pubs.acs.

org/Langmuir
2009 American Chemical Society

Amine-Bearing Mesoporous Silica for CO2 and H2S Removal from Natural
Gas and Biogas
Youssef Belmabkhout, Guy De Weireld, and Abdelhamid Sayari*,

Department of Chemistry, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada, and Thermodynamics
Department, Facult
e Polytechnique, Universit
e de Mons, 31 bd Dolez, Mons 7000, Belgium
Received July 3, 2009. Revised Manuscript Received October 20, 2009

Triamine-grafted pore-expanded mesoporous silica (TRI-PE-MCM-41) exhibited high CO2 and H2S adsorption
capacity as well as high selectivity toward acid gases versus CH4. Unlike physical adsorbents such as zeolites and
activated carbons, the presence of moisture in the feed enhanced the CO2 removal capability of TRI-PE-MCM-41
without altering its H2S adsorption capacity. Thus, depending on the feed composition, CO2 and H2S may be removed
over TRI-PE-MCM-41 simultaneously or sequentially. These findings are suitable for acid gas separation from CH4containing mixtures such as natural gas and biogas.

Introduction
To be transported in pipelines or as a vehicle fuel, methane
originating from natural gas and biogas must meet strict specifications with respect to carbon dioxide and hydrogen sulfide
content. In pipeline-grade methane, CO2 and H2S must be
removed typically to below 2% and 1 ppm, respectively.1 The
most common method of acid gas removal is liquid-phase
chemical scrubbing with amines. However, this technology suffers
from inherently high regeneration cost and inefficiency.2,3 Adsorption is recognized to be an energy efficient technology for
CO2 and H2S removal provided that a stable material with high
adsorption capacity and selectivity toward acid gases is available.
In recent years, considerable research effort was made to develop
novel CO24-13 and H2S12,14,15 adsorbents with such attributes.
However, only a few investigations focused on the development of
adsorbents with the ability to remove selectively both CO2 and
H2S. Recently, Ma et al.12 reported that polyethylenimine (PEI)impregnated MCM-41 or SBA-15 is capable of the selective
adsorption of both CO2 and H2S. However, not only was the
adsorption of CO2 found to be diffusion-limited, but the optimum
*Corresponding author. E-mail: abdel.sayari@uottawa.ca.
(1) Industrial Gas Processing; Haring, H. W., Ed.; Wiley-VCH: Weinheim,
Germany, 2008.
(2) Veawab, A.; Tontiwachwuthikul, P.; Chakma, A. Ind. Eng. Chem. Res. 1999,
38, 3917.
(3) Hook, R. J. Eng. Chem. Res. 1997, 36, 1779.
(4) Lee, K. B.; Beaver, M. G.; Caram, H. S.; Sircar, S. Ind. Eng. Chem. Res. 2008,
47, 80488062.
(5) Llewellyn, P. L.; Bourrelly, S.; Serre, C.; Vimont, A.; Daturi, M.; Hamon, L.;
De Weireld, G.; Chang, J. S.; Hong, D. Y.; Hwang, Y. K.; Jhung, S. H.; Ferey, G.
Langmuir 2008, 24, 72457250.
(6) Hicks, J. C.; Drese, J. D.; Fauth, D. J.; Gray, M; Qi, G.; Jones, C. W. J. Am.
Chem. Soc. 2008, 130, 29022903.
(7) Chaffee, A. L.; Knowles, G. P.; Liang, Z.; Zhang, J.; Xiao, P.; Webley, P. A.
Int. J. Greenhouse Gas Control 2007, 1, 1118.
(8) Gray, M. L.; Champagne, K. J.; Fauth, D.; Baltrus, J. P.; Pennline, H. Int.
J. Greenhouse Gas Control 2008, 2, 38.
(9) Yue, M. B.; Sun, L. B.; Cao, Y.; Wang, Y.; Wang, Z. J.; Zhu, J. H. Chem.;
Eur. J. 2008, 14, 34423451.
(10) Belmabkhout, Y.; Sayari, A. Adsorption 2009, 15, 318128.
(11) Serna-Guerrero, R.; Dana, E.; Sayari, A. Ind. Eng. Chem. Res. 2008, 47,
47614766.
(12) Ma, X.; Wang, X.; Song, C. J. Am. Chem. Soc. 2009, 131, 57775783.
(13) Couck, S.; Denayer, J. F. M.; Baron, G.; Remy, T.; Gascon, J.; Kapteijn, F.
J. Am. Chem. Soc. 2009, 131, 63266327.
(14) Wang, X.; Ma, X.; Xu, X.; Sun, L.; Song, C. Top. Catal. 2008, 49, 108117.
(15) Hamon, L.; Serre, C.; Devic, T.; Loiseau, T.; Millange, T.; Ferey, G.;
De Weireld, G. J. Am. Chem. Soc. 2009, 131, 87758777.

Langmuir 2009, 25(23), 1327513278

temperature for CO2 adsorption was 348 K versus 298 K for H2S.
In addition, although CO2 did not adsorb on this material at room
temperature, it strongly inhibited H2S adsorption. Thus, the
development of materials that adsorb both CO2 and H2S, either
simultaneously or sequentially, is of prime importance for some
applications such as the upgrading of biogas and natural gas. It is
worth mentioning that the adsorption capacity of CO2 and CH4
and the selectivity toward CO2 over methane at low CO2
concentration were seldom reported in the literature. Among
the rare examples, Huang et al.16 reported a CO2 capacity of 1.14
mmol/g at 5% CO2 in helium and significantly lower CH4
adsorption capacity on amine-grafted MCM-48, but the CO2
versus CH4 selectivity was not reported. Sircar and Golden17
found the CO2 versus CH4 selectivity in the Henrys law region at
303 K to be 195.6 and 2.5 for the 5A zeolite and BPL activated
carbon, respectively. A literature survey of CO2 versus CH4
selectivity data at high CO2 concentration on materials such as
zeolites, carbons, and periodic mesoporous silicas was reported
elsewhere.18 The current work addresses CO2 and, for the first
time, H2S removal over triamine-grafted pore-expanded mesoporous silica TRI-PE-MCM-41 at room temperature. Using the
gas-solid approach, it is expected to combine the high selectivity
and reversibility of the acid gas-amine chemistry and the lowenergy requirements for regenerating the solid adsorbent. Adsorption isotherms for dry CO2, H2S, and CH4 were determined
for the amine-functionalized material as well as for the PE-MCM41 silica support. Column breakthrough measurements were used
to determine the selectivity of CO2 adsorption over CH4 at
concentrations akin to those of natural gas and biogas. In
addition, the effect of moisture on the CO2 adsorption capacity
and selectivity was investigated. Recycling of TRI-PE-MCM-41
in the presence of CO2 and H2S is also reported and discussed.

Experimental Section
The detailed preparation procedure and structural characterization of TRI-PE-MCM-41 was reported in earlier conributions.10,19
Single gas adsorption equilibrium measurements for CO2, H2S,
(16) Huang, H. Y.; Yang, R. T.; Chinn, D.; Munson, C. L. Ind. Eng. Chem. Res.
2003, 42, 24272433.
(17) Sircar, R.; Golden, T. C. Sep. Sci. Technol. 2000, 35, 667687.
(18) Belmabkhout, Y; Sayari, A. Chem. Eng. Sci. 2009, 64, 37293735.
(19) Harlick, P. J. E.; Sayari, A. Ind. Eng. Chem. Res. 2007, 46, 446458.

Published on Web 10/29/2009

DOI: 10.1021/la903238y

13275

Letter

Belmabkhout et al.

Figure 1. Adsorption isotherms for pure CO2, H2S, and CH4 on

TRI-PE-MCM-41 at 298 K using a vacuum temperature regeneration operation at 348 K.
and CH4 were performed using a Rubotherm gravimetric-densimetric apparatus.10,20,21 The experimental setup used for column
breakthrough measurements was also described elsewhere.22 All
experimental details are provided in Supporting Information.

Results and Discussion

The BET surface area, pore volume, and pore size for poreexpanded mesoporous PE-MCM-41 silica used as support for
amine grafting were 1141 m2/g, 2.24 cm3/g, and 9.2 nm, respectively. The current adsorbent, TRI-PE-MCM-41, was obtained
by surface functionalization of PE-MCM-41 using triaminesilane,
leading to strongly attached triamine species onto the silica
surface. The BET surface area, pore volume, pore size, and
particle density of TRI-PE-MCM-41 were as follows: 367 m2/g,
0.87 cm3/g, 9.4 nm, and 0.93 g/cm3. The typical amine loading of
TRI-PE-MCM-41 determined by TGA was 7.9 mmol g-1.
Pure Gas Adsorption. Figure 1 shows the adsorption isotherms up to 1.5 bar for pure CO2, H2S, and CH4 at 298 K on
TRI-PE-MCM-41. Before adsorption measurements, the material was pretreated under vacuum (5  10-4 mbar) at 348 K. In the
low-pressure range (i.e., < 0.4 bar), TRI-PE-MCM-41 had a
higher capacity for CO2 than for H2S but an exceedingly small
adsorption capacity for CH4 (Figure S4, Supporting Information). Interestingly, the CO2-H2S sequence was switched in
favor of H2S at pressure higher than 0.4 bar whereas the
adsorption of CH4 remained extremely low.
In earlier contributions,10,11 we demonstrated that the adsorption of CO2 at low pressure on amine-modified material is
dominated by chemisorption as a result of direct interaction
between the adsorbate and amine groups. Whether under dry
or humid conditions, there was a striking similarity with
CO2-amine chemistry in solution.11 It is thus inferred that the
interaction takes place via the zwetterion mechanism for carbamate formation.23,24 The occurrence of carbamate is consistent
with the appearance of a 13C NMR peak at 165 ppm (not
shown).25 The occurrence of chemisorption is also consistent
with the very high initial isosteric heat of CO2 adsorption on
(20) Belmabkhout, Y; Serna-Guerrero, R.; Sayari, A. Chem. Eng. Sci. 2009, 64,
37213728.
(21) Hamon, L.; Frere, M.; De Weireld, G. Adsorption 2008, 14, 493499.
(22) Serna-Guerrero, R.; Sayari, A. Environ. Sci. Technol. 2007, 41, 47614766.
(23) Lu, J. G.; Zheng, Y. F.; He, D. L. Sep. Purif. Technol. 2006, 52, 209217.
(24) Mandal, B.; Bandyopadhyay, S. Environ. Sci. Technol. 2006, 40, 60766084.
(25) Drage, T. C.; Arenillas, A.; Smith, K. M.; Snape, C. E. Microporous
Mesoporous Mater. 2008, 116, 504512.

13276 DOI: 10.1021/la903238y

TRI-PE-MCM-41, reported to be 92 kJ/mol.10 However, as seen

in Figure 1, at high pressure, the CO2 adsorption isotherm does
not level off, indicating that additional physisorption took place.
This is consistent with the adsorption of CO2 on plain PE-MCM41 silica (Figure S3 in Supporting Information).
Being an acid gas, the behavior of H2S as an adsorbate is
qualitatively similar to that of CO2. At low pressure, H2S
chemisorbs on the amine groups via the proton-transfer mechanism. Using a single data point at different temperatures, we
estimated the isosteric heat of H2S adsorption on TRI-PEMCM41 to be ca. 35 kJ/mol. Although lower than the heat of
adsorption for CO2, this result is consistent with the occurrence of
H2S chemisorption at low pressure. It is also consistent with the
fact that H2S is a weaker acid than CO2 and is in line with
literature data regarding the experimental heat of absorption in
amine solution.26 Interestingly, the shape of the H2S adsorption
isotherm (Figure 1) at higher pressure indicates that H2S physisorption is more strongly dependent on pressure than CO2. This is
also consistent with the relative slopes of the H2S and CO2
adsorption isotherms on pure PE-MCM-41 silica (Figure S3 in
Supporting Information). Thus, the combination of weaker
chemisorption and stronger physisorption for H2S compared to
CO2 is at the origin of the crossover that takes place at ca. 0.4 bar
(Figure 1).
Regenerability and Stability of TRI-PE-MCM-41 in the
Presence of CO2 and H2S. It is worth noticing that the
equilibrium adsorption capacity of CO2 at 0.15 bar partial
pressure and 328 K on TRI-PE-MCM41 using a vacuum temperature regeneration operation at 423 K is ca. 73 mg/g (0.21 mol
CO2/N).10 versus ca. 0.4 mol CO2/N for the saturation absorption
capacity of 15 wt % MEA,27 30 wt % DEA,28 and 30 wt %
MDEA29 at 328 K and 0.15 bar CO2 partial pressure. However,
not only are H2S and CO2 adsorption capacity and selectivity of
TRI-PE-MCM-41 important, but also the stability of the adsorbent and its ability to be recycled repeatedly without adverse
effects are also equally important. As reported earlier,10 TRI-PEMCM-41 exhibited excellent stability during CO2 adsorption-desorption cycling using vacuum temperature regeneration.10 Figure 2 shows pure H2S uptake on TRI-PE-MCM-41
at 298 K and 1 bar during adsorption-desorption cycling with
regeneration under vacuum (5  10-3 mbar) at 348 K. It is seen
that the material does not undergo any loss of performance,
indicating that H2S adsorption is completely reversible and that
TRI-PEMCM-41 is equally stable under H2S adsorptiondesorption cycling. Interestingly, the adsorption of pure CO2 at
298 K after H2S cycling led to an uptake of 2.53 mmol/g, identical
to the CO2 uptake (2.56 mmol/g) over fresh TRI-PE-MCM-41.
This finding indicates that while competing for adsorption sites
they may coadsorb simultaneously or sequentially without any
loss of performance.
Adsorption of CO2/CH4 Mixtures under Dry and Humid
Conditions. To assess the effect of competition between CO2
and CH4 for adsorption on TRI-PE-MCM-41, column breakthrough experiments were carried out using CH4/CO2 mixtures with concentrations akin to those of natural gas and
biogas. Figure 3 shows the breakthrough curves of CH4
and CO2 using a 1:99 CO2/CH4 mixture at 298 K and 1 bar
in the presence of TRI-PE-MCM-41 pretreated with nitrogen
at 423 K.
(26) Le Bouchelec, E. B.; Mougin, P.; Bareau, A.; Solimando, R. Energy Fuels
2007, 21, 20442055.
(27) Dugas, R.; Rochelle, G. Energy Procedia 2009, 1, 11631169.
(28) Seo, D. J.; Hong, W. H. J. Chem. Eng. Data 1996, 41, 258260.
(29) Shen, K. P.; Li, M. H. J. Chem. Eng. Data 1992, 37, 96100.

Langmuir 2009, 25(23), 1327513278

Belmabkhout et al.

Figure 2. Pure H2S adsorption-desorption cycles on TRI-PEMCM-41 at 298 K and 1 bar using vacuum temperature regeneration at 348 K.

Figure 3. Column breakthrough curves for 1:99 CO2/CH4 at

Letter

Figure 4. Column breakthrough curves for 50:50 CO2/CH4 at

298 K and 1 bar over TRI-PE-MCM-41.

Figure 5. Column breakthrough curves for 50:50 CO2/CH4 with

27% RH over TRI-PE-MCM-41 at 298 K and 1 bar.

298 K and 1 bar.

As expected from the adsorption isotherms of pure CO2 and

CH4, methane appeared downstream in the column immediately
after the process had started, suggesting that the adsorption
capacity for CH4 is very small or nonexistent. It is hence inferred
that the selectivity of CO2 over CH4 is extremely high. No CO2
was detected downstream in the column until a breakthrough
time of 24 min. Figure 3 shows that the breakthrough curve of
CO2 is very steep, reaching 80% of the upstream concentration in
a few seconds. After 40 min, the bed was completely saturated,
corresponding to a final CO2 dynamic adsorption capacity of
1.89 ( 0.19 mmol/g. Figure S5 in Supporting Information shows
that the gravimetric capacity obtained at a similar partial pressure
was somewhat smaller (ca. 1.6 mmol/g). Because of the very low
CH4 adsorption capacity (Figures 1 and 3), an estimation of CH4
uptake with any degree of accuracy is hardly possible. Thus, no
attempt was made to quantify the extremely high selectivity of the
material for CO2 versus CH4. Similar work dealing with H2S/CH4
mixtures is in progress. However, on the basis of Figures 1 and S4
(Supporting Information), it inferred that the H2S versus
CH4 selectivity is equally high. The above findings indicate that
TRI-PE-MCM-41 is a viable material for the removal of acid
gases from methane-rich mixtures such as natural gas.
Figure 4 shows the breakthrough profile for CH4 and CO2 in
the presence of 50:50 CO2/CH4 at 298 K and 1 bar over TRI-PEMCM-41 pretreated with helium at 373 K. Here also, methane
Langmuir 2009, 25(23), 1327513278

was detected in the column downstream within seconds, indicating that the adsorption capacity for CH4 is very small or nonexistent. No CO2 was detected up to 55 s. The breakthrough of
CO2 was abrupt and the saturation of the packed column corresponded to a CO2 adsorption capacity of 2.23 ( 0.22 mmol/g, in
good agreement with the equilibrium capacity shown in Figure 1
(2.33 mmol/g at 0.5 bar), although the activation temperature was
slightly lower. Figures 3 and 4 show that the selectivity of CO2 over
CH4 was very high regardless of the CO2 concentration and the
regeneration temperature, within the range of thermal stability of
the adsorbent. Thus, TRI-PE-MCM-41 may also be used for the
bulk separation of CO2 from biogas.
Whether dealing with biogas or natural gas, the feed is likely to
contain different amounts of moisture. To assess the effect of
humidity on CO2 removal, column breakthrough experiments
were carried using 50:50 CO2/CH4 with 27% relative humidity
(RH) at 298 K and 1 bar (Figure 5) and were compared to the
corresponding data under dry conditions shown in Figure 4.
Notice that both experiments were carried out using the same
adsorption column. As under dry conditions, CH4 was detected
downstream in the column immediately, consistent with a negligible adsorption capacity for CH4 and a high selectivity of CO2
over CH4 in the presence of moisture. The breakthrough time of
CO2 was 60 s versus 55 s for the experiment under dry conditions.
Saturation of the bed occurred at a final CO2 dynamic adsorption
capacity of ca. 2.28 ( 0.23 mmol/g. Although a minute amount of
DOI: 10.1021/la903238y

13277

Letter

Belmabkhout et al.

in a forthcoming contribution. Nevertheless, the CO2 versus H2S

selectivity is not mandatory to demonstrate the proof of concept
of using amine-grafted PE-MCM-41 for the removal of both acid
gases. Depending on their relative concentration, CO2 and H2S
may be removed simultaneously or sequentially. Because the
concentration of CO2 is typically much larger than that of H2S
and its interaction with the surface is stronger, it is conceivable
that CO2 is removed first and H2S is removed in an additional
column when the CO2 concentration decreases.

Conclusions

Figure 6. Adsorption isotherm for H2O vapor over TRI-PEMCM-41 at 298 K.

water was detected since the beginning of the experiment, no

significant concentration of water vapor was observed in the
effluent before ca. 1100 s. The total uptake of water vapor was ca.
2.7 mmol/g, in good agreement with the water adsorption
capacity at equilibrium (2.96 mmol/g) determined gravimetrically
as shown in Figure 6. These findings clearly show that the CO2
and CH4 adsorption properties of TRI-PE-MCM-41 were not
affected by the presence of moisture with an increasing tendency
in terms of CO2 adsorption capacity.11 Literature data16 also
indicate that moisture has no effect on the adsorption of H2S.
Thus, water can be removed along with CO2 and eventually H2S.
Unlike physical adsorbents such as zeolites and activated
carbons, the CO2 adsorption on amine-grafted PE-MCM-41 is
enhanced in the presence of moisture.11,19 The promoting effect
may be explained on the basis of the generally accepted reaction
mechanisms between CO2 and amines (i.e., the formation of
carbamate under dry conditions and the partial formation of
bicarbonate in the presence of moisture).11,16 As reported in our
previous contribution using monoamine-grafted PE-MCM-41,
CO2 uptake increased with the RH of the feed stream.11
CO2/H2S Selectivity. Although CH4 may be regarded as a
nonadsorbing gas, it is clear that because CO2 and H2S interact
strongly with the amine groups they will compete for adsorption
on the same sites. Thus, a knowledge of the CO2 versus H2S
selectivity on TRI-PE-MCM-41 in a relevant range of conditions
is necessary for process design. Whether the CO2 versus H2S
selectivity is higher or lower than 1 will certainly affect the design
of biogas and natural gas upgrading processes using TRI-PEMCM-41. Detailed work on CO2/H2S selectivity will be reported

13278 DOI: 10.1021/la903238y

On the basis of single-component CO2, H2S, and CH4 adsorption measurements, it has been demonstrated that TRI-PEMCM-41 exhibits high affinity toward both CO2 and H2S versus
CH4. It was also found that TRI-PE-MCM-41 exhibits extremely
high CO2 selectivity over methane, regardless of the CO2 concentration and the occurrence of moisture. Moreover, on the basis
of single-component adsorption data, it was inferred that (i) the
affinity toward H2S versus CH4 is also very high, (ii) there exists a
pressure beyond which the adsorbent affinity switches in favor of
H2S versus CO2, (iii) like CO2, H2S adsorption is completely
reversible using vacuum temperature regeneration at 348 K, and
(iv) depending on the feed composition, H2S and CO2 may be
removed over TRI-PE-MCM-41 simultaneously or sequentially.
These are key findings for the treatment of methane-containing
industrial gases such as biogas and natural gas. Such applications
require highly selective adsorbents for both CO2 and H2S with
tunable relative affinity toward these adsorbates, such as TRI-PEMCM-41.
Acknowledgment. We thank co-workers Rodrigo SernaGuerrero for providing the column breakthrough data and Lomig
Hamon (Universite de Mons, Belgium) for providing the H2S
adsorption-desorption data. The financial support of the Natural Science and Engineering Council of Canada (NSERC) is
acknowledged. Y.B. thanks the NSERC for a postdoctoral
fellowship, and A.S. thanks the Federal Government of Canada
for the Canada Research Chair in Nanostructured Materials for
Catalysis and Separation (2001-2015).
Supporting Information Available: Details of material
synthesis, experimental methods, adsorption isotherms for
H2S, CO2, and CH4 on amine-functionalized and non-functionalized PE-MCM-41, and adsorption isotherms for H2S
and CO2 on TRI-PE-MCM-41 at different regeneration
temperatures. This material is available free of charge via
the Internet at http://pubs.acs.org.

Langmuir 2009, 25(23), 1327513278