Stresses at the Interface of Micro with Nano Nicholas Leventis,* ,† Sudhir Mulik, † Xiaojiang Wang, † Amala Dass, † Chariklia Sotiriou-Leventis, † and Hongbing Lu ‡ Department of Chemistry, UniVersity of MissourisRolla, Rolla, Missouri 65409, and Department of Mechanical and Aerospace Engineering, Oklahoma State UniVersity, Stillwater, Oklahoma 74078 Received June 2, 2007; E-mail: leventis@umr.edu Bicontinuous materials consisting of an open interconnected macroporous system with mesoporous walls create a wide range of possibilities for sorption, catalysis, drug delivery, electrochemical energy storage, and separations (filtration, HPLC). 1,2 In monoliths, macropores provide low hydraulic resistance, hence easy access to the bulk by pressure-driven flow, whereas the large mesoporous surface area is always a very short distance away from everywhere and is accessed quickly by diffusion. 3 Here we report that in dry form those materials have innate stresses built in at the interface of the two size regimes. Relief and stabilization against collapse is possible by coating all internal surfaces with polymer. Materials with ordered hexagonal mesopores can be made through a modified sol-gel process that involves a surfactant such as Pluronic P123 (a PEO 20 PPO 70 PEO 20 triblock copolymer) as a structure directing agent (template). That is, at the right concentra- tion, Pluronic P123 self-assembles in closely packed hexagonal stacks of cylindrical pillars, and subsequently a sol-gel step adds silica filling the voids around them. Post-gelation removal of Pluronic P123 leaves behind a sol-gel material perforated by a hexagonal arrangement of tubes (SBA-15 type materials). 4 The diameter of the tubes can be increased by using a swelling agent such as 1,3,5-trimethylbenzene (TMB), 4,5 but if the concentration of TMB is increased above a certain threshold, the result is a mesoporous cellular foam (MCF), namely, a bicontinuous macro- porous system with random mesoporous walls. 6 If the concentration of Pluronic P123 by itself or in combination with TMB is adjusted carefully, the result is a macroporous silica with walls consisting of regular (organized) mesopores. 7 For several applications, wet gels of these materials may need to be dried, and the question is whether macro- and mesopores shrink the same. If not, stresses are included at the interface of the two fractal levels, causing the inherent structural instability associated with a potential collapse or at the very least a dimensional change of the material upon drying. Addressing this question, however, is not trivial because it requires stabilization of the internal stresses in order to produce dry forms preserving the dimensions of wet gels. In other words, the bicontinuous porous system should become capable of resisting shrinkage during drying. This was accomplished by coating all internal surfaces with a diisocyanate-derived polymer. As it turns out, shrinkage is indeed different at the nano and the micro size regimes. For our purposes, bicontinuous macro/mesoporous silica was prepared by Nakanishi’s modification of Stucky’s method where mesoporous gels rather than precipitates are obtained by reducing the volume of the sol. 7 Typically, Pluronic P123 (P, 4 g) was dissolved in 1.0 M aqueous HNO 3 (12 g), and TBA (T, 0.4 g) was added at room temperature. After stirring for 30 min, samples were cooled to 0 °C, TMOS (5.15 g) was added, and stirring continued for another 30 min. The mixture was poured into molds, which were kept at 60 °C for gelation (110 min). Samples were aged for 5× the gelation time. Following Nakanishi’s notation, 7 the material is referred to as native MP4-T045. However, preparation of dry self-standing monoliths is not straightforward: Nakanishi’s pro- cedure that calls for drying at 60 °C under ambient pressure followed by burning Pluronic P123 off at 600 °C led to coarse powders. Structural collapse, with or without prior removal of Pluronic P123 (vide infra), occurs during drying rather than during calcination. Therefore, reasoning that surface tension forces on the skeletal framework by the residing vapor/liquid interface might be responsible for shrinkage, 8 we resorted to an aerogel-like workup strategy: after the template and swelling agent were removed from the wet gels by a Soxhlet extraction (CH 3 CN), pore filling solvents were exchanged with liquid CO 2 , which was taken out supercriti- cally. This approach did produce monoliths which, as opposed to typical silica aerogels, still shrink significantly (29%) relative to the dimensions of their wet gels. In order to “lock” the skeletal framework at the wet gel stage, immediately after Soxhlet extraction, wet gels were exposed to an aliphatic diisocyanate solution (Desmodur N3200 from Bayer) that reacts both with the surface -OH groups forming urethane and with adsorbed water forming amines; in turn, amines react with more diisocyanate from the pores, yielding tethers of polyurea. 9 Thus, the bulk density of dry monoliths increases by a factor of 2.05, but now they shrink only by 13% relative to the molds versus 29% of native samples. The diisocyanate-treated material is referred to as X-MP4-T045 (“X” for cross-linked). Figure 1 compares native and X-monoliths at the two size extremes. The presence of the small-angle XRD pattern suggests that the mesoporous systems of † University of MissourisRolla. ‡ Oklahoma State University. Figure 1. Left: Photographs of typical dry native (MP4-T045; Fbulk ) 0.367 ( 0.003 g cm -3 ; Fskeletal ) 1.935 ( 0.002 g cm -3 ) and diisocyanate- treated (X-MP4-T045; Fbulk ) 0.755 ( 0.017 g cm -3 ; Fskeletal ) 1.279 ( 0.001 g cm -3 ) bicontinuous meso/macroporous silica monoliths, made using Pluronic P123 as structure directing agent, 1,3,5-trimethylbenzene as swelling agent, and supercritical fluid removal of pore solvents. Both gels were prepared using 1.04 cm diameter molds, so size differences reflect different shrinkage upon drying. The diameter of the native samples is 0.743 ( 0.005 cm and of the X-samples it is 0.909 ( 0.007 cm. Right: Powder XRD patterns of samples as indicated. The X-sample shows a smaller diffraction angle, indicative of a larger spacing between ordered features. Published on Web 08/14/2007 10660 9 J. AM. CHEM. SOC. 2007, 129, 10660-10661 10.1021/ja074010c CCC: $37.00 © 2007 American Chemical Society