temperature. (a) Results obtained without scaling the diffusivity (τ
= 1), and (b) results obtained by scaling diffusivity to 3.7% of its original
value (τ = 0.037). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.1 Comparison of the pilot plant data (squares) (Tontiwachwuthikul et al.,

1992) and the result of our model (curves) for Run T22 for the absorption
of CO2 in an 18wt% solution of MEA. The dashed curves represent the
results obtained without scaling the diffusivity (τ = 1) and the continuous
curves the results obtained by scaling the CO2 diffusivity in the liquid phase
to 3.7% of its original value (τ = 0.037). a) Temperature profile for the
liquid phase, b) gas-phase CO2 concentration profile, and c) liquid-phase
CO2 loading. Stage 50 corresponds to the bottom of the column. . . . . . . 91
5.2 Sensitivity analysis of the mass transfer correlation on the profiles pre-
dicted with our model compared with the experimental pilot plant data of
Run T22 (Tontiwachwuthikul et al., 1992): Mass-transfer correlations from
Onda et al. (1968a,b) (continuous curve), mass-transfer correlations from
Rocha et al. (1993, 1996). (a) Temperature profile for the liquid phase, (b)
gas phase CO2 concentration profile, and (c) liquid phase CO2 loading. . . 93
13
5.3 Comparison of the pilot plant data (squares) and the predictions with our
model (continuous curves) for run T13 taken from the work of Tontiwach-
wuthikul et al. (1992). (a) Temperature profile for the liquid phase, (b)
gas phase CO2 concentration profile, and (c) liquid phase CO2 loading. . . 95
5.4 Comparison between pilot plant data (squares) and models predictions
(continuous curves) for run T14 taken from the work of (Tontiwachwuthikul
et al., 1992). (a) Temperature profile for the liquid phase, (b) gas phase
CO2 concentration profile and, (c) liquid phase CO2 loading. . . . . . . . . 96
(continuous curves) for run T15 taken from the work of (Tontiwachwuthikul
et al., 1992). (a) Temperature profile for the liquid phase, (b) gas phase
(continuous curves) for run T16 taken from the work of Tontiwachwuthikul
et al. (1992). (a) Temperature profile for the liquid phase, (b) gas phase
14
5.12 Run T22 of the pilot plant data from Tontiwachwuthikul et al. (1992).
Comparison of the predictions of our model for: (a) temperature of the
bulk liquid phase (continuous curve), the temperature at the vapour-liquid
interface (dashed curve) and the temperature for the bulk vapour phase
(dot-dashed curve); (b) the gas phase CO2 concentration in the bulk vapour
phase (continuous curve) and at the interface (dashed curve); (c) the liquid
phase CO2 loading in the bulk liquid phase (continuous curve) and at the
interface (dashed curve). . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
5.13 Run T18 of the pilot plant data from Tontiwachwuthikul et al. (1992).
Comparison of the predictions of our model for: (a) temperature of the
bulk liquid phase (continuous curve), the temperature at the vapour-liquid
interface (dashed curve) and the temperature for the bulk vapour phase
(dot-dashed curve); (b) the gas phase CO2 concentration in the bulk vapour
phase (continuous curve) and at the interface (dashed curve); (c) the liquid
phase CO2 loading in the bulk liquid phase (continuous curve) and at the
interface (dashed curve). . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
15
5.14 Sensitivity analysis of the CO2 diffusivity in the liquid phase on the profiles
predicted with our model in terms of the scaling parameter τ compared with
the experimental pilot plant data of Run T22 (Tontiwachwuthikul et al.,
1992). τ = 0.037 (continuous curve), τ = 0.074 (dashed curve), τ = 0.0185
(dot-dashed curve). (a) Temperature profile for the liquid phase, (b) gas
phase CO2 concentration profile, and (c) liquid phase CO2 loading. . . . . 111
5.15 Sensitivity analysis of the liquid viscosity on the profiles predicted with our
model compared with the experimental pilot plant data of Run T22 (Ton-
tiwachwuthikul et al., 1992): Nominal value (continuous curve), viscosity
doubled (dashed curve), viscosity halved (dot-dashed curve). (a) Temper-
ature profile for the liquid phase, (b) gas phase CO2 concentration profile,
and (c) liquid phase CO2 loading. . . . . . . . . . . . . . . . . . . . . . . . 112
5.16 Sensitivity analysis of the vapour-liquid surface tension on the profiles pre-
dicted with our model compared with the experimental pilot plant data
of Run T22 (Tontiwachwuthikul et al., 1992): Nominal value (continuous
curve) (Vazquez et al., 1997), surface tension doubled (dashed curve), sur-
face tension halved (dot-dashed curve), surface tension correlation from
data from Jayarathna et al. (2013) (dotted curve). (a) Temperature profile
for the liquid phase, (b) gas phase CO2 concentration profile, and (c) liquid
phase CO2 loading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.17 Sensitivity analysis of the heat of absorption of CO2 on the profiles pre-
dicted with our model in terms of the correction factor δHCO2 compared
with the experimental pilot plant data of Run T22 (Tontiwachwuthikul
et al., 1992). δHCO2 = 50 kJ mol−1 (continuous curve), δHCO2 = 100
kJ mol−1 (dashed curve), δHCO2 = 25 kJ mol−1 (dot-dashed curve). (a)
Temperature profile for the liquid phase, (b) gas phase CO2 concentration
profile, and (c) liquid phase CO2 loading. . . . . . . . . . . . . . . . . . . . 115
16
5.18 Sensitivity analysis of the vapour-phase heat-transfer coefficient on the pro-
files predicted with our model compared with the experimental pilot plant
data of Run T22 (Tontiwachwuthikul et al., 1992). Nominal value (continu-
ous curve), coefficient value increased ten times (dashed curve), coefficient
value reduced ten times (dot-dashed curve). (a) Temperature profile for
the liquid phase, (b) gas phase CO2 concentration profile, and (c) liquid
phase CO2 loading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.19 Sensitivity analysis of the enthalpy of vaporisation of water on the profiles
predicted with our model compared with the experimental pilot plant data
of Run T22 (Tontiwachwuthikul et al., 1992). Nominal value (continu-
ous curve), enthalpy doubled (dashed curve), enthalpy halved (dot-dashed
curve). (a) Temperature profile for the liquid phase, (b) gas phase CO2
concentration profile, and (c) liquid phase CO2 loading. . . . . . . . . . . . 117
5.20 Sensitivity analysis of the amount of water in the inlet flue gas on the
profiles predicted with our model compared with the experimental pilot
plant data of Run T22 (Tontiwachwuthikul et al., 1992). Nominal value
(continuous curve), yH2 O = 0.08 (dashed curve), yH2 O = 0 (dot-dashed
curve). (a) Temperature profile for the liquid phase, (b) gas phase CO2
concentration profile, and (c) liquid phase CO2 loading. . . . . . . . . . . . 119
6.1 Diagram of the desorber . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

6.2 Diagram of the reboiler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
6.3 Diagram of the condenser . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
7.1 Topology of the solvent regeneration unit . . . . . . . . . . . . . . . . . . . 136
17
7.2 Comparison between the model predictions and the experimental data from
Tobiesen et al. (2008) (a) Loading at the bottom of the desorber, (b) load-
ing at the reboiler outlet, (c) temperature at the bottom of the desorber,
(d) temperature at the reboiler outlet, (e) condensate flowrate out of the
condenser and, (f) CO2 flowrate out of the condenser. This last value is
not used to assess the validity of the model (see text). . . . . . . . . . . . . 138
7.3 Run 12 of the pilot plant data from Tobiesen et al. (2008). Comparison of
the predictions of our model for: (a) temperature of the bulk liquid phase
(continuous curve), the temperature at the vapour-liquid interface (dashed
curve) and the temperature for the bulk vapour phase (dot-dashed curve);
(b) the gas phase CO2 concentration in the bulk vapour phase (continuous
curve) and at the interface (dashed curve); (c) the liquid phase CO2 loading
in the bulk liquid phase (continuous curve) and at the interface (dashed
curve). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
7.4 Sensitivity analysis of the heat of absorption of CO2 on the profiles pre-
dicted with our model in terms of the correction factor δHCO2 compared
with the experimental pilot plant data of Run 12 from Tobiesen et al.
(2008). δHCO2 = 80 kJ mol−1 (continuous curve), δHCO2 = 160 kJ mol−1
(dashed curve), δHCO2 = 40 kJ mol−1 (dot-dashed curve). (a) Temperature
profile for the liquid phase, (b) gas phase CO2 concentration profile and,
(c) liquid phase CO2 loading. . . . . . . . . . . . . . . . . . . . . . . . . . 144
7.5 Sensitivity analysis of the heat of absorption of CO2 on the predictions of
our model in terms of the correction factor δHCO2 compared with the ex-
perimental pilot plant data of 6 selected runs from Tobiesen et al. (2008).
δHCO2 = 80 kJ mol−1 (squares), δHCO2 = 160 kJ mol−1 (upward triangles),
δHCO2 = 40 kJ mol−1 (downward triangles). (a) Loading at the bottom of
the desorber, (b) loading at the reboiler outlet, (c) temperature at the bot-
tom of the desorber, (d) temperature at the reboiler outlet, (e) condensate
flowrate out of the condenser and, (f) CO2 flowrate out of the condenser . 145
18
7.6 Sensitivity analysis of the enthalpy of vaporisation of water on the pro-
files predicted with our model compared with the experimental pilot plant
data of Run 12 from Tobiesen et al. (2008). Nominal value (continuous
curve), enthalpy doubled (dashed curve), enthalpy halved (mixed curve).
(a) Temperature profile for the liquid phase, (b) gas phase CO2 concentra-
tion profile and, (c) liquid phase CO2 loading. . . . . . . . . . . . . . . . . 147
7.7 Sensitivity analysis of the enthalpy of vaporisation of water on the predic-
tions of our model compared with the experimental pilot plant data of 6
selected runs from Tobiesen et al. (2008). Nominal value (squares), en-
thalpy doubled (upward triangles), enthalpy halved (downward triangles).
(a) Loading at the bottom of the desorber, (b) loading at the reboiler out-
let, (c) Temperature at the bottom of the desorber, (d) temperature at
the reboiler outlet, (e) Condensate flowrate out of the condenser, (f) CO2
flowrate out of the condenser . . . . . . . . . . . . . . . . . . . . . . . . . . 148
7.8 Sensitivity analysis of the reboiler heat duty on the predictions of our model
compared with the experimental pilot plant data of 6 selected runs from
Tobiesen et al. (2008). Nominal value (squares), heat duty doubled (up-
ward triangles), heat duty halved (downward triangles). (a) Loading at the
bottom of the desorber, (b) loading at the reboiler outlet, (c) Temperature
at the bottom of the desorber, (d) temperature at the reboiler outlet, (e)
Condensate flowrate out of the condenser and, (f) CO2 flowrate out of the
condenser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
19
7.9 Impact of the reduction of the pressure in the reboiler on the predictions
of our model compared with the experimental pilot plant data of 6 selected
runs from Tobiesen et al. (2008). Nominal value, P≃0.2 Mpa (squares),
reduced pressure P=0.14 Mpa (downward triangles). (a) Loading at the
bottom of the desorber, (b) loading at the reboiler outlet, (c) Temperature
at the bottom of the desorber, (d) temperature at the reboiler outlet, (e)
Condensate flowrate out of the condenser and, (f) CO2 flowrate out of the
condenser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
8.1 Schematic representation of the complete absorber-desorber flowsheet . . . 157

8.2 Stream tables for the reference run . . . . . . . . . . . . . . . . . . . . . . 158
8.3 Model predictions for a complete absorber-desorber flowsheet. (a) Ab-
sorber temperature profile for the liquid phase, (b) absorber gas phase
CO2 concentration profile, (c) absorber liquid phase CO2 loading (d) des-
orber temperature profile for the liquid phase, (e) desorber gas phase CO2
concentration profile, and (f) desorber liquid phase CO2 loading. . . . . . . 159
8.4 Impact of addition of MEA in the system via the solvent make-up on the
make−up
model predictions for a complete absorber-desorber flowsheet. LMEA =0
make−up
mol s−1 (continuous line), LMEA = 10−7 mol s−1 (dashed line), Lmake−up
MEA
= 10−6 mol s−1 (dot-dashed line) and, Lmake−up

MEA = 10−5 mol s−1 (dotted
line). (a) Absorber temperature profile for the liquid phase, (b) absorber
gas phase CO2 concentration profile, (c) absorber liquid phase CO2 loading
(d) desorber temperature profile for the liquid phase, (e) desorber gas phase
CO2 concentration profile and, (f) desorber liquid phase CO2 loading. . . . 162
8.5 Stream tables of the process model with Lmake−up
MEA = 10−7 mol s−1 . . . . . 163
MEA = 10−6 mol s−1 . . . . . 164
MEA = 10−5 mol s−1 . . . . . 165
20
(continuous curves) for Run 1 taken from the work of Notz et al. (2012).
(a) Temperature profile for the liquid phase and, (b) mass fraction of CO2
in the liquid phase. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
(continuous curves) for Run 2 taken from the work of Notz et al. (2012).
(a) Temperature profile for the liquid phase and, (b) mass fraction of CO2
in the liquid phase. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
21
List of Tables
2.1 Overview of models of CO2 absorber using MEA solutions. . . . . . . . . . 46

2.2 Experimental pilot plant data of CO2 absorber using MEA solutions. . . . 47
3.1 The SAFT-VR parameters characterising the pure component models stud-
ied in our current work: the number of segments mi , the diameter of the
spherical core σii , the depth εii and range λii of the dispersive square well
potential, the type and the number of hydrogen–bonding sites. . . . . . . . 59
3.2 Binary interaction parameters for the mixtures relevant to our current
work. kij characterises the strength of the dispersion interaction between
molecules of types i and j, see equation 3.3. . . . . . . . . . . . . . . . . . 60
3.3 The a−b site-site association energies εHB
ab,ij for MEA, H2 O and CO2 (Figure
3.3). The interaction matrix is symmetric, i.e., εHB HB

eH,ii = εHe,ii ; the lower
diagonal part has been left blank and is implied. The unlike association
between sites of the same type is assumed to be symmetric, i.e., εHB
eH,ij =
εHB HB HB
He,ij = εeH,ji = εHe,ji (Mac Dowell et al., 2010c; Rodrı́guez et al., 2012). . . 60
22
3.4 The a − b site-site range parameters rc;ab,ij for mixtures containing MEA,
H2 O and CO2 (Figure 3.3). The interaction matrix is symmetric, i.e., rc;eH,ii
= rc;He,ii and so the lower diagonal part has been left blank. The unlike
association between sites of the same type is assumed to be symmetric, i.e.,
rc;eH,ij = rc;He,ij = rc;eH,ji = rc;He,ji (Mac Dowell et al., 2010c; Rodrı́guez
et al., 2012). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.5 List of specified variables for the simulation of the enthalpy of absorption
of CO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.1 Correlations used in the model. . . . . . . . . . . . . . . . . . . . . . . . . 85
5.1 Inputs to absorber column model for runs T13 to T22 (Tontiwachwuthikul
et al., 1992). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
5.2 Characteristics of the column and the packing . . . . . . . . . . . . . . . . 89
5.3 Sensitivity analysis of key parameters in the model . . . . . . . . . . . . . 109
7.1 Inputs for the simulated runs for the desorber model . . . . . . . . . . . . 135
7.2 Desorber column properties . . . . . . . . . . . . . . . . . . . . . . . . . . 136
7.3 Sensitivity analysis of key parameters in the model . . . . . . . . . . . . . 140
7.4 Heat consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
7.5 Standalone reboiler simulation for Run T13 . . . . . . . . . . . . . . . . . 152
8.1 Inputs to the complete process model . . . . . . . . . . . . . . . . . . . . . 160

8.2 Liquid solvent flowrate, MEA concentration and, CO2 recovery rate for
different flowrates of MEA addition . . . . . . . . . . . . . . . . . . . . . . 161
8.3 Inputs to the absorber model . . . . . . . . . . . . . . . . . . . . . . . . . 167
1 SPF file used for the SAFT-VR foreign object . . . . . . . . . . . . . . . . 209
23
Chapter 1
Introduction
1.1 Motivation for CO2 capture
Carbon dioxide (CO2 ) emissions are generally considered to play a major role in climate
change and particularly in global warming. Fossil fuel power plants are the major fixed
point-source emitters of CO2 . In response to global warming, the Roadmap for 2050 set
by the European Commission in 2011 suggested reducing greenhouse gas emissions in
Europe by 25% by 2020 and by 80% by 2050 (European Union, 2011). In this context,
the development of carbon capture systems must be addressed in the short term, and
amine-based post-combustion capture processes are seen as the most promising near-
term technology in terms of development and applicability. In this technique, absorption
is achieved both physically and chemically, so that significant CO2 removal can take
place even at low partial pressures of the greenhouse gas. The major advantage of this
technology is that it can be retrofitted to existing power plants (e.g., see Rao and Rubin
(2002); Figueroa et al. (2008); Mac Dowell et al. (2010a)).
There are however several concerns with this technology, in particular the large energy
requirements associated with solvent regeneration, the degradation of the solvent, which is
exacerbated by the presence of oxygen in the flue gas, and the environmental and health
24
1. Introduction 25
impact that may result from solvent losses and solvent degradation products. These
issues are particularly important because of the scale of deployment required to have a
meaningful impact on CO2 emissions. There are significant experimental programmes
to identify new solvents (Paul et al., 2008; Mangalapally et al., 2009; Bardow et al.,
2010; Barzagli et al., 2012; Mangalapally et al., 2012; Salkuyeh and Mofarahi, 2012) and
several pilot plant studies are under way (Tontiwachwuthikul et al., 1992; Dugas, 2006;
Tobiesen et al., 2007; Gabrielsen et al., 2007; Godini and Mowla, 2008; Cottrell et al.,
2009; Notz et al., 2012). Modelling studies can play an invaluable and complementary role
in addressing some of these issues, including the choice of solvent and operating conditions
that yield optimal performance.
A key challenge in realising the benefits of a model-based approach is the development of
models that can accurately predict the behaviour of the process under different conditions
and for a range of solvents. This is particularly difficult in CO2 absorption due to the
complex reaction chemistry that occurs and the large number of ionic species present in the
process. For example, in the case of absorption of CO2 using the most common solvent, an
aqueous solution of monoethanolamine (MEA, HOC2 H4 NH2 ), the key reactions, excluding
the speciation of water, are (Astarita, 1967; Hikita et al., 1977; Danckwerts, 1979; Laddha
and Danckwerts, 1981; Astarita et al., 1983; Penny and Ritter, 1983; Blauwhoff et al.,
1984):
CO2 + HOC2 H4 NH2 ↽HOC2 H4 NH+

⇀ −
2 CO2 (1.1)
HOC2 H4 NH+ −
2 CO2 + HOC2 H4 NH2
⇀
↽HOC2 H4 NHCO− +
2 + HOC2 H4 NH3 (1.2)
HOC2 H4 NH+ −
2 CO2 + H2 O
⇀
↽HOC2 H4 NHCO−
2 + H3 O
+
(1.3)
HOC2 H4 NHCO−
2 + H2 O
⇀
↽HOC2 H4 NH2 + HCO−
3 (1.4)
HOC2 H4 NH+ −
2 CO2 + H2 O ↽HOC2 H4 NH+
⇀ −
3 + HCO3 (1.5)
CO2 + OH− ⇀
↽HCO−
3 (1.6)
HOC2 H4 NH+
3 + H2 O ↽HOC2 H4 NH2 + H3 O+
⇀ (1.7)
⇀ +
CO2 + H2 O ↽HCO−
3 +H (1.8)
1. Introduction 26
The main reaction products are therefore the zwitterion (HOC2 H4 NH+ −
2 CO2 ), the car-
+
bamate (HOC2 H4 NHCO− −
2 + HOC2 H4 NH3 ), and bicarbonate (HCO3 ). The elucidation
and characterization of the speciation, reaction mechanism, equilibria and kinetics for
mixtures relevant to CO2 chemisorption typically requires extensive experimental inves-
tigations before a detailed model can be built. This presents a significant barrier to the
rapid development of improved processes for carbon capture. There is a need to develop
models that offer reasonable predictive capabilities without extensive reliance on thermo-
dynamic and kinetic data, and that can provide quantitative insight into the behaviour
of the process.
1.2 Objectives of this work
The aim of this work is to develop a reliable and predictive model of an amine based CO2
absorber-desorber process, capable of screening a wide range of operating conditions, us-
ing different solvents in order to gain a good understanding of the process and reduce its
energy penalty.
To realise this objective, a single thermodynamic approach is used to mediate both the
phase and the chemical equilibrium, limiting the reliance on experimental data. An ex-
tensive validation of the proposed models of the different unit operations constituting the
CO2 capture process is performed using different pilot plant data (Tontiwachwuthikul
et al., 1992; Tobiesen et al., 2008).
Insights on the behaviour of the process under different operating conditions and on the
relative importance of the different model parameters are obtained via a sensitivity anal-
ysis.
1. Introduction 27
1.3 Outline of thesis
This thesis is organised as follows:
• In chapter 2, a brief description of some of the most advanced CO2 techniques are
presented, comparing their characteristics and performances. The carbon capture
processes using amine based solvent are presented in more detail. The different
techniques to model these processes are described and compared. Different thermo-
dynamic theories and mass transfer models are presented.
• Chapter 3 first focuses on the evolution of thermodynamic theories applicable to

carbon capture. Then the SAFT-VR equation of state used in this thesis is in-
troduced and the molecular model for the mixtures of MEA, CO2 , H2 O and N2 is
presented in detail.
• In chapter 4, a predictive model of a rate-based absorber column incorporating the

SAFT-VR model is described. Due to the radically different approach followed in
this study, all the equations of the model are presented explicitly.
• In chapter 5, the rate based model developed in chapter 4, incorporating the ther-
modynamic treatment introduced in chapter 3, is validated against pilot plant data.
The key parameters of the model are identified and their impact assessed via a
sensitivity analysis.
• In chapter 6, the modelling of the solvent regeneration unit, composed of a desorber,

a reboiler, and a condenser, is presented. The absorber model previously developed
is transferred to the desorber model. The equations that differ from those presented
chapter 4 are listed explicitly. The models for the condenser and the reboiler are
proposed.
• In chapter 7, the predictive capabilities of the solvent regeneration unit model de-
veloped in chapter 6 are assessed using pilot plant data. The key parameters of the
1. Introduction 28
model are identified and their impact assessed via a sensitivity analysis.
• In chapter 8, the models developed in chapters 3, 4 and 6 are linked together to

form a complete flowsheet of an absorber- desorber process. An initial study of the
complete process is presented.
• Finally, in chapter 9, the conclusions of this work and suggestions for future work
are presented.
Chapter 2
Techniques for CO2 capture
In the introduction, it has been shown that carbon capture and storage (CCS) is a must
to reduce CO2 emissions and meet the target set by the European Union. The transport
and storage of CO2 are great challenges but nonetheless they are not insurmountable. In
this study, the focus will be placed on the capture of CO2 from flue gas. A carbon capture
method should be carefully selected by taking into account its applicability, its industrial
readiness, and its capital and operating costs. The environmental impact should be ad-
dressed too, as there is no point in swapping one pollutant for another. The fossil fuel
used to produce electricity in power plant are either gas or coal. It is the combustion of
these fossil fuels that results in the formation of CO2 . The flue gas from a natural gas-fired
power plant typically consists of 2-3% CO2 (Amrollahi et al., 2012), while a coal-fired one
consists of 10-20% CO2 (Nogueira and Mamora, 2008). There are also H2 O, O2 , N2 and,
impurities like NOx and SOx. These components and their potential effects on the carbon
capture process need to be considered when selecting a technology.
As explained previously, there are several carbon capture technologies, but only five of
them will be presented subsequently as they are already being used or likely to be used
by the industry. These key technologies are: calcium-looping, oxy-fuel combustion, pre-
combustion gasification combined cycle, pressure-swing adsorption, and amine solvent
29
2. Techniques for CO2 capture 30
(Figueroa et al., 2008; Rao and Rubin, 2002; Mac Dowell et al., 2010a; Wang et al., 2011).
In the next section, the five technologies are described in detail. The following section is
focused on amine solvent processes and presents the modelling effort to date.
2.1 CO2 capture technologies
2.1.1 Calcium looping
Among the high temperature solid sorbents available for CO2 capture, calcium oxide
(CaO) really stands out. This material can be derived from natural limestone, which is
widely available and comes at low cost. The process is often referred to as carbonate or
calcium looping in the literature. It uses the solid-gas reversible reaction between calcium
oxide and carbon dioxide to form calcium carbonate.
CaO(s) + CO2(g) ⇀
↽ CaCO3(s) (2.1)
The application of this process to carbon capture was first introduced by Shimizu et al.
(1999). This process is suitable for large scale industrial applications and compared to
other processes, the overall penalty efficiency is extremely competitive (Mac Dowell et al.,
2010a). The major energy penalty of this technology is the desorption of the CO2 that
needs to be done at temperature as high as 1200 K (Blamey et al., 2010). Moreover, it
is important to note that the sorbent derived from natural limestone loses its capacity
to capture CO2 after multiple capture and release cycles (Grasa et al., 2008). A side
advantage of this process is the potential link to the cement industry which is an important
CO2 emitter (Hadlington, 2011). After multiple CO2 absoprtion cycles, the spent CaO
could be sold to the cement industry and thus suppress the need for this industry to
produce it. The production of CaO is responsible for about half the CO2 emissions from
the cement industry, the coupling of these processes could de-carbonate both of them.
This type of technology exists at pilot plant scale and is now moving to demonstration
scale (Sánchez-Biezma et al., 2011). With positive feedback, this technology should be
considered for practical deployment in the medium to long term.
2.1.2 Oxy-fuel combustion
In oxy-fuel combustion, coal is burned with a gas mixture of O2 and recycled gas (CO2
and H2 O) from the combustion (Abraham et al., 1982). The proportion of oxygen in
this gaseous mixture is 30% (Buhre et al., 2005), significantly higher than that in air
(21%) and the gas does not contain N2 . The exhaust gas of this combustion will be
pure CO2 and H2 O which can be recovered by condensation. This technology allows for
nearly 100% of the CO2 to be captured (Figueroa et al., 2008). The oxy-fuel combustor
is similar to the traditional combustor using air (Wall et al., 2009). Thus, this process
could easily be retrofitted to an existing power-plant. The major energy penalty comes
from the separation of N2 from the air to produce O2 . This penalty is comparable to that
of an amine-based scrubbing process. Different technologies exist to produce the oxygen.
The industry standard is currently cryogenic air separation, but the energy needed to
produce cryogenic conditions is high (Buhre et al., 2005; Figueroa et al., 2008; Wall et al.,
2009). Coal-fired oxy-fuel combustion was first introduced by Abraham et al. (1982)
but received only mild interest. There has been a regain in interest for the last decade.
Buhre et al. (2005) presented a review of oxy-fuel combustion technology for coal-fired
power generation. They identified four issues for oxy-fuel combustion that need to be
addressed before further development: the heat transfer in the gas; the environmental
issues, namely the control of gaseous emissions of SO2 ; the ash related issues and the
control of the combustion, including the flame stability.
2.1.3 Integrated gasification combined cycle
The integrated gasification combined cycle (IGCC) process consists of the gasification
of coal (or another fossil fuel) into synthesis gas with a sub-stoichiometric amount of
oxygen. The synthesis gas is composed of H2 and CO. Water is added to this mixture
and reacts with CO in the so-called water-gas shift reaction: CO+H2 O⇀
↽CO2 +H2 . The
CO2 can then be separated, leaving a product rich in H2 . The current benchmark to
remove the CO2 at this point in the process is the Selexol process (Pennline et al., 2008;
Figueroa et al., 2008). The major advantages of IGCC are the low energy penalty needed
to separate the CO2 from the H2 and the fact that the volume of gas to handle in the
combustor is small, as there is no nitrogen. However, there is still an energy penalty for
the shift reaction. Compared to an amine based post-combustion process, the IGCC offers
better performance in term of operating costs (Gibbins and Chalmers, 2008). Moreover,
this process produces H2 which can be either burnt directly to produce electricity or used
for other applications (Holt et al., 2003; Minchener, 2005). However, this method could
not be retrofitted to an existing power-plant and so the investment costs are higher than
for post-combustion chemical absorption (Damen et al., 2006; Figueroa et al., 2008; Mac
Dowell et al., 2010a), which is a major drawback. To deploy such a process, new power
generation plants and more complex systems should be built. Many existing power plants
are still to be used for decades. Since the target set by the European Union involves
short term reductions, this pre-combustion method will not be considered in this work.
Nevertheless, pre-combustion capture might be preferable over post-combustion as a long
term solution.
2.1.4 Pressure swing adsorption
Pressure swing adsorption (PSA) is a gas separation process in which the adsorbent is
regenerated by reducing the partial pressure of the adsorbed component. It was intro-
duced first by Skarstrom (1958) and Guerin de Montgareuil and Domine (1957). The first
applications of this system to CO2 capture showed only limited results on the product
purity due to the low selectivity between CO2 and CH4 on activated carbon (Doong and
Yang, 1986).
One of the key aspect in PSA is the choice of adsorbent (Yoo et al., 1995). There were
originally three main adsorbent type that were used for this process: activated carbon,
carbon molecular sieves and synthetic zeolites. The main characteristics of an adsorbent
in this context are the selectivity, the effective adsorption amount, the mass transfer ki-
netics, and the heat of adsorption. Zeolites have high selectivity for CO2 , but also a higher
heat of adsorption than carbon sorbent, thus rendering the desorption difficult (Kikkinides
et al., 1993; Yoo et al., 1995; Siriwardane et al., 2001). Kikkinides et al. (1993) obtained
a high purity of CO2 (99.97%) from flue gas using activated carbon and concluded that
this process was superior to the one using molecular sieves. Later on Yoo et al. (1995)
compared activated carbon and synthetic zeolite 13X and concluded that zeolite 13X is
a better adsorbent for bulk separation of CO2 from flue gas. Although zeolite 13X is
now considered as the benchmark sorbent for PSA (Sumida et al., 2012), a new family of
sorbents has gained interest recently; the metal-organic frameworks (MOFs) (Deng et al.,
2010; Britt et al., 2009; Yazaydin et al., 2009; Bae and Snurr, 2011; Sumida et al., 2012).
Extensive research is ongoing towards the design of these new sorbents for the application
in carbon capture using PSA.
Another crucial aspect in PSA processes is the operation procedure. A typical fixed-bed
adsorption system is operated in a cyclic manner, undergoing different steps; pressuriza-
tion, adsorption, blowdown, and desorption (Biegler et al., 2005). Several improvements
to this basic cycle have been introduced. For example, Berlin (1996); Marsh et al. (1964);
Wagner (1969) proposed a pressure equalization stage during which the beds are con-
nected and the pressures in the two chambers are equating, resulting in a saving in overall
process energy consumption since less mechanical energy is required.
The PSA systems are particularly suited for medium scale processes but existing systems
are not suitable for large scale industrial carbon capture (Wang et al., 2011).
2.1.5 Chemical absorption
In amine-based CO2 capture, the CO2 rich flue gas is contacted with a liquid amine solvent
in solution. The amine solvent reacts reversibly with the carbon dioxide and creates weak
bonds resulting in the formation of water-soluble salts. This bond formation is explained
in the next chapter. The solvent is then regenerated by heating, which breaks the bonds.
A flowsheet of a typical amine-based carbon capture process is represented in figure 2.1.
The flue gas is fed to the bottom of the absorber where it will travel upward and be in
contact with the liquid solvent flowing downward. The gaseous CO2 is reacting with the
liquid solvent and is transfer to the liquid phase. The cleaned flue gas escapes at the top of
the absorber, while the solvent rich in CO2 leaves at the bottom. This solvent is fed to the
top of the desorber, flowing downwards while in contact with a gaseous mixture of H2 O
and CO2 . At the bottom of the desorber there is a reboiler to heat up the rich solvent,
separate the CO2 from the solvent and transfer it to the gas phase along with some H2 O.
This gas mixture is fed back to the desorber and flows upward along the column. The
solvent leaving the reboiler is stripped from the CO2 and sent back to the absorber. At the
top of the desorber column, there is condenser to condense the H2 O back to the desorber
and obtain a high CO2 content gas stream. Three main classes of amines could be used
as chemical solvents. These are primary, secondary and tertiary amines, named after the
number of hydrogen atoms substituted by an alkyl group on the amine group. The primary
amines, such as monoethanolamine (MEA, ), react strongly and quickly with
CO2 . They are most suitable for low partial pressures of CO2 . However, these types of
amines are subject to degradation due to the oxygen contained in the flue gas (Bello and
Idem, 2006; Uyanga and Idem, 2007; Davidson, 2007). Secondary and tertiary amines
such as diethanolamine (DEA, ) and methyl diethanolamine (MDEA,
) respectively, have a lower reactivity (Laddha and Danckwerts, 1981;

Blauwhoff et al., 1984; Barth et al., 1984) which eases the regeneration of the solvent as
the association bonds are weaker. Less heating is needed in the desorber. Nevertheless,
as the rate of absorption is lower for these types of amines, they are unsuitable to work
Figure 2.1: Flowsheet of an amine-based carbon capture process
at low CO2 partial pressure. In the context of carbon capture for power plant generation
where low CO2 partial pressures are involved, MEA is considered as the benchmark. The
problem of degradation can be addressed by using additives to the solvent to prevent the
reactions with oxygen (Bello and Idem, 2006; Sexton and Rochelle, 2009). An important
advantage of amine-based CO2 capture is its potential applicability to the majority of
the existing large fixed point CO2 emitters as it could be retrofitted to an existing power
plant (Rao and Rubin, 2002; Figueroa et al., 2008; Mac Dowell et al., 2010a).
In addition to the ionic speciation equilibria due to the dissociation of CO2 and the
amines in aqueous solution, the principal reaction of interest between CO2 and a primary
or secondary amine (in water) is the formation of a carbamate:
+
CO2 + 2R1 R2 NH ⇀
↽ R1 R2 NCO−
2 + R1 R2 NH2 (2.2)
This reaction is the combination of the two chemical reactions 1.1 and 1.2 presented in
chapter 1. It can be seen that the overall stoichiometry of the reaction is such that each
molecule of CO2 is eventually associated with two amine molecules.
The five technologies presented in this section show promising capacities for carbon cap-
ture and are all in constant development. Early work on solvent-based carbon capture
processes date back from 1980, with notably the contributions from Hikita and Konishi
(1983), Takahashi et al. (1984), Yonemoto et al. (1984), Takeshita and Kitamoto (1988).
Rao and Rubin (2002) presented 4 reasons why the amine-based CO2 absorption processes
are the most suitable technologies for the short and medium term. (1) This technology is
suitable for dilute CO2 stream, a flue gas from a coal-fired plant typically contains 10-20
% CO2 . (2) It is a proven technology that is commercially used today. (3) Amine-based
systems can be retrofitted to an existing power plant, in the same way as other end-of-pipe
environmental control systems. (4) As can be seen by the high number of articles being
published on the model development to improve this technology, one can expect future
technological advances.
The different modelling approaches that have been developed for this technology are dis-
cussed in the remainder of this chapter.
2.2 Amine-based processes
2.2.1 Importance of the thermodynamic and the transport mod-

els
In order to design a process for carbon capture, one should develop a model which incor-
porates both the thermodynamics (Gani and O’Connell, 2001) and the transport phenom-
ena (Taylor and Krishna, 1993) relevant to the process. Most non-trivial process models
make use of physical properties, such as the chemical potential, density and enthalpy,
which need to be determined accurately with a thermodynamic model. The choice of
the thermodynamic model is of high importance as it is linked directly to the accurate
determination of the vapour-liquid equilibrium that is the key to the description of the
separation process. This choice is described subsequently. A model for the transport
phenomena then has to be developed to describe the exchange of mass and energy within
and between the different phases. A goal of our process model development is to address
concerns relating the large scale deployment of this technology, including energy require-
ments, solvent degradation and the environmental and health impact resulting from the
loss of solvent and solvent degradation products. Modelling studies can play a useful role
in addressing some of these issues and identifying the choice of solvent and operating
conditions that yield the best performance. A key challenge is to develop models that can
predict accurately the behaviour of the process and are robust enough to span the entire
optimisation domain.
Any model considered for the design of CO2 capture processes must be robust and pre-
dictive. The fluids comprising CO2 absorbed in aqueous amine solvents considered in this
study are highly complex, exhibiting extensive ionic speciation, intermolecular associa-
tion via hydrogen bonds and chemical reactions. Thus, a model specifically designed for
these reactive fluids is required. In this section, the methods typically used for modelling
such fluids are reviewed, followed by a focus on the use of SAFT theory in process models.
There is an extensive body of literature concerning the process modelling and simula-
tion of CO2 absorption in packed columns. The modelling approaches that have been
proposed to date differ in the choice of thermodynamic and kinetic models, and, where
appropriate, heat- and mass-transfer models. Most of the effort has focused on the use
of aqueous monoethanolamine solutions, due to their widespread industrial use and the
availability of pilot plant data, although there have been been some models developed for
other solvents, notably aminomethylpropanol (AMP) (Aboudheir et al., 2006; Gabrielsen
et al., 2006, 2007; Choi et al., 2009). The models presented here are applicable to the
absorber column; the specificities of the models developed for the desorber column are
presented in Chapter 6.
2.2.2 Column description
The amine-based carbon capture process involves an absorber and a desorber (cf. Figure
2.1). The absorber and the desorber are countercurrent vapour-liquid multistage separa-
tion columns, with a liquid feed at the top stage and a vapour feed at the bottom stage.
A vapour product comes off the top stage and a liquid product off the bottom stage. The
inside of the column is filled with an inert packing material designed for maximum mass
transfer between the vapour and the liquid and for low pressure drop. The vapour and
liquid compositions vary continuously with packing height rather than discretely as in a
tray column. The modelling of such columns could be either discrete or continuous in the
vertical direction (Taylor and Krishna, 1993; Khoury, 2005; Gáspár and Cormo, 2011).
2.2.3 The concept of a stage
The column is usually divided into hypothetical stages, each representing a (sometimes
infinitesimal) section of packing in the column (Taylor and Krishna, 1993; Khoury, 2005).
A larger number of stages yields better results (Taylor and Krishna, 1993). Taken to the
limit, using an infinite number of stages would be equivalent to modelling the column
continuously. However, experience shows that above a certain number of stages, the dif-
ference in the results is negligible. A large number of segments needs more computational
time, so a suitable compromise should be found. A discrete model of the column is suffi-
cient to provide good predictive capability, compared to a more complex continuous model
that would involve a significantly higher number of equations and differential equations.
The use of discrete stages is the most common choice for the modelling of such processes
(Taylor and Krishna, 1993; Khoury, 2005). Although both approaches could be used, the
discrete model is chosen for this study.
Figure 2.2: Diagram of a non-equilibrium stage. This stage represents a section of packing
in a packed column.
2.2.4 Rate-based vs Equilibrium models
Each stage can be modelled using either an equilibrium or a rate-based model. In an

equilibrium model, vapour-liquid equilibrium is assumed on each stage, everywhere on
the stage. A rate-based model accounts for limitations due to transport phenomena. For
this model, the assumption is made that the two phases are distinct, and each phase is
perfectly mixed on each stage. Figure 2.2 presents a diagram of a non-equilibrium stage.
Vapour from the stage below is brought into contact with liquid from the stage above and
the two phases exchange mass and energy through their common interface represented in
the diagram by the wavy line. In a rate-based model, separate mass balances are written
for each phase. The two phases are in contact through their interface where the material
lost by the vapour phase is gained by the liquid phase and vice versa. The heat transfer
is treated in a similar way. There is one energy balance in each phase, and the phases are
linked together via the rate of energy transferred across the phase interface. The compo-
sitions of the two phases at the interface are determined assuming equilibrium conditions
at the interface.
The rate-based model is a more precise representation of the physical process than the
equilibrium model. For chemisorption processes with fast reaction kinetics, as is the case
for the process of interest (Blauwhoff et al., 1984; Ying and Eimer, 2013), a rate-based
process is more reliable. Indeed, Lawal et al. (2009) compared the two approaches using
the same physical property model and concluded that the rate-based model gave a better
description of pilot plant (Dugas, 2006) temperature profiles.
2.2.5 Two film theory and enhancement factor
When using a rate-based approach, an important aspect in model development is the

choice of approach to treat heat- and mass-transfer phenomena. The concentration and
temperature profiles across hypothetical films in a two-film model can be imposed, taking
into account the effect of chemical reactions on mass transfer with an enhancement fac-
tor, defined as the ratio of the amount of gas absorbed in a reacting liquid to the amount
which would be absorbed if there were no reaction (Danckwerts, 1970; DeCoursey, 1982;
van Swaaij and Versteeg, 1992). A liquid-gas interface between the two films is sketched in
Figure 2.3, where the profiles for composition of component i, temperature and pressure
are represented.
The enhancement factor varies along the length of the column and can often be ad-
justed to pilot plant data. This is the route followed in most models of CO2 absorption
(e.g., Pandya (1983); Tontiwachwuthikul et al. (1992); Pintola et al. (1993); Pacheco and
Rochelle (1998); Tobiesen et al. (2007); Faramarzi et al. (2010); Lawal et al. (2010)).
In an improvement over film theory, Tobiesen et al. (2007) developed a penetration model,
where the two films at the interface are described continuously. This was found to describe
their own pilot plant data well. A more rigorous approach is that followed by Kucka et al.
(2003a), in which the Maxwell-Stefan formalism is used together with film discretisation
to account for the effect of the chemical reaction on the mass transfer. This more detailed
model leads to better predictions of concentration and temperature profiles at the pilot
Figure 2.3: A schematic of the two-film model of a column stage. δ V and δ L represent
the thickness of the gas and liquid films, respectively. P V , P I , and P L are the pressure
in the bulk vapour phase, at the gas-liquid interface and in the bulk liquid, respectively.
As can be seen an isobaric profile is assumed. The temperatures of the bulk vapour
phase, at the gas-liquid interface, and of the bulk liquid are denoted by T V , T I , and T L ,
respectively. Finally, yi and xi are the mole fractions of component i in the bulk vapour
and liquid phases, respectively, and yiI and xIi are the mole fractions of component i at
the vapour-liquid interface in the vapour and the liquid phases, respectively.
plant scale (Tontiwachwuthikul et al., 1992) than other models, without the need to fit
any parameters to pilot plant data (Kucka et al., 2003a). However, in keeping with the
objective of developing a simple model using an implicit treatment of the chemical reac-
tions, this last approach is unnecessary.
2.2.6 Mass transfer correlations
In all rate-based models, empirical mass-transfer correlations are required to take into
account the type of packing used and the operating conditions. The correlations of Onda
et al. (1968a,b) and Bravo and Fair (1982) were developed specifically for random packing,
whereas the correlations of Rocha et al. (1996) were developed for structured packing, but
can be applied to random packing by using an equivalence relation linking the random
packing characteristics to the structured packing ones. Correlations developed by Billet
and Schultes (1999) are also available as they apply to both structured and random pack-
ing. In a detailed comparison applied to a model of a CO2 capture pilot plant presented
by Faramarzi et al. (2010), it appears that the main operating conditions to consider when
choosing a mass-transfer correlation are the flowrates of the flue gas and the lean solvent.
2.2.7 Phase equilibrium and chemical reactions
Another important component in modelling CO2 absorption is the representation of the

chemical reactions and fluid phase equilibria of the mixture of MEA, CO2 and H2 O. Most
models proposed to date rely on the use of reaction kinetics derived from experimental
data specific to each reaction (e.g., Austgen et al. (1989, 1991); Kucka et al. (2002); Noeres
et al. (2003); Kenig et al. (2003); Kucka et al. (2003a,b)). In the earliest absorber models,
the thermodynamics of the gas and liquid phases were described with the assumption of
ideal gas and ideal solution behaviour. This is the case for example of the model developed
by Pandya (1983) and later used by Tontiwachwuthikul et al. (1992). However, this model
was too simplified to describe the complex interactions between the CO2 and the solvent,
and might not be suitable when transferred to other solvents. The combination of the
electrolyte-NRTL (eNRTL) approach (Chen and Evans, 1986; Austgen et al., 1989) with
a Henry’s constant to describe CO2 solubility is a method which has attracted much
interest and is typically used in recent studies (Lawal et al., 2010). The eNRTL approach
has recently been corrected for inconsistencies by Bollas et al. (2008). The major drawback
of this method is that it contains a large number of parameters, requiring a great deal of
experimental data for vapour-liquid equilibrium (VLE) and reaction kinetics, thus making
it difficult to use with different solvents. A similar approach (Kucka et al., 2003a) is to
combine the Soave-Rechlich-Kwong (SRK) equation of state (Poling et al., 2001) with
the eNRTL model. Tobiesen et al. (2007) adapted an activity coefficient model from
Hoff (2003) to fit VLE experimental data, using an experimentally-derived equilibrium
constant (Weiland et al., 2004).
As an alternative to reduce the number of parameters to be fitted and the computational
effort needed to solve the vapour-liquid equilibrium, Gabrielsen et al. (2005) developed
their own thermodynamic model to determine the VLE of MEA, CO2 and H2 O. They
used a simple correlation derived from experimental data of CO2 solubility in aqueous
MEA. This correlation is valid only for MEA and over the conditions used in fitting:
loading of CO2 varying from 0 to 0.5 at two temperatures, 313K and 393K.
Although the combined reaction and activity coefficient models developed to date to
represent the MEA + CO2 + H2 O mixture provide an accurate representation of the
behaviour of this complex mixture over the range of conditions of interest (Isaacs et al.,
1980; Jou et al., 1982; Lal et al., 1985), they require extensive parametrisation and are
clearly not transferable to other solvents.
2.2.8 Integration of SAFT into process models
Recently, an alternative approach to modelling the phase and chemical equilibria of mix-
tures of alkanolamines, CO2 and water has emerged (Mac Dowell et al., 2010c; Rodrı́guez
et al., 2012). The approach has been shown to offer good transferability from one alka-
nolamine to another (Rodrı́guez et al., 2012), thereby requiring limited experimental data
as new solvents are investigated. It is based on a version of the molecular statistical
associating fluid theory (SAFT) equation of state (Jackson et al., 1988; Chapman et al.,
1989, 1990), in which a square-well intermolecular potential of variable range is used – this
version is referred to as SAFT-VR (Gil-Villegas et al., 1997; Galindo et al., 1998; Kontoge-
orgis and Folas, 2010; McCabe and Galindo, 2010). In the SAFT formalism, the chemical
reactions of the type seen in the mixtures of interest here can be mediated via short-range
association sites incorporated in the molecular models, thanks to an inter-site interaction
potential. In cases where chemical equilibrium holds, such an approach has been shown
to be equivalent to the chemical models used to date in which the aggregated chemical
species (e.g., carbamate, zwitterion) are modelled explicitly (Economou and Donohue,
1991). The SAFT family of equations has been successfully used for complex fluid phase
equilibrium (Müller and Gubbins, 2001; Tan et al., 2008; McCabe and Galindo, 2010).
Due to the strong basis of the equation of state in molecular thermodynamics and sta-
tistical mechanics, SAFT-VR parameters can often be transferred from one compound to
another compound on the basis of similarities in molecular structure and the predictions
obtained are valid over wide ranges of compositions, temperatures and pressures.
The use of the SAFT-VR thermodynamic approach within process models of CO2 absorp-
tion has recently started being explored for high pressure physical absorption in alkanes
(Pereira et al., 2011). In an early study (Mac Dowell, 2010; Mac Dowell et al., 2010b)
the modelling of an amine-based absorption process was considered and explored in the
context of solvent blend design. More recently, a dynamic model of a CO2 absorber, that
makes use of the SAFT-VR thermodynamic treatment, has been proposed (Mac Dowell
et al., 2013) and has been used for control (Arce et al., 2012) and economic (Mac Dowell
and Shah, 2013) investigations of post-combustion CO2 capture processes. Qualitative

agreement with data from two pilot plant runs from (Tontiwachwuthikul et al., 1992) was
found for the model used in these last three publications. The column profiles obtained
by Mac Dowell et al. (2013) are somewhat difficult to interpret as the location of the plant
data points as reported in the paper does not match the location of the sensors reported
in the physical pilot plant. Nevertheless, the results from Mac Dowell et al. (2010b) and
Mac Dowell et al. (2013) indicate that a physical approach to modelling the chemical and
physical equilibria allows one to capture the process behaviour accurately with a limited
number of parameters.
2.2.9 Summary of existing process models
The different absorber models cited in this section are listed in Table 2.1, together with
a summary of their main characteristics. Overall, the models reported to date are able
to represent the general behaviour of the absorber column. However, only the model
of Tontiwachwuthikul et al. (1992) provides accurate descriptions of the temperature of
the rich solvent at the outlet, and in all published models, the bulge in the temperature
profile, which is a well-known characteristic of this absorption process (Kvamsdal and
Rochelle, 2008), is reproduced qualitatively but not quantitatively. Existing models are
able to reproduce either its magnitude or its location along the column, but not both.
2.2.10 Overview of available pilot plant
All the absorber models presented in this study rely on experimental data for their val-
idation. There are only a few detailed sets of pilot plant data available for the capture
of CO2 in aqueous amine solutions. Some of the key contributions are listed in table 2.2.
Table 2.1: Overview of models of CO2 absorber using MEA solutions.
Source Thermodynamic model Mass transfer model Validation

Pandya (1983) ideal gas and ideal solutions Onda et al. (1968b)
2. Techniques for CO2 capture
Tontiwachwuthikul et al. Pandya (1983) Onda et al. (1968a,b); Cho Tontiwachwuthikul et al.
(1992) (1987) (1992)
Pintola et al. (1993) Pandya (1983) Onda et al. (1968a,b) Tontiwachwuthikul et al.
(1992); Pintola et al. (1993)
Kucka et al. (2003a) SRK and eNRTL Billet and Schultes (1999) Tontiwachwuthikul et al.
(1992)
Tobiesen et al. (2007) Tobiesen et al. (2007) Rocha et al. (1996) Tobiesen et al. (2007)
Lawal et al. (2009, 2010) eNRTL Onda et al. (1968a,b) Dugas (2006)
Zhang et al. (2009) eNRTL AspenPlus Dugas (2006)
Faramarzi et al. (2010) Gabrielsen et al. (2005) Onda et al. (1968a,b); Rocha Tontiwachwuthikul et al.
et al. (1996); Billet and (1992)
Schultes (1999)
Mores et al. (2011, 2012) Peng and Robinson (1976) Onda et al. (1968a,b); Bravo Tontiwachwuthikul et al.
and Fair (1982) (1992); Dugas (2006)
Gáspár and Cormo (2011) Wilson-NRF Rocha et al. (1996) Tobiesen et al. (2007)
46
Table 2.2: Experimental pilot plant data of CO2 absorber using MEA solutions.
Source Packing Height of Internal

packing (m) diameter (m)
Tontiwachwuthikul et al. (1992) Random 6.6 0.1
Dugas (2006) Structured; 6.1 0.43
Random
Tobiesen et al. (2007) Structured 4.36 0.15
Notz et al. (2012) Structured 4.2 0.125
Available pilot plant data for desorber columns are presented in Chapter 6. The first
thing to consider is the size of these pilot plants. All of these absorbers have a packing
height of the same order of magnitude (5m). This should be compared to an industrial
scale absorber with height of tens of meters. These absorbers also have the same diameter,
except for the one from Dugas (2006) which is larger. Once again, this is one order of
magnitude smaller than future industrial scale CO2 absorber. However, these pilot plant
are among the largest available at the time of writing. Demonstration and industrial scale
absorbers are being commissioned and to the best of our knowledge, no experimental data
have been published yet. The second important factor is the type of packing used. A
random packing is usually used on small installations and structured packing for larger
scale processes. Although structured packing are more expensive than random packings,
they have better pressure drop and mass transfer characteristics than random packings
and are, therefore, more and more used in new columns or for revamping older columns to
improve performance (Taylor and Krishna, 1993). It is highly probable that an industrial
CO2 absorber will have structured packing. The operating conditions at which these pilot
plants are runs are also important factors. The pilot plants from Tontiwachwuthikul et al.
(1992) and Dugas (2006) have a concentration of CO2 in the flue gas around 15 to 20%,
corresponding to the proportion in the flue gas from a coal-fired power plant. The pilot
plants from Tobiesen et al. (2007) and Notz et al. (2012) have a much lower concentration
of CO2 in the flue gas at 2 to 3%, representing a gas-fired power plant. Once a model is
validated, data from different pilot plant could be used to test the range of predictability
of this model. The works from Tontiwachwuthikul et al. (1992) and Dugas (2006) are
both well documented and most of all, there are values of temperature, and CO2 con-
centration in the gas phase and the liquid phase all along the column. These values are
essential to validate a model as the outlet values alone are not sufficient. Tobiesen et al.
(2007) presented graphs of these results but did not publish tabulated values. Notz et al.
(2012) present detailed measurement along the column but only two different runs are
documented. Moreover, it was published after the development of the model used in this
study. It should be considered as a reliable source of experimental data for future model
development. Although the pilot plant used in the work presented by Dugas (2006) is
larger than the other pilot plants, it should be treated with caution as there is reported in-
accuracy in the measurements of the gas flow rate (Dugas, 2006; Kvamsdal and Rochelle,
2008). For this study, the results presented by Tontiwachwuthikul et al. (1992) are used.
2.2.11 Concluding remarks
In the present study, a rate-based model for a CO2 absorber-desorber process using MEA
is developed. The SAFT-VR thermodynamic model is used together with mass transfer
correlations from Onda et al. (1968b). In the SAFT-VR thermodynamic model, the
reactions are treated implicitly, with the products of the chemical reaction treated as
aggregates of the reactant molecules, so that there is no need for explicit rates of reaction.
It is also proposed not to use an enhancement factor as the reactions are implicitly treated
within the thermodynamic model. The extent to which SAFT-VR can be integrated in a
process model is assessed.
Chapter 3
Thermodynamic theory and

molecular model
When absorbing or desorbing CO2 in an amine solvent, the gas is experiencing both a
change of phase and a chemical reaction. These two phenomena have to be precisely
described with a molecular model in order to accurately predict the process behaviour.
As discussed in chapter 2, different thermodynamic approaches have been proposed and
used for the modelling of carbon capture in a absorber desorber process (Chen and Evans,
1986; Austgen et al., 1989, 1991; Kucka et al., 2002; Noeres et al., 2003; Kenig et al., 2003;
Kucka et al., 2003a,b). A description of the process relies on the use of a thermodynamic
model for the VLE and on reaction kinetics derived from experimental data specific to
each reaction for the chemical reaction. In our current study, the SAFT-VR thermo-
dynamic framework is used to represent both the phase and the chemical equilibrium.
The theoretical background of the different thermodynamic models is described first; the
discussion is the centred on the SAFT equation-of-state and its integration in the CO2
capture process model.
49
3. Thermodynamic theory and molecular model 50
3.1 Thermodynamic theory
The treatment of the chemical-reaction equilibria relevant to the absorption process is a

key aspect of the modelling strategy adopted in our work. The vast majority of models
presented to date follow a chemical approach, in which each reaction product is modelled
as an independent species. This requires a large amount of experimental data on speciation
and its dependence on thermodynamic state (temperature, pressure, and composition)
which limits the transferability of the model to other solvents. In the physical approach
followed here, the chemical and physical interactions are treated on an equal footing,
within the SAFT-VR equation of state. Before presenting the SAFT-VR equation of
state, we review the different thermodynamic models.
There are three basic modelling approaches, historically categorised as chemical, quasi-
chemical or physical. These three approaches are discussed briefly, followed by a more
detail description of a specific physical approach: SAFT.
A critical assessment and comparison of chemical, quasichemical and physical approches
has been made by Economou and Donohue (1991): the associated-perturbed-anisotropic-
chain theory (APACT) (Ikonomou and Donohue, 1986, 1988; Economou and Donohue,
1990; Elliott et al., 1990), a lattice physical approach based on the quasichemical theory,
and SAFT, a continuum physical approach, were compared with traditional chemical
approaches. They showed that the functional form of the expressions derived from APACT
and SAFT are essentially identical, and very similar in form to those derived with chemical
equilibrium theories, reinforcing the thermodynamic equivalence of the various approaches
to associating fluids. The intermolecular potential parameters of the physical approach
can be used to provide an explicit state dependence of the equilibrium constant.
3.1.1 Chemical approach
The chemical approach was developed by Dolezalek (1908), where association between two
molecules can be treated as a chemical reaction. The n-mers resulting from the chemical
reactions are considered as distinct ideal chemical species. The contribution to the free
energy is thus completely entropic corresponding to xi lnxi for each chemical species i.
The model developed is considered suitable for mixtures that are close to ideal (Acree,
1984; Prausnitz et al., 1999). This method should be applied with care as it could lead to
an incorrect description. One could cite the example given by Dolezalek (1918), where the
application of a chemical approach lead to the conclusion that argon was forming a poly-
mer. Another disadvantage of this method is that if more than one associating component
is handled, then a large number of parameters are required as association products are
treated as distinct species and the use of temperature dependent equilibrium constants
implies a significant amount a experimental data, which could be difficult to obtain or
unavailable.
3.1.2 Quasi chemical approach
In the quasi-chemical approach, the non-idealities in the fluids are considered to arise from
non-random mixing at the molecular level. The formation of distinct association com-
pounds is no longer considered explicitly. As the strong interactions found in associating
systems bias the random mixing expected in simple fluid mixtures, the quasi-chemical
theories account for association by assigning large energetic contributions to the actual
associating interactions. In a random mixture, the probability of pair of components i and
j being in close proximity is taken to be proportional to the product of their mole fractions
xi xj , while in a non-random treatment such as the so-called quasi-chemical approach de-
veloped by Guggenheim (1952), the probability is weighted by an appropriate Boltzmann
factor: xi xj exp(−Eij /kT ), Eij being the energy characterising the interaction. This the-
ory is the basis for several well known treatments of liquid mixtures like the Non-Random
Liquid Theory (NRTL) (Renon and Prausnitz, 1968), the UNIversal QUAsi-Chemical
(UNIQUAC) (Abrams and Prausnitz, 1975), and the corresponding group contribution
UNIversal Functional Activity Coefficient (UNIFAC) (Fredenslund et al., 1975) meth-
ods. These models have shown to successfully describe the thermodynamic properties of
mixtures of non-polar, polar and associating liquids well below their critical points, but
cannot be applied as successfully to gases as the approaches are generally cast as a lattice
description of the fluid which is not appropriate for less dense fluids phases by not taking
into account the changes in pressure. For example, the NRTL method does not account
for compressibility of the fluid, hence an equation of state (typically a cubic equation) is
used separately to describe the gas phase.
Brandani (1983) has developed a succesful method to determine the association equi-
librium constants for pure components to give a representation of enthalpy of hydrogen
bonding. In a second paper Brandani and Evangelista (1984) used the UNIQUAC ap-
proach (an implementation of the quasichemical theory) to assess mixtures including
alcohols and amines. They showed that UNIQUAC can be applied in a rigorous manner
when one of the components in the mixture is associative and the other is not. When
mixtures of two associative components are examined one has to determine a much larger
set of interaction parameters (at least 7), limiting the practical applicability of the model.
3.1.3 Molecular perturbation approach, SAFT
One of the first continuum statistical-mechanical treatments of associating fluids were

presented by Andersen (1973, 1974) and by Dahl and Andersen (1983). In this approach,
association sites are located inside a repulsive core to mediate short-ranged and highly
attractive forces. The major innovation of this theory is the introduction of the density of
un-associated monomers as well as the usual total density. This multi-density formalism
is the key to the success of further developments of this method. It should be noted that
in this original approach one could handle only one bond per association site. A ma-
jor contribution was made shortly afterwards by Wertheim (1984a,b, 1986a,b), who used
graphical summation techniques to derive an approximation for the degree of association
and free energy of molecules with highly directional, strong, short-range attractive forces.
The statistical associating fluid theory (SAFT) family of equations of state (Müller and
Gubbins, 2001; Tan et al., 2008; McCabe and Galindo, 2010) stems from Wertheim’s first-
order thermodynamic perturbation theory (TPT1) (Wertheim, 1984a,b, 1986a,b; Jackson
et al., 1988; Chapman et al., 1988). In SAFT approaches molecules are modelled as chain
of bonded spherical segments with embedded short-range association sites incorporated to
mediate hydrogen bonding which lead to aggregate formation (speciation). The equation
of state is developed using a perturbation approach, such that the free energy of a chain
is obtained with respect to the properties (free energy and radial distribution function)
of a reference monomeric (un-bonded) system. The original version of SAFT (Chapman
et al., 1989, 1990) has been revised and modified by a number of researchers so that
several versions are now available (Huang and Radosz, 1990, 1991; Blas and Vega, 1997;
Gil-Villegas et al., 1997; Galindo et al., 1998; Gross and Sadowski, 2001, 2002; Lafitte
et al., 2006, 2007). The intermolecular potential defining the interactions between spher-
ical segments (monomers) varies from version to version. In the SAFT-VR equation of
state for potentials of variable range (VR) first introduced by Gil-Villegas et al. (1997)
and Galindo et al. (1998), a square-well (SW) potential is used to describe the monomer
reference fluid. Fixed forms of the intermolecular potential were used in the development
of earlier versions of SAFT; the SAFT-VR formulation allows for a generic description of
the effect of the range of the interaction and has been implemented for the square-well
(Gil-Villegas et al., 1997), Sutherland (Gil-Villegas et al., 1997), Lennard-Jones (Davies
et al., 1998), Yukawa (Davies et al., 1999), and Mie (generalised Lennard-Jones) (Lafitte
et al., 2006; Avendaño et al., 2013) potentials.
Specific details of the SAFT-VR equation of state employed in our current work are dis-
cussed in the next section.
(a) (b) (c)
Figure 3.1: The square-well potentials for a monomer, Φmono , and association between
the sites, Φassoc , employed in the SAFT-VR approach. (a) Φmono is defined by a hard-core
of diameter σ, range of attraction λσ and depth −ε. (b) Φassoc is defined by an off-centre
potential of depth −εab and of range rc;ab . (c) The centre of the site is at a distance rd
from the centre of the segment.
3.1.4 SAFT-VR
In the SAFT-VR SW approach, a molecule i is modelled as a homonuclear chain of

mi bonded spherical segments of diameter σii . The interactions between two identical
segments are described by a square-well potential of range λii and depth εii (cf. Figure
3.1). For each molecule i, the number of site types Ns,i must be defined, as well as the
number of sites of each type a, Ns,ia . The sites are characterised by site-site energetic
εHB
ab,ij and range rc;ab,ij parameters (cf. Figure 3.1).
In the SAFT framework, the Helmholtz free energy A of a model mixture of associating
chain molecules can be written as:
A Aideal Amono Achain Aassoc

= + + + (3.1)
N kT N kT N kT N kT N kT
where N is the number of molecules, k the Boltzmann constant, and T the absolute
temperature. The superscript ideal, mono, chain and assoc denote the contribution to
the Helmholtz free energy due to the ideal, the monomer, the chain, and the association
contributions, respectively. The first two terms are obtained from an augmented van
der Waals equation of state for a fluid of attracting hard-sphere molecules obtained with
a Barker and Henderson high-temperature perturbation theory (Barker and Henderson,
1967). The third term represents the effect on the free energy of irreversibly fusing hard
spheres into chains of m segments. The final term represents the contribution due to the
formation of bonds between association sites.
In order to model mixtures, combining rules based on the Lorentz-Berthelot approach (cf.
Haslam et al. (2008)) are used to describe the cross interaction between segments on two
different molecules i and j:
σii + σjj
σij = (3.2)
2
√
εij = (1 − kij ) εii εjj (3.3)
σii λii + σjj λjj
λij = (3.4)
σij
where kij is an adjustable parameter that characterises the unlike dispersion attractive
interaction.
In recent work (Mac Dowell et al., 2010c; Rodrı́guez et al., 2012) the reactions involved
in aqueous amine solutions of CO2 have been treated implicitly within the SAFT-VR
framework, with the products of the chemical reaction represented as aggregates of the
reactant molecules. To do so, the overall set of reactions is reduced to (Blauwhoff et al.,
1984):
CO2 + HOC2 H4 NH2 ↽ [HOC2 H4 NH+

⇀ −
2 + CO2 ]
[HOC2 H4 NH+ −
2 + CO2 ] + HOC2 H4 NH2 ↽ HOC2 H4 NH+
⇀ −
3 + HOC2 H4 NHCO2
Association sites that allow the complexation of CO2 and MEA are introduced in our
SAFT-VR molecular models. The reaction products can thus be modelled as neutral
aggregates of CO2 and MEA, bonded at association sites as shown in Figure 3.2. An
accurate representation of the vapour-liquid phase equilibria of MEA + CO2 + H2 O can
Figure 3.2: Schematic representation of the association scheme between MEA and CO2
showing two reaction products.
be obtained in this manner by estimating the molecular parameters from experimental

fluid phase equilibria data, for the pure components and mixtures. One important im-
plication of the physical treatment of chemical equilibrium is that there is no need to
explicitly specify a reaction scheme or reaction products. The types of products formed
(e.g., carbamates or bicarbonates) are dictated by the association scheme chosen (number
of sites and strength of their interactions), and the relative extent of formation of the
different products depends on the temperature-independent parameters that describe the
association energies. The fraction of molecules associated or un-associated at a given site
is an output of the SAFT-VR model (and others based on the Wertheim formalism) and
the distribution of reaction products can be determined from a statistical analysis of the
molecules not-bonded at given sites at the thermodynamic state of interest (Economou
and Donohue, 1991). Thus, although no speciation data are used in deriving the SAFT-VR
parameters, the speciation equilibria can nonetheless be described successfully (Rodrı́guez
et al., 2012).
A SAFT approach greatly reduces the number of parameters needed to describe the mix-
ture compared to explicit approaches such as eNRTL, in which every species must be
described as a separate entity and temperature-dependent equilibrium constants must
be derived for all the relevant reactions. The SAFT-VR approach has been shown to
be applicable to the absorption of CO2 for a number of aqueous alkanolamine solutions

(Rodrı́guez et al., 2012). In many cases, it is possible to transfer parameters from one
alkanolamine to another based on molecular similarity, further reducing the need for ex-
perimental data.
One key assumption in adopting this type of physical approach is that all reactions are
assumed to be at equilibrium, which is only applicable for processes in the physical regime,
i.e., where mass transfer is the rate limiting process. In the case where a specific treat-
ment of the charged electrolytic species is required, the SAFT-VRE equation approach
(Galindo et al., 1999; Gil-Villegas et al., 2001; Patel et al., 2003; Behzadi et al., 2005;
Paricaud et al., 2010) can be adopted following a chemical approach as appropriate. It
is also possible to represent some or all species explicitly so that any chemical reaction
that is not at equilibrium can then be modelled via a separate kinetic model. We do not
follow such an approach here, but instead test the validity of the chemical equilibrium
assumption.
3.2 Molecular model
The molecules considered in this study are MEA, H2 O, CO2 , and N2 . A schematic of the
molecular models used with the SAFT-VR equation is presented in figure 3.3.
The values of all parameters are listed tables 3.1, 3.2, 3.3 and 3.4.
(a) (b)
(c) (d)
Figure 3.3: A schematic of the molecular model used in the SAFT-VR method. (a) H2 O,
(b) MEA, (c) CO2 and (d) N2 . The sites e represent the electron lone pairs on the oxygen
atom, the sites e∗ correspond to a lone pair on the nitrogen atom, the sites H correspond
to the hydroxyls, the sites H ∗ correspond to the amines hydrogens, and the sites α1 and
1 α2 are acceptor sites.
Table 3.1: The SAFT-VR parameters characterising the pure component models studied in our current work: the number
of segments mi , the diameter of the spherical core σii , the depth εii and range λii of the dispersive square well potential,
the type and the number of hydrogen–bonding sites.
i mi σii (Å) εii /k (K) λii e-sites H-sites e∗ -sites H ∗ -sites α1 -sites α2 -sites source
H2 O 1.0 3.03420 250.00 1.78890 2 2 0 0 0 0 Clark et al. (2006)
3. Thermodynamic theory and molecular model
MEA 2.0 3.57229 305.00 1.58280 2 1 1 2 0 0 Mac Dowell et al. (2010c)

CO2 2.0 2.78640 179.27 1.51573 0 0 0 0 1 1 Rodrı́guez et al. (2012)
N2 1.4 3.07357 74.587 1.58795 0 0 0 0 0 0 Mac Dowell (2010)
59
Table 3.2: Binary interaction parameters for the mixtures relevant to our current work.
kij characterises the strength of the dispersion interaction between molecules of types i
and j, see equation 3.3.
i + j kij Source
MEA + CO2 0.47878 Rodrı́guez et al. (2012)
MEA + H2 O 0.01000 Mac Dowell et al. (2010c)
MEA + N2 0.03000 Mac Dowell (2010)
CO2 + H2 O -0.06000 Mac Dowell et al. (2010c)
CO2 + N2 -0.0599 Mac Dowell (2010)
H2 O + N2 -0.3635 Mac Dowell (2010)
Table 3.3: The a − b site-site association energies εHBab,ij for MEA, H2 O and CO2 (Figure
3.3). The interaction matrix is symmetric, i.e., εHB HB
eH,ii = εHe,ii ; the lower diagonal part has
been left blank and is implied. The unlike association between sites of the same type is
assumed to be symmetric, i.e., εHB HB HB HB
eH,ij = εHe,ij = εeH,ji = εHe,ji (Mac Dowell et al., 2010c;
Rodrı́guez et al., 2012).
εHB
ab,ij /k (K)
MEA H2 O CO2
b\a e H e∗ H ∗
e H α1 α2
e 0 2357.79 0 900.00 0 1780.71 0 0
H - 0 550.00 0 1780.71 0 0 0
MEA
e∗ - - 0 960.00 0 1517.10 5200.00 3982.66
H∗ - - - 0 1517.10 0 0 0
e - - - - 0 1400.00 0 0
H2 O
H - - - - - 0 0 0
α1 - - - - - - 0 0
CO2
α2 - - - - - - 0 0
Table 3.4: The a − b site-site range parameters rc;ab,ij for mixtures containing MEA, H2 O
and CO2 (Figure 3.3). The interaction matrix is symmetric, i.e., rc;eH,ii = rc;He,ii and so
the lower diagonal part has been left blank. The unlike association between sites of the
same type is assumed to be symmetric, i.e., rc;eH,ij = rc;He,ij = rc;eH,ji = rc;He,ji (Mac
Dowell et al., 2010c; Rodrı́guez et al., 2012).
rc;ab,ij (Å)
MEA H2 O CO2
b\a e H e∗ H∗ e H α1 α2
e 0 2.08979 0 2.65064 0 2.10763 0 0
H - 0 2.65064 0 2.10763 0 0 0
MEA
e∗ - - 0 2.32894 0 2.22626 1.97978 1.96999
H∗ - - - 0 2.22626 0 0 0
e - - - - 0 2.10822 0 0
H2 O
H - - - - - 0 0 0
α1 - - - - - - 0 0
CO2
α2 - - - - - - 0 0
The MEA molecule is represented with 2 tangent spherical segments and has 2 sites of
type e (electron lone pairs on the oxygen atom), 1 site of type e∗ (corresponding to the
lone pair on the nitrogen atom), 1 site of type H (hydroxyl), and two sites of type H ∗
(amine hydrogens) (Mac Dowell et al., 2010c). The H2 O is represented with 1 spherical
segment and has 2 e sites and 2 H sites (Clark et al., 2006). The CO2 model comprises
2 segments and has 1 α1 site and 1 α2 site (acceptor sites) that interact only with the
e∗ sites of MEA (Rodrı́guez et al., 2012). The N2 model is non-spherical, with an aspect
ratio of m = 1.4, and does not have association sites as it is chemically inert and does not
have strong directional interactions (Paricaud et al., 2004; Mac Dowell, 2010). A site of
type e/e∗ can bond with a site H/H ∗ and vice versa. For example, when an e site from
HB
MEA and an H site from water come within the cut-off range rc,eH,MEA−H 2O
of each other,
there is a site-site hydrogen-bonding associative interaction of energy -εHB
c,eH,MEA−H2 O . In
the standard Wertheim TPT1 treatment commonly used in SAFT, association into linear-
chain, branched-chain, and network aggregates are considered. Association into ring-like
structures (Sear and Jackson, 1994a; Ghonasgi et al., 1994; Sear and Jackson, 1996a;
Galindo et al., 2002) (and even double bonding (Sear and Jackson, 1994b) and bond co-
operativity (Sear and Jackson, 1996b)) can be taken into account but this is not found to
be necessary for the systems described in our current work. The values of the parameters
are listed in tables 3.1-3.4, and the file used in the gPROMS software containing these
parameters is in the appendix.
The SAFT-VR thermodynamic model is implemented as a foreign object in the gPROMS
software (Process Systems Enterprise Ltd.; Kakalis et al., 2004). The thermodynamic
properties needed to determine the fluid phase (and implicit chemical) equilibrium, namely
the pressure, the chemical potential and the enthalpy, are obtained in the usual manner
as the corresponding derivatives of the Helmholtz free energy.
3.3 Validation of the thermodynamic model
The thermodynamic model is used to calculate the VLE for the ternary mixture MEA-
CO2 -H2 O. Rodrı́guez et al. (2012) presented a comparison of the SAFT-VR description
of the VLE and the experimental data for three isotherms (313, 353 and 393 K) as shown
in Figure 3.4. A good description of the partial pressure of CO2 as a function of the CO2
loading, which is defined as the number of moles of CO2 absorbed in the liquid phase
per mole of MEA, is obtained for both low temperature (313 K) corresponding to the
absorber temperature range and high temperature (393 K) corresponding to the desorber
temperature range.
Another important aspect in the model validation is the description of the degree of spe-
ciation of the system. The main products of reactions in the model are the carbamate
(MEACOO− ) and the bicarbonate (HCO−
3 ) (cf. Equations 1.1-1.8). Rodrı́guez et al.
(2012) explained that the concentration of these products can be deduced from the num-
ber of bonded and unbonded α sites on the CO2 . The analysis of the degree of speciation
for MEA + H2 O + CO2 at 313 K and 353 K is presented in Figure 3.5. SAFT-VR pro-
vides very good predictions of the composition in carbamate and bicarbonate at different
10000
PCO (kPa)
100
2
1
SAFT-VR T = 313 K
Jou et al. T = 313 K
SAFT-VR T = 353 K
0.01 SAFT-VR T = 393 K
0 0.2 0.4 0.6 0.8 1
θCO
2
Figure 3.4: Solubility of CO2 in a 30wt% solution of MEA. SAFT-VR model predictions
(continuous curves) from Rodrı́guez et al. (2012) and experimental data (squares) from
Jou et al. (1995)
(a)
313 K
0.08
Carbamate (SAFT-VR)
Carbamate (Bottinger et al.)
Mole fraction
Bicarbonate (SAFT-VR)
0.06 Bicarbonate (Bottinger et al.)
0.04
0.02
0
0 0.2 0.4 0.6 0.8 1
θCO
2
(b)
353 K
0.08
Carbamate (SAFT-VR)
Carbamate (Bottinger et al.)
Mole fraction
Bicarbonate (SAFT-VR)
0.06 Bicarbonate (Bottinger et al.)
0.04
0.02
0
0 0.2 0.4 0.6 0.8 1
θCO
2
Figure 3.5: Predicted mole fraction of carbamate and bicarbonate in the ternary mixture
of MEA + H2 O + CO2 with a concentration of MEA in the liquid phase of 30 wt%. (a)
T = 313 K and (b) T = 353 K. The symbols (squares for carbamate and diamonds for
bicarbonate) correspond to the experimental data (Böttinger et al., 2008) and the curves
(continuous for carbamate and dashed for bicarbonate) to the SAFT-VR calculations
(Rodrı́guez et al., 2012).
temperatures.
3.4 Enthalpy of absorption
In the context of carbon capture processes, another important aspect to consider is the
enthalpy of absorption of CO2 , as the major source of heat in the system is the heat
released when CO2 is absorbed by MEA. In this section, the SAFT-VR model predictions
of the enthalpy of absorption of CO2 are compared to experimental data (Kim and Svend-
sen, 2007). It is shown that despite providing a very good description of the experimental
VLE and speciation data, SAFT-VR leads to value of the heat of absorption of CO2 ,
abs
∆HCO 2
, that are too low compared to experimental values (Kim and Svendsen, 2007).
Most equations of state do not provide an accurate description of the caloric properties
of complex fluids, although more recent developments can lead to a marked improvement
(Lafitte et al., 2006, 2013). A parameter is introduced in the model to correct the enthalpy
of absorption of CO2 .
3.4.1 Methodology
The experimental setup described by Kim and Svendsen (2007) has been modelled using
the gPROMS software (Process Systems Enterprise Ltd.). The setup consists of a 2 dm3
fixed volume equilibrium cell. A given number of moles of MEA, CO2 , H2 O and N2 is
placed in the cell at a given temperature. At the appropriate conditions the mixture
separates into a liquid phase and a gas phase. The total enthalpy of the cell is obtained
as the sum of the enthalpies from each coexisting phase using SAFT-VR. The procedure
is then repeated for a different concentration of CO2 and the enthalpy is then obtained as
a function of the loading of CO2 in the liquid phase. The difference in enthalpies divided
by the difference in the number of moles of CO2 between the two simulations gives the
enthalpy of absorption. In this model, the feed and equilibrium temperatures are not
Table 3.5: List of specified variables for the simulation of the enthalpy of absorption of
CO2 .
Variable Value Unit

Ntotal 6.5705164 ×10−5 mole
MEA Loading 30 wt%
CO2 loading 0.002
Tequ 313 K
Volume total 2 L
xin (N2 ) 2.8687698 ×10−4
equal as is the case in the experimental setup. However, the enthalpy resulting from the
heat capacity of the mixture and the temperature difference is negligible compared to the
enthalpy of absorption of CO2 . The specified values in the system are listed in Table 3.5.
3.4.2 Results
The comparison between the SAFT-VR predictions and the experimental data of the en-
thalpy of absorption of CO2 in MEA is depicted in Figure 3.6. At 313 K, which is a typical
temperature for CO2 absorbers, SAFT-VR underestimates the enthalpy of absorption of
CO2 by about 50 kJ mol−1 . This underestimation is 80 kJ mol−1 at 393 K, a typical
temperature for CO2 desorbers. To mitigate for this underestimation, the same strat-
egy as in other thermodynamic approaches is adopted and a correction of the enthalpy
of absorption δHCO2 is implemented (Kucka et al., 2003a; Gabrielsen et al., 2007; Plaza
et al., 2010a). When operating in the temperature range of the absorber, the value of this
parameter is fixed to δHCO2 = 50 kJ mol−1 . When operating in the temperature range of
the desorber, the value of this parameter is fixed to δHCO2 = 80 kJ mol−1 . The corrected
enthalpies of absorption of CO2 are presented in Figure 3.6.
120
-∆HCO (kJ mol )
-1
100
80
60
2
40
20
0
0 0.2 0.4 0.6 0.8
θCO
2
Figure 3.6: Enthalpy of absorption of CO2 in a 30 wt% solution of MEA: SAFT-VR
model predictions with no correction at 313 K (dashed curve), SAFT-VR model with a
correction of δHCO2 = 50 kJ mol−1 at 313 K (continuous curve), and experimental data
at 313 K (squares) from Kim and Svendsen (2007); SAFT-VR model predictions with no
correction at 393 K (dotted curve), SAFT-VR model with a correction of δHCO2 = 80 kJ
mol−1 at 393 K (dot-dashed curve) and experimental data at 393 K (void squares) from
Kim and Svendsen (2007).
3.5 Concluding remarks
The different thermodynamic models have been introduced in this chapter based on their
background theory: chemical, quasi-chemical and physical. From this last category, the
SAFT equation of state, is described in more details, focusing on SAFT-VR SW, the ther-
modynamic model used in the carbon capture process model which is the topic of this
dissertation. The molecular models of the mixture of MEA, CO2 , H2 O, and N2 is then
detailed. Both the vapour-liquid equilibrium and the chemical equilibrium are treated
within the SAFT-VR thermodynamic framework, ensuring a consistent and accurate rep-
resentation of the physical interactions in the system under the assumption that reaction
kinetics are not rate determining. This physical approach lends itself to extension to
other solvents, as a consequence of the transferable nature of the SAFT molecular models
and the relatively small number of parameters and data required to develop them. A
correction on the enthalpy of absorption of CO2 is proposed to compensate for an under
estimation with the theory.
Chapter 4
Development of an absorber model

using SAFT
4.1 Introduction
The role of the absorber in a solvent based carbon capture process and the need for a
transferable model of absorber have been explained in chapter 2. The existing models
and their performance have also been discussed. It appears that there are some reliable
models in the literature, some of them being entirely predictive at the process scale but
they always require a lot of kinetic and equilibrium data. This is a problem for solvent
selection, which is an important issue in CO2 capture. The alternative strategy used to
describe the VLE and the chemical reactions has been presented in chapter 3. The goal
of this study is to assess the applicability of this alternative strategy in a process model
to get good predictions and some insights into the process behaviour (e.g., the cause
of the formation of a temperature bulge). This chapter describes the development of a
rate-based model of a CO2 absorber that uses the approach introduced in chapter 3. A
single thermodynamic theory is used to describe the liquid and the gas phases, their VLE,
69
4. Development of an absorber model using SAFT 70
their chemical equilibrium, and their enthalpy. Another distinctive feature of the model
developed here is the implicit treatment of the chemical reactions, taking into account
only the reactants, hence limiting the compounds considered explicitly in this model to
only MEA, CO2 , H2 O, and N2 . These aspects of the modelling represent a radically
different approach to absorber modelling (Pandya, 1983; Tontiwachwuthikul et al., 1992;
Pintola et al., 1993; Kucka et al., 2003a; Tobiesen et al., 2007; Lawal et al., 2009, 2010;
Faramarzi et al., 2010; Mores et al., 2011, 2012; Gáspár and Cormo, 2011). The use of
the SAFT-VR thermodynamic theory for the development of a rate-based process model
of a CO2 absorber is presented in this chapter.
4.2 Non-equilibrium absorber model
The absorber is a counter-current vapour-liquid multistage separation column, with a

liquid feed at the top stage and a vapour feed at the bottom stage. The vapour product
comes off the top stage and the liquid product off the bottom stage. The inside of the
column is filled with an inert packing material designed for a maximum mass transfer
between the vapour and the liquid and for a low pressure drop. The vapour and liquid
compositions vary continuously with packing height rather than discretely as in a tray
column.
The modelling of such a column can be either discrete or continuous in the vertical
direction. In a discrete model, the column is divided in hypothetical stages, each of
which represents a section of packing in the packed column. The larger the number of
stages the better the description (Taylor and Krishna, 1993). Taken to the limit, using an
infinite number of stages is equivalent to modelling the column continuously. Although
both models could be used, we chose the discrete approach for our study.
The modelling of the stages can be equilibrium or rate-based. In an equilibrium model,
it is assumed that vapour-liquid equilibrium is achieved at each stage, everywhere on
the stage. The rate-based model accounts for limitations due to transport phenomena.
Figure 4.1: Diagram of a non-equilibrium stage. This stage represents a section of packing
in a packed column.
Due to the complexity of the process modelled in this case, a rate-based approach is
chosen over an equilibrium one in order to capture some of the key physical phenomena
(Krishnamurthy and Taylor, 1985a,b).
For this model, the assumption is made that the two phases are distinct, and each phase
is perfectly mixed at each stage. A diagram of a non-equilibrium stage is presented in
Figure 4.1.
Vapour from the stage below is brought into contact with liquid from the stage above and
the two phases exchange mass and energy through their common interface represented in
the diagram by the wavy line. In a rate-based model, separate mass balances are written
for each phase. The two phases are in contact through their interface where material lost
by the vapour phase is gained by the liquid phase. The heat transfer is treated in a similar
way. There is one energy balance in each phase, and the phases are linked via the rate of
energy transferred across the phase interface. The compositions of the two phases at the
interface are determined assuming equilibrium conditions at the interface.
The behaviour at the phase interface in the rate-based stages is described with a two-film
theory (Krishnamurthy and Taylor, 1985a). Each phase is split into two parts: the bulk
phase; and the film in which the reaction and heat and mass transfer occur. In the bulk
Figure 4.2: A schematic of the two-film model of a column stage. δ V and δ L represent
the thickness of the gas and liquid films, respectively, P V , P I , and P L are the pressure
in the bulk vapour phase, at the gas-liquid interface and in the bulk liquid, respectively.
As can be seen an isobaric profile is assumed. The temperatures of the bulk vapour
phase, at the gas-liquid interface and of the bulk liquid are denoted by T V , T I , and T L ,
respectively. Finally, yi and xi are the mole fractions of component i in the bulk vapour
and liquid phases, respectively, and yiI and xIi are the mole fractions of component i at
the vapour-liquid interface in the vapour and the liquid phases, respectively.
phase, the concentration, pressure, and temperature are assumed to be uniform, whereas
there could be a gradient in composition and temperature in the film. A liquid–gas
interface between the two films is sketched in Figure 4.2, where the profiles for composition
of component i, temperature and pressure are represented.
The following assumptions are made:
• The model is steady state.
• The bulk phases and films are at chemical equilibrium everywhere (i.e., the reac-
tion rates are faster than the mass transfer rates) (Laddha and Danckwerts, 1981;
Blauwhoff et al., 1984).
• The interface is at phase and chemical equilibrium (Economou and Donohue, 1991).
• The interfacial surface area is the same for both heat and mass transfer.
• The effective area is equal to the wetted area.
• The absorption column is considered to be adiabatic.
• There is no pressure drop along the column.
4.3 Model equations
In this subsection, we present the equations used to model a packed column subsection
(stage). The equations are grouped into different categories: energy and mass balances,
rate equations, equilibrium equations, mass-transfer correlations, heat-transfer correla-
tions, and diffusion correlation. We chose to list explicitly all of the equation of the model
due to the radically different approach followed here. One important model parameter, τ ,
is introduced in this section to account for the scaling of the CO2 diffusivity. This param-
eter will be estimated from experimental data. In the following equations, the subscript
i refers to the ith component and j to the j th stage. c is the total number of components
and Ns is the number of stages in the column. It is fixed to 50 which is amply sufficient
to assure numerical convergence without significantly impacting the computational time.
Heat and mass balances
The liquid and vapour phase mole balances for component i of stage j are given by
L
Lj−1 xi,j−1 + Ni,j = Lj xi,j i = 1, 2..., c; j = 1, 2..., Ns (4.1)
V
Vj+1 yi,j+1 − Ni,j = Vj yi,j i = 1, 2..., c; j = 1, 2..., Ns (4.2)
where Lj and Vj are the total liquid and vapour molar flowrate leaving stage j respectively
(mol s−1 ), xi,j and yi,j are the bulk liquid and vapour mole fractions of component i of
L
stage j (-), Ni,j is the net gain of species i in the liquid phase due to interphase transport
V
for stage j (mol s−1 ), and Ni,j is the net loss of species i in the vapour phase due to
interphase transport for stage j (mol s−1 ).
The mole fractions in the streams leaving each stage must sum to unity:
c
X c
X
xi,j = 1 and yi,j = 1 j = 1, 2..., Ns (4.3)
i=1 i=1
The energy balances for the liquid and vapour phases respectively are
L L
Lj−1 Hj−1 + NCO2 ,j
δHCO2 + EjL = Lj HjL j = 1, 2..., Ns (4.4)
V
Vj+1 Hj+1 − EjV = Vj HjV j = 1, 2..., Ns (4.5)
where HjL and HjV are the enthalpies of the liquid and gas phase respectively of stage j
(J mol−1 ), δHCO2 is the term correcting the enthalpy of absorption of CO2 (J mol−1 ) as
introduced in section 3.1. EjL is the net gain of energy of the liquid phase through the
interface at stage j (W), and EjV is the net loss of energy from the vapour phase through
the interface at stage j (W). HjL and HjV are determined from SAFT-VR as functions of
the temperature of the bulk phases, TjL and TjV (K), the molar volumes of the phases, VjL
and VjV (cm3 mol−1 ), and the composition of the phases, ~xj and ~yj .
There is no accumulation of mass or energy at the vapour–liquid interface:
L V
Ni,j = Ni,j i = 1, 2..., c; j = 1, 2..., Ns (4.6)
EjL = EjV i = 1, 2..., c; j = 1, 2..., Ns (4.7)

Rate equations: mass transfer
According to the film model for mass transfer, we define the mass transfer rates as follows:
L L I,L L
Ni,j = ki,j aT,j (Ci,j − Ci,j ) i = 1, 2..., c; j = 1, 2..., Ns (4.8)
V V V I,V
Ni,j = ki,j aT,j (Ci,j − Ci,j ) i = 1, 2..., c; j = 1, 2..., Ns (4.9)
where aT,j is the total interfacial area on stage j available for heat or mass transfer (m2 ),
L V
ki,j is the liquid-phase mass-transfer coefficient of stage j for component i (m s−1 ), ki,j
L
is the vapour-phase mass-transfer coefficient of stage j for component i (m s−1 ), Ci,j and
V
Ci,j are the concentrations of component i in the bulk liquid and vapour phases of stage j
I,L I,V
(mol m−3 ), and Ci,j and Ci,j are the concentrations of component i at the vapour-liquid
interface in the liquid and vapour phases of stage j (mol m−3 ).
The total area for heat and mass transfer is given by the product of the effective specific
area and the stage volume:
aT,j = a′j Asection Dz j = 1, 2..., Ns (4.10)
where a′j is the interfacial area density of stage j (m2 m−3 ), Asection is the cross sectional
area of the column (m2 ), and Dz is the stage height (m) given by hpacking /Ns , with hpacking
as the total packing height (m).
Rate equations: heat transfer
The energy fluxes in the liquid and gas phases at stage j are given by the sum of the
conductive and diffusive heat fluxes:
EjL = QLcond,j + QLdif f,j j = 1, 2..., Ns (4.11)

EjV = QVcond,j + QVdif f,j j = 1, 2..., Ns (4.12)
where QLcond,j and QLdif f,j are the conductive and diffusive heat fluxes at the interface into
the liquid phase of stage j (W), and QVcond,j and QVdif f,j are the conductive and diffusive
heat fluxes at the interface out of the vapour phase of stage j (W).
The conductive heat transfer rate into the liquid phase and out of the gas phase of stage j
are driven by temperature gradients between the bulk phases and the interface and given
by
QLcond,j = hLT,j aT,j (TjI − TjL ) j = 1, 2..., Ns (4.13)
QVcond,j = hVT,j aT,j (TjV − TjI ) j = 1, 2..., Ns (4.14)
where hLT,j and hVT,j are the heat transfer coefficients in the liquid and gas phases of stage
j, respectively (W m−2 K−1 ) , and TjI is the temperature at the vapour-liquid interface
of stage j (K).
The diffusive heat fluxes account for the variation in enthalpy in the liquid and gas streams
associated with the transfer of mass from one phase to another. This is obtained based on
the individual component fluxes and the corresponding enthalpies of the pure components:
c
X
QLdif f,j = L
Ni,j H(TjL , Vi,j
∗L
, ~zi ) j = 1, 2..., Ns (4.15)
i=1
Xc
QVdif f,j = V
Ni,j H(TjV , Vi,j
∗V
, ~zi ) j = 1, 2..., Ns (4.16)
i=1
where H(TjL , Vi,j

∗L
, ~zi ) and H(TjV , Vi,j
∗V
, ~zi ) are the enthalpies of a pure component in the
bulk liquid and gas phase, respectively (J mol−1 ). Strictly speaking, one needs to use par-
tial molar enthalpies instead of a pure component enthalpies however, to keep the model
simple this assumption has been deemed acceptable. ~zi denotes the molar composition of
a stream consisting of pure component i defined by zk,i = 1 if i = k and zk,i = 0 otherwise.
∗L ∗V
Vi,j and Vi,j are the molar volumes of pure component i in stage j in the bulk liquid and
gas phases respectively (cm3 mol−1 ) given by
PjL = P (TjL , Vi,j

∗L
, ~zi ) i = 1, 2..., c; j = 1, 2..., Ns (4.17)
PjV = P (TjV , Vi,j

∗V
, ~zi ) i = 1, 2..., c; j = 1, 2..., Ns (4.18)
where PjL and PjV are the pressure of the bulk liquid and gas phases of stage j respectively
(MPa).
When equating the vapour and liquid fluxes, one obtains an expression which includes
the difference between the enthalpies in the vapour phase and in the liquid phase for a
pure component i; this is the enthalpy of vaporisation of this component. As we will see,
the enthalpy of vaporisation of water plays a significant role in the process. We assume
that the mass transfer of N2 between the two phases is small and that the enthalpy of
vaporisation of CO2 is negligible as the operating conditions are close or above the critical
point. In the case where only one phase is stable for a pure component at the conditions of
interest, as is typically the case for CO2 , one can access the hypothetical phase information
by providing an adequate initial guess for the molar volume to the SAFT-VR equation of
state.
Equilibrium relations
Both chemical and phase equilibrium are assumed to prevail at the interface so that the
conditions of equality of pressure, temperature, and chemical potential must be satisfied:
µ(TjI , VjI,V , ~yjI ) = µ(TjI , VjI,L , ~xIj ) = µ

~ I,L
j ~ I,V
=µ j j = 1, 2..., Ns (4.19)
P (TjI , VjI,V , ~yjI ) = P (TjI , VjI,L , ~xIj ) = PjI j = 1, 2..., Ns (4.20)
~ I,L
where µ j ~ I,V
and µ j are the vector of chemical potential on stage j in the liquid and gas
phases respectively (J mol−1 ), VjI,L and VjI,V are the molar volumes of the liquid phase
and the gas phases at the vapour-liquid interface of stage j (cm3 mol−1 ), and PjI is the
pressure at the liquid-vapour interface of stage j (MPa). The chemical potentials and the
molar volume are determined using the SAFT-VR foreign object.
The mole fractions at the interface must sum to unity:
c
X c
X
xIi,j =1 and I
yi,j =1 j = 1, 2..., Ns (4.21)
i=1 i=1
Each stage is assumed to be at mechanical equilibrium, so that
PjV = PjL = PjI j = 1, 2..., Ns (4.22)
where
PjV = P (TjV , VjV , ~yj ) j = 1, 2..., Ns (4.23)
PjL = P (TjL , VjL , ~xj ) j = 1, 2..., Ns (4.24)
Finally, the column pressure drop is assumed to be negligible
V
Pj−1 = PjV j = 2..., Ns (4.25)
This assumption can easily be removed at a later stage of the model development. Prelim-
inary results have indicated that the pressure drop has a negligible effect on the process
used for the validation of our model.
The equations presented in the remainder of this section apply to each stage. The sub-
script j has been omitted for the purpose of clarity.
Mass transfer correlations
The interfacial area, the liquid phase mass-transfer coefficient and the gas phase mass-
transfer coefficient are obtained using the correlations from Onda et al. (1968a,b). Other
correlations exist, the most commonly used being the ones developed by Rocha et al.
(1996) and Billet and Schultes (1999). Faramarzi et al. (2010) compared performances of
the correlations and found that all three correlations provide a good description. They
expressed a slight preference for the correlations of Rocha et al. (1996) but could not draw
a general conclusion which is applicable to all operating conditions. The correlations of
Rocha et al. (1996) apply to structured packing and have to be adapted for use in random
packing, whereas those developed by Onda et al. (1968a,b) have been developed specifi-
cally using random Berl saddle packing, the same as the one employed in the pilot plant by
Tontiwachwuthikul et al. (1992). This last set of correlations is sometimes reparametrised
when used in commercial software like Aspen or ProTreat. The correlations developed by
Onda et al. (1968a,b), as reported by Treybal (1981), are implemented in our model of
the absorber column. The liquid phase mass transfer coefficient for component i is given
by (Treybal, 1981)
1/3
ηLg

′
kiL = 0.0051 (ReL )2/3 (ScLi )−1/2 (ap Lp )0.4 i = 1, 2..., c (4.26)
ρL
where ρL is the density of the liquid phase (kg m−3 ), η L is the dynamic viscosity of the liq-
′
uid phase (kg m−1 s−1 ), g is the gravitational acceleration (m s−2 ), ReL is the liquid-phase
Reynolds number based on the interfacial area, ScL is the Schmidt number of component
i in the liquid phase, ap is the specific surface area of the packing (m2 m−3 ), and Lp is
the nominal packing size (m).
The Reynolds and the Schmidt numbers for component i in the liquid phase are
′ ρ L uL
ReL = (4.27)
a′ η L
ηL

ScLi = i = 1, 2..., c (4.28)
ρL DiL
where DiL is the diffusion coefficient of component i in the liquid phase (m2 s−1 ), uL is
the liquid velocity (m s−1 ), and a′ is the interfacial area density defined in equation (4.32).
The vapour-phase mass-transfer coefficient is obtained from
0.7 1/3
kiV = 2ap DiV ReV ScVi (ap Lp )−2 i = 1, 2..., c (4.29)
where DiV is the diffusivity of component i in the vapour-phase (m2 s−1 ), ReV is the
vapour-phase Reynolds number, and ScVi is the Schmidt number of component i in the
vapour phase.
The Reynolds and the Schmidt numbers for the vapour phase are
ρ V uV
ReV = (4.30)
ap η V
ηV

ScVi = i = 1, 2..., c (4.31)
ρV DiV
where ρV is the density of the vapour phase (kg m−3 ), η V is the dynamic viscosity of the
vapour phase (kg m−1 s−1 ), and uV is the vapour velocity (m s−1 ).
The interfacial area density a′ is obtained from (Onda et al., 1968a,b)
a′ σc
= 1 − exp(−1.45( )0.75 (ReL )0.1 (F rL )−0.05 (W eL )0.2 ) (4.32)
ap σ
where σc is the critical surface tension of the packing material (N m−1 ), σ is the vapour-
liquid surface tension (N m−1 ), ReL is the liquid-phase Reynolds number based on the
specific surface area, F rL is the liquid-phase Froude number, and W eL is the liquid-
phase Weber number. The interfacial area density, together with the three dimensionless
numbers used in equation 4.32, are properties of the mixture, not properties of individual
components.
The expressions of the dimensionless numbers are
ρ L uL
ReL = (4.33)
η L ap
2
L a p uL
Fr = (4.34)
g
2
ρ L uL
W eL = (4.35)
ap σ
The velocities of each phase are
v̇ L
uL = (4.36)
Asection
v̇ V
uV = (4.37)
Asection
where v̇ L and v̇ V are the volumetric flowrate of the liquid and gas phase respectively (m3
s−1 ).
Heat transfer correlations
For the gas and liquid heat-transfer coefficients the correlations presented in Treybal’s
Mass Transfer Operations (Treybal, 1981) are used. Once again these expressions apply
to each stage. The expression for the liquid-phase heat-transfer coefficient is
0.45
hLT ds

ds Lspec
= 25.1 (P rL )0.45 (4.38)
λLT ηL
where hLt is the liquid-phase heat-transfer coefficient (W m−2 K−1 ), ds is the diameter of
a sphere of the same surface area as a single packing particle (not the same as ap ) (m),
λLT is the liquid thermal conductivity (W m−1 K−1 ), Lspec is the specific liquid flowrate
(kg s−1 ), and P rL is the liquid phase Prandtl number. Its expression is given by
CpL η L
P rL = (4.39)
λLT
where CpL is the specific heat capacity of the liquid phase (J kg−1 K−1 ). The vapour-phase
heat-transfer coefficient is given by
−0.36
hVT

2/3 ds Vspec
V
P rV = 1.195 V
(4.40)
Cp Vspec η (1 − ǫL0 )
where hVT is the vapour-phase heat transfer coefficient (W m−2 K−1 ), CpV is the specific
heat capacity of the vapour phase (J kg−1 K−1 ), P rV is the Prandtl number for the gas
phase, Vspec is the specific vapour flowrate (kg s−1 ), and ǫL0 is the operating void space
in the packing. It is assumed to be equal to the void fraction ǫ. The Prandtl number for
the gas phase is
CpV η V
P rV = (4.41)
λVT
where λVT is the vapour thermal conductivity (W m−1 K−1 ).
Vapour-phase diffusion coefficient
The vapour-phase diffusion coefficient used in equation (6.5) is obtained from the Fuller
expression reported by Poling et al. (2001). It is assumed that the components are dif-
fusing through air (Poling et al., 2001).
1.75
0.00143T V
DiV = 10 −4
1/2 1/3 1/3
i = 1, 2..., c (4.42)
10P V Mi−air [Σv i + Σv air ]
where Σv is the atomic diffusion volume (Å), and Mi−air is given by
Mi−air = 2[(1/Mair ) + (1/Mi )]−1 i = 1, 2..., c (4.43)

where Mi is the molecular weight of component i (g mol−1 ).
Liquid-phase diffusion coefficient
The liquid-phase binary diffusion coefficients are obtained through different correlations.
The mutual diffusion coefficients (cm2 s−1 ) of CO2 at very low concentration in pure MEA
and pure H2 O are derived from the Wilke-Chang correlation (Poling et al., 2001):
◦ 7.4 × 10−8 (φk Mk )1/2 T L

DCO2 ,k
= k = H2 O, MEA (4.44)
ηkL (VCO
m 0.6
2
)
where φk is the ”association factor” of solvent k (-), and Vim is the molar volume of
component i at its normal boiling temperature (cm3 mol−1 ).
These mutual diffusion coefficients are used in the Takahashi correlation (Takahashi et al.,
1982) to determine the diffusivity of CO2 in a liquid mixture of H2 O and MEA:
1/3 1/3
VL ηkL
X
L −4 ◦
DCO = τ 10 xk DCO 2 ,k
(4.45)
2
ηL k=H2 O,MEA
Vk∗L
where we have introduced the scaling factor τ to the correlation. The factor 10−4 is
converting the diffusion coefficient from cm2 s−1 to m2 s−1 . Its value is discussed in
section 5. A graph depicting the variation of the diffusion coefficient of CO2 in the liquid
phase with regard to the temperature is presented in Figure 4.3.
Finally, the Perkins and Geankoplis method (Perkins and Geankoplis, 1969) is used to
calculate the diffusivity of H2 O, MEA and N2 in the liquid phase:
1/2
7.4 × 10−8 (φM )i T
DiL = 10−4 i = H2 O, MEA, N2 (4.46)
1000η L (Vim )0.6
where the term (φM )i represents the “association factor” of component i in the liquid
(a)
1e-08
8e-09
DCO (m s )
2 -1
6e-09
4e-09
2
2e-09
0
280 300 320 340 360 380
T (K)
(b)
5e-10
4e-10
DCO (m s )
2 -1
3e-10
2e-10
2
1e-10
0
280 300 320 340 360 380
T (K)
Figure 4.3: Variation of the diffusion coefficient of CO2 in the liquid phase with regard
to the temperature. (a) Results obtained without scaling the diffusivity (τ = 1), and (b)
results obtained by scaling diffusivity to 3.7% of its original value (τ = 0.037).
Table 4.1: Correlations used in the model.
Physical property Component Source

H2 O Westmeier (1977)
Liquid phase viscosity MEA Leibush and Shorina (1947)
Mixture Maham et al. (2002)
Vapour phase viscosity Kestin et al. (1966)
Surface tension Vazquez et al. (1997)
Heat conductivity Haynes (2011)
H2 O Laliberte (2009)
Liquid heat capacity
MEA Mundhwa and Henni (2007)
H2 O Riddick et al. (1986)
Vapour heat capacity N2 Vargaftik (1972)
CO2 Bender et al. (1981)
mixture (g mol−1 ) and is derived from
n
X
(φM )i = x j φ j Mj i = H2 O, MEA, N2 (4.47)
j=1
j6=i
The other correlations of experimental data used in our model are listed table 4.1.
In contrast with other CO2 -MEA absorber models, this model does not uses enhancement
factor, defined as the ratio of the amount of gas absorbed in a reacting liquid to the amount
which would be absorbed if there were no reaction (Danckwerts, 1970; DeCoursey, 1982;
van Swaaij and Versteeg, 1992) It is generally accepted that for absorption within the
fast pseudo first-order reaction regime, as is the case for CO2 absorption with MEA
(Laddha and Danckwerts, 1981; Blauwhoff et al., 1984) the enhancement factor could be
approximated by the Hatta number (Danckwerts, 1970; Kucka et al., 2003b; Tobiesen
et al., 2007) and is much greater than one. This means that the enhancement factor is
increasing the value of the mass transfer coefficient and by doing so, favouring the mass
transfer. It should be noted that the enhancement factor is not the same as the CO2
diffusivity scaling parameter τ introduced in this chapter. They do not relate to the same
physical properties as they stem from fundamentally differential approach towards the
modelling of the chemical reactions. The enhancement factor is introduced in a modelling

approach using an explicit treatment of the chemical reactions whereas τ is introduced in
an implicit reaction modelling approach.
4.4 Conluding remarks
A rate-based CO2 absorber model incorporating a SAFT-VR thermodynamic treatment

has been developed. The model combines physical and chemical interactions within the
consistent thermodynamic framework (SAFT-VR) introduced in chapter 3, under the
assumption that the chemical reactions are at equilibrium. The chemical reactions are
treated implicitly in this formalism. Thus there is no need to incorporate an enhancement
factor or to use experimental data on the rate of reaction. This greatly simplifies the
description and provides a considerably more transferable model. At this point, it is
more transferable in principle. Chapter 5 explores whether it is valid. Its transferability
to other solvents is left for future work. A simple scaling of the CO2 diffusivity in the
liquid phase is proposed to account for the formation of carbamate. The correction of the
enthalpy of absorption of CO2 introduced in chapter 3 has been integrated in the energy
balance. In the next chapter, this model is validated against pilot plant data and the
model parameter τ , accounting for the scaling of the CO2 diffusivity in the liquid phase,
is adjusted using pilot plant data.
Chapter 5
Absorber model results
5.1 Validation of the model
The absorber model developed in this study is validated by adjusting the model parameter
τ , as it appears in equation (4.45), to a small subset of the pilot plant experimental data
obtained by Tontiwachwuthikul et al. (1992). These include measurements from one run
at different points along the column of the temperature in the liquid phase, the mole
fraction of CO2 in the vapour phase, and the CO2 loading in the liquid phase, defined
as θCO2 = xCO2 /xMEA , which have been successfully modelled by other authors (Kucka
et al., 2003a; Gabrielsen et al., 2006; Faramarzi et al., 2010; Mac Dowell, 2010; Mac Dowell
et al., 2013). The absorption column has an internal diameter of 0.1 m and a total packed
height of 6.6 m; care has to be taken not to mistake the total column height of 7.2 m
with the packed height, which is the relevant dimension in modelling the absorber. The
absorption column internals are randomly packed 12.7 mm ceramic Berl saddles. The gas
inlet stream is assumed to be free of MEA and the liquid solvent inlet stream free of N2 .
The input values used are listed in Table 5.1 and the characteristics of the column and
the packing are listed in Table 5.2.
87
5. Absorber model results
Table 5.1: Inputs to absorber column model for runs T13 to T22 (Tontiwachwuthikul et al., 1992).
Input variable T13 T14 T15 T16 T17 T18 T19 T20 T21 T22
Inlet gas flow (mol.m−2 .s−1 ) 17.54 17.54 18.4 17.51 17.54 18.29 16.72 18.32 13.72 18.3
Inlet gas pressure (MPa) 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
Inlet gas temperature (K) 292.15 292.15 292.15 292.15 293.15 293.15 292.15 292.15 292.15 292.15
Inlet gas CO2 mole fraction 0.153 0.156 0.195 0.155 0.156 0.191 0.115 0.192 0.191 0.191
Inlet gas H2 O mass fraction 0.0058 0.0058 0.0058 0.0058 0.0058 0.0058 0.0058 0.0058 0.0058 0.0058
Inlet liquid flow (m3 .m−2 .h−1 ) 13.5 13.5 13.5 9.5 13.5 9.5 13.5 9.5 9.5 9.5
Inlet liquid temperature (K) 292.15 292.15 292.15 292.15 293.15 293.15 292.15 292.15 292.15 292.15
Inlet MEA concentration (kmol.m−3 ) 2.00 2.00 2.03 2.08 3.08 2.00 2.00 2.55 2.00 3.00
Lean loading, θCO2 10−4 0.118 10−4 10−4 0.237 10−4 10−4 10−4 10−4 10−4
88
5. Absorber model results 89
Table 5.2: Characteristics of the column and the packing
Name Value Source

Asection (m2 ) 0.00785 Tontiwachwuthikul et al. (1992)
ap (m2 .m−3 ) 466 Treybal (1981)
σc (N m−1) 61 Perry and Green, 2008, pg. 18-34, table 18-11
Lp (m) 0.0127 Tontiwachwuthikul et al. (1992)
ǫ 0.63 Treybal, 1981, pg. 198, table 6.3
Dz (m) 0.132 Tontiwachwuthikul et al. (1992)
Ns 50
ds (m) 0.31622 Treybal, 1981, pg. 206, table 6.5
Some of the data needed for the simulation of the process were not explicitly reported by
Tontiwachwuthikul et al. (1992): in particular, the temperature of the flue gas; and the
quantity of water in the flue gas. Close inspection of an earlier paper (Tontiwachwuthikul
et al., 1989) detailing the experimental apparatus reveals that the temperature of the flue
gas is controlled to be that of the lean solvent with a thermostatic bath. The inlet vapour
stream is composed of ambient air and bottled CO2 . Typical values of relative humidity of
air lie between 20% to 70%, which, at 20◦ C and 1 bar, corresponds to an amount of H2 O
in the flue gas between ωHV 2 O ≃ 0.0025 and ωHV 2 O ≃ 0.01, where ωiV is the mass fraction of
component i in the vapour phase.
A sensitivity analysis of the extent of humidity indicates that the variation of ωHV 2 O be-
tween 0.0025 and 0.01 has a negligible impact on the temperature and composition profiles.
The effect of the amount of water in the flue gas is discussed in more detail in the next
section. For our working model, the amount of water in the inlet flue gas is fixed to ωHV 2 O
= 0.0058, corresponding to 50% humidity at 20◦ C and 1 bar.
The validation is performed using the operating conditions of Run T22 (Tontiwach-
wuthikul et al., 1992). In preliminary calculations with unadjusted model parameters
(i.e., τ =1), the rate of absorption of CO2 along the column was found to be over-predicted.
In order to investigate the nature of the over-prediction of the rate of absorption, a sen-
sitivity analysis is performed and the most influential parameter for the description of
the composition profile is found to be the diffusivity of CO2 in the liquid phase. In our
reaction-implicit model, the amount of CO2 in any phase represents a combination of

free CO2 , bicarbonate and carbamate. The relative amount of each component can be
extracted from the SAFT-VR treatment of the association aggregates (Rodrı́guez et al.,
2012), although this information is not used explicitly in the process model. In practice,
the CO2 in the liquid phase is mostly carbamate, a much larger molecular species than
CO2 , which is expected to have a reduced diffusion in the medium compared to free CO2 .
The CO2 diffusivity in the liquid phase is therefore scaled down to account for the high
proportion of carbamate; a good estimate of the rate of absorption of CO2 is thereby
obtained. The extent of rescaling can be estimated from the experimental data obtained
from Run T22 (Tontiwachwuthikul et al., 1992) for the gas and liquid composition pro-
files. On this basis, the value of the liquid-phase CO2 diffusivity in our model that results
in the minimum deviation with the pilot plant data corresponds to a scaling to 3.7% of its
original value (i.e., τ = 0.037). The scaling parameter τ is related to mass transfer effects
so it was deemed appropriate to favour the composition profile over the temperature for
the determination of its value. Han et al. (2013) recently presented a molecular dynamics
study of diffusion of CO2 in pure MEA, where CO2 is treated as a combination of free
CO2 and CO2 -containing aggregates. They suggested a reduction of the liquid CO2 dif-
fusion coefficient by a factor of 0.025. Although no water was included in this theoretical
investigation, leading to a comparatively high liquid viscosity, it is interesting to note that
the relative change in diffusivity was found to be commensurate with that propsed here.
The liquid-phase temperature, gas-phase CO2 concentration, and liquid-phase CO2 con-
centration profiles for Run T22, determined with and without a scaling of the diffusivity,
are presented in Figure 5.1.
Because the parameter τ is estimated from pilot plant data the value found for τ may
be correcting any potential errors in the chosen mass transfer correlation rather than the
diffusivity only. To the best of our knowledge, there is no published experimental data
for the diffusion of carbamate in MEA and H2 O mixtures. In order to assess the validity
of the proposed scaling and the impact of choice of the mass-transfer correlation on its
value, we implemented the mass-transfer correlations of Rocha et al. (1993, 1996) in our
(a))
330
320
TL (K)
310
300
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.1: Comparison of the pilot plant data (squares) (Tontiwachwuthikul et al., 1992)
and the result of our model (curves) for Run T22 for the absorption of CO2 in an 18wt%
solution of MEA. The dashed curves represent the results obtained without scaling the
diffusivity (τ = 1) and the continuous curves the results obtained by scaling the CO2
diffusivity in the liquid phase to 3.7% of its original value (τ = 0.037). a) Temperature
profile for the liquid phase, b) gas-phase CO2 concentration profile, and c) liquid-phase
CO2 loading. Stage 50 corresponds to the bottom of the column.
current model of the absorber with the same scaling of the diffusivity (τ = 0.037). The
results for the liquid-phase temperature, gas-phase CO2 concentration, and liquid-phase
CO2 concentration profiles for Run T22 with the two different mass transfer correlations
are compared in Figure 5.2. It can be seen that the profiles obtained are very similar:
with the correlation of Rocha et al. (1993, 1996) there is a slight underestimation of the
composition of CO2 in the liquid and gas phases, but the temperature profile obtained
with the correlation of Rocha et al. (1993, 1996) is more faithful to the pilot plant data
than that of Onda et al. (1968a,b). In conclusion, the value of τ is essentially found to
be independent of the choice of the mass-transfer correlation, confirming our hypothesis
that a scaling of the diffusivity is required.
(a))
330
320
TL (K)
310
300
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.2: Sensitivity analysis of the mass transfer correlation on the profiles predicted
with our model compared with the experimental pilot plant data of Run T22 (Tontiwach-
wuthikul et al., 1992): Mass-transfer correlations from Onda et al. (1968a,b) (continuous
curve), mass-transfer correlations from Rocha et al. (1993, 1996). (a) Temperature profile
for the liquid phase, (b) gas phase CO2 concentration profile, and (c) liquid phase CO2
loading.
5.2 Model predictions
All the runs using MEA from the pilot plant absorber are simulated using the diffusivity
parameter (τ =0.037) obtained during the validation step (previous section). The predic-
tions for the temperature profile in the liquid phase, gas phase CO2 concentration, and
liquid phase CO2 concentration for the different Runs T13 to T21 (Tontiwachwuthikul
et al., 1992) are presented in Figures 5.3 to 5.11. These runs represent a wide range of
operating conditions in terms of amine concentration, inlet liquid phase CO2 loading, inlet
gas phase CO2 concentration, and gas-to-liquid flow-rate ratio (see Table 5.1).
(a))
320
315
310
TL (K)
305
300
295
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.3: Comparison of the pilot plant data (squares) and the predictions with our
model (continuous curves) for run T13 taken from the work of Tontiwachwuthikul et al.
(1992). (a) Temperature profile for the liquid phase, (b) gas phase CO2 concentration
profile, and (c) liquid phase CO2 loading.
(a))
320
315
310
TL (K)
305
300
295
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.4: Comparison between pilot plant data (squares) and models predictions (con-
tinuous curves) for run T14 taken from the work of (Tontiwachwuthikul et al., 1992). (a)
Temperature profile for the liquid phase, (b) gas phase CO2 concentration profile and, (c)
liquid phase CO2 loading.
(a))
320
315
310
TL (K)
305
300
295
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
tinuous curves) for run T15 taken from the work of (Tontiwachwuthikul et al., 1992). (a)
(a))
320
315
310
TL (K)
305
300
295
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
tinuous curves) for run T16 taken from the work of Tontiwachwuthikul et al. (1992). (a)
(a))
315
310
TL (K)
305
300
295
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
(a))
330
325
320
TL (K)
315
310
305
300
295
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.6
0.5
2
0.4
θCO
0.3
0.2
0.1
0
0 10 20 30 40 50
Stage
(a))
320
315
310
TL (K)
305
300
295
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
(a))
330
325
320
TL (K)
315
310
305
300
295
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
2
0.3
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Good agreement is found between the model predictions and the pilot plant data for
Runs T13, T14, T16, T17, T19, and T20 in relation to the liquid-phase temperature
and the liquid- and gas-phase composition profiles along the entire length of the column.
This demonstrates that the scaling procedure can be applied in a transferable manner to
different operating conditions. There is a slight over-prediction of the temperature and
the composition profile for runs T15 and T21, though the compositions at the top and
the bottom of the column are accurately described. The outlet liquid temperature in run
T18 is overestimated by about 12K. There is a good match between the model predictions
and the experimental values for the composition profiles. Although there is a slight over
prediction in Run T16, a good representation of the temperature bulges in Runs T16 and
T20 can be seen in terms of location along the column, and the amplitude is accurately
predicted in run T20. To the best of our knowledge, no other published models provide
a description of the temperature bulge to this level of accuracy (Kucka et al., 2003a;
Gabrielsen et al., 2006; Faramarzi et al., 2010; Mac Dowell et al., 2013). In keeping with
our objective to develop a predictive model to support solvent design activities, the CO2
diffusivity scaling parameter τ is not readjusted for each of the pilot plant runs. This is
in stark contrast with models based on an enhancement factor (van Swaaij and Versteeg,
1992), where the model has to be adjusted for the different operating conditions. The
behaviour of runs T15, T18 and T21 could be represented more accurately if we undertook
such a refinement.
In all runs but T18, flat profiles are found toward the top of the column. This plateau
means that a maximum in the absorption is reached, meaning that equilibrium is achieved
not only at the liquid-phase interface but also between the bulk liquid and the bulk vapour
phases. As a results the profiles would not be affected by making the column any higher.
This plateau is depicted in Figure 5.12 for run T22. It can be seen that the temperature
of the liquid, the vapour phase, and the vapour-liquid interface are all equal between
stages 1 and 15 (i.e., the top of the column). Similarly, the CO2 gas phase composition in
the bulk vapour phase and at the vapour interface are equal, as are the CO2 loadings in
the liquid phase and at the interface between stages 1 and 15. Additionally, the profiles
(a))
320
315
310
TL (K)
305
300
295
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
for the temperature of the liquid phase and the temperature at the interface are almost
identical, suggesting that all the heat-transfer resistance is in the vapour film.
In this set of runs, T18 stands out as it does not exhibit a plateau in the profiles as
found for the other operating conditions, meaning that the whole length of the column is
used for absorption. This is illustrated in Figure 5.13. The conditions for equilibrium are
briefly met at the bottom of the column (large values of the stages) for the compositions
profiles but there is no clear levelling off of the profiles. From Table 5.1, it appears that
Run T18 has the lowest amount of MEA in the lean solvent and the highest amount of
CO2 in the flue gas compared to the other runs. This explains why the whole column is
used for absorption. In the other runs, the totality of CO2 is absorbed as the gas travels
upwards between stages 50 and 15. So the lean solvent flowing at the top of the absorber
(stages 1 to 15) has nothing to absorb. Run T18 is the only run where there is still CO2
in the gas stream leaving the absorber, so the lean solvent starts absorbing CO2 as soon
as it enters the column at the top.
Overall the model exhibits very good prediction capabilities over a very wide range of
operating conditions.
(a))
335
330
325
320
TL (K)
315
310
305
300
295
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
2
θCO
0.3
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.12: Run T22 of the pilot plant data from Tontiwachwuthikul et al. (1992).
Comparison of the predictions of our model for: (a) temperature of the bulk liquid phase
(continuous curve), the temperature at the vapour-liquid interface (dashed curve) and
the temperature for the bulk vapour phase (dot-dashed curve); (b) the gas phase CO2
concentration in the bulk vapour phase (continuous curve) and at the interface (dashed
curve); (c) the liquid phase CO2 loading in the bulk liquid phase (continuous curve) and
at the interface (dashed curve).
(a))
335
330
325
320
TL (K)
315
310
305
300
295
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
2
θCO
0.3
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.13: Run T18 of the pilot plant data from Tontiwachwuthikul et al. (1992).
Comparison of the predictions of our model for: (a) temperature of the bulk liquid phase
(continuous curve), the temperature at the vapour-liquid interface (dashed curve) and
the temperature for the bulk vapour phase (dot-dashed curve); (b) the gas phase CO2
concentration in the bulk vapour phase (continuous curve) and at the interface (dashed
curve); (c) the liquid phase CO2 loading in the bulk liquid phase (continuous curve) and
at the interface (dashed curve).
5.3 Sensitivity analysis
The relative importance of several mass- and heat-transfer parameter is now assessed
using a sensitivity analysis. Different key parameters in the model are altered to observe
their impact on the predicted profiles. The values tested are listed in Table 5.3.
5.3.1 Mass transfer
The mass transfer in the model is mediated via mass-transfer coefficients (Onda et al.,
1968a,b). These coefficients are highly dependent on the viscosity, the diffusivity, and
the surface tension of the fluid considered; these three properties are thus considered as
key parameters in the model and their influence on the temperature and composition
profiles are assessed. The value of each parameter in each phase is doubled and halved
in order to determine the effect on the temperature and composition profiles. Varying
the diffusivity or viscosity in the gas phase has no visible effect on the liquid temperature
profile or the composition profiles. However, varying these parameters in the liquid phase
results in a significant variation in the profiles. The effect on the profiles due to the
variation of the liquid viscosity and the diffusivity in the liquid phase is represented in
Figures 5.14 and 5.15. Varying the surface tension also produces a significant variation
in the profiles as can be seen in Figure 5.16. The data that were originally considered
for the surface tension correlation in our model is for a mixture of only H2 O and MEA
(Vazquez et al., 1997). Jayarathna et al. (2013) published very recently experimental data
of the surface tension of liquid mixtures of H2 O, MEA and CO2 , finding an increase in
the vapour-liquid interfacial tension on absorption of CO2 . A new correlation from these
data is also implemented in our process model. The comparisons of the temperature and
composition profiles obtained with the two different correlations is shown in Figure 5.16.
The difference in the profiles is negligible so changing the surface tension correlation is
not considered necessary.
Table 5.3: Sensitivity analysis of key parameters in the model
Parameter description Variation Effect

Vapour diffusivity doubled negligible effect
halved
Liquid diffusivity doubled significant effect

halved see figure 5.14
Vapour viscosity doubled negligible effect

halved
Liquid viscosity doubled significant effect

Surface tension doubled significant effect

Heat transfer liquid × 10 negligible effect

÷10
Heat transfer vapour × 10 limited effect

÷10 see figure 5.18
Enthalpy of absorption doubled significant effect

Heat of vaporisation of water doubled significant effect

It should be noted that the values of the profiles at the outlets are not affected by the
variations of the surface tension or the viscosity and diffusivity in the liquid phase. It
would be safe to assume that under these operating conditions, where the residence time
is large enough for the system to reach equilibrium, the outlet values are determined
solely by thermodynamic quantities while the composition profiles are mostly dependent
on mass transfer.
5.3.2 Heat transfer
The major source of heat in the absorption process is a consequence of the exothermic
reaction between MEA and CO2 , which occurs in the liquid phase. This heat can then
be transferred to the vapour phase. Another major element in the energy balance is the
heat associated with the transfer of water from one phase to the other. In the model,
the heat transfer is mediated via the heat-transfer coefficients (Treybal, 1981). Four key
thermal parameters can be identified in the model: the heat of absorption of CO2 , the
heat-transfer coefficient for the liquid phase, the heat-transfer coefficient for the gas phase,
and the heat of vaporisation of H2 O. Kvamsdal and Rochelle (2008) and Faramarzi et al.
(2010) identified the heat capacities as sensitive parameters for the temperature profiles,
this is due to the fact that in their model, the heat capacities are included in the energy
balance as they are used explicitly to determine the enthalpies. This is not the case in
our model as the enthalpies are derived from the SAFT-VR equation of state. Although
the heat capacities are used inside SAFT-VR to determine the ideal contribution toward
the enthalpies, they appear explicitly in the gPROMS model only in the heat-transfer
coefficient correlations, hence the choice of these coefficients as key parameters. The values
of the four parameters (the heat of absorption of CO2 , the heat transfer coefficient for the
liquid phase, the heat transfer coefficient for the gas phase and the heat of vaporisation of
H2 O) are hence varied to observe the impact on the liquid phase temperature profile and
(a))
330
320
TL (K)
310
300
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.14: Sensitivity analysis of the CO2 diffusivity in the liquid phase on the pro-
files predicted with our model in terms of the scaling parameter τ compared with the
experimental pilot plant data of Run T22 (Tontiwachwuthikul et al., 1992). τ = 0.037
(continuous curve), τ = 0.074 (dashed curve), τ = 0.0185 (dot-dashed curve). (a) Tem-
perature profile for the liquid phase, (b) gas phase CO2 concentration profile, and (c)
(a))
330
320
TL (K)
310
300
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.15: Sensitivity analysis of the liquid viscosity on the profiles predicted with our
model compared with the experimental pilot plant data of Run T22 (Tontiwachwuthikul
et al., 1992): Nominal value (continuous curve), viscosity doubled (dashed curve), viscosity
halved (dot-dashed curve). (a) Temperature profile for the liquid phase, (b) gas phase
CO2 concentration profile, and (c) liquid phase CO2 loading.
(a))
330
320
TL (K)
310
300
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.16: Sensitivity analysis of the vapour-liquid surface tension on the profiles pre-
dicted with our model compared with the experimental pilot plant data of Run T22 (Ton-
tiwachwuthikul et al., 1992): Nominal value (continuous curve) (Vazquez et al., 1997),
surface tension doubled (dashed curve), surface tension halved (dot-dashed curve), surface
tension correlation from data from Jayarathna et al. (2013) (dotted curve). (a) Temper-
ature profile for the liquid phase, (b) gas phase CO2 concentration profile, and (c) liquid
phase CO2 loading.
the composition profiles. For our particular model, the sensitivity of the heat of absorption
of CO2 can be assessed by varying the value of the correction factor δHCO2 . Although the
heat of absorption of CO2 and the heat of vaporisation of water are known thermodynamic
quantities, their variation provides insights into the behaviour of the column. Their values
were doubled and halved, and the values of the heat transfer coefficients were multiplied
by 10 and divided by 10 (Table 5.3). The variation of the liquid heat-transfer coefficient
had no visible effect on the liquid-phase temperature and composition profile. The effects
on the profiles resulting from varying the correction factor of the heat of absorption, the
vapour heat-transfer coefficient, and the heat of vaporisation of water are presented in
Figures 5.17, 5.18 and 5.19, respectively.
The first observation that can be made from these graphs is that varying the thermal
parameters does not have a significant impact on the composition profiles. The effect is
visible only on the liquid-phase temperature profile. The heat of absorption of CO2 does
have a significant effect on the outlet temperature value. However, its variation has a
negligible impact on the amplitude of the temperature bulge, as can be seen in Figure
5.17. The effect of the variation of the vapour-phase heat-transfer coefficient is illustrated
in Figure 5.18. Despite a variation of one order of magnitude in the parameter value,
the effect on the temperature profile is very limited. Furthermore, the effect due to the
variation of the liquid-phase heat-transfer coefficient is equally negligible. It can be con-
cluded that for the operating conditions under consideration, the model is insensitive to
the heat-transfer coefficients. This is why we have not refined the heat-transfer correla-
tions in our current study.
The value of heat of vaporisation of H2 O is known experimentally but the sensitivity anal-
ysis allows to investigate the importance of water phase changes. Its variation is found
to have a significant impact on the liquid phase temperature profile and a lesser impact
on the composition profiles; it has virtually no effect on the end-point values. The effect
on the temperature profile is concentrated on the temperature bulge. An increase in the
(a))
350
340
330
TL (K)
320
310
300
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.17: Sensitivity analysis of the heat of absorption of CO2 on the profiles predicted
with our model in terms of the correction factor δHCO2 compared with the experimental
pilot plant data of Run T22 (Tontiwachwuthikul et al., 1992). δHCO2 = 50 kJ mol−1 (con-
tinuous curve), δHCO2 = 100 kJ mol−1 (dashed curve), δHCO2 = 25 kJ mol−1 (dot-dashed
curve). (a) Temperature profile for the liquid phase, (b) gas phase CO2 concentration
profile, and (c) liquid phase CO2 loading.
(a))
330
320
TL (K)
310
300
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.18: Sensitivity analysis of the vapour-phase heat-transfer coefficient on the pro-
files predicted with our model compared with the experimental pilot plant data of Run
T22 (Tontiwachwuthikul et al., 1992). Nominal value (continuous curve), coefficient value
increased ten times (dashed curve), coefficient value reduced ten times (dot-dashed curve).
(a) Temperature profile for the liquid phase, (b) gas phase CO2 concentration profile, and
(c) liquid phase CO2 loading.
(a))
340
330
TL (K)
320
310
300
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.19: Sensitivity analysis of the enthalpy of vaporisation of water on the profiles
predicted with our model compared with the experimental pilot plant data of Run T22
(Tontiwachwuthikul et al., 1992). Nominal value (continuous curve), enthalpy doubled
(dashed curve), enthalpy halved (dot-dashed curve). (a) Temperature profile for the liquid
phase, (b) gas phase CO2 concentration profile, and (c) liquid phase CO2 loading.
heat of vaporisation magnifies the amplitude of the bulge, whereas a reduction eliminates
the bulge. This suggests that the heat of vaporisation of H2 O is responsible for the rate
at which the liquid heats up and cools down. The key elements to predict the tempera-
ture bulge accurately are therefore the enthalpy of absorption of CO2 and the enthalpy
of vaporisation of water. The absorption of CO2 releases energy that heats both phases.
As the gas temperature increases upon entering the column, water is transferred from the
liquid phase to the gas phase through evaporation to maintain saturation; this results in
an absorption of energy, and a consequent decrease in temperature.
In conclusion, the competition between those two thermal effects (absorption of CO2 ,
increases temperature and vaporisation of water, decreases temperature) are responsible
for the observed and predicted temperature bulge. This interpretation is consistent with
that reported in Mac Dowell et al. (2013).
5.3.3 Extent of water in the flue gas
The amount of water in the flue gas is considered as important in determining the position
of the temperature bulge in the absorber (Kvamsdal and Rochelle, 2008; Kvamsdal et al.,
2010). In the pilot plants of Tontiwachwuthikul et al. (1992) and Dugas (2006), the inlet
flue gas is not saturated in water, while the flue gas is saturated in the pilot plants of
Tobiesen et al. (2007) and Gabrielsen et al. (2007). In order to assess the impact of
humidity in our model of the absorber, we have varied the amount of water in the inlet
flue gas from yH2 O = 0 to yH2 O = 0.08 (above saturation). The effect on the profiles is
represented in Figure 5.20. It is clear from the figure that varying the amount of water in
the flue gas does not have a significant impact on the composition profiles. The effect is
visible only for the liquid-phase temperature profile, where the outlet temperature value
is higher by approximately 5 K for the saturated gas yH2 O = 0.08 than for the dry gas
yH2 O = 0. This is in contrast to the work of Mac Dowell et al. (2013) who found a large
(a))
350
340
330
TL (K)
320
310
300
290
0 10 20 30 40 50
Stage
(b))
0.2
0.15
2
yCO
0.1
0.05
0
0 10 20 30 40 50
Stage
(c))
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 5.20: Sensitivity analysis of the amount of water in the inlet flue gas on the profiles
predicted with our model compared with the experimental pilot plant data of Run T22
(Tontiwachwuthikul et al., 1992). Nominal value (continuous curve), yH2 O = 0.08 (dashed
curve), yH2 O = 0 (dot-dashed curve). (a) Temperature profile for the liquid phase, (b) gas
phase CO2 concentration profile, and (c) liquid phase CO2 loading.
effect of the humidity of the inlet gas on the profiles in the mass transfer zone.
5.4 Conluding remarks
The performances of the absorber model developed in chapter 3 and 4 have been assessed.
Model validation has been carried out by comparison with published experimental pilot
plant data (Tontiwachwuthikul et al., 1992). A central finding of our current study is that
diffusivity of CO2 in the liquid phase required adjustment in order to provide an accurate
representation of the column profiles; this was done based on one pilot plant run, and used
transferably to successfully model all other runs for different operating conditions. Very
good predictions are obtained for the liquid-phase temperature profiles and liquid- and
gas-phase compositions along the column. This highlights the importance of capturing
the physical and chemical equilibria reliably; in addition the assumption that the process
is not limited by reaction kinetics appears to be valid.
With a careful sensitivity analysis we show that the liquid viscosity and diffusivity, and
vapour-liquid surface tension, are key properties for the prediction of the composition
profiles. The heat generated by the absorption of CO2 affects the magnitude of the tem-
perature variation in the absorber and the vaporisation of water affects the presence and
amplitude of the temperature bulge.
The transferability of this model to other operating conditions, more specifically its ap-
plicability to the desorption process is discussed in the next two chapters.
Chapter 6
Model of the regeneration module
6.1 Introduction
The solvent based carbon capture processes consist of two main parts: an absorber where
the CO2 is transferred from the gas phase to the liquid phase and reacts with the solvent
(here MEA) to form a carbamate, and a solvent regeneration part where the rich solvent
is heated up to break this carbamate and transfer the CO2 to the gas phase. The lean
solvent is then recycled back to the absorber. This second part is composed of a desorber
column, similar to the absorber column, a reboiler at the bottom of the desorber to pro-
vide the heat required for the separation of the CO2 from the solvent, and a condenser at
the top of the desorber to condense the water back to the process and ensure a quasi-pure
gas stream of CO2 . The heat requirement of the reboiler is the source of the high en-
ergy penalty in solvent-based carbon capture processes. The decarbonisation of a flue gas
produced by a coal-fired power plant can represent an extra consumption of fuel of 20%
(Parson and Keith, 1998). The energy penalty can be reduced in a number of ways, either
by modifying the topology of the process to allow heat integration (Mimura et al., 1997;
Desideri and Paolucci, 1999; Romeo et al., 2008; Jassim and Rochelle, 2006; Harkin et al.,
2010; Ahn et al., 2013) or by changing the solvent (Mangalapally et al., 2009; Salkuyeh
122
6. Model of the regeneration module 123
and Mofarahi, 2012; Barzagli et al., 2012). Although some of these changes can reduce
the energy penalty by up to 50%, there is still a need to reduce it further.
A number of models have been proposed for the amine-based CO2 desorption. Most mod-
els are similar to those developed for the absorber: the heat- and mass-transfer models
are identical, the thermodynamic model is similar but requires different parameter values
based on experimental data for the temperature-dependent ones. These aspects have been
covered in chapter 2. One important aspect of a desorber modelling is the treatment of
the chemical reactions. In a desorber, the relevant temperature range is larger and there
are limited data available for the rates of the reactions of interest at these temperatures
so in some models, the chemical reactions are assumed to be instantaneous, i.e. chemical
equilibrium prevails everywhere in the liquid phase, where the reactions are assumed to
take place (Tobiesen et al., 2008; Lawal et al., 2010). If one wants to take into account the
reaction kinetics then new rates of reactions are needed; this is the route followed by Greer
et al. (2010); Plaza et al. (2010b); Gáspár and Cormo (2011) based on Aboudheir et al.
(2003) experimental data. In all these models, the reboiler and condenser are modelled
as equilibrium stages.
Tobiesen et al. (2008) validated their model against the CO2 loading at the bottom of
the desorber of their own pilot plant (Tobiesen et al., 2007, 2008). Their predictions ex-
hibit an over prediction at higher loading. Although they reported experimental data for
temperatures and flowrates in the flowsheet, they did not report any simulation results
of these values.
Greer et al. (2010) did not validate their process model on pilot plant data.
Plaza et al. (2010b) validated their model using only one run from the pilot plant of
Dugas (2006). Their model shows a good agreement for the temperature profile along
the desorber column. However, due to the unreliability in some of the experimental data,
they had to adjust some of them in order to close mass and energy balances and match
CO2 removal. An interfacial area factor was also introduced as an adjustable parameter
in this model to correct the calculated interfacial area. Due to all these adjustments and
the single run used for the validation, the predictive capabilities of their model is not
demonstrated.
Lawal et al. (2010) also validated their results on the pilot plant from Dugas (2006). They
presented two temperature profiles for two different runs with acceptable agreement of
the trend between the model predictions and the experimental data. However, actual
temperature values have a maximum error of about 5 K. There are no results on the CO2
loading.
Gáspár and Cormo (2011) validated their model against pilot plant data from Tobiesen
et al. (2007, 2008). They obtained a good agreement between the simulated values and
the experimental values of the flowrate of the stream of CO2 leaving the system. However,
they simulated only 13 of the 19 runs performed by Tobiesen et al. (2008) and they did
not validated the thermal predictions from their model.
Although it is fair to say that desorber validation is especially challenging because of the
data uncertainties (Dugas, 2006; Tobiesen et al., 2008), no study to date has presented a
comprehensive validation of an amine-based CO2 desorber including compositions, tem-
peratures and flowrates. This is the ambition of the model presented in this chapter and
validated in the next chapter, using pilot plant data from Tobiesen et al. (2008).
The regeneration module is composed of three units: a desorber, a condenser and a
reboiler. The models developed in this chapter are adapted from the absorber model
presented earlier. These models can be divided in two parts: a thermodynamic model, in
which the SAFT-VR equation of state is used, and a mass transfer model. The thermo-
dynamic model used is identical to the one used for the absorber. The desorber column is
described via rate-based stages with a different mass transfer correlation to model the dif-
ferent type of packing. The condenser and the reboiler are modelled as equilibrium stages.
Figure 6.1: Diagram of the desorber
6.2 Model of the desorber column
The model developed for the absorber is used for the desorber. However, in relation to
specific pilot plant data, this model should be modified to be applicable to structured
packing. The only change required are in the correlations for the mass transfer. They
are presented in this section. A schematic of the desorber is presented in Figure 6.1.
The interfacial area, the liquid phase mass transfer coefficient and the gas phase mass
transfer coefficient are obtained using the correlations from Rocha et al. (1996). These
correlations have been developed specifically for structured packing, as used in the pilot
plant by Tobiesen et al. (2008). The liquid phase mass transfer coefficient for component
i is given by
1/2
Ce DiL uLe

kiL =2 i = 1, 2..., c (6.1)
πS
where Ce is a factor slightly less than unity to account for those parts of the packed
bed that do not encourage rapid surface renewal (-), DiL is the diffusion coefficient of
component i in the liquid phase (m2 s−1 ), uLe is the effective liquid-phase velocity (m s−1 ),
S is the side dimension of corrugation (m).
The effective liquid-phase velocity is defined by
uL
uLe = (6.2)
εht sin θ
where uL is the superficial liquid velocity (m s−1 ), ε is the packing void fraction (-), ht is
the liquid holdup (-), and θ is the angle with respect to the horizontal for falling film or
corrugation channel.
The vapour phase mass transfer coefficient is obtained from
0.8
DV ρV S|uLe + uVe |

0.33
kiV = 0.054 i ScVi
!
i = 1, 2..., c (6.3)
S ηV
where DiV is the diffusion coefficient of component i in the vapour phase (m2 s−1 ), ρV is
the density of the vapour phase (kg m−3 ), uVe is the effective vapour-phase velocity (m
s−1 ), η V is the dynamic viscosity of the vapour phase (kg m−1 s−1 ), ScVi is the Schmidt
number of component i in the vapour phase.
The effective vapour-phase velocity is defined by
uV
uVe = (6.4)
ε(1 − ht ) sin θ
where uV is the superficial vapour velocity (m s−1 ). The Schmidt number for the vapour
phase is
ηV

ScVi = i = 1, 2..., c (6.5)
ρV DiV
The interfacial area density a′ is obtained from the third part of the correlations from
Rocha et al. (1996)
a′ 29.12(W eL F rL )0.15 S 0.359

= FSE (6.6)
ap ReL 0.2 ε0.6 (1 − 0.93 cos γ)(sin θ)0.3
where the factor FSE accounts for the variations in surface enhancements (lancing, fluting,
etc.) and the contact angle γ accounts for surface wettability. ReL is the liquid-phase
Reynolds number based on the specific surface area, F rL is the liquid-phase Froude num-
ber, and W eL is the liquid-phase Weber number. The interfacial area density, together
with the three dimensionless numbers used in this expression, are not component depen-
dent.
The equations for the dimensionless numbers are:
ρ L uL
ReL = (6.7)
η L ap
2
L a p uL
Fr = (6.8)
g
2
ρ L uL
W eL = (6.9)
ap σ
where uL is the liquid superficial velocity (m s−1 ), ρL is the density of the liquid phase
(kg m−3 ) and η L is the dynamic viscosity of the liquid phase (kg m−1 s−1 ).
The superficial velocities of each phase are
v̇ L
uL = (6.10)
Asection
v̇ V
uV = (6.11)
Asection
where v̇ L and v̇ V are the volumetric flowrates of the liquid and gas phase respectively (m3
s−1 ), and Asection is the sectional area of the column (m2 ).
The holdup expression is:
2/3 1/3
3µL uL

4Ft
ht = (6.12)
S ρL εgef f sin θ
Figure 6.2: Diagram of the reboiler
where Ft stands for:
a′
Ft = (6.13)
ap FSE
and gef f is the effective gravity defined by:
ρL − ρV
gef f = g (6.14)
ρL
where g is the gravitational acceleration (m s−2 ).
6.3 Reboiler
The role of the reboiler is to heat up the incoming liquid solvent and vaporise the CO2
and H2 O to obtain a lean stream of solvent. The prevailing effect is the phase separation
and not the mass transfer as is the case in the desorber, hence a rate-based description
is not necessary. The reboiler is modelled as an equilibrium stage with a liquid inlet, a
liquid outlet and a vapour outlet (cf. Figure 6.2).
The component mass balance is:
Lin in out out out out

r xr,i = Lr xr,i + Vr yr,i i = 1, 2..., c (6.15)
where Lin −1 out

r is the liquid molar flowrate entering the reboiler (mol s ), Lr and Vrout are
the liquid and vapour molar flowrates leaving the reboiler respectively (mol s−1 ), xin
r,i is
the molar fraction of component i in the liquid stream entering the reboiler and xout
r,i and
out
yr,i are the molar fractions of component i in the liquid and the vapour streams leaving
the reboiler, respectively.
The mole fractions in the streams leaving the reboiler must sum to unity:
c
X c
X
xout
r,i =1 and out
yr,i =1 (6.16)
i=1 i=1
The net gain of species i in the liquid phase due to interphase transport in the reboiler
L
Nr,i (mol s−1 ) is given by:
L
Nr,i out
= −Vrout yr,i i = 1, 2..., c (6.17)
The enthalpy balance for the reboiler is given by:
Lin L,in
r Hr
L
+ Qr + Nr,CO2
δHCO2 (Tr ) = Lout
r Hr
L,out
+ Vrout HrV,out (6.18)
where HrL,in is the enthalpy of the liquid phase entering the reboiler (J mol−1 ), HrL,out and
HrV,out are the enthalpies of the liquid and vapour phase leaving the reboiler respectively (J
mol−1 ), and Qr is the reboiler heat duty (W). The reboiler operates at high temperature
(around 393 K), so the enthalpy of absorption of CO2 as defined in chapter 3 takes the
value of δHCO2 (Tr ) = 80 kJ mol−1 .
Both chemical and phase equilibrium are assumed to prevail in the reboiler. Thus, the
conditions of equality of pressure, temperature and chemical potential must be satisfied:
µLr,i = µVr,i i = 1, 2..., c (6.19)
and only one temperature, Trout and one pressure, Prout are defined in the reboiler. µLr,i
and µVr,i are the chemical potential of component i in the reboiler in the liquid and gas
Figure 6.3: Diagram of the condenser
phases respectively (J mol−1 ), obtained using the SAFT-VR foreign object:
~ Lr = µ
µ ~ (Trout , VrL,out , ~xout
r ) (6.20)
µ ~ (Trout , VrV,out , ~yrout )

~ Vr = µ (6.21)
where VrL,out and VrV,out are the molar volumes of the liquid phase and the gas phases,
respectively in the reboiler (m3 kmol−1 ), also obtained via SAFT-VR:
Prout = P (TrL,out , VrL,out , ~xout

r ) (6.22)
Prout = P (TrV,out , VrV,out , ~yrout ) (6.23)
6.4 Condenser
The condenser is modelled as an equilibrium stage with only a vapour inlet, a vapour
outlet and a liquid outlet (cf. Figure 6.3).
The component mole balance is:
Vcin xin out out out out

c,i = Lc xc,i + Vc yc,i i = 1, 2..., c (6.24)
where Vcin is the vapour molar flowrate entering the condenser (mol s−1 ), Lout
c and Vcout
are the liquid and vapour molar flowrates leaving the condenser respectively (mol s−1 ),
xin
c,i is the molar fraction of component i in the liquid stream entering the condenser and
xout out
c,i and yc,i are the molar fractions of component i of the liquid and the vapour streams
leaving the condenser, respectively.

The mole fractions in the streams leaving the condenser must sum to unity:
c
X c
X
xout
c,i =1 and out
yc,i =1 (6.25)
i=1 i=1
The net gain of species i in the liquid phase due to interphase transport in the condenser
L
Nc,i (mol s−1 ) is given by:
L
Nc,i = Lout out
c xc,i i = 1, 2..., c (6.26)
The enthalpy balance for the condenser is given by:
Vcin HcV,in + Qc + Nc,CO

L
2
δHCO2 (Tr ) = Lout
c Hc
L,out
+ Vcout HcV,out (6.27)
where HcV,in is the enthalpy of the vapour phase entering the condenser (J mol−1 ), HcL,out
and Hcv,out are the enthalpies of the liquid and vapour phase leaving the condenser re-
spectively (J mol−1 ) and Qc is the condenser heat duty (W). The condenser operates at
both high and low temperature (around 393 K and 313 K), so assigning the value the
enthalpy of absorption of CO2 as defined in chapter 3 based on the temperature does not
make sense here. However, there is almost no MEA in the condenser so the enthalpy of
absorption of CO2 is not a sensitive parameter here. We chose to assign it the value of
δHCO2 (Tr ) = 80 kJ mol−1 , corresponding to the higher temperature (393 K).
Both chemical and phase equilibrium are assumed to prevail in the condenser. Thus, the
conditions of equality of pressure, temperature and chemical potential must be satisfied:
µLc,i = µVc,i i = 1, 2..., c (6.28)
and only one temperature, Tcout and one pressure, Pcout are defined in the condenser. µLc,i
and µVc,i are the chemical potentials of component i in the condenser in the liquid and gas
phases respectively (J mol−1 ), obtained using the SAFT-VR foreign object:
~ Lc = µ(Tcout , VcL,out , ~xout

µ c ) (6.29)
~ Vc = µ(Tcout , VcV,out , ~ycout )

µ (6.30)
where VcL,out and VcV,out are the molar volumes of the liquid phase and the gas phases,
respectively, in the condenser (m3 kmol−1 ), also obtained via SAFT-VR:
Pcout = P (TcL,out , VcL,out , ~xout

c ) (6.31)
Pcout = P (TcV,out , VcV,out , ~ycout ) (6.32)
The model for the solvent regeneration unit presented in this chapter is developed with
transferability as the key requirement. The desorber column model is transferred directly
from the absorber. It uses the same thermodynamic model and the same approach to-
wards heat and mass transfer. The scaling in the diffusivity of CO2 introduced in chapter
4 is conserved intact. The condenser and reboiler are developed using the same ther-
modynamic model and the heat- and mass-transfer are simplified as these process units
are considered to be equilibrium stages. The predictive capabilities of this model are
presented in the next chapter.
Chapter 7
Regeneration module results
The scope of this chapter is to assess the transferability of the absorber model developed
in chapter 3 and 4 and adapted to the desorption conditions in chapter 6. The predictive
capabilities of this solvent regeneration model are assessed using extensive pilot plant
data.
7.1 Model predictions
The desorber model developed in the current work is validated by comparing the model
predictions with the pilot plant experimental data obtained by Tobiesen et al. (2008).
These include measurements of the temperature and the CO2 loading in the liquid phase
at the bottom of the desorber and at the reboiler outlet, and flowrate of the liquid and gas
streams leaving the condenser. The experimental results for the loading at the bottom
of the desorber have been successfully reproduced by the model developed by Tobiesen
et al. (2008). The experimental values of the flowrate of the gas stream of CO2 leaving
the condenser have been successfully reproduced by the model developed by Gáspár and
Cormo (2011) for 13 out of 19 runs. New pilot plant data have been recently published by
133
7. Regeneration module results 134
Notz et al. (2012). They presented detailed experimental pilot plant data for a complete
absorber-desorber process. However, only two runs are presented and the topology used
is slightly more complex due to the addition of a washer section at the top of the desorber
column. Moreover, the development and validation of the desorber model presented in
our study was already well under way at the time of their publication. These are the
reasons why the work from Notz et al. (2012) has not been chosen for the validation of
the desorber model.
The topology of the regeneration system including the desorber, the reboiler and the
condenser is represented in Figure 7.1. It should be noted that in this particular system,
the liquid condensate is fed back directly to the reboiler. In most similar systems, the
condensate is fed back to the top of the desorber (Dugas, 2006; Notz et al., 2012). The
desorption column has an internal diameter of 0.1 m and a total packed height of 3.89
m. The desorption column internals are structured packing Mellapack 250Y. The input
values used are listed in Table 5.1 and the characteristics of the column and the packing
are listed in Table 5.2. The outputs of the simulation are the temperature and the CO2
loading at the bottom of the desorber column, the temperature and the CO2 loading at
the reboiler outlet, the flowrate of condensed water leaving the condenser and the flowrate
of the CO2 gas stream leaving the condenser. This model is implemented and simulated
using the gPROMS software (Process Systems Enterprise Ltd.).
All runs have been successfully simulated. The predictions for the temperature and
liquid phase CO2 loading at the bottom of the desorber and at the reboiler outlet, and
the flowrates of the liquid and gas streams out of the condenser are compared with the
experimental values from Tobiesen et al. (2008) in Figure 7.2. These runs represent a wide
range of operating conditions in terms of flowrates, liquid compositions, heat duties, and
temperatures (cf. Table 7.1). Good agreement is found between the model predictions
and the pilot plant data in relation to the temperature and the loading of the desorber
and the reboiler, and for the liquid flowrate out of the condenser. Although there is
a slight over-prediction of the CO2 loading and the temperature in both process units,
these deviations may be explained on the basis of the experimental measurements error
Table 7.1: Inputs for the simulated runs for the desorber model
Run Rich solvent Desorber Rich solvent MEA Rich solvent Reboiler Reboiler Condenser Condenser
flowrate pressure temperature concentration CO2 loading duty pressure temperature pressure
(m3 h−1 ) (MPa) (◦ C) (kmol m−3 ) (kW) (MPa) (◦ C) (MPa)
1 0.54 0.198 115 5.1 0.316 11.6 0.198 15 0.1965
2 0.24 0.197 115 5.0 0.315 11.6 0.197 13 0.1955
7. Regeneration module results
3 0.24 0.198 115 4.9 0.313 11.5 0.198 15 0.1965

4 0.24 0.195 115 4.9 0.309 11.6 0.195 15 0.1935
5 0.24 0.2065 116 4.9 0.264 13.5 0.2065 15 0.205
6 0.36 0.2135 115 4.6 0.365 11.6 0.2165 16 0.212
7 0.36 0.202 115 4.7 0.295 11.8 0.202 18 0.2005
8 0.36 0.217 118 4.7 0.312 13.8 0.217 19 0.2155
9 0.18 0.2046 115 5.0 0.329 9.6 0.2046 12 0.2036
10 0.18 0.2036 115 5.0 0.338 9.7 0.2036 12 0.2026
11 0.18 0.2070 112 5.0 0.396 5.8 0.207 11 0.206
12 0.36 0.205 113 5.0 0.392 11.6 0.205 11 0.203
13 0.18 0.2094 106 5.1 0.445 3.9 0.2094 12 0.2084
14 0.54 0.2154 109 5.1 0.457 11.6 0.2154 22 0.2124
15 0.36 0.2021 103 5.1 0.45 5.8 0.2021 12 0.2001
16 0.54 0.2012 104 5.1 0.451 7.9 0.2012 15 0.1992
17 0.54 0.2058 111 5.1 0.429 13.4 0.2058 31 0.2038
18 0.54 0.2047 113 5.2 0.407 11.7 0.2047 19 0.2027
19 0.54 0.2067 116 5.2 0.350 13.6 0.2067 22 0.2047
135
Figure 7.1: Topology of the solvent regeneration unit
Table 7.2: Desorber column properties
Name Value Source

Dsection (m) 0.1 Tobiesen et al. (2008)
θ π/4 Rocha et al. (1996)
Fse 350 Rocha et al. (1996)
Ce 0.9 Rocha et al. (1996)
ap (m2 m−3 ) 250 Rocha et al. (1996)
ε 0.95 Rocha et al. (1996)
described by Tobiesen et al. (2008). In relation to the CO2 loading, they explained that
it was extremely hard to obtain sufficient accuracy in the desorber outlet loading reading
to justify a detailed comparison with simulated values. Errors in the loading have a very
strong effect on the flowrate of CO2 leaving the condenser. For example, in run 11, the
measured loading varies from 0.392 at the top of the desorber to 0.291 at the reboiler
outlet. However, the simulated value for the loading at the reboiler outlet is 0.335, which
represent a 15% error. This error is directly repercuted on the CO2 flowrate leaving the
condenser which is simulated at 4.5 kg h−1 instead of 8.3 kg h−1 as measured. This explains
the deviation between simulated results and experimental data observed in Figure 7.2.
This error magnification was highlighted by Tobiesen et al. (2008) and they concluded
that comparing simulated results of the flowrate of gaseous CO2 leaving the condenser
to experimental data was irrelevant. The same approach is followed here. Hence, these
measurements are not used to assess the predictive capabilities of the model.
With regard to the temperature measurements, the exact position of the temperature
probe at the bottom of the desorber or in the reboiler has a significant impact on the
temperature reading and can lead to an error in the reboiler temperature reading of
about 2-2.5 K and to an even higher error in the desorber temperature reading (Tobiesen
et al., 2008). The error bars for the temperatures in figure 7.2 represent only a 2.5 K
error. The desorber temperature probe is placed at the bottom of the column, lower than
the bottom of the packing, this difference between the total height of the column (4.1 m)
and the total packing height (3.89 m) is not taken into account in the model.
Despite the uncertainties, this demonstrates that the model developed in our current work
is suitable for the operating conditions under which a CO2 desorber operates, and that
the scaling procedure devised under the absorber operating conditions can be applied in
a transferable manner to different operating conditions.
(a) (b)
0.5 0.5
0.45
0.4 0.4
Sim
Sim
0.35
2
2
θCO
θCO
0.3 0.3
0.25
0.2 0.2
0.15
0.2 0.3 0.4 0.5 0.2 0.3 0.4 0.5
Exp Exp
θCO θCO
2 2
(c) (d)
130 130
TSim ( C)
TSim ( C)
120 120
o
110 110
100 100
100 110 120 130 100 110 120 130
o o
TExp ( C) TExp ( C)
(e) (f)
Simulated Rate (kg/h)
12
14
12 10
10 8
8 6
6
4
4
2 2
0 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12
Experimental Rate (kg/h) Experimental Rate (kg/h)
Figure 7.2: Comparison between the model predictions and the experimental data from
Tobiesen et al. (2008) (a) Loading at the bottom of the desorber, (b) loading at the
reboiler outlet, (c) temperature at the bottom of the desorber, (d) temperature at the
reboiler outlet, (e) condensate flowrate out of the condenser and, (f) CO2 flowrate out of
the condenser. This last value is not used to assess the validity of the model (see text).
7.2 Sensitivity analysis
The relative importance of several mass- and heat-transfer parameter is now assessed
using a sensitivity analysis. Different key values in the model are altered to observe
their impact on the predicted profiles along the column of the temperature in the liquid
phase, the mole fraction of CO2 in the vapour phase and the CO2 loading in the liquid
phase. The impact is also observed on the temperature and CO2 loading at the bottom
of the desorber column and in the reboiler, and the flowrates of the streams leaving the
condenser. The values tested are listed in Table 7.3.
7.2.1 Mass transfer
As explained in the absorber model, the mass transfer in the desorber is mediated via
mass transfer coefficients (Rocha et al., 1996). These coefficients are highly dependent on
the viscosity, the diffusivity and the surface tension of the fluid considered; thus, these
three properties are considered as key parameters in the model and their influence on the
temperature and composition profiles are therefore assessed. The value of each parameter
in each phase was doubled and halved in order to determine their effect on the tempera-
ture and composition profiles. As obtained in the absorber model, varying the diffusivity
or viscosity in the gas phase has no visible effect on the liquid temperature profile or the
composition profiles. However, neither does the variation of the diffusivity of CO2 , the
viscosity of the liquid phase or the surface tension of the liquid phase. We explain this
result by fact that the differences in composition between the bulk phase and the film
for the gas and the liquid phases are very low, as presented in Figure 7.3. Hence, this
counteracts any variation in the mass transfer coefficients.
The mass transfer in the desorber is very limited. For example on run 12, the loading
varies from 0.392 to 0.382. This is to be compared with the values in the absorber, where
the loading varies from 0 to 0.5.
Table 7.3: Sensitivity analysis of key parameters in the model
Parameter description Variation Effect

Vapour diffusivity doubled negligible effect
halved
Liquid diffusivity doubled negligible effect

halved
Vapour viscosity doubled negligible effect

halved
Liquid viscosity doubled negligible effect

halved
Surface tension doubled negligible effect

halved
Heat transfer liquid × 10 negligible effect

÷10
Heat transfer vapour × 10 negligible effect

÷10
Enthalpy of absorption doubled limited effect

of CO2 halved see Figures 7.4 and 7.5
Heat of vaporisation of water doubled limited effect

halved see Figures 7.6 and 7.7
(a) )
400
395
TL (K)
390
385
380
0 10 20 30 40 50
Stage
(b) )
0.4
0.3
2
yCO
0.2
0.1
0
0 10 20 30 40 50
Stage
(c) )
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 7.3: Run 12 of the pilot plant data from Tobiesen et al. (2008). Comparison of the
predictions of our model for: (a) temperature of the bulk liquid phase (continuous curve),
the temperature at the vapour-liquid interface (dashed curve) and the temperature for the
bulk vapour phase (dot-dashed curve); (b) the gas phase CO2 concentration in the bulk
vapour phase (continuous curve) and at the interface (dashed curve); (c) the liquid phase
CO2 loading in the bulk liquid phase (continuous curve) and at the interface (dashed
curve).
The loading is further reduced in the reboiler from 0.382 to 0.335, indicating that the
majority of the mass transfer between phases takes place in the reboiler, where the mass
transfer coefficients do not have any impact as the reboiler is modelled as an equilibrium
stage.
In accordance with the results on the profiles, the variation of the three mass transfer
properties (viscosity, diffusivity and surface tension) has no noticeable effect on the tem-
perature and CO2 loading at the bottom of the desorber column and in the reboiler, and
the flowrates of the streams leaving the condenser.
7.2.2 Heat transfer
The major heat consumption in the system is the reverse endothermic reaction to sep-
arate MEA and CO2 , which occurs in the liquid phase. This separation needs a high
temperature to take place, hence the need for a reboiler. There can be heat transfer
between the liquid phase and the gas phase. In the model, the heat transfer is mediated
via the heat transfer coefficients (Treybal, 1981). Four key parameters can be identified
in the model: the heat of absorption of CO2 , the heat of vaporisation of water, the heat
transfer coefficient for the liquid phase and the heat transfer coefficient for the gas phase.
The values of these four parameters are varied to observe the impact on the liquid phase
temperature profile and the composition profiles. For this particular model, the sensitiv-
ity of the heat of absorption of CO2 was assessed by varying the value of the correction
parameter δHCO2 of the enthalpy of absorption of CO2 . Although the heat of absorption
of CO2 and the heat of vaporisation of water are known thermodynamic quantities, their
variation provides insights into the behaviour of the column. Their values were doubled
and halved as before and the values of the heat transfer coefficients were multiplied by 10
and divided by 10 (Table 7.3).
The variation of the liquid and vapour heat transfer coefficients has no visible effect on
the liquid phase temperature and composition profiles.

The variation of the heat of absorption of CO2 has no visible impact on the desorber profile
but only on the liquid outlet temperature as can be seen in Figure 7.4. This contrasts with
the absorber model where this parameter was a key factor in predicting the temperature
profile of the column but its qualitative effect is similar, as its impact is mostly on the
outlet liquid temperature at the bottom of the column. The effect of the variation of the
heat of vaporisation on the temperature and liquid phase CO2 loading at the bottom of
the desorber and at the reboiler outlet, and the flowrates out of the condenser for selected
runs is represented in Figure 7.5. These runs were selected based on the loading of the rich
solvent fed to the desorber. Tobiesen et al. (2008) used three loading ranges during the
pilot plant campaign: low CO2 loading, θCO2 = 0.26-0.32, medium CO2 loading, θCO2 =
0.33-0.40, and high CO2 loading, θCO2 = 0.41-0.46. Two runs of each loading range have
been selected. The heat of absorption of CO2 has a very limited effect on the temperature
and loading at the bottom of the desorber and at the reboiler outlet. However, it has a
noticeable impact on the flowrate of the streams leaving the condenser.
The variation of the heat of vaporisation of water has a very limited impact on the
liquid phase temperature profile and the gas phase composition profile, and a negligible
impact on the CO2 liquid loading profile as represented in Figure 7.6. Halving the heat of
vaporisation of water reduces the outlet temperature of the desorber by only about 2-3 K.
The effect of the variation of the enthalpy of vaporisation of water on the temperature and
the CO2 loading at the bottom of the desorber and in the reboiler, and on the flowrates
of water and CO2 leaving the condenser for several selected runs is represented in Figure
7.7. While the limited impact on the CO2 liquid loading is confirmed for the desorber,
there is a significant impact on the loading in the reboiler for the lower loading runs. The
negligible impact on the temperature is confirmed for the different runs. The impact is
significant for the flowrate of water from the condenser, but it is almost negligible for the
lower flowrate while doubling it or halving it at higher flowrates. The variation of the
enthalpy of vaporisation of water has a consistent effect on the CO2 leaving the condenser:
the lower the heat of vaporisation, the higher the flowrate and the better the match be-
(a) )
400
395
TL (K)
390
385
380
0 10 20 30 40 50
Stage
(b) )
0.4
0.3
2
yCO
0.2
0.1
0
0 10 20 30 40 50
Stage
(c) )
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 7.4: Sensitivity analysis of the heat of absorption of CO2 on the profiles predicted
with our model in terms of the correction factor δHCO2 compared with the experimental
pilot plant data of Run 12 from Tobiesen et al. (2008). δHCO2 = 80 kJ mol−1 (continuous
curve), δHCO2 = 160 kJ mol−1 (dashed curve), δHCO2 = 40 kJ mol−1 (dot-dashed curve).
(a) Temperature profile for the liquid phase, (b) gas phase CO2 concentration profile and,
(c) liquid phase CO2 loading.
(a) (b)
0.5 0.5
0.45
0.4 0.4
Sim
Sim
0.35
2
2
θCO
θCO
0.3 0.3
0.25
0.2 0.2
0.15
0.2 0.3 0.4 0.5 0.2 0.3 0.4 0.5
Exp Exp
θCO θCO
2 2
(c) (d)
130 130
TSim ( C)
TSim ( C)
120 120
o
110 110
100 100
100 110 120 130 100 110 120 130
o o
TExp ( C) TExp ( C)
(e) (f)
12
14
12 10
10 8
8 6
6
4
4
2 2
0 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12
Figure 7.5: Sensitivity analysis of the heat of absorption of CO2 on the predictions of
our model in terms of the correction factor δHCO2 compared with the experimental pilot
plant data of 6 selected runs from Tobiesen et al. (2008). δHCO2 = 80 kJ mol−1 (squares),
δHCO2 = 160 kJ mol−1 (upward triangles), δHCO2 = 40 kJ mol−1 (downward triangles). (a)
Loading at the bottom of the desorber, (b) loading at the reboiler outlet, (c) temperature
at the bottom of the desorber, (d) temperature at the reboiler outlet, (e) condensate
flowrate out of the condenser and, (f) CO2 flowrate out of the condenser
tween experimental data and model predictions. Essentially, the extra energy available
in the reboiler due to lower heat of vaporisation of H2 O is used to boil more water, hence
the increase of water flowrate, and to desorb more CO2 , explaining the lower loading in
the reboiler and the higher CO2 flowrate leaving the condenser.
The composition of the streams leaving the condenser are not normally affected by the
heat of vaporisation of water. Indeed, under the conditions of phase separation, the gas
stream will be almost exclusively CO2 and the liquid stream water; this holds whatever
the temperature used in the condenser.
(a) )
400
395
TL (K)
390
385
380
0 10 20 30 40 50
Stage
(b) )
0.4
0.3
2
yCO
0.2
0.1
0
0 10 20 30 40 50
Stage
(c) )
0.5
0.4
0.3
2
θCO
0.2
0.1
0
0 10 20 30 40 50
Stage
Figure 7.6: Sensitivity analysis of the enthalpy of vaporisation of water on the profiles
predicted with our model compared with the experimental pilot plant data of Run 12
from Tobiesen et al. (2008). Nominal value (continuous curve), enthalpy doubled (dashed
curve), enthalpy halved (mixed curve). (a) Temperature profile for the liquid phase, (b)
gas phase CO2 concentration profile and, (c) liquid phase CO2 loading.
(a) (b)
0.5 0.5
0.45
0.4 0.4
Sim
Sim
0.35
2
2
θCO
θCO
0.3 0.3
0.25
0.2 0.2
0.15
0.2 0.3 0.4 0.5 0.2 0.3 0.4 0.5
Exp Exp
θCO θCO
2 2
(c) (d)
130 130
TSim ( C)
TSim ( C)
120 120
o
110 110
100 100
100 110 120 130 100 110 120 130
o o
TExp ( C) TExp ( C)
(e) (f)
18 12
16
10
14
12 8
10
6
8
6 4
4
2
2
0 0
0 2 4 6 8 10 12 14 16 18 0 2 4 6 8 10 12
Figure 7.7: Sensitivity analysis of the enthalpy of vaporisation of water on the predictions
of our model compared with the experimental pilot plant data of 6 selected runs from
Tobiesen et al. (2008). Nominal value (squares), enthalpy doubled (upward triangles),
enthalpy halved (downward triangles). (a) Loading at the bottom of the desorber, (b)
loading at the reboiler outlet, (c) Temperature at the bottom of the desorber, (d) tem-
perature at the reboiler outlet, (e) Condensate flowrate out of the condenser, (f) CO2
flowrate out of the condenser
7.3 Energy requirements in the reboiler
The reboiler is the major consumer of external energy in the process. This is due to three
main physical phenomena taking place in this unit: the reverse carbamate formation when
the CO2 -MEA bonds are broken and the CO2 is desorbed, the vaporisation of water, and
the increase in temperature of the liquid stream. The energy required by these phenomena
is brought into the system via the reboiler; this is the heat duty of the reboiler. This
section details the relative energy consumption of each of the three phenomena described
above.
The energy balance on the reboiler can be written as follows:
Qr + QCO2 + QH2 O + Q∆T = 0 (7.1)
Here we chose not to include the heat of mixing as it is negligible. where QCO2 is the
power for the desorption of CO2 (W), QH2 O is the power of the vaporisation of H2 O (W),
and Q∆T is the power for the temperature increase of the liquid stream. The power for
the desorption of CO2 is given by:
L
QCO2 = Nr,CO2
∆HCO2 (7.2)
where ∆HCO2 is the enthalpy of absorption of CO2 (J mol−1 ). The power for the vapori-
sation of H2 O is given by
QH2 O = Nr,H2 O ∆Hvap,H2 O (7.3)
Each of these contributions is expressed as a percentage of the total reboiler heat duty
for a set of selected runs in Table 7.4. It is clear from this table that there is a correlation
between the proportion of the heat consumed for the CO2 desorption and the vaporisation
and H2 O and the loading range of CO2 . The higher the loading, the more important the
share of energy for the desorption of CO2 and the lesser the share of energy for the
Table 7.4: Heat consumption
Run Loading range CO2 desorption % H2 O vaporisation % Liquid stream ∆T

1 0.26-0.32 18 67 15
5 0.26-0.32 13 71 16
6 0.33-0.40 26 59 15
12 0.33-0.40 30 53 17
14 0.41-0.46 38 26 36
16 0.41-0.46 38 21 41
vaporisation of H2 O.
The heat duty of the reboiler was an input in the pilot plant experimental campaign and
is also an input in the model so its value on each run should be fixed to the value reported
by Tobiesen et al. (2008). However, for the sake of understanding further the process, its
value has been doubled and halved and the effect of its variation on the temperature and
liquid phase CO2 loading at the bottom of the desorber and at the reboiler outlet, and the
flowrates out of the condenser for selected runs is represented in Figure 7.8. Increasing
the heat duty of the reboiler reduces the CO2 loading at the bottom of the desorber and
in the reboiler, hence resulting in a better match between the experimental data and the
simulated results. A direct consequence of this better prediction of the CO2 loading is the
better prediction of the CO2 flowrate leaving the condenser. However, an increase in the
heat duty also results in an increase in the flowrate of H2 O leaving the condenser, leading
to an important over estimation. The effect on the temperature is somewhat limited as it
increases the predictions by about only 3 K. Doubling the reboiler heat duty gives better
predictions for the CO2 composition and flowrate across the whole process but degrades
the quality of the predictions for the temperature and the flowrate of H2 O.
(a) (b)
0.5 0.5
0.45
0.4 0.4
Sim
Sim
0.35
2
2
θCO
θCO
0.3 0.3
0.25
0.2 0.2
0.15
0.2 0.3 0.4 0.5 0.2 0.3 0.4 0.5
Exp Exp
θCO θCO
2 2
(c) (d)
130 130
TSim ( C)
TSim ( C)
120 120
o
110 110
100 100
100 110 120 130 100 110 120 130
o o
TExp ( C) TExp ( C)
(e) (f)
20 12
18
16 10
14 8
12
10 6
8
6 4
4 2
2
0 0
0 2 4 6 8 10 12 14 16 1820 0 2 4 6 8 10 12
Figure 7.8: Sensitivity analysis of the reboiler heat duty on the predictions of our model
compared with the experimental pilot plant data of 6 selected runs from Tobiesen et al.
(2008). Nominal value (squares), heat duty doubled (upward triangles), heat duty halved
(downward triangles). (a) Loading at the bottom of the desorber, (b) loading at the
reboiler outlet, (c) Temperature at the bottom of the desorber, (d) temperature at the
reboiler outlet, (e) Condensate flowrate out of the condenser and, (f) CO2 flowrate out of
the condenser
7.4 Effect of the pressure in the reboiler
From the sections above, it appears that a large part of the desorption of CO2 takes
place in the reboiler. Being an equilibrium stage, the predictive capabilities of the model
rely on the precision of the VLE obtained with the SAFT-VR approach. In Section 7.1,
it was shown that the complete desorber process model slightly over predicted the CO2
loading in the liquid phase at the bottom of the desorber and in the reboiler. This over
prediction could be the result of uncertainties in the experimental data as explained in
Section 7.1; it could also arise from the mass-transfer model but this is unlikely as the
sensitivity analysis showed that the desorber model was almost insensitive to changes in
key mass-transfer related model properties. This leaves the precision of the SAFT-VR
approach as possible source of error.
In order to assess the precision of the thermodynamic approach in this process model,
the reboiler is isolated and simulated as a stand alone unit. The operating conditions of
Run 13 (Tobiesen et al., 2008) are chosen as the condensate rate is null, simplifying the
determination of the composition of the stream entering the reboiler. The composition of
the inlet liquid stream is deduced from the MEA concentration and the reported loading
at the bottom of the desorber. The temperature and pressure of the stream and the heat
duty of the reboiler are also known experimentally. The flowrate is estimated on the basis
that it decreases only slightly along the desorber column. The simulated values of the CO2
loading and the temperature in the reboiler are compared with the experimental values
(cf. Table 7.5). It can be seen that the reboiler loading and temperature are too high.
Varying the inlet flowrate has only a limited impact on the loading and temperature and
Table 7.5: Standalone reboiler simulation for Run T13
Original operating Reduced pressure Experimental values

conditions
Pressure (MPa) 0.2094 0.14 0.2094
CO2 loading 0.398 0.374 0.375
Temperature (◦ C) 117.8 108.4 111.7
varying the heat duty would either increase the loading and decrease the temperature or
vice versa. However, although the inlet pressure is a given experimental value, reducing it
from 0.209 Mpa to 0.14 MPa leads to a significant improvement in the agreement between
the simulated and the experimental values for both the loading and the temperature in
the reboiler as presented in Table 7.5. This confirms the hypothesis that errors in the
thermodynamic model lead to significant deviation between the process model predictions
and pilot plant data at higher temperatures.
The influence of the pressure is assessed on the complete desorber model. It appears that
reducing the pressure from about 0.2 MPa to 0.14 MPa in the desorber column or in the
condenser has a negligible effect and as expected reducing the pressure in the reboiler has
a major effect on the CO2 loading and temperature at the bottom of the desorber and
in the reboiler and on the flowrates of the streams leaving the condenser as presented
in Figure 7.9. Albeit improving the prediction for the loading and the CO2 gas stream
flowrate, this pressure reduction impairs the quality of the predictions for the temperature
of the desorber. In conclusion, the potential error in the thermodynamic model does have
an impact on the predictive capability of the model but this is not the only significant
factor.
(a) (b)
0.5 0.5
0.45
0.4 0.4
Sim
Sim
0.35
2
2
θCO
θCO
0.3 0.3
0.25
0.2 0.2
0.15
0.2 0.3 0.4 0.5 0.2 0.3 0.4 0.5
Exp Exp
θCO θCO
2 2
(c) (d)
130 130
TSim ( C)
TSim ( C)
120 120
o
110 110
100 100
100 110 120 130 100 110 120 130
o o
TExp ( C) TExp ( C)
(e) (f)
18 12
16
10
14
12 8
10
6
8
6 4
4
2
2
0 0
0 2 4 6 8 10 12 14 16 18 0 2 4 6 8 10 12
Figure 7.9: Impact of the reduction of the pressure in the reboiler on the predictions
of our model compared with the experimental pilot plant data of 6 selected runs from
Tobiesen et al. (2008). Nominal value, P≃0.2 Mpa (squares), reduced pressure P=0.14
Mpa (downward triangles). (a) Loading at the bottom of the desorber, (b) loading at the
reboiler outlet, (c) Temperature at the bottom of the desorber, (d) temperature at the
reboiler outlet, (e) Condensate flowrate out of the condenser and, (f) CO2 flowrate out of
the condenser
The performance of the rate-based CO2 desorber model developed and extended from the
absorber model (cf. Chapters 4 and 5) has been assessed. Model validation has been
carried out by comparison with published experimental pilot plant data (Tobiesen et al.,
2008). It is shown that good predictive capabilities can be obtained over a wide range
of operating conditions without any need for modifying the modelling approach between
the absorber and the desorber. The model parameter scaling the diffusivity of the CO2
in the liquid phase and adjusted based on the absorber operating conditions appears to
be transferable to the desorber conditions. The precision of the predictions were not as
good as those obtained for the absorber model. This is due to the low desorption rate of
the chosen pilot plant (Tobiesen et al., 2008) where a higher level of accuracy is needed.
The model is expected to give better predictions for a pilot plant with a higher rate of
desorption.
With a careful sensitivity analysis we show that the model was almost insensitive to
variation in key mass-transfer related model properties. It was highlighted that the higher
the CO2 loading in the process, the more important the share of reboiler heat duty for
the desorption of CO2 and the lesser the share of reboiler heat duty for the vaporisation
of H2 O. Finally, it was shown that the pressure in the reboiler has a significant impact on
the amount of CO2 being desorbed in the reboiler.
These two models can now be combined to provide a complete model of the absorber
desorber CO2 capture process. This is done in the next chapter.
Chapter 8
Application of the model
In this chapter, we combine the absorber model presented in chapters 4 and 5 with the
desorber model presented in chapter 6 and 7 in order to represent a complete process of
CO2 capture.
8.1 Closed flowsheet
A schematic of the complete process model is presented in Figure 8.1. The configuration
presented by Tobiesen et al. (2008) of desorber, reboiler and condenser is used. The outlet
stream of rich solvent from the absorber is fed to the top of the desorber and the outlet
stream of lean solvent from the desorber is fed to the top of the absorber. A pump and a
heater are added on each of these two streams to adjust their pressure and temperature.
The model of the pump is simple, corresponding to specifying a pressure (1 bar for the
absorber and 2 bar for the desorber), while the heater changes the temperature via a
specified heat duty. In Table 8.1 the model conditions that have to be specified for the
simulated run are given. Due to numerical constraint on gPROMS and the SAFT-VR for-
eign object, there must be H2 O, MEA, CO2 and N2 in all the streams, liquid or gas. This
156
8. Application of the model 157
Figure 8.1: Schematic representation of the complete absorber-desorber flowsheet
is why in the process model there is a trace of MEA in the flue gas (yMEA = 1.4575647 ×
10−6 ). The high precision of the numbers reflect the difficulties of obtaining a converged
solution from gPROMS (Process Systems Enterprise Ltd.). The absorber column char-
acteristics are similar to those used in Chapter 5, the desorber column characteristics are
similar to those used in Chapter 7. The stream tables for this run are presented in Figure
8.2. The liquid-phase temperature, gas-phase CO2 concentration, and liquid-phase CO2
concentration profiles in the absorber and the desorber columns are presented in Figure
8.3. The temperature and compositions profiles for the absorber column are really similar
to those obtained in Chapter 5 when reproducing the pilot plant results from Tontiwach-
wuthikul et al. (1992). The temperature profile of the desorber presents a sharp drop
in stage 1. This suggests that there is a flash of the rich solvent stream upon entering
the desorber. This rich solvent is at equilibrium when leaving the absorber. As it passes
through the pump its pressure is increased, moving it away from equilibrium into the liq-
uid phase. It then passes through a heater. The resulting increase in temperature could
be enough to vaporise some of the liquid. This phenomen is observed in this case. The
difference in loading between the rich and the lean solvent is much more important in this
simulation than in the experimental data from Tobiesen et al. (2008).
Figure 8.2: Stream tables for the reference run

(a) ) (d) )
320 397
396
310
TL (K)
TL (K)
395
300
394
290
393
280 392
0 10 20 30 40 50 0 10 20 30 40 50
Stage Stage
(b) ) (e) )
0.15 0.15
0.1 0.1
2
2
yCO
yCO
0.05 0.05
0 0
0 10 20 30 40 50 0 10 20 30 40 50
Stage Stage
(c) ) (f) )
0.5 0.5
0.4 0.4
0.3 0.3
2
2
θCO
θCO
0.2 0.2
0.1 0.1
0 0
0 10 20 30 40 50 0 10 20 30 40 50
Stage Stage
Figure 8.3: Model predictions for a complete absorber-desorber flowsheet. (a) Absorber
temperature profile for the liquid phase, (b) absorber gas phase CO2 concentration profile,
(c) absorber liquid phase CO2 loading (d) desorber temperature profile for the liquid
phase, (e) desorber gas phase CO2 concentration profile, and (f) desorber liquid phase
CO2 loading.
Table 8.1: Inputs to the complete process model
Input variable
Inlet gas flow (mol m−2 s−1 ) 30
Inlet gas pressure (MPa) 0.1
Inlet gas temperature (K) 292.15
Inlet gas CO2 mole fraction 0.1264
Inlet gas H2 O mole fraction 0.014
Inlet gas MEA mole fraction 1.4575647×10−6
Cooler outlet temperature (K) 285.846
Heater outlet temperature (K) 394.949
Desorber pressure (MPa) 0.2067
Reboiler heat duty (kW) 14.95
Condenser temperature (K) 295.15
The CO2 removal rate Ψ (%) is defined as
in out
Ψ = VCO 2
/VCO 2
(8.1)
in out
where VCO 2
is the CO2 molar flowrate in the flue gas and VCO 2
is the CO2 molar flowrate
in the gas stream leaving the condenser (this is the captured CO2 ). For the selected run
Ψ = 99.98 %. This rate is exceptionally high compared to pilot plant data. Cottrell
et al. (2009) state that a typical removal rate is 85-95 % while Notz et al. (2012) reported
removal rates of 75.9 % and 51.3 %. The liquid flowrate entering the absorber is equal to
L = 1.07 mol s−1 and the concentration in MEA of the stream entering the desorber is
16.9 wt%.
The concentration in MEA in the liquid solvent and the liquid circulation rate are not
specified but determined in the simulation. In order to gain control on these parameters,
we added a solvent make-up in the liquid loop, just before the liquid inlet of the absorber
to add more degrees of freedom in the system. It should be noted that having a solvent
make-up is compulsory for real processes to compensate for the MEA and H2 O losses
in the clean flue gas stream and the CO2 gas stream. In the model these losses are
compensated by the H2 O and the trace of MEA present in the flue gas stream.
The flowrate of this solvent make-up has been varied and the effect of this variation on
Table 8.2: Liquid solvent flowrate, MEA concentration and, CO2 recovery rate for different
flowrates of MEA addition
Lmake−up
MEA (mol s−1 ) 0 10−7 10−6 10−5
Liquid flowrate (mol s−1 ) 1.07 0.62 0.47 0.33
MEA concentration (wt%) 16.9 21.20 33.60 58.92
Ψ (%) 99.98 99.95 99.98 99.95
the behaviour of the process is discussed in the remainder of this section.
8.1.1 Addition of MEA
The composition of the make-up solvent is nearly pure MEA (the mole fraction of H2 O,
CO2 , and N2 are set to 10−8 ). The flowrate is increased from Lmake−up
MEA = 0 to 10−5 mol
s−1 . The temperature and pressure are set to be equal to the one of the main liquid
stream entering the absorber. The impact of this addition of MEA on the temperature
and composition profiles along the absorber and desorber columns is presented in Figure
8.4. It appears that an ever so slight addition of MEA in the system leads to tremendous
changes in the temperature profiles. The liquid flowrate, MEA concentration and, CO2
recovery rate for different values of solvent make-up flowrates are listed in Table 8.2.
Complete stream tables are presented in Figures 8.2 and 8.5 to 8.7.
It is observed that an addition of MEA in the process via the solvent make-up decreases
the liquid flowrate and increases the MEA concentration. The case where Lmake−up = 10−6
mol s−1 is not feasible in reality as the MEA concentration is too high and this would
lead to severe corrosion of the columns. The CO2 recovery rate is only slightly affected
and stays consistently very high. Looking at both Figure 8.4 and Table 8.2, it can be
seen that the lower the liquid flowrate, the higher and wider the temperature bulge. This
is explained by the fact that there is less liquid to cool the column down, so there is an
accumulation of energy in the absorber, leading to the build up of the temperature bulge.
Although the behaviour of the temperature profile with changes in the flowrate of make-
(a) ) (d) )
360 410
340 405
TL (K)
TL (K)
320 400
300 395
280 390
0 10 20 30 40 50 0 10 20 30 40 50
Stage Stage
(b) ) (e) )
0.1 0.1
2
2
yCO
yCO
0.05 0.05
0 0
0 10 20 30 40 50 0 10 20 30 40 50
Stage Stage
(c) ) (f) )
0.5 0.5
0.4 0.4
0.3 0.3
2
2
θCO
θCO
0.2 0.2
0.1 0.1
0 0
0 10 20 30 40 50 0 10 20 30 40 50
Stage Stage
Figure 8.4: Impact of addition of MEA in the system via the solvent make-up on the model
predictions for a complete absorber-desorber flowsheet. Lmake−up
MEA = 0 mol s−1 (continuous
make−up make−up
line), LMEA = 10−7 mol s−1 (dashed line), LMEA = 10−6 mol s−1 (dot-dashed line)
make−up −5 −1
and, LMEA = 10 mol s (dotted line). (a) Absorber temperature profile for the
liquid phase, (b) absorber gas phase CO2 concentration profile, (c) absorber liquid phase
CO2 loading (d) desorber temperature profile for the liquid phase, (e) desorber gas phase
CO2 concentration profile and, (f) desorber liquid phase CO2 loading.
make−up
Figure 8.5: Stream tables of the process model with LMEA = 10−7 mol s−1
make−up
make−up
up MEA is monotonic (the higher the flowrate, the higher the temperature), this is not
the case for the loading in both columns and for the CO2 composition in the gas phase in
the desorber. For example in the desorber, as the flowrate of make-up MEA is increased,
the mole fraction of CO2 in the gas phase and the CO2 liquid loading first reduce but
then increase again, highlighting the highly non-linear behaviour of the process.
8.2 Application to an other pilot plant
The newest detailed pilot plant data available in the literature are presented by Notz et al.
(2012). As explained in Chapter 7, these data have not been used for the validation of the
model for two reasons: firstly, they were published after the development of the absorber
model and during the development of the desorber model, and secondly, only two runs are
presented, which in our opinion is not sufficient to fully validate a process model. However,
their study presents extensive experimental data for a complete absorber-desorber MEA-
based CO2 capture process. The topology is modified by the addition of a water washer
at the top of each column. We chose to confirm the transferability of the model developed
in our study by applying it to the pilot plant data from Notz et al. (2012). The results
presented in this section are not yet final as this is part of a work in progress study.
8.2.1 Absorber
The absorber column used in the work of Notz et al. (2012) is 4.20 m high and is filled
with the Sulzer Mellapack 250.YTM structured packing. Its diameter is 0.125 m. The flue
gas stream and the liquid lean solvent stream characteristics for the two simulated runs
are listed in Table 8.3. The comparisons between the simulated and experimental values
for the liquid phase temperature and the mass fraction of CO2 in the liquid phase profiles
are presented in Figure 8.8 and 8.9.
Table 8.3: Inputs to the absorber model
Input variable Run 1 Run 2

Inlet gas flow (kg h−1 ) 72 72.4
Inlet gas pressure (MPa) 0.1 0.1
Inlet gas temperature (K) 321.16 321.3
Inlet gas CO2 mass fraction 0.085 0.165
Inlet gas H2 O mass fraction 0.071 0.069
Inlet gas MEA mass fraction 4.5 × 10−6 4.5 × 10−6
Liquid solvent flow (kg h−1 ) 200.1 200.0

Liquid solvent temperature (K) 313.18 313.33
Liquid solvent CO2 mass fraction 0.052 0.063
Liquid solvent H2 O mass fraction 0.673 0.653
Liquid solvent MEA mass fraction 0.275 0.284
(a) ) (b)
360 0.1
0.08
340
TL (K)
CO2
0.06
320
L
ω
0.04
300
0.02
280 0
0 10 20 30 40 50 0 10 20 30 40 50
Stage Stage
tinuous curves) for Run 1 taken from the work of Notz et al. (2012). (a) Temperature
profile for the liquid phase and, (b) mass fraction of CO2 in the liquid phase.
(a) ) (b)
360 0.1
0.08
340
TL (K)
CO2
0.06
320
L
ω
0.04
300
0.02
280 0
0 10 20 30 40 50 0 10 20 30 40 50
Stage Stage
tinuous curves) for Run 2 taken from the work of Notz et al. (2012). (a) Temperature
profile for the liquid phase and, (b) mass fraction of CO2 in the liquid phase.
Very good agreement is found between the model predictions and the pilot plant data for
Runs 1 and 2 in relation to the profile for the mass fraction of CO2 in the liquid phase
along the entire length of the column. This is a good confirmation of the transferability
of the absorber model. In relation to the temperature profile in the liquid phase, the
model prediction of the outlet temperature value is in acceptable agreement with the
experimental data. However, the model predictions are underestimating the temperature
profile for the liquid phase. This underprediction should be investigated via a sensitivity
analysis in order to assess the relative importance of the process model parameters. In
particular, the heat-transfer related parameters might have a significant effect under these
operating conditions as the flue gas flowrate ratio to the liquid solvent flowrate is higher
in this pilot plant (Notz et al., 2012) than in the pilot plant used for the absorber model
validation (Tontiwachwuthikul et al., 1992).
This chapter first presented the application of the model developed in this study to a
complete MEA based absorber-desorber CO2 capture process. The high sensitivity of
the process to the flowrate of MEA in the solvent make-up was highlighted. The carbon
capture model has also been employed to reproduce experimental results from a recently
published pilot plant study (Notz et al., 2012). Very good agreement was found between
the model predictions and the pilot plant data for different runs of absorber in relation to
the mass fraction of CO2 in the liquid phase along the entire length of the column. The
predictions for the liquid temperature profile were not as good and additional work on
the model is required to assess the cause for the temperature deviation.
Chapter 9
Concluding remarks
9.1 Summary of thesis
The focus of this thesis has been the integration of the SAFT-VR thermodynamic ap-
proach in the development of an absorber-desorber process model for carbon capture. The
aim of the model is to be as predictive as possible, limiting the reliance on experimental
data and bearing in mind that it must ultimately be used for solvent design.
The development of the process model has been done in two steps: we first presented
a model for an absorber column and secondly we transferred this model to a desorber
column and added a condenser and a reboiler.
In both columns the heat- and mass-transfer is described with rate-based equations. Both
the vapour-liquid equilibrium and the chemical equilibrium are treated within the SAFT-
VR thermodynamic framework, ensuring a consistent and accurate representation of the
physical interactions in the system under the assumption that reaction kinetics are not
rate-determining. This approach lends itself to extension to other solvents, as a conse-
quence of the transferable nature of the SAFT molecular models and the relatively small
number of parameters and data required to develop them.
Model validation has been carried out for the absorber model by comparison with pub-
170
9. Concluding remarks 171
lished experimental pilot plant data (Tontiwachwuthikul et al., 1992). A central finding
of our current study is that diffusivity of CO2 in the liquid phase required adjustment in
order to provide an accurate representation of the column profiles; this was done based
on different operating conditions. Very good predictions are obtained for the liquid-phase
temperature profiles and liquid- and gas-phase compositions along the column. This
highlights the importance of capturing the physical and chemical equilibria reliably; in
addition the assumption that the process is not limited by reaction kinetics appears to be
valid.
With a careful sensitivity analysis we show that the liquid viscosity and diffusivity, and
vapour-liquid surface tension, are key properties for the prediction of the composition
profiles. The heat generated by the absorption of CO2 affects the magnitude of the tem-
perature variation in the absorber and the vaporisation of water affects the presence and
amplitude of the temperature bulge.
The CO2 desorber model has been developed and extended from the absorber model.
Model validation has been carried out by comparison with published experimental pilot
plant data (Tobiesen et al., 2008). It is shown that good predictive capabilities can be
obtained over a wide range of operating conditions without any need for modifying the
modelling approach between the absorber and the desorber. The model parameter scaling
the diffusivity of the CO2 in the liquid phase and adjusted based on the absorber operat-
ing conditions appears to be transferable to the desorber conditions.
With a careful sensitivity analysis we show that the model is almost insensitive to vari-
ations in key mass-transfer related model properties. It was highlighted that the higher
the CO2 loading in the process, the more important the share of reboiler heat duty for
the desorption of CO2 and the lesser the share of reboiler heat duty for the vaporisation
of H2 O. Finally, it was shown that the pressure in the reboiler has a significant impact on
the amount of CO2 being desorbed in the reboiler.
Once both the absorber and the desorber model are developed and validated, an appli-
cation of this model to simulate a complete MEA based absorber-desorber CO2 capture
process is presented. The carbon capture model is also employed to reproduce experi-
mental results from a recently published pilot plant study (Notz et al., 2012). Although
additional work on the model is required to assess the cause for the temperature deviation
between the model predictions and the pilot plant data, good agreement is found for the
CO2 composition profile confirming the transferable aspect of the carbon capture model
developed in this study.
9.2 Key contributions
The contributions of this thesis include:
• The development of a predictive and transferable rate-based model of absorber-

desorber carbon capture process based on the SAFT-VR thermodynamic approach.
• An extensive validation of the model on different pilot plants and for a wide range
of operating conditions.
• A demonstration that the concept of implicit reactions and the dominance of mass
transfer limitations works.
• A detailed understanding of the impact of the mass- and heat-transfer phenomena

in the process model.
• A complete model that can be used for process optimisation and solvent selection.
9.3 Directions for future work
In this thesis, a complete model of amine based CO2 capture process has been developed.
The directions for future work arising from this study can be divided across several objec-
tives: increase the accuracy of the model, validate it further and, use the model for process
design, and treatment of impurities. The last paragraph provides a wider perspective on
the use of the model in the global effort to reduce global warming.
9.3.1 Increase the model accuracy
The first direction would be to reassess the accuracy of the SAFT-VR thermodynamic ap-
proach in the temperature range of the desorber. When developing the molecular model,
the aim was to have one single model applicable to different temperatures. Due to the
deviation of the model predictions compared to experimental data at higher temperatures,
this would need to be reviewed and one should explore whether adding a dependence in
temperature on the parameters describing the MEA+CO2 interaction is improving the
accuracy of the results when used in the process model.
A second direction to improve the accuracy would be to assess the heat transfer correla-
tion sensitivity under different operating conditions in order to gain better predictions of
the temperature profiles.
9.3.2 Validate further
An experimental study of the diffusion of carbamate in MEA would be very useful to

obtain a diffusion correlation and so suppressing the model parameter introduced in this
study.
More advanced version of SAFT are available and integrating them in the process would
be beneficial. For example, SAFT-VR Mie is thought to provide a better description of
the heat related parameters, hence potentially suppressing the need for a correction of
the heat of absorption of CO2 . The SAFT-VR Mie molecular model of H2 O, MEA and
CO2 and their cross interactions are still under development.
9.3.3 Process design
A potential application of this model is its use with different solvents. It would be rel-
atively straight-forward to transfer from MEA to AMP for example. In this case one
should reassess the assumption that the process is mass-transfer limited due to the slower
kinetics of AMP. The model could also be applied to blends of solvents but this would
require the determination of new SAFT-VR parameters to mediate the solvent-solvent
interactions. Using a blend of solvents can improve the characteristics of the absorption
process (e.g. reduce the energy penalty) while improving the behaviour of the solvent
(less degradation).
SAFT-γ is a group contribution approach which can be used for molecular design to find
new solvents. A process-wide optimisation of the carbon capture system including solvent
design could be performed.
9.3.4 Extending the model to account for impurities
The model developed in this study does not account solvent degradation or impurities. As
explained in chapter 2, an important limitation of amine-based carbon capture processes is
the degradation of the solvent with oxygen. In order to take into account this phenomena
in the model, future work should focus on developing a model for oxygen and its interaction
with the amine solvent considered (for example MEA). A set of SAFT parameters should
be derived from solubility data of O2 in a mixture of H2 O and MEA.
The best strategy to address the impurities problem would be to use this model as a
tool to explore different configurations and solvents for the carbon capture process. Once
a promising solvent and topology have been selected, a more detailed model including
the effect of impurities can be used to validate further the characteristics of a certain
combination of solvent and configuration.
9.3.5 Potential use of this model
The major advantage of this model is that it can reduce the number of experimentation
needed for the design and the optimisation of an amine based carbon capture process. This
model is a tool to explore different process configurations and identify suitable solvents in a
cheaper and faster way than a full-scale experimentation like pilot or demonstration plant.
The importance of improving the solutions available to reduce the emissions of CO2 in
the atmosphere have been explained in the first chapter of this thesis. It is reconfirmed by
the report recently published by the UN’s climate change panel (IPCC, 2013) highlighting
the urgent need for “substantial and sustained reductions of greenhouse gas emissions”.
By providing researchers a tool capable of understanding and improving carbon capture
processes, the model developed in this study is playing a part in the common effort needed
to stop global warming.
9.4 Publications and presentations
9.4.1 Publications
C. V. Brand, J. Rodrı́guez, A. Galindo, G. Jackson, C. S. Adjiman, Validation of an ab-

sorber model of carbon dioxide capture in an aqueous amine solvent developed based on
the SAFT-VR framework, Proceedings of the 11th Internatinoal Symposium on Process
Systems Engineering, July 2012, Singapore.
C. V. Brand, J. Rodrı́guez, A. Galindo, G. Jackson, C. S. Adjiman, Validation of a process

model of CO2 capture in an aqueous solvent, using an implicit molecular based treatment
of the reactions, GHGT-11, Energy Procedia 2013.
C. V. Brand, J. Rodrı́guez, A. Galindo, G. Jackson, C. S. Adjiman, Development and

validation of an MEA carbon capture absorber process integrating advanced thermody-
namics and process modelling, Submitted to Energy & Environmental Science.
9.4.2 Conference presentations and posters
C. V. Brand, C. S. Adjiman, A. Galindo, G. Jackson, Development and validation of an

MEA carbon capture absorber process integrating advanced thermodynamics and process
modelling, (poster), UK Carbon Capture and Storage Consortium (UKCCSC) Biannual
meeting, September 2011, Heriot-Watt University, Edinburgh, UK.
C. V. Brand, J. Rodrı́guez, A. Galindo, G. Jackson, C. S. Adjiman, Application of the

SAFT-VR framework to the development of a carbon dioxide absorber using aqueous
amine solutions, (talk), UKCCSC Early Careers Winter School, January 2012, University
of Cambridge, UK.
C. V. Brand, J. Rodrı́guez, A. Galindo, G. Jackson, C. S. Adjiman, Validation of an

absorber model of carbon dioxide capture in an aqueous amine solvent developed based
on the SAFT-VR framework, (poster), Centre for Process System Engineering PhD sym-
posium, March 2012, Imperial College London, UK.
C. V. Brand, J. Rodrı́guez, A. Galindo, G. Jackson, C. S. Adjiman, Validation of an

absorber model of carbon dioxide capture in an aqueous amine solvent developed based
on the SAFT-VR framework, (talk), 11th Internatinoal Symposium on Process Systems
Engineering, July 2012, Singapore.

of the reactions, (talk), 2012 AIChE Annual Meeting, October 2012, Pittsburgh, USA.

of the reactions, (poster), GHGT-11, November 2012, Kyoto, Japan.
C. V. Brand, J. Rodrı́guez, A. Galindo, G. Jackson, C. S. Adjiman, CO2 capture using

mono solutions: Development and validation of a process model based on the SAFT-VR
equation of state, (talk), IC4S Annual meeting, April 2013, Imperial College London,
UK.
Nomenclature
δHCO2 Correction to the enthalpy of absorption of CO2 (J mol−1 )
abs
∆HCO 2
Enthalpy of absorption of CO2 (J mol−1 )
v̇ L Volumetric flowrate of the liquid phase (m3 s−1 )
v̇ V Volumetric flowrate of the vapour phase (m3 s−1 )
ǫ Void fraction (-)
ǫL 0 Operating void space in the packing (-)
ηL Dynamic viscosity of the liquid phase (kg m−1 s−1 )
ηV Dynamic viscosity of the vapour phase (kg m−1 s−1 )
γ Contact angle (deg)
λLT Liquid thermal conductivity (W m−1 K−1 )
λVT Vapour thermal conductivity (W m−1 K−1 )
178
Nomenclature 179
λij Parameter of the attractive range of the intermolecular potential between two seg-
ments i and j (-)
µLi Chemical potential of component i in the liquid phase (J mol−1 )
µVi Chemical potential of component i in the gas phase (J mol−1 )
ωiV Mass fraction of component i in the vapour phase
φM Association factor for the liquid mixture (-)
φj Association factor of solvent j (-)
Ψ CO2 removal rate (%)
ρL Density of the liquid phase (kg m−3 )
ρV Density of the vapour phase (kg m−3 )
σ Vapour-liquid surface tension (N.m−1 )
σc Critical surface tension of the packing material (N m−1 )
Σv Atomic diffusion volume (Å)
σii Diameter of segments forming molecule i (Å)
τ Scaling factor for the liquid diffusivity of CO2 in the solvent (-)
θ Angle with horizontal for falling film or corrugation channel (deg)
θCO2 CO2 loading in the liquid phase (-)

Nomenclature 180
εab,ij Depth of the interaction potential between association site a on molecule i and site
b of molecule j (J K−1 )
εij Depth of the intermolecular potential between two segments i and j (J K−1 )
~x Liquid mole fraction (-)
~y Vapour mole fraction (-)
~zi Molar composition of a stream consisting of pure component i (-)
a′ Interfacial area density (m2 m−3 )
ap Specific surface area of the packing (m2 m−3 )
Asection Cross sectional area of the column (m2 )
aT Total interfacial area available for heat or mass transfer (m2 )
c Total number of components
CL Concentration in the liquid phase (mol m−3 )
CV Concentration in the vapour phase (mol m−3 )
C I,L Concentration at the liquid-vapour interface in the liquid phase (mol m−3 )
C I,V Concentration at the liquid-vapour interface in the vapour phase (mol m−3 )
Ce Correaction factor for surface renewal (-)
CpL Specific heat capacity of the liquid phase(J kg−1 K−1 )
CpV Specific heat capacity of the vapour phase(J kg−1 K−1 )

Nomenclature 181
DiL Diffusion coefficient of component i in the liquid phase (m2 s−1 )
DiV Diffusion coefficient of component i in the vapour phase (m2 s−1 )
◦
Di,j Mutual diffusion coefficient of solute i at very low concentrations in solvent j (cm2
s−1 )
ds Diameter of a sphere of the same surface as a single packing particle (m)
Dz Stage height (m)
EL Net gain of energy of the liquid phase (W)
EV Net loss of energy from the vapour phase (W)
FSE Factor for surface enhancement (-)
F rL Liquid-phase Froude number (-)
g Gravitational acceleration (m s−2 )
gef f Effective gravitational acceleration (m s−2 )
HL Enthalpy of the liquid phase (J mol−1 )
HV Enthalpy of the vapour phase (J mol−1 )
ht Liquid holdup (-)
hLT Liquid-phase heat-transfer coefficient (W m−2 K−1 )
hVT Vapour-phase heat-transfer coefficient (W m−2 K−1 )
hpacking Total packing height (m)

Nomenclature 182
kL Liquid-phase mass-transfer coefficient (m s−1 )
kV Vapour-phase mass-transfer coefficient (m s−1 )
kij Binary interaction parameter used to compute the strength of the interactions
between segments i and j (-)
Lj Liquid molar flowrate leaving stage j (mol s−1 )
Lp Nominal packing size (m)
Lspec Specific liquid flowrate (kg s−1 )
Mi Molecular weight of component i (g mol−1 )
mi Number of segments in the molecule i (-)
NiL Net gain of species i in the liquid phase due to interphase transport (mol s−1 )
NiV Net loss of species i in the vapour phase due to interphase transport (mol s−1 )
Ns Number of stage in the column (-)
Ns,ia Number of sites of each type a on molecule i
Ns,i Number of site types for molecule i
PI Pressure at the vapour-liquid interface (MPa)
PL Pressure in the liquid phase (MPa)
PV Pressure in the vapour phase (MPa)
P rL Prandtl number for the liquid phase (-)

Nomenclature 183
P rV Prandtl number for the vapour phase (-)
QLcond Conductive heat flux in the liquid phase (W)
QLdif f Diffusive heat flux in the liquid phase (W)
QVcond Conductive heat flux in the vapour phase (W)
QVdif f Diffusive heat flux in the vapour phase (W)
Qr Reboiler heat duty (W)
Q∆T Power for the temperature increase of the liquid stream (W)
QCO2 Power for the desorption of CO2 (W)
QH2 O Power for the vaporisation of H2 O (W)
rc;ab,ij Attractive range of the interaction potential between association site a on molecule
i and site b of molecule j (Å)
ReL Liquid-phase Reynolds number based on the specific surface area (-)
ReV Vapour-phase Reynolds number (-)
S Side dimension of corrugation (m)
ScLi Schmidt number of component i in the liquid phase (-)
ScVi Schmidt number of component i in the vapour phase (-)
TI Temperature at the vapour-liquid interface (K)
TL Temperature of the liquid phase (K)

Nomenclature 184
TV Temperature of the vapour phase (K)
uL Liquid velocity (m s−1 )
uLe Effective liquid-phase velocity (m s−1 )
uV Vapour velocity (m s−1 )
uVe Effective vapour-phase velocity (m s−1 )
VL Molar volume of the liquid phase (cm3 mol−1 )
VV Molar volume of the vapour phase (cm3 mol−1 )
Vi∗L Molar volume of the pure component i in the bulk liquid phase (cm3 mol−1 )
Vi∗V Molar volume of the pure component i in the bulk vapour phase (cm3 mol−1 )
V I,L Molar volume of the liquid phase at the vapour-liquid interface (cm3 mol−1 )
V I,V Molar volume of the gas phase at the vapour-liquid interface (cm3 mol−1 )
Vim Molar volume of solute i at its normal boiling temperature (cm3 mol−1 )
Vj Vapour molar flowrate leaving stage j (mol s−1 )
Vspec Specific vapour flowrate (kg s−1 )
W eL Liquid-phase Weber number (-)
′
ReL Liquid-phase Reynolds number based on the interfacial area (-)
Bibliography
A. Aboudheir, P. Tontiwachwuthikul, and R. Idem. Rigorous Model for Predicting the

Behavior of CO2 Absorption into AMP in Packed-Bed Absorption Columns. Industrial
& Engineering Chemistry Research, 45(8):2553–2557, April 2006.
A. Aboudheir, P. Tontiwachwuthikul, A. Chakma, and R. Idem. Kinetics of the re-

active absorption of carbon dioxide in high CO2 -loaded, concentrated aqueous mo-
noethanolamine solutions. Chemical Engineering Science, 58(23-24):5195–5210, De-
cember 2003.
B. M. Abraham, J. G. Asbury, E. P. Lynch, and A. P. S. Teotia. Coal-oxygen process

provides CO2 for enhanced recovery. Oil & Gas Journal, 80(11):68–70, 1982.
D. S. Abrams and J. M. Prausnitz. Statistical thermodynamics of liquid-mixtures - new

expression for excess gibbs energy of partly or completely miscible systems. AIChE
Journal, 21(1):116–128, 1975.
W. E. Acree. Thermodynamic properties of nonelectrolyte solutions. New York, 1984.
H. Ahn, M. Luberti, Z. Liu, and S. Brandani. Process configuration studies of the amine
capture process for coal-fired power plants. International Journal of Greenhouse Gas
Control, 16:29–40, August 2013.
185
References 186
Z. Amrollahi, P. A. M. Ystad, I. S. Ertesvåg, and O. Bolland. Optimized process configu-

rations of post-combustion CO2 capture for natural-gas-fired power plant Power plant
efficiency analysis. International Journal of Greenhouse Gas Control, 8:1–11, May 2012.
H. C. Andersen. Cluster expansions for hydrogen-bonded fluids. 1. Molecular association

in dilute gases. Journal Of Chemical Physics, 59(9):4714–4725, 1973.
H. C. Andersen. Cluster expansions for hydrogen-bonded fluids. 2. Dense liquids. Journal

Of Chemical Physics, 61(12):4985–4992, 1974.
A. Arce, N. Mac Dowell, N. Shah, and L. F. Vega. Flexible operation of solvent regener-
ation systems for CO2 capture processes using advanced control techniques: Towards
operational cost minimisation. International Journal of Greenhouse Gas Control, 11:
236–250, 2012.
G. Astarita. Mass Transfer with Chemical Reaction. Elsevier, 1967.
G. Astarita, D. W. Savage, and A. Brito. Gas Treating with Chemical Solvents. John
Wiley & Sons, 1983.
D. M. Austgen, G. T. Rochelle, X. Peng, and C. C. Chen. Model of vapor liquid equi-

libria for aqueous acid gas alkanolamine systems using the electrolyte NRTL equation.
Industrial & Engineering Chemistry Research, 28(7):1060–1073, 1989.
D. M. Austgen, G. T. Rochelle, and C. C. Chen. Model of Vapor-Liquid Equilibria for

Aqueous Acid Gas-Alkanolamine Systems. 2. Representation of H2 S and CO2 Solubility
in Aqueous MDEA and CO2 Solubility in Aqueous Mixtures of MDEA with MEA or
DEA. Industrial & Engineering Chemistry Research, 30:543–555, 1991.
C. Avendaño, T. Lafitte, C. S. Adjiman, A. Galindo, E. A. Müller, and G. Jackson.

SAFT-γ force field for the simulation of molecular fluids: 2. Coarse-grained models of
greenhouse gases, refrigerants, and long alkanes. The Journal of Physical Chemistry.
B, 117(9):2717–2733, March 2013.
References 187
Y.-S. Bae and R. Q. Snurr. Development and evaluation of porous materials for carbon
dioxide separation and capture. Angewandte Chemie (International ed. in English), 50
(49):11586–11596, December 2011.
A. Bardow, K. Steur, and J. Gross. Continuous-molecular targeting for integrated solvent

and process design. Industrial & Engineering Chemistry Research, 49(6):2834–2840,
2010.
J. A. Barker and D. Henderson. Perturbation theory and equation of state for fluids II.
A successful theory of liquids. Journal of Chemical Physics, 47(11):4714–4721, 1967.
D. Barth, C. Tondre, and J.-J. Delpuech. Kinetics and mechanisms of the reactions of
carbon dioxide with alkanolamines: a discussion concerning the cases of MDEA and
DEA. Chemical Engineering Science, 39(12):1753–1757, January 1984.
F. Barzagli, M. Di Vaira, F. Mani, and M. Peruzzini. Improved solvent formulations

for efficient CO2 absorption and low-temperature desorption. ChemSusChem, 5(9):
1724–1731, September 2012.
B. Behzadi, B. Patel, A. Galindo, and C. Ghotbi. Modeling electrolyte solutions with the
SAFT-VR equation using Yukawa potentials and the mean-spherical approximation.
Fluid Phase Equilibria, 236(1-2):241–255, September 2005.
A. Bello and R. O. Idem. Comprehensive Study of the Kinetics of the Oxidative Degra-
dation of CO2 Loaded and Concentrated Aqueous Monoethanolamine (MEA) with and
without Sodium Metavanadate during CO2 Absorption from Flue Gases. Industrial &
Engineering Chemistry Research, 45(8):2569–2579, 2006.
R. Bender, K. Bier, and G. Maurer. Measurement of the specific heat and the Joule-
Thomson coefficient of carbon dioxide at temperatures from -40 to 200 C and pressures
up to 15 bar. Berichte der Bunsengesellschaft für physikalische Chemie, 85:778–784,
1981.
N. H. Berlin. Method for Providing an Oxygen-Enriched Environment, 1996.

References 188
L. T. Biegler, L. Jiang, and V. G. Fox. Recent advances in simulation and optimal design
of pressure swing adsorption systems. Separation & Purification Reviews, 33(1):1–39,
2005.
R. Billet and M. Schultes. Prediction of mass transfer columns with dumped and arranged
packings - Updated summary of the calculation method of Billet and Schultes. Chemical
Engineering Research & Design, 77(A6):498–504, 1999.
J. Blamey, E. J. Anthony, J. Wang, and P. S. Fennell. The calcium looping cycle for
large-scale CO2 capture. Progress in Energy and Combustion Science, 36(2):260–279,
2010.
F. J. Blas and L. F. Vega. Thermodynamic behaviour of homonuclear and heteronuclear

Lennard-Jones chains with association sites from simulation and theory. Molecular
Physics, 92(1):135–150, 1997.
P. M. M. Blauwhoff, G. F. Versteeg, and W. P. M. Van Swaaij. A study on the reaction

between CO2 and alkanolamine in aqueous solutions. Chemical Engineering Science,
39:207–225, 1984.
G. M. Bollas, C. C. Chen, and P. I. Barton. Refined electrolyte-NRTL model: Activity

coefficient expressions for application to multi-electrolyte systems. AIChE Journal, 54
(6):1608–1624, 2008.
W. Böttinger, M. Maiwald, and H. Hasse. Online NMR spectroscopic study of species

distribution in MEA-H2 O-CO2 and DEA-H2 O-CO2 . Fluid Phase Equilibria, 263:131–
143, February 2008.
V. Brandani. A continuous linear association model for determining the enthalpy of

hydrogen-bond formation and the equilibrium association constant for pure hydrogen-
bonded. Fluid Phase Equilibria, 12:87–104, 1983.
References 189
V. Brandani and F. Evangelista. The UNIQUAC associated-solution theory: vapor-liquid

equilibria of binary systems containing one associating and one inert or active compo-
nent. Fluid Phase Equilibria, 17:281–302, 1984.
J. L. Bravo and J. R. Fair. Generalized correlation for mass-transfer in packed distillation-

columns. Industrial & Engineering Chemistry Process Design And Development, 21(1):
162–170, 1982.
D. Britt, H. Furukawa, B. Wang, T. G. Glover, and O. M. Yaghi. Highly efficient sepa-

ration of carbon dioxide by a metal-organic framework replete with open metal sites.
Proceedings of the National Academy of Sciences of the United States of America, 106
(49):20637–20640, December 2009.
B. J. P. Buhre, L. K. Elliott, C. D. Sheng, R. P. Gupta, and T. F. Wall. Oxy-fuel com-

bustion technology for coal-fired power generation. Progress in Energy and Combustion
Science, 31(4):283–307, 2005.
W. G. Chapman, G. Jackson, and K. E. Gubbins. Phase equilibria of associating fluids:

Chain molecules with multiple bonding sites. Molecular Physics, 65(5):1057–1079, 1988.
W. G. Chapman, K. E. Gubbins, G. Jackson, and M. Radosz. SAFT - equation-of-state

solution model for associating fluids. Fluid Phase Equilibria, 52:31–38, 1989.
W. G. Chapman, K. E. Gubbins, G. Jackson, and M. Radosz. New reference equation

of state for associating liquids. Industrial & Engineering Chemistry Research, 29(8):
1709–1721, 1990.
C.-C. Chen and L. B. Evans. A local composition model for excess Gibbs energy of
aqueous electrolyte systems. Institute of Chemical Engineering Journal, 32(3):444–454,
1986.
J. S. Cho. Gas absorption in a countercurrent packed tower : (1) Absorption with simul-
taneous chemical reaction (2) absorption into varying viscous solutions. PhD thesis,
Oregon State University, 1987.
References 190
W.-J. Choi, J.-B. Seo, S.-W. Cho, S.-W. Park, and K.-J. Oh. Simultaneous absorption of
CO2 and SO2 into aqueous AMP/NH3 solutions in binary composite absorption system.
Korean Journal of Chemical Engineering, 26(3):705–710, 2009.
G. N. I. Clark, A. J. Haslam, A. Galindo, and G. Jackson. Developing optimal Wertheim-

like models of water for use in Statistical Associating Fluid Theory (SAFT) and related
approaches. Molecular Physics, 104(22-24):3561–3581, November 2006.
A. J. Cottrell, J. M. McGregor, J. Jansen, Y. Artanto, N. Dave, S. Morgan, P. Pearson,

M. I. Attalla, L. Wardhaugh, H. Yu, A. Allport, and P. H. M. Feron. Post-combustion
capture R&D and pilot plant operation in Australia. Energy Procedia, 1(1):1003–1010,
February 2009.
L. W. Dahl and H. C. Andersen. Cluster expansions for hydrogen-bonded fluids. 3. Water.

Journal Of Chemical Physics, 78(4):1962–1979, 1983.
K. Damen, M. van Troost, A. Faaij, and W. Turkenburg. A comparison of electricity

and hydrogen production systems with CO2 capture and storage. Part A: Review and
selection of promising conversion and capture technologies. Progress In Energy And
Combustion Science, 32(2):215–246, January 2006.
P. V. Danckwerts. Gas-Liquid Reactions. McGraw-Hill, 1970.
P. V. Danckwerts. The reaction of CO2 with ethanolamines. Chemical Engineering Sci-

ence, 34:443–446, 1979.
R. Davidson. Post-combustion carbon capture from coal fired plants solvent scrubbing.
Profiles, IEA Clean Coal Centre, 2007.
L. A. Davies, A. Gil-Villegas, and G. Jackson. Describing the Properties of Chains of

Segments Interacting Via Soft-Core Potentials of Variable Range with the SAFT-VR
Approach. International Journal of Thermophysics, 19(3):675–686, 1998.
References 191
L. A. Davies, A. Gil-Villegas, and G. Jackson. An analytical equation of state for chain

molecules formed from Yukawa segments. The Journal of Chemical Physics, 111(18):
8659–8665, 1999.
W. J. DeCoursey. Enhancement factors for gas absorption with reversible reaction. Chem-
ical Engineering Science, 37(10):1483–1489, 1982.
H. Deng, C. J. Doonan, H. Furukawa, R. B. Ferreira, J. Towne, C. B. Knobler, B. Wang,

and O. M. Yaghi. Multiple functional groups of varying ratios in metal-organic frame-
works. Science, 327:846–850, February 2010.
U. Desideri and A. Paolucci. Performance modelling of a carbon dioxide removal system

for power plants. Energy Conversion and Management, 40(18):1899–1915, December
1999.
F. Dolezalek. Theory of binary mixtures and concentrated solutions. Z. Phys. Chem., 64:
747, 1908.
F. Dolezalek. Theory of binary mixtures. 5. Vapour tension and constitution of liquid

argon and argon nitrogen mixtures. Z. Phys. Chem., 93:585, 1918.
S.-J. Doong and R. T. Yang. Bulk separation of multicomponent gas mixtures by pressure
swing adsorption: pore/surface diffusion and equilibrium models. AIChE Journal, 32
(3):397–410, 1986.
R. E. Dugas. Pilot Plant Study of Carbon Dioxide Capture by Aqueous Monoethanolamine.

PhD thesis, The University of Texas at Austin, 2006.
I. G. Economou and M. D. Donohue. Chemical, quasi-chemical and perturbation theories

for associating fluids. AIChE Journal, 37(12):1875–1894, 1991.
I. G. Economou and M. D. Donohue. Mean field calculations of thermodynamic properties

of supercritical fluids. AIChE Journal, 36:1920–1925, December 1990.
References 192
J. R. Elliott, S. J. Suresh, and M. D. Donohue. A simple equation of state for non-

spherical and associating molecules. Industrial & Engineering Chemistry Research, 29:
1476–1485, 1990.
European Union. Roadmap for moving to a low-carbon economy in 2050, 2011. URL
http://ec.europa.eu/clima/policies/roadmap/index en.htm.
L. Faramarzi, G. M. Kontogeorgis, M. L. Michelsen, K. Thomsen, and E. H. Stenby.

Absorber model for CO2 capture by monoethanolamine. Industrial & Engineering
Chemistry Research, 49(8):3751–3759, 2010.
J. D. Figueroa, T. Fout, S. Plasynski, H. McIlvried, and R. D. Srivastava. Advances in CO2

capture technologyThe U.S. Department of Energy’s Carbon Sequestration Program.
International Journal of Greenhouse Gas Control, 2(1):9–20, January 2008.
A. Fredenslund, R. L. Jones, and J. M. Prausnitz. Group-contribution estimation of

activity coefficients in nonideal liquid mixtures. AIChE Journal, 21(6):1086–1099, 1975.
J. Gabrielsen, M. L. Michelsen, E. H. Stenby, and G. M. Kontogeorgis. A Model for esti-

mating CO2 solubility in aqueous alkanolamines. Industrial & Engineering Chemistry
Research, 44(9):3348–3354, April 2005.
J. Gabrielsen, M. L. Michelsen, E. H. Stenby, and G. M. Kontogeorgis. Modeling of CO2

absorber using an AMP solution. AIChE Journal, 52(10):3443–3451, 2006.
J. Gabrielsen, H. F. Svendsen, M. L. Michelsen, E. H. Stenby, and G. M. Kontogeorgis.

Experimental validation of a rate-based model for CO2 capture using an AMP solution.
Chemical Engineering Science, 62(9):2397–2413, 2007.
A. Galindo, L. A. Davies, A. Gil-Villegas, and G. Jackson. The thermodynamics of

mixtures and the corresponding mixing rules in the SAFT-VR approach for potentials
of variable range. Molecular Physics, 93(2):241–252, February 1998.
References 193
A. Galindo, A. Gil-Villegas, G. Jackson, and A. N. Burgess. SAFT-VRE : Phase Behavior

of Electrolyte Solutions with the Statistical Associating Fluid Theory for Potentials of
Variable Range. Journal of Physical Chemistry B, 103(46):10272–10281, 1999.
A. Galindo, S. J. Burton, G. Jackson, D. P. Visco, and D. A. Kofke. Improved models

for the phase behaviour of hydrogen fluoride: chain and ring aggregates in the SAFT
approach and the AEOS model. Molecular Physics, 100(14):2241–2259, July 2002.
R. Gani and J. P. O’Connell. Properties and CAPE: from present uses to future challenges.
Computers & Chemical Engineering, 25(1):3–14, January 2001.
J. Gáspár and A.-M. Cormo. Dynamic modeling and validation of absorber and desorber
columns for post-combustion CO2 capture. Computers & Chemical Engineering, 35
(10):2044–2052, October 2011.
D. Ghonasgi, V. Perez, and W. G. Chapman. Intramolecular association in flexible hard

chain molecules. The Journal of Chemical Physics, 101(8):6880–6887, 1994.
J. Gibbins and H. Chalmers. Carbon capture and storage. Energy Policy, 36(12, Sp. Iss.
SI):4317–4322, 2008.
A. Gil-Villegas, A. Galindo, P. J. Whitehead, S. J. Mills, and G. Jackson. Statistical

associating fluid theory for chain molecules with attractive potentials of variable range.
Journal of Chemical Physics, 106(10):4168–4186, 1997.
A. Gil-Villegas, A. Galindo, and G. Jackson. A statistical associating fluid theory for

electrolyte solutions (SAFT-VRE). Molecular Physics, 99(6):531–546, March 2001.
H. R. Godini and D. Mowla. Selectivity study of H2 S and CO2 absorption from gaseous
mixtures by MEA in packed beds. Chemical Engineering Research and Design, 86(4):
401–409, April 2008.
G. Grasa, J. Abanades, M. Alonso, and B. Gonzalez. Reactivity of highly cycled particles

References 194
of CaO in a carbonation/calcination loop. Chemical Engineering Journal, 137(3):561–

567, April 2008.
T. Greer, A. Bedelbayev, J. M. Igreja, J. F. Gomes, and B. Lie. A simulation study on the

abatement of CO2 emissions by de-absorption with monoethanolamine. Environmental
technology, 31(1):107–15, January 2010.
J. Gross and G. Sadowski. Perturbed-chain SAFT: An equation of state based on a

perturbation theory for chain molecules. Industrial & Engineering Chemistry Research,
40(4):1244–1260, 2001.
J. Gross and G. Sadowski. Application of the perturbed-chain SAFT equation of state to

associating systems. Industrial & Engineering Chemistry Research, 41(22):5510–5515,
2002.
P. Guerin de Montgareuil and D. Domine. Process for separating a binary gaseous mixture
by adsorption, 1957.
E. A. Guggenheim. Mixtures. Oxford University Press, 1952.
S. Hadlington. Chemistry World, 2011. URL

http://www.rsc.org/chemistryworld/News/2011/April/20041102.asp.
B. Han, Y. Sun, M. Fan, and H. Cheng. On the CO2 Capture in Water-Free Mo-
noethanolamine Solution: An ab Initio Molecular Dynamics Study. The Journal of
Physical Chemistry. B, 117:5971–5977, May 2013.
T. Harkin, A. Hoadley, and B. Hooper. Reducing the energy penalty of CO2 capture and
compression using pinch analysis. Journal Of Cleaner Production, 18(9, Sp. Iss. SI):
857–866, 2010.
A. J. Haslam, A. Galindo, and G. Jackson. Prediction of binary intermolecular potential

parameters for use in modelling fluid mixtures. Fluid Phase Equilibria, 266(1-2):105–
128, April 2008.
References 195
W. N. Haynes. Handbook of Chemistry and Physics. CRC Press, 2011.
H. Hikita and Y. Konishi. The Absorption of SO2 into aqueous NA2 CO3 solutions accom-
panied by the desorption of CO2 . Chemical Engineering Journal and The Biochemical
Engineering Journal, 27(3):167–176, 1983.
H. Hikita, S. Asai, H. Ishikawa, and M. Honda. The Kinetics of Reactions of Carbon

Dioxide with Monoethanolamine, and Triethanolamine by a Rapid Mixing Method
Diethanolamine. The Chemical Engineering Journal, 13:7–12, 1977.
K. A. Hoff. Modeling and experimental study of carbon dioxide absorption in a membrane

contactor. PhD thesis, Norges Teknisk-Naturvitenskapelige Universitet, 2003.
N. Holt, G. B. Epri, and D. Todd. A summary of recent IGCC studies of CO2 capture
for sequestration. In The Gasification Technologies Conference San Fransisco, 2003.
S. H. Huang and M. Radosz. Equation of state for small, large, polydisperse, and as-
sociating molecules. Industrial & Engineering Chemistry Research, 29(11):2284–2294,
August 1990.
S. H. Huang and M. Radosz. Equation of state for small, large, polydisperse, and as-
sociating molecules: extension to fluid mixtures. Industrial & Engineering Chemistry
Research, 30(8):1994–2005, August 1991.
G. D. Ikonomou and M. D. Donohue. Thermodynamics of hydrogen-bonded molecules:

The associated perturbed anisotropic chain theory. AIChE Journal, 32:1716–1725,
October 1986.
G. Ikonomou and M. Donohue. Extension of the associated perturbed anisotropic chain

theory to mixtures with more than one associating component. Fluid phase equilibria,
39:129–159, 1988.
IPCC. IPCC Report, 2013. URL http://www.ipcc.ch/.

References 196
E. E. Isaacs, F. D. Otto, and A. E. Mather. Solubility of mixtures of H2 S and CO2 in a

monoethanolamine solution at low partial pressures. Journal of Chemical Engineering
Data, 25:118–120, 1980.
G. Jackson, W. G. Chapman, and K. E. Gubbins. Phase-equilibria of associating flu-

ids: Spherical molecules with multiple bonding sites. Molecular Physics, 65(1):1–31,
September 1988.
M. S. Jassim and G. T. Rochelle. Innovative Absorber / Stripper Configurations for CO2

Capture by Aqueous Monoethanolamine. Industrial & Engineering Chemistry Research,
45:2465–2472, 2006.
S. A. Jayarathna, A. Weerasooriya, S. Dayarathna, D. A. Eimer, and M. C. Melaaen.

Densities and surface tensions of CO2 loaded aqueous monoethanolamine solutions with
r = (0.2 to 0.7) at T = (303.15 to 333.15) K. Journal of Chemical & Engineering Data,
Accepted, 2013.
F. Y. Jou, A. E. Mather, and F. D. Otto. Solubility of H2 S and CO2 in aqueous

methyldiethanolamine solutions. Industrial and Engineering Chemical Process Design,
21:539–544, 1982.
F. Y. Jou, A. E. Mather, and F. D. Otto. The solubility of CO2 in a 30 mass percent

monoethanolamine solution. The Canadian Journal of Chemical Engineering, 73:140–
147, 1995.
N. M. P. Kakalis, A. I. Kakhu, and C. C. Pantelides. Implementing SAFT-based thermo-

dynamics in process modelling tools. In Proceedings of the 6th international conference
on foundations of computer-aided process design, pages 537–540. Michigan: CACHE
Corporation, 2004.
E. Y. Kenig, L. Kucka, and A. Górak. Rigorous modeling of reactive absorption processes.

Chemical Engineering & Technology, 26(6):631–646, 2003.
References 197
J. Kestin, Y. Kobayashi, and R. T. Wood. The viscosity of four binary, gaseous mixtures
at 20 and 30 C. Physica, 32:1065–1089, 1966.
F. M. Khoury. Multistage Separation Processes. CRC Press, 2005.
E. S. Kikkinides, R. T. Yang, and S. H. Cho. Concentration and recovery of carbon

dioxide from flue gas by pressure swing adsorption. Industrial & Engineering Chemistry
Research, 32:2714–2720, 1993.
I. Kim and H. F. Svendsen. Heat of absorption of carbon dioxide (CO2 ) in mo-

noethanolamine (MEA) and 2-(aminoethyl)ethanolamine (AEEA) solutions. Industrial
& Engineering Chemistry Research, 46(17):5803–5809, 2007.
G. M. Kontogeorgis and G. K. Folas. Thermodynamic Models for Industrial Applications:

From Classical and Advanced Mixing Rules to Association Theories. Wiley, 2010.
R. Krishnamurthy and R. Taylor. A nonequilibrium stage model of multicomponent

separation processes. 1. model description and method of solution. AIChE Journal, 31
(3):449–455, 1985a.
R. Krishnamurthy and R. Taylor. A nonequilibrium stage model of multicomponent

separation processes. 2. comparison with experiment. AIChE Journal, 31(3):456–465,
1985b.
L. Kucka, E. Y. Kenig, and A. Górak. Kinetics of the gas-liquid reaction between carbon
dioxide and hydroxide ions. Industrial & Engineering Chemistry Research, 41:5952–
5957, 2002.
L. Kucka, I. Muller, E. Y. Kenig, and A. Górak. On the modelling and simulation of

sour gas absorption by aqueous amine solutions. Chemical Engineering Science, 58
(16):3571–3578, 2003a.
L. Kucka, J. Richter, E. Y. Kenig, and A. Górak. Determination of gas-liquid reaction

References 198
kinetics with a stirred cell reactor. Separation and Purification Technology, 31:163–175,
2003b.
H. M. Kvamsdal and G. T. Rochelle. Effects of the temperature bulge in CO2 absorp-

tion from flue gas by aqueous monoethanolamine. Industrial & Engineering Chemistry
Research, 47(3):867–875, 2008.
H. M. Kvamsdal, J. Hetland, G. Haugen, H. F. Svendsen, F. Major, V. Kårstad, and

G. Tjellander. Maintaining a neutral water balance in a 450MWe NGCC-CCS power
system with post-combustion carbon dioxide capture aimed at offshore operation. In-
ternational Journal of Greenhouse Gas Control, 4(4):613–622, July 2010.
S. S. Laddha and P. V. Danckwerts. Reaction of CO2 with ethanolamines: kinetics from

gas-absorption. Chemical Engineering Science, 36(3):479–482, January 1981.
T. Lafitte, D. Bessières, M. M. Piñeiro, and J.-L. Daridon. Simultaneous estimation of

phase behavior and second-derivative properties using the statistical associating fluid
theory with variable range approach. The Journal of Chemical Physics, 124(2):024509,
January 2006.
T. Lafitte, M. M. Piñeiro, J.-L. Daridon, and D. Bessières. A comprehensive description

of chemical association effects on second derivative properties of alcohols through a
SAFT-VR approach. Journal of Physical Chemistry B, 111(13):3447–3461, 2007.
T. Lafitte, A. Apostolakou, C. Avendaño, A. Galindo, C. S. Adjiman, E. A. Muller, and

G. Jackson. Accurate statistical associating fluid theory for chains of MIE soft-core
segments (SAFT-VR Mie) for the description of vapour-liquid equilibria and thermo-
dynamic derivative properties. 2013.
D. Lal, F. D. Otto, and A. E. Mather. The solubility of H2 S and CO2 in a diethanolamine

solution at low partial pressures. The Canadian Journal of Chemical Engineering, 63:
681–685, 1985.
References 199
M. Laliberte. A model for calculating the heat capacity of aqueous solutions, with updated
density and viscosity data. Journal of Chemical & Engineering Data, 54(6):1725–1760,
June 2009.
A. Lawal, M. Wang, P. Stephenson, and H. Yeung. Dynamic modelling of CO2 absorption

for post combustion capture in coal-fired power plants. Fuel, 88(12):2455–2462, 2009.
A. Lawal, M. Wang, P. Stephenson, G. Koumpouras, and H. Yeung. Dynamic modelling

and analysis of post-combustion CO2 chemical absorption process for coal-fired power
plants. Fuel, 89(10):2791–2801, 2010.
A. G. Leibush and E. D. Shorina. Physico-chemical properties of ethanolamines. Zhurnal

Prikladnoi Khimii, 20:69–76, 1947.
N. Mac Dowell. The integration of advanced molecular thermodynamics and process mod-
elling for the design of amine-based CO2 capture processes. PhD thesis, Imperial College
London, 2010.
N. Mac Dowell and N. Shah. Identification of the cost-optimal degree of CO2 capture: An
optimisation study using dynamic process models. International Journal of Greenhouse
Gas Control, 13:44–58, 2013.
N. Mac Dowell, N. Florin, A. Buchard, J. Hallett, A. Galindo, G. Jackson, C. S. Adjiman,

C. K. Williams, N. Shah, and P. S. Fennell. An overview of CO2 capture technologies.
Energy & Environmental Science, 3(11):1645–1669, 2010a.
N. Mac Dowell, A. Galindo, G. Jackson, and C. S. Adjiman. Integrated solvent and process
design for the reactive separation of CO2 from flue gas. In S. Pierucci and B. G. Ferraris,
editors, Computer Aided Process Engineering, volume 28 of Computer-Aided Chemical
Engineering, pages 1231–1236, 2010b.
N. Mac Dowell, F. Llovell, C. S. Adjiman, G. Jackson, and A. Galindo. Modeling the

fluid phase behavior of carbon dioxide in aqueous solutions of monoethanolamine us-
References 200
ing transferable parameters with the SAFT-VR approach. Industrial & Engineering
Chemistry Research, 49(4):1883–1899, 2010c.
N. Mac Dowell, N. J. Samsatli, and N. Shah. Dynamic modelling and analysis of an

amine-based post-combustion CO2 capture absorption column. International Journal
of Greenhouse Gas Control, 12:247–258, January 2013.
Y. Maham, C. N. Liew, and A. E. Mather. Viscosities and excess properties of aqueous

solutions of ethanolamines from 25 to 80C. Journal of Solution Chemistry, 31(9):743–
756, 2002.
H. P. Mangalapally, R. Notz, S. Hoch, N. Asprion, G. Sieder, H. Garcia, and H. Hasse.

Pilot plant experimental studies of post combustion CO2 capture by reactive absorption
with MEA and new solvents. In Energy Procedia, volume 1 of Energy Procedia, pages
963–970, 2009.
H. P. Mangalapally, R. Notz, N. Asprion, G. Sieder, H. Garcia, and H. Hasse. Pilot

plant study of four new solvents for post combustion carbon dioxide capture by reactive
absorption and comparison to MEA. International Journal of Greenhouse Gas Control,
8:205–216, May 2012.
W. D. Marsh, F. S. Pramuk, R. C. Hoke, and C. W. Skarstrom. Pressure Equalization

Depressurization in Heatless Adsorption, 1964.
C. McCabe and A. Galindo. Applied Thermodynamics of Fluids, Chapter 8 SAFT Asso-

ciating Fluids and Fluid Mixtures. RCS Publishing, 2010.
T. Mimura, H. Simayoshi, T. Suda, M. Iijima, and S. Mituoka. Development of en-

ergy saving technology for flue gas carbon dioxide recovery in power plant by chemical
absorption method and steam system. Energy Conversion and Management, 38(96):
57–62, January 1997.
A. J. Minchener. Coal gasification for advanced power generation. Fuel, 84(17):2222–2235,

December 2005.
References 201
P. Mores, N. Scenna, and S. Mussati. Post-combustion CO2 capture process: Equilibrium

stage mathematical model of the chemical absorption of CO2 into monoethanolamine
(MEA) aqueous solution. Chemical Engineering Research and Design, 89(9):1587–1599,
September 2011.
P. Mores, N. Scenna, and S. Mussati. A rate based model of a packed column for CO2
absorption using aqueous monoethanolamine solution. International Journal of Green-
house Gas Control, 6:21–36, January 2012.
E. A. Müller and K. E. Gubbins. Molecular-based equations of state for associating

fluids: A review of SAFT and related approaches. Industrial & Engineering Chemistry
Research, 40(10):2193–2211, 2001.
M. Mundhwa and A. Henni. Molar heat capacity of various aqueous alkanolamine solutions
from 303.15 K to 353.15 K. Journal of Chemical & Engineering Data, 52(2):491–498,
2007.
C. Noeres, E. Y. Kenig, and A. Górak. Modelling of reactive separation processes: reactive

absorption and reactive distillation. Chemical Engineering and Processing, 42:157–178,
2003.
M. Nogueira and D. D. Mamora. Effect of flue-gas impurities on the process of injection

and storage of CO2 in depleted gas reservoirs. Journal of Energy Resources Technology,
130(1), 2008.
R. Notz, H. P. Mangalapally, and H. Hasse. Post combustion CO2 capture by reactive

absorption: Pilot plant description and results of systematic studies with MEA. Inter-
national Journal of Greenhouse Gas Control, 6:84–112, January 2012.
K. Onda, E. Sada, and H. Takeuchi. Gas absorption with chemical reaction in packed
columns. Journal of Chemical Engineering of Japan, 1:62–66, 1968a.
K. Onda, H. Takeuchi, and Y. Okumoto. Mass transfer coefficient between gas and liquid
phases in packed columns. Journal of Chemical Engineering of Japan, 1:56–62, 1968b.
References 202
M. A. Pacheco and G. T. Rochelle. Rate-based modeling of reactive absorption of CO2 and

H2 S into aqueous methyldiethanolamine. Industrial & Engineering Chemistry Research,
37(10):4107–4117, 1998.
J. D. Pandya. Adiabatic gas-absorption and stripping with chemical-reaction in packed

towers. Chemical Engineering Communications, 19(4-6):343–361, 1983.
P. Paricaud, A. Galindo, and G. Jackson. Modeling the Cloud Curves and the Solubility
of Gases in Amorphous and Semicrystalline Polyethylene with the SAFT-VR Approach
and Flory Theory of Crystallization. Industrial & Engineering Chemistry Research, 43
(21):6871–6889, October 2004.
P. Paricaud, L. Tazi, and J. Borgard. Modeling the phase equilibria of the HIx mixture
using the SAFT-VRE equation of state: Binary systems. International Journal of
Hydrogen Energy, 35(3):978–991, February 2010.
E. A. Parson and D. W. Keith. Fossil The Fuels Without CO2 Emissions. Science, 282:
1053–1054, 1998.
B. H. Patel, P. Paricaud, A. Galindo, and G. C. Maitland. Prediction of the Salting-Out

Effect of Strong Electrolytes on Water + Alkane Solutions. Industrial & Engineering
Chemistry Research, 42(16):3809–3823, August 2003.
S. Paul, A. K. Ghoshal, and B. Mandal. Theoretical studies on separation of CO2 by single

and blended aqueous alkanolamine solvents in flat sheet membrane contactor (FSMC).
Chemical Engineering Journal, 144(3):352–360, November 2008.
D.-Y. Peng and D. B. Robinson. A New Two-Constant Equation of State. Industrial &
Engineering Chemistry Fundamentals, 15(1):59–64, February 1976.
H. W. Pennline, D. R. Luebke, K. L. Jones, C. R. Myers, B. I. Morsi, Y. J. Heintz,

and J. B. Ilconich. Progress in carbon dioxide capture and separation research for
gasification-based power generation point sources. Fuel Processing Technology, 89(9):
897–907, 2008.
References 203
D. E. Penny and T. J. Ritter. Kinetic study of the reaction between carbon dioxide
and primary amines. Journal of the Chemical Society, Faraday Transactions 1, 79(9):
2103–2109, 1983.
F. E. Pereira, E. Keskes, A. Galindo, G. Jackson, and C. S. Adjiman. Integrated sol-

vent and process design using a SAFT-VR thermodynamic description: High-pressure
separation of carbon dioxide and methane. Computers & Chemical Engineering, 35(3):
474–491, 2011.
L. R. Perkins and C. J. Geankoplis. Molecular diffusion in a ternary liquid system with

the diffusing component dilute. Chemical Engineering Science, 24(7):1035–1042, July
1969.
R. H. Perry and D. W. Green. Perry’s Chemical Engineers’ Handbook. McGraw Hill, 6th
edition, 2008.
T. Pintola, P. Tontiwachwuthikul, and A. Meisen. Simulation of pilot plant and industrial

CO2 -MEA absorbers. Gas Separation and Purification, 7(1):47–52, 1993.
J. M. Plaza, E. Chen, and G. T. Rochelle. Absorber intercooling in CO2 absorption

by piperazine-promoted potassium carbonate. AIChE Journal, 56(4):905–914, April
2010a.
J. M. Plaza, D. Van Wagener, and G. T. Rochelle. Modeling CO2 capture with aqueous
monoethanolamine. International Journal of Greenhouse Gas Control, 4(2):161–166,
March 2010b.
B. E. Poling, J. M. Prausnitz, and J. P. O’Connell. The Properties of Gases and Liquids.

McGraw-Hill Book Co, 5th edition, 2001.
J. M. Prausnitz, R. N. Lichtenthaler, and E. G. de Azevedo. Molecular thermodynamics

of fluid phase equilibria. 1999.
References 204
Process Systems Enterprise Ltd. gPROMS ModelBuilder. URL

http://www.psenterprise.com/.
A. B. Rao and E. S. Rubin. A technical, economic, and environmental assessment of amine-

based CO2 capture technology for power plant greenhouse gas control. Environmental
science & technology, 36(20):4467–4475, October 2002.
H. Renon and J. M. Prausnitz. Local compositions in thermodynamic excess functions

for liquid mixtures. AIChE Journal, 14(1):135–144, 1968.
J. A. Riddick, W. B. Bunger, and T. K. Sakano. Organic Solvents. Wiley, New York, 4th
edition, 1986.
J. A. Rocha, J. L. Bravo, and J. R. Fair. Distillation Columns Containing Structured

Packings:. Industrial & Engineering Chemistry Research, 32(4):641–651, 1993.
J. A. Rocha, J. L. Bravo, and J. R. Fair. Distillation columns containing structured pack-

ings: A comprehensive model for their performance. 2. Mass-transfer model. Industrial
& Engineering Chemistry Research, 35(5):1660–1667, 1996.
J. Rodrı́guez, N. Mac Dowell, F. Llovell, C. S. Adjiman, G. Jackson, and A. Galindo. Mod-

elling the fluid phase behaviour of aqueous mixtures of multifunctional alkanolamines
and carbon dioxide using transferable parameters with the SAFT-VR approach. Molec-
ular Physics, 110(11-12):1325–1348, 2012.
L. M. Romeo, I. Bolea, and J. M. Escosa. Integration of power plant and amine scrubbing
to reduce CO2 capture costs. Applied Thermal Engineering, 28(8-9):1039–1046, June
2008.
Y. K. Salkuyeh and M. Mofarahi. Reduction of CO2 capture plant energy requirement

by selecting a suitable solvent and analyzing the operating parameters. International
Journal of Energy Research, 2012.
References 205
A. Sánchez-Biezma, J. C. Ballesteros, L. Diaz, E. de Zárraga, F. Álvarez, J. López,

B. Arias, G. Grasa, and J. C. Abanades. Postcombustion CO2 capture with CaO.
Status of the technology and next steps towards large scale demonstration. Energy
Procedia, 4:852–859, January 2011.
R. P. Sear and G. Jackson. Thermodynamic perturbation theory for association into

chains and rings. Physical Review E, 50(1):386–394, 1994a.
R. P. Sear and G. Jackson. Thermodynamic perturbation theory for association into

doubly bonded dimers. Molecular Physics, 82(5):1033–1048, 1994b.
R. P. Sear and G. Jackson. The ring integral in a thermodynamic perturbation theory for
association. Molecular Physics, 87(2):517–521, 1996a.
R. P. Sear and G. Jackson. Thermodynamic perturbation theory for association with

bond cooperativity. The Journal of Chemical Physics, 105(3):1113–1120, 1996b.
A. J. Sexton and G. T. Rochelle. Catalysts and inhibitors for MEA oxidation. Energy
Procedia, 1(1):1179–1185, February 2009.
T. Shimizu, T. Hirama, H. Hosoda, K. Kitano, M. Inagaki, and K. Tejima. A twin

fluid-bed reactor for removal of CO2 from combustion processes. Chemical Engineering
Research & Design, 77(A1):62–68, 1999.
R. V. Siriwardane, M. S. Shen, E. P. Fisher, and J. A. Poston. Adsorption of CO2 on

molecular sieves and activated carbon. Energy & Fuels, 15:279–284, 2001.
C. W. Skarstrom. Method and apparatus for fractionating gaseous mixtures by adsorption,

1958.
K. Sumida, D. L. Rogow, J. A. Mason, T. M. McDonald, E. D. Bloch, Z. R. Herm, T.-H.

Bae, and J. R. Long. Carbon dioxide capture in metal-organic frameworks. Chemical
reviews, 112(2):724–781, February 2012.
References 206
M. Takahashi, Y. Kobayashi, and H. Takeuchi. Diffusion Coefficients and Solubilities of

Carbon Dioxide in Binary Mixed Solvents. Journal of Chemical & Engineering Data,
27(3):328–331, 1982.
S. Takahashi, K. Song, and R. Kobayashi. Experimental vapor-liquid equilibria in the

CO2 -diethylene glycol-H2 O and CO2 -triethylene glycol-H2 O systems at feasible absorp-
tion temperatures and pressures. Journal Of Chemical And Engineering Data, 29(1):
23–28, 1984.
K. Takeshita and A. Kitamoto. Chemical equilibria of absorption of CO2 into nonaqueous

solution of amine. Journal of chemical engineering of Japan, 21(4):411–417, August
1988.
S. P. Tan, H. Adidharma, and M. Radosz. Recent advances and applications of statistical

associating fluid theory. Industrial & Engineering Chemistry Research, 47(21):8063–
8082, November 2008.
R. Taylor and R. Krishna. Multicomponent Mass Transfer. John Wiley & Sons, 1993.
F. A. Tobiesen, H. F. Svendsen, and O. Juliussen. Experimental validation of a rigorous

absorber model for CO2 postcombustion capture. AIChE Journal, 53(4):846–865, 2007.
F. A. Tobiesen, O. Juliussen, and H. F. Svendsen. Experimental validation of a rigorous

desorber model for CO2 post-combustion capture. Chemical Engineering Science, 63
(10):2641–2656, May 2008.
P. Tontiwachwuthikul, A. Meisen, and C. J. Lim. Novel pilot plant tchniques for sizing
gas absorbers with chemical reactions. The Canadian Journal of Chemical Engineering,
67(4):602–607, 1989.
P. Tontiwachwuthikul, A. Meisen, and C. J. Lim. CO2 absorption by NaOH, mo-

noethanolamine and 2-amino-2-methyl-1-propanol solutions in a packed column. Chem-
ical Engineering Science, 47(2):381–390, 1992.
References 207
R. E. Treybal. Mass-Transfer Operations. McGraw-Hill, 3rd edition, 1981.
I. J. Uyanga and R. O. Idem. Studies of SO2 - and O2 -Induced Degradation of Aqueous

MEA during CO2 Capture from Power Plant Flue Gas Streams. Industrial & Engi-
neering Chemistry Research, 46(8):2558–2566, 2007.
W. P. M. van Swaaij and G. F. Versteeg. Mass transfer accompanied with complex

reversible chemical reactions in gasliquid systems: an overview. Chemical Engineering
Science, 47(13-14):3181–3195, September 1992.
N. B. Vargaftik. Dictionary of Thermophysical Properties of Gases and Liquids.

MOSKVA, pages 433–434, 1972.
G. Vazquez, E. Alvarez, J. M. Navaza, R. Rendo, and E. Romero. Surface tension of

binary mixtures of water plus monoethanolamine and water plus 2-amino-2-methyl-1-
propanol and tertiary mixtures of these amines with water from 25 degrees C to 50
degrees C. Journal of Chemical & Engineering Data, 42(1):57–59, 1997.
J. L. Wagner. Selective Adsorption Processes, 1969.
T. Wall, Y. Liu, C. Spero, L. Elliott, S. Khare, R. Rathnam, F. Zeenathal, B. Moghtaderi,

B. Buhre, C. Sheng, R. Gupta, T. Yamada, K. Makino, and J. Yu. An overview on
oxyfuel coal combustionState of the art research and technology development. Chemical
Engineering Research and Design, 87(8):1003–1016, August 2009.
M. Wang, A. Lawal, P. Stephenson, J. Sidders, and C. Ramshaw. Post-combustion CO2

capture with chemical absorption: A state-of-the-art review. Chemical Engineering
Research and Design, 89(9):1609–1624, September 2011.
R. H. Weiland, M. Rawal, and R. G. Rice. Stripping of carbon dioxide from mo-

noethanolamine solutions in a packed column. AIChE Journal, 28(6):963–973, 2004.
M. S. Wertheim. Fluids with highly directional attractive forces. 1. Statistical thermody-

namics. Journal of Statistical Physics, 35(1-2):19–34, 1984a.
References 208
M. S. Wertheim. Fluids with highly directional attractive forces. 2. Thermodynamic

perturbation-theory and integral-equations. Journal of Statistical Physics, 35(1-2):35–
47, 1984b.
M. S. Wertheim. Fluids with highly directional attractive forces. 3. Multiple attraction

sites. Journal of statistical physics, 42(3-4):459–476, February 1986a.
M. S. Wertheim. Fluids with highly directional attractive forces. IV. Equilibrium poly-
merization. Journal of statistical physics, 42(3-4):477–492, February 1986b.
S. Westmeier. Excess enthalpy, free excess enthalpy, excess volume and viscosity of selected
binary-liquid mixtures. 4. Systems water-tertiary butanol and water- secondary butanol.
Chemische Technik, 29(4):218–222, 1977.
A. O. Yazaydin, A. I. Benin, S. A. Faheem, P. Jakubczak, J. J. Low, R. R. Willis, and

R. Q. Snurr. Enhanced CO2 adsorption in metal-organic frameworks via occupation
of open-metal sites by coordinated water molecules. Chemistry of Materials, 21(13):
1425–1430, 2009.
J. Ying and D. A. Eimer. Determination and measurements of mass transfer kinetics of

CO2 in concentrated aqueous monoethanolamine solutions by a stirred cell. Industrial
& Engineering Chemistry Research, 52(7):2548–2559, February 2013.
T. Yonemoto, T. Shindo, H. Inomata, and T. Tadaki. Effect of ionic migration on ab-

sorption rate of CO2 into NaOH solutions. Kagaku Kogaku Ronbushu, 10(4):415–420,
1984.
J. Yoo, S. H. Cho, and T. Yang. Comparison of Activated Carbon and Zeolite 13X for
CO2 Recovery from Flue Gas by Pressure Swing Adsorption. Industrial & Engineering
Chemistry Research, 34(2):591–598, 1995.
Y. Zhang, H. Chen, and C. Chen. Rate-based process modeling study of CO2 capture with
aqueous monoethanolamine solution. Industrial & Engineering Chemistry Research, 48
(20):9233–9246, 2009.
Appendix 209
Table 1: SPF file used for the SAFT-VR foreign object

4 4 components, H2O, MEA, CO2 and N2
18.015d0 MW, molecular weight for H2O
0 IFLAG (Spectra data are not available)
32.24255d0 1.923835d-3 1.055492d-5 -3.596461d-9 Cp (J/molK) Constants obtained from 3rd ed. Props of Gases & Lqds, Reid & Prousnitz
1.0d0 SPHERES
3.03420d-10 SIGMA
1.78890d0 LAMBDA
250.00d0 EPSILON/K
2 number of site types (2 for H2O)
1 2 2 ”O” sites
2 2 2 ”H” sites
1 2 1400.00d0 2.10822d-10 EHB and Rc (K) (Angstroms)
0 0 0 0 Terminator
61.083d0 MW, molecular weight for MEA
0 IFLAG (Spectra data are available)
9.311d0 3.009d-1 -1.818d-4 4.656d-8 Cp (J/molK) Constants obtained from 3rd ed. Props of Gases & Lqds, Reid & Prousnitz
2.0d0 SPHERES
3.57229d-10 SIGMA
1.58280d0 LAMBDA
305.00d0 EPSILON/K
4 number of site types (4 for MEA)
1 2 2 ”A” sites - the lone pair on the alkanol oxygen
2 1 1 ”B” sites - the hydrogen on the alkanol
3 1 1 ”C” sites - the lone pair on the alkylamine nitrogen
4 2 1(2?) ”D” sites - the hydrogen on the alkylamine
1 2 2357.79d0 2.08979d-10 EHB and Rc for the alkanol - alkanol self-association interaction ”A-B”
3 4 960.00d0 2.32894d-10 EHB and Rc for the amine - amine self-association interaction ”C-D”
1 4 900.00d0 2.65064d-10 EHB and Rc for the alkanol - amine cross-self-association interaction
2 3 550.00d0 2.65064d-10 EHB and Rc for the amine - amine self-association interaction ”C-B”
0 0 0 0 Terminator
44.01d0 MW, molecular weight for CO2
0 IFLAG (Spectra data are not available)
1.978d1 7.339d-2 -5.598d-5 1.714d-8 Cp constants for CO2
2.0d0 SPHERES
2.7864d-10 SIGMA
1.51573d0 LAMBDA
179.27d0 EPSILON/K
2 number of site types (2 for CO2)
1 1 1 ”a 1” site
2 1 1 ”a 2” site
0 0 0 0 Terminator
28.014d0 MW, molecular weight for N2
0 IFLAG (Spectra data are available)
3.115d1 -1.357d-2 2.680d-5 -1.168d-8 Cp (J/molK) Constants obtained from 3rd ed. Props of Gases & Lqds, Reid & Prousnitz
1.4d0 SPHERES
3.07357783d-10 SIGMA
1.58795239d0 LAMBDA
74.5872388d0 EPSILON/K
0 number of site types (0 for N2)
0 0 0 0 Terminator
1 2 2 Binary interaction parameters between H2O & MEA, parameter 2: kij
0.01
1 2 3 MEA+H2O, hydroxyl - H2O associative interaction, parameter 3 eHBij = 1780.7121 K
1780.71d0 1 2
1 2 3 MEA+H2O, hydroxyl - H2O associative interaction, parameter 3 eHBij = 1780.7121 K
1780.71d0 2 1
1 2 3 MEA+H2O, amine - H2O associative interaction, parameter 3 eHBij = 1517.1049 K
1517.10d0 1 4
1 2 3 MEA+H2O, amine - H2O associative interaction, parameter 3 eHBij = 1517.1049 K
1517.10d0 2 3
1 2 4 MEA+H2O, hydroxyl - H2O, parameter 4: rcij = 2.143735218(Angst)
2.10763d-10 1 2
1 2 4 MEA+H2O, hydroxyl - H2O, parameter 4: rcij = 2.143735218(Angst)
2.10763d-10 2 1
1 2 4 MEA+H2O, amine - H2O, parameter 4: rcij (Angst)
2.22626d-10 1 4
1 2 4 MEA+H2O, amine - H2O, parameter 4: rcij (Angst)
2.22626d-10 2 3
1 3 2 Binary interaction parameters between H2O and CO2 - parameter 2: kij
-0.06d0 kij for H2O+CO2 = -0.06
2 3 2 Binary interaction parameters between MEA & CO2, parameter 2: kij
0.47878
2 3 3 Bond energy between MEA and CO2, eHBij = 5200d0
5200d0 3 1
2 3 4 rcij between MEA+CO2 (Angst)
1.97978d-10 3 1 Cross-association rc (Angsrtoms)
2 3 3 Bond energy between MEA and CO2, eHBij = 3982.66d0
3982.66d0 3 2
2 3 4 rcij between MEA+CO2 (Angst)
1.96999d-10 3 2 Cross-association rc (Angsrtoms)
1 4 2 Binary interaction parameters between H2O and N2 - parameter 2: kij
-0.3635d0 kij for H2O+N2 = -0.3635
2 4 2 Binary interaction parameters between MEA and N2 - parameter 2: kij
0.03d0 kij for MEA+N2 = 0.03
3 4 2 Binary interaction parameters between CO2 and N2 - parameter 2: kij
-0.05990 kij for CO2+N2 = -0.0599
0 0 0 End of Input File

HC E155q

Uploaded by

Copyright:

Available Formats

HC E155q

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

HC E155q

Uploaded by

Copyright:

Available Formats

CO2 capture using monoethanolamine solutions:

Development and validation of a process model based

Charles Victor Brand

A thesis submitted for the degree of

Centre for Process Systems Engineering

2 Techniques for CO2 capture 29

3 Thermodynamic theory and molecular model 49

4 Development of an absorber model using SAFT 69

5 Absorber model results 87

6 Model of the regeneration module 122

7 Regeneration module results 133

8 Application of the model 156

9 Concluding remarks 170

2.1 Flowsheet of an amine-based carbon capture process . . . . . . . . . . . . 35

4.1 Diagram of a non-equilibrium stage. This stage represents a section of

5.1 Comparison of the pilot plant data (squares) (Tontiwachwuthikul et al.,

6.1 Diagram of the desorber . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

7.1 Topology of the solvent regeneration unit . . . . . . . . . . . . . . . . . . . 136

8.1 Schematic representation of the complete absorber-desorber flowsheet . . . 157

= 10−6 mol s−1 (dot-dashed line) and, Lmake−up

2.1 Overview of models of CO2 absorber using MEA solutions. . . . . . . . . . 46

3.3). The interaction matrix is symmetric, i.e., εHB HB

4.1 Correlations used in the model. . . . . . . . . . . . . . . . . . . . . . . . . 85

8.1 Inputs to the complete process model . . . . . . . . . . . . . . . . . . . . . 160

1 SPF file used for the SAFT-VR foreign object . . . . . . . . . . . . . . . . 209

1.1 Motivation for CO2 capture

CO2 + HOC2 H4 NH2 ↽HOC2 H4 NH+

1.2 Objectives of this work

1.3 Outline of thesis

This thesis is organised as follows:

• Chapter 3 first focuses on the evolution of thermodynamic theories applicable to

• In chapter 4, a predictive model of a rate-based absorber column incorporating the

• In chapter 6, the modelling of the solvent regeneration unit, composed of a desorber,

• In chapter 8, the models developed in chapters 3, 4 and 6 are linked together to

Techniques for CO2 capture

2.1 CO2 capture technologies

2.1.1 Calcium looping

2.1.2 Oxy-fuel combustion

2.1.3 Integrated gasification combined cycle

2.1.4 Pressure swing adsorption

2.1.5 Chemical absorption

) respectively, have a lower reactivity (Laddha and Danckwerts, 1981;

Figure 2.1: Flowsheet of an amine-based carbon capture process

2.2 Amine-based processes

2.2.1 Importance of the thermodynamic and the transport mod-

2.2.2 Column description

2.2.3 The concept of a stage

2.2.4 Rate-based vs Equilibrium models

Each stage can be modelled using either an equilibrium or a rate-based model. In an

2.2.5 Two film theory and enhancement factor

When using a rate-based approach, an important aspect in model development is the

2.2.6 Mass transfer correlations

2.2.7 Phase equilibrium and chemical reactions

Another important component in modelling CO2 absorption is the representation of the

2.2.8 Integration of SAFT into process models

and Shah, 2013) investigations of post-combustion CO2 capture processes. Qualitative

2.2.9 Summary of existing process models

2.2.10 Overview of available pilot plant

Source Thermodynamic model Mass transfer model Validation