Microporous and Mesoporous Materials 152 (2012) 246–252

Contents lists available at SciVerse ScienceDirect

Microporous and Mesoporous Materials

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

Adsorption of CO, CO2 and CH4 on Cu-BTC and MIL-101 metal organic frameworks:Effect of open metal sites and adsorbate polarity

Pradip Chowdhury a, Samuel Mekala a, Frieder Dreisbach b, Sasidhar Gumma a,⇑a Department of Chemical Engineering, Indian Institute of Technology Guwahati, Guwahati 781039, Indiab Rubohtherm GmbH, Universitätsstraße 142, D-44799 Bochum, Germany

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

Article history:Received 10 September 2011Received in revised form 9 November 2011Accepted 14 November 2011Available online 20 November 2011

Keywords:Cu-BTCMIL-101Virial-Langmuir modelDual Site Langmuir modelElectrostatic interactions

1387-1811/$ - see front matter � 2011 Elsevier Inc. Adoi:10.1016/j.micromeso.2011.11.022

⇑ Corresponding author. Tel.: +91 361 2582261; faxE-mail address: s.gumma@iitg.ernet.in (S. Gumma

A comparative adsorption study of three gases viz. CO, CO2 and CH4 on two adsorbents viz. Cu-BTC (orHKUST-1) and Cr-BDC (or MIL-101) is reported in this article. The gravimetric adsorption equilibriummeasurements on the samples were performed in a Rubotherm magnetic suspension balance at three dif-ferent temperatures: 295, 318 and 353 K and pressures ranging between 0 and 100 bar. Virial-Langmuirmodel was used to model the experimental data on Cu-BTC, whereas Dual Site Langmuir (DSL) model wasused for adsorption on MIL-101. For all gases the enthalpy of adsorption at low loading was higher onMIL-101 than that on Cu-BTC, indicating the availability of open metal sites in case of MIL-101. Moreover,a sharp decrease in enthalpy of adsorption is observed in case of MIL-101, whereas only a moderatedecrease is observed in case of Cu-BTC. CO has a large Henry’s constant on MIL-101, whereas at higherpressures, the solid exhibits better capacity for CO2. In case of Cu-BTC, CO2 has a higher capacity onthe adsorbent as compared to the other two gases throughout the entire range of pressures studied.All the experimental data is critically analyzed by examining the role of open metal centers, adsorbatepolarity and the effect of temperature on the electrostatic interactions.

� 2011 Elsevier Inc. All rights reserved.

1. Introduction

‘‘Metal Organic Frameworks’’ or MOFs is a class of novel adsor-bents that is being widely investigated in recent years. The diver-sity in their structures results from co-ordination betweeninorganic metal atoms and organic ligands or linkers. Appropriatechoice of metal atoms and organic ligands impart key characteris-tics to MOFs including tunable pore size, pore volume and geome-try. Attempts have also been made to use functionalized organicligands to meet specific applications (CO2 adsorption for example)[1–6]. Potential applications that are being investigated by the re-search community include adsorptive gas separation/purification[7–12], high pressure gas storage [13–23] and catalysis [24–30].

The MOFs used in this work Cu-BTC and MIL-101 (Cr-BDC) havebeen widely investigated in literature. They were first reported byChui et al. [30] and Férey et al. [31], respectively. Earlier studies onthese frameworks have focused on structural aspects of theseadsorbents [32], improvement of synthesis procedures [25,33],H2 adsorption [17–23] and adsorption of gases like Ar [2,33], N2

[2–4,33,34], O2 [2,3,34], CO [3,35], CO2 [1–6], CH4 [1,3,4,6,16,34,36,37], C3H8 [2,6] and SF6 [2,6]. Molecular simulations explaining‘‘host–guest’’ interactions were explained in detail by various re-search groups [7,8,11,12].

ll rights reserved.

: +91 361 2582291.).

In our previous works [2,6], we have reported adsorption equi-libria of several gases on Cu-BTC and Cr-BDC frameworks up tomoderate pressures (about 10 bar). Both the frameworks showeddifferent characteristics with MIL-101 showing marked heteroge-neity due to presence of coordinatively unsaturated metal centers.While such centers are also present in Cu-BTC, they are not readilyaccessible most likely due to left-over solvent molecules from syn-thesis procedure [33]; treatment procedures were not effective toremove all these solvent molecules.

In this work, we present a comparative study of CH4, CO2 andCO adsorption on these two frameworks i.e. Cu-BTC and MIL-101up to high pressures and at three different temperatures. The ratio-nale behind choice of gases was twofold. On one hand, they havemarkedly different polarities; CH4 is non-polar, CO2 has a quadru-pole moment while CO has permanent dipole (Table 1). Differencesin electrostatic nature of the gases and frameworks should beinteresting from a fundamental aspect. On the other hand, thegases are of interest from an industrial view point. Separation ofCO2/CH4 mixture is important for natural gas processing; whilemixtures of CO/CH4 and/or CO2 are found in a variety of industrialoff gases, and in process streams from steam reforming, coal gasi-fication, partial oxidation of hydrocarbons, etc. [38]. In addition, sofar, only limited data was available in literature for CO adsorptionon MOFs; high pressure experimental data for CO is rare. Wanget al. [3] carried out CO adsorption study on Cu-BTC at 295 K andup to 1 bar pressure and compared the results with CO2, C2H4

Table 1Adsorptive gas properties.

Gas Mol. wt. Liquid molar volumea (cm3 mol�1) Kinetic diameter (Å) Polarizability (�10�25 cm3) Dipole moment(�1018 esu-cm)

Quadruple moment(�10�26 esu-cm2)

CO 28 33.0 3.76 19.5 0.112 2.5CO2 44 33.3 3.3 26.5 0.00 4.30CH4 16 37.7 3.80 26.0 0.00 0.00

a At normal boiling point.

P. Chowdhury et al. / Microporous and Mesoporous Materials 152 (2012) 246–252 247

and CH4. Karra and Walton [38] studied the effect of open metalsites of Cu-BTC on adsorption of polar (CO) and non-polar (CH4)molecules using Grand Canonical Monte Carlo (GCMC) simulations.In a recent paper [35] they also report CO, CO2 and N2 on variousMOFs both from experiments and GCMC simulations. High pres-sure adsorption data of CO in MIL-101 is unavailable in literatureto the best of our knowledge.

While some of the data is available in literature (especially at lowpressures), a systematic investigation on the role of adsorbate polar-ity and the structural properties of these two frameworks (coordin-atively unsaturated metal centers, pore volume, etc.) is unavailable.We believe that such studies are essential in order to fine tune andsynthesize MOFs with desired adsorption properties.

2. Experimental

2.1. Synthesis and characterization

Standard recipes from literature were followed for synthesizingMOF samples. Cu-BTC was synthesized following the publishedwork of Liu et al. [33]. MIL-101 was synthesized following the ori-ginal work of Férey et al. [31]. Detailed synthesis and post treat-ment procedures along with various characterization steps werealready reported in our previous articles for Cu-BTC (sample B)[2] and MIL-101 [6]. Table 2 summarizes the results of N2 adsorp-tion isotherm at 77 K on both adsorbents.

6

9

12

unt a

dsor

bed,

N /

mm

ol g

-1

2.2. Adsorption equilibrium measurements

The adsorption isotherms at three different temperatures viz.295, 318 and 353 K were measured gravimetrically in a Rubothermmagnetic suspension balance. The purity of all the gases used forthis study was above 99.9% or better. We followed the usual proce-dure for adsorption equilibrium measurements which is well doc-umented in literature [39]. The samples were activated at about403 K after carrying out detailed thermo-gravimetric analysis[2,6] while the initial activation was done for about 8 h, subse-quent activations (in between isotherm measurements) wereshorter (about 2–3 h each). It was however ensured that the origi-nal weight of sample is restored in each case. The excess amountadsorbed was calculated from the raw measurements using usualbuoyancy corrections. The buoyancy volume required for the cor-rections was obtained using helium as a non-adsorbing referencegas at 295 K. The density of the bulk gas needed for buoyancy cor-rections was obtained using the weight change of the sinker pro-vided in the balance.

Table 2Isotherm properties of the adsorbents from N2 experiment at 77 K.

Adsorbents BET surface area (m2 g�1) Isotherm type (IUPAC)

Cu-BTC (HKUST-1) 1663 ICr-BDC (MIL-101) 2674 IV

3. Results and discussion

3.1. Isotherms

Figs. 1–3 show the isotherms for CH4, CO2 and CO on Cu-BTC,respectively; the corresponding isotherms on MIL-101 are shownin Figs. 4–6. Following the usual procedure to account for bulk-phase non-ideality [39], fugacity was used to represent these iso-therms instead of pressure. Even at pressures as high as 50 barand 295 K, the loading for CH4 on both Cu-BTC and MIL-101 ishigher than those on activated carbons (BPL, Norit) [40], purelysiliceous zeolite like silicalite [41,42] and polar zeolites like 13X[43,44], 5A [43], NaY [45] and MgY [45]. These results for Cu-BTCalso compare well and are close to those reported earlier in litera-ture [34]. However, for MIL-101, the values are about 10–15%lower than those reported by Llewellyn et al. [1] on their sampleMIL-101b.

The CO2 loading on both samples is higher than those on bench-mark zeolites like 13X at around ambient temperature [5]. Thistrend holds even at higher pressures. CO2 loading on MIL-101 mea-sured in this work, was about 15–35% lower than that reported onMIL-101 samples (a, b and c) by Llewellyn et al. [1]. CO adsorptionon MIL-101 at about ambient temperature, is better than that onpurely siliceous zeolite (7 bar) [46], and comparable to that on po-lar zeolite like 5A (10 bar) [47]. CO adsorption on Cu-BTC is betterthan that on MIL-101, above atmospheric pressures. CO loading onCu-BTC measured in this work is slightly better than that reportedearlier by Wang et al. [3] for their sample b at 295 K; while theymeasured 0.8 mmol g�1 of loading (1 bar); loading on our samplewas 1.4 mmol g�1 at 1.04 bar. Our results for both CO and CH4 alsocompare well with GCMC simulations of Karra and Walton [38] onCu-BTC for pressures up to 10 bar. At higher pressures, GCMC sim-ulations over predict the adsorption of methane and under predictthe adsorption of CO.

0

3

0 20 40 60 80 100

Am

o

Fugacity, f / bar

Fig. 1. Adsorption isotherms for CH4 on Cu-BTC. Symbols: experimental points;lines: Virial-Langmuir model fits; squares: 295 K; triangles: 318 K; circles: 353 K.

0

3

6

9

12

15

18

0 10 20 30 40 50

Am

ount

ads

orbe

d, N

/ mm

ol g

-1

Fugacity, f / bar

Fig. 2. Adsorption isotherms for CO2 on Cu-BTC. Symbols: experimental points;lines: Virial-Langmuir model fits; squares: 295 K; triangles: 318 K; circles: 353 K.

0

2

4

6

8

10

12

0 20 40 60 80

Am

ount

ads

orbe

d, N

/ mm

ol g

-1

Fugacity, f / bar

Fig. 3. Adsorption isotherms for CO on Cu-BTC. Symbols: experimental points;lines: Virial-Langmuir model fits; squares: 295 K; triangles: 318 K; circles: 353 K.

0

2

4

6

8

10

12

14

0 20 40 60 80 100

Am

ount

ads

orbe

d, N

/ mm

ol g

-1

Fugacity, f / bar

Fig. 4. Adsorption isotherms for CH4 on MIL-101. Symbols: experimental points;lines: Dual Site Langmuir model fits; squares: 295 K; triangles: 318 K; circles:353 K.

0

5

10

15

20

25

0 10 20 30 40 50 60

Am

ount

ads

orbe

d, N

/ m

mol

g-1

Fugacity, f / bar

-2

-1

0

1

2

0 5 10 15 20 25

ln [

(f /

bar

) / (

N/ m

mol

g-1

)]

Amount adsorbed, N / mmol g-1

Fig. 5. Adsorption isotherms for CO2 on MIL-101 (in both conventional and virialdomains). Symbols: experimental points; lines: Dual Site Langmuir model fits;squares: 295 K; triangles: 318 K; circles: 353 K.

0

1

2

3

4

5

6

7

0 20 40 60 80

Am

ount

Ads

orbe

d, N

/ mm

ol g

-1

Fugacity, f / bar

Fig. 6. Adsorption isotherms for CO on MIL-101. Symbols: experimental points;lines: Dual Site Langmuir model fits; squares: 295 K; triangles: 318 K; circles:353 K.

248 P. Chowdhury et al. / Microporous and Mesoporous Materials 152 (2012) 246–252

3.2. Modeling

Isotherms of all three gases on Cu-BTC studied in this work ap-proach saturation in the high pressure region. Virial-Langmuirequation is often helpful to model behavior in such systems [48]:

f ¼ NmaxNbðNmax � NÞ

expðbN þ cN2Þ ð1Þ

where b is Henry constant, Nmax is saturation capacity, b and c arethe second and third virial coefficients, respectively. The usual tem-perature dependency was considered for these parameters:

b ¼ b0expb1

T

� �ð2Þ

P. Chowdhury et al. / Microporous and Mesoporous Materials 152 (2012) 246–252 249

b ¼ b0 þb1

Tð3Þ

c ¼ c0 þc1

Tð4Þ

The saturation capacity Nmax is also expressed by a similar tem-perature dependency:

Nmax ¼ g0 þg1

Tð5Þ

In case of adsorption on MIL-101 the virial domain plots for allthe three gases show a knee in the low loading region, which is anindication of heterogeneity of the adsorbent. As an example, virialdomain plot for CO2 on MIL-101 is shown in Fig. 5. However, thiseffect is obviously less pronounced for non-polar molecule likeCH4. A Dual Site Langmuir (DSL) model equation [49] was usedto model the isotherms for all three gases on MIL-101:

N ¼ Nmax1 b1P

1þ b1Pþ Nmax

2 b2P1þ b2P

ð6Þ

where Nmaxi and bi denotes saturation capacity and affinity parame-

ters for sites of type i. The temperature dependency is includedthrough affinity parameters via:

bi ¼ b0i exp

�ei

R1T� 1

T0

� �� �ð7Þ

where b0i is the affinity at reference temperature T0 and ei is the en-

thalpy of adsorption on site i with respect to temperature T0. TheHenry’s constant b in this case is given by:

b ¼ Nmax1 b1 þ Nmax

2 b2 ð8Þ

The results of the modeling along with the fit parameters are gi-ven in Tables 3 and 4, respectively.

3.3. Enthalpy of adsorption

The enthalpy of adsorption can be readily derived from the iso-therm models. Accordingly, Eqs. (9) and (10) were used for calcu-

Table 3Virial-Langmuir model fit parameters for CH4, CO2 and CO adsorption on Cu-BTC.

Parameters Gases

CH4 CO2 CO

ln(b0) –6.39 –7.48 –8.89b1 (K) 1931 2774 2810b0 (mmol�1 g) 0.24 0.45 �0.13b1 (mmol�1 g K) �59.3 �157.5 77.24c0 (mmol�2 g2) �0.068 �0.069 �0.086c1 (mmol�2 g2 K) 17.2 21.8 23.4g0 (mmol g�1) �4.02 �8.29 �8.98g1 (mmol g�1 K) 4635 7643 6003

Table 4Dual Site Langmuir model fit parameters for CH4, CO2 and CO adsorption on MIL-101.

Parameters Gases

CH4 CO CO2

Nmax1 (mmol g�1) 0.072 0.608 0.974

Nmax2 (mmol g�1) 15.68 9.62 34.04

b01 (bar�1) 6.48 66.2 14.04

b02 (bar�1) 0.025 0.03 0.061

�Dhð1Þads (kJ mol�1) 30.56 48.16 34.67

�Dhð2Þads (kJ mol�1) 13.45 14.84 20.79

Reference temperature T0 = 283 K for all gases.

lation of enthalpy using Virial-Langmuir and Dual Site Langmuirmodel:

Dhads

R¼ b1 þ b1N þ c1N2 þ g1

Nmax �g1

Nmax � Nð9Þ

Dhads

R¼ e1Nmax

1 b1ð1þ b2PÞ2 þ e2Nmax2 b2ð1þ b1PÞ2

Nmax1 b1ð1þ b2PÞ2 þ Nmax

2 b2ð1þ b1PÞ2ð10Þ

The enthalpy at zero loading for both frameworks is shown inFig. 7. For the three gases considered, the enthalpy of adsorptionin this region is higher on MIL-101 than that on Cu-BTC. This isdue to presence of coordinatively unsaturated metal centers incase of MIL-101. CO with its dipole, strongly interacts with thesemetal centers in MIL-101 and exhibits markedly higher adsorptionenthalpy than either CO2 or CH4. While such centers are also pres-ent in case of Cu-BTC framework [50], they are either not open ormay be hindered due to presence of left-over solvent moleculesfrom the synthesis procedure. This results in only small differencesin adsorption enthalpies for the three gases in case of Cu-BTC.

Fig. 8 shows variation of enthalpy of adsorption at 295 K for CH4,CO2 and CO on Cu-BTC. The enthalpy of adsorption for CO2 matcheswith the earlier reports of Wang et al. [3] (�25 kJ mol�1) at highloading, although they reported a value of about �35 kJ mol�1 at

Fig. 7. Comparison of enthalpy of adsorption at zero loading of CO, CO2 and CH4 onCu-BTC and MIL-101.

-30

-25

-20

-15

-10

-5

00 3 6 9 12

Δhad

s/

kJ m

ol-1

Amount Adsorbed, / mmol g-1

Fig. 8. Variation of enthalpy of adsorption with loading on Cu-BTC at 295 K.

-50

-40

-30

-20

-10

00 1 2 3 4

Δh a

ds/

kJ m

ol-1

Amount adsorbed, N / mmol g-1

CO

CO2

CH4

Fig. 9. Variation of enthalpy of adsorption with loading on MIL-101 at 295 K.

0

0.4

0.8

1.2

1.6

2

0 1 2 3

Am

ount

ads

orbe

d, N

/ m

mol

g-1

Fugacity, f / bar

Fig. 11. CO adsorption on Cu-BTC (closed symbols) and MIL-101 (open symbols) at295 K in the low pressure region.

250 P. Chowdhury et al. / Microporous and Mesoporous Materials 152 (2012) 246–252

zero loading. Both CH4 and CO2 only show modest variation in en-thalpy with loading; CO on the other hand shows a considerable de-crease in enthalpy of adsorption from about �23.4 kJ mol�1 at zeroloading to about �17.4 kJ mol�1 at a loading of ca. 6 mmol g�1.Although electrostatic interactions dominate the low loading region,as these sites are progressively filled up, the enthalpy of adsorptiondrops down sharply.

As more open metal sites are available [1], the above observationsfor enthalpy variations are pronounced in case of MIL-101 (Fig. 9)compared to Cu-BTC. In the low loading region, CO has the highestenthalpy due to its dipole, followed by CO2. As the open metal sitesfill up, a sharp decline in adsorption enthalpy is prominent; by about1.2 mmol g�1 this value drops to �15.1, �13.7 and �24.2 kJ mol�1

for CO, CH4 and CO2, respectively.

3.4. Effect of adsorbate polarity and electrostatic interactions onHenry’s constants

3.4.1. Comparison of Henry’s constants on MIL-101 and Cu-BTCFig. 10 shows a plot of Henry’s constant (b) against temperature

for the three gases on both the adsorbents. Several interesting fea-tures can be observed. First, at all temperatures b for CH4 is higherfor Cu-BTC than on MIL-101, indicating that the van der Waalsinteractions are stronger in case of Cu-BTC. The difference betweenthe Henry’s constant of methane on the two frameworks is fairly

-4

-2

0

2

4

2.8 2.9 3 3.1 3.2 3.3 3.4 3.5

ln H

, H

in m

mol

g-1

bar

-1

1000 / T , K-1

Fig. 10. Henry constant vs. temperature. Circles (CO2); triangles (CH4); squares(CO); open symbols MIL-101; closed symbols Cu-BTC. Lines are drawn as a guide tothe eye. Note that temperature scale is increasing from right to left.

constant with temperature. On the other hand, CO2 has aquadrupole and its electrostatic interactions with the open metalsites in MIL-101 result in a higher value of b at 295 K (comparedto Cu-BTC). But as temperature increases, these electrostatic inter-actions become weaker and dispersion interactions dominate; thisresults in a higher value of b for CO2 on Cu-BTC at 353 K (comparedto MIL-101). The electrostatic interactions of CO (due to its dipole)with the MIL-101 framework are stronger than that of CO2. Due tothis reason, even at 353 K CO has a larger Henry’s constant on MIL-101 than on Cu-BTC; nevertheless, as in case of CO2, as tempera-ture increases the difference between the Henry’s constants ofCO on the two frameworks decreases.

3.4.2. Comparison of Henry’s constants for different gases on MIL-101In case of MIL-101, at the lowest temperature (295 K) among

the three gases considered, CO has the largest Henry’s constantdue to its dipole, closely followed by CO2 (Fig. 10). As temperatureincreases electrostatic interactions become weaker and by 353 K,CO2 has a larger value of Henry’s constant compared to CO. Onthe other hand, CH4 has negligible electrostatic interactions withthe open metal sites in MIL-101 and has a smaller Henry’s constantcompared to the other two gases. However, with increase in tem-perature, electrostatic interactions (of CO2 and CO) weaken andthe difference in Henry’s constants of CH4 and CO2 (or CO)decreases.

3.4.3. Comparison of Henry’s constants for different gases on Cu-BTCIn case of Cu-BTC, although open metal sites exist in theory,

they may be hindered or not accessible to the gas molecules as dis-cussed earlier; hence the effect of electrostatic interactions is notpronounced as in case of MIL-101. Nevertheless, at the lowest tem-perature (295 K) CO exhibits slightly higher Henry’s constant thanCH4 (Fig. 11); the difference however reduces as temperatureincreases.

3.5. Effect of adsorbate polarity and electrostatic interactions onloading at higher pressures

While the previous section discusses the effect of electrostaticinteractions on the two frameworks in the low pressure region, thissection focuses on the loading at higher pressures. An attempt willbe made highlight the role of electrostatic interactions in theadsorption behavior on these two MOF frameworks; this will beexplained by comparing the loading (in the high pressure region)of a select gas on both the frameworks and also by comparingthe loading of the three gases on each one of the MOF frameworks.

-3

-1

1

3

0 4 8 12 16

ln [

(f/b

ar)

/ (N

/ mm

ol g

-1)]

Amount Adsorbed, N / mmol g-1

Fig. 12. Comparison of adsorption isotherms of CO2 (circle), CO (squares) and CH4

(triangles) on Cu-BTC framework at 295 K (virial domain plot).

-3

-2

-1

0

1

2

3

0 5 10 15 20 25

ln[(

f / b

ar)

/ (N

/ m

mol

g-1

)]

Amount Adsorbed, N / mmol g-1

Fig. 13. Comparison of adsorption isotherms of CO2 (circle), CO (squares) and CH4

(triangles) on MIL-101 framework at 295 K (virial domain plot).

P. Chowdhury et al. / Microporous and Mesoporous Materials 152 (2012) 246–252 251

3.5.1. Comparison of adsorption of a gas on MIL-101 and Cu-BTCAs discussed earlier, Henry constant for CH4 on Cu-BTC is larger

at 295 K indicating that adsorption capacity of Cu-BTC is larger nearthe zero pressure limit; this trend continues even in the high pres-sure region up to fugacity as high as 80 bar (Figs. 1 and 4). However,the difference between the adsorption capacities of the two frame-works at high pressures is smaller. While the adsorption on Cu-BTCseems to approach saturation, loading on MIL-101 does not; this isto be expected since MIL-101 has a large open porous structure,with a unit cell volume of 702,000 Å3 and its measured pore volumefrom N2 experiments was 1.38 cm3 g�1 [32]. Cu-BTC on the otherhand has much smaller pore volume (ca. �0.75 cm3 g�1). Earlierworks indicate the importance of pore volume at high loadings [32].

In case of CO, Henry’s constants (Fig. 11) indicate that thecapacity in the low loading region will be higher on MIL-101 dueto electrostatic interactions. However, as the metal centers areoccupied, the contribution of electrostatic interactions decreasesand that of dispersion forces dominates; hence, the isotherms ofCO on Cu-BTC and MIL-101 at 295 K cross each other at about aloading of 0.75 mmol g�1. Thereafter, loading of CO on Cu-BTC isgreater (Fig. 11) compared to that of MIL-101. Unlike in case ofCH4, adsorption of CO on Cu-BTC (and MIL-101) does not seem toapproach saturation even at 70 bar (Figs. 3 and 6); this is due tosmaller polarizability of CO. In absence of electrostatic interactions,the smaller polarizability results in its lower capacity for CO com-pared to that of CH4, at the same condition.

Finally, in case of CO2 the behavior in the low loading region is sim-ilar to that of CO. Henry constant value on MIL-101 is higher due to itsquadrupole moment. However, as metal centers are occupied in theMIL-101 framework, the contribution of electrostatic interactions de-creases and dispersion interactions become dominant. Above0.05 bar (at a loading of ca. 0.35 mmol g�1) capacity on Cu-BTC ishigher that on MIL-101, until 16 bar. By about 16 bar, loading onCu-BTC approaches saturation. However, due to availability of largepore volume, loading on MIL-101 does not approach saturation andkeeps increasing. In fact, at a fugacity of 36 bar loadings on MIL-101and Cu-BTC are 21.3 and 15.5 mmol g�1, respectively.

Similar trends are observed at other temperatures also; how-ever, as noted earlier the effect of electrostatic interactions isprominent at lower temperatures and hence isotherms at 295 Kwere chosen to illustrate the behavior.

3.5.2. Comparison of adsorption the three gases on Cu-BTCFig. 12 shows loading for the three gases on Cu-BTC at 295 K, in

the virial domain. Throughout the experimental data range adsorp-tion of CO2 on the framework is higher (bottom most curve in thevirial domain plot). However, between CO and CH4, as discussedearlier the Henry’s constant for CO is higher at 295 K than that

on CH4 due to its dipole moment; as sites with which electrostaticinteractions occur are progressively filled up, dispersion interac-tions dominate and larger polarizability of CH4 results in its highercapacity; the isotherms for CO and CH4 cross each other at aloading of about 5 mmol g�1. As the influence of electrostatic inter-actions at higher temperature is weaker, the loading at which COand CH4 isotherms cross each other shifts to left (at 353 K this va-lue is about 0.22 mmol g�1).

3.5.3. Comparison of adsorption the three gases on MIL-101This effect was also observed in case of MIL-101 framework

(Fig. 13). Since the electrostatic interactions with the MIL-101framework are stronger as discussed earlier, CO has highest Henryconstant value amongst the three gases. However at higher load-ings, the capacity for CO drops significantly due to its low polariz-ability and its isotherms cross the isotherms of both CO2 and CH4 at295 K. The difference between the Henry’s constants of the threegases decreases as temperature increases to 353 K; however, theelectrostatic interactions of CO with the framework are still strongenough to have a high Henry constant value even at this tempera-ture compared to the other two gases. As in case of Cu-BTC, theloadings at which CO isotherm at 353 K, crosses the isothermsfor CH4 and CO2 shifts slightly to left (compared to that of 295 K).

4. Conclusions

In this work, we have presented a comparative high pressureadsorption study of industrially relevant gases viz. CH4, CO2 andCO at three different temperatures: 295, 318 and 353 K on Cu-BTC and MIL-101 frameworks. Virial-Langmuir model was usedto model data on Cu-BTC, whereas Dual Site Langmuir model ex-plains the adsorption behavior on MIL-101 framework, which isknown to have two distinct sites for adsorption. For all gases theenthalpy of adsorption at low loading was higher on MIL-101 thanthat on Cu-BTC. MIL-101 possesses coordinatively unsaturated me-tal centers and the difference between enthalpies of these threegases is due to differences in interaction of the adsorbates withthese metal centers. While such centers are also present in caseof Cu-BTC framework, they are either not open and may be hin-dered due to presence of left over solvent molecules from the syn-thesis procedure. The availability of more open metal sitespronounces the variation of enthalpy with loading in MIL-101 ascompared to Cu-BTC. At all temperatures Henry constant for CH4

was higher on Cu-BTC, whereas for CO it was higher on MIL-101.For CO2, Henry’s constant was higher on MIL-101 at 295 K, whereas

252 P. Chowdhury et al. / Microporous and Mesoporous Materials 152 (2012) 246–252

at 353 K it was higher on Cu-BTC. Several key features observed incomparison of isotherms of the three gases on the two frameworkswere accounted by examining the role of adsorbate polarity andtemperature on electrostatic interactions.

Acknowledgements

We gratefully acknowledge Rubotherm GmbH for financialsupport. We also acknowledge help from Dr. Biswanath Saha(Scientist, R&D division BPCL, Faridabad, India) for surface areaanalysis. We also acknowledge corrections suggested by Prof.R. Krishna on an unedited version of this manuscript.

References

[1] P.L. Llewellyn, S. Bourrelly, C. Serre, A. Vimont, M. Daturi, L. Hamon, G.D. Weireld,J.-S. Chang, D.-Y. Hong, Y.K. Hwang, S.H. Jhung, G. Férey, Langmuir 24 (2008) 7245.

[2] P. Chowdhury, C. Bikkina, D. Meister, F. Dreisbach, S. Gumma, Micropor.Mesopor. Mater. 117 (2009) 406.

[3] Q.M. Wang, D. Shen, M. Bülow, M.L. Lau, S. Deng, F.R. Fitch, N.O. Lemcoff, J.Semanscin, Micropor. Mesopor. Mater. 55 (2002) 217.

[4] D. Farrusseng, C. Daniel, C. Gaudillere, U. Ravon, Y. Schuurman, C. Mirodatos, D.Dubbeldam, H. Frost, R.Q. Snurr, Langmuir 25 (2009) 7383.

[5] Z. Liang, M. Marshall, A.L. Chaffee, Energy Fuels 23 (2009) 2785.[6] P. Chowdhury, C. Bikkina, S. Gumma, J. Phys. Chem. C 113 (2009) 6616.[7] T. Düren, R.Q. Snurr, J. Phys. Chem. B 108 (2004) 15703.[8] B. Liu, B. Smit, Langmuir 25 (2009) 5918.[9] S. Keskin, D.S. Sholl, J. Phys. Chem. C 111 (2007) 14055.

[10] A. Vishnyakov, P.I. Ravikovitch, A.V. Neimark, M. Bülow, Q.M. Wang, Nano Lett.3 (2003) 713.

[11] S. Wang, Q. Yang, C. Zhong, Sep. Purif. Technol. 60 (2008) 30.[12] Q. Yang, C. Xue, C. Zhong, J.-F. Chen, AIChE J. 53 (2007) 2832.[13] M. Eddaoudi, J. Kim, N. Rosi, D. Vodak, J. Wachter, M. O’Keeffe, O.M. Yaghi,

Science 295 (2002) 469.[14] S. Bourrelly, P.L. Llewellyn, C. Serre, F. Millange, T. Loiseau, G. Férey, J. Am.

Chem. Soc. 127 (2005) 13519.[15] A.R. Millward, O.M. Yaghi, J. Am. Chem. Soc. 127 (2005) 17998.[16] I. Senkovska, S. Kaskel, Micropor. Mesopor. Mater. 112 (2008) 108.[17] N.L. Rosi, J. Eckert, M. Eddaoudi, D.T. Vodak, J. Kim, M. O’Keeffe, O.M. Yaghi,

Science 300 (2003) 1127.[18] J.L.C. Rowsell, A.R. Millward, K.S. Park, O.M. Yaghi, J. Am. Chem. Soc. 126 (2004)

5666.[19] A.G. Wong-Foy, A.J. Matzger, O.M. Yaghi, J. Am. Chem. Soc. 128 (2006) 3494.[20] L. Pan, M.B. Sander, X. Huang, J. Li, M. Smith, E. Bittner, B. Bockrath, J. Karl

Johnson, J. Am. Chem. Soc. 126 (2004) 1308.

[21] G. Férey, M. Latroche, C. Serre, F. Millange, T. Loiseau, A. Percheron-Guégan,Chem. Commun. (2003) 2976.

[22] Y. Li, R.T. Yang, AIChE J. 54 (2008) 269.[23] B. Panella, M. Hirscher, H. Putter, U. Muller, Adv. Funct. Mater. 16 (2006)

520.[24] F.X. Llabres i Xamena, A. Abad, A. Corma, H. Garcia, J. Catal. 250 (2007) 294.[25] K. Schlichte, T. Kratzke, S. Kaskel, Micropor. Mesopor. Mater. 73 (2004) 81.[26] B. Gomez-Lor, E. Gutierrez-Puebla, M. Iglesias, M.A. Monge, C. Ruiz-Valero, N.

Snejko, Chem. Mater. 17 (2005) 2568.[27] C. Janiak, Dalton Trans. (2003) 2781.[28] S.-H. Cho, B. Ma, S.T. Nguyen, J.T. Hupp, T.E. Albrecht-Schmitt, Chem. Commun.

(2006) 2563.[29] L. Alaerts, E. Séguin, H. Poelman, F. Thibault-Starzyk, P.A. Jacobs, D.E. De Vos,

Chem. Eur. J. 12 (2006) 7353.[30] S.S.-Y. Chui, S.M.-F. Lo, J.P.H. Charmant, A.G. Orpen, I.D. Williams, Science 283

(1999) 1148.[31] G. Férey, C. Mellot-Draznieks, C. Serre, F. Millange, J. Dutour, S. Surblé, I.

Margiolaki, Science 309 (2005) 2040.[32] O.I. Lebedev, F. Millange, C. Serre, G. Van Tendeloo, G. Férey, Chem. Mater. 17

(2005) 6525.[33] J. Liu, J.T. Culp, S. Natesakhawat, B.C. Bockrath, B. Zande, S.G. Sankar, G.

Garberoglio, J. Karl Johnson, J. Phys. Chem. C 111 (2007) 9305.[34] E. García-Pérez, J. Gascón, V. Morales-Flórez, J.M. Castillo, F. Kapteijn, S. Calero,

Langmuir 25 (2009) 1725.[35] J.R. Karra, K.S. Walton, J. Phys. Chem. C 114 (2010) 15735.[36] G. Garberoglio, A.I. Skoulidas, J. Karl Johnson, J. Phys. Chem. B 109 (2005)

13094.[37] S. Wang, Energy Fuels 21 (2007) 953.[38] J.R. Karra, K.S. Walton, Langmuir 24 (2008) 8620.[39] O. Talu, Adv. Colloid Interf. Sci. 76–77 (1998) 227.[40] Y. Belmabkhout, G. De Weireld, M. Frère, J. Chem. Eng. Data 49 (2004) 1379

.[41] T.C. Golden, S. Sircar, J. Colloid Interf. Sci. 162 (1994) 182.[42] J.A. Dunne, R. Mariwala, M. Rao, S. Sircar, R.J. Gorte, A.L. Myers, Langmuir 12

(1996) 5888.[43] J. Vermesse, D. Vidal, P. Malbrunot, Langmuir 12 (1996) 4190.[44] M.M.K. Salem, P. Braeuer, M.V. Szombathely, M. Heuchel, P. Harting, K.

Quitzsch, M. Jaroniec, Langmuir 14 (1998) 3376.[45] O. Talu, S.-Y. Zhang, D.T. Hayhurst, J. Phys. Chem. 97 (1993) 12894.[46] F. Dreisbach, R. Staudt, J.U. Keller, Adsorption 5 (1999) 215.[47] S. Pakseresht, M. Kazemeini, M.M. Akbarnejad, Sep. Purif. Technol. 28 (2002)

53.[48] F.R. Siperstein, A.L. Myers, AIChE J. 47 (2001) 1141.[49] P.M. Mathias, R. Kumar, J.D. Moyer, J.J.M. Schork, S.R. Srinivasan, S.R. Auvil, O.

Talu, Ind. Eng. Chem. Res. 35 (1996) 2477.[50] A.O. Yazaydin, R.Q. Snurr, T.-H. Park, K. Koh, J. Liu, M.D. LeVan, A.I. Benin, P.

Jakubczak, M. Lanuza, D.B. Galloway, J.J. Low, R.R.J. Willis, J. Am. Chem. Soc.131 (2009) 18198.