c h e m i c a l e n g i n e e r i n g r e s e a r c h a n d d e s i g n 8 6 ( 2 0 0 8 ) 401–409

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/cherd

Selectivity study of H2 S and CO2 absorption from gaseous

mixtures by MEA in packed beds

Hamid Reza Godini, Dariush Mowla ∗

Chemical and Petroleum Engineering Department, School of Engineering,
Shiraz University, Shiraz, Iran

a r t i c l e i n f o a b s t r a c t

Article history: In this work, an experimental and theoretical investigation of selectivity was performed
Received 28 April 2007 for simultaneous absorption of H2 S and CO2 from gaseous streams into an aqueous mono-
Accepted 28 November 2007 ethanol amine (MEA) solution in a packed bed. The operating parameters playing a vital role
and have been considered here, are amine concentration and pressure, while the structural
and hydrodynamic parameters such as packing size and L/G ratio were found to be impor-
Keywords: tant as well. The qualitative effects of these parameters on absorption performance were
Simultaneous absorption analyzed via evaluating of absolute and relative removal efficiency. In the proposed math-
MEA ematical model for theoretical investigation of the absorption performance, the two-film
Selectivity theory of mass transfer was used. In order to check the validity of the proposed model, in
Acid gas compounds experimental section, a pilot-scale plant was constructed and variety of experimental data
Structural parameters were obtained in a wide range of operating conditions. In this plant, construction of packed
Packed beds beds in several sections of the column, gives the possibility of changing the structural and
hydrodynamic parameters in a flexible way. Also the proposed mathematical model is able
to consider the effects of vital parameters on the selectivity of absorption. The predicted
results by the proposed mathematical model were compared to the experimental data. This
comparison shows a good agreement between these data and so verifies the validity of
the proposed model. The results reveal that the effects of pressure, amine concentration,
packing size and L/G ratio on the selectivity performance of H2 S absorption into an MEA
solution are significant. Results show that, the relative efficiency could be increased near
20% when L/G is manipulated properly. As a side effect it reduces the absolute efficiency.
Using the packing size 1 in. instead of 0.5 in., relative efficiency could be increased more
than 5% while absolute efficiency is reduced. Then the experimental results testified the
benefit of any operating procedure appropriately involving these parameters for selective
H2 S absorption.
In general, since CO2 is a weak competitor against H2 S in being absorbed by MEA
solution, under difficult conditions for absorption process, the weakness of CO2 is more
pronounced. Therefore, the selectivity of absorption is increased, while the absolute effi-
ciency is decreased. In particular, increasing the gas flowrate and decreasing the L/G ratio,
increases the selectivity. Also when higher packing size is used, selectivity is increased. In
contrast, when large amount of amine are contacted with a special quantity of acid gas in
the unit volume of packed bed (generally lower packing size and higher L/G ratio), due to
the strength of MEA, the amine solution absorbs nearly all gaseous acidic compounds and
reduces the selectivity.
© 2007 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

∗
Corresponding author. Tel.: +98 711 62865315x3143; fax: +98 711 6287294.
E-mail addresses: dmowla@shirazu.ac.ir, dmowla@gmail.com (D. Mowla).
0263-8762/$ – see front matter © 2007 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.cherd.2007.11.012
402 c h e m i c a l e n g i n e e r i n g r e s e a r c h a n d d e s i g n 8 6 ( 2 0 0 8 ) 401–409

Nomenclature 1. Introduction

a specific surface area of the column (m2 /m3 ) Removal of H2 S from gaseous mixtures by use of amine solu-
A cross-sectional area of the tower (m2 ) tions is a conventional technique (variety of processes such as
C liquid concentration (kmol/m3 ) SCOT and natural gas sweetening are using this technique).
Ci liquid concentration of component i (kmol/m3 ) Usually these gaseous mixtures contain 1–5% CO2 which is a
Cp specific heat of mixture at constant pressure close competitor of H2 S (less than 1.5%) in being absorbed by
(kJ/(kmol ◦ C)) amine solutions. So the behavior of selective H2 S absorption by
dp packing diameter (m) amine solution is very interesting from economical and oper-
dv volume of each volume element (m3 ) ational points of view. It should be noted that the targets for
dz height of each volume element modeling of this system in different situations, might be com-
Di diffusivity of component i in liquid phase (m2 /s) pletely different. For example in SCOT process which the inlet
G gas phase flow rate (m3 /s) gas contains considerable amount of CO2 , even many times as
G mass flux of gas (kg/(m2 s)) existing H2 S, the purpose of modeling is to study the process
Hi Henry’s constant of component i in aqueous with emphasis on the selectivity of absorption. This subject
MEA solution has been studied in different aspects by different investigators
j volume element counter (Zheng et al., 2006; Lyddon and Nguyen, 1999; Yasavage et al.,
kCO2 kinetic reaction constant of CO2 with MEA 2001; Rascol et al., 1997; Pant and Srivastava, 2007; Glasscock
(m3 /(kmol s)) and Rochelle, 1993). The basic idea in majority of these stud-
kG gas phase mass transfer coefficient (kg/(m2 s)) ies is the use of the sterically hindered type of amines as a
kL liquid phase mass transfer coefficient (m/s) chemical factor toward improving the absorption selectivity.
K equilibrium constant In some researches, physical parameters have been investi-
L liquid flow rate (m3 /s) gated to show that they have such a potential for developing
L mass flux of liquid (kg/(m2 s)) the new practical parameters toward achieving more selectiv-
L/G MEA solution–gas ratio (mol MEA/mol gas) ity in chemical absorption. In this manner, it has been verified
n number of data in statistical analysis that, although MEA is not a sterically hindered amine, but it
Ni mass transfer flux of component i (kmol/(m2 s)) could be a selective chemical absorber (Pant and Srivastava,
Pi partial pressure of component i in gas phase 2007).
(kPa) For a long time MEA has been the widest industrially used
qi heat of absorption of component i (kJ/kmol) amine (especially in oil refineries). Therefore, any success-
ri consumption (reaction) rate of component i ful effort toward optimum usage of this amine (even a little
(kmol/s) progress) leads to considerable improvement. MEA is a strong
Sc Schmidt number amine and under well prepared conditions in absorption pro-
Ss standard deviation of statistical data for popu- cess, it absorbs almost all available gaseous acidic compounds
lation without serious competition (alkalinity potential). By use of
SE standard deviation of statistical data for sam- secondary and tertiary amines, difference potential between
ples absorption of H2 S and CO2 will be more pronounced due to the
t confidence coefficient in statistical analysis weakness of these amines. These types of amines are naturally
T temperature (K) selective and when they are involved in the liquid mixture, the
TL liquid bulk temperature (K) observed improvements followed from the key parameters in
Vi molecular volume (m3 /kmol) this article are not noticeable beside the effects of the intrinsic
x direction toward depth of liquid phase selectivity of these amines.
xi mole fraction of component i in liquid phase The main orientation of this study is the controllable
(mol i/mol liquid) employing the existing potential of selectivity for absorption
yi mole fraction of component i in gas phase (mol of H2 S and CO2 into aqueous MEA solutions. In order to exam-
i/mol gas) ine the pure effects of proposed parameters on the observed
selectivity, the unselective amine like MEA was chosen as a
Greek letters
case study here.
˛ confidence limit in statistical analysis
In spite of numerous researches in this field, little con-
ıc thickness for zone of liquid phase which there,
siderations have been given to the mechanisms and related
the reaction of CO2 with MEA takes place (m)
parameters of selective amine absorption. For example, the
ε packed bed void fraction (m3 /m3 )
structural parameters such as packing size and the hydrody-
i absolute efficiency of absorption of component
namic parameters such as L/G ratio have not been considered
i
widely as effective parameters on mechanisms and selectiv-
rel relative efficiency which is equal to (H2 S /CO2 )
ity of absorption for these compounds. It is the purpose of
 viscosity (kg/ms)
this work to consider these parameters in details. To do this
1 normal population mean of each data set
a combination of chemistry and transport phenomena is con-
˚ association factor for solvent
sidered as a basis of the mathematical model. The resulted
˚0 degree of freedom in two way ANOVA test
trends let us conclude useful guidelines for appropriate use of
these parameters to achieve more selective performances.
The pilot-scale plant used in the experimental study was
constructed in sulfur recovery unit (SRU) in Shiraz oil refinery.
It should remained as close as possible to the industrial condi-
 c h e m i c a l e n g i n e e r i n g r e s e a r c h a n d d e s i g n 8 6 ( 2 0 0 8 ) 401–409 403

tions in the amine unit in order to optimize the operating con-
ditions and cost. In this manner it became possible to use the
existing regenerator of amine unit. Therefore, the conditions
(temperature and concentration) of lean amine were fixed.
1.1. Mathematical modeling
Different mathematical models have been proposed for simul-
taneous absorption of CO2 and H2 S in reacting solutions
(Brettschneider et al., 2004; Gabrielsen et al., 2007; Mandal
and Bandyopadhyay, 2005; Saha et al., 1999; Ali et al., 2002).
In this section a general mathematical model is proposed for
the case of MEA solution. It is based on the chemistry and
transport phenomena relations, applying over the typical area
where two films of liquid and gas contact each other. In this
model, the orientation is to diversify the physical and chemi-
cal mechanisms in simultaneous absorption of acidic gaseous
compounds in to MEA solution for possible usage in improving
the process selectivity.
In this area the gas and liquid films are moved in a counter
current way. In Fig. 1, five different zones that could be visu-
alized for this process are shown. These are as follows:
Zone 1: Bulk of upward gas stream.
Zone 2: Diffusing gas film through which the acidic compo-
nents (CO2 and H2 S) are diffusing toward liquid–gas interface.
Zone 3: Liquid–gas interface.
Zone 4: Diffusing liquid film in which CO2 is diffusing into
the depth of liquid. In this zone H2 S instantaneously and
CO2 gradually react with MEA.
Zone 5: Bulk downward liquid stream.
The basic components, which are considered in liquid
phase, are; CO3 2− , CO2 , H2 S, HS− , MEA and MEAH+ and the
main reactions occurring in the zone 4 are as follows:
MEA + H2 S ↔ MEAH+ + HS− (instantaneous)
2MEA + CO2 + H2 O → 2MEAH+ + CO3 2− (fast)
ally same as those which have been reported in other studies
(Brettschneider et al., 2004). They have been obtained based
on the proposed absorption mechanisms, chemical reactions
and governing equations. In this manner, H2 S concentration
profile is sharply reduced at the interface due to its instan-
taneous rate of reaction with MEA and then H2 S diffuses in
equilibrium state (slight amount) through the zone 4 toward
the bulk part of the falling liquid film.
CO2 profile is reduced due to fast reaction with MEA and
the carbon dioxide diffuses toward the bulk part of the liquid
film until complete conversion.
Due to instantaneous reaction of MEA with H2 S (at the
gas–liquid interface) and the fast reaction with CO2 in the
whole reaction-diffusion zone (zone 4), the MEA concentration
is reduced from its amount in bulk part of the liquid phase
until neighborhood of gas–liquid interface where it sharply
reduces due to the reaction with H2 S.
Due to the important phenomena taking place in the liq-
uid film and their important role in the process performance,
the attention is focused on this zone (zone 4). Therefore, the
mathematical model was constructed based on what is occur-
ring in this zone. As such, the mass balance equation for each
component is applied on the typical liquid shell which was
considered in zone 4. The resulted relations are the process
governing equations (Eqs. (3)–(8)). Based on the kinetic of reac-
tions occurring in this system and different resistances that
exist in this process, the governing absorption mechanisms
for various components are different. H2 S reacts with MEA
at the liquid–gas interface instantaneously while reaction of
CO2 with MEA is relatively slower and takes place gradually in
whole of the zone 4. For H2 S due to low concentration (equilib-
rium concentration) and for HS− due to low reactivity poten-
tial, the governing equations are Fick’s law. But for CO3 2− , CO2 ,
MEA and MEAH+ due to reaction of CO2 with MEA, the combi-
nation of reaction rate and diffusion would be the governing
equations describing their material balances over the zone 4.
Based on the above mentioned descriptions, the material bal-
ance equations for six basic components are given as follows:

In order to propose a mathematical model explaining this d2 CCO2 kCO2 CCO2 CMEA
+ =0 (3)
system, the following assumptions were made: dx2 DCO2

(1) Absorption of H2 S in MEA solution is more sensitive to dif- d2 CCO 2− kCO2 CCO2 CMEA
3
− =0 (4)
fusion steps (diffusion controlled mechanism) than CO2 dx2 DCO 2−
3
absorption.
(2) No chemical reaction occurs in the bulk of liquid (equilib-
d2 CMEA kCO2 CCO2 CMEA
rium). + =0 (5)
dx2 DMEA
(3) Electrical neutrality in liquid phase.
(4) Henry’s law is applicable at gas–liquid interface.
d2 CHS−
(5) Evaporation of water and MEA is neglected. =0 (6)
dx2

The reaction of CO2 with MEA is a one-side reaction with a

d2 CH2 S
rate given as =0 (7)
dx2

rCO2 = kCO2 CCO2 CMEA (1)

d2 CMEAH+ kCO2 CCO2 CMEA
− =0 (8)
dx2 DMEAH+
In which the rate constant kCO2 is given by (Riesenfield,
1998):
The thickness of the zone 4 in which the reaction of CO2
with MEA takes place is defined by the following equation
21.95
ln kCO2 = (2) (Rascol et al., 1997):
T

The concentration profiles of the components are also DCO2

ıc = (9)
shown in different zones of Fig. 1. These profiles are gener- kL
404 c h e m i c a l e n g i n e e r i n g r e s e a r c h a n d d e s i g n 8 6 ( 2 0 0 8 ) 401–409

Fig. 1 – Conceptual diagram of the two-film theory in the proposed model.

Also the following equations are used to define the fluxes This B.C. is based on the assumption of electroneutrality at
of mass transfer (Rascol et al., 1997): the surface.

dCi CMEAH + CHS−

Ni = Di |x−0 (liquid phase mass transfer) (10) (B.C.12) at x = ıc , C H2 S = (23)
dz CMEA KC

Ni = kGi Yi (gas phase mass transfer) (11) In the above expressions j denotes the number of the lon-
gitudinal (volume) elements of the tower (each has height
In order to solve the above Eq. (12), boundary conditions are dz = 0.0005 m; j = 1, 2, . . ., 3600).
needed. Considering the situation depicted in Fig. 1 and based The following expressions are employed for calculation of
on the assumptions in the mathematical model, the following gas and liquid phase mass transfer coefficients (Riesenfield,
boundary conditions (B.C.s) can be proposed for the edges of 1998; Treybal, 1980) and the general diffusion coefficients in
zone 4 of liquid phase (which are addressed here by x = ıc and liquid and gas phases (Perry and Green, 1984). The equilib-
x = 0). rium constant and the Henry’s constants were obtained from
reference Kent and Eisenberg (1976).
(B.C.1) at x = ıc , CCO2 = 0 (12)  0.45
dp L DL 0.5
kL = 25.1 Sc (24)
PCO2 L dp L
(B.C.2) at x = 0, CCO2 |0 = (13)
HCO2
 −0.36
dp G G −2/3
(B.C.3) at x = 0, CMEA = 0 (14) kG = 1.195 Sc (25)
G (1 − ε) PBM G

j+1 j NCO2 a dz A
(B.C.4) at x = ıc , C =C + (15) 7.4e−8 (˚B MB )
0.5
T
CO3 2− CO3 2− L DAB = (26)
0.6
B VA

j+1 j ⎡ ⎤−1
(B.C.5) at x = ıc , CMEAH+ = CMEAH+
  xj 
Di,m = ⎣ ⎦ (27)
a dz A(NH2 S + 2NCO2 ) Dij
+ (16) j
L

Other required information such as diffusivities in gas phase

j+1 j a dz A(NH2 S + 2NCO2 )
(B.C.6) at x = ıc , CMEA = CMEA − (17) and, etc., were obtained from Perry’s chemical engineering
L
handbook (Perry and Green, 1984).
(B.C.7) at x = 0, CCO 2− |0 − CCO2 |0 = CCO 2− |ıc (18)
3 3
1.2. Solution of the model
PH 2 S
(B.C.8) at x = 0, CH2 S |0 = (19) In order to solve the Eqs. (3)–(8) by simultaneous use of the
HH2 S
boundary conditions (12)–(23) and auxiliary Eqs. (2), (9)–(11)
j+1 j NH2 S a dz A and (24)–(27), the bed is first divided into 3600 small vol-
(B.C.9) at x = ıc , CHS− = CHS- + (20) ume elements of height dz = 0.0005 m. Then using an iterative
L
procedure through the bed, all required information for deter-
ıc mining the performance of the whole bed, are calculated. In
(B.C.10) at x = 0, CHS− = CHS− |ıc + NH2 S (21)
DHS− the general loop of calculation (outer loop), the outlet con-
centrations of CO2 and H2 S from the top of the packed tower
(B.C.11) at x = 0, CHS− + 2CCO 2− = CMEAH+ (22)
3 are assumed as initial guesses. In this manner, calculation
 c h e m i c a l e n g i n e e r i n g r e s e a r c h a n d d e s i g n 8 6 ( 2 0 0 8 ) 401–409
is initiated from the top of the bed (which the concentra-
tions in leaving gas stream and entering liquid stream are
known). Gradually all unknown parameters are calculated in
each volume element by applying the before mentioned gov-
erning equations. To start the calculations in each volume
element (steps), an initial pair of guesses for fluxes of CO2
and H2 S in each volume element is assumed since they are
needed for expressing the B.C.s. In this manner, the govern-
ing equations (in zone 4) of each volume element could be
solved numerically and the liquid phase concentration profiles
in that element could be obtained. The initial guesses (mass
transfer fluxes of CO2 and H2 S) for each volume element are
adjusted by new results of mass transfer fluxes, which are cal-
culated based on the determined concentration profiles and
Eq. (10) (inner loop). In each volume element, the mass trans-
fer from gas phase should be equal to the mass transfer in to
the liquid phase. Therefore, in this adjustment procedure the
corresponding gas phase CO2 and H2 S concentrations entering
and leaving each element, are rechecked by use of gas phase
mass transfer coefficient (Eqs. (11) and (25)). This procedure
is continued for all of the elements, up to the bottom of the
bed. In outer loop, the known inlet gas phase concentration is
compared with the calculated gas concentration (in previous
step) entering to the lowest volume of bed. In case of contrast
the initial guesses required in the general loop of calculation
(the outlet gas phase concentrations of CO2 and H2 S from top
of the bed) are adjusted by shooting method. Then the new
pair of guesses is obtained. Through this procedure, the inlet
and outlet liquid and gas phase concentration of components
into/from each bed-volume element are correlated according
to the following equations.
a
CGout = CGin − NiG dv
G
a
CLout = CLin + NiL dv
L For checking the validity of the proposed mathematical model,
A computer program in MATLAB Code developed to per-
form the calculations. Based on the initial guesses, the total
required time for the solution of this model via Pentium 4 pro-
cessor, is at most 5 min. Also multiplicity is not a case here and
convergence behavior of the system is not a critical issue.
The final results of calculations are the concentration pro-
files in liquid and gas phases through the tower. They should
be used to determine the performance of the process in simul-
taneous absorption of CO2 and H2 S.
2. Experimental
In order to investigate the effects of vital parameters exper-
imentally, and also checking the accuracy of the proposed
model, a pilot-scale plant was erected in Shiraz oil refinery.
This plant could be used for simultaneous studying the effects
of different parameters such as L/G ratio, pressure and pack-
ing size. A schematic and a picture of the pilot-scale plantare
shown in Fig. 2.
The 25 cm diameter tower was made of carbon steel with
total height of 370 cm. The tower consists of three 60 cm sec-
tion beds. In each (60 cm) section of the tower, sample points
for gas and liquid were erected. The rate of absorption was
measured by analyzing the gaseous samples for H2 S and CO2
concentration in each section. It was doing via Testo-350 gas
analyzer which time to time was rechecked by detector tubes.
405
In each experiment the gas mixture (SRU tail gas) which
is composed mainly of nitrogen and contains some H2 S and
CO2 , is first pressurized by a reciprocating compressor to
the desired pressure (1–10 bar). Then after mixing with the
required amount of CO2 and H2 S in order to reach the desired
composition in each experiment, it is entered to the bottom
of the tower. The 15 wt% lean MEA solution (low alkalinity
solution is preferred for selectivity study) is fed to the tower
from the top of the bed. In fact, there were three points avail-
able for feeding the amine solution above each section of the
bed. Gas samples were taken and analyzed after each sec-
tion. The experimental results obtained under constant feed
gas composition, 5% CO2 and 1.5% H2 S, temperature of 30 ◦ C.
Low temperature is preferred since it helps physical absorp-
tion step. The vital parameters were total pressure (1–10 bar),
molar L/G (0.1–0.4 mol MEA/mol gas) and packing size (0.5 and
1 in. Raschig ring). Volumetric gas flow rate was around 0.01
(N m3 /s). Mass flow rate of MEA solution was in the range of
0.025–0.1 kg/s. Corresponding average pressure drop for 0.5 in.
packing size was around 1500 Pa/m and for 1 in. packing size
was 600 Pa/m (under 1 bar pressure). Another fixed parameter
was the lean amine concentration, 15 wt%. On the main out-
let gas stream, a globe valve was used for manipulating the
operating pressure. The amine solution flowrate were tuned
to get the desired amount of L/G. The samples of gas stream
were taken in order to analyze the quality and quantity of
absorption under different conditions. Using observed input
and output concentration of gas stream, the efficiencies could
be calculated for different height of the bed since inlet MEA
solution could be entered the tower from each of three avail-
able points.
(28)
3. Results and discussion
(29)
the results predicted by the model are compared with those
obtained by experimentation in the pilot-scale plant for the
same operating conditions.
For ease of comparison and qualitative investigation, the
new parameters, absolute efficiency () and relative efficiency
(rel ) are defined as follows:
Cin − Cout
= (30)
Cin
H2 S
rel = (31)
CO2
Since the qualitative verification of the effects of proposed
parameters (packing size, L/G, . . .) is desirable here, only effi-
ciencies were compared in different states. Especially effects
of different parameters on the selectivity of process were
targeted. Also it should be noted that the possible errors in
prediction of absolute efficiencies will be out due to error dis-
tributing mechanism through evaluation of relative efficiency.
Furthermore, the results of these expressions are comparable
with the published data such as references Zheng et al. (2006)
and Mandal et al. (2004). In order to evaluate the efficiencies
under each set of conditions, only gas phase concentrations
(in entering and leaving gas streams) are required. In this way
the following results were obtained.
406 c h e m i c a l e n g i n e e r i n g r e s e a r c h a n d d e s i g n 8 6 ( 2 0 0 8 ) 401–409

Fig. 2 – A schematic (a) and picture (b) of the pilot-scale plant (right hand side view).

3.1. Effect of operating pressure
Figs. 3 and 4 represent the experimental and predicted results
related to the effects of pressure on absolute and relative
removal efficiency ( and rel ).
As it could be seen in these figures, increasing of pres-
sure increases the absolute efficiency in absorption of both
components. It also increases the relative efficiency of the
process toward H2 S absorption. The observed results reflect
the importance and sensitivity of physical absorption step in
the selective H2 S absorption. Comparison of calculated and
experimental results indicates the validity of the proposed
model in predicting the qualitative and quantitative effect of
pressure.
3.2. Effect of L/G
Figs. 5 and 6 represent the experimental and predicted results
related to the effects of L/G ratio on absolute and relative
removal efficiency ( and rel ). As it has been shown in Fig. 5,
increasing of L/G (increasing L while G is kept constant) has
a positive effect on absolute removal efficiency for absorption
of both H2 S and CO2 .
 c h e m i c a l e n g i n e e r i n g r e s e a r c h a n d d e s i g n 8 6 ( 2 0 0 8 ) 401–409 407

Fig. 3 – Effect of pressure on absolute efficiency of Fig. 5 – Effect of L/G on absolute efficiency of absorption (a)
absorption, (a) for H2 S (b) for CO2 (gas flow for H2 S, (b) for CO2 (gas flow rate = 0.01 N m3 /s, liquid
rate = 0.01 N m3 /s, liquid flowrate = 0.06 kg/s). flowrate range = 0.025–0.0975 kg/s, bottom
pressure = 7.65 bar, top pressure = 7.5 bar).

On the other hand as it was mentioned earlier, the rate

of H2 S absorption is more sensitive to gas phase resistance.
Instead, absorption of CO2 is controlled by both of reaction and
gas phase resistances. In this manner, although increasing
of L/G increases the rate of absorption for both components,
but its effect is more pronounced for CO2 . Therefore, the rela-
tive efficiency is decreased with increasing L/G and vice versa.
Typical results revealing this phenomenon are shown in Fig. 6.
It should be noted that the range of variation of L/G is within
permissible hydrodynamic range, namely between dryness
and flooding regions.
In literature, a few references like Zheng et al. (2006) and
Mandal et al. (2004) have reported such a results and could Fig. 6 – Effect of L/G on relative efficiency of absorption (gas
be compared with this study. For example, Fig. 6 in reference flow rate = 0.01 N m3 /s, liquid flowrate
Zheng et al. (2006) and Fig. 1 in reference Mandal et al. (2004), range = 0.025–0.0975 kg/s, bottom pressure = 7.65 bar, top
“effect of gas flow rate on absorption performance”, reveal that pressure = 7.5 bar).
increasing of gas flowrate, increases the absorption selectiv-
ity. In the fixed liquid flow rate, increasing the gas flow rate
means decreasing the L/G. Therefore, here in this study (cur-
rent work) qualitatively the same results have been reported
but quantitative comparison needs lots of information such
as gas phase composition which is totally different in these
researches.
Therefore, the quantitative comparison is not possible here
because exact experimental conditions such as packing size,
gas and liquid flowrate, packed bed height, temperature, pres-
sure, liquid and gas concentration have to be known about
each set of data. Especially these data might be found rarely
for pilot-scale plants.

3.3. Effect of packing size

Fig. 4 – Effect of pressure on relative efficiency of absorption Using smaller packing size causes higher available mass trans-
(gas flow rate = 0.01 N m3 /s, liquid flowrate = 0.06 kg/s). fer area per bed volume. In this manner the amount of
408 c h e m i c a l e n g i n e e r i n g r e s e a r c h a n d d e s i g n 8 6 ( 2 0 0 8 ) 401–409
Fig. 7 – Effect of packing size on relative efficiency of
absorption (gas flow rate = 0.01 N m3 /s, liquid
flowrate = 0.06 kg/s).
absorption per bed volume is increased for both components,
but it is more pronounced for CO2 which is the weaker com-
petitor for being absorbed by MEA solution. This has been
shown in Fig. 7.
The effect of packing size on the process performance,
strongly depends on the states of other conditions. Therefore,
this parameter was studied in variety of operating condi-
tions. Typical results of “two side statistical t-test” have been
reported in Table 1 as a test of significance for this parameter.
The test based on the calculated and experimental data, verify
that packing size parameter has a significant effect on relative
efficiency of simultaneous absorption.
This could be followed through the experimental infor-
mation of this research, which has been reported in Table 1.
The test shows that, there is a highly significant difference
between the means (relative efficiencies data) for these two
sets of experiments (0.5, 1 in.). This behavior is similar to the
effect of L/G. In this manner when large amount of amine are
facing the special quantity of acid gas in the unit volume of
packed bed (lower packing size and higher L/G), due to the
strength of MEA, it absorbs nearly all H2 S and CO2 and reduces
the selectivity.
Although the results show that using smaller packing size
decreases the absorption selectivity of H2 S, but it should be
noted that decreasing of packing size causes increasing of spe-
cific surface and reduces the required volume of the contactor
for a specific dusty. Therefore, there is an optimum packing
size. This leads to the design of experimentations to obtain
the optimum value of packing size to satisfy the appropriate
Table 1 – The t-test of significance on experimental data
for verification of the effect of packing size
Specimen Test 1, X1 Test 1, X2 (packing
(packing 2.5 cm) 1.125 cm)
1 (p = 1 bar) 1.282 1.24
2 (p = 3 bar) 1.303 1.265
3 (p = 4.5 bar) 1.316 1.275
4 (p = 5 bar) 1.319 1.278
5 (p = 7.5 bar) 1.323 1.284
Total 6.543 6.342
Mean 1.3086 1.2684
n1 = n2 = 5, ˚0 = 8, Ss = 0.017, SE = .0107, (˛ = 0.05, t = 2.31), (˛ = 0.01,
t = 3.36), 2 − 1 = (mean test 2) − (mean test 1) ± 3.36 (0.0107). There
is significant difference between two means.
Fig. 8 – Effect of MEA concentration on relative efficiency of
absorption (gas flow rate = 0.01 N m3 /s, liquid
flowrate = 0.06 kg/s, bottom pressure = 7.65 bar, top
pressure = 7.5 bar).
objective function under different operating conditions. This
is the subject of next research.
3.4. Effect of amine concentration
Another parameter, which is considered in this study, is the
inlet concentration of amine solution. Generally speaking,
increasing of amine concentration, increases the driving force
for absorption of both H2 S and CO2 and so their absolute
removal efficiencies. But since the rate of CO2 absorption is
controlled by a chemical reaction in the liquid phase where
amine is one of the reactant, so increasing of amine concen-
tration has negative effect on selectivity or relative efficiency.
This fact has been verified for different amines in other stud-
ies (Mandal et al., 2004) and is shown in Fig. 8 which has been
observed for MEA solution in this study. It should be men-
tioned that in low amine concentration, a physical mechanism
of absorption is governing and the relative efficiency is more
sensitive to amine concentration (the curve is steepest in low
amine concentration). In higher amount of amine concen-
tration due to the availability of amines for both of H2 S and
CO2 the selectivity is reduced and there is no real competition
between H2 S and CO2 for being absorbed in to MEA solution.
In this research the observed trends and qualitative behav-
ior is comparable with other researches such as Zheng et al.
(2006) and Mandal et al. (2004). The conceptual comparison
confirms the observed results in this research.
4. Conclusions
Simultaneous absorption of H2 S and CO2 from gaseous
streams into an aqueous MEA solution using a packed bed
d = (X2 − X1 ) pilot-scale contactor has been studied theoretically and exper-
imentally. As a first conclusion, the packing size affects the
relative absorption (selectivity) of CO2 and H2 S in MEA solu-
0.042
0.038
tion as a new proposed structural parameter. Smaller packing
0.041 size leads to decrease in selectivity of H2 S absorption. The
0.041 results also show that, increasing of total pressure, which
0.039 helps the diffusion steps in absorption, increases the H2 S
0.2 absorption selectivity and rate. It also could be seen that
0.04 increasing L/G as a potential for liquid phase events, decreases
the H2 S absorption selectivity while increases the absolute
rate of absorption. It is the same, when amine concentra-
tion is increased. Increasing the amine concentration, L/G and
decreasing the packing size conceptually means the increase
 c h e m i c a l e n g i n e e r i n g r e s e a r c h a n d d e s i g n 8 6 ( 2 0 0 8 ) 401–409
in absorption potential per bed volume which decreases the
selectivity.
In general the proposed model is able to predict the effects
of different parameters. Based on the experimental and calcu-
lated data, the model predicts the effect of packing size within
7%, the effect of pressure on absolute efficiency with max-
imum 14% and relative efficiency within range of 6% error.
In the range of operating conditions which has been studied
here, the minimum errors occurred in prediction of effect of
L/G and the maximum error occurred in prediction of the effect
of pressure.
Based on the system behaviors, considerable improvement
due to erecting the proposed parameters could be expected. It
could be concluded that by decreasing the L/G in its observed
range, the relative efficiency is increased near 20% while abso-
lute efficiency is reduced. It also could be concluded using
packing size 1 in. instead of 0.5 in.; the relative efficiency is
increased near 5% while absolute efficiency is reduced.
It should be noted that all the above findings are due to
the fact that in simultaneous absorption o H2 S and CO2 , CO2
is a relatively weak competitor for H2 S for being absorbed in
MEA solution. By optimal preparing the conditions in such a
way that weakness of CO2 is more pronounced, selectivity of
absorption could be increased. So in the cases such as SCOT
process in which H2 S absorption selectivity is the major goal of
the process (while larger amount of weaker component (CO2 )
are exist), suitable procedure could be developed based on
these results, as a guidelines for design of absorption tower
in the industrial plants.
Acknowledgements
The authors would like to acknowledge Shiraz oil Refinery
Company for the financial and technical supports. Special
thanks to Mr. Shadravan, Mr. Ghanadi and Mr. Kar-andish for
their assistance in running the experiments.
references
Ali, S.H., Merchant, S.Q. and Fahim, M.A., 2002, Reaction kinetics
of some secondary alkanolamines with carbon dioxide in
aqueous solutions by stopped flow technique. Sep Purif
Technol, 27: 121–136.
409
Brettschneider, O., Wozny, G., et al., 2004, Experimental
investigation and simulation of the chemical absorption in a
packed column for the system NH3 –CO2 –H2 S–NaOH–H2 O. Sep
Purif Technol, 39: 139–159.
Gabrielsen, J., et al., 2007, Experimental validation of a rate-based
model for CO2 capture using an AMP solution. Chem Eng Sci,
62: 2397–2413.
Glasscock, D.A. and Rochelle, G.T., 1993, Approximate simulation
of carbon dioxide and hydrogen sulfide absorption into
aqueous alkanol amine solutions. AIChE J, 39(8): 1389–
1397.
Kent, R.L. and Eisenberg, B., 1976, Better data for amine
sweetening. Hydrocarbon Process, 87: 87–92.
Lyddon, L. and Nguyen, H., 1999, Analysis of various flow
schemes for sweetening with amines, In Proceedings of the
Seventy-Eighth GPA Annual Convention (Gas Processors
Association, Nashville, TN), pp. 177–184.
Mandal, B.P. and Bandyopadhyay, S.S., 2005, Simultaneous
absorption of carbon dioxide and hydrogen sulfide into
aqueous blends of 2-amino-2-methyl-1-propanol and
diethanolamine. Chem Eng Sci, 60: 6438–6451.
Mandal, B.P., Biswas, A.K. and Bandyopadhyay, S.S., 2004,
Selective absorption of H2 S from gas stream containing H2 S
and CO2 into aqueous solutions of N-methyl diethanolamine
and 2-amino-2-methyl-1-propanol. Sep Purif Technol, 35:
191–202.
Pant, K.K. and Srivastava, V.K., 2007, Carbon dioxide absorption
into monoethanolamine in a continuous film contactor. Chem
Eng J, 133: 229–237.
Perry, R.H. and Green, D., (1984). Perry’s Chemical Engineers’
Handbook (6th ed.). (McGraw-Hill, New York).
Rascol, E., Meyer, M., Hour, M.H. and Prevost, M., 1997, Modeling
and simulation of the absorption of H2 S and CO2 into mixed
alkanol amine solutions. Hung J Ind Chem, 25: 11–16.
Riesenfield, F.C., (1998). Gas Purification. (McGraw-Hill, New York).
Saha, A.K., Biswas, A.K. and Bandyopadhyay, S.S., 1999,
Absorption of CO2 in a sterically hindered amine: modeling
absorption in a mechanically agitated contactor. Sep Purif
Technol, 15: 101–112.
Treybal, R.E., (1980). Mass Transfer Operations. (McGraw-Hill, New
York).
Yasavage, V.F., Al-Bagheli, N.A., Pruess, S.A. and Selim, M.S., 2001,
A rate based model for the design of gas absorbers for the
removal CO2 and H2 S using aqueous solutions of MEA and
DEA. Fluid Phase Equilib, 185: 31–43.
Zheng, Y., et al., 2006, Selective absorption of H2 S from gas
mixtures into aqueous solutions of blended amines of
methyl-di-ethanol-amine and
2-tertiarybutylamino-2-ethoxyethanol in a packed column.
Sep Purif Technol, 52: 209–217.