A Mechanistic Study of the Formation of Mesoporous Structures from

Aug 16, 2007 - Anionic surfactant templated mesoporous silicas (AMSs). Lu Han , Shunai Che. Chem. Soc. Rev. 2013 42, 3740-3752 ...
1 downloads 0 Views 443KB Size
Download PDF
Langmuir 2007, 23, 9875-9881

9875

A Mechanistic Study of the Formation of Mesoporous Structures from in Situ AC Conductivity Measurements Alfonso E. Garcia-Bennett,*,† Ulrika Brohede,† Robert P. Hodgkins,‡ Niklas Hedin,‡ and Maria Strømme† Nanotechnology and Functional Materials, Department of Engineering Sciences, The A° ngstro¨m Laboratory, Uppsala UniVersity, Box 534, SE-751 21 Uppsala, Sweden, and Materials Chemistry Research Group, Department of Physical, Inorganic and Structural Chemistry, Arrhenius Laboratory, Stockholm UniVersity, SE-106 91 Stockholm, Sweden ReceiVed March 28, 2007. In Final Form: June 1, 2007 The purpose of this work is to study the kinetics of self-assembly in the formation mechanism of anionic templated mesoporous solids (AMS-n) during the first few seconds of the synthesis as well as to demonstrate the use of alternating ion current (AIC) conductivity measurements to follow the self-assembly in complex hybrid systems. The formation of different AMS-n caged-type mesostructures through the delayed addition of the silica source is demonstrated and explained in terms of the interaction between the co-structure-directing agent (CSDA) and the oppositely charged surfactant headgroup regions. Our findings, supported by transmission electron microscopy, 29Si magic angle spinning nuclear magnetic resonance (MAS NMR) spectroscopy, dynamic light scattering (DLS) measurements, and powder X-ray diffraction suggest that the interaction of the CSDA with the surfactant headgroup occurs within seconds after its addition to the synthesis gel leading to interaction between the polymerizing CSDAs and the oppositely charged micelle and to an increase in the micelle-CSDA aggregate size. Both DLS and AIC measurements agree that this process occurs within the first 1000 s after addition of the CSDA to the synthesis gel at room temperature. In addition to the mechanistic study it was found that the intermediate materials are comprised of a three-layer entity. Timedependent 29Si MAS NMR studies reveal that an organo-silica layer forms around the micelles prior to a condensed outer inorganic shell of silica.

Introduction Silica-based mesoporous materials1 have become a popular matrix for the incoporation and support of functional species. These include metallic nanoparticles, oxide catalysts, and funtional biorganic molecules such as proteins and enzymes.2 The diversity of structures and morphologies available offers control over textural properties that allows tailoring the function of the support with unprecedented precision. For example, in catalysis greater selectivity is expected in many reactions from the use of geometrically designed pores.3 The formation of new porous mesostructures (denoted AMS-n mesoporous materials) combining the use of anionic surfactants and alkoxysilane co-structure-directing agents (CSDA) was recently reported.4-6 The use of CSDAs in the self-assembly of the surfactant species through interactions allows the functionalization of the internal surface in one direct step. This study focuses on structures prepared using N-lauroyl glutamic acid (C12-GlutA) and aminopropyl triethoxysilane (APES) as the CSDA. Novel mesocaged (Fd3hm, AMS-8), tetragonal (P42/mnm, AMS-9), and cylindrical bicontinuous (Pn3hm, AMS-10) structures synthesized with this surfactant have been prepared and * To whom correspondence should be addressed. E-mail: alfonso.garcia@ angstrom.uu.se. † Uppsala University. ‡ Stockholm University. (1) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710. (2) Schu¨th, F.; Schmidt, W. AdV. Mater. 2002, 14, 629. (3) Taguchi, A.; Schu¨th, F. Microporous Mesoporous Mater. 2005, 77, 1. (4) Che, S.; Garcia-Bennett, A. E.; Yokoi, T.; Sakamoto, K.; Kunieda, H.; Terasaki, O.; Tatsumi, T. Nat. Mater. 2003, 2, 801. (5) Garcia-Bennett, A. E.; Che, S.; Tatsumi, T.; Terasaki, O. Chem. Mater. 2004, 16, 813. (6) Garcia-Bennett, A. E.; Kupferschmidt, N.; Sakamoto, Y.; Che, S.; Terasaki, O. Angew. Chem., Int. Ed. 2005, 44, 2.

characterized extensively.7,8 Recently, we have discovered the possibility of controlling the final structural charactersitics of the mesoporous silicates produced by delaying the addition of the silica source, (tetraethyl orthosilicate, TEOS) with respect to the CSDA. It is clear from previous studies on the family of N-acyl amino acid surfactants that these amphiphiles are extremely sensitive to temperature and pH, thermal history, and the presence of electrolytes.9,10 The chemical structure of anionic surfactant, N-lauroyl glutamic acid, and CSDA, aminopropyl triethoxysilane, are schematically represented in Scheme 1. Mesoporous solid AMS-10 is of particular interest as the last remaining bicontinuous cubic structure to be replicated in silica form. Analogous to the D periodic minimal surface, it was prepared through the neutralization of the surfactant headgroup by addition of sodium hydroxide to the synthesis gel. A phase transition from tetragonal P42/mnm to cubic Fd3hm to twodimensional (2D) hexagonal p6mm and finally to the novel Pn3hm (AMS-10) has been observed8 and explained in terms of a decrease in the effective headgroup area of surfactant molecules, within the micelle, and with increasing neutralization through addition of sodium hydroxide. Such micelle curvature transitions have been described previously in the surfactant literature.11,12 In our previous work, the mesocaged tetragonal structure AMS-9 was prepared through variations in the time of addition of the silica (7) Garcia-Bennett, A. E.; Miyasaka, K.; Che, S.; Terasaki, O. Chem. Mater. 2004, 16, 3597. (8) Gao, C.; Sakamoto, Y.; Sakamoto, K.; Terasaki, O.; Che, S. Angew. Chem., Int. Ed. 2005, 45, 26. (9) Sakamoto, K. In Protein-Based Surfactants; Xia, J., Nnanna, I. A., Eds.; Surfactant Science Series. Vol. 101; Dekker: New York, 2001; Chapter 10, pp 261-280 (10) Mhaskar, S. Y.; N. Prasad, R. B.; Lakshminarayana, G. J. Am. Oil Chem. Soc. 1990, 12, 1215. (11) Lord, E. A.; Mackay, A. L. Curr. Sci. 2003, 85, 346. (12) Zeng, X.; Ungar, G.; Liu, Y.; Percec, V.; Dulcey, A. E.; Hobbs, J. K. Nature 2004, 428, 157.

10.1021/la700899s CCC: $37.00 © 2007 American Chemical Society Published on Web 08/16/2007

9876 Langmuir, Vol. 23, No. 19, 2007 Scheme 1. Schematic Representation of the Formation of Charged Species During the Synthesis of AMS-n Type Mesoporous Solids: (a) Deprotonation of N-Lauroyl Glutamic Acid upon Addition of APES, (b) Hydrolysis of Alkoxide Groups in APES, (c) Neutralization of Surfactant Headgroup through Interaction and Condensation of Silanol Groups

source with respect to the CSDA. A transition between cubic Fd3hm, cubic Pm3hn, and tetragonal P42/mnm mesocaged structures was found finally returning to the Fd3hm structure. In addition, a further structure containing extensive modulations was also identified, possibly containing variations of all three caged structures.7 It is unclear how such a transition occurs although an increased micellar curvature is implied from the phase transformation.13 In order to understand the interaction between surfactant headgroup, CSDA, and the growing inorganic silica wall in more detail, we have performed in situ conductivity measurements during the synthesis of AMS-n materials using C12-GlutA as surfactant and APES as CSDA, under synthesis conditions similar to those previously reported for the preparation of AMS-9 and AMS-8. Conductivity measurements are typically used for the study of amphiphilic systems. The electric current conducted by a liquid containing ions provides a direct measure of the number of charge carriers (ions) present in the liquid. Below the critical micellar concentration (CMC), the addition of surfactant to an aqueous solution causes an increase in the number of charge carriers and, consequently, an increase in the conductivity. Above the CMC, further addition of surfactant increases the micelle concentration, whereas the monomer concentration remains approximately constant. Since a micelle is much larger than its monomer surfactant ions, it diffuses more slowly through solution and is, hence, a less efficient charge carrier. A plot of conductivity against surfactant concentration is thus expected to show a sudden (13) Israelachvili, J. Intermolecular and Surface Forces; Academic Press: London, 2004.

Garcia-Bennett et al.

change of slope at the CMC. Surfactant N-lauroyl glutamic acid, will behave as an electrolyte over a wide pH range, obeying Ohm’s law. The two pKa values for the R- and γ-acid groups in the headgroup are aproximately 2.10 and 4.07, respectively. The CSDA, aminopropyl triethoxysilane, has been added to the synthesis in order to increase the interaction between the anionic surfactant and the growing inorganic wall. As a primary amine, APES has a pKa of 9.80, and hence a maximum interaction between headgroup and the protonated amine moiety will occur below this pH value. Although the primary ammonium cation may behave as an electrolyte, partial neutralization with the oppositly charged headgroup will occur. The ethoxy groups of the APES undergo hydrolysis, leading to negatively charged species, and hence contribute to the number of charge carriers. As it is well-known however, the reaction will involve hydrolyzed species of mono-, di-, or trihydroxyl forms, leading to polymeric chains (referred to as condensation).14 When the CSDA starts to polymerize and interact with the micelles, one expects behavior similar to that in polyelectrolyte-surfactant complexes. The formation of such may have a strong entropic free energy contribution.15 The overall reaction that occurs is highly dependent on variables such as the availability of water, the chemical functionality of the silane, and the pH of the original solution.16 Under the conditions employed for the synthesis of AMS-n materials (that is, high water content, a pH of 9.40, and temperatures between room temperature and 80 °C), the hydrolysis of the APES is expected to be slow, and condensation to oligomeric species, fast. Substituted silanes have been shown on the basis of 1H NMR spectroscopy to have slower rates of hydrolysis than tetraethyl orthosilicate (TEOS), a possible explanation based on the hindrance of the C-Si group toward nucleophilic attack.17 Upon the addition of TEOS to the micellarCDSA assembly in the synthesis of AMS-n mesoporous materials, the kintetics of alkoxide hydrolysis will be dominated by the hydrolysis of TEOS, which should produce charge carriers contributing to the electrolyte conductivity. Due to the complex mechanism involved in the synthesis of mesoporous materials in general and the various processes that must occur for the cooperative self-assembly of the surfactant template and the silica species, few in situ studies have been conducted. It is clear from the overall literature that there is not one definitive mechanism for the formation mechanism of mesoporous materials, since the synthesis conditions may vary from one route to another, as highlighted in a recent review.18 The formation mechanism of mesoporous MCM-41 and SBA15 have, however, been studied in situ by electron spin resonance and 1H NMR spectroscopies.19-21 In the former, it was concluded that the formation of the hexagonal (p6mm) mesostructure occurs in two stages: a rapid stage lasting for a few minutes, depending on the temperature where the formation of hexagonal ordered surfactant domains encapsulated by silica oligomers occurs, and a slow condensation stage of a few hours where crosslinking occurs and the formation of the inorganic wall can be discerned. In the case of the large-pore MCM-41 analogue SBA-15, the (14) Iler, R. K. The Chemistry of Silica; Wiley: New York, 1979. (15) Michael, A. S. Ind. Eng. Chem. 1965, 57, 32. (16) Raman, N. K.; Anderson, M. T.; Brinker, C. J. Chem. Mater. 1996, 8, 1682. (17) Icopini, G. A.; Brantley, S. L.; Heaney, P. J. Geochim. Cosmochim. Acta 2005, 69, 293. (18) Berggren, A.; Palmqvist, A. E. C.; Holmberg, K. Soft Matter 2005, 1, 219. (19) Ottaviani, M. F.; Galarneau, A.; Displatier-Giscard, D.; Di Renzo, F.; Fajula, F. Langmuir 2006, 22, 1493. (20) Zhang, J.; Luz, Z.; Goldfarb, D. J. Phys. Chem. B 1997, 101, 7087. (21) Firouzi, A.; Jumar, D.; Bull, L. M.; Besier, T.; Huo, Q.; Walker, S. A.; Zasadzinski, J. A.; Glinka, C.; Nicol, J.; Margolese, D.; Stucky, G. D.; Chmelka, B. F. Science 1995, 267, 1138.

Mechanistic Study of Mesoporous Structures

Langmuir, Vol. 23, No. 19, 2007 9877 few minutes before addition of the CSDA (APES, 0.37 g). The electric current I conducted by a liquid containing singly charged ions can be expressed as23 U j 2 I ) 2XNAFµ L

Figure 1. Digital photograph of the alternating ion current (AIC) measurement cell. Mesoporous materials were directly prepared in the measurement cell, which consists of a cubic cavity (4 × 4 × 4 cm) in a poly(vinyl carbonate) (PVC) unit, with stainless steel electrodes mounted on two of its sides. A lid covers the cell in order to prevent evaporation during measurements. The cell is situated on top of a magnetic stirrer and placed inside an oven to allow temperature control.

preparation conditions are considerably different since a higher concentration of surfactant is utilized and the synthesis is conducted under acidic conditions. The study concludes that the polymerization of the silica leads to attractive forces between micelles, thus forming precipitated micellar flocs. In situ microcalorimetry studies have also been performed in order to follow the restructuring of silica-surfactant composites under hydrothermal conditions, concluding that restructuring of phases (from hexagonal to lamellar) is entropically driven.22 The aim of the present work is to employ an in situ alternating ionic current (AIC) technique (originally developed to assess the functionality of drug delivery systems) for a detailed analysis of interactions between surfactant headgroup, CSDA, and silica source under similar conditions as previously reported for the preparation of AMS-8 and AMS-9 structures. More specifically, and in combination with 29Si MAS NMR spectroscopy, X-ray diffraction (XRD), transmission electron microscopy (TEM), and dynamic light scattering (DLS), we derive information about the kinetics of the self-assembly mechanism taking place. Experimental Section The synthesis of mesoporous solid AMS-8 and related materials has been described elsewhere.6,7 For the purpose of our study, the synthesis was directly conducted in the AIC measurement cell, which consists of a cubic cavity (4 cm × 4 cm × 4 cm) in a poly(vinyl chloride) (PVC) unit with stainless steel electrodes mounted on two of its sides; see Figure 1. To prevent loss of solvents through evaporation, a polypropylene lid was placed above the cell. Reactants were added directly to the cell. A function generator (HP 3325A) applied an alternating voltage (1 Vrms, 10 kHz) to the electrodes of the cell, and the conductance through the cell was calculated from measurements of the voltage across the cell and the current that passed through it using two digital multimeters (Agilent 34401A). All reagents were obtained from Sigma-Aldrich and used as received. Anionic N-lauroyl amino acid-derived surfactants were obtained from Nanologica AB (Uppsala, Sweden). In a typical experiment 0.20 g of the surfactant was mixed at temperatures between room temperature and 60 °C, in 20 g of distilled water. Preliminary conductance measurements determined that complete dissolution of the surfactant was achieved after a period of approximately 5 h or until no deviation in conductivity was observed. After dissolution, the AIC measurements were typically restarted a (22) Gross, A. F.; Yang, S.; Navrotsky, A.; Tolbert, S. H. J. Phys. Chem. B 2003, 107, 2709.

(1)

where NA is the Avogadro constant, F is the Faraday constant, X denotes the amount (number of moles) of ions in solution, µ j is the average ion mobility, U is the applied potential, and L is the length of the conduction path, i.e. the distance between the electrodes. After addition of the CSDA, measurements were either continued without addition of TEOS (1.5 g) or addition of the silica source after the required period of time. The final molar composition in the synthesis gel was C12GlutA/H2O/APES/TEOS, 1:1828:2.75:11.9, respectively. Measurements were continued for 24 h prior to hydrothermal treatment at 100 °C for 48 h. The remaining solution was filtered and dried without washing. Samples were calcined in an oven at 550 °C in a flow of nitrogen and oxygen, in order to remove the surfactant. For comparison, measurements were conducted also with a non-silane analogue of APES, such as pentylamine. Interaction between anionic headgroup and the protonated amine moiety of APES will occur below a pH value of 9.80 (pKa of APES). The surfactant solution becomes distinctively transparent in the first few seconds after addition of the APES, presumably due to the higher solubility incurred from the extra negative charge. The pKa1 of Si(OH)4 is 9.8 at pH ) 10.0.16 XRD patterns were recorded on a Philips PANalytical X’pert Pro powder diffractometer equipped with Cu radiation source (45 kV, 40 mA) at the rate of 1.0 deg/min over the range of 0.8-6.0°. TEM was conducted with a JEOL-3010 microscope, operating at 300 kV (Cs 0.6 mm, resolution 1.7 Å). Images were recorded using a CCD camera (model Keen View, SIS analysis, size 1024 × 1024, pixel size 23.5 µm × 23.5 µm) at 30 000-100 000x magnification using low-dose conditions on as-synthesized samples. DLS experiments were conducted via a Zetasizer instrument (Nano ZS, Malvern Instruments). The scattering intensity autocorrelation decays were analyzed with the manufacturers program, DTS (Nano). The respective stoichiometric quantities of dissolved anionic surfactant and CSDA were mixed at the desired temperature and stirred until the suspension turned transparent. The mixture was rapidly pipetted into a cuvette (1-cm path length) and added to the DLS apparatus. Before the experiments were conducted, the chamber was primed to the desired temperature to enable probing of the early time dependency. This early dependency was found to be sensitive to temperature variations. The time delay between subsequent measurements was limited by the signal-to-noise ratio and set to 3 min. Experiments were conducted until phase separation occurred after which the mixture suddenly became opaque. As a micellar size distribution is known to be relatively narrow,24 we applied a rather small but constant regularization parameter in the data analysis. Refractive indices and temperature-dependent solvent viscosities were used to transfer the autocorrelation time constant distributions into hydrodynamic radii via the Stokes-Einstein relationship. The scattering intensity was employed in the calculation of the particle size distribution; no scaling to the volumetric distribution was applied. Quantitative 29Si NMR experiments were carried out using a Chemagnetics Infinity 400 spectrometer equipped with a 9.4 T widebore magnet operating at 79.50 MHz. A 6-mm double resonance probe head was used to spin the sample at a magic angle spinning (MAS) rate of 8.5 kHz. The 29Si NMR chemical shift scale was externally referenced with tetramethylsilane (TMS), and the integral intensities were estimated using Spinsight software. Special care was taken not to saturate the spectra. A 4.8-µs excitation pulse corresponding to a 70° pulse was used. About 250 transients were added in a blocked manner. The recycling delay was held to 1000 s, more than 5T1, to avoid spectral distortion from uneven saturation (23) Frenning, G.; Ek, R.; Strømme, M. J. Pharm. Sci. 2002, 91, 776. (24) Safran, S. A. Statistical Thermodynamics of Surfaces, Interfaces, and Membranes; Addison-Wesley: Reading, MA, 1994.

9878 Langmuir, Vol. 23, No. 19, 2007

Figure 2. Dynamic light scattering (DLS) profile of C12GlutA/ H2O/APES at room temperature. The addition of the APES costructure-directing agent at t ) 0 s can be clearly seen to cause an increase in micellar size. The increase in micellar size is over within 3000 s after addition of the APES. At about 10 000 s the organo-silica/surfactant mesophase precipitates, and the solution becomes opaque. patterns. The large fraction of organic moieties in these hybrid materials, the broad 29Si NMR lines together with a low natural abundance and long T1 of the 29Si nuclei prohibited in situ studies of the silica condensation rates via 29Si NMR spectroscopy. As in situ measurements are not possible due to the obstacles presented above, we captured the time-dependent changes in the silica chemistry in the following way: Selected samples were taken out from the reaction medium at 20 and 40 min and 6 and 23 h intervals, after addition of the CSDA and precipitation of the (organo)-silica/ surfactant composites. The wet solid phase was separated from the liquid phase through centrifugation for 5 min and the liquid decanted. The concentrated precipitated phase was rapidly frozen in liquid nitrogen and subsequently freeze-dried to halt ongoing silica chemistry.

Results and Discussion The surfactant concentration utilized for the synthesis of AMS-8 is above the CMC, as determined previously.9 The formation of mesocaged structures during these syntheses suggests the presence of spherical micelles in the initial synthesis medium. This has been confirmed, for the first time, by DLS experiments. Figure 2 shows the changes in hydrodynamic diameter over the first 10 000 s after addition of APES (t ) 0 s) in the system C12GlutA/H2O/APES at room temperature. A rapid but small increase in micellar size is observed within the first 3000 s and is consistent, considering the hydrocarbon chain length and amino acid headgroup, with the formation of spherical micelles. In the mechanistically very different fluoride-assisted formation of MSU-type mesoporous silica, a similar behavior was found.25 The increase in size from 3 to 6 nm suggests an assembly between the polymerizing APES and the oppositely charged micelle. Interestingly, these APES micellar species are stable in size for a period of up to 10 000 s, after which aggregation and precipitation (not shown) of larger organo-silica/surfactant clusters is observed. At somewhat higher temperatures precipitation occurs rapidly after the CSDA addition. Conductance measurements show the surfactant dissolution to take approximately 5 h at 40 °C (see Supporting Information) for C12GlutA/H2O, observed as a stabilization of the conductance after this period. Figure 3 shows a typical plot of conductance versus time for the system C12GlutA/H2O/APES/TEOS at room temperature where the TEOS addition was delayed for 600 s with respect to the addition (25) Lesaint, C.; Lebeau, B.; Marichal, C.; Patarin, J.; Zana, R. Langmuir 2005, 21, 8923.

Garcia-Bennett et al.

Figure 3. Typical plot of conductance versus time for the system C12GlutA/H2O/APES/TEOS at RT. Initially the surfactant and distilled water are mixed until no changes in the conductance are observed. The time of addition of the APES is t ) 0 s. No other species are added to the solution for a period of 600 s at which point the necessary amount of TEOS is added.

of the APES (t ) 0 s). An immediate increase in the conductance is observed upon addition of APES due to the creation of charged moieties in the solution, reaching a maximum after t ) 65 s. From t ) 65 to 600 s there is a continuous decay in the conductance. This is consistent with the time taken for the increase in micellar size from 3 to 6 nm, observed by DLS. Mobility and, thus, also conductivity is expected to decrease with increasing micellar size, and a consumption of charged species is in addition expected from the condensation of hydrolyzed CSDA silanol groups. If no other species are added to the solution, the conductance rapidly decreases for the first 3 000 s after which it remains stable with no further change, as suggested by conductance plots where no TEOS was added to the remaining solution (see Supporting Information). When TEOS is added to the solution, at t ) 600 s in Figure 3, another sharp rise in conductance is observed. This effect is somewhat surprising at first sight due to the neutral uncharged nature of TEOS, slow hydrolysis under basic conditions and an initial low miscibility.14 Hydrolysis of the silica source may be catalyzed by the presence of condensed or partially condensed APES-micellar species. A rapid decrease in conductance thereafter indicates consumption of charges through the formation of larger more bulky silicasurfactant agglomerates. The rate of decrease in the conductance is faster after TEOS addition (>600 s) than that observed after the addition of the co-structure directing agent APES (t ) 65600 s). This must relate to a different consumption of charges associated with aggregate size changes. DLS measurements complement and extend the conductivity data with regards to the growth of micelles on addition of the CSDA as a function of time. This is consistent with the formation of an organo-silica layer around the micelles still in solution. Solid-state NMR spectroscopy studies show this to be the case at earlier stages after TEOS addition. The same technique shows that an inorganic layer forms at much later stages, i.e. there are strong indications for a three-layer composite: an inner organic core, a middle organo-silica, and finally an outer inorganic silica shell. 29Si MAS NMR spectroscopy is a good tool to differentiate between silica and organo-silica. The central inorganic Si atom is referred to Qn, depending on the degree of condensation in a pure inorganic SiOX environment. The central Si atom is termed Tn, depending on the degree of condensation in a hybrid organosilica composite. To further study the ratio of organo-silica to silica solid state 29Si MAS NMR spectroscopy was conducted. Synthesis gels with C12GlutA/H2O/APES/TEOS were quenched

Mechanistic Study of Mesoporous Structures

Langmuir, Vol. 23, No. 19, 2007 9879

and freeze-dried at different times after addition of the TEOS silica source. Figure 4 shows direct polarization (DP) NMR spectra quenched after 1200 s, 2400 s, 21 600 s (6 h) and 23 h. The organo-silica moieties are identified as T2 and T3 species from their chemical shifts at -60 to -70 ppm.26 The formation of Q4 peaks are associated with fully condensed inorganic silica species, at a chemical shift of -110 ppm, and the partially condensed Q3 inorganic silica have a 29Si NMR signature at a chemical shift of -100 ppm. Within Figure 4 it is clearly discernible that the CSDA has condensed around the micelle before the onset of TEOS hydrolyzation/condensation. Figure 4a shows strongly condensed organo-silica species with approximately 60-70% T3 groups. Note the relative absence of any Qn groups, i.e. the TEOS has yet to hydrolyze/condense. Figure 4b-d shows how the relative concentration of the inorganic silica is increased with time in the surfactant/organo-silica hybrid material. In Figure 4 the increasing fraction of inorganic silica (Qn /(Qn + Tn)) most probably reflects a slow hydrolyzation and fast condensation under basic conditions. After 23 h, Figure 4d, the composite consists of a high fraction of condensed inorganic silica. Initial stages of the precipitated matter reveal that the organo-siloxane is highly condensed around the inner micelle, and only at later stages does the inorganic silica form the outer shell of the threelayer hybrid inorganic/organo-silica/surfactant final material. Characterization of quenched samples by XRD (See Supporting Information) reveals the presence of mesoscale order after merely 1200 s. In addition, a shift toward lowered 2θ values is also observed with a higher degree of condensed inorganic silica, as samples were quenched at a later stage. High-resolution TEM images were recorded on as-synthesized quenched samples; however, only samples quenched after 6 or 23 h showed ordered periodicity. It was not possible from either TEM or XRD to uniquely determine the mesostructure characteristics of quenched (freeze-dried) samples. Typical TEM images of as-synthesized

mesoporous silicate prepared by quenching the condesation 23 h after addition of TEOS are shown in the Supporting Information. Figure 5 shows structural characterization of mesoporous material obtained after filtration and drying for the system C12GlutA/H2O/APES/TEOS at room temperature and where the addition of the TEOS was delayed for 600 s. The sample shows typical mesoporous peaks between 1 and 5° in 2θ. From XRD alone it is not possible to identify the mesoporous structure; however, TEM observations, contrast patterns, and Fourier transform diffractograms are consistent with a cubic unit cell with space group Fd3hm, where aXRD ) 158.8 Å and aTEM ) 155.7 Å. This structure has been previously termed AMS-8. A typical SEM image is shown in Figure 5c, where particle agglomerates containing many facets are clearly observed. Conductivity measurements were also carried out at 60 °C. Figure 6 shows conductance plots versus time for synthesis prepared at 60 °C through addition of the TEOS at t ) 0, and 150 s. A full data set of these plots (from -1000 to 30 000 s) is included in the Supporting Information. It is clear that the conductance curves follow different paths during the initial stages of the synthesis. For the preparation where TEOS was added simultaneously (t ) 0 s) with the APES hydrolysis, which results in an increased amount of charge species, conductance is taken over by neutralization and condensation steps since the conductance curve decreases steeply only 80 s after the addition of the APES and TEOS. When the addition of TEOS is delayed for 150 s, the conductance during the first 150 s after addition of APES increases above that observed for the APES-TEOS system of experiment t ) 0 s. The XRD patterns of as-synthesized samples in the system C12Glut/H2O/APES/TEOS prepared at 60 °C after addition of the TEOS at t ) 0, 150, 300, and 800 s show typical mesoscale diffraction peaks (Figure 7). High-resolution TEM images recorded along main zone axes of some of these samples are shown in Figure 8. Both XRD and TEM evidence are consistent with the formation of caged-type cubic and tetragonal structures at 60 °C. Cubic mesoporous materials AMS-8 (Fd3hm, aTEM ) 130 Å) and AMS-2 (Pm3hn, aTEM ) 94.4 Å) are observed at t ) 0 s (Figure 8a) and t ) 300 s (Figure 8c), with a modulated AMS-8 mesostructure (aTEM ) 118.3 Å) observed at t ) 150 s (Figure 8b). When TEOS is added after 800 s, tetragonal AMS-9 (P42/mnm, aTEM ) 168.0 Å and cTEM ) 83.0 Å) is formed (Figure 8d). From a geometrical approach, an overall decrease in the surfactant packing parameter, g, must occur to explain the formation of well-defined spherical surfactant micelles. This is consistent with the wide observation of mesocaged structures such as AMS-8, AMS-2, and AMS-9 in the C12GlutA/H2O/ APES/TEOS system. In the case of AMS-10 mesoporous materials synthesized previously by Che et al.8 through the neutralization of the surfactant headgroup, the repulsion between the surfactant headgroups decreases and, hence, the g parameter increases. This would explain the micellar transition from spheres to cylinders. It is not possible at this stage to explain through the packing parameter alone the phase transitions observed within the C12GlutA/H2O/APES/TEOS system described here, since structures produced are templated by packing of spherical micelles in primitive, face-centered, and tetragonal arrangements with small curvature changes. Navrotsky et al. have summarized the formation of mesoporous materials,27 based partly on calorimetry data of the formation of hexagonal SBA-15 and MCM-41 structures, in two stages: (1)

(26) Ek, S.; Iiskola, E. I.; Niinisto, L.; Vaittinen, J.; Pakkanen, T. T.; Root, A. J. Phys. Chem. B 2004, 108, 11454.

(27) Trofymluk, O.; Levechenko, A. A.; Tolbert, S. H.; Navrotsky, A. Chem. Mater. 2005, 17, 3772.

Figure 4. 29Si MAS NMR spectra acquired on freeze-dried threelayer surfactant/organo-silica/silica solids. The samples were quenched at 1200 s, 2400 s, 6 h, and 23 h after TEOS addition. Initially, the spectra reveal that the organo-siloxane is highly condensed; only at the later stages the inorganic silica forms an outer shell of a threelayer hybrid material.

9880 Langmuir, Vol. 23, No. 19, 2007

Garcia-Bennett et al.

Figure 6. Typical conductivity plots for synthesis conducted at 60 °C through addition of the TEOS after t ) 0 s (lower curve) and t ) 150 s (upper curve). It is clear in the initial part of the experiment and after addition of the TEOS that the reaction follows different paths. The change in conductance is faster on simultaneous addition of the APES and TEOS. Mesostructures shown are consistent with cubic Pm3hn where aTEM ) 94.4 Å, obtained at t ) 0 s and with a modulated cubic Fd3hm where aTEM ) 118.3 Å, obtained at t ) 150 s.

Figure 5. Structural characterization of as-synthesized mesoporous AMS-n material prepared in the system C12GlutA/H2O/APES/TEOS at RT where the addition of TEOS has been delayed for t ) 600 s: (a) indexed XRD pattern consistent with a cubic unit cell where a ) 158.8 Å and (b) typical TEM image recorded along the [110] orientation of the cubic unit cell. Inset in (b) shows a Fourier transform diffractogram taken over a thin region of the image shown. XRD and TEM suggest the structure to be consistent with that of AMS-8 (Fd3hm). A typical SEM image of an as-synthesised sample is also shown in (c), showing highly faceted agglomerated particles.

mesoscopic self-assembly of the inorganic precursor and CSDA, followed by framework condensation and (2) structural rearrangement due to dehydroxylation and post-condensation reactions during aging (curing/hydrothermal treatment) and calcination. They conclude that at each of these stages the energetic and structure of the framework assembly differ. It is clear from our study here that the formation mechanism of mesoporous AMS-n materials is more complex and may involve at least one more step, including the interaction of the CSDA with the surfactant headgroup, neutralization of the headgroup charges, and hydrolysis and condensation of the silanol groups. Scheme 1 shows a representation of the initial interactions that may take place during the first few seconds of the self-assembly process

Figure 7. X-ray diffraction patterns of AMS-n as-synthesized samples prepared in the system C12GlutA/H2O/APES/TEOS through addition of TEOS after t ) 0, 60, 150, 300, and 800 s at 60 °C.

in the AMS-n system presented here. Neutralization of the surfactant headgroup, and possibly charge reversal of the formed interface, takes place preferentially to condensation since electrostatic interactions will favor repulsion between hydrolyzed silanol groups and surfactant headgroup. From our conductance plots and from DLS studies presented here it is possible to conclude that steps (a) and (b) in Scheme 1, occur within 65 and 50 s at room temperature and at 60 °C, respectively. At this stage and as a result of the neutralization and condensation stages, depicted in (c), structural rearrangement occurs only after approximately 1000 s or through the addition of TEOS, where a return to a mesoscopic assembly of the inorganic precursor with the surfactant-CSDA is observed, followed by condensation of rearrangement through curing as suggested previously by Navrotsky et al. It is difficult at this stage to determine precisely the relationship between conductivity and micellar “state” (shape and size) formation. The micellar packing differs very slightly between the different cubic mesostructures obtained here, and such phases are reminiscent of the intermediate cubic phases often found in cubic lipid systems.28 The precise shape and size

Mechanistic Study of Mesoporous Structures

Langmuir, Vol. 23, No. 19, 2007 9881

distinct types of micellar aggregates.6 Our future work will concentrate on modeling the transitions between the spherical micellar systems presented here, as well as the study of other mechanisms for the preparation of mesoporous solids using the AIC method described.

Conclusion

Figure 8. HRTEM images of AMS-n as-synthesized samples prepared in the system C12GlutA/H2O/APES/TEOS through addition of TEOS after t ) 0 s (a), 150 s (b), 300 s (c), and 800 s (d) at 60 °C. Images are consistent with cubic structures: (a) AMS-8 along [110], (b) AMS-8 along [100], (c) AMS-2 along [100], and (d) tetragonal AMS-9 along [001].

of micellar aggregates in such intermediate phases is still debated in the surfactant literature, as in the case of the cubic Fd3hm structure.29 In our system however, the neutralization of the surfactant headgroup by the CSDA obviously changes the dimensions of spheroid micellar aggregates and their packing, resulting in a lower overall conductance of the solution and a subsequent change in final mesoporous structure. The Fd3hm mesophase is generally accepted to be formed of a bimodal set of spherical micelles. In contrast, the Pm3hn mesophase is thought to be composed of bimodal micelles where one type of micelle is elongated or disclike.30 Early crystallographic work on the P42/mnm mesostructure hints that it may be composed of three (28) Fontell, K. Colloid Polym. Sci. 1990, 268, 264. (29) Delacroix, H.; Gulik-Krzywicki, T.; Seddon, J. M. J. Mol. Biol. 1996, 258, 88. (30) Anderson, M. W.; Egger, C. C.; Tiddy, G. J. T.; Casci, J. L.; Brakke, K. A. Angew. Chem. 2005, 44, 3243.

In conclusion, we have demonstrated the potential of AIC measurements in order to study the formation mechanism for mesoporous materials. In the case of materials prepared with C12GlutA in the context of AMS-n materials, there is clear evidence that structural changes occur upon addition of the CSDA within the first few seconds of the self-assembly process. This leads to a variety of at least three different phases, depending on the time of addition of TEOS, the synthesis temperature, and the pretreatment of the surfactant. A precise model is not possible through the evaluation of packing parameter alone; however, from structural analysis of final products we can conclude that the phase transition is governed by a reduction in g and an increase in micellar surface curvature. Furthermore, the three-layer hybrid inorganic/organo-silica/surfactant material forms from the organic core to the outer inorganic shell via an organo-silica layer. This potentially allows a programmable functionalization of these AMS-type materials as well as prediction of the final product for large-scale synthesis of mesoporous solids. Acknowledgment. We are grateful to Prof. Osamu Terasaki and Prof. Lennart Bergstro¨m (Stockholm University) for helpful discussions. We thank Prof. Andreas Fischer (Kungliga Tekniska Ho¨gskolan) for access to XRD facilities. The Swedish Research Council, The Knut and Alice Wallenberg Foundation, The Go¨ran Gustafsson Foundation, and The Swedish Foundation for Strategic Research are acknowledged for their financial support of this project. One of the authors (M.S.) is a Swedish Royal Academy of Sciences Research Fellow and thanks the Academy for their support. Supporting Information Available: Supporting conductance plots and structural characterization of freeze-dried (quenched) mesoporous as-synthesized samples based on SEM, TEM, and XRD data. This material is available free of charge via the Internet at http: //pubs.acs.org. LA700899S