Biocompatibility of a Monolayer Protected Nanostructured Material. Daniel S. Albrecht, Jacob T. Lee, Nick Molby, Steven D. Rhodes, Hieu Minh Dam, Jason L.
Mater. Res. Soc. Symp. Proc. Vol. 1063 © 2008 Materials Research Society
1063-OO06-01
Functionalized Porous Silicon in a Simulated Gastrointestinal Tract: Modeling the Biocompatibility of a Monolayer Protected Nanostructured Material Daniel S. Albrecht, Jacob T. Lee, Nick Molby, Steven D. Rhodes, Hieu Minh Dam, Jason L. Siegel, and Lon A. Porter Chemistry, Wabash College, 301 W. Wabash Ave., Crawfordsville, IN, 47933 ABSTRACT Owing to its photoluminescent properties and high surface area, porous silicon (por-Si) has shown great potential toward a myriad of applications including optoelectronics, chemical sensors, biocomposite materials, and medical implants. However, the native hydride-termination is only metastable with respect to surface oxidation under ambient conditions. Por-Si samples oxidize and degrade even more quickly when exposed to saline aqueous environments. Borrowing from solution phase synthetic methods, a selection of hydrosilylation reactions has been recently reported for functionalizing organic groups onto oxide-free, hydride-terminated porous silicon surfaces. Monolayers, bound through direct silicon-carbon bonds, are produced via thermal, microwave, Lewis acid, and carbocation mediated pathways. All of these wet, benchtop methods result in the formation of stable monolayers which protect the underlying silicon surface from ambient oxidation and chemical attack. However, no direct comparison of monolayer stability resulting from these diverse mechanisms has been reported. A variety of alkyl monolayers were prepared on porous silicon using the diverse hydrosilylation routes describe above and then immersed into a sequence of simulated gastric and intestinal fluids to replicate the conditions of potential por-Si biosensors or medicinal delivery systems in the human gastrointestinal tract. Degradation of the organic monolayers and oxidation of the underlying por-Si surfaces were monitored using both qualitative and semiquantitative transmission mode Fourier transform infrared spectroscopy (FTIR). Our initial results indicate that methods employing chemical catalysts often incorporate these species within the monolayer as defects, producing less robust surfaces compared to catalyst-free reactions. Regardless, monolayer protected por-Si samples demonstrated superior durability as opposed to the unfunctionalized controls. INTRODUCTION Nanostructured materials such as porous silicon (por-Si) continue to entice the materials research community with the promise of numerous practical applications as well as advancing fundamental understanding of structure-property relationships [1,2]. Although interfacial regions have proven to play a dominant role in the materials properties of these systems [3,4], only recently has their chemical reactivity and stability been explored in detail. In the case of por-Si, its native hydride-terminated surface structure is only metastable with respect to oxidative degradation under ambient and aqueous environments. However, a number of bench top hydrosilylation reactions are reported to protect the integrity of the por-Si surface against oxidation through the formation of stable organic monolayers [5]. In this study, we utilized microwave, thermal, Lewis acid, and carbocation mediated routes to produce alkyl monolayers on por-Si, bound through direct, covalent silicon-carbon linkages (Figure 1). The alkylfunctionalized por-Si samples were then exposed to simulated gastric and intestinal fluids to
monitor the stability of functionalized por-Si samples under demanding aqueous environments of varying pH and ionic strength. Qualitative and semiquantitative transmission mode Fourier transform infrared spectroscopy (FTIR) was used to monitor sample degradation. The following work presents our initial results indicating that while monolayer protected por-Si samples demonstrate superior resistance to oxidation as opposed to the unfunctionalized controls, hightemperature, catalyst-free hydrosilylation pathways produce more durable samples when compared to reactions employing catalysts at room temperature. Microwave (2.45 GHz, 450 Watts), 180o C, 30 min
H Si
H Si
H Si
Hydride-terminated por-Si
R EtAlCl2 in THF, 25oC, 12 h
R
R
200oC, 12 h
Si
H Si
Si
Alkyl-functionalized por-Si
(C6H5)3CBF4 in CH2Cl2, 25o C, 12 h
Figure 1. (top to bottom) Microwave, thermal, Lewis acid, and carbocation hydrosilylation routes to alkyl-functionalized porous silicon EXPERIMENTAL DETAILS Hydride-terminated por-Si samples are prepared via anodic electrochemical etching under ambient conditions. Polished monocrystalline silicon wafers (B-doped, 1.0–5.0 Ω-cm, (100) orientation, 1.1 cm2 exposed area) are secured in a Teflon cell and etched with 1:1 (v:v) 48% HF (aq)/ethanol in the absence of light under galvanostatic conditions of 7.1 mA/cm2 for 30 min [6]. After etching, samples are rinsed with ethanol, pentane, and dried under nitrogen prior to characterization via FTIR. Infrared spectra (64 scans, 4 cm−1 resolution) are acquired utilizing a Nicolet Avatar 370 spectrophotometer. The etching cell also functions as an IR cell, allowing semiquantitative analysis of the hydrosilylation reaction and subsequent degradation study. Freshly prepared, hydride-terminated por-Si samples are functionalized via microwave [7], thermal [8], Lewis acid [6], or carbocation [9] mediated hydrosilylation of linear primary olefins under inert atmosphere. Microwave hydrosilylation is accomplished using a multi-mode Milestone START Labstation reactor operating at 2.450 GHz. Two por-Si samples and ~10 mL of 1:1 (v:v) neat olefin/mesitylene are loaded into a specially constructed Pyrex pressure vessel. A maximum power of 450 W is used to ramp the temperature of the reaction mixture to 180°C within ten minutes, where this temperature is held for an additional thirty minutes. The temperature is monitored continuously through the use of an integrated infrared sensor. After cooling the reaction to room temperature, the por-Si samples are washed with tetrahydrofuran, dichloromethane, ethanol, and pentane. Thermal hydrosilylation is similarly carried out in a Pyrex pressure vessel containing hydride-terminated por-Si samples immersed into ~10 mL of 1:1 (v:v) neat olefin/mesitylene. Using an oil bath, the reaction vessel is heated to 200°C for approximately twelve hours. Once cooled to room temperature, the samples are washed in the same sequence of solvents listed for the microwave reaction. Both Lewis acid and carbocation mediated hydrosilylation reactions share a generalized procedure. Approximately 1.0 mL of catalyst solution, followed by ~1.0 mL of neat olefin is added to the hydride-terminated por-Si sample, still contained within the Teflon cell. A 1.0 M solution of ethylaluminum dichloride (C2H5AlCl2) in tetrahydrofuran is used as the catalyst solution in the case of the Lewis acid
reaction, while a 0.01 M solution of triphenylcarbenium tetrafluoroborate ((C6H5)3CBF4) in dichloromethane serves as the catalyst solution for the carbocation reaction. Both reactions proceed for twelve hours at room temperature and are then washed thoroughly with tetrahydrofuran, dichloromethane, ethanol, and pentane prior to FTIR analysis. Stability studies are carried out by immersion of both control and functionalized por-Si samples in enzyme-free simulated gastric fluid (SGF) for four hours at 37°C (SGF contains 0.2% sodium chloride and hydrochloric acid; pH of the final solution is ~1.2). After a brief DI water, ethanol, and pentane wash, samples are characterized by FTIR and immediately immersed into enzyme-free simulated intestinal fluid (SIF) for eighteen hours 37°C (SIF contains 0.68% monobasic potassium phosphate and sodium hydroxide; pH of the final solution is 7.5). Once removed from solution and washed with DI water, ethanol, and pentane, final infrared spectra are obtained. RESULTS & DISCUSSION FTIR evidence for the successful preparation of hydride-terminated por-Si is shown in Figure 2(a-1). The spectrum obtained for freshly etched por-Si displays a tripartite signal centered at 2110 cm−1, indicative of the SiHx (x = 1-3) stretching modes. In addition, peaks at 913 cm−1 and 627 cm−1 are evidence for surface SiHx bending modes [2,5-9]. Surface oxidation is kept to a minimum by quickly transferring the freshly etched por-Si samples to an inert atmosphere glove box and passing all olefins through alumina to remove peroxides. Upon reaction with 1octadecene via thermal, microwave, Lewis acid, or carbocation hydrosilylation pathways, the appearance of strong aliphatic CHx stretching bands in the region of 2850-2960 cm−1 is observed (Figure 2, spectra b-2 through e-2). These peaks remain after washing with copious amounts of solvent and withstand sonication. Olefinic groups, indicative of unreacted or physisorbed 1octadecene, are absent from all four FTIR spectra. Similarly, consumption of silicon-hydride is supported by diminished and broadened SiHx stretching and bending modes [5-9]. While all four hydrosilylation reactions effectively yield alkyl-functionalized por-Si, the infrared data provide evidence that these diverse pathways do not produce identical surfaces. While preliminary semiquantitative analysis by integration of the ν(SiHx) region (2000-2200 cm−1) before and after hydrosilylation reveals no statistical difference in the percent consumption of SiHx surface moieties, analysis of the ν(CHx) region indicates a large variation in the amount of surface-bound alkyl groups for the four reactions shown in Figure 2. Integration of the aliphatic CHx stretching region (2800-3000 cm−1) reveals the thermal and microwave reactions both display a statistically larger incorporation of alkyl groups, when compared to the Lewis and carbocation methods. The simplest explanation for this result is that the higher temperatures maintained by the thermal and microwave reactions result in greater yield than the room temperature, catalytic routes. Another key difference can be gleaned by analysis of the CHx stretching region. The infrared spectrum for the carbocation mediated hydrosilylation reaction exhibits weak, yet persistent aromatic ν(CH) bands at 3087, 3066, and 3029 cm−1 that are not diminished by solvent washes or sonication. These peaks are interpreted as evidence for the chemisorption of triphenylcarbenium cation and/or its degradation products within the por-Si subarchitecture. If these sterically bulky groups are incorporated into the organic monolayer, they are expected to interfere with the packing of the long-chain alkyl groups. Higher resolution analysis of the ν(CHx) region provides insight into the alkyl chain packing within the monolayer. Consider,
Figure 2. Transmission mode FTIR spectra for hydride-terminated por-Si (a-1) functionalized with 1-octadecene (C18H36) via thermal (b-1), microwave (c-1), Lewis acid (d-1), and carbocation (e-1) mediated hydrosilylation; the second column (a-2 through e-2) displays the infrared spectra obtained following sample immersion in simulated gastric fluid (SGF) for four hours, while the final column spectra (a-3 through e-3) were taken after sample immersion in simulated intestinal fluid (SIF) for eighteen hours for example, the CH2 stretches of a model hydrocarbon such as polyethylene in which the antisymmetric (νas) and symmetric (νs) methylene stretches appear as relatively broad bands at νas(CH2) = 2828 cm−1 and νs(CH2) = 2856 cm−1 when the polymer is dissolved in solution (i.e. poorly packed and random oriented alkyl chains). In crystalline form, however, these bands
sharpen and shift to νas(CH2) = 2820 cm−1 and νs(CH2) = 2850 cm−1 [10]. Similar information can be gleaned from the FTIR spectra of alkyl monlayers [11]. Although the surface of por-Si is not flat, the long alkyl chains (C18) bound to its surface through the thermal, microwave, and Lewis acid hydrosilylation reactions appear well-packed as evidenced by the νas(CH2) = 2822 cm−1 and νs(CH2) = 2852 cm−1 values observed for all three reactions. In the case of the alkyl-functionalized por-Si prepared by carbocation mediated hydrosilylation, these values shift to νas(CH2) = 2823 cm−1 and νs(CH2) = 2854 cm−1 indicating slightly more disorder, due to the incorporation of the trityl species. This interference with chain packing is more pronounced at shorter alkyl chain lengths. Hydrosilylation of 1-hexene (C6H12) via thermal, microwave, and Lewis acid reactions yield νas(CH2) = 2823 cm−1 and νs(CH2) = 2854 cm−1, while the carbocation mediated pathway produces values of νas(CH2) = 2826 cm−1 and νs(CH2) = 2857 cm−1.
Figure 3. Integration of the ν(OSiO) signal for an unfunctionalized, hydride-terminated por-Si control and functionalized samples following hydrosilylation of 1-octadecene (C18H36) onto hydride-terminated porous silicon via thermal, microwave, Lewis acid, and carbocation reactions; the second table displays the integration data obtained following sample immersion in simulated gastric fluid (SGF) for four hours, while the final table was compiled after immersion in simulated intestinal fluid (SIF) for eighteen hours Degradation studies reveal that the unfunctionalized, hydride-terminated por-Si control exhibit minimal oxidation when immersed into the simulated gastric fluid (SGF) for four hours, but significant oxidation when subsequently exposed to the higher pH simulated intestinal fluid (SIF) for eighteen hours (Figure 2, a-2 and a-3). Integration of the ν(OSiO) signal confirms the qualitative analysis (Figure 3). Similarly, the ν(SiHx) and δ(SiHx) peaks are not observed to diminish until the sample is immersed into the SIF. The 1-octadecene functionalized por-Si samples prepared by the thermal and microwave hydrosilylation pathways produce well-packed monolayers that exhibit excellent resistance to oxidation in both SGF and SIF. No statistical difference is observed for any of the major spectral features following immersion into either solution (Figure 2, b,c-2 and b,c-3). The Lewis acid prepared, alkyl-functionalized por-Si also prove durable under the conditions tested. While the ν(CHx) region remains undiminished (Figure 2, d-2 and d-3), a slight increase in the ν(OSiO) integrated area (Figure 3) is observed, indicating a small amount of surface oxidation upon exposure to SIF. This is attributed to the smaller percentage of alkyl incorporation in comparison to the thermal and microwave reactions. Finally, the carbocation mediated functionalized por-Si samples also retain intensity in the ν(CHx) region when exposed to both SGF and SIF (Figure 2, e-2 and e-3), yet experience the greatest oxidation when immersed into SIF of the hydrosilylation methods (Figure 3). Fewer
incorporated alkyl groups and an increased degree of alkyl chain packing disorganization, attributed to incorporated trityl species, are the most plausible explanations. CONCLUSIONS The preliminary results presented here confirm that alkyl-functionalized por-Si samples can be effectively produced via microwave, thermal, Lewis acid, and carbocation mediated hydrosilylation reactions and demonstrate superior resistance to oxidation as opposed to the unfunctionalized controls when immersed into SIF. High temperature, catalyst-free hydrosilylation pathways appear to incorporate more alkyl groups and produce more densely packed monolayers when compared to room temperature reactions employing catalysts. Furthermore, evidence is provided that some manner of trityl species is incorporated into the monolayer during carbocation mediated hydrosilylation. These species produce defect sites that increase alkyl packing disorder and facilitate more rapid oxidation in SIF. Future studies will focus on the affect of olefin chain length on monolayer packing and sample durability as well as exploration of functionalized por-Si stability in additional aqueous and organic environments. ACKNOWLEDGMENTS We thank Wabash College for generous support through the Department of Chemistry, the Englehardt Startup Fund, and Byron K. Trippet Summer Research Funds. Additional support was provided by the Camille and Henry Dreyfus Foundation through a startup award (2003) and a grant provided by the Petroleum Research Fund (ACS PRF# 44993-GB5). REFERENCES 1. A. Inoue and K. Hashimoto, Amorphous and Nanocrystalline Materials: Preparation, Properties, and Applications, (Springer, 2001). 2. L. T. Canham, Properties of Porous Silicon, (INSPEC, 1997). 3. M. E. Davis, Nature 417, 813-821 (2002). 4. M. Rosoff, Nano-Surface Chemistry, (Marcel Dekker, 2002). 5. L. A. Porter, Jr. and J. M. Buriak, “Harnessing synthetic versatility toward intelligent interfacial design: organic functionalization of nanostructured silicon surfaces,” Chemistry of Nanostructured Materials (World Scientific, 2003) pp. 227-259; J. M. Buriak, Chem. Rev. 102, 1271-1308 (2002). 6. J. M. Buriak, M. P. Stewart, T. W. Geders, M. J. Allen, H. C. Choi, J. Smith, D. Raftery, and L. T. Canham, J. Am. Chem. Soc. 121, 11491-11502 (1999). 7. R. Boukherroub, A. Petit, A. Loupy, J.-N. Chazalviel, and F. Ozanam, J. Phys. Chem. B 107, 13459-13462 (2003). 8. J. E. Bateman, R. D. Eagling, D. R. Worrall, B. R. Horrocks, and A. Houlton, Angew. Chem. Int. Ed. 37, 2683-2685 (1998); R. Boukherroub, S. Morin, D. D. M. Wayner, and D. J. Lockwood, Phys. Stat. Sol. A 182, 117-121 (2000). 9. J. M. Schmeltzer, L. A. Porter, Jr., M. P. Stewart, and J. M. Buriak, Langmuir 18, 2971-2974 (2002). 10. R. G. Snyder, H. L. Strauss, and C. A. Elliger, J. Phys. Chem. 86, 5145-5150 (1982). 11. L.C. Sander, J. B. Callis, and L. R. Field, Anal. Chem. 55, 1068-1075 (1983).
