Energy & Fuels 2001, 15, 279-284 279

Adsorption of CO2 on Molecular Sieves and Activated

Carbon
Ranjani V. Siriwardane,* Ming-Shing Shen, Edward P. Fisher, and
James A. Poston
U.S. Department of Energy, National Energy Technology Lab, 3610 Collins Ferry Road,
P.O. Box 880, Morgantown, West Virginia 26507-0880

Received October 25, 2000. Revised Manuscript Received December 18, 2000

Pressure swing adsorption (PSA) and temperature swing adsorption (TSA) are some of the
potential techniques that could be applicable for removal of CO2 from high-pressure fuel gas
streams. Molecular sieves and activated carbons are some of the sorbents that could be utilized
in the PSA process. Volumetric adsorption studies of CO2, N2, or H2 on molecular sieve 13X,
molecular sieve 4A, and activated carbon were conducted at 25 °C up to a pressure of 300 psi
(∼2× 106 Pa). Preferential adsorption of CO2 was observed with all three sorbents. The adsorption
capacity of CO2 for molecular sieve 13X was higher than that for molecular sieve 4A at all
pressures up to 300 psi. At low pressures (<25 psi) the adsorption capacity for CO2 of activated
carbon was lower than that of molecular sieve 13X, but at higher pressures (>25 psi) activated
carbon exhibited significantly higher CO2 capacities than were found for molecular sieves.
Competitive adsorption of CO2 from gas mixtures also indicated that both molecular sieve 13X
and activated carbon can be utilized for separation of CO2 from gas mixtures.

Introduction In this work, three sorbents, namely, molecular sieve

Fossil fuels supply more than 98% of the world’s 13X, molecular sieve 4A, and activated carbon from Süd
energy needs. However, the combustion of fossil fuels Chemie Inc., that can be utilized in the PSA process for
is one of the major sources of the greenhouse gas CO2. recovery of CO2 were studied. Volumetric adsorption
It is necessary to develop technologies that will allow studies of CO2, N2, and H2 on the three sorbents were
us to utilize the fossil fuels while reducing the emissions conducted at 25 °C and up to a pressure of 300 psi (∼2
of greenhouse gases. Commercial CO2 capture technol- × 106 Pa) to determine the equilibrium adsorption
ogy that exists today is very expensive and energy capacity of these materials. The adsorption isotherms
intensive. Improved technologies for CO2 capture are of these materials at pressures up to 300 psi are not
necessary to achieve low energy penalties. Pressure reported in the literature. Studies on competitive ad-
swing adsorption (PSA) is one of the potential tech- sorption of CO2 from gas mixtures were conducted
niques that could be applicable for removal of CO2 from utilizing an atmospheric micro reactor. Elemental char-
gas streams. acterizations of these sorbents were conducted utilizing
PSA processes are based on preferential adsorption advanced analytical techniques such as X-ray photo-
of the desired gas (e.g., CO2) on a porous adsorbent at electron spectroscopy and scanning electron microscopy/
high pressure, and recovery of the gas at low pressure. energy dispersive spectroscopy.
Thus, the porous sorbent can be reused for subsequent
adsorption. PSA technology has gained interest due to Experimental Section
low energy requirements and low capital investment
The three sorbents, namely, Zeochem-Z10-02/molecular
costs.1,2 Applications of PSA processes to separation/ sieve 13X (8 × 12 mesh), Zeochem Z4-01/molecular sieve 4A
capture of carbon dioxide are also reported.3,4 Low (4 × 8 mesh), and G-32 H activated carbon (4 × 10 mesh),
recovery rate of CO2 is one of the problems reported with were obtained from Süd Chemie Inc. Adsorption isotherms at
the PSA processes. Development of regenerable sorbents 25 °C of pure CO2 (99.5%, Jackson Welding Supply Co.), N2
that have high selectivity, high adsorption capacity, and (99.5%, Jackson Welding Supply Co.), and H2 (99.95%, Butler
high desorption rate for CO2 adsorption/desorption is Gas Products Co.) on the three sorbents were obtained up to
critical for the success of the PSA process. an equilibrium pressure of about 300 psi (∼2 × 106 Pa)
utilizing a volumetric adsorption apparatus. Approximately
* Author to whom correspondence should be addressed. Tel: 304- 10 mL of the sorbent materials were placed in the sample
285-4513. E-mail: rsiriw@netl.doe.gov.
(1) Skarstrom, C. W. U.S. Patent 2,944,627, 1960.
chamber and were evacuated to ∼5 × 10-5 Torr. The amount
(2) Guerrin de Montgareuil, P.; Domine, D. U.S. Patent 3,155,468, of CO2 adsorbed was calculated utilizing the pressure mea-
1964. surements before and after the exposure of the sample
(3) Cheu, K.; Jong-Nam, K.; Yun-Jong, Y.; Soon-Haeng, C. Funda- chamber to CO2. Baseline data with CO2 were obtained
mentals of Adsorption. Proc. Int. Conf.; LeVan, D., Ed.; Kluwer
Academic Publishers: Boston, Massachusetts, 1996; pp 203-210.
utilizing 10 mL of glass beads (2 mm diameter). Nine adsorp-
(4) Dong, F.; Lou, H.; Goto, M.; Hirose, T. Sep. Purif. Technol. 1990, tion cycles were performed with each sorbent. After each cycle
15, 31-40. the sorbent was evacuated to about 5 × 10-5 Torr overnight.

10.1021/ef000241s CCC: $20.00 © 2001 American Chemical Society

Published on Web 02/07/2001
280 Energy & Fuels, Vol. 15, No. 2, 2001 Siriwardane et al.

Figure 2. Adsorption isotherms of CO2, N2, and H2 on

Figure 1. Adsorption isotherms of CO2, N2, and H2 on molecular sieve 4A at 25 °C.
molecular sieve 13X at 25 °C.
adsorption isotherms of both nitrogen and hydrogen
Cycles 1,2,3,4,7, and 10 were performed with CO2, cycles 5 and were lower than those of CO2. Both nitrogen and
6 were performed with nitrogen, and cycles 8 and 9 were hydrogen isotherms were also completely reversible.
performed with hydrogen. Preferential adsorption of CO2 on molecular sieve 13X
BET nitrogen surface areas of the materials were measured indicates that this material can be used for separation
utilizing Micromeritics surface area and pore volume analyzer. of CO2 from some gas mixtures. The final adsorption
X-ray photoelectron spectroscopy (XPS) spectra were recorded isotherm (cycle 9 which was obtained after the comple-
with a Physical Electronics Model SAM 590 equipped with a
tion of adsorption experiments with nitrogen and hy-
model 15-255 GAR analyzer and a 15 kV X-ray source from
Physical Electronics.5 The system was routinely operated drogen) was conducted with CO2 and it was very similar
within a pressure range of 1.3 × 10-7 to 1.3 × 10-6 Pa (10-9- to the previous adsorption isotherms with CO2. This
10-8 Torr). The instrument was calibrated using the photo- indicated that the sorbent was not affected by the
emission lines EB(Cu 2p3/2) ) 932.4 eV and EB (Au 4f7/2) ) 83.8 adsorption of nitrogen and hydrogen.
eV. The binding energies were referenced to the C(1s) level at The volumetric adsorption isotherms for molecular
284.6 eV for adventitious carbon. All intensities reported are sieve 4A are shown in Figure 2. The adsorption iso-
experimentally determined peak areas divided by the instru- therms indicate that the uptake of CO2 was higher than
mental sensitivity factors. Spectra were obtained with about
that of the other two gasessnitrogen and hydrogen. This
five different samples and the intensity and binding energy
values reported are the averages of values obtained from all
preferential uptake of CO2 indicates that this sorbent
sets of data. X-ray microanalysis was carried out using JEOL may be suitable for the separation of CO2 from gaseous
840A model scanning electron microscope (SEM). The electron mixtures. The adsorption isotherm of CO2 for molecular
microscope is equipped with a Noran Instruments energy- sieve 4A during the first cycle was higher than that of
dispersive spectrometer.6 The SEM was interfaced to a Noran the following cycles, indicating that the adsorption was
Instruments Voyager-IV X-ray analysis and image processing not completely reversible. The adsorption at the first
unit. Resolution based on the Mn KR energy line was 149 eV. cycle was the highest. The average pore diameter of
The competitive adsorption studies at 25 °C were conducted molecular sieve 4A is 4 Å, while the average pore
utilizing Micromeritics Autochem 2910 atmospheric micro diameter of molecular sieve 13X is about 10 Å. The
reactor. The composition of the two gas mixtures utilized were kinetic diameter of the CO2 molecule is 3.9 Å, which is
14.8% CO2 and 85.2% N2, and 10% CO2, 28% H2, and 62% He.
only slightly lower than the pore diameter of molecular
The analysis of the outlet gas stream was conducted utilizing
a Pfeiffer Vacuum Thermostar mass spectrometer. The samples sieve 4A.7 So the mobility of the CO2 molecule may be
were pre- treated at 250 °C for 1 h in the micro reactor under restricted inside the pores of molecular sieve 4A. The
helium prior to the exposure to the gas mixtures containing incomplete desorption of CO2 after the first adsorption
CO2. cycle in molecular sieve 4A may be due to the smaller
pore size. However, the amount of CO2 adsorbed was
Results and Discussion constant after the second cycle. This indicates that after
the second cycle, a constant amount of CO2 may be
Adsorption Isotherms. Volumetric adsorption iso- desorbed at each cycle. The uptake of CO2 for molecular
therms of CO2, N2, and H2 on molecular sieve 13X at sieve 4A was lower than that of molecular sieve 13X at
25 °C are shown in Figure 1. The shape of the isotherms all equilibrium pressures up to 300 psi. However, both
appears to be similar to type 1(Langmuir). The CO2 nitrogen and hydrogen adsorption isotherms of molec-
adsorption increased rapidly when the pressure was ular sieve 4A appear to be similar to those of molecular
increased up to 50 psi. The increase in CO2 adsorption sieve 13X. The adsorption of nitrogen and hydrogen did
after 50 psi appears to be gradual. The adsorption not appear to adversely affect the adsorption capacity
isotherms for repeated cycles were very similar. This of the sorbent for CO2 adsorption.
indicated that the adsorption is fully reversible and The adsorption isotherms of CO2 on activated carbon
complete regeneration can be obtained by evacuation are shown in Figure 3. It is interesting to note that all
of the material after adsorption. At all pressures, the isotherms were extremely reproducible, which in-
dicates the excellent reversibility of adsorption. The CO2
(5) Siriwardane, R. V.; Poston, J. P.; Fisher, E. P.; Shen, M.; Miltz,
A. Appl. Surf. Sci. 1999, 152, 219-236.
(6) Siriwardane, R. V.; Poston, J. P.; Grover, E., Jr. Ind. Eng. Chem. (7) Choudary, V. R.; Mayadevi, S.; Singh, A. P. J. Chem. Soc.,
Res. 1994, 33, 2810-2818. Faraday Trans. 1995, 91, 2934-2944.
CO2 Adsorption on Molecular Sieves and Activated C Energy & Fuels, Vol. 15, No. 2, 2001 281

Figure 3. Adsorption isotherms of CO2, N2, and H2 on Figure 5. Adsorption isotherms of CO2 on three sorbents at
activated carbon at 25 °C. 25 °C.

Figure 4. Ratios of amounts of CO2/N2 adsorption on three

sorbents. Figure 6. Adsorption isotherms of CO2 on three sorbents at
25 °C.
uptake of activated carbon was lower than that for the
Isosteric heats of adsorption as a function of surface
two molecular sieves at lower pressures (<30 psi), but
coverage were calculated using adsorption isotherms at
at higher pressures the CO2 uptake of activated carbon
different temperatures (30, 40, 50 °C). The Clausius-
was higher than that of the molecular sieves. Nitrogen
Clapeyron equation was utilized in these calculations.8
and hydrogen adsorption isotherms on activated carbon
Isosteric heats of adsorption of molecular sieve 13X
were also very reproducible. The final CO2 adsorption
decreased from 50 to 10 kJ mol-1 with increasing
cycle (cycle 10) was very similar to the first cycle, which
coverage from 1 to 5 mol/kg. For the activated carbon,
indicated that the CO2 adsorption capacity of the
the isosteric heats of adsorption decreased from 28 to
sorbent was not affected by the adsorption of other
11 kJ/mol with increasing coverage. The wide range of
gases.
activation energies observed for the molecular sieve 13X
The ratio of the moles of adsorbed CO2 to nitrogen is
indicated that there is a wide range of sites with varying
shown in Figure 4. The ratio of CO2/nitrogen decreased
adsorption energies present in molecular sieve 13X.
with increasing pressure and appeared to remain con-
Characterization of Adsorbents. Since molecular
stant above 150 psi in all three sorbents. This indicates
sieve 13X and activated carbon showed reversible
that the selectivity of the sorbents for CO2 adsorption
adsorption most of the characterization work was done
is better at lower pressures even though the total CO2
with these two sorbents. The nitrogen BET surface
adsorption was higher at higher pressures. At lower
areas of molecular sieves 13X and activated carbon were
pressures, this selectivity molar ratios appears to be
506 and 897 m2/g, respectively. The high surface area
higher for the molecular sieves than that for the
of activated carbon may have contributed to the high
activated carbon but was similar at higher pressures.
CO2 uptake at higher pressures. The CO2 uptake values
Similar observations were made with the adsorbed mole
per unit surface area are shown in Figure 7. It is
ratios of CO2/hydrogen on the three sorbents.
interesting to note that the moles of CO2 adsorbed per
Adsorption isotherms of CO2 (best isotherm selected
unit area is higher on molecular sieve 13X than that
from the repeated cycles) with adsorption denoted in
on activated carbon. These values were similar on both
mol/kg and mol/L, for three adsorbents are shown in
materials only at about 250-300 psi. This indicates that
Figures 5 and 6, respectively. Both molecular sieves 13X
the surface of molecular sieve 13X has a better affinity
and 4A exhibited higher uptake of moles of CO2 per kg
for CO2 than that of activated carbon.
of the sorbent at pressures less than 30 psi, but at
The cross sectional area9 of the CO2 molecule is 19.5
higher pressures activated carbon showed higher CO2 2
Å . If close packing of CO2 is assumed, at a monolayer
uptake. When the CO2 uptake was in mol/L, the
activated carbon showed higher CO2 uptake at pres- (8) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area and
sures above 100 psi. Porosity; Academic Press: New York, 1967.
282 Energy & Fuels, Vol. 15, No. 2, 2001 Siriwardane et al.

× 10-6 mol/m2. This indicates that a close pack mono-

layer of CO2 is also formed at saturation with the
activated carbon. So both these materials utilize the
complete surface area to form a close packed monolayer
at higher pressures.
The skeletal and bulk densities of molecular sieve 13X
as measured by mercury porosimetry were 1.87 and 1.06
g/cm3, respectively, and for activated carbon these
values were 1.14 and 0.94 g/cm3, respectively. So the
density of activated carbon was lower than that of the
molecular sieve 13X.
The elemental analysis of the bulk and the surface of
the sorbents are listed in Table 1. The major elements
Figure 7. Adsorption isotherms of CO2 at 25 °C based on the observed at the surface of the molecular sieves were Si,
surface area of the sorbents. Al, and Na. Trace amounts of K, Ca, and Fe were also
observed. The ratio of the K/Si of molecular sieve 13X
Table 1: Elemental Ratios at the Bulk and the Surface
was considerably higher at the surface than that at the
bulk surf. bulk surf. bulk surf. bulk. This indicates that most of the K resides on the
13X 13X 4a 4a act. C act. C
surface of the molecular sieve 13X. Surface analysis of
Al/Si 0.75 1.08 0.77 0.74 Al/C 0.0001 0.023 the activated carbon indicated that there were Al, K,
Na/Si 0.60 0.69 0.54 0.78 Na/C 0.001 0.005
and trace amounts of Cl, Na, S, and Si on the surface.
K/Si 0.004 0.03 0.04 0.04 K/C 0.007 0.041
Both Al/Si and K/Si ratios were substantially higher at
coverage the surface should occupy about 8.5 × 10-6 the surface of activated carbon than at the bulk as
mol/m2. From the adsorption isotherms it was deter- shown in Table 1. This indicated that the potassium
mined that the molecular sieve 13X has about 9.9 × 10-6 resides mainly on the surface of the activated carbon,
moles of CO2 per square meter at saturation coverage similar to the observations with molecular sieve 13X.
(250-300 psi) and this is very close to the value Elemental distributions of the cross sections of the pellet
corresponding to a close packed monolayer coverage of of molecular sieve 13X are shown in Figure 8. It is clear
CO2. The saturation CO2 coverage of activated carbon that all the elements including the trace elements are
as determined by the adsorption isotherms is about 9.5 well distributed in the pellet. Since the chemical com-

Figure 8. Elemental maps of fresh molecular sieves.

CO2 Adsorption on Molecular Sieves and Activated C Energy & Fuels, Vol. 15, No. 2, 2001 283

Figure 9. Micro reactor studies on CO2 separation with Figure 11. Micro reactor studies on CO2 separation with
molecular sieve 13Xs 14.8% CO2/85.2% N2 mixture. molecular sieve 13X (10% CO2/28% H2/62% He mixture).

Figure 10. Amount of CO2 adsorption from CO2/N2 mixture Figure 12. Micro reactor data on CO2 separation with
as a function of time on molecular sieve 13X. activated carbons14.8% CO2/85.2% N2 mixture.

positions of molecular sieve 13X and activated carbon a gas mixture containing 10% CO2, 28% H2, and 62%
are completely different, it is difficult to relate the He as shown in Figure 11. It was also possible to obtain
adsorption properties to the chemical composition of the an excellent separation of CO2 from the gas mixture
surfaces. containing 14.8% CO2 and 85.2% N2, with activated
Competitive Adsorption Studies from Gas Mix- carbon (0.5 g) as shown in Figure 12, but it was
tures Containing CO2. Results of the competitive gas necessary to lower the flow rate (5 cm3/min) to increase
adsorption studies conducted utilizing 14.8% CO2 and the contact time to utilize the full equilibrium adsorp-
85.2% N2 on molecular sieve 13X in the atmospheric tion capacity for CO2 adsorption. This indicates that the
micro reactor are shown in Figure 9. The gas mixture rate of adsorption of CO2 on activated carbon at 1 atm
was introduced to 1.5 g of molecular sieve 13 X at a flow was lower than that for the molecular sieve 13X.
rate of 15 cm3/min and at 25 °C. After the introduction
of the gas mix to the molecular sieve 13X, the CO2 Conclusions
concentration decreased to almost zero while the nitro-
gen concentration increased to 100%, until the break- Molecular sieve 13X, 4A, and activated carbon showed
through. This indicates that an excellent separation of preferential adsorption of CO2 at 25 °C and pressures
CO2 from a gas mixture of N2 and CO2 can be obtained up to 300 psi. Adsorption of CO2 was found to be
with molecular sieve 13X. The amount of CO2 adsorbed reversible on both molecular sieve 13X and activated
per kilogram of the sorbent as a function of time, carbon. Equilibrium adsorption data from volumetric
calculated from the data is shown in Figure 10. There adsorption experiments indicated that the amount of
is a linear uptake of CO2 with time and the total amount CO2 adsorbed on the molecular sieves is more than that
of CO2 adsorbed at the breakthrough or saturation was of activated carbon at pressures less than 50 psi, but
about 2.1 mol/kg of the sorbent. This value is very activated carbon showed significantly higher adsorption
similar to the amount of CO2 adsorbed at 1 atm from than molecular sieves at pressures above 50 psi. The
volumetric equilibrium adsorption studies, as shown in equilibrium adsorption capacities of activated carbon,
Figure 1. This indicates that the full capacity of the molecular sieve 13X, and molecular sieve 4A at 300 psi
molecular sieve 13X was utilized for CO2 adsorption and at 25 °C are about 8.5, 5.2, 4.8 mol of CO2 /kg of
during competitive gas adsorption from a CO2 and N2 the sorbent, respectively. At 300 psi, both molecular
mixture. A similar observation was made with molec- sieve 13X and activated carbon utilized the full capacity
ular sieve 13X during competitive CO2 adsorption from of the surface to form a close packed monolayer at the
surface. Isosteric heats of adsorption calculations indi-
(9) Lowell, S.; Shields, J. E. Powder Surface Area and Porosity, 3rd cated that the molecular sieve 13X has a wide distribu-
ed.; Chapman and Hall, New York, 1991; p 41 tion of active sites for CO2 adsorption. Potassium is
284 Energy & Fuels, Vol. 15, No. 2, 2001 Siriwardane et al.

preferentially distributed at the surface of the molecular of the gases with activated carbon than that with
sieve 13X and activated carbon. Excellent separation of molecular sieve 13X to achieve the excellent separation.
CO2 was obtained from gas mixtures containing both
Acknowledgment. The Authors thank Mr. Bill
CO2 /N2 and CO2 /H2/He utilizing both molecular sieve
Turner at Süd Chemie Inc. for providing the sorbents
13X and activated carbon during competitive gas ad-
for this work.
sorption studies performed with the micro reactor.
However, it was necessary to use higher contact times EF000241S