Adsorption of alkanes, alkenes, and their mixtures in

single-walled carbon nanotubes and bundles

Sven Jakobtorweihen, Frerich J. Keil

To cite this version:

Sven Jakobtorweihen, Frerich J. Keil. Adsorption of alkanes, alkenes, and their mixtures in
single-walled carbon nanotubes and bundles. Molecular Simulation, 2009, 35 (01-02), pp.90-99.
�10.1080/08927020802378936�. �hal-00515053�

HAL Id: hal-00515053

https://hal.science/hal-00515053
Submitted on 4 Sep 2010

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci-
entific research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
 Fo
Adsorption of alkanes, alkenes, and their mixtures in single-
walled carbon nanotubes and bundles
rP

Journal: Molecular Simulation/Journal of Experimental Nanoscience

Manuscript ID: GMOS-2008-0099.R1

ee

Journal: Molecular Simulation

Date Submitted by the

21-Jul-2008
Author:
rR

Complete List of Authors: Jakobtorweihen, Sven; Hamburg University of Technology, Institute

of Chemical Reaction Engineering
Keil, Frerich J.; Hamburg University of Technology, Institute of
Chemical Reaction Engineering
ev

carbon nanotubes, adsorption, alkanes, alkenes, molecular

Keywords:
simulations
ie
w
On
ly

http://mc.manuscriptcentral.com/tandf/jenmol
Page 1 of 23

1
2
3 Adsorption of alkanes, alkenes, and their mixtures in
4
5
6 single-walled carbon nanotubes and bundles
7
8
9
S. Jakobtorweihen and F. J. Keil∗
10 Hamburg University of Technology, Institute of Chemical Reaction Engineering,
11
12 Eissendorfer Str. 38, D-21073 Hamburg, Germany
13
14
15 Abstract
16 Monte Carlo simulations are employed to calculate pure component adsorption isotherms of
Fo

17
18 linear alkanes (C2−C12), alkenes (C2−C4), and some of their binary mixtures (ethane-ethene,
19
20 propane-propene, cis-2-butene-trans-2-butene, propene-1-butene) in single-walled carbon nan-
rP

21
22 otubes. The zigzag structures of carbon nanotubes of various diameters ((10,0), (20,0), (30,0), and
23
24 (40,0)) are used. Furthermore, Henry coefficients and isosteric heats of adsorption are calculated.
ee

25
26 The dependence of theses properties as a function of chain length (carbon number) is presented.
27
28 The relation of the critical parameters and the isosteric heats of adsorption, observed earlier
rR

29 for zeolites, could be confirmed for carbon nanotubes. The adsorption behaviour of 1-butene,
30
31 cis-2-butene, and trans-2-butene are compared in detail. Radial density profiles of 1-butene in
32
33 a (40,0) nanotube for various pressures reveal a built-up of three layers inside the pores with
ev

34
35 increasing pressure. For all investigated binary mixtures, one of the component isotherms shows a
36
distinct maximum owing to a entropic effect and non-idealities of the bulk gas phase behaviour.
iew

37
38
39 Additionally, adsorption in carbon nanotube bundles in hexagonal arrangement is studied.
40
41 Depending on the pore arrangements, pore diameters, and pressures, a fraction of the adsorbed
42 gases is located in the interstitial space.
43
On

44
45
46 Keywords: carbon nanotube; adsorption; alkanes; alkenes; molecular simulations
47
ly

48
49
50
51
52
53
54
55
56
57
58
59 ∗
corresponding author, e-mail: keil@tu-harburg.de
60

http://mc.manuscriptcentral.com/tandf/jenmol
 Page 2 of 23

1
2
3 1. INTRODUCTION
4
5
6 After the discovery of carbon nanotubes (CNTs) by Iijima [1], these materials were con-
7
8 sidered for adsorption of various gases. Besides hydrocarbons [2–6], adsorption of other
9
10 molecules and mixtures on single-walled carbon nanotubes (SWCNTs) were investigated,
11
12 like hydrogen [7–9], nitrogen [10–13], nitrogen/oxygen [14, 15], water [16], neon [17], he-
13
14
lium [18], and xenon [19]. Cao et al. [3] investigated the adsorption of methane on triangular
15 arrays of SWCNTs at room temperature with the grand canonical Monte Carlo (GCMC)
16
Fo

17 method. The carbon atoms on the tubular wall were structured according to the armchair
18
19 arrangement, and the “site-to-site method” was used to calculate the interaction between a
20
rP

21 methane molecule inside the tube and a carbon atom on the tubular wall. The gap between
22
23
the tubes was varied in order to find the optimum gap with respect to the amount of ad-
24 sorbed methane. Jiang et al. [4] studied adsorption and separation of linear (C1-C5) and
ee

25
26 branched (C5 isomers) alkanes on (10,10) SWCNT bundles at 300 K using configurational-
27
28 bias Monte Carlo (CBMC) simulations. For pure linear alkanes, the limiting adsorption
rR

29
30 properties at zero coverage exhibit a linear relation with the alkane carbon number; long
31
32
alkanes are more adsorbed at low pressures, but the reverse is found for short alkanes at
33 high pressures. For pure branched alkanes, the linear isomer adsorbs to a greater extent
ev

34
35 than its branched counterparts. For a five-component mixture of C1-C5 linear alkanes, the
36
long alkane adsorption first increases and then decreases with increasing pressure, but the
iew

37
38
39 short alkane adsorption is continuesly increasing and progressively replaces the long alkanes
40
at high pressures due to the size entropy effect. This phenomenon was already detected be-
41
42 fore in Refs. [20, 21] for silicalite and by Heyden et al. [22] for binary mixtures of methane,
43
On

44 ethane, propane, and tetrafluormethane in SWCNTs by using CBMC simulations. At high

45
46 loadings, a maximum occurs with increasing pressure in the absolute adsorption isotherm of
47
one or both adsorbing species. It was detected that there exist two fundamentally different
ly

48
49
reasons for this maximum. First, the size entropy effect [23] and second, non-ideality effects
50
51 of the gas phase [22] can be made responsible for this maximum. If, owing to non-ideality
52
53 effects of the gas phase, the fugacity of one component does not increase as steeply with
54
55 pressure as the other component, a maximum can occur in the absolute adsorption isotherm
56
57 of this component [22]. It must be stressed that the displaced component is not necessarily
58 the larger molecule. Similar effects were also observed by Jakobtorweihen et al. [24] for
59
60

http://mc.manuscriptcentral.com/tandf/jenmol
Page 3 of 23

1
2
3 adsorption of alkenes in various zeolites.
4
5 Kondratyuk et al. [5] found three well-defined adsorption sites on opened single-walled
6
7 carbon nanotubes (tube diameter 13.6 Å) by temperature programmed desorption measure-
8
9 ments for several alkanes. On the basis of hybrid Monte Carlo simulations, the two highest
10
11
binding energy adsorption site correspond to adsorption inside tubes, and to groove sites
12 between adjacent nanotubes in bundles. The third highest binding energy corresponds to
13
14 sites on the exterior nanotube surface. Burde et al. [2] investigated the kinetics of gas up-
15
16 take on different regions of carbon nanotube bundles by means of the kinetic Monte Carlo
Fo

17
18 scheme. On both external and internal sites of a nanotube bundle, equilibration times are
19
20
observed to decrease linearly as the coverage increases toward monolayer completion; the
rP

21 rate at which this occurs strongly depends on the ratio between the binding energy and
22
23 the temperature. For low coverages, very long equilibration times can be observed. The
24
adsorption in pore-like phases is typically 2 orders of magnitude slower than that of external
ee

25
26
27 phases.
28
A survey of computational techniques for carbon nanotubes was presented by Rafii-
rR

29
30 Tabar [25].
31
32 The motivation of the this paper is the investigation of alkene adsorp-
33
ev

34 tion in SWCNTs and the separation potential of linear alkanes and alkenes
35
36 in carbon nanotubes. Furthermore, the effect of pore diameters on adsorp-
iew

37
tion and the adsorption in the interstitial space in CNT-bundles is evaluated.
38
39 Hysteresis is not investigated in this work. Jiang et al. [6] have found hysteresis for alkane
40
41
adsorption inside a CNT having a larger diameter as the widest pore employed in this study,
42 but the dependence of hysteresis on tube size was not studied.
43
On

44 The outline of this article is as follows: In the next section the models underlying our
45 simulations are explained and details of the simulations are given. In section 3, the results
46
47 are shown and discussed. We end with a conclusion in the last section.
ly

48
49
50
51 2. SIMULATION DETAILS
52
53
54 Intermolecular interactions were modelled with the truncated Lennard-Jones (LJ)
55
56 12-6 potential. The fluid molecules were modelled as united-atoms. For the in-
57
58
tramolecular and the fluid-fluid intermolecular interactions the TraPPE force field was
59 used [26, 27], where the bond angle bending is described by a harmonic potential and
60

http://mc.manuscriptcentral.com/tandf/jenmol
 Page 4 of 23

1
2
3
4
5 Table 1. Lennard-Jones parameters and their source.
6
7 centre σ [Å] ǫ/kB [K] Ref.
8
9 CH4 3.73 148 [26]
10
CH3 (sp3 ) 3.75 98 [26]
11
12 CH2 (sp3 ) 3.95 46 [26]
13
14 CH2 (sp2 ) 3.675 85 [27]
15
16 CH(sp2 ) 3.73 47 [27]
Fo

17
18 carbon 3.4 28 [28]
19
20
rP

21
22 the bond torsion with a 3-cosine-fourier potential. For trans- and cis-2-butene an har-
23
24 monic torsional potential is used. In the TraPPE force field the bond lengths are fixed.
ee

25 To check the influence of this assumption, some systems were re-simulated with an harmonic
26
27 bond potential, we found no influence on the adsorption isotherms.
28 The carbon LJ parameters were taken from Ref. [28]. Lorentz-
rR

29
30 Berthelot mixing rules were used to determine the interactions of un-
31
32 like molecular centres. The LJ parameters are listed in table 1.
33
ev

For the LJ interactions a cutoff radius of 14 Å was used in combination with the usual tail
34
35 corrections [29]. Note that Macedonia and Maginn [30] have shown that tail corrections do
36
not apply to adsorption in nanoporous materials as tail corrections assume a homogenous
iew

37
38
39 distribution of the molecules behind the cutoff. As the TraPPE force field includes tail
40
41
corrections and to date no force field that was explicitly developed for hydrocarbon-CNT
42 interactions is available (lake of experimental data prevent devolping such parameters), tail
43
On

44 corrections were used in this study. Moreover, Macedonia and Maginn [30] have shown that
45
46 the influence of tail correction on calculated adsorption isotherms for butane in silicalite is
47
small. The influence on the systems studied in this work is also minor. For example, ethane
ly

48
49
50
at a temperature of 300 K and a pressure of 1 bar inside a (20,0) CNT shows an adsorption
51 capacity of 3.69 mol/kg, this value reduces to 3.34 mol/kg when tail corrections are not
52
53 used.
54 The size and helicity of carbon nanotubes is specified by two integer numbers [31] n1 and
55
56 n2 , so they can be unambiguously labelled as (n1 ,n2 ). All results in this work were obtained
57
58 with single-walled carbon nanotubes in a zigzag structure. For n2 =0 tubes have the zigzag
59
60

http://mc.manuscriptcentral.com/tandf/jenmol
Page 5 of 23

1
2
3 structure, at which one carbon-carbon bond is parallel to the pore axis. The investigated
4
5 tubes have the following inner diameters: 7.8 Å (10,0), 15.7 Å (20,0), 23.5 Å (30,0), and 31.3
6
7 Å (40,0). Note that, although deformations can occur especially for larger tubes, we have
8
9 simulated all nanotubes as “perfect” single-walled nanotubes. As framework flexibility has a
10
11
vanishing influence on adsorption isotherms [32] we modelled the nanotubes as rigid (carbon
12 atoms fixed on their positions). Note that carbon nanotube flexibility can have a dramatic
13
14 influence on the dynamics of guest molecules [33]. The rigid framework assumption allows
15
16 the use of interpolation grids for the calculation of the fluid-nanotube interactions. A grid
Fo

17
18 fineness of 0.065 Å was used. The nanotubes length was set to 102 Å. However, effectively
19
20
an infinitely long tube was simulated as usual periodic boundaries [34] were used for the
rP

21 direction parallel to the pore axis.

22
23 Grand canonical Monte Carlo simulations were performed to calculate adsorption
24
isotherms. In the grand canonical ensemble the number of particles can fluctuate
ee

25
26
27 whereas the chemical potentials, the temperature, and the volume are constant. For
28
details about the GCMC method see Ref [29]. For effective simulations of molecules
rR

29
30 the configurational bias Monte Carlo method was used, which allows the growth of
31
32 a molecule atom by atom [35, 36]. The following Monte Carlo trials were performed
33
ev

34 during a GCMC simulations: displacement of a molecule, rotation of a molecule

35
36 around its centre of mass, partial and total regrowth of a molecule, and molecule
iew

37
exchange with a reservoir. Additionally, for simulating mixtures particle identity
38
39 changes were carried out. All acceptance rules are summarised in an earlier publication [37].
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57
58
59
60

http://mc.manuscriptcentral.com/tandf/jenmol
 Page 6 of 23

1
2
3
4
The number of generated configurations depends on the system. For all studied systems more
5 than 1.5 million Monte Carlo trials were performed for equilibration and more than 8 million
6
7 trials were carried out for sampling. For the molecule exchange acceptance rules either
8
9 the bulk chemical potentials or the bulk fugacities must be specified [37]. The fugacities
10
11 were calculated with the Peng-Robinson equation of state using parameters taken from
12
Ref [38]. To determine the Henry coefficients and the isosteric heats of adsorption Monte
13
14 Carlo simulations in the canonical ensemble in combination with the CBMC technique were
15
16 carried out for the zero loading limit [39] (only one fluid molecule). Additionally, a separate
Fo

17
18 simulation in the ideal gas state was necessary. The Henry coefficient can then be calculated
19
20 as
rP

21 hWi
KH = β , (1)
22 hWIG i
23
24 where β = 1/(kB T ) and W is the normalised Rosenbluth factor calculated according to the
ee

25
26
CBMC approach [29, 35, 36]. The subscript, IG, is used to indicate the ideal gas state. The
27 angular brackets denote ensemble averages. The term kB is the Boltzmann constant and T
28
rR

29 the temperature. The isosteric heat of adsorption has been calculated from
30
31
0
32 qst = hUiIG − hUi + kB T, (2)
33
ev

34
35 where U is the total energy.
36
iew

37
38 3. RESULTS AND DISCUSSION
39
40 The first results in this section are for adsorption inside a single carbon nanotubes, there-
41
42 after results for CNT bundles are shown. All presented isotherms are absolute adsorption
43
On

44 isotherms.
45 The pure component adsorption isotherms are presented by plotting the amount of moles
46
47 adsorbed per unit weight of adsorbent as a function of pressure. Error bars have not been
ly

48 included in the figures since they are, in almost all cases, smaller than the symbols. In
49
50 figures 1 and 2 the pure component adsorption isotherms of ethane, ethene, propane, and
51
52 propene for various single-walled carbon nanotubes at 300 K are presented.
53
54 As can be seen from figure 1, at low pressures more alkane molecules are adsorbed than
55
56 alkenes, and the adsorption in narrow pores is stronger than in wider pores. The interaction
57 of the adsorbed molecules is stronger in narrow pores, and the isotherms level off at lower
58
59 pressures than in larger pores. As expected, at higher pressures more molecules can be
60

http://mc.manuscriptcentral.com/tandf/jenmol
Page 7 of 23

1
2
3 16
4 ethane (20,0) CNT
ethane (30,0) CNT
5 14 ethane (40,0) CNT
ethene (20,0) CNT

amount adsorbed [mol kg-1]

6 12
ethene (30,0) CNT
ethene (40,0) CNT
7
8 10

9 8
10
6
11
12 4
13
2
14
15 0
0.01 0.1 1 10 100 1000
16
Fo
pressure [bar]
17
18
19 Figure 1. Adsorption isotherms of ethane and ethene in different carbon nanotubes at 300
20
rP

21 K. Lines are added to guide the eye.

22
23
12
24 propane (20,0) CNT
ee

propane (30,0) CNT

25 propane (40,0) CNT
10 propene (20,0) CNT
26
amount adsorbed [mol kg-1]

propene (30,0) CNT

propene (40,0) CNT
27
8
28
rR

29
6
30
31
4
32
33
ev

2
34
35 0
36 0.001 0.01 0.1 1 10 100 1000
iew

37 pressure [bar]
38
39 Figure 2. Adsorption isotherms of propane and propene in different carbon nanotubes at
40
41 300 K. Lines are added to guide the eye.
42
43
On

44
adsorbed inside the larger pores. A comparison between figures 1 and 2 shows that at
45
46 low pressures more propane molecules are adsorbed than ethane, but at high pressure the
47
ly

48 number of adsorbed ethane molecules is higher as the required space is smaller than for
49
50 propane. At low pressures both ethene and propene are adsorbed in a smaller amount
51
52 than the respective alkanes, at high pressures it is the opposite. This is caused by the larger
53
number of interaction sites of alkanes. At high pressure the smaller size of alkenes is decisive.
54
55 In figure 3 the adsorption isotherms of 1-butene, cis-butene, and trans-butene at 300 K
56
57 in (20,0) and (40,0) SWCNTs are presented. As before, at low pressures more molecules
58
59 in the narrow pores are adsorbed, and at high pressure it is the opposite. The adsorption
60

http://mc.manuscriptcentral.com/tandf/jenmol
 Page 8 of 23

1
2
3 10
4 1-butene (20,0) CNT
1-butene (40,0) CNT
5 cis-butene (20,0) CNT
cis-butene (40,0) CNT

amount adsorbed [mol kg-1]

8
6 trans-butene (20,0) CNT
trans-butene (40,0) CNT
7
8 6

9
10 4
11
12
2
13
14
15 0
10-4 10-3 10-2 10-1 100 101 102 103
16
Fo
pressure [bar]
17
18
19 Figure 3. Adsorption isotherms of 1-butene, cis-2-butene, and trans-2-butene in different
20
rP

21 carbon nanotubes at 300 K. Lines are added to guide the eye.

22
23
24 behaviour of the three alkenes is quite similar.
ee

25
26 The development of adsorption layers for 1-butene is shown in detail in figure 4. RD
27
28 profiles of 1-butene inside a (40,0) SWCNT at 300 K and four pressures (0.1, 0.4, 0.6, and
rR

29
30
100 bar) are presented. Additionally, the adsorption isotherm is given, where the state point
31 for which the RD profile is shown is marked with a cross. At 0.1 bar the first adsorption
32
33 layer, at a distance of 0.4 nm from the wall, is formed. At a slightly higher pressure of 0.4
ev

34
35 bar the formation of a second layer is initiated. The centre part of the tube (1.2−1.6 nm) is
36
still empty. This changes at 0.6 bar where the third layer is slowly built up and the second
iew

37
38
39
layer gets more distinctive. Eventually, at 100 bar the third layer is formed and the centre
40 part of the tube is filled.
41
42 Figure 5 shows the calculated Henry coefficients of n-alkanes as a function of their carbon
43
On

44 numbers at 300 K in SWCNTs of various diameters. These simulations were executed for the
45
46 zero loading limit. The very strong interaction between the (10,0) carbon nanotube and the
47
alkanes is demonstrated by the rather large Henry coefficients compared to the values for the
ly

48
49 wider tubes. The alkenes (see figure 6) show a similar behaviour. A comparison of the Henry
50
51 coefficients of 1-butene, cis-2-butene, and trans-2-butene in (10,0) carbon tubes reveals that
52
53 trans-2-butene adsorbs strongest and the comparatively bulky cis-2-butene adsorbs weakest.
54
55 For both alkanes and alkenes the differences between the Henry coefficients become smal-
56
ler the larger the tube diameters are. This is also supported by the isosteric heats of
57
58 adsorption in figures 7 and 8. The heats of adsorption show a linear dependence on the
59
60

http://mc.manuscriptcentral.com/tandf/jenmol
Page 9 of 23

1
2
3 pressure [bar] pressure [bar]
4 10-4
10-2
10 0
10 2
10-4
10-2 100 102
5 40 10 40 10

amount adsorbed [mol kg-1]

amount adsorbed [mol kg-1]

6 RD
35 35
particle density [nm-3]

particle density [nm-3]

isotherm 8 8
7 30 30
8 25 25
6 6
9 20 (a) p = 0.1 bar 20 (b) p = 0.4 bar
10 4 4
15 15
11
12 10 10
2 2
13 5 5
14 0 0 0 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
15
16 distance from CNT wall [nm] distance from CNT wall [nm]
Fo
pressure [bar] pressure [bar]
17
10-4 10-2 100 102 10-4 10-2 100 102
18 40 10 40 10

amount adsorbed [mol kg-1]

amount adsorbed [mol kg-1]

19 35 35
particle density [nm-3]

particle density [nm-3]

20 8 8
rP

30 30
21
25 6 25 6
22
23 20 (c) p = 0.6 bar 20 (d) p = 100 bar
15 4 15 4
24
ee

25 10
2
10
2
26 5 5
27 0 0 0 0
28 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
rR

29 distance from CNT wall [nm] distance from CNT wall [nm]
30
31
32 Figure 4. Radial density (RD) profiles represent 1-butene distributions inside a (40,0)
33
ev

34
CNT at 300 K. The adsorption isotherm of 1-butene at 300 K is plotted on the secondary
35 axes, the state point for which the RD profile is shown is marked with a cross. RD profiles
36
iew

37 for four different pressures are shown: (a) 0.1 bar, (b) 0.4 bar, (c) 0.6 bar, and (d) 100 bar.
38
39
40
1024
41 (10,0) CNT
1022
42 (20,0) CNT
Henry coefficient [mol m-3 Pa-1]

20
10
43 (30,0) CNT
On

1018
44 10 16 (40,0) CNT

45 10 14

46 1012
47 1010
ly

48 108
106
49
104
50
102
51
100
52 10-2
53 2 4 6 8 10 12

54 carbon number of n-alkane

55
56 Figure 5. Henry coefficients of n-alkanes as functions of their carbon numbers at 300 K in
57
58 different carbon nanotubes. Lines represent exponential fits.
59
60

http://mc.manuscriptcentral.com/tandf/jenmol
 Page 10 of 23

1
2
3 108
4 (10,0) CNT trans-2-butene (10,0) CNT
107 (20,0) CNT
5

Henry coefficient [mol m-3 Pa-1]

(30,0) CNT
106
6 (40,0) CNT

7 105

8 104
9 103
10
102 cis-2-butene (10,0) CNT
11
101
12
13 100

14 10-1
15 10-2
2 3 4
16
Fo
carbon number of alkene
17
18
19 Figure 6. Henry coefficients of alkenes as functions of their carbon numbers at 300 K in
20
rP

21 different carbon nanotubes. Lines represent exponential fits. Additionally, the results for
22
23 cis-2-butene and trans-2-butene in the (10,0) carbon nanotube are shown (open symbols).
24
ee

25 220
26 (10,0) CNT
isosteric heat of adsorption [kJ/mol]

200
27 (20,0) CNT
180
28 (30,0) CNT
rR

160 (40,0) CNT

29
140
30
31 120

32 100

33 80
ev

34 60
35 40
36 20
iew

37 0
2 4 6 8 10 12
38
carbon number of n-alkane
39
40
41 Figure 7. Isosteric heats of adsorption of n-alkanes as function of their carbon numbers for
42
43 the zero loading limit at 300 K in different carbon nanotubes. Lines represent linear fits.
On

44
45
46 chain length for alkanes and alkenes. Note that both, the chain length dependency of the
47
ly

48 Henry coefficients and the chain length dependency of the heats of adsorption were also
49
50 observed experimentally for alkanes in zeolites [40]. Furthermore, these dependencies were
51
52 already observed with molecular simulation methods for the adsorption of alkanes in CNT
53
bundles [4] and, in agreement with experiments, for the adsorption in silicalite [20, 41].
54
55 In table 2 a comparison of Henry coefficients, KH , and isosteric heats of adsorption, qst ,
56
57 for the butenes inside carbon nanotubes of different sizes is presented. Unlike the (10,0)
58
59 tube (see also figures 6 and 8), the larger nanotubes show the following sequence in the KH
60

10

http://mc.manuscriptcentral.com/tandf/jenmol
Page 11 of 23

1
2
3 90
4 (10,0) CNT trans-2-butene (10,0) CNT

isosteric heat of adsorption [kJ/mol]

5 80 (20,0) CNT
(30,0) CNT
6 70
(40,0) CNT
7
8 60
cis-2-butene (10,0) CNT
9 50
10
40
11
12 30
13
20
14
15 10
2 3 4
16
Fo
carbon number of alkene
17
18
19 Figure 8. Isosteric heats of adsorption of alkenes as function of their carbon numbers for
20
rP

21 the zero loading limit at 300 K in different carbon nanotubes. Lines represent linear fits.
22
23 Additionally, the results for cis-2-butene and trans-2-butene in the (10,0) carbon nanotube
24
are shown (open symbols).
ee

25
26
27
28
and qst values: 1-butene < cis-2-butene < trans-2-butene. In the (10,0) nanotube the order
rR

29 is cis-2-butene < 1-butene < trans-2-butene. In the narrow (10,0) tube the structure of the
30
31 1-butene molecule is changed compared to the structure in the ideal gas state, whereas the
32
33 structure of 1-butene in the wider tubes is similar to the structure in the ideal gas state. The
ev

34
35 structure of 1-butene in a (10,0) CNT is closer to trans-2-butene than in the other tubes.
36
Moreover, in the (10,0) nanotube 1-butene is never in the cis state, whereas in the ideal gas
iew

37
38 state a small fraction of 1-butene molecules can be found in the cis configuration. Thus,
39
40 the 1-butene molecule structure is adjusted to better fit these molecules into the narrow
41
42 channels.
43
On

44 For silicalite-1 and highly de-aluminated Y zeolite (US-Ex), Thamm et al. [42] found a
45
46
linear relation between the experimental initial heats of adsorption of n-alkanes and cyclo-
47 hexane and their critical parameters
ly

48
49 0 TC
50 qst =D , (3)
p0.5
C
51
52 where D represents the proportional constant. These authors showed that the isosteric heat
53
54 of adsorption of ethene and 1-butene is also determined by this relation. Jakobtorweihen
55
56 et al. [24] confirmed these findings by simulations of a broad variety of alkanes and alkenes
57
58
in silicalite-1. Bruce et al. [43] found this result also for graphitized black powders. In
59 figure 9 the relation given in equation 3 is confirmed for linear alkanes and some alkenes
60

11

http://mc.manuscriptcentral.com/tandf/jenmol
 Page 12 of 23

1
2
3
4
5 Table 2. Comparison of the Henry coefficients, KH , and the isosteric heats of adsorption,
6 qst , for the butenes in different carbon nanotubes. All results correspond to a temperature
7
8 of 300 K. Errors are given in the subscripts.
9
10 species CNT KH qst
11
12 [mol m3 Pa−1 ] [kJ mol−1 ]
13
14 1-butene (10,0) 4.733×105 0.158 76.260.04
15
16 cis-2-butene (10,0) 6.047×101 0.603 54.920.10
Fo

17
18 trans-2-butene (10,0) 3.938×107 0.071 85.620.03
19
20 1-butene (20,0) 14.2140.027 34.040.01
rP

21
22
cis-2-butene (20,0) 16.6940.051 35.370.01
23 trans-2-butene (20,0) 22.0920.055 36.360.02
24
ee

25 1-butene (30,0) 1.1270.002 28.070.01

26
27 cis-2-butene (30,0) 1.3190.004 29.480.02
28
rR

29 trans-2-butene (30,0) 1.8790.006 30.720.01

30
31 1-butene (40,0) 0.4170.001 25.970.01
32
33 cis-2-butene (40,0) 0.4860.001 27.340.02
ev

34
35 trans-2-butene (40,0) 0.6910.002 28.640.02
36
iew

37
38
39 in SWCNTs of various diameters. Only the bulky cis-2-butene shows a somewhat larger
40
41 deviation in the very narrow (10,0) tube.
42
In figure 10a simulated adsorption isotherms of the binary mixture of ethane and ethene
43
On

44 inside a (20,0) carbon nanotube at 300 K for a fixed equimolar bulk phase composition
45
46 are presented. Additionally the selectivity is shown. The selectivity of component 1 with
47
respect to component 2 is defined as
ly

48
49
50 x1 /y1
51 S12 = , (4)
x2 /y2
52
53
54
where x and y are the mole fractions of the adsorbed phase and the bulk phase, respectively.
55 The lines represent results obtained by the ideal adsorption solution theory (IAST) [44].
56
57 For these calculations the pure component adsorption isotherms detailed in this work
58
59 were used. To take into account the non-ideality of the gas phase, we have used the
60

12

http://mc.manuscriptcentral.com/tandf/jenmol
Page 13 of 23

1
2
3 250
4 (10,0) CNT

isosteric heat of adsorption [kJ/mol]

5 (20,0) CNT dodecane
200 (30,0) CNT
6
(40,0) CNT
7
linear fit
8 150
9
10 100
11 trans-2-butene

12 C2H6 cis-2-butene
50
13 C2H4

14
15 0
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
16
Tc pc-0.5 [K Pa-0.5]
Fo

17
18
19 Figure 9. Dependence of the simulated isosteric heats of adsorption for the zero loading
20
rP

21 limit at 300 K in different carbon nanotubes on the critical data for hydrocarbons. Lines
22
23 represent linear fits, where the fit constants are: D10 =396.64, D20 =179.66, D30 =150.66,
24
and D40 =140.66.
ee

25
26
27
28 fugacities rather than the pressure for calculating the IAST isotherms. Like for the pure
rR

29
30
component adsorption, at low pressure more ethane than ethene is adsorbed. At high
31 pressures the entropic effect [23] leads to a reversal. Thus, we observe a distinct maximum
32
33 in the ethane isotherm, which is also calculated by the IAST. A further reason for the
ev

34
35 maximum of the isotherms is a non-ideality effect of the gas phase. Details are given by
36
Heyden et al. [22]. In figure 10b the component fugacities of ethane and ethene in the gas
iew

37
38
39
phase as calculated by the Peng-Robinson equation of state show a difference between the
40 fugacities from about 25 bar on. Close to this value is the maximum of the adsorption
41
42 isotherm of ethane. These maxima were also observed by Jiang et al. [4] and Heyden et
43
On

44 al. [22] for carbon nanotubes and by Jakobtorweihen et al. [24] for zeolites, among others.
45 Note that Jiang et al. [4] have also observed distinct maxima in the selectivities for mixtures
46
47 of C5 isomers.
ly

48 The selectivity drops from about two at low pressures to about one at high pressures.
49
50 A qualitatively similar behaviour can be observed in figure 11 for a propane and propene
51
52 binary mixture. The selectivities for this mixture are closer to one. The larger the hydrocar-
53
54
bons the weaker the influence of the double bond. Hence, the higher the carbon number the
55 higher the similarity of an alkane and an 1-alkene having the same carbon number. It can
56
57 be concluded, that for these kind of alkane-alkene mixtures the separation is more difficult
58
59 the higher the carbon number.
60

13

http://mc.manuscriptcentral.com/tandf/jenmol
 Page 14 of 23

1
2
3 6 2.4
4 5.5
(a)
2.2
5 5 2

amount adsorbed [mol kg-1]

6 4.5 1.8
7 4 1.6
8

selectivity
3.5 1.4
9 3 1.2
10 2.5 1
11 2 0.8
12 1.5 0.6
total amount adsorbed
13 1 ethane 0.4
ethene
14 0.5 selectivity 0.2
IAST
15 0 0
0.01 0.1 1 10 100 1000
16
Fo
total bulk pressure [bar]
17 25
18 (b)

19
20
20
component fugacity [bar]
rP

21
22 15
23
24 10
ee

25
26
5
27
28 ethane
rR

ethene
29 0
0 20 40 60 80 100
30 total bulk pressure [bar]
31
32
33 Figure 10. (a) Simulated adsorption isotherm of a binary mixture of ethane and ethene
ev

34
35 inside a (20,0) carbon nanotube for a fixed equimolar bulk phase composition at 300 K and
36
selectivity of ethane with respect to ethene. Lines represent results of the ideal adsorption
iew

37
38
solution theory. (b) The corresponding component fugacities, calculated with the
39
40 Peng-Robinson equation of state, are plotted against total bulk pressure. Lines are added
41
42 to guide the eye.
43
On

44
45 Figure 12 shows the adsorption isotherms of cis-2-butene and trans-2-butene inside a
46
47 (20,0) SWCNT at 400 K for a fixed equimolar bulk phase composition. The lines present
ly

48
49 again the IAST results. Initially the adsorption isotherms are very similar. At very high
50
51 pressures a configurational entropic effect occurs and the isotherms start to diverge. An
52
effective separation of these isomers is therefore, if at all, only possible at very high pressures.
53
54 A better separation can be achieved for a binary mixture of propane and 1-butene (see
55
56 figure 13).
57
58 All results presented above dealt with adsorption inside, and only inside, a single CNT.
59
60

14

http://mc.manuscriptcentral.com/tandf/jenmol
Page 15 of 23

1
2
3 5 2
4 (a)
4.5 1.8
5

amount adsorbed [mol kg-1]

4 1.6
6
7 3.5 1.4

selectivity
3 1.2
9 2.5 1
10 2 0.8
11
1.5 0.6
12
1 total amount adsorbed 0.4
13 propane
propene
14 0.5 selectivity 0.2
IAST
15 0 0
0.001 0.01 0.1 1 10 100 1000
16
Fo
total bulk pressure [bar]
17 7
18 (b)

19 6
20
component fugacity [bar]
rP

5
21
22 4
23
24 3
ee

25
2
26
27 1
28 propane
rR

propene
29 0
0 20 40 60 80 100
30 total bulk pressure [bar]
31
32
33 Figure 11. (a) Simulated adsorption isotherm of a binary mixture of propane and propene
ev

34
35 inside a (20,0) carbon nanotube for a fixed equimolar bulk phase composition at 300 K and
36
selectivity of propane with respect to propene. Lines represent results of the ideal
iew

37
38
adsorption solution theory. (b) The corresponding component fugacities, calculated with
39
40 the Peng-Robinson equation of state, are plotted against total bulk pressure. Lines are
41
42 added to guide the eye.
43
On

44
45 In order to investigate the adsorption in the interstitial space of bundles of SWCNTs,
46
47 hexagonally arranged bundles of SWCNTs with various diameters were investigated. The
ly

48
49 distance a (see figure 14) was set to 0.36 nm. In figure 15 the fraction of ethane molecules
50
51 located in the interstitial space of different SWCNT bundles at 300 K is shown.
52
As the interstitial space is small, no ethane molecules adsorb in the interstitial space of
53
54 (10,0) tubes in a hexagonal arrangement. Inside (20,0) tubes the interstitial space is just
55
56 large enough to adsorb ethane molecules (see figure 16), but a pressure of at least 1 bar is
57
58 necessary to initiate a measurable adsorption.
59
60

15

http://mc.manuscriptcentral.com/tandf/jenmol
 Page 16 of 23

1
2
3 4 1.4
4
5 3.5 1.2

amount adsorbed [mol kg-1]

6 3
1
7
8 2.5

selectivity
0.8
9 2
10 0.6
1.5
11
0.4
12 1
total amount adsorbed
13 cis-butene
0.2
0.5 trans-butene
14 selectivity
IAST
15 0 0
0.01 0.1 1 10 100 1000
16
Fo
total bulk pressure [bar]
17
18
19 Figure 12. Simulated adsorption isotherm of a binary mixture of cis-butene and
20
rP

21 trans-butene inside a (20,0) carbon nanotube for a fixed equimolar bulk phase composition
22
23 at 400 K and selectivity of cis-butene with respect to trans-butene. Lines represent results
24
of the ideal adsorption solution theory.
ee

25
26
27 4 1
total amount adsorbed
28 propene
rR

3.6 0.9
1-butene
29 selectivity
amount adsorbed [mol kg-1]

3.2 0.8
30 IAST
2.8 0.7
31
32 2.4 0.6 selectivity
33
ev

2 0.5
34 1.6 0.4
35 1.2 0.3
36
0.8 0.2
iew

37
38 0.4 0.1

39 0 0
0.01 0.1 1 10 100 1000
40 total bulk pressure [bar]
41
42
43 Figure 13. Simulated adsorption isotherm of a binary mixture of propene and 1-butene
On

44
45 inside a (20,0) carbon nanotube for a fixed equimolar bulk phase composition at 400 K and
46
47 selectivity of propene with respect to 1-butene. Lines represent results of the ideal
ly

48 adsorption solution theory.

49
50
51
52 For the (30,0) and (40,0) tubes in hexagonal arrangement, ethane molecules at low pres-
53
sure preferentially adsorb in the interstitial space. Due to the wall surrounding this is the
54
55 energetically most favourable position (see also figure 16). From about 0.01 bar on, the
56
57 interstitial space for (30,0) and (40,0) tubes in hexagonal arrangement are nearly filled, and
58
59 adsorption inside the tubes starts (see figure 15).
60

16

http://mc.manuscriptcentral.com/tandf/jenmol
Page 17 of 23

1
2
3
4
5
6
7
8
9
10 a
11
12
13 a
14
15
16
Fo

17 Figure 14. Carbon nanotubes in a hexagonally arranged bundle. The x-y plan of the
18
19 bundle of (20,0) nanotubes are shown as it was used in the simulations. For all
20
rP

21 investigated bundles a was set to 0.36 nm.

22
23
1
24 (10,0) CNT hexa. bundle
(20,0) CNT hexa. bundle
ee

25
fraction in the interstitial space [-]

(30,0) CNT hexa. bundle

(40,0) CNT hexa. bundle
26 0.8

27
28 0.6
rR

29
30 0.4
31
32
0.2
33
ev

34
35 0

36 10-4 10-3 10-2 10-1 100 101 102

pressure [bar]
iew

37
38
39 Figure 15. Fraction of ethane located in the interstitial space of a hexagonally arranged
40
41 carbon nanotube bundle at a temperature of 300 K. All bundles contain only one type of
42
43 nanotubes. Lines are added to guide the eye.
On

44
45
46 In this study the distance of the tubes a=0.36 nm (see figure 14) was chosen according to the
47
ly

48 employed force field parameters (σC =0.34 nm and the position of the minimum 21/6 σC =0.38
49
50 nm). Note that in other studies [45] the tube-tube distance is revealed to be smaller than
51
52 0.36 nm. Whether or not ethane adsorbs in the interstitial space is strongly dependent on
53
54
the tube-tube distances and the nanotube radii.
55 However, the recently produced CNT-membranes contain CNTs encapsulated in a mat-
56 rix [46, 47], so that the interstitial space is not accessible for adsorption. Note that for a
57
58 comparison between measured adsorption isotherms and simulations the real tube diameters
59
60

17

http://mc.manuscriptcentral.com/tandf/jenmol
 Page 18 of 23

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Fo

17 Figure 16. The snapshot displays an ethane molecule in the interstitial space of a
18
19
hexagonal arranged bundle of (20,0) carbon nanotubes.
20
rP

21
22 and tube arrangements are needed. Furthermore, the tubes have to be open at both ends
23
24
which was obviously not always the case in previous measurements [48].
ee

25
26
27 4. CONCLUSION
28
rR

29
30 Molecular simulations have been performed to evaluate the adsorption isotherms of linear
31
32 alkanes, alkenes, and some of their binary mixtures. As expected, at low pressures the ad-
33
ev

34 sorption in narrow pores is stronger than in wider pores, and more alkanes than alkenes are
35
36 adsorbed under these conditions. At higher pressures more molecules inside larger pores
iew

37
38
can be adsorbed. At low pressures more alkanes are adsorbed than the corresponding al-
39 kenes, at high pressures it is the opposite. Short and long alkanes show, in principle, the
40
41 same adsorption isotherm shapes. The adsorption isotherms of 1-butene, cis-2-butene, and
42
43 trans-2-butene are quite similar. The Henry coefficients and isosteric heats of adsorption of
On

44
45 all alkanes and alkenes investigated revealed a very strong interaction with the (10,0) CNT
46
47
(dCN T =7.8 Å). For both alkanes and alkenes the differences between the Henry coefficients
ly

48 and isosteric heats of adsorption become smaller the larger the tube diameters are. The re-
49
50 lation between critical parameters and the isosteric heats of adsorption by Thamm et al. [42]
51
52 could be confirmed for adsorption in CNTs. For each of the binary mixtures, one component
53
54 isotherm shows a distinct maximum. This could be attributed to a configurational entropy
55
effect inside the pores and non-idealities in the bulk phase. The IAST calculations also
56
57 showed these maxima, and coincided with the simulation results. The selectivity of the
58
59 ethane-ethene mixture is higher at low pressures than for propane-propene. Cis- and trans-
60

18

http://mc.manuscriptcentral.com/tandf/jenmol
Page 19 of 23

1
2
3 2-butene can only be separated, if at all, at very high pressures. The larger the difference in
4
5 the carbon numbers of the co-adsorbing species the larger the selectivity. The selectivities
6
7 are low when both components (n-alkane and 1-alkene) have the same number of carbons.
8
9 At low pressures, a considerable amount of ethane is adsorbed in the interstitial space of
10
11
the here investigated hexagonal arrangements (tube-tube distance a=0.36 nm) of (30,0) and
12 (40,0) CNTs.
13
14 Acknowledgment. This work was supported by the Deutsche Forschungsgemeinschaft
15
16 (DFG) in priority program SPP 1155. We thank Niels Hansen for help with IAST calcula-
Fo

17
18 tions.
19
20
rP

21
22
23
24
ee

25
26
27
28
rR

29
30
31
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57
58
59
60

19

http://mc.manuscriptcentral.com/tandf/jenmol
 Page 20 of 23

1
2
3
4
5
6
7 [1] S. Iijima, Helical microtubules of graphitic carbon, Nature, 354 (1991), pp. 56–58.
8
9 [2] J. T. Burde and M. M. Calbi, Physisorption kinetics in carbon nanotube bundles, J. Phys.
10
11
Chem. C, 111 (2007), p. 5057.
12
[3] D. Cao, X. Zhang, J. Chen, W. Wang, and J. Yun, Optimization of single-walled carbon
13
14 nanotube arrays for methane storage at room temperature, J. Phys. Chem. B, 107 (2003), p.
15
16 13286.
Fo

17
18 [4] J. Jiang, S. I. Sandler, M. Schenk, and B. Smit, Adsorption and separation of linear and
19
20 branched alkanes on carbon nanotube bundles from configurational-bias monte carlo simulation,
rP

21
22 Phys. Rev. B, 72 (2005), p. 045447.
23
24
[5] P. Kondratyuk, Y. Wang, J. K. Johnson, and J. T. Yates, Observation of a one-dimensional
ee

25
adsorption site on carbon nanotubes: Adsorption of alkanes of different molecular lengths, J.
26
27 Phys. Chem. B, 109 (2005), p. 20999.
28
rR

29 [6] J. Jiang, S. I. Sandler, and B. Smit, Capillary phase transitions of n-alkanes in a carbon
30
31 nanotube, Nano Letters, 4 (2004), pp. 241–244.
32
33 [7] F. Darkim, P. Melbrunot, and G. P. Tartaglia, Review of hydrogen storage by adsorption in
ev

34
35 carbon nanotubes, Int. J. Hydrogen Energy, 27 (2002), p. 193.
36
[8] V. Meregalli and M. Parinello, Review of theoretical calculations of hydrogen storage in carbon-
iew

37
38
based materials, Appl. Phys. A, 72 (2001), p. 143.
39
40 [9] G. Froudakis, Hydrogen interaction with carbon nanotubes: a review of ab initio studies, J.
41
42 Phys. Condens. Matter, 14 (2002), pp. R453–R465.
43
On

44 [10] J. I. Paredes, F. Suárez-Garcı́a, S. Villar-Rodil, A. Martı́nez-Alonso, J. M. D. Tascón, and

45
46 E. J. Bottani, n2 physisorption on carbon nanotubes: Computer simulation and experimental
47
results, J. Phys. Chem. B, 107 (2003), pp. 8905–8916.
ly

48
49
50 [11] J. Jiang and S. I. Sandler, Nitrogen adsorption on carbon nanotube bundles: Role of the
51
external surface, Phys. Rev. B, 68 (2003), p. 245412.
52
53 [12] J. Jiang, N. J. Wagner, and S. I. Sandler, A monte carlo simulation study of the effect of carbon
54
55 topology on nitrogen adsorption on graphite, a nanotube bundle, c60 fullerite, c168 schwarzite,
56
57 and a nanoporous carbon, Phys. Chem. Chem. Phys., 6 (2004), pp. 4440–4444.
58
59 [13] G. Arora, N. J. Wagner, and S. I. Sandler, Adsorption and diffusion of molecular nitrogen in
60

20

http://mc.manuscriptcentral.com/tandf/jenmol
Page 21 of 23

1
2
3 single wall carbon nanotubes, Langmuir, 20 (2004), pp. 6268–6277.
4
5 [14] J. Jiang and S. I. Sandler, Nitrogen and oxygen mixture adsorption on carbon nanotube bundles
6
7 from molecular simulation, Langmuir, 20 (2004), p. 10910.
8
9 [15] G. Arora and S. I. Sandler, Air separation by single wall carbon nanotubes: Thermodynamics
10
11 and adsorptive selectivity, J. Chem. Phys., 123 (2005), p. 044705.
12
13 [16] A. Striolo, A. A. Chialvo, K. E. Gubbins, and P. T. Cummings, Water in carbon nanotubes:
14
Adsorption isotherms and thermodynamic properties from molecular simulation, J. Chem.
15
16 Phys., 122 (2005), p. 234712.
Fo

17
18 [17] M. C. Gordillo, L. Brualla, and S. Fantoni, Neon adsorbed in carbon nanotube bundles, Phys.
19
20 Rev. B, 70 (2004), p. 245420.
rP

21
22 [18] B. Marcone, E. Orlandini, F. Toigo, and F. Ancilotto, Condensation of helium in interstitial
23
24 sites of carbon nanotubes bundles, Phys. Rev. B, 74 (2006), p. 085415.
ee

25
26 [19] V. V. Simonyan, J. K. Johnson, A. Kuznetsova, and J. T. Yates, Molecular simulation of xenon
27
adsorption on single-walled carbon nanotubes, J. Chem. Phys., 114 (2001), pp. 4180–4185.
28
rR

29 [20] T. J. H. Vlugt, R. Krishna, and B. Smit, Molecular simulations of adsorption isotherms for
30
31 linear and branched alkanes and their mixtures in silicalite, J. Phys. Chem. B, 103 (1999), pp.
32
33 1102–1118.
ev

34
35 [21] Z. Du, G. Manos, T. J. H. Vlugt, and B. Smit, Molecular simulation of adsorption of short
36
linear alkanes and their mixtures in silicalite, AIChE J., 44 (1998), pp. 1756–1764.
iew

37
38
39 [22] A. Heyden, T. Düren, and F. J. Keil, Study of molecular shape and non-ideality effects on
40
mixture adsorption isotherms of small molecules in carbon nanotubes: A grand canonical
41
42 monte carlo simulation study, Chem. Eng. Sci., 57 (2002), pp. 2439–2448.
43
On

44 [23] M. Schenk, S. L. Vidal, T. J. H. Vlugt, B. Smit, and R. Krishna, Separation of alkane isomers
45
46 by exploiting entropy effects during adsorption on silicalite-1: A configurational-bias monte
47
ly

48 carlo simulation study, Langmuir, 17 (2001), pp. 1558–1570.

49
50 [24] S. Jakobtorweihen, N. Hansen, and F. J. Keil, Molecular simulation of alkene adsorption in
51
52 zeolites, Molec. Phys., 103 (2005), pp. 471–489.
53
[25] H. Rafii-Tabar, Computational Physics of Carbon Nanotubes, Cambridge University Press,
54
55 Cambridge, 2008.
56
57 [26] M. G. Martin and J. I. Siepmann, Transferable potentials for phase equilibria. 1. United-atom
58
59 description of n-alkanes, J. Phys. Chem. B, 102 (1998), pp. 2569–2577.
60

21

http://mc.manuscriptcentral.com/tandf/jenmol
 Page 22 of 23

1
2
3 [27] C. D. Wick, M. G. Martin, and J. I. Siepmann, Transferable potentials for phase equilibria. 4.
4
5 United-atom description of linear and branched alkenes and alkylbenzenes, J. Phys. Chem. B,
6
7 104 (2000), pp. 8008–8016.
8
9 [28] W. A. Steele, The Interaction of Gases with Solid Surfaces, Pergamon, Oxford, 1974.
10
11 [29] D. Frenkel and B. Smit, Understanding Molecular Simulations, From Algorithms to Applica-
12
13 tions, 2nd ed., Academic Press, San Diego, 2002.
14
[30] M. D. Macedonia and E. J. Maginn, A biased grand canonical monte carlo method for simu-
15
16 lating adsorption using all-atom and branched united atom models, Molec. Phys., 96 (1999),
Fo

17
18 pp. 1375–1390.
19
20 [31] R. Saito, M. Fujita, G. Dresselhaus, and M. S. Dresselhaus, Electronic structure of chiral
rP

21
22 graphene tubules, Appl. Phys. Lett., 60 (1992), pp. 2204–2206.
23
24 [32] T. J. H. Vlugt and M. Schenk, Influence of framework flexibility on the adsorption properties
ee

25
26 of hydrocarbons in the zeolite silicalite, J. Phys. Chem. B, 106 (2002), pp. 12757–12763.
27
[33] S. Jakobtorweihen, M. G. Verbeek, C. P. Lowe, F. J. Keil, and B. Smit, Understanding
28
rR

29 the loading dependence of self-diffusion in carbon nanotubes, Phys. Rev. Lett., 95 (2005),
30
31 p. 044501.
32
33 [34] M. P. Allen and D. J. Tildesley, Computer Simulation of Liquids, Clarendon Press, Oxford,
ev

34
35 1987.
36
[35] J. I. Siepmann and D. Frenkel, Configurational bias monte carlo: a new sampling scheme for
iew

37
38
39 flexible chains, Molec. Phys., 75 (1992), pp. 59–70.
40
[36] D. Frenkel, G. Mooij, and B. Smit, Novel scheme to study structural and thermal properties
41
42 of continuously deformable molecules, J. Phys.: Condens Matter, 4 (1992), pp. 3053–3076.
43
On

44 [37] S. Jakobtorweihen, N. Hansen, and F. J. Keil, Combining reactive and configurational-bias

45
46 monte carlo: Confinement influence on the propene metathesis reaction system in various
47
ly

48 zeolites, J. Chem. Phys., 125 (2006), p. 224709.

49
50 [38] B. E. Poling, J. M. Prausnitz, and J. P. O’Connell, The Properties of Gases and Liquids, 5th
51
52 ed., McGraw-Hill, New York, 2000.
53
[39] B. Smit and J. I. Siepmann, Computer simulations of the energetics and sitting of n-alkanes
54
55 in zeolites, J. Phys. Chem., 98 (1994), pp. 8442–8452.
56
57 [40] J. F. Denayer, W. Souverijns, P. A. Jacobs, J. A. Martens, and G. V. Baron, High-temperature
58
59 low-pressure adsorption of branched c5 -c8 alkanes on zeolite beta, zsm-5, zsm-22, zeolite y, and
60

22

http://mc.manuscriptcentral.com/tandf/jenmol
Page 23 of 23

1
2
3 mordenite, J. Phys. Chem. B, 102 (1998), pp. 4588–4597.
4
5 [41] B. Smit and J. I. Siepmann, Simulating the adsorption of alkanes in zeolites, Science, 264
6
7 (1994), pp. 1118–1120.
8
9 [42] H. Thamm, H. Stach, and G. I. Berezin, Anfangsadsorptionswärmen von Gasen und Dämpfen
10
11 an den SiO2 -Molekularsieben US-Ex und Silicalit, Z. Chem., 24 (1984), pp. 420–421.
12
13 [43] C. D. Bruce, T. R. Rybolt, H. E. Thomas, T. E. Agnew, and B. S. Davis, Two-surface virial
14
analysis of alkane adsorption on carbopack c with and without hydrogen treatment, J. Colloid
15
16 Interface Sci., 194 (1997), pp. 448–454.
Fo

17
18 [44] A. L. Myers and J. M. Prausnitz, Thermodynamics of mixed gas adsorption, AIChE J., 11
19
20 (1965), pp. 121–127.
rP

21
22 [45] J.-C. Charlier, X. Gonze, and J.-P. Michenaud, First-principles study of carbon nanotube
23
24 solid-state packings, Europhys. Lett., 29 (1995), pp. 43–48.
ee

25
26 [46] J. K. Holt, H. G. Park, Y. wang, M. Stadermann, A. B. Artyukhin, C. P. Grigoropoulos,
27
A. Noy, and O. bakajin, Fast mass transport through sub-2-nanometer carbon nanotubes, Sci-
28
rR

29 ence, 312 (2006), pp. 1034–1037.

30
31 [47] B. J. Hinds, N. Chopra, T. Rantell, R. Andrews, V. Gavalas, and L. G. Bachas, Aligned
32
33 multiwalled carbon nanotube membranes, Science, 303 (2004), pp. 62–65.
ev

34
35 [48] S. Agnihotri and F.-L. Yun, Fundamental structural properties of single-walled carbon nan-
36
otubes: Reopening the debate over hydrogen storage, AIChE Spring National Meeting, Con-
iew

37
38
39 ference Proceedings, Americal Institute of Chemical Engineers, New York, N.Y., 2006, p.
40
P43844.
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57
58
59
60

23

http://mc.manuscriptcentral.com/tandf/jenmol