Process Design

---------------------------------------------------------------------------------------------------------------------
---------------------------------------------------------------------------------------------------------------------
Process Design:
Absorption-Desorption Columns
---------------------------------------------------------------------------------------------------------------------
---------------------------------------------------------------------------------------------------------------------
Edited by
Chakib BOUALLOU
Centre Energtique et Procds
MINES ParisTech
60, boulevard Saint-Michel 75272 Paris Cedex 06, France

E-mail : chakib.bouallou@mines-paristech.fr http://www.mines-paristech.fr
CONTENTS
Summary 2
Chapter 1. Kinetics of Absorption of CO2 in Concentrated Aqueous

Methyldiethanolamine Solutions in the Range 296 K - 343 K. 4
F. Pani, A. Gaunand, R. Cadours, C. Bouallou and D. Richon
Chapter 2. Absorption of H2S by an Aqueous Methyldiethanolamine

Solution at 296 and 343K 25
F. Pani, A. Gaunand, D. Richon, R. Cadours, and C. Bouallou
Chapter 3. Kinetics of absorption of carbon dioxide in aqueous solutions of N-

methyldiethanolamine + ethanolamine or + diethanolamine at 296 K or 343 K. 42
F. Pani; C. Bouallou; R. Cadours; A. Gaunand; D. Richon
Chapter 4. Kinetics of CO2 desorption from highly concentrated and CO2 loaded MDEA
Aqueous Solutions in the Range 312 K - 383 K. 51

R. Cadours, C. Bouallou, A. Gaunand and D. Richon
Chapter 5. Rigorous Simulation Of Gas Absorption Into Aqueous Solutions 74

R. Cadours and C. Bouallou
Chapter 6. Acid gas partial vapor pressure over aqueous MDEA solutions 102
B. Lemoine, Yi-Gui Li, R. Cadours, C. Bouallou, D. Richon
Chapter 7. Representation of regeneration of CO2-Loaded MethylDiEthanolAmine Aqueous

solutions in the range 313-383 K 120
1
Chapter 8. Modeling of acid gases absorption column using alkanolamine solutions
D.-J. Vinel and C. Bouallou 131
Chapter 9. Kinetics of Carbonyl Sulphide (COS) Absorption with Aqueous Solutions of

DiEthanolAmine and MethyDiEthanolAmine 158
F. Amararene, C. Bouallou
2
Summary
This work includes seven parts. In the first part, the kinetics of CO2 absorption by
aqueous solutions of Methyl DiEthanol Amine (MDEA) were measured in the temperature range
296 343 K and MDEA concentration range : (830 - 4380) mol. m-3 (10 - 50 mass %). A
thermoregulated constant interfacial area Lewis-type cell was operated by recording the pressure
drop during batch absorption. The kinetic results are in agreement with a fast regime of
absorption according to film theory. MDEA depletion at the interface has a significant effect on
the kinetics at the CO2 pressures (100 to 200 kPa) studied in this work, especially at low
temperatures and low MDEA concentrations. Considering only the reaction between CO2 and
MDEA, the CO2 absorption appears as a first order reaction with respect to MDEA.
In the second one, data obtained for H2S absorption into a 50 mass %
methyldiethanolamine (MDEA) aqueous solution with various initial H2S loadings are reported
for the following conditions: H2S loadings from 0 to 0.44 mol H2S per mol MDEA and two
temperatures: 296 and 343 K. The H2S experimental absorption rates have been used together
with a model, based on mass transfer and involving an instantaneous reversible reaction, to
determine the MDEA diffusion coefficient.
In the third part, kinetics measurements of CO2 absorption by aqueous amine solutions
are reported for several operating conditions : temperature, 296 K or 343 K; CO2 loadings, (0 to
0.38) mol gas/mol amines, - amines used and concentrations: 50 wt. % methyldiethanolamine
(MDEA)- 5 wt. % monoethanolamine (MEA), 50 wt. % MDEA - 5 wt. % diethanolamine
(DEA), 45 wt. % MDEA - 5 wt. % DEA and 30 wt. % MDEA - 20 wt. % DEA.
In the fourth part, kinetics of CO2 desorption from CO2 loaded methyldiethanolamine
(MDEA) aqueous solutions were measured in the following conditions : 312 - 383 K, 25 - 50 wt
% MDEA aqueous solutions, CO2 loadings from 5 to 85 %. A thermoregulated constant
interfacial area reaction cell was operated by measuring the pressure over the solution.
Producing a very slight depression in the cell, the time dependent equilibrium pressure recovery
is accurately recorded during batch desorption. Kinetics are in agreement with a fast reaction-
regime of desorption according to the film theory. For CO2 loadings below 0.50 mol of gas per
mol of amine, desorption rates are well predicted by using the kinetic constant and orders
3
determined from absorption experiments for the reaction between CO2 and MDEA. Some
discrepancies pointed out for loadings above 0.50 mol of gas per mol of amine.
In the fifth part, a general model based on the film theory is used to set up the
diffusion-reaction partial differential equations which describe the absorption of gases with
multiple reversible reactions. This model includes the electrostatic terms that arise from
differences in diffusivities. The system of equations made up of a set of partial differential
equations and an algebraic equation is solved by finite-difference iterative method. The model
developed in this article, combined with appropriate literature parameters will be used to
model acid gas treating systems in the general case. Optimisation using a Levenberg-
Marquardt algorithm provides the rate coefficients for different kinetic mechanisms, from
experimental data. The utility of this model is illustrated by applying it to the absorption of
carbon dioxide into MDEA aqueous solutions, the comparison of different kinetic
mechanisms shows a significant influence of the reaction between CO2 and OH-. The
contribution of this reaction seems to be generally underestimated in the previous works using
analytical or numerical methods.
In the sixth part, a new apparatus is presented to measure low partial vapor pressures
of acid gas. New measurements are given for carbon dioxide and hydrogen sulfide into
aqueous solutions of MDEA. Large discrepancies among the measured VLE data and the
slopes of their solubility curves are found between the authors due to systematic experimental
errors, and particularly in the low acid loading regions. Different modeling approaches are
proposed herein.
In the seventh part, an analytical method using thermodynamic and kinetic

approximations was used to represent the CO2 desorption rates from MDEA aqueous solutions.
The kinetic model developed in the fifth part, combined with appropriate diffusivity parameters
and a thermodynamic model using apparent constants were used to determine the kinetics
parameters of a CO2-water-MDEA mechanism from absorption rate data . In this work, this
numerical tool was used to predict desorption rates. We have selected desorption experiments
which were consistent with the thermodynamic model used in the model.
4
The eighth part is devoted to the improvement of acid gases removal process by
alkanolamine solutions, especially DiEthanolAmine (DEA) and MethyDiEthanolAmine
(MDEA). A new rigorous model treating reactive absorption based on modified two-film
theory is developed as a first step. This model uses Nernst-Planck equations for the liquid
film and Stefan-Maxwell equations for the gas film. The program was developed to handle
either kinetically controlled or instantaneous chemical reactions in the liquid film. In a second
step, the model is introduced within a simulator of industrial column; the new numerical tool
thus developed makes it possible to represent successfully the cases of industrial absorption
columns in term of absorption rates and enhancement factors.
In the last part, a constant gas-liquid interface reactor was used to measure carbonyl
sulphide (COS) absorption rates in alkanolamine aqueous solutions. Two alkanolamines were
studied : a secondary alkanolamine, DiEthanolAmine (DEA) and a tertiary alkanolamine, N-
MethylDiEthanolamine (MDEA) in the temperature range : 313 to 353 K and for
alkanolamine mass fractions going from 0.05 to 0.50. The reactor used is of Lewis type,
modified with a non-rotating valve shortens considerably the cell loading time. For COS
absorption in DEA aqueous solutions, the limiting step of the reaction is the deprotonation of
the zwitterion. The kinetics data for COS absorption into MDEA aqueous solutions is
interpreted using the base catalysed mechanism for the hydrolysis of COS. The new kinetic
data show that previously reported results concerning the COS absorption by MDEA aqueous
solutions were over-estimated.
5
Chapter 1.
Kinetics of Absorption of CO2 in Concentrated Aqueous Methyldiethanolamine Solutions

in the Range 296 K - 343 K
F. Pani, A. Gaunand, R. Cadours, C. Bouallou and D. Richon
Introduction
Absorption by aqueous solutions of alkanolamines is the dominant industrial process

for removing acid gases, mainly CO2 and H2S, from natural gas. Such washing processes are
also used in petroleum refining, coal gasification and hydrogen production. Instead of
ethanolamine (MEA) and diethanolamine (DEA) solutions, industry would prefer to use less
corrosive and more advanced solvent systems which could be formulated along with the plant
design and operation, according to the feed and exit streams specifications of plants.
Methyldiethanolamine (MDEA) and blends with primary or/and secondary amines, and
sterically hindered amines are new systems that have been studied in the laboratory and
sometimes used on industrial scale. In spite of an abundance of literature (Barth et al., 1981,
1984), Blauwhoff et al., 1981, Yu et al., 1985, Critchfield 1988, Haimour et al., 1987,
Versteeg and van Swaaij, 1988, Littel et al., 1990, Rinker et al., 1995), only few works
(Tomcej and Otto, 1989, Toman and Rochelle, 1989, Xu et al., 1992) deal with absorption
kinetics for CO2 + MDEA + H2O with solutions more concentrated than 3.10-3 mol.m-3
MDEA in water. High MDEA concentrations seem advantageous for absorption kinetics, in
spite of the increasing viscosities which lead to decreasing diffusivities and physical transfer.
Only Tomcej and Otto (1989), and Xu et al. (1992) investigated absorption of carbon dioxide
in aqueous MDEA solution with MDEA concentrations higher than 3400 mol.m-3 solutions
and temperatures higher than 340 K.
Rate expressions generally accepted (Yu et al., 1985, Critchfield 1988, Haimour et al.,
1987, Versteeg and van Swaaij, 1988, Littel et al., 1990, Rinker et al., 1995, Tomcej and Otto,
1989, Toman and Rochelle, 1989) for the forward chemical reaction between CO2 and MDEA
are first order with respect to the concentrations of each of these species, while the estimated
energies of activation lie between (33.1 and 71.6) kJ.mol-1.
6
This work describes the very efficient apparatus developed to measure absorption
kinetics of acid gases into amines solutions. Experiments of CO2 absorption by aqueous
MDEA solutions, starting with CO2 pressures ranging from (100 - 200) kPa, and in an
extended range of MDEA concentrations and temperatures have been made with this
apparatus.
Experimental Section
Chemicals
Twice-distilled water and reagent-grade MDEA are used. MDEA is from different
origins : from Aldrich, with a certified minimum purity of 99 mass %, from Merck, with a
certified minimum purity of 98 mass %, and from Alfa with a certified minimum purity of 98
mass %. Carbon dioxide is from L'Air Liquide, with a certified purity of 99.995 vol %.
Experimental set-up and mode of operation
The 6 x 10-2 m internal diameter thermostated glass reactor (Lewis type, Lewis 1954)
shown in Figure 1 is provided with a six-bladed Rushton turbine in its lower part, a 4 x 10-2 m
diameter propeller in its upper part, and four equally spaced vertical PTFE baffles to prevent
vortexing. An horizontal PTFE plate and a ring are put at midway of the bottom and the top of
the cell to set both level and area of the gas-liquid interface, and to make sure of its stability
during stirring. The shafts are maintained vertically between two pivots supported by two
sapphire bearings fixed in the seats machined on the inside parts of the two stainless steel
flanges on the flanges of the cell and the horizontal PTFE plate, they are driven magnetically
by adjustable speed motors. This technique avoids leaking, friction and heat generation due to
shafts passing through the envelope of the cell by means of packings. The impeller speed is
checked with a stroboscope, it remains constant within 1 rpm during each test. The
temperature in the reactor is known within 0.05 K through a 100 W platinum probe,
calibrated against a 25 W STHPB platinum probe from LYON ALEMAND LOUYOT. The
temperature is controlled by circulating a thermostatic fluid through the glass double jacket.
The whole cell is placed inside a thermoregulated air bath. A tube allows either to degas or to
introduce CO2 into the cell. The total volume available for gas and liquid is (0.3504 0.0005)
7
x 10-3 m3 and the gas-liquid interfacial area A is (11.72 0.05) x 10-4 m2. Uncertainties are
total estimated
Kinetics of gas absorption are measured by recording the absolute pressure drop
through a SEDEME pressure transducer, working in the range (0 to 200) kPa. This transducer
is thermostated at a temperature slightly higher than the experiment temperature to avoid
liquid condensation in its measuring chamber. For each temperature investigated, it is
calibrated within 200 Pa against a mercury manometer. A microcomputer fitted with a data
acquisition card is used to convert the pressure transducer signal directly into pressure units
(Pa), using calibration constants previously determined, and record it as a function of time.
Water and MDEA are degassed independently and aqueous solutions are prepared
under a vacuum. The mass of water and MDEA are known by differential weighings to within
10-2 g. In the worst cases, uncertainties on masses lead to uncertainties on concentrations than
0.05%.The flask containing the solution is connected to the reactor to allow the solution to
transfer by gravity under vacuum. Accurate weighings of the flask before and after transfer
yield the mass of solution actually present in the cell, and the liquid phase volume was
calculated using the density measurements from Al-Ghawas et al. (1989) for MDEA aqueous
solutions. At a given temperature, and under solution vapor pressure PI, pure CO2 is
introduced during a very short time (about 2 s) in the upper part of the cell, the volume of
which is noted as VG. The resulting pressure P0 is between 100 and 200 kPa. Then stirring is
started, and the pressure drop resulting from absorption is recorded. The kinetics of CO2
absorption into MDEA+water solutions with no initial CO2 loading are adequately measured
within one hour.
Results
Measurements have been carried out at three temperatures : 296, 318 and 343 K, for
six MDEA compositions. All rough results of pressure-time absorption data are compiled as
a supplementary material. An example of such a table is given in Appendix I-A. Table I-1
gives in column 1 the reference number of experiment, column 2 the temperature of the
experiment, column 3 the concentration of MDEA, column 4 the initial pressure, column 5
the inert pressure, column 6 the CO2 partial pressure, column 7 vapor phase volume inside the
8
Lewis cell, column 8 the mass transfer coefficient in the liquid side, column 9 the slope b and
column 10 the flux.
The influence on absorption kinetics of all chemical reactions between dissolved CO2
and reactants in solution is usually expressed by an "enhancement factor" E over physical
absorption :
d PC O 2 RT
= k L A E C C O 2 ,i (1)
dt VG
where kL is the liquid-side mass transfer coefficient of unreacted CO2, R is the gas constant
8.3143 J.K-1.mol-1, T is the absolute temperature. VG is the volume of gas. The gas phase is
assumed ideal and the concentration of CO2 is very small in the bulk liquid compared to its
concentration CCO2,i at the interface.
At the interface, vapor-liquid equilibrium is assumed. The partial pressure PCO2 is

related to the concentration of unreacted dissolved CO2 by Henry's law :
PCO 2 = H C CO 2 ,i (2)
where H is the molar scale Henry's law constant.
PCO2 is obtained from the measured pressures :
PCO2 = P PI (3)
Where P is the total pressure. PI is the total vapor pressure over the MDEA+water solution
before CO2 loading, it is assumed constant during an experiment (the absence of leak was
verified by checking for pressure stability before pure CO2 was introduced in the cell).
Initial absorption rates are measured for a pressure range of 10 kPa from the initial
total pressure P0 : for this small pressure drop, the concentration of CO2 resulting from
9
absorption does not change much the composition of the solution, so that kL, H and E remain
constant with time, and integration of equation (1) yields :
P PI
ln = b t (4)
P0 PI
where :
RT
b= k LA E (5)
VG H
The regression of experimental data shows that eq (4) is verified with a root mean
square average error on the left member lower than 0.01. Slopes b are given for each
experiment in Table 1. MDEA is from Aldrich except for some experiments pointed out in the
footnotes of Table 1.
For 4380 mol.m-3 MDEA solutions, at T = 296.45 K and N = 100 rpm, reproducibility
tests have been performed (Experiments 8-13). The relative standard deviation on slope b for
the six experiments is less than 8 %, within the reproducibility range, which shows that no
significant difference can be found when solutions are prepared with MDEA supplied by
either Alfa, Aldrich or Merck, despite their slight differences in purity. For 1980 mol.m-3
MDEA solutions, at T = 296.45 K and N = 100 rpm, and widely varying initial pressures P0, a
reproducibility of 8 % is obtained (Experiments n 3 and 4). For 1980 mol.m-3 MDEA
solutions, at T = 342.6 K and N = 100 rpm, a reproducibility of b better than 3.5 % is obtained
(Experiments n 25 to 29). Therefore, all chemical reactions between dissolved CO2 and
reactants in solution are first order with respect to the amount of dissolved CO2.
With a 4380 mol.m-3 MDEA solution, at 296.45 K, changing the liquid-side stirring
speed N from 50 (Experiment n 14) to 100 rpm (Experiments n 7-9) leads to an increase on
CO2 absorption rate values of about 15 %, but no increase is found between N = 100 rpm and
N = 150 rpm (Experiment n15). In conditions of faster absorption kinetics, and higher vapor
pressure of the inert components, for example at 343.45 K, 1940 mol.m-3 MDEA and N = 100
rpm, where gas-side mass transfer resistance may appear, two experiments (n 30 and 31)
10
have been made with the gas phase stirred at 680 rpm. For these two experiments, slopes b are
similar within 1.2%. Another test of the influence of gas phase stirring was done with a 830
mol.m-3 MDEA solution (Experiment n 23). The differences between slopes b from
experiments with and without gas-side stirring are found to be within the reproducibility
range.
At a given temperature, the slopes b show that absorption kinetics become faster when
MDEA concentration increases from 840 to 1900 mol.m-3, but decrease again for MDEA
concentrations from 1900 to 4300 mol.m-3 (Figure 2). This behaviour comes from opposite
influences of MDEA concentration upon chemical kinetics, CO2 diffusion and CO2 solubility.
Additionaly, for given a MDEA concentration, the absorption kinetics always increases with
temperature in the range 296 K - 343 K.
Interpretation of results in terms of reaction kinetics.
For each experiment, the enhancement factor E is obtained from b using eq 5 and
estimates of data for the molar based Henry's law constant H and the liquid-side mass-transfer
coefficient kL of dissolved CO2.
The data and correlations of Al-Ghawas et al. (1989) are used to estimate the CO2
Henrys constant H.
The mass-transfer coefficient kL is calculated using the correlation between

dimensionless numbers presented in Appendix B. This correlation has been established for
our apparatus from N2O absorption experiments by similar MDEA+water solutions, and
correlations of density, viscosity and diffusion coefficient from previous studies cited.
At a given temperature and stirring speed, because of both increase of the molar based
CO2 Henrys constant H and decrease of the mass-transfer coefficient kL with the
concentration of MDEA, the enhancement factor increases with the concentration of MDEA.
The values of E are always higher than 3, indicating a fast regime of absorption.
Kinetics of the reaction between CO2 and MDEA
11
Most of previous works on CO2 absorption by solutions of tertiary amines R1R2R3N
consider base catalysis for the reaction of CO2 with water, as initially proposed by Donaldson
and Nguyen (1980). Fast hydrogen bonding between the nitrogen electron pair of the tertiary
amine and a water molecule would increase the reactivity of water with dissolved CO2
according to the mechanism :
R 1 R 2 R 3 N + H 2 O == R 1 R 2 R 3 N , HOH (I)

CO 2 + R 1 R 2 R 3 N , HOH
R 1 R 2 R 3 N H +
+ HCO 3 (II)
Equation I assumed to be at equilibrium. Such a mechanism, for water concentration

far higher than the MDEA concentration, leads to the generally accepted rate expression :
r = k C MDEA C CO2 (6)
Hereafter, the first order with respect to the MDEA concentration is tested and the Arrhenius
law for the kinetic constant k is obtained.
For the fast regime of absorption found, whatever the hydrodynamics at the interface
and the mass-transfer model chosen, the enhancement factor E is expressed as a function of
chemical kinetics at the interface :
1
E= ( k ov D CO 2 ) 1/ 2 (7)
kL
where
k ov = k C MDEA ,i (8)
if it is assumed that CO2 only reacts with MDEA according to the rate, eq 7. Subscript i
indicates the interfacial concentration of MDEA.
12
Using Brian's approximate film theory (Brian et al., 1961), which assumes that a
single irreversible reaction between CO2 and MDEA is responsible for the enhancement of
absorption, leads to :
p D CO 2
1/ 2

C MDEA ,i = C MDEA ,T 1

CO 2
( E 1)

(9)
HC MDEA ,T D MDEA
CMDEA,T is the total MDEA concentration. The estimation of the diffusion coefficient
DCO2 of CO2 is based on the work of Versteeg and van Swaaij (1988b), also used by
Glasscock et al. (1991), for temperatures between (293 and 333) K, and MDEA
concentrations between (240 and 2900) mol.m-3. These authors also give a correlation to
estimate the ratio of diffusion coefficients of CO2 and MDEA (see Appendix B). Significant
discrepancies appear between diffusion coefficient data sets, here between those from
Versteeg and van Swaaij (1988b) and those from Al-Ghawas et al. (1989). Both values are
based on experiments with N2O and the analogy between N2O and CO2, but Versteeg and van
Swaaij (1988b) use a Lewis type cell, similar to ours, while Al-Ghawas et al. (1989) use a
wetted sphere and a laminar-jet apparatus. We chose the correlation given by Versteeg and
van Swaaij (1988b).
For each experiment, the interfacial MDEA concentration is calculated using eq 9, for
the partial pressure of CO2 after a decrease of 5 kPa, and the initial MDEA concentration. For
the lowest concentrations of MDEA used in this work, at all temperatures, the MDEA
depletion reached 36%.
Taking H from Al-Ghawas at al. (1989) and assuming that the kinetic constant k
varies with temperature according to an Arrhenius law, a multiple regression of ln kOV upon
ln(CMDEA,i) and 1/T yields for our results in the range (293 to 343) K and (840 to 2900)
mol.m-3 MDEA (Figure 3) :
ln k ov = 135 (
. + 0.934 ln C MDEA ,i / mol. m 3 ) (T5454
/ K)
(10)
13
With eq 10, the relative root mean square average error on the estimate of slopes b is
3.6 %. Using the regression of ln[kov/CMDEA,i] upon 1/T, leads to the Arrhenius expression for
the kinetic constant (Figure 4) :
5461
k = 4.68 105 exp - (11)
( T / K)
with 3.9 % as the relative root mean square average error on the estimate of b, very close to
the previous one.
Using for H the work of Versteeg and van Swaaij (1988b), the multiple regression of
ln kOV upon ln(CMDEA,i) and 1/T yields for our results in the range 293 to 323 K and 840 to
2960 mol.m-3 MDEA :
( ) (T4957
ln k ov = 12.2 + 0.879 ln C MDEA,i / mol. m3
/ K)
(12)
with 3.8 % for the relative root mean square average error on the estimate of slopes b. The
slight difference between the temperatures of activation in eqs 10 and 12 shows the
significant influence of the thermodynamic data chosen on the Arrhenius equation for the
CO2+MDEA reaction, and how important is the reliability of both diffusion and equilibrium
data. The two sets of Henrys constants lead to ln kov values different by a maximum of 22%.
Comparison with Previous Data and Analysis
Versteeg and van Swaaij (1988), Haimour et al. (1987), Yu et al. (1985), Critchfield
(1988) have already studied CO2 absorption by MDEA solutions in stirred cells with a known
flat interface similar to our own. Tomcej and Otto (1989) used a single-sphere absorber. Yu et
al. used pure CO2 at one atmosphere, Versteeg used CO2 pressures lower than 1 atm. The
ranges of temperatures and MDEA concentrations, as well as partial pressures of CO2
investigated, are presented in Table 2.
14
The values of k of eq 11 for the temperatures investigated are in very good agreement
with those of recent work (Tomcej and Otto (1989), and Littel et al. (1990)) (Table 2). But,
because the MDEA depletion at the interface has been taken into account in this work, they
are higher than those calculated by the first authors cited, whose data are limited to less
concentrated MDEA solutions and low temperatures. Table 2 also shows that there is a very
good agreement with the recent work for the activation energy from eq 11.
Conclusion
Kinetics of CO2 absorption by aqueous solutions of MDEA have been studied in the extended
ranges of temperatures from 296 K to 343 K and of MDEA concentrations from 10 to 50 mass
%, and compared with previous studies. They were measured though recording the pressure
drop during absorption in a batch thermoregulated Lewis-type cell. Mass-transfer properties
corresponding to this cell have been characterized by considering absorption measurements
with the MDEA+water+N2O system, leading to a correlation between the classical
dimensionless criteria.
Kinetics constants are in good agreement with the Tomcej and Otto (1989), and Littel
et al. (1990) data as well as for activation energy. Data treatment which involves additional
data is very dependent on the Henrys constants and diffusion coefficients for which careful
and accurate measurements would be of the utmost interest.
ACKNOWLEDGEMENT- The authors gratefully acknowledge the financial support of the

Gas Research Institute.
15
Literature cited
Al-Ghawas, H.A.; Hagewiesche, D.P.; Ruiz-Ibanez, G. ; Sandall, O.C. Physicochemical

properties important for carbon dioxide absorption in aqueous methyldiethanolamine. J.
Chem. Eng. Data 1989, 34, 385-391.
Barth, D.; Tondre C.; Lappai, G.; Delpuech, J.J. Kinetic study of carbon dioxide reaction with
tertiary amines in aqueous solutions. J. Phys. Chem. 1981, 85, 3660-3667.
Barth, D.; Tondre, C; Delpuech, J.J., Kinetics and mechanisms of the reactions of carbon
dioxide with alkanolamines : a discussion concerning the cases of MDEA and DEA. Chem.
Eng Sci. 1984, 39, 1753-1757.
Blauwhoff, P.M.M.; Versteeg, G.F; van Swaaij, W.P.M. A study on the reaction between CO2
and alkanolamines in aqueous solutions. Chem. Eng Sci. 1984, 39, 207-225.
Bosch, H.; Versteeg, G.F.; van Swaaij, W.P.M. Gas-liquid mass transfer with parallel
reversible reactions-I. Absorption of CO2 into solutions of sterically hindered amines. Chem.
Eng Sci. 1989, 44, 2732-2734.
Brian, P.L.T.; Hurley, J.F. ; Hasseltine, E.H. Penetration theory for gas absorption
accompanied by a second order chemical reaction. AIChE J. 1961, 7, 226.
Critchfield, J.E. CO2 absorption/desorption in methyldiethanolamine solutions promoted with

monoethanolamine and diethanolamine : mass transfer and reaction kinetics. PhD
dissertation. The University of Texas at Austin. 1988,
Donaldson, T.L.; Nguyen, Y.N. Carbon dioxide reaction and transport into aqueous amine
membranes. Ind. Eng Chem. Fundam. 1980, 19, 260-266.
16
Glasscock, D.A.; Critchfield, J.E.; Rochelle, G.T. CO2 absorption/desorption in mixtures of
methyldiethanolamine with monoethanolamine or diethanolamine. Chem. Eng Sci. 1991, 46,
2829-2845.
Haimour, N.; Bidarian, A.; Sandall, O. Kinetics of the reaction between carbon dioxide and
methyldiethanolamine. Chem. Eng Sci. 1987, 42, 1393-1398.
Lewis, J.B. The mechanism of mass-transfer of solutes across liquid-liquid interfaces-I.

Chem. Eng. Sci. 1954, 3, 248-259.
Littel, R.J.; van Swaaij, W.P.M.; Versteeg, G.F. Kinetics of carbon dioxide with tertiary
amines in aqueous solution. AIChE J. 1990, 36, 1633-1640.
Reid, R.C.; Prausnitz, J.M.; Sherwood, T.K. The properties of gases and liquids, 3rd ed. Mc
Graw-Hill, 1977.
Rinker E. B.; Oelschlager D.W.; Colussi T.; Henry K.R.; Sandall O. C. Viscosity, density and
surface tension of binary mixtures of water and N-methyldiethanolamine and water and
diethanolamine and tertiary mixtures of these amines with water over the temperature range
20-100C. J. Chem. Eng. Data 1994, 39, 392-395.
Rinker E. B.; Ashour S. S.; Sandall O. C. Kinetics and modelling of carbon dioxide
absorption into aqueous solutions of N- methyldiethanolamine. Chem. Eng. Science 1995, 50,
755-768.
Toman, J.J.; Rochelle, G.T. Carbon dioxide absorption rates and physical solubility in 50%
aqueous methyldiethanolamine partially neutralized with sulfuric acid. 1989, paper n56c,
AIChE Nat. Meeting, Houston, Texas.
Tomcej, R.A.; Otto, F.D. Absorption of CO2 and N2O into aqueous solutions of
methyldiethanolamine. AIChE J. 1989, 35, 861-864.
17
Versteeg, G.F.; van Swaaij, W.P.M. On the kinetics between CO2 and alkanolamines both in
aqueous and non-aqueous solutions-II. Tertiary amines. Chem. Eng Sc. 1988, 43, 587-591.
Versteeg, G.F.; van Swaaij, W.P.M. Solubility and diffusivity of acid gases (CO2, N2O) in
aqueous alkanolamine solutions. J. of Chem. and Eng Data 1988b, 33, 29-34.
Wilke, C.R.; Chang, P. Correlation of diffusion coefficients in dilute solutions. Am. Inst.
Chem. Eng. J. 1955, 1, 264-270.
Xu G-W.; Zhang C-F.; Qin S-J.; Wang Y-W. Kinetics study on absorption of carbon dioxide
into solutions of activated methyldiethanolamine. Ind. Eng. Chem. Res. 1992, 31, 921-927.
Yu, W.C.; Astarita, G.; Savage, D.W. Kinetics of carbon dioxide absorption in solutions of
methyldiethanolamine. Chem. Eng Sc. 1985, 40(8), 1585-1590.
18
Table 1 : Conditions of CO2 absorption kinetics by MDEA aqueous solutions in the Lewis cell.
Reference T/K CMDEA,T/mol.m-3 P0/Pa PI/Pa PCO2,i/Pa Vg/10-6 m3 kL/10-5 m.s-1 b/10-4 s-1
number
1 295.95 844.4 177608 2790 174818 181.73 1.319 3.19
2 (d) 295.85 1272.6 178836 2840 175996 180.35 1.102 3.39
3 296.15 1984.1 188084 2730 185354 179.00 0.804 3.58
4 295.95 1984.2 158324 2720 155604 179.27 0.799 3.89
5 (d) 295.75 2583.2 190252 2660 187592 182.22 0.591 3.36
6 (d) 295.25 3477.4 175169 2710 172459 181.30 0.365 2.8
7 296.45 4379.1 114916 2830 112086 177.84 0.238 2.66
8 296.45 4379.1 135463 2424 133039 181.12 0.238 2.84
9 296.45 4379.1 134244 5730 128514 179.00 0.238 2.59
10 296.45 4379.1 127511 5890 121332 179.00 0.238 2.6
11 (g) 296.45 4379.1 133356 4690 128666 179.00 0.238 2.47
12 (a) 296.45 4379.1 198841 2820 196021 181.49 0.238 2.22
13 (a) 296.45 4379.1 178580 2816 175764 181.29 0.238 2.36
14 (b) 296.45 4379.1 118458 2500 116048 179.65 0.150 2.28
15 (c) 296.45 4379.1 167527 2300 165227 181.61 0.312 2.3
16 (d) 317.75 838.0 173484 9650 163834 183.30 2.168 5.36
17 (d) 318.25 1262.1 183039 9600 173439 180.21 1.985 5.84
18 317.55 1966.6 166516 9415 157101 177.48 1.493 6.6
19 318.35 2556.7 176797 9460 167337 179.52 1.197 6.28
20 318.15 3435.8 172549 9440 163109 179.60 0.816 5.75
21 317.65 4324.1 113864 8942 104922 179.83 0.543 5.22
22 (d) 343.45 826.7 178912 31800 147112 178.25 3.619 9.62
23 (e,f) 343.45 826.7 191684 32630 159054 168.83 3.619 9.74
24 (d) 343.55 1245.0 181652 32000 149652 179.04 3.205 10.65
25 342.55 1939.2 173218 33950 139268 176.87 2.527 12.32
26 342.65 1939.0 175590 33660 141930 178.88 2.531 11.86
27 342.65 1939.0 176608 34550 142058 174.32 2.531 11.89
28 342.65 1939.0 172657 31250 141407 177.33 2.531 11.13
29 342.55 1939.2 172043 33750 138293 176.61 2.527 11.46
30 (e,f) 343.55 1937.9 182505 31745 150760 170.27 2.571 12.34
31 (e,f) 343.55 1937.9 182365 31800 150565 168.38 2.571 12.49
32 (d) 342.75 2519.8 182503 29220 153283 180.31 2.080 10.93
33 (d) 342.95 3381.4 188045 29060 158985 180.22 1.534 10.56
34 342.25 4251.2 127278 28520 100070 179.76 1.098 10.25
(a) : MDEA from Alfa

(b) : liquid-side stirring speed = 50 RPM
(c) : liquid-side stirring speed = 150 RPM
(d) : volume of cell Vg+Vl = 350.8 cc
(e) : volume of cell Vg+Vl = 351.4 cc; others with Vg+Vl =350.4 cc
(f) : gas phase stirred at 680 RPM
(g) : MDEA from Merck
19
Figure 1. The Lewis type stirred reactor and its flow diagram. (AB) air bath, (B) baffles, (MP)
micro-computer, (MR) magnetic rod, (MS) magnetic stirrer, (P) propeller, (PP) platinum
probe, (PT) pressure transducer, (RLI) temperature control liquid inlet, (RLO) thermostatic-
liquid outlet, (RT) Rushton turbine, (TED) thermal electronic display, (TDE) transparent
thermostated double envelope, (TJ) thermostated jacket, (Vi) shut-of valve i, (VP) vacuum
pump.
20
Table 2 : Predicted kinetic constants at 296 K, 318 K and 343 K from literature
constants and energies of activation. Comparison to values from equation (11).
Reference T/K CMDEA/103 mol.m-3 PCO2/105 Pa Ea/kJ.mol-1 k/10-3 m3.mol-1.s-1
296 K 318 K 343 K
Yu et al. 313-333 0.2-2.5 1 38.5 - 15.4 -
(1985)
Haimour et al. 288-308 0.85-1.7 1 71.6 2.02 14.6 -
(1987)
Critchfield 282-350 1.7 1 56.9 2.24 11.0 51.6
(1988)
Versteeg and 293-333 0.17-2.7 <1 42.3 4.02 13.0 -
van Swaaij
(1988)
Toman and 298-308 4.3 0.02-0.12 - - - -
Rochelle
(1989)
Tomcej et al. 298-308 1.7-3.47 0.95 42.7 4.91 16.0 -
(1989)
Littel et al. 293-333 0.17-2.7 <1 48.1 4.70 17.8 -
(1990)
Rinker et al. 293-342 0.85 1 37.8 5.57 16.2 46.4
(1995)
This work 296-343 0.84-4.4 1-1.7 44.3 4.55 16.3 57.0
14
12
10
8
b / (s-1)
0
0 1000 2000 3000 4000 5000
-3
CMDEA,T (mol.m )
Figure 2 : Initial slopes b as a function of the total MDEA concentration at 296 K, 318 K and
343 K from experiments present in Table 1
21
160
140
120
100
kov / (s-1)
80
60
40
20
0
0 1000 2000 3000
CMDEA,T (mol.m-3)
Figure 3 : Influence of the interfacial MDEA concentration on kov at 296 K, 318 K and 343
K.
22
-2.5
-3
-3.5
ln (k / m3.mol-1.s-1)
-4
-4.5
-5
-5.5
0.0029 0.003 0.0031 0.0032 0.0033 0.0034
1/(T/K)
Figure 4 : Arrhenius plot.
23
Appendix A
CO2 absorption kinetics
T = 295.95 K, solvent = MDEA + H2O
Initial concentration of MDEA = 0.8444mol/l
initial CO2 loading : 0 mol CO2 / mol MDEA
final CO2 loading : 0.0522 mol CO2 / mol MDEA
t/s P/Pa t/s P/Pa t/s P/Pa
0.00 177608.0 360.31 158309.0 720.02 141850.0

10.49 176473.0 370.26 157872.0 730.51 141414.0
20.38 175863.0 380.14 157392.0 740.40 140983.0
30.32 175194.0 390.08 156862.0 750.29 140513.0
40.21 174673.0 399.97 156365.0 760.17 140139.0
50.15 174134.0 410.46 155965.0 770.00 139756.0
60.04 173628.0 420.40 155454.0 780.49 139285.0
69.98 173022.0 430.29 154996.0 790.33 138901.0
80.47 172471.0 440.23 154523.0 800.21 138412.0
90.36 171892.0 450.12 154050.0 810.10 138010.0
100.24 171371.0 460.06 153625.0 819.99 137684.0
110.18 170906.0 469.95 153196.0 830.48 137117.0
120.07 170407.0 480.44 152614.0 840.36 136668.0
130.01 169826.0 490.38 152195.0 850.19 136240.0
140.01 169282.0 500.26 151647.0 860.08 135964.0
150.44 168783.0 510.21 151256.0 869.97 135496.0
160.33 168262.0 520.09 150819.0 880.46 135038.0
170.27 167681.0 530.03 150348.0 890.35 134669.0
180.16 167246.0 540.52 149834.0 900.23 134137.0
190.10 166718.0 550.41 149508.0 910.12 133761.0
199.99 166223.0 560.35 149025.0 920.01 133474.0
210.48 165684.0 570.24 148557.0 930.44 132924.0
220.36 165259.0 580.13 148091.0 940.33 132570.0
230.31 164704.0 590.01 147633.0 950.21 132087.0
240.19 164189.0 599.95 147202.0 960.10 131757.0
250.13 163727.0 610.45 146674.0 969.99 131237.0
260.08 163209.0 620.39 146366.0 980.48 130862.0
269.96 162733.0 630.27 145765.0 990.36 130456.0
280.40 162220.0 640.16 145386.0 1000.20 130141.0
290.34 161659.0 650.05 145035.0 1010.08 129643.0
300.28 161277.0 659.99 144524.0 1019.97 129314.0
310.17 160751.0 670.48 144025.0 1030.46 128874.0
320.11 160235.0 680.42 143540.0 1040.46 128446.0
330.00 159743.0 690.31 143148.0 1050.29 128036.0
340.49 159310.0 700.19 142682.0 1060.18 127687.0
350.43 158847.0 710.08 142288.0 1070.06 127312.0
1079.95 126986.0 1600.37 107828.0 2120.46 91469.0

1090.49 126430.0 1610.25 107463.0 2130.34 91176.7
1100.38 126074.0 1620.20 107138.0 2140.28 90922.3
1110.27 125699.0 1630.14 106828.0 2150.17 90623.6
1120.10 125359.0 1640.08 106496.0 2160.11 90325.1
24
1130.04 124945.0 1649.96 106197.0 2170.05 90018.9
1140.48 124553.0 1660.40 105872.0 2179.94 89796.2
1150.36 124103.0 1670.34 105473.0 2190.43 89363.9
1160.25 123737.0 1680.23 105153.0 2200.32 89105.8
1170.14 123439.0 1690.17 104822.0 2210.31 88838.9
1180.02 123012.0 1700.06 104482.0 2220.20 88524.6
1190.51 122580.0 1709.94 104168.0 2230.09 88216.6
1200.40 122212.0 1720.43 103834.0 2240.03 87987.9
1210.29 121716.0 1730.38 103492.0 2250.52 87661.4
1220.17 121443.0 1740.26 103232.0 2260.41 87401.8
1230.06 121001.0 1750.15 102890.0 2270.35 87127.9
1239.95 120737.0 1760.09 102585.0 2280.23 86889.1
1250.44 120334.0 1769.98 102231.0 2290.18 86598.3
1260.27 119940.0 1780.47 101898.0 2300.06 86354.0
1270.16 119499.0 1790.35 101580.0 2310.00 86059.8
1280.04 119205.0 1800.24 101328.0 2320.50 85761.3
1290.53 118784.0 1810.18 100949.0 2330.38 85451.4
1300.42 118467.0 1820.07 100721.0 2340.32 85185.9
1310.31 118094.0 1829.96 100331.0 2350.21 84912.4
1320.19 117685.0 1840.45 99996.7 2360.15 84639.6
1330.08 117334.0 1850.39 99708.2 2370.09 84392.3
1339.97 116976.0 1860.33 99242.7 2379.98 84070.7
1350.40 116590.0 1870.22 99103.9 2390.47 83823.4
1360.29 116226.0 1880.10 98698.8 2400.36 83565.4
1370.23 115851.0 1890.04 98499.0 2410.30 83327.4
1380.12 115446.0 1900.54 98144.4 2420.24 83086.7
1390.00 115117.0 1910.48 97817.8 2430.13 82731.6
1399.94 114783.0 1920.36 97478.7 2440.07 82414.4
1410.43 114431.0 1930.25 97164.5 2449.95 82252.8
1420.32 114094.0 1940.19 96920.9 2460.45 81998.4
1430.21 113802.0 1950.08 96554.5 2470.39 81721.1
1440.15 113400.0 1960.02 96304.5 2480.27 81430.0
1450.04 113086.0 1970.51 95916.2 2490.22 81170.8
1460.53 112632.0 1980.40 95626.8 2500.10 80931.1
1470.41 112308.0 1990.28 95379.6 2509.99 80680.9
1480.30 112001.0 2000.28 95070.3 2520.48 80378.6
1490.24 111572.0 2010.17 94728.9 2530.42 80136.2
1500.13 111263.0 2020.11 94417.2 2540.31 79926.3
1510.07 110897.0 2029.99 94119.0 2550.25 79618.6
1519.96 110589.0 2039.94 93827.0 2560.14 79394.5
1530.45 110174.0 2050.43 93505.7 2570.02 79160.8
1540.33 109879.0 2060.37 93255.1 2579.96 78844.7
1550.28 109557.0 2070.25 92979.3 2590.45 78607.4
1560.16 109104.0 2080.20 92668.3 2600.34 78366.5
1570.10 108848.0 2090.08 92364.8 2610.28 78093.9
1579.99 108510.0 2100.02 92127.0 2620.17 77930.6
1590.48 108135.0 2109.97 91752.4 2630.11 77618.1
2640.00 77286.1
25
APPENDIX B.
The Lewis cell correlation for mass transfer.
A mass-transfer correlation between dimensionless numbers has been established for

our apparatus from N2O absorption experiments by the same MDEA-water solutions:
Sh = 0.340 Re 2/ 3 Sc1/ 3 (B1)
2
d N D Ag
Re = is the stirrer Reynolds number

Sc = is the Schmidt number,
d Dj
k L j DT
Sh = is the Sherwood number.
Dj
The transfer correlation (B1) is established for similar ranges of Re and Sc than those
for experiments of CO2 absorption by MDEA.
DAg (m) is the stirrer diameter, DT (m) is the internal diameter of the Lewis cell. kLj (m.s-1)
and Dj (m2.s-1) are the mass transfer and the diffusion coefficient of a species j.
The viscosity (Pa.s) and the density d of the mixture are estimated using the
correlations of Al-Ghawas et al. (1989).
As did Glasscock et al. (1991), the work and correlations of Versteeg and van Swaaij
(1988b), based on the analogy between CO2 and N2O, were used to estimate the diffusion
coefficient DCO2 (m2.s-1) of CO2, although their study is limited to temperatures between
(293 and 333) K, and MDEA concentrations between (840 and 2960) mol.m-3. Their
measurements were achieved in a stirred vessel with horizontal gas-liquid interface, similar to
26
our apparatus, and not a wetted sphere or a laminar-jet apparatus like in Al-Ghawas et al.
(1989). The latter study, achieved with MDEA concentrations between (0 and 4400) mol.m-3,
is limited to temperatures between (288 and 323) K. It must be pointed out that, for the
maximum temperatures and concentrations in our study for which Versteeg and Al-Ghawas
3
works are both available, i.e. 2600 mol.m- MDEA and 296 K or 318 K, diffusion coefficients
from Versteeg are half the values of Al-Ghawas et al (1989).
The work of Versteeg and van Swaaij (1988) also allows an estimate of the ratio of
CO2 to MDEA diffusion coefficients, used to calculate the MDEA depletion in the interfacial
boundary layer :
0 .2
DCO2 DCO2 Water
= (B2)
D MDEA Aq sol D MDEA Water Aq sol
As did Bosch et al. (1989), in agreement with Wilke and Chang (1955) correlation, the ratio
of diffusion coefficients at T in water (subscript w) is related to the ratio of molal volumes of
species at their normal boiling points :
0.6
DCO2 v
= MDEA (B3)
D MDEA Water v CO2
From Le Bas tables in Reid et al. (1977) :
vCO2 = 34 10-6 m3.mol-1 and vMDEA = 148.9 10-6 m3.mol-1.
Thence :
DCO2
= 2.43
D MDEA Water
27
Chapter 2.
Absorption of H2S by an Aqueous Methyldiethanolamine Solution at 296 and 343 K
F. Pani, A. Gaunand, D. Richon, R. Cadours, and C. Bouallou
Introduction
The removal of acid gases, such as H2S and CO2, from natural and industrial gases is a
frequently encountered operation in the process industry. A common removal method is the
chemical absorption into an alkanolamine solution. Among the industrially important
alkanolamines are the tertiary amines such as the N-methyldiethanolamine (MDEA). MDEA
has been found to be an efficient chemical solvent for a selective absorption of H2S from gas
mixtures containing CO2.
The following workers have measured the absorption kinetics of carbon dioxide and
hydrogen sulphide in aqueous methyldiethanolamine solutions : Littel et al. (1992),
Blauwhoff et al. (1984), Xu et al. (1992), Haimour et al. (1987) and Haimour and Sandall
(1987).
The objective of this work was to study the absorption of H2S into fresh and loaded
MDEA aqueous solutions of initial MDEA composition of 50 mass %. At this concentration,
for which no literature results are available, we measured H2S absorption rates at 296 and 343
K, for H2S loadings up to 0.44 mol of H2S per mol of MDEA. These results were used to
determine MDEA diffusion coefficients.
Experimental apparatus
The reaction cell, shown in Figure 1, with a known interfacial area A = (11.72 0.05)
10-4 m2, is the same as that used by Pani et al. (1997). It is composed of a double walled
28
glass cylinder closed at both ends by two metallic flanges. The upper flange holds a SEDEME
pressure transducer with a range from 0 to 2 bars, and a tube for introducing the acid gases.
The lower part of the cell is equipped with four vertical baffles, two concentric PTFE rings
placed midway between the bottom and the top of the cell, and a Rushton turbine [diameter,
Dag, = (4.25 0.02) 10-2 m] which is magnetically driven by an adjustable rotating field.
The lower flange holds a special inlet closed by means of a septum which allows the
introduction of a needle to feed the cell with the solvent and to load the solvent with H2S. The
total volume of the cell available for the gas and liquid phases is (0.3504 0.0005) 10-3 m3.
A thermostatic liquid is circulated inside the double walled glass cylinder to control the
experimental temperature within 0.05 K. The pressure transducer is thermostated at a
temperature slightly higher than the experimental temperature (by about 2 K) to avoid any
liquid condensate in its measuring chamber. The temperature is measured within 0.05 K by
means of a platinum probe going through the lower metallic flange as seen on Figure 1
(chapter 1.). The whole absorption cell is contained inside an air thermostat.
A microcomputer equipped with a Strawberry Tree Computers data acquisition card

is used to record pressure as a function of time. The electrical signal from the pressure
transducer is directly converted into pressure units (Pa) by the computer. Pressures are
measured within 200 Pa through the pressure transducer calibrated at each working
temperature.
To work with loaded solutions, the equipment is fitted with a High Pressure Reservoir,
HPR (see Figure 1, Chapter 1.), containing the acid gas to be added to the absorption cell.
Experimental procedure
Water and MDEA are degassed independently and aqueous solutions are prepared under
a vacuum. The respective masses of water and methyldiethanolamine are determined by
differential weighings to within 10-2 g. This uncertainty on weighings leads to uncertainties
in concentrations less than 0.05%.
The flask containing the degassed MDEA aqueous solution is connected to the
29
absorption cell by means of a needle introduced through the septum situated at the bottom of
the cell (see Figure 1a). The tube between the gas cylinder and the cell, the cell and the tubing
up to the Mohrs clip, MC1, are evacuated, then valves V1 and V3 are closed and Mohrs clip,
MC1, is opened to let the aqueous alkanolamine solution flow down by gravity under a
vacuum into the cell. Weighing the flask with tube and needle before and after transfer,
allows the determination of the exact mass of solvent transferred into the cell.
Once amine aqueous solution is loaded and temperature equilibrated, the inert gas
pressure PI corresponding mainly to the solvent vapor pressure plus eventual residual inert
gases is measured. Then H2S is introduced during a very short time (about 2 s) in the upper
part of the cell, the resulting pressure being slightly higher than atmospheric pressure. Stirring
is started at 100 rpm as soon as valve V1 is closed. Pressure decay versus time is recorded as a
result of the H2S absorption through the horizontal gas-liquid interface. This procedure
corresponds to the study of an initial absorption kinetic. Then, after reaching equilibrium the
liquid solvent is loaded with H2S. For these purposes the solvent loading circuit, a, is removed
and replaced by the H2S loading circuit, b. This loading circuit is evacuated prior the transfer
by opening MC2 and MC3 while valve V4 is closed and the extremity of the needle is closed
by the septum media. Then MC3 is closed, V4 is opened and the extremity of the needle is
pushed inside the liquid phase. H2S is bubbled inside the liquid phase, through the needle,
while the Rushton turbine is rotated at full speed, to improve the gas dispersion and
consequently the solubilization rate. An estimate of the absorbed H2S mass is controlled by
continuously weighing the high pressure cell. At the end of the process, the HPR cell is
disconnected from the absorption cell and weighed within 10-2 g. This mass is compared with
that before transfer to get an accurate value of the mass of H2S transferred. A new absorption
experiment for this given loading is done consecutively at the normal Rushton turbine
rotation speed.
Theory
It seems generally accepted that reactions which only consist in the exchange of a
hydrogen ion between a solute molecule and the solvent or another solute molecule, are
sufficiently fast to be considered as instantaneous with respect to mass transfer (Maddox et al.
30
1987). The reaction between H2S and aqueous alkanolamine involves only one proton transfer
:
MDEA + H 2 S MDEAH + + HS - (I)
This reversible reaction can be considered to be infinitely fast and hence the absorption
rate is entirely mass transfer controlled under practical conditions. Equilibrium is then
reached everywhere in the solution :
C MDEAH + C HS
K eq = (1)
C MDEA C H 2S
Keq is used as the equilibrium constant corresponding to reaction I but defined in terms
of concentrations instead of activities. Cj are the concentrations of species j. Using the
stagnant boundary layer model for hydrodynamic at gas-liquid interface and Ficks law for
mass transfer of the species involved in reaction I leads to :
d 2 C MDEA d 2 C H 2S d 2 C MDEAH + d 2 C HS
D MDEA = D H 2S = D MDEAH + = D HS (2)
d2z d2z d2z d2z
where Dj are the diffusivities of species j and z the distance from the gas-liquid interface
in the boundary layer. As the gas phase inside the cell is almost a pure H2S atmosphere, the
resistance in the gas phase can be neglected. Assuming that the mean diffusivities of ionic
species are equal to MDEA diffusivity, solving eq. 2 with the equilibrium eq. 1 assumed all
through the layer leads to :
D
( )
H 2S = k L C H 2S,i C H 2S,b 1 + MDEA

D H 2S C H 2S ,i
(3)
Where Cj,i and Cj,b are interfacial and bulk concentrations of species j, kL is the liquid
side mass transfer coefficient for our apparatus and :
31
1
1
(
= C H 2S,i K eq ) 1
+ 4C H 2S ,i K eq C MDEA ,b K eq C H 2S,i
2 2
(4)
2 2
A mass transfer correlation between dimensionless numbers has been established for
our apparatus from the Pani et al. (1997) data of N2O absorption in MDEA aqueous solutions:
Sh = 0.25 Re 0.63 Sc 0.42 (5)
Sh=kLDT/DCO2 is the Sherwood number, Re=dNDAg2/m is the stirrer Reynolds number

and Sc=m/dDCO2 is the Schmidt number. N is the rotation speed of the Rushton turbine: 1.67
Hz; DT is the cell internal diameter (6.000.02) x 10-2 m; d is the density and the viscosity.
The correlation, presented in Pani et al. (1997), included absorption data of N2O by MDEA
aqueous solutions and of CO2 by water. As CO2 absorption may be enhanced by its reaction
with hydroxide ions, only the N2O absorption in MDEA solutions experiments have been
considered here, leading to Eq. 5. For Pani et al. (1997) results, taking one or the other of the
two mass transfer correlation is of no influence on their conclusions because CO2 absorption
in MDEA aqueous solutions takes place according to a fast regime of reaction, insensitive to
the liquid side mass transfer resistance.
Viscosity and density were determined using the correlations proposed by Al-Ghawas et
al. (1989). The diffusivity of H2S into aqueous MDEA solutions was calculated with the
Haimour and Sandall (1984) correlation for the diffusivity of H2S into water on the
temperature range (288 - 303 K) :
(D H 2S / m2 .s-1 ) ( / Pa.s) 0.74
= 3.4756 1014 (6)

(T / K)
The interfacial H2S concentration is at equilibrium with the vapor phase
PH 2S
C H 2 S ,i = (7)
H H 2S
where HH2S is the apparent Henrys law constant of H2S. When no experimental data are
32
available, van Krevelen and Hoftijzer (1948) suggest the use of eq. 8 to estimate the effect of
salt concentration on the solubility by means of a contribution method. For our solutions :
H H 2S , 0
log
H H 2S
( )
= h MDEAH + + h HS + h H 2S C MDEAH + (8)
HH2S is the Henrys constant of the gas in the actual electrolyte solution as compared
to that in pure water HH2S,0. The parameters : hMDEAH+ and hHS- are specific to the cation and
the anion, and are considered as temperature independent. In some cases, estimations holds
very well even up to salt concentrations as high as 8 kmol.m-3 while, in other cases,
significant deviations occur at concentrations lower than 1 kmol.m-3, e.g. because of the
concentration dependency of the dissociation degree. Generally, the salting-out effect tends to
be overestimated at high electrolyte concentrations, i.e. the predicted values for H tend to be
higher than values obtained without taking into account the ionic strengh influence. For the
absorption of H2S into the amine solution, an instantaneous reaction planar region is formed
near the gas-liquid interface as a result of the reaction between H2S and MDEA. At the gas-
liquid interface, MDEAH+ and HS- ions are present. The salting-out parameter for MDEAH+
is assumed to be zero as done by Haimour and Sandall (1987) following Danckwerts (1970).
The value of the salting-out parameter for HS- is assumed to have a similar value to that of
OH-, i.e. 0.066 m3.kmol-1. The salting-out parameter for gaseous H2S is taken as - 0.033
m3.kmol-1 [Danckwerts (1970)].
The bulk concentrations are calculated from the total MDEA concentration and the
initial H2S loading of each experiment. Keq is estimated using the correlation of Kent and
Eisenberg (1976). The equilibrium constants used at 296 and 343 K are respectively 44.6 and
19.3.
Taking into account the uncertainty on several of the above parameters for concentrated
and loaded solutions, it was decided to use the absorption rate data obtained with our cell to
determine the MDEA diffusion coefficient :
33
C H 2 S ,i H 2S
D MDEA = D H 2S 1 (9)
(2 2
)
k L C H S ,i C H S , b
The H2S absorption rate H2S was determined from pressure-time data by :
H2S RTA dPT

= = (10)
VG dt
is the slope of the pressure vs time curve, at the beginning of the absorption process.
Results and discussion
Reagent grade H2S is supplied by Alpha Gaz with a purity certified higher than 99.7%.
Water has been distilled twice and MDEA was from two different sources : Aldrich (purity
certified higher than 99%) and Merck (purity certified higher than 98%). Liquids are carefully
degassed prior to use.
The conditions of the absorption experiments are given in Table 1. All rough results of
pressure-time absorption data (25 tables) are compiled as Supporting Information. The first
table (table A1) is given in the Appendix. Table 1 gives in column 1 the reference number of
experiment, column 2 the temperature of experiment, column 3 the initial pressure, column 4
the solvent+inert gas pressure, column 5 the initial H2S partial pressure, column 6 the vapor
phase volume inside the absorption reactor, column 7 the H2S loading and column 8 the H2S
absorption rate H2S derived from eq 10.
A good reproducibility for initial absorption is found between two experiments at 296 K
with unloaded 50 mass % MDEA solutions using MDEA from Aldrich and two others with
MDEA from Merck (see tables A1, A2, A3 and A4). No important deviation is found between
MDEA from the two suppliers, despite of their slight differences in purity. Reproducibility
has been tested and demonstrated at 343 K with MDEA provided by Aldrich (see tables A6
and A7).
From the results at 296 and 343 K, we note that the H2S absorption rate is an increasing
function of temperature (see tables A2 and A7).
34
Table 2 shows the values of the MDEA diffusion coefficient in a 50 mass % solution,
for loadings ranging from 0 to 0.44 mol H2S per mol MDEA. Figure 2 displays the slight
influence of H2S loading on MDEA diffusivity at 296 K and for 50 mass % MDEA.
The diffusion coefficients of MDEA in water were measured by Rowley et al. (1997) at
20, 35 and 50 mass % MDEA and at 298, 323 and 348 K. The diffusion coefficients of
MDEA in water have also been measured by Snijder et al. (1993). Figure 3 shows a
comparison between these literature values and ours. This comparison reveals that the three
diffusion coefficient sources are in good agreement at 298 K. The correlation used to estimate
DH2S is given for a 15 K temperature range below 303 K. Then our extrapolation up to 343 K
is very doubtful and it is not surprising if there is a strong disagreement between the direct
data from the literature and our indirect evalutions (eq. 9). The estimation method of
diffusivities in aqueous amine solutions requiring a lot of assumptions appears as far from
reliable and is certainly the point for which supplementary work is necessary.
Conclusion
Acid gas absorption kinetics have been determined using an especially designed
equipment where total pressure is recorded as a function of time as in Laddha and Danckwerts
(1982). This cell can be also used as a perfectly stirred reactor to quickly increase the acid gas
loading of the investigated solution and then determine a series of isothermal absorption
kinetics in a short period of time. Reproducibility of measurements is quite good (better than
10%). The cell can be used either for CO2 or H2S absorption and whatever the solvent: water
with single or mixed amines for a large range of concentrations.
Acknowlegments - The authors gratefully acknowledge the financial support of the Gas
Research Institute.
35
Literature cited
Al-Ghawas, H.A.; Hagewiesche, D.P.; Ruiz-Ibanez, G.; Sandall, O.C. Physicochemical

Properties Important for Carbon Dioxide Absorption in Aqueous Methyldiethanolamine, J.
Chem. Eng. Data 1989, 34, 385-391.
Blauwhoff, P. M. M.; Versteeg, G. F.; van Swaaij W. P. M., A Study of the Reaction between
CO2 and Alkanolamines in Aqueous Solutions, Chem. Eng. Sci. 1984, 39, 207-225.
Danckwerts, P.V. Gas Liquid Reactions, McGraw-Hill Book Co, New York, NY, 1970.
Haimour, N.; Bidarian, A.; Sandall, O. C. Simultaneous Absorption of H2S and CO2 into
Aqueous Methyldiethanolamine, Separation Science and Technology 1987, 22, 921-947.
Haimour, N.; Sandall, O. C. Molecular Diffusivity of Hydrogen Sulfide in Water, J. Chem.

Eng. Data. 1984, 29, 20-22.
Haimour, N.; Sandall, O. C. Absorption of H2S into Aqueous Methydiethanolamine, Chem.

Eng. Comm. 1987, 59, 85-93.
Kent, R. L.; Eisenberg, B. Better Data for Amine Treating, Hydrocarbon Processing 1976,
55, 87-90.
van Krevelen, D. W.; Hoftijzer, P. J., Sur la solubilit des gaz dans les solutions aqueuses,
Chimie et industrie : numro spcial du XXIe Congrs International de Chimie Industrielle,
Bruxelles, 1948, 168-173.
Laddha, S.S.; Danckwerts, P.V. The Absorption of CO2 by Amine-Potash Solutions, Chem.
Eng. Sc. 1982, 37, 665-667.
Littel, R. J.; Versteeg, G. F.; van Swaaij, W. P. M., Kinetics of CO2 with Primary and
Secondary Amines in Aqueous Solutions I. Zwitterion Deprotonation Kinetics for DEA
and DIPA in Aqueous Blends of Alkanolamines, Chem. Eng. Sci. 1992, 47, 2027-2035.
36
Maddox, R. N.; Mains, G. J.; Rahman, M. A. Reactions of Carbon Dioxide and Hydrogen
Sulfide with Some Alkanolamines, Ind. Eng. Chem. Res. 1987, 26, 27-31.
Pani, F.; Gaunand, A.; Cadours, R.; Bouallou, C.; Richon, D. Kinetics of Absorption of CO2
in Concentrated Aqueous Methyldiethanolamine Solutions in the Range 296-343 K, J. Chem.
Eng. Data 1997, 42, 353-369.
Rowley, R.L.; Adams, M.E.; Marshall, T.L.; Oscarson, J.L.; Wilding, W.V.; Anderson, D.J.
Measurement of Diffusions Coefficients Important in Modeling the Absorption Rate of
Carbon Dioxide into Aqueous N-Methyldiethanolamine, J. Chem. Eng. Data 1997, 42, 310-
317.
Snijder, E.D.; te Riele, M.J.M.; Versteeg, G.F.; van Swaaij, W.P.M. Diffusion Coefficients of
Several Aqueous Alkanolamine Solutions, J. Chem. Eng. Data 1993, 38, 475-480.
Xu, G.-W.; Zhang, C.-F.; Qin, S.-J.; Wang, Y.-W. Kinetics study on absorption of carbon
dioxide into solutions of activated methyldiethanolamine, Ind. Eng. Chem. Res. 1992, 31,
921-927.
37
Table 1 : Conditions of H2S absorption kinetics by MDEA solutions
ref no. T/K P0 / Pa PI / Pa PH2S,i / Pa VG / 10-6 m3 a / molH2S/molMDEA H2S / mol.m-2.s-1
1 296.05 154877 2750 152127 181.58 0 0.01155

2 295.85 173044 2700 170344 182.26 0 0.01253
3 295.95 190584 2780 187804 180.91 0 0.01284
4* 295.95 169062 2600 166462 181.84 0 0.01201
5 295.95 159321 2810 156511 184.26 0 0.01168
6** 343.05 196055 36800 159255 178.35 0 0.02930
7** 343.05 176141 38400 137741 179.54 0 0.02606
8 296.05 68473 2750 65723 181.58 0.0116 0.00720

9* 295.95 172462 2780 169682 180.91 0.0118 0.01172
10 295.85 169275 2810 166465 184.26 0.0829 0.01142
11 295.95 98046 2810 95236 184.26 0.1217 0.00793
12 295.75 163939 2810 161129 184.26 0.1736 0.00972
13 295.85 161414 2810 158604 184.26 0.2365 0.00751
14 295.85 175902 2810 173092 184.26 0.3246 0.00699
15 295.85 198543 2810 195733 184.26 0.3796 0.00610
16 295.75 179846 2810 177036 184.26 0.4421 0.00550
17** 343.05 197858 38400 159458 179.54 0.011 0.01864
18** 343.05 191696 38400 153296 179.54 0.0197 0.01777
19** 343.05 197581 38400 159181 179.54 0.027 0.01446
20** 343.05 197765 38400 159365 179.54 0.0509 0.00593
21** 343.05 171268 38400 132868 179.54 0.0541 0.00266
22** 343.05 172605 38400 134205 179.54 0.0548 0.00289
23** 343.05 197072 38400 158672 179.54 0.0555 0.00621
24** 343.05 179629 38400 141229 179.54 0.0581 0.00278
25** 343.05 198160 38400 159760 179.54 0.0665 0.00334
* : MDEA from Merck instead of Aldrich

** : Volume of cell VG + VL = 343.5 cm3 instead of 350.4 cm3.
Table 2 : Diffusion coefficient of MDEA in 50 mass % MDEA solutions at 296 K
ref no. T/K kL / 10-5 m2.s-1 a / molH2S/molMDEA DMDEA / 10-9 m2.s-1
1 296.05 0.373 0 0.25

2 295.85 0.371 0 0.27
3 295.95 0.372 0 0.27
4 295.95 0.372 0 0.26
5 295.95 0.372 0 0.25
6 343.05 1.112 0 1.26
7 343.05 1.112 0 1.18
8 296.05 0.373 0.0116 0.20
9 295.95 0.372 0.0118 0.25
10 295.85 0.371 0.0829 0.27
11 295.95 0.372 0.1217 0.23
12 295.75 0.370 0.1736 0.26
13 295.85 0.371 0.2365 0.22
14 295.85 0.371 0.3246 0.23
15 295.85 0.371 0.3796 0.21
16 295.75 0.370 0.4421 0.18
38
0.40
DMDEA / 10-9 m2.s-1
0.20
0.00
0 0.1 0.2 0.3 0.4 0.5
/ molH2S/molMDEA
Figure 2: Loading influence on MDEA diffusivity
1.60
1.20
DMDEA / 10-9 m2.s-1
0.80
0.40
0.00
280 300 320 340 360 380
T/K
Figure 3: Comparison of DMDEA in 50 mass % MDEA aqueous solution with previous works
{ This work, z Rowley et al. (1997), Snijder et al. (1993)
39
Appendix. Rough Results
TABLE A1
H2S absorption kinetics
T = 296.05 K, solvent = MDEA + H2O
initial concentration of MDEA = 4.39 mol/l
initial H2S loading : 0 mol H2S / mol MDEA
final H2S loading : 0.0116 mol H2S / mol MDEA
0.00 154877 22.30 151140 44.55 147502

0.66 154719 22.91 150994 45.15 147351
1.27 154532 23.51 150945 45.81 147303
1.87 154477 24.12 150820 46.42 147145
2.53 154288 24.78 150728 47.02 147091
3.14 154253 25.38 150622 47.63 147006
3.74 154162 25.98 150496 48.28 146921
4.34 153958 26.59 150409 48.89 146785
5.00 153817 27.19 150326 49.49 146746
5.61 153731 27.85 150205 50.15 146616
6.21 153681 28.46 150104 50.76 146582
6.82 153539 29.12 150060 51.36 146448
7.47 153499 29.72 149927 51.96 146361
8.08 153368 30.32 149814 52.62 146252
8.68 153337 30.93 149712 53.23 146089
9.29 153180 31.53 149530 53.83 146057
9.95 153052 32.19 149542 54.44 145882
10.55 152903 32.80 149372 55.10 145744
11.15 152784 33.40 149381 55.70 145782
11.76 152869 34.06 149184 56.30 145672
12.36 152676 34.66 149059 56.96 145482
13.02 152644 35.27 148984 57.57 145436
13.63 152589 35.87 148962 58.17 145391
14.23 152404 36.48 148729 58.78 145270
14.83 152299 37.13 148643 59.43 145159
15.49 152180 37.74 148603 60.04 145038
16.10 152091 38.34 148466 60.64 145032
16.70 152035 39.00 148474 61.25 144854
17.31 151913 39.61 148264 61.91 144717
17.97 151754 40.21 148169 62.51 144595
18.57 151698 40.87 148029 63.11 144592
19.17 151542 41.47 147965 63.72 144522
19.78 151501 42.08 147918 64.38 144394
20.44 151381 42.68 147846 64.98 144244
21.04 151185 43.34 147734 65.64 144113
21.65 151208 43.95 147644 66.24 144104
40
66.85 143944 95.36 139484 123.86 135168

67.45 143878 96.01 139433 124.47 135052
68.06 143856 96.62 139278 125.07 134940
68.72 143727 97.22 139176 125.73 134866
69.32 143578 97.88 139128 126.33 134715
69.92 143421 98.49 139050 126.94 134733
70.53 143414 99.09 138813 127.54 134518
71.13 143283 99.69 138803 128.20 134405
71.79 143180 100.35 138717 128.81 134402
72.40 143096 100.96 138664 129.41 134222
73.00 143017 101.56 138514 130.01 134183
73.60 142938 102.17 138477 130.67 134051
74.26 142862 102.77 138354 131.28 134041
74.87 142722 103.43 138314 131.88 133929
75.47 142639 104.03 138180 132.49 133797
76.08 142549 104.64 138132 133.14 133745
76.68 142499 105.30 137931 133.75 133625
77.34 142291 105.90 137876 134.35 133511
77.94 142223 106.51 137766 134.96 133428
78.60 142098 107.11 137636 135.56 133324
79.21 141995 107.77 137598 136.22 133191
79.81 141920 108.37 137459 136.82 133157
80.42 141882 108.98 137461 137.43 133096
81.07 141727 109.64 137362 138.03 132995
81.68 141617 110.24 137210 138.69 132918
82.28 141546 110.84 137136 139.30 132773
82.89 141529 111.45 137059 139.90 132661
83.55 141401 112.11 136811 140.50 132608
84.15 141258 112.71 136789 141.16 132502
84.75 141164 113.32 136719 141.77 132353
85.41 141127 113.92 136603 142.37 132364
86.02 140896 114.58 136564 143.03 132253
86.62 140936 115.18 136467 143.63 132115
87.28 140752 115.79 136351 144.24 132047
87.89 140707 116.45 136277 144.90 131892
88.49 140514 117.05 136236 145.50 131865
89.09 140551 117.66 136030 146.11 131784
89.70 140383 118.26 135958 146.71 131613
90.41 140255 118.92 135803 147.37 131604
91.07 140121 119.52 135832 147.97 131479
91.68 140107 120.13 135714 148.58 131356
92.28 139923 120.79 135669 149.18 131264
92.88 139920 121.39 135478 149.84 131151
93.54 139886 121.99 135462 150.45 131074
94.15 139629 122.65 135254 151.05 130944
94.75 139540 123.26 135244 151.71 130914
41
152.31 130749 180.76 126495 209.22 122346

152.92 130734 181.42 126426 209.88 122110
153.58 130621 182.03 126293 210.48 122137
154.18 130559 182.63 126230 211.08 122031
154.78 130357 183.24 126119 211.69 121945
155.39 130255 183.90 126020 212.35 121840
156.05 130154 184.50 125974 212.95 121786
156.65 130091 185.10 125893 213.56 121635
157.26 130087 185.71 125759 214.16 121518
157.92 129955 186.37 125758 214.82 121450
158.52 129861 186.97 125569 215.42 121367
159.12 129751 187.58 125524 216.03 121310
159.73 129699 188.18 125420 216.69 121149
160.33 129479 188.78 125325 217.29 121119
160.99 129537 189.44 125229 217.89 121002
161.60 129298 190.05 125152 218.50 120937
162.20 129271 190.65 124997 219.16 120758
162.86 129219 191.31 124847 219.76 120721
163.46 129112 191.91 124880 220.37 120643
164.07 128985 192.52 124753 220.97 120588
164.67 128880 193.12 124739 221.63 120481
165.28 128832 193.73 124591 222.23 120414
165.93 128740 194.39 124538 222.84 120230
166.54 128649 194.99 124426 223.50 120146
167.14 128492 195.59 124299 224.10 120096
167.80 128370 196.20 124220 224.70 120030
168.41 128378 196.86 124106 225.31 119963
169.01 128249 197.46 124069 225.97 119830
169.67 128070 198.07 123835 226.57 119743
170.27 128106 198.67 123873 227.18 119733
170.88 127984 199.27 123733 227.78 119579
171.48 127878 199.93 123679 228.44 119485
172.14 127719 200.54 123594 229.04 119416
172.75 127724 201.14 123526 229.65 119300
173.35 127612 201.80 123347 230.31 119268
173.95 127580 202.41 123267 230.91 119150
174.56 127465 203.01 123179 231.52 119037
175.22 127339 203.67 123166 232.12 118991
175.82 127224 204.27 123003 232.72 118856
176.43 127117 204.88 122918 233.38 118798
177.08 127089 205.54 122845 233.99 118666
177.69 126974 206.14 122772 234.59 118588
178.29 126860 206.74 122636 235.25 118517
178.95 126645 207.40 122564 235.85 118416
179.56 126736 208.01 122453 236.46 118302
180.16 126489 208.61 122399 237.06 118240
42
237.72 118140 266.17 114050 294.68 110004

238.33 118101 266.83 113921 295.28 109903
238.93 117979 267.44 113840 295.94 109876
239.59 117847 268.04 113762 296.55 109776
240.19 117704 268.65 113722 297.15 109625
240.80 117625 269.25 113642 297.76 109626
241.46 117628 269.91 113565 298.41 109553
242.06 117510 270.51 113402 299.02 109404
242.67 117424 271.12 113319 299.62 109347
243.27 117302 271.78 113269 300.23 109224
243.87 117237 272.38 113172 300.89 109148
244.53 117139 272.98 112979 301.49 109054
245.14 117133 273.64 112874 302.09 108984
245.74 116933 274.25 112882 302.75 108856
246.40 116915 274.85 112821 303.36 108879
247.00 116819 275.46 112670 303.96 108763
247.61 116680 276.12 112637 304.62 108695
248.27 116696 276.72 112527 305.23 108550
248.87 116627 277.32 112397 305.83 108481
249.48 116440 277.98 112307 306.43 108387
250.08 116411 278.59 112228 307.04 108279
250.68 116366 279.19 112241 307.70 108234
251.34 116141 279.85 112123 308.30 108135
251.95 116117 280.45 112060 308.91 107997
252.55 116042 281.06 111921 309.51 107961
253.16 115907 281.66 111861 310.17 107805
253.76 115773 282.32 111788 310.77 107815
254.42 115671 282.93 111619 311.38 107645
255.02 115597 283.53 111565 311.98 107543
255.63 115505 284.13 111528 312.64 107538
256.23 115444 284.79 111368 313.24 107431
256.89 115398 285.40 111362 313.85 107345
257.50 115309 286.00 111226 314.45 107298
258.10 115122 286.66 111091 315.06 107215
258.70 115094 287.27 111087 315.72 107139
259.36 115016 287.87 110923 316.32 106903
259.97 115011 288.53 110811 316.92 107001
260.57 114807 289.13 110767 317.58 106861
261.23 114830 289.74 110740 318.19 106789
261.83 114713 290.34 110721 318.79 106643
262.44 114590 290.95 110532 319.40 106525
263.10 114461 291.60 110420 320.06 106466
263.70 114399 292.21 110310 320.66 106412
264.31 114356 292.87 110325 321.26 106231
264.97 114190 293.47 110202 321.92 106212
265.57 114081 294.08 110087 322.53 106138
43
323.13 106002 340.49 103708 357.84 101348

323.79 105954 341.09 103582 358.45 101253
324.39 105772 341.70 103449 359.05 101192
325.00 105825 342.36 103403 359.66 101048
325.60 105746 342.96 103348 360.32 100961
326.21 105626 343.56 103264 360.92 100902
326.87 105517 344.22 103171 361.52 100771
327.47 105445 344.83 103057 362.18 100752
328.07 105365 345.43 102980 362.79 100627
328.68 105261 346.09 102927 363.39 100622
329.34 105179 346.69 102795 364.05 100498
329.94 105059 347.30 102678 364.66 100482
330.55 105046 347.90 102627 365.26 100356
331.21 104909 348.51 102568 365.86 100222
331.81 104920 349.17 102509 366.52 100140
332.41 104798 349.77 102382 367.13 100068
333.07 104706 350.37 102301 367.73 99942
333.68 104609 350.98 102252 368.34 99845
334.28 104562 351.64 102161 368.99 99887
334.89 104432 352.24 102127 369.60 99704
335.54 104330 352.85 101954 370.20 99654
336.15 104177 353.51 101863 370.86 99584
336.75 104223 354.11 101832 371.47 99477
337.41 104055 354.71 101779 372.07 99452
338.02 104050 355.37 101722 372.67 99389
338.62 103934 355.98 101563
339.28 103782 356.58 101418
339.88 103719 357.19 101429
44
Chapter 3.
Kinetics of absorption of carbon dioxide in aqueous solutions of N-

methyldiethanolamine + ethanolamine or + diethanolamine at 296 K or 343 K.
F. Pani; C. Bouallou; R. Cadours; A. Gaunand; D. Richon
Introduction
Aqueous alkanolamine solutions are commonly used for the removal of acid gases from
natural gases. Accurate absorption data are necessary for the design of extraction processes.
N-methyldiethanolamine (MDEA) is a non-corrosive and non-volatile compound which
makes it a good candidate for acid gas removal. Furthermore it presents a high selectivity
with respect to H2S and CO2. Nevertheless, in some cases, the rate of acid gas absorption can
be improved by adding small quantities of a primary or a secondary amine, the latter acting as
an activator.
The aim of this work is to provide the engineer with tools for designing the most suitable
processes according to client specifications. For this purpose, we have measured the kinetics
of CO2 absorption in aqueous solutions of blends of the tertiary alkanolamine, N-
methyldiethanolamine [N-methyl-bis(2-hydroxyethyl)amine, MDEA] with a primary
alkanolamine, monoethanolamine [2-aminoethanol, MEA], or a secondary alkanolamine,
diethanolamine [bis(2-hydroxyethyl)amine, DEA].
Apparatus and procedure
We used a thermostated glass reactor of the Lewis type [LEWJ2540] with a constant gas-
liquid interfacial area of (13.64 0.05) 10-4 m2 [PANF1970]). The total volume available
for gas and liquid is (0.3504 0.0005) 10-4 m3. The reactor is provided with a Rushton
turbine in its lower part and with a propeller in its upper part. The turbine speed is checked
45
with a stroboscope, it remains constant within 1 rpm during each test. The temperature in the
reactor is known within 0.05 K through a 100 W platinum probe, calibrated against a 25 W
STHP-B platinum probe from LYON ALEMAND LOUYOT (Lyon, France). The
temperature is controlled by circulating a thermostatic fluid through the glass double jacket. A
tube allows either to evacuate or to introduce CO2 into the cell.
Kinetics of gas absorption are measured by recording the absolute pressure drop through a
SEDEME (Les Ulis, France) pressure transducer, working in the range (0 to 200) 103 Pa. For
each temperature T investigated, it is calibrated within 200 Pa against a mercury manometer.
A microcomputer equipped with a data acquisition card is used to convert the pressure
transducer signal directly into pressure P units, using calibration constants previously
determined, and record it as a function of time, t.
Water and amines are degassed independently and aqueous solutions are prepared under a
vacuum. The masses of water and amines are known by differential weighings to within 0.1
104 kg. Accurate weighings of the flask before and after transfer yield the mass of solution
actually present in the cell. The liquid phase volume was calculated using the density
correlation from [GLAD0900] for the aqueous solutions of amines. At a given temperature,
and under solution vapor pressure PI, pure CO2 is introduced during a very short time (about 2
s) in the upper part of the cell, the volume of which is noted as VG. The resulting pressure P0
is between 100 and 200 kPa. Then stirring is started and the pressure drop resulting from
absorption is recorded.
In order to determine the physical mass transfer coefficient kL,i of a species i, a mass-
transfer correlation between dimensionless quantities has been established for our apparatus
from N2O absorption experiments by the same MDEA-water solutions:
Sh = 0.25 Re0.63 Sc0.42
N D 2Ag
Re = is the stirrer Reynolds number

46

Sc = is the Schmidt number,
Di
k L DT
Sh = is the Sherwood number.
Di
DAg = (4.25 0.02) 10-2 m is the Rushton turbine diameter, DT = (6.00 0.02) 10-2 m
is the internal diameter of the Lewis cell, and Di (m2.s-1) the diffusion coefficient of a species
i.
To establish this mass transfer correlation, the viscosity (Pa.s) and the density r of the
mixture were estimated using the correlations [ALGH1890]. The diffusion coefficients were
obtained with the correlations proposed by [VERG3880].
A test of the influence of the gas phase stirring was done by [PANF1970]. They concluded
that the gas phase agitation has no significant influence, and they also neglected any gas
phase resistance.
Materials
H2O Water. Ordinary bidistilled water was used.
CO2 Carbon dioxide. Air Liquide( Saint Quentin en Yvelines, France), certified purity of
99.995 vol %, used as received.
C2H7NO 2-Aminoethanol (Ethanolamine, MEA). B. Merck AG (Darmstadt, Germany),

material of certified GLC purity, 98 %, used as received; r (298.15K)/kg.m-3 = 935.9 (1014.7
at 293.15 K [RIDJ0860]).
C4H11NO2 2,2-dihydroxyethylamine (Diethanolamine, DEA). B. Merck AG (Darmstadt,

Germany), material of certified GLC purity, 98 %, used as received; r (293.15K)/kg.m-3 =
1007.4 (1096.9 [RIDJ0860]).
C5H13NO2 2,2-Dimethoxy-N-methylamine N-Methyldiethanolamine, MDEA). Aldrich

Chem. Co Inc. (Milwaukee, WI, USA) material of stated purity 99 mole %, used as received;
r (293.15K)/kg.m-3 = 1043.1 (1038.0 at 298.15 K [LIDD1960]).
47
Densities were measured with an Anton-Paar (Graz, Austria) vibrating tube densimeter,
Model DMA-512.
Results
Our experimental time-pressure (t-P) measurements at 296 K and 343 K for CO2
absorption in aqueous solutions of N-methyl-bis(2-hydroxyethyl)amine (N-
methyldiethanolamine, MDEA) + 2-aminoethanol (ethanolamine, MEA) (from 402 to 1590
data points) or + bis(2-hydroxyethyl)amine (diethanolamine, DEA) (from 340 to 1755 data
points) are all saved on disk as Standard ELDATA Files PANF1960.001 and
PANF1960.002. Table I gives the experimental conditions of our measurements, temperature
T, mass fractions of amines w(amine), initial pressure P0, initial solution vapor pressure
PI,and initial loading Rc, i.e. the amount (mol) of CO2 divided by the total amount (mol) of
amines.
Discussion and Conlusion
Experiments achieved without initial loadings at 296 K can be compared to those achieved
by [PANF1970] with aqueous solutions only containing MDEA. The respective time-pressure
acquisition data shows that the addition of a small amount of MEA or DEA to an MDEA
aqueous solution significantly increases the CO2 absorption rate. For example, at 296.45 K
and under a total pressure of ca. 178 103 Pa, 41.4 % of CO2 are absorbed after 1800 s by a
blend of MDEA and MEA, instead of 35.9 % when using a 50 wt % MDEA solution (ref no.
13 in [PANF1970]).
When CO2 is absorbed into an aqueous solution of a tertiary alkanolamine R1R2R3N, and a
primary or a secondary R4R5N, the following reactions may occur:
CO 2 + R 1 R 2 R 3 N + H 2 O R 1 R 2 R 3 NH + + HCO 3 (1)
CO 2 + OH HCO 3 (2)
CO 2 + R 4 R 5 NH R 4 R 5 NCOO + H + ( 3)
HCO 3 = CO 3= + H + (4)
R 1 R 2 R 3 N + H + = R 1 R 2 R 3 NH + (5)
R 4 R 5 NH + H + = R 4 R 5 NH 2+ ( 6)
OH + H + = H 2 O (7)
R 4 R 5 NH 2+ + R 1 R 2 R 3 N = R 4 R 5 NH + R 1 R 2 R 3 NH + (8 )
48
Where R1=CH3 and R2=R3=C2H4OH for MDEA; R4=H and R5=C2H4OH for MEA; and
R4=R5=C2H4OH for DEA.
Reactions (1)-(3) are considered to be reversible with finite reaction rates, whereas
reactions (4)-(8) are considered to be instantaneous, with respect to mass transfer, i.e. at
equilibrium since they involve only a proton transfer. Only the first six equilibrium constants
Ki are independent. The remaining two can be obtained by an appropriate combination of the
first six ones :
K 2 K5
K7 =
K1
K5
K8 =
K6
For penetration or surface renewal theories. The physical mass transfer coefficients are
expressed as :
D CO
kL = 2 2
(penetration theory)

kL = D CO s (surface renewal theory)
2
The transfer with multiple chemical reactions is then represented by a system of partial
differential equations, where n is the number of components :
F Ci
D i 2 Ci z i Di ( Ci ) = + R i [ C]
kT t
where is the electrical potential gradient :
kT i i i
z D C
= i =1
n
F
z2i DiCi
i =1
For the film theory, the preceding partial differential equations turn into stationary ones :
49
F
D i 2 Ci zi Di ( Ci ) = R i [C]
kT
where the physical mass transfer coefficient is given by :
D CO
kL = 2
Such systems are usually solved by means of a numerical model [GLAD0900] [GLAD0910],
[HAGD0950], [CADR0970]. The comparison of measured absorption rates with calculated
ones obtained through these models allows to find the kinetic parameters of the previous
reactions.
50
Literature cited
[ALGH1890] - Al-Ghawas, H.A.; Hagewiesche, D.P.; Ruiz-Ibanez, G. ; Sandall, O.C.

Physicochemical properties important for carbon dioxide absorption in aqueous
methyldiethanolamine. J. Chem. Eng. Data 1989, 34, 385-391.
[CADR0970] - Cadours, R.; Bouallou, C. Simulation of Gas Absorption into Aqueous
Solutions, ECCE-1, May 1997, Florence, Italy.
[GLAD0900] - Glasscock D.A. Modelling and Experimental Study of Carbon Dioxide
Absorption into Aqueous Alkanolamines, Ph.D. dissertation, The University of Texas at
Austin, 1990.
[GLAD0910] - Glasscock D.A.; Critchfield J.E.; Rochelle G.T. CO2
Absorption/Desorption in Mixtures of Methyldiethanolamine with Monoethanolamine or
Diethanolamine, Chem. Eng. Sc. 1991, 46, 2829-2845.
[HAGD1951] - Hagewiesche D.P.; Ashour S.S.; Al-Ghawas H.A.; Sandall O.C.
Absorption of Carbon Dioxide into Aqueous Blends of Monoethanolamine and N-
Methyldiethanolamine, Chem. Eng. Sc. 1995, 50, 1071-1079.
[LEWJ2540] - Lewis, J. B. The mechanism of mass-transfer of solutes across liquid-liquid
interfaces, Chem. Eng. Sc., 1954, 3, 248-259.
[LIDD1960] - Lide, D. R. CRC Handbook of Chemistry and Physics, 77th Edition (Lide,
D. R. Editor). CRC Press. Inc., Boca Raton, Florida, USA (ISBN 0-8493-0477-6) 1996.
[PANF1970] - Pani, F.; Gaunand, A.; Cadours, R.; Bouallou, C.; Richon, D. Kinetics of
absorption of CO2 in concentrated aqueous methyldiethanolamine solutions in the range 296
K to 343 K, J. Chem. Eng. Data 1997, 42, 353-359.
[RIDJ0860] - Riddick, J. A.; Bunger, W. B.; Sakane, T. K. Techniques of Chemistry, Vol.
II Organic Solvents, ath Ed. John Wiley & Sons (ISBN 0-471-08467-0) 1986.
[VERG3880] - Versteeg, G.F.; van Swaaij, W.P.M. Solubility and diffusivity of acid gases
(CO2, N2O) in aqueous alkanolamine solutions. J. Chem. Eng. Data 1988, 33, 29-34.
51
Notations
Ci concentration/mol.m-3.
DAg Rushton turbine diameter/m.
Di diffusion coefficient/m.s-1.
DT internal diameter of the reactor/m.
F Faradays constant.
Kj equilibrium constant
k Boltzmanns constant
kL mass transfer coefficient/m.s-1.
N liquid side stirring speed/rpm.
n number of components
P0 initial pressure at t=0, Pa.
PI Inert pressure, Pa.
Re Reynolds number.
Ri reaction term for species i/mol.m-3.s-1.
s surface renewal rate/s-1.
Sc Schmidt number.
Sh Sherwood number.
T temperature/K.
t time/s.
VG gas phase volume/m3.
z charge number.
film thickness/m.
electrical potential/V.
contact time/s.
viscosity/Pa.s.
density/kg.m-3.
52
Table 1 : Experimental condtions of CO2 absorption kinetics by aqueous solutions of blends
of alkanolamines : temperature T, mass fraction of amine w(amine), initial pressure P0, initial
solution vapor pressure PI, loading Rc.
File T w(MDEA) w(MEA) w(DEA) P0 PI Rc

K wt % wt % wt % Pa Pa
156 296.55 50 5 - 179097 2760 0.0000
157 296.65 50 5 - 175446 2760 0.0000
158 296.65 50 5 - 177333 2760 0.0159
159 296.65 50 5 - 177713 2760 0.0320
160 296.65 50 5 - 177102 2760 0.0482
161 296.65 50 5 - 188677 2760 0.0905
162 296.65 50 5 - 197956 2760 0.1464
163 296.65 50 5 - 183235 2760 0.2645
164 296.65 50 5 - 185473 2760 0.3753
165 296.85 50 - 5 186289 2780 0.0000
166 296.65 50 - 5 176494 2350 0.0000
167 296.65 50 - 5 176157 2350 0.0092
168 296.65 50 - 5 179267 2350 0.0182
169 296.55 50 - 5 177440 2350 0.0288
170 296.65 50 - 5 177142 2350 0.1134
171 296.65 50 - 5 197522 2350 0.2260
172 296.65 50 - 5 186998 2350 0.3296
173 342.75 50 - 5 183017 27600 0.0000
174 342.55 50 5 - 184252 27700 0.0000
175 296.85 45 - 5 184018 2840 0.0000
176 342.75 45 - 5 181642 29500 0.0000
177 296.85 30 - 20 181117 2880 0.0000
178 342.75 30 - 20 188566 29800 0.0000
53
Chapter 4.
Kinetics of CO2 desorption from highly concentrated and CO2 loaded MDEA Aqueous
Solutions in the Range 312 K - 383 K.
R. Cadours, C. Bouallou, A. Gaunand and D. Richon
Introduction
Absorption by aqueous alkanolamine solutions is the dominant industrial process for

acid gases removal, in particular CO2 and H2S, from natural gas. Such washing processes are
also used in petroleum refining, coal gasification and hydrogen production.
Methyldiethanolamine and its blends with primary or secondary amines, or with sterically
hindered amines are the main systems of interest, already studied at the laboratory and
sometimes industrial scales. Number of workers have measured absorption kinetics of acid
gases in various alkanolamine solutions.
Although desorption processes often act as regenerative processes for absorption

systems, studies devoted to desorption are not as numerous as those concerning absorption.
The regeneration often uses steam or heated inert gas as stripping agents, then an accurate
design based on proper chemical and mass-transfer data could lead to significant energy
savings. Among the few works dealing with CO2 desorption from aqueous alkanolamine
solutions, Critchfield (1988), Bosch et al. (1990) and Xu et al. (1995) have studied the
desorption of CO2 from MDEA solutions at various temperatures in a stirred cell reactor or in
packed columns. These investigations are limited to CO2 loadings lower than 0.50 mol of CO2
per mol of amine, and temperatures lower than 343 K. Experiments and analyses show that
the absorption theory can also be applied to desorption in these ranges of temperatures,
MDEA concentrations and CO2 loadings. There are no data at temperatures between 343 K
and the temperatures found in strippers, i.e. 383 K.
54
This work describes the apparatus developed to measure acid gases desorption
kinetics from amine solutions. Experiments of CO2 desorption from MDEA aqueous solutions
have been achieved with CO2 loadings ranging 0.05 - 0.85 mol of gas per mol of amine, in an
extended range of temperatures, 313 - 383 K, and with 25 to 50 wt % MDEA aqueous
solutions.
The (6.20 0.02) x 10-2 m internal diameter thermostated glass reactor shown in
Figure 1 is provided with a (4.25 0.02) x 10-2 m diameter six-bladed Rushton turbine in its
lower part, a (4.00 0.02) x 10-2 m diameter propeller in its upper part, and four equally
spaced vertical PTFE baffles to prevent vortexing. An horizontal PTFE plate and a ring are
put at midway of the bottom and the top of the cell to set both level and area A of the gas-
liquid interface, and to make sure of their stability during stirring. A is (13.64 0.05) x 10-4
m2. The shafts are maintained vertically between two pivots supported by two sapphire
bearings on the flanges of the cell and the horizontal PTFE plate; they are driven magnetically
by adjustable speed motors. This technique avoids leaking, friction and heat generation
appearing when using stems crossing the cell bottom and top. The Rushton turbine speed is
checked with a stroboscope; it remains constant within 1 rpm during each test. The total
volume available for gas and liquid is (0.3648 0.0005) x 10-3 m3. The temperature in the
reactor is known within 0.05 K through a 100 W platinum probe, calibrated between 273 and
398 K by the Laboratoire National dEssais. It is controlled through to a thermostatic fluid
circulated into the glass double jacket. The upper flange is kept at a temperature slightly
higher than the temperature in the reactor, in order to avoid water condensation. The whole
cell is placed inside a thermoregulated air bath. A tube settled through the upper flange allows
either to degas the cell or to connect the cell to a variable volume device originally under a
vacuum. The lower flange is equipped with an hypodermic needle for liquid and gas loading.
Kinetics of gas desorption are measured, just after performing a slight depression
inside the cell, by recording the absolute pressure increase up to equilibrium pressure through
a SEDEME pressure transducer, working in the range : 0 to 200 kPa. This transducer is
thermostated at 408 K, temperature higher than that in the reaction cell, in order to avoid
liquid condensation in its measuring chamber. It is calibrated within 10 Pa against a
55
mercury manometer. A microcomputer fitted with a data acquisition card is used to record the
pressure transducer and platinum probe signals.
Procedure
Water and MDEA are degassed independently and aqueous solutions are prepared
under a vacuum. The amounts of water and MDEA are known separately by differential
weighings within 10-2 g. The flask containing the solution is connected to the reaction cell
tube in order to transfer the solution by gravity under a vacuum. Accurate weighings of the
flask before and after transfer yield the mass of solution present in the cell, and the liquid
phase volume through the density correlation used in Glasscock (1990).
At a given temperature, and under the solution vapor pressure PI, pure CO2 from the
gas cylinder (see figure1) is bubbled into the solution through the hypodermic needle, under
stirring to speed-up the loading, and reach the vapor-liquid equilibrium. The CO2 amount
absorbed nCO2,abs is known from weighings of the gas cylinder before and after loading and
from the equilibrium CO2 partial pressure PCO2 :
m CO2 PCO2 VG
n CO2 abs = (1)
M CO2 R TK
PCO2 is obtained from the difference between the total pressure at equilibrium PT,eq and
the solution vapor pressure PI :
PCO 2 = PT,eq PI (2)
Then stirring is started, a vacuum is created in the small external volume and this one
is connected during a short time to the upper part of the reaction cell. The resulting small
pressure drop leads water to evaporate and CO2 to desorb, to go back to the initial pressure;
the corresponding pressure increase is recorded, up to the vapor-liquid equilibrium value.
Small pressure drops are achieved in order to avoid bubble desorption, and to keep the initial
liquid composition constant during the whole experiment. Water evaporation takes place in a
56
very short time at the beginning of experiment, and CO2 desorption rates are carefully
measured once this first step is finished.
The apparatus and the method chosen allow to use small amounts of amine solution
and gas, compared to semi-batch or batch apparati generally used. Kinetics of CO2 desorption
from loaded MDEA-water solutions are adequately measured during a period ranging
between 500 and 20000 s (figure 2).
Materials
Twice-distilled water and reagent-grade MDEA are used. MDEA is from Aldrich,
with a 99 wt % certified minimum purity. Carbon dioxide is from L'Air Liquide, with a
certified purity of 99.995 vol %.
Theory
Most of previous works on CO2 absorption or desorption by and from tertiary

alkanolamine solutions consider the reaction mechanism proposed by Donaldson and Nguyen
(1980) :
k a1
MDEA + CO2 + H 2 O MDEAH + + HCO 3 (I)
k d1
Such a mechanism leads to the generally accepted rate expression :
r = k a1 C MDEA C CO 2 k d1 C MDEAH + C HCO (3)

3
In the frame of the stagnant film theory, using Ficks law for CO2 transfer leads to the
usual stationary balance at distance x from the interface :
d 2 C CO 2
D CO 2 =r (4)
dx 2
57
For each experiment, the interfacial gradients of all species concentrations except
CO2 can be easily estimated from the corresponding measured initial desorption rate,
diffusivities, CO2 Henrys law constant and CO2 mass transfer coefficient : these gradients are
always lower than 3 % and will be disregarded. Eq 4 becomes :
d 2 C CO 2
D CO 2
dx 2
[
= k a 1 C MDEA, eq C CO 2 - C CO 2 ,eq ] (5)
where
C MDEAH + ,eq C HCO - ,eq
C CO 2 ,eq = 3
(6)
K 1 C MDEA,eq
is the CO2 bulk concentration at equilibrium with the bulk concentrations of the three other
species. K1 is the apparent equilibrium constant of reaction I. Using Ficks law at x = 0, the
analytical solution of eq 5 yields the CO2 desorption rate :
(
0 = k L ,CO 2 C CO 2 ,in C CO 2 ,eq ) Ha
th( Ha)
(7)
which is negative in the case of the desorption. Ha is the dimensionless Hatta number :
1
Ha = k a1 D CO 2 C MDEA ,eq (8)
k L ,CO 2
From the experiments achieved at various stirring speed, it is assumed that there is no
mass transfer resistance in the gas phase. Then :
PCO 2
C CO 2 ,in = (9)
H CO 2
where HCO2 is the apparent Henrys law constant in the concentration scale. As amounts of
CO2 transfered to the gas phase are very small compared to amounts of bicarbonate and
58
protonated or molecular MDEA, the bulk concentration CCO2,eq does not vary significantly
during an experiment. It is calculated from Henrys law before the pressure drop. Then :
k L ,CO 2
0 =
H CO 2
Ha
(
P
th ( Ha ) CO 2 ,in
PCO 2 ,eq ) (10)
Solving the mass balance for CO2 in the gas phase
d n CO 2
= 0 A (11)
dt
PCO 2 VG
where n CO 2 = is the number of CO2 moles in the gas phase at time t, with eq 10,
RT
yields :
PT PT ,eq
Y = ln (
= U t t dep
PT ,dep PT,eq
) (12)
where
RT Ha
U = k L,CO 2 (13)
VG H CO 2 th ( Ha )
Linear regression of Y over t shows that relation 12 is fulfilled for all experiments. It
is then possible to calculate the initial CO2 desorption rates as :
VG
0 = U P (14)
RTK A
where
P = PT ,dep PT,eq (see figure 3) (15)
59
Results
Experiments have been carried out in the temperature range 313 - 383 K, 25 and 50 wt
% MDEA aqueous solutions, and for CO2 loadings from 0.05 to 0.85 mol of gas per mol of
amine. The different experimental conditions and the corresponding initial desorption rates
obtained from eq 14 are shown in tables 1 and 2. For given temperature and MDEA
concentration, the constant U defined by equation 13 is independant of the depression created.
So the desorption rate is proportional to DP (see for instance experiments 93 to 101). The
influence of DP must be taken into account to observe the influence of the experimental
parameters and the ratio will be used to establish further observations :
0 H CO 2
= (16)
P
Using c, for given concentrations, temperatures and CO2 loadings, several experiments
have usually been performed. The reproducibility of initial CO2 desorption rates is better than
10 % (see figure 4).
We have found that gas phase stirring does not influence the CO2 desorption rate; this
is shown by experiments, at given temperature, MDEA concentration and CO2 loading
(experiments 34 to 39 with gas phase stirring and experiment 40 without gas phase stirring).
These experiments agree with the previous assumption about the absence of mass transfer
resistance in the gas phase. With a 2189 mol.m-3 MDEA solution, at 313.6 K, and 47 % CO2
loading, reducing the liquid stirring speeds on both liquid and gas sides from 100 to 50 rpm,
experiments n12, 13, and 14, 15 respectively, does not lead to significant decrease of CO2
desorption kinetic rates, within the reproducibility range. These results agree with high values
of Ha, between 7 and 89 for our experiments, for which kL,CO2 has no influence on desorption
rate.
At given temperature and total MDEA concentration, increasing the CO2 loading leads
to a decrease of the CO2 desorption rate (see figure 4). This observation agree with eq 8 : at
given experimental conditions except loading, the Hatta number characterizes the
60
enhancement of the desorption processus by the chemical reaction. Increasing the CO2
loading leads to decreasing CMDEA and then the Hatta number.
Figure 4 clearly shows the influence of temperature on CO2 desorption from MDEA
solutions : the CO2 desorption rate increases with the temperature. This result agrees with the
conclusions of Bosch et al. (1990) and Xu et al. (1995). The same observations are obtained
with 25 wt % MDEA solutions.
Discussion
For each experiment at 313, 323 and 333 K, the desorption rates were estimated using
correlations of previous works which are given in the Appendix A. The CO2 diffusion
coefficient is obtained with a relation established for unloaded solutions, the correlation used
for HCO2 was extrapolated up to 333 K. The Arrhenius law is taken from Pani et al. (1997)
who studied CO2 absorption by aqueous MDEA solutions in the temperature range 296 to 343
K:
5461
k a1 = 4.68 10 5 exp (17)
TK
Let W be the ratio of the predicted desorption rate on the measured desorption rate.
For 25 wt % MDEA aqueous solutions, CO2 experimental desorption rates are well predicted
at low loadings (see figures 5 and 6). This conclusion agrees with the results of Bosch et al
(1990) or Xu et al. (1995). But for loadings higher than 0.5 mol of CO2 per mol of MDEA for
which no data are available in the literature, predicted desorption rates are significantly higher
than experimental ones. The deviation observed increases with temperature. Same results are
observed on figures 7 and 8 for 50 wt % MDEA solutions.
The deviation pointed out at 333 K can be explicated by the extrapolation of the
correlations used to calculate the CO2 Henrys law constant (see appendix A). In the same
way, the correlation used for the CO2 diffusivity has been established for unloaded MDEA
solutions. The most important deviation is found at high loadings and with 50 wt % MDEA
aqueous solutions. In these cases, the ionic species concentrations are quite high,
61
unfortunately, the correlations leading to diffusivity and solubility data did not consider the
effect of electrolyte species.
At 363 and 383 K, no data are available in the literature, and CO2 diffusivities and
Henrys constants are missing. Consequently, it is not possible to interpret our experimental
desorption rates with the model.
Conclusion
We designed a thermoregulated constant interfacial area reactor for accurate

desorption measurements up to 383 K. Desorption rates of CO2 from loaded aqueous
solutions of methyldiethanolamine were measured in the temperature range 313 - 383 K, for
25 and 50 wt % MDEA solutions, and CO2 loadings from 0.05 to 0.85 mol of gas per mol of
amine.
For loadings lower than 0.5 mol of CO2 per mol of MDEA, the CO2 desorption rates
calculated with our model are comparable to experimental ones. These conclusions agree with
Bosch et al. (1990) or Xu et al. (1995) results. For higher loadings for which the correlations
used for diffusivities or Henrys law constant are extrapolated, the deviations between
experimental and predicted desorption rates are significant. These deviations are more
important with 50 wt % MDEA solutions than with 25 wt % MDEA solutions. The lack of
diffusivity and Henrys constant data at temperatures higher than 343 K prevent us to
compare the experimental desorption rates measured at 363 and 383 K with the results
obtained with our model. We supposed that neglecting the influence of electrolyte species on
the diffusivity and the Henrys constant could be the principal source of discrepancies. We
need a rigorous thermodynamic model taking into account the electolyte species for the
representation of the desorption rates measured with high loadings and high temperatures..
62
APPENDIX A
For the calculation of the density and the viscosity of the solution, we used the
correlations proposed by Glasscock (1990). The viscosity of unloaded solution is obtained by
:
B1 = 19.52 23.40 w MDEA 31.24 w 2MDEA + 36.17 w 3MDEA

B 2 = 3912 + 4894 w MDEA + 8477 w 2MDEA 8358 w 3MDEA (A1a)
B 3 = 0.02112 + 0.03339 w MDEA + 0.02780 w MDEA 0.04202 w MDEA
2 3
B
aq sol, =0 = exp B1 + 2 + B 3 TK 10 3 (A1b)
TK
This correlation is considered to be reasonable for 0 to 50 wt % total amine, and a

temperature range of 290 to 320 K. It is used to estimate the viscosity of loaded solution
through the relative viscosity :
r = 1.000 + 0.8031 + 0.35786 2 (A2a)
aq sol, 0
= 1.0 + 2.0 ( r 1) w MDEA (A2b)
aq sol, = 0
The density of loaded solution is given by :
1000
d
[ ] [
. 10 3 exp 0.000344 (TK T0 ) + w MDEA 0.918 10 -3 exp 0.000528 (TK T0 )
= w water 101 ]
[
+ w CO 2 0.0636 10 -3 exp 0.0036 (TK T0 ) ]
(A3)
The diffusion coefficient of CO2 was estimated using the data and the N2O analogy of
Versteeg and van Swaaij (1988). It was estimated for unloaded solutions through :
63
0.8
2119 water
D CO 2 = 2.35 10 6
exp (A4)
TK aq sol
The relation proposed by Pani et al. (1997) was used to estimate the MDEA
coefficient diffusion :
0. 2
D CO 2 water
= 2.43 (A5)
D MDEA aq sol aq sol
The equilibrium constant for equation I is obtained with the ratio of the equilibrium
constants of the following reactions :
CO2 + 2 H 2O H 3O + + HCO 3 (II)
MDEAH + + H 2O MDEA + H 3O + (III)
The equilibrium constant K2 is given by Kent et Eisenberg (1976) for temperatures

lower than 413 K :
. 1011 2.96445 10 13
53.6855 10 4 4.8123 10 8 194
K 2 = exp 241818
. + + (A6a)
TR TR2 TR3 TR4
The equilibrium constant of reaction III is obtained through the correlation used by
Rinker et al. (1995) over the range of 298 - 333 K :
log 10 ( K 3 ) = 14.01 + 0.018 TK (A6b)
The CO2 Henry constant was determined with the correlation proposed by Al-Ghawas
et al. (1989) for temperatures between 288 K and 323 K and for MDEA compositions up to
50 wt % MDEA :
64
B4 = 2.01874 2.37638 101 w MDEA , = 0 + 2.90092 102 w 2MDEA , = 0 4.80196 102 w 3MDEA , = 0

B5 = 313549
. .10 + 154931
3
. 10 w MDEA , = 0 183987
4
. 10 w MDEA , = 0 + 3.00562 10 w MDEA , = 0
5 2 5 3

B6 = 8.13702.10 + 2.48081 10 w MDEA , = 0 + 2.92013 10 w MDEA , = 0 4.70852 10 w MDEA , = 0
5 6 7 2 7 3
(A7a)
B5 B 6
H CO 2 , = 0 = 101.325 exp B 4 + + (A7b)
TK TK2
A mass transfer correlation between dimensionless numbers has been established for
our apparatus from the same N2O absorption experiments by MDEA-water solutions used by
Pani et al. (1997) :
Sh = 0.25 Re 0.63 Sc 0.42 (A8)
Sh=kLDT/DCO2 is the Sherwood number, Re=dNDAg2/m is the stirrer Reynolds number

and Sc=m/dDCO2 is the Schmidt number. In Pani et al. (1997), the correlation presented
includes N2O-aqueous MDEA solutions and CO2-water absorption experiments. As CO2
absorption may be enhanced by its reaction with hydroxide ions, only the N2O absorption in
MDEA solutions experiments were considered here, leading to the above correlation.
However, the previous choice is of no consequence on Pani et al. (1997) results as their
absorption experiments of CO2 in MDEA solutions took place according to a fast regime of
reaction, insensitive to the liquid side mass transfer resistance.
65
Notations
A interfacial area, m2
C concentration, mol.m-3
d density of the solution, kg.m-3
Di diffusivity of species i in liquid phase, m2.s-1
Dag Rushton turbine diameter, m
DT internal cell diameter, m
Ha Hatta number
HCO2 Henry's law constant in the concentration scale, Pa.m3.mol-1
K equilibrium constant
ka1, kd1 reaction rate constants for the reaction between CO2 and MDEA, m3.mol-1.s-1
kL liquid side mass-transfer coefficient of dissolved CO2, m.s-1
N stirring speed, s-1
mCO2 CO2 mass introduced in the reactor, kg
M molecular weight, kg
nCO2, abs CO2 moles absorbed in liquid phase
P pressure, Pa
r rate of reaction between CO2 and MDEA, mol.m-3.s-1
R gas constant, 8.3143 J.K-1.mol-1
Re dimensionless Reynolds number
Sc dimensionless Schmidt number
Sh dimensionless Sherwood number
t time, s
T temperature, K or R
U defined by eq 17,
V volume, m3
w wt %.
x spatial variable measured from th gas-liquid interface, m
z dimensionless spatial variable
66
Greek letters
loading of the solution, molCO2/molMDEA

laminar film thickness, m
P pressure drop, Pa
0 CO2 desorption rate at the interface, mol.m-2.s-1
viscosity, Pa.s-1
r reduced viscosity
defined by eq 16, m.s-1.
W (Predicted desorption rate)/(Experimental desorption rate)
Subscripts
aq sol values in aqueous solution

CO2 carbone dioxide
dep depression
eq equilibrium
G gas
HCO3- bicarbonate
in interface
I inert
K for temperature, in K
MDEA methyldiethanolamine
MDEAH+ protoned methyldiethanolamine
R for temperature, in R
T total
water values in pure water
=0 unloaded solution
0 loaded solution
67
Literature cited
Al-Ghawas, H.A.; Hagewiesche, D.P.; Ruiz-Ibanez, G. ; Sandall, O.C. Physicochemical

Properties Important for Carbon Dioxide Absorption in Aqueous Methyldiethanolamine. J.
Chem. Eng. Data, 1989, 34, 385-391.
Bosch, H.; Versteeg, G.F.; van Swaaij, W.P.M. Desorption of Acid Gases (CO2 and H2S)
from Loaded Alkanolamine Solutions. In Gas Separation Technology; Vansant, E.F.,
Dewolfs, R., eds.; Elsevier Science Publishers B.V.: Amsterdam, 1990, 505-512.
Critchfield, J.E. CO2 Absorption/Desorption in Methyldiethanolamine Solutions Promoted

with Monoethanolamine and Diethanolamine : Mass Transfer and Reaction Kinetics. PhD
dissertation. The University of Texas at Austin. 1988.
Donaldson, T.L.; Nguyen, Y.N. Carbon Dioxide Reaction and Transport into Aqueous Amine
Membranes. Ind. Eng. Chem. Fundam., 1980, 19, 260-266.
Glasscock, D.A. Modelling and Experimental Study of Carbon Dioxide Absorption into
Aqueous Alkanolamines. PhD dissertation. The University of Texas at Austin. 1990.
Kent R.L.; Eisenberg B. Better Data for Amine Treating. Hydrocarbon Proc., 1976, 55, 87-
90.
Mahajani V.V., Danckwerts P.V. The Stripping of CO2 from Amine-promoted Potash
Solutions at 100C, Chem. Eng. Sc.,1983, 38 (2), 321-327.
Pani, F.; Gaunand, A.; Cadours, R.; Bouallou, C.; Richon, D. Kinetics of absorption of CO2 in
Concentrated Aqueous Methyldiethanolamine Solutions in the Range 296 K to 343K, J.
Chem. Eng. Data, 1997, 42 (2), 353-359.
68
Rinker E. B.; Ashour S. S.; Sandall O. C. Kinetics and modelling of carbon dioxide
absorption into aqueous solutions of N- methyldiethanolamine. Chem. Eng. Sc., 1995, 50 (5),
755-768.
Versteeg, G.F.; van Swaaij, W.P.M. Solubility and Diffusivity of Acid Gases (CO2, N2O) in
Aqueous Alkanolamine Solutions, J. Chem. Eng. Data, 1988, 33 (1), pp29-34.
Xu G-W.; Zhang C-F.; Qin S-J.; Zhu B.-C. Desorption of CO2 from MDEA and Activated
MDEA Solutions. Ind. Eng. Chem. Res. 1995, 34 (3), 874-880.
69
Table 1 : Conditions of CO2 desorption from 25 wt % MDEA solutions.
N Temperature CMDEA,total a VG DP Experimental desorption Calculated desorption
rate rate
K mol.m-3 molCO2/molMDEA 10-6 m3 Pa mol.m-2.s-1 mol.m-2.s-1
1 312.3 2109 0.16 181.30 98 -3.63E-06 -3.40E-06
2 312.3 2108 0.27 181.17 185 -4.15E-06 -5.85E-06
3 312.3 2108 0.27 181.17 145 -3.42E-06 -4.58E-06
4 312.3 2107 0.27 181.17 169 -3.83E-06 -5.34E-06
5 312.3 2107 0.30 181.14 169 -3.80E-06 -5.21E-06
6 313.5 2193 0.24 178.72 151 -4.47E-06 -5.00E-06
7 313.6 2193 0.24 178.72 87 -3.39E-06 -2.88E-06
8 313.6 2193 0.24 178.71 148 -4.74E-06 -4.93E-06
9 (b) 313.5 2190 0.43 178.50 333 -1.15E-05 -9.31E-06
10 (b) 313.5 2190 0.43 178.50 305 -8.41E-06 -8.51E-06
11 313.6 290 0.43 178.50 276 -8.66E-06 -7.71E-06
12 313.5 2189 0.47 178.45 343 -8.44E-06 -9.10E-06
13 313.5 2189 0.47 178.45 355 -7.25E-06 -9.42E-06
14 (c) 313.6 2189 0.47 178.44 266 -5.78E-06 -7.07E-06
15 (c) 313.6 2189 0.47 178.44 340 -7.09E-06 -9.04E-06
16 313.5 2188 0.58 178.32 474 -9.42E-06 -1.11E-05
17 313.5 2188 0.58 178.32 592 -1.11E-05 -1.38E-05
18 313.6 2186 0.71 178.17 1278 -1.62E-05 -2.47E-05
19 313.6 2186 0.71 178.16 1470 -1.83E-05 -2.84E-05
20 313.6 2185 0.76 178.11 1779 -1.85E-05 -3.12E-05
21 313.6 2184 0.83 178.03 2148 -2.07E-05 -3.26E-05
22 313.6 2184 0.83 178.03 2220 -1.96E-05 -3.37E-05
23 322.4 2195 0.16 175.26 115 -3.65E-06 -4.89E-06
24 322.4 2191 0.45 174.91 118 -4.09E-06 -3.89E-06
25 322.4 2191 0.45 174.91 740 -1.92E-05 -2.44E-05
26 322.4 2189 0.57 174.76 1157 -2.06E-05 -3.32E-05
27 322.4 2189 0.57 174.76 1202 -1.87E-05 -3.45E-05
28 322.6 2188 0.66 174.63 1732 -2.83E-05 -4.37E-05
29 322.3 2187 0.72 174.59 2490 -3.12E-05 -5.71E-05
30 322.3 2187 0.72 174.58 2829 -3.11E-05 -6.49E-05
31 322.4 2186 0.78 174.51 3178 -4.33E-05 -6.58E-05
32 322.4 2186 0.78 174.50 3198 -3.94E-05 -6.62E-05
33 322.3 2185 0.85 174.58 3445 -3.98E-05 -6.16E-05
34 321.8 2086 0.32 179.76 387 -1.13E-05 -1.42E-05
35 321.8 2086 0.32 179.76 347 -1.09E-05 -1.28E-05
36 321.9 2086 0.32 179.76 307 -1.23E-05 -1.13E-05
37 322.2 2086 0.32 179.74 249 -6.81E-06 -9.20E-06
38 322.0 2086 0.32 179.75 393 -1.01E-05 -1.45E-05
39 322.0 2086 0.32 179.75 443 -1.17E-05 -1.63E-05
40 (a) 322.0 2086 0.32 179.75 399 -1.12E-05 -1.47E-05
41 333.7 2322 0.20 174.05 341 -1.50E-05 -1.74E-05
42 333.6 2321 0.31 173.91 489 -2.31E-05 -2.26E-05
43 333.6 2321 0.31 173.91 686 -2.28E-05 -3.17E-05
44 333.9 2320 0.31 193.88 673 -2.04E-05 -3.12E-05
45 333.5 2319 0.43 173.76 1081 -3.01E-05 -4.46E-05
46 333.6 2319 0.43 173.75 1188 -3.80E-05 -4.91E-05
47 333.6 2319 0.43 173.75 1185 -3.38E-05 -4.90E-05
48 333.5 2317 0.52 173.63 1848 -4.45E-05 -6.87E-05
49 333.6 2316 0.62 173.48 2481 -4.95E-05 -8.18E-05
50 333.6 2316 0.62 173.48 2789 -4.48E-05 -9.20E-05
51 333.4 2314 0.70 173.39 4564 -5.49E-05 -1.33E-04
52 364.2 2216 0.08 177.50 190 -1.21E-05 -
53 364.2 2214 0.21 177.32 900 -3.93E-05 -
54 382.9 2013 0.06 174.03 1680 -1.11E-04 -
55 382.9 2013 0.06 174.03 1180 -7.68E-05 -
56 383.0 2079 0.10 176.76 1711 -1.08E-04 -
57 (a) 383.0 2079 0.10 176.76 752 -6.30E-05 -
All experiments realised with liquid and gas phase agitation at 100 rpm excepted :
(a) N= 100 rpm - No gas phase agitation
(b) N= 175 rpm - No gas phase agitation
(c) N= 50 rpm - Gas phase and liquid phase agitation
(d) N= 165 rpm - No gas phase agitation
70
Table 2 : Conditions of CO2 desorption from 50 wt % MDEA solutions.
N Temperature CMDEA,total a VG DP Experimental desorption Predicted desorption rate
rate mol.m-2.s-1
K mol.m-3 molCO2/molMDEA 10-6 m3 Pa mol.m-2.s-1
58 312.8 4412 0.11 175.54 81 -2.18E-06 -2.34E-06
59 312.8 4412 0.11 175.54 58 -1.49E-06 -1.66E-06
60 312.8 4412 0.11 175.54 69 -1.71E-06 -1.99E-06
61 312.8 4412 0.11 175.54 77 -1.29E-06 -2.23E-06
62 312.7 4407 0.20 175.34 194 -4.48E-06 -5.14E-06
63 312.7 4407 0.20 175.34 181 -4.48E-06 -4.81E-06
64 312.7 4407 0.20 175.34 173 -3.42E-06 -4.59E-06
65 312.9 4400 0.33 175.01 342 -6.95E-06 -8.03E-06
66 312.9 4400 0.33 175.01 375 -7.16E-06 -8.81E-06
67 312.7 4394 0.44 174.77 618 -1.19E-05 -1.28E-05
68 312.7 4386 0.59 175.28 1369 -1.79E-05 -2.34E-05
69 312.7 4381 0.67 175.28 2106 -2.32E-05 -3.16E-05
70 (a) 313.5 4000 0.15 180.81 158.1 -5.13E-06 -4.77E-06
71 (a) 313.5 4000 0.15 180.81 152.9 -5.22E-06 -4.62E-06
72 (a) 313.5 4000 0.15 180.81 167.4 -6.04E-06 -5.06E-06
73 313.5 3993 0.32 180.46 381.8 -9.25E-06 -9.92E-06
74 322.1 4104 0.28 173.03 453 -1.42E-05 -1.50E-05
75 322.1 4104 0.28 173.03 679 -1.45E-05 -2.25E-05
76 322.1 4104 0.28 173.03 703 -1.55E-05 -2.33E-05
77 322.4 4098 0.39 172.75 1383 -3.28E-05 -4.11E-05
78 322.4 4098 0.39 172.75 1243 -3.26E-05 -3.70E-05
79 322.4 4092 0.52 172.45 2204 -2.66E-05 -5.66E-05
80 322.4 4092 0.52 172.46 2353 -3.49E-05 -6.04E-05
81 322.6 4089 0.57 172.32 2249 -2.97E-05 -5.45E-05
82 322.6 4089 0.57 172.32 2907 -3.25E-05 -7.05E-05
83 322.4 4088 0.59 172.27 2102 -1.79E-05 -4.88E-05
84 322.5 4086 0.63 172.18 3660 -2.93E-05 -8.07E-05
85 322.6 4085 0.65 172.13 4425 -3.36E-05 -9.50E-05
86 332.9 4630 0.23 174.72 811 -2.24E-05 -3.70E-05
87 (d) 332.3 4226 0.08 177.13 233 -1.09E-05 -1.18E-05
88 (d) 332.3 4226 0.08 177.13 270 -1.03E-05 -1.37E-05
89 (d) 332.3 4226 0.08 177.13 179 -8.00E-06 -9.03E-06
90 332.3 4222 0.14 176.97 459 -1.59E-05 -2.19E-05
91 332.3 4222 0.14 176.97 495 -1.46E-05 -2.37E-05
92 332.3 4222 0.14 176.97 507 -1.50E-05 -2.43E-05
93 332.4 4217 0.22 176.77 884 -2.35E-05 -3.94E-05
94 332.4 4217 0.22 176.77 852 -2.15E-05 -3.80E-05
95 332.4 4217 0.22 176.77 840 -2.28E-05 -3.74E-05
96 332.4 4217 0.22 176.77 658 -1.67E-05 -2.94E-05
97 332.5 4217 0.22 176.76 601 -1.53E-05 -2.68E-05
98 332.5 4217 0.22 176.76 601 -1.52E-05 -2.68E-05
99 332.5 4217 0.22 176.76 422 -1.24E-05 -1.88E-05
100 332.5 4217 0.22 176.76 307 -9.71E-06 -1.37E-05
101 332.6 4217 0.22 176.76 233 -7.69E-06 -1.05E-05
102 332.3 4215 0.28 176.63 1624 -3.34E-05 -6.84E-05
103 332.4 4215 0.28 176.62 1449 -2.98E-05 -6.11E-05
104 332.4 4215 0.28 176.62 1111 -2.97E-05 -4.70E-05
105 332.4 4211 0.33 176.50 1622 -2.66E-05 -6.49E-05
106 332.5 4209 0.38 176.38 2201 -3.86E-05 -8.42E-05
107 332.6 4206 0.42 176.28 2732 -4.24E-05 -1.01E-04
108 332.6 4206 0.44 176.21 2913 -4.81E-05 -1.05E-04
109 364.5 4067 0.05 167.68 982 -3.96E-05 -
110 364.3 4065 0.10 167.49 1854 -4.65E-05 -
111 364.2 4062 0.15 167.38 4013 -7.22E-05 -
112 364.3 4061 0.16 167.31 3157 -5.69E-05 -
113 382.7 4073 0.02 176.47 1029 -4.85E-05 -
114 382.7 4071 0.05 176.38 1287 -5.07E-05 -
All experiments realised with liquid and gas phase agitation at 100 rpm excepted :
(a) N= 100 rpm - No gas phase agitation
(b) N= 175 rpm - No gas phase agitation
(c) N= 50 rpm - Gas phase and liquid phase agitation
(d) N= 165 rpm - No gas phase agitation
71
DA VV
TJ
VP
PT
HR
MS
Vapor
phase
MR
Liquid
phase
DE
S
MS
Pt 100
Experimental apparatus
Figure 1 : Experimental apparatus. (DA) data acquisition, (DE) transparent thermostated

double envelope, (HR) heating rod, (MR) magnetic rod, (MS) magnetic stirrer, (N) needle, (Pt
100) platinum probe, (S) septum, (PT) pressure transducer, (TJ) thermostated jacket, (VP)
vacuum pump, (VV) variable volume.
72
Pressure, Pa Temperature, K
182000 384
180000 383,8
178000 383,6
176000 383,4
174000 383,2
172000 383
170000 382,8
0 500 1000 1500 2000 2500 3000
Time, s
Figure 2 : Typical CO2 pressure versus time profiles at high temperature, CMDEA = 2079
mol.m3 and = 0.10 molCO2/molMDEA
Total pressure (Pa) Temperature (K)
31400 334.8
Peq
31300
31200
P
31100
Pdep
31000
334.3
30900
30800
30700
30600 tdep
30500 333.8
0 1000 2000 3000 4000
Time (s)
Figure 3 : Treatment of a desorption acquisition curve
73
, molCO2 /molMDEA
0,0 0,2 0,4 0,6 0,8
0
-0,5
-1
, 10 m.s
-1
-3
-1,5
-2
Figure 4 : Influence of temperature and loading on CO2 desorption rate from 50 wt % MDEA
solutions. 313 K, + 323 K, 333 K, 364 K, 383 K
0
Predicted desorption rate, 10-5 mol.m-2.s-1
-4
-8
-12
-16
-20
-16 -12 -8 -4 0
-5 -2 -1
Measured desorption rate, 10 mol.m .s
Figure 5 : Comparison of experimental and predicted rates of desorption of CO2 from 25 wt %

MDEA solutions. 313 K, + 323 K, 333 K
74
2,5
1,5
0,5
0
0 0,2 0,4 0,6 0,8 1
, molCO2/molMDEA
Figure 6 : Influence of the loading on the deviation between experimental and predicted rates
of desorption from 25 wt % MDEA solutions. 313 K, + 323 K, 333 K
0
Predicted desorption rate, 10-5 mol.m-2.s-1
-3
-6
-9
-12
-12 -9 -6 -3 0
Measured desorption rate, 10-5 mol.m-2.s-1
Figure 7 : Comparison of experimental and predicted rates of desorption of CO2 from 50 wt %

MDEA solutions
75
3
2,5
1,5
0,5
0
0 0,2 0,4 0,6 0,8
, molCO2/molMDEA
Figure 8 : Influence of the loading on the deviation between experimental and predicted rates
of desorption from 50 wt % MDEA solutions. 313 K, + 323 K, 333
76
Chapter 5.
Rigorous Simulation Of Gas Absorption Into Aqueous Solutions
Introduction
Processes in which mass transfer is enhanced by chemical reaction are frequently

encountered in process industry. For both energy and environmental purposes, acid gases
must be removed from natural gas, as well as in petroleum refining, coal gasification and
hydrogen production. For such a purification, absorption with aqueous solutions of
alkanolamines remains the dominant industrial technology. The design of absorbers and
strippers requires information about transfer and kinetic mechanisms, and the related rate
expressions. Much work has been done with respect to modelling and calculating the rate of
mass transfer in absorption of gases followed by complex chemical reactions.
In the cases of single reversible or irreversible first order reaction, instantaneous or

not, analytical expressions based on film or penetration theory have been presented for the
mass transfer rate (Secor and Beutler, 1967). The major drawback of these models is that their
validity are limited due to the assumption involved. When no analytical or approximate
solution is available, it is necessary to use numerical methods to estimate the effect of
reaction rate on absorption, or to estimate a rate constant from absorption data.
Cornelisse et al. (1980) have studied the simultaneous absorption of CO2 and H2S in a
secondary amine. The CO2 absorption rate was enhanced by reaction of amine to form
carbamate, and the H2S reaction with the amine was assumed to be instantaneous. They used
a Newton-Raphson method to linearize the nonlinear reaction terms. This technique
decoupled the simultaneous differential equations and allowed them to be solved as a set of
linear equations. Littel et al. (1991) have used fundamentally the same method to describe the
simultaneous absorption of H2S and CO2 in a solution of a primary or secondary amine and a
tertiary amine.
77
In the model presented by Glasscock and Rochelle (1989) for the absorption of CO2
into MDEA aqueous solutions, multiple reactions and the diffusion of ionic species were
taking into account. The coupling between positive and negative ions were described by
Nernst-Planck equation. For gas absorption accompanied by second-order reversible reaction
and the CO2 absorption in MDEA comparisons were made between the enhancement factors
obtained for film theory, Higbies penetration theory, Danckwerts surface renewal theory and
simplified eddy diffusivity theory. For the steady-state theories, the equations are transformed
into a large set of coupled, non linear algebraic equations. For the unsteady state theories, a
spatial discretization results in a set of coupled ordinary differential/algebraic equations
which are integrated through time by DDASSL (Petzold, 1983).
Rinker et al. (1995) presented a study in order to determine the significant reactions
involved in the absorption of CO2 into MDEA aqueous solutions, in particular the effect of
including or neglecting the CO2/OH- reaction on the estimation of the forward rate coefficient
of the MDEA catalysed hydrolysis reaction. The method of lines was used to transform each
partial differential equations into a system of ordinary differential equations which were then
solved by DDASSL (Petzold, 1983). The same numerical method was used by Hagewiesche
et al. (1995) to estimate the rate coefficient of the reaction between CO2 and MEA at 313 K
from experimentally measured absorption rates of CO2 into blends of MDEA and MEA.
We developed a general model for the diffusion/reaction processes. We have

considered the most rigorous approach of taking into account the electrostatic potential
gradient terms with unequal diffusion coefficients for the ionic species even if much more
computational effort is required. This model allows us to consider reactions with finite rate
and instantaneous reactions with respect to mass transfer. We assume reversibility,
generalised reaction rate expressions and generalised stoichiometry. The applicability of our
model is not limited by the type of reactions modelled. The main advantage lives in the
simpleness of the numerical treatment because no numerical method is needed to transform
each partial differential equations into a system of ordinary differential equations.
Optimisation using a Levenberg-Marquardt algorithm provides the rate coefficients for
different kinetic mechanisms, from experimental data. In this work, the model has been
78
applied to investigate the discrepancies between different works reported in the literature for
the CO2 absorption into MDEA aqueous solutions, especially at high temperatures.
Numerical model
In this work, the phenomenon of mass transfer accompanied by chemical reaction was
modeled according to the film theory. We use a fictitious time to find a steady-state solution
corresponding to the film theory. The effect of the electrostatic potential gradient on the
diffusion of ionic species was also taken into account. This yields for each component i a
partial differential equation :
C i ( x , t ) C i2 ( x, t ) F ( ( x, t ) C i ( x, t ))
= Di zi Di + R i ( x, t ) (1)
t x 2
RT x
In Eq. 1, represents the electrostatic potential gradient, which couples the diffusion
of ionic species. Under the assumption of dynamic electroneutrality and the Nernst-Einstein
equation, the electrostatic potential can be expressed as a function of ion concentrations and
ion diffusivities :
NC C q ( x , t )
RT
z q Dq
x
( x, t ) =
q =1
NC
(2)
F
z
q =1
2
q D q C q ( x, t )
Ri is the production term :
NR NC
R i ( x, t ) = i , j k j C q q , j

(3)
j=1 q =1
79
For one of the ionic components, the partial differential equation is replaced by the
static electroneutrality condition in order to maintain electroneutrality throughout the mass
transfer zone :
NC
z C ( x, t ) = 0
i =1
i i (4)
Together with appropriate initial and boundary conditions, the set of equations gives
an unique solution.
Initial condition :
For the volatile species, 1 i NG, we consider linear profiles taking into account the
boundary conditions at the gas-liquid interface and their bulk concentrations. For the non
volatile species, the concentrations are equal to their bulk concentrations.
x P x
For volatile species C i ( x, 0) = 1 i + C i , bulk
Hi 0 x (5)
For non volatile species C i ( x, 0) = C i , bulk

Boundary Conditions :
x=0
* For the volatile species, the boundary conditions at the gas-liquid interface in
the isothermal case were obtained by considering the mass transfer rate in the gas near the
interface equal to the mass transfer rate in the liquid near the interface. This leads to the
following boundary condition :
C i
k Gi ( Pi H i C i ) = D i (6)
x x=0
80
Di Ci Pi
C i (0, t ) = + t > 0 (7)
k Gi H i x x=0
Hi
* For the non-volatile species, the fluxes are equal to zero :
Ci
=0 (8)
x x=0
x=
* For all chemical species i, the concentrations are equal to their liquid bulk
concentrations :
C i ( , t ) = C i , bulk (9)
There are a great number of advantages in using non dimensional variables in

numerical work, though it is not essential to do so. A good deal of arithmetic is involved in
our problem. A whole set of solutions with different physical parameters can be obtained
from one basic solution in non dimensional variables with considerable economy of computer
time. The fundamental parameters are high-lighted and analogies with physically different
systems become clearer. To obtain a dimensionless system, we consider the following
reduced variables :
~ x ~t = t D 1 ~ D ~ F
x= Di = i = (10)
2 D1 RT
Where D1 is the diffusion coefficient of the first volatile component.
~ H ~ 2 Hi
For volatile species Ci = i Ci Ri = Ri
Pi Di Pi
Ci 2 (11)
~ ~
For non - volatile species Ci = Ri = Ri
Ci , bulk Di Ci , bulk
81
Eqs. 1 and 4 become :
~ ~ ~ ~
Ci ~ 2 Ci ~ ~ Ci ~ ~ ~
= Di zi Di zi Di Ci ~ + R i (12)
~t ~
x2 ~x x
NC
~
z
i =1
i Ci = 0 (13)
Initial condition :
~ C i , bulk H i
For volatile species Ci (~ x , 0) = 1 + 1 ~
x ~
Pi x 0 (14)
~
For non volatile species C i ( ~
x , 0) = 1
Boundary Conditions :
~x = 0
~
~ ~ Di Ci
For volatile species C i (0, t ) = + 1
k Gi H i ~
x ~ ~
t > 0
x =0
~ (15)
Ci
For non volatile species =0
~x ~
x =0

~x = 1
~ C i , bulk H i
For volatile species C i (1, ~t ) = ~
Pi t > 0 (16)
~ ~
For non volatile species C i (1, t ) = 1
Numerical treatment
The production terms are non-linear. Consequently equations render a set of non-
linear coupled partial differential equations. An implicit centered finite difference iterative
method was employed to solve the system of coupled differential equations subject to the
82
boundary and initial conditions given above. The implicit method ensures stability for the
system of equations. From the initial concentrations field specified above, the concentrations
gradients are obtained by numerical derivation. Then the electrostatic potential and its
gradient are deduced. They enter as source terms in the transfer equations which are solved
using the finite-difference iterative method giving the solution for the next iteration. The
calculations are repeated until the convergence criterion :
~ ~
Ci,n j Ci,n-1j
Max ~n . 6
510 (17)
1 i NC
1 j j
C i, j
max
~
is satisfied. C in, j is the concentration of species i at point j and at iteration n.
The fictitious adimensional time was chosen in order to realise a compromise : it must be
large enough to limit roundoff error and also sufficiently small for convergence of the small
grid size. The steady-state solution is found by iterating upon an index (n). In this way, an
efficient and consistent algorithm is obtained for the set of equations. The model developed in
this work, combined with appropriate literature parameters will be used to model acid gas
treating systems in the general case. A Levenberg-Marquardt fitting procedure was used to
infer kinetic rate constants from the experimental data.
Validation
A first validation was established with the case of H2S absorption by

monoethanolamine presented by Danckwerts (1970). H2S at 0.1 atm is absorbed in a 1 M
solution of MEA at 298.15 K. The value of the equilibrium constant K is 275. The solubility
of H2S is taken as 0.10 gmole.l-1.atm-1. The diffusivity of monoethanolamine is 0.64 times
that of H2S. The reaction between H2S and MEA is instantaneous.
H 2S + MEA MEAH + + HS
83
The equilibrium is available everywhere in the film, and the concentrations of the
products HS- and MEAH+ are taken to be equal to one another everywhere. In this case, it is
possible to get an analytical solution for the absorption rate :
D
( )
= k L CH 2 S,int CH 2 S, bulk 1 + MEAH

+

DH 2 S CH 2 S,int CH 2 S, bulk
(18)
with
0.5
D DHS
2

HS

1 K CMEA , bulk CH 2 S, bulk + C K

1 DMEAH + DMEA H 2 S,int
=
2 DHS DMEA
+ 4 CH 2 S,int K C + K CMEA , bulk CH 2 S, bulk (19)
DMEA DMEAH + MEA , bulk
1 D D
HS + 1 K CMEA , bulk CH 2 S, bulk + HS CH 2S,int K
2 DMEAH + DMEA
We assume that the bulk liquid is free from dissolved H2S. We also assume that the
diffusivities of the products are equal to that of the amine. Using the same parameters for
numerical and analytic computations, we obtained the same result within a deviation of 0.6
% : we obtain 4.9710-2 mol.m-2.s-1 with the numerical model even though the analytical
solution proposed by Danckwerts (1970) gives 510-2 mol.m-2.s-1. As suggested by Glasscock
and Rochelle (1989), we treated the instantaneous reaction between H2S and MEA by making
rate constants large enough so that the reaction is in equilibrium everywhere in the boundary
layer. The concentration profiles for each species are represented in Fig. 1. These profiles
clearly shows the depletion of the rapid reacting MEA near the gas-liquid interface.
A second verification of the model was carried out by comparison with the numerical
solution given by Glasscock and Rochelle (1989) for a four reactions mechanism involving
two finite rate reactions, reactions I and II, and two instantaneous equilibria involving a
proton transfer, reactions III and IV.
CO2 + MDEA + H 2 O MDEAH + + HCO3 (I)
84
HCO3 CO2 + OH (II)
CO23 + H 2 O = HCO3 + OH (III)
MDEA + H 2 O = MDEAH + + OH (IV)
At the conditions of the simulation, the concentration of hydronium ion is very low,
we have neglected the water dissociation reaction :
2 H 2 O = OH + H 3 O + (V)
We use the same parameters as Glasscock and Rochelle (1989) (Table 1). The
relationship between the physical mass transfer coefficient and the film thickness is :
D1
= (20)
kL
The results in term of concentrations profiles are in good agreement in the case of a low
CO2 partial pressure (Fig. 2). Some differences are observed between the two models at high
CO2 partial pressure (Fig. 3). They are explained by the choice of the convergence criterion
(Eq 18), which seems to be more hard in our case. The hydroxide depletion is also
overestimated by the model of Glasscock and Rochelle (1989) and then, their model would
give an underestimation of the influence of the reaction II. With this mechanism, we checked
that we did not introduce any perturbation by replacing for one of the ionic species the
differential equation of diffusion by the algebraic equation of static electroneutrality.
Results and Discussion
The emission of carbon dioxide is considered one of the biggest cause of the global
warming. Reduction of CO2 emitted as flux gases from a large scale stationary source has
become a worldwide problem. Removal of CO2 from process gas has been achieved in
85
industry by gas absorption processes employing alkanolamines. From the view point of
economics, high level regeneration of the absorbent is indispensable. Since the regeneration
step includes heating of the absorbent, the energy requirement will significantly increase
when treating a large volume of absorbent. Therefore, the reduction of heating energy
becomes a determining factor for realising gas absorption processes.
CO2 absorption into alkanolamines aqueous solutions (MEA, DEA, MDEA and their
blends) have been extensively studied. However, there has been a great confusion and
disagreement in the literature concerning the magnitude of the rate coefficient of each
reaction involved in the considered mechanisms for absorption in aqueous solutions of
alkanolamines.
The reaction between CO2 and aqueous MDEA has been the subject of many
experimental studies. However, they are widely varying opinions in the literature regarding
the mechanism of this reaction. In general, the reaction mechanism between CO2 and MDEA
proposed by Donaldson and Nguyen (1980) is considered. This mechanism is a base-
catalysed hydration reaction which implies that MDEA do not react directly with CO2
(reaction I). The first investigations of the kinetics of this reaction was reported in literature
by Blauwhoff et al. (1984), Barth et al. (1984). Their results were limited at 298 K. The
determination of the kinetic parameters at various temperatures have been reported by Yu et
al. (1985), Haimour et al. (1987), Critchfield (1988), Versteeg and van Swaaij (1988a),
Toman and Rochelle (1989), Tomcej and Otto (1989), Littel et al. (1990), Rinker et al. (1995)
and recently Pani et al. (1997) reported a complete review of the data on the reaction between
CO2 and aqueous MDEA. The significant discrepancies in the results presented by these
authors can be explained by the significant effect of the reaction between CO2 and OH-,
especially when estimating the rate coefficient of the reaction I for unloaded MDEA aqueous
solutions. Rinker et al. (1995) have considered over and above the reaction I the reactions II,
III, IV and V. They have shown clearly that neglecting the CO2/OH- reaction can result in
large errors in the rate coefficients for the MDEA catalysed hydrolysis reaction, especially at
high temperatures.
Pani et al. (1997) have developed an efficient apparatus to measure absorption kinetics
of acid gases into amines solutions. A thermoregular constant interfacial area Lewis-type cell
86
was operated by recording the pressure drop during batch absorption. The experimental data
obtained by these authors are well adapted to be treated with the film theory.
For the calculation of absorption rate and also of kinetic parameters, the values of the
diffusivities of all species and the Henrys law constant for the CO2 are required. The
equilibrium constants of all chemical reactions are also needed.
The diffusivity of CO2 in amine solutions is estimated with the N2O-analogy presented
by Versteeg and van Swaaij (1988b) :
(D N2O 0.8 ) aq sol

(
= D N 2 O 0.8 ) water
(21)
The diffusion coefficient of CO2 in the amine solution is then calculated using :
DN2O DN O
= 2 (22)
DCO 2 aq sol DCO 2 water
The diffusion coefficient for N2O and CO2 in water are calculated according to the
modified Stokes-Einstein relation (Versteeg and van Swaaij, 1988b):
(D ) CO 2 water
2119
= 2.35 106 exp
T
(23)
(D ) N 2 O water
2371
= 5.07 106 exp

T
(24)
The diffusion coefficient of MDEA is given by the correlation proposed by Pani et al.
(1997) :
0 .2
DCO 2 aq sol
DMDEA = (25)
2.43 water
87
The diffusion coefficients of the ionic species are assumed to be equal to those of
MDEA in order to compare our results with the literature results.
The viscosities of water and MDEA aqueous solutions, and the Henrys law constant
for the CO2 are calculated from the correlations of Al-Ghawas et al. (1989).
When the reaction II is considered, the values of the backward rate were calculated
from the correlation proposed by Pinsent et al. (1956) for the temperature range 273 - 313 K :
( )
log10 k II , b = 10.635
2895
T
(26)
For all the equilibrium constants, we use the same correlations previously used by
Rinker et al. (1995). Olofsson and Hepler (1975) reported a correlation for the water
dissociation constant KV for the temperature range of 293 - 573 K :
142613.6
log10 ( KV ) = 8909.483 4229.195 log10 ( T)
T (27)
+ 9.7384 T 0.0129638 T2 + 115068
. 10 5 T3 4.602 10 9 T4
A correlation for the determination of the values of (KIIKV) is obtained from the data
reported by Read (1975) for the temperature range of 273 - 523 K :
K 7495.441
log10 V = 179.648 + 0.019244 T 67.341log10 ( T) (28)
KII T
The equilibrium constant of the reaction III is obtained from the correlation of
Danckwerts and Sharma (1966) over the temperature range of 273 - 323 K :
K 2902.4
log10 V = 6.498 0.0238 T (29)
KIII T
88
The correlation of Barth et al. (1984) established over the temperature range of 298 -
333 K is used for the calculation of the equilibrium constant of the reaction IV :
K
log10 V = 14.01 + 0.018 T (30)
KIV
All these correlations are extrapolated up to 343 K, the greatest experimental

temperature reached by Pani et al. (1997).
In a first time, the reaction I is treated as a reversible second order reaction and the
effect of the reactions II, III and IV is neglected (mechanism 1). This mechanism, proposed
by Donaldson and Nguyen (1980), is generally presented in literature as reasonable. It is a
base catalysis of CO2 hydration reaction. We have determined with our model the kinetic rate
of the forward reaction (I) for each experimental data of Pani et al. (1997) (Table 2). These
results are correlated with the following Arrhenius law :
5332.8
k I , f = 2.96 105 exp (31)
T
The values of Ea for the temperatures investigated are in good agreement with those
of recent works (Table 3). In this table, the energy activation for Rinker et al. (1995) was
calculated from the results obtained by these authors with a numerical model treating their
own experimental data with the reaction I as a pseudo-first order irreversible reaction. We
observe a good agreement for the activation energy with the previous works. The value
obtained by Haimour et al. (1987) is higher than those reported by the other authors. By
considering a pseudo first order reaction, Haimour et al. (1987) have neglected the MDEA
depletion in the boundary layer for the same experimental conditions for which Pani et al.
(1997) have observed a significant depletion. By the assumption of a pseudo first order
reaction, Haimour et al. (1987) have overestimated the activation energy of the reaction. This
result show that the hydrolysis rate of CO2 in MDEA solutions is second order, first order
with respect to CO2 and MDEA concentrations. Fig 4 shows the comparison of the kinetic
rates of three assumptions : a pseudo first order irreversible reaction tested by Rinker et al.
(1995), a second order irreversible reaction used by Pani et al. (1997), and a second order
89
reversible reaction treated in this work. The pseudo first order mechanism leads to an
overestimate of the kinetic rate, especially at high temperatures.
In his work, Rinker et al. (1995) have observed a significant influence of the reactions
neglected in mechanism 1. Taking into account this observation, mechanism 2 composed with
two reversible reactions, I and II, and two chemical equilibria , III and IV is considered. We
determined the kinetic rate of the forward reaction (I) for each experiments of Pani et al.
(1997) (Table 2). These results are fitted by the following Arrhenius law :
31311
.
k I , f = 163
. 102 exp (32)
T
The results (table 2) shows the significant influence of the reactions II, III and IV on
the estimates of kI,f. The estimates of the kinetic rates with a four reactions mechanism are
significantly lower than those obtained from single reaction mechanism considered above,
especially at high temperatures. This indicates that reaction II has a significant effect and
should not be neglected. This observation agrees with the conclusions of Rinker et al. (1995).
Figure 5 shows the difference between the results of Rinker et al. (1995), and those presented
in this work. We observe a greater influence of the reactions II, III and IV than Rinker et al.
(1995) did. Consequently, the forward kinetic rate of the reaction I calculated with our model
is lower than those of Rinker et al. (1995) (Fig 5). In order to estimate the influence of the
reaction II, we consider the mechanism 3 with a reversible reaction, reaction II, and an
instantaneous equilibrium, reaction IV. The experimental absorption rates measured by Pani
et al. (1997) and the predicted absorption ones obtained with our model using the mechanism
3 described above are listed in Table 4. At 298 K, we observe a great deviation between the
experiments and the simulations. This deviation decreases with temperature and, at 343 K, the
absorption phenomenon could be describe practically with the reaction II and IV.
Figure 6 shows that the concentration profiles of CO2 and MDEA obtained with the
mechanism 2 tend to those obtained with the mechanism 3 at 343 K. So, at high temperatures,
neglecting the reactions II and IV when estimating the forward reaction rate for the reaction I
leads to overestimate the contribution of this reaction in the total absorption rate. This
observation appears clearly on figure 6 : the MDEA depletion at the gas-liquid interface is
90
more important with the mechanism 1 than with the mechanism 3. The difference between the
model of Rinker et al. (1995) and our can be explain by an overestimation of the hydroxide
depletion at the gas-liquid interface. These overestimation leads to underestimate the
CO2/OH- reaction contribution.
Conclusion
The model presented makes it possible to calculate mass transfer of a typical, highly
complicated chemical absorption reaction system without convergence or stability problem.
The results model are in good agreement with analytical and numerical solutions of previous
authors. As an example, we have treated the case of CO2 absorption in MDEA aqueous
solutions. We have considered three kinetic mechanisms. The comparison of the kinetic laws
obtained with the mechanisms 1 (a single reversible reaction CO2/MDEA) and 2 (two
reversible reactions, CO2/MDEA and CO2/OH-, and two equilibrii, the MDEA protonation
and the CO2 second acidity) shows that the reaction between the CO2 and the hydroxide ion
can not be neglected. In this work, we found a greater influence of this reaction than Rinker et
al. (1995) did and the absorption of CO2 into MDEA aqueous solutions can be represented by
the mechanism 3 (a reversible reaction CO2/OH- and an equilibrium, the MDEA protonation)
at high temperatures.
This model can be used to estimate kinetic parameters which were obtained by
interpreting absorption or desorption rate data according to a realistic kinetic mechanism. It
will be used to interpret the CO2 desorption data obtained by Cadours et al. (1997) with
highly concentrated and CO2 loaded MDEA. If it is necessary to take into account the
influence of the ionic species on the equilibrium parameters at high loading, this model will
be coupled to rigorous thermodynamic representation to predict the desorption rates.
Acknowledgements
This work was supported by the Agence de lEnvironnement et de la Matrise de lEnergie
(ADEME) under Grant number 9674093 and by Elf Exploration Production under Grant
number EAP9611.
91
92
Notations
C concentration (mol.m-3)
D diffusivity (m2.s-1)
Ea energy of activation of the reaction between MDEA and CO2 (J.mol-1)
F Faraday constant (96,489 C.mol-1)
H Henrys law constant in the concentration scale, Pa.m-3.mol-1
k rate coefficient
K equilibrium constant
kG gas side resistance mass transfer (mol.m-2.s-1.Pa-1)
kL liquid side resistance mass transfer, (m.s-1)
NC total number of components
NG total number of gas
NR total number of reactions
P pressure (Pa)
R ideal gas constant (8.314 J.K-1.mol-1)
Ri production term for component i, (mol.m-3.s-1)
t time (s)
T temperature (K)
x distance from interface (m)
zi ion charge for component i
reaction order
defined by eq 19
film thickness (m)
l stoichiometric coefficient
viscosity (Pa.s)
absorption rate (mol.m-2.s-1)
electrostatic potential (V.m-1)
93
Subscripts and superscripts
aq sol aqueous solutions

b backward
bulk bulk
cal calculated
exp experimental
f forward
int interface
water water
~ dimensionless notation
Components abbreviations
CO2 carbon dioxide

2-
CO3 carbonate
HCO3- bicarbonate
H2O water
H3O+ hydronium
H2S hydrogen sulfide
HS- hydrosulfide
MDEA methyldiethanolamine
MDEAH+ protoned methyldiethanolamine
MEA monoethanolamine
MEAH+ protoned monoethanolamine
N2O nitrous oxide
OH- hydroxide
94
Literature cited
Al-Ghawas, H.A.; Hagewiesche, D.P.; Ruiz-Ibanez, G.; Sandall O.C. Physicochemical

Properties Important for Carbon Dioxide Absorption in Aqueous Methyldiethanolamine. J.
Chem. Eng. Data 1989, 34, 385.
Astarita, G.; Savage, D.W.; Bisio, A. Gas Treating with Chemical solvents; John Wiley &
Sons : New-York, 1983.
Barth, D.; Tondre, C.; Delpuech, J.J. Kinetics and Mechanisms of the Reactions of Carbon
Dioxide with Alkanolamines : a Discussion Concerning the Cases of MDEA and DEA. Chem.
Eng. Sc. 1984, 39, 1753.
Blauwhoff, P.M.M.; Versteeg, G.F.; van Swaaij, W.P.M. A Study on the Reaction between
CO2 and Alkanolamines in Aqueous Solutions. Chem. Eng. Sc. 1984, 39, 207.
Cadours, R.; Bouallou, C.; Gaunand, A.; Richon, D. Kinetics of CO2 desorption from highly
concentrated and CO2 loaded MDEA aqueous solutions in the range 312 K - 383 K. Ind. Eng.
Chem. Res., in press.
Cornelisse, R.; Beenackers, A.A.C.M.; van Beckum, F.P.H.; van Swaaij, W.P.M. Numerical
Calculation of Simultaneous Mass Transfer of Two Gases Accompanied by Complex
Revesible Reactions. Chem. Eng. Sc. 1980, 35, 1245.
Critchfield, J.E. CO2 Absorption/Desorption in Methyldiethanolamine Solutions Promoted

with Monoethanolamine and Diethanolamine : Mass Transfer and Reaction Kinetics. Ph.D.
Dissertation, University of Texas, Austin, 1988.
Critchfield, J.E.; Rochelle, G.T. CO2 Absorption into Aqueous MDEA and MDEA/MEA
Solutions, AIChE National Meeting, Houston, march 1987; paper 43E.
Danckwerts, P.V. Gas Liquid Reaction; McGraw-Hill : New-York, 1970.
95
Danckwerts, P.V.; Sharma, M.M. The absorption of carbon dioxide into solutions of alkalis
and amines (with some notes on hydrogen sulfide and carbonyl sulfide). Chem. Engr. 1966,
October, CE244-CE280.
Donaldson, T.L.; Nguyen, Y.N. Carbon Dioxide Reaction Kinetics and Transport in Aqueous
Amine membranes. Ind. Eng. Chem. Fundam. 1980, 19, 260.
Glasscock, D.A.; Rochelle, G.T. Numerical Simulation of Theories for Gas Absorption with
Chemical Reaction. AIChE J. 1989, 35, 1271.
Hagewiesche, D.P.; Ashour, S.S.; Al-Ghawas, H.A.; Sandall, O.R. Absorption of carbon
Dioxide into Aqueous Blends of Monoethanolamine and N-Methyldiethanolamine. Chem.
Eng. Sc. 1995, 50, 1071.
Haimour, N.; Bidarian, A.; Sandall, O.C. Kinetics of the Reaction between Carbon Dioxide
and Methyldiethanolamine. Chem. Eng. Sc. 1987, 42, 1393.
Kigoshi, K.; Hashitani, T. The Self-Diffusion Coefficients of Carbon Dioxide, Hydrogen

Carbonate Ions and Hydrogen Carbonate Ions in Aqueous Solutions. Bull. Chem. Soc. Japan
1963, 36, 1372.
Littel, R.J.; Filmer, B.; Versteeg, G.F.; van Swaaij, W.P.M. Modelling of Simultaneous
Absorption of H2S and CO2 in Alkanolamine Solutions : the Influence of Parallel and
Consecutive Reversible Reactions and the Coupled Diffusion of Ionic Species. Chem. Eng.
Sc. 1991, 46, 2303.
Littel, R.J.; Van Swaaij, W.P.M.; Versteeg, G.F. Kinetics of Carbon Dioxide Absorption into
Aqueous Solutions of N-Methyldiethanolamine. AIChE J. 1990, 36, 1633.
Newman ,J.S. Electrochemical Systems, Prentice-Hall, Inc: Englewood Cliffs, NJ, 1973.
Oloffson, G.; Hepler, L.G. Thermodynamics of ionization of water over wide ranges of
temperature and pressure. J. Soln. Chem. 1975, 4, 127.
96
Pani, F.; Gaunand, A.; Cadours, R.; Bouallou, C.; Richon, D. Kinetics of Absorption of CO2
in Concentrated Aqueous Methyldiethanolamine Solutions in the Range 296 K to 343 K. J.
Chem. Eng. Data 1997, 42, 353.
Petzold, L.R. A decription of DDASSL : a differential/algebraic system solver. In Scientific

Computing, Stepleman R.S. et al. (eds.), IMACS/North-Holland Publishing Co. : Amsterdam,
1983.
Pinsent, B.R.W.; Pearson, L.; Roughton, F.J.W. The kinetics of combination of carbon
dioxide with hydroxide ions. Trans. Faraday Soc. 1956, 52, 1512.
Read, A.J. The first ionization constant of carbonic acid from 25 to 250 and to 2000 bar. J.
Soln. Chem. 1975, 4, 53.
Rinker, E.B.; Ashour, S.S.; Sandall, O.C. Kinetics and Modelling of Carbon Dioxide
Absorption into Aqueous Solutions of N-Methyldiethanolamine. Chem. Eng. Sci, 1995, 50,
755.
Secor, R.M.; Beutler, J.A. Penetration Theory for Diffusion Accompanied by a Reversible
Chemical Reaction with Generalized Kinetics. AIChE J., 1967, 13, 365.
Toman, J.J.; Rochelle, G.T. Carbon dioxide absorption rates and physical solubility in 50 %
aqueous methyldiethanolamine partially neutralized with sulfuric acid, AIChE National
Meeting, Houston, Texas, 1989; paper 56C.
Tomcej, R.A.; Otto, F.D. Absorption of CO2 and N2O into aqueous solutions of
methyldiethanolamine. AIChE J. 1989, 35, 861.
Versteeg, G.F. Mass Transfer and Chemical Reaction Kinetics in Acid Gas Treating
Processes. Ph.D. Dissertation, University of Twente, The Netherlands, 1986.
97
Versteeg, G.F.; van Swaaij, W.P.M. On the Kinetics between CO2 and Alkanolamines both in
Aqueous and Non-aqueous Solutions - II. Tertiary Amines. Chem. Eng. Sc. 1988a, 43, 587.
Versteeg, G.F.; van Swaaij, W.P.M. Solubility and Diffusivity of Acid Gases (CO2, N2O) in
Aqueous Alkanolamine Solutions. J. Chem. Eng. Data 1988b, 33, 29.
Yu, W.-C.; Astarita, G. Kinetics of Carbon Dioxide Absorption in Solutions of

Methyldiethanolamine. Chem. Eng. Sc. 1985, 40, 1585.
98
Table 1 : Parameters for MDEA system at 318 K used by Glasscock and Rochelle (1989).
Parameter Value Reference
KI 132 Critchfield and Rochelle
(1987)
KII 4.18 10-8 kmol.m-3 Critchfield and Rochelle
(1987)
KIII 5.96 10-4 kmol.m-3 Critchfield and Rochelle
(1987)
KIV non independent Critchfield and Rochelle
(1987)
kI.f 10 m3.kmol-1.s-1 Critchfield and Rochelle
(1987)
kII.b 3.47 104 m3.kmol-1.s-1 Astarita and al. (1983)
HCO2 50 atm.m3.kmol-1 Critchfield and Rochelle
(1987)
DCO2 1.62.10-9 m2.s-1
DMDEA 0.75 10-9 m2.s-1 Versteeg (1986)
DHCO3- 0.94 10-9 m2.s-1 Kigoshi and Hashitani
(1963)
DCO3= 0.7 10-9 m2.s-1 Kigoshi and Hashitani
(1963)
DOH- 4.5 10-9 m2.s-1 Newman (1973)
loading 0.005 molCO2/molMDEA -
[MDEA]T 2.0 kmol.m-3 -
99
Table 2 : Identification of the forward rate coefficients of the reaction I for mechanism 1 and 2 from the data of Pani et al.
(1997).
T CMDEA,T kI,f (Mechanism 1) kI,f (Mechanism 2)

-3 3 -1 -1
K mol.m m .mol .s m3.mol-1.s-1
295.95 844.4 6.33 4.70
295.85 1272.6 5.10 3.85
296.15 1984.1 3.82 3.11
295.95 1984.2 3.96 3.77
295.75 2583.2 3.52 3.37
295.25 3477.4 3.30 3.14
296.45 4379.1 5.92 5.76
296.45 4379.1 5.77 5.52
296.45 4379.1 4.37 4.10
296.45 4379.1 4.43 4.14
296.45 4379.1 3.87 3.60
296.45 4379.1 3.13 2.94
296.45 4379.1 3.62 3.42
296.45 4379.1 4.61 4.35
296.45 4379.1 3.11 2.89
317.75 838.0 19.93 6.71

317.55 1966.6 13.65 10.25
318.35 2556.7 13.17 9.80
318.15 3435.8 13.59 10.30
317.65 4324.1 16.80 12.78
343.45 826.7 51.08 13.21

343.45 826.7 46.81 11.49
343.55 1245.0 50.07 13.45
342.55 1939.2 58.65 22.27
342.65 1939.0 55.14 18.99
342.65 1939.0 52.32 16.02
342.55 1939.2 49.43 12.19
343.55 1937.9 54.27 17.96
343.55 1937.9 54.32 17.96
342.75 2519.8 47.87 12.61
342.95 3381.4 51.66 27.92
342.25 4251.2 63.20 36.28
100
Table 3 : Comparison with energies of activation from previous works
References T CMDEA PCO2 Ea
K 103 mol.m-3 105 Pa kJ.mol-1
Yu et al.(1985) 313-333 0.2-2.5 1 38.5
Haimour et al.(1987) 288-308 0.85-1.7 1 71.6
Critchfield (1988) 282-350 1.7 1 56.9
Versteeg and van Swaaij (1988a) 293-333 0.17-2.7 <1 42.3
Tomcej and Otto (1989) 298-308 1.7-3.47 0.95 42.7
Littel et al.(1990) 293-333 0.17-2.7 <1 48.1
Rinker et al.(1995) 293-342 0.85 1 46.1
Pani et al. (1997) 296-343 0.84-4.4 1-1.7 45.4
This work 296-343 0.84-4.4 1-1.7 44.3
101
Table 4 : Comparison between measured absorption rates from Pani et al. (1997) and predicted ones from the mechanism 3
T CMDEA,T exp cal with mechanism 3 Deviation
-3 -3 -2 -1 -3 -2 -1
K mol.m 10 mol.m .s 10 mol.m .s %
295.95 844.4 3.41 2.16 36.7

295.85 1272.6 3.63 2.18 39.9
296.15 1984.1 4.01 2.14 46.6
295.95 1984.2 3.64 1.87 48.6
295.75 2583.2 3.88 1.84 52.6
295.25 3477.4 2.95 1.20 59.3
296.45 4379.1 1.75 0.70 60.0
296.45 4379.1 2.28 0.84 63.2
296.45 4379.1 1.98 0.82 58.6
296.45 4379.1 1.88 0.79 58.0
296.45 4379.1 1.89 0.82 56.6
296.45 4379.1 2.66 1.07 59.8
296.45 4379.1 2.53 1.00 60.5
296.45 4379.1 1.59 0.71 55.4
296.45 4379.1 2.32 1.02 56.0
317.75 838.0 5.32 4.43 16.7

317.55 1966.6 5.76 3.68 31.1
318.35 2556.7 5.90 3.77 36.1
318.15 3435.8 5.27 3.34 36.6
317.65 4324.1 3.03 1.90 37.3
343.45 826.7 7.28 7.23 0.7

343.45 826.7 7.57 7.57 0.0
343.55 1245.0 8.24 7.98 3.2
342.55 1939.2 8.77 7.80 11.1
342.65 1939.0 8.70 7.80 10.3
342.65 1939.0 8.51 7.92 6.9
342.55 1939.2 8.08 7.76 4.0
343.55 1937.9 9.15 7.19 21.4
343.55 1937.9 9.14 8.42 7.9
342.75 2519.8 8.75 7.53 13.9
342.95 3381.4 8.77 6.76 22.9
342.25 4251.2 5.25 4.14 21.1
102
-3 [MEA] - [MEAH+] - [HS
[H2S] (mol.m )
(mol.m-3)
12,5 1250
[MEA]
10 1000
7,5 750
5 500
2,5 [MEAH+] 250

[H2S] [HS-]
0 0
0 0,25 0,5 0,75 1 1,25
Dimensionless distance from the interface
Figure 1 : Concentrations profiles for H2S absorption into MEA aqueous solution.
10 [CO2]
[MDEA]
Concentrations (kmol.m-3)
[MDEAH+]
10-1
[HCO3-]
[CO3--]
[OH-]
10-3
[CO2]*
[MDEA]*
[MDEAH+]*
10-5
[HCO3-]*
[CO3--]*
-7 [OH-]*
10
0 0.2 0.4 0.6 0.8 1
Figure 2 : Comparison between the concentrations profiles from this work ([-]) and from Glasscock and
Rochelle (1989) ([-]*) for the MDEA system at low CO2 partial pressure (CCO2,int = 10-4 kmol.m-3).
103
10 [CO2]
[MDEA]
Concentrations (kmol.m-3)
[MDEAH+]
10-1 [HCO3-]
[CO3--]
[OH-]
10-3
[CO2]*
[MDEA]*
[MDEAH+]*
10-5
[HCO3-]*
[CO3--]*
-7
10 [OH-]*
0 0.2 0.4 0.6 0.8 1

Figure 3 : Comparison between the concentrations profiles from this work ([-]) and from Glasscock and
Rochelle (1989) ([-]*) for the MDEA system at high CO2 partial pressure (CCO2,int= 10-2 kmol.m-3)
0.1
Pseudo first order irreversible reaction, (Rinker et al., 1995)
Second order irreversible reaction (Pani et al., 1997)

kinetic rate (m3.mol-1.s-1)
0.075 Second order reversible reaction (this work)
0.05
0.025
0
295 305 315 325 335 345
Temperature (K)
Figure 4 : Comparison of three kinetic assumptions for the mechanism 1.
104
0.06
Mechanism 1 (this work)
Kinetic rate (m .mol .s )
-1 -1
Mechanism 2 (Rinker et al., 1995)

0.04
Mechanism 2 (this work)
3
0.02
0
295 305 315 325 335 345
Temperature (K)
Figure 5 : Comparison of kinetic mechanisms.
CCO2 (mol.m-3) CMDEA (mol.m-3)

25 840
20 820
CO2 MDEA
15 800
Mechanism 1
Mechanism 2
10 780
Mechanism 3
5 760
0 740
0 0.2 0.4 0.6 0.8 1
Figure 6 : Concentration profiles from the different mechanisms for the CO2/MDEA system at 343K.
105
Chapter 6.
Acid gas partial vapor pressure over aqueous MDEA solutions
B. Lemoine, Yi-Gui Li, R. Cadours, C. Bouallou, D. Richon
Introduction
There have been a great number of investigations of the solubility of CO2 and H2S in
MDEA aqueous solutions. Only a few studies have reported experimental data for very low
gas loadings. Jou et al. [1, 11, 13], Rogers et al.[15] have reported CO2 and H2S solubility in
35 wt% MDEA solutions at 313 and 373 K, and H2S solubility in 50 wt% MDEA solution at
313 K for acid gas loadings lower than 0.1 mole of gas per mole of amine.
Experimental
The equilibrium apparatus (see figure 1) is composed of a glass cylinder closed at each
extremity by two stainless steel flanges. Two Viton gaskets ensure adequate sealing and avoid
glass-metal contact. The cell is provided with two propellers supported by a shaft maintained
vertically by two sapphire bearings on the metallic flanges. The shaft is driven magnetically
and the liquid phase and the vapor phase are agitated simultaneously. This technique avoids
leaking, friction and heat generation, which appear when using stems crossing the cell bottom
and top. The propellers speed is maintained constant at 100 rpm. The upper flange is equipped
with a septum for liquid and gas loading. This septum is covered with mercury in order to
avoid gas diffusion and then leaking. A tube passing through this flange is used to degas the
cell. The cell is provided with a differential pressure transducer, working in the range 0-1500
Pa. The pressure transducer is from MKS, type 315B 0-11 mm Hg full scale, it is fitted with a
MKS type 272 temperature controller and a MKS type 270 signal conditioner. It has a
calibration certificate within 0.03 Pa from MKS.
106
Figure 1 : Flow diagram of the apparatus for small differential vapor pressure measurements
107
ASM: Variable speed motor
CI : component inlet
DPT differential pressure
transducer
EC : equilibrium cell
GC : glass cylinder
Hg : mercury
LB : liquid bath
LP : liquid phase
MR : magnetic rod
MWG : movable washer gasket
O : Oven
PA : pressure actuator
PP : platinum probe
S : septum
TR : temperature regulation
Va P : vacuum pump
VP : vapor phase
108
In our laboratory we have performed some additional calibrations against a mercury
manometer and we have confirmed the accuracy of the MKS calibration within the mercury
manometer accuracy, ie within 210-2 mm Hg which is our own limiting accuracy and do not
affect the real precision announced by the manufacturer within the considered measuring
range. The transducer is thermostated at 407 K, in order to avoid liquid condensation in the
measuring chamber. A microcomputer fitted with an acquisition card is used to record the
pressure transducer signal. The lower flange is equipped with a movable container to trap a
liquid sample, which will act as a reference. The sealing of this reference container is
obtained by a Viton disk. The volume of the sample trapped in this way is about 1 % of the
total volume of the aqueous solution.
The whole cell is placed in a thermostated liquid bath. The liquid bath temperature and
the equilibrium temperature are measured by a platinum probe. The injection of various
volumes of gas in the cell yields the total volume available for gas and liquid in the
equilibrium cell. This volume corresponding to the cell volume and to the pressure transducer
volume is estimated at (68.90.1)10-6 m3. The transducer volume is given by the constructor:
4.010-6 m3.
This apparatus presents several advantages. First, the experimental procedure avoids
analysis of the gas and liquid phases. The gas loading is calculated from the volumes and the
amounts of gas and liquid introduced in the cell. Second, no inert gas was added to measure
acid gas solubility for low partial pressures. This technique avoids gas phase resistance and
diminishes the time to reach the equilibrium conditions. Third, we used a small liquid volume,
between 12 and 15 cm3. This volume is small in comparison with the liquid volumes,
generally encountered in the literature, above 100 cm3. By avoiding sampling methods, the
liquid volume in our apparatus remains constant during the experiment. This technique avoids
liquid reinjection during the experiments to maintain the liquid volume constant (Austgen et
al. [2]). The most important advantage of this apparatus is that the acid gas partial pressures
are directly obtained in the range 0-11 mm Hg. The gas phase composition at low sour gas
partial pressure is generally obtained by chromatographic analyses in the literature. This
technique implies the addition of an inert gas to get a sufficient pressure allowing sampling
the gas phase, and introduces uncertainties through the chromatographic measurements [3].
109
Water and MDEA are degassed independently, and mixed under a vacuum in desired
proportions. The amounts of water and MDEA are known separately by differential
weighings within 0.01 g. The aqueous mixture is introduced inside the degassed equilibrium
cell by gravity and under a vacuum through a hypodermic needle introduced in the septum.
Accurate differential weighings within 210-2 g of the glass flask before and after the transfer
yields the mass of solution present in the cell. The accuracy of the solvent composition is
within 10-4 in weight fractions.
The liquid phase volume, Vintroduced is obtained through the density correlation of Al-
Ghawas et al. [4]. Then stirring is started, about 2100 rpm at the beginning, to reach
thermodynamic equilibrium and reduced at 100 rpm during measurements. When thermal
equilibrium is reached, a liquid sample is isolated in the movable washer gasket moved
upward by a pneumatic actuator. One of the entrances of the differential transducer is then
closed. In this way, the closed arm of the pressure transducer works as a reference for water-
MDEA mixture. We assume that due to low acid gas content the differential pressure which is
measured is the acid gas partial pressure.
Given volumes of acid gases are then introduced through the septum with calibrated
chromatographic syringes. The amount of acid gas introduced is known from the volume
introduced at room temperature T0 and atmospheric pressure P0 :
P0 Vintroduced
N acid gases introduced = (1)
R T0
At the beginning, the differential pressure is very high, but decreases as acid gases
dissolve with the help of vigorous stirring. The pressure decrease down to its equilibrium
value is recorded as a function of time. The equilibrium is reached within 4 hours with CO2
injection, and 45 minutes with H2S injection. This time difference is explained by the
different kinetic mechanisms involved in each case. The reaction between MDEA and H2S
implies a proton transfer. It is generally considered instantaneous compared to the transfer
mechanism. CO2 cannot react directly with a tertiary amine. The Donaldson and Nguyen [5]
mechanism is the most accepted one to describe the reaction involved in the CO2 absorption
into aqueous solutions of tertiary amines. This theory leads to a finite rate reaction
mechanism. The different reaction rates between the MDEA and CO2 or H2S explain the
110
selectivity of aqueous solutions of MDEA with respect to hydrogen sulfide. The amount of
acid gas absorbed is then obtained from :
P Vgas phase P Vtransducer
N acid gases absorbed = N acid gases intoduced (2)
R Tequilibrium cell R Ttransducer
The thermal gradient between the equilibrium cell and the transducer is also neglected.
The equilibrium solution loading is then obtained by :
N acid gases absorbed

loading = (3)
Vliquid C MDEA
where Vliquid and CMDEA are the liquid phase volume and the MDEA concentration in
the aqueous solution.
The relative accuracy on the acid gas loading, in the worst case, is 510-2 for CO2 or
H2S.
The solubilities of CO2 in 23.63 wt% MDEA aqueous solutions at 298 K are given in
table 1. At this temperature the CO2 partial pressures were from 12 to 1636 Pa.
The solubilities of H2S in 23.63 wt% at 313 K and 11.83 wt% MDEA aqueous
solutions at 298 K are given in table 2. At this temperature the H2S partial pressures were
from 23 to 1611 Pa.
Thermodynamic expression
Li and Mather [6] used the simplified Clegg-Pitzer equations [7] to correlate the
solubility data for MDEA-CO2-H2O and MDEA-H2S-H2O systems over a wide range of
solvent concentration (12-50 wt% MDEA aqueous solutions), temperature (298-393 K),
partial pressure of acid gases (0.001-5290 kPa) and acid gas loading (0.001-1.0 mol CO2 or
H2S/mol MDEA) from different authors. They have shown that the simplified Clegg-Pitzer
[7] equations are capable of correlation and prediction of the solubility of an acid gas in a
mixed solvent system with chemical equilibria. They found the experimental VLE data
measured from different authors in the same temperature and the same initial MDEA
111
concentration were not coincident with each other due to the measurement errors, which
decreased the accuracy of the correlation and prediction by use of these equations. In this
work, we regress the Clegg-Pitzer parameters by incorporating our experimental data with
those from other authors in the whole temperature and acid gas loading ranges. The results
from three methods are compared.
The chemical equilibrium constants for the reactions and the Henry constants for CO2
and H2S used here are the same as in the paper of Li and Mather [6] and are only temperature
dependent. In our material balance we considered the ionic species of MDEAH+, HCO3- and
HS- and the neutral species of H2O, MDEA, CO2 and H2S and neglected the ionic species of
CO3=, S=, OH- and H3O+ in the aqueous phase. (From Posey-Rochelle [8] the equilibrium
concentrations of CO3= even in 50% MDEA solution and acid loading near 1 at 313 K are less
than 1mol % in liquid phase, explicitly the concentration of CO3= is 100 times lower than the
concentration of HCO3- for loadings lower than 0.3, which is our interest range). In our
calculation of the activity coefficients of the components in aqueous phase, the neutral species
of dissolved free gases (CO2 and H2S) are also neglected, because their concentrations are
low. We also neglect the non-ideality of the gas phase as in Li-Mathers paper [6]. Only the
activity coefficients for H2O, MDEA, MDEAH+, HCO3- and HS- in aqueous phase are
involved in our calculation for the non-ideality of the liquid phase. In this work, both water
and MDEA are treated as solvents. The reference state (or standard state) associated with
each solvent is the pure liquid at the system temperature and pressure. The adopted reference
state for the ionic species and molecular solutes (CO2 and H2S) is the ideal, infinitely dilute
aqueous solution (without MDEA) at the system temperature and pressure.
Thus, we can use the Henry constants of CO2 and H2S in pure water without any
additional approximate assumption. The chemical equilibrium constants in this work should
adopt the same reference states for each component in equilibrium. According to our
simplified Clegg-Pitzer equations, the activity coefficient expressions for each component are
as follows :
ln 1 = 2 Ax I x /(1 + I x ) x M x X BMX exp(I x

3/ 2 1/ 2 1/ 2 2
) + A12 (1 2 x1 ) x2 + 2 A21 x1 x2 (1 x1 )
+ (1 x1 ) x I W1,MX x2 x I W2,MX
(4)
112
ln 2 = 2 Ax I x /(1 + I x ) x M x X BMX exp(I x
3/ 2 1/ 2 1/ 2 2
) + A21 (1 2 x 2 ) x1 + 2 A12 x1 x 2 (1 x 2 )
+ (1 x 2 ) x I W2, MX x1 x I W1, MX
(5)
ln M = z M Ax [(2 / ) ln(1 + I x ) + I 1x / 2 (1 2 I x / z M ) /(1 + I x )] + x x BMX g (I x

* 2 1/ 2 2 1/ 2 1/ 2
)
x M x X BMX [ z M g (I x ) / 2 I x + (1 z M / 2 I x ) exp(I x
2 1/ 2 2 1/ 2
)] 2 x1 x 2 ( A12 x 2 + A21 x1 )
+ x1 (1 x I )W1, MX + x 2 (1 x I )W2,MX W2, MX
(6)
if z M = z X ,
ln X = ln M
* *
Here the subscripts 1, 2, 3, M and X represent MDEA, H2O, CO2 (or H2S), MDEAH+
and HCO3- (or HS-), respectively. In the above equations the concentrations for the
equilibrated components in the liquid phase are expressed as mole or ionic fractions (x).
Superscript * represents the infinitely dilute aqueous solution without amine as the reference
state. The values without superscript * represent the pure solvent liquid as reference state.
x I = x M + x X = 1 x1 x 2
g ( x) = 2[1 (1 + x) exp( x)] / x 2
2 2
I x = ( xM z M + x X z X ) / 2
Ax = A ( C n )1 / 2
n
A = 0.391(78.54 * 298.15 / DmT ) 3 / 2 (d m / 0.99702)1 / 2

= 2150(18.02d m C n / 1000 DmT )1 / 2
n
Dm = n Dn
n
n = C nVn / C nVn
n
I x = ( z i C i )1 / 2
1/ 2 2
= 2I 1/ 2
(7)
i
Here Ix and I are the ionic strength expressed in ionic fraction and molarity,
respectively. Ax is the Debye-Hckel parameter on a mole fraction basis and A is the original
Debye-Hckel parameter (Pitzer, [9]), which is a function of temperature (T), density (dm)
and dielectric constant (Dm) of the mixed solvents (MDEA + H2O). The parameter is related
to the hard-core collision diameter, or distance of closest approach, of ions in solution. n is
the volume fraction of solvent n and Vn is the molar volume of the pure solvent n. The
113
temperature and concentration (weight percentage) dependence of the density of aqueous
MDEA solution was re-regressed in this work as polynomial expression from the
experimental density data of Teng et al. [10]. The temperature dependence of the dielectric
constants for pure H2O and pure MDEA were listed in the paper of Li and Mather [6]. The
molar concentration is expressed as C in aqueous phase.
The equations for the material balance and the calculation of the chemical equilibria
are shown in appendix.
Data regression
For each ternary system, the regressed interaction parameters are BMX, W1,MX and
W2,MX, which are all temperature dependent as B (or W or A)= a + b/T. The interaction
parameters A12 and A21, which are also temperature dependent, are parameters between
solvent molecules MDEA and H2O. The definitions of these interaction parameters are
expressed in Clegg and Pitzers original paper [7]. The objective function used in this study is
as follows:
F = (1/ n) ( pcal pex ) / pex

n
114
Table 1 : CO2 solubility into MDEA aqueous solutions (wt % MDEA = 23.63)
T Loading PCO2 T Loading PCO2

(K) (molCO2/molMDEA) (Pa) (K) (molCO2/molMDEA) (Pa)
297.70 0.0171 20 297.72 0.1295 477
297.70 0.0342 62 297.73 0.1361 546
297.72 0.0512 114 297.72 0.1617 716
297.71 0.0648 165 297.70 0.2042 1075
297.71 0.0853 256 297.72 0.2550 1593
297.71 0.0971 304 297.67 0.2625 1636
297.70 0.1126 402
Table 2 : H2S solubility into MDEA aqueous solutions
T wt % MDEA Loading PH2S T wt % MDEA Loading PH2S

(K) (mol H2S/molMDEA) (Pa) (K) (mol H2S/molMDEA) (Pa)
313.08 23.63 0.0151 40 298.15 11.83 0.0101 31
313.09 23.63 0.0301 113 298.16 11.83 0.0202 66
313.11 23.63 0.0452 224 298.16 11.83 0.0606 176
313.13 23.63 0.0601 355 298.15 11.83 0.1010 329
313.13 23.63 0.0754 509 298.17 11.83 0.1208 423
313.12 23.63 0.1049 887 298.13 11.83 0.1610 661
313.14 23.63 0.1346 1352 297.95 11.83 0.2011 945
313.12 23.63 0.1494 1611 298.16 11.83 0.2411 1278
313.11 23.63 0.0356 183 298.16 11.83 0.2610 1468
313.27 23.63 0.0497 291 299.34 11.83 0.0154 23
313.25 23.63 0.0638 416 298.16 11.83 0.0233 49
313.19 23.63 0.0920 735 298.17 11.83 0.0308 71
313.25 23.63 0.1201 1128 298.13 11.83 0.0463 129
298.16 11.83 0.0770 218
298.17 11.83 0.1384 482
298.15 11.83 0.2299 1091
exp. data (this work)
10 2 exp. data [11]
exp. data [14] 298 K (this work)
exp. data [3] 103 313, 343, 373, 393 K [14]
exp. data [15]
298, 313, 343, 373, 393 K [11]
101 correlation 298 K (this work)
313 K, [15]
pressure of H2S (kPa)
10 3 298, 313, 343, 373, 393

102K [11] correlation
313, 343, 373, 393 K [14]
313 K [3]
partial pressure of CO2 (kPa)
10 2 313 K [15] 1
1030 393 K 10
partial pressure of CO2 (kPa)
10
393K 373 K 343 K
exp. data [13] 10 1 393 K 373 K
313 K
correlation 100 343 K
102 373 K 103 313, 343, 373, 393 K [14]
-1 10 0
10
313 K 313 K 313, 343, 373, 393 K [11]
298 K
partial pressure of CO2partial
10-12 298 K
313 K [1]
(kPa)
1 393 K 10 -1 10 313 K [15]

10
correlation
partial pressure of H2S (kPa)
10-2 353 K 10-2

10 -2 101
MDEA(23wt% )-H2S-H2O M D EA (23w t% )-C O 2 -H 2 O MDEA(50wt%)-CO2-H2O
100
313 K 393 K
-3
10 -3 10
100 373 K
10-3 298 K10 -3 10 -2 10 -110-3 10 0-2
10343 K 313 K10
-1
100
10-1 -2
10 10 -1 C O10
2
0
loading (m ol C O 2 /m ol M DEA)
CO2 loading (mol CO2/mol MDEA)
10-1
H2S loading (mol H2S/mol MDEA)
Figure
10-2 3. Comparison of correlated and experimental Figure 4. Comparison of correlated and experimental
MDEA(50wt%)-H 2S-H2O
30wt%MDEA-CO2-H2O 10-2
values of CO2 partial pressure for 30% MDEA-CO2-H2O values of CO2 partial pressure for 50% MDEA-CO2-H2O
system with new parameters system with new parameters
10-3
10-3 10-2 10-1 100 10-3 10-2 10-1 100
115
CO2 loading (mol CO2/mol MDEA) H2S loading (mol H2S/mol MDEA)
Table 3. Experimental solubility data used to re-regress the interaction parameters and the correlation results.
MDEA-CO2-H2O ternary system

Reference MDEA conc, temp (K) data points %(corr)
wt%
This work 23.63 298 13 8.28
This work 50.0 298 13 44.0
MacGregor- 23.3 313 4 20.7
Mather [3]
Jou et al. [13] 30.0 298, 313, 353, 393 24 26.6
Jou et al. [1] 35.0 313, 373 32 23.5
Jou et al. [11] 23.3 313, 343, 373, 393 27 27.7
Huang-Ng [14] 23.1 313, 343, 373, 393 21 21.5
Huang-Ng [14] 50.0 313, 343, 373, 393 34 17.3
Rogers et al. [15] 23.0 313, 323 20 62.7
Rogers et al. [15] 50.0 313 14 43.5
Total 202 28.7
MDEA-H2S-H2O ternary system
Reference MDEA conc, temp (K) data points %(corr)
wt%
This work 11.83 298 16 5.34
This work 23.63 313 13 27.7
MacGregor-
Figure 3 23.3 313 20 23.7
Mather [3]
Jou et al. [1] 35.0 313, 373 28 Figure 4. 20.4
Jou et al. [1] 48.8 313 12 19.5
Jou et al. [11] 23.3 313, 373 17 55.1
Huang-Ng [14] 23.1 313, 373, 393 9 25.4
Huang-Ng [13] 50.0 313, 343, 373, 393 26 17.6
Rogers et al. [15] 50.0 313 10 22.3
Total 151 24.1
Table 4. Re-fitted values of interaction parameters B (or W) = a + b/T
MDEA-CO2-H2O system a b, K B (or W)(at 313 K)

BMX 106.302 -59515.65 -83.75275
W1,MX 15.43094 -4443.233 1.242103
W2,MX -7.014513 1100.45 -3.500382
MDEA-H2S-H2O system a b, K B (or W)(at 313 K)
BMX 536.4005 -188076.2 -64.19418
W1,MX 8.648682 -2366.862 1.090444
W2,MX 2.077459 -1182.494 -1.698668
Subscripts: 1 = MDEA, 2 = H2O, M = MDEAH+, X = HCO3-.
Generally, the industry requires parameters, which are valid in the whole temperature,
acid gas loading and amine concentration ranges. So we have to re-regress the parameters by
use of the experimental data covering the wide ranges of temperature from 298 to 393 K and
the5.acid
Figure gas loading
Comparison from 0.001
of correlated and to about 1.0 and the MDEA concentration from 10 to 50 wt%.
experimental values of H2S partial pressure for 23% Figure 6. Comparison of correlated and experimental
We used the method of Weiland
MDEA-H2S-H2O system with new parameters. et al. [12] to discard
values ofsome bad pressure
H2S partial data points. HereMDEA-H
for 50% we only
2S-
H2O system with new parameters.
considered the Debye-Hckel term and the A12 and A21 terms for the non-ideality of MDEA-
116
H2O system. The average deviation of the correlation is still very large due to systematic
experimental errors. So we have to plot a lot of equilibrium curves at the same temperatures
and the same MDEA initial concentrations but from different authors and decided to discard
some sets of data far away from the most of the data. Our third option is to re-regress the
Clegg-Pitzer parameters with the selected experimental data from Jou et al. [1, 11, 13],
MacGregor-Mather [3], Huang-Ng [14], Rogers et al. [15] and ours. The correlation results
are listed in Table 3 and are shown in Figures 2-6. The interaction parameters with their
temperature coefficients thus regressed are listed in Table 4. The interaction parameters
between solvents A12 and A21 are the same as those in Li-Mathers paper [6] and are not listed
here again. From the tables and figures it can be seen that the correlation results are fair for
our data and are poor for the Rogers et al. [15] data. But they are still good for the Jous data
[1, 13]. From these figures it can be seen also that the discrepancies between the experimental
data of Jou et al. [11] and ours are serious in low acid loading particularly in 50% MDEA
solution. The slope of the experimental curves from Rogers et al. [15] is located between
those from Huang-Ng [14] and Jou et al. [11]. The slope from ours is located near that of
Huang-Ng [14].
Conclusion
An original apparatus has been designed and used here to measure very low partial
vapor pressures of acid gases over MDEA aqueous solutions. Data obtained within 1 Pa are
compared with data available from literature. There are large discrepancies among the
measured VLE data and the slopes of their solubility curves from different authors due to
systematic experimental errors, particularly in the low acid loading regions. The original Li-
Mather parameters [6] can be used to predict our experimental VLE data in low acid loading
regions (<0.35) except for the 50wt% MDEA-CO2-H2O system at 298 K. The simplified
Clegg-Pitzer equation can be used only to correlate our own experimental low acid loading
and low temperature VLE data with good accuracy. But the parameters thus obtained cannot
be used to predict the VLE data in high temperature range (up to 373 K) and high acid
loading (>0.3). In this case, the prediction capability will be limited. Our experimental VLE
data can be incorporated with those from MacGregor and Mather [3], Jou et al. [1, 11, 13],
Huang and Ng [14] and Rogers et al. [15] to re-regress the interaction parameters with
acceptable correlation results in the very wide ranges of temperature (298-393 K), acid
117
loading (0.001-1.0) and MDEA concentration (11.83-50.0 wt%), which is very useful to the
gas purification process.
Appendix
The chemical equilibria considered in this work are the same as those of Li and
Mather [6]:

CO2 + 2 H 2 O H 3O + + HCO3
K1
H 2 S + H 2 O H 3 O + + HS
K2
H 2 O + MDEAH + H 3 O + + MDEA
K3
The equations for the material balance and the calculation of the chemical equilibria For
MDEA-CO2H2O ternary system:
CO2 + MDEA + H 2 O MDEAH + + HCO3

K4
K 4 = K 1 / K 3 = x M M x X X / x1 1 x 2 2 x3 3
* * *
K 4 X = x M x X / x1 x 2 x3 = K 4 1 2 3 / M X = C M C X (C1 + C 2 + C 3 + C M + C X ) / C1C 2 C 3
* * *
0
C1 = C1 C M
0
C2 = C2 CM
C 3 = C1 C M
0
C X = CM
( K 4 X 1)C M + [C 2 + C1 (1 + )](1 K 4 X )C M + [C 2 + (C1 + C 2 ) ]K 4 X C1 C M C1 C 2 K 4 X = 0
3 0 0 2 0 0 0 0 02 0
K4X is the apparent chemical equilibrium constant, which changes with the concentration of
each species. C0 is the initial concentration in aqueous solution. is the CO2 loading (total
moles of CO2/moles of original MDEA) in liquid phase. The equilibrated concentrations of
118
MDEAH+ and the free dissolved CO2 can be obtained by solving the above cubic equation
with Newton iterative method based on the chemical equilibrium and material balance.
pCO2 = H CO2 x3 3 = H CO2 xM M xX X / K 4 x1 1 x2 2

* * *
K 4 = K1 / K3
In the expression of K4, we used the Henry's constant of CO2 (HCO2) instead of the x33* term.
So we did not calculate the activity and activity coefficient of CO2 in the liquid phase.
For MDEA-H2S-H2O ternary system :

H 2 S + MDEA MDEAH + + HS
K5
K 5 = K 2 / K 3 = x M M x X X / x1 1 x3 3
* * *
K 5 X = x M x X / x1 x3 = K 5 1 3 / M X = C M C X / C1C 3
* * *
0
C1 = C1 C M
0
C2 = C2
C 3 = C1 C M
0
C X = CM
K5X is the apparent chemical equilibrium constant, which is dependent on the concentration of
each species. is the H2S loading (total moles of H2S/ moles of original MDEA) in liquid
phase. The equilibrated concentrations of MDEAH+ and the free dissolved H2S can be
obtained by solving the following quadratic equation.
(1 K 5 X )C M + K 5 X C1 (1 + )C M + K 5 X C1 = 0
2 0 02
p H 2 S = H H 2 S x3 3 = H H 2 S x M M x X X / K 5 x1 1
* * *
We also used the Henry constant of H2S (HH2S) instead of the x33* term. So we did not
calculate the activity and activity coefficient of H2S in liquid phase directly.
As shown above, only the activity coefficients for H2O, MDEA, MDEAH+, HCO3- and HS-
are involved in our calculation for non-ideal mixed solvent system. Thus our material
balances are simpler than the traditional Kent-Eisenberg model [16].
Notations
119
Ax = Debye-Hckel parameter on a mole fraction basis
A = interaction parameter between and among neutral molecules
a, b = coefficients for interaction parameters
B = interaction parameter between ions
F = objective function
Ix = ionic strength on mole fraction basis
P = partial pressure, Pa or kPa as noted
T = absolute temperature, K
W = interaction parameter between and among neutral and ionic species
X = liquid phase mole fraction based on the true species, molecular and ionic
Z = valence of an ion
Greek Letters
= Pitzer universal constant
= CO2 or H2S loading in the liquid phase, mol of CO2 or H2S/mol of MDEA
= activity coefficient in mole fraction base
= average relative deviation of partial pressure, %
= Pitzer parameter relating to the hard-core collision diameter between ions
Superscripts
* = unsymmetric convention
Subscripts
1 = MDEA
2 = H2O
M = MDEAH+
X = HCO3- or HS-
cal = calculated value
corr = correlation
ex = experimental value
pred = prediction
Literature cited
120
[1] F.-Y. Jou, J.J. Carroll, A.E. Mather, F.D. Otto, Can. J. Chem. Eng. 71 (1993) 264-268.
[2] D.M. Austgen, G.T. Rochelle, C.-C Chen,. Ind. Eng. Chem. Res. 30 (1991) 543-555.
[3] R.J. MacGregor, A.E.Mather, Can. J. Chem. Eng. 69 (1991) 1357-1366.
[4] H.A. Al-Ghawas, D.P. Hagewiesche, G. Ruiz-Ibanez, O.C. Sandall, J. Chem. Eng. Data 34
(1989) 385-391.
[5] T.L. Donaldson, Y.N. Nguyen Ind. Eng. Chem. Fundam. 19 (1980) 260-266
[6] Y.-G. Li, A.E. Mather, Ind. Eng. Chem. Res. 36 (1997) 2760-2765.
[7] S.L. Clegg, K.S. Pitzer, J. Phys. Chem. 96 (1992) 3513-3520.
[8] M.L. Posey, G.T. Rochelle, Ind. Eng. Chem. Res., 36(9) (1997) 3944-3953.
[9] K.S. Pitzer, Ion Interaction Approach Theory and Data Correlation. In Activity
Coefficients in Electrolyte Solutions, 2nd ed.; Pitzer, K.S., Ed.; CRC Press: Boca Raton, FL,
1991; Chapter 3.
[10] T.-T. Teng, Y. Maham, L.G. Hepler, A.E. Mather, Physical Properties of Aqueous
Solutions of MDEA, DEA and Their Mixtures. Proceedings of the second International
Symposium on Thermodynamics in Chemical Engineering and Industry. Beijing, China,
1994.
[11] F.-Y. Jou, A.E. Mather, F.D. Otto, Ing. Eng. Chem. Process Des. Dev. 21 (1982) 539-
544.
[12] R.H. Weiland, T. Chakravarty, A.E. Mather, Ind. Eng. Chem. Res. 32(7) (1993) 1419-
1430.
[13] F.-Y. Jou, F.D. Otto, A.E. Mather, Ind. Eng. Chem. Research, 33 (1994) 2002-2005
[14] S.-H. Huang, H.-J.. Ng, Final report to the Gas Research Institute, GRI-95/0126, Feb.
1996.
[15] W.J. Rogers, J.A. Bullin, R.R. Davidson, AIChE J. 44(11) (1998) 2423-2430.
[16] R.L. Kent, B. Eisenberg, Hydrocarbon Process. 55(2) (1976) 87-90.
121
Chapter 7.
Representation of regeneration of CO2-Loaded MethylDiEthanolAmine Aqueous

solutions in the range 313-383 K
Introduction
For both energy and environmental purposes, acid gases must be removed from natural gas, as
well as in petroleum refining, coal gasification and hydrogen production. For such a
purification, absorption with aqueous solutions of alkanolamines remains the dominant
industrial technology. The design of absorbers and strippers requires information about
transfer and kinetic mechanisms, and the related rate expressions. Although desorption
processes often act as regenerative processes for absorption systems, studies devoted to
desorption are not as numerous as those concerning absorption. The regeneration often uses
steam or heated inert gas as stripping agents, then an accurate design based on proper
chemical and mass-transfer data could lead to significant energy savings. Aqueous solutions
of alkanolamines are commonly used for acid gases removal from gaseous streams. H2S and
CO2 are absorbed into the solution and desorbed from the solution by steam stripping.
Monoethanolamine (MEA) and diethanolamine (DEA) have been the most widely used
amines in gas treating processes. Methyldiethanolamine (MDEA) solutions with the
advantages of H2S selective removal, and lower enthalpy of reaction and vapor pressure than
primary or secondary alkanolamines solutions has found an increased use. The design of such
separation equipment also requires the knowledge of vapor-liquid equilibrium as deviation
from equilibrium provides the force to drive the kinetically controlled process.
Cadours et al. [1] have designed a batch reactor (modified Lewis type cell) : from
equilibrium, total pressure over the liquid solution is slightly decreased by a rapid
withdrawing of a small vapor volume, then the system is left to reach new equilibrium
conditions, pressure increase due to desorption of the acid gas is recorded as a function of
time. The experimental procedure is the following : the aqueous alkanolamine solution is
introduced at a given composition into the reactor. The acid gas is bubbled into the solution to
prepare loaded solutions. At equilibrium, a controlled pressure drop inside the Lewis type cell
lead to gas desorption. The controlled pressure drop is obtained by connecting, for a short
122
period, the gas phase of the cell with an adjustable volume which is initially under a vacuum.
Very small pressure drops are performed to prevent any gas nucleation and then keep constant
the liquid-vapor interface area. This technique gives fast measurements (less than 3 hours for
one desorption curve) with a small quantity of solvent : about 190 cc. Typical reproducibility
measurements have shown that the dispersion of desorption rates is about 10 percent. Rates of
carbon dioxide desorption have been measured in aqueous solutions of methyldiethanolamine.
Data are now available over the temperature range : 293 - 383 K, for 25 - 50 weight percent
MDEA-Water solvents and different CO2 loadings ranging from 0.10 to 0.80 mole CO2 / mole
MDEA.
Cadours and Bouallou [2] have developed a general model for the diffusion/reaction
processes. They have considered the most rigorous approach of taking into account the
electrostatic potential gradient terms with unequal diffusion coefficients for the ionic species
even if much more computational effort is required. This model allows us to consider
reactions with finite rate and instantaneous reactions with respect to mass transfer. We
assume reversibility, generalized reaction rate expressions and generalized stochiometry. The
applicability of our model is not limited by the type of reactions modeled. Optimization using
a Levenberg-Marquardt algorithm provides the rate coefficients for different kinetic
mechanisms, from experimental data.
Equations render a set of non-linear coupled partial differential equations. The main advantage of our
method lives in the simpleness of the numerical treatment because no numerical method is needed to transform
each partial differential equations into a system of ordinary differential equations. An implicit centered finite
difference iterative method was employed to solve the system of coupled differential equations subject to the
boundary and initial conditions given above. The implicit method ensures stability for the system of equations.
From the initial concentrations field, the concentrations gradients are obtained by numerical derivation. Then the
electrostatic potential and its gradient are deduced. They enter as source terms in the transfer equations which
are solved using the finite-difference iterative method giving the solution for the next iteration. The calculations
are repeated until the convergence criterion is satisfied. The steady-state solution is found by iterating upon an
index (n). In this way, an efficient and consistent algorithm is obtained for the set of equations. The model
developed in previous works, combined with appropriate parameters will be used to model acid gas treating
systems in the general case.
Results and Discussion
123
The numerical tool described above was used to represent the regeneration of CO2-loaded
methyldiethanolamine aqueous solutions. We have considered a four-reactions mechanism
involving two finite rate reactions, reactions I and II, and two instantaneous equilibria
involving a proton transfer, reactions III and IV.
CO2 + MDEA + H 2 O MDEAH + + HCO3 (I)
HCO3 CO2 + OH
(II)
CO23 + H 2 O = HCO3 + OH
(III)
MDEA + H 2 O = MDEAH + + OH
(IV)
In the case of CO2-loaded methyldiethanolamine aqueous solutions, the consistency

between the kinetic and thermodynamic models is essential for the representation of the
desorption phenomenon. It is necessary to have a thermodynamic model giving a good
representation of the gas-liquid equilibrium, and to have the reaction rate constants associated
with this thermodynamic model. The CO2 solubility in loaded MDEA aqueous solutions is
calculated using Henrys law. The salting-out effect needs to be known because of its
influence on the estimation of the mass transfer rate. Weiland [3] suggested the following
model :
H CO 2 , 0
log 10 = ( h+ + h + hG ) I (1)
H CO , = 0
2
where I is the ionic strength
1
I=
2
Ci z2i (2)
The parameters h+ and h- are specific to the cation and the anion, respectively, and were
considered temperature independent. Different sets of the gas-specific constants hG were
provided for each temperature. Until now, the assumption made by Weiland et al. [4]
124
consisting in using the salt coefficient of NH4+ for a secondary amine and the salt coefficient
of CO32- for HCO3- were generally used to estimate the influence of the ionic strength on the
Henry's law constant. This very reducing assumption led to an unrealistic estimate of the
influence of the ionic strength. We will use the values recapitulated in table I :
Table I.: Salt coefficient, 10-3 m3.mol-1 [3]

Temperature
Species
313 K 323 K 333 K
MDEAH+ 0.041 0.041 0.041
HCO3- 0.073 0.073 0.073
CO2 -0.026 -0.029 -0.016
First, we have selected desorption experiments which were consistent with the
thermodynamic model used in the general mass transfer model (Tableau II.). The actual lack
of thermodynamic parameters prevent us to represent the CO2 desorption rate data obtained at
high temperatures. The selected experiments were also limited to CO2 loading ranging from
0.2 to 0.4 molCO2/molMDEA. For each selected experiment at 313, 323 and 333 K, the
desorption rates were estimated using previous works correlations. For the calculation of the
density and the viscosity of the solution, we used the correlations proposed by Glasscock [5].
The diffusion coefficient of CO2 was estimated using the data and the N2O analogy of
Versteeg and van Swaaij [6]. The equilibrium constants for equation I to IV were similar to
those used by Cadours and Bouallou [2]. The CO2 Henry constant in unloaded MDEA
aqueous solutions was calculated with the correlation proposed by Al-Ghawas et al. [7] for
temperatures between 288 K and 323 K and for MDEA compositions up to 50 wt % MDEA.
The mass transfer correlation established by Cadours et al. [1] was used to estimate the liquid
side mass transfer resistance. The values of the backward rate of reaction II were calculated
from the correlation proposed by Pinsent et al. [8]. The Arrhenius law of reaction I is taken
from Cadours and Bouallou [2] :
3131.1
k I,f = 1.63 10 2 exp (3)
T
125
For each experiment, by taking again the kinetic law (3) for the reaction I between CO2
and the MDEA, we determined experimental rate corresponding to the regeneration of CO2-
loaded methyldiethanolamine aqueous solutions. The comparison between experimental
desorption rate and predicted desorption rate reveals that a same tool can be used to represent
the phenomenon of absorption as well as that of desorption.
126
Table II. : stripping of CO2
N Temperature CMDEA.T P Experimental Predicted
desorption rate desorption rate
K mol.m-3 molCO2/molMDEA Pa Mol.m-2.s-1 mol.m-2.s-1
1 312.3 2107 0.30 169 3.80 10-06 3.69 10-06

2 321.8 2086 0.32 387 1.13 10-05 0.93 10-05
-05
3 321.8 2086 0.32 347 1.09 10 0.82 10-05
4 322.2 2086 0.32 249 6.81 10-06 5.52 10-06
-05
5 322.0 2086 0.32 393 1.01 10 0.94 10-05
6 322.0 2086 0.32 443 1.17 10-05 1.08 10-05
7 322.0 2086 0.32 399 1.12 10-05 0.96 10-05
-05
8 333.6 2321 0.31 686 2.28 10 1.83 10-05
9 333.9 2320 0.31 673 2.04 10-05 1.80 10-05
-06
10 312.7 4407 0.20 194 4.48 10 3.32 10-06
11 312.7 4407 0.20 181 4.48 10-06 3.08 10-06
-06
12 312.7 4407 0.20 173 3.42 10 2.93 10-06
13 322.1 4104 0.28 679 1.45 10-05 1.25 10-05
-05
14 322.1 4104 0.28 703 1.55 10 1.29 10-05
15 332.9 4630 0.23 811 2.24 10-05 1.98 10-05
-05
16 332.4 4217 0.22 884 2.35 10 2.15 10-05
17 332.4 4217 0.22 852 2.15 10-05 2.07 10-05
18 332.4 4217 0.22 840 2.28 10-05 2.04 10-05
-05
19 332.4 4217 0.22 658 1.67 10 1.58 10-05
20 332.5 4217 0.22 601 1.53 10-05 1.44 10-05
-05
21 332.5 4217 0.22 601 1.52 10 1.44 10-05
22 332.5 4217 0.22 422 1.24 10-05 1.00 10-05
-06
23 332.5 4217 0.22 307 9.71 10 7.10 10-06
24 332.6 4217 0.22 233 7.69 10-06 5.26 10-06
-05
25 332.3 4215 0.28 1624 3.34 10 3.54 10-05
26 332.4 4215 0.28 1449 2.98 10-05 3.15 10-05
-05
27 332.4 4215 0.28 1111 2.97 10 2.40 10-05
28 332.4 4211 0.33 1622 2.66 10-05 3.18 10-05
-05
29 332.5 4209 0.38 2201 3.86 10 3.87 10-05
It is noted that the experimental desorption rates are well predicted by our numerical tool
taking into account coherence between the kinetic and thermodynamic data used. The
deviations observed for each test are presented in table III. Twenty one tests out of the twenty
nine is predicted with an uncertainty lower than that corresponding to kinetic-thermodynamic
coherence. The influence of the solution loading on regeneration is shown on figure 2. : at
given temperature, by increasing the solution loading, regeneration is slowed down. For a
given solvent concentration, the increase of the temperature accelerates regeneration (see
figure 3). For the plant design, a compromise between these parameters seems to be
necessary.
127
Table III. : relative deviation between experimental and predicted desorption rates
test n deviation ( % ) test n deviation ( % ) test n deviation ( % )
1 2.9 11 31.3 21 5.3
2 17.7 12 14.3 22 19.3
3 24.8 13 13.8 23 26.9
4 18.9 14 16.7 24 31.6
5 6.9 15 11.6 25 6.0
6 6.8 16 8.5 26 5.7
7 14.3 17 3.7 27 19.2
8 19.7 18 10.5 28 19.5
9 11.8 19 5.4 29 0.3
10 25.9 20 5.9
5,0E-05
4,0E-05
3,0E-05
Prediction
2,0E-05
1,0E-05
0,0E+00
0,0E+00 1,0E-05 2,0E-05 3,0E-05 4,0E-05 5,0E-05
Experimental
Figure 1. : Comparison between experimental and predicted regeneration of CO2-loaded

methyldiethanolamine aqueous solutions
128
3,0E-08
2,5E-08
Desorption rate/ P
2,0E-08
1,5E-08
1,0E-08
5,0E-09
0,0E+00
0,15 0,3 0,45
Loading
Figure 2. : Influence of the loading T= 333K, CMDEA,T = 4200 mol.m-3 MDEA
3,0E-08
2,5E-08
Desorption rate/ P
2,0E-08
1,5E-08
1,0E-08
25 wt% MDEA
5,0E-09
50 wt% MDEA
0,0E+00
310 315 320 325 330 335
Temperature
Figure 3. : Influence of temperature on regeneration
129
1,015
[CO2]/18,48
[MDEA]/2611,47
[MDEAH+]/1597,53
Dimensionless concentration
1,01 [HCO3-]/1563,93
[CO3=]/16,79
[OH-]/0,02
1,005
0,995
0 0,2 0,4 0,6 0,8 1
Dimensionless film thickness
Figure 4. : Adimensionnal concentration profiles obtained for CO2-loaded MDEA aqueous

solution
(CMDEA,T = 4209 mol.m-3, T = 332,5 K, = 0,38 molCO2/molMDEA).
As an example, Figure 4 represents the concentrations profiles of the various species
in solution in the case of the regeneration of CO2-loaded methyldiethanolamine aqueous
solutions (Test n 29). One notes there is no concentrations profiles for the species in
solution, except for the carbon dioxide.
Concluding remarks
In this work, we have considered a general model of mass transfer with chemical
reactions. The choice of the modified film theory , associated with an original method of
resolution, guarantees a precise representation of the considered phenomena. While
guaranteeing a coherence between the kinetic and thermodynamic parameters, this model was
used to represent regeneration of CO2-loaded methyldiethanolamine aqueous solutions.
However the regeneration of solvents by stripping uses considerable water vaporization in the
assessments matters. So the phase gas mass-transfer resistance is of primary importance in
stripper design. From a practical point of view, it is essential to take into account this
130
phenomenon in the present numerical tool to improve the regeneration columns design. In
addition, within the framework of stripper design, the solvent quality is of primary importance
in order to guarantee the standard specifications on treated gas. Thus, it will be necessary to
associate to our numerical tool a thermodynamic model available for the representation of the
low loading data such as those obtained by Lemoine et al. [9].
Notations
Ci Concentration of species i, mol.m-3

HCO2 Henry's law constant in the concentration scale, Pa.m3.mol-1
kf forward reaction rate constants for the reaction between CO2 and MDEA,
m3.mol-1.s-1
T temperature, K
loading of the solution, molCO2/molMDEA
P Depression from equilibrium pressure, Pa.
CO2 carbon dioxide
=0 unloaded solution
0 loaded solution
Literature cited
1 Cadours, R., Bouallou, C., Gaunand, A. & Richon, D. (1997) Kinetics of CO2
Desorption from Highly Concentrated and CO2-Loaded Methyldiethanolamine Aqueous
Solutions in the Range 312-383 K, Ind. Eng. Chem. Res, 36, 5384.
2 Cadours, R. & Bouallou, C.(1998) Rigorous Simulation of Gas Absorption into
Aqueous Solutions, Ind. Eng. Chem. Res., 37, 1063.
3 Weiland, R.H. (1996) Physical properties of MEA, DEA, MDEA and MDEA-Based
blends loaded with CO2, GPA Research Report RR-152.
4 Weiland R.H., Rawal M. & Rice R.G. (1982) Stripping of carbon dioxide from
monoethanolamine solutions in a packed column, AIChe J., 28, 963.
131
5 Glasscock, D.A. (1990) Modelling and experimental study of carbon dioxide absorption
into aqueous alkanolamines. Ph.D. Dissertation. University of Texas, Austin.
6 Versteeg, G.F. & van Swaaij, W.P.M. Solubility and diffusivity of acid gases (CO2, N2O)
in aqueous alkanolamine Solutions, J. Chem. Eng. Data, 33, 29.
7 Al-Ghawas, H.A., Hagewiesche, D.P., Ruiz-Ibanez, G. & Sandall, O.C. (1989)
Physicochemical properties important for carbon dioxide absorption in aqueous
methyldiethanolamine, J. Chem. Eng. Data, 34, 385.
8 Pinsent, B.R.W., Pearson, L. & Roughton, F.J.W. (1956) The kinetics of combination
of carbon dioxide with hydroxide ions, Trans. Faraday Soc., 52, 1512.
9 Lemoine, B., Yi-Gui, L., Cadours, R., Bouallou C. & Richon, D. (2000) Partial vapor
pressure of CO2 and H2S over aqueous methyldiethanolamine solutions, Fluid Phase
Equilib., 172, 261.
132
Chapter 8. Modeling of acid gases absorption column using alkanolamine solutions
D.-J. Vinel, C. Bouallou
Introduction
The packed columns used in the gas treatment processes by aqueous alkanolamine
solutions have specificities complicating their modeling. Mass and heat transfer are associated
chemical reactions within the liquid phase and in the particular case of regeneration, the
whole of the gas compounds transfers from one phase to the other. The chemical reactions are
instantaneous or kinetically controlled. These absorption and regeneration columns present
many analogies with the reactive distillation columns. This analogy makes it possible to use
the abundant literature concerning reactive distillation columns and to take as a starting point
the modeling of these columns. The type of column studied will be in priority the stage
column which is used industrially, nevertheless experimental measurements having been
realized on columns with packing, this type of column will have also to be studied. One
review will be made on the various manners of apprehending the stages, either at equilibrium
or non equilibrium with or without effectiveness. The sizes which intervene in the modeling
of a stage and column make it possible to establish the link between an elementary model
who determines the heat and molar fluxes through an interfacial unit area and a stage whose
interfacial area is known. The modeling of the stages columns at thermodynamic equilibrium
is frequently used or quoted in the articles treating reactive separations (Perez-Cisneros et al.
[1], Scenna et al. [2], Sneesby et al. [3], Taylor and Krishna [4], Baur et al. [5]). The set of
equations defining the equilibrium stage models is known under the name of "MESH"
(Material, Equilibrium, Summation, and Heat). Several authors (Isla et al. [6], Lee et al. [7])
used a modified theoretical stage model by incorporating an effectiveness term for the stage.
The definition most used for the stage j effectiveness with respect to a species i is that of
Murphee.
y ij,O y ij,I
E = j,
j
i
y i y ij,I (1)
In (1), y ij,O , y ij,I and y ij, represent respectively the average composition in the gas
phase leaving the stage j, the composition in the gas phase entering the stage j and the
133
composition of the gas phase at equilibrium with the liquid phase leaving the stage j. This
approach has the simple and fast advantage of being but the disadvantage of not considering
with precision the transfer phenomena.
The philosophy of this modeling is that the gas phase and the liquid phase are in
equilibrium only at the gas-liquid interface and nowhere elsewhere. This representation is not
completely satisfactory. The kinetics as well as the equilibrium constants of reactions depends
on the local concentrations and the temperatures which can vary in an important way along
the stage. Thus Higler et al. [8, 9, 10] developed a model of cells for the non equilibrium
stage. The idea is to divide the stage into a certain number of contact cells. These cells
represent a small zone of the stage where the liquid and gas compositions can be assumed
homogeneous and by choosing in an adequate way the channeling and recirculation of fluids,
the hydrodynamic problems and dead space can be apprehended. The columns with packing
are very interesting in all the processes requiring large interfacial surface between two phases.
The packing inside the columns can be arranged, thus offering a regular geometrical structure
throughout the column, or not arranged as the Raschig rings which are distributed by chance
within the column. In General, the arranged packing is preferred because the pressure loss is
less and the column hydrodynamic parameters are accessible.
Modeling of references absorption column
The modeling of the reference absorption columns is based on the use of the Murphee
mass effectiveness, E ij and of the thermal effectiveness, ETHj (Figure 1).
The temperature and concentration profiles for each stage of the column and the liquid
and the vapor composition at outlet are obtained by iteration on the gas molar fractions in
column overheading ( y1i ). A total assessment on each species then makes it possible to obtain
the liquid in bottom
Vn +1 y in +1 + L 0 x i0 = V1 y1i + L n x in (2)
134
V1
L0
G L
Vj Tj Lj-1 Tj1
Stage j
Eij et ETHj
L
Vj+1 TjG+1 Lj Tj
Ln
Vn+1
Figure 1: Diagram of the column and the theoretical stage
The calculation of the column is carried out since last stage N to first stage 1. For a
given stage j, the liquid outlet L j and the inlet vapor Vj+1 are known. An outlet theoretical
vapor is given by considering that the vapor at outlet of the stage is at equilibrium with the
liquid L j at outlet.
H Si .x ij .C Ltot, j
y ij,* = (3) The
Ptotj
real outlet gas ( y ij ) is expressed using the theoretical outlet gas at equilibrium by introducing
a coefficient of effectiveness E ij .
Vj y ij = Vj+1 .(y ij+1 E ij .( y ij+1 y ij,* ) ) (4)
The liquid inlet L j1 is deduced by an assessment mass on the stage j.
L j1 .x ij1 = L j .x ij + Vj .y ij Vj+1 .y ij+1 (5)
From a mass profile given for the column, the effectiveness E ij is calculated using the
enhancement factors Ea i of each compound i.
1
E ij = 1
Ptotj
exp Kog.A. (6)

0 . 5 .(V j + V )
j+1

135
1 1 Hi
= +
Kog k G k L Ea i (7)
Kog , A and Ptotj represent respectively the gas side mass transfer coefficient, the
interfacial area and total pressure at the stage j.

Enhancement factors Ea i for each compound i which transfers from the gas phase to
the liquid phase are calculated thanks to the Hatta number (Ha).
Ha
Ea i =
tanh(Ha ) (8)
L
D CO (k OH C OH + k am C A min es )
Ha = 2
In the case of CO2: kL (9)
the terms kOH and kam represent the constant kinetics of the reactions between CO2 and
OH- and CO2 and the alkanolamine. The composition of the outgoing vapor of the first stage
is compared with the initial composition at the overhead then this one is modified until
obtaining convergence of the column on the following criterion:
V1 y1i,init V1 y1i,calc

V1 y1i,init (10)
The calculation of the thermal profile of the absorption column uses a thermal
effectiveness E thj at the stage j. The calculation of the temperatures is carried out since the
columns bottom while going up towards the overhead. For a stage j, the temperature of the
gas phase outgoing TjG is calculated by:
(
TjG = TjG+1 + E thj . TjL TjG+1 ) (11)
E thj =1 is equivalent considering the liquid and outgoing gas in heat balance. By an
assessment enthalpic on the stage j, the enthalpy of the inlet liquid H Lj1 (TjL1 ) on the stage j is
given.
( ) ( ) ( ) ( )
L j1 . H Lj1 TjL1 + Vj+1 . H Gj+1 TjG+1 = L j . H Lj TjL + Vj . H Gj TjG (12)
136
From the enthalpy of the liquid H Lj1 (TjL1 ) , the composition of the liquid L j1 being
known, the temperature of the liquid TjL1 is deduced.
The convergence of the column temperature profiles is regarded as attack when the
criterion translating the equality between the initial temperature at the columns overhead and
the temperature at the overhead recomputed is satisfied.
T1G ,init T1G ,calc

(13)
T1G ,init
The modeling of the columns with the non equilibrium stage has the advantage of
retranscribing reality. This modeling is very general-purpose since all the types of packing or
stage columns and can be modeled easily. Many authors (Perez-Cisneros et al. [1], Scenna et
al. [2], Sneesby et al. [3]) use the MESH equations in reactive separations cases. However
reactive separations process is a having many similarities with the absorption and
regeneration columns which we want to model.
The reference simulation is based on the stage model, which describes the column as a
succession of stages at thermodynamic equilibrium. The transfer model is used to calculate an
enhancement factor as well as a thermal effectiveness. This approach has the simple and
effective advantage of being simpler. Our transfer model based on the double film theory
makes it possible to know the flux transferred to the interface which can be expressed in the
term of an enhancement factor and thus we can introduce our model within the industrial
simulator easily.
Model description
The first stage of this work is the development of a precise elementary model holding
account of the mass transfer of multiconstituant as well as heat transfer in liquid film and the
gas film (Figure 2). The theory used is that of the modified film of Chang and Rochelle [11]
which gives results equivalent to the penetration theory while being simpler.
Bulk gas Gas film Liquid film Bulk liquid
137
C bulk
k
liquid
Ni
C k (x)
Pibulk gas
Piint C int
i
C ibulk liquid
0
C i (x ) G + L x
Figure 2: Schematic diagram
G of the two films model
Mass transfer
We make the assumption of an ideal and diluted liquid medium where water is the
stagnant majority species. In this case, the Nernst-Planck equations are applied, in liquid film
for x [ G ; G + L ] . NC components are considered. NR reactions take place of which NRI
instantaneous reactions.
d 2Ci (x) F
L rday d ( ( x ) * C i ( x ))
D L
i Z i D i + R i (x ) = 0 (14)
dx 2 RT dx
In this equation, the variation of the diffusion coefficient according to the position in
liquid film is neglected. R i ( x ) represents the production term.
NR
R i ( x ) = i ,k .(rk ( x ) r k ( x ) ) (15)
k =1
i ,k , rk , r k represent respectively the stoechiometric coefficient of compound i and
the reaction rate constants. In the case of reversible reactions, like the reaction between CO2
and OH ions -, the reaction rate constants are expressed according (16) and (17). The
relationship between the constant kinetics k d and k d is the equilibrium constant of reaction
K eq
k (18).
NC
rk ( x ) = k d (Ci ( x ) )
i , k
(16)
i =1
i , k >0
138
NC
r k ( x ) = k d (C i ( x ) )
i , k
(17)
i =1
i , k < 0
kd
k =
K eq
k d (18)
In the case of instantaneous reactions, like an exchange of proton, the constant kinetics
k d and k d do not exist. An extent of reaction k ( x ) for the instantaneous reactions is
defined. The relation (15) is broken up into two terms according to the instantaneous nature or
not of the chemical reactions considered.
N RI NR
R i (x ) = i,k . k ( x ) + .(rk ( x ) rk ( x ) )
i ,k
k =1 k = N RI +1
(19)
For one of the ionic components, the Nernst-Planck equation is replaced by the static
electroneutrality condition in order to maintain electroneutarlity throughout the mass-transfer
zone (Cadours and Bouallou [12]; Littel et al. [13]).
NC
Z C (x) = 0
i =1
i i (20)
At the gas-liquid interface ( x = G ), for each species i.
dC ( x )
N i . A = A.D iL i (21)
dx x =G
For the non volatile species, the fluxes are equal to zero.
dC ( x )
0 = D iL i
dx x =G (22)
Within the liquid phase, the relation (27) is applied to all the ionic species and amines
for which the saturated vapor tension is neglected, but also for the water which is regarded as
a stagnant solvent not transferring towards the gas phase. For the volatile species, like CO2
and H2S, the condition (26) is applied.
At the gas-liquid interface ( x = G ), the molar fractions of the volatile species of the
gas phase are connected to the concentrations of the liquid phase by the Henrys law.
Ptot Yi ( G ) = H i C i ( G ) (23)
139
At liquid bulk zone ( x = G + L ), the medium is homogeneous.
Concentration C i (x = G + L ) imposed on the differential equations system is calculated by
an assessment on the liquid bulk zone of volume V ZML .

C iE C Si
L E . E + R i (x = G + L ) * V ZML
+ N i ( x = G + L ). A = L S . S
L
C tot C tot (24)

Data input L E and C iE are respectively the molar flux and the species i concentration
entering the bulk zone and are known. Concentration C Si outgoing is imposed as boudary
condition at x = G + L .
In particular cases of absorption, when CO2 and H2S represent only one minority of
the gaseous medium compared to CH4 who does not transfer, it is possible to apply the first
Ficks law for the gas phase ( x [0; G ] ).
D Gi d 2 y i ( x )
0=
C tot dx 2 (25)
The concentration profiles are linear in the gas phase. But, put besides this particular
case, when we consider a system of NC components in the gas phase, the transfer
multiconstituant with drive represents the general case and requires the use of the equations of
Stefan-Maxwell.
dy i N C y i N j y j N i
= (26)
dx j=1 C Gtot D i , j
This equation makes it possible to obtain the concentration profile yi in the gas film
according to NC molar fluxes Ni.
At the gas bulk zone ( x = 0 ), the medium is considered homogeneous and at
equilibrium. The molar fractions of the gas bulk zone are imposed like boundary conditions
(27) to the systems of equations (26).
yi (0, t ) = yibulk (27)
At the gas-liquid interface ( x = G ), the continuity of flux (21) as well as the Henrys
law (23) are applied.
Heat transfer
140
Within liquid film ( x [ G ; G + L ] ), the heat transfer takes place by conduction. The
Fourier analysis is applied with an additional source term holding account of the heat
generated or absorbed by the chemical reactions.
d 2 T L (x)
0 = L + Q raction ( x ) (28)
dx 2
The heat of reactions Qreaction is expressed as follows.
NR
Q raction ( x ) = R k ( x ) R H k (29)
k =1
Molar enthalpy of reaction k R H k is calculated using the chemical equilibrium
constants K eq
k and Van't Hoff relation.
(
d Ln (K eq
k ) )
H
= R k2 (30) At
dT R. T L ( )
the liquid bulk: T L ( G + L ) = T L,bulk (31)
At the gas-liquid interface ( x = G ), the boundary conditions are the continuity of the
temperature and the continuity of the heat fluxes. The boundary condition (33) must take into
account the enthalpic contribution brought by molar fluxes of the species which transfer
through the interface.
T G (x = G ) = T L (x = G ) (32)
dT G dT L
+ N i .(H Gi H iL ) = L
NG
G (33)
dx x =G i =1 dx x =G
The term H iG represents enthalpy of compound i in the gas phase with as reference
pure compound i at 273.15 K. The term H iL represents enthalpy of compound i in the liquid
phase. For example, for the water transfer from the gas phase to the liquid phase, the term
(H L
i H Gi ) will be the liquefaction latent heat and for the transfer of CO2 or H2S, this term
will represent the solubilization latent heat.
Within the gas film ( x [0; G ] ), the heat transfer takes place by conduction and
convection. The Fourier analysis must be supplemented by a term holding account of the
enthalpy transported by molar flux and as reference the pure gas at 273.15 K.
Ncomp

0 = q + H i N i
(34)
i =1
The heat flux q is expressed as follows.
141
q = G T G (35)
G is equivalent conductivity holding account of the conductive and the convective
contribution. This coefficient is given using the Chilton-Colburn analogy which defines
empirical factors jh and jd respectively for the heat and mass transfer.
Nu hG C p G 32
f
jH = =( )( ) =
1
G C pV G 2 (36)
Re Pr 3
Sh k G .R.T G . p BM 2
f
jD = =( )( G G ) 3 =
1
Ptot .V G D 2 (37)
Re Sc 3
At the gas bulk: T G (0, t ) = T G ,bulk (38)
At the gas-liquid interface ( x = G ), the boundary conditions are the continuity of the
temperature and the continuity of the heat fluxes.
142
Numerical resolution
In liquid film, the difficulty of the mass transfer modeling is based on the taking into
account of the slow and instantaneous chemical reactions. In our work, the stability and the
adaptability of the resolution to any kinetic system are of primary importance. An implicit
centered finite-difference iterative method was employed to solve the system of coupled
differential equations in the liquid phase. In the gas phase, the Taylor and Krishna [14]
algorithm was used to solve the Stefan-Maxwell equations and to obtain molar fluxes at the
gas-liquid interface. A Newton-Raphson procedure was used to determine the volatile species
concentration at the gas-liquid interface. In the gas film, the equations of Stefan-Maxwell will
be solved by an iterative diagram (Taylor and Krishna [14]) in order to obtain molar fluxes
which are used as boundary condition at the interface to solve the Nernst-Planck equations in
the liquid phase. The calculations are repeated until the convergence criterion
y int, ini
y ini , calc
R int
i = i
int,ini
i
10 8
yi (39)
is satisfied. In this way, a consistent algorithm is obtained for the set of equations.
Modeling of absorption column
The most important stage in the validation of the new model is the comparison with
the reference simulation for column profiles. Our new model is introduced within the
industrial simulator. Then, during the columns simulation, it is possible to choose between
our new model and the reference model in order to simulate the columns. This choice enables
us to compare interface flux profiles as well as enhancement factors on the whole of the
stages of the columns. Simulations of absorption columns were carried out on two industrial
cases. The first case is CO2 and H2S absorption by an aqueous solution of DEA. The second
studied case is a CO2 absorption by an aqueous solution of DEA. For the two studied cases,
the molar flux profiles transferred between the gas and liquid phases for each model are
compared as well as the enhancement factor profiles.
143
CO2 and H2S absorption by DEA solution
The studied industrial case comprises a particular absorption column.The absorption

column presented on Figure 3 has two inlets for amines solutions. The first inlet is carried out
at stage 1 and the second at stage 11. The column overheading is fed with an aqueous solution
of entirely regenerated DEA, while stage 11 comprises an inlet of an aqueous solution of
semi-regenerated DEA.
The interest of this technique is to save energy during the regeneration phase. In
effect, a share of the DEA aqueous solution is regenerated only partly. This solution is
introduced within the absorption column at stage 11 in order to absorb CO2 and H2S in the gas
phase. The introduction of an aqueous solution of DEA completely regenerated to the level of
stage 1 makes it possible to finalize the CO2 absorption and H2S in order to obtain the
specifications desired on outlet gas.
Treated gas
Lean DEA
Stage with central outfall
Lean DEA inlet
Stage with side outfall

Semi-regenerated DEA inlet
Gas inlet
Rich DEA
Figure 3: Diagram of the absorption column with studied stages

The kinetic mechanism considered in the liquid phase is specified by the equations (I),
(II), (III), (IV), (V), (VI) and (VII). The reaction (I) is based on the zwitterion mechanism.
144
CO 2 + 2 DEA DEACOO + DEAH + (I)
Constant kinetics correlated by Rinker et al. [14] is used. The other chemical reactions
considered are:
CO 2 + OH
HCO 3 (II)

HCO 3 CO 32 + H + (III)
DEAH + DEA + H + (IV)

H 2 O H + + OH (V)
H 2 S HS + H + (VI)
HS S 2 + H + (VII)
The thermophysical parameters used and in particular solubility are detailed in [15], the other
parameters used for simulations are given in Table 1. For stage 3, Figures 4, 5, 6 present the
concentration profiles of all species present in the liquid phase. The flux transferred between
the gas phase and the liquid phase is weak.
6
CO 2 = 5.07 10
N int mol.m 2 .s 1 ; N int
H 2S = 1.49 10
5
mol.m 2 .s 1
For stage 21, Figures 7, 8, 9 present the concentration profiles of the same species present in
the liquid phase. The flux transferred between the gas phase and the liquid phase is important.
3
CO 2 = 3.53 10
N int mol.m 2 .s 1 ; N int
H 2S = 1.30 10
2
mol.m 2 .s 1
Comparison of these profiles is interesting because stage 3, located in top of the column, is
fed by an aqueous solution of DEA slightly acid gas loaded and the gas phase then contains a
weak molar fraction of CO2 and H2S, while stage 21 is fed by an aqueous solution of DEA
strongly acid gas loaded; then the gas phase contains a proportion much more important of
CO2 and of H2S.
145
Table 1: Parameters used
Gas to be Lean Semi-regenerated
treated amine amine
5
Pression (10 Pa) 67.5 66.9 61.0
Temperature (K) 325.15 324.64 324.50
Composition y CO2 = 0.04 x CO2 = 2.9 x CO2 = 2 103
y H 2S = 0.18 x H 2S = 6 1 x H 2S = 67 104
y CH 4 = 0.72 x H 2O = 0.89 x H 2O = 0.9103
x DEA = 0.10 x DEA = 0.081
Fluxrate (mol.s-1) 3778 6501 8228
The comparison between Figure 4 and Figure 8 shows a H2S profile in the liquid phase
completely different. For stage 3, the liquid is slightly loaded, the concentration in hydronium
ions is important at the interface; it is much higher than the H2S concentration.
3
H 2S = 3.83 10
C int mol.m 3 C int
OH
= 1.42 mol.m 3
;
The H2S absorption is then entirely under kinetic control, which explains the shape of
the very similar curve to that of CO2 absorption. For stage 21, the aqueous DEA solution is
strongly acid gases loaded and H2S flux at the interface is important. The concentration in
hydronium ions at the interface is of the same order of magnitude as the H2S concentration.
1
H 2S = 2.63 10
C int mol.m 3 C int
OH
= 2.16 10 1 mol.m 3
;
The H2S absorption depends then only on the OH- ions diffusion within liquid film,
which explains the linear profile of the H2S concentration within liquid film.
For stage 3, CO2 absorbed is transformed by the chemical reactions into carbonates and
hydrogen carbonates ions in weak concentrations. On the other hand, for the stage 21, CO2 is
transformed mainly into carbamates ions, while the concentrations in carbonates and
hydrogen carbonates ions are close.
On Figures 10, 11, 12, CO2 and H2S molar fluxes as well as the CO2 and H2S
enhancement factors are compared for the reference model within the simulator and our
model. The visible change of slope in all the curves between stage 10 and stage 11
corresponds to the introduction within the column of aqueous semi-regenerated DEA
solution. Figure 10 presents the comparison of molar CO2 and H2S fluxes between our new
model and the reference model. A general shape of the molar flux curves according to N of
stage is identical between the two models. Nevertheless, differences between CO2 fluxes are
146
on average respectively 15 % for the first 10 stages and 22 % for the following stages. For
H2S, the average deviation between the curves is 20 %. Figure 13 presents the comparison of
molar flux between our new model and the reference model. The average deviation between
the CO2 flux calculated with our new model and the CO2 flux calculated with the reference
model is 12 %. For H2S the variation is 20% on average. For H2S, this variation remains the
same one. Figure 14 presents the CO2 and H2S fluxes at the gas-liquid interface.
The comparison of the CO2 and H2S enhancement factors between our new model and
the reference model is presented in Figures 11 and 12. The first observation is the important
difference between the enhancement factors calculated with our new model and those
calculated with the reference model. For CO2 (Figure 11 and Figure 14), for the first 10
stages, the average deviation is 30 % and this variation goes up to 35 % between stages 11 to
27. For H2S (Figure 12 and Figure 14), the average deviation between the enhancement factor
curves of each model is 42 %. The enhancement factor is calculated using interfacial flux but
also using the gradient of concentration between the interface and the liquid bulk. However
this gradient of concentration is different between the two models what explains why the
differences between the enhancement factors of the two models are higher than the difference
between fluxes at the interface. The tendencies of enhancement factor curves can nevertheless
be compared between the two models. With regard to the CO2 enhancement factor, the
profiles of the curves are similar between our new model and the reference model. On the
other hand, for H2S, the profile of the enhancement factor curve for our new model is
different from that of the reference model. Indeed, the enhancement factor with our new
model increases much for the first column stages where the liquid phase is slightly loaded,
which is more realistic than the almost stage profile of the enhancement factor curve with the
reference model for the first 10 stages. The average deviation of 20 % noted between the new
model and the reference model has several causes. We already noted that with the same
transfer model, an average deviation of 10 % appears when a different kinetic model for the
reaction between CO2 and the DEA is considered. On the other hand, it appears clearly that
the successive iterations on the simulation of the column modify the fluxes values only of 2 %
at most. The remainder of the difference is then directly ascribable with the transfer model.
147
0.14 3.830
0.13 3.830
CO 2 concentration
0.12 3.829
H 2 S concentration
0.11 3.829
-3
H2S concentration 10 mol.m
CO2 concentration 10-3 mol.m-3
0.10 3.828
-3
0.09 3.828
0.08 3.827
0.07 3.827
0.06 3.826
0.05 3.826
0.04 3.825
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Dimensionless liquid film
Figure 4: CO2 and H2S concentration profiles in liquid film for stage 3
1.4210 5.2484
5.2480
1.4209
-
OH concentration
H+ concentration 5.2476
H+ concentration mol.m-3
OH- concentration mol.m-3
1.4208
5.2472
1.4207
5.2468
1.4206
5.2464
1.4205 5.2460
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Figure 5: H+ and OH-concentration profiles in liquid film for stage 3
148
37.816 3884.54
8.4530
37.812
+
DEAH+ concentration
8.4528
DEA concentration
DEACOO- concentration
-3
-
DEA mol.m
37.808 3884.53
-3
-3
mol.m
mol.m
-
8.4526
-
DEACOO
37.804
+
DEAH
37.800 3884.52
8.4524
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Dimensionless liquid film 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0.7092 2.0178
23.030 8.4019
-3
10 mol.m
8.4018
-3
0.7088 mol.m 2.0172
23.025
HCO3- concentration HS- concentration
-2
CO32- concentration S2- concentration
3
--
8.4017
CO3
-
S
-3
HS mol.m
-3
HCO3 mol.m
0.7084 2.0166 23.020
-
8.4016
-
-
0.7080 2.0160 23.015 8.4015

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Dimensionless liquid film Dimensionless liquid film
Figure 6: concentration profiles in liquid film (stage 3)
149
3290 300 54
-
DEACOO concentration
53 -
3270
-3
-3
DEACOO mol.m
-3
DEA mol.m
DEAH mol.m
280
-
+
3250 52
DEA concentration
+
DEAH concentration
3230 260 51
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
9
225 0.168
-
-
HS concentration
HCO3 concentration 2-
S concentration
2-
- CO3 concentration
-3
mol.m
0.167
215
-3
2-
HS mol.m
HCO3 , CO3
-3
6
S mol.m
0.166
-
--
205
0.165
3 195 0.164
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Figure 7: concentration profiles in liquid film (stage 21)
150
200 270
180
CO2 concentration
260
160 H2S concentration
140 250
CO2 concentration 10-3 mol.m-3
120
H2S concentration 10-3 mol.m-3

240
100
230
80
60 220
40
210
20
0 200
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Figure 8: CO2 and H 2S concentration profiles in liquid film (stage 21)

0.26 5.5
-
OH concentration
H+ concentration
5.3
0.24
-3
OH- concentration mol.m-3
+
H concentration 10 mol.m
5.1
-7
0.22
4.9
+
0.20 4.7
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
- +
Figure 9: OH and H concentration profiles in the liquid film (stage 21)
151
1.00E+00
CO 2 (our model)
1.00E-01 CO 2 (reference model)
H 2 S (our model)
H 2 S (reference model)
1.00E-02
1.00E-03
-1
Flux mol.m .s
-2
1.00E-04
1.00E-05
1.00E-06
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Stage number
Figure 10: Comparison of molar fluxes (gas and liquid phases) for each stage
141
121
101
Enhancement factor
CO2 enhancement factor (our model)

CO2 enhancement factor (reference
81
61
41
21
1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Stage number
Figure 11: Comparison of the CO2 enhancement factor for each stage
152
100000
10000
Enhancement factor
1000
100
H2S enhancement factor (our model)

H2S enhancement factor (reference model)
10
1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Stage number
Figure 12: Comparison of the H2S enhancement factor for each stage
1.00E+00
1.00E-01
CO 2 (our model)
H 2S (our model)
CO 2 (reference model)
1.00E-02
H 2S ( reference model)
-1
Flux mol.m .s
-2
1.00E-03
1.00E-04
1.00E-05
1.00E-06
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Stage number
Figure 13: Comparison of molar fluxes between our new model and the reference model
153
140 100000
120
Enh ancem ent factor CO 2 (our m odel)
Enh ancem ent factor CO 2 (reference m odel)
100 Enh ancem ent factor H 2S (our m odel)

CO2 enhancement factor
Enh ancem ent factor H 2S (reference m odel) 10000
H2Senhancement factor
80
60
1000
40
20
0 100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Stage num ber
Figure 14: Comparison of the enhancement factor.
CO2 Absorption by DEA solution

The absorption column considered now is a column with only one food to absorb CO2
by DEA aqueous solution. The kinetic mechanism considered is specified by the equations
(I), (II), (III), (IV) and (V).Figure 15 presents a comparison of the CO2 flux for our new
model and the reference model. Profiles of molar CO2 flux have similar tendencies. The
difference between the curves remains important (27 % with the stage 25 and 31% with stage
31). The whole of the parameters used for this simulation is given in Table 2.
Table 2: Parameters used

Gas to be treated Lean amine
5
Pressure (10 Pa) 54.4 54.1
Temperature (K) 299.15 329.22
Composition
x CO2 = 0.0011 yCO2 = 0.0167
x H 2S =0 y H 2S = 0.0
x H 2O = 0.895 y CH 4 = 0.981
x DEA = 0.102
Fluxrate (mol.s-1) 7264 3527
154
0.08
0.07
CO2 ( our model)
0.06
CO2 (reference model)
0.05
0.04
Flux mol.m-2.s-1
0.03
0.02
0.01
0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
-0.01
Stage number
Figure 15: Comparison of CO2 flux

Conclusion
This study shows that the profiles obtained by our model are in accordance with the awaited
results. At the columns overhead, the liquid is slightly acid gas loaded and CO2 and H2S absorption is
entirely under kinetic control. In columns bottom, the liquid is strongly acid gases loaded, the H2S
absorption is then under diffusional control. The comparison of the CO2 and H2S enhancement factor
profiles given by the two models show very consequent variations (respectively 30 and 40 %),
whereas the difference between absorption rate is only 20 % on average. This important variation can
be allotted to the difference between the gradients of concentration of the liquid phase for the two
models. A variation of 12 % for the CO2 absorption rate and of 20 % for the H2S absorption rate
between our model and the reference model confirms the good agreement between the two models.
Our computer code also represents CO2 absorption in an aqueous solution of DEA with an average
deviation of 20 % with respect to the reference model. This variation is explained by the difference
between the models in the taking into account of heat and mass transfer phenomena. Experiments on
a pilot column are in hand. These measurements of the CO2 and H2S absorption by the DEA will make
it possible to compare the results provided by our model but also those provided by the reference
model with the experimental results and thus to determine which models is able to represent the
absorption columns well. The industrial interest of our model is the taking into account in a more
realistic and more complete way of the heat and mass transfer phenomena compared to the reference
model. Its principal advantages are to be able to model any whole of chemical reactions. The taking
account of new chemical compounds, like the mercaptans or new reactions is easy.
155
Notations
A ...................Interfacial area (m2 / m3)
Ci
..................Molar concentration, mol.m-3
C v ..................Heat capacity of component i, J.mol-1.K-1
di
...................Driving force, m-1
D Si or D i ......Diffusivity, m2.s-1
D i , j ................Stefan-Maxwell diffusivity, m2.s-1
Ea i
................Enhancement factor
E ij ..................Mass effectiveness
E thj
.................Thermal effectiveness
F ....................Faradays constant 96500 C/mol
Ha
.................Hatta number
H Si or H i ......Henrys law constant in the concentration scale, Pa.m3.mol-1
H Gj
.................Molar enthalpy, J.mol-1
Ji
...................Molar diffusion flux of species i, mol.m-2.s-1
K eq
k ................Equilibrium constant of reaction k
k G .................Gas side mass transfer coefficient, mol-1.m-2.Pa-1.s-1
kL
..................Liquid side mass transfer coefficient, m.s-1
k.....................Rate coefficient
L ....................Liquid molar flux, mol.s-1
Mi
.................Molar mass, g.mol-1
N C .................Number of diffusing species
N G .................Number of volatiles species
NR
.................Total number of reactions
N tot ................Molar flux, mol.m-2.s-1
156
Nu ..................Nusselt number
P ...................Pressure, Pa
q.....................Energy flux, en W.m-2

Q raction ..........Heat of reactions, W.m-3
Re
.................Reynolds number
Ri
..................Rate of production of i due to chemical reaction en mol.s-1
R ij
..................Residual (Newton-Raphson method)
Sc...................Schmidt number
Sh...................Sherwood number
t ....................Time, s
T ...................Temperature, K
ui
...................Velocity, m.s-1
VS...................Molar volume, m3.mol-1
Vi
..................Vapor fluxrate, mol.s-1
x ....................Coordinate, m
x ij ..................Liquid mole fraction, mol/mol
y ij ..................Gas mole fraction, mol/mol
Zi
..................Ionic charge of component i
157
Greek Letters:
R Hi
.............Molar Enthalpy of reaction i, J.mol-1
....................Difference
.....................Film thickness, m
i , j .................Kronecker delta ( i , j =1 for i = j, i , j =0 for i#j)
C q ( x)
RT
Z q D qL x
...................Electrostatic potential, V.m-1 q
F
Z q
2
q D qL C q ( x)
....................Thermal conductivity, W.m-1.K-1

....................Viscosity, 10-3 Pa.s (cP)
i ,k .................Stoichiometric coefficient
.....................Density, g.cm-3
Subscripts:
Am.................Amine
calc.calculed
ini.. initial
j..Stage j parameter
tot...................Total
Superscripts:
int, .................Interface parameter
bulk................Bulk phase
L ....................liquid
G ....................gas
*.....................Equilibrium value
158
Literature cited
1. Perez-Cisneros, E.; Schenk, M.; Gani, R.; Pilavachi, P.A. Aspects of simulation, design and
analysis of reactive distillation operations. Comp. Chem. Eng. 1996, 20, S267.
2. Scenna, N. J.; Ruiz, C.A.; Benz, S.J. Dynamic simulation of startup procedures of reactive
distillation columns. Comp. Chem. Eng. 1998, 22, S719.
3. Sneesby M. G.; Tad, M. O.; Smith, T. N. Steady-state transitions in the reactive
distillation of MTBE. Comp. Chem. Eng. 1998, 22, 879.
4. Taylor, R.; Krishna, R. Modelling reactive distillation. Chem. Eng. Sci. 2000, 55, 5183.
5. Baur, R.; Taylor, R.; Krishna, R. Dynamic behaviour of reactive distillation columns
described by a non equilibrium stage model. Chem. Eng. Sci. 2001, 56, 2085.
6. Isla, M. A.; Irazoqui, H. A. Modeling, analysis, and simulation of a methyl tert-butyl ether
reactive distillation column. Ind. Eng. Chem. Res. 1996, 35, 2396.
7. Lee, J. H.; Dudukovic, M. P. A comparison of the equilibrium and non equilibrium models
for a multicomponent reactive distillation column. Comp. Chem. Eng. 1998, 23, 159.
8. Higler, A. P.; Krishna R.; Taylor R.; A non equilibrium cell for multicomponent reactive
separation processes. Am. Inst. Chem. Eng. J. 1999, 45, 2357.
9. Higler, A. P., Krishna, R.; Taylor, R. A non equilibrium cell model for packed distillation
columns. The influence of distribution. Ind. Eng. Chem. Res. 1999, 38, 3988.
10. Higler, A. P.; Krishna, R.; Taylor R. The influence of mass transfer and liquid mixing on
the performance or reactive distillation tray column. Chem. Eng. Sci. 1999, 54, 2873.
11. Chang, S.; Rochelle, G. T. Mass transfer enhanced by equilibrium reaction. Ind. Eng.
Chem. Fund. 1982, 21, 379.
12. Cadours, R.; Bouallou, C. Rigorous simulation of gas absorption into aqueous solutions.
Ind. Eng. Chem. Res. 1998, 37, 1063.
13. Little, R. J.; Filmer, B.; Versteeg, G. F.; van Swaaij, W. P. M. Modelling of simultaneous absorption
of H2S and CO2 in alkanolamine solutions: the influence of parallel and consecutive reversible
reactions and the coupled diffusion of ionic species. Chem. Eng. Sc. 1991, 46, 9, 2303.
14. Rinker, E. B.; Ashour, S. S.; Sandall, O. C. Kinetics and modeling of carbon dioxide absorption into
aqueous solutions of diethanolamine. Ind. Eng. Chem. Res. 1996, 35, 1107.
15. Vinel, D-J.; Bouallou, C. Proprits physico-chimiques des solutions aqueuses dalcanolamines
utilises dans le traitement du gaz. Scientific Study & Research 2004, V(1-2), 11.
159
Chapter 9. Kinetics of Carbonyl Sulphide (COS) Absorption with Aqueous Solutions of
DiEthanolAmine and MethyDiEthanolAmine
F. Amararene, C. Bouallou
Introduction
Industries and urbanization are widely contribute to air pollution. The sulphur compounds
produced by refineries and other operations of the energy industrial sector have serious
consequences on the environment. Because of its economic and ecological advantages,
natural gas becomes more attractive for many countries. However, after extraction, natural
gas can neither be transported nor used commercially. First, it must be processed to remove
acid gases (CO2, H2S) and impurities such as COS and RSH (mercaptans). Sulphur
compounds are harmful toxic and corrosive gases, therefore, the industry has developed strict
specifications for their transportation, limiting the total content of sulphur to 35-50 ppm.
Separation of the various sulphur compounds is often achieved by absorption by a chemical
solvent such as alkanolamines. A wide variety of alkanolamines such as monoethanolamine
(MEA), diethanolamine (DEA), N-methyldiethanolamine (MDEA), DiGlycolAmine (DGA)
have been used commercially.
In spite of abundant literature, only few works 1, 3, 4, 5, 6, 7 address the absorption kinetics for
COS/DEA + H2O or COS/MDEA + H2O systems. The data that exist mainly concerns of the
kinetics in gas-unloaded solutions.
The first study devoted to COS by Sharma1 was carried out with several primary and
secondary alkanolamines, but only at one temperature (T = 298 K) and only at one
concentration (C = 1000 mol.m-3). It suggested that COS reaction mechanisms with primary
and secondary alkanolamines, were similar to that with CO2. Singh and Bullin2 studied COS
absorption, using DGA aqueous solutions in a reactor which simulate a stage of a column
absorption. The temperature varied from 307 K to 322 K, the DGA concentration from 3220
to 5630 mol.m-3, and the COS partial pressure from 37 to 394 kPa. From kinetic experiments
using wetted sphere apparatus Al-Ghawas et al.3 provide data for COS Henrys constant and
diffusion coefficient in MDEA aqueous solutions in a temperature range from 298 to 313 K,
and with MDEA concentrations from 1259 to 2599 mol.m-3. The kinetic results are well
represented by a base catalysed COS hydratation. Additional data were obtained by Little et
al.4 who used a stirred cell reactor, with primary and secondary alkanolamines, over a
temperature range of 283 to 333 K. These authors used COS-N2O analogy to determine the
COS physico-chemicals parameters. They found that the reaction of COS with primary and
160
secondary alkanolamines is well represented by the zwitterion mechanism, the zwitterion
deprotonation being the limiting step. Other studies (Little et al.5, 6) were dedicated to COS
absorption in aqueous solutions of tertiary alkanolamines and proposed absorption models.
The temperature was fixed at 303 K for all experiments except those with MDEA aqueous
solutions (293 K323 K). The alkanolamine concentration was varied from 153 mol.m-3 to
1011 mol.m-3. Recently, the study by Hinderaker and Sandall7 provided kinetics data of COS
absorption by 5 wt % to 25 wt % DEA aqueous solutions, and at temperatures between 298
and 348 K. These authors, carried out measurements of solubility and diffusivity of COS and
N2O in aqueous solutions of polyethylene glycol at 298 K and confirmed the validity of the
COS-N2O analogy. They found that the limiting step in the reaction of COS absorption in
DEA aqueous solutions, is the deprotonation of the zwitterion resulting in overall third order
kinetics.
The apparatus (Figure 1) is a thermostated Lewis type reactor with a constant gas-liquid
interface (15.34 0.05) 10-4 m2. The temperature is controlled by circulating a thermostatic
fluid through the glass double jacket. The reactor is closed at both ends by two metallic
flanges. The liquid and gas phases are agitated respectively by a 4.25 10-2 m diameter six-
bladed Rushton turbine and 4.10-2m diameter propeller, and they are driven magnetically by
variable speed motor. A DRUCK pressure (0 to 250 kPa) is mounted on the upper flange, and
is calibrated at a temperature higher than the experiment temperature to avoid liquid
condensation in its measuring chamber. A tube is mounted through the upper flange and
allows either to degas the cell or to connect it to a tank of COS gas. The lower flange is
equipped with a temperature probe, and a non rotating stem valve. The standard reactor
configuration was modified to expedite the loading of the solutions. In a previous study (Pani
et al.8) which used the same apparatus, the solution was introduced by means of a hypodermic
needle, and the loading time was significant. Also, four vertical baffles are placed inside the
cell in order to avoid the formation of a vortex. The pressure transducer is calibrated within
42 Pa against a pressure calibration device. The temperature in the reactor is known within
0.02 K, it is calibrated against a 25 platinum probe. A microcomputer is used to record both
pressure and temperature signals.
Water and alkanolamine are degassed independently and aqueous solutions are prepared
under vacuum. The masses of water and alkanolamine are known by differential weighing.
161
The flask containing the solution is connected to the reactor to allow the solution to transfer
by gravity under vacuum. Accurate weighing of the flask before and after transfer yield the
mass of solution actually introduced in the cell, and the liquid phase volume was calculated
using the density.
At a given temperature and vapor pressure PI, pure COS is introduced during a very short
time in the upper part of the cell, the volume of which is noted as Vg. Then stirring is started,
and the pressure drop resulting from absorption is recorded. The estimated maximum
experimental error in the COS absorption rate is 8 %.
Chemicals
Twice-distilled water and reagent-grade MDEA are used. DEA is from Aldrich, with a
certified minimum mass purity of 99 %. MDEA is from Fluka, with a certified minimum mass
purity of 98 %. Carbonyl sulphide is from L'Air Liquide, with a certified volume purity of
99.997 %.
Results and discussion

Based on mass balance of COS for gas phase, the influence on absorption kinetics of all
chemical reactions between dissolved COS and reactants in solution is expressed by an
enhancement factor E over physical absorption :
Vg dPCOS
k L EC COS,int a = (1)
RT dt
The gas phase is assumed ideal and COS is completely consumed by the chemical reaction.
At the interface, vapor-liquid equilibrium is assumed. The partial pressure PCOS is related to
the concentration of unreacted dissolved COS by Henry's law :
PCOS int
C COS,int = (2)
H COS
PCOS is obtained from the measured pressures :
PCOS = PT PI (3)
162
The concentration of COS resulting from absorption does not change much the composition
of the solution, so that kL, HCOS and E remain constant with time, and integration of Eq.1
yields :
P PI
Ln T = (t t 0 ) (4)

PT ,0 PI
where :
k L EaRT
= (5)
VgH COS
For each experiment, the enhancement factor E is obtained from using Eq.5. The density
and the viscosity of the DEA and MDEA aqueous solutions are estimated using Hsu and Li9,10
correlations. The DEA diffusion coefficient is estimated from the correlation supplied by
Hikita et al.11. The correlation given by Pani et al.8 is used for estimates the MDEA diffusion
coefficient. The COS-N2O analogy provides estimates of COS diffusion coefficient and
Henry's law constant in alkanolamines aqueous solutions. The COS diffusion coefficient in
water is obtained from Al-Ghawas et al.3 correlation. The N2O diffusion coefficient in water
and in the aqueous solutions is found using Versteeg and van Swaaij12 correlations. Henry'
law constant of COS in water is calculated from Little et al.13 correlation. The model of Wang
et al.14 is used for the calculation of N2O Henry's law constant in alkanolamines solutions
with the parameters of binary interactions given by Tsai et al.15. The mass-transfer coefficient
kL is calculated using the following correlation established for our apparatus from N2O
absorption experiments by MDEA aqueous solutions :
Sh=0.352 Re0.618Sc0.434 (6)
Where the dimensionless Sherwood (Sh), Reynolds (Re) and Schmidt (Sc) numbers are
expressed by :
163
k L d cell dNd ag2
Sh = , Re = , Sc =
D N 2O D N 2O
Kinetics for the COS/DEA+water system

According to literature the reaction of COS in DEA aqueous solutions is generally
represented by the zwitterion mechanism :
COS+(C2H4OH)2NH (C2H4OH)2NH+COS-
k 2 , k 2
(C2H4OH)2NH+COS-+b (C2H4OH)2NCOS-+bH+
k k
b , H 2 O ; b , DEA
b is for any base (alkanolamine, H2O).
The zwitterion deprotonation is often considered as being the limiting step, implying
the participation of a second molecule of DEA. Therefore, it was that assumed the reaction of
COS with DEA aqueous solutions irreversible and of first order with respect to COS and
second order with respect to DEA. In these conditions, the rate of reaction between COS and
DEA is given by:
rCOS = k COS DEA C COS C 2DEA (7)
and the enhancement factor is equal to Hatta number:
k COSDEA C 2DEA D COS

E Ha = (8)
k 2L
All the reactions were studied with a large excess of alkanolamine over COS and gave good
pseudo first order according to:
rCOS = k obs C COS (9)
The determination of the enhancement factor allows the calculation of the rate constants :
164
E 2 k 2L
k obs = (10)
D COS
The rate constants kCOS-DEA is calculated by the relation :
k obs
k COS DEA = (11)
C 2DEA
Only the results for DEA and MDEA concentrations ranging from 5 to 25 wt %, are used
for the calculation of the rate constants that ensure the validity of the COS-N2O analogy.
A comparison between our experimental rate constants, and the data available in the literature shows
that our results (Figure 2) are consistent with the data of Hinderaker and Sandall7, Little et al. 4 at low
DEA concentrations. The difference between our kinetic results and those of Hinderaker and Sandall7
becomes more significant from a concentration of 1400 mol.m-3 (T = 313 K), beyond they reported
values are higher. For a temperature of 333 K our results are slightly lower compared with the data of
Little et al.4 whereas those of Hinderaker and Sandall7 are highest. The origin of these deviations can
be very probably attributed to an underestimation of the COS solubility and an overestimation of the
diffusivity values in rather concentrated amine solutions. In addition, the COS-N2O analogy is only
valid for relatively diluted aqueous solutions. The results of Hinderaker and Sandall7 are somewhat
higher than those found in the literature. On the one hand, the values for the COS solubility used by
these authors are on average 6% smaller than those used in this study. On the other hand, the
average deviation between the various diffusivity values used in the comparative studies here can be
estimated at 13%. The combined influence of these two parameters on the determination of (Eq. 5)
1 D cos H cos
thus results in an uncertainty of 12.5 % ( = + ). Thus while considering the
2 D cos H cos
most unfavorable case at 313 K, the relative deviation between our result and that of Hinderaker and
Sandall7 in term of the rate constants for COS absorption in aqueous DEA which was 35.0 % become
14.5 %.
The regression of the experimental rate constants (Table 1) obtained by the Eq.10 according
to 1/T for the three studied concentrations was represented in Figure 3. The rate constants
depend on the concentration of DEA. This observation does not appear in the literature that
often presents kinetic laws for the complete concentration range. Consequently, we developed
Eq.12 that allows taking into account the influence at the same time of the temperature and
the concentration :
165
5785
k COS DEA = (2121 0.4 C DEA ) exp (12)
T
This equation was developed by extrapolating the physico-chemicals and thermodynamic

parameters. We have observed (Table 2) that the activation energy obtained in this study is in
agreement with that determined by Hinderaker et Sandall7.
Assuming the participation of water as a second base in the zwitterion deprotonation, the
rate constant becomes:
C DEA
k obs ,cal = (13)
1 1
+
k 2 k H 2O C H 2O + k DEA C DEA
The rate constants kH2O and kDEA, being defined by:
k 2 k b ,H 2 O
k H 2O = (14)
k 2
k 2 k b ,DEA
k DEA = (15)
k 2
kH2O, kDEA represent the zwitterion deprotonation rate constants (m6.mol-2.s-1) which were
determined by adjusting the experimental constants (kobs) with those given by Eq.13.
This procedure gave the following equations for zwitterion deprotonation rate constants:
5772
k DEA = 9.63 10 2 exp (16)
T
6333
k H 2O = 4.55 10 2 exp (17)
T
The term 1/k2 can be neglected (values of k2 were in the order of 108 m3.mol1.s-1). The rate
constants kDEA obtained here agree with activation energy values of Little et al.4 , whereas the
rate constants kH2O obtained by these same authors from only two points is lower than ours.
166
The restricted number of points retained in the case of Little et al.4 explains this difference.
Moreover the influence of DEA in the deprotonation of the zwitterions was more important
compared with the influence of water (Figure 4).
Kinetics for the COS/MDEA+water system
Most of the previous studies on COS absorption by aqueous solutions of tertiary

alkanolamines consider base catalysis for the reaction of COS with water :
COS + C5H13NO2 + H2O C5H13O2NH+ + HCO2S-
The order with regard to MDEA is assumed 1. For COS/MDEA+water system, conditions
of the pseudo first order are not satisfied for all experiments, notably those realized at lower
concentrations and lower temperatures. We have considered therefore a general form of the
enhancement factor :
Ha
E= (18)
thHa
The rate constants kobs is calculated from Eq.10. Summary of the results of COS absorption
in aqueous MDEA solutions are presented in Table 3. The plot of Ln (kCOS-MDEA) against (1/T)
for the three studied concentrations (0.05, 0.15 and 0.25) is presented in Figure 3.
We have not been able to determine in this case a general kinetic law taking into account
the influence of MDEA concentration and temperature. We have proposed the following
kinetic law for each of studied concentrations :
B
k COS MDEA = ACOS MDEA exp COS MDEA (19)
T
ACOS-MDEA and BCOS-MDEA were given in Table 4.

We have compared our experimental results with those of Al-Ghawas et al.3 at 313 K. As
shown in Figure 5, the rate constants given by Al-Ghawas et al.3 are higher than our
experimental data. The study carried out by Little et al.5 also showed that the data of Al-
167
Ghawas et al.3 are high. Theses discrepancies are explained by the presence of contaminant
(primary and secondary alkanolamines) in MDEA. Our results are in good agreement in terms
of activation energy with Little et al.5.
By considering the same operating conditions we have compared COS rate absorption in
aqueous solutions of DEA and aqueous solutions of MDEA. For a given concentration, COS
rate absorption in DEA aqueous solutions is higher compared with that in MDEA aqueous
solutions. This is explained by the high reactivity of DEA compared with MDEA.
Conclusion
Kinetics of COS absorption by DEA and MDEA aqueous solutions have been studied in the
extended range of temperatures from 296 K to 353 K, and of DEA concentrations from 5 to
40 wt %, and of MDEA from 5 to 50 wt%.
In the case of COS/DEA+water system, kinetic results obtained were consistent with the
literature data for temperatures between 313 and 333 K. By assuming a partial order of two
with regard to DEA, we have determined a general kinetic law taking into account the
influence of DEA concentration and the temperature.
The results of COS absorption in DEA aqueous solutions show that the limiting step in the
reaction is the deprotonation of the zwitterion. Rate constants of the zwitterion deprotonation
kDEA and kH2O were compared in this study, and we found that the influence of water in the
deprotonation of the zwitterions is lower than that of the DEA.
The limited data available in the open literature regardless the COS absorption in MDEA
aqueous solutions enabled us to confirm that the rate constants given by Al-Ghawas et al.3 are
over estimated. Our kinetic results are consistent in term of activation energy with those of
Little et al.5, but we observed a significant difference in the values of the absolute rate
constants. A rigorous transfer models with chemical reactions will certainly make it possible
to explain these discrepancies which can have as origin the presence of secondary reactions,
impurities in solvent, the validity of the considered mechanism.
Notations:
a interface area (m2)
C concentration (mol.m-3)
D diffusion coefficient (m2.s-1)
168
dag stirrer diameter (m)
dcell internal diameter of the Lewis cell (m)
E enhancement factor (-)
H molar scale Henrys law constant (Pa.m3.mol-1)
Ha Hatta number (-)
k reaction rate constants
kL liquid-side mass transfer coefficient of unreacted COS (m.s-1)
P pressure (Pa)
PI vapor pressure over the aqueous alkanolamine solutions (Pa)
rcos rate of the reaction between COS and DEA (mol.m-3.s-1)
R gas constant (8.3143 J.K-1.mol-1)
Nd ag2
Re = Reynolds number (-)

T absolute temperature (K)
t time (s)
V Volume (m3)

Sc = Schmidt number (-)
D N 2 O
k L d cell
Sh = Sherwood number (-)
D N 2O
Greek letters
slope (s-1)
viscosity of the aqueous solutions (Pa.s)
density of the aqueous solutions (g.cm-3)
Subscripts
cal calculated
exp experimental
g gas
int interface
obs observed overall rate constants
169
T total
0 initial
Literature cited
(1) Sharma, M. M. Kinetics of Reactions of Carbonyl Sulphide and Carbon Dioxide with
Amines and Catalysis by Bronsted Bases of the Hydrolysis of COS. Transactions of the
Faraday Society. 1965, 61, 681.
(2) Singh, J.; Bullin, A. Determination of Rate Constants for the Reaction between DiGlycolAmine and
Carbonyl Sulphide. Gas separation and Purification. 1988, 2, 131.
(3) Al-Ghawas, H. A.; Ruiz-Ibanez, G.; Sandall, O. C. Absorption of Carbonyl Sulphide in

Aqueous MethylDiEthanolAmine. Chem. Eng. Sci. 1989, 44, 631.
(4) Littel, R. J.; Versteeg, G. F.; van Swaaij, W. P. M. Kinetics of COS with Primary and
Secondary Amines in Aqueous Solutions. AIChE J. 1992, 38, 244.
(5) Little, R. J.; Versteeg, G. F.; van Swaaij, W. P. M. Kinetics Study of COS with Tertiary
Alkanolamine Solutions. 1. Experiments in an Intensely Stirred Batch Reactor. Ind. Eng.
Chem. Res. 1992, 31, 1262.
(6) Little, R. J.; Versteeg, G. F.; van Swaaij, W. P. M. Kinetics Study of COS with Tertiary
Alkanolamine Solutions. 2. Modeling and Experiments in a Stirred Cell Reactor. Ind. Eng.
Chem. Res. 1992, 31, 1269.
(7) Hinderaker, G.; Sandall, O.C. Absorption of Carbonyl Sulphide in Aqueous

Diethanolamine. Chem. Eng. Sci. 2000; 55, 5813.
(8) Pani, F.; Gaunand, A.;Cadours, R.; Bouallou, C.; Richon, D. Kinetics of Absorption of
CO2 in Concentrated Aqueous MethylDiEthanolAmine Solutions in the Range 296K to 343K.
J. Chem. Eng. Data. 1997, 42, 353.
(9) Hsu, C. H.; Li, M.H. Viscosities of aqueous blended amines. J. Chem. Eng. Data. 1997,
42, 714.
(10) Hsu, C. H.; Li, M.H. Densities of aqueous blended amines. J. Chem. Eng. Data. 1997,
42, 502.
170
(11) Hikita, H.; Ishikawa, H.; Uku, K.; Murakami, T. Diffusivity of Mono-Di and
TriEthanolAmines in Aqueous Solutions of Amine. J .Chem. Eng. Data. 1980, 25, 324.
(12) Versteeg, G.F ; van Swaaij, W. P. M. Solubility and Diffusivity of Acid Gases (CO2,
N2O) in Aqueous Alkanolamine Solutions. J. Chem. Eng. Data. 1988, 33, 29.
(13) Littel, R. J.; Versteeg, G. F.; van Swaaij, W. P. M. Solubility and Diffusivity Data for the
Absorption of COS, CO2,and N2O in Amine Solutions. J. Chem. Eng. Data. 1992, 37, 49.
(14) Wang, Y. W.; Xu, S.; Otto, F. D.; Mather, A. E. Solubility of N2O in Alkanolamines and
in Mixed Solvents. Chem. Eng. J. 1992, 48, 31.
(15) Tsai, T. C.; Ko, J. J.; Wang, H. M.; Lin, C. Y.; Li, M. H. Solubility of Nitrous Oxide in
Alkanolamine Aqueous Solutions. J. Chem. Eng. Data. 2000, 45, 341.
(16) Sandall, O. V. Kinetics of Sulphur Species Hydrocarbon Aqueous Amine Systems. Gas
Processors Association, Project 962, RR-182, 2002.
171
Table 1. Conditions of COS absorption kinetics by DEA aqueous solutions.
T CDEA Vg PT,0-PI E kobs
/K /mol m-3 /m3 /Pa /s-1
313.88 474.17 181.83 115761 3.00 4.62
313.82 474.18 183.60 115248 2.86 4.26
323.68 491.16 183.47 111776 4.06 9.20
333.35 498.33 181.10 108922 4.85 14.21
333.45 498.30 182.20 101563 4.85 14.15
352.90 183.54 180.59 74567 7.73 41.21
313.98 1433.06 179.98 113706 8.08 28.30
323.75 1422.54 183.06 103780 10.01 48.69
333.35 1425.50 182.62 97232 12.01 77.45
333.54 1425.36 181.59 98818 12.98 89.27
353.11 1409.64 181.09 77670 19.16 226.11
313.94 2425.56 183.40 101154 13.85 70.65
323.73 2388.52 182.98 108683 16.92 117.57
333.42 2392.72 182.63 99454 20.27 185.99
333.55 2412.32 182.68 99953 20.51 189.84
352.91 2359.66 180.31 59268 32.21 555.67
352.90 2375.63 180.38 66023 31.36 526.58
Table 2. Comparison with activation energies of COS absorption in aqueous DEA from previous
works.
References T CDEA Ea
/K /mol m-3 /kJ mol-1
Hinderaker and Sandall7 293-348 476-2440 52.35
Our results 313-353 474-2400 48.10
Table 3. Conditions of COS absorption kinetics by MDEA aqueous solutions.

T CMDEA Vg PT,0-PI E kobs
/K / mol m-3 / m3 / Pa / s-1
313.86 445.08 183.57 117484 1.25 0.40
313.80 445.09 181.94 112456 1.21 1.17
323.66 471.53 183.57 112206 1.53 0.98
333.21 469.30 181.68 106711 1.89 1.92
352.91 464.00 180.51 82879 3.14 6.73
352.73 464.05 180.36 76586 3.02 6.22
313.98 1432.82 184.07 133011 1.54 0.80
323.60 1426.09 182.75 114238 2.09 2.00
333.57 1418.48 182.01 108944 2.56 3.46
352.94 1401.85 180.40 79453 4.25 11.06
313.80 2408.22 180.40 136760 1.84 1.09
313.90 2408.10 183.86 118046 2.15 1.23
323.60 2395.62 183.35 110440 2.60 2.77
333.46 2381.99 181.76 112683 3.21 4.63
333.54 2381.87 182.34 105536 3.36 5.09
352.68 2323.98 179.44 83390 6.47 22.55
353.00 2352.16 181.70 81930 6.40 22.46
172
Table 4. Kinetic laws of COS reaction according to MDEA concentration.
T CMDEA ACOS-MDEA BCOS-MDEA
/K / mol m-3 / s-1 /K
313.83 445.08
323.66 471.53 4.17 107 7688
333.21 469.30
352.82 464.03
313.98 1432.82
323.60 1426.09 9.25 106 7358
333.57 1418.48
352.94 1401.85
313.85 2408.16
323.64 2395.62 1.77 108 8357
333.54 2381.87
352.84 2338.07
Pressuretransducer
Heating
rode Vaccum pump
Data
acquisition Degassed
solution
Vapor phase
Gas cylinder
Liquid phase
Micro-computer
Temperature probe Magnetic stirrer Tube of load
Figure 1: Flow diagram of the apparatus.
173
350
Our results, T = 313 K

300
Hinderaker and Sandall 7 T = 313 K
Our results, T = 333 K
250
Hinderaker and Sandall 7 , T = 333 K
kobs (s-1)
Little et al.4 T = 333 K

200
150
100
50
0
0 400 800 1200 1600 2000 2400 2800
CDEA (mol.m-3)
Figure 2. Rate constants of COS absorption in aqueous DEA. Comparison with previous data.
-1
10
kCOS-DEA (m6mol-2s-2), kCOS-MDEA (m3mol-1s-1)
-2
10
-3
10
-4
10
W DEA=0.05
W DEA=0.15
-5 W DEA=0.25
10
W MDEA=0.05
W MDEA=0.15
W MDEA=0.25
-6
10
-3 -3 -3 -3 -3
2.8x10 2.9x10 3.0x10 3.1x10 3.2x10
1/T (K-1)
Figure 3. Arrhenius law for the COS absorption in aqueous DEA and aqueous MDEA
solutions.
174
-3
10
-4
kDEA, kH2O (m6 mol-2 s-1) 10
-5
10
-6
10
-7
10
kDEA, our results
10
-8 kDEA, Little et al. 4
kH2O, our results
kH2O, Little et al.4
-9
10
-3 -3 -3 -3 -3 -3
2.8x10 2.9x10 3.0x10 3.1x10 3.2x10 3.3x10
1/T (K-1)
Figure 4. Comparison of the zwitterions deprotonation rate constants with the previous data .
10
1
kobs (s-1)
0.1
Our results
Al-Ghawas et al.3
Littel et al.7
Sandall16
0.01
500 1000 1500 2000 2500 3000 3500
-3
CDEA (mol m )
Figure 5. The observed rate constants of COS absorption in aqueous MDEA.
175

Process Design

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Process Design

Hochgeladen von

Copyright:

Verfügbare Formate

---------------------------------------------------------------------------------------------------------------------

60, boulevard Saint-Michel 75272 Paris Cedex 06, France

Chapter 1. Kinetics of Absorption of CO2 in Concentrated Aqueous

Chapter 2. Absorption of H2S by an Aqueous Methyldiethanolamine

Chapter 3. Kinetics of absorption of carbon dioxide in aqueous solutions of N-

Aqueous Solutions in the Range 312 K - 383 K. 51

Chapter 5. Rigorous Simulation Of Gas Absorption Into Aqueous Solutions 74

Chapter 7. Representation of regeneration of CO2-Loaded MethylDiEthanolAmine Aqueous

Chapter 9. Kinetics of Carbonyl Sulphide (COS) Absorption with Aqueous Solutions of

In the seventh part, an analytical method using thermodynamic and kinetic

Kinetics of Absorption of CO2 in Concentrated Aqueous Methyldiethanolamine Solutions

F. Pani, A. Gaunand, R. Cadours, C. Bouallou and D. Richon

Absorption by aqueous solutions of alkanolamines is the dominant industrial process

energies of activation lie between (33.1 and 71.6) kJ.mol-1.

Experimental set-up and mode of operation

At the interface, vapor-liquid equilibrium is assumed. The partial pressure PCO2 is

where H is the molar scale Henry's law constant.

PCO2 is obtained from the measured pressures :

Interpretation of results in terms of reaction kinetics.

The mass-transfer coefficient kL is calculated using the correlation between

Equation I assumed to be at equilibrium. Such a mechanism, for water concentration

r = k C MDEA C CO2 (6)

Comparison with Previous Data and Analysis

ACKNOWLEDGEMENT- The authors gratefully acknowledge the financial support of the

Al-Ghawas, H.A.; Hagewiesche, D.P.; Ruiz-Ibanez, G. ; Sandall, O.C. Physicochemical

Critchfield, J.E. CO2 absorption/desorption in methyldiethanolamine solutions promoted with

Lewis, J.B. The mechanism of mass-transfer of solutes across liquid-liquid interfaces-I.

(a) : MDEA from Alfa

Figure 4 : Arrhenius plot.

t/s P/Pa t/s P/Pa t/s P/Pa

0.00 177608.0 360.31 158309.0 720.02 141850.0

1079.95 126986.0 1600.37 107828.0 2120.46 91469.0

The Lewis cell correlation for mass transfer.

A mass-transfer correlation between dimensionless numbers has been established for

Sh = 0.340 Re 2/ 3 Sc1/ 3 (B1)

From Le Bas tables in Reid et al. (1977) :

vCO2 = 34 10-6 m3.mol-1 and vMDEA = 148.9 10-6 m3.mol-1.

Absorption of H2S by an Aqueous Methyldiethanolamine Solution at 296 and 343 K

F. Pani, A. Gaunand, D. Richon, R. Cadours, and C. Bouallou

A microcomputer equipped with a Strawberry Tree Computers data acquisition card

MDEA + H 2 S MDEAH + + HS - (I)

Sh = 0.25 Re 0.63 Sc 0.42 (5)

Sh=kLDT/DCO2 is the Sherwood number, Re=dNDAg2/m is the stirrer Reynolds number

(D H 2S / m2 .s-1 ) ( / Pa.s) 0.74

= 3.4756 1014 (6)

The interfacial H2S concentration is at equilibrium with the vapor phase

H2S RTA dPT

Results and discussion

Al-Ghawas, H.A.; Hagewiesche, D.P.; Ruiz-Ibanez, G.; Sandall, O.C. Physicochemical

Haimour, N.; Sandall, O. C. Molecular Diffusivity of Hydrogen Sulfide in Water, J. Chem.

Haimour, N.; Sandall, O. C. Absorption of H2S into Aqueous Methydiethanolamine, Chem.

ref no. T/K P0 / Pa PI / Pa PH2S,i / Pa VG / 10-6 m3 a / molH2S/molMDEA H2S / mol.m-2.s-1

1 296.05 154877 2750 152127 181.58 0 0.01155

8 296.05 68473 2750 65723 181.58 0.0116 0.00720

* : MDEA from Merck instead of Aldrich

Table 2 : Diffusion coefficient of MDEA in 50 mass % MDEA solutions at 296 K

ref no. T/K kL / 10-5 m2.s-1 a / molH2S/molMDEA DMDEA / 10-9 m2.s-1

1 296.05 0.373 0 0.25

Figure 2: Loading influence on MDEA diffusivity

t/s P/Pa t/s P/Pa t/s P/Pa

0.00 154877 22.30 151140 44.55 147502

66.85 143944 95.36 139484 123.86 135168

152.31 130749 180.76 126495 209.22 122346

237.72 118140 266.17 114050 294.68 110004