ARTICLE pubs.acs.

org/JPCC

Single-Walled Aluminosilicate Nanotubes with Organic-Modified Interiors

Dun-Yen Kang, Ji Zang, Christopher W. Jones,* and Sankar Nair*
School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive NW, Atlanta, Georgia 30332-0100, United States
S b Supporting Information

ABSTRACT: A methodology for modifying the interior of single-walled metal oxide (aluminosilicate) nanotubes by covalently immobilizing organic functional entities on the interior surface of the nanotube structure is reported. Characterization of the modied nanotubes by a range of solid-state characterization techniquesincluding nitrogen physisorption, thermogravimetric analysis, transmission electron microscopy (TEM), powder X-ray diraction (XRD), and solid-state NMR strongly indicates that the organic entities are immobilized on the inner surface of the nanotubes by reaction with the silanol groups on the interior wall. The resulting organic-modied single-walled nanotubes (SWNTs) show higher hydrophobicity than bare nanotubes based upon water adsorption measurements. Furthermore, a mechanistic understanding of water adsorption in the modied SWNTs is developed, by interpretation of the water adsorption data with a multilayer adsorption model. The degree of interior surface silanol substitution is estimated, with up to 35% of the silanols being substituted through the present modication chemistry. This methodology of immobilizing various functional entities at the inner wall of aluminosilicate nanotubes opens up a range of previously inaccessible molecular recognition-based applications for nanotube materials in areas such as catalysis, molecular encapsulation, sensing, and separation.

1. INTRODUCTION Single-walled nanotubes (SWNTs) have been considered important building blocks in the development of nanotechnology for more than a decade. Synthetic carbon SWNTs have been investigated extensively for potential applications18 based upon their unique dimensions and structure. The necessity of developing processing routes to carbon-based SWNT nanostructures and devices has led to intensive study of surface functionalization/modication of both the outer and inner surfaces of carbon SWNTs. While outer-surface modication is usually intended to increase the compatibility of the nanotube with other solidor liquid-phase materials,914 interior modicationespecially immobilization of functional groups inside the nanotube by covalent bondscould open up an array of new applications based upon molecular recognition, such as molecular separation, molecular storage, catalysis, and drug delivery. The inltration of lipids,15 metals,16,17 and C60 beads18,19 into carbon SWNT channels has been reported. Functional entities have also been grafted at the tips of carbon SWNTs.10,20,21 However, the covalent functionalization of the inner surfaces of SWNTs has remained a long-standing challenge. The inner wall of carbon SWNTs has relatively low reactivity and also suers from steric and transport limitations in delivering potentially reactive functional entities to the desired sites in the carbon SWNTs.21,22 On the other hand, synthetic metal oxide/hydroxide SWNTs2327 can be expected to possess properties quite dierent from carbon
r 2011 American Chemical Society

nanotubes. Such SWNTs can be used to address the problem of interior functionalization because the metal oxide/hydroxide inner surfaces are more reactive and thus amenable to surface modication than the graphitic sheets of carbon nanotubes. More specically, synthetic aluminosilicate SWNTs, which were rst synthesized in 197728 and thereafter well characterized regarding their dimensions,2835 structure,3640 surface composition,4043 bundling characteristics,29,31,40,44,45 and formation mechanisms,31,4650, have attracted substantial interest in recent years. This SWNT consists of an aluminum(III) hydroxide sheet on the outer surface and is lined with pendant silanol groups on the inner surface (Figure 1). These silanols can potentially be functionalized in a manner analogous to the well-known techniques for functionalization of porous silicas.5158 The capability to control the chemistry of the inner surface of the aluminosilicate SWNTs via introduction of desired functional groups could thus have signicant implications for nanotube science and engineering. There have been several reports of the outer surface modication of single-walled aluminosilicate nanotubes.5962 However as in the case of carbon nanotubesthe inner wall modication has proven to be much more dicult. The direct synthesis of aluminosilicate SWNTs containing covalently attached organic
Received: February 1, 2011 Revised: March 16, 2011 Published: March 28, 2011
7676
dx.doi.org/10.1021/jp2010919 | J. Phys. Chem. C 2011, 115, 76767685

The Journal of Physical Chemistry C

ARTICLE

Figure 1. Structure of the as-synthesized single-walled aluminosilicate nanotube.

groups is one potential approach; however, it is only recently that Bottero et al.43 have reported a successful synthesis of SWNTs containing methyl groups on their inner walls via introduction of methylsiloxane precursors in the reactant solution. The ecacy of this method for introduction of more complex organic functional groups is a subject of considerable future interest. Ackerman et al.63 reported that a silane reagent was used to modify the inner surface of the aluminosilicate SWNT, but no detailed characterization of the resulting material was presented. Modication of the aluminosilicate SWNT interior is impeded by its high hydrophilicity at ambient conditions,40 due to its high inner surface silanol density (9.1 OH/nm2).43 Therefore, a successful interior modication may not be achieved without a high degree of dehydration of the SWNT samples. Our previous SWNT dehydration study showed that a heat treatment at 250300 C under vacuum, which removes the physisorbed water inside the SWNT channels while preserving the nanotube structure, may be an optimal pretreatment allowing for SWNT interior modication.40 In this report, we describe a general strategy for the interior modication of aluminosilicate SWNTs, as illustrated by three reagents: acetyl chloride, methyltrimethoxysilane, and trichlorosilane. The modied SWNTs are assessed by a combination of solid-state techniques including nitrogen physisorption, thermogravimetric analysis (TGA), powder X-ray diracton (XRD), transmission electron microscopy (TEM), solid-state NMR, and water adsorption. On the basis of these results, we demonstrate that the organic functional groups are immobilized on the inner wall of the aluminosilicate SWNT by condensation with the inner-surface silanols. Furthermore, a detailed study of water adsorption yields an understanding of adsorption mechanisms in the bare and modied SWNT samples and demonstrates the variation of SWNT surface hydrophilicity resulting from interior modication. The reported methodology for modifying SWNTs with various organic groups can be broadly applied to make the SWNTs functional for a range of applications involving selective interactions of the nanotube with molecules based on their shape, size, and chemical properties.

transferred to a nitrogen glovebox, and ca. 5 mL of hexane solvent was added into the flask. The functionalizing reagent (acetyl chloride, trimethylmethoxysilane, or trichlorosilane) was then transferred into the flask, with the reagent to SWNT hydroxyl group molar ratio being 2. The mixture was allowed to stir under nitrogen for 24 h. The flask was then connected to the vacuum line and treated at 180 C for 24 h to remove the solvent and unreacted reagent. The resulting powder samples were used for characterization studies. The label NT denotes the bare SWNT, whereas NT-A, NT-M, and NT-T denote the SWNT treated by acetyl chloride, methyltrimethoxysilane, and trichlorosilane, respectively. 2.2. Solid-State NMR. The SWNT sample was first packed into a 7 mm rotor. 13C, 27Al, and 29Si MAS NMR experiments were carried out on a Bruker DSX 300 spectrometer at frequencies of 75.5, 78.1, and 59.6 MHz. For 13C cross-polarization (CP) MAS NMR studies, the sample was spun at 5 kHz, and a single pulse of /2 and repetition time of 4 s was used. The sample was spun at 56 kHz for 27Al MAS NMR tests, for which a single pulse of /6 and a repetition time of 0.1 s was used. For 29Si MAS NMR, directpolarization (DP) and cross-polarization (CP) tests were performed with repetition times of 10 and 5 s, respectively, at /2 single pulse and 5 kHz spinning rate. The chemical shifts of 13C, 27Al, and 29 Si were referenced to adamantane, aluminum trichloride, and 3-(trimethylsilyl)-1-propanesulfonic acid sodium salt, respectively. 2.3. X-ray Diffraction (XRD). Powder X-ray diffraction (XRD) was performed on a PAnalytical Xpert Pro diffractometer operating with a Cu KR source. The high-resolution diffraction data were collected with a diffracted-beam collimator and a proportional detector, scanning from 2 to 30 two theta with a step size of 0.05. 2.4. Transmission Electron Microscopy (TEM). Approximately 5 mg of SWNT sample was first dispersed in 10 mL of deionized water. The resulting dispersion was sonicated for 10 min. Around 5 drops of the sonicated SWNT dispersion were added on 300-mesh copper grids coated with Formvar layers. Transmission electron microscopy (TEM) images were recorded on a Hitachi HF2000 field emission gun TEM operated at 200 kV. 2.5. Thermogravimetric Analysis (TGA). The experiment was performed with a Netzsch STA409 instrument. Approximately 20 mg of powder sample was heated under nitrogendiluted air from 25 to 900 C with a ramp rate of 10 C/min. 2.6. Nitrogen Physisorption. Nitrogen physisorption measurements were carried out on a Micromeritics Tristar II at 77 K. The sample was placed in an analysis tube and degassed under 15 mTorr at 200 C for 12 h before the physisorption measurement. 2.7. Water Adsorption. Water adsorption measurements were performed on IGAsorp (Hiden Analytical, Warrington, U.K.) at 25 C. The sample was outgassed at 200 C for 8 h prior to recording the isotherm.

3. RESULTS AND DISCUSSION

3.1. Porosity, Structure, and Organic Loading. The nitrogen physisorption isotherms (Figure 2) of the as-made and the three modified SWNT samples all show the characteristics of IUPAC type I isotherms,64 suggesting that the pore channels of the modified SWNT samples are microporous, as expected. More detailed information can then be extracted by employing the BET model65 and t-plot method66 to these isotherm data. The BET model yields the total surface area (SBET), contributed by both interior and outer surfaces of the SWNT.67 (The BET model for interpreting nitrogen physisorption isotherms from
7677
dx.doi.org/10.1021/jp2010919 |J. Phys. Chem. C 2011, 115, 76767685

2. EXPERIMENTAL SECTION
2.1. Interior Modification of Aluminosilicate SWNT. The synthesis and purification of the as-synthesized SWNT sample is reported in our previous work.40 For SWNT interior modification, 500 mg of as-synthesized SWNT powder was first placed in a flask connected to a 15 mTorr vacuum line and heat treated at 250 C for 24 h, after which it is considered fully dehydrated based upon our previous study. The heat-treated SWNT sample was then

The Journal of Physical Chemistry C

ARTICLE

Figure 2. Nitrogen physisorption isotherms of as-synthesized and modied SWNTs, where NT denotes the bare SWNT, NT-A denotes SWNT treated by acetyl chloride, NT-M denotes SWNT treated by methyltrimethoxysilane, and NT-T denotes SWNT treated by trichlorosilane.

Figure 3. XRD patterns of as-synthesized and modied SWNTs.

Table 1. SWNT Sample Porosity Derived from Nitrogen Physisorption Data

BET method SBET sample NT NT-A NT-M NT-T (m2/g-Al2O3SiO2) 418 256 153 260 Vmp (cm3/g-Al2O3SiO2) 0.17 0.11 0.06 0.11 t-plot method Sext (m2/g-Al2O3SiO2) 10.1 15.9 11.3 14.0

microporous materials (pore size <2 nm) should be used with caution, since the concept of monolayer adsorption is not definitive when the window size is close to the size of adsorbates. Rouquerol et al.50 have assessed the applicability of the BET model for microporous materials and have suggested the BET model is valid and yields a reliable monolayer adsorption quantity if two criteria are satisfied: (1) n(P0 P) should increase as P/P0 in the applicable region, and (2) the fitted linear correlation should have a positive intercept, where n is the adsorption quantity, P the pressure of nitrogen, and P0 the saturation pressure of nitrogen at 77 K. In our calculations, we chose a pressure range of 0.005 < P/P0 <0.05 wherein the two criteria are fulfilled. Finally, the obtained monolayer adsorption quantities were converted into surface areas under an assumption that the distances between any two adjacent adsorbed nitrogen molecules are identical.) On the other hand, the t-plot method is well-known for differentiating mesoporosity from microporosity present in the same sample.66,6870 Specifically, the linear fitting of the isotherms with HJ correlations71,72 allows us to estimate the external surface area (Sext) contributed by the outer surface of the SWNT and the micropore volume (Vmp) due to the pore volume in the SWNT channels. The derived t-plots are presented in the Supporting Information, and the values of SBET, Vmp, and Sext are summarized in Table 1. For the as-synthesized SWNT, SBET is about 42 times larger than Sext, suggesting there is significantly larger accessible surface area at the interior of the SWNT in comparison to the outer surface.

The relatively small accessible external surface area of the SWNT is likely due to the packing of SWNTs into bundles. After treatment with the three different reagents, all the samples show substantial decreases in both Vmp and the internal surface area SBET Sext, thereby providing direct evidence that most of the surface modification reaction takes place at the interior of the SWNTs and that the introduced organic entities are immobilized in the SWNT channels. However, the amount of decrease in SBET Sext and Vmp for the three modified SWNT samples is strongly related to the molecular size of the reagent and the fractional silanol substitution at the SWNTs inner surface (which can also be considered the loading of the reagent). A quantitative analysis of the fractional silanol substitution is discussed later in this report. A deviation of Sext from bare to modied SWNTs is also observed, due to the variation in SWNT bundling characteristics between samples. In particular, the Sext for as-synthesized SWNT from ve batches shows an average of 12.2 with a standard deviation of 4.5 m2/g-Al2O3SiO2 (Supporting Information). On the other hand, Vmp and SBET of the as-made SWNT from ve batches show averages of 0.168 cm3/g-Al2O3SiO2 and 417 m2/g-Al2O3SiO2 with relatively small standard deviations of 0.008 cm3/g-Al2O3SiO2 and 18 m2/g-Al2O3SiO2, respectively. Hence, the dierences of Sext between bare and modied SWNTs listed in Table 1 are within the statistical variation, whereas the deviations of Vmp and SBET Sext from bare to modied SWNTs are statistically meaningful. As a consequence, the analysis from nitrogen physisorption measurements reveals that the modied SWNT samples possess signicantly lower pore volumes and total surface areas (dominated by the inner surface area of nanotubular channels) than the bare SWNTs, whereas no statistically signicant deviation in the external surface areas is observed, hence clearly suggesting that the surface modication takes place in the interior of the SWNT. While nitrogen physisorption analysis elucidates the porosity and surface area of the as-made and modied SWNTs, X-ray diraction (Figure 3) gives information on the morphology and bundling of the SWNTs. XRD patterns of nanotubular materials have been theoretically and experimentally studied in detail in previous reports.40,7376 It is clear that the diraction patterns of nanotubes forming small bundles are not dominated by Bragg diraction but by X-ray scattering, as opposed to ordered porous materials with onedimensional channels such as 1D-channel zeolites, MCM-41, or SBA-15. Previous XRD studies on single-walled carbon nanotubes
7678
dx.doi.org/10.1021/jp2010919 |J. Phys. Chem. C 2011, 115, 76767685

The Journal of Physical Chemistry C

ARTICLE

Figure 4. TEM images of as-synthesized and modied SWNT bundles. The scale bars (in black or white) represent 20 nm. The pairs of red arrows represent individual SWNTs, with approximately 2 nm diameter.

have elucidated the diraction pattern for samples containing isolated nanotubes.7476 The bundling of nanotubes leads to additional characteristic shoulder-like peaks. Comparisons of experimental and simulated XRD patterns have provided denitive characterizations of the bundling of carbon SWNTs. Similarly, our previous XRD simulation study on the aluminosilicate SWNTs showed that the patterns of Figure 3 unequivocally represent the structure of SWNTs organized in bundles. The as-synthesized SWNT sample shows high purity, as reected in the sharply dened SWNT characteristic peaks (between 3 and 19). Similarly, the three organic-modied SWNT samples show well-dened XRD patterns nearly identical to that of the bare SWNT, with no evidence of structure amorphization or alteration. The nanotubular structure and the SWNT bundling characteristics therefore remain unchanged after the surface modication. While XRD reveals the nanotubular structure and high bulk purity of the bare and modied SWNTs, TEM images provide localized visual information on the samples and conrm the XRD results. The TEM image from the as-synthesized SWNT sample (Figure 4a) clearly shows bundles of aligned nanotubes with a ca. 2 nm line-to-line distance representing the outer diameter of the SWNT. After treatment with the organic reagents, the channels of the modied SWNT samples remain intact as shown in Figures 4b4d. Although the SWNTs form dense bundles on the TEM grid after evaporation of the solvent, one can also occasionally observe isolated nanotubes (as seen in Figure 4d). Thermogravimetric analysis was employed to investigate the mass losses associated with heating the SWNTs in diluted air, including losses associated with physisorbed water, surface

Figure 5. Dierential TGA curves of as-synthesized and modied SWNTs.

hydroxyl groups, and grafted organic groups. The rst-derivative TGA curves are summarized in Figure 5, and the original TGA traces are presented in the Supporting Information. For the bare
7679
dx.doi.org/10.1021/jp2010919 |J. Phys. Chem. C 2011, 115, 76767685

The Journal of Physical Chemistry C Table 2. Normalized Weight of Water/Hydroxyl Groups/ Organic Groups in As-Synthesised and Modied SWNT Samples Determined by TGA
normalized mass physisorbed water (g-H2O/g-Al2 sample NT NT-A NT-M NT-T O3SiO2 %) 30.9 15.3 12.5 13.3 hydroxyl group (g-OH/g-Al2 O3SiO2) 21.3 19.4 14.5 13.6 organic group (g-organic/gAl2O3SiO2 %) -12.6 9.2 8.2

ARTICLE

Scheme 1. Reactions at the SWNT Interior by Various Reagents: Acetyl Chloride, Methyltrimethoxysilane, and Trichlorosilane, Denoted by NT-AcCl, NT-MTMS, and NTTClS, Respectively

SWNT, two pronounced peaks, one ranging from 25 to 250 C and the other from 250 to 450 C, are observed. They are assigned to the loss of physisorbed water and hydroxyl groups, respectively, following our previous work.40 An additional peak in the 450600 C region is observed in all three modied SWNTs and is assigned to the loss of the introduced organic entities. Furthermore, considering the low boiling points of the reagents used (52 C for acetyl chloride, 102 C for methyltrimethoxysilane, and 32 C for trichlorosilane), the relatively high temperature (450600 C) at which the organic groups are lost implies that they are covalently bonded to the surface and not physisorbed on the inner surfaces of the SWNTs. The mass loadings of physisorbed water, hydroxyl groups, and organic groups in the bare and modied SWNTs, all normalized by the mass of dry aluminosilicate (Al2O3SiO2) at the end of TGA measurement (900 C), are summarized in Table 2. For the three modied SWNT samples, decreases of the physisorbed water loading in comparison to the as-made SWNT suggest that all the surface modications yield a lower hydrophilicity in the SWNT. A more detailed hydrophilicity analysis, based on water adsorption measurements, is presented below. Apart from the decrease of physisorbed water loading, a decrease of the hydroxyl group loading accompanying the organic group loading is also observed in the three modied samples. This clearly shows that the surface silanols in SWNTs are partially substituted by the surface modication reagents. 3.2. Surface Reaction Schemes. Upon the basis of the results reported above, it is likely that the reagents (acetyl chloride, methyltrimethoxysilane, and trichlorosilane) react with surface silanols in the SWNT interior and are therefore immobilized on the surface. Therefore, we propose reaction schemes for the treatment of the SWNT with different reagents (Scheme 1) in analogy to the reported surface modifications of silicate materials by acid halides,77,78 methoxysilanes,52,53,58,79 and chlorosilanes,8083 respectively. The surface products associated with the proposed reaction schemes are then examined by 29Si and 13C solid-state NMR. The 29Si CP-MAS NMR spectra (Figure 6a and 6b) provide a molecular-level characterization of the silicon environment in the bare and modied SWNTs. For as-synthesized SWNTs, the most pronounced peak is at 79 ppm and is assigned to the unique Q3(6Al) silicon framework36,40 which is not commonly found in other aluminosilicate frameworks. The small and relatively broad peak at 90 ppm is assigned to Q4(6Al) and considered the contribution from defect sites involving a small number of condensed silanols present in as-made SWNT samples.36,40 The 29Si CP-MAS spectrum from the as-prepared SWNT sample implies the high purity of the sample. For modied SWNT

samples, slight increases of the Q4(6Al)/Q3(6Al) ratio are likely due to the heat treatment at 250 C prior to the modication reaction40 as well as the immobilization of the introduced silanes on the SWNT inner surface, specically for NT-M and NT-T. Moreover, the grafted acetyl group is likely to have a negligible eect on Q4(6Al)/Q3(6Al) ratio since our previous study suggests that the silicon chemical shifts of the SiOC and SiOH environments are not signicantly dierent.55 For NT-M, there are several additional peaks besides the Q3(6Al) and Q4(6Al) in the spectrum shown in Figure 6b. These peaks are assigned to T0 (45 ppm), T1 (53 ppm), T2 (59 ppm), and T3 (63 ppm), respectively, and these assignments are in good agreement with previous reports.79,84 Furthermore, the agreement of the chemical shifts of the T1, T2, and T3 species with those reported for silane modications on porous silicas also suggests that the reaction takes place on the SWNT inner surface (composed of Q3 silicon). In contrast, a reaction on the SWNT outer surface (composed of octahedrally coordinated aluminum)
7680
dx.doi.org/10.1021/jp2010919 |J. Phys. Chem. C 2011, 115, 76767685

The Journal of Physical Chemistry C

ARTICLE

Figure 6. (a) 29Si CP/MAS NMR spectra of SWNT samples. (b) A detailed view of 29Si CP/MAS NMR spectra for NT-M and NT-T.

would signicantly alter the silicon chemical shifts of T1, T2, and T3 species due to the formation of SiOAl linkages.8588 On the other hand, we also assigned the peaks in the 29Si CP/MAS spectrum (Figure 6b) for T0 groups present in the NT-T sample: T0 3 at 87 ppm, T0 2 at 73.8 ppm, and T0 1 at 64.9 ppm based upon a previous report89 wherein it has been suggested that the T0 3 chemical shifts are in the range of 70 to 85 ppm, T0 2 in the range of 50 to 68 ppm, and T0 1 in the range of 30 to 40 ppm. However, with regard to the disagreements in the peak assignments, we performed a control experiment by modifying a well-known mesoporous silicate material (SBA-15) with trichlorosilane (Supporting Information). The 29Si CP-MAS NMR spectrum for this control sample shows good consistency of the chemical shifts of T0 groups with those seen in the NT-T sample. Finally, it should be cautioned that the relative peak areas in the CP-MAS spectra do not provide reliable quantitative information on the molar ratio of dierent T groups. A quantitative analysis can be performed using 29Si direct polarization (DP) spectra (Supporting Information) and is discussed below. While 29Si CP/MAS NMR probes the silicon coordination, 13 C CP-MAS NMR (Figure 7) is an excellent probe of the immobilized organic entities. For the NT-A sample, there are two peaks clearly assigned to C1 (the methyl group at 22.9 ppm) and C2 (the carbonyl group at 180.8 ppm), thereby strongly supporting the proposed scheme for acetyl chloride immobilized on the SWNT inner surface. On the other hand, the two peaks observed in the spectrum of NT-M are assigned to the C3 (the methyl group at 4.5 ppm) and C4 (the methoxy group at 48.8 ppm),79 in good agreement with the proposed scheme for modication by methyltrimethoxysilane. 27Al NMR (Figure 8) spectroscopy was then used to investigate the possible alteration of the SWNT outer surface during surface modication. The peak at 4 ppm for the as-made SWNT is assigned to octahedral aluminum.38,40,90,91 The 27Al spectra remain unchanged, and no additional peaks are observed in the three modied samples. This implies that the nanotube wall remains intact during surface modication, since previous studies suggest that any partial degradation of the outer wall of the nanotube is accompanied by the appearance of tetrahedral or pentacoordinated aluminum.38,40

Figure 7.

13

C CP/MAS NMR spectra of NT-A and NT-M.

No signicant shift of the octahedral aluminum peak is seen, suggesting that the silane reagents do not modify the SWNT outer surface. 3.3. Water Adsorption. The water adsorption isotherms measured at 25 C are summarized in Figure 9, with the adsorption quantity normalized by the amount of dry aluminosilicate (Al2O3SiO2) as obtained from the TGA results (Table 2). After surface modification by the three reagents, the water uptake capacity of the SWNTs decreases substantially to about 6075% of the bare SWNT capacity, suggesting that the modified samples become more hydrophobic. However, a decrease of water capacity in the modified SWNT can also be rationalized by a lower pore volume (verified by nitrogen physisorption) as well as the variation of surface hydrophilicity after modifications. A mechanistic model is necessary to gain physical insight into the water adsorption isotherms in SWNTs. Upon the basis of Grand Canonical Monte Carlo (GCMC) simulation results (Figure 10a) of water adsorption in bare SWNTs (carried out in a manner identical to our previous study),40,92,93 it is clear that the water molecules can form
7681
dx.doi.org/10.1021/jp2010919 |J. Phys. Chem. C 2011, 115, 76767685

The Journal of Physical Chemistry C

ARTICLE

Figure 8.

27

Al NMR of as-synthesized and modied SWNT samples.

Figure 10. (a) Simulated bare SWNTwater models with dierent water loading obtained by grand canonical Monte Carlo (GCMC) simulations. (b) Proposed water adsorption mechanisms in the channels of bare and modied SWNTs.

monolayer and subsequent multilayer adsorption. Usually, the constant C is several orders of magnitude larger than unity64,94 for N2 adsorption at 77 K as well as in our water adsorption fitting results. Hence eq 1 can be simplified to P=P0 1 1 P=P0 nm nm C n1 P=P0 2

Figure 9. Water adsorption isotherms at 25 C of as-synthesized and modied SWNT samples.

multiple layers inside the SWNT, beginning from a monolayer at low chemical potential. On the basis of our previous work, it is clear that the first layer of water molecules is primarily formed by hydrogen bonding between water molecules and surface silanol groups, whereas the subsequently adsorbed water layers form by hydrogen bonding between water molecules. Therefore, we propose a model for water adsorption in modified SWNTs as illustrated in Figure 10b. The proposed mechanism includes two phenomena: (1) formation of the first adsorbed water layer, with the water molecules only hydrogen bonding on the sites at which the surface silanols have not been substituted by organic groups; and (2) a subsequent water layer forms adjacent to the first water layer by hydrogen bonding between two adjacent water molecules. The BET model,64,65,94 which captures multilayer adsorption phenomena, can be used to model the water adsorption mechanism in the SWNT P=P0 C 1 1 P=P0 nm C nm C n1 P=P0 1

where P is the pressure of water vapor; P0 is the saturated water vapor pressure at a given temperature; n is the adsorption quantity (g-water/g-Al2O3SiO2); nm is the monolayer coverage (g-water/gAl2O3SiO2); and C is the ratio of the equilibrium constants for the

Based on eq 2, a plot of (P/P0)/(n(1(P/P0)) vs (P/P0) is the wellknown BET plot for multilayer adsorption phenomena and is applicable in the moderate relative pressure region. The applicable pressure region for the BET plot is well-defined for nitrogen physisorption at 77 K (0.05 < P/P0 < 0.35 for mesoporous materials64 and P/P0 < 0.05 for microporous materials95). However, some reports have suggested that the BET plot for water adsorption can be applied in the relative pressure range of 0.05 < P/P0 < 0.5.9699 We chose data in the range 0.1 < P/P0 < 0.35, wherein the four BET plots show high linearity, to fit eq 2 (Figure 11). The fitted linear correlations all have positive intercepts, implying that it is feasible to apply the BET model in the assumed pressure region.67,68,95 The fitted slope of the BET equation gives the monolayer water coverage nm, and these values are summarized in Table 3. A decrease of nm between bare and modified SWNTs clearly suggests that a certain fraction of silanols in the SWNT interior are substituted during surface modification and are hence unavailable for monolayer adsorption of water. Furthermore, the introduced reagents create hydrophobic regions in the SWNT. These two factors are responsible for a lower water uptake capacity of the modified SWNTs in both the intermediate and high-pressure regions. In contrast, the lowpressure region shows negligible differences since the Henrys constant for initial water adsorption on available silanol sites remains essentially unaffected. 3.4. Fractional Silanol Substitution. In this section, we estimate the fractional surface silanol substitution after interior modifications by three introduced reagents, from the results of the different characterization techniques including nitrogen physisorption, TGA, 29Si NMR, and water adsorption. The fractional silanol substitution is physically equivalent to the surface coverage of the introduced organic entities on the inner
7682
dx.doi.org/10.1021/jp2010919 |J. Phys. Chem. C 2011, 115, 76767685

The Journal of Physical Chemistry C

ARTICLE

Figure 11. BET plots derived from water adsorption isotherms for assynthesized and modied SWNT samples.

Table 3. Monolayer Coverage Fitted from Water Adsorption BET Plots

sample NT NT-A NT-M NT-T nm (g-H2O/g-Al2O3SiO2) 0.174 0.118 0.107 0.131

Figure 12. Illustration of SWNT modied by various reagents.

Table 4. Fractional Silanol Substitution in Bare and Modied SWNTs

fractional silanol substitution (# of silanols being substituted in modied SWNT/ # of silanols in bare SWNT) N2 physisorption sample TGA NT-A NT-T 0.35 0.26 (liquid density) 0.25 0.24 0.28 N2 physisorption (molecular size) 0.34 0.37 0.30 water adsorption 0.32 0.38 0.25

NT-M 0.33

wall of the SWNT. The results are summarized in Table 4, and the methodology is described as follows. The estimate of fractional silanol substitution from the nitrogen physisorption data is based upon the assumption that the lower micropore volume of modified SWNTs is due to the introduced organic entities in SWNT channels. Two approaches are then possible to interpret the micropore volume decreases in terms of the surface coverage of organic groups. The first approach assumes that the molar density of the organic groups embedded inside the SWNT channels is unchanged from its value in a pure liquid phase at 25 C. The mass of the organic groups in the SWNT can then be calculated. This quantity can finally be converted into the molar loading (mol organic group/g-Al2O3SiO2) using an average molecular weight for the different immobilized types of groups (for NT-M and NT-T, based upon the molar ratio between T groups derived from 29Si DP spectra as presented in the Supporting Information). In the second approach, one can estimate the sizes of the organic groups grafted in the modified SWNTs (see Supporting Information) and directly interpret the

pore volume decrease for modified SWNTs in terms of the surface coverage. The surface coverage can also be derived from TGA data. We can utilize the organic loading from TGA data (Table 2) to approximate the SWNT inner surface coverage. This approach converts the gravimetric data into a molar loading by using an average molecular weight derived from the NMR results (similar to the approach used for the nitrogen physisorption data in the case of NT-M and NT-T). Finally, the decrease of monolayer water coverage (summarized in Table 4) in modified SWNTs can be considered equivalent to the decrease of interior surface silanol loading due to the modifications and hence allows calculation of the organic loading. Table 4 shows the results of the estimates from the four different methods. It is clear that 2535% of the interior silanols have been substituted for NT-A, 2438% for NT-M, and 2530% for NT-T. It is hypothesized that the surface reaction in the SWNT interior may ultimately be limited by the slow diffusion of the organic entities due to partial blockage of the SWNT pores by already immobilized organic groups. This aspect requires further study to reach a conclusive characterization. Considering the large stoichiometric excess (8:1) between the surface modification reagents and the silanol groups on the inner surface of SWNTs, we find that less than 5% of the reagent is successfully immobilized in the NT-A, NT-M, and NT-T materials.

4. CONCLUSIONS We have developed a general method for functionalizing the inner surface of single-walled aluminosilicate nanotube materials with organic reagents (as illustrated in Figure 12) and presented the rst unambiguous and comprehensive characterization to reveal the occurrence, extent, and structural details of the innersurface functionalization in such modied SWNTs. Furthermore, a comprehensive investigation of the resulting solids using nitrogen physisorption, powder X-ray diraction (XRD), transmission electron microscopy (TEM), thermogravimetric analysis (TGA), 29Si and 13C solid-state NMR, and water adsorption provides a detailed understanding of the porosity, structure, and surface chemistry of the functionalized nanotubes. In particular,
7683
dx.doi.org/10.1021/jp2010919 |J. Phys. Chem. C 2011, 115, 76767685

The Journal of Physical Chemistry C SWNTs modied with three reagents preserve their nanotube structure, and the variation in size and type of reagents allows for the control of the pore volume of the SWNT. We have also demonstrated that dierent types of organic groups, including alkyl, alkoxy, or carbonyl groups, can be immobilized at the SWNT inner surface. Water adsorption in the SWNTs is interpreted by the BET model to elucidate the adsorption mechanism in both bare and modied SWNTs. As an example of the potential applications, we show that the functionalization approach enables us to control the surface hydrophilicity as well as the water uptake of the SWNT. Finally, estimates of fractional surface silanol substitution for the three modied SWNTs are achieved via dierent characterization techniques, and consistent results are obtained. The present study provides a clear basis for addressing the challenging problem of adding organic functionalities to the interiors of SWNT materials and thereby greatly expands their potential applications. For example, by introducing appropriate functional groups, the SWNT can become an excellent candidate for size- and shape-selective catalytic reactions, sensing, molecular encapsulation, and molecular separations.

ARTICLE

ASSOCIATED CONTENT
S b

Supporting Information. Supporting data on the organic-modied nanotube materials, including XRD patterns, TGA traces, NMR spectra, details of nitrogen physisorption data analysis, and details regarding the estimates of the size of the immobilized groups. This material is available free of charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION
Corresponding Author

*E-mail: sankar.nair@chbe.gatech.edu; cjones@chbe.gatech.edu.

ACKNOWLEDGMENT This work was supported by ConocoPhillips Company. The authors also acknowledge Dr. J. Leisen (Georgia Tech), Dr. K. C. McCarley (ConocoPhillips), and Prof. M. Tsapatsis (U. Minnesota) for useful discussions. REFERENCES
(1) Iijima, S. Nature 1991, 354, 5658. (2) Popov, V. N. Mater. Sci. Eng. R-Rep. 2004, 43, 61102. (3) Ismail, A. F.; Goh, P. S.; Sanip, S. M.; Aziz, M. Sep. Purif. Technol. 2009, 70, 1226. (4) Allen, B. L.; Kichambare, P. D.; Star, A. Adv. Mater. 2007, 19, 14391451. (5) Wang, J.; Chen, Y.; Blau, W. J. J. Mater. Chem. 2009, 19, 74257443. (6) Zhang, H.; Cao, G. P.; Yang, Y. S. Energy Environ. Sci. 2009, 2, 932943. (7) Dai, H. J. Acc. Chem. Res. 2002, 35, 10351044. (8) Dai, H. J.; Javey, A.; Pop, E.; Mann, D.; Kim, W.; Lu, Y. R. NANO 2006, 1, 113. (9) Singh, S.; Kruse, P. Int. J. Nanotechnol. 2008, 5, 900929. (10) Hirsch, A.; Vostrowsky, O. In Functional Molecular Nanostructures; Springer: Berlin, 2005; Vol. 245, p 193-237. (11) Ahir, S. V.; Huang, Y. Y.; Terentjev, E. M. Polymer 2008, 49, 38413854. (12) Bokobza, L. Polymer 2007, 48, 49074920.

(13) Moniruzzaman, M.; Winey, K. I. Macromolecules 2006, 39, 51945205. (14) Grossiord, N.; Loos, J.; Regev, O.; Koning, C. E. Chem. Mater. 2006, 18, 10891099. (15) Kameta, N.; Masuda, M.; Minamikawa, H.; Goutev, N. V.; Rim, J. A.; Jung, J. H.; Shimizu, T. Adv. Mater. 2005, 17, 27322376. (16) Li, J.; Moskovits, M.; Haslett, T. L. Chem. Mater. 1998, 10, 19631967. (17) Abbaslou, R. M. M.; Tavassoli, A.; Soltan, J.; Dalai, A. K. Appl. Catal., A 2009, 367, 4752. (18) Kawasaki, S.; Matsuoka, Y.; Yokomae, T.; Nojima, Y.; Okino, F.; Touhara, H.; Kataura, H. Carbon 2005, 43, 3745. (19) Kawasaki, S.; Hara, T.; Yokomae, T.; Okino, F.; Touhara, H.; Kataura, H.; Watanuki, T.; Ohishi, Y. Chem. Phys. Lett. 2006, 418, 260263. (20) Nishino, T.; Ito, T.; Umezawa, Y. Anal. Chem. 2002, 74, 42754278. (21) Lin, T.; Bajpai, V.; Ji, T.; Dai, L. M. Aust. J. Chem. 2003, 56, 635651. (22) Kyotani, T.; Nakazaki, S.; Xu, W. H.; Tomita, A. Carbon 2001, 39, 782785. (23) Tenne, R.; Seifert, G. Annu. Rev. Mater. Res. 2009, 39, 387413. (24) Rao, C. N. R.; Govindaraj, A. Adv. Mater. 2009, 21, 42084233. (25) Son, S. J.; Bai, X.; Lee, S. B. Drug Discovery Today 2007, 12, 650656. (26) Bar-Sadan, M.; Kaplan-Ashiri, I.; Tenne, R. Eur. Phys. J. Spec. Top. 2007, 149, 71101. (27) Hu, S.; Ling, X.; Lan, T.; Wang, X. Chem.Eur. J. 16, 18891896. (28) Farmer, V. C.; Fraser, A. R.; Tait, J. M. J. Chem. Soc., Chem. Commun. 1977, 462463. (29) Vandergaast, S. J.; Wada, K.; Wada, S. I.; Kakuto, Y. Clays Clay Miner. 1985, 33, 237243. (30) Tani, M.; Liu, C.; Huang, P. M. Geoderma 2004, 118, 209220. (31) Mukherjee, S.; Bartlow, V. A.; Nair, S. Chem. Mater. 2005, 17, 49004909. (32) Bursill, L. A.; Peng, J. L.; Bourgeois, L. N. Philos. Mag. A 2000, 80, 105117. (33) Konduri, S.; Mukherjee, S.; Nair, S. Phys. Rev. B 2006, 74, art. # 033401. (34) Guimaraes, L.; Enyashin, A. N.; Frenzel, J.; Heine, T.; Duarte, H. A.; Seifert, G. ACS Nano 2007, 1, 362368. (35) Zhao, M. W.; Xia, Y. Y.; Mei, L. M. J. Phys. Chem. C 2009, 113, 1483414837. (36) Barron, P. F.; Wilson, M. A.; Campbell, A. S.; Frost, R. L. Nature 1982, 299, 616618. (37) Goodman, B. A.; Russell, J. D.; Montez, B.; Oldeld, E.; Kirkpatrick, R. J. Phys. Chem. Miner. 1985, 12, 342346. (38) Mackenzie, K. J. D.; Bowden, M. E.; Brown, I. W. M.; Meinhold, R. H. Clays Clay Miner. 1989, 37, 317324. (39) Ildefonse, P.; Kirkpatrick, R. J.; Montez, B.; Calas, G.; Flank, A. M.; Lagarde, P. Clays Clay Miner. 1994, 42, 276287. (40) Kang, D.-Y.; Zang, J.; Wright, E. R.; McCanna, A. L.; Jones, C. W.; Nair, S. ACS Nano 2010, 4, 48974907. (41) Gustafsson, J. P. Clays Clay Miner. 2001, 49, 7380. (42) Theng, B. K. G.; Russell, M.; Churchman, G. J.; Partt, R. L. Clays Clay Miner. 1982, 30, 143149. (43) Bottero, H.; Bonelli, B.; Ashbrook, S. E.; Wright, P. A.; Zhou, W.; Tagliabue, M.; Armandi, M; Garrone, E. Phys. Chem. Chem. Phys. 2011, 13, 744750. (44) Kajiwara, K.; Donkai, N.; Fujiyoshi, Y.; Inagaki, H. Makromol. Chem., Macromol. Chem. Phys. 1986, 187, 28952907. (45) Farmer, V. C.; Adams, M. J.; Fraser, A. R.; Palmieri, F. Clay Miner. 1983, 18, 459472. (46) Abidin, Z.; Matsue, N.; Henmi, T. J. Comput. Aided Mater. Des. 2007, 14, 518. (47) Yang, H. X.; Wang, C.; Su, Z. H. Chem. Mater. 2008, 20, 44844488. (48) Su, C.; Harsh, J. B. Geochim. Cosmochim. Acta 1994, 58, 16671677. (49) Su, C.; Harsh, J. B. Geochim. Cosmochim. Acta 1996, 60, 42754277.
7684
dx.doi.org/10.1021/jp2010919 |J. Phys. Chem. C 2011, 115, 76767685

The Journal of Physical Chemistry C

(50) Hu, J.; Kamali Kannangara, G. S.; Wilson, M. A.; Reddy, N. J. Non-Cryst. Solids 2004, 347, 224230. (51) Kupiec, K.; Konieczka, P.; Namiesnik, J. Crit. Rev. Anal. Chem. 2009, 39, 6069. (52) McMorn, P.; Hutchings, G. J. Chem. Soc. Rev. 2004, 33, 108122. (53) De Vos, D. E.; Dams, M.; Sels, B. F.; Jacobs, P. A. Chem. Rev. 2002, 102, 36153640. (54) Sayari, A.; Hamoudi, S. Chem. Mater. 2001, 13, 31513168. (55) Cheng, C. H.; Bae, T. H.; McCool, B. A.; Chance, R. R.; Nair, S.; Jones, C. W. J. Phys. Chem. C 2008, 112, 35433551. (56) Hicks, J. C.; Drese, J. H.; Fauth, D. J.; Gray, M. L.; Qi, G. G.; Jones, C. W. J. Am. Chem. Soc. 2008, 130, 29022903. (57) Yu, K. Q.; Sommer, W.; Weck, M.; Jones, C. W. J. Catal. 2004, 226, 101110. (58) Borrego, T.; Andrade, M.; Pinto, M. L.; Silva, A. R.; Carvalho, A. P.; Rocha, J.; Freire, C.; Pires, J. J. Colloid Interface Sci. 2010, 344, 603610. (59) Johnson, L. M.; Pinnavaia, T. J. Langmuir 1990, 6, 307311. (60) Johnson, L. M.; Pinnavaia, T. J. Langmuir 1991, 7, 26362641. (61) Yamamoto, K.; Otsuka, H.; Wada, S.; Takahara, A. Chem. Lett. 2001, 11621163. (62) Yamamoto, K.; Otsuka, H.; Wada, S. I.; Sohn, D.; Takahara, A. Polymer 2005, 46, 1238612392. (63) Ackerman, W. C.; Hua, D. W.; Kim, Y. W.; Huling, J. C.; Smith, D. M. Charact. Porous Solids Iii 1994, 87, 735744. (64) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.; Rouquerol, J.; Siemieniewska, T. Pure Appl. Chem. 1985, 57, 603619. (65) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309319. (66) Lippens, B. C.; Deboer, J. H. J. Catal. 1965, 4, 319323. (67) Rouquerol, J.; Llewellyn, P.; Rouquerol, F.; Llewellyn, P.L.; F. R.-R., J. R.; Seaton, N. In Stud. Surf. Sci. Catal.; Elsevier: New York, 2007; Vol. 160, pp 4956. (68) Scherdel, C.; Reichenauer, G.; Wiener, M. Microporous Mesoporous Mater. 2010, 132, 572575. (69) Cape, J. A.; Kibby, C. L. J. Colloid Interface Sci. 1990, 138, 515520. (70) Kakei, K.; Ozeki, S.; Suzuki, T.; Kaneko, K. J. Chem. Soc., Faraday Trans. 1990, 86, 371376. (71) Harkins, W. D.; Jura, G. J. Am. Chem. Soc. 1944, 66, 13661373. (72) Deboer, J. H.; Lippens, B. C.; Linsen, B. G.; Broekhof, Jc; Vandenhe, A; Osinga, T. J. J. Colloid Interface Sci. 1966, 21, 405414. (73) Thess, A.; Lee, R.; Nikolaev, P.; Dai, H. J.; Petit, P.; Robert, J.; Xu, C. H.; Lee, Y. H.; Kim, S. G.; Rinzler, A. G.; Colbert, D. T.; Scuseria, G. E.; Tomanek, D.; Fischer, J. E.; Smalley, R. E. Science 1996, 273, 483487. (74) Rols, S.; Almairac, R.; Henrard, L.; Anglaret, E.; Sauvajol, J. L. Eur. Phys. J. B 1999, 10, 263270. (75) Bendiab, N.; Almairac, R.; Rols, S.; Aznar, R.; Sauvajol, J. L.; Mirebeau, I. Phys. Rev. B 2004, 69, 195415. (76) Cambedouzou, J.; Pichot, V.; Rols, S.; Launois, P.; Petit, P.; Klement, R.; Kataura, H.; Almairac, R. Eur. Phys. J. B 2004, 42, 3145. (77) Bachmann, S.; Wang, H.; Albert, K.; Partch, R. J. Colloid Interface Sci. 2007, 309, 169175. (78) Gorlov, Y. I.; Nesterenko, A. M.; Chuiko, A. A. Colloids Surf., A 1996, 106, 8388. (79) El Rassy, H.; Pierre, A. C. J. Non-Cryst. Solids 2005, 351, 16031610. (80) Hair, M. L.; Hertl, W. J. Phys. Chem. 1969, 73, 23722378. (81) Low, M. J. D.; Severdia, A. G.; Chan, J. J. Colloid Interface Sci. 1982, 86, 111118. (82) Sindorf, D. W.; Maciel, G. E. J. Am. Chem. Soc. 1981, 103, 42634265. (83) Sindorf, D. W.; Maciel, G. E. J. Am. Chem. Soc. 1983, 105, 37673776.

ARTICLE

(84) Ek, S.; Iiskola, E. I.; Niinisto, L.; Vaittinen, J.; Pakkanen, T. T.; Root, A. J. Phys. Chem. B 2004, 108, 1145411463. (85) Fyfe, C. A.; Feng, Y.; Grondey, H.; Kokotailo, G. T.; Gies, H. Chem. Rev. 1991, 91, 15251543. (86) Merzbacher, C. I.; McGrath, K. J.; Higby, P. L. J. Non-Cryst. Solids 1991, 136, 249259. (87) Lippmaa, E.; Magi, M.; Samoson, A.; Tarmak, M.; Engelhardt, G. J. Am. Chem. Soc. 1981, 103, 49924996. (88) Field, L. D.; Sternhell, S. Analytical NMR; Wiley: Chichester, 1989. (89) Clark, J. C.; Barnes, C. E. Chem. Mater. 2007, 19, 32123218. (90) Shimizu, H.; Watanabe, T.; Henmi, T.; Masuda, A.; Saito, H. Geochem. J. 1988, 22, 2331. (91) Wilson, M. A.; Wada, K.; Wada, S. I.; Kakuto, Y. Clay Miner 1988, 23, 175190. (92) Zang, J.; Chempath, S.; Konduri, S.; Nair, S.; Sholl, D. S. J. Phys. Chem. Lett. 2010, 1, 12351240. (93) Konduri, S.; Tong, H. M.; Chempath, S.; Nair, S. J. Phys. Chem. C 2008, 112, 1536715374. (94) Vannice, M. A. Kinetics of catalytic reactions; Springer: New York, 2005. (95) Cejka, J.; van Bekkum, H. Zeolites and ordered mesoporous materials: progress and prospects; the 1st FEZA School on Zeolites, Prague, Czech Republic, August 2021, 2005; 1st ed.; Elsevier: Amsterdam, 2005. (96) Mooney, R. W.; Keenan, A. G.; Wood, L. A. J. Am. Chem. Soc. 1952, 74, 13671371. (97) Mooney, R. W.; Keenan, A. G.; Wood, L. A. J. Am. Chem. Soc. 1952, 74, 13711374. (98) Stromme, M.; Mihranyan, A.; Ek, R.; Niklasson, G. A. J. Phys. Chem. B 2003, 107, 1437814382. (99) Tomczak, E.; Kaminski, W. Drying Technol. 2009, 27, 12861291.

7685

dx.doi.org/10.1021/jp2010919 |J. Phys. Chem. C 2011, 115, 76767685