Synthesis and Tuning of Bimodal Mesoporous Silica by Combined

Mar 26, 2009 - Lexington, Kentucky 40506-0046, and ‡Department of Occupational and ... Engineering, The Ohio State University, 125A Koffolt Laborato...
0 downloads 0 Views 3MB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Synthesis and Tuning of Bimodal Mesoporous Silica by Combined Hydrocarbon/Fluorocarbon Surfactant Templating 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, 222 IREH, Iowa City, Iowa 52242. § Current address: Department of Chemical and Biomolecular Engineering, The Ohio State University, 125A Koffolt Laboratories, 140 West 19th Ave, Columbus, Ohio 43210-1185 Received January 9, 2009. Revised Manuscript Received February 21, 2009 Hydrocarbon and fluorocarbon surfactants show highly nonideal mixing that under some conditions results in demixing of the two types of surfactants into distinct populations of fluorocarbon-rich and hydrocarbon-rich aggregates. This also occurs in materials prepared by cooperative assembly of hydrolyzed tetraethoxysilane with mixtures of cetyltrimethylammonium chloride (CTAC) and 1,1,2,2-tetrahydro-perfluorodecylpyridinium chloride (HFDePC). Here, we report conditions under which demixed micelles lead to bimodal mesoporous materials (including specific concentrations of ammonia and salt in the synthesis solution) and show that the sizes of the hydrocarbontemplated and fluorocarbon-templated pores can be finely and independently controlled by adding lipophilic or fluorophilic oils, respectively. Nitrogen sorption isotherms and transmission electron microscopy provide clear evidence for a single phase of demixed but disordered wormhole-like pores.

Introduction Molecular aggregates of amphiphilic compounds (e.g., micelles) are currently being used in diverse areas ranging from storage, transport, and controlled release of drugs (or other active agents) to templating the pores of metal oxide materials.1-4 Demixed micelles are a relatively new type of system in which micelles differing in characteristics such as size, shape, and chemical compositions can be used as segregated matrices for multifunctional applications, for instance, in drug delivery.2 These demixed micelles sometimes form when incompatible binary hydrocarbon and fluorocarbon surfactant molecules coexist in an aqueous solution, thus leading to two populations of selforganized but segregated aggregates with different compositions, hydrocarbon-rich and fluorocarbon-rich.5,6 These demixed micelles have unique properties that are both useful and fundamentally interesting. Mixed hydrocarbon/fluorocarbon templating has received increasing attention in recent years because the two templates provide controlled populations of different nanoscale aggregates. Until now, mixed hydrocarbon/fluorocarbon surfactants have been employed not only to achieve high hydrothermal stability,7 to generate hierarchical pore systems,8,9 to control morphology,10

to explore unknown phase behavior,11 and to tailor the particle size, but also to facilitate the synthesis of novel biphasic materials with long-range ordering. Demixed layers of fluorinated surfactants have been used both as hollow macropore templates12 and for particle size and morphology control.13 Although demixed micellar aggregates have been actively investigated for advanced material design and synthesis, their potential to be selectively swollen with different organic additives has not been reported. Selectively swollen micelles are expected to act as templates for the synthesis of controlled bimodal porous materials. Here, we report a facile synthetic approach to well-defined bimodal porous silica using mixed cetyltrimethylammonium chloride (CTAC) and 1,1,2,2-tetrahydro-perfluorodecylpyridinium chloride (HFDePC) surfactants as templates, and also test the hypothesis that the hydrocarbon-templated and fluorocarbon-templated pores can be finely and independently controlled within the range of 3-7 nm by adding lipophilic or fluorophilic oils, respectively. This pair of surfactants is chosen because its aggregation and mixing behavior have been studied extensively; both surfactants have similar critical micelle concentrations and the same counterions, and they have been shown to demix into CTAC-rich micelles and HFDePC-rich micelles over a range of compositions in dilute solution.14-16

*To whom correspondence should be addressed. E-mail: srankin@ engr.uky.edu. Telephone: 1-859-257-9799. (1) Kotzev, A.; Laschewsky, A.; Adriaensens, P.; Gelan, J. Macromolecules 2002, 35, 1091. (2) Lodge, T. P.; Rasdal, A.; Li, Z.; Hillmyer, M. J. Am. Chem. Soc. 2005, 127, 17608. (3) Xing, R.; Rankin, S. E. Microporous Mesoporous Mater. 2008, 108, 65. (4) Tan, B.; Lehmler, H.-J.; Vyas, S. M.; Knutson, B. L.; Rankin, S. E. Adv. Mater. 2005, 17, 2368. (5) Mukerjee, P. A.; Yang, Y. S. J. Phys. Chem. 1976, 80, 1388. (6) Asakawa, T.; Johten, K.; Miyagishi, S.; Nishida, M. Langmuir 1985, 1, 347. (7) Li, D.; Han, Y.; Song, J.; Zhao, L.; Xu, X.; Xiao, F.-S. Chem.;Eur. J. 2004, 10, 5911. (8) Areva, S.; Boissiere, C.; Grosso, D.; Asakawa, T.; Sanchez, C.; Linden, M. Chem. Commun. 2004, 1630. (9) Groenewolt, M.; Antonietti, M. Langmuir 2004, 20, 7811. (10) Han, Y.; Ying, J. Y. Angew. Chem., Int. Ed. 2005, 44, 288. ::  (11) Wang, K.; Oradd, G.; Almgren, M.; Asakawa, T.; Bergenstahl, B. Langmuir 2000, 16, 1042.

6486

DOI: 10.1021/la9000939

Experimental Section Materials. Cetyltrimethyammonium chloride, CTAC (98%+), and tetraethylorthosilicate, TEOS (98%), were purchased from Sigma. 1H,1H,2H,2H-perfluorodecylpyridinium chloride, HFDePC, was first synthesized by alkylation of pyridine with 1H,1H,2H,2H-perfluorodecyl iodide followed by ion exchange, as described previously.17 The molecular structures of CTAC (12) Djojoputro, H.; Zhou, X. F.; Qiao, S. Z.; Wang, L. Z.; Yu, C. Z.; Lu, G. Q. J. Am. Chem. Soc. 2006, 128, 6320. (13) Han, Y.; Ying, J. Y. Angew. Chem., Int. Ed. 2004, 44, 288. (14) Asakawa, T.; Hisamatsu, H.; Miyagishi, S. Langmuir 1995, 11, 478. (15) Kadi, M.; Hansson, P.; Almgren, M.; Furo, I. Langmuir 2002, 18, 9243. (16) Almgren, M.; Garamus, V. M. J. Phys. Chem. B 2005, 109, 11348. (17) Tan, B.; Dozier, A.; Lehmler, H.-J.; Knutson, B. L.; Rankin, S. E. Langmuir 2004, 20, 6981.

Published on Web 3/26/2009

Langmuir 2009, 25(11), 6486–6492

Xing et al. and HFDePC are shown in Figure 1. Concentrated aqueous ammonia (29 wt % NH4OH, Merck), deionized ultrafiltered (DIUF) water (Fisher Scientific), and NaCl (Merck KGaA) were used as received for material synthesis. For swelling studies, the hydrogenated solvent 1,3,5-trimethylbenzene, TMB (neat), and fluorinated solvent perfluorodecalin, PFD (95 wt%), were purchased from Aldrich. Concentrated aqueous HCl (ACS grade, Fisher Scientific) and anhydrous ethanol (Aaper Alcohol and Chemical) were used for surfactant extraction. Silica Material Synthesis. The synthesis of mesoporous silica materials was carried out in dilute solution of CTAC, HFDePC, and TEOS under mild basic conditions. For all samples, the total molar amount of mixed CTAC/HFDePC surfactants remained constant. The initial molar composition of reactants for the synthesis of silica materials was TEOS/H2O/ HFDePC/CTAC/NH3/NaCl/TMB (or PFD) = 1:148:0.12x: 0.12(1 - x):10:2.8:yTMB (or PFD). The specific initial molar ratios of reactants will be described in the following sections. In a typical synthesis procedure, the calculated amounts of CTAC, HFDePC, and NaCl were first mixed with DIUF water and concentrated aqueous ammonia. The mixture was vigorously stirred at room temperature for at least 30 min to completely dissolve and equilibrate the surfactants. If used, a swelling solvent (TMB or PFD) was then added and stirred for another 2 h to attain equilibrium. The required amount of TEOS was slowly added, and the solution was aged for 24 h at room temperature with gentle stirring (∼100 rpm). To allow direct comparison, the size of the reactor vessel, the stir bar, the stirring speed, and the TEOS addition rate were kept the same for each sample. The precipitate was isolated by filtration and dried in air at room temperature. The mixed surfactants were removed by washing the materials twice with an acidic mixture of 6% concentrated HCl and 94% ethanol. The washing time for each step was 24 h. Characterization. X-ray diffraction (XRD) patterns were obtained with a Siemens 5000 diffractometer using Cu KR radiation (λ = 1.54098 A˚) and a graphite monochromator. Scanning electron microscopy (SEM) studies were performed on a Hitachi S-900 microscope. Solid samples were loaded onto 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 isopropyl alcohol solution by sonication, followed by deposition on 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 analysis. The equilibration time for each pulse of nitrogen was 20 s, and equilibrium at a given pressure was defined by a pressure change of at most 0.01% of the average over an 11-point pressure measurement window. The pore size distributions were calculated using the modified Barrett-Joyner-Halenda (BJH) method of Kruk, Jaroniec and Sayari (the KJS method) from the adsorption branch of the isotherms assuming cylindrical pore shapes for all samples.18,19 The thermal stability was analyzed with a Pyris 1 thermogravimetric analyzer. The as-made samples were analyzed by first drying at 85 °C under vacuum overnight, followed by thermogravimetric analysis (TGA) carried out with a ramp of 10 °C/min in nitrogen from 30 to 1000 °C.

Article

Figure 1. Molecular structures of the surfactants used for materials synthesis.

Figure 2. (a) Nitrogen sorption isotherm plots of the series of samples a-1, a-2 (upshifted 100 cm3/g), and a-3 (upshifted 200 cm3/ g) made with different ratios of CTAC to HFDePC. (b) Pore size distributions of this series of samples after extraction, calculated using the KJS method assuming cylindrical pore geometry.

Results and Discussion Synthesis of Bimodal Mesoporous Silica. A series of samples was prepared with varying molar ratios of CTAC to HFDePC. The initial molar ratio of reactants was TEOS/H2O/ HFDePC/CTAC/NH3/NaCl = 1:148:0.12x:0.12(1 - x):10:2.8,

where x = 1/3, 1/2, or 2/3 for samples a-1 through a-3, respectively. The nitrogen sorption isotherms and calculated pore size distributions of this series of samples are shown in panels a and b, respectively, of Figure 2. All samples exhibit typical reversible type IV isotherms without hysteresis loop.20 For each sample, there are two distinct capillary condensation steps at relative pressure (P/P0) ranges of 0.12-0.25 and 0.300.40, indicating the formation of bimodal mesoporous silica. The pore size distribution of each sample is bimodal with two clear peaks at 3.3 and 3.8 nm, which are sizes consistent with mesopores templated by HFDePC alone and CTAC alone under the same conditions, respectively. In addition, the intensity of each peak is correlated with the relative amounts of the two surfactants (an increasing volume of larger pores is observed as more CTAC is used for templating). All samples showed both high surface area and large pore volume. Some key structure parameters, such as the Brunauer-Emmett-Teller (BET) surface area, total mesopore volume Vp, total surface area St, and external surface area Sex, were calculated as shown in Table 1.

(18) Barrett, E. P.; Joyner, L. G.; Halenda, P. J. Am. Chem. Soc. 1951, 73, 373. (19) Sayari, A.; Liu, P.; Kruk, M.; Jarioniec, M. Chem. Mater. 1997, 9, 2499.

(20) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L. R.; Pierotti, A.; Rouquerol, J.; Siemieniewska, T. Pure Appl. Chem. 1985, 57, 603.

Langmuir 2009, 25(11), 6486–6492

DOI: 10.1021/la9000939

6487

Article

Xing et al.

Table 1. Structure Parameters of the Mixed-Surfactant-Templated Mesoporous Silica Materials Derived by Analysis of Nitrogen Sorption Dataa name

SBET (m2/g)

Vp (cm3/g)b

V@P/P0=0.95 (cm3/g)

Vm (cm3/g)b

WKJS (nm)

St (m2/g)b

Sex (m2/g)b

a-1 a-2 a-3 b-1 b-2 b-3 b-4 b-5 c-1 c-2 c-3 c-4 c-5

809.4 0.65 0.69 0.022 3.3, 3.8 862.3 46.1 934.2 0.63 0.70 0.025 3.3, 3.8 788.8 44.1 865.8 0.69 0.73 0.018 3.3, 3.8 872.4 48.2 871.9 0.65 0.71 0.019 3.3, 4.2 731.1 42.0 890.5 0.75 0.79 0.014 3.4, 4.7 744.7 24.7 900.2 0.86 0.91 0.022 3.7, 5.2 770.3 35.6 938.7 1.04 1.11 0.031 3.7, 6.5 809.6 46.5 882.4 0.87 1.00 0.020 3.2, 6.5 761.6 92.5 859.7 0.66 0.71 0.011 3.9 713.2 33.4 857.2 0.69 0.74 0.024 3.9 732.3 35.2 854.1 0.70 0.78 0.0061 3.9, 5.2 714.5 36.0 852.1 0.74 0.80 0.011 3.9, 5.8 716.1 39.1 863.2 0.75 0.79 0.0067 3.9, 5.6 718.0 28.2 a 24 SBET = BET surface area, VP/P0=0.95 = the total adsorbed 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, and Sex = external specific surface area. b Calculated using Rs comparative nitrogen adsorption plots.25

Figure 3a shows a representative TEM image of sample a-2. As can be seen, the silica particles are composed of wormholelike pores with sizes ranging from ∼3 to 4 nm. The wormholelike pores seem to be completely interconnected in a sponge-like manner. While the pore size distribution cannot readily be determined by TEM, Figure 3a is consistent with pore templating in this material by a mixture of wormlike micelles, where the micelles differ in chemical composition of surfactants (as suggested by the bimodal pore size distribution in Figure 2). By TEM, separate ordered domains with constant pore size could not be found despite extensive searching. This suggests that the bimodal mesopores templated by either CTAC-rich or HFDePC-rich micelles coexist within one large particle. Figure 3b shows a representative SEM image of this sample, which shows that it consists of individual globular particles with uniform sizes of ∼100 nm that have flocculated into larger grapelike aggregates. However, there is no indication of segregation of HFDePC and CTAC into precipitated particles with an obvious difference in size or shape, suggesting that the surfactants are uniformly distributed among particles. The observation of disordered wormhole-like pore arrays is due not only to the mixture of surfactants used but also to the synthesis conditions, especially the salt concentration. Adding NaCl during mesoporous silica synthesis is known to reduce longrange order for a single cationic surfactant such as CTAB,21 but in the case of a 1:1 mixture of CTAB and HFDePC it also causes a change in overall morphology from hexagonal mesoporous silica with large hollow voids to the mixed disordered pore structure discussed here. The salt and surfactant ratio effects will be discussed in a separate contribution from our group. Tuning of Bimodal Mesopore Size Distribution. Samples a-1 through a-3 show that bimodal mesoporous silica can be successfully prepared with different molar ratios of CTAC to HFDePC in the presence of NaCl. The bimodal mesopore size distribution is consistent with preserving separate populations of CTAC-rich and HFDePC-rich micelles in solution and after templating. In contrast to the hydrogenated core of CTAC-rich micelles, the fluorinated core of HFDePC is not only hydrophobic but also lipophobic. Thus, differing affinities of organic addictives in the cores of the two types of micelles provide opportunities to independently tune the bimodal pore size distribution by using a “controlled micelle swelling” approach. Figure 4 shows the hypothesized selective swelling schemes for CTAC-rich and HFDePC-rich micelles by using different (21) Yu, J.; Shi, J.-L.; Chen, H. R.; Yan, J.-N.; Yan, D.-S. Microporous Mesoporous Mater. 2001, 46, 153.

6488

DOI: 10.1021/la9000939

Figure 3. Representative (a) TEM and (b) SEM images of sample a-2 after extraction.

Figure 4. Schematic representation of proposed selective swelling of demixed micelles composed of CTAC-rich and HFDePC-rich surfactant in the initial sols.

organic solvents. For example, TMB, a hydrogenated oil, is expected to preferentially dissolve in the hydrocarbon core of CTAC-rich micelles and to swell them.22,23 On the other hand, PFD, a fluorophilic oil, is expected to preferentially swell HFDePC-rich micelles. Assuming that this preferential swelling occurs also in the silica/surfactant particles that precipitate (22) Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.; Chmelka, B. F.; Stucky, G. D. Science 1998, 279, 548. (23) Ottaviani, M. F.; Moscatelli, A.; Desplantier-Giscard, D.; DiRenzo, F.; Kooynab, P.; Galarneau, A. J. Phys. Chem. B 2004, 108, 12123. (24) Brunauer, S.; Emmett, P. H.; Teller, E. J. J. Am. Chem. Soc. 1938, 60, 309. (25) Kruk, M.; Jaroniec, M. Langmuir 1999, 15, 5410.

Langmuir 2009, 25(11), 6486–6492

Xing et al.

from the solution, tuning of the bimodal pore size distribution will result. To test this hypothesis, samples with different amounts of TMB are first discussed. The initial composition of the reactants is TEOS/H2O/HFDePC/CTAC/NH3/NaCl/TMB = 1:148:0.06: 0.06:10:2.8:y. In this section, we discuss a series of samples prepared with y = TMB/TEOS = 0, 0.054, 0.228, 0.3, 0.44, and 0.61 for samples a-2 and b-1 through b-5, respectively. Figure 5 shows nitrogen sorption isotherms and calculated pore size distributions for this series of samples after surfactant extraction. All samples have type IV isotherms, and samples b-1 through b-5 have type H2 hysteresis loop at a relative pressure range of 0.35-0.7. Two capillary condensation steps can be clearly identified, indicating that all samples exhibit a bimodal mesopore distribution. Sample b-1, prepared with a minimum amount of TMB, shows a shift of the second inflection to higher relative pressure (approaching P/P0 = 0.4) and a small hysteresis loop. This suggests that TMB molecules swell CTAC-rich micelles, which leads to nonuniform pores or branching and hence to the observation of type H2 hysteresis.22 With more TMB loaded, the second adsorption step gradually shifts to the even higher relative pressure of 0.65. The hysteresis loop grows in area and is associated with the second adsorption step, which is consistent with increased swelling of CTAC-rich micelle templated pores. With increasing TMB addition, the relative pressure corresponding to the first adsorption step increases only a little, indicating that TMB has little solubility inside the fluorocarbon cores. The calculated pore size distributions confirm the bimodal mesoporous distribution for all samples. Figure 6 compares the change of both pore sizes and the corresponding pore volumes for this series of samples. This figure clearly shows that the CTAC-rich micelle templated pore diameter increases from 3.8 to 6.5 nm as the loading amount of TMB increases, while the HFDePC-rich templated pore diameter increases only slightly. Correspondingly, the CTACrich micelle templated pore volume increases as the amount of TMB increases while the HFDePC-rich micelle templated pore volume remains constant. The plateau in the degree of swelling suggests that the capacity to incorporate TMB into CTAC-rich micelles is about 0.44 mol (per TEOS). Excessive TMB appears to only broaden the CTAC-templated pore size distribution (sample b-5). Other structure parameters of this series of samples are compared in Table 1. The total mesopore volume (Vp) increases from 0.63 cm3/g to a maximum of 1.04 cm3/g as the amount of TMB increases. The BET surface area SBET, Vp, and St all reach maxima in sample b-4. An excessive amount of TMB (beyond 0.44 TMB/TEOS) decreases the pore volume and total surface area of the final products because the porous network becomes nonuniform, which may make the material susceptible to partial pore collapse during drying and calcination. The thermal behavior of as-made samples a-2 and b-3 were determined by TGA as shown in Figure 7. For direct comparison, TGA results of samples templated by CTAC alone and HFDePC alone are also presented in this figure. For samples a-2 and b-3, TGA curves show four main stages of weight loss. The first weight loss that occurs below 120 °C is associated with the loss of water absorbed due to the exposure of the samples to air during the sample preparation. The second weight loss region (120-250 °C) corresponds to the degradation of both cationic head groups, that is, the pyridinium ring of HFDePC and perhaps the trimethylamonium of CTAC (although little weight loss is observed at this stage for materials templated with CTAC alone). Compared to a-2, the large weight loss in the second stage of b-3 is attributed to the additional degradation and Langmuir 2009, 25(11), 6486–6492

Article

Figure 5. (a) Nitrogen sorption isotherm plots of the series of samples a-2, b-1 (upshifted 100 cm3/g), b-2 (upshifted 200 cm3/g), b-3 (upshifted 300 cm3/g), b-4 (upshifted 400 cm3/g), and b-5 (upshifted 600 cm3/g) made with addition of different amounts of TMB. Dashed lines indicate the shift of capillary condensation steps corresponding to CTAC-templated mesopores and HFDePC-templated mesopores. (b) Pore size distributions of this series of samples after extraction, calculated using the KJS method assuming cylindrical pore geometry.

Figure 6. Calculated pore width (WKJS) and mesopore volume V as a function of the amount of TMB added to samples a-2 and b-1 through b-5.

evaporation of TMB, suggesting that TMB is incorporated into as-made silica particles. The third weight loss region (250-310 °C) is associated with the removal of CH2 and CF2 groups of CTAC and HFDePC. However, it is not clear which group starts to degrade earlier. In the final weight loss region (after 310 °C), the hydrocarbon and fluorocarbon residues may continue to slowly degrade, and continued silica curing may lead to loss of water produced by condensation. DOI: 10.1021/la9000939

6489

Article

Xing et al.

Figure 7. TGA of as-made samples a-2 and b-3. For direct comparison, samples prepared with CTAC alone and HFDePC alone are also shown.

Figure 8 shows the XRD patterns of this series of samples as a function of the amount of TMB. All samples have broad peaks indicating poor long-range order. For samples a-2 and b-1, the broad peak includes two reflections, which correspond to the average pore-pore distance of CTAC-rich micelle templated pores and HFDePC-rich micelle-templated pores, respectively. With the addition of TMB, the reflection of CTAC-templated pore shifts to a lower 2θ value due to the swelling effect, and eventually the reflection moves outside of the detected range, whereas the reflection of HFDePC-templated pores shifts only slightly to lower 2θ as shown in samples b-2 through b-4. In order to see how the addition of TMB affects the morphology of mesopores, sample b-4 (with the second largest TMB loading) was studied using TEM and scanning TEM (STEM) as shown in the Supporting Information. Figure S-1a (Supporting Information) shows a representative TEM image of sample b-4, illustrating nonuniform and disordered pores throughout the entire sample. The pore sizes appear to be consistent with the range determined from the nitrogen sorption isotherm of this sample. Figure S-1b (Supporting Information) shows a darkfield scanning transmission electron micrograph collected with a high resolution, 2 nm radius probe. The depth of field for this image is between 50 and 100 nm. The STEM image shows 3D interconnected spherical and disordered wormhole-like pores within a single particle. Due to the small difference in pore sizes between CTAC-templated mesopores and HFDePC-templated mesopores, it is difficult to clearly estimate the pore size distribution from these images. The effects of the amount of TMB loaded into this series of samples on the morphology of the particles is studied using SEM. Similar to sample a-2 in Figure 3b, all samples are composed of smooth round particles that are flocculated into large aggregates. However, the particles’ shape and size change as more TMB is added. Increasing the amount of TMB causes the morphology to change from uniformly spherical particles to irregular particles, with an accompanying increase in particle size. The increase of particle size can be explained by a reduction in the rate of precipitation as TMB is added, perhaps because TEOS associates with excess TMB. For this series of samples, 6490

DOI: 10.1021/la9000939

Figure 8. XRD patterns for the series of extracted samples a-2 and b-1 through b-4.

it is also qualitatively observed that the onset of turbidity starts later and the yield of particles decreases as more TMB is added. Some representative SEM images of samples b-2 and b-4 are shown in the Supporting Information (Figure S-2). In addition to adding TMB, the selective swelling of fluorophilic domains with PFD is investigated. The composition of the initial reactants is TEOS/H2O/HFDePC/CTAC/NH3/ NaCl/PFD = 1:148:0.06:0.06:10:2.8:y. In this section, we discuss a series of samples prepared with y = PFD/TEOS = 0, 0.02, 0.36, 0.56, 0.86, and 1.08 for samples a-2 and c-1 through c-5, respectively. Figure 9 shows nitrogen sorption isotherms and pore size distributions for this series of extracted samples. All samples have type IV isotherms. Samples a-2, c-3, c-4, and c-5 clearly show two adsorption steps, while samples c-1 and c-2 appear to have only one condensation step. Two dashed lines on the isotherm plots indicate the shift of capillary condensation steps corresponding to CTAC-templated mesopores and HFDePCtemplated mesopores. Without the addition of PFD, the size of pores templated by HFDePC-rich micelles is below that of pores templated by CTAC-rich micelles. However, as PFD is added, this inflection shifts to larger relative pressure, and eventually an inflection at a relative pressure larger than the step for CTACtemplated pores appears. The capillary condensation step corresponding to CTAC-templated mesopores shifts only slightly to higher relative pressure as PFD is added. This suggests that the added PFD preferentially swells the fluorinated HFDePCrich micelles and that these swollen micelles are preserved in the templated materials. It is difficult to be sure that swelling of the HFDePC-templated pores is responsible for the overlap of pore sizes in samples c-1 and c-2 (rather than surfactant mixing), but a separate control experiment showed that PFD does not significantly swell the wormhole-like pores of silica templated with CTAC alone and prepared under the conditions used here (results not shown for brevity). Along with the crossover of the two dashed lines due to the preferential swelling of fluorinated micelles by PFD, type H2 hysteresis loops start to appear Langmuir 2009, 25(11), 6486–6492

Xing et al.

Article

Figure 10. Calculated pore width (WKJS) and volume (V) as a function of amount PFD added to samples a-2 and c-1 through c-5.

Figure 9. (a) Nitrogen sorption isotherm plots of the series of samples a-2, c-1 (upshifted 150 cm3/g), c-2 (upshifted 300 cm3/g), c3 (upshifted 450 cm3/g), c-4 (upshifted 550 cm3/g), and c-5 (upshifted 700 cm3/g) made with addition of different amounts of PFD. Dashed lines indicate the shift of capillary condensation steps corresponding to CTAC-templated mesopores and HFDePC-templated mesopores. (b) Pore size distributions of this series samples after extraction, calculated using the KJS method assuming cylindrical pore geometry.

in sample c-1 and gradually increase in area as PFD distorts the final pore channels. The calculated pore size distributions confirm our analysis of the isotherms. Figure 10 compares the sizes and volumes of each population of pores in this series. Adding PFD increases the diameter of HFDePC-rich micelle-templated pores in the range from 3.3 to 5.8 nm, whereas the CTAC-templated pore diameter only increases slightly, from 3.7 to 3.9 nm. Correspondingly, the HFDePC-rich micelle-templated pore volumes increase while the CTAC-rich micelle templated pore volumes remain almost constant as PFD increases. From this series of samples, it was concluded that only a small amount of PFD might be incorporated into CTAC-rich micelles, while almost 0.86 mol of PFD (per TEOS) could be incorporated into HFDePC-rich micelles. In addition, the peak of the pore size distribution in sample c-2 in Figure 10b is sharper than that in sample c-1, indicating that the pore size of HFDePC-templated mesopores most closely matches that of CTAC-rich templated mesopores when 0.36 mol of PFD is used. The other pore properties are summarized in Table 1. The main observation is that the total mesopore volume increases from 0.6 to 0.74 cm3/g as more PFD is used in the synthesis, mainly due to HFDePC-rich micelle swelling. XRD patterns for members of this series after extraction are shown in Figure 11. All samples show a poorly ordered wormhole-like structure. In contrast to the sample a-2, the addition of a small amount of PFD improves the ordering, as indicated by more intense primary and secondary reflections in samples c-1 and c-2. The strong peaks at the lowest angle represent the Langmuir 2009, 25(11), 6486–6492

Figure 11. XRD patterns of samples a-2 and c-2 through c-4 after extraction.

average pore-pore distance and gradually shift to lower 2θ as more PFD is used. Presumably, the strong peak in each sample includes two reflections, from mesopores templated by CTACrich micelles and HFDePC-rich micelles. The position of the two reflections is too close to be resolved for samples a-2 and c-1 through c-2, however. Continued swelling of HFDePC-rich micelles eventually shifts the reflection from mesopores templated by HFDePC-rich micelles to an angle low enough to resolve, as shown in sample c-3. The lower-angle peaks of sample c-3 give a spacing of 5.3 nm between HFDePC-rich micelle templated pores, and the higher peak indicates 4.6 nm between CTAC-rich micelle-templated pores. These results are consistent with the calculated pore size distributions. The pore structure and morphology of particles were examined by TEM and SEM. Generally, the particle size increases as more PFD is used, and the pore structure of this series of samples shows poorly ordered wormhole-like pores. Representative TEM and SEM images of sample c-3 are shown in the Supporting Information (Figure S-3). DOI: 10.1021/la9000939

6491

Article

Xing et al.

Conclusions Well-defined bimodal mesoporous silica was synthesized by cooperative assembly of hydrolyzed tetraethoxysilane with mixtures of CTAC and HFDePC. The bimodal pore size could be finely and independently tailored by adding lipophilic or fluorophilic oils. As hypothesized, the results are consistent with trimethylbenzene selectively partitioning into and swelling the CTAC-rich micelles and perfluorodecalin preferentially swelling the HFDePC-rich micelles. This led to selective increases in the sizes and volumes of pores templated by each type of surfactant. The principle demonstrated here can be used to selectively incorporate active metal or metal oxide clusters inside the pore channels, providing new possibilities for creating heterogeneous bifunctional catalysts especially useful in performing reactions in

6492

DOI: 10.1021/la9000939

series. In addition, the materials prepared here may provide unique molecular sieving opportunities in materials with high accessibility, pore volumes, and surface areas. Acknowledgment. This report is based on work partially funded by the National Science Foundation (NSF) under Grant No. DMR-0210517 and Grant No. CBET-0348234. We thank Prof. W. S. Winston Ho at The Ohio State University for providing access to the TGA instrument. Supporting Information Available: Additional electron micrographs of selected samples discussed in the text. This material is available free of charge via the Internet at http:// pubs.acs.org.

Langmuir 2009, 25(11), 6486–6492