Morphological characterization of mesoporous silicate–carbon nanocomposites

Morphological characterization of mesoporous silicate–carbon nanocomposites

Microporous and Mesoporous Materials 80 (2005) 85–94 www.elsevier.com/locate/micromeso Morphological characterization of mesoporous silicate–carbon n...

416KB Sizes 0 Downloads 47 Views

Microporous and Mesoporous Materials 80 (2005) 85–94 www.elsevier.com/locate/micromeso

Morphological characterization of mesoporous silicate–carbon nanocomposites ´ kos Kukovecz, Tı´mea Kanyo´, Zolta´n Ko´nya *, Imre Kiricsi A Department of Applied and Environmental Chemistry, University of Szeged, H-6720 Szeged, Rerrich Be´la te´r 1, Hungary Received 24 September 2004; received in revised form 23 November 2004; accepted 29 November 2004 Available online 20 January 2005

Abstract Mesoporous MCM-41-carbon nanocomposites were prepared in several Si:C ratios from amorphous active carbon, graphite and multi-wall carbon nanotube carbon sources. A complete morphological characterization of the samples was performed using XRD, TEM, SEM, N2 adsorption and surface hydrophilicity measurements. Specific surface area, pore size distribution, surface fractal dimension and the dimension of capillary condensation were calculated and the adsorption excess isotherms of the ethanol–cyclohexane system were measured. Our results show that the three carbon types give rise to rather different composite morphological features. On this basis the fine tuning of carbon–silicate composite properties by design appears feasible. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Mesoporous silica; Silicate–carbon nanocomposites; Morphological characterization

1. Introduction The preparation of various inorganic composite materials has been in the focus of materials science since decades. In most cases the new substances show peculiar, surprisingly novel properties. Special interest has been devoted recently to carbon based composites. This effort has been partially initiated by the paper of Schmitt [1] who summarized a 10 years development of carbon molecular sieves. The achievement established was found to be unsatisfactory. At that time researchers could not transplant the molecular sieve properties of zeolites to any synthetic micro- or mesoporous carbon derivative. Some years later Korean researchers observed shape selective catalytic effect of palladium supported on activated carbon fibers [2]. Very recently Schlo¨gl and co-workers reported a novel application of graphite and carbon filaments as selective catalysts in oxidative dehydrogena-

*

Corresponding author. Tel.: +36 62 544 620; fax: +36 62 544 619. E-mail address: [email protected] (Z. Ko´nya).

1387-1811/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2004.11.025

tion of ethylbenzene [3]. The authors attributed the catalytic activity to the OH groups on the carbon surface that are continuously regenerated by adsorption of oxygen. A novel alternative was the preparation of composite materials from carbon and oxides. Moreno-Castilla et al. [4] prepared carbon–alumina and –silica based sol–gel composites. They showed that the silica containing composites had much more developed micro- and mesoporosity than the alumina containing ones. Lukens and Stucky reported on the synthesis of mesoporous carbon foams [5]. They used microemulsion-polymerized polystyrene micro-spheres as templates for producing composite materials using resorcinol–formaldehyde polymers. Upon heat treatment these spheres were transformed into mesoporous carbon. Another template application was reported in the recent work of Pang et al. who obtained nanoporous carbon by utilizing a sol–gel silica–carbon composite intermedier [6]. Another sol–gel approach resulting in silica/C composites was described by Aguado-Serrano et al. [7]. Classical molecular sieves (zeolites) have also been synthesized with carbon to yield composite samples [8] using the conventional

86

A´. Kukovecz et al. / Microporous and Mesoporous Materials 80 (2005) 85–94

synthesis path for ZSM-5. Zeolite–carbon membranes have been studied by Smith et al. [9] and tested in separation processes [10]. Sterte and co-workers reported the preparation of continuous silicate films on inorganic fibers including carbon type material [11]. Carbon–oxide composites are also used in catalysis [12]. Carbon nanotube–silica composites prepared by the sol–gel method showed a variety of capacitance and electron transfer rates depending on the type and ratio of silane precursor and carbon nanotubes [13]. Singlewall carbon nanotubes embedded into silica matrix showed photochemical activity [14] opening the way towards novel heterogeneous photocatalysts. In a paper from the Topsøe laboratory the synthesis of carbon nanotube–zeolite composite is reported [15]. Their aim was to obtain zeolite samples with mesopores to accelerate the diffusion rates of reactants and products. The presence of mesopores in microporous zeolite structures has been exhaustively studied as well [16]. It was established that more acid sites are accessible for reactant molecules in mordenites possessing mesopores and the diffusion limitation has also been decreased. The aim of this work was to undertake the morphological characterization of various carbon–mesoporous silicate nanocomposites with different composition and variable hydrophobic–hydrophilic properties. The latter are especially important for the successful design of novel high performance adsorbents [17] and may also improve the selectivity of heterogeneous catalytic reactions through the ability to tailor metal-support interactions. We report results obtained in the study of graphite-, active carbon- and multiwall carbon nanotube-MCM-41 mesoporous silicate composites. Amorphous active carbon exhibits a high specific surface area (674 m2/g) and broad pore size distribution featuring a high micropore content. Graphite has a small specific surface area (73 m2/g) and layered structure in which the graphene sheets are in regular arrangement, 0.35 nm apart from each other, building up a crystalline phase. Multiwall carbon nanotubes (MWCNTs) possess an intermediate specific surface area (237 m2/g) and exhibit a structure in which the rolled-up graphene sheets are in the same distance from each other as they are in the graphite. The uniqueness of the present study is that results obtained by characterization methods stretching over several orders of magnitude in lateral resolution (normal TEM level–nanoscale surface level–molecular level) are discussed in a coherent manner.

2. Experimental 2.1. Synthesis of the starting materials Active carbon and graphite were purchased from Aldrich and used as received.

The MWCNT sample was synthesized in our laboratory using the catalytic chemical vapor deposition (CCVD) procedure [18]. Briefly summarizing, Co–Fe supported on alumina was prepared with 2.5–2.5% loading of cobalt- and iron acetate. The catalyst was placed in a quartz boat into the reactor tube preheated to 970 K. After purging the system with pure nitrogen at 373 K the gas stream was switched to acetylene–nitrogen (1:10) mixture. The flow rate of the acetylene was 15 ml min 1. After 1 h the acetylene stream was closed and the system was cooled to ambient temperature in nitrogen flow. The boat was removed and the carbon deposit was purified by a two step procedure. The sample was first leached in conc. NaOH solution to dissolve the alumina support and then treated in conc. HCl solution to remove the metallic catalyst particles. The last step of nanotube preparation was oxidation in 0.1 M H2SO4 solution by the stepwise addition of a total of 0.1 M equivalent (relative to the amount of carbon) of KMnO4 at room temperature. This is a standard procedure in our laboratory which reduces the amount of amorphous carbon contaminants while leaving carbon nanotubes intact. After washing and drying the sample was ready for the composite synthesis. Synthesis of MCM-41 mesoporous silicate was carried out using the procedure described previously [19,20]. Briefly, 30.3 g cetyltrimethyl ammonium bromide (CTMABr) template was added to 180.2 g dist. water under stirring at 313 K. 37.03 g sodium silicate solution was added dropwise to the solution at room temperature, followed by 10.14 g 10 wt.-% aqueous sulphuric acid solution. The molar ratio of the synthesis gel components was 1SiO2:0.5CTMABr:63H2O. After 150 min of stirring the pH was adjusted to 10.0 by drops of 50 wt.-% aqueous sulphuric acid solution. The crystallization was performed in Teflon lined autoclaves at 353 K for 40 h. After washing the sample with dist. water and drying in air at 393 K the sample was heat treated at 813 K in order to burn off the template. 2.2. Synthesis of the composites Prior to the composite preparation, active carbon and graphite were oxidized in 1 M HNO3 solution at 40 °C for 3 h in order to increase the amount of surface functional groups like –OH, –COOH, –[email protected] [21]. Presence of these polar groups [22,23] is considered advantageous for the bond strength between the hydrophilic and hydrophobic partners of composites [24,25]. The carbon component was mixed into the prepared reactive MCM41 precursor solution (see above) before the pH adjustment step under vigorous stirring for 30 min. Various Si–MCM-41–carbon composites with MCM-41:carbon mass ratios of 10:1, 5:1, 2.5:1, 1:1, 1:2.5, 1:5, 1:10 and 1:100 were synthesized (mass ratios are used instead of molar ones because the unit ‘‘1 mol’’ is ill-defined

A´. Kukovecz et al. / Microporous and Mesoporous Materials 80 (2005) 85–94

for the studied carbon species). The gelation process was monitored by regular TG/DTA and FT-IR measurements. The most customary MCM-41 template removal procedure, that is, heat treatment in air above 813 K could not be applied to the composites since under these conditions the carbon content would have been removed as well. Therefore, two alternative methods were tested. Either we treated the composites at 813 K in nitrogen stream for 6 h or performed solvent extraction with ethanol containing 2 v/v.-% acetic acid at room temperature for 1 h. The efficiency of the template removal was monitored by observing the intensity of the IR bands in the region of mC–H vibrations. We found that template free material can be achieved by both methods. Nevertheless, in the present contribution only results measured on thermally treated samples are reported since they were generally found to be of better crystallinity (i.e., they contained less amorphous contaminants). 2.3. Characterization methods The progress of the synthesis reactions was monitored by IR spectroscopy and thermal analysis. Infrared spectra were recorded on a Mattson Genesis 1 FT-IR spectrometer in 0.1 wt.-% KBr pellets. Derivatographic measurements were performed using a DerivatographQ Paulik–Paulik type equipment. Samples were placed in a ceramic sample holder and the TG, DTG, DTA profiles were recorded using 10 K/min heating ramp from room temperature to 1270 K in air. Transmission electron microscope (TEM) images were recorded on a Phillips CM10 instrument. Diluted suspensions of composite samples in methanol were prepared and drop-dried on holey carbon coated copper TEM grids. Scanning electron microscopy (SEM) was performed on a Hitachi S 2400 instrument on goldcoated samples. Powder XRD profiles were taken with a DRON 3 diffractometer operating with Cu Ka radiation in the 0.5– 10 and 20–30 Bragg angle ranges.

87

Nitrogen adsorption isotherms were measured at 77 K with a QuantaChrome Nova 2000 surface area analyzer. Samples were outgassed at 423 K for 1 h to remove adsorbed contaminants prior to the measurement. The correctness of using N2 adsorption for the study of porous carbonaceous systems was validated by the theoretical work of Cascarini de Torre and Bottani [26]. The specific surface area (AS) was calculated using the multipoint BET method [27] on six points of the adsorption isotherm near monolayer coverage. Pore size distribution (PSD) curves were calculated from the adsorption branch of the isotherms using the Barrett– Joyner–Halenda (BJH) method [28]. The adsorption branch was favored over the desorption branch in order to avoid the tensile strength effect (TSE) artifact which very often complicates the PSD determination in mesoporous systems [29]. Surface fractal dimension (DS) was calculated using the Frenkel–Halsey–Hill (FHH) method [30] from adsorption data near monolayer coverage. Additionally, the Neimark–Kiselev method [31] was also utilized to calculate a fractal dimension value from the capillary condensation region of the adsorption isotherm (DCC). For a more detailed theoretical treatment of the interpretation of N2 adsorption isotherms we refer to the work of Ehrburger–Dolle who compared several methods for calculating the surface fractal dimension of silica and various carbonaceous adsorbents [32] and to a recent review of the fractal geometry concept in physical absorption on solids by Terzyk et al. [33]. The standard error of the N2 adsorption isotherms was approximately 1% for all measurements reported herein. Therefore, in Tables 1–3. the AS values are accurate to zero significant decimal digits while the DS, DCC and dpore values are accurate to two significant decimal digits. Adsorption excess isotherms of ethanol–cyclohexane mixtures were studied in a static solvent adsorption system at 293 K and the adsorption capacity (Schay–Nagy extrapolation method) was used for the calculation of the equivalent specific surface area, which in turn was used to derive the ratio between the hydrophilic and

Table 1 Selected morphological characteristics of the activated MCM-41–carbon composites Sample

AS (m2/g), BET surface area

DS, Surface fractal dimension

DCC, Capillary condensation dimension

dpore (nm), BJH pore radius

Pure MCM-41 MCM41–C (10:1) MCM41–C (5:1) MCM41–C (2.5:1) MCM41–C (1:2.5) MCM41–C (1:5) MCM41–C (1:10) Pure active carbon

1030 884 826 759 561 618 622 674

2.36 2.46 2.56 2.57 2.68 2.73 2.77 2.81

13.21 12.31 11.89 10.92 5.30 3.72 3.26 3.57

2.82 2.63 2.79 2.64 2.62 2.50 2.45

a

Microporous sample, out of the range accessible by BJH method.

a

A´. Kukovecz et al. / Microporous and Mesoporous Materials 80 (2005) 85–94

88

Table 2 Selected morphological characteristics of the MCM-41–graphite composites Sample Pure MCM-41 MCM41–graphite MCM41–graphite MCM41–graphite MCM41–graphite MCM41–graphite MCM41–graphite Pure graphite a

(10:1) (5:1) (2.5:1) (1:2.5) (1:5) (1:10)

AS (m2/g), BET surface area

DS, Surface fractal dimension

DCC, Capillary condensation dimension

dpore (nm), BJH pore radius

1030 957 862 671 272 179 71 65

2.36 2.52 2.55 2.53 2.54 2.54 2.54 2.80

13.21 12.89 14.48 14.23 15.44 11.96 10.42

2.82 2.67 2.86 2.58 2.77 2.79 2.52

a

a

Because of the lamellar structure of graphite, capillary condensation and BJH-accessible mesoporosity do not apply to pure graphite.

Table 3 Selected morphological characteristics of the MCM-41–MWCNT composites Sample Pure MCM-41 MCM41–MWCNT MCM41–MWCNT MCM41–MWCNT MCM41–MWCNT Pure MWCNT

(100:1) (10:1) (1:1) (1:10)

AS (m2/g), BET surface area

DS, Surface fractal dimension

DCC, Capillary condensation dimension

dpore (nm), BJH pore radius

1030 923 909 433 318 237

2.36 2.45 2.47 2.51 2.50 2.53

13.21 12.37 13.36 7.05 3.33 3.74

2.82 2.79 2.79 2.82 2.50 2.32

the hydrophobic surface portions. The procedure was described in detail previously [34,35].

3. Results and discussion 3.1. Large scale morphology Representative TEM images of the individual composite elements are presented in Fig. 1. Amorphous carbon (part a) consists of particles about 5–50 nm in diameter that do not exhibit any characteristic morphological features. On the other hand, the large (over 1 lm diagonally) layers of graphite (part b) and the high aspect ratio tubular structure (12–16 nm diameter, over 400 nm in length) of nanotubes (part c) are clearly identifiable. Part d shows the ordered hexagonal mesoporous channel system (apparent diameter 2.6 nm) of Si–MCM-41. In Fig. 2 selected TEM images of the prepared composites are depicted. The pictures are dominated by features originating from MCM-41. In the case of the MCM-41–active carbon materials both at 1:1 (part a) and also at 1:10 (part b) the characteristic features of the silicate material are clearly seen. The amorphous carbon phase is completely covered by MCM-41. Parts c and d show the 1:1 and 1:10 MCM-41–graphite systems, respectively. By increasing the graphite content it was possible to switch the dominant morphology from a tubular one into a layered one. Nevertheless, even at low silicate content the hexagonal arrays could be identified in between the graphene sheets. Satisfactory nano-

tube coverage by the silicate part could not be established in the MCM-41–MWCNT 1:1 material (part e), presumably because of MCM-41 segregation. At smaller silicate contents (1:10 composite shown in part f) the coverage became better. The thickness of the MCM-41 layer enveloping the carbon nanotubes varied between 12 and 30 nm. SEM images of the amorphous carbon and graphite containing composites lacked characteristic features identifiable by our microscope. On the other hand, carbon nanotubes were quite visible even at low concentration as depicted for the MCM-41–MWCNT 10:1 in Fig. 3a. Increasing the nanotube contents ratio to 1:10 resulted in the material shown in Fig. 3b. MWCNTs cover fully the surface of silicate crystallites. The samples were studied by XRD to characterize their phases. As far as the pure single components are concerned (not shown), graphite and MCM-41 are highly ordered materials with sharp and intense reflections at 2H = 26.5° and a group of peaks at 2H = 2.5°, 4.5° and 5°, respectively. Active carbon is an essentially amorphous matter, exhibiting only a small reflection arising from its graphitic domains. Multiwall carbon nanotubes show a broad reflection at around 2H = 26°, close to the characteristic value of graphite. In Fig. 4 XRD profiles of the synthesized active carbon (a), graphite (b) and nanotube (c) based MCM-41 composites are presented. In agreement with the TEM results, the MCM-41 contribution is pronounced even at low silicate loadings for all three carbon types. The carbon signal intensity changed according to the compo-

A´. Kukovecz et al. / Microporous and Mesoporous Materials 80 (2005) 85–94

89

Fig. 1. TEM images of the pure materials used for building the composites: (a) active carbon, (b) graphite, (c) multiwall carbon nanotubes, (d) Si– MCM-41.

nent concentrations, even though this is evident only for the graphite-based composites due to the less intense reflections of both the MWCNT and amorphous carbon. Neither peak position shifts nor relative intensity changes were observed in the XRD profiles of the composites. This indicates that the silicate and the composite parts preserved their original properties and did not form any new XRD-observable phase. Such conservative behavior may prove advantageous in the design phase of fine tuning composite performance for industrial applications. 3.2. Nanoscale morphology The detailed analysis of N2 adsorption isotherms can be used to collect information about the nanoscale morphology of mesoporous carbonaceous samples [36]. For each material the following quantities were determined: (i) the specific surface area (AS), (ii) the surface fractal dimension (DS), (iii) the dimension of capillary condensation (DCC) and (iv) the position of the maximum of the pore size distribution curve (dpore). All calculated parameters are given in Tables 1–3 for the active carbon, graphite and nanotube based composites, respectively. Graphical representation of the data is provided in Fig. 5, where the height of the columns corresponds to BET area, squares denote DS values (left y) and triangles denote DCC values (right y axis).

Let us first examine the descriptors of the MCM-41– active carbon system (Table 1, Fig. 5a). As long as there is silicate in these composites, the maximum of the pore size distribution curve is determined by the MCM-41. The observed 0.35 nm downshift of dpore can be tentatively assigned to the partial blocking of the MCM-41 pores by the increased carbon content. The DS = 2.36 value obtained for pure MCM-41 is in excellent agreement with data reported in the literature for MCM-41 surface fractal dimension measured by N2 adsorption [37,38]. The surface fractal dimension of the other samples increases linearly from 2.36 (pure MCM-41) to 2.8 (pure active carbon) and appears to be a weighted linear combination of the values of the individual components. Therefore, composites with comparable silicate and carbon loadings show surface fractal behavior, while those dominated by either component do not classify as surface fractals. Both MCM-41 and amorphous active carbon are high surface area materials, therefore, all composites possess high AS values. It is interesting to note that the specific surface area vs. composition curve runs through a minimum at 1:2.5 silicate to carbon ratio. This somewhat unexpected finding can be correlated with the changes observed in the dimension of the capillary condensation. The phenomenological interpretation of this quantity––as discussed recently by Pomonis and co-workers [39]––is that higher DCC values correspond to more uniform pore size distributions.

90

A´. Kukovecz et al. / Microporous and Mesoporous Materials 80 (2005) 85–94

Fig. 2. TEM images of MCM-41–carbon composites: (a) MCM-41–active carbon 1:1, (b) MCM-41–active carbon 1:10, (c) MCM-41–graphite 1:1, (d) MCM-41–graphite 1:10, (e) MCM-41–MWCNT 1:1, (f) MCM-41–MWCNT 1:10.

Indeed, MCM-41 rich samples down to 2.5:1 silicate to carbon ratio have DCC > 10, whereas the amorphous active carbon with its mixed mesoporous–microporous channel system shows DCC < 4. The correlation with the anomalous BET area change lies within the sharp drop in the DCC curve at the 1:2.5 MCM-41–carbon composite. This indicates that the pore system experiences major changes as the composite is switched from a silicate based one into an active carbon based one. We assume that the pore system becomes less welldefined and mostly determined by local fluctuation effects in this transition regime which results in a smaller specific surface area than expected. The behavior of the MCM-41–graphite system (Table 2, Fig. 5b) differs considerably from the active carbon based one. The drop in dpore is less pronounced because the graphite sheets and the silicate channels are sterically hindered from such intimate mixing as that responsible for the narrowing of the MCM-41 pores in the previous material. Another indicator of the less thorough mixing is the regular change of the BET surface area which ap-

pears to be a linear combination of the surfaces of the individual components. On the other hand, both the surface fractal dimension and the dimension of the capillary condensation remain quasi constant over a large composition range from 10:1 to 1:10 silicate to graphite ratio. These composites show DS 2.53 and therefore, they classify as surface fractal systems. A possible explanation for the conservation of this property over such a wide range is that both MCM-41 crystals decorated with graphite facelets and graphite sheets holding MCM-41 islands can be self-similar. Since both MCM-41 and graphite possess very well-defined structures and their intermixing is restricted, it is rather self-explanatory that the DCC value of all composites was above 10, indicating a uniform, ordered pore system. A third type of nanoscale morphology can be deduced from the parameters of the MCM-41–MWCNT composites (Table 3, Fig. 5c). Similarities with the graphite based systems are: (i) dpore does not decrease with increasing carbon content, and (ii) the BET surface area appears to be a linear combination of the specific

A´. Kukovecz et al. / Microporous and Mesoporous Materials 80 (2005) 85–94

91

Fig. 3. Representative SEM images of the 10:1 (a) and 1:10 (b) MCM41–MWCNT composites.

surfaces of the individual components. Both observations can be traced back to the comparable sizes of nanotubes and MCM-41 channels which keep them from intimate mixing. The surface fractal dimension of the composites increases linearly with the increasing nanotube content, and practically all composites classify as surface fractals. Unlike in the case of graphite-based materials, the surface self-similarity does not seem to arise from the interaction of the two components, rather, it comes from the intrinsic surface fractality of multiwall carbon nanotubes as observed by several authors before [40]. Finally, the DCC vs. composition function experiences a sharp drop at 1:1 loading from DCC > 10 for silicate rich materials down to DCC < 4 for nanotube rich samples. This finding agrees well with the known broadness of the MWCNT pore size distribution curve, which in turn originates from the inability of the CCVD nanotube preparation method to produce monodisperse MWCNT samples. 3.3. Surface chemical properties Mixtures of ethanol and cyclohexane were adsorbed on the single components and on selected composites. We have recently shown [34,35] that this test monitors well the hydrophilic–hydrophobic character of the samples, or––in wider consideration––the surface chemistry of the composite. Pure MCM-41 is a hydrophilic mate-

Fig. 4. Characteristic XRD profiles of the synthesized active carbon (a), graphite (b) and nanotube (c) based MCM-41 composites. Profiles are identified with the MCM-41: carbon mass ratio of the corresponding composite.

rial, characterized by a steep EtOH adsorption excess isotherm running in the positive region up to high nsEtOH values. Fig. 6a shows the isotherms for MCM41, active carbon, and their 10:1 composite sample. The surface of active carbon possesses roughly equivalent amounts of hydrophilic and hydrophobic portions. By introducing a small amount of amphiphilic active carbon into the MCM-41 it was possible to shift the MCM-41 isotherm downwards along the y axis, and the composite material exhibited hydrophobic character not present in the original MCM-41. It seems reasonable to assume that by using a higher amount of active carbon, the degree of hydrophobicity could be fine tuned to some extent. The adsorption excess isotherms of the MCM41–graphite system (Fig. 6b) exhibit a more complex behavior. Pure graphite shows an amphiphilic isotherm indicating only minor adsorption capacity, most probably due to restrictions applying to adsorption between the carbon layers. The isotherm of the composite sample indicates that both hydrophilic and hydrophobic surface portions are present in the 10:1 MCM-41–graphite

92

A´. Kukovecz et al. / Microporous and Mesoporous Materials 80 (2005) 85–94

Fig. 5. N2 adsorption based morphological descriptors of the (a) MCM-41:active carbon, (b) MCM-41:graphite and (c) MCM41:MWCNT composites. In all three figures, the height of the columns corresponds to the specific surface area (exact value written onto the column), DS denotes surface fractal dimension (j, left y axis) and DCC denotes the dimension of capillary condensation (m, right y axis). The lines connecting the symbols are guides for the eye.

material. It is remarkable that the small amount of added graphite was able to donate considerable hydrophobicity to the system. A possible explanation for this phenomenon can be offered on the basis of the established nanoscale morphology (see Fig. 5b) of this composite. Contrary to the active carbon and the MWCNT based systems, the intermixing of the carbon and the silicate component of the MCM-41–graphite composites is restricted even at low graphite loadings. Therefore, the studied material contains graphitic ‘‘islands’’ (the presence of which could also be deduced from the surface fractal dimension calculations) obstructing the preferential adsorption of ethanol on the composite. This hindrance manifests as a hydrophobic portion in the adsorption excess isotherm at high ethanol concentrations.

Fig. 6. Adsorption excess isotherms measured in ethanol–cyclohexane mixture for active carbon (a), graphite (b) and nanotube (c) based MCM-41 composites. In all three figures, the isotherms of pure MCM41, pure carbon component and MCM-41:carbon = 10:1 composites are denoted by m, j and d, respectively. The lines connecting the symbols are guides for the eye.

Multiwall carbon nanotubes by themselves show a dominantly hydrophilic character (Fig. 6c) which can be explained by the oxidative pre-treatment they were subjected to. Albeit a small hydrophobic character originating from the graphitic walls can also be observed in the 0.7 < xEtOH < 1 region for MWCNTs, the excess isotherm of the 10:1 MCM-41–MWCNT composite reveals no hydrophobic part at all. This indicates that the small amount of nanotubes introduced into the composite was

A´. Kukovecz et al. / Microporous and Mesoporous Materials 80 (2005) 85–94 Table 4 Distribution of the hydrophilic–hydrophobic surface portion in selected samples Sample

H1 (%) Hydrophilic surface portion

H2 (%) Hydrophobic surface portion

Pure MCM-41 Pure active carbon Pure graphite Pure MWCNT MCM41–C (10:1) MCM41–graphite (10:1) MCM41–MWCNT (10:1)

85 39 23 62 72 39 72

15 61 77 38 28 61 28

not able to change the hydrophilic nature of the silicate matrix. It is interesting to note here that the oxidative treatment was noticeably more effective in adding hydrophilic character to the carbon surface in the case of nanotubes than in the case of graphite (Fig. 6b). This observation can be explained by considering the enhanced chemical reactivity of the MWCNT surface as compared to pure graphite [41]. The presented isotherms were also analyzed quantitatively using the Schay–Nagy extrapolation method to determine the total hydrophilic and hydrophobic surface portion of each composite discussed above. The obtained results (summarized in Table 4) are in good agreement with the qualitative analysis and provide numerical evidence for the composite nature of the samples.

4. Conclusions A huge research effort has been invested recently into the field of composite materials, particularly in nanocomposite matters. The change in the composition of a binary composite like the ones we discuss here may shift some substantial properties of the material. We investigated MCM-41–carbon composites in varied compositions utilizing three different carbon sources: amorphous active carbon, graphite and multiwall carbon nanotubes. Our purpose was to characterize the complete morphology of the composites from the micrometer dimension down to the surface chemistry level. Even though no new phases were detected by XRD, we have shown that by choosing the right carbon source and carbon to silicate ratio it is possible to tailor the fundamental properties of carbon-containing MCM41 composites. Morphological parameters we could change by design are: (i) macroscopic morphology, (ii) specific surface area, (iii) characteristic pore size, (iv) surface fractal dimension and (v) pore system homogeneity. In each case the observed variations could be explained on the basis of sample composition. Additionally, we showed that surface chemical properties––as monitored by the ethanol adsorption excess isotherm––can also be fine-tuned. Surveying the governing

93

principles of MCM-41–carbon composite surface chemistry work is in progress in our laboratory. The presented materials may find industrial applications as e.g. tunable selective adsorbents or novel heterogeneous catalyst supports.

Acknowledgments The authors thank the financial support of the Hungarian Ministry of Education (OTKA T037952, F038249, F046361). Z.K. and A.K. acknowledge the support of Ja´nos Bo´lyai and Zolta´n Magyary postdoctoral fellowships, respectively.

References [1] J.L. Schmitt, Carbon 29 (1991) 743. [2] H. Jin, S.-E. Park, J.M. Lee, S.K. Ryu, Carbon 34 (1996) 429. [3] G. Mestl, N.I. Maximova, N. Keller, V.V. Roddatis, R. Schlo¨gl, Angew. Chem., Int. Ed. 40 (2001) 2066. [4] C. Moreno-Castilla, F.J. Maldonado-Hodar, Phys. Chem. Chem. Phys. 2 (2000) 4818. [5] W.W. Lukens, G.D. Stucky, Chem. Mater. 14 (2002) 1665. [6] J. Pang, Q. Hu, Y. Wu, J.E. Hampsey, J. He, Y. Lu, Micropor. Mesopor. Mater. 74 (2004) 73. [7] J. Aguado-Serrano, M.L. Rojas-Cervantes, A.J. Lopez-Peinando, V. Gomez-Serrano, Micropor. Mesopor. Mater. 74 (2004) 111. [8] J. Garcia-Martinez, D. Cazorla-Amoros, A. Linares-Solano, Y.S. Lin, Micropor. Mesopor. Mater. 42 (2001) 255. [9] S.P.J. Smith, V.M. Linkov, R.D. Sanderson, L.F. Petrik, C.T. OÕConnor, K. Keiser, Micropor. Mater. 4 (1995) 385. [10] Y. Yan, M.E. Davis, G.R. Gavalas, J. Membr. Sci. 126 (1997) 53. [11] V. Valtchev, J. Hedlund, B.J. Schoeman, J. Sterte, S. Mintova, Micropor. Mesopor. Mater. 8 (1997) 93. [12] G.D. Yadav, J.J. Nair, Langmuir 16 (2000) 4072. [13] V.G. Gavalas, R. Andrews, D. Bhattacharyya, L.G. Bachas, Nano Lett. 1 (2001) 719. [14] M. Alvaro, P. Atienzar, J.L. Bourdelande, H. Garcia, Chem. Commun. (2002) 3004. [15] I. Schmidt, A. Boisen, E. Gustavsson, K. Stahl, S. Pegrson, S. Dahl, A. Carlsson, C.J.H. Jacobsen, Chem. Mater. 13 (2001) 4416. [16] M. Tromp, J.A. van Bokhoven, M.T.G. Oostenbrink, J.H. Bitter, K.P. de Jong, D.C. Koningsberger, J. Catal. 190 (2000) 209. [17] R. Leboda, V.V. Turov, B. Charmas, J. Skubiszewska-Zieba, V.M. GunÕko, J. Coll. Int. Sci. 223 (2000) 112. ´ . Kukovecz, Z. Ko´nya, N. Nagaraju, I. Willems, A. Tama´si, A. [18] A Fonseca, J.B. Nagy, I. Kiricsi, Phys. Chem. Chem. Phys. 2 (2000) 3071. [19] J.S. Beck, J.C. Vartuli, G.J. Kennedy, C.T. Kresge, W.J. Roth, S.E. Schramm, Chem. Mater. 6 (1994) 1816. ´ . Fudala, Z. Ko´nya, Y. Kiyozumi, S.-I. Niwa, M. Toba, F. [20] A Mizukami, P. Lentz, J.B. Nagy, I. Kiricsi, Micropor. Mesopor. Mater. 35 (2000) 631. [21] M. Holzinger, O. Vostrowsky, A. Hirsch, F. Hennrich, M. Kappes, R. Weiss, F. Jellen, Angew. Chem., Int. Ed. 40 (2001) 2001. ´ . Kukovecz, Ch. Kramberger, M. Holzinger, H. Kuzmany, J. [22] A Schalko, M. Mannsberger, A. Hirsch, J. Phys. Chem. B 106 (2002) 6374. [23] S. Banerjee, S.S. Wong, J. Phys. Chem. B 106 (2002) 12144.

94

A´. Kukovecz et al. / Microporous and Mesoporous Materials 80 (2005) 85–94

[24] S.J. Park, M.K. Seo, T.J. Ma, D.R. Lee, J. Coll. Int. Sci. 252 (2002) 249. [25] A.A. Mamedov, N.A. Kotov, M. Prato, D.M. Guldi, J.P. Wicksted, A. Hirsch, Nat. Mater. 1 (2002) 190. [26] L.E. Cascarini de Torre, E.J. Bottani, Langmuir 13 (1997) 3499. [27] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309. [28] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373. ´ rez, Micropor. Mesopor. [29] J.C. Groen, L.A.A. Peffer, J. Pe´rez-Ramı Mater. 60 (2003) 1. [30] P. Pfeifer, Y.J. Wu, M.W. Cole, J. Krim, Phys. Rev. Lett. 62 (1989) 1997. [31] A. Neimark, Physica A 191 (1992) 258. [32] F. Ehrburger-Dolle, Langmuir 15 (1999) 6004. [33] A.P. Terzyk, P.A. Gauden, P. Kowalczyk, Arab. J. Sci. Eng. 28 (2003) 133.

[34] T. Kanyo´, Z. Ko´nya, F. Berger, I. De´ka´ny, I. Kiricsi, Chem. Commun. (2003) 2746. ´ . Kukovecz, F. Berger, I. De´ka´ny, I. [35] T. Kanyo´, Z. Ko´nya, A Kiricsi, Langmuir 20 (2004) 1656. [36] P. Kowalczyk, V.M. GunÕko, A.P. Terzyk, P.A. Gauden, H.Q. Rong, Z.Y. Ryu, D.D. Do, Appl. Surf. Sci. 206 (2003) 67. [37] K. Morishige, H. Fujii, M. Uga, D. Kinukawa, Langmuir 13 (1997) 3494. [38] C.G. Sonwane, S.K. Bhatia, N.J. Calos, Langmuir 15 (1999) 4603. [39] G.S. Armatas, C.E. Salmas, M. Louloudi, G.P. Androutsopoulos, P.J. Pomonis, Langmuir 19 (2003) 3128. [40] P. Staszczuk, M. Matyjewicz, E. Kowalska, J. Radomska, P. Byszewski, M. Kozlowski, Rev. Adv. Mater. Sci. 5 (2003) 471. [41] S. Niyogi, M.A. Hamon, H. Hu, B. Zhao, P. Bhowmik, R. Sen, M.E. Itkis, R.C. Haddon, Acc. Chem. Res. 35 (2002) 1105.