Ind. Eng. Chem. Res.

1992, 31, 921-927

92 1

GENERAL RESEARCH
Kinetics Study on Absorption of Carbon Dioxide into Solutions of
Activated Methyldiethanolamine
Guo-Wen Xu, Cheng-Fang Zhang,* Shu-Jun &in, and Yi-Wei Wang
Research Institute of Inorganic Chemical Technology, East China University of Chemical Technology,
Shanghai 200237, People's Republic of China

A disk column was used to investigate the kinetics of carbon dioxide absorption in activated methyldiethanolamine (MDEA) solutions which contain piperazine as an activator. Experiments were
conducted within the temperature range 30-70 "C, MDEA concentration 1.75-4.21 kmol/m3, and
piperazine concentration 0.0414.21 kmol/m3. The kinetic data agree with a proposed mechanism
which can be regarded as a rapid pseudo-first-order reversible reaction between COzand piperazine
in parallel with the reaction between C02 and MDEA. The absorption rate can be expressed as
NCO= HCO,[DCO
(k&, + k,Cp)11/2(p~o,
- pco?), where k2 = 5.86 X lo6 exp(-3984/T) and k, = 2.98
x 1611 exp(-6424/~).
Introduction
The BASF activated methyldiethanolamine (MDEA)
technology for recovery of carbon dioxide from gas mixture
was developed in the 1970s, and it was well-known as a
low-energy process (Meissner and Wagner, 1983). The
energy consumption of the BASF activated MDEA technology is only one-third that of conventional Benfield
technology, and the activated MDEA solution is more
stable, is not subject to degradation, is noncorrosive to
carbon steel, and has a low minimizing solubility of hydrocarbon.
In general, the overall reaction between COBand MDEA
can be written as
COz + HzO + R3N F= R3NH+ + HC03-

(1)

The first investigation of the kinetics of C 0 2reaction

with MDEA was reported in the literature by Barth et al.
(1981). They studied the kinetics of C02 reaction with
MDEA by using a stopped-flow technology, and the contribution of MDEA to the absorption was assumed only
through the neutralization of the species formed by carbon
dioxide hydration.
Blauwhoff et al. (1983) studied the reaction between COP
and various alkanolamines including MDEA at 25 "C.
Their results showed a significant base catalytic effect of
MDEA on hydration of COz.
In light of the findings by Blauwhoff et al. (1983), Barth
et al. (1984) reexamined their previous results for the reaction of COPand MDEA (Barth et al., 1981). They found
that their new results can be very satisfactorilyinterpreted
by the mechanism supported by Blauwhoff et al. (1983).
Yu and Astarita (1985) have studied the kinetics of the
reaction of carbon dioxide in aqueous MDEA solution in
a range of MDEA concentrations from 0.25 to 2.5 kmol/m3
and temperatures from 40 to 60 "C. In this study, the
authors concluded that MDEA acts as a homogeneous
catalyst for COzhydrolysis and then proposed a possible
zwitterion mechanism to account for the catalytic effect.
Versteeg and Van Swaaij (1988) studied the kinetics of
the reaction between COz and aqueous MDEA solutions

in a range of temperature from 20 to 60 "C. The results

showed that the kinetics is in good agreement with the
proposed reaction mechanism which is a pseudo-fmt-order
irreversible reaction.
Wang (1988) studied the kinetics of C02 reaction with
MDEA in a disk column at atmospheric pressure. The
absorption rates of COPinto aqueous MDEA solution were
measured in a range of MDEA concentrations from 1.75
to 4.28 kmol/m3 and temperatures from 30 to 70 "C. All
experimental data of absorption rates in those studies can
be described by a relational expression for a rapid pseudo-first-order reversible reaction.

Nco,= ~ C O z ~ ~ C O z ~ z ~ a m 1 1 ' 2-@PCOZ*)

C02

(2)

where

kz = 5.86

lo6 exp(-3984/T)

(3)

The activation energy E2 is 33.13 kJ/mol.

Tomcej and Otto (1989) investigated the rate of absorption of COz into aqueous solutions of MDEA by using
a single sphere absorber. Kinetic rate constants were
calculated from the absorption rate data by a pseudofirst-order reaction mechanism.
An activator must be added into the absorbent in order
to accelerate the absorption or desorption rate. Piperazine
was used as an activator in the BASF activated MDEA
technology (Appl et al., 1982). It was reported that piperazine would be more effective than the conventional
absorption accelerators. In this paper a study on kinetics
of COz absorption into solutions of MDEA activated by
piperazine is presented.
Theoretical Analysis
Absorption Reactions. MDEA has a base catalytic
effect on the C02 hydrolytic reaction. According to the
zwitterion mechanism reported by Yu and Astarita (1985),
MDEA may react with C02 in liquid film to form an
unstable weakly bonded C02-nitrogen atom complex as
follows:
R3N + COZ R3NCOO
(4)

0888-5885/9212631-0921$03.00/0 0 1992 American Chemical Society

922 Ind. Eng. Chem. Res., Vol. 31, No. 3, 1992

Then the hydrolytic reaction of R3NCO0 takes place in

the liquid phase as follows:
R3NCO0 + H20 + R3NH++ HC03(5)
Reaction 5 is a homogeneous hydrolytic reaction in
equilibrium, and reaction 4 is the dominant reaction of
absorption of C02 into activated MDEA solution and
generally can be regarded as rapid pseudo first order with
respect to R3N if the partial pressure of C02 is not very
high and the free concentration of MDEA is not very low.
Simultaneously,the activator piperazine may react with
C02 in liquid film to form an intermediate as follows:
R(NH)2 + 2C02 R(NHC00)2
(6)
Reaction 6 is a rapid reaction which is parallel with reaction 2. The hydrolytic reaction of R(NHC00)2 also
takes place in equilibrium in the liquid phase as follows:
R(NHC00)2 + 2Hz0 R(NH2+) + 2HC03- (7)

Equilibrium Relationship in the Liquid Phase.

Piperazine acta as an effective activator in MDEA solution.
Its main contribution to absorption of C02is the fact that
C02 can be transferred through the intermediate Rto MDEA
R(NHC00)2 + 2R3N + R(NH)2 + 2R3NCO0 (8)
R(NHCOO)2can be renewed to R(NH)2 by reaction 8
in equilibrium.
[R3NC00I2[R(NH)2]
Ka =
[R(NHC00),][R3NI2

MDEA and piperazine can be obtained if X and W are

fixed.
Absorption Rate. A kinetics study (Wang, 1988) indicated that the absorption of C02 into MDEA solution
is a rapid pseudo-first-order reaction with respect to free
MDEA. In activated MDEA solution, the concentration
of free piperazine will be depleted if C02 does not transfer
to MDEA. However, the rate of transfer reaction 8 is so
fast that the concentration of free piperazine only depends
on the equilibrium of reactions defined by eq 12 and is not
subject to depletion. Therefore, the model of reaction can
be regarded as two parallel rapid pseudo-first-order reversible reactions and the overall reaction rate of carbon
dioxide is
r = (k2C, + kpCp)(Cco,- CCO2*)
(16)
where C, and C, are the free concentration of piperazine
and MDEA, respectively; they can be obtained from eqs
13 and 15.
When the diffusion equation in liquid film is combined
with a rapid pseudo-first-orderparallel reversible reaction,
the chemical absorption rate can be expressed as follows:
The absorption rate coefficient k which regards the
partial pressure as the driving force can be expressed as
follows:
or

For reaction 5, the equilibrium relationship of R3NCO0

hydrolysis is

For reaction 7, the equilibrium relationship of Rhydrolysis is

Combining eqs 9-11, we may obtain a relationship as

follows:

Consider an aqueous solution initially containing C,

of MDEA and Cpo of piperazine; the conversions of MDEA
to R3NH+ and piperazine to R(NH2+I2are yamand yp,
respectively. We assume the intermediate concentrations
of R,NCOO and R(NHC00)2 are very low and can be
C, = [R3N] = (1neglected; then [R3NH+]= y,C,,
y,)C,O,
[R(NH2+)21= ypCpo,and Cp = [R(NH),] = (1
- y,)Cpo. Equation 12 may be rearranged to give
1- Y p

-=w(i..)
YP

1-Y,

where

W = K&b2/K,
(14)
Equation 13 shows the relationship of conversions dispensed between MDEA and piperazine. Therefore, the
total content of C02 absorbed in liquid is
= yc,o,
+ 2ypcpo
(15)

When eqs 13 and 15 are combined, the conversions of

Physical Properties
The physicochemical properties needed to interpret the
absorption data are the solubility and diffusivity of C02
in aqueous activated MDEA solutions. Because the content of added piperazine is very low, we use the values of
Hco, and Dc in MDEA solution directly to substitute for
activated M8EA solution. These physical properties of
MDEA solution were previously reported by Haimour et
al. (1987).
During the course of absorption, electrolyte ions, Le.,
HC03-, R3NH+,and R(NH2+)2,would be formed and affect the solubility of C02. It can be estimated by the
method of Van Krevelen and Hoftijizer (Danckwerts,
1970). In order to estimate the coefficient for R3NH+and
R(NH2+)2,the value for NH4+ is used for R3NH+ and
R(NH2+),. The diffusivity data of C02in aqueous MDEA
was correlated by Haimour et al. (1987). The solution
viscosity was measured by using a capillary viscosity meter.
Partial pressure of water vapor and C02 in activated
MDEA solutions was measured by Wang (1988).
Experimental Section
A disk column, developed by Stephens and Morris
(1951), was used for the kinetic measurements. The sketch
of the experimental setup is presented in Figure 1. The
principal dimensions of the disk column are shown in
Table I.
Carbon dioxide from cylinders with a purity of 99.0%
was passed through a molecular sieve adsorber and then
a soap film meter to measure the flow rate. After passing
through a heating tube, the gas was introduced to the
bottom of the disk column. The outgas flow rate was also
measured by a soap film meter.
The activated MDEA solution was heated to the desired
temperature and then pumped to the top of the disk

"':"
Ind. Eng. Chem. Res., Vol. 31, No. 3,1992 923

0 70 OC

- 7

a
1

'

Figure 1. Schematic diagram of the apparatus used for absorption

experiments: 1, pump; 2, stirred flask; 3, magnetic stirrer; 4,liquid
sample valve; 5, heating tape; 6,disk column; 7, molecular sieve
adsorber; 8, COPcylinder; 9, soap film meter.
Table I. The Principal Dimensions of the
number of disks
disk diameter
disk thickness
tube diameter
free space (dry)
absorption surface
mean perimeter for liquid flow
equivalent diameter for gas flow

Disk Column
40
14.4 mm
4.3 mm
30 mm
85 %
0.02057 m2
35.97 mm
19.3 mm

column. After direct contact with the gas, the solution was
introduced into a stirred flask. The solution was recycled
to the column, so that the content of C02 absorption in
solution increased with run time. We measured the absorption rate at intervals of minutes and took liquid sample
to analyze simultaneously. The absorption rate data at
any content of C02 in solution can be obtained within a
shorter run period and consume only a small amount of
activated MDEA solution. The flow rate of solution was
fixed at 5 L/h during all experiments.
The absorption rate of C02was calculated from inlet and
outlet flow rates of C02 measured by soap film meters.
The concentration of C 0 2absorption in solution was analyzed by chemical method.
The MDEA concentrations in solution used in the kinetic studies were 1.75, 3.04, and 4.21 kmol/m3, and the
piperazine concentrations in solution were 0.041,0.10, and
0.21 kmol/m3. The experiments took place at 30,40,55,
and 70 "C.
We measured the liquid-film mass-transfer coefficient
of this disk column using a pure carbon dioxide-water
system at 20 OC. Experimental results were well correlated
as follows:

Results and Discussion

Absorption Rate Coefficient k. The absorption rate
of C02into activated MDEA solution was monitored under
different conditions. Figure 2 shows the changes in absorption rate coefficient k at various temperatures. The
absorption rate coefficient increases with temperature,
since both reaction rate constant and diffusivity increase
with temperature. The absorption rate coefficient decreases with increasing conversion for C, C, and Hco,
decreases with conversion.
Comparing the absorption rate coefficient of activated
MDEA aqueous solution and that of MDEA aqueous solution without any piperazine given by Wang (1988), we
find that a little piperazine will obviously improve the
absorption rate. That means the higher the absorption

2
P
Y

*Y
4

0 1-75 k m l / m 3

A 3.04 kmol/m3

!%
W

l n 4

L
s 32

4.21 kml/m3

924 Ind. Eng. Chem. Res., Vol. 31, No.3, 1992

t-:
3

3
P

r(

z
m

u)

z
0

2
m

co

Ind. Eng. Chem. Res., Vol. 31, No.3,1992 925

00

c?

*
aj

m
m

m
N

N
*

v)
v)

(D

%:

2
m

rl
m

m
*

,.

.e

0B
.-f

&

3
8
.-a

926 Ind. Eng. Chem. Res., Vol. 31, No. 3,1992

Ln

xya 6oo
500

0.04

0.08

0.20

0.16

0.12

0.24

p.1

0.2

0.3 0 . 4 0 . 5 0.6
Yam

Figure 4. Absorption rate coefficient vs concentration of piperazine

in 4.21 kmol/m3 MDEA solution at 55 "C.

160

-A

140

100 -

100

0.21

0.6

120

80

, kmol/J

0.041

0 0.10
0

Figure 6. k , vs concentration of piperazine at 55 "C.

2.4

1.8

1.2

, ,c

'.6

3.0

4.2

kmol/m3

Figure 7. k, vs concentration of MDEA at 40 "C.

3000

2000

-<

1000

y1

"

0.02

0.06

0.10

0.14

0.18

0.22

700
500

cP,

kmol/m3

Figure 5. Apparent rate coefficient vs concentration of piperazine

in 4.21 kmol/m3 MDEA solution at 55 "C.

Apparent Rate Constant k,,,,,. According to eq 19,

the apparent rate constant k,, can be obtained from the
The conversions of
values of k, Hcoz, Dcoz, k2, andPC,.
MDEA and piperazine are determined by eqs 13 and 15.
In order to simplify the model, we assumed that W is
hardly changed with temperature and can be regarded as
a constant. Then we try to calculate the value of k, from
W within the range 0.8-1.2. It is found that the experimental results are in good agreement with the proposed
reaction mechanism if W = 1.0. The conversions and free
concentrations of MDEA and piperazine corresponding to
a certain content of COz in solution can be calculated
according to eqs 13 and 15 by a trial and error method.
In this paper, we used the value of k 2 given by eq 3
(Wang, 1988). The calculation results are shown in Table
11.
The apparent first-order rate constants are calculated
and listed in Figure 5. k o m Figure 5, a linear dependence
of the apparent rate constant ,k
with concentration of
piperazine is observed. Therefore, it can be concluded that
the reaction between piperazine and C02is fmt order with
respect to piperazine and is in good agreement with the
proposed reaction mechanism.
Second-OrderRate Constant k,. The values of the
second-order rate constant k, are shown in Figures 6 and
7. From Figures 6 and 7, k, is nearly constant at the same
temperature under various concentrations of MDEA and
piperazine. It can be concluded that the reaction between
piperazine and COzis a rapid pseudo-fmt-order reversible
reaction, and it is parallel with the reaction between
MDEA and C02. An Arrhenius plot of the values of the

h
E

300

200

2.7

2.9
l/T*103,

3.1

5.5

3.3

k-l

Figure 8. Arrhenius plc for k,.

h

//

el0
u1
N.

r 4

2 4

10

k* I O 5 , kmol/( m2, s.Mpa)

Experiment value

Figure 9. Comparison between value of experiment and model.

second-order rate constant k, is shown in Figure 8. An

activation energy E of 53.41 kJ/mol is indicated by the
data. The fact of
> Ez shows that the higher the absorption temperature, the higher the contribution of piperazine on absorption rate as shown in Figure 2.
The rate constant k, can be represented by

k, = 2.98 X 10" exp(-6424/T)

(21)

According to eq 18 or 19, we use the values of k2 and k,

Ind. Eng. Chem. Res. 1992,31,927-934

to calculate the absorption rate coefficient and then compare it with the experimental values. The results are
presented in Figure 9 with mean error no greater than
10%.

Acknowledmnent
This work was supported by the Science Foundation of
National Education Committee of China.
Nomenclature
C = concentration in liquid phase, kmol/m3
Co = initial component concentration, kmol/m3
C* = concentration in equilibrium condition, kmol/m3
D = diffueivity in liquid phase, mz/s
E = activation energy, kJ/mol
H = solubility coefficient, kmol/(m3.MPa)
K,, Kb,K, = equilibrium constants
k = absorption rate coefficient, kmol/ (m2.s.MPa)
kL = mass-transfer coefficient in liquid film, m/s
k2, k p = reaction rate constant of reactions 2 and 7,respectively, m3/(kmol*s)

kzapp= apparent firsborder rate constant, kZapp= kzC,,

k
= apparent first-order rate constant, kp,app= kpCp,
fi:Pabsorption rate, kmol/(m2.s)
p = partial pressure, MPa
p * = partial pressure in equilibrium condition, MPa
r = reaction rate, kmol/(m3-s)
T = absolute temperature, K
X = content of COz absorbed in liquid phase, kmol/m3
y = conversion
Greek Symbols
r = liquid flow rate per unit width of surface, kg/(ms)
p = density of liquid, kg/m3
ec. = viscosity of liquid, Pa/s
Subscripts
COz = corresponding to COz
am = corresponding to MDEA

927

p = corresponding to piperazine
Registry No. MDEA, 105-59-9;COz, 124-38-9;piperazine,
110-85-0.

Literature Cited
Appl, M.; Wagner, U.; Henrici, H. J.; Kuessner, K.; Voldamer, K.;
Fuerst, E. Removal of C02And/or H a And/or COS From Gases
Containing These Constituents. U S . Patent 4,336,233,1982.
Barth, D.; Tondre, C.; Lappai, G.; Delpuech, J. J. Kinetic Study of
Carbon Dioxide 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 Reaction of Carbon Dioxide with Alkanolamines: A Discussion Concerning the Cases of MDEA and DEA. Chem. Eng. Sci.
i984,39,i753-i757.
Blauwhoff, P. M. M.; Versteeg, G. F.; Van Swaaij, W. P. M. A Study
on the Reaction Between CO, and Alkanolamines in Aaueous
Solutions. Chem. Eng. Sci. 1983,38,1411-1429.
Danckwerta, P. V. Gas-Liquid Reactions; McGraw-Hill: New York,
1970;p 19.
Haimour, N.; Bidarian, A.; Sandall, 0. C. Kinetics of the Reaction
Between Carbon Dioxide and Methyldiethanolamine. Chem. Eng.
Sci. 1987,42,1393-1398.
Meissner, R. E.; Wagner, U. Low-energy Process Recovers C02. Oil
Gas J. 1983,Feb, 55-58.
Stephens, E. J.; Morris, G. A. Determination of Liquid-Film Absorption Coefficients. Chem. Eng. Prog. 1951,May, 232-242.
Tomcej, R.A.; Otto, F. D. Absorption of C 0 2 and N20 Into Aqueous
Solutions of Methyldiethanolamine. AIChE J. 1989,35,861-864.
Versteeg, G. F.; Van Swaaij, W. P. M. On The Kinetics between C02
and alkanolamines both in aqueous and non-aqueous solutions11. Tertiary Amines. Chem. Eng. Sci. 1988,43,587-591.
Wang, Y. W. Kinetics of C02 absorption in Activated Methyldiethanolamine. M.S. Thesis, East China University of Chemical
Technology, 1988.
Yu, W. C.; Astarita, G. Kinetics of Carbon Dioxide Absorption in
Solutions of Methyldiethanolamine. Chem. Eng. Sci. 1985,40,
1585-1590.
I

Received for review March 13, 1991

Revised manuscript received September 17,1991
Accepted November 8,1991

Diffusion of Phenylacetic Acid and Vanillin in Supercritical Carbon

Dioxide
Tony Wells,* Neil R. Foster, and Rodney P. Chaplin
School of Chemical Engineering and Industrial Chemistry, University of New South Wales, P.O.Box 1,
Kensington 2033, N S W Australia

Binary diffusion coefficients, Ol2, of phenylacetic acid and vanillin have been determined in supercritical carbon dioxide using the Taylor-Aris tracer response technique. DI2values are reported
for temperatures ranging from 308 to 318 K and densities between 600 and 850 kg/m3. The binary
diffusion coefficients, D12,have a magnitude of
m2/s. The influence of pressure, temperature,
and carbon dioxide density on the D12values is examined. The applicability of Stokes-Einstein
based correlations and the free volume diffusion model is also evaluated.
Introduction
The advantages of utilizing solvents in a supercritical
state to perform extractions and achieve separations are
documented
williams lg81;
et
1983)'
The use of supercritical COz,the most commonly employed
supercritical fluid (SCF),has considerable potential in the
*,'

*Address correspondence to this author at Faculty of Chemical

Engineering, Delft University of Technology, P.O.Box 5045,2600
GA Delft, The Netherlands.

0888-5885/92/2631-0927$03.00/0

flavor and fragrance industry as it is inexpensive, nontoxic,

and nonflammable. Additionally C 0 2is an effective solvent at ne^ ambient conditions (30-60 oc), and conse-

quently the
for thermal degradation, characteristic of alternative distillation
is minimized.
The potential for the use of SCFs in the extraction of
flavor and fragrance
has
in n u m e r o ~investigations of their solubility in COz (e.g., Wells et al.,
1990; Vitzthum and Hubert, 1978; Schutz e t al., 1984).
However, there has been little or no information published
on the diffusion and mass-transfer characteristics of such
0 1992 American Chemical Society