Sei sulla pagina 1di 9

International Journal of Greenhouse Gas Control 11 (2012) 6472

Contents lists available at SciVerse ScienceDirect

International Journal of Greenhouse Gas Control


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

Carbon dioxide absorption characteristics of aqueous amino acid salt solutions


Ho-Jun Song a , Sangwon Park a , Hyuntae Kim a , Ankur Gaur a , Jin-Won Park a, , Seung-Jong Lee b
a
b

Department of Chemical and Biomolecular Engineering, Yonsei University, 262 Seongsanno, Seodaemun-gu, Seoul 120-749, South Korea
Plant Engineering Center, Institute for Advanced Engineering, Ajou University, Suwon 443-749, South Korea

a r t i c l e

i n f o

Article history:
Received 28 February 2011
Received in revised form 11 July 2012
Accepted 14 July 2012
Available online 24 August 2012
Keywords:
Carbon dioxide (CO2 )
Alkanolamine
Amino acid salt
Absorption
Desorption

a b s t r a c t
In the present study, the aqueous potassium salts of 16 common amino acids and some blends with
piperazine (PZ) were experimentally screened. Critical solution concentration and surface tension were
measured to assess the aqueous solutions as CO2 absorbents for membrane-gas absorption (MGA). Cyclic
CO2 absorption and desorption were conducted on the prepared absorbent solution, at 40 C with a
continuous feed of 15 kPa CO2 (N2 balance) and at 80 C with a N2 feed, respectively. The cyclic CO2
absorption performances were evaluated by computing and comparing net cyclic capacity as well as
initial rates of absorption and desorption. From the experimental results, it was found that (i) longer
distance between amino group-CO2 -binding site-and other functional group and (ii) bulkier functional
group of amino acid, would result in higher net cyclic capacity as well as faster CO2 desorption. Of the
16 amino acids, the alanine (ALA), serine (SER), and -aminobutyric acid (AABA) salts had relatively fast
initial rates of absorption and desorption, resulting in high net cyclic capacity. A small amount of PZ added
as a rate promoter to the ALA, SER, and AABA salts increased net cyclic capacity by more than 25%. This
result indicates that such salts could be energy-efcient alternates for monoethanolamine (MEA). Due
to their high surface tension values, thus lower membrane pore wettability, ALA and SER salt solutions
with PZ addition could be utilized as CO2 absorbing liquids in membrane contactors.
2012 Elsevier Ltd. All rights reserved.

1. Introduction
such
as
monoethanolamine
(MEA),
Alkanolamines
diethanolamine (DEA), N-methyldiethanolamine (MDEA), and
2-amino-2-methyl-1-propanol (AMP) are the most frequently
utilized absorbents in chemical absorption used to capture carbon
dioxide from various ue gases (Aroonwilas and Veawab, 2007;
Ismael et al., 2009). MEA is widely used in various processes
due to its fast reaction kinetics with CO2 , low solvent cost, low
hydrocarbon solubility, and other advantageous characteristics.
However, MEA introduces the disadvantages of O2 - and SO2 induced degradation, metal-corrosion, high vapor loss, and high
absorbent regeneration energy (4.0 GJ/ton CO2 ) (Sakwattanapong
et al., 2005; Soosaiprakasam and Veawab, 2008; Supap et al.,
2009). MDEA has a higher CO2 loading capacity than MEA at high
CO2 partial pressure. It also has low absorbent regeneration energy
and fast reaction kinetics with H2 S. However, MDEA reacts slowly
with CO2 , restricting its wide application (Rinker et al., 1995). AMP
is a sterically hindered primary amine with relatively fast reaction
kinetics, low regeneration energy, low metal corrosivity, and a
greater CO2 loading capacity than other primary amines due to

Corresponding author. Tel.: +82 2364 1807; fax: +82 2312 6401.
E-mail address: jwpark@yonsei.ac.kr (J.-W. Park).
1750-5836/$ see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijggc.2012.07.019

unstable carbamate formation (Xu et al., 1996; Sakwattanapong


et al., 2005; Sartori and Savage, 1983; Veawab et al., 1997).
Amino acid salts have advantages over conventional alkanolamine. An amino acid salt has low volatility due to its ionic
structure and may resist oxidative degradation (Hook, 1997). The
reactivities of amino acid salts are similar to those of alkanolamines
due to the presences of identical amino functional groups in their
molecules. Some amino acid salts, particularly the potassium salts
of glycine, sarcosine, and proline, have faster reaction kinetics with
CO2 than does MEA (Van Holst et al., 2009). The high surface tension of aqueous amino acid salts makes them suitable for use in
membrane-gas absorption (MGA) systems as sour gas absorbing
liquids. Kumar et al. (2002) and Yan et al. (2007) revealed that
the high surface tensions of aqueous taurine and glycine potassium
salts can lower the pore wettabilities of hollow ber membranes
(i.e., increase the breakthrough pressure of the MGA system). This
resists the mass transfer of CO2 from outer ue gas into the inner
absorbing liquid bulk. Recently, Van Holst et al. (2009) conducted
screening tests of eight common amino acid salts by plotting the
et al. (2009) synoverall reaction rate constant versus pKa . Munoz
thesized piperazine ring-containing amino acids and measured CO2
solubilities in the aqueous solutions of common and synthesized
amino acid salts. Although amino acid salts confer many advantages, only a few investigations have screened various amino acid
salts for CO2 capture.

H.-J. Song et al. / International Journal of Greenhouse Gas Control 11 (2012) 6472

In the present study, the CO2 absorption performances of the


potassium salts of 16 common amino acids were assessed using
cyclic absorption and desorption experiments. Suggested amino
acids were categorized into four groups: linear amino acids (glycine
(GLY), taurine (TAU), -alanine (BALA), and -aminobutyric acid
(GABA)); sterically hindered amino acids (alanine (ALA), aminobutyric acid (AABA), -methyl alanine (AMALA), serine (SER),
and cysteine (CYS)); cyclic amino acids (proline (PRO), 4-hydroxy
proline (HYPRO), and pyroglutamic acid (PGA)); and poly amino
acids (asparagine (ASN), glutamine (GLN), diglycine (DIGLY), and
arginine (ARG)). The molecular structures and CAS number of these
16 amino acids and reference amines (monoethanolamine, piperazine) are shown in Table 1.
The 16 amino acids were suggested based on the following
scheme. The CO2 absorption and desorption values of amines were
predicted based on their molecular structures. As depicted in Fig. 1,
the intermediate was formed via nucleophilic attack of the amine
on CO2 . Intermediate 2 is a reaction product derived from the
sterically hindered- or bulky side group-substituted amine and
CO2 . Intermediate 2 undergoes strong steric repulsion between
substituents (R) and the two oxygen atoms of bound CO2 . Intermediate 2 is sterically unstable and the bound CO2 can easily desorb,
resulting in energy savings for CO2 desorption of CO2 -saturated
absorbents. For amino acids, R is either the substituent or its carboxylic group. Sterically unhindered primary amines such as MEA
and glycine salt form Intermediate 1, which is stable due to weaker
repelling power of the hydrogen atom compared to that of Intermediate 2 substituents (R). More thermal energy is needed to cleave
the NC bond and high absorbent regeneration energy is required. A
molecules steric hindrance increases its CO2 removal capacity due
to ease of CO2 desorption. Steric hindrance also decreases the accessibility of CO2 to the amino group, decreasing the initial absorption
rate. The initially slow absorption of a sterically hindered amine
can be enhanced by adding a rate promoter such as piperazine (PZ).
Solubility in pure water, glass transition temperature, number and
position of amino groups, and the number and position of methyl
groups were additional properties considered for candidate amino
acids.
After the rst cyclic CO2 absorption and desorption screening
test of the 16 amino acid salts, three amino acid salts were selected
for additional experiments. The amino acid salts were blended with
a small amount of the rate promoter piperazine (Bougie et al., 2009;
Oexmann et al., 2008) and identical screening tests were conducted.
It was previously reported that adding a small amount of PZ to
pure water would only slightly decrease surface tension (Derks
et al., 2005). The critical concentrations for precipitation during
CO2 absorption were also experimentally studied for the screened
amino acid salts. The precipitate could pose operational problems,
such as fouling and plugging the equipment or decreasing liquid
phase gas transfer (Kumar et al., 2003). The surface tensions of some
screened amino acid salt solutions were measured to explore their
applicability to the MGA system.

2. Experimental
2.1. Materials
All of the absorbent chemicals used in this study, including MEA
and PZ, were purchased from Alfa-Aesar at purities greater than
98%. The amino group of amino acids in the aqueous phase loses its
electro-neutrality, resulting in the loss of binding power with CO2 .
KOH (8 kmol/m3 standard aqueous solution, Acros, CAS No.: 131058-3) was slowly added to the solutions to recover amino group
neutrality. The molar concentrations of prepared aqueous amino
acid salt solutions were checked by titration with 0.1 kmol/m3

65

H2 SO4 (Duksan Chemical Co., CAS No.: 7664-93-9). The measured


molar concentrations were accurate to 0.5%.
2.2. Experimental set-up and procedure
An experimental apparatus was equipped to assess the rates and
capacity of each amino acid salt for CO2 absorption and desorption
(Fig. 2). The apparatus and procedure are similar to those used by
Aronu et al. (2010).
The simulated feed gas was prepared by diluting CO2 with N2 .
The desired CO2 volumetric concentration was attained in a saturator placed just before the reactor by using a mass ow controller
(TSC-120, MKP) for each gas ow rate. The designated CO2 concentration (15 kPa) simulated ue gas from a conventional coal-red
power plant. The actual CO2 ow rate was continuously monitored
by a Bronkhurst High-Tech El-Flow mass ow meter. CO2 concentrations in the simulated feed gas were veried before each run
by bypassing the mixed feed gas through the IR CO2 analyzer (Gas
Master, Woori System), which had an analyzing span of 025% CO2
and a resolution of 0.01% CO2 . CO2 absorption was then initiated
by feeding and bubbling simulated gas into the aqueous absorbent
solution, where the CO2 ow rate was maintained at 500 ml/min.
The volume of the absorbent solution was 400 ml. The reactor
was made of Pyrex glass, and its diameter and height were 7.5 cm
and 16 cm, respectively. To assess the absorption performance of
each amino acid salt at identical concentrations, each absorbent
solution concentration was 1.0 kmol/m3 . Relatively low molarity
(1.0 kmol/m3 ) was chosen to prevent precipitate formation during CO2 absorption. The feed gas was spread to the absorbent
solution bulk via bubbling through a 1/4 in. stainless tube without a sintered glass tube to prevent plugging the gas injection
tube by precipitate formation during absorption. The temperature was maintained within 0.1 C of 40 C using a thermostatic
water circulator. Cooling water (4 C) was circulated in a condenser
attached to the reactor top to inhibit water vapor loss and to maintain the solution concentration. For all salts except the potassium
salt of ARG, the outlet CO2 concentration reached that of the feed
gas (15 kPa) after 2 h of absorption. This indicates that the aqueous absorbent solutions were saturated with CO2 . Simulated gas
ow was then stopped and a high purity inert gas (99.999% N2 ,
Gas Valley) was purged through the whole system in a short time
(<3 min) to remove trace CO2 gas. The solution-containing reactor was then quickly immersed in a hot thermostating water bath
(80 C) to thermally desorb the absorbed CO2 from the absorbent
solution. Although CO2 stripping in the actual amine process
occurred at around 120 C, a lower stripping temperature (80 C)
was shown to be efcient in screening CO2 absorbents in a short
time (Singh and Versteeg, 2008; Aronu et al., 2010). The outlet
CO2 concentration was measured and recorded in 30 s time intervals using an IR CO2 analyzer for the duration of the cyclic process
(240 min).
A number of amino acid salts were selected for further characterization to compare cyclic performances. A simple apparatus
similar to that described by Majchrowicz et al. (2009) was equipped
to experimentally study precipitate formation during reactions
between CO2 and the aqueous potassium salts of amino acids.
Nine glass reactors were installed in a transparent thermostating air bath. The reactors were connected in parallel with exible
Tygon tube so the simulated gas could simultaneously feed to
the whole reactor. Absorbent solutions were prepared at specic solution concentrations. At intervals of 0.5 kmol/m3 , 60 ml
of aqueous amino acid salts were added to a 100 ml glass reactor. The equipment and procedure used for preparing the 15 kPa
CO2 feed gas were nearly identical to those used in the cyclic
absorption and desorption operations. The CO2 volumetric ow
rate was maintained at 200 ml/min using a mass ow controller.

66

H.-J. Song et al. / International Journal of Greenhouse Gas Control 11 (2012) 6472

Table 1
Molecular structures and CAS numbers of suggested amino acids and amines.

HN

OH
Monoethanolamine
(MEA), 141-43-5

H 2N

Piperazine (PZ), 110-85-0

NH

O
Glycine (GLY), 56-40-6

H2N

OH
S

H2N

OH

Taurine (TAU), 107-35-7

NH2

-Alanine (BALA),
107-95-9

OH

-Aminobutyric acid
(GABA), 56-12-2

NH2
OH

Alanine (ALA),
302-72-7

OH

-Aminobutyric acid
(AABA), 2835-81-6

OH

NH2

NH2
O

NH2
OH
-Methyl alanine
(AMALA), 62-57-7

Serine (SER), 302-84-1

OH

HO

NH2

OH
HS

Cysteine (CYS), 52-90-4

OH

Proline (PRO), 609-36-9

N
H
O

NH2
HO

OH

OH

4-Hydroxy proline
(HYPRO), 51-35-4

N
H

Pyroglutamic acid (PGA),


149-87-1

N
H
O

NH2

H2N

Asparagine (ASN),
70-47-3

OH

OH

NH2

Glutamine (GLN), 56-85-9

O
NH2

OH

NH

H
N
H2N

O
O

Diglycine (DIGLY),
556-50-3

H2N

N
H

OH
NH2

Arginine (ARG), 7200-25-1

H.-J. Song et al. / International Journal of Greenhouse Gas Control 11 (2012) 6472

67

Fig. 1. Simple scheme for predicting CO2 absorption and desorption performances of amines.

The reaction temperature was maintained at 40 1 C in a thermostating air bath. Experiments began by simultaneously owing
and bubbling the feed gas into nine absorbent solutions. The procedure was terminated when the outlet CO2 concentration was
the same as that of the feed gas. To precipitate ne particles,
samples were aged for 12 h with no gas feed while maintaining the temperature at 40 C. The rst absorbent concentration
was prepared at 1.5 kmol/m3 because no precipitate was visually
detected in the 1.0 kmol/m3 experiment, as demonstrated by an
earlier cyclic experiment. Solution concentrations were gradually
increased and the experiments were repeated until precipitation
occurred.
The surface tension values of the selected 2.5 kmol/m3 aqueous amino acid salt solutions and the 0.25 kmol/m3 aqueous
PZ-promoted solutions of some amino acid salts were directly
measured at 25 C using a commercial bubble pressure meter
(BP2, Krss). The measuring times were greater than 50,000 ms
to obtain the exact equilibrium surface tension value. A relatively
high solution concentration (2.5 M) was adopted to observe
the direct or inverse proportion of concentration to surface
tension.

3. Results and discussion


3.1. Cyclic CO2 absorption and desorption
In the present study, three major criteria were computed for
performance comparison: initial absorption rate, initial desorption
rate, and net cyclic capacity. Many researchers have adopted these
metrics as parameters for selecting a novel efcient CO2 absorbent,
and the method has been shown to be effective and rapid (Mamun
et al., 2007; Singh et al., 2010). The faster the absorption rate,
the smaller the required absorber size in the real absorption process. The faster the desorption rate, the lower the thermal energy
required to regenerate the saturated absorbent solution. The larger
the net cyclic capacity, the lower the required L/G ratio, eventually
resulting in operational cost savings.
For quantitative analysis, CO2 loading () was dened as moles
of CO2 absorbed in a unit mole of amine. In this study, the concentration of the aqueous solutions is 1.0 kmol/m3 , so that the CO2
loading values are identical whether the unit is mol/mol or mol/L.
To express the reactivity of CO2 absorbent itself, the mole-based
unit, that is, unit-moles of CO2 per moles of absorbent would be

Fig. 2. Schematic diagram of experimental set up for cyclic CO2 absorption and desorption operations.

68

H.-J. Song et al. / International Journal of Greenhouse Gas Control 11 (2012) 6472

1.2

CO2 loading (mol CO2 / mol amine)

Absorption

Desorption
1.0 kmol/m3 MEA
1.0 kmol/m3 PZ
1.0 kmol/m3 KGLY

1.0

0.8

0.6

0.4

0.2
The slopes to determine the rate
0.0
0

60

120

180

240

Operating time (min)


Fig. 3. Example of CO2 loading curve for cyclic absorption and desorption.

more suitable. Moles of CO2 absorbed were calculated as the difference in CO2 concentration between feed gas and outlet gas. CO2
loading at a given time (t) was computed by the following equation:
t =

respectively. These results indicate that GLY salt had CO2


absorption performance similar or slightly superior to that of
MEA. The initial CO2 desorption rate and net cyclic capacity of GLY salt were 1.84 E02 mol CO2 /(mol amine min) and
0.465 mol/mol, respectively. The comparable values for MEA were
2.00 E02 mol CO2 /(mol amine min) and 0.483 mol/mol, respectively. These results indicate that GLY salt had worse CO2
desorption performance than MEA. Although the initial CO2 absorption rate for TAU salt was less than that of GLY salt, the initial
CO2 desorption rate for TAU salt was superior to that of GLY salt.
This resulted in a greater net cyclic capacity of 0.483 mol/mol for
TAU salt, identical to that of MEA. TAU salt is thought to have a
faster initial desorption rate than GLY salt, resulting in higher net
cyclic capacity and a slower initial absorption rate. This is probably
due to the bulkiness of the sulfonic groups (SO3 H) compared the
carboxylic group (CO2 H) of GLY.
The relationship between chain length (or distance between
amino- and carboxyl-groups) and performance may be examined
by comparing the results of linear amino acid salts. The salts
of GLY, BALA, and GABA have similar molecular structures, with
one amino group and one carboxylic group. There are 1, 2, and
3 carbons between the two functional groups of these amino
acids, respectively. Generally, the closer the amino and carboxylic
groups, the stronger the steric repulsion between the CO2 -bound

Moles of CO2 absorbed at t [mol]


Moles of amine in solution [mol]

(CO2 ow rate of feed gas [cm3 / min]) (CO2 molar volume [mol/cm3 ])



t (Inlet CO2 conc.Outlet CO2 conc.)


dt
Inlet CO2 conc.
0


[min]

Moles of amine in solution [mol]


(1)

Table 2 presents the experimental results of the absorbents tested


with a continuous 15 kPa CO2 feed. An example of the computed
rates and net cyclic capacity are shown in Fig. 3. The initial absorption rate was the slope of the CO2 loading curve in a time interval of
05 min during CO2 absorption (see the dashed tangent line on the
PZ curve). Hence, rate units were mol CO2 /(mol amine min). In the
initial ve minutes of CO2 desorption, it was difcult to compute
the linear slope while heating the absorbent solution to the desorption temperature (80 C). Therefore, the initial desorption rate
was the absolute value of the slope in the 510 min time interval
of CO2 desorption, which can be used to calculate linear slope. Rich
CO2 loading (rich ) was dened as the maximum CO2 loading after
the absorption operation, i.e., the CO2 loading at 120 min. Lean CO2
loading (lean ) was determined at 60 min of the desorption operation instead of at 120 min. The CO2 loadings of some amino acid
salts after 60 min of CO2 desorption were close to zero, eliminating the driving force to desorb CO2 and resulting in difculty with
quantitative comparison.
Table 3 presents values of CO2 solubility in the aqueous solutions of MEA, PZ, and the potassium salt of GLY drawn from the
literature under similar experimental conditions. The data were
collected at 40 C for a 1.0 kmol/m3 solution under CO2 partial pressure of approximately 15 kPa. As shown in Tables 2 and 3, the rich
CO2 loadings of PZ and GLY salt in this study corroborated the literature values. Data for 1.0 kmol/m3 MEA was scarce and comparison
to literature results was not straightforward. However, the experimental results for 1.0 kmol/m3 MEA in this study may be reasonable
because there is a tendency for higher CO2 partial pressures and
lower solution concentrations to result in higher CO2 loadings.
3.1.1. Linear amino acids (GLY, TAU, BALA, GABA)
As shown in Table 2, the initial CO2 absorption rate and rich
CO2 loading for GLY salt were 3.93 E02 mol CO2 /(mol amine min)
and 0.738 mol/mol, respectively. The same values for MEA
were 3.84 E02 mol CO2 /(mol amine min) and, 0.736 mol/mol,

amino group and carboxyl group, the slower the initial absorption rate (BALA < GABA) and the faster the initial desorption rate
(GLY > BALA > GABA). All salts of the non-hindered linear amino
acids in this study had one amino group and attained rich CO2 loading at 0.70.75 mol/mol. TAU salt was the exception because it has
a bulky sulfonic group that results in low accessibility of CO2 to the
amino group.

3.1.2. Sterically hindered amino acids (ALA, AABA, AMALA, SER,


CYS)
Some sterically hindered amino acids were tested, using GLY as a
standard non-hindered amino acid for comparison. ALA, AABA, and
AMALA molecules were in the form of -methylated, -ethylated,
and two -methylated GLY, respectively. Unexpectedly, the bulkier
-ethylated (AABA) salt had a faster initial absorption rate and a
slower initial desorption rate than the -methylated (ALA) salt.
Although the AABA salt had a lower initial desorption rate than
the ALA salt, its rich CO2 loading was greater. The net cyclic capacity of the AABA salt was equivalent to that of the ALA salt. AMALA
is the substituted form of glycine with two methyl groups at the
-carbon. The structural relationship between AMALA and GLY
is similar to that of two methylated AMPs and non-substituted
MEA. A slow absorption rate was expected due to the steric hindrance of AMALA. However, the two methyl groups bound to the
-carbon fully converted carbamate to bicarbonate. This resulted in
free amine release and higher CO2 loading (Hook, 1997). In all cases,
CO2 absorption was terminated at 120 min. However, AMALA salt
did not obtain completely saturated CO2 loading due to lasting and
vigorous CO2 absorption. The initial desorption rate of AMALA salt
was very slow for reasons we were unable to determine. Although
the CO2 desorption capacity of AMALA salt was very high in 2 h
( = 0.7 mol/mol), its slower initial desorption rate is a drawback
for the utilization of AMALA salt in CO2 capture processes. The real

H.-J. Song et al. / International Journal of Greenhouse Gas Control 11 (2012) 6472

69

Table 2
Experimental results for cyclic CO2 absorption and desorption operations of aqueous solutions of reference amines and amino acid salts.
Absorbent

Concentration
(kmol/m3 )

Initial absorption rate


(mol CO2 /(mol amine min))

Reference amines
3.84 E02
MEA
1.0
PZ
1.0
4.58 E02
Salts of amino acids
Linear amino acids
1.0
3.93 E02
GLY
3.17 E02
TAU
1.0
BALA
1.0
3.65 E02
GABA
1.0
3.87 E02
Sterically hindered amino acids
1.0
3.16 E02
ALA
1.0
3.76 E02
AABA
2.43 E02
AMALA
1.0
SER
1.0
3.16 E02
3.18 E02
CYS
1.0
Cyclic amino acids
4.26 E02
PRO
1.0
1.0
4.03 E02
HYPRO
2.41 E02
PGA
1.0
Poly amino acids
1.0
2.79 E02
ASN
1.0
3.06 E02
GLN
1.0
2.99 E02
DIGLY
3.56 E02
ARG
1.0
Addition of PZ to selected sterically hindered amines
1.0 + 0.1
3.44 E02
ALA + PZ
AABA + PZ
1.0 + 0.1
3.70 E02
SER + PZ
1.0 + 0.1
3.47 E02

Initial desorption rate


(mol CO2 /(mol amine min))

Rich CO2 loading


(mol CO2 /mol amine)

Lean CO2 loading


(mol CO2 /mol amine)

Net cyclic capacity


(mol CO2 /mol amine)

2.00 E02
2.74 E02

0.736
1.022

0.253
0.317

0.483
0.705

1.84 E02
2.26 E02
1.80 E02
1.54 E02

0.738
0.573
0.721
0.749

0.273
0.090
0.252
0.309

0.465
0.483
0.469
0.440

1.98 E02
1.87 E02
1.20 E02
2.26 E02
2.46 E02

0.670
0.728
0.750
0.619
0.485

0.135
0.188
0.234
0.061
0.000

0.535
0.540
0.516
0.558
0.485

1.48 E02
1.55 E02
0.38 E02

0.746
0.655
0.224

0.334
0.261
0.155

0.412
0.394
0.069

2.34 E02
2.08 E02
2.18 E02
2.19 E02

0.573
0.600
0.510
1.107

0.026
0.065
0.043
0.540

0.547
0.535
0.467
0.567

2.33 E02
2.23 E02
2.58 E02

0.734
0.797
0.723

0.128
0.188
0.092

0.606
0.609
0.631

process used to regenerate absorbent involves short residence time


of the absorbent solution in the stripper.
SER and CYS have molecular structures similar to that of AABA,
with the addition of a hydroxyl group (OH) and a sulfhydryl group
(SH), respectively. Generally, for AABA, SER, and CYS, the bulkier
the substituent group that exhibited stronger steric repulsion, the
slower the initial absorption rate (AABA > SER, CYS) and the faster
the initial desorption rate (AABA < SER < CYS). The net cyclic capacity of SER salt was greater than that of TAU salt, though the two had
comparable initial desorption rates. This is because the rich loading
of SER salt was greater than that of TAU salt. Although AABA and SER
have similar molecular structures and molecular weights, large differences were observed in their initial rates of CO2 absorption and
desorption.

less rich CO2 loading than PRO salt. The desorption driving force
may have decreased due to lower rich CO2 loading of HYPRO salt.
For PGA salt, absorption and desorption performances may severely
deteriorate due to the existence of a carbonyl oxygen adjacent to
the amino group. It is thought that PGA salt is not able to absorb CO2 ,
because valence electrons around the amino group, which conduct
nucleophilic attack on CO2 were almost delocalized to the carbonyl
oxygen.
3.1.4. Poly amino acids (ASN, GLN, DIGLY, ARG)
Although ASN, GLN, and DIGLY are di-amino acids, their salts
have lower rich CO2 loadings (0.50.6 mol/mol) than some salts
of mono-amino acids such as GLY, BALA, GABA, AABA, and PRO
(rich = 0.70.75 mol/mol). As with PGA salt, the strong tendency
for electron delocalization by the carbonyl oxygen may completely
incapacitate the adjacent amino group or affect both amino groups
of the di-amino acid. Apart from electron delocalization, steric
hindrance effects of the bulky carbonyl oxygen in di-amino acids
may be demonstrated in the results for ASN and GLN salts. The
carbonyl oxygen in ASN applies more steric repulsion to the CO2 bound amino group near the carboxyl group than that of GLN does.
This is due to the shorter carbon chain length of ASN and results
in a higher initial desorption rate (ASN > GLN) and a lower initial absorption rate (ASN < GLN). For di-amino acids, the presence
of carbonyl oxygen caused higher desorption performances and

3.1.3. Cyclic amino acids (PRO, HYPRO, PGA)


HYPRO and PGA are substituted forms of PRO with a substituted hydroxyl group and carbonyl group, respectively. As shown
in Table 2, PRO salt had the highest initial absorption rate among
tested amino acid salts. This is because the protruding PRO amino
group led to easy CO2 access to the amino group. Alternatively,
PRO salt requires more energy to desorb CO2 due to lower steric
repulsion of the reaction intermediate. The HYPRO hydroxyl group
exhibits steric repulsion, resulting in a higher initial desorption
rate. However, the HYPRO salt has a lower net cyclic capacity and
Table 3
Solubilities of CO2 in aqueous solutions of absorbents.
Absorbent

Concentration
(kmol/m3 )

Temperature
( C)

CO2 partial
pressure (kPa)

CO2 loading
(mol CO2 /mol absorbent)

Reference

MEA
MEA
MEA
PZ
PZ
Potassium glycinate
Potassium glycinate

1.0
2.5
2.5
1.0
1.0
1.0
1.0

40
40
40
40
40
40
40

9.5
19.9
22.5
9.0
18.1
8.0
30.5

0.593
0.562
0.602
1.046
1.058
0.721
0.846

Aronu et al. (2011)


Lee et al. (1974)
Jones et al. (1959)
Aroua and Salleh (2004)
Aroua and Salleh (2004)
Portugal et al. (2009)
Portugal et al. (2009)

H.-J. Song et al. / International Journal of Greenhouse Gas Control 11 (2012) 6472

0.050
Linear amino acid
Sterically hindered amino acid
Cyclic amino acid
PRO
Poly amino acid
HYPRO
Trend
GLY

0.045

0.040

PZ

MEA
AABA

GABA
BALA

ARG

0.035
CYS ALA SER
TAU

0.030

GLN
DIGLY
ASN
PGA

0.025

AMALA

0.020
0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Initial desorption rate (mol CO2 / (mol amine min))

Initial absorption rate (mol CO 2 / (mol amine min))

70

0.030
Linear amino acid
Sterically hindered amino acid
CYS
Cyclic amino acid
ASN
SER
TAU
Poly amino acid
ARG
Trend
DIGLY GLN

0.025

0.020

GLY MEA
HYPRO

0.015
PRO

BALA

PZ

ALA
AABA

GABA
AMALA

0.010

PGA

0.005

0.000
0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Net cyclic capacity (mol CO2 / mol amine)

Net cyclic capacity (mol CO2 / mol amine)


Fig. 4. Initial absorption rate versus net cyclic capacity.

Fig. 5. Initial desorption rate versus net cyclic capacity.

signicant decreases in initial absorption rate. A similar tendency


was reported by Nagao et al. (1998), in which some -amides with
carbonyl oxygen have good desorption capacity. ARG salt showed
the highest rich CO2 loading and moderately fast initial rates of
absorption and desorption. ARG salt has two primary and two secondary amino groups compared to the two secondary amino groups
of PZ salt. The lower net cyclic capacity of ARG salt may be attributed
to strong bonds between CO2 and the two ARG primary amino
groups.

amino acids could be upgraded by adding a small amount of PZ


rate promoter to improve the absorption performances.
The experimental procedure was identical to those conducted
for the 16 amino acid salts. As shown in Table 2, an approximate
9% increase in initial absorption rate (except AABA), a 17% increase
in initial desorption rate, and a 13% increase in net cyclic capacity
were achieved by adding a small amount of PZ into the sterically
hindered amino acid salts. The initial absorption rate for AABA salt
was not enhanced. The net cyclic capacities of PZ-promoted sterically hindered amino acid salts were at least 25% greater than that of
MEA, indicating that they required a lower absorbent regeneration
energy than MEA to remove the unit quantity of CO2 from ue gas.
Furthermore, adding a small amount of PZ to the amino acid salts
increased their absorption rates, desorption rates, and capacity.

3.1.5. Performance comparisons


The initial rates of CO2 absorption and desorption and the net
cyclic capacities of 16 amino acid salts were previously discussed
with regard to steric repulsion and steric hindrance. Some general rules related amino acid salt performances to their molecular
structures. Specically, with closer carboxyl and CO2 -bound amino
groups and a bulkier substituted side group, the amino acid tended
to have better desorption performance and a slower initial absorption rate. The cyclic amino acid salts, except PGA, showed poor
desorption performances and fast initial reaction rates due to the
protruding structures of the amino groups. The di-amino acid salts
had amino group electron delocalization due to the presence of carbonyl oxygen, resulting in low rich CO2 loading. They also showed
high initial desorption rates due to steric hindrance. In some cases,
it was difcult to provide plausible reasons for unexpected experimental results, such as the poor desorption rates of AMALA and
PGA salts.
Figs. 4 and 5 plot the initial rates of CO2 absorption and desorption versus net cyclic capacity to explore the overall performances
of the suggested amino acid salts. Net cyclic capacity is a better
estimator of performance than rich CO2 loading because net cyclic
capacity is a similar parameter to the actual amount of CO2 removal
in real cyclical CO2 capture processes. The approximate order of
decreasing initial desorption rates was sterically hindered amino
acids = poly amino acids > linear amino acids > cyclic amino acids.
The order of increasing net cyclic capacity was nearly identical,
although the initial absorption rate followed the inverse order.
3.2. Promotion of some sterically hindered amino acids (ALA,
AABA, SER) via PZ addition
Salts of some sterically hindered amino acids, such as ALA, AABA,
and SER, showed moderately fast initial absorption rates and fast
initial desorption rates, resulting in greater net cyclic capacities,
as shown in Figs. 4 and 5. The salts of these sterically hindered

3.3. Critical concentration


Of 16 amino acid salts, eight were selected for further characterization based on cyclic CO2 absorption and desorption results.
Critical concentrations, the solution concentrations at which precipitation occurs, were measured for GLY, TAU, BALA, GABA, ALA,
AABA, SER, and PRO salts. In a real cyclical absorption process, the
precipitate induces operational problems such as fouling and plugging, and also increases solution viscosity, resulting in a decrease
in the liquid phase gas transfer. In the present study, aqueous solutions of amino acid salts were fully saturated with CO2 . Kumar et al.
(2003) reported that the critical CO2 loading for which crystallization occurred was related to solution concentration and solubility
of the amino acid in water. Fig. 6 depicts critical concentrations
at 40 C and the corresponding solubilities in water at 25 C (Lide,
2008). Higher solubility of an amino acid in water may produce
a greater critical concentration. The low critical concentrations of
ALA, AABA, and SER salts is a serious drawback for their application in CO2 capture processes because lower concentrations of
absorbent solution led to a higher L/G ratio and increased operation
costs.
3.4. Surface tension
In the present study, the surface tensions of 2.5 kmol/m3 aqueous solutions of some amino acid salts at 25 C were measured to
investigate their applicability with the membrane contactor. The
results are shown in Table 4. If the pores are wetted with absorbent
solution, CO2 is partially concentrated near the pore and the mass

H.-J. Song et al. / International Journal of Greenhouse Gas Control 11 (2012) 6472

Critical concentration (kmol/m3)

6
GLY

5
BALA

PRO

4
GABA

3
TAU ALA
AABA

Linear amino acid


Sterically hindered amino acid
Cyclic amino acid

2
SER

1
0

10

12

14

16

Solubility of amino acid into water (mol / kg water)


Fig. 6. Critical concentrations versus amino acid solubility.
Table 4
Surface tensions of 2.5 kmol/m3 (+ 0.25 kmol/m3 PZ) aqueous solutions of amino
acid salts at 25 C.
Amino acid salt

Surface tension (mN/m)

GLY
TAU
ALA
BALA
AABA
GABA
SER
PRO
ALA + PZ
AABA + PZ
SER + PZ

76.7
77.0
74.8
76.5
67.9
74.2
79.2
65.8
72.2
64.9
76.7

4L cos 
dmax

during the cyclic CO2 absorption and desorption experiments.


Smaller distances between amino and carboxyl groups and bulkier
substituted groups resulted in faster initial desorption rates,
increased net cyclic capacities, and slower initial absorption rates.
Adding a small amount of PZ to salts of some sterically hindered
amino acids, such as ALA, AABA, and SER, enhanced initial desorption rates and initial absorption rates. The net cyclic capacities
of PZ-promoted ALA, AABA, and SER were at least 25% greater
than that of MEA, showing that they could be used as energyefcient absorbents. Promoted ALA and SER salts have potential
for energy savings and low MGA pore wettability due to their low
surface tension. However, they precipitate at low solution concentrations during CO2 absorption. Recent work by Aronu et al.
(2010) prompted us to investigate a novel method to reduce precipitate. Our future work will include investigations on decreasing
precipitation during CO2 absorption through neutralization of the
protonated amino group with various counter ions, including metal
and organic ions. We will also study the additions of various salts
to control the salting-in and -out effect.
Acknowledgement
DSME CO., LTD and Norway Sargas supported this work. The
National Research Foundation of Korea (NRF) grant funded by the
Korean government (MEST) (No. 2011-0029161) also supported
this work.
References

transfer of sour gas would be restricted. The LaplaceYoung equation governs membrane pore wetting:
P =

71

(2)

where P is the breakthrough pressure of the microporous membrane,  L is the surface tension of the absorbent solution,  is the
contact angle between the solution and the membrane, and dmax is
the maximum diameter of the pore;  and dmax are dependent on
the properties of the polymer membrane. The higher the surface
tension of the absorbent solution, the lower the pore wettability.
Salts of GLY, TAU, ALA, BALA, GABA, and SER had greater surface
tensions than pure water (72.0 mN/m at 25 C). However, AABA
and PRO salts had weaker surface tensions than water, making
them inappropriate for low cost commercial membranes, such as
polyolen. The aqueous SER salt solution had the greatest surface
tension because the hydroxyl group of the SER molecule provided
high adhesive strength to the water molecule and adjacent SER
molecules. The addition of a small amount of PZ, with a 0.1 molar
ratio of PZ to amino acid salt, only slightly decreased the solutions
surface tension. This result indicated that PZ could be an appropriate rate promoter for the absorbing liquid adopted in a membrane
system.
4. Conclusions
In the present study, 16 common amino acid salts were
investigated with respect to cyclic CO2 absorption performance,
precipitate formation characteristics, and surface tension. A number of general rules based on molecular structure were observed

Aronu, U.E., Svendsen, H.F., Hoff, K.A., 2010. Investigation of amine amino acid salts
for carbon dioxide absorption. International Journal of Greenhouse Gas Control
4 (5), 771775.
Aronu, U.E., Hoff, K.A., Svendsen, H.F., 2011. CO2 capture solvent selection by
combined absorptiondesorption analysis. Chemical Engineering Research and
Design 89 (8), 11971203.
Aroonwilas, A., Veawab, A., 2007. Integration of CO2 capture unit using singleand blended-amines into supercritical coal-red power plants: Implications
for emission and energy management. International Journal of Greenhouse Gas
Control 1 (2), 143150.
Aroua, M.K., Mohd Salleh, R., 2004. Solubility of CO2 in aqueous piperazine and its
modeling using the KentEisenberg approach. Chemical Engineering and Technology 27 (1), 6570.
Bougie, F., Lauzon-Gauthier, J., Iliuta, M., 2009. Acceleration of the reaction
of carbon dioxide into aqueous 2-amino-2-hydroxymethyl-1,3-propanedisol
solutions by piperazine addition. Chemical Engineering Science 64 (9),
20112019.
Derks, P.W., Hogendoorn, K.J., Versteeg, G.F., 2005. Solubility of N2 O in and density, viscosity, and surface tension of aqueous piperazine solutions. Journal of
Chemical Engineering and Data 50 (6), 19471950.
Hook, R., 1997. An investigation of some sterically hindered amines as potential
carbon dioxide scrubbing compounds. Industrial and Engineering Chemistry
Research 36 (5), 17791790.
Ismael, M., Sahnoun, R., Suzuki, A., Koyama, M., Tsuboi, H., Hatakeyama, N., Endou,
A., Takaba, H., Kubo, M., Shimizu, S., Del Carpio, C.A., Miyamoto, A., 2009. A
DFT study on the carbamates formation through the absorption of CO2 by AMP.
International Journal of Greenhouse Gas Control 3 (5), 612616.
Jones, J.H., Froning, H.R., Claytor Jr., E.E., 1959. Solubility of acidic gases in aqueous monoethanolamine. Journal of Chemical and Engineering Data 4 (1),
8592.
Kumar, P.S., Hogendoorn, J.A., Feron, P.H.M., Versteeg, G.F., 2002. New absorption liquids for the removal of CO2 from dilute gas streams using membrane contactors.
Chemical Engineering Science 57 (9), 16391651.
Kumar, P.S., Hogendoorn, J.A., Feron, P.H.M., Versteeg, G.F., 2003. Equilibrium solubility of CO2 in aqueous potassium taurate solutions. Part 1. Crystallization
in carbon dioxide loaded aqueous salt solutions of amino acids. Industrial and
Engineering Chemistry Research 42 (12), 28322840.
Lee, J.I., Otto, F.D., Mather, A.E., 1974. The solubility of H2 S and CO2 in aqueous
monoethanolamine solutions. Canadian Journal of Chemical Engineering 52 (6),
803805.
Lide, D.R., 2008. Handbook of Chemistry and Physics, 88th ed. CRC Press, Boca Raton,
FL.
Majchrowicz, M.E., Brilman, D.W.F.(W.), Groeneveld, M.J., 2009. Precipitation regime
for selected amino acid salts for CO2 capture from ue gases. Energy Procedia 1
(1), 979984.
Mamun, S., Svendsen, H.F., Hoff, K.A., Juliussen, O., 2007. Selection of new absorbents
for carbon dioxide capture. Energy Conversion and Management 48 (1),
251258.

72

H.-J. Song et al. / International Journal of Greenhouse Gas Control 11 (2012) 6472

Munoz,
D.M., Portugal, A.F., Lozano, A.E., De La Campa, J.G., De Abajo, J., 2009. New
liquid absorbents for the removal of CO2 from gas mixtures. Energy and Environmental Science 2 (8), 883891.
Nagao, Y., Hayakawa, A., Suzuki, H., Mitsuoka, S., Iwaki, T., Mimura, T., Suda, T., 1998.
Comparative study of various amines for the reversible absorption capacity of
carbon dioxide. Studies in Surface Science and Catalysis 114, 669672.
Oexmann, J., Hensel, C., Kather, A., 2008. Post-combustion CO2 -capture from coalred power plants: Preliminary evaluation of an integrated chemical absorption
process with piperazine-promoted potassium carbonate. International Journal
of Greenhouse Gas Control 2 (4), 539552.
Portugal, A.F., Sousa, J.M., Magalhes, F.D., Mendes, A., 2009. Solubility of carbon
dioxide in aqueous solutions of amino acid salts. Chemical Engineering Science
64 (9), 19932002.
Rinker, E.B., Ashour, S.S., Sandall, O.C., 1995. Kinetics and modeling of carbon dioxide absorption into aqueous solutions of N-methyldiethanolamine. Chemical
Engineering Science 50 (5), 755768.
Sakwattanapong, R., Aroonwilas, A., Veawab, A., 2005. Behavior of reboiler heat duty
for CO2 capture plants using regenerable single and blended alkanolamines.
Industrial and Engineering Chemistry Research 44 (12), 44654473.
Sartori, G., Savage, D.W., 1983. Sterically hindered amines for CO2 removal from
gases. Industrial and Engineering Chemistry Fundamentals 22 (2), 239249.
Singh, P., Versteeg, G.F., 2008. Structure and activity relationships for CO2 regeneration from aqueous amine-based absorbents. Process Safety and Environment
Protection 86 (5), 347359.

Singh, P., Brilman, D.W.F.(W.), Groeneveld, M.J., 2010. Evaluation of CO2 solubility
in potential aqueous amine-based solvents at low CO2 partial pressure. International Journal of Greenhouse Gas Control 5 (1), 6168.
Soosaiprakasam, I.R., Veawab, A., 2008. Corrosion and polarization behavior of
carbon steel in MEA-based CO2 capture process. International Journal of Greenhouse Gas Control 2 (4), 553562.
Supap, T., Idem, R., Tontiwachwuthikul, P., Saiwan, C., 2009. Kinetics of sulfur
dioxide- and oxygen-induced degradation of aqueous monoethanolamine solution during CO2 absorption from power plant ue gas streams. International
Journal of Greenhouse Gas Control 3 (2), 133142.
Van Holst, J., Versteeg, G.F., Brilman, D.W.F., Hogendoorn, J.A., 2009. Kinetic study
of CO2 with various amino acid salts in aqueous solution. Chemical Engineering
Science 64 (1), 5968.
Veawab, A., Tontiwachwuthikul, P., Bhole, S.D., 1997. Studies of corrosion
and corrosion control in a CO2 -2-amino-2-methyl-1-propanol (AMP) environment. Industrial and Engineering Chemistry Research 36 (1), 264
269.
Xu, S., Wang, Y.W., Otto, F.D., Mather, A.E., 1996. Kinetics of the reaction of carbon
dioxide with 2-amino-2-methyl-1-propanol solutions. Chemical Engineering
Science 51 (6), 841850.
Yan, S.P., Fang, M.X., Zhang, W.F., Wang, S.Y., Xu, Z.K., Luo, Z.Y., Cen, K.F., 2007.
Experimental study on the separation of CO2 from ue gas using hollow ber
membrane contactors without wetting. Fuel Processing Technology 88 (5),
501511.

Potrebbero piacerti anche