Three-Dimensional Opal-Like Silica Foams - Langmuir (ACS

May 6, 2006 - The fine-tuning of both the liquid foam's fraction and the bubble size allows a rational design over the macroscopic cell morphologies (...
19 downloads 10 Views 696KB Size
Langmuir 2006, 22, 5469-5475

5469

Three-Dimensional Opal-Like Silica Foams Florent Carn,*,† Hassan Saadaoui,† Pascal Masse´,† Serge Ravaine,† Beatriz Julian-Lopez,‡,# Cle´ment Sanchez,‡,# Herve´ Deleuze,§ Daniel R. Talham,⊥ and Re´nal Backov*,†,# Centre de Recherche Paul Pascal, UPR 8641-CNRS, 115 AVenue Albert Schweitzer, 33600 Pessac, France, Laboratoire de Chimie de la Matie` re Condense´ e, (UMR CNRS 7574), UniVersite´ Pierre et Marie Curie, 4, Place Jussieu, 75252, Paris, France, Laboratoire de Chimie Organique et Organome´ tallique, UMR 5802-CNRS, UniVersite´ Bordeaux 1, 351 Cours de la Libe´ ration, 33045 Talence Cedex, France, and Department of Chemistry, UniVersity of Florida, GainesVille, Florida 32611-7200 ReceiVed January 23, 2006. In Final Form: April 7, 2006 The synthesis of novel meso-/macroporous SiO2 monoliths by combining a nano-building-blocks-based approach with the confined geometry of a tailored air-liquid foam structure is described. The resulting macrostructure in which ordered close-packed colloidal silica nanoparticles constitute the monolith’s scaffolds very closely resembles the tailored periodic air-liquid foam template. The void spaces between adjacent particles create textural mesoporosity; therefore, the as-prepared silica networks are characterized by hierarchical porosity at the macroscopic and mesoscopic length scales. The fine-tuning of both the liquid foam’s fraction and the bubble size allows a rational design over the macroscopic cell morphologies (shape, Plateau border’s length, and width). Striking results of this approach are the weak shrinkage of the as-synthesized opal-like scaffolds during the thermally induced sintering process and, in contrast with previous studies, the formation of closed-cell structures. Particle organization and the foam film surface roughness are investigated by atomic force microscopy (AFM), showing the influence of the liquid flow, within the foams’ Plateau borders and films, on the final assemblies.

1. Introduction The synthesis of micro- and/or mesoporous inorganic materials is a competitive research area with an extensive range of synthetic strategies described in the literature.1,2 In addition to nanoscale design, the properties of these porous materials also rely on the way the solid is distributed at the macroscopic length scale.3 The shaping of macroscopic monoliths in the form of open-cell macrocellular networks with well-defined topology, morphology, and cell dimension is a crucial task to reach that strongly influences suitability toward potential applications in areas such as tissue engineering,4 thermal and/or acoustic insulation,3 chromatography,5 and heterogeneous catalysis.6 Materials with complex architectures can be prepared by different routes,7 including selfassembly,8,9 shape-directed assembly,10 or layer-by-layer assemblies.11 Among them, a promising approach to generate hierarchically organized materials with macropore diameters ranging from 5 to 600 µm involves the use of complex fluids such as biliquid12-14 or air-liquid foams.15-19 However, the * Corresponding author. Phone: 33 (0)5 56 84 56 30. Fax: 33 (0)5 56 84 56 00. E-mail: [email protected] (R.V.B); [email protected] (F.C.). † Centre de Recherche Paul Pascal. ‡ Universite ´ Pierre et Marie Curie. § Universite ´ Bordeaux 1. ⊥ University of Florida. # Authors belonging to the European “FAME” network of excellence (NoE). (1) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710. (2) Soler-Illia, G. J. A. A.; Sanchez, C.; Lebeau, B.; Patarin, J. Chem. ReV. 2002, 102, 4093. (3) Gibson, L. J.; Ashby, M. F. Cellular Solids: Structure and Properties; Cambridge University Press: Cambridge, U.K., 2001. (4) Freynman, T. M.; Yannas, I. V.; Gibson, L. J. Prog. Mater. Sci. 2000, 46, 273. (5) Siouffi, A.-M. J. Chromatogr., A 2003, 1000, 80. (6) Hlatky, G. G. Chem. ReV. 2000, 100, 1347. (7) Mann, S.; Burkett, S. L.; Davis, S. A.; Fowler, C. E.; Mendelson, N. H.; Sims, S. D.; Walsh, D.; Wilton, N. T. Chem. Mater. 1997, 9, 2300. (8) Xia, Y.; Gates, B.; Yin, Y.; Lu, Y. AdV. Mater. 2000, 12, 693. (9) Wong, S.; Kitaev, V.; Ozin, G. A. J. Am. Chem. Soc. 2003, 125, 15589.

rational control of the macropore architecture and connectivity associated with improved structural stability toward moisture, thermal treatment, and mechanical compression remains an experimental challenge. In the present work, we propose to approach these issues by achieving the very close transcription of a tailored air-liquid foam using a confined colloidal crystallization that leads to an opal-like silica foam structure. The first part of this paper is dedicated to the design of the macroscopic void space morphologies tuned by both the foam liquid fraction and the bubble size. Final morphologies and thermal behaviors of the monoliths are compared with materials synthesized via the more conventional alkoxide polymerization under acidic conditions.20 At the mesoscopic length scale, the influence of film drainage on trapped particles in this confined geometry is explored by atomic force microscopy (AFM) measurements, while the mesoporous texture is investigated by nitrogen adsorption-desorption experiments. Furthermore, we present some results concerning the effect of (10) (a) Velev, O. D.; Jede, T. A.; Lobo, R. F.; Lenhoff, A. M. Nature 1997, 389, 448. (b) Davis, S. A.; Breulmann, M.; Rhodes, K. H.; Zhang, B.; Mann, S. Chem. Mater. 2001, 13, 3218. (c) Smatt, J. H.; Schunk, S.; Linden, M. Chem. Mater. 2003, 15, 2354. (d) Moon, J. H.; Kim, S.; Yi, G. R.; Lee, Y. H.; Yang, S. M. Langmuir 2004, 20, 2033. (e) Kamp, U.; Kitaev, V.; Freinmann, G.; Ozin, G. A.; Mabury, S. A. AdV. Mater. 2005, 17, 438. (f) Yan, F.; Goedel, W. A. Angew. Chem., Int. Ed. 2005, 44, 2084. (g) Li, F.; Badel, X.; Linnros, J.; Wiley: J. B. J. Am. Chem. Soc. 2005, 127, 3268. (11) Masse, P.; Ravaine, S. Chem. Mater. 2005, 17, 4244. (12) Imhof, A.; Pine, D. J. AdV. Mater. 1998, 10, 697. (13) Binks, B. P. AdV. Mater. 2002, 14, 1824. (14) Carn, F.; Colin, A.; Achard, M. F.; Deleuze, H.; Birot, M.; Backov, R. J. Mater. Chem. 2004, 14, 1. (15) Chandrappa, G. T.; Steunou, N.; Livage, J. Nature 2002, 416, 702. (16) Maekawa, H.; Esquena, J.; Bishop, S.; Solans, C.; Chmelka, B. F. AdV. Mater. 2003, 15, 591. (17) Huerta, L.; Guillem, C.; LaTorre, J.; Beltran, A.; Beltran, D.; Amoros, P. Chem. Commun. 2003, 1448. (18) Carn, F.; Colin, A.; Achard, M. F.; Deleuze, H.; Saadi, Z.; Backov, R. AdV. Mater. 2004, 16, 140. (19) Walsh, D.; Kulak, A.; Aoki, K.; Ikoma, T.; Tanaka, J.; Mann, S. Angew. Chem., Int. Ed. 2004, 43, 6691. (20) Brinker, C. J.; Sherer, G. W. Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing; Academic Press: San Diego, CA, 1990.

10.1021/la060220b CCC: $33.50 © 2006 American Chemical Society Published on Web 05/06/2006

5470 Langmuir, Vol. 22, No. 12, 2006

Carn et al.

evaporation temperature, flocculation process (particle concentration or salt effect), and particle sizes on the induced mesoporosities. Finally, a particle assembly mechanism is discussed, and some results concerning the monolith behavior toward moisture are presented. 2. Experimental Section 2.1. Materials. Tetraethyl orthosilicate (TEOS, 98%), ammonia solution (28%), hydrochloric acid (37%), and absolute ethanol were purchased from VWR Prolabo; tetradecyltrimethylammonium bromide (TTAB, 98%) was purchased from Fluka Chemica; sodium chloride (99%) and a Ludox HS-40 colloidal silica (with an average particle size of 12 nm) 40 wt % suspension in water were purchased from Aldrich; perfluoro-compound FC-72 was purchased from Acrosorganics; and aluminum chloride hexahydrate (99%) was purchased from Avocado. All chemicals were used as received in their reagent grade and were used without further purification. 2.2. Methods. Silica particles (70 and 200 nm) were synthesized following the well-known Sto¨ber-Fink-Bohn method.21 A 100 mL portion of absolute ethanol and 7.5 mL of ammonia were introduced in a three-neck round flask of 250 mL equipped with a refrigerating system. The mixture was stirred at 300 rpm to homogenize and heated to 60 °C. Then, 3 mL of TEOS was added into the solution, and the reaction occurred during 24 h under permanent stirring. The 200 nm silica particles were synthesized through a similar procedure at room temperature. Foams were obtained by bubbling perfluorohexane-saturated nitrogen through a porous glass disk (porosity: 100-160 µm, 40100 µm, 16-40 µm, 10-16 µm) into the foaming solution. The reaction took place inside a Plexiglas column. To minimize the foam destabilization by the drainage (i.e., the flow of the liquid between the bubbles due to gravity), we wet the foam from above with the sol solution. A stationary regime was reached, and the amount of solution injected at the top of the foam compensated the amount of liquid evacuated at the bottom. This strategy, allows preparing foams with a homogeneous liquid fraction from the top to the bottom of the column with a good controllability. The liquid fraction (F) is the ratio of the volume of liquid present in the foam divided by the total volume of the foam. This technique allows one to tune the liquid fraction by varying the flux of sol at the foam’s top (Q). High flux induces a high liquid fraction. The foam’s liquid fraction can be checked using conductivity measurements. The metastable foams were taken off at the top of the column with a spatula and stocked into a beaker. After being allowed to dry for 3 days in air at room temperature, the resulting hybrid organicinorganic monolith-type materials were then thermally treated at 650 °C to remove the organic part. The temperature increase rate was 2 °C‚min-1 with a first plateau at 200 °C for 2 h. The cooling process was uncontrolled and driven by the oven inertia. 2.3. Characterization. SEM observations were performed with a JEOL JSM-840A scanning electron microscope operating at 10 kV. The specimens were carbon-coated prior to examination. Surface areas and pore characteristics at the mesoscale were obtained after the sample was treated at 473K in a vacuum with a Micromeritics ASAP 2010 employing the Brunauer-Emmett-Teller (BET) and the Barrett-Joyner-Halenda (BJH) theories. AFM was performed using a Nanoscope IIIa instrument (Veeco Instruments) in soft tapping mode using the electronic extender module, allowing simultaneous phase detection, amplitude, and height imaging. We used Si tips with a resonant frequency of about 280 kHz and scan rates of 0.2 Hz. The free oscillation amplitude, A0, of the oscillating cantilever was around 30 nm. SiO2 foam’s film (diameters between 2 and 3 mm) were pasted with epoxy on mica substrate freshly cleaved prior to experimentation with adhesive tape. All experiments were performed in air at room temperature.

3. Results and Discussion 3.1. Control of the Macropore Morphology and Sintering Process. The basic scheme of our synthetic procedure is as (21) Sto¨ber, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26, 62.

Figure 1. Monolith before (a) and after (b) thermal treatment (T ) 650 °C). (c) SEM visualization of the foam Plateau borders and films. (d) Focused SEM picture of a foam’s film region composed of 200 nm silica particles.

follows: First, we prepared the foaming solution using an homogeneous dispersion of concentrated silica colloid (40 wt %) with a cationic surfactant (TTAB) in an aqueous media at pH 9. In a second step, the foam is generated by using a continuous bubbling of perfluorohexane-saturated nitrogen through a porous glass disk. During the bubbling procedure, the foam liquid fraction is controlled by a continuous and constant wetting by forced drainage. After 10 h of drying under ambient conditions and free drainage, a macroporous hybrid material is obtained in a monolithic state (Figure 1a), while a further thermal treatment at 650 °C leads to a pure inorganic monolith (Figure 1b). In contrast to our previous approach,18 there is no lyophilization step in the synthesis. Figure 1a,b shows that there is low shrinkage endured by the hybrid scaffolds during the thermal treatment and the associated sintering effect. The present shrinkage is considerably reduced (around 7%) when compared with materials obtained via alkoxides as oxide precursors (around 23%).18 This expected behavior12 is attributed to the high diameters of the building blocks as well as their good organization within the foam scaffold. The monodispersity of the particle size allows better accommodation of the elementary nanoscale units and good organization within the foam scaffold (Figure 1c,d). Indeed, the foam-confined media promotes close-packing of the particles before applying any thermal treatment. A consequence is that there is only minor evolution of texture before and after thermal treatment, like in the case of aerogels, but no supercritical conditions are employed. This feature reduces cracking, leading to more robust entities of higher dimensions. The overall foam architecture can be described using three structural elements: the films, the Plateau borders (i.e., liquid channels formed at the meeting of three soap films), and the nodes (i.e., the region where the Plateau borders are joined fourfold). The continuous interconnected network formed by these three elements represents the confinement media for the particle assembly. We propose to tune the macroscopic cell morphologies (Plateau border’s length, width, and curvature) by adjusting the starting air-liquid foam liquid fractions (F) defined as the volume of liquid that constitutes the foam divided by the total foam volume. The liquid fraction was modified by adjusting the continuous flux (Q) of the initial solution at the top of the growing foam.

3D Opal-like Silica Foams

Langmuir, Vol. 22, No. 12, 2006 5471

Figure 2. SEM visualization of SiO2 macroporosity for three different sol flux values (Q): (a) Q ) 0 g‚s-1, (b) Q ) 0.07 g‚s-1, and (c) Q ) 0.15 g‚s-1. The scale bars represent 100 µm.

Figure 3. SEM visualization of SiO2 macroporosity for three apparatus pore sizes: (a) P ) 25 µm, (b) P ) 70 µm, and (c) P ) 130 µm. The scale bars represent 250 µm.

By modifying this drainage property, either polygonal (Figure 2a) or spherical (Figure 2b,c) SiO2 macroscopic cell morphologies can be reached. Beyond the inorganic scaffold macrocellular shapes, the Plateau border thicknesses can be modulated by varying the foam’s liquid fraction or imposed sol flux. The inorganic (thermally treated) Plateau border’s thickness is strongly related to the wetting strength applied at the top of the growing air-liquid foams (Figure 2a-c). To tune the average macroscopic cell sizes, that is, the Plateau border’s lengths, we varied the porous disk employed at the bottom of the column during the bubbling process (Figure 3). We can observe that the Plateau borders’ lengths, defined as the node-to-node distance, can be tuned from 60 µm (Figure 3a) to 200 µm (Figure 3c). Moreover, for the same cationic surfactant and the same concentration, previous approaches based on inorganic alkoxide precursor polymerization in acidic conditions leads to open-cell structures (Figure 4b), while closed-cell structures can be obtained when colloidal flocculation in basic solution (Figure 4a) is used. The difference observed in the structures is obviously related to the pH and the formation mechanism of the foam scaffold. In the synthesis via alkoxides, the acidic medium promotes a fast hydrolysis of the silica precursors,20 which condense within the foam’s continuous aqueous part. On the other hand, considering the synthesis with the SiO2 particles under basic conditions, the negatively charged nanoparticles interact with the cationic surfactant and, under specific conditions of concentration, a

cooperative surface interaction takes place, leading to nanoparticle rearrangement along the foam pattern. 3.2. Particle Assembly at the Mesoscopic Length Scale. In addition to design at the macroscopic length scale, a critical point is to evaluate how the nano-building blocks assemble at the mesoscopic length scale to understand and ultimately tune the macroscopic properties. Herein, the particle packing and associated structural aspects are investigated by AFM measurements, while the void space textures are explored by nitrogen adsorption-desorption experiments. We attempt to modify this mesoporosity by adjusting either the evaporation temperature, the effect of flocculation induced by concentration effects or flocculating agents, and the particle size (12, 70, 200 nm). Figure 5 shows typical AFM images of monodisperse silica spheres (diameter ≈ 12 nm) assembled in the foam film region. These experiments were performed on various isolated films (diameter ≈ 2-3 mm) stuck on mica substrate. A thick, mechanically strong film is required to avoid sample contamination by capillary effects, so we chose to work with films obtained from foams with the highest liquid fraction (Q ) 0.15 g‚s-1). The first observation is the heterogeneous character of the film surface, as seen in Figure 5. Regions of smooth surface (Figure 5d) coexist with rough areas (Figure 5a,b). Along with heterogeneous surface roughness, there is heterogeneity of the degree of organization. Indeed, in some regions, we observe the presence of channels with preferential orientation over a large length scale (Figure 5d) in addition to completely disorganized regions (Figure 5a). The

5472 Langmuir, Vol. 22, No. 12, 2006

Carn et al. Scheme 1. Bidimensional Pattern of a Fluid Flow Distribution Inside a Foam’s Filma

a

Figure 4. SEM visualization of SiO2 macroporosity for two synthetic approaches: (a) nano-building blocks at pH 8.5 and (b) alkoxide (TEOS) polymerization under strong acidic conditions.

observed surface roughness could be associated with either disorganized regions (Figure 5a) or areas for which two preferential orientations cross each other (Figure 5b; see white arrows). On the other hand, smooth regions (Figure 5d) are always associated with a one preferential orientation (Figure 5d; see black arrow). The transitions between the smooth and rough regions are quite sudden, as can be observed in Figure 5c. This abrupt variation in the film’s topology is certainly governed by the complex liquid flow present within the foam film. It has been shown that foam films with Plateau border lengths of 2.5 mm present particular patterns of flow inside the film.22 According to Carrier et al.,22 in some cases, the film periphery in contact with the Plateau borders is pinched, while the central part is swollen by a liquid flow from the upper node to the lower one. According to these AFM observations, the heterogeneous character of the film surface can be interpreted as a direct expression of the liquid flow within the air-liquid foam film before the inorganic particle organization has taken place. The smooth region should be situated along the main axis of the liquid flow (i.e., the central part of the film, Scheme 1) where the shear rate applied to the particles is high, while roughness may result from intermediate regions where the liquid is stagnant and the associated shear rate is weak (Figure 5a) and without any existing preferential orientation. When the particle organization is perturbed by the flow emerging from different directions, as suggested by Figure 5b, the roughness is still important, but this time it is associated with some degree of preferential orientation. A basic scheme that defines the flow dynamic within a foam’s film is proposed in Scheme 1. Beyond these topological investigations, it is important to characterize the nanoparticle arrangement and its impact on surface area. A two-dimensional (2D) Fourier transform spectrum of smooth areas of the foam allows quantification of the particle organization (Figure 6). Over a large surface area, the Fourier transform spectrum confirms the opal-like organization with a periodicity of 13 nm, (22) Carrier, V.; Destouesse, S.; Colin, A. Phys. ReV. E 2002, 65, 061404.

Mainly extracted from ref 22.

in agreement with the average particle size of 12 nm. This organization is due to the monodispersity of the particle size. As a direct consequence of this optimized particle packing, the monolith shrinkage induced by sintering is low (Figure 1a,b). On the other hand, if we aim to obtain textural porosity, it appears to be important to create a certain amount of defects by varying the flocculation processes. Overall, the flocculation of a colloidal suspension is an out-of-equilibrium phenomenon leading to flocs with a fractal structure for very dilute systems, while gel-like structures can be obtained for higher concentration.23 In the current study, the suspension concentration fixed at 40 wt % yields the best transcription of the metastable foam pattern to the robust monolithic state. Therefore, to modify the gel mesoporous texture, we studied the influence of several other parameters, such as evaporation temperature, flocculation process (salt effect), and particle sizes. First, nitrogen adsorption-desorption experiments were performed on thermally treated samples prepared with the same procedure starting with colloidal silica of 12 nm diameter and without further salt addition. For this first set of experiments, the evaporation temperature was monitored at 25 °C during 48 h in one case (Figure 7a) and 60 °C during 48 h in the other case (Figure 7b). The nitrogen adsorption-desorption isotherms (Figure 7a,b) depict a type IV behavior with an H1-type hysteresis according to IUPAC classification,24,25 indicating the presence of a well-defined mesoporosity with an average size of 7 nm considering the narrow pore size distribution obtained via the original density functional theory associated with the BJH method applied to the desorption curve. The strong similarity between panels a and b of Figure 7 reveals the low impact of water evaporation kinetics on the floc mesostructure. Moreover, this result establishes the reproducibility of the data. Second, the salt addition effect is revealed by Figure 7a,c,d. The isotherms presented in Figure 7c,d for materials obtained with colloidal silica of 12 nm flocculated via the addition of a monovalent salt (NaCl) and a trivalent salt (AlCl3), respectively, show a type IV behavior, indicating the presence of mesoporosity, but this time the pore distribution is broader, ranging from 4 to 20 nm according to the BJH result. This increase of pore size polydispersity is combined with a slight increase in the cumulative pore volume from 0.24 cm3‚g-1 for the sample flocculated without destabilizing agent to 0.31 cm3‚g-1 for the sample flocculated by salt (23) Gunes, D. Z.; Munch, J. P.; Dorget, M.; Knaebel, A.; Lequeux, F. J. Colloid Interface Sci. 2005, 286, 564. (24) 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. (25) Greg, S. J.; Sing, K. S. W. Adsorption, Surface Area and Porosity; Academic Press: New York, 1982.

3D Opal-like Silica Foams

Langmuir, Vol. 22, No. 12, 2006 5473

Figure 5. AFM three-dimensional (3D) or 2D images and section profiles of foam film surfaces, showing typical morphologies observed on the film surfaces of different foams.

Figure 6. AFM 2D images and the corresponding 2D Fourier transform spectrum performed on a smooth part of the foam film.

incorporation (Table 1). This expected result reveals the salt’s capability to introduce disorder in the well-arranged floc structure. It is worth noticing that no significant effect over the specific surface area is observed for monoliths obtained either with NaCl or without salt addition (SBET ≈ 150 m2‚g-1), while the addition of a trivalent salt induces a soft surface area increase up to 188 m‚g-1. Finally, using larger sizes of silica colloids (70 nm, Figure 7e, and 200 nm, Figure 7f) flocculated by concentration changes

without a destabilizing agent results in quasi non-mesoporous materials with type II behavior of the nitrogen isotherm. Overall, the BJH analysis of the nitrogen desorption curve indicates the presence of meso-/macropores in the range of 30150 nm. This elimination of the floc mesostructure leads to a very low specific surface area (SBET ≈ 22.5 m2‚g-1). Main parameters related to the mesoporosity of each silica scaffold synthesized in this study are proposed in Table 1.

5474 Langmuir, Vol. 22, No. 12, 2006

Carn et al.

Figure 7. Nitrogen adsorption-desorption experiments. On the right side, the pore size distribution as calculated from the original density functional theory model is presented, and the nitrogen adsorption-desorption isotherms are figured on the left side (× adsorption, O desorption): (a) flocculation of 12 nm particles without salt at room temperature, (b) flocculation of 12 nm particles without salt at 60 °C, (c) flocculation of 12 nm particles with NaCl at room temperature, (d) flocculation of 12 nm particles with AlCl3 at room temperature, (e) flocculation of 70 nm particles without salt at room temperature, and (f) flocculation of 200 nm particles without salt at room temperature. Table 1. Different Surface Areas Obtained by Nitrogen Adsorption-Desorption Measurements

Ludox (25 °C) Ludox (60 °C) Ludox + NaCl (25 °C) Ludox + AlCl3 (25 °C) 70 nm particle (25 °C) 200 nm particle (25 °C) a

BET surface area [m2‚g-1]

BJHa surface area [a] [m2‚g-1]

BJHa cumulative pore volume [m3‚g-1]

158 146 152 188 22 23

184 200 187 221 18 17

0.25 0.23 0.31 0.32 0.06 0.08

Calculated from the desorption curve.

In summary, AFM analysis reveals the heterogeneous character of the film surface roughness, where well-arranged smooth regions coexist with disordered rough parts. A correlation between surface morphology and liquid flow in the confined geometry of the foam pattern is proposed. Nitrogen adsorption-desorption measurements reveal the presence of textured mesoporosity, which emerges from interparticle void spaces. The mesopore

size distributions and the specific surface areas can be modified by varying both the flocculation process and the particle sizes. Furthermore, we show that particles of diameter greater than 70 nm favor a second macroporosity texture rather than mesopores, thus the specific mesoscopic surface area is drastically diminished. 3.3. Particle Aggregation Mechanism. The starting point of our approach is based on the use of a dynamic formulation (continuous wetting) using kinetically stable air-liquid foams (several minutes) with a continuous phase composed of a large amount of colloidal particles dispersed in water. This concentrated continuous liquid phase must exhibit a low viscosity, allowing both liquid flow within the foam (i.e., drainage) and bubble formation by gas injection. Moreover, the positioning of the particles in the confined geometry of the foam’s charged channel indicates cooperative interaction between the charged substrate and the charged particles. All these requirements can be achieved following different routes. Herein, we work with a concentrated silica colloidal suspension (40 wt %) negatively charged at pH 8.5. Considering this initial point, the choice of the surfactant charge and concentration are key parameters that govern the

3D Opal-like Silica Foams

stability of both the foam air-liquid interface and the colloidal dispersion. We choose a cationic surfactant (TTAB) at a concentration of 10 critical micelle concentration (CMC) to saturate the air-liquid interface with surfactant, thus stabilizing the foam toward coalescence. Also, the TTAB surfactant allows the promotion of the partial hydrophobic character of the silica particles, thus balancing both the attractive substrate-particle interactions and the repulsive particle-particle interactions that favor closed-packed particle assemblies. Similar experiments performed with less concentrated silica suspensions (10 and 20 wt %) do not allow the formation of a solid monolith after either evaporation or lyophilization, but result in the formation of a powder rather than monoliths, meaning that the foams collapse during the drying processes. In fact, the use of a less concentrated surfactant solution certainly promotes both the particle flocculation, induced by a depletion effect, and a decrease in the air-liquid interface stability. In the first stage of the bubbling process under forced drainage conditions, the interfacial membrane does not act as a rigid barrier and allows the regulation of both the volume liquid fraction and the quantity of up-coming nanoparticles. This feature allows control of the Plateau border’s width, as shown previously. In the second step, when the continuous wetting is stopped (i.e., free drainage condition), the rate of the liquid flow decreases significantly within the Plateau border regions, thus leading to a lower shear rate that favors the aggregation of colloidal particles and giving rise to an enhanced interfacial rigidity. As a result, particle flow blocking should appear in the node region, while water drainage remains with slower rate. The monodisperse silica particles are gradually concentrated and arranged into an ordered lattice. This soft and gradual aggregation process explains why, in contrast to a previous approach,18 no congelation/lyophilization procedure is necessary in this liquid-solid-state transition. 3.4. Behavior in the Presence of Water. Silica aerogels have the strong tendency to both deteriorate and collapse with time when exposed to ambient humidity because adsorbing water from air both physically and chemically enhances the condensation of residual silanol groups.26 As a consequence, in actual use for insulation, aerogels must either be encapsulated or have their surfaces passivated using hydrophobic tails to cap surface silanol sites.27 An additional strategy consists of mimicking the lotus effect by increasing the surface roughness to enhance the hydrophobic properties.28 We propose herein to compare the behavior of an open-cell structure with large surface area (∼1000 m2‚g-1) obtained via alkoxide (TEOS) polymerization in acidic conditions with that of a closed-cell network depicting low surface area (∼200 m‚g-1) obtained via the colloidal flocculation route. The experiment consists of immersing an inorganic monolith of several centimeters (Figure 1b) within a demineralized water (26) Koyano, K. A.; Tatsumi, T.; Tanaka, Y.; Nakata, S. J. Phys. Chem. B 1997, 101, 9436. (27) Miner, M. R.; Hosticka, B.; Norris, P. M. J. Non-Cryst. Solids 2004, 350, 285. (28) (a) Barthlott, W.; Neinhuis, C. Planta 1997, 202, 1. (b) Patankar, N. A. Langmuir 2004, 20, 8209.

Langmuir, Vol. 22, No. 12, 2006 5475

Figure 8. Nanotextured SiO2 monoliths after being immerged in water for 2 weeks. The scale bars represent 2.5 cm.

bath and wetting the sample from above with demineralized water. The monolith obtained via the alkoxide polymerization collapses completely in the first seconds of the experiment (without wetting), while the opal-like monolith (Figure 8) remains almost undamaged after 2 weeks under those conditions. In the first case, we can argue that the microstructure fractal organization leads to both a low scaffold cohesion and a large surface area. This surface provides a large amount of silanol sites favoring dissolution of the scaffold. Moreover, the simultaneous presence of continuous micro-, meso-, and open macroscopic void spaces drastically enhances the water diffusion by capillary rising. In the second case (Figure 8), the absence of micropores, according to Figure 7 and Table 1, combined with low mesoporosity leads to low surface area (∼150 m2‚g-1) and thus a low quantity of accessible silanol sites. The closedcell network remains an additional limitation for water diffusion into the foam’s scaffold. The curved foam film surface topography may also influence the hydrophobicity and play an additional role in the stability under aqueous conditions.

Conclusion We present an easy method for the fabrication of meso/ macroporous nanostructured silica foams based on the rearrangement of SiO2 nano-building blocks along the confined geometry of a tailored air-liquid foam structure acting as a macroscopic pattern. In contrast with previous approaches, the rational design of the air-liquid foam leads to better control over the macropores characteristics: topology (open- or closed-cell macrocellular networks), morphology (spherical or polygonal cell structure), and dimensions (Plateau border’s length and width). At the same time, the inhibition of cracks during the sintering process and moisture resistance are drastically enhanced. Moreover, a correlation between liquid flow and solid foam film surface roughness is presented. This feature suggests that drainage exists within the foam’s film and might play a key role to explain the roughness observed in final mineralized scaffolds. Acknowledgment. We thank Odile Babot (LCOO, Talence) for the nitrogen adsorption-desorption experiments. This work is part of the FAME-MIOH 6-PCRDT program. LA060220B