Sintered Silica Colloidal Crystals with Fully Hydroxylated Surfaces

Department of Chemistry, The UniVersity of Arizona, Tucson, Arizona 85721. ReceiVed March 30, 2007. In Final Form: May 16, 2007. Silica colloidal crys...
0 downloads 0 Views 658KB Size
8554

Langmuir 2007, 23, 8554-8559

Sintered Silica Colloidal Crystals with Fully Hydroxylated Surfaces Thai Van Le, Eric E. Ross, Tomika R. C. Velarde, Michael A. Legg, and Mary J. Wirth* Department of Chemistry, The UniVersity of Arizona, Tucson, Arizona 85721 ReceiVed March 30, 2007. In Final Form: May 16, 2007 Silica colloidal crystals require multiple processing steps before they are useful materials in analytical applications, such as chemical separations, microarrays, sensors, and total internal reflection microscopy. These chemical processing steps include calcination, sintering, surface rehydroxylation, and chemical modification, but these steps have not been fully characterized for colloidal crystals. Silica particles of nominally 200 nm in diameter were prepared, and FTIR, SEM, UV-visible spectroscopy, and refractive index measurements were used to study the changes in chemical composition, particle size, and particle density throughout the process. The final material is shown to be a durable, crack-free crystal of solid particles bearing a fully hydroxylated surface of silanols, which can then be chemically modified.

Introduction Silica particles of controlled submicron size and low polydispersity1 can be deposited on solid substrates to form highly ordered face-centered cubic crystals2 with thicknesses ranging from two layers to several hundred layers.3 As a photonic material, a colloidal crystal gives strong Bragg diffraction at a wavelength determined by the particle diameter, refractive index, and crystal thickness.3,4 There has been widespread interest in silica colloidal crystals as photonic materials for many years, and this has recently been reviewed.5 New applications of colloidal crystals are now emerging, such as supports for lipid bilayers,6 high-surface-area substrates for microarrays,7 patterned supports for microarrays,8,9 microreactors,10 and media for chemical separations.11-15 Chemical modification of silica colloidal crystals has been reported using silylation12 and silylation followed by atom-transfer radical polymerization.16 These many emerging applications place new demands on the chemical characteristics of colloidal crystals. For example, their use as separation media and as substrates for microarrays and lipid bilayers require that the surfaces have controlled adsorptivity. Silica is well-established in the field of chemical separations, including chromatographic silica gel and silica capillaries, because (1) Stober, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26 (1), 62 ff. (2) Miguez, H.; Meseguer, F.; Lopez, C.; Mifsud, A.; Moya, J. S.; Vazquez, L. Langmuir 1997, 13 (23), 6009-6011. (3) Jiang, P.; Bertone, J. F.; Hwang, K. S.; Colvin, V. L. Chem. Mater. 1999, 11 (8), 2132-2140. (4) Bertone, J. F.; Jiang, P.; Hwang, K. S.; Mittleman, D. M.; Colvin, V. L. Phys. ReV. Lett. 1999, 83 (2), 300-303. (5) Koenderink, A. F.; Johnson, P. M.; Lopez, J. F. G.; Vos, W. L. C. R. Phys. 2002, 3 (1), 67-77. (6) Brozell, A. M.; Muha, M. A.; Sanii, B.; Parikh, A. N. J. Am. Chem. Soc. 2006, 128 (1), 62-63. (7) Zheng, S.; Zhang, H.; Ross, E.; Le, T. V.; Wirth, M. J. Anal. Chem. 2007, 79, 3867-3872. (8) Park, J.; Moon, J.; Shin, H.; Wang, D.; Park, M. J. Colloid Interface Sci. 2006, 298 (2), 713-719. (9) Choi, W. M.; Park, O. O. Colloids Surf., A 2006, 277 (1-3), 131-135. (10) Das, M.; Zhang, H.; Kumacheva, E. Annu. ReV. Mater. Res. 2006, 36, 117-142. (11) Newton, M. R.; Bohaty, A. K.; White, H. S.; Zharov, I. J. Am. Chem. Soc. 2005, 127 (20), 7268-7269. (12) Zheng, S. P.; Ross, E.; Legg, M. A.; Wirth, M. J. J. Am. Chem. Soc. 2006, 128 (28), 9016-9017. (13) Newton, M. R.; Morey, K. A.; Zhang, Y. H.; Snow, R. J.; Diwekar, M.; Shi, J.; White, H. S. Nano Lett. 2004, 4 (5), 875-880. (14) Zeng, Y.; Harrison, D. J. Anal. Chem. 2007, 79, 2289-2295. (15) Cichelli, J.; Zharov, I. J. Am. Chem. Soc. 2006, 128 (25), 8130-8131. (16) Schepelina, O.; Zharov, I. Langmuir 2006, 22 (25), 10523-10527.

of its controlled adsorptivity. However, simply having a substrate composed of silica does not guarantee high-performance separations. Subtle differences in surface treatment can result in severe zone broadening in separations, and nonspecific adsorption to the substrate is always an issue with microarrays.17 It was a breakthrough made in the 1980s that led to the high-quality surface on silica gel that is used today.18 Considerable processing is also required to improve the surface quality of silica capillaries before use in electrophoresis.19 To date, there have been no studies specifically addressing the surface properties of silica particles per se. For preparing colloidal crystals, it is known that these silica particles must be calcined to shrink their volumes prior to deposition to avoid large cracks in the final crystal.20 Furthermore, calcination at different temperatures results in particles with different refractive indices, and the particles with a lower refractive index are known to slowly absorb solvent.21 One can reasonably assume that the particles have to be solid to prevent unwanted adsorption. Heating to 1050 °C, which causes sintering, results in particles with refractive indices comparable to that of fused silica,21 indicating that they are solid silica, but the resulting crystal has an unacceptable amount of cracking without prior calcinations of the particles. It is not known whether one can first calcine the particles to prevent cracks and then sinter the resulting crystal to achieve solid particles. If so, then these materials would likely exhibit high performance in chemical separations due to reduction in mechanical defects and in optical applications due to higher refractive indices of the particles. The purpose of this work is to investigate whether the silica particles approach the density of fused silica when crack-free sintered colloidal crystals are made. We combine FTIR, SEM, optical microscopy, UV-visible spectroscopy, and refractive index measurements using index matching fluids to characterize the materials. We investigate the conditions needed to completely rehydroxylate the surfaces of the particles. Experimental Section The Sto¨ber method was used to synthesize the particles.1 A 500 mL round-bottom flask and a 250 mL beaker were cleaned by (17) Kohler, J.; Chase, B. D.; Farlee, R. D.; Vega, A. J.; Kirkland, J. J. J. Chromatogr. 1986, 352, 275-305. (18) Kohler, J.; Kirkland, J. J. J. Chromatogr. 1987, 385, 125-150. (19) Kohr, J.; Engelhardt, H. J. Chromatogr., A 1993, 652 (2), 309-316. (20) Chabanov, A. A.; Jun, Y.; Norris, D. J. Appl. Phys. Lett. 2004, 84 (18), 3573-3575. (21) Garcia-Santamaria, F.; Miguez, H.; Ibisate, M.; Meseguer, F.; Lopez, C. Langmuir 2002, 18 (5), 1942-1944.

10.1021/la700929e CCC: $37.00 © 2007 American Chemical Society Published on Web 06/27/2007

Fully Hydroxylated Sintered Si Colloidal Crystals immersion in a saturated KOH/isopropanol solution for 24 h. The glassware was thoroughly rinsed with water (deionized, 18 MΩ cm), followed by a single ethanol rinse, then dried at 120 °C in an oven for 2 h. Solution A was made with 50 mL 2 M ammonium hydroxide, 20 mL deionized H2O, and 23 mL ethanol filtered through a Nalgene syringe filter (polytetrafluoroethylene, 25 mm diameter, 0.2 µm) into a 500 mL round-bottom flask. A magnetic stir bar was used to spin this solution at a slow, constant rate. Solution B composed of 100 mL ethanol and 7.2 mL tetraethoxysilane was filtered through a Nalgene syringe filter into the 250 mL beaker. The solution was sonicated (VWR, model 75T) for 2 min to thoroughly mix the ethanol and TEOS, and was then poured into the round-bottom flask containing solution A. The reaction was allowed to run at room temperature for 3 h with constant stirring. After the reaction, the silica nanoparticles were separated from the supernatant by centrifugation at 10 000 rpm for 15 min. The supernatant was removed, and the colloidal pellet was rinsed and resuspended in pure ethanol by sonication. This procedure was performed three times. Silica particles were calcined in a Pyrex beaker covered with a Coors ceramic crucible prior to crystallization. This was done to promote the formation of siloxane bonds and to drive off the ethanol and water trapped within the particles. Specifically, the particles were heated to 300, 450, and 550 °C, in succession, for 12 h at each temperature. To minimize particle aggregates, following each calcination step, the particles were dispersed in ethanol by sonicating them for 4 h. After the three calcination steps were complete, the particles were resuspended in ethanol by sonication, and the suspension was allowed to rest at ambient temperature for 24 h to allow any remaining aggregates to settle. We prepared the colloidal crystal by evaporative deposition of a colloidal slurry.22 The substrates were fused silica slides that had been cleaned by dipping into boiling methanol followed by placement in a UV-ozone plasma cleaner (Novascan Technologies, PSD-UV) at 20 W/cm2 for 10 min. A colloidal suspension was prepared by sonication of calcined silica particles in ethanol for 2 h (0.1 g colloid in 20 mL ethanol). The cleaned slides were placed vertically into 30 mL glass beakers containing approximately 20 mL of the colloidal suspension. The suspension was then incubated under a 100 W incandescent lamp for 20 h to accelerate the evaporation of ethanol relative to lab ambient conditions. For FTIR analysis, polished silicon wafers were used as substrates, and the same procedures as for fused silica substrates were used. The fused silica slides bearing the silica colloidal crystals were sintered atop a 1/4-in quartz plate in a furnace at 1050 °C for 12 h. The quartz plate was used to prevent the fused silica slide from warping during lengthy sintering. The crystal was then allowed to cool gradually over a period of 3-4 h within the furnace. The oven door remained closed during cooling to avoid sudden, large changes in temperature, which might cause cracks in the material. Sintering causes condensation of surface silanols into siloxane bonds,23,24 and the completely dehydroxylated silica surfaces are chemically unreactive to silylating agents. Treatment with either an aqueous base or acid solution will achieve rehydroxylation on a time scale ranging from a few hours to a few days through catalysis of the cleavage of siloxane bonds and the formation of silanols.18,25 To rehydroxylate the crystals, the sintered material was placed in a solution of tetrabutylammonium hydroxide of pH 9.5 at 60 °C for 24 h, followed by rinsing with deionized H2O, 1 M nitric acid, methanol, and then deionized H2O, in succession. To demonstrate that the surface was chemically reactive, a brush layer of polyacrylamide with nominal thickness of 10 nm was grown by atom-transfer radical polymerization.26-29 Prior to chemical modification, the sintered material was cleaned in hot methanol for 2 h, rinsed with deionized water, and dried under nitrogen in a tube furnace at room temperature. The clean sintered material was placed (22) Dimitrov, A. S.; Nagayama, K. Langmuir 1996, 12 (5), 1303-1311. (23) Iler, R. K. The Chemistry of Silica, 2nd ed.; Wiley: New York, 1979. (24) Tripp, C. P.; Hair, M. L. Langmuir 1993, 9 (12), 3523-3529. (25) Kirkland, J. J.; Kohler, J. Porous silica microspheres having a silanol enriched surface. U. S. Patent 4,874,518, 1989.

Langmuir, Vol. 23, No. 16, 2007 8555 in a 250 mL flask that contained 1.0 mL of (chloromethylphenylethyl)dimethylchlorosilane (Gelest, Morrisville, PA) and 1.0 mL of pyridine in 100 mL of dicholoromethane. The reaction proceeded at reflux temperature overnight. After the reaction, the silica gel was rinsed with dichloromethane, toluene, and methanol, then dried in an oven at 110 °C for 1.0 h. The colloidal crystal was modified with polyacrylamide by free radical polymerization with a CuCl/CuCl2/ Me6TREN catalytic system at room temperature, using a previously reported procedure.29 A Schlenk flask was charged with 49.5 mg (500 µmol) of CuCl and 6.7 mg (50 µmol) of CuCl2. The flask was sealed with a rubber stopper was cycled between vacuum and argon three times to remove oxygen. A 100 mL solution of 3 M acrylamide in N,N-dimethylformamide was bubbled with argon for 2 h and then transferred into the flask via a syringe. After the catalyst was completely dissolved, 0.17 mL (120 µmol) of Me6TREN was injected into the flask. The reaction solution was then transferred to another flask containing the silica colloidal crystal, which was sealed with a rubber stopper, then cycled between vacuum and argon at least three times to remove oxygen. Then, the flask was placed in a water bath at room temperature and allowed to react for 10 h. After reaction, the material was rinsed with N,N-dimethylformamide. Infrared spectra were obtained for colloidal crystals on silicon using a Nicolet 4700 FTIR from Thermo Electron Corporation. All spectra were obtained using 256 scans, with a bare silicon slide used as the blank. The spectra were acquired in transmission mode at 55° incidence with vertical polarization to avoid interference fringes. UV-visible absorbance spectra were obtained at normal incidence using a blank silica slide as the reference, using an Agilent 8453 spectrophotometer. Optical micrographs were obtained using a Nikon Eclipse TE2000-U microscope with a Nikon model C-SHG1 mercury lamp power supply, and a Cascade 512B CCD camera from Photometrics. Field-emission SEM images were obtained using a Hitachi S-4500 using Thermo-Norm Digital Imaging/EDS. The method used to determine the refractive indices for the various silica particles was a modified version of the one used by Nussbaumer and co-workers30 to determine the refractive indices of planer glass and quartz samples. Absolute ethanol obtained from Decon Labs Inc. and 99.5% pure, ACS-grade toluene obtained from EMD were used as solvents. Both were used as received, without further purification. A series of 13 standard solutions was created by mixing the ethanol and toluene in volume fractions of toluene ranging from 0 to 1. The refractive indices of these standards were measured using a Bausch and Lomb Abbe 3L Refractometer cooled to 20 °C with recirculating water, and these measurements were used to create the calibration curve. All refractive indices are reported at the wavelength of the sodium D-line (589 nm). For each silica sample, 25.0 ( 0.5 mg of silica was suspended in 8 mL of solution. To maintain consistency, both solvents were maintained at 20 °C in a water bath while the samples were being prepared. Initially, the ethanol was added to the particles and then the sample was sonicated for ∼1 min to break up any aggregated particles. After the aggregates were dispersed, the toluene was added. For each sample, the % transmission at 589 nm was measured against a blank of solvent with the same volume fraction of toluene over an appropriate range of concentrations to determine a peak value. This peak value corresponds to the point where the refractive index of the silica particles matches that of the solvent and scattering from the particles is at a minimum. The % transmission measurements were taken on an Agilent 8453 UV-vis spectrophotometer.

Results and Discussion Figure 1a shows an optical micrograph for a sintered colloidal crystal, which had been previously calcined. To demonstrate (26) Huang, X.; Wirth, M. J. Macromolecules 1999, 32 (5), 1694-1696. (27) Huang, X. Y.; Doneski, L. J.; Wirth, M. J. Anal. Chem. 1998, 70 (19), 4023-4029. (28) Huang, X. Y.; Wirth, M. J. Anal. Chem. 1997, 69 (22), 4577-4580. (29) Xiao, D. Q.; Wirth, M. J. Macromolecules 2002, 35 (8), 2919-2925. (30) Nussbaumer, R. J.; Halter, M.; Tervoort, T.; Caseri, W. R.; Smith, P. J. Mater. Sci. 2005, 40 (3), 575-582.

8556 Langmuir, Vol. 23, No. 16, 2007

Van Le et al.

Figure 2. FTIR spectra of silica colloidal crystals made using particles with no calcination (as-made), using calcined particles, and using calcined particles that were sintered after the crystal was formed.

Figure 1. Optical micrographs of (a) sintered silica colloidal crystal with three prior calcinations steps and (b) colloidal crystal with neither calcinations nor sintering, shown to illustrate cracks.

that cracks would be evident on this scale, Figure 1b shows the optical micrograph for a colloidal crystal made from the same particles but without any prior calcination. The latter crystal was not sintered; therefore, the cracks evident in the image were the result of shrinkage at room temperature. The triply calcined materials of Figure 1a, even after sintering, show no cracks that could be resolved by the optical microscopy. The underlying processes that occur upon calcination and sintering were monitored by FTIR measurements. The infrared spectra corresponding to the as-made, calcined, and sintered particles are shown in Figure 2. The peaks for the reagents are reduced with progressive heat treatment. The large, broad peak in the O-H stretching region, with a maximum near 3400 cm-1, corresponds to water on the surface or trapped within the colloid. On the basis of the relative intensities at this frequency, the calcination step eliminates approximately three times as much water as the sintering step. By the end of the sintering step, there is virtually no detectable water. The C-H stretches from 2800 to 2900 cm-1 correspond to the ethoxy groups of TEOS. This peak is small, and it disappears after calcination. The broad peak for the hydrogen-bonded silanols (3600 cm-1) decreases with calcination and then decreases even further upon sintering. The peak for the isolated silanols (3745 cm-1) becomes sharper and more prominent as the sample is calcined and then sintered. The siloxane peaks resulting from condensation of the silanols are from 1868 to 1990 cm-1 in the infrared spectrum,31 and these

are shown in Figure 2 to increase upon calcination and then further upon sintering. Overall, the sintered material has lost almost all water and associated silanols, and its spectrum is dominated by the siloxane peaks, indicating that it has approached the chemical composition of fused silica. The progressive loss of water shown in the infrared spectrum upon heat treatment should correlate with a corresponding decrease in the sizes of the particles, and this was confirmed by SEM. Figure 3 illustrates, from a small section of two SEM images, how scanning electron microscopy was used to measure colloid size. A comparison of as-made and calcined particles shows that the latter are distinctly smaller, as indicated by putting an arrow of the same size on each of the images. For the [111] plane of the fcc lattice, there are rows of particles along one crystal axis that lie along a line. The crystal domains are sufficiently large and sufficiently crystalline that many such lines of five particles in length can be found and measured to determine the colloid diameter with a precision of a few nanometers. The sizes of the particles, as determined by SEM, at each step in the heating process are listed in Table 1. There is a greater decrease in diameter going from as-made to calcined than there is from calcined to sintered. The calculated change in colloid volume is approximately three times higher upon calcination than it is upon sintering, which is in agreement with the relative changes in the intensities of the water peaks in the infrared spectra. A decreased colloid size is expected to be associated with an increase in colloid density, and therefore an increase in refractive index. The data bear out part of this expectation, as shown in Figure 4. The as-made particles exhibit a distribution of refractive indices, indicating that the material is heterogeneous on the optical scale. The calcined particles have a lower and better defined refractive index, 1.440, compared to the as-made colloid. The decrease in refractive index suggests that, while the calcination removes water, at least part of the volume is filled by air rather than by silica. Upon sintering, the refractive index increases to a value of 1.454, which approaches the refractive index of fused silica, 1.458. Garcia-Santamaria et al. showed a similar result,21 but the particles were not treated sequentially in their case; hence, it was not previously known whether the lower-index particles obtained upon calcinations could be converted to higher-index particles by the higher-temperature sintering. Our result shows that the conversion to the higher-index colloid is achieved. Using (31) Gorski, D.; Klemm, E.; Fink, P.; Horhold, H. H. J. Colloid Interface Sci. 1988, 126 (2), 445-449.

Fully Hydroxylated Sintered Si Colloidal Crystals

Langmuir, Vol. 23, No. 16, 2007 8557

Figure 4. Optical transmission measured at 589 nm for colloid slurries as a function of the refractive index of the solvent used for the slurry. Four different preparations of particles were studied, as denoted by the labels in the inset.

Figure 3. SEM images of colloidal crystal made from a) as-made particles, and b) calcined particles. The arrow size is the same in each panel to indicate that the particles in panel b are smaller. Table 1. Summary of Data for Particles after the Indicated Treatment

material

refractive indexa (λ ) 589 nm)

colloid diameter

predicted Bragg peak

as-made

1.44-1.46

205-215 nm 448-473 nm

calcined sintered rehydroxylated

1.439 1.457 1.457

193 nm 188 nm

a

422 nm 415 nm

observed Bragg peak 454 nm (broad) 425 nm 414 nm 409 nm

Note that fused silica has an index of refraction of 1.4585 at 589

nm.

an extension of the Lorentz-Lorenz relation for refractive indices of mixtures,32 the refractive index of the calcined particles indicates that they are 5% air by volume. The sintered particles are only 0.7% air by volume, meaning that they are nearly equivalent to fused silica. Sintering differs from calcination in an important way: sintering is performed above 1000 °C, which is sufficient to cause flow of silica on the surface. The existence of fusion points among the particles after sintering is established in the SEM (32) Li, W. B.; Segre, P. N.; Gammon, R. W.; Sengers, J. V.; Lamvik, M. J. Chem. Phys. 1994, 101 (6), 5058-5069.

Figure 5. Field-emission SEM micrographs of a sintered silica colloidal crystals showing the attachment points after breaking the crystal.

image of Figure 5, which shows the divots and blebs remaining after a sintered colloidal crystal was snapped in half. The entire material becomes quite rugged after sintering, capable of surviving ultrasonic cleaning for hours. The many attachment points among particles explain why the material is now durable. The flow of silica on the surface during sintering might by responsible for filling in the free volume in the particles. An independent assessment of refractive index can be obtained by measuring the photonic band gap, which is a sharp band in the absorption spectra due to Bragg diffraction. For fcc crystals, the lattice spacing, d111, is 0.816‚d, where d is the particle diameter.4 The lattice spacing is related to the peak wavelength for Bragg diffraction, λpeak. In eq 1, m is the order of diffraction, 2

m‚λpeak ) 2‚d111(n2eff- sin θ)1/2

(1)

which is 1 in this case; θ is the angle between the incident light and the normal to the diffraction planes, which is 0° in this case; and neff is the mean refractive index of the colloidal crystal. Figure 6 shows a comparison of the visible transmission spectra

8558 Langmuir, Vol. 23, No. 16, 2007

Van Le et al.

Figure 6. Transmission spectra of silica colloidal crystals, including as-made, calcined, and sintered, as indicated. The positions of the Bragg peaks are denoted.

for a colloidal crystal using as-made, calcinated, and sintered particles. The Bragg peak is evident in each spectrum, indicating crystalline order. The Bragg peak is broad for the as-made particles, which is consistent with wider distribution of colloid sizes. The Bragg peak shifts 29 nm to the blue upon calcination, and it shifts another 11 nm further to the blue upon sintering. According to eq 1, this shift is qualitatively consistent with the progressive shrinkage of the particles with heat treatment. One can assess the quantitative agreement between the transmission spectra, the colloid sizes, and the refractive index data calculating the position of the Bragg peak. The refractive indices of the particles, nsilica, and air, nair, combine to give neff, where the volume fraction of colloid in an fcc lattice is 0.74.33 The positions

n2eff ) 0.74‚n2silica + 0.26‚n2air

(2)

of the observed and calculated Bragg peaks are listed in Table 1, and they agree well. The observed positions of the Bragg peaks provide independent confirmation of the refractive indices of the particles. Figure 6 showed that sintering slightly reduces the height of the Bragg peak, indicating a small decrease in crystallinity. To assess how crystalline the material is, SEM images for the sintered material are shown in Figure 7over a wide field of view. The material was snapped in half to allow a cross-sectional view. From Figure 7a, it is evident that the material is quite crystalline, and the crystal planes extend through the depth of the crystal. Figure 7b shows the material over a wider range, where it is now evident that there are lines in the image revealing slight gaps. These gaps are likely caused by the small shrinkage of the particles, and their presence is consistent with the slightly attenuated Bragg peak. Optimal chemical modification of the silica surface via reactive silanes requires complete rehydroxylation of surface siloxanes to surface silanols.18 This process requires a relatively aggressive chemical treatment involving extended treatment in a mild base at elevated temperatures. To demonstrate that the sintered silica crystal particles can be suitably modified in such a manner, rehydroxylation of the sintered colloidal crystal was performed to convert surface siloxanes into surface silanols. Figure 8 includes infrared spectra before and after rehydroxylation of a sintered colloidal crystal. The infrared spectrum for the sintered colloidal crystal is shown for reference. The spectra confirm that rehydroxylation provides a large gain in the peak area for (33) Mittleman, D. M.; Bertone, J. F.; Jiang, P.; Hwang, K. S.; Colvin, V. L. J. Chem. Phys. 1999, 111 (1), 345-354.

Figure 7. SEM images at two different magnifications for a sintered colloidal crystal snapped in half: (a) higher magnification to show the crystal planes; (b) lower magnification to show the outlines of domain boundaries.

Figure 8. Infrared spectra for a colloidal crystal after sintering, after rehydroxylation, and after growth of a polyacrylamide brush layer, as labeled.

hydrogen-bonded silanols, centered at 3600 cm-1. The amount of water adsorbed to this newly hydrophilic surface also increases, as shown by the broad band centered at 3300 cm-1. The peak for the isolated silanols has become virtually undetectable, indicating that the surface is now fully rehydroxylated. The infrared spectra thus establish that the surface silanols are regenerated, and these surface silanols are presumably available

Fully Hydroxylated Sintered Si Colloidal Crystals

Figure 9. UV-visible transmission spectra for a colloidal crystal after sintering (solid line), after base treatment to rehydroxylate the surface (dashed line), and after growth of a polyacrylamide brush layer (dotted line). The small spikes, e.g., at 490 nm, are artifacts.

for reactions to chemically modify the colloidal crystal surfaces for end applications. To demonstrate that these particles can be chemically modified without compromising the integrity of the crystal, a chlorosilane bearing a benzyl chloride group was reacted with the surface to serve as an initiator for atom-transfer radical polymerization of acrylamide.26-28 The resulting infrared spectrum of the polyacrylamide-modified colloidal crystal is shown in Figure 8. New peaks related to the polymer include the C-H stretch region near 2900 cm-1, which arise from the polymer backbone, are evident, as well as a peak from the N-H stretch at 3200 cm-1. The spectrum indicates that there is water solvating the hydrophilic polymer chains. To our knowledge, this is the first published account to detail the chemical modification of sintered silica colloidal crystals. The Bragg peak of a colloidal crystal after rehydroxylation and after chemical modification were monitored, and the spectra are shown in Figure 9. The UV-visible spectrum for a sintered

Langmuir, Vol. 23, No. 16, 2007 8559

colloidal crystal before chemical modification is also shown. Upon rehydroxylation, the Bragg peak shifts to the blue. This shift can either be due to a decrease in the refractive index of the particles caused by base etching the insides of the particles or due to a slight decrease on colloid size, but not density, caused by removal of silica from the particle exterior. To distinguish the cause of the blue shift, the refractive indices of the particles were measured, and the results are included in Figure 4. These show no detectable reduction in the refractive index after rehydroxylation, indicating that the base only reacts with the surface of the silica particles, not the bulk. Using eq 2, the 5 nm shift corresponds to a drop in the volume fraction of silica from 0.74 for an fcc lattice to 0.72 upon removal of silica from the spheres. This corresponds to etching 5 nm of silica from the surface. The spectrum for polymer-modified colloidal crystals indicates that this loss is offset by addition of a 5 nm polyacrylamide layer,27 which has a refractive index similar to that of silica.34

Conclusions The silica particles in sintered silica colloidal crystals are shown to have a refractive index approaching the composition of fused silica, despite being made of calcined particles having a lower refractive index. Rehydroxylation of the surface is shown to restore the surface silanols for subsequent silylation, without etching the interiors of the particles. Overall, the chemical modification of silica particles required many steps in processing, and the Bragg peak in the final material demonstrates that crystalline order remains intact after all of these steps. This ability to chemically modify the silica extends the range of applications for silica colloidal crystals, including their use as separation media and high-surface-area capture materials for microarrays. Acknowledgment. This work was supported by NIH under grant 1 R01 GM65980. We are grateful for fellowship support from Merck, Inc., for M.A.L. LA700929E (34) Polymer handbook. CRC Press: 1995.