Characterization of single wall carbon nanotubes by nonane preadsorption

Transcription

1 Carbon 44 (2006) Characterization of single wall carbon nanotubes by nonane preadsorption Oleg Byl a, Jie Liu b, John T. Yates Jr. a, * a Department of Chemistry, Surface Science Center, University of Pittsburgh, Pittsburgh, PA 15260, United States b Department of Chemistry, Duke University, Durham, NC 27708, United States Received 11 November 2005; accepted 18 January 2006 Available online 20 March 2006 Abstract The preferential blocking of the interior adsorption sites of single walled carbon nanotubes (SWNTs) by n-nonane is demonstrated. Following adsorption of n-nonane and evacuation for 24 h at, it was found that interior sites with diameters less than 14 Å remained filled with n-nonane, blocking the physical adsorption of N 2 on these sites at 77.3 K. We demonstrate that nonane blocking is a very useful technique for nanotube porosity characterization. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Carbon nanotubes; Adsorption; Surface area; Porosity 1. Introduction Recently much attention has been directed to single wall carbon nanotubes (SWNTs) as potential adsorbents. Carbon nanotubes have tubular structures formed by graphene sheets rolled into seamless cylinders with diameters of a few nanometers and lengths that can reach micrometers. Adsorption in the nanotube interior can provide a significant advantage over adsorption on other forms of carbon used today. A molecule adsorbed in the nanotube interior is bound more strongly to the concave walls than to a flat carbon surface. Nitrogen adsorption at 77 K has been used over decades for sorbent characterization. A number of methods have been developed for N 2 isotherm analysis to obtain specific surface area, external and internal surface areas, microporosity and mesoporosity volumes, pore size distribution etc. There are several reports on N 2 adsorption studies on nanotubes [1 16]. Both theoretical calculations [1,5,12,15] and experimental estimations [13,14,17,18] give the binding * Corresponding author. Fax: address: jyates@pitt.edu (J.T. Yates Jr.). energy for nitrogen adsorbed in nanotubes at zero coverage within the range of kj/mol. This binding energy is greater by about 50% than on planar graphite [17]. The adsorption energy of the outer convex surface is the same or less than that for graphite. Theoretical local adsorption isotherms calculated for nitrogen adsorption in the interior of small nanotubes belong to type I in the IUPAC classification [5,12,19], and are characteristic for microporous materials [20]. If the nanotube diameter is larger than 20 Å then this type of condensation can be referred to as classic capillary condensation described by the Kelvin equation. For smaller nanotubes the concept of capillary condensation cannot be applied due to the lack of precise definition of the meniscus in such small pores [21]. The complexity of nanotube materials requires a reliable way of porosity and surface area characterization. One widely used parameter to characterize N 2 adsorption is a S = n/n 0.4, where n = number of molecules adsorbed at some equilibrium relative pressure, and n 0.4 = number of molecules adsorbed at P/P 0 = 0.4. Traditionally a S plots, n-nonane preadsorption and Deryagin Radushkevich (DR) methods were used for characterization of microporous materials [20]. The recent development of high resolution a S plots makes this method very useful for /$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi: /j.carbon

2 2040 O. Byl et al. / Carbon 44 (2006) discrimination of overlapping micro- and mesoporosities [22 24]. In this paper we report the result of our study of nitrogen adsorption on single walled carbon nanotubes where the interior sites have been blocked by the preadsorption of n-nonane. 2. Experimental section 2.1. Single wall carbon nanotubes Single walled carbon nanotubes were obtained from the Liu group, Duke University, and were prepared by a CVD method with a Fe/Mo catalyst supported on alumina aerogel. Briefly, an alumina boat which was loaded with a certain amount of catalyst was placed into a quartz tube. The tube was heated to 850 C under Ar flow; then the Ar flow was shut off and CO was introduced at 850 C. After 40 min growth time the gas was switched back to Ar and the system was cooled to room temperature. The product (100 mg) was mixed with 5 ml of HF(aq) (52%) and 15 ml of distilled water to dissolve the alumina support. It was stirred for 18 h and then HF was removed. The sample still contained many metal particles coated by carbon layers. To further purify the product, the sample was heated in a 15% air/ar mixture at 440 C for 1 h, and then washed in 5 ml of concentrated HCl(aq) (37%) to dissolve the metal particles formed during the oxidation process. The final purified product consists of SWNTs in bundles, which agglomerate into sheets. TEM images show that the sample still contains a small amount of amorphous carbon. The nanotube diameter distribution is broad ranging from 9 Å to 25 Å as shown in Fig. 1. Thermogravimetric studies indicate that the purified SWNTs also contain a small amount of metal oxide. The material remaining after full combustion in air is less than 5% of the original weight. We estimate that the purity of the SWNT sample is over 90% Nitrogen adsorption Nitrogen adsorption isotherms were measured by the conventional volumetric method on a 15 mg sample. Nanotubes in the form of compact buckypaper were weighed and stored through the whole experiment in a small quartz thimble used to achieve high accuracy in sample weighing. The thimble with a known mass of nanotubes was inserted into a 12 mm OD quartz adsorption tube and annealed at 1073 K for 3 h to remove carbon oxygen functionalities from the nanotubes [25 33]. Upon annealing, the sample was repeatedly weighed and returned to the tube for immediate initial isotherm measurement. A filler rod was utilized to decrease the dead volume of the adsorption tube and to enhance the accuracy of defining of the adsorbed amount of nitrogen. The isotherm measurements were carried out on a Quantachrome Autosorb 1 MP instrument equipped with 1 mtorr pressure transducer. In total, 132 points were recorded starting at 10 6 relative pressure. The temperature transpiration effect was accounted for automatically by the instrument Nonane preadsorption n-nonane (99+% purity) was purchased from Aldrich. An aliquot of nonane was placed into a small bulb and several freeze pump thaw cycles were used to remove air from the liquid. After that nonane was distilled under vacuum into a bulb with zeolite to remove possible traces of water. An IR spectrum of the vapor taken after this purification did not show any impurities. Fig. 1. TEM image of single wall carbon nanotube. The nanotube bundled into ropes of various sizes. Nanotube diameters vary in the 9 25 Å range. A small bar in the inset in the left corner is 10 Å long.

3 O. Byl et al. / Carbon 44 (2006) The adsorption tube containing nanotubes in the thimble was connected to the gas handling system and outgased at 573 K for more than 12 h. After cooling to room temperature the sample was exposed to the vapor of n-nonane at about 4.7 Torr for 30 min. After this the tube was immersed into -N 2 for 5 10 min to condense n-nonane on the wall of the adsorption tube. As the tube warmed the condensed n-nonane melted and filled the thimble. The nanotubes were left for 10 min in liquid n-nonane after this the n-nonane was desorbed at chosen temperatures for more that 24 h. The n-nonane desorption was carried at,,, and 523 K. Following n-nonane desorption, the adsorption tube with the sample was purged with dry nitrogen and immediately connected to the measuring port of the Quantachrome Autosorb 1 MP instrument for N 2 adsorption High resolution a S plots High resolution a S -plots were obtained by plotting the measured amount of nitrogen adsorbed on the nanotubes versus the corresponding a S. We used the values of a S from Ref. [24] where a standard adsorption isotherm was measured on a Cabot BP 280 nongraphitazed carbon black over a relative pressure range from 10 6 to 0.99 resulting in 104 points. The a S at a particular relative pressure was obtained according to the formula a S = n/n 0.4, where n and n 0.4 are the amounts of adsorbed nitrogen from the isotherm of the standard Cabot BP 280 carbon sample at this relative pressure and at P/P 0 = 0.4, respectively. A more detailed explanation about the application of the a S -plot method in N 2 adsorption studies can be found elsewhere [20,22,23,34] DFT pore size distribution The pore size distribution was estimated applying a nonlinear density functional theory (NDFT) kernel supplied by Quantachrome as a part of their software package. The distribution was calculated by fitting the experimental isotherm with a sum of theoretical isotherms that were computed for the carbon slit pore model. The details of this computation are described elsewhere [35 37]. In this paper we present the pore volume histograms and cumulative pore volumes versus pore sizes. 3. Results Fig. 1 shows a TEM image of nanotubes. Nanotubes diameters vary in the 9 25 Å range; however nanotubes with Å diameters make the main fraction. Nanotubes are bundled into ropes of different diameters. Fig. 2 shows nitrogen adsorption isotherms on SWNTs measured at different stages of n-nonane preadsorption. For convenience, labels -, 298-, 323-, 348- and 373- are used on the plots to indicate SWNTs which are unfilled or filled with n-nonane and outgased at various temperatures. Fig. 2(A) shows the isotherms on a semilog N 2 Adsorbed Amount, cm 3 g (STP) A B Relative Pressure Fig. 2. Representative nitrogen adsorption desorption isotherms for nanotubes at different stages of n-nonane preadsorption. (A) Semilog scale; (B) normal scale. relative pressure scale to expand the low pressure area of the isotherms. As can be seen, N 2 adsorption starts at P/ P for the -isotherm whereas the beginning of the 298-isotherm is delayed to P/ P The amount of adsorbed nitrogen at relative pressures below 0.01, i.e. in the microporous regime, is several times higher for the -isotherm than that for the 298-isotherm. All isotherms have a similar hysteresis loop in the relative pressure range on the adsorption branch of the isotherm, as seen from Fig. 2(B). Desorption branches increase their deviations from the adsorption lines from P/P 0 = 1.0 down to P/P then sharply approach the adsorption branches. For the No nonane-isotherm the hysteresis completely closes around P/P whereas desorption and adsorption lines run almost in parallel for 298- and 323-isotherms in Fig. 2(A) down to P/P The mismatch of adsorption and desorption branches below P/P can be explained by partial trapping of nitrogen at higher relative pressures either in the interstitial channels of nanotube bundles or inside nanotubes when filled with n-nonane. In the first case a swelling of nanotube bundles may occur which allows N 2 molecules into the enlarged interstitial channels. In the second case in addition to bundle swelling, nitrogen can mix with n-nonane inside nanotubes. In both cases nitrogen trapping is a kinetic effect related to a potential barrier a molecule needs to overcome in order to desorb; this barrier is higher for nitrogen trapped by mixing

4 2042 O. Byl et al. / Carbon 44 (2006) N 2 Adsorbed Amount, cm 3 /g (STP) Small micropore volume Pore Volume, cm 3 /g A Cumulative Pore Volume α S Fig. 3. High resolution a S -plots obtained for the sample at different stages of nonane preadsorption. The intercepts of the dashed and straight lines give the total pore volume and micropore volumes, respectively. with n-nonane. All the N 2 isotherms presented in Fig. 2 can be classified as type VI according to the IUPAC classification. Fig. 3 shows the high resolution a S -plots for the sample at different levels of n-nonane preadsorption. An upward swing below a S 0.2 is observed for -, 373-, and 348-a S -plots with the highest value for the -a S -plot. This swing originates from the N 2 monolayer adsorption enhanced by the micropore field and is called the filling swing [22]. There is no filling swing in the 298- and 323-a S -plots. Straight line sections of -, 373-, 348-, and 323-a S -plots in the region of a S correspond to monolayer filling in mesopores and macropores and can be used for apparent specific surface area calculations. The slope of this straight line is proportional to the external specific surface area and the intercept is equal to the micropore volume. The next upward swing observed for all a S -plots above a S = 0.6 is called the cooperative swing and corresponds to filling of residual pore volumes left upon N 2 monolayer completion [22]. This process cannot be considered as classic capillary condensation due to the uncertainty of the definition of the meniscus for very narrow pores. In all a S -plots presented here the increase of the cooperative swing slows after a S = 1.6. Fig. 4(A) and (B) shows the cumulative pore volume (CPV) plots and pore size distribution (PSD) histograms, Pore Volume, cm 3 /(gxå) Pore Size Distribution Histogram B Pore Size, Å Fig. 4. (A) Density functional theory cumulative pore volume plots; (B) pore size distribution histograms. respectively, for adsorption isotherms on SWNTs measured at different stages of n-nonane preadsorption. The 298- and 323-CPV plots as well as correspondent PSD histograms start above 10 Å indicating that n-nonane is retained by 10 Å micropores up to. The other regions of the CPV plots and PSD histograms have very similar profiles for all stages of n-nonane preadsorption. This corroborates the previous conclusion that n-nonane is retained only by small micropores with diameters no more than 10 Å. Above 10 Å two families of peaks are observed in the PSD histograms that correspond to large micropores (10 20 Å) and mesopores (20 50 Å). The Å and the Å pores have 0.2 cm 3 /g and 0.5 cm 3 /g volumes and form 10% and 50% of the total pore volume, respectively as calculated for the -sample. Table 1 summarizes the characteristics of the nanotube material studied. The total a S surface area of the sample Table 1 Summarized characteristic of studied nanotube material (the data in the table is for the sample) Method Pore volumes, cm 3 /g Specific surface area, m 2 /g Small micropores (nanotube diameters 614 Å) Large micropores (nanotube diameters Å) Mesopores (20 50 Å pore sizes) Total External Total a S DFT BET + nonane

5 O. Byl et al. / Carbon 44 (2006) was obtained from the slope of the line drawn from the origin and tangent to the -a S -plot at a S = The total a S volume was calculated from the amount of adsorbed N 2 at P/P 0 = 0.98 of the -isotherm. The DFT small micropore and total pore volume were obtained from the -CPV plots for 10 Å and 200 Å pores, respectively. The BET specific surface area was calculated in the region of the isotherm. The external BET surface area was assumed to be equal to the surface area obtained from the 323- isotherm. The small micropore volume was calculated from the difference of nitrogen amounts for the isotherm and 323-isotherm at P/P 0 = (at higher relative pressures all isotherms coincide) where all nanotubes with diameters below 14 Å should be filled [19]. 4. Discussion Nonane blocking of micropores has been extensively used for porosity characterization of different carbons [20,34]. It was shown that nonane, if subjected to pumping below, is retained in different carbons containing micropores narrower than 7 Åor Å when p-orbitals are taken into account [38]. Temperature programmed desorption (TDP) studies of n-nonane adsorption on SWNTs conducted by Kondratyuk and Yates [39] indicate that nonane can be retained in nanotubes with a mean diameter of 13.6 Å upon outgasing at. n-nonane blocking is a kinetic effect so n-nonane can be eventually desorbed; however, N 2 adsorption isotherms were measured at 77 K, and at this temperature the residence time for n-nonane inside carbon nanotubes is infinite on the time scale of the experiment. Fig. 4 presents pore size distributions for carbon nanotubes calculated employing the slitpore model. The 323-PSD shows that nonane is retained by 10 Å pores (or Å diameter nanotubes) that corroborates the TPD experiment. It is known that adsorption in a cylindrical pore occurs at lower pressures relative to a slit pore of comparable size [40]. Therefore a nanotube size distribution calculated adopting the slit-pore model should underestimate nanotube diameters. The origin of this coincidence is not clear and the results obtained from slit-pore DFT pore size distribution are included here only for comparison with other micropore estimation methods. Fig. 4(B) shows that there are three pore size regions in PSD histograms. The first region is below 10 Å and covers small micropores that can be assigned mostly to the interior of small diameter (<14 Å) nanotubes. A rough estimate of the volume of these nanotubes is specified in the Table 1 and is not larger than 0.1 cm 3 /g constituting only 10% of the total pore volume. The reduced contribution of small diameter nanotubes to the total pore volume is corroborated by low nitrogen uptake at P/P 0 <10 5. Below this value of relative pressure nanotubes with diameters below 14 Å should be filled with N 2 at 77 K whereas a 15.2 Å nanotube is only 50 % filled at P/P [19]. The second region covers Å pores or the interior of nanotubes with Å diameters and has 0.2 cm 3 /g volume. Our TEM studies (Fig. 1) show that the fraction of Å nanotubes is lower relative to the nanotubes with diameters below 14 Å. The higher volume of the second region is explained by larger volume of bigger diameter nanotubes, for instance, a 20 Å tube is 2.5 or 6.3 times more capacious than 14 or 10 Å one. The third region encompasses Å mesopores exhibiting 0.5 cm 3 /g volume. A TEM study indicates that the fraction of nanotubes with diameters above 25 Å is not significant. This mesoporosity may result from the voids in the complex net structure formed by nanotube bundles and individual nanotubes. All together nanotubes form only 30% of the total pore volume (as calculated for <20 Å pores from the DFT pore volume plot of the sample in Fig. 4(A)) that, in combination with the high value of the external specific surface area shown in Table 1, indicates that mild air oxidation is not sufficient to open the majority of nanotubes in the sample. This contradicts our assumption made in the earlier report [33] that due to a high BET surface area the main fraction of nanotubes is open. This study shows that adsorption sites other than nanotube interior may significantly contribute to the BET surface area and that the n-nonane blocking method can be very helpful for determining a fraction of opened tubes for carbon nanotube materials with tube diameters below 14 Å. 5. Summary Single wall carbon nanotubes have been characterized by nitrogen adsorption complemented with n-nonane preadsorption. It has been shown that adsorbed n-nonane is retained only on small diameter (<14 Å) nanotubes after desorption at for 24 h. The nanotube interior forms only 30% of the total pore volume and the external surface area contributes substantially to the total surface area indicating that mild air oxidation at 400 C for 1 h is not sufficient for opening the majority of carbon nanotubes. Acknowledgements We gratefully acknowledge the Army Research Office for the support of this work. We also acknowledge the interest and support from Dr. Chris Karwacki at the Aberdeen Proving Ground, and the loan of the Quantachrome Autosorb instrument. References [1] Maddox MW, Gubbins KE. Molecular simulation of fluid adsorption in buckytubes. Langmuir 1995;11(10): [2] Cinke M, Li J, Chen B, Cassell A, Delzeit L, Han J, et al. Pore structure of raw and purified HiPco single-walled carbon nanotubes. Chem Phys Lett 2002;365(1 2): [3] Eswaramoorthy M, Sen R, Rao CNR. A study of micropores in single-walled carbon nanotubes by the adsorption of gases and vapors. Chem Phys Lett 1999;304(3 4):

6 2044 O. Byl et al. / Carbon 44 (2006) [4] Inoue S, Ichikuni N, Suzuki T, Uematsu T, Kaneko K. Capillary condensation of N 2 on multiwall carbon nanotubes. J Phys Chem B 1998;102(24): [5] Paredes JI, Suarez-Garcia F, Villar-Rodil S, Martinez-Alonso A, Tascon JMD, Bottani EJ. N 2 physisorption on carbon nanotubes: computer simulation and experimental results. J Phys Chem B 2003;107(34): [6] Martinez MT, Callejas MA, Benito AM, Cochet M, Seeger T, Anson A, et al. Sensitivity of single wall carbon nanotubes to oxidative processing: structural modification, intercalation and functionalisation. Carbon 2003;41(12): [7] Hu YH, Ruckenstein E. Pore size distribution of single-walled carbon nanotubes. Industrial and Engineering Chemistry Research 2004;43(3): [8] Bossuot C, Bister G, Fonseca A, Nagy JB, Pirard J-P. Textural properties of raw carbon nanotubes by nitrogen adsorption and mercury porosimetry. In: AIP Conf Proc 2001, p [9] Du WF, Wilson L, Ripmeester J, Dutrisac R, Simard B, Denommee S. Investigation of the pore structure of as-prepared and purified HiPco single-walled carbon nanotubes by N 2 /Ar adsorption implication for H 2 storage. Nano Lett 2002;2(4): [10] Yang CM, Kaneko K, Yudasaka M, Iijima S. Effect of purification on pore structure of HiPco single-walled carbon nanotube aggregates. Nano Lett 2002;2(4): [11] Fujiwara A, Ishii K, Suematsu H, Kataura H, Maniwa Y, Suzuki S, et al. Gas adsorption in the inside and outside of single-walled carbon nanotubes. Chem Phys Lett 2001;336(3 4): [12] Ohba T, Kaneko K. Internal surface area evaluation of carbon nanotube with GCMC simulation-assisted N 2 adsorption. J Phys Chem B 2002;106(29): [13] Yoo DH, Rue GH, Chan MHW, Hwang YH, Kim HK. Study of nitrogen adsorbed on open-ended nanotube bundles. J Phys Chem B 2003;107(7): [14] Yoo DH, Rue GH, Hwang YH, Kim HK. Study of nitrogen adsorbed on single-walled carbon nanotube bundles. J Phys Chem B 2002;106(13): [15] Arab M, Picaud F, Devel M, Ramseyer C, Girardet C. Molecular selectivity due to adsorption properties in nanotubes. Phys Rev B 2004;69(16): [16] Li FX, Wang Y, Wang DZ, Wei F. Characterization of single-wall carbon nanotubes by N 2 adsorption. Carbon 2004;42(12 13): [17] Wei BY, Hsu MC, Yang YS, Chien SH, Lin HM. Gases adsorption on single-walled carbon nanotubes measured by piezoelectric quartz crystal microbalance. Mater Chem Phys 2003;81(1): [18] Wilson T, Tyburski A, DePies MR, Vilches OE, Becquet D, Bienfait M. Adsorption of H 2 and D 2 on carbon nanotube bundles. J Low Temp Phys 2002;126(1 2): [19] Agnihotri S, Mota JPB, Rostam-Abadi M, Rood MJ. Structural characterization of single-walled carbon nanotube bundles by experiment and molecular simulation. Langmuir 2005;21(3): [20] Gregg SJ, Sing KSW. Adsorption, surface area, and porosity. 2nd ed. London, New York: Academic Press; [21] Naguib N, Ye HH, Gogotsi Y, Yazicioglu AG, Megaridis CM, Yoshimura M. Observation of water confined in nanometer channels of closed carbon nanotubes. Nano Lett 2004;4(11): [22] Kaneko K, Ishii C, Kanoh H, Hanzawa Y, Setoyama N, Suzuki T. Characterization of porous carbons with high resolution alpha (s)- analysis and low temperature magnetic susceptibility. Adv Colloid Interface Sci 1998;77: [23] Jaroniec M, Kaneko K. Physicochemical foundations for characterization of adsorbents by using high-resolution comparative plots. Langmuir 1997;13(24): [24] Kruk M, Jaroniec M, Gadkaree KP. Nitrogen adsorption studies of novel synthetic active carbons. J Colloid Interface Sci 1997;192(1): [25] Mawhinney DB, Naumenko V, Kuznetsova A, Yates Jr JT, Liu J, Smalley RE. Surface defect site density on single walled carbon nanotubes by titration. Chem Phys Lett 2000;324(1 3): [26] Kuznetsova A, Mawhinney DB, Naumenko V, Yates Jr JT, Liu J, Smalley RE. Enhancement of adsorption inside of single-walled nanotubes: opening the entry ports. Chem Phys Lett 2000;321(3 4): [27] Kuznetsova A, Yates Jr JT, Liu J, Smalley RE. Physical adsorption of xenon in open single walled carbon nanotubes: observation of a quasi-one-dimensional confined Xe phase. J Chem Phys 2000;112(21): [28] Kuznetsova A, Yates Jr JT, Simonyan VV, Johnson JK, Huffman CB, Smalley RE. Optimization of Xe adsorption kinetics in single walled carbon nanotubes. J Chem Phys 2001;115(14): [29] Mawhinney DB, Naumenko V, Kuznetsova A, Yates Jr JT, Liu J, Smalley RE. Infrared spectral evidence for the etching of carbon nanotubes: ozone oxidation at. J Am Chem Soc 2000;122(10): [30] Byl O, Kondratyuk P, Yates Jr JT. Adsorption and dimerization of NO inside single-walled carbon nanotubes an infrared spectroscopic study. J Phys Chem B 2003;107(18): [31] Byl O, Kondratyuk P, Forth ST, FitzGerald SA, Chen L, Johnson JK, et al. Adsorption of CF 4 on the internal and external surfaces of opened single-walled carbon nanotubes: a vibrational spectroscopy study. J Am Chem Soc 2003;125(19): [32] Kondratyuk P, Yates Jr JT. Nanotubes as molecular sponges: the adsorption of CCl 4. Chem Phys Lett 2004;383(3 4): [33] Byl O, Liu J, Yates Jr JT. Etching of carbon nanotubes by ozone a surface area study. Langmuir 2005;21(9): [34] Rouquerol F, Rouquerol J, Sing KSW. Adsorption by powders and porous solids: principles, methodology, and applications. San Diego: Academic Press; [35] Gelb LD, Gubbins KE, Radhakrishnan R, Sliwinska-Bartkowiak M. Phase separation in confined systems. Reports on Progress in Physics 1999;62(12): [36] Seaton NA, Walton JPRB, Quirke N. A new analysis method for the determination of the pore size distribution of porous carbons from nitrogen adsorption measurements. Carbon 1989;27(6): [37] Ravikovitch PI, Vishnyakov A, Russo R, Neimark AV. Unified approach to pore size characterization of microporous carbonaceous materials from N 2, Ar, and CO 2 adsorption isotherms. Langmuir 2000;16(5): [38] Carrott PJM, Drummond FC, Kenny MB, Roberts RA, Sing KSW. Desorption of n-nonane from microporous carbons. Colloids Surf 1989;37:1 13. [39] Kondratyuk P, Yates Jr JT. Desorption kinetic detection of different adsorption sites on opened carbon single walled nanotubes: the adsorption of n-nonane and CCl 4. Chem Phys Lett 2005;410(4 6): [40] Balbuena PB, Gubbins KE. The effect of pore geometry on adsorption behavior. In: Rouquâerol J, Rodriguez-Reinoso F, Sing KSW, Unger KK, editors. Characterization of porous solids III: proceedings of the IUPAC Symposium (COPS III), Marseille, France, May 9 12, Amsterdam, New York: Elsevier; p