Langmuir 2009, 25, 973-976
973
Study of a Butane Monolayer Adsorbed on Single-Walled Carbon Nanotubes Dinesh S. Rawat, T. Furuhashi, and A. D. Migone* Department of Physics, Southern Illinois UniVersity, Carbondale, Illinois 62901 ReceiVed September 18, 2008. ReVised Manuscript ReceiVed NoVember 14, 2008 We present the results of a study of first-layer butane films adsorbed on single-walled carbon nanotubes. We measured 12 isotherms between 180 and 311 K. Butane molecules bind more strongly than shorter alkanes to the nanotubes. We measured a value of 391 meV for the low-coverage isosteric heat of butane. This value is 1.21 times larger than that for butane adsorbed on planar graphite and 1.27 times larger than the value for ethane on nanotubes at comparable coverages. We also compared the characteristics of the adsorption isotherms for butane with those we determined for ethane at the same relative temperatures. This comparison allowed us to infer that there is a change in the adsorption behavior of linear alkanes which occurs as a function of increasing carbon chain length. While ethane isotherms display two substeps in the first layer (corresponding to adsorption on different groups of adsorption sites), one of these steps is significantly smeared for butane isotherms, becoming essentially impossible to resolve above 220 K.
Introduction The study of gas adsorption on carbon nanotubes has been the subject of intense attention in recent years.1 Several experimental and theoretical studies have been performed with the aim of understanding how gases adsorbed on these novel materials behave. Although adsorption studies of small and spherical molecules have been extensively performed on nanotubes,2-10 reports of adsorption of hydrocarbons (with the exception of methane), especially of linear molecules, have been rather limited.11-15 Interest in these systems stems also from their widespread use in the petrochemical industry. A single-walled carbon nanotube (SWNT) can be viewed as a single graphene sheet rolled over itself and closed seamlessly, with hemispherical caps at each end.16,17 The individual nanotubes assemble into bundles due to van der Waals interactions between * To whom correspondence should be addressed. E-mail: aldo@ physics.siu.edu. Phone: (618) 453-1053. Fax: (618) 453-1056. (1) For a review of recent literature on gas adsorption on carbon nanotubes, see: Migone, A. D.; Talapatra, S. In Encyclopedia of Nanoscience and Nanotechnology; Nalwa, H. S., Ed.; American Scientific Publishers: Los Angeles, 2004; Vol. 4, pp 749-767. (2) Calbi, M. M.; Gatica, S. M.; Bojan, M. J.; Cole, M. W. J. Chem. Phys. 2001, 115, 9975. (3) Calbi, M. M.; Cole, M. W.; Gatica, S. M.; Bojan, M. J.; Stan, G. ReV. Mod. Phys. 2001, 73, 857. (4) Calbi, M. M.; Cole, M. W. Phys. ReV. B 2002, 66, 115413. (5) Talapatra, S.; Zhambano, A. J.; Weber, S. E.; Migone, A. D. Phys. ReV. Lett. 2000, 85, 138. (6) Talapatra, S.; Migone, A. D. Phys. ReV. Lett. 2001, 87, 206106. (7) Talapatra, S.; Rawat, D. S.; Migone, A. D. J. Nanosci. Nanotechnol. 2002, 2, 467. (8) Rawat, D. S.; Heroux, L.; Krungleviciute, V.; Migone, A. D. Langmuir 2006, 22, 234. (9) Bienfait, M.; Zeppenfeld, P.; Dupont-Pavlovsky, N.; Muris, M.; Johnson, M. R.; Wilson, T.; DePies, M.; Vilches, O. E. Phys. ReV. B 2004, 70, 035410. (10) Muris, M.; Dupont- Pavlovsky, N.; Bienfait, M.; Zeppenfeld, P. Surf. Sci. 2001, 492, 67. (11) Mao, Z.; Sinnott, S. B. Phys. ReV. Lett. 2002, 89, 278301. (12) Jiang, J.; Sandler, S. I.; Schenk, M.; Smit, B. Phys. ReV. B 2005, 72, 045447. (13) Kondratyuk, P.; Wang, Y.; Johnson, J. K.; Yates, J. T., Jr J. Phys. Chem. B 2005, 109, 20999. (14) Funk, S.; Burghaus, U.; White, B.; O’Brien, S.; Turro, N. J. J. Phys. Chem. C 2007, 111, 8043. (15) Rawat, D. S.; Migone, A. D. Phys. ReV. B 2007, 75, 195440. (16) Ajayan, P. M.; Ebbesen, T. W. Rep. Prog. Phys. 1997, 60, 1024. (17) Saito, R.; Dresselhaus, G.; Dresselhaus, M. S. Physical Properties of Carbon Nanotubes; Imperial College Press: London, 1998.
them. Four possible binding sites have been identified in these bundles: (i) the interior of the individual nanotubes, which is accessible only if the ends of the nanotubes are uncapped and unblocked; (ii) the space between the individual nanotubes at the interior of the bundle, i.e., the interstial channels (ICs); (iii) the convex “valley” formed in the space where two neighboring tubes, on the periphery of the bundles, come closest together, i.e., the grooves; (iv) the cylindrical outer surface of the individual nanotubes that lie at the external surface of the bundles or the external surface sites. For spherical molecules, a consensus has emerged which indicates that, for closed-ended nanotubes, adsorption occurs mainly on the grooves, on the outer surface sites, and on the few large-diameter, defect-induced ICs present;2-10 the much more numerous nondefect ICs are inaccessible for adsorption. Asproduced nanotubes have capped ends, and thus, the interior of the individual nanotubes is not accessible. However, recent studies have suggested that, in the case of HiPco (high-pressure carbon monoxide) process nanotubes, a fraction of open-ended tubes may be present in as-produced samples.18 Purified (i.e., chemically treated) samples have a measurable fraction of nanotubes with their ends uncapped. However, as a result of the chemical treatment, the ends of the nanotubes are blocked by the functional groups, and thus, access to the interior sites is not possible. Studies have shown that these functional groups can be removed by heat treatment, thus enabling adsorption on the interior sites.19,20 Here we present the results of an adsorption study of butane on purified HiPco SWNTs. The measurements were conducted to investigate how the adsorption characteristics for alkanes evolve with increasing carbon chain length. We compared the results for butane on SWNTs to those of a previous study of ethane on the same sample.15 Our results show an increase in the strength of the interaction between the adsorbate and the substrate (i.e., an increase in the value of the binding energy) with increasing carbon chain length. (18) Matranga, C.; Bockrath, B. J. Phys. Chem. B 2005, 109, 4853. (19) Kuznetsova, A.; Yates, J. T., Jr.; Liu, J.; Smalley, R. E. J. Chem. Phys. 2000, 112, 9590. (20) Kuznetsova, A.; Mawhinney, D. B.; Naumenko, V.; Yates, J. T., Jr.; Liu, J.; Smalley, R. E. Chem. Phys. Lett. 2000, 321, 292.
10.1021/la8030705 CCC: $40.75 2009 American Chemical Society Published on Web 12/22/2008
974 Langmuir, Vol. 25, No. 2, 2009
More notably, there appears to be a qualitative difference in the characteristics of monolayer adsorption isotherms for alkanes as a function of the chain length. Ethane films exhibit two fairly well-defined substeps in the first layer. These features correspond to adsorption on groups of sites of two different binding energies. By contrast, in butane isotherms one of these two substeps (the one which corresponds to the higher pressure adsorption that occurs on the weaker sites) is significantly smeared. Above 220 K this substep is no longer resolvable in the butane data. This likely arises due to the size entropy effect, which also affects the value of the effective surface area measured with these two different adsorbates. (Size entropy results when a threedimensional pore, a 2D surface, or a 1D line is packed with linear molecules: linear molecules are not able to pack into these spaces as efficiently as spherical molecules. As a consequence, there are unoccupied sites left and a resulting increase in the entropy. The effect becomes more important as the total space available decreases, i.e., as more molecules occupy the volume, surface, or line.)
Experimental Methods The single-walled carbon nanotube sample employed in this study was purchased from CNI (Carbon Nanotechnologies Inc.). The nanotubes were produced using the HiPco process, and they were purified by the manufacturer. The reported purity of the sample is 91%. Since we did not subject the sample to any vacuum heat treatment, any uncapped tube ends present will likely be blocked by functional groups. Thus, access to the tubes’ interior is not possible in our sample.19,20 The mass of the nanotubes used in our measurements was 0.325 g. The sample was placed inside a copper cell, which was evacuated to a pressure below 10-6 Torr for at least 48 h, at room temperature, prior to the performance of the adsorption measurements. The inhouse-built, automated, volumetric apparatus, used for these adsorption measurements, has been described previously.21 All the pressures were measured using 1, 10, or 1000 Torr full-scale capacitance gauges. Complete monolayer adsorption isotherms were measured at ten different temperatures between 180 and 260 K, while two additional temperaturess301 and 311 Kswere used to conduct low-coverage measurements.
Results and Discussion We studied full monolayer isotherms to investigate and understand the behavior of an adsorbed butane film on SWNTs. We also compared the results of the butane adsorption measurements on SWNTs with those obtained in our previous study of ethane adsorption on the same sample. The main motivation for the current study was to understand how the adsorption characteristics of linear molecules evolve with the molecular length. In Figure 1 we present the result of isotherms taken at 12 different temperatures (the isotherm temperatures increase from left to right). The amount adsorbed (cm3 Torr at 273 K, with 1 cm3 Torr at 273 K ) 3.54 × 1016 molecules) is plotted as a function of the natural logarithm of pressure (Torr), a quantity which is directly proportional to the chemical potential of the adsorbed film. As we have discussed previously,15 the presence of adsorption sites with different binding energies manifests itself as discernible substeps in an adsorption isotherm. In Figure 1 these substeps are shown by arrows. The bottom arrow indicates the presence of a group of high-binding-energy sites in the nanotube sample. (21) Shrestha, P.; Alkhafaji, M. T.; Lukowitz, M. M.; Yang, G.; Migone, A. D. Langmuir 1995, 10, 3244.
Rawat et al.
Figure 1. Isotherms for butane adsorption on single-walled nanotubes at 180, 185, 190, 195, 200, 220, 230, 240, 250, 260, 301, and 311 K (isotherm temperatures increase from left to right). The coverage, on the y axis (cm3 Torr at 273 K, with 1 cm3 Torr at 273 K equal to 3.54 × 1016 molecules) is presented as a function of the natural logarithm of pressure (Torr) (x axis).
Figure 2. Isotherms for ethane (160 K) and butane (220 K) adsorption on single-walled nanotubes. The coverage (cm3 Torr at 273 K) (y axis) is presented as a function of the natural logarithm of pressure (Torr) (x axis). The arrows indicate the presence of substeps in two isotherms.
As can be seen, this substep is present in all isotherms taken at different temperatures between 180 and 311 K. The second substep (indicated by the top arrow in Figure 1), however, is only clearly present for isotherms taken at 180 and 190 K, becoming barely resolvable for isotherms between 190 and 220 K and impossible to resolve above this temperature. A study of ethane adsorption15 on the same sample had indicated the presence of two different groups of binding energy sites in the first layer. In that study, the lower substep was assigned to adsorption occurring on the grooves and, possibly, on higher binding energy defect-induced, large-diameter sites present between nanotubes of different diameters (i.e., defect-induced ICs), while the higher pressure substep was assigned to adsorption occurring on the outer surface sites present on the surface of individual nanotubes of the bundles. The same identification is made for butane (both substeps are resolvable at low temperatures). In Figure 2 we present comparative plots for the results of ethane and butane adsorption on SWNTs. The isotherms are
Butane Monolayer Adsorbed on SWNTs
chosen at the same relative temperatures (as measured with respect to the corresponding critical temperatures) to account for any effects resulting from differences in the isotherm temperatures. The bottom arrows (shown as a group in Figure 2) indicate lowpressure substeps that correspond to adsorption occurring on high-binding-energy sites, i.e., the grooves and the few defectinduced large-diameter interstitials present in the SWNTs. The high-pressure substep (indicated by the top arrows and also shown as a group), however, is present in a clearly resolvable fashion only for the ethane adsorption on SWNTs. The feature, if present, is completely smeared for the butane isotherm at 220 K. In fact, for temperatures in the range 190-220 K there is only a smeared signature of this substep, and for temperatures above 220 K this step is not discernible in the data. By contrast, this higher pressure, lower binding energy substep was clearly visible over a wide range of temperatures for ethane on the same sample. The smearing of this isotherm feature in the butane data is related to both an increasing contribution from the size entropy and an increased contribution from the intermolecular interaction between butane molecules to the free energy of the adsorbed film as the coverage increases. Longer molecules are not able to pack as efficiently as shorter or spherical molecules on the surface of the bundles. As a consequence of the unoccupied sites left, there is an increase in the entropy of the film (size entropy). In Figures 1 and 2 the horizontal axis is proportional to the chemical potential of the film. If the entropy per particle remains reasonably unchanged as the coverage increases, the chemical potential will remain fairly constant, as well, when a group of sites of similar binding energies is being filled (and, thus, a substep will appear in the isotherm, as is the case for ethane in Figure 2). If this is not the case, then the chemical potential will not be essentially constant either, and no substep will result as the similar binding energy sites get filled. This effect will appear as a smearing of the isotherm substep. Smearing resulting from the size entropy will be more pronounced at higher temperatures. For higher temperatures the entropy effects make a more significant contribution to the free energy. Smearing due to size effects and intermolecular interactions will also be more significant for the weaker group of adsorption sites. The (homogeneous) term due to the interaction energy between the adsorbate and the substrate over a group of weaker sites will make a smaller contribution to the free energy; hence, the smearing effects will be more noticeable. These characteristics are observed in the experimental data. Specific Surface Area Determination. We used the “point B” method,22 as illustrated in Figure 3, to compare the relative uptake of ethane and butane by SWNTs and measure the specific surface area of the sample using these adsorbates. The point B method allows the identification of monolayer completion in the adsorption isotherm. After we determine the number of molecules constituting the monolayer using the point B method (i.e., the monolayer capacity), we multiply this by the specific molecular area of the adsorbate. We then divide this product by the mass of the sample. We thus obtain the experimental value of the specific surface area of the nanotubes, as measured with the given adsorbate. The ethane and butane monolayer completions (indicted by arrows in Figure 3) were found at, respectively, 27 625 and 16 575 cm3 Torr at 273 K. If the specific area per molecule for ethane is taken to be 21 Å2/molecule,23 the measured monolayer capacity corresponds to a specific surface area of 643 m2/gram. For butane, (22) Gregg, S. J.; Sing, K. S. Adsorption, Surface Area and Porosity; Academic Press: London, 1967; pp 54-56. (23) Newton, J. C.; Taub, H. Surf. Sci. 1996, 364, 273.
Langmuir, Vol. 25, No. 2, 2009 975
Figure 3. Monolayer isotherms for ethane (160 K) and butane (220 K) adsorption on single-walled carbon nanotubes. The coverage [(cm3 Torr at 273 K)/g] (y axis) is presented as a function of the pressure (Torr) (x axis). After a low-pressure region of steep coverage increase, there is a region where the coverage increases linearly with the pressure; the lowest pressure value where the isotherm deviates from linearity marks point B, as indicated by the arrows parallel to the x axis in the figure. This point is taken to coincide with monolayer completion. The monolayer completions are shown by the arrows in the figure. The top arrow corresponds to the ethane monolayer, while the bottom arrow shows butane monolayer completion.
using a specific surface area of 32.7 Å2/molecule,24 we obtained a value of 557 m2/g for the specific surface area of the same sample. To determine an estimate for the uncertainty in the value of the specific surface area, we also calculated the adsorptive capacities for butane at 190, 200, and 240 K. The value for the specific surface area was found to be 557 ( 4.21 m2/g for the above temperature range. The decrease in the specific surface area measured with the longer linear alkanes is a manifestation of the size entropy effect12 on the surface of the nanotubes. Size entropy effects were used to understand differences in the packing in the pore spaces at the interior of open-ended nanotubes in ref 12. The idea here is entirely analogous: shorter molecules are able to pack more efficiently on the surface of the bundle and, hence, result in a larger value for the specific surface area. (To obtain similar results for the specific surface area of the substrate using adsorption isotherms for linear molecules of different sizes, one must calculate the area using a model that explicitly takes into account the fact that there will be sites in the substrate that cannot be occupied by linear molecules and that the number of these sites will vary with the length of the molecule.25) Binding Energy and Isosteric Heat Results. The isosteric heat of adsorption, qst, is the amount of heat released when an atom gets adsorbed on a substrate. Experimentally this quantity can be calculated from adsorption isotherms by using the following expression:26
( ∂ ∂Tln P )
qst ) KBT 2
N
(1)
In eq 1, kB is Boltzmann’s constant, N is the coverage of the adsorbed gas, ln P is the logarithm of the pressure of the coexisting 3D gas (that is, of the gas present in the vapor phase, above the (24) Herwig, K. W.; Newton, J. C.; Taub, H. Phys. ReV. B 1994, 50, 15287. (25) Roma, F.; Ramirez-Pastor, A. J.; Riccardo, J. L. Surf. Sci. 2005, 583, 213. (26) Wiechert, H. In Excitations in Two-Dimensional and Three-Dimensional Quantum Fluids; Wyatt, A. G. F., Lauter, H. J., Eds.; Plenum Press: New York, 1991; p 499.
976 Langmuir, Vol. 25, No. 2, 2009
Rawat et al.
adsorbed film, inside the cell), and T is the isotherm temperature. In practice, the partial differentials are approximated by the difference between the logarithm of the pressures divided by the difference between the isotherm temperatures, with all quantities evaluated at the same fractional coverage. The relation between the low-coverage value of the isosteric heat of adsorption and binding energy for an adsorbed film is given by27
qst ) E + γKBT
(2)
In eq 2, E is the binding energy and γ is a constant that depends on the dimensionality of the adsorbed film (γ ) 2 in the case of 1D adsorption). Repeated application of eq 1 helps evaluate the variation of the isosteric heat with fractional coverage. For low fractional coverages, we found a value of 391 ( 13.18 meV for isosteric heat.28 In the case of SWNTs the values of the isosteric heat at the lowest fractional coverages provide the binding energy of the most attractive sites present in the sample through the use of eq 2. We used the isotherm results of Figure 1 to calculate the binding energy of butane on SWNTs. The average value of the (low-coverage) binding energy of butane on SWNTs was found to be 345 meV. A previous study of butane adsorption on open-ended SWNTs14 reported a value of 349 meV for the “grooves” site in the nanotubes and a value of 238 meV for the external surface sites; that study used temperature-desorption spectroscopy to determine these values. For our measurements, we found a value of 259 meV for the outer surface sites. The results for the coverage dependence of qst are presented in Figure 4. As is the case for several other adsorbates on SWNTs, the isosteric heat is a generally decreasing function of coverage. The solid line in this figure corresponds to the latent heat of sublimation (218 meV) for bulk butane. A previous computer simulation study of alkane adsorption on SWNTs predicted an increase in the binding energy with increasing molecular chain length similar to the increase we have observed here.12 Although the simulations were performed for open-ended nanotubes, there is general agreement between our experimental results and the simulation predictions. Longer alkanes have more atoms in contact with adsorption sites on the substrate. Hence, the increase in the size of the alkane molecule results in an increase in the interaction between the substrate and molecule. At low fractional coverages this increase leads to higher binding energies. The simulations predicted a linear relation for the increase in the isosteric heat with the number of carbon atoms in the alkane molecule. (27) Wilson, T.; Tyburski, A.; DePies, M. R.; Vilches, O. E.; Becquet, D.; Beinfait, M. J. Low Temp. Phys. 2002, 126, 403. (28) To calculate the uncertainty in the value of the binding energy for the lowest measured fractional coverages, we calculated the uncertainty in the isosteric j )/N]1/2/(N - 1)1/2. heat for those coverages by using the relation ∆qst,i ) [∑(Xi - X Here Xi is the value of the isosteric heat obtained from a pair of temperatures using j is the average value of the isosteric heat for that given fractional coverage eq 1, X using all different combinations of temperature pairs, and N is the number of pairs used. Once we find the uncertainty in the value of the isosteric heat for each lowest measured fractional coverage, the uncertainty in the value of the isosteric heat is calculated by using the relation ∆qst ) (∑∆qst,i)/S, where S is the total number of coverages.
Figure 4. Coverage dependence of the isosteric heat of adsorption on the monolayer fractional coverage. The isosteric heat (meV) (y axis) is presented as a function of the monolayer fractional coverage (layers) (x axis).
We note that some of our results (specifically, the increase in the binding energy, as well as the decrease in the values of the specific surface area measured with increasing chain length) suggest that it would be of interest to perform similar measurements using alkanes longer than butane.
Conclusions We conducted adsorption isotherm measurements of butane on purified HiPco SWNTs. Comparing the results obtained for this alkane with prior results which we had obtained for the shorter molecule ethane, we found strong evidence that the adsorption characteristics depend on the size of the adsorbed molecules. This was manifested in the binding energy values we measured, in the specific surface areas we determined, and in the overall features observed in the isotherms. We determined the binding energy of butane for the grooves and external surface sites in the SWNTs. The binding energy for grooves was found to be 345 meV; for the external surface site, it was 259 meV, in good agreement with a previous determination of these quantities.14 These values are larger than 308 and 167 meV, respectively, measured for ethane. The specific surface area measured with butane, using the point B method, was 557 m2/g. The value measured by the same approach, on this same sample, using ethane was 643 m2/g. Finally, we observed a smearing of the substep in the isotherm corresponding to adsorption on the outer surface of the bundle. We attribute these last two phenomena to size entropy effects and, in the case of the smearing, also to the intermolecular interactions, both of which become more pronounced for longer molecules. Acknowledgment. A.D.M. acknowledges support provided for this study by the National Science Foundation through Grant No. DMR-0705077. LA8030705
