Tuning Stability of Mesoporous Silica Films under Biologically

Xinxin Li, Sutapa Barua, Kaushal Rege, and Bryan D. Vogt*. Department of Chemical Engineering, Arizona State UniVersity, Tempe, Arizona 85284. ReceiVe...
1 downloads 0 Views 2MB Size
Download PDF
Langmuir 2008, 24, 11935-11941

11935

Tuning Stability of Mesoporous Silica Films under Biologically Relevant Conditions through Processing with Supercritical CO2 Xinxin Li, Sutapa Barua, Kaushal Rege, and Bryan D. Vogt* Department of Chemical Engineering, Arizona State UniVersity, Tempe, Arizona 85284 ReceiVed June 12, 2008. ReVised Manuscript ReceiVed July 30, 2008 Mesoporous materials have been proposed for use in numerous biological environments such as substrates for cell culture and controlled release for drug delivery. Although mesoporous silica synthesis is facile, recent reports (Dunphy et al. Langmuir 2003, 19, 10403; Bass et al. Chem. Mater. 2007, 19, 4349) have demonstrated instability (dissolution) of pure mesoporous silica films under biologically relevant conditions. In this work, we demonstrate a simple processing handle (pressure) to control the dissolution of mesoporous silica films that are synthesized using preformed template films and supercritical CO2. Spectroscopic ellipsometry is utilized to quantify changes in both the film thickness and porosity; these properties provide insight into the dissolution mechanism. The pore size increases as the films are exposed to phosphate-buffered saline (PBS) through preferential dissolution at the pore wall in comparison to the film surface; a mechanism reminiscent of bulk erosion of scaffolds for drug delivery. Thin mesoporous silica film lifetimes can be extended from several hours using traditional sol-gel approaches to days by using CO2 processing for identical film thickness. Osteoblast attachment and viability on these films was found to correlate with their increased stability. This enhanced stability opens new possibilities for the utilization of mesoporous silica for biological applications, including drug delivery and tissue engineering.

Introduction Mesoporous silicas have fascinated scientists and engineers ever since their reported synthesis by Mobil in 1992.1 Sol-gel chemistry2 has proved to be an effective means to the synthesis of templated mesostructured films3,4 with tunable physical properties.5 The physicochemical properties of mesoporous silicas including pore wall chemistry,6 pore morphology,7 porosity,8 pore size distribution9 and mechanical robustness10 define which materials are viable for a given application including photovoltaic devices,11 catalysis,12 separation membranes,13 sensors,14 and low-k dielectric for microelectronics.15 In addition to these more traditional applications, mesoporous silicas have been proposed * To whom correspondence should be addressed. E-mail: bryan.vogt@ asu.edu. (1) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359(6397), 710–712. (2) Brinker, C. J.; Scherer, G. W. Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing; Academic Press: San Diego, 1990. (3) Lu, Y. F.; Ganguli, R.; Drewien, C. A.; Anderson, M. T.; Brinker, C. J.; Gong, W. L.; Guo, Y. X.; Soyez, H.; Dunn, B.; Huang, M. H.; Zink, J. I. Nature 1997, 389(6649), 364–368. (4) Nicole, L.; Boissiere, C.; Grosso, D.; Quach, A.; Sanchez, C. J. Mater. Chem. 2005, 15(35-36), 3598–3627. (5) Sanchez, C.; Boissiere, C.; Grosso, D.; Laberty, C.; Nicole, L. Chem. Mater. 2008, 20(3), 682–737. (6) Yang, P. D.; Zhao, D. Y.; Margolese, D. I.; Chmelka, B. F.; Stucky, G. D. Nature 1998, 396(6707), 152–155. (7) Alberius, P. C. A.; Frindell, K. L.; Hayward, R. C.; Kramer, E. J.; Stucky, G. D.; Chmelka, B. F. Chem. Mater. 2002, 14(8), 3284–3294. (8) Imperor-Clerc, M.; Davidson, P.; Davidson, A. J. Am. Chem. Soc. 2000, 122(48), 11925–11933. (9) Fang, H.; Zhang, M.; Shi, W. H.; Wan, T. L. J. Non-Cryst. Solids 2006, 352(21-22), 2279–2283. (10) Fan, H.; Hartshorn, C.; Buchheit, T.; Tallant, D.; Assink, R. A.; Simpson, R.; Kissel, D. J.; Lacks, D. J.; Torquato, S.; Brinker, C. J. Nat. Mater. 2007, 6, 418–423. (11) Coakley, K. M.; McGehee, M. D. Appl. Phys. Lett. 2003, 83(16), 3380– 3382. (12) Corma, A. Chem. ReV. 1997, 97(6), 2373–2419. (13) Lai, Z. P.; Bonilla, G.; Diaz, I.; Nery, J. G.; Sujaoti, K.; Amat, M. A.; Kokkoli, E.; Terasaki, O.; Thompson, R. W.; Tsapatsis, M.; Vlachos, D. G. Science 2003, 300(5618), 456–460. (14) Stein, A. AdV. Mater. 2003, 15(10), 763–775. (15) Yang, S.; Mirau, P. A.; Pai, C.-S.; Nalamasu, O.; Reichmanis, E.; Lin, E. K.; Lee, H.-J.; Gidley, D. W.; Sun, J. Chem. Mater. 2001, 13, 2762–2764.

for use in biological applications such as drug release,16 bioadsorption and biosensing,17 bone tissue regeneration, and environmental remediation.18 For drug delivery, the pore size and geometry structure are critical for the design of controlled release delivery systems.16,19 These initial results in the biological arena suggest that mesostructured materials have significant potential in biological and biomedical applications. However, one drawback of mesoporous silica films in biological applications is their limited hydrolytic stability.20-22 Bass et al. reported that templated silica films with spherical pores lose mechanical stability after only 90 min exposure to phosphate buffered saline (PBS) due to dissolution of the silica.23 The dissolution rate can be decreased by ∼40% by increasing the calcination temperature, which also increases the film density.23 Alternatively, addition of organic or other transition metal oxides into the mesoporous silica framework during material synthesis can increase the film stability on the order of several days.20-24 However, this synthetic change alters the composition and chemistry of the pore wall, which also impacts the biological interaction with the material. Since processing (calcination temperature and sol-gel composition) impact the stability of mesoporous silica films, examination of several alternative methods for the fabrication of mesoporous silica films outside of the typical one pot synthesis may yield further improvements to their biostability. These routes utilize preformed templates and include delivery of the silica (16) Vallet-Regi, M.; Ramila, A.; del Real, R. P.; Perez-Pariente, J. Chem. Mater. 2001, 13(2), 308–311. (17) Hartmann, M. Chem. Mater. 2005, 17(18), 4577–4593. (18) Mercier, L.; Pinnavaia, T. J. AdV. Mater. 1997, 9(6), 500–&. (19) Doadrio, A. L.; Sousa, E. M. B.; Doadrio, J. C.; Pariente, J. P.; IzquierdoBarba, I.; Vallet-Regi, M. J. Controlled Release 2004, 97(1), 125–132. (20) Mokaya, R. J. Phys. Chem. B 2000, 104(34), 8279–8286. (21) Das, D.; Tsai, C. M.; Cheng, S. F. Chem. Commun. 1999, (5), 473–474. (22) Chen, L. Y.; Jaenicke, S.; Chuah, G. K. Microporous Mater. 1997, 12(4-6), 323–330. (23) Bass, J. D.; Grosso, D.; Boissiere, C.; Belamie, E.; Coradin, T.; Sanchez, C. Chem. Mater. 2007, 19(17), 4349–4356. (24) Dunphy, D. R.; Singer, S.; Cook, A. W.; Smarsly, B.; Doshi, D. A.; Brinker, C. J. Langmuir 2003, 19(24), 10403–10408.

10.1021/la801849n CCC: $40.75  2008 American Chemical Society Published on Web 09/17/2008

11936 Langmuir, Vol. 24, No. 20, 2008

precursor to the template via its vapor pressure,25 supercritical CO226 or a simple sol-gel.27 These synthetic routes provide new processing opportunities to enable improved long-range order,28 perpendicular alignment of cylindrical mesopores,29 and direct device level patterning.30 Recently, we reported that the morphology of silica films synthesized using supercritical carbon dioxide (CO2) can be tuned simply by fabricating the mesoporous films at different CO2 pressures.31 Additionally, water condensation within the mesopores is inhibited for films synthesized at pressures just above the critical pressure for CO2. This property suggests that these mesoporous silica films will respond differently under biologically aqueous conditions in comparison to mesoporous silica films fabricated using evaporation induced selfassembly.32 Here we report on the impact of CO2 pressure during synthesis on the stability of mesoporous silica films.

Experimental Section Materials and Synthesis. Mesoporous silica films were synthesized using preformed polymeric template films prepared from solutions containing poly(ethylene oxide)-b-poly(propylene oxide)b-poly(ethylene oxide) (Pluronic F108, BASF), poly(p-hydroxystyrene) (PHOSt, Mn ) 8000 g/mol, DuPont Electronic Materials), and p-toluenesulfonic acid (p-TSA, Aldrich) dissolved in a mixture of ethanol (Aldrich) and deionized water. Films from these solutions were formed by spin coating (2000 rpm, 55 s) on clean silicon wafers. These films were reactively modified at 60 °C by exposure to tetraethylorthosilicate (TEOS, 10 µL) and deionized water (20 µL) either by their vapor pressure or by dissolution in carbon dioxide as described previously.31 For what we term vapor phase condensation, no CO2 was added and the reaction was allowed to proceed via the vapor pressure of the precursors. Otherwise, the vessel was pressurized slowly with CO2 to desired pressure. The template was removed via calcination at 450 °C for 5 h at a heating rate of 1 °C/min in air to yield a mesoporous silica film. As determined previously, the films synthesized with F108 and PHOSt templates exhibit a BCC structure (Im3jm).31 Aqueous stability was determined by immersion of the calcined silica samples in pH 7.4 PBS solutions (140 mM NaCl, 10 mM Na2HPO4, and 2.7 mM KCl). At regular intervals, the samples were removed from their solutions, washed with 5 mL water and dried at 120 °C for 30 min. The structure was then interrogated using spectroscopic ellipsometry. Characterization. Film thickness and refractive index were determined using a UV-visible-NIR (240-1700nm) variable angle spectroscopic ellipsometer (VASE M-2000, J.A. Woollam Co.). To fit the data, the Cauchy model was utilized for the optical properties of the polymeric template, the silica-polymer nanocomposite after reaction and the mesoporous film. The Bruggemann effective medium approximation (BEMA) model based on a two component system of amorphous silica and voids (air) was used to calculate the film porosity (P). A fixed refractive index for glass silica skeleton33 was assumed for these calculations. (25) Tanaka, S.; Nishiyama, N.; Oku, Y.; Egashira, Y.; Ueyama, K. J. Am. Chem. Soc. 2004, 126(15), 4854–4858. (26) Pai, R. A.; Humayun, R.; Schulberg, M. T.; Sengupta, A.; Sun, J. N.; Watkins, J. J. Science 2004, 303, 507–510. (27) Hayward, R. C.; Chmelka, B. F.; Kramer, E. J. AdV. Mater. 2005, 17(21), 2591–+. (28) Tirumala, V. R.; Pai, R. A.; Agaarwal, S.; Testa, J. J.; Bhatnagar, G.; Romany, A. H.; Chandler, C.; Gorman, B. P.; Jones, R. L.; Lin, E. K.; Watkins, J. J. Chem. Mater. 2007, 19(24), 5868–5874. (29) Nagarajan, S.; Li, M.; Pai, R. A.; Bosworth, J. K.; Busch, P.; Smilgies, D. M.; Ober, C. K.; Russell, T. P.; Watkins, J. J. AdV. Mater. 2008, 20(2), 246– 251. (30) Nagarajan, S.; Bosworth, J. K.; Ober, C. K.; Russell, T. P.; Watkins, J. J. Chem. Mater. 2008, 20(3), 604–606. (31) Li, X. X.; Vogt, B. D. Chem. Mater. 2008, 20(9), 3229–3238. (32) Brinker, C. J.; Lu, Y. F.; Sellinger, A.; Fan, H. Y. AdV. Mater. 1999, 11(7), 579–585. (33) Palik, E. D. Handbook of optical constants; Academic Press: Orlando, 1985; Vol. 1.

Li et al. The pore size distribution (PSD) of the film was determined using VASE34 utilizing toluene (Aldrich) as the probe solvent. Condensation of toluene within the pores results in a significant increase in refractive index. Both adsorption and desorption isotherms were measured. To calculate the PSD, the data were analyzed on the basis of the change in refractive index as a function of relative pressure.34 A Bruker D8 diffractometer with Cu KR (λ ) 0.1542 nm) radiation in a θ/2θ geometry was utilized to characterize the structure of the mesoporous films over an angular range between 0.5° to 5° with an increment of 0.01°. The primary d-spacing was calculated using Bragg’s law while correcting for the critical angle of the sample due to the low diffraction angles for the mesopores.35 TEM micrographs were obtained using JEOL 2010F operating at an accelerating voltage of 200 keV. TEM cross sections were prepared by manual polishing of a cut section of the film/substrate. Cell Culture. MC3T3 murine osteoblasts were cultured in a 5% CO2 incubator in Dulbeccos’s Modified Eagle’s Medium (DMEM; BioWhittaker) containing 4.5 g/L glucose and L-glutamine, supplemented with 10% fetal bovine serum (Invitrogen, CA) and 1% penicillin/streptomycin (Invitrogen, CA). Mesoporous silica films were placed in individual wells of a 24 well plate following which osteoblasts were added at a density of 75 000 cells/film for 24 h. The medium was removed from the wells and films were lightly washed with PBS phosphate-buffered saline (PBS) to remove floating cells that were not well adhered. Cells were stained with 150 µL of a solution containing 2 µM calcein-AM (Invitrogen) and 4 µM ethidium homodimer-1 (Invitrogen) dyes that stain living and dead cells, respectively. Following incubation at room temperature for 45 min, the films were mounted on microscope cover glass (Fisher) and visualized using a Zeiss Axio Observer D1 (Zeiss, Germany) fluorescence microscope. The microscope was equipped with an epifluorescence setup; excitation/emission setting of 488/530 and 530/580 nm were used to image living and dead cells, respectively. Two different films were imaged for each synthesis condition and at least three different fields of view were imaged for every film.

Results and Discussion For a baseline comparison, a mesoporous silica film synthesized utilizing the vapor pressure of TEOS is examined; we suspect that the behavior of this film will mimic those examined by Bass et al.23 using evaporation induced self-assembly to synthesize mesoporous silica films. The impact of exposure to PBS on the structure of the mesoporous silica film was assessed using ellipsometry. Figure 1 shows the relative film thickness change in comparison to the initial film thickness (640 nm) for the vapor phase condensation synthesis. Over the first 90 min in PBS, there is no significant change in film thickness but the porosity increases tremendously from approximately 31% to greater than 70%, suggesting that dissolution occurs within the film at the pore wall/solution interface; this is similar to a bulk erosion process that occurs in some polymeric materials for drug delivery.36 This mechanism would act to progressively increase the pore size. After ∼1.5 h of exposure to PBS, the film thickness begins to decrease rapidly, while the porosity only changes marginally. This suggests a change in the mechanism to surface erosion37 where dissolution of the films occurs at the film surface/solution interface. After 3 h, the porosity of the film decreases as the film thickness continues to decrease. This suggests a transport limitation during the initial ‘bulk erosion’ that results in a gradient in porosity through the film with larger pores initially near the (34) Baklanov, M. R.; Mogilnikov, K. P.; Polovinkin, V. G.; Dultsev, F. N. J. Vac. Sci. Technol. B 2000, 18(3), 1385–1391. (35) Tanaka, S.; Tate, M. P.; Nishiyama, N.; Ueyama, K.; Hillhouse, H. W. Chem. Mater. 2006, 18(23), 5461–5466. (36) Vogel, B. M.; Mallapragada, S. K. Biomaterials 2005, 26(7), 721–728. (37) Quick, D. J.; Macdonald, K. K.; Anseth, K. S. J. Controlled Release 2004, 97(2), 333–343.

Tuning Stability of Mesoporous Silica Films

Langmuir, Vol. 24, No. 20, 2008 11937

Figure 1. (a) Fraction silica remaining in the film (•) synthesized at vapor phase condensation and the relative thickness of the film (0) as a function of exposure to PBS at 37 °C and (b) porosity of the film determined with ellipsometry. The line is provided as a guide to the reader.

Figure 2. TEM image of a F108 templated silica mesoporous film synthesized using vapor phase condensation (a) initially after fabrication and (b) after immersion in PBS buffer at 37 °C for 1 h. A significant increase in the pore size, especially near the free surface, is observed.

free surface in comparison to the pores near the substrate. As the film dissolves from the surface, progressively smaller pores comprise the film; hence decrease in the apparent film porosity. From the changes in the porosity and film thickness, the relative fraction of silica remaining is calculated. After 3 h at 37 °C, ellipsometric analysis indicates that 90% of the silica film was dissolved. The stability for this pure silica film templated with Pluronic F108 with vapor phase reaction is similar to the film templated with Pluronic F127 determined by Bass at al.23 Both films had a BCC morphology, thus this suggests that the structural details between the vapor phase reaction and EISA films would be similar. However, previous studies have indicated a surface capping layer for films synthesized with vaporized precursors.35 Figure 2 clearly shows a surface capping layer (50 nm) on top of the mesoporous films formed by the vapor exposure. After the film is immersed in PBS at 37 °C for 1 h, the thickness of the film is only slightly reduced with the capping layer decreasing to 20-30 nm thick. Interestingly, the pores near the free surface are significantly enlarged from 3.6 to 9.1 nm in diameter, but the diameter of the pores in the lower part of the film only slightly increases to 5.4 nm. The gradient in pore sizes during the film dissolution predicted from the changes in porosity and film thickness during exposure to PBS buffer is thus confirmed by TEM. The ellipsometry data can be utilized to estimate the dissolution kinetics from the rate of silica loss from the film

(Figure 1a). Previously, Bass et al. used a simple reaction limited dissolution model based on the surface area of the silica to extract a first order rate constant.38 However, the dissolution of mesoporous silica results in gradients in the pore size (Figure 2), thus diffusion limitations are present and a reaction-limited model is not valid for these films. For this reason, no reaction kinetic parameters are extracted from this data set. To quantify as to how the pore size in mesoporous silica films changes after exposed to PBS, adsorption and desorption isotherms were obtained by monitoring changes in the refractive index of the film with ellipsometry upon exposure to varying partial pressure of toluene. The sorption behavior of toluene in the film synthesized with vapor phase condensation upon exposure to PBS for 0, 15, 60, and 150 min is shown in Figure 3. In Figure 3a, the steep increase in the refractive index due to filling of mesopores by capillary condensation is seen to occur at progressively higher partial pressures upon exposure. The PSD is calculated from a modified form of the Kelvin equation, rp ) rk + t, where rk is Kelvin radius and t is the thickness of the adsorbed layer of vapor in the pores before capillary condensation occurs.39 Upon exposure to PBS, the average radius of the pores (38) Bass, J. D.; Grosso, D.; Boissiere, C.; Belamie, E.; Coradin, T.; Sanchez, C. Chem. Mater. 2007, 19, 4349–4356. (39) Baklanov, M. R.; Mogilnikov, K. P.; Polovinkin, V. G.; Dultsev, F. N. J. Vac. Sci. Technol. B 2000, 18(3), 1385–1391.

11938 Langmuir, Vol. 24, No. 20, 2008

Li et al.

Figure 3. (a) Toluene adsorption-desorption isotherms and (b) respective pore size distribution of the silica film calculated using the Kelvin equation for the desorption loop prepared by vapor phase condensation as a function of exposure time to PBS buffer solution at 37 °C. Shown are data (•) before exposure, (O) after 15 min, (]) after 60 min and (+) after 150 min in PBS. The gray bar in Figure 2b is a PSD determined from TEM in which the average pore radius is 3.3 nm, consistent with the result 3.5 nm calculated using Kelvin equation. The refractive index is given for λ ) 632 nm.

Figure 5. Fraction of silica remaining in films synthesized at different CO2 pressures, 77 (•), 84 (O), 125 (9), and 146 bar (2), upon exposure to PBS buffer solution. This fraction is calculated from the changes in porosity and the film thickness. To improve the readability of the figure, sparse data markers are utilized except for the data at 84 bar. All films were measured at the same number of conditions.

Figure 4. (a) Comparison of film relative thickness change and (b) porosity for the mesoporous silica films as a function of exposure time to PBS at 37 °C for films synthesized at CO2 pressures of 77 (•), 84 (O), 125 (9), and 146 bar (2). To improve the readability of the figure, sparse data markers are utilized except for the data at 84 bar. All films were measured at the same number of conditions.

increases. Also, the PSD broadens, also indicative of nonuniform dissolution through the film. After the film exposed in PBS for 60 min, the average pore diameter is 7.1 nm from toluene desorption isotherm which is close to the average pore diameter determined from TEM micrographs, 6.6 nm. Moreover, the broad pore size distribution indicated by the toluene desorption isotherm is consistent with the local PSD obtained from analysis of TEM micrographs as shown in Figure 3b at 60 min exposure to PBS. The poor stability of the vapor phase reacted mesoporous silica film is consistent in many aspects with pure silica EISA films. However, the films synthesized in CO2 may behave differently as CO2 acts to increase the extent of condensation within the film26 and more recently we have demonstrated that synthesis near its critical pressure can lead to difficulty in condensing both hydrophilic and hydrophobic solvents within the pores.31 Figure 4 illustrates that the films synthesized with CO2 degrade much slower than the film prepared with vapor phase reaction. For the film prepared at 77 bar, the thickness is slightly reduced by 10% but the pore volume fraction increases significantly from 25% to almost 75% in the first 18 h. This is similar to the bulk erosion

like behavior of the vapor phase film examined previously except this behavior is extended from 1.5 h to more than 18 h. To compare the difference in dissolution, for the film synthesized with vapor phase condensation 90% of the silica was lost in first 3 h; the worst performing film synthesized using CO2 examined here requires ∼2 days of immersion to reach the same fraction of silica removed as shown in Figure 5. Although the film synthesized with CO2 at 77 bar possesses an initial pore size (1.8 nm from desorption in radius)31 similar to those prepared by vapor phase condensation (1.6 nm in pore radius), a much slower dissolution rate is observed. Moreover, the total porosity of the film synthesized with CO2 is greater than without. Thus, the pore size and surface area do not appear to be the primary properties that determine the film dissolution behavior. We suspect that the use of CO2 alters the silica network due to rapid removal of the ethanol byproduct during condensation,40 which leads to the improved hydrolytic stability. To further assess the impact of CO2 on the film stability in PBS, several other processing conditions were examined, 84, 125 and 146 bar. For these films, their initial pore size (3.81-3.9 nm), porosity (57-64%) and wall thickness (0.9-1.1 nm) are similar as demonstrated previously.31 Despite these similarities, the dissolution behavior is distinct for each film. Increasing the pressure slightly to 84 bar results in a slightly faster initial dissolution rate in PBS compared to the film synthesized at 77 bar. Thus, increasing the solvent power of the CO2 (through (40) Pai, R. A.; Watkins, J. J. AdV. Mater. 2006, 18(2), 241–245.

Tuning Stability of Mesoporous Silica Films

Langmuir, Vol. 24, No. 20, 2008 11939

Figure 6. Toluene adsorption-desorption isotherms for the films prepared at CO2 pressure of 77 (a), 84 (c), 125 (e) and 146 bar (g), as a function of exposure time to PBS buffer solution at 37 °C. Shown are data (•) before exposure, (2) after 75 min and (O) after 300 min in PBS solution. The refractive index is given for λ ) 632 nm. The respective pore size distribution [(b), (d), (f), and (h)] of the films are calculated using the Kelvin equation from the desorption isotherms.

increased density) does not necessarily always improve the stability of the film in PBS. To elucidate the origins of this effect, a closer examination of the initial morphology is necessary. From XRD (see supplemental data), the d-spacings are similar between films synthesized at 77 and 84 bar: 8.57 and 8.52 nm, respectively. However, the pore radius is significantly altered with an increase from 1.80 to 3.87 nm by changing the synthesis conditions from 77 to 84 bar. By geometric arguments, the surface area for the film at 84 bar is 4.6× of the surface area for the film prepared at 77 bar. Since the initial dissolution appears to be internal at

the pore wall interface, an increase in surface area would be expected to increase the dissolution rate. However, note that the dissolution rate for the film synthesized at 84 bar is only slightly faster than for 77 bar and is significantly more stable than for the vapor phase film. Thus, surface area is an important factor in determining the rate of silica loss, but the improvements in the stability by processing with CO2 are significantly greater than the simple surface area effect. Increasing the pressure to 125 bar results in a more hydrolytically stable film with only 30% silica fraction reduction in

11940 Langmuir, Vol. 24, No. 20, 2008

Figure 7. Probing filling of mesopores by toluene capillary condensation. Only a portion of the pores are filled with liquid toluene upon exposure to saturated vapor for the films synthesized at a CO2 pressure of 84 (O), 125 (9), and 146 bar (2). This property is largely independent of prior immersion in PBS, which tends to increase the average pore size.

the first 12 h upon the exposure to PBS, and an almost zero thickness change. During this time, the porosity continuously increases from 60% to 80%. However after 12 h, the silica amount in the film is nearly unchanged for another 20 h showing that pure silica films can be quite stable in PBS with CO2 processing. Compared to the silica films doped with titania and zirconia investigated by Bass et al.,23 mesoporous mixed metal oxide films are more hydrostable. However, intermediate stability between the order of several hours and one week are not readily obtainable with the mixed metal oxide, but can be obtained by simple variations in CO2 pressure during synthesis of pure mesoporous silica films. At higher CO2 pressures examined (146 bar), similar changes in the porosity of the film to the film at 125 bar are observed, but the thickness decreases by 30% for the film prepared at 146 bar in the first 12 h, resulting in a 60% silica reduction. Therefore further increasing the solvent quality of CO2 did not appear to increase the stability. However, there are two factors to consider. First, the density of CO2 at 60 °C does not significantly change between 125 and 146 bar. Maxima in property variations with CO2 pressure have been observed as solvation and hydrostatic pressure commonly have competing effects.41 Second, there is a degradation in the ordering of the mesoporous silica for synthesis at 146 bar with the TEOS loadings utilized.31 A decrease in long-range order of mesoporous films is known to lead to a significant decrease in the mechanical properties.42 The combination of a minimal increase in solvent strength of CO2 with a decrease in mechanical properties of the film is likely the cause of the decreased stability observed for the film synthesized at 146 bar. However again note that decrease in the hydrostability of the film is small and results in properties similar to the film synthesized at 77 bar. To confirm the mechanism of the silica loss, PSD were calculated from toluene adsorption-desorption isotherms in the films as shown in Figure 6. The average pore size continuously increases as exposed to PBS for longer time. This is the same behavior as observed for the vapor phase reacted film except the time frame for the pore size expansion is extended. One other interesting phenomenon in the toluene adsorption-desorption isotherms is the lack of condensation within some of the pores. This was previously reported for mesoporous silica films synthesized near the critical pressure.31 These are identified by the smaller change in refractive index upon exposure to saturated toluene vapor then predicted if all the pores are filled with toluene, in which case the refractive index should be between 1.458 and 1.497, which is the refractive index for silica and toluene, respectively. The porosity of the film calculated from the BEMA (41) Pham, J. Q.; Johnston, K. P.; Green, P. F. J. Phys. Chem. B 2004, 108(11), 3457–3461. (42) Li, X.; Song, L. Y.; Vogt, B. D. J. Phys. Chem. C 2008, 112(1), 53–60.

Li et al.

model using only the refractive index of the neat mesoporous film and that calculated from the change in refractive index upon exposure to saturated toluene can be compared to approximately determine the fraction of pores that can be filled. From this analysis, almost all the pores for the films synthesized with vapor phase condensation and at a CO2 pressure of 77 bar are filled by toluene (Figure 3a, Figure 6a) irrespective of the time immersed in PBS. However, increasing the CO2 pressure above 84 bar during synthesis results in a significant fraction of the pores remaining unfilled after exposure to saturated toluene. This phenomenon continues for these films even after exposure to PBS at 37 °C. The final refractive indices of the films prepared at 84, 125 and 146 bar are far below 1.458-1.497 in Figure 6c, e, g. Figure 7 illustrates the percentage of pore volume filled by toluene according to the time exposed to PBS. The film synthesized at 84 bar has the lowest pore accessibility to toluene condensation which is about 10-15% in the first 6 h time period exposed to PBS. The fraction of the pores filled appears to not change significantly upon exposure to PBS, thus suggesting that this property might be associated with the film stability. However, the film synthesized at 77 bar is much more stable than the one prepared at atmospheric pressure although the pores in both films are almost equally accessible to condensation of toluene. Therefore, ability of small molecules to condense within the mesopores does not appear to be the primary reason for the enhanced stability of mesoporous silica films synthesized using CO2. Instead we hypothesize that the molecular structure of the wall silica is altered by synthesis in CO2. Silica dissolution mechanism at near neutral conditions is reported to be a result of hydrolytic attack on the Si-O-Si bond at the surface. The energy associated with this reaction will be dependent upon the local bonding. For example, the Si-O bond in mesoporous silica synthesized via sol-gel route is known to be strained;43 this strained structure could alter the hydrolytic stability of the silica through modification of the activation energy required for the hydrolysis of Si-O-Si. Conversely, the structure of silica can be altered depending upon the synthesis conditions to form linear, cyclic, or cubic species. The cluster size of the silica glass can also be modified by synthesis conditions. These local structural changes could alter the accessibility of the Si-O-Si bonds. Additionally, CO2 in the system during condensation of TEOS acts to extract the ethanol byproduct from the film into the fluid phase thereby driving the reaction toward completion. This byproduct extraction could lead to a decrease in the number of dangling silanol groups prior to calcination and a decrease in the mechanical strain in the silica network. As a result, the contraction of the films during calcination decreases monotonically as the CO2 synthesis pressure is increased.31 The stability of the films in PBS is however not monotonic with pressure and thus structural changes such as those suggested previously are likely responsible for the enhanced stability of films synthesized with CO2. Future work using X-ray absorption spectroscopy is necessary to quantify differences in the local structure between mesoporous silica films synthesized with and without CO2. The ability to tune film stability and morphology by varying processing conditions is a promising approach for generating nanoscale materials for biological applications. To explore the implications of the silica stability on cell viability, these films were investigated as scaffolds for supporting the culture of murine osteoblasts (bone cells) in vitro. Figure 8 shows representative images of osteoblasts cultured on these films and stained with calcein-AM (living cells; green) and ethidium homodimer (dead (43) Yamamoto, T.; Mori, S.; Kawaguchi, T.; Tanaka, T.; Nakanishi, K.; Ohta, T.; Kawai, J. J. Phys. Chem. C 2008, 112(2), 328–331.

Tuning Stability of Mesoporous Silica Films

Langmuir, Vol. 24, No. 20, 2008 11941

Figure 8. Representative fluorescence microscopy images of murine osteoblasts cultured on mesoporous silica films synthesized at a CO2 pressure of (a) 77, (b) 84, (c) 125, and (d) 146 bar.

cells; red). Films processed using 77 and 125 bar CO2 supported the adhesion and viability of osteblasts as seen from the high number of cells that fluoresce green. Interestingly, viable multicellular spheorids of osteoblasts were found in these films; formation of bone cell spheroids has been demonstrated to result in the upregulation of bone-related proteins leading to the formation of microcrystalline bone in vitro.44 In contrast, films synthesized using 84 and 146 bar CO2 did not promote cell adhesion and a low fraction of adhered cells was viable in case of these films. These results are in agreement with the film stability properties described previously; films that are stable over the period of one day (Figure 4) demonstrated high osteoblast viabilities; however, films that disintegrated over this time frame did not support osteoblast growth. The most important factor for cell viability was the stability of the surface (minimal thickness change). The fractional loss of silica from the films synthesized at 77 and 146 bar are comparable (Figure 5), but the nature of the silica dissolution appears to be critical for cell adhesion. This dependence might be expected as dissolution of the surface of the film would interfere with cell adhesion. These results are significant in that, mesoporous films that support cell growth can now be generated in a facile manner simply by varying processing conditions. Implications of these studies could be important for a number of biological applications including tissue engineering and drug delivery.

Conclusions The impact of immersion of mesoporous silica films prepared at different CO2 pressure in PBS buffer was examined through monitoring changes in porosity, thickness, remaining silica fraction, and pore size. In all cases, the dissolution of the silica (44) Kale, S.; Biermann, S.; Edwards, C.; Tarnowski, C.; Morris, M.; Long, M. W. Nat. Biotechnol. 2000, 18(9), 954–958.

proceeds initially from the internal pore walls in a manner reminiscent of bulk erosion for polymer drug delivery. At long times, the free surface begins to erode as well. Without using CO2 (vapor phase reaction), nearly 90% of the silica in pure silica films is removed in 3 h. The hydrolytical stability can be significantly improved from hours to days by using CO2 processing. For a film synthesized at 125 bar, only 80% of the silica is removed after 3 days of exposure to PBS. Outside of CO2 pressure during synthesis, long-range order of the mesopores and the surface area also appear to be factors in controlling the dissolution rate of silica in PBS. This tunable stability using CO2 opens new possibilities for the utilization of mesoporous silica for biological applications as demonstrated by osteoblast viability on films that possess high stabilities. Acknowledgment. The authors acknowledge support from the National Science Foundation (ENG-0746664) and partial financial support from the State of Arizona. We thank DuPont Electronic Materials (Jim Sounik and Michael Sheehan) for donation of the poly(4-hydroxystyrene). The authors thank Dr. Christine Pauken, Harrington Department of Bioengineering, for providing the osteoblasts for the cell studies. We acknowledge the use of facilities in the LeRoy Erying Center for Solid State Science. The authors thank Jerry Y.S. Lin for access to XRD instrumentation. We acknowledge the assistance of Barry O’Brien at the ASU Flexible Display Center for assistance with FTIR measurements. Supporting Information Available: Figures showing XRD and FTIR data for the mesoporous silica films for the different synthesis conditions. This material is available free of charge via the Internet at http://pubs.acs.org. LA801849N