Hydrophobic and Hydrophilic Behavior of Micelle-Templated

---

Langmuir 1997, 13, 2773-2778

2773

Hydrophobic and Hydrophilic Behavior of Micelle-Templated Mesoporous Silica Anne Cauvel, Daniel Brunel, and Francesco Di Renzo* Laboratoire de Mate´ riaux Catalytiques et Catalyse en Chimie Organique, UMR 5618 CNRS, ENSCM, 8 rue Ecole Normale, 34296 Montpellier Cedex 5, France

Edoardo Garrone and Bice Fubini* Dipartimento di Chimica Inorganica, Chimica Fisica e Chimica dei Materiali, Universita` degli Studi di Torino, via P. Giuria 7, 10125 Torino, Italy Received November 29, 1996. In Final Form: February 18, 1997X Adsorption of water on M41s systems has been studied by means of microcalorimetry and IR spectroscopy. After thermal treatment at only 423 K, the samples show two types of surface patches, one hydrophobic (characterized by isolated silanols not interacting with water) and the other highly hydrophilic (showing the highest heat of reversible adsorption observed so far for the H2O/SiO2 system). Thermal treatment at 1023 K causes an extensive dehydration, partially recovered by dissociative adsorption. Both the propensity for silanol condensation (and the opposite reaction of hydrolysis of siloxane bridges) and the very high heat of reversible adsorption are related to the structural properties of M41s.

Introduction The disclosure of M41s materials, formed by precipitation of amorphous silica around surfactant micelles ordered in regular arrays,1 has provided an ideal method for the preparation of mesoporous solids. The texture of the surfactant mesophases can be easily controlled, and the thermal decomposition of the templating micelles sets free a corresponding regular pore system: for instance, hexagonal pores in a hexagonal pattern in the case of the largely studied MCM-41.2 Mesoporous silica has been proposed as a reference material for the study of adsorption processes in mesopores.3-6 Moreover, the potential interest of solids featuring monodispersed pores larger than the micropores of zeolites is witnessed by the number of proposed applications, including catalysis of the conversion of bulky molecules,7 catalysis of reactions with unstable products,8 separation of biologically-active molecules, and transport of biological tracers.9 The knowledge of the surface properties of mesoporous silica is of paramount interest both for the preparation of active and stable catalysts10,11 and for the tuning of their sorption properties. The SiO4 tetrahedra of M41s do not present a longrange order,2 but their degree of local ordering is currently X

Abstract published in Advance ACS Abstracts, April 1, 1997.

(1) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710. (2) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, C. T.-W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.; Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 10834. (3) Branton, P. J.; Hall, P. G.; Sing, K. S. W. J. Chem. Soc., Chem. Commun. 1993, 1257. (4) Llewellyn, P. L.; Grillet, Y.; Schu¨th, F.; Reichert, H.; Unger, K. K. Microporous Mater. 1994, 3, 345. (5) Kloetstra, K. R.; Zandbergen, H. W.; van Koten, M. A.; van Bekkum, H. Catal. Lett. 1995, 33, 145. (6) Schmidt, R.; Sto¨cker, M.; Hansen, E.; Akporiaye, D.; Ellestad, O. H. Microporous Mater. 1995, 3, 443. (7) Armengol, E.; Cano, M. L.; Corma, A.; Garcı´a, H.; Navarro, M. T. J. Chem. Soc., Chem. Commun. 1995, 519. (8) Bellussi, G.; Perego, C.; Carati, A.; Peratello, S.; Previde Massara, E.; Perego, G. Stud. Surf. Sci. Catal. 1994, 84, 85. (9) Mollo, L.; Levresse, V.; Ellouk-Achard, S.; Jaurand, M. C.; Ottaviani, F. Environ. Health Perspect., in press. (10) Brunel, D.; Cauvel, A.; Fajula, F.; Di Renzo, F. Stud. Surf. Sci. Catal. 1995, 97, 173. (11) Chen, J.; Li, Q.; Xu, R.; Xiao, F. Angew. Chem., Int. Ed. Engl. 1995, 34, 2694.

S0743-7463(96)02059-8 CCC: $14.00

under investigation.12 The analysis of the water-surface interactions can provide useful information on the nature and local arrangement of the surface groups, which cause either hydrophobic or hydrophilic behavior, characterized by a heat of adsorption smaller or larger than the heat of liquefaction of water (44 kJ mol-1), respectively. Calorimetric and volumetric data about water adsorption can easily differentiate, for instance, crystalline and amorphous pyrogenic silica (produced by burning silicon tetrachloride).13-17 This kind of data about micelle-templated silicas are still lacking. A recently published paper only reports isosteric heats of adsorption of water on the surface of MCM-41 outgassed at 423 K.18 A hydrophobic waterMCM-41 interaction has been observed at low coverage, followed by a more exothermic adsorption, which attains 58 kJ mol-1. As IR spectroscopy shows the presence of isolated silanols on the observed samples,11,18 the two kinds of adsorption have been justified by a model envisaging adsorption on isolated silanols at low coverage and further formation of water clusters around the initial adsorption sites. Isolated silanols on a silica surface outgassed at low temperature are usually found only on pyrogenic silicas. Their presence on MCM-41 and their influence on the adsorption of water deserve further characterization. In the present work, we present and discuss microcalorimetric and IR data on the adsorption of water vapor on samples of mesoporous silica pre-equilibrated and outgassed under different conditions. In the Discussion, the peculiarities of the surface properties of mesoporous silica (12) Froba, M.; Behrens, P.; Wong, J.; Engelhardt, G.; Haggenmu¨ller, Ch.; van de Goor, G.; Rowen, M.; Tanaka, T.; Schwieger, W. Mater. Res. Soc. Symp. Proc. 1995, 371, 99. (13) Bolis, V.; Marchese, L.; Coluccia, S.; Fubini, B. Adsorpt. Sci. Technol. 1989, 5, 239. (14) Bolis, V.; Fubini, B.; Marchese, L.; Martra, G.; Costa, D. J. Chem. Soc., Faraday Trans. 1991, 87, 497. (15) Bolis, V.; Cavenago, A.; Fubini, B. Langmuir, in press. (16) Fubini, B.; Bolis, V.; Cavenago, A.; Ugliengo, P. J. Chem. Soc., Faraday Trans. 1992, 88, 277. (17) Fubini, B.; Bolis, V.; Cavenago, A.; Garrone, E.; Ugliengo, P. Langmuir 1993, 9, 2712. (18) Llewellyn, P. L.; Schu¨th, F.; Grillet, Y.; Rouquerol, F.; Rouquerol, J.; Unger, K. K. Langmuir 1995, 11, 574.

© 1997 American Chemical Society

2774 Langmuir, Vol. 13, No. 10, 1997

Cauvel et al.

Table 1. Characterization of materials by elemental analysis, X-ray diffraction and N2 sorption Sample A A B

Al/(Si+Al) Toutgas d100 SBET Vmp DBJH D4v/S (M) (K) (Å) (m2‚g-1) (ML‚g-1) (Å) (Å) 0.000 0.000 0.004

423 1073 423

45 40 42

702 673 907

0.56 0.49 0.73

30 28 30

32 29 32

are highlighted by comparison with literature data on the surface behavior of other amorphous and crystalline silicas. Experimental Section Materials. Two samples of micelle-templated silica have been prepared, with the same pore size but different thicknesses of the walls between mesopores.19 The reagents used were Aerosil 200V (Degussa, for sample A) or Zeosil 175MP (Rhoˆne-Poulenc, 0.17 wt % Al, for sample B) silica, cetyltrimethylammonium (CTMA) bromide (Aldrich), sodium hydroxide, and deionized water. The composition of the synthesis mixture was 0.13 NaOH, 0.10 CTMABr, SiO2, 32 H2O for sample A and 0.28 NaOH, 0.10 CTMABr, SiO2, 32 H2O for sample B. In both cases, reagents were mixed under stirring at 343 K and heated in a stirred autoclave at 393 K for 16 h. After synthesis, the solid phase was washed with water and ethanol and dried at 353 K in air. The organic template was decomposed by calcination at 823 K in air flow for 7 h, and the solids were re-equilibrated with ambient moisture at 298 K for at least 2 days. These samples were outgassed in vacuo at 423 K (2 h) in order to desorb molecular water without substantial condensation or hydrolisis. To observe the effects of thermal treatment at higher temperature, the calcined sample A was also outgassed at 1073 K in vacuum. Both outgassed samples were used as such in the calorimetric experiments and, after re-equilibration with ambient moisture at 298 K, in X-ray diffraction and N2 sorption experiments. Elemental compositions were determined by atomic absorption spectroscopy, and the mole fraction XAl ) Al/(Si + Al) is reported in Table 1. The small amount of aluminum observed in sample B comes from the source of silica used. Powder X-ray diffraction (XRD) experiments have been carried out by using a CGR Theˆta-60 diffractometer with monochromated Cu KR radiation and 0.25, 0.40, 0.40, and 0.25 mm slots. All samples featured the characteristic low-angle XRD pattern of mesoporous silica. The interplanar spacing of the most intense line, attributed to the 〈100〉 diffraction of the hexagonal array, is reported in Table 1 for the activated samples. The decrease of lattice parameter and pore diameter with thermal treatment corresponds to the condensation of silanols inside the walls between the pores.20 Textural properties have been characterized by N2 sorption at 77 K on calcined samples outgassed at 423 K. N2 sorption experiments have been carried out in a Micromeritics ASAP 2000 apparatus. All samples featured the characteristic type IV reversible N2 sorption isotherm described for MCM-41 materials. No micropore contribution was observed by t-plot analysis of the isotherms, and the BET surface area SBET, measured by assuming that an adsorbed N2 molecule corresponds to an area of 0.162 nm2, is reported in Table 1. The mesopore volume Vmp measured at the top of the type IV adsorption step is also reported in Table 1. The difference in mesopore volume between samples A and B is accounted for by the different wall thickness and the resulting different void fraction. The results of two different methods of pore size measurement are given in Table 1. DBJH is evaluated from the pressure of the step of the type IV isotherm by applying the Kelvin equation modified for the thickness of preadsorbed layers.21 D4V/S ) 4Vmp/ SBET is calculated as the ratio between measured volume and the surface of cylindrical pores. (19) Coustel, N.; Di Renzo, F.; Fajula, F. J. Chem. Soc., Chem. Commun. 1994, 967. (20) Chen, C. Y.; Li, H. X.; Davis, M. E. Microporous Mater. 1993, 2, 17. (21) De Boer, J. H.; Lippens, B. C.; Linsen, B. C.; Broekhoff, J. C. P.; van den Heuvel, A.; Osinga, Th. J. J. Colloid Interface Sci. 1966, 21, 405.

Water used in adsorption experiments was distilled several times in vacuo and rendered gas free by several “freeze-pumpthaw” cycles. Methods. Heats have been determined by means of a TianCalvet microcalorimeter (Setaram) connected to a volumetric apparatus which allowed simultaneous measurement of adsorbed amount (uptake, na), heat released (Q), and equilibrium pressure (p) for small increments of water vapor dosed to the silica sample. The procedure has been thoroughly described in previous papers.16,22 Before the adsorption of water vapor, each sample was outgassed for 2 h at a given temperature. The temperatures adopted, chosen on the basis of previous works,14,16 have been (i) 423 K, in order to remove contaminating adsorbed molecules and physisorbed water, and (ii) 1073 K, the temperature at which most amorphous silica is virtually hydrophobic and crystalline silicas exhibit a marked decrease in the surface activity toward water. The temperature of the calorimeter was maintained at 303 K throughout the adsorption experiment. A typical adsorption sequence comprised two runs, with the following procedure: (i) dosing successive amounts of water vapor to the sample up to a pressure of typically 5-10 Torr (Ads I), (ii) desorption at 303 K under vacuum, and (iii) readsorption of doses in order to evaluate the reversible adsorption (Ads II). The high surface area of the samples allowed us to minimize the relative extent of the adsorption of water on the frame glass walls, making comparisons between the various samples reasonably error-free. IR spectra were recorded on a FT-IR Perkin Elmer 1760-X at room temperature. A self-supported wafer of the calcined sample was outgassed under vacuum at 423 K before IR measurements. Controlled amounts of water vapor were admitted to the sample, and the equilibrium pressure in the IR cell was recorded after each dose. The sample was evacuated at room temperature after water adsorption.

Results Calorimetric data. The isotherms of adsorption of water vapor on mesoporous silica samples A and B outgassed at 423 K are shown in Figure 1a. Uptakes increase with pressure with a negative curvature up to a change in the slope at p ∼ 6 Torr (p/p° ∼ 0.19), beyond which the isotherms are concave. Both samples yield the same results, both in the first and the second adsorption run. Differences between the first and the second run on each sample are within the experimental uncertainty, and therefore the adsorption appears to be reversible, although further evidence shows that this is not strictly so. There is no evidence of differences due to aluminum sites on the adsorption of water on sample B. The integral heats of adsorption on both samples A and B outgassed at 423 K in the first and second runs are reported in Figure 1b, plotted against the equilibrium pressure (plots hereafter referred to as calorimetric isotherms). For both samples, the heat evolved in the second run is slightly higher. The corresponding differential heats of adsorption as a function of water pressure are shown in Figure 1c. The initial heat is about 90 kJ mol-1, but beyond 0.05 Torr (coverage 0.1 µmol m-2) a plateau at 60-70 kJ mol-1 is observed. At pressure values higher than 6 Torr (coverage higher than 3 µmol m-2), corresponding to the inflection point of the isotherm, the enthalpy of adsorption slowly increases. The effect of severe thermal treatment is visible in Figure 2, where the isotherms and differential heats are compared for the first runs on the sample A outgassed at 423 and 1073 K. Uptakes on the sample outgassed at 1073 K are much lower (Figure 2a), and the isotherm becomes completely concave. The first dose only suggests the possible presence of a convex portion at very low (22) Fubini, B. Thermochim. Acta 1988, 135, 19.

Micelle-Templated Mesoporous Silica

Langmuir, Vol. 13, No. 10, 1997 2775

Figure 2. (a) Volumetric isotherms (na vs p/p°) and (b) differential heat of adsorption (q vs p/p°) for H2O in the first adsorption run on mesoporous silica sample A outgassed at 423 (dots) and 1073 K (squares).

Figure 1. (a) Volumetric isotherms (na vs p/p°); (b) calorimetric isotherms (Q vs p/p°); and (c) differential heat of adsorption (q vs p/p°) for H2O in the first (filled symbols) and second (empty symbols) adsorption runs on mesoporous silica samples A (dots) and B (triangles) outgassed at 423 K.

pressure. Such a point, however, is bound to be affected by a significant error. The corresponding differential heats of adsorption are reported in Figure 2b as a function of water pressure. After the first dose, adsorbed with an enthalpy of about 90 kJ mol-1, the heat of adsorption on the sample outgassed at 1073 K presents a plateau at about 35 kJ mol-1, corresponding to a hydrophobic surface. At a pressure of water vapor of about 6 Torr (p/p° ∼ 0.19), corresponding to coverages lower than about 1 µmol m-2, the differential heat increases and attains the same values observed for the sample outgassed at 473 K. The first and second adsorption runs on sample A outgassed at 1073 K are compared in Figure 3. Opposite to what is usually found,22 in the second run, uptakes at the same “equilibrium” pressure are higher (Figure 3a). Furthermore, the isotherm is definitely convex. In this respect, the first adsorption run cannot be considered an equilibrium curve. The calorimetric isotherms are compared in Figure 3b. The enthalpy of adsorption is much higher for the second run, suggesting that irreversible modifications slowly occurring in the first run consistently increased the hydrophilic character of the surface. The differential heats of the first and second runs are reported in Figure 3c. The heat evolved in the second run

presents a plateau at pressure lower than 6 Torr (coverage higher than ∼1 µmol m-2). The molar enthalpy of the plateau is about 50 kJ mol-1, much higher than the 35 kJ mol-1 for the first run. Uptake and enthalpy data are consistent with a surface no longer hydrophobic after the first run. Infrared Spectroscopy. The IR spectra obtained by dosing water on sample B are reported in Figure 4. In order to investigate in more detail the frequency range characteristic of the free and terminal silanols, the range 3650-3800 cm-1 is reported at higher magnification in the inset. The sample outgassed at 423 K (curve a) presents a sharp double band with maxima at 3747 and 3742 cm-1, a weaker band at 3715 cm-1, a shoulder at 3653 cm-1, and a broad signal centered at 3536 cm-1. Curve b corresponds to the sample equilibrated with a pressure of water vapor of 1.35 Torr (p/p° ) 0.04). The two bands at 3747 and 3742 cm-1 are not significantly modified; the intensity of the band at 3715 cm-1 decreases, while the bands at 3653 and 3536 cm-1 increase, this latter becoming distinctly asymmetrical due to the lower frequency contribution of adsorbed molecular water. The bending vibrations of molecular water can be observed at 1628 cm-1. Curve c corresponds to the solid equilibrated with a vapor pressure of 9.0 Torr (p/p° ) 0.28). Other intermediate spectra are available but not reported. The band at 3742 cm-1 is reduced to a shoulder of the 3747 cm-1 band, which is unaffected; the band at 3715 cm-1 has disappeared, while the bands at 3653 and 3536 cm-1 are merged in the broad band due to O-H stretches of molecular water with a maximum at about 3300 cm-1. After evacuation at room temperature the spectrum is represented by curve d. The curve is neatly equivalent to curve a, corresponding to the sample before water sorption in the range 3670-3770 cm-1 (see inset). In the lower frequency range (main figure) a slight overall

2776 Langmuir, Vol. 13, No. 10, 1997

Cauvel et al.

Discussion

Figure 3. (a) Volumetric isotherms (na vs p/p°); (b) calorimetric isotherms (Q vs p/p°); and (c) differential heat of adsorption (q vs p/p°) for H2O in the first (filled squares) and second (empty squares) adsorption run on mesoporous silica sample A outgassed at 1073 K.

Figure 4. IR spectra of mesoporous silica sample B (curve a) outgassed at 423 K, (curve b) equilibrated with water vapor at 1.35 Torr (p/p° ) 0.04) and (curve c) 9.0 Torr (p/p° ) 0.28), and (curve d) evacuated at room temperature.

increase in adsorption is observed: the adsorption is basically reversible, the rehydration of the surface by water contact being marginal.

The thermal history of the sample deeply influences the reactivity of the surface toward water. In fact, samples are basically hydrophilic when outgassed at 423 K and hydrophobic when outgassed at 1073 K. In the following, we discuss the properties of the two kinds of surface and their interconversion, and make the comparison with other kinds of silica surfaces (Table 2). The main features of the interaction of water with the samples outgassed at 423 K are as follows: (i) The adsorption is essentially reversible. Minor irreversible effects are only evidenced by the integral adsorption heats and, to some extent, by IR spectra. (ii) Adsorption occurs below the inflection points of the isotherms close to values measured on largely hydrophobic pyrogenic silica (Aerosil 50 in Table 2).13,14,16 (iii) Heat values are higher than those found on hydrophilic crystalline silica (cristobalite in Table 2).13,14,16 (iv) A remarkably extended plateau occurs at a constant heat of about 60 kJ mol-1 in the pressure range 0.1-6 Torr (0.03 < p/p° < 0.19). (v) The heat of adsorption is slightly enhanced for p > 6 Torr (p/p° > 0.19). (vi) There is a very limited amount of high-energy sites: an interaction energy higher than 65 kJ mol-1 is only observed in the pressure range below 0.1 Torr (p/p° 0.03). A larger range of heterogeneity is present on previously examined silica samples.13-17 Hydrophilicity of Micelle-Templated Silica. IR spectroscopy indicates the presence of several kinds of silanols on a sample outgassed at 423 K (Figure 4). Isolated silanols are present (3747 cm-1 band) as well as mutually H-bonded species: the 3715 and 3536 cm-1 bands can be attributed to terminal and H-bonded partners of pairs of hydrogen-bonded silanols, respectively, nearly freely interacting, like those coming from the opening by water of three-membered siloxane rings.23 Very likely, the 3742 cm-1 band can be attributed to terminal silanols interacting with neighbor groups in a less efficient way. Upon adsorption of water, the terminal silanols at 3715 cm-1 disappear first, followed by those at 3742 cm-1. When the solid is at equilibrium with 9 Torr of vapor pressure (p/p° ) 0.28), the band of isolated silanols at 3747 cm-1 is unaffected. This suggests a different acidity for three species, in the order 3747 < 3742 < 3715 cm-1. The occurrence of H bonding renders the terminal silanols more acidic, in agreement with calculations on model systems.17 The main conclusion is that isolated silanols are not prone to engage in H-bonding with water, even under 9 Torr pressure, and that adsorption takes place on species already mutually interacting. The adsorption of water on pairs of H-bonded silanols represents a good model for the adsorption of up to one molecule per site. If chains or patches of H-bonded silanols are present on the surface, each water molecule can be hydrogen-bonded to two silanols, and each silanol can interact with two water molecules, as schematized in Figure 5. The presence of two hydrogen bonds per molecule is needed to justify the high differential heat measured, about 60 kJ mol-1, much higher than that found on pyrogenic silica at the same equilibrium pressure. From the isotherm of Figure 1a on the solid outgassed at 423 K, we obtain that 4.3 µmol m-2 has been adsorbed at 9 Torr, corresponding to 2.6 molecules nm-2. If the adsorption of one water molecule per two silanols is assumed, as well as the usual density of silanols on the surface of an amorphous silica (about 4.6 sites nm-1), the coverage of 2.6 molecules nm-2 arrived at roughly corresponds to a monolayer. In this case, the inflection of (23) Morrow, B. A.; Cody, L. A.; Lee, L. S. M. J. Phys. Chem. 1976, 80, 2761.

Micelle-Templated Mesoporous Silica

Langmuir, Vol. 13, No. 10, 1997 2777

Table 2. Water Uptake and Molar Heat of Adsorption at p ) 5 Torr (p/p° ) 0.156) on Silica Samples of Different Origin Outgassed at 423 and 1073 K

sample cristobalitea Aerosil 50b mesoporous silicac a

outgassed at 423 K 2nd run q na (µmol‚m-2) (kJ‚mol-1) 8.10 2.00

outgassed at 1073 K 1st run na (µmol‚m-2)

q (kJ‚mol-1)

6.45 0.75 69

44 40 0.75

53 40 2.70

2nd run na q (µmol‚m-2) (kJ‚mol-1) 3.80 0.40 43

47 30 0.9153

Data from ref 16. b Data from ref 17. c Sample A from this work.

Figure 5. Scheme of the adsorption of water molecules on chains of silanols, accounting for the adsorption of one molecule per site at the unusually high differential heat of 60 kJ mol-1.

the isotherm beyond 6 Torr would correspond to the onset of a second layer. This interpretation, however, does not take into account the IR evidence, showing the presence under 9 Torr of water of a large amount of isolated silanols not engaged in H-bonding. Being that a significant fraction of the surface is still free from water at a coverage of 2.6 molecules nm-2, water adsorption has taken place at higher coverage than one molecule per two silanols on the rest of the surface. In conclusion, sample A outgassed at 423 K seems to be made up of two kinds of surfaces, one hydrophobic, carrying isolated silanols untouched by water, and the other highly hydrophilic, carrying rather strongly held clusters of water molecules. The inflection of the isotherm and the increase of the interaction enthalpy beyond 6 Torr (Figure 1) must therefore be amened to the adsorption of other layers of water on the portions of the surface already covered by a monolayer. The formation of clusters involving several water molecules per silanol could represent the cooperative effect needed to account for the increase in adsorption heat at high coverage, unaccountable for in simple multilayer adsorption. Hydrophobicity and Nonequilibrium Effects. The low differential heat of adsorption on the sample outgassed at 1073 K (Figure 2b) can be related to the physisorption of water on the nearly dehydroxilated surface. During this process, however, some dissociation of water takes place, with the corresponding generation of new adsorption sites by hydrolysis of siloxane bridges. This process accounts for the increase of differential heat in the first adsorption run beyond a coverage of 0.9 µmol m-2 (0.6 molecules nm-2, p ∼ 6 Torr) and for the second run isotherm being higher than the first one (Figure 3a). The modification of the surface is by no means complete at the end of the first run. In fact, the increase in differential heat in the second run beyond a coverage of 1.1 µmol m-2 (0.7 molecules nm-2, p ∼ 6 Torr, Figure 3c) can again be assigned to hydrolysis of siloxane bridges on a surface not completely equilibrated with water vapor. In this respect, both adsorption runs cannot be considered equilibrium curves. Data are consistent with the chemisorption of water on the siloxane-rich surface of mesoporous silica outgassed at high temperature being an activated process, whose rate only becomes significant beyond p ) 6 Torr (p/p° ) 0.19) under our measurement conditions.

Nonequilibrium phenomena are also present, although to a much lesser extent, on the sample outgassed at 423 K. Equilibration of the samples with moist air after calcination is a slow process, and the amount of isolated silanols present in the sample outgassed at 423 K (Figure 4) clearly indicates that it has by no means been completed. Probably calcination or any high-temperature treatment creates a wide distribution of siloxanes with different geometries. Some of them are more stable and require higher water pressure to be hydrolyzed, as observed by Llewellyn et al.18 The hydrophilic or hydrophobic behavior at low vapor pressure influences the shape of the adsorption isotherm. The initial stage of the isotherms on mesoporous silica outgassed at 423 and 1073 K, reported in Figure 2a, clearly indicates that a concave isotherm is only observed for the sample outgassed at the highest temperature. Also in this last case, adsorption of water vapor modifies the slope of the isotherm, which becomes convex, as shown in Figure 3a. As a consequence, the isotherm of water adsorption is type V only after treatment of the sample at high temperature. In the case of samples pre-equilibrated with water vapor, a type IV adsorption isotherm is observed. Peculiarity of the Adsorption Behavior of MicelleTemplated Silicas. It is informative to compare the adsorption behavior of silicas templated by surfactant micelles with that of silicas obtained by more conventional ways of synthesis or formation. In Table 2 are collected water uptake and molar enthalpy at a reference equilibrium pressure p ) 5 Torr (p/p° ) 0.156) for representative samples of crystalline, amorphous, and micelle-templated silica (respectively, cristobalite from Fubini et al.,16 Aerosil 50 from Fubini et al.,17 and sample A from this work). The conditions of outgassing and measurement closely matched for all samples. Data are given for second adsorption runs, considered as more representative of reversible phenomena. In the case of samples outgassed at high temperature (1073 K), data for first adsorption runs are also given, to point out differences in the rehydration patterns. Mesoporous silica differs from previously characterized amorphous and crystalline silicas. A first comparison can be made for the samples outgassed at 423 K (Table 2). The adsorbed amounts on micelle-templated silica are close to the values measured on hydrophobic pyrogenic silica, while the interaction energy is higher than the highest value so far observed on any silica system. These data can be accounted for by the simultaneous presence of a very hydrophilic fraction of the surface, constituted of dense patches of interacting silanols, and a hydrophobic part, formed of siloxane bridges and isolated silanols. The amount adsorbed reversibly is probably proportional to the extent of the hydrophilic surface, while the hydrophobic patches remain essentially water-free at low vapor pressure. The exceptionally high measured values of adsorption enthalpy are closely reminiscent of the high isosteric heats observed in the case of the adsorption of cyclopentane

2778 Langmuir, Vol. 13, No. 10, 1997

on MCM-41.24,25 This strong interaction has been explained by an effect of the curvature of the surface, as isosteric heat increases at smaller pore size.25 A similar effect can contribute to the adsorption enthalpies observed in the case of water. When the solids are outgassed at higher temperature (1073 K), adsorbed amounts always decrease, albeit to a larger extent for pyrogenic amorphous silicas than in the case of cristobalite. The molar adsorption heat is less affected from the thermal treatment in the case of cristobalite, due to the stiffness of the lattice, which prevents silanols from condensing. In the case of all amorphous silicas, the adsorption heat decreases at values