274 Energy & Fuels 2001, 15, 274-278

Study of CO2 Absorption and Desorption in a Packed

Column
James T. Yeh* and Henry W. Pennline
National Energy Technology Laboratory, U.S. Department of Energy, P.O. Box 10949,
Pittsburgh, Pennsylvania 15236-0940

Kevin P. Resnik
Parsons Project Services, Inc., P.O. Box 10940, Pittsburgh, Pennsylvania 15236

Received October 25, 2000. Revised Manuscript Received January 4, 2001

See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Carbon dioxide, a major greenhouse gas, may need to be removed from flue gas produced by
combustion of fossil fuels in order to manage future climate changes. Although conventional wet
scrubbing techniques exist for removal of carbon dioxide from gas streams, the wet scrubbing
techniques must improve to process large volumes of flue gas at acceptable thermal efficiencies
and minimal costs. Amine scrubbing is one such technique to remove CO2. To make the process
Downloaded via GAZI UNIV on March 1, 2022 at 08:26:01 (UTC).

more efficient, various areas of improvement were investigated: gas-liquid contacting area, the
type of reactant, and dilution of the aqueous fraction with organic liquids. Pertaining to the former,
various absorption tests with conventional packing material and structured packings were
conducted with mono-ethanolamine (MEA), a traditional solvent, as well as with novel amines.
Significant improvements in CO2 removal were obtained with the structured packing. With respect
to the amine investigation, a key to improved efficiency is the ease of regeneration of the CO2-
loaded absorbent. Testing with a sterically hindered amine, 2-amino-2-methyl-1-propanol, revealed
that, although absorption was somewhat less as compared to MEA, thermal regeneration was
far easier. In addition, the effect of organic diluent on CO2 absorption capacity of MEA solution
is determined. The impact of various process parameters on the absorption and desorption steps
will also be discussed.

Introduction even before disposal costs are added. These results are
consistent with an EPRI study on CO2 capture and
Chemical solvent absorption is based on reactions disposal.3 Combustion with pure oxygen requires rede-
between CO2 and one or more basic absorbents, such sign of the entire combustion and boiler system, and
as aqueous solutions of mono-, di-, or tri-ethanolamine. therefore, cannot be retrofitted. The chemical solvent
An advantageous characteristic of absorption is that it absorption process for CO2 capture can be retrofitted
can be reversed by sending the CO2-rich absorbent to a in existing boiler systems and provides a rationale to
desorber (or stripper) where the temperature is raised. pursue near-term evolutionary capture techniques.
In the case of physical absorption, where solvents Emerging and future electricity generation technologies
include methanol, poly(ethylene glycol), dimethyl ether, and novel CO2 capturing methods have the potential to
and others, CO2 is absorbed under pressure, and the significantly reduce electricity costs if the control of CO2
gas desorption can be achieved at reduced pressure. The emissions is mandated.
regenerated absorbent is then returned to the absorber,
In a more recent study,7 it has been projected that
thereby creating a continuous recycling process. The
the energy penalty of amine scrubbing of flue gas
disadvantages of chemical absorption processes include
their limited loadings and high energy requirements
(1) Aroonwilas, A.; Toniwachwuthikul, P. Mass Transfer of High
resulting from the reaction stoichiometry and the heats Performance Structured Packing for CO2 Separation Processes. Pro-
of absorption, respectively. There are also problems of ceedings of the 3rd International Conference on Carbon Dioxide
Removal; Pergamon Press: Elmsford, NY; Sept. 9-11, 1996, Cam-
corrosion and degradation. bridge, MA; pp S75-S80.
All currently available CO2 separation processes are (2) The Capture, Utilization and Disposal of Carbon Dioxide from
energy intensive. In ranking the energy penalty of Fossil Fuel-Fired Power Plants; Herzog, H., Ed.; DOE/ER-30194, 1993;
Vol. 1.
various capture processes, combustion with pure oxygen (3) Booras, G. S.; Smelser, S. C. An Engineering and Economic
is the least energy intensive (about 30% energy penalty), Evaluation of CO2 Removal from Fossil-Fired Power Plants. Energy
1991, 16, pp 1295-1305.
and is followed by chemical solvent absorption processes (4) Mimura, T. S.; Iijima, M.; Mitsuoka, S. Development on Energy
(about 35% energy penalty).2 Also, it has been concluded Saving Technology Saving Technology for Flue Gas Carbon Dioxide
that even the current most efficient technology will Recovery by Chemical Absorption Method and Steam System in Power
Plant. Proceedings of the 4th International Conference on Greenhouse
reduce energy efficiency of utility steam plants by about Gas Control Technologies; Pergamon Press: Elmsford NY; Interlaken,
30% and will increase the price of electricity by 80%, Switzerland, Aug. 30sSept. 2, 1988; pp 71-76.

10.1021/ef0002389 CCC: $20.00 © 2001 American Chemical Society

Published on Web 02/20/2001
CO2 Absorption and Desorption in a Packed Column Energy & Fuels, Vol. 15, No. 2, 2001 275

Figure 1. Flowsheet for the recovery of CO2 from flue gas

using chemical absorption.
Figure 2. Simplified flowsheet for CO2 scrubbing.
produced in a coal-fired utility can be reduced to 15%
through various improvements to the process. Some of distribution to the packing material. The spray nozzle, which
these improvements other than process integration were delivers fine mists, is placed 2.5 cm above the packing. The
outlined by Plasynski and Chen.8 liquid is electrically preheated to the reactor control temper-
Increased contacting areas per unit volume will ature in the stainless steel inlet line. The baseline absorber
enhance gas-liquid interactions and thus mass trans- temperature is normally set at 38 °C during CO2 absorption.
fer. Regeneration of the spent amine solution is an Liquid flow rate is measured by a MicroMotion mass flowme-
energy-intensive step. As compared to conventional ter, and controlled by a metering pump. Flow rate and
temperature data are continuously stored in a computer at a
monoethanolamine (MEA), sterically hindered amines
predetermined time interval.
require lower heats of regeneration. In the MEA process,
Simulated flue gas enters from the bottom of the absorber.
the aqueous fraction is about 80% of the solution, and Gas flow rate is controlled by a mass flow meter controller
during the regeneration step, this must be heated. manufactured by Tylan General. Gas flow rate data are also
Certain organic diluents have specific heats about 50% stored in computer files. The baseline gas composition is 15%
lower than water, and thus the possibility exists that CO2 and 85% N2. Coal-fired flue gas nominally consists of 15%
replacing a portion of the water in the MEA solution CO2 on a dry basis when the combustion takes place with 20%
with organic diluent may reduce the energy penalty excess air. Sulfur oxides, hydrogen chloride, nitrogen oxides,
during the regeneration heating cycle. It is the intent and oxygen are not included in the simulated flue gas to avoid
of the present study to investigate improvements in the possible interferences with the test objectives for this series
amine scrubbing process for CO2 capture. of experiments. These gases are known to cause degradation
of the solvents.
The temperature of the experiment is maintained by
Experimental Section controllers on the solvent inlet heater, the gas inlet heater,
and the water jacket heater. For all of the tests described
Absorption. Figure 1 shows amine scrubbing, a typical below, the temperature setpoints for these three streams are
liquid chemical absorption process for CO2 capture. The equivalent at each test condition. The actual temperatures
flowsheet represents a continuous absorption/regeneration typically vary from the setpoint by less the 0.5 °C. The gas
cycling process. CO2 is captured in the absorber at ap- temperature at the exit of the packing is measured at a point
proximately 38 °C and is released from the regenerator at 1 cm above the solvent nozzle. It is generally greater than the
approximately 121 °C at a much higher CO2 concentration. gas inlet temperature due to the exothermic absorption; the
In the present investigation, the experimental apparatus magnitude of the temperature rise depends on the amount of
consists of a packed-column absorber to promote gas-liquid CO2 absorbed. During this series of tests, the temperature
contact and reaction in a counter-current flow pattern. Figure variations along the height of the packing were not measured
2 shows the schematic diagram of the packed column absorber. owing to the cooling water flowing in the outer jacket of the
The glass absorber is 7.6-cm i.d. and packing height varied double-jacketed packed-column reactor. The double jacket was
from 18 to 71 cm during this series of tests. The reactor has designed to counter the exothermic heat of reaction between
an exterior jacket with hot water flowing in this outer jacket carbon dioxide and absorbents. In the future the temperature
for absorber temperature control. The heat-jacketed section variation along the center core of the packing with and without
is 83 cm high. CO2 absorbent enters from the top of the the cooling water flowing in the outer jacket will be monitored.
absorber through a spray nozzle to ensure good initial liquid
The initial objectives of the experiments are (1) to obtain
first-hand data on the effect of structured packing versus
(5) Freon, P. H. M.; Jansen, A. E. Techno-economic assessment of traditional random (or dumped) packing on the CO2 capture
membrane gas absorption of the production of carbon dioxide from flue
gas. Proceedings of the 4th International Conference on Greenhouse Gas
rate; (2) to obtain CO2 capture rate data by mono-ethanolamine
Control Technologies; Pergamon Press: Elmsford, NY; Interlaken, (MEA) and by a sterically hindered amine; and (3) to compare
Switzerland, Aug. 30-Sept. 2, 1998; pp 53-58. CO2 stripping rates between CO2-rich MEA and CO2-rich
(6) Chakma, A. CO2 Capture ProcessessOpportunities For Improved sterically hindered amine. The sterically hindered amine used
Energy Efficiencies. Proceedings of the 3rd International Conference
on Carbon Dioxide Removal; Pergamon Press: Elmsford, NY; Cam-
during this test series was 2-amino-2-methyl-1-propanol (AMP).
bridge, MA, Sept. 9-11, 1996; pp S51-S56. AMP reacts with CO2 at a slower rate than MEA. Though AMP
(7) Herzog, H.; Drake, E.; Adams, E. CO2 Capture, Reuse, and reacts with CO2 at a slower rate, less energy is required to
Storage Technologies for Mitigating Global Climate Change. DOE drive out CO2 from its CO2-rich solution1. There is a wide
Final Report DE-AF22-96PC01257; January 1997.
(8) Plasynski, S. I.; Chen, Z.-Y. Review of CO2 Capture Technologies
selection of sterically hindered amines. Recently, in Japan,
and Some Improvement Opportunities. Prepr. Pap.sAm. Chem. Soc., Kansei Electric Power Company, in conjunction with Mitsub-
Div. Fuel Chem. August 2000, 45, (4), 650-655. ishi Heavy Industries Limited, has developed a sterically
276 Energy & Fuels, Vol. 15, No. 2, 2001 Yeh et al.

hindered amine specifically for CO2 recovery from flue gas.4 Table 1. Absorber Efficiency and Absorbent/CO2
TNO, in The Netherlands, also is developing liquid absorbents Molar Ratioa
for flue gas applications.5 However, the identities of those molar absorbent
absorbents were not disclosed. ratio efficiency, % packing utilization, %
The basic reaction chemistry for an aqueous monoethano-
1.4 61.5 ( -0.7 BX gauze 43.9 ( 0.5
lamine solution and CO2 is represented by the following 2.2 98.1 ( 0.8 BX gauze 43.6 ( 0.6
reversible reaction:2 1.4 41.4 Flexipac 29.6
1.4 40.1 Random Saddle 28.6
Cold
C2H4OH NH2 + CO2 + H2O [\
Hot
] C2H4OHNH3+ + HCO3- a Absorbent: 20 wt % MEA; absorber temp: 38 °C; molar ratio:

mole absorbent inlet/mole CO2 inlet; absorber height: 53 cm.

This is an exothermic reaction and 72 KJ of thermal energy
is released per mole of CO2 absorbed in MEA solution. an actual process. The regeneration temperature is usually
Absorption usually takes place near 50 °C. During regenera- set at 121 °C under slightly elevated pressure in the carbon
tion, thermal energy (about 165 kJ/mole of CO2) is added to dioxide industry. In the present investigation, the uniqueness
the solution to release the CO2, because a large amount of of the laboratory reactor is that the absorber was used as the
water in the 20 wt % amine solution must be heated to regenerator during the regeneration phase of the absorption/
regeneration temperature. Regeneration usually takes place regeneration cycle. The only difference is that higher temper-
near 120 °C. It has been estimated that up to 80% of total ature is maintained in the packed column in order to drive
cost in the CO2 absorption/regeneration cycle is due to the away CO2 from the rich amine solution. The structured
regeneration procedure.6 packing accelerates the CO2 release from the CO2-rich amine
Three types of column packings are compared for their CO2 solution. Prior to regeneration, the amine solution is saturated
absorption rates at identical test conditions: Intalox saddle with CO2. The CO2-rich solution, which is sprayed into the
(ceramic, random packing), 1.9-cm and the smallest available reactor, is trickling in a thin film down the extensive surface
from the supplier (Norton Chemical Process Products); Flexi- area provided by the packing. No purge gas is required during
pac structured packing supplied by Koch Engineering Com- the regeneration; the recovered CO2 is pure after condensing
pany; and BX Gauze structured packing also supplied by Koch out the vapor. A bag meter is used to measure the total mass
Engineering Company. Structured packing provides more of CO2 evolving from a known amount of CO2-rich amine
gas-liquid contact surface area per unit packed volume than solution; the time at every 2830 cm3 (0.1 ft3) advance at the
random packings. Thus the overall CO2 capturing capacity and bag meter is manually recorded.
rate by the liquid absorbent is increased. The packing height Evaluation of Organic Diluent. The purpose of using
ranged approximately 18 cm to 71 cm during this series of organic diluent to replace some portions of water in the amine
tests. solution is to reduce the thermal energy required to heat the
CO2 Absorption Rate Determination. Inlet nitrogen CO2-saturated absorbent. If part of the water is replaced by
flow, inlet CO2 concentration (vol %), and outlet CO2 concen- an organic diluent, an amount of thermal energy could be
tration (vol %) are used to calculate instantaneous absorption saved during the solvent heating cycle for releasing CO2. Two
rates every 10 s during an absorption test. The nitrogen flow organic diluents, methanol and ethylene glycol, were studied.
is measured by a mass flow controller. The inlet and outlet Methanol has a density of 0.79 g/mL and a normal boiling point
CO2 concentrations are measured by an Horiba infrared gas of 64.7 °C; ethylene glycol has a density of 1.11 g/mL and a
analyzer, which is calibrated immediately before the test. The normal boiling point of 198 °C.
inlet CO2 flow rate is calculated using the following equation:
Results and Discussions
Gi ) [(F × D/M)/(1 - yi)]yi (1)
Absorber Efficiencies. Effects of absorbent (MEA)
where Gi ) CO2 inlet flow rate (lb-mol/h), F ) nitrogen gas flow rate and packing type on CO2 absorption efficien-
flow rate (ft3/h), D ) density of nitrogen (lb/ft3) at standard cies are compared. The test results are shown in Table
conditions, M ) molecular weight of nitrogen, and yi ) CO2 1. Higher absorbent flow rate increases CO2 absorption
mole fraction at inlet of absorber (vol %). efficiency as expected. Absorbent utilization is defined
The inlet CO2 concentration, yi, is averaged for the 20 data as the removal efficiency divided by the stoichiometric
points (190 s) recorded immediately before solvent flow is ratio. If 1 mole of MEA has a maximum capacity in
initiated and is assumed constant throughout the test dura- removing 1 mole of CO2, the stoichiometric ratio of MEA
tion. The CO2 inlet flow rate is controlled by a mass flow
to CO2 is equal to the mole ratio divided by 1.
controller. The calculated flow rate, Gi, is used in the absorp-
tion calculation for data accuracy purposes. Since nitrogen gas Three types of packings were studied for their ef-
is inert in the absorber, the outlet CO2 flow rate is calculated fectiveness in CO2 absorption: BX wire gauze, Flexipac
using the nitrogen flow rate and outlet CO2 concentration 1Y, and random saddle (ceramic 0.95 cm) packings. BX
using the following equation: gauze and Flexipac structured packings are the products
of Koch Engineering Company. The specific surface area
Go ) [(F × D/M)/(1 - yo)]yo (2) for BX Gauze, Flexipac, and random packings are 500,
400, and 66 m2/m3, respectively. At an absorbent-to-CO2
where Go ) CO2 outlet flow rate(lb-mol/h), and yo ) CO2 mole mole ratio of 1.4, BX gauze improves packed column
fraction at outlet of absorber (vol %). absorber efficiency by about 50% over the use of random
The CO2 absorption rate is the difference between the inlet saddle packing. Flexipac minimally improved efficiency
and outlet CO2 flow rates. Total CO2 absorbed in the absorber over random saddle packing. Also, the increase in molar
can be obtained by integrating the instantaneous rates over a
ratio increases the removal efficiency. The variability
selected time period. Efficiency of CO2 absorption is defined
as [(yi - yo)/yi] × 100% at steady state.
in the calculated efficiencies is approximately ( 0.8%.
Regeneration. The CO2-rich liquid chemical absorbent can Comparison of CO2 Absorption Rates. Table 2
be regenerated by heating. CO2 evolves from the rich liquid shows that the sterically hindered amine, AMP, at-
absorbent during the heating. This regenerated liquid absor- tained near-equal CO2 absorption rate using structured
bent is CO2-lean and recirculated to the absorber for reuse in packing as compared to random saddle packing for MEA
CO2 Absorption and Desorption in a Packed Column Energy & Fuels, Vol. 15, No. 2, 2001 277

Table 2. Packing Type and Absorbent Effects on CO2

Absorption Ratea
CO2 absorption rate,
kg-mol/s × 106 absorbent packing
2.65 MEA BX gauze
1.78 MEA Random Saddle
1.88 AMP BX gauze
1.23 AMP Flexipac
1.15 AMP Random Saddle
a Absorbent: 20 wt % MEA; absorber temp: 38 °C; absorbent:

29.2 wt % AMP; absorber height: 53 cm; molar ratio 1.4 (for both
MEA and AMP).

Table 3. Rate of CO2 Regeneration, AMP versus MEAa

CO2
regeneration
MEA AMP rate ratio
(20 wt %) (29.25 wt %) AMP/MEA
CO2 regeneration 0.56 1.03 1.83
rate, kg-mol/s × 106
regeneration temp °C 93 93 N/A
CO2 saturated 10.7 10.7 NA
absorbent flow, lb/h
a Structured packing height: 53 cm (BX gauze). Figure 3. CO2 removal efficiency vs temperature at various
processing. In these tests, 29.2 wt % of AMP solution is packing heights (20% MEA solution).
used while only 20 wt % of MEA solution is used, Table 4. Effect of Organic Diluent (methanol) on CO2
because the molecular weight of AMP is larger than Absorption by MEA Solutiona
MEA. Thus equal molar flowrates (0.09 lb-mol/h) are density of CO2
maintained for both MEA and AMP in the absorbent overhead removal
solution, respectively. Flexipac structured packing did test MEA, methanol, water, condensate, efficiency,
not significantly improve the CO2 absorption rates. no. wt % wt % wt % g/mL %
Rate of CO2 Regeneration: AMP vs MEA. To 1 20 60 20 0.815 96.5
compare CO2 regeneration rate between MEA and AMP, 2 20 40 40 0.858 94.7
3 20 20 60 0.881 92.7
CO2-saturated MEA and AMP solutions are prepared. 4 20 0 80 0.989 93.9
The calculated saturation results for AMP and MEA
a Absorber temperature: 37 °C; absorber height: 53 cm; mole
revealed that after bubbling with CO2, the molar ratio
MEA inlet/mole CO2 inlet: 2.0.
of solvent to CO2 is approximately 1:1 when saturated
using the standard procedure (consecutive absorption above a certain packed column height. The effect of
tests with a 15% CO2 gas mixture until no further reaction temperature on CO2 removal efficiency is
absorption is seen, then bubbling with pure CO2 over- relatively insignificant in the range between 38 and
night). During regeneration, the packed absorber, used 54 °C.
as the regenerator, is only heated to 93 °C, since in the Effects of Organic Diluent. Although the assump-
present reactor setup, this is the highest temperature tion was made that organic diluents would reduce the
that can be safely maintained to prevent the water or energy requirements of regeneration, experiments were
solution from boiling. CO2-saturated AMP or MEA conducted to determine if the organic diluent would
solution is sprayed on the top of the packing, while the have any impact during the absorption step. Table 4
CO2 releasing rate is recorded with the aid of a bagme- shows the effect of methanol on CO2 absorption by MEA
ter. The data in Table 3 show that CO2 releasing rate solution during testing. It indicates that if methanol is
from AMP solution is about 80% faster than from MEA substituted for water, there is no appreciable difference
solution. No purge gas is used. in CO2 absorption by MEA solution at 20%, 40%, and
Parametrics Tests. The effects of height of the 60% methanol concentrations. However, the overhead
packed column and column temperature on CO2 capture methanol carryover was substantial. The fraction of
efficiency using 20 wt % MEA were evaluated. The methanol in the overhead condensate trap in the gas
height of packed column parameter varied from a exit line was approximately calculated from the known
minimum 18 cm to a maximum 71 cm; the column density of pure methanol (0.79 g/mL), the measured
temperature study ranged from 29 to 54 °C. Sections of density of the methanol-free MEA solution (0.99 g/mL),
BX Gauze structured packing that were approximately and the measured density of overhead condensate (0.82
18 cm in length were used. An O-ring was inserted g/mL) for the 60% (or at 0.6 fraction) methanol in MEA
between adjacent structured packing to prevent liquid solution. The fraction of methanol in the overhead
channeling along the wall of the packed column. The condensate was thus calculated at 0.89. The concentra-
results are plotted in Figure 3. The molar ratio of MEA tion of methanol increased from the original 60% in the
solution flowrate to CO2 gas flow rate is 2.0. At a MEA solution to 88.8% in the overhead condensate, an
constant temperature, increasing the height of the increase of 48%. The low boiling point of 64.7 °C for the
packing increases the CO2 removal. The relationship methanol caused this preferential vapor carryover.
between the height and CO2 removal is almost expo- Although, methanol would be unsuitable for use as an
nential. The increase in CO2 removal becomes marginal organic diluent in an atmospheric pressure application,
278 Energy & Fuels, Vol. 15, No. 2, 2001 Yeh et al.

Table 5. Effect of Organic Diluent (ethylene glycol) on Conclusions

CO2 Absorption by MEA Solutiona
density of CO2
Investigations of amine-based scrubbing for CO2
ethylene overhead removal capture were performed to elucidate ways to improve
test MEA, glycol, water, condensate, efficiency, this chemical absorption process. Increasing the gas-
no. wt % wt % wt % g/mL % liquid contacting area has a major impact on scrubbing.
1 20 60 20 0.981 80.1 Absorption tests revealed that for a particular amine,
2 20 40 40 0.992 85.5 structured packing improves the absorber efficiency and
3 20 20 60 0.985 88.8 absorption rate as compared to the more traditional
4 20 0 0 0.989 92.4
random packing. The type of amine is also a consider-
a Absorber temperature: 37 °C; absorber height: 53 cm; mole
ation. At the same bed geometry, the conventional MEA
MEA inlet/mole CO2 inlet: 2.0. performed much better during absorption studies than
the sterically hindered amine, AMP. However, in the
it will be suitable in an higher pressure system, such regeneration step, the CO2 releasing rate from the
as IGCC, since the high pressure raises the boiling point saturated AMP solution is over 80% greater than from
of the methanol. saturated MEA solution. By extracting information from
The CO2 solubility in methanol was also measured the above results, it can be speculated that if the more
using the same packed column and only 5% CO2 easily regenerable AMP is substituted for MEA, an
removal efficiency was obtained by pure methanol from overall process benefit will be obtained if a structured
the simulated flue gas. packing is used as compared to the random packing.
Table 5 shows the effect of ethylene glycol on CO2 Additionally, organic diluents can reduce the sensible
absorption by MEA solution. In contrast to methanol, heat duty in the regeneration step, but experimental
ethylene glycol as diluent suppresses CO2 absorption results indicate that the volatility of the organic diluent
by MEA solution. The higher the concentration of the and its impact on the absorption step are important
diluent the lower the CO2 removal efficiency. Also in considerations. Other techniques to improve the amine-
contrast to methanol experiments, the density of over- based scrubbing will be investigated in the future.
head condensate stayed virtually unchanged at different
diluent concentrations. This was not surprising, since Disclaimer
pure ethylene glycol has a boiling point of 198 °C. This
Reference in this report to any specific commercial
shows that ethylene glycol could be a viable organic
process, product or service is to facilitate understanding
diluent at atmospheric pressure. However, an economic
and does not necessarily imply its endorsement or
study will dictate if the lower regeneration cost can more
favoring by the United States Department of Energy.
than compensate for the lower CO2 removal efficiency
of the chemical scrubbing system. EF0002389