Langmuir 2005, 21, 8923-8929
8923
Fluorescence Probing Investigation of the Mechanism of Formation of MSU-type Mesoporous Silica Prepared in Fluoride Medium Ce´dric Lesaint,† Be´ne´dicte Lebeau,† Claire Marichal,*,† Joe¨l Patarin,† and Raoul Zana‡ Laboratoire de Mate´ riaux a` Porosite´ Controˆ le´ e (UMR CNRS 7016), E.N.S.C.Mu., UHA, 3 rue A. Werner, 68093 Mulhouse-Cedex, France, and Institut C. Sadron (UPR CNRS 22), ULP, 6 rue Boussingault, 67000 Strasbourg, France Received April 8, 2005. In Final Form: June 28, 2005 The mechanism of formation of a MSU-type siliceous material from tetraethyl orthosilicate (TEOS) in the presence of the nonionic surfactant tergitol T-15-S-12, sulfuric acid, and sodium fluoride has been investigated using mainly fluorescence probing techniques and, to a lesser extent, dynamic light scattering (DLS) and 29Si NMR spectroscopy. The tergitol micelles present in the systems obtained by progressively generating the reaction mixture giving rise to the mesostructured material by adding to an appropriate tergitol solution sulfuric acid, TEOS, and NaF were characterized by fluorescence probing (micelle aggregation number, micropolarity, and microviscosity) and also by dynamic light scattering (apparent micelle diameter). 29Si NMR experiments were also performed on selected systems after hydrolysis of the TEOS. The fluorescence probing techniques were also used to follow the changes of micelle characteristics with time during the evolution of the full reaction mixture from a limpid solution to a system containing a minor amount of condensed siliceous material. The synthesized solid material was characterized by X-ray diffraction and nitrogen adsorption-desorption analyses. The micelle aggregation number N was found to change only little, and the micropolarity remained constant when going from the tergitol solution to the full reaction mixture. The results of DLS measurements agree with this finding. Besides, while the condensation of silica took place after addition of NaF, the N value increased only very little with time up to the point where a small amount of mesostructured material precipitated out. These results indicate that the interaction between tergitol micelles and the siliceous species formed in the system by the hydrolysis of TEOS and also between micelles and the growing siliceous species must be very weak. As in our previous studies of the mechanism of formation of MCM41-type material from sodium silicate in the presence of cetyltrimethylammonium bromide, it appears that the locus of formation of the mesostructured material is not the micelle surface but the bulk phase. Micelles only act as reservoirs of surfactant providing surfactant monomer that binds to the growing siliceous species.
1. Introduction 1,2
Since their first synthesis in 1992, using cationic surfactants, the mesoporous materials of the M41S family have been attracting worldwide interest due to their potential uses as catalysts and adsorbents. The mode of synthesis was rapidly extended to mesoporous molecular sieves templated by nonionic surfactants.3,4 Much research has focused on understanding the mechanisms of formation of these new materials. Zhao and co-workers5 proposed that these mesostructures result from a cooperative process involving electrostatic interactions between the positively charged quaternary ammonium surfactant micelles (S+) and the negatively charged silicate or aluminosilicate framework (I-). This cooperative selfassembly mechanism was extended to all the pathways leading to the formation of organized mesoporous solids * Author to whom correspondence should be addressed. † Laboratoire de Mate ´ riaux a` Porosite´ Controˆle´e. ‡ Institut C. Sadron. (1) 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. (2) Beck, J. S.; Vartuli, J. C.; Kennedy, G. J.; Kresge, C. T.; Roth, W. J.; Schramm, S. E. Chem. Mater. 1994, 6, 1816. (3) Tanev, P. T.; Pinnavaia, T. J. Science 1995, 267, 865. (4) Bagshaw, S. A.; Prouzet, E.; Pinnavaia, T. J. Science 1995, 269, 1242. (5) Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. J. Am. Chem. Soc. 1998, 120, 6024.
in the presence of ionic surfactants: electrical charge density matching between the inorganic polymer (I-) or (I+) and the surfactant (S+ or S-) and counterion-mediated synthesis (S+X-I+, X- ) halide anion or S-M+I-, M+ ) alkali metal cation). The other studies also mainly concerned the S+I- pathway and clearly showed that techniques permitting in situ investigations of the systems were necessary in order to get a better insight into the formation of these materials. This topic has been reviewed in particular by Ying et al.6 in 1999 and Patarin et al.7 in 2002. An essential assumption in the above mechanism of formation of mesotructured materials is that the micelles interact with the inorganic precursor of the material. This interaction is responsible for the change of micelle shape that must occur if, starting from spherical micelles, one ends up with a material with a hexagonal array of pores. These assumptions have been strongly questioned in recent studies of the formation of MCM-41-type siliceous materials, using sodium silicate (water glass) as the precursor, in the presence of the cationic surfactant cetyltrimethylammonium bromide (CTAB).8 Fluorescence probing studies permitted Frasch et al.8 to show that the (6) Ying, J. Y.; Mehnert, C. P.; Wong, M. S. Angew. Chem., Int. Ed. 1999, 38, 56. (7) Patarin, J.; Lebeau, B.; Zana, R. Curr. Opin. Colloid Interface Sci. 2002, 7, 107. (8) Frasch, J.; Lebeau, B.; Soulard, M.; Patarin, J. Langmuir 2000, 16, 9049.
10.1021/la0509347 CCC: $30.25 © 2005 American Chemical Society Published on Web 08/09/2005
8924
Langmuir, Vol. 21, No. 19, 2005
surfactant micelles do not interact with the inorganic precursor of the material, the siliceous species do not bind to the micelles, and the micelle size and shape undergo hardly any change up to the point where precipitation of the mesostructured material is about to start. It was concluded that the condensation of silica does not occur at the micelle surface (no templating) but in the aqueous phase and that the free surfactant provided by the micelles binds in increasing amounts and increasingly cooperatively to the growing polymeric siliceous species. This process goes on until the surfactant/siliceous polymer complex becomes very large and precipitates out as mesostructured material. The micelles act as reservoirs of surfactant. This mechanism bears many similarities with the precipitation of mesomorphous complexes when surfactants are added to oppositely charged polyelectrolyte solutions.8 Several recent studies9-17 deal with the mechanism of formation of mesoporous materials prepared via a nonionic pathway. Of particular interest is the study of the mechanism of formation of MSU-type materials which are characterized by a wormlike porous structure4 prepared by a two-steps synthesis with tetraethyl orthosilicate (TEOS) as the precursor, in the presence of a nonionc surfactant, the tergitol T-15-S-12, hydrochloric acid for hydrolyzing TEOS, and sodium fluoride to catalyze the condensation.10 The proposed mechanism assumes an accumulation of siliceous species resulting from the hydrolysis of TEOS around the nonionic micelles and leading to a change of micelle size. In view of the previous studies performed in our group we felt it worthwhile to extend our fluorescence probing studies to this system. Our aim in doing so was 2-fold: (i) bring additional information on the mechanism of formation of MSU-type silica prepared by the nonionic pathway and (ii) check the degree of generality of the conclusions reached in our previous study on the formation of the MCM-41-type siliceous material. Recall that fluorescence probing has been extensively used to characterize surfactant-containing systems.18,19 Indeed, the use of appropriate fluorescence probes permits one to obtain information on the micropolarity and microviscosity sensed by these probes in surfactant aggregates. Also time-resolved fluorescence quenching permits the determination of the number (N) of surfactants making up an aggregate (aggregation number, which allows determining the size and shape of micelles).18,19 Besides, fluorescence lifetime measurements can be used to study the exchange between two different counterions (bromide and silicate, for instance) if one of the ions (bromide) is a quencher of the fluorescence of the probe used (pyrene).8,20,21 (9) Blin, J. L.; Leonard, A.; Su, B. L. Chem. Mater. 2001, 13, 3542. (10) Boissie`re, C.; Larbot, A.; Bourgaux, C.; Prouzet, E.; Bunton, C. A. Chem. Mater. 2001, 13, 3580. (11) Ruthstein, S.; Frydman, V.; Kababya, S.; Landau, M.; Goldfarb, D. J. Phys. Chem. B 2003, 107, 1739. (12) Ruthstein, S.; Frydman, V.; Goldfarb, D. J. Phys. Chem. B 2004, 108, 9016. (13) Zhao, D.; Fenf, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.; Chmelka, B. F.; Stucky, G. D. Science 1998, 279, 549. (14) Flodstrom, K.; Teixeira, C. V.; Amenitsch, H.; Alfredsson, V.; Linden, M. Langmuir 2004, 20, 4885. (15) Flodstrom, K.; Wennerstrom, H.; Alfredsson, V. Langmuir 2004, 20, 680. (16) Flodstrom, K.; Wennerstrom, H.; Teixeira, C. V.; Amenitsch, H.; Linden, M.; Alfredsson, V. Langmuir 2004, 20, 10311. (17) Yu, C.; Fan, J.; Tian, B.; Zhao, D. Chem. Mater. 2004, 16, 889. (18) Thomas, J. K. Acc. Chem. Res. 1977, 10, 133. (19) Zana, R. Surfactant Solutions. New Methods of Investigation; M. Dekker Inc.: New York, 1987; Chapter 5. (20) Abuin, E.; Lissi, E.; Bianchi, N.; Miola, L.; Quina, F. H. J. Phys. Chem. 1983, 97, 5166.
Lesaint et al.
In the present work spectrofluorimetry with the fluorescent probes pyrene and dipyrenylpropane (DPyP) was used to obtain information on the micelle micropolarity and microviscosity, respectively, at different stages of generation of the full reaction mixture (precursor system) that leads to the mesostructured material and also, after addition of the catalyst to the precursor system in order to induce the condensation, during the time the reaction mixture remained limpid. Time-resolved fluorescence quenching of the pyrene fluorescence by cetylpyridinium chloride was used to determine the micelle aggregation number N in the same conditions. The fluorescence studies were complemented by dynamic light scattering (DLS) and 29Si NMR spectroscopy measurements that provided information on the micelle diameter at different stages of generation of the full reaction mixture and on the siliceous species present in the systems after hydrolysis of TEOS, respectively. The mesotructured material was characterized by X-ray diffraction analysis and nitrogen adsorption-desorption isotherms. Whatever the system, under our experimental conditions the results showed hardly any effect on the micelle size. It appears that there is little or no interaction between silica and the micelles in the early stage of formation of the mesostructured material, up to the point where a small amount of the material precipitates out. The condensation of silica takes place in the aqueous phase and not at the micelle surface. 2. Experimental Section 2.1. Preparation of the Materials. The MSU-type materials were synthesized using the procedure proposed by Boissie`re et al.10 slightly modified to fulfill our experimental constraints. The first step of the synthesis consists of preparing a micellar solution of the nonionic surfactant tergitol T-15-S-12 (C15H31(OCH2CH2)12OH, from Sigma; critical micellar concentration:22 cmc ) 5.4 × 10-5 mol dm-3 at the concentration C ) 0.02 mol dm-3 (this concentration closely corresponds to the molar composition 0.125 tergitol; 330 H2O). Therefore, 1.476 g of T-15-S-12 was dissolved in 95 mL of distilled water. After 1 h of stirring, 3.328 g of tetraethyl orthosilicate (TEOS, Si(OC2H5)4, 98%, from Aldrich) was added, under magnetic stirring, to the T-15-S-12 solution, resulting in a solution with a silica concentration of about 0.16 mol dm-3 and a [Si]/[surfactant] molar ratio of 8. The mixture turned immediately into a white emulsion. Its pH was adjusted to a value of 2 by addition of 5 mL of a 0.2 mol dm-3 aqueous solution of sulfuric acid (96%, from Carlo Erba). The emulsion progressively clarified, became limpid, and was kept one night at room temperature. The second step of the reaction was then started by adding a small amount of sodium fluoride ([NaF]/[Si] molar ratio ranging from 0.001 to 0.04) and heating the solution at 313 K, bringing about the condensation reaction.23 The solution remained clear for a variable time much dependent on the [NaF]/[Si] molar ratio, after which it became increasingly cloudy, indicating the formation of a solid material. The system was kept at 313 K for 72 h. The solid was recovered by filtration, washed, dried at 343 K overnight, and calcined at 893 K for 6 h. The fluorescence and NMR studies also involved reference systems where the tergitol T-15-S-12 was replaced by a poly(ethylene glycol) of molecular weight 600 (PEG-600) that closely corresponds to the EO12 headgroup of the tergitol. It should be noted that this polymer alone does not give rise to micelles. All of the studied systems are listed in Table 1. (21) Abuin, E.; Lissi, E. J. J. Colloid Interface Sci. 1991, 143, 97. (22) Sanchez-Camano, M.; Rodriguez-Cruz, S.; Sanchez-Martin, M. J. Environ. Sci. Technol. 2003, 37, 2758. (23) Corriu, R. J. P.; Leclercq, D.; Vioux, A.; Pauthe, M.; Phalippou, J.; Mackenzie, J. D.; Ultrich, D. R. Ultrastructure Processing of Advanced Ceramics; Wiley: New York, 1988.
Formation of Mesoporous Silica
Langmuir, Vol. 21, No. 19, 2005 8925
Table 1. Molar Composition of the Investigated Systems and Values of the Aging Time t, Micelle Aggregation Number N, Pyrene Intensity Ratio I1/I3 (Micropolarity), Dipyrenylpropane Intensity Ratio IM/IE, and of the Quantity Q (Proportional to Microviscosity) at 313 K ta (h)
system
molar compositions
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16c
0.125 T-15-S-12; 330 H2O 0.125 T-15-S-12; 0.062 H2SO4; 0.001 NaF; 330 H2O 0.125 T-15-S-12; 0.062 H2SO4; 1 TEOS; 330 H2O 0.125 T-15-S-12; 4 EtOH; 0.001 NaF; 330 H2O 0.125 T-15-S-12; 0.062 H2SO4; 1 TEOS; 0.001 NaF; 330 H2O 0.125 T-15-S-12; 0.062 H2SO4; 1 TEOS; 0.001 NaF; 330 H2O 0.125 T-15-S-12; 0.062 H2SO4; 1 TEOS; 0.001 NaF; 330 H2O 0.125 T-15-S-12; 0.062 H2SO4; 1 TEOS; 0.001 NaF; 330 H2O 0.125 T-15-S-12; 0.062 H2SO4; 1 TEOS; 0.001 NaF; 330 H2O 0.125 T-15-S-12; 0.062 H2SO4; 1 TEOS; 0.001 NaF; 330 H2O 0.125 T-15-S-12; 0.062 H2SO4; 2 TEOS; 330 H2O 0.125 T-15-S-12; 8 EtOH; 0.001 NaF; 330 H2O 0.125 PEG-600; 0.062 H2SO4; 1 TEOS; 330 H2O 2.5 PEG-600; 0.062 H2SO4; 1 TEOS; 330 H2O 6.25 PEG-600; 0.062 H2SO4; 1 TEOS; 330 H2O “aged” 0.125 T-15-S-12; 0.062 H2SO4; 1 TEOS; 330 H2O
0 1.2 1.75 2.75 3.75 4.4
N
I1/I3
IM/IE
Qb(ns)
98 94 90 85 90 92 94 95 95 96 80 81
1.24
2.53
168
1.24 1.24 1.24
2.40 2.38
159 159
1.24 1.24 1.24 1.80 1.78 1.79 1.23
2.25
The time t ) 0 corresponds to about 3 h after the addition of NaF to the system. The values of t are expressed in hours. b Q ) τE(IM/IE), and τE ≈ 66.5 ns. c System 3 aged for three months. a
2.2. Methods. 2.2.1. XRD. X-ray diffraction patterns were recorded using a Philips PW1800 diffractometer (copper anticathode radiation, λKR ) 1.54178 Å) in order to check the structure of the synthesized materials. Measurements were achieved for 2θ angle values in the 1-10° range. 2.2.2. Nitrogen Adsorption-Desorption. The measurements were performed on the calcined samples (see above) at 77 K with a Micromeritics ASAP 2010 apparatus. Prior to the measurement, the sample was degassed at 573 K overnight. The specific surface area was determined using the BET method in the 0.02-0.2 relative pressure range.24 2.2.3. DLS. The size of the surfactant aggregates in various systems was determined at 293 and 313 K before the condensation step by DLS measurements with a Malvern HPPS-ET instrument. DLS curves could be fitted according to the cumulants method (information given by Malvern) to a single exponential function indicating that the micelles are nearly monodispersed in size. Hydrodynamic diameters of micelles were calculated from the measured diffusion coefficient by using the Stokes-Einstein relationship with the appropriate values of the solvent viscosity (see below). 2.2.4. 29Si NMR. The spectra of selected silica-containing solutions were recorded on a Bruker DSX400 spectrometer operating at a Larmor frequency of 79.49 MHz with a pulse duration of 3.75 µs corresponding to a flip angle of π/4 and a recycle delay of 60 s using a Bruker 7 mm MAS probe in static conditions. 2.2.5. Fluorescence Probing. Pyrene and dipyrenylpropane (DPyP) were used as fluorescent probes and cetylpyridinium chloride (CPC) as a quencher of the pyrene fluorescence. Timeresolved fluorescence quenching (TRFQ) experiments were carried out on a single photon counting apparatus operating at 335 nm for the excitation of the pyrene fluorescence and 381 nm for the fluorescence emission with and without CPC. The micelle aggregation numbers (N) were determined from the values of the parameters obtained from the fittings according to the theoretical expressions of the fluorescence intensity decay curves in the absence and presence of quencher (Figure 1).8,25-27 Curve 1 shows an example of probe fluorescence decay. When the quencher is added, the fluorescence decay curve is modified. Part 3 of the curve shows a rapid decrease of the intensity which corresponds to micelles containing a probe and a quencher while part 2 of the curve corresponds to micelles containing only a probe but no quencher (as curve 1). (24) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309. (25) Alargova, R. G.; Kochijashky, I. I.; Sierra, M. L.; Zana, R. Langmuir 1998, 14, 5412. (26) Alargova, R. G.; Kochijashky, I. I.; Zana, R. Langmuir 1998, 14, 1575. (27) Zana, R.; Frasch, J.; Soulard, M.; Lebeau, B.; Patarin, J. Langmuir 1999, 15, 2603.
Figure 1. Examples of fluorescence decay curves of pyrene in a tergitol micellar solution (Ctergitol ) 0.02 mol dm-3) without quencher (1) and with the quencher cetylpyridinium chloride (2 and 3). (Cquencher ≈ 2 × 10-4 mol dm-3, Cpyrene ≈ 2.5 µmol dm-3.) The fluorescence emission spectra were recorded using a Hitachi F4010 spectrofluorometer operating at an excitation wavelength of 335 nm. The emission spectrum of pyrene was used to determine the intensities of the first (I1) and the third (I3) vibronic peaks. The ratio I1/I3 gives a measure of the micelle micropolarity at the solubilization site of pyrene.18,19 The emission spectra of DPyP at an excitation wavelength of 346 nm were used to determine the value of the ratio IM/IE from the intensities of the monomer (IM at about 378 nm) and excimer (IE at about 485 nm) emissions. The product of this ratio by the lifetime of the DPyP excimer (τE, measured using the same single photon counting apparatus) is proportional to the micelle microviscosity at the solubilization site of DPyP (Figure 2).28-30 The solutions were prepared and degassed following a previously reported procedure.21 All the fluorescence measurements were performed at 313.0 ( 0.3 K.
3. Results and Discussion 3.1. Optimization of the Experimental Conditions for Fluorescence Measurements. The TRFQ measurements performed under our experimental conditions require that the investigated solutions remain clear during at least 4 h after addition of the sodium fluoride. This delay is necessary for the equilibration, degassing, and (28) Turley, W. D.; Offen, H. W. J. Phys. Chem. 1986, 90, 1967. (29) Turley, W. D.; Offen, H. W. J. Phys. Chem. 1985, 89, 2933. (30) Zana, R.; In, M.; Levy, H. Langmuir 1997, 13, 5552.
8926
Langmuir, Vol. 21, No. 19, 2005
Figure 2. Examples of fluorescence emission spectra of pyrene and dipyrenylpropane in a tergitol micellar solution (Ctergitol ) 0.02 mol dm-3, Cprobe ≈ 2.5 µmol dm-3).
Figure 3. Variation of the time tclear during which the reaction mixture remains limpid after addition of NaF with the [NaF]/ [Si] molar ratio.
regassing of the solution with oxygen-free nitrogen. The system 1 TEOS; 0.125 T-15-S-12; 0.062 HCl; 0.04 NaF; 330 H2O used by Boissie`re et al.10 was observed to remain limpid for a time tclear of only 20 min before the mesostructured material precipitated out. This led us to investigate the variation of tclear with the [NaF]/[Si] molar ratio. The results reported in Figure 3 show that as the [NaF]/[Si] molar ratio is decreased from 0.04 to 0.001, the value of tclear increases from 20 min to 9 h. The fluorescence experiments were therefore performed using the molar ratio [NaF]/[Si] ) 0.001. Another modification of the system imposed by the fluorescence quenching studies was the replacement of the hydrochloric acid previously used 10 by sulfuric acid. Indeed, preliminary experiments showed that HCl was partly removed from the system during the degassing of the solution performed prior to the fluorescence quenching experiments. The composition of the precursor system used in these fluorescence studies was the following: 1 TEOS; 0.125 T-15-S-12; 0.062 H2SO4; 0.001 NaF; 330 H2O. 3.2. Characterization of the Final Materials. The XRD patterns of the calcined samples obtained after synthesis at different [NaF]/[Si] molar ratios were all similar, irrespective of the [NaF]/[Si] molar ratio. As an example, the XRD pattern of the material obtained with [NaF]/[Si] ) 0.001 is given in Figure 4. Only one broad peak is observed at low angles, characteristic of a wormlike structure,4,31 and corresponding to a d-spacing of 4.62 nm. The N2 adsorption-desorption isotherms of the sample prepared with a molar ratio [NaF]/[Si] ) 0.001 are shown (31) Prouzet, E.; Pinnavaia, T. J. Angew. Chem., Int. Ed. 1997, 36, 516.
Lesaint et al.
Figure 4. Powder XRD pattern of the MSU-type materials prepared with H2SO4 and with a molar ratio of [NaF]/[Si] ) 0.001.
Figure 5. Nitrogen adsorption (circle)-desorption (square) isotherms of the MSU-type material prepared with a molar ratio of [NaF]/[Si] ) 0.001 (STP: standard temperature and pressure).
in Figure 5. According to the IUPAC classification,32 these isotherms are of type Ib. They present an increase of the adsorbed volume at low relative pressure (p/p0 < 0.25) followed by a plateau.33 The presence of a “knee” before the plateau (p/p0 ) 0.25) reveals the existence of large micropores (secondary micropores). The SBET surface area of the sample was found to be 815 m2/g. 3.3. Fluorescence Studies. The values of the surfactant aggregation number, N, and of the polarity ratio, I1/I3, were determined in a series of solutions, from the aqueous tergitol solution in the absence of any additive to the precursor system containing all the components giving rise to the mesostructured material. Besides, the value of N in the precursor system was determined as a function of time (t) until the mesostructured solid started to precipitate out. The results are listed in Table 1. For the aqueous T-15-S-12 solution at C ) 0.02 mol dm-3 (system 1), the measured aggregation number N ) 98 reveals that the micelles are probably very close to being spherical. Indeed this value is only slightly larger than that for the maximum spherical micelles, Nmax ) 83, formed by a surfactant with a pentadecyl alkyl chain, calculated using the values of the length (2.05 nm) and volume (0.432 nm3) of a pentadecyl chain obtained using Tanford’s expressions34 (Note that the assumption of a (32) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.; Rouque´rol, J.; Siemieniewska, T. Pure Appl. Chem. 1985, 57, 603. (33) Voegtlin, A. C.; Matijasic, A.; Patarin, J.; Sauerland, C.; Grillet, Y.; Huve, L. Microporous Mater. 1997, 10, 137. (34) Tanford, C. J. Phys. Chem. 1972, 76, 3020.
Formation of Mesoporous Silica
protrusion35 of the alkyl chain out of the micelle surface by about 0.1 nm would yield Nmax ) 96). The addition of sulfuric acid and fluoride (system 2) is seen to lead to a small decrease of N to 94, which is within the experimental error ((5%). The addition of TEOS at a molar concentration ratio [TEOS]/[tergitol] ) 8 in the presence of sulfuric acid (system 3) leads to a more significant decrease of N to 90. This effect is most likely due to the presence of ethanol arising from the hydrolysis of TEOS (4 ethanol molecules are formed upon hydrolysis of each TEOS molecule present in the system). To check this point TEOS was replaced by its equivalent of ethanol (system 4; the comparison of systems 3 and 5 shows that the introduction of NaF has no effect on N). The micelle aggregation number in this system is seen to be close to that for system 3. This point was further checked by investigating system 11 with the molar ratio [TEOS]/[tergitol] ) 16 and system 12 where TEOS is replaced by its ethanol equivalent, that is, with a molar ratio [ethanol]/[tergitol] ) 64. The results listed in Table 1 show that systems 11 and 12 are characterized by identical values of N and that this value is lower than for systems 3 and 4 which contain less ethanol. The decrease of micelle aggregation number in the presence of ethanol is well documented for other surfactants, and it has been discussed.36 Aging of the full precursor system 5 resulted in a small increase of the value of N (see systems 6 to 10). These values all remained consistent with a spherical shape of the micelles present in the system. Indeed, even after 4.4 h of aging (system 10), the aggregation number was found to be 96. It is interesting to note that system 10 contained a small amount of precipitated solid. XRD analyses showed this solid to be a MSU-type material. The above results thus indicate that under the experimental conditions used, the tergitol micelles undergo very little change in shape and size when the various compounds required for the formation of the mesostructured material are progressively introduced into the solution. They also change very little when the formation of a small amount of condensed material can be visually observed. Measurements were not performed at later stages of evolution of the system. Indeed as more and more solid material is formed it becomes impossible to calculate reliable values of the micelle aggregation number because unknown amounts of surfactant, probe, and quencher are likely to be bound to the solid material. The pyrene fluorescence emission spectra were recorded for the systems 1, and 3-5, that is, before occurrence of condensation, and systems 10-12. The value of the pyrene polarity ratio I1/I3 remained the same at 1.24 for all these systems. This result confirms that the value of I1/I3 is essentially determined by the nature of the tergitol micelles present in the systems and that these micelles undergo very little change when adding the various compounds that give rise to the mesostructured material. The pyrene molecules are probably located in the inner part of the hydrophilic corona of the micelles formed by the tergitol EO12 headgroups, partly embedded in the hydrophobic core (see Figure 6).37 The constancy of the I1/I3 ratio indicates that there is little, if any, penetration of silica in the corona. Besides, I1/I3 measurements on systems 13, 14, and 15, which are the equivalent of system 3 but where the tergitol was replaced by increasing amounts of poly(ethylene glycol) PEG-600, yielded values of 1.79 ( 0.01, irrespective of the amount of polymer (see (35) Aniansson, G. E. A. J. Phys. Chem. 1978, 82, 2805. (36) Zana, R. Adv. Colloid Interface Sci. 1995, 57, 1. (37) Zana, R.; Lianos, P.; Lang, J. J. Phys. Chem. 1985, 89, 41.
Langmuir, Vol. 21, No. 19, 2005 8927
Figure 6. Scheme of the localization of pyrene in the tergitol micelle.
Table 1). This value is the same within the experimental error as that measured in pure water at 313 K, i.e., 1.80. Indeed, all these systems are in fact relatively dilute aqueous solutions containing no micelles, and the solubilizates (siliceous species, sulfuric acid, and NaF) are all hydrophilic and thus do not interact with pyrene. The fluorescence emission spectra of dipyrenylpropane were recorded and the excimer lifetimes τE measured for the systems 1, 3, and 4. From the values of the IM/IE ratio and of τE, the values of the quantity Q ) τE(IM/IE), which is proportional to the microviscosity were calculated (see Table 1).28-30 The value of Q shows a rather small decrease from 168 to 159 ns in going from the pure tergitol solution (system 1) to the systems 3 and 4. It is noteworthy that the value of Q remains unchanged when the TEOS is replaced by its ethanol equivalent (systems 3 and 4). Taken all together the fluorescence probing results show that the presence of siliceous species has no influence or a very little effect on the properties of tergitol micelles. To check the temporal stability of system 3, a solution without catalyst (NaF) was prepared and kept in the dark for 3 months (system 16). After this delay, a small amount of precipitate was observed which revealed that part of the silica had condensed. XRD showed this solid to be a MSU-type material. The system was filtrated. The values of the I1/I3 and IM/IE ratios for the supernatant were found to be close to those for system 3 (see Table 1). Clearly, the aging of the solution had no effect on the micelle micropolarity and microviscosity. This result gives further support to the conclusion above that there is very little or no interaction between silica and the surfactant micelles up to the point where a small amount of condensed material is generated in the solution. 3.4. 1H-Decoupled 29Si NMR. The system 1 TEOS; 0.062 H2SO4; 330 water, and the systems 3 and 15 were studied by 1H-decoupled 29Si NMR. Three similar-looking spectra were obtained (see Figure 7). Although the signalto-noise ratio of these spectra is very low, two resonances at -91 and -102 ppm can be observed, corresponding to the Q2 [O2Si(OH) 2] and Q3 [O3Si(OH)] species, respectively. The amounts of Q2 and Q3 species are seen to be almost the same, in agreement with the results reported by Boissie`re et al.10 This indicates that the main part of the silica that results from the hydrolysis of TEOS at pH 2 is under the form of small unconnected oligomers. Consequently, before the condensation, the presence of PEG-600 or of tergitol micelles affects neither the hydrolysis of TEOS nor the formation of oligomeric species (step 2 of the synthesis). It is noteworthy that due to the time necessary to obtain a decent signal-to-noise ratio (24 h), it was not possible to follow the evolution of system 5 (surfactant in water, TEOS, sulfuric acid, and NaF) as a function of time, as precipitation occurred after 4-5 h.
8928
Langmuir, Vol. 21, No. 19, 2005
Lesaint et al.
Table 2. Values of the Cloud Temperature (TC) and Apparent Micelle Diameter (D) for Selected Systems and of the Viscosity (η) Used for Calculating D system
a
molar compositions
[silica]/[tergitol] molar ratio
T (K)
TC (K)
ηa (mP/m)
D (nm)
293 313 293 293 293 293 293 313 293 313 293 313 293 313
359.5
1.01 0.655 1.01 1.04 1.07 1.14 1.13 0.701 1.12 0.700 1.29 0.775 1.25 0.752
7.4 7.7 7.4 7.5 7.2 7.4 9.2 8.1 7.6 7.0 9.1 8.5 7.3 7.5
17
0.125 T-15-S-12; 0.062 H2SO4; 330 H2O
0
18 19 20 21 3
0.125 T-15-S-12; 0.062 HCl; 330 H2O 0.125 T-15-S-12; 0.062 HCl; 0.25 TEOS; 330 H2O 0.125 T-15-S-12; 0.062 HCl; 0.5 TEOS; 330 H2O 0.125 T-15-S-12; 0.062 HCl; 1 TEOS; 330 H2O 0.125 T-15-S-12; 0.062 H2SO4; 1 TEOS; 330 H2O
0 2 4 8 8
22
0.125 T-15-S-12; 0.062 H2SO4; 4 EtOH; 330 H2O
0
11
0.125 T-15-S-12; 0.062 H2SO4; 2 TEOS; 330 H2O
16
23
0.125 T-15-S-12; 0.062 H2SO4; 8 EtOH; 330 H2O
0
ndb ndb ndb ndb 366.5 366.0 ndb ndb
Viscosities were determined from surfactant-free systems using a capillary viscosimeter. b The term “nd” means “not determined”.
Figure 7. 1H-decoupled 29Si NMR spectra of three representative solutions.
3.5. Dynamic Light Scattering. Two series of measurements of apparent micelle diameter were performed by DLS. The word apparent is used because the data were not extrapolated to the cmc of tergitol, and the measured diameter may be affected by the intermicellar interactions. It is noteworthy that the temperatures at which the measurements were performed (293 and 313 K) were much lower (by over 40 K) than the clouding temperatures (TC) of the investigated systems, obtained by visual observation (see Table 2, systems 3, 17, and 22). Thus, it can be reasonably assumed that diffusion in the investigated systems was not affected by the critical fluctuations that occur at temperatures close to TC.38,39 The first series of measurements, performed at 293 K, allowed us to check that the substitution of hydrochloric acid by sulfuric acid leaves the micelle diameter unchanged (in Table 2 compare systems 17 and 18). When TEOS is introduced in the system, a slight increase is observed for the sulfuric acid-containing system (compare systems 3 and 17). These measurements also showed that the micelle diameter remained constant when the molar concentration ratio [Si]/[tergitol] was increased from 0 to 8 for the hydrochloric acid-containing systems (Table 2, systems 18 to 21) and from 8 to 16 for the sulfuric acid-containing systems (Table 2, systems 3 and 11). Note that the calculations of the micelle diameter from the micelle diffusion coefficient determined by DLS used the values of the viscosity of the medium surrounding the tergitol micelles, that is, an aqueous solution of sulfuric or hydrochloric acid, with or without TEOS or its equivalent (38) Corti, M.; Degiorgio, V. J. Phys. Chem. 1981, 85, 1442. (39) Corti, M.; Degiorgio, V. Phys. Rev. Lett. 1980, 45, 1085.
in ethanol, depending on the investigated system. The viscosity values used in the calculation of the micelle diameter from the micelle diffusion coefficient measured by DLS are reported in Table 2. The second series of DLS measurements was performed at 313 K specifically for comparing for systems 3 and 11 the DLS and TRFQ results and for checking the effect of the substitution of TEOS by its equivalent of ethanol on the micelle diameter (see systems 22 and 23, Table 2). The values of the aggregation number N ) 90 and 80, obtained by TRFQ for systems 3 and 11 (see Table 1), respectively, were used to calculate the diameter of tergitol micelles in these systems. The later was determined using the volume of the surfactant pentadecyl chain (0.43 nm3)34 and of the ethylene oxide unit (0.062 nm3).40 With the use of an average hydration number of 4 for the ethylene oxide unit,41-46 the values of the micelle diameter were calculated to be 7.6 and 7.4 nm for systems 3 and 11, respectively, as compared to 8.1 and 8.5 nm from DLS. The difference between the values of the diameter obtained from two such drastically differing methods is rather small. At this stage it is recalled that DLS measurements at a finite concentration only yield an apparent diffusion coefficient that depends on the micelle size and also on the intermicellar interactions.47 One cannot exclude that the intermicellar interactions are slightly modified in the presence of siliceous species. However, according to our results, it seems that there is no effect on the micelle properties. 4. Conclusions The mechanism of formation of the MSU-type siliceous material in the presence of tergitol T-15-S-12 surfactant was investigated before the precipitation of silica by fluorescence measurements, 29Si in situ NMR, and DLS measurements. In our synthesis conditions, it appears that there is little or no interaction between silica and the tergitol micelles in the early stages of formation of the (40) Lepori, L.; Mollica, V. J. Polym. Sci. Polym. Phys. Ed. 1978, 16, 1123. (41) Matsuura, H.; Fukuhara, K. Bull. Chem. Soc. Jpn. 1986, 59, 763. (42) Paradies, H. H. J. Phys. Chem. 1980, 84, 599. (43) Garti, N.; Aserin, A.; Ezrahi, S.; Tiunova, I.; Berkovic, G. J. Colloid Interface Sci. 1996, 178, 60. (44) Nilsson, G.; Lindman, B. J. Phys. Chem. 1983, 87, 4756. (45) Schefer, J.; McDaniel, R.; Schoenbom, B. P. J. Phys. Chem. 1988, 92, 729. (46) Carlstrom, G.; Halle, B. J. Chem. Soc., Faraday. Trans. 1989, 85, 1049. (47) Candau, S. J. In Surfactant Solutions: New Methods of Investigation; M. Deckker Inc: New York, 1987; Chapter 3.
Formation of Mesoporous Silica
MSU-type material, before the precipitation of a significant amount of condensed siliceous material. The results of these three different techniques agree well. In particular, the aggregation number N of the tergitol micelles varies very little during the evolution of the reaction mixture with time until the visual observation of a small amount of condensed silica. The small changes in the value of N measured by TRFQ have been shown to arise essentially from the ethanol formed by the hydrolysis of the TEOS. At this stage we recall that similarly to us Sattler and Hoffmann in a small angle neutron scattering study of the formation of mesoporous silica from Si(OCH2CH2OH)4 in the presence of tetradecyltrimethylammonium bromide observed no change of micelle size till the gelation of the system occurred.48 Thus, as in the case of the generation of mesostructured MCM-41 silica in the presence of cetyltrimethylammonium bromide using sodium silicate as reactant, the present study shows that the formation of the mesostructured material does not take place at the interface of the micelles initially presents. The micelles do not bind the precursor material on their surface as this would have resulted in significant changes of micellar properties. The same conclusion as developed in our previous study8 and, more recently, by Ne et al.49 concerning the role of micelles in the synthesis of these mesostructured materials may also apply in this case: the micelles could simply act as reservoir of surfactant and provide monomeric surfactant that is adsorbed in increasing amounts and increasingly cooperatively (i.e., under the form of aggregates) by the growing siliceous polymers (condensed material), as (48) Sattler, K.; Hoffmann, H. Prog. Colloid Polym. Sci. 1999, 112, 40. (49) Ne, F.; Testard, F.; Zemb, T.; Grillo, I. Langmuir 2003, 19, 8503.
Langmuir, Vol. 21, No. 19, 2005 8929
Figure 8. Scheme of the proposed mechanism of formation of the MSU-type mesoporous material prepared in fluoride medium.
polymerization proceeds in the system. At some stage of the condensation, when the siliceous polymers have grown large enough and have bound enough surfactant, the polymer-bound micelle-like aggregates undergo a change of shape from spherical to elongated. The polymer that binds such aggregates can be templated by the aggregates and start organizing into the mesotructured material. At some point, the polymer/surfactant complex becomes so large that it precipitates out. In this model organization occurs just before or during precipitation. The proposed mechanism is schematized in Figure 8. LA0509347
