Characterization of the hydrophobicity of mesoporous silicas and clays with silica pillars by water adsorption and DRIFT

Journal of Colloid and Interface Science 317 (2008) 206–213 www.elsevier.com/locate/jcis

Characterization of the hydrophobicity of mesoporous silicas and clays with silica pillars by water adsorption and DRIFT João Pires a,∗ , Moisés Pinto a , Juncal Estella b , Jesús C. Echeverría b a Departamento de Química e Bioquímica da Faculdade de Ciências de Lisboa, Centro de Química e Bioquímica, Edifício C8, Campo Grande, Lisboa, Portugal b Departamento de Química Aplicada, Universidad Pública de Navarra, Campus Arrosadía, 31006 Pamplona, Spain

Received 21 June 2007; accepted 14 September 2007 Available online 18 September 2007

Abstract The hydrophobic–hydrophilic properties of a solid are related to the material chemistry and, often, these properties are relevant to the applications of a particular material. Contrarily to what happens with other properties, such as specific surface areas or pore volumes, the methodologies to ascertain on the hydrophilicity of a porous material are not well defined. In this work, we discuss and relate the information on the hydrophobicity degree obtained from water adsorption isotherms and from diffuse reflectance infrared Fourier transform (DRIFT), in a set of porous materials. The studied materials were mainly mesoporous solids, namely of MCM-41 and SBA-15 types, two xerogels and also different porous clays heterostructures. Both techniques were informative on the hydrophobic–hydrophilic properties of the studied samples, but the correlation between the information obtained by each technique was not straightforward. Water adsorption isotherms are much more sensitive to the differences of the studied materials than the DRIFT spectra. For silica-based mesoporous materials with similar surface chemistry, the water adsorption process and hence, the hydrophobic–hydrophilic properties, is mainly dependent on the pore diameters. However, water adsorption is much more sensitive to changes in the nature of the adsorbent surface than to changes in the pore diameter. © 2007 Elsevier Inc. All rights reserved. Keywords: Water adsorption; DRIFT; Mesoporous materials; Silicas; Hydrophilicity; Hydrophobicity

1. Introduction Mesostructured silica and silica-like materials, either unmodified or surface modified, have a wide range of applications in adsorption and catalysis [1–3]. Textural properties such as specific surface area, or pore size distributions and pore volumes are the most relevant properties of these materials in relation with their potential uses. However, for some applications as, for instance, in the abatement of volatile organic compounds (VOCs) [4], in the controlled drug release [5], or even for the stability of some silica structures [6,7], the hydrophobic nature of the material is also a relevant property. There are in the literature various proposal to access the hydrophobicity degree of porous materials. Some of these proposals are based on the water loss at different temperatures using * Corresponding author. Fax: +351 217 5000 088.

E-mail address: [email protected] (J. Pires). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.09.035

thermogravimetry [8], and the use of data from thermogravimetry and nitrogen adsorption [9] or heats of immersion [10] were also suggested. In the case of zeolites, the use of competitive adsorption of water and hydrocarbons was also considered [11]. The degree of hydrophobicity of the surface reflects the material chemistry and, by its very nature, can be related with the interaction with water molecules and, therefore, it is expected that water adsorption isotherms can be highly informative in this context. A considerable number of studies exist in the literature related to the adsorption of water in mesoporous silica or related materials [7,12–15]. However, these studies normally concern more about the characterization of a given surface, or the structural stability toward water, and less the hydrophobicity degree of a given sequence of materials. One of the few studies that proposed the assessment of the hydrophobic–hydrophilic properties of a series porous solids [16], with different material chemistry, from the water adsorption isotherms used the energetic parameter of the Dubinin–Asthakov equation [17]. More recently, the use of the diffuse reflectance infrared Fourier

J. Pires et al. / Journal of Colloid and Interface Science 317 (2008) 206–213

transform (DRIFT) spectroscopy, was proposed for the quantification of the degree of hydrophobicity [6]. Nevertheless, to our knowledge, the information on the hydrophobicity degree obtained for a given set of materials from the water adsorption isotherms and DRIFT experiments was not yet compared and discussed. In the present work we intend to discuss the information on the hydrophobicity degree obtained, from water adsorption isotherms and from DRIFT spectra, in a set of porous solids with different material chemistry. The studied solids are mainly mesoporous (pore widths between 2 and 50 nm) [18], and two of them, the MCM-41 [19] and the SBA-15 materials, present a well defined mesoporous structure [14,20]. Additionally, two amorphous xerogels prepared by sol–gel process in acid media and aged in NH3(aq) , as well as various porous clays heterostructures (PCHs) were also studied. PCH solids are relatively new and are obtained by surfactant direct assembly of open framework silica in the galleries of different types of clays [21–23]. PCHs have pore sizes in the transition between micro to small mesopores, a particularity that make them potential catalysts for transformations that involve molecules with sizes larger than the pores of conventional zeolites [24]. 2. Experimental 2.1. Materials MCM-41 was synthesized according to procedures described in the literature [25]. Briefly, 5.28 g of cetyltrimethylammonium bromide—CTAB (Aldrich), was dissolved in 264 cm3 of deionized water and aqueous ammonia was added. The silica source, tetraethoxysilane—TEOS (Aldrich, 98%) was then added (22.4 cm3 ) under stirring. The solid was dried at 90 ◦ C and heated at 550 ◦ C for five hours, after a ramp of 1 ◦ C min−1 . The preparation of SBA-15 was adapted from original procedures [20,26] by adding 126 cm3 of a 1.6 M HCl solution to 4.0 g of poly(ethylene glycol)-block-poly(propylene glycol)-block-poly(ethylene glycol) (from Aldrich, Mn ∼ 5.800). The mixture was stirred until it became colourless and TEOS (9.1 cm3 ) was then added also under stirring. The mixture was sealed and kept in a drying oven at 35 ◦ C for 24 h and at 100 ◦ C for an additional period of 24 h. The solid was filtered, dried, and calcined at 550 ◦ C for five hours, after a ramp of 1 ◦ C min−1 . Xerogels were synthesized using the sol–gel process as detailed described elsewhere [27]. TEOS (Fluka, 98%) and ethanol (Merck, GR absolute) were mixed in a 1:4.75 molar ratio with a magnetic stirrer. While stirring, the amount of water (Milli-Q quality) needed to obtain a TEOS:water molar ratio of 1:5.5 was added drop by drop, and the pH of the solution was adjusted at 4.5 by adding 0.05 M HCl (Merck) with an automated burette (Titrino 702 SM, Metrohm, Herisau, Switzerland). Sample containers were closed and kept until gelation in an orbital shaker. Alcogels were covered with 5 ml of the aging solvent, which was 0.5 or 2.0 M aqueous ammonia for XG 0.5 and XG 2 samples, respectively, and allowed to age for 7 days. Then the alcogels were dried at atmospheric pressure and room

207

temperature to obtain xerogels. Glass containers were covered with parafilm with several needle holes to allow for the evaporation of the aging solvent. The PCHs were prepared from a Portuguese clay (structural formula: (Si3.70 Al0.30 )IV (Al1.16 Fe0.51 Mg0.45 )VI (Ca/2 , K, Na)0.39 ) previously characterised [28]. To the fraction <63 µm, the carbonates were removed and the sample was then washed in a dialysis tube, until a conductivity lower than 1 mS m−1 . In this work, four different porous clays heterostructures were studied, which will be labelled from PCH-1 to PCH-4. The preparation of these samples was based in the literature [21–23,29–31]. For the samples PCH-1 and PCH-2, a suspension of the clay (1 g in 100 cm3 of water) was firstly equilibrated with a 0.5 M solution of an ionic surfactant, the cetyltrimethylammonium bromide—CTAB (Aldrich)—under stirring during one night at 50 ◦ C. The solid was then separated from solution by centrifugation and washed with demineralised water until pH ≈ 7 and air-dried. A given amount of neutral amine was then added: decylamine (Aldrich, 95%) or dodecylamine (Aldrich, 98%) for PCH-1 and PCH-2, respectively. The amine was added under stirring, after which, TEOS was added as silica source, in a molar proportion amine:TEOS of 20:150 (PCH-1) or 20:120 (PCH-2). After 3 h of mixing the solid was air-dried and calcined at 650 ◦ C for 5 h with a ramp of 1 ◦ C min−1 . For PCH-3 and PCH-4, 2.5 or 25% (molar in relation to TEOS), respectively, of 3-aminopropyltriethoxysilane—APTES (Aldrich, 99%) was added. After addition of dodecylamine and TEOS, as already described for PCH-1 and PCH-2, the air-dried solids were solvent extracted with a solution of 1 M HCl in ethanol in reflux during 24 h. 2.2. Methods Nitrogen adsorption isotherms at −196 ◦ C were determined in an automatic apparatus (Micromeritics, mod. ASAP 2010). Water adsorption isotherms were also measured in a automatic apparatus (Coulter, mod. Omnisorp 100Cx), using a fixed vapour dosing method as described in more detail elsewhere [16]. The adsorption temperature (30 ± 0.1 ◦ C) was maintained with a recycle bath equipped with a temperature controller (Eurotherm, mod. 2216L). Water was bi-distilled, de-ionised, and purified by freeze-vacuum-thaw cycles. The reproducibility of the water adsorption isotherms was, as more detailed discussed in a previous work [16], checked against data obtained in a manual installation, and was better than 5%. A similar value was obtained for the nitrogen adsorption isotherms also. X-ray diffractograms were determined in a Philips PX 1710 instrument using the CuKα radiation. In the case of the PCHs samples, oriented and non-oriented mounts [32] were used but no diffraction peaks were detected. This experimental fact was already reported in the literature [23], and can be related to the poor long-range order presented by some types of porous clays heterostructures [23,24]. The diffuse reflectance infrared Fourier transform (DRIFT) spectra of the samples were collected on a Nicolet 6700 at 2 cm−1 resolution using the Smart Diffuse Reflectance acces-

208

J. Pires et al. / Journal of Colloid and Interface Science 317 (2008) 206–213

sory, at room temperature with a DTGS TEC detector. The samples were prepared by mixing with KBr in 1% weight and powdered in an agate mortar. Each collected spectrum was an average of 128 scans of the sample subtracted by the average of 64 background scans using only KBr in the sample container. These conditions allowed obtaining spectral absorbance in the range for the application of the Kubelka–Munk transformation [33]. Within each sample, the most intense absorption band, observed at ∼1090 cm−1 was used for the normalization of the band intensities. Only for the observation of the APTES absorption bands the spectra (not shown) of PCH-3 and PCH4 were obtained without the dilution with KBr. The bands of APTES were observed at 2927–2857 cm−1 due to the C–H stretching modes, at 1650 cm−1 due to N–H bending, and the bands between 1500 and 1400 cm−1 due to the aliphatic C–H bends [34,35]. 3. Results and discussion

published by other authors [25,37]. In the case of the xerogels, the samples presented pore size distributions which were broader than for the remaining materials and the maximum appeared around 4.5 nm for xerogel aged in 0.5 M NH3(aq) and 6.0 nm for xerogel aged in 2.0 M NH3(aq) . For the PCHs, the maxima in the mesopore size distributions are near 3.0 nm, but the distributions are broader for samples PCH-3 and PCH-4 in which 3-aminopropyltriethoxysilane was added in the synthesis. 3.2. Water adsorption isotherms and DRIFT spectroscopy The water adsorption isotherms at 30 ◦ C, with the adsorbed amounts in each material per unit of surface area (mmol/m2 ), in the studied samples are given in Fig. 3. In the case of the MCM-41 material, the shape of the curve can be considered of type V according to the IUPAC recommendations [18,40], and it is indicative of an hydrophobic character. In some way this

3.1. Textural properties The textural properties of the samples were evaluated from the N2 adsorption isotherms at −196 ◦ C (Fig. 1) and the respective specific surface areas, micro, meso, and total pore volumes are given in Table 1. The curves obtained for MCM-41, SBA-15, and both xerogels are of type IV according to the IUPAC classification [18], although for MCM-41, due to the lack of hysteresis, the isotherm should be classified type IVc [36]. These isotherms are characteristic of mesoporous solids and agree with literature results [7,19,37]. In the cases of the PCHs the total adsorbed amounts are lower than those found for the materials already mentioned and the isotherms in Fig. 1 present mixed characteristics, of types I and IV, as a most probable consequence of the pore sizes being in the transition range from micro to mesopores, as already discussed by several authors [23,24,31]. From the low temperature nitrogen adsorption data, the pore size distributions were obtained by the Broekhoff–de Boer method in a version that also uses Frenkel–Halsey–Hill equation (BdB–FHH) [38]. This methodology gives more reliable pore size distributions than the more currently used Barret– Joyner–Halenda (BJH) method [39] and, as discussed elsewhere [13], the results obtained by the BdB–FHH method can give mesopore size distributions that match well those obtained by more elaborated methods as those based on the density functional theory (DFT). The pore size distributions so obtained are given in Fig. 2 and the maxima for the more well-known MCM-41 (3.2 nm) and SBA-15 (6.4 nm) are within the values

Fig. 1. Nitrogen adsorption isotherms at −196 ◦ C in the indicated materials.

Table 1 Specific surface area (ABET ), microporous (Vmicro ) and total pore volumes (Vtotal ) obtained from the N2 adsorption isotherms at −196 ◦ C (m2 /g)

ABET Vmicro a (cm3 /g) Vtotal b (cm3 /g)

PCH-1

PCH-2

PCH-3

PCH-4

MCM-41

SBA-15

XG 0.5

XG 2

665 0.29 0.43

908 0.38 0.57

677 0.27 0.39

916 0.37 0.50

1087 0 0.86

728 0.06 0.87

666 0.22 0.45

517 0.17 0.55

a From the t -method [36]. b From the amount adsorbed at the relative pressure of 0.97 [36].

J. Pires et al. / Journal of Colloid and Interface Science 317 (2008) 206–213

Fig. 2. Pore size distributions obtained by the BdB–FHH method for the studied samples from the nitrogen adsorption isotherms at −196 ◦ C.

209

samples, where only TEOS was used, the sigmoid character (type V isotherms) can still be observed. But for PCH-3 and PCH-4 samples, where TEOS and APTES were used, the shape of the curves reveals an enhanced interaction between the material and the adsorbed molecules at low relative pressures. According to shape of water adsorption surface concentration, the material can be divided into three groups. PCH-3 and PCH-4 presented a sharp initial rise, increasing the pressure the curves reach a plateau at p/p0 < 0.60. Samples MCM-41, XG 0.5, PCH-1, and PCH-2 present sigmoid curves. SBA-15 and XG 2 samples have reduced isotherms that are convex toward the relative pressure axis in the highest range of pressures. SBA-15 presented values very close to the MCM-41 until p/p 0 near 0.5, but from this point onwards the curves diverge. The water adsorption process seems to be dependent on the surface chemistry of the sample, surface coverage and pore size. For PCH samples that have similar pore size, the different behaviour of PCH-3 and PCH-4 at low pressures, in relation to PCH-1 and PCH-2, can be explained by the presence of amine groups on the surface. For materials with pores similar to those of MCM-41 or smaller, the water adsorption proceeds by a process of surface coverage, being completely covered at values about 0.02 to 0.035 mmol/m2 . It is interesting to note that these values are comparable with those reported by Dubinin et al. in activated carbons [42] and by us in a previous work with various types of materials [43]. In the cases of materials with wider pores like xerogels, the water capillary condensation at high relative pressures can also play an important role [42]. In fact, in the cases of materials with wider pores the water adsorption process seems to be dependent not only on the surface coverage but also on a superimposed process of capillary condensation, as reported by other authors [42]. The above mentioned observations denote, in a qualitative way, the different behaviours of the various samples toward the adsorption of water, that is, their different hydrophobic– hydrophilic properties. One approach to quantify the hydrophobic–hydrophilic properties of the materials from the water adsorption isotherms could be the analysis of the Henry constants (K). These values, determined from the initial slopes of the isotherms (as p/p 0 → 0) are given in Table 2. The PCH-1 and PCH-2 presented low values of K, close to the value obtained for the MCM-41 sample, but the SBA-15 material presented a somewhat high value. The high values obtained for the xerogels, when comparing with MCM-41 and SBA-15 could Table 2 Hydrophobicity parameters obtained from the water adsorption isotherms (K and Amax ) and the DRIFT spectra (A(Si–Od ) ) Water adsorption

Fig. 3. Water adsorption isotherms at 30 ◦ C expressed by surface unit of each adsorbent material.

curve approaches the sigmoid shape of the isotherms obtained in highly hydrophobic materials such as activated carbons [41]. In the case of SBA-15, and also for both studied xerogels, the convexity toward the relative pressure axis is extended, in relation to what happened to the MCM-41. For PCH-1 and PCH-2

MCM-41 SBA-15 XG 0.5 XG 2 PCH-1 PCH-2 PCH-3 PCH-4

DRIFT

K (µmol/m2 )

Amax (kJ/mol)

A(Si–Od ) (%)

9.2 19.2 39.0 40.6 7.5 9.9 63.5 45.8

1.5 0.4 1.0 0.6 1.7 1.5 3.4 3.7

4.7 2.7 6.6 5.2 3.8 7.0 8.4 6.8

210

J. Pires et al. / Journal of Colloid and Interface Science 317 (2008) 206–213

eventually be related with a highest number of silanols in the latter materials. Nevertheless, what is more important to observe in the context of the present work is the fact that the values of K would lead to the observation that the xerogels would be more hydrophilic than the MCM-41 or the SBA-15 samples. In the low pressure region (p/p 0 below 0.1) the adsorbed amounts are very sensitive to the presence of hydrophilic groups, even if the number of these groups is small and does not change the properties of the materials surface. Therefore the analysis based on the K values could be misleading since it does not reflect the main behaviour of the material toward water adsorption. A different approach to quantify these observations is applying the well known concept of adsorption potential A (A = RT ln(p 0 /p)) to determine the value of A where the adsorption isotherm changes the curvature (from convex to concave) in relation to the pressure axis. Mathematically this point corresponds to the inflexion point of the adsorption curve. It can be more easily determined from the maximum of the numerical derivative of the adsorbed amounts with respect to A (the adsorption potential distribution curves). This particular point Amax depends on the given adsorbent material and is associated with the position of the sharp rise in the adsorption isotherm, since mathematically it also corresponds to the point of the curve where the maximum slope is observed. Although in the present work the determination of Amax is independent of any adsorption model assumption, this parameter is analogous to the energetic parameter of the Dubinin–Asthakov equation [17,44]. In fact, it was previously shown [16] that this parameter can coherently express the evolution of the hydrophobic– hydrophilic properties of various materials although, the extension of this treatment to mesoporous materials was not, to our knowledge, previously made. This methodology can, in principle, be applied to several types of materials and water adsorption isotherms. In theory, Amax can vary between ∞ and 0. A high value of Amax will correspond to a material with a high affinity to water, whereas a low value of Amax is characteristic of a hydrophobic material. In terms of the shape of the water adsorption isotherms, this means that only the isotherms that are highly concave to the pressure axis (type I isotherms [18]) will have a high value of Amax . Conversely, the isotherms that are convex (type III [18]) or that change the curvature from convex to concave (type V [18]) will present low Amax values. The Amax values obtained for the materials studied in this work are given in Table 2. Amax values for PCH-3 and PCH-4 materials are 3.4 and 3.7 kJ/mol, respectively, being in this way the more hydrophilic samples of the series. The materials that present sigmoid isotherms (MCM-41, PCH-1, and PCH-2) have similar Amax values in the range 1.5–1.7 kJ/mol. The SBA-15, XG 0.5 and XG 2 samples, for which the isotherm is convex toward the relative pressure axis in the highest range of pressures for the series of studied materials, present Amax values 1.0 kJ/mol. The enhanced hydrophilicity of PCH-3 and PCH-4, in relation to the other studied materials, is related to the presence of the amine groups, due to the use of APTES, as discussed more detailed below. In fact, in a first approximation the Amax val-

Fig. 4. Relation between the Amax values and the pore diameters of the indicated materials.

ues for these samples could be compared with the Gibbs energy for the hydration (h G0 ) of the propylamine. In this way, using the values of Plyasunov et al. [45], and making the correction for the same standard state (298.15 K and 0.1 MPa), the h G0 value for propylamine is −10.4 kJ/mol. This can be compared with the negative Amax values, corrected for the same standard state, for PCH-3 and PCH-4 which are about −11 kJ/mol. The similarity of these values is in favour that in the PCH-3 and PCH-4 cases the relative pressure at which the inflexion point in the isotherms occurs is related with the specific interaction between the amine groups and the water. These interactions, which seem to be comparable to the hydration interactions, determine the hydrophilic properties of these materials. It is highly informative if we consider at this point the data in Fig. 4 where the Amax values are plotted against the pore diameters corresponding to the maxima in the pore size distributions (Fig. 2). As can be seen, Amax values are linearly related to pore diameters, except for the samples PCH-3 and PCH-4, which have the amine groups. For a series of materials with a similar chemical nature of the surface, Amax values are linearly related to the pore diameters. Fig. 4 also shows that the Amax parameter is much more sensitive to changes in the nature of the adsorbent surface than to changes in the pore diameter. In fact, the Amax values vary from 0.6 kJ/mol at 6.5 nm for the SBA-15 to 1.7 kJ/mol at 3 nm for PCH-1, in contrast with the change from about 1.5 to 3.5 kJ/mol due to the presence of the amine groups in solids with similar pore sizes. As already mentioned, a different approach to characterize the hydrophilic–hydrophobic properties of silica based materials by using DRIFT spectra has been proposed [6]. Being eventually more available than the water adsorption isotherms, this methodology could have some practical advantages. The DRIFT spectra for the various studied materials in the region 1500–700 cm−1 are given in Fig. 5. These spectra agree with other previously obtained, in the cases where literature data exists, namely for the MCM-41 [46] and the SBA-15 [47] materials. In Fig. 5 the strong and broad absorbance band in the 1300–1000 cm−1 region is assigned to the asymmetric stretching modes of Si–O–Si vibrational moiety—ν as (Si–O–Si) [6,46–49]. The band near 800 cm−1 is assigned to the symmetric stretching mode ν s (Si–O–Si) [37,47–49], and the band near 950 cm−1 is assigned to the dangling ν(Si–Od ) due to Si–OH and Si–O− groups [6,50]. This latter band is particularly relevant in the context of the present work since the Si–OH and

J. Pires et al. / Journal of Colloid and Interface Science 317 (2008) 206–213

211

Fig. 5. DRIFT spectra between 1500 and 700 cm−1 for the indicated samples.

Si–O− are both hydrophilic and it was proposed [6] that the ratio of these groups to the total silica could be an indication of the hydrophilicity degree of the material (the A(Si–Od ) parameter in Table 2). This could be then calculated (in %) by the ratio between the area of the ν(Si–Od ) band and the sum of the areas of the ν as (Si–O–Si) and ν(Si–Od ) bands [6]. To calculate the areas of the ν as (Si–O–Si) and ν(Si–Od ) bands, a deconvolution of the normalized DRIFT spectra was made by fitting Gaussian functions, using the nonlinear least squares method. This deconvolution took into account that the asymmetric mode (band around 1300–1000 cm−1 ) composes of two major components [6,48], due to the in-phase and the out-of-phase motion of two adjacent oxygen atoms in relation to the central silicon atom and, moreover, each of these components is associated with a longitudinal-optic and a transverseoptic vibrational modes [48]. Therefore, the band around 1300– 1000 cm−1 was fitted with four Gaussians, the best fitting being obtained when maxima around 1220, 1160, 1090, and 1060 cm−1 were used. For the band at 950–900 cm−1 two Gaussians were used with maxima around 960 and 940 cm−1 . The band around 800 cm−1 was fitted with one Gaussian with the purpose of improving the global fit of the spectral region. As an example, Fig. 6 shows the deconvolution fittings for (a) MCM-41 and (b) PCH-4. The absorption maxima (ν˜ ) and relative areas (%A) of the respective bands are given in Table 3. The hydrophilicity degrees—A(Si–Od ) —(in %) are given in Table 2 where these values can be compared with the K and Amax values. Concerning the ν(Si–Od ) vibration mode, and according to the assignments described in the literature [6,50], the absorption maxima listed in Table 3 due to the Si–OH species appear between 966 (PCH-3 and PCH-4) and 974 (XG 0.5) and those due to Si–O− between 902 (PCH-2) and 951 cm−1

Fig. 6. Deconvolution of spectral region between 1300 and 700 cm−1 for (a) MCM-41 and (b) PCH-4.

(SBA-15 and XG 0.5). The contribution to the total area of the ν(Si–Od ) from the Si–O− species in PCHs samples, is 7.3% for PCH-3 and 5.4% for PCH-4, while for the other PCHs is always lower than 2.5%. This aspect, that has consequences on the hydrophobic–hydrophilic properties of PCH-3 and PCH-4, as discussed below, may be due the presence of APTES in these samples. In fact, the terminal amine groups of the APTES molecule can form an inter-chain structure of type Si– O− · · ·H· · ·NH+ 2 –CH2 –CH2 –CH2 –Si [51] thus enhancing the relative amount of the Si–O− type species. It can be noticed that the organo-functionalization of various types of silica-like materials with aminoalkoxysilanes is important either for reinforcing the surfaces [51] or as catalysts and catalysts supports [52]. At this point, the information on the hydrophobic–hydrophilic nature of the various materials, accessed from the water adsorption isotherms and from the DRIFT spectra, can be discussed. For this, the A(Si–Od ) values (in %) can be compared with those found for the Amax parameter in each solid (Table 2). As denoted in this table the relation between the information obtained by the two methodologies used is not clear. Several reasons can justify this situation. For instance, in cases such as the PCH-2 sample the contribution to the total area of the ν(Si– Od ) band is predominantly due to species type Si–OH (band at 967 cm−1 ) and less from species type Si–O− (Table 3). The former species, which always contribute to the infrared absorption, may not be strong adsorption sites for the water molecule. Therefore, the DRIFT analysis can overestimate the hydrophilic character of the solid. Another reason is that the DRIFT spectra account for all the Si–OH groups but part of these groups

212

J. Pires et al. / Journal of Colloid and Interface Science 317 (2008) 206–213

Table 3 Maximum absorption wavenumber ν˜ (cm−1 ) and relative area (%A) obtained by deconvolution of the bands observed in the spectral region 1300–700 cm−1 PCH-1

PCH-2

PCH-3

PCH-4

MCM-41

SBA-15

XG 0.5

XG 2

ν˜ %A

1225 4.6

1219 5.6

1221 4.2

1219 4.0

1234 5.5

1216 7.3

1224 5.8

1221 6.2

ν˜ %A

1165 28.1

1143 36.9

1169 26.6

1166 24.0

1162 31.4

1139 39.1

1162 36.8

1162 36.6

ν˜ %A

1094 3.9

1082 32.8

1090 2.9

1089 3.9

1090 6.2

1085 31.3

1087 44.8

1091 46.1

ν˜ %A

1081 58.0

1041 16.4

1080 56.1

1080 60.3

1076 50.5

1045 17.3

1045 3.3

1048 3.7

ν(Si–Od ) (Si–OH)

ν˜ %A

967 1.6

967 6.6

966 1.0

966 1.2

968 2.1

973 0.7

974 1.0

973 0.7

(Si–O− )

ν˜ %A

935 2.1

902 0.3

946 7.3

946 5.4

938 2.5

951 1.9

951 5.5

950 4.4

ν˜ %A

805 1.6

806 1.4

801 1.8

800 1.1

804 1.7

809 2.4

802 2.7

803 2.4

ν as (Si–O–Si)

ν s (Si–O–Si)

can be inaccessible to the water molecules and hence, in practice, they do not contribute to the hydrophilicity of the sample via adsorbent–adsorbate interactions. In this situation the Amax values would be lower than expected on the basis of the DRIFT analysis. The possibility that this situation also arises when the contribution for the hydrophilicity comes more from the Si–O− species is lowest since the formation of these species implies some type of opening of the silica rings and, therefore, rending these groups less hindered to the water molecules. In the same line to what was discussed for PCH-2, in the xerogel samples the hydrophilicity accounted by DRIFT may be overestimated in relation to the information from water adsorption. A possible justification for this is that the Ostwald ripening may occur to some extent. In this situation the probability of the existence of larger particles is higher and, therefore, a significant number of terminal groups would not be able to interact with the water molecules. Moreover, the different amorphous or crystalline character of the various studied samples, which can be for instance accounted by the band near 1100 cm−1 [53] may difficult the comparison of the information obtained by the DRIFT and the water adsorption. An important additional reason to justify the lack of relation between the information obtained from water adsorption and DRIFT is related to the relatively high standard deviation associated with the determination of the A(Si–Od ) parameter, when compared for instance with the standard deviation for the Amax values. In fact, in the MCM-41 and SBA-15 cases the standard deviation for A(Si–Od ) is very high, reaching 40% of the estimated value. This is due to the relatively low area of the ν(Si–Od ) band that makes the determination of the area of this absorption band very sensitive to the goodness of the fit of the two Gaussians and also to the experimental noise in the data.

the correlation between the information obtained by each technique, was not straightforward. Water adsorption isotherms are much more sensitive to the differences of the studied materials than the DRIFT spectra. For silica based mesoporous materials with similar surface chemistry, the water adsorption process and hence, the hydrophobic–hydrophilic properties, is mainly dependent on the pore diameters. The lack of a more direct relation between the hydrophilicity degrees measured by DRIFT and by water adsorption can be partially related to the existence of terminal groups that are accounted by DRIFT but may not be accessible to interact with the water molecules and that, in fact, do not contribute to the hydrophobic–hydrophilic properties of the material. Additionally the standard deviation associated with the determination of the area of the ν(Si–Od ) band can be very high for some materials. As a consequence of reasons discussed above, the DRIFT analysis, due to their experimental simplicity when compared with the determination of water adsorption isotherms, could eventually be used as a first approach, particularly in materials that are structurally closed related but differ in the nature of the surface chemistry. However, to access the hydrophilic– hydrophobic properties of the materials and to relate these properties with the structure and the surface chemistry, in a way that different materials can be compared, the water adsorption, in particular through the analysis of the Amax parameter, should be preferred.

4. Conclusions

References

Both techniques used in this work, that is, the water adsorption isotherms and the DRIFT, were informative on the hydrophobic–hydrophilic properties of the studied samples but

Acknowledgments Thanks are due to FCT for the plurianual funding to CQB and to the project POCTI/CTM/56192/2004. M. Pinto thanks FCT for a post-doctoral grant.

[1] A. Taguchi, F. Schüth, Micropor. Mesopor. Mater. 77 (2005) 1. [2] F. Fajula, A. Galarneau, F. Di Renzo, Micropor. Mesopor. Mater. 82 (2005) 227.

J. Pires et al. / Journal of Colloid and Interface Science 317 (2008) 206–213

[3] A. Corma, Chem. Rev. 97 (1997) 2373. [4] A. Giaya, R.W. Thompson, R. Denkewiz Jr., Micropor. Mesopor. Mater. 40 (2000) 205. [5] J.C. Doadrio, E.B. Sousa, I. Izquierdo-Barba, A.L. Doadrio, J. PerezPariente, M. Vallet-Regi, J. Mater. Chem. 16 (2006) 462. [6] A. Fidalgo, L. Ilharco, Chem. Eur. J. 10 (2004) 392. [7] M.M.L.R. Carrott, A.J.E. Candeias, P.J.M. Carrott, K.K. Unger, Langmuir 15 (1999) 8895. [8] M. Anderson, W.J. Klinowski, J. Chem. Soc. Faraday Trans. 82 (1986) 1449. [9] A. Giaya, R.W. Tompson, R. Denkewicz Jr., Micropor. Mesopor. Mater. 40 (2000) 205. [10] T. Kawai, K. Tsutsumi, Colloid Polym. Sci. 270 (1992) 711. [11] J. Stelzer, M. Paulus, M. Hunger, J. Weitkamp, Micropor. Mesopor. Mater. 22 (1998) 1. [12] S. Komarneni, R. Pidugu, V.C. Menon, J. Porous Mater. 3 (1996) 99. [13] J. Pires, A. Carvalho, M. Pinto, J. Rocha, J. Porous Mater. 13 (2006) 107. [14] V.M. Gun’ko, V.V. Turov, V.M. Bogatyrev, B. Charmas, J. SkubiszewskaZieüba, R. Leboda, S.V. Pakhovchishin, V.I. Zarko, L.V. Petrus, O.V. Stebelska, M.D. Tsapko, Langmuir 19 (2003) 10816. [15] K.E. Collins, A.B. Dimiras, V.R. Camargo, C.H. Collins, Micropor. Mesopor. Mater. 89 (2006) 246. [16] J. Pires, M.L. Pinto, A. Carvalho, M.B. Carvalho, Adsorption 9 (2003) 303. [17] M.M. Dubinin, Carbon 27 (1989) 457. [18] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska, Pure Appl. Chem. 57 (1985) 603. [19] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, T. Kresge, K. Schmitt, C.T.-W. Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen, J.B. Higgins, J.L. Schlenkert, J. Am. Chem. Soc. 114 (1992) 10834. [20] D. Zhao, Q. Huo, J. Feng, B.F. Chmelka, G.D. Stucky, J. Am. Chem. Soc. 120 (1998) 6024. [21] A. Galarneau, A. Barodawalla, T.J. Pinnavaia, Nature 374 (1995) 529. [22] A. Galarneau, A. Barodawalla, T.J. Pinnavaia, Chem. Commun. 1661 (1997). [23] J. Pires, A.C. Araújo, A.P. Carvalho, M.L. Pinto, J.M. González-Calbet, J. Ramírez-Castellanos, Micropor. Mesopor. Mater. 73 (2004) 175. [24] M. Polverejan, Y. Liu, T.J. Pinnavaia, Chem. Mater. 14 (2002) 2283. [25] M. Grün, K.K. Unger, A. Matsumoto, K. Tsutsumi, in: B. McEnaney, T.J. Mays, J. Rouquerol, F. Rodríguez-Reinoso, K. Sing, K.K. Unger (Eds.), Characterisation of Porous Solids IV, The Royal Society of Chemistry, London, 1997, p. 81.

213

[26] W.J. Roth, J.C. Vartuli, in: J. Cejka, H. van Beckkum (Eds.), Synthesis in Studies in Surface Science and Catalysis, vol. 157, Elsevier, Amsterdam, 2005, pp. 91–110. [27] J. Estella, J.C. Echeverría, M. Laguna, J.J. Garrido, Micropor. Mesopor. Mater. 102 (2007) 274. [28] M.B. Carvalho, J. Pires, A.P. Carvalho, Micropor. Mater. 6 (1996) 65. [29] M. Polverejan, T.R. Pauly, T.J. Pinnavaia, Chem. Mater. 12 (2002) 2698. [30] M. Pichowicz, R. Mokaya, Chem. Commun. 2100 (2001). [31] R. Ishii, M. Nakatsuji, K. Ooi, Micropor. Mesopor. Mater. 79 (2005) 111. [32] B. Velde, Introduction to Clay Minerals—Chemistry, Origin Uses and Environmental Significance, first ed., Chapman & Hall, London, 1992. [33] P. Kubelka, F.Z. Munk, Tech. Phys. 12 (1931) 593. [34] A.S.M. Chong, X.S. Zhao, J. Phys. Chem. B 107 (2003) 12650. [35] A.B. Boulinos, D.D. Jiang, E.P. Gianelis, Chem. Mater. 16 (2004) 2404. [36] F. Rouquerol, J. Rouquerol, K. Sing, Adsorption by Powders and Porous Solids, Academic Press, San Diego, 1999. [37] M.S. Morey, S. O’Brien, S. Schwarz, G.D. Stucky, Chem. Mater. 12 (2000) 898. [38] W.J. Lukens Jr., P. Schmidt-Winkel, D. Zhao, J. Feng, G.D. Stucky, Langmuir 15 (1999) 5403. [39] E.P. Barret, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373. [40] P.L. Llewellyn, F. Schüth, Y. Grillet, F. Rouquerol, J. Rouquerol, K.K. Unger, Langmuir 11 (1995) 574. [41] F. Stoeckli, T. Jabukov, A. Lavanchy, J. Chem. Soc. Faraday Trans. 90 (1994) 783. [42] M.M. Dubinin, E.D. Zaverina, V.V. Serpinsky, J. Chem. Soc. (1955) 1760. [43] J. Pires, M.L. Pinto, A. Carvalho, M.B. Carvalho, Adsorption 9 (2003) 303. [44] R.C. Bansal, J.B. Donnet, H.F. Stoeckli, Active Carbon, Dekker, New York, 1988. [45] A.V. Plyasunov, J.P. O’Connell, R.H. Wood, E.L. Shock, Fluid Phase Equil. 183–184 (2001) 133. [46] D.C. Calabro, E. Valyocsik, F.X. Ryan, Micropor. Mater. 7 (1996) 243. [47] S. Shen, F. Chen, P.S. Chow, P. Phanapavudhikul, K. Zhu, R.B.H. Tan, Micropor. Mesopor. Mater. 92 (2006) 300. [48] C.T. Kirk, Phys. Rev. B 38 (1988) 1255. [49] F.L. Galeener, Phys. Rev. B 19 (1979) 4292. [50] A. Chmel, E.K. Mazurina, V.S. Shashkin, J. Non-Cryst. Solids 122 (1990) 285. [51] I. Shimizu, H. Okabayashi, K. Taga, E. Nishio, C.J. O’Connor, Vibr. Spectrosc. 14 (1997) 113. [52] A. Taguchi, F. Scüth, Micropor. Mesopor. Mater. 77 (2005) 1. [53] A. Fidalgo, L.M. Ilharco, Micropor. Mesopor. Mater. 84 (2005) 229.