Fluorocarbon

Sep 21, 2010 - We report a study of the coassembly behavior of mixed hydrocarbon surfactant cetyltrimethylammonium chloride (CTAC) and the fluorocarbo...
0 downloads 0 Views 3MB Size
Download PDF
17390

J. Phys. Chem. C 2010, 114, 17390–17400

Demixed Micelle Morphology Control in Hydrocarbon/Fluorocarbon Cationic Surfactant Templating of Mesoporous Silica Rong Xing,†,‡ Hans-Joachim Lehmler,§ Barbara L. Knutson,† and Stephen E. Rankin*,† Chemical and Materials Engineering Department, UniVersity of Kentucky, 177 F.P. Anderson Tower, Lexington, Kentucky 40506-0046, and Department of Occupational and EnVironmental Health, UniVersity of Iowa, 221 IREH, Iowa City, Iowa 52242 ReceiVed: June 4, 2010; ReVised Manuscript ReceiVed: August 21, 2010

We report a study of the coassembly behavior of mixed hydrocarbon surfactant cetyltrimethylammonium chloride (CTAC) and the fluorocarbon surfactant 1H,1H,2H,2H-perfluorodecylpyridinium chloride (HFDePC) with precipitated silica at room temperature. This pair of surfactants is known to segregate into demixed fluorocarbon-rich and hydrocarbon-rich micelles in dilute solution for a range of compositions. Synthesis parameters (including the molar composition of the surfactant mixture, the ammonia concentration, and the addition of NaCl or ethanol to the initial sols) are varied to show how the mixed or demixed micelles found in dilute aqueous solutions act as templates to form mesoporous silica. Four distinct types of pore structures are found. The first are particles with a single mesopore size and structure whose morphology is influenced by the addition of a second surfactant. Adding HFDePC to CTAC for this series of samples induces a transition from 2D hexagonal mesopores to disordered mesopores to mesh phase pores. At intermediate HFDePC compositions, large voids (templated by HFDePC vesicles) are introduced. A second class of particles, with a biphasic 2D hexagonal/mesh phase structure, is formed when the amount of ammonia increases. The third class of particles, with separate fluorocarbon and hydrocarbon micelles combined in a single disordered phase, forms when a moderate amount of salt is added, most likely due to a reduction in the rates of coassembly and precipitation. Finally, a sample consisting of two types of particles with differing shape and pore structure is prepared by adding ethanol to enhance the rate of precipitation. The conditions explored here can be used for the design and synthesis of mesoporous silica with controlled pore structures and morphologies. 1. Introduction Fluorocarbon (FC) surfactants are of growing interest due to their unique properties compared with ordinary hydrocarbon (HC) surfactants.1,2 For example, in contrast to the relative flexibility of alkyl chains, fluoroalkyl chains prefer a stiff helical conformation because of their torsional potential.3 Also, due to the large electronegativity of fluorine, FC surfactants are characterized by very strong intramolecular C-F bonds and weak intermolecular interactions. These properties give rise to the well-known chemical and thermal stability, low friction, and nonstick properties of polymers such as poly(tetrafluoroethylene).4 Moreover, FC surfactants allow cosolubilization of different solvents with strongly opposed affinities, such as water and perfluoroalkanes or supercritical CO2.5,6 Recently, the use of pure FC surfactants as templates for mesoporous metal oxides has been reported to give products with features consistent with the self-assembly behavior of the surfactants.7-18 However, mixing together colloidal templates creates tremendous new opportunities to tune the size, shape, symmetry, and functionality of mesoporous materials. For instance, hierarchically organized porous metal oxides prepared from latex/surfactant mixtures have been developed for adsorption19,20 and separations.21 Mixing together miscible HC * To whom correspondence should be addressed. E-mail: srankin@ engr.uky.edu. Phone: 1-859-257-9799. † University of Kentucky. ‡ Current address: Department of Chemical Engineering, University of Massachusetts, 215 Goessmann Lab, 686 North Pleasant Street, Amherst, MA 01003-9303. § University of Iowa.

surfactant templates for mesoporous ceramics has also been used to tune pore sizes and wall thickness,22 to functionalize the pore surface,23 and to stabilize otherwise metastable structures.24,25 Mixtures of FC and HC surfactants have not been utilized as often, but they have significant potential for novel templating. Because FCs are not only severely hydrophobic but also lipophobic, FC and HC surfactants exhibit nonideal mixing, which often leads to micelle demixing. The triphasic nature of demixed surfactants in a polar matrix may lead to rich phase behavior analogous to triblock copolymers,26 or to completely demixed mesophases.27 Due to micelle demixing, at least four types of particles might be anticipated when a binary HC-FC surfactant system is used as the pore template, as illustrated in Figure 1. The pores of type I particles are templated primarily by one surfactant, while the second sits at the particle interface to control particle morphology. This particle type has previously been generated by using mixed FC and block polymer surfactant templating to produce silica with high hydrothermal stability28 and hierarchical pore systems.29,30 Examples of type I particles have been reported in which demixed layers of FC surfactants act either as hollow macropore templates31 or for particle morphology control.32 Type II particles show biphasic structures in which separate FC-rich and HC-rich domains form, but both are present within single particles. Type III particles contain a single phase composed of separate FC and HC micelles. The stiffness of bulky FC tails causes FC surfactants to prefer aggregates of low curvature (rods and discs) and a novel “intermediate” phase which can lead to as yet unexplored phase behavior in type III

10.1021/jp1051563  2010 American Chemical Society Published on Web 09/21/2010

Demixed Micelle Morphology Control

Figure 1. Illustration of possible arrangements of demixed hydrocarbon micelles (unfilled circles) and fluorocarbon micelles (filled circles) in precipitated particles: (I) primary pores templated by HC and particle structure dictated by FC at the surface, (II) biphasic domains of HC and FC micelles within one particle, (III) single intermixed FC and HC micelles constituting one mesophase, and (IV) HC and FC micelles present separately in two different types of particles.

particles.33 Type IV particles have FC-rich and HC-rich phases demixed into separate particles, which may form if the rates of precipitation of the two phases are very different. Despite the great variety possible in demixed micelle-containing particles, the formation of particles using a combination of cationic surfactants that are both capable of coassembling with silica into ordered phases has not yet been reported. Here, we discuss cases in which all four types of demixing occur in precipitated silica templated with mixed HC and FC surfactants. Different cases are realized by varying synthesis parameters, including the molar composition of mixed surfactants, the ammonia concentration, and the addition of NaCl or ethanol to the initial sols. Cetyltrimethylammonium chloride (CTAC) and 1H,1H,2H,2H-perfluorodecylpyridinium chloride (HFDePC) surfactants are employed as templates (see Supporting Information Figure S1). Both surfactants have similar values of critical micelle concentration (CMC) and the same counterion. This pair has been studied extensively in dilute solution and shown to demix into CTAC-rich micelles and HFDePC-rich micelles in a range of compositions.34-36 This is consistent with many reports of micelle demixing in dilute HC/FC surfactant mixtures, although the demixed HC/FC micelles in dilute solutions are often found to mix together in concentrated lyotropic liquid crystal phases.37,38 To the best of our knowledge, mixing of CTAC and HFDePC in lyotropic mesophases has not been investigated, and despite the number of investigations of their micellization behavior, it is not known whether using them for pore templating induces mixing or the formation of type I particles (as has been observed for nonionic FC/HC mixtures27). Here, we demonstate that the synthesis of mesostructured silica formed using CTAC and HFDePC presents a novel approach to the controlled hierarchical porous materials, and also provides evidence of micelle demixing in a concentrated solution of silica and surfactant. Conditions are reported for the full variety of demixed micelle architectures depicted in Figure 1 for HC/FC surfactant templated silica. 2. Experimental Section 2.1. Materials. The hydrocarbon surfactant CTAC (98%+) and tetraethyl orthosilicate, TEOS (98%), were purchased from Sigma. The fluorocarbon surfactant HFDePC was synthesized

J. Phys. Chem. C, Vol. 114, No. 41, 2010 17391 by alkylation of pyridine with 3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10heptadecafluorodecyl iodide followed by ion exchange, as described previously.10 Concentrated aqueous ammonia (29 wt % NH4OH, Merck), deionized ultrafiltered (DIUF) water, anhydrous ethanol (Aaper Alcohol and Chemical), and NaCl (Merck KGaA) were used as received for material synthesis. Concentrated aqueous HCl (ACS grade, Fisher Scientific) and anhydrous ethanol were used for surfactant extraction. 2.2. Silica Material Synthesis. The synthesis of mesoporous silica materials was carried out in dilute solutions of CTAC, HFDePC, and TEOS under mild basic conditions. We prepared four series of samples to investigate the effects of key synthetic parameters including the molar composition of the mixed surfactant system, the amount of ammonia, the addition of NaCl, and the addition of ethanol. For all samples, the total molar amount of mixed surfactant was kept constant. The initial molar composition of reactants for the synthesis of silica materials can be generalized as follows: TEOS:H2O:HFDePC:CTAC:NH3: NaCl:C2H5OH ) 1:148:0.12x:0.12(1 - x):y:s:z. The specific initial molar ratios of reactants will be described in the following sections. A typical synthesis procedure is as follows: The calculated amounts of CTAC and HFDePC were mixed with DIUF water and concentrated ammonium hydroxide. If needed, the appropriate amount of NaCl or C2H5OH was also added at this time. The mixture was vigorously stirred for at least 30 min to completely dissolve and equilibrate the surfactants. The calculated amount of TEOS was then slowly added, and the solution was aged for 24 h at room temperature with gentle stirring (∼100 rpm). To permit direct comparison, the size of the reactor vessel, the stir bar, the stirring speed, the mixing time, and the TEOS addition rate were kept the same for all of the macroscopically well-mixed samples. The precipitate was isolated by filtration and dried in air, and the mixed surfactants were removed by twice washing with an acidic mixture of 6% concentrated HCl and 94% ethanol. The washing time for each step was 24 h. 2.3. Characterization. X-ray diffraction (XRD) patterns were obtained with a Siemens 5000 diffractometer using Cu KR radiation (λ ) 1.54098 Å) and a graphite monochromator. Scanning electron microscopy (SEM) was performed on a Hitachi S-900 microscope. Solid samples were loaded on PELCO carbon tabs, and then coated with gold under vacuum conditions for SEM imaging. Transmission electron microscope (TEM) images were collected with a JEOL 2010F electron microscope operating at 200 kV. Solid samples were dispersed in an isopropanol solution by sonication and then deposited onto lacey carbon grids for TEM observation. Nitrogen sorption measurements were performed with a Micromeritics Tristar 3000 system. All samples were degassed at 120 °C for 4 h under flowing nitrogen prior to measurements. For different types of pore geometries, the pore size distributions (PSDs) were calculated using the appropriate variant of the Kelvin equation (discussed below) but always from the adsorption branch of the isotherms. FTIR spectra were obtained with a desiccated and sealed ThermoNicolet Nexus 470 infrared spectrometer with a DTGS detector and a nitrogen-purged sample compartment. Samples were finely ground and diluted to 1 wt % with KBr powder before being pressed into translucent pellets with a hand press. 3. Results and Discussion 3.1. Effects of Molar Composition of CTAC and HFDePC. A series of samples with different molar ratios of HFDePC to CTAC was synthesized with initial reactant molar ratios of 1 TEOS:0.12x HFDePC:0.12(1 - x) CTAC:10 NH3:148 H2O with

17392

J. Phys. Chem. C, Vol. 114, No. 41, 2010

Figure 2. FTIR spectra of KBr pellets pressed with 1 wt % of (a) the reagent CTAC, (b) the reagent HFDePC, (c) sample A-3 as synthesized, and (d) sample A-3 after extraction.

x ) 0, 1/3, 1/2, 2/3, or 1. The products will be called sample A-1 through A-5 corresponding to the order from x ) 0 to x ) 1. For comparison with the other series of samples, A-3 is denoted as the base sample. Infrared spectroscopy (Figure 2) confirms that both HC and FC surfactants are incorporated into the as-made silica samples, and that both can be extracted by washing with acidic ethanol. Results are shown for sample A-3 as a representative example. The FTIR spectra of the pure surfactants contain several bands concentrated in two regions from 3100 to 2700 cm-1 and from 1500 to 400 cm-1. For CTAC, the bands at 2920 and 2850 cm-1 are attributed to CH2 asymmetric and symmetric stretching, respectively.39 Bands around 1486 cm-1 are attributed to CTAC headgroup deformation modes.40 For HFDePC, the bands in the high wavenumber range are associated with C-H stretching in the pyridinium ring of the headgroup and CH2 stretching in the spacers between pyridinium and the fluorocarbon tail. The low wavenumber bands are primarily attributed to C-F vibrations including CF2 symmetric stretching (1204 and 1151 cm-1) and asymmetric stretching at 1246 cm-1.41 The sharp band at 1490 cm-1 is associated with pyridinium. The strongest bands from CTAC and HFDePC can be clearly resolved in the assynthesized sample (Figure 2c), although they shift slightly compared to the pure compounds (from 2920/2850 to 2926/ 2855 cm-1 and from 1151 to 1150 cm-1, respectively). After twice washing with acidic alcohol, all of the surfactant bands are absent from the FTIR spectra (Figure 2d). The band at 959 cm-1 is attributed to Si-OH stretching.42 The sharp band at 1070 cm-1 is attributed to asymmetric Si-O-Si stretching43 and shifts to 1085 cm-1 after extraction, which suggests enhanced sol-gel condensation during extraction. Powder XRD patterns for this series of samples after extraction are shown in Figure 3. Samples A-1 and A-2 show consistent patterns that are indexed to well-ordered 2D HCP structures: three well-resolved intense (100), (110), and (200) reflections and one weak (210) reflection. These HCP patterns suggest that, at low molar fraction of HFDePC, the mesopore structures of the products are governed by CTA+-rich/silica aggregates. With increasing molar fraction of HFDePC in the template for sample A-3, the material displays a low angle XRD pattern with one strong (100) peak and a broad signal of relatively low intensity that can be attributed to overlapping (110) and (200) reflections of a 2D HCP pattern. The broadening of the XRD peaks in sample A-3 indicates a degradation in the

Xing et al.

Figure 3. XRD patterns for the series of extracted samples A-1 through A-5. The samples were prepared with 1:0, 2:1, 1:1, 1:2, or 0:1 mol ratios of CTAC:HFDePC, respectively.

order of the hexagonal phase of CTAB-templated materials, consistent with the TEM observations of less-ordered and wormhole-like pores (see below). Sample A-4 exhibits no reflection, indicating that it has a disordered pore structure. Further increasing the molar fraction of HFDePC causes a gradual transition from a disordered structure to a new phase with two reflections. At x ) 5/6, only one intense XRD reflection is observed, and at x ) 0.95, we can see two reflections by XRD similar to sample A-5 (not shown for brevity). The XRD patterns of sample A-5 exhibit peaks indexed to (001) and (002) reflections of a random mesh phase.10 We expect that there may be one reflection due to the pillar spacing in the random mesh phase below 2θ of 1.6, which cannot be observed because of instrumental limitations. Taken together, the XRD patterns of this series of samples indicate transitions from hexagonal to disordered 2D HCP to disorder to random mesh phase as the molar fraction of HFDePC increases in the A-x series. Nitrogen sorption isotherm plots and calculated pore size distributions of this series of samples are shown in Figure S2 (Supporting Information). All samples have a reversible type IV isotherm,44 indicating that they have uniform mesopores. As the molar fraction of HFDePC increases, a small type H3 hysteresis loop appears in sample A-2, and enlarges to a maximum area in sample A-3 before reducing in samples A-4 and A-5. The size of the hysteresis loop correlates with the amount of hollow cells observed by TEM (see below). An upturn at high relative pressure also becomes stronger as the molar fraction of HFDePC increases, indicating that the textural porosity between clusters of a particle is enhanced. The increase of textural porosity is consistent with the observation of smaller and more elongated particles by SEM (see below). In addition, the inflection points corresponding to capillary condensation in mesopores shift to lower relative pressure, suggesting the formation of smaller pores, as the molar fraction of HFDePC increases. For all samples, the pore size distributions were calculated using the BJH method with modified Kelvin equation and the Harkins-Jura equation for film thickness (also known here as KJS pore size distributions).45 The first four samples have a 2D HCP mesophase and the pore size distributions were calculated assuming cylindrical pore geometry, but because sample A-5 has a random mesh phase structure, the pore size

Demixed Micelle Morphology Control

J. Phys. Chem. C, Vol. 114, No. 41, 2010 17393

TABLE 1: Structure Parameters of the Mixed-Surfactant-Templated Mesoporous Silica Materialsa sample name

SBET (m2/g)

VP/P0)0.95 (cm3/g)

WKJS (nm)

Vpb (cm3/g)

Vmb (cm3/g)

Stb (m2/g)

Sexb (m2/g)

A-1 A-2 A-3 A-4 A-5 B-1 B-3 C-2 C-3 C-4 D-2 D-3 D-4

947 963 856 712 895 901 906 912 934 550 1089 1143 1174

0.76 0.86 0.76 0.74 0.73 0.63 0.75 0.62 0.70 0.29 0.89 0.61 0.67

3.8 3.7 3.7 3.5 2.4 3.6 2.4, 3.8 3.3, 3.8 3.3, 3.8 3.7 3.0, 3.5 3.0 3.0

0.68 0.69 0.52 0.36 0.39 0.61 0.56 0.53 0.63 0.27 0.65 0.62 0.60

0.033 0.021 ∼0 ∼0 0.0038 0.022 0.012 0.0098 0.025 0.14 0.0075 ∼0 ∼0

796 817 699 571 680 818 885 729 788 214 864 843 813

55.0 117.9 158.5 250.2 248.3 152.6 158.2 65.6 44.1 12.1 159.5 59.8 36.4

a SBET ) BET surface area,62 VP/P0)0.95 ) total pore volume, WKJS ) pore diameter at peak of KJS pore size distribution, Vp ) total mesopore volume, Vm ) micropore volume ) I × 0.001547 (cm3), where I represents the Y-intercept in the Rs plot, St ) total specific surface area, Sex ) external specific surface area. b Calculated using Rs comparative nitrogen adsorption plots.47

distribution was calculated assuming slit-shaped pores. By analyzing Rs plots generated from the isotherms as described by Sayari et al.,46 additional pore texture characteristics were calculated, including the primary mesopore volume Vp, total surface area St, and external surface area Sex (Table 1). The standard reduced nitrogen adsorption isotherm data (Rs) for the reference material, LiChrospher Si-1000 silica, were taken from Jaroniec et al.47 The PSDs of all samples in this series are unimodal, and for samples A-1 through A-3 centered at 3.7 nm, suggesting that the mesopore size is governed by CTAC micelles when the molar fraction of HFDePC is less than or equal to 1/2. For samples A-4 and A-5, the pore size decreases, suggesting that the pore size is governed by HFDePC micelles. In addition, the PSDs for this series of samples become broader as more fluorinated surfactant is introduced. The mesopore volume and specific surface area in sample A-4 are minimized within this series due to its disordered pore structure. Representative TEM images of this series of samples are compared in Figure 4. Sample A-1, prepared with only CTAC, shows uniformly ordered domains of perfect 2D hexagonal columnar phase (HCP) cylindrical pores. Both stripe and spot patterns are observed, corresponding to two different views, edge-on and end-on, respectively. This sample shows predominantly round rough particle morphology in the TEM images. Sample A-2, prepared with almost equal weights of HFDePC and CTAC, also contains large domains of ordered cylindrical pores similar to sample A-1. However, some hollow cells with sizes on the order of tens of nanometers are observed in this sample, and the hollow cells appear to be captured within larger particles to form a bimodal pore structure. The image in Figure 4 is representative of the sparse, random distribution of voids in sample A-2, although not every particle observed contained voids. Similar hollow cells were reported previously when just dilute HFDePC was used as a template under similar conditions. From TEM analysis, the formation mechanism of hollow cells was proposed to be coalescence of individual vesicle-like hollow silica particles.11,12 If only CTAC surfactant is employed as a template, there are no vesicle-like hollow cells formed under similar conditions.48 The formation of coexisting domains of 2D HCP phase and vesicle-like hollow cells suggests that CTAC and HFDePC do not mix when they are combined together with almost equal weight fraction. The demixed structure of sample A-2 suggests that there exist two populations of cationic micelles composed of CTA+-rich and HFDePy+-rich surfactant in bulk solution, in agreement with the results reported by Almgren et al.34 After TEOS is added, CTA+-rich and HFDePy+-rich

Figure 4. Representative TEM images for the series of extracted samples A-1 through A-5. The white scale bar on the image of sample A-1 is 20 nm wide, and all other white scale bars are 100 nm wide. The white circle on the image of sample A-1 indicates hexagonal pores.

micelles coassemble with silicate species separately and precipitate to form particles with bimodal pores. In this sample, most particles are found by TEM to be globular with incorporated hollow cells, although some elongated particles can also be seen in the low magnification TEMs (Supporting Information Figure S3). Further increase of the molar fraction of HFDePC in sample A-3 leads to primarily elongated particles with a large number

17394

J. Phys. Chem. C, Vol. 114, No. 41, 2010

density of hollow cells. Multiple distinct hollow cells captured within the larger particles are highly curved and assembled along the particle axes. The increase in the number density of hollow cells corresponds with the increased molar fraction of HFDePC, consistent with the idea that hollow cells form by merging together vesicle-like silica particles templated by HFDePy+rich surfactants. In this sample, CTA+-rich micelle templated domains exhibit a less-ordered mesopore structure than that of sample A-2, and some regions with deformed 2D HCP and even wormhole-like mesopores are found (Supporting Information Figure S4). This suggests that CTA+-rich micelle templated mesopores follow the tendency of CTA+ micelles to go from lamellar to hexagonal to wormhole-like phase as the surfactant concentration decreases.49,50 Increasing the molar fraction of HFDePC to 2/3 in sample A-4 (Supporting Information Figure S3) leads to particles possessing disordered mesopores. For this sample, all particles show irregular shape, and no hollow cells were observed from TEM. Also, no evidence of demixed domains could be found in this sample even after extensive searching by TEM, which suggests that CTA+ surfactant molecules have some degree of solubility within HFDePy+rich micelles. This result is similar to the report by Asakawa et al.,34 who estimate a solubility of about 17 mol % CTAC in HFDePC micelles when the total concentration is above the second CMC of the mixture, 2.6 mM. In addition, the structure of the final material seems to be governed by HFDePy+ micelles in this sample because the disordered channel arrangement follows the phase transition sequence of HFDePC templated mesostructures from vesicular particles to disordered porous particles as the HFDePC concentration increases.12 Sample A-5, prepared with HFDePC alone, consists of elongated particles with random mesh phase structure. The mesh phase layers orient perpendicular to the particle axis in a way consistent with the detailed characterizations reported by Tan et al.10 Representative SEM images of members of the A-x series of samples are shown in Figure 5. The particle size dramatically decreases from micrometer to nanometer scale as the fraction of HFDePC increases, which can be explained by the higher surface activity of HFDePC relative to CTAC; for example, at 25 °C, the surface tensions of aqueous HFDePC and CTAC solutions above their CMC are 26.1 and 42.3 mN/m, respectively.34 As a consequence, the addition of fluorocarbon surfactant in the template favors the formation of small silica particles. In addition, sample A-1 consists of predominantly rough spherical particles, while sample A-3 contains a mix of spherical and elongated particles. The appearance of some elongated particles in Figure 5 for sample A-1 is thought to be due to agglomeration of particles, while the elongated particles in sample A-3 are found (by TEM) to be hollow, elongated particles with multiple hollow voids. Similarities between the two images suggest that aggregation of particles occurs in both cases, and may play a role in the formation of the elongated particles in sample A-3. With a further increase of HFDePC in sample A-5, only small elongated particles with uniform size are found. From this series of samples, we find that different phases or a mixture of phases can be obtained depending on the molar ratio of mixed surfactants. All of the particles are classified as type I particles in which pores are templated by micelles rich in one surfactant and the second aggregates at external or internal surfaces of the particle. The external surface area and amount of textural porosity depend on the particle size. As more HFDePC is added, a sequence of pore structures from ordered 2D HCP to less ordered 2D HCP to disorder to random mesh phase is observed, which exactly follows the phase behavior of

Xing et al.

Figure 5. Representative SEM images for extracted samples A-1 (scale bar: 3 µm), A-3 (scale bar: 500 nm), and A-5 (scale bar: 500 nm). A-1 consists primarily of globular particles or aggregates, A-3 of a mix of spherical and elongated particles, and A-5 of small elongated particles.

mesoporous materials templated with only CTA+ aggregates as the surfactant concentration decreases. For pores templated by HFDePy+-rich aggregates, a sequence from close-packed cylinders to vesicular to disordered pores to mesh phase is observed as the HFDePC concentration increases, which differs from the transitions seen for templating with HFDePC alone in the appearance of the disordered pore structure before the mesh phase.31 The disordered structure most likely forms due to mixing with CTAC. 3.2. Effects of Ammonia Concentration. From the molar ratio study, we observe that CTA+ and HFDePy+ show strong demixing behavior when equal mole fractions of the surfactants are used. Next, we discuss a series of samples prepared with this surfactant ratio and variable ammonia concentration, to observe the effect of the sol-gel catalyst on the nature of demixing. The reactants have initial molar ratios of 1 TEOS:0.06 HFDePC:0.06 CTAC:y NH3:148 H2O with y ) 5, 10, and 15 for samples B-1, B-2, and B-3, respectively. Sample B-2 is the base sample in this series. Powder XRD patterns of this series of samples after extraction are shown in Figure 6. The pattern of sample B-1, synthesized at the lowest ammonia concentration, shows two strong reflections and one weak reflection. Together with nitrogen adsorption isotherm and TEM results (see below), this pattern is interpreted as coexisting domains of deformed 2D HCP and wormholelike phases. The three diffractions in this sample can be indexed from left to right as (100), (100), and (110). The (100) and (110) diffractions come from the deformed 2D HCP phase formed by CTA+/silica aggregates, while the second (100) diffraction comes from separate regions of wormhole phase formed by HFDePy+/silica aggregates. With more ammonia in the synthesis solution (sample B-2), the XRD pattern shows one strong

Demixed Micelle Morphology Control

J. Phys. Chem. C, Vol. 114, No. 41, 2010 17395

Figure 6. XRD patterns for the series of extracted samples B-1 through B-3 prepared with a 1:1 mol ratio of CTAC:HFDePC and 5, 10, or 15 mol of NH3/mol of TEOS in the synthesis solution, respectively.

reflection of (100) and one broad peak that includes the (110) and (210) reflections of the HCP phase. Further increase of ammonia content (sample B-3) yields a product with at least six distinct reflections, possibly indicating a novel two- or three-dimensional pore structure. An initial hypothesis was that three-dimensional Pm3jn cubic structure might have formed in this sample, given that the mixed CTAC/ HFDePC surfactant system can produce samples with Pm3jn cubic structure in concentrated acid-catalyzed solutions (unpublished results). However, careful XRD indexing, N2 sorption, and STEM analysis together rule out this possibility (see below) and suggest that the structure consists of coexisting but highly ordered phases. Four peaks, indexed as (100), (110), (200), and (210) reflections, are from a 2D HCP phase with a unit cell parameter of a ) 4.4 nm. The first reflection to the left of (100) is interpreted as a characteristic diffraction from silica micropillars between layers in a random mesh phase,10 and the shoulder to the right of (100) is indexed as the (001) reflection from mesh phase layers with a layer spacing of 3.3 nm. However, the (002) reflection expected from the mesh phase can not be clearly resolved because it overlaps with the (210) reflection. Because no more reflections could be found to indicate a periodic three-dimensional arrangement of silica micropillars into an ordered mesh phase, we assign the (001) and low-angle pillar reflections to a random mesh phase. The set of reflections for sample B-3 could not be indexed to any other single known mesophase. Nitrogen sorption isotherms for this series of samples after extraction are shown in Figure S5 of the Supporting Information. All samples have type IV isotherms with upturns at high relative pressure. In contrast to samples B-1 and B-3, sample B-2 shows a hysteresis loop due to the hollow vesicular nature of the base sample (see above). For sample B-3, two distinct capillary condensation steps can be observed at relative pressures (P/P0) ranging from 0.12-0.25 and 0.32-0.4, which indicates a bimodal pore size distribution. The PSDs of this series of samples were calculated by the KJS method. For samples B-1 and B-2, cylindrical pore geometry was assumed in the PSD calculation. For sample B-3, because XRD indicates the coexistence of two different types of pore geometries, i.e., slit pores and cylindrical pores, we calculated the PSDs for both shapes. All the results are shown in Figure 7. As we can see,

Figure 7. Pore size distributions of the series of sample B-1 through B-3 calculated using the KJS method. For sample B-3, PSDs were calculated by the KJS method for both cylindrical and slit-shaped pores but the curves are solid for the region where the geometry is believed to apply.

samples B-1 and B-2 show unimodal distributions of mesopores but sample B-3 clearly shows a bimodal mesopore size distribution. In addition, the peaks become sharper as the ammonia concentration increases. For B-3, the PSDs calculated assuming both cylindrical and slit-shaped pores are both shown in Figure 7, but the part of each PSD that we believe accurately represents the dominant pore shape for that region is represented with a solid curve, while the rest of each PSD is dashed. The 2D HCP mesophase component (with cylindrical pore geometry) has a peak in the PSD of around 3.8 nm, which is the same as the pore size of (CTA+-templated) sample A-1. The random mesh phase component of sample B-3 with slit pore geometry has a peak in the PSD of 2.4 nm, which is consistent with sample A-5 prepared with only HFDePC surfactant. The other structure parameters for the B-x series of samples after extraction are shown in Table 1. Two representative TEM images of sample B-1 are shown in Figure 8. Sample B-1 contains coexisting rough spherical and elongated particles, perhaps indicating weak micelle demixing in the precipitated particles. On the basis of the analysis above, the spherical particles are formed by CTA+/silica aggregates, while the elongated particles are formed by HFDePy+/ silica aggregates. In the spherical particles, cylindrical pores can be found with short-range order (see Figure 8a), indicating weak 2D HCP ordering. For the elongated particles, wormholelike mesopores are found, as shown in Figure 8b. The d-spacings measured from TEM are consistent with XRD results. Sample B-2 is the base sample, which has improved 2D HCP ordering of the CTA+ aggregates and vesicular elongated particles. High magnification TEM images of sample B-3 confirm the presence of coexisting ordered domains of type II demixed phases. Figure S6 of the Supporting Information shows a brightfield TEM image of this sample, suggesting that particles are composed of distinct domains with different pore orientations, some of which appear to be oriented parallel to the domain boundary while others are perpendicular to the domain boundary. The calculated d100 and d001 spacings from the TEM images are

17396

J. Phys. Chem. C, Vol. 114, No. 41, 2010

Figure 8. Representative TEM images of particles from sample B-1 showing (a) globular particles with HCP cylindrical pores (an example of which is circled) and (b) elongated particles with wormhole-like pores. Both scale bars are 50 nm wide.

Figure 9. Representative STEM image of sample B-3. The two squares on the image are expanded in the insets to show coexisting ordered 2D HCP phase regions (upper right) and mesh phase regions (lower right). The scale bar is 100 nm wide in the main image and 10 nm wide in the insets.

consistent with the XRD interpretation of a biphasic, ordered mesostructure. On the basis of the XRD spacings and bimodal PSD, we conclude that they are templated by CTA+-rich and HFDePy+-rich micelles separately. The coexisting distinct domains of mesostructure can be more easily observed in the dark field STEM images shown in Figure 9. The entire particle consists of many ordered domains, and the insets show examples of 2D HCP cylindrical pores (both edge-on and end-on views) and mesh phase pores. Pillars between the silica layers are apparent in the mesh phase image. Figure S7 of the Supporting Information shows the SEM images of the extracted samples B-1 and B-3. Sample B-1 consists of both spherical and elongated particles, which is consistent with TEM results. Sample B-3 is composed of rough particles with heterogeneous size and shape distribution. From this series of samples, we conclude that the ammonia concentration significantly influences the distribution of phases in the final products. The formation of biphasic domains was observed for all samples in this series. The biphasic domains

Xing et al. change from deformed 2D HCP/wormhole-like to 2D HCP/ vesicle to well-ordered 2D HCP/mesh as the ammonia concentration increases. Ammonia is expected to accelerate hydrolysis and precipitation of TEOS, which apparently allows the rapid formation of small, separate demixed domains that are better preserved than they are when hydrolysis and precipitation are more gradual. The mechanism of forming a biphasic mesostructure by increasing ammonia concentration requires detailed investigation, but the observations discussed above are consistent with competitive precipitation between HFDePy+/silica and CTA+/silica aggregates. At low ammonia concentration, the precipitation rate of HFDePy+/silica and CTA+/silica aggregates is different due to low availability of hydrolyzed TEOS, and the aggregates of HFDePy+/silica are most likely able to precipitate first. At medium ammonia concentration, HFDePy+/ silica aggregates may still precipitate first but more extensive charge screening of HFDePy+ micelles by hydrolyzed silica can explain the formation of vesicles. At high ammonia concentration, HFDePy+/silica and CTA+/silica may precipitate separately but at similar rates, leading to the formation of small ordered particles that aggregate into type II biphasic mesh/2D HCP particles. 3.3. Effects of Adding NaCl. A series of samples was prepared to investigate the effect of adding NaCl on the nature of demixing during templating with equal moles of HFDePC and CTAC. The initial reactants have molar ratios 1 TEOS:0.06 HFDePC:0.06 CTAC:10 NH3:148 H2O:s NaCl with s ) 0, 0.28, 2.8, and 5.6 for samples C-1 through C-4, respectively. Sample C-1 is the base sample in this series. Powder XRD patterns for this series of extracted samples are shown in Figure S8 of the Supporting Information. As discussed above, the base sample (C-1) shows two reflections corresponding to weak hexagonal ordering. Sample C-2, prepared with a small amount of NaCl, shows an XRD pattern similar to that of the disordered surfactant-templated silica reported by Ryoo et al.51 which possesses shorter-range order than the 2D HCP mesophase. The intensity and resolution of higher order peaks are lower than those for sample C-1, indicating diminished hexagonal ordering. Thus, the addition of NaCl induces a transition toward wormhole-like pores, which is consistent with previous reports on the effects of adding NaCl to chloride-based cationic surfactant solutions.52,53 Large enough concentrations of salts can shorten cylindrical micelles to the point that they eventually transition into globular micelles.54 In addition, the (100) peak shifts to a lower angle, suggesting that the average pore-pore distance increases. Increasing the NaCl concentration further in sample C-3 causes a further shift of the reflection to lower angle. Sample C-4, prepared with the highest NaCl concentration (2 M), is completely disordered, consistent with previous reports that a high NaCl concentration disrupts micelle ordering in mesoporous silica.55 Figure 10 shows the nitrogen adsorption isotherms and calculated pore size distributions of this series of samples after extraction. For direct comparison, the pore size distributions of all samples were calculated by the KJS method assuming cylindrical pore geometry. Sample C-2 shows two distinct capillary condensation steps at relative pressure (P/P0) ranging from 0.12-0.25 and 0.30-0.40, indicating that bimodal mesoporous materials are formed in the presence of NaCl. The pore size distribution of this sample is weakly bimodal with two peaks at 3.3 and 3.8 nm. Compared to sample C-1, the hysteresis loop disappears and the upturn at high relative pressure decreases, showing that the addition of NaCl reduces both the number of vesicle cavities and the textural porosity

Demixed Micelle Morphology Control

J. Phys. Chem. C, Vol. 114, No. 41, 2010 17397

Figure 10. (a) Nitrogen adsorption isotherms of samples C-1, C-2 (upshifted 150 cm3/g), C-3 (upshifted 300 cm3/g), and C-4 (upshifted 600 cm3/g) made with a 1:1 mol ratio of CTAC:HFDePC and 0, 0.28, 2.8, or 5.6 mol of NaCl per mol of TEOS, respectively. (b) Pore size distribution of samples C-1 through C-4 calculated using the KJS method assuming cylindrical pore geometry. The solid arrows indicate the first inflection, and the dashed arrows indicate the second inflection.

Figure 11. Representative TEM images of sample C-2 through C-4. The black arrows in sample C-2 indicate small silica particles surrounding the larger particles. The left two images have 50 nm scale bars, and the right two images have 20 nm scale bars.

between clusters of particles. The isotherm for sample C-3 more clearly shows two capillary condensation steps and a bimodal pore size distribution with well-defined peaks at 3.3 and 3.8 nm. To our knowledge, this is the first reported example of welldefined bimodal mesoporous silica with such a small pore size difference (only 0.5 nm). Sample C-4, prepared with a large amount of NaCl, shows only one capillary condensation step and a single PSD peak. The Rs plot of this sample shows the largest micropore volume of any sample reported here. The other structure parameters of this series of samples are shown in Table 1. This series shows that the addition of NaCl can induce the translation of the mesoporous matrix from unimodal to bimodal and then back to unimodal. The pore sizes in the mesopore distribution remain almost constant with increasing NaCl. In addition, the BET surface area and total mesopore volume all reach maxima in sample C-3. The bimodal PSD suggests templating with two large populations of separate CTAC-rich and HFDePC-rich micelles that coassemble with negatively charged silica to form an intimately mixed single phase. We have demonstrated that each type of micelles can be selectively swelled by adding lipohilic or fluorophilic oils to the initial sols, thus making independently tunable bimodal mesopore silica.56 With excessive NaCl (sample C-4), effective micelle templating cannot occur due to electrostatic screening and the resulting materials exhibit low surface area and considerable reduction in mesophase structure. Representative TEM images of samples C-2 through C-4 are compared in Figure 11. Sample C-2 consists predominantly of ∼100 nm spherical particles. Unlike the base sample C-1, it contains no elongated particles with hollow cells. The high magnification TEM images of this sample do not show separate ordered domains, but a distribution of weakly ordered pores can be discerned that is consistent with the XRD and pore size distribution data. We conclude that sample C-2 possesses a bimodal mixture of wormhole-like mesopores as compared to sample C-1 with 2D hexagonal mesopores. The loss of vesicular HFDePy+-templated chambers is accompanied with the addition of smaller 3.3 nm pores in the PSD, which are presumably

templated by HFDePy+-rich micelles. However, because the pores are not organized into separate phases, these are type III particles formed from a disordered phase consisting of coexisting HFDePy+-rich and CTA+-rich aggregates. The larger particles of around 100 nm in this sample are surrounded by a layer of small silica particles, which presumably are deposited late in the precipitation process. With a further increase of NaCl in sample C-3, the TEM image shows larger spherical particles than sample C-2, without a layer of small particles coating them. Sample C-4 shows mainly disordered wormhole-like pores. The formation of a unimodal distribution of wormhole-like mesopores may indicate enhanced mixing of CTAC and HFDePC surfactants in the presence of a sufficient amount of NaCl. This result is consistent with the effect of salt on mixing of hydrocarbon and fluorocarbon surfactants in dilute solution determined by fluorescence quenching.57 Figure S9 of the Supporting Information shows the effect of the NaCl amount on the morphology of this series of extracted samples. Unlike sample C-1, the SEM image of sample C-2 consists of only rough spherical particles with sizes near 100 nm, coated with smaller secondary spherical particles with sizes less than 30 nm. Sample C-3 consists of smooth spherical particles. Sample C-4 consists of large irregular-shaped microparticles along with a small amount of sheet-like silica particles, all of which are flocculated into larger aggregates. The difference in morphology should be due to screening of electrostatic interactions between surfactants and silica. Increased screening would be expected to decrease the difference in the rate of precipitation of CTA+-rich and HFDePy+-rich micelles with silica, leading to more homogeneous and uniform particles. The appropriate amount of NaCl results in predominantly globular silica/surfactant aggregates with sizes near 100 nm. From this series of samples, we conclude that adding NaCl significantly affects the structure of both particles and pores by influencing the coassembly behavior of demixed micelles with silicate species in the solution. On one hand, the addition of a large amount of NaCl may greatly widen the molar concentration range over which demixed micelles composed of HFDePy+rich and CTA+-rich micelles coexist. For example, Asakawa et

17398

J. Phys. Chem. C, Vol. 114, No. 41, 2010

Figure 12. Representative TEM images of sample D-2 (100 nm scale bar) and D-4 (500 nm scale bar) prepared with a 1:1 mol ratio of CTAC:HFDePC and 10 or 30 mol of added ethanol per mol of TEOS, respectively.

al.58 found that the addition of LiCl causes almost perfect demixing for the lithium perfluorooctanonate (LiPFN)/lithium dodecyl sulfate (LiDS) system, with micelles dividing into one population with 5 mol % fluorinated surfactant and another with 99 mol %. On the other hand, adding NaCl introduces an equivalent increase in Cl- at cationic micellar interfaces,59 which would shield the electrostatic interactions between cationic surfactant micelles and negatively charged silica and thus slow down precipitation. This shielding would allow strongly demixed CTA+-rich and HFDePy+-rich micelles in solution to simultaneously coassemble with silica and precipitate together, leading to the formation of materials with intimately mixed bimodal pores. 3.4. Effects of Adding Ethanol. A series of samples was prepared to investigate the effects of adding ethanol on the nature of demixing during templating with equivalent moles of HFDePC and CTAC. The initial reactants have the molar ratios 1 TEOS:0.06 HFDePC:0.06 CTAC:10 NH3:148 H2O:z with z ) 0, 10, 20, and 30 for sample D-1 through D-4, respectively. D-1 is the base sample. Figure 12 shows the representative TEM images of samples D-2 and D-4. Sample D-2 shows coexisting, clearly differentiated spherical and elongated silica particles, suggesting that CTA+-rich and HFDePy+-rich micelles may separately coassemble with silica species to form independent particles with

Xing et al. different morphologies. Both the spherical and elongated particles have wormhole-like pore structure. Upon further increase of the ethanol content, the elongated particles are no longer found, and all particles are predominantly spherical (as in sample D-4). SEM shows a mixed spherical and elongated shape for sample D-2 and uniform spherical shape for D-4 (Figure S10, Supporting Information). The nitrogen adsorption isotherms of members of this series prepared with the addition of different amounts of ethanol are shown in Figure S11a (Supporting Information). Generally, all samples in this series exhibit type IV isotherms. For samples D-1 and D-2, the capillary condensation takes place at P/P0 between 0.2 and 0.4. Upon further increase of the amount of ethanol in samples D-3 and D-4, the capillary condensation happens at lower relative pressure, P/P0 between 0.15 and 0.3. Without ethanol, the isotherm of sample D-1 shows an upturn at high relative pressure and a type H3 hysteresis loop. The upturn at high relative pressure indicates textural porosity between clusters of particles. When ethanol is added, the H3 type hysteresis loop disappears, indicating that ethanol disrupts the formation of vesicle-like particles templated by HFDePy+rich micelles. With the addition of a large amount of ethanol in samples D-3 and D-4, the upturns disappear completely. Figure S11b (Supporting Information) compares pore size distributions of this series of samples. We observe that the average pore size decreases as the amount of ethanol increases, suggesting that ethanol acts as a cosolvent in this case to reduce the micelle (pore) size by decreasing the micelle aggregation number.60 Additional pore texture parameters are shown in Table 1. It is found that, with the increase of ethanol amount, both BET surface area and mesopore volume increase, while the external surface area decreases significantly, possibly due to the changing particle morphology. Figure S12 (Supporting Information) presents the XRD patterns for this series of samples after extraction. Sample D-1 has a poorly ordered hexagonal mesophase as mentioned above. With the addition of 10 mol of ethanol/mol of TEOS in sample D-2, a new reflection close to the original (100) peak is found. This new reflection may appear to indicate an Ia3jd cubic phase, given that the phase transformation from the hexagonal phase to the Ia3jd cubic phase can be induced by increasing the ethanol concentration in TEOS-CTAB-ethanol-ammonia-water mixtures.61 However, careful XRD indexing, N2 adsorption isotherms, and TEM imaging together rule out this possibility and suggest that the structure actually consists of two coexisting microphase-separated domains. Upon further increase of ethanol to 20 mol/mol of TEOS in sample D-3, the XRD shows only one broad peak, indicating a wormhole-like structure. Similar to that of sample D-3, the XRD pattern of sample D-4 only shows one reflection. From this series of samples, we conclude that adding ethanol can induce mesophase transitions by separately altering surfactant packing parameters within demixed surfactant micelles. A small amount of ethanol promotes demixing of HC and FC micelles, leading to the formation of type IV demixed particles with a bimodal pore size distribution. With further increase of ethanol amount, the miscibility of HC and FC surfactants is enhanced, leading to a unimodal pore size distribution and uniform particle morphology. 4. Conclusions Mesoporous silica particles with diverse mesophases and pore size distributions have been synthesized using combined CTAC and HFDePC as templates. The structure and pore size distribu-

Demixed Micelle Morphology Control tions are adjusted by changing the synthetic parameters including the molar ratio of CTAC to HFDePC, the ammonia concentration, the amount of NaCl added, and the amount of ethanol added. Evidence for surfactant demixing can be observed under many conditions in precipitated CTA+/HFDePy+/silica aggregates, and all four types of demixed surfactant-based particles described in Figure 1 can be observed. Under the “base case” conditions (moderate ammonia concentration without addition of salt or ethanol), the structure of the final materials changes from ordered 2D HCP to less-ordered 2D HCP to disordered to mesh phase as the molar fraction of HFDePC increases in the mixture. At the low molar fraction of HFDePC, the phase structure is governed by CTAC. When the molar fraction of HFDePC increases above 50%, the phase structure is governed by HFDePC. Thus, the particles prepared under the “base case” conditions can be classified as type I particles where one type of micelle determines the predominant mesophase. At equal molar fractions of CTAC and HFDePC where strong demixing is observed, biphasic silica particles are observed with different ammonia concentrations. As the ammonia concentration increases, a transition of the biphasic structure can be found from less-ordered 2D HCP/wormholelike to less-ordered 2D HCP/vesicle to well-ordered 2D HCP/ mesh. A large amount of ammonia causes the formation of ordered biphasic 2D HCP/mesh materials with well-defined bimodal mesoporosity corresponding to type II particles. The addition of an appropriate amount of NaCl leads to the formation of type III bimodal mesoporous particles with wormhole-like pore structure, presumably due to the incorporation of demixed CTA+-rich and HFDePy+-rich micelles into single particles. Finally, the addition of a small amount of ethanol leads to the formation of type IV demixed particles, in which FC-rich and HC-rich micelles segregate into separate particles, as indicated by the differing morphologies of the mixture of particles in the product. This work shows not only the use of the sol-gel approach to verify mixing or demixing but also the ability to control the demixed micelle architectures in concentrated mixtures of HC/ FC surfactants in precipitated silica. In addition, a facile method for the synthesis of hierarchical porous silica with well-defined biphasic mesostructure and bimodal mesoporosity is demonstrated by using CTAC and HFDePC surfactants that are known to demix in dilute solution. The triphasic nature of demixed surfactants in a silica matrix provides opportunities to selectively tune bimodal pore sizes,56 and also to utilize the separate HC and FC domains for controlled deposition of two different catalytically active transitional metal oxides into different domains of intimately mixed mesopore channels. The applications of these materials are under ongoing investigation in our group. Acknowledgment. This work is partially supported by the National Science Foundation (NSF) under Grant Nos. DMR0210517, CTS-0348234, and EAR-0521405. We thank Dr. Sandhya Vyas for the synthesis of the fluorinated surfactant and Dr. Alan Dozier for the assistance with TEM imaging. Supporting Information Available: Molecular structures of surfactants used; nitrogen sorption results for the A-x, B-x, and D-x series; additional TEM images of samples A-2 through A-4; brightfield TEM of sample B-3; SEM images of samples B-1, B-3, C-2 through C-4, D-2, and D-4; and XRD patterns of the C-x and D-x series. This material is available free of charge via the Internet at http://pubs.acs.org.

J. Phys. Chem. C, Vol. 114, No. 41, 2010 17399 References and Notes (1) Hoffmann, H.; Wu¨rtz, J. J. Mol. Liq. 1997, 72, 191. (2) Kissa, E. Fluorinated Surfactants and Repellents; Hubbard, A. T., Ed.; Marcel Dekker: New York, 2001. (3) Ro¨thlingsberger, U.; Laasonen, K.; Klein, M. L.; Sprik, M. J. Chem. Phys. 1996, 104, 3692. (4) Sprik, M.; Ro¨thlingsberger, U.; Klein, M. L. Mol. Phys. 1999, 97, 355. (5) Ravey, J. C.; Ste´be´, M. J. Prog. Colloid Polym. Sci. 1987, 73, 127. (6) Mathis, G.; Leempoel, P.; Ravey, J. C.; Selve, C.; Delpuech, J. J. Am. Chem. Soc. 1984, 106, 6162. (7) Blin, J. L.; Lesieur, P.; Ste´be´, M. J. Langmuir 2004, 20, 491. (8) Blin, J. L.; Ste´be´, M. J. J. Phys. Chem. B 2004, 108, 11399. (9) Rankin, S. E.; Tan, B.; Lehmler, H. J.; Knutson, B. L. Microporous Mesoporous Mater. 2004, 73, 197. (10) Tan, B.; Dozier, A.; Lehmler, H.-J.; Knutson, B. L.; Rankin, S. E. Langmuir 2004, 20, 6981. (11) Tan, B.; Lehmler, H.-J.; Vyas, S. M.; Knutson, B. L.; Rankin, S. E. AdV. Mater. 2005, 17, 2368. (12) Tan, B.; Vyas, S. M.; Lehmler, H.-J.; Knutson, B. L.; Rankin, S. E. J. Phys. Chem. B 2005, 109, 23225. (13) Tan, B.; Lehmler, H. J.; Vyas, S. M.; Knutson, B. L.; Rankin, S. E. Chem. Mater. 2005, 17, 916. (14) Tan, B.; Lehmler, H. J.; Vyas, S. M.; Knutson, B. L.; Rankin, S. E. Nanotechnology 2005, 16, 502. (15) Ghosh, K.; Lehmler, H. J.; Rankin, S. E.; Knutson, B. L. Langmuir 2005, 21, 6145. (16) Blin, J. L.; Ste´be´, M. J. Stud. Surf. Sci. Catal. 2005, 156, 97. (17) Blin, J. L.; Ste´be´, M. J. Microporous Mesoporous Mater. 2005, 87, 67. (18) Blin, J. L.; Bleta, R.; Ghanbaja, J.; Ste´be´, M. J. Microporous Mesoporous Mater. 2006, 94, 74. (19) Sen, T.; Tiddy, G. J. T.; Casci, J. L.; Anderson, M. W. Chem. Mater. 2004, 16, 2044. (20) Liu, S.; Kong, L.; Yan, X.; Li, Q.; He, A. Stud. Surf. Sci. Catal. 2005, 156, 379. (21) Valle´, K.; Belleville, P.; Pereira, F.; Sanchez, C. Nat. Mater. 2006, 5, 107. (22) Song, M.-G.; Kim, J.-K.; Cho, S.-H.; Kim, J.-D. Langmuir 2002, 18, 6110. (23) Zhang, Q.; Ariga, K.; Okabe, A.; Aida, T. Stud. Surf. Sci. Catal. 2003, 146, 465. (24) Chen, F.; Huang, L.; Li, Q. Chem. Mater. 1997, 9, 2685. (25) Pang, J.; Hampsey, J. E.; Hu, Q.; Wu, Z.; John, V. T.; Lu, Y. F. Chem. Commun. 2004, 682. (26) Hillmayer, M. A.; Lodge, T. P. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, 1. (27) Blin, J. L.; Henzel, N.; Ste´be´, M. J. J. Colloid Interface Sci. 2006, 302, 643. (28) Li, D.; Han, Y.; Song, J.; Zhao, L.; Xu, X.; Di, Y.; Xiao, F.-S. Chem.sEur. J. 2004, 10, 5911. (29) Groenewolt, M.; Antonietti, M. Langmuir 2004, 20, 7811. (30) Areva, S.; Boissiere, C.; Grosso, D.; Asakawa, T.; Sanchez, C.; Linden, M. Chem. Commun. 2004, 1630. (31) Djojoputro, H.; Zhou, X. F.; Qiao, S. Z.; L Wang, Z. C.; Yu, Z.; Lu, G. Q. J. Am. Chem. Soc. 2006, 128, 6320. (32) Han, Y.; Ying, J. Y. Angew. Chem., Int. Ed. 2005, 44, 288. (33) Wang, K.; Ora¨dd, G.; Almgren, M.; Asakawa, T.; Bergenståhl, B. Langmuir 2000, 16, 1042. (34) Asakawa, T.; Hisamatsu, H.; Miyagishi, S. Langmuir 1995, 11, 478. (35) Kadi, M.; Hansson, P.; Almgren, M.; Furo´, I. Langmuir 2002, 18, 9243. (36) Almgren, M.; Garamus, V. M. J. Phys. Chem. B 2005, 109, 11348. (37) Tamori, K.; Esumi, K.; Meguro, K. J. Colloid Interface Sci. 1991, 142, 236. (38) Ravey, J. C.; Gherbi, A.; Ste´be, M. J. Prog. Colloid Polym. Sci. 1989, 79, 272. (39) Hind, A. R.; Bhargava, S. K.; Grocott, S. C. Langmuir 1997, 13, 6255. (40) Calabro, D. C.; Valyocsik, E. W.; Ryan, F. X. Microporous Mater. 1996, 7, 243. (41) Tsao, M. W.; Hoffmann, C. L.; Rabolt, J. F. Langmuir 1997, 13, 4317. (42) Innocenzi, P. J. Non-Cryst. Solids 2003, 316, 309. (43) Scheuing, D. R.; Weers, J. G. Colloids Surf. 1991, 55, 41. (44) 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. (45) Kruk, M.; Jaroniec, M. Langmuir 1997, 13, 6267. (46) Sayari, A.; Liu, P.; Kruk, M.; Jaroniec, M. Chem. Mater. 1997, 9, 2499. (47) Jaroniec, M.; Kruk, M. Langmuir 1999, 15, 5410.

17400

J. Phys. Chem. C, Vol. 114, No. 41, 2010

(48) Ikari, K.; Suzuki, K.; Imai, H. Langmuir 2004, 20, 7811. (49) Vartuli, J. C.; Schmitt, K. D.; Kresge, C. T.; Roth, W. J.; Leonowicz, M. E.; McCullen, S. B.; Hellring, S. D.; Beck, J. S.; Schlenker, J. S.; Schlenker, J. L.; Olson, D. H.; Scheppard, E. W. Chem. Mater. 1994, 6, 2317. (50) Pen˜a, M. L.; Kan, Q.; Corma, A.; Rey, F. Microporous Mesoporous Mater. 2001, 44-45, 9. (51) Ryoo, R.; Kim, J. M.; Ko, C. H.; Shi, C. H. J. Phys. Chem. 1996, 100, 17718. (52) Imae, T.; Ikeda, S. Colloid Polym. Sci. 1987, 265, 1090. (53) Magid, L.; Han, Z.; Warr, G.; Cassidy, M.; Bulter, P.; Hamilton, W. J. Phys. Chem. B 1997, 101, 7919. (54) Matsumoto, M.; McNamee, C. E.; Bossev, D. P.; Nakahara, M.; Ogawa., T. Colloid Polym. Sci. 2000, 278, 619. (55) Yu, J.; Shi, J.-L.; Chen, H.-R.; Yan, J.-N.; Yan, D.-S. Microporous Mesoporous Mater. 2001, 46, 153.

Xing et al. (56) Xing, R.; Lehmler, H.-J.; Knutson, B. L.; Rankin, S. E. Langmuir 2009, 25, 6486. (57) Asakawa, T.; Hisamatsu, H.; Miyagishi, S. Langmuir 1996, 12, 1204. (58) Asakawa, T.; Fukita, T.; Miyagishi, S. Langmuir 1991, 7, 2112. (59) Romsted, L. S. Langmuir 2007, 23, 414. (60) Anderson, M. T.; Matin, J. E.; Odinck, J. G.; Neweomer, P. P. Chem. Mater. 1998, 10, 311. (61) Liu, S.; Cool, P.; Collart, O.; Voort, P. V. D.; Vansant, E. F.; Lebedev, O. I.; Tendeloo, G. V.; Jiang, M. J. Phys. Chem. B 2003, 107, 10405. (62) Brunauer, S.; Emmett, P. H.; Teller, E. J. J. Am. Chem. Soc. 1938, 60, 309.

JP1051563