Different Functionalization of the Internal and External Surfaces in

Dec 9, 2010 - Ghodsi Mohammadi Ziarani , Zahra Hassanzadeh , Parisa Gholamzadeh ... Patrick V. Almeida , Mohammad-Ali Shahbazi , Ermei M?kil? , Martti...
1 downloads 0 Views 2MB Size
pubs.acs.org/Langmuir © 2010 American Chemical Society

Different Functionalization of the Internal and External Surfaces in Mesoporous Materials for Biosensing Applications Using “Click” Chemistry Bin Guan,† Simone Ciampi,† Guillaume Le Saux,† Katharina Gaus,‡ Peter J. Reece,§ and J. Justin Gooding*,† †

School of Chemistry, The University of New South Wales, Sydney, NSW 2052, Australia, ‡Centre for Vascular Research, The University of New South Wales, Sydney, NSW 2052, Australia, and §School of Physics, The University of New South Wales, Sydney, NSW 2052, Australia Received June 28, 2010. Revised Manuscript Received October 12, 2010

We report the use of copper(I)-catalyzed alkyne-azide cycloaddition reaction (CuAAC) to selectively functionalize the internal and external surfaces of mesoporous materials. Porous silicon rugate filters with narrow line width reflectivity peaks were employed to demonstrate this selective surface functionalization approach. Hydrosilylation of a dialkyne species, 1,8-nonadiyne, was performed to stabilize the freshly fabricated porous silicon rugate filters against oxidation and to allow for further chemical derivatization via “click” CuAAC reactions. The external surface was modified through CuAAC reactions performed in the absence of nitrogen-based CuI-stabilizing species (i.e., ligand-free reactions). To subsequently modify the interior pore surface, stabilization of the CuI catalyst was required. Optical reflectivity measurements, water contact angle measurements, Fourier transform infrared spectroscopy (FTIR), and X-ray photoelectron spectroscopy (XPS) were used to demonstrate the ability of the derivatization approach to selectively modify mesoporous materials with different surface chemistry on the exterior and interior surfaces. Furthermore, porous silicon rugate filters modified externally with the cell-adhesive peptide Gly-Arg-Gly-Asp-Ser (GRGDS) allowed for cell adhesion via formation of focal adhesion points. Results presented here demonstrate a general approach to selectively modify mesoporous silicon samples with potential applications for cell-based biosensing.

1. Introduction Mesoporous materials are attracting considerable research attention for a range of applications as diverse as catalysis,1 filtration and separation,2 gas adsorption and storage,3 enzymes immobilization,4 biomedical tissue regeneration,5 drug delivery,6,7 and chemical/biochemical sensing.8-10 The ability to precisely tune the optical, electrochemical, and/or biological properties of a porous structure among these applications is often achieved via selective chemical routes.9,11 Particularly desirable are chemical strategies that allow for different chemical functionalities to be presented either at the inner pore walls or on the exterior surfaces (i.e., selective functionalization). For example, in an envisioned porous device with ultrahigh surface area for drug delivery applications, the inner pore walls would require a chemistry *Corresponding author. E-mail: [email protected]. (1) De Vos, D. E.; Dams, M.; Sels, B. F.; Jacobs, P. A. Chem. Rev. 2002, 102, 3615–3640. (2) Liu, X.; Du, Y.; Guo, Z.; Gunasekaran, S.; Ching, C.-B.; Chen, Y.; Leong, S. S. J.; Yang, Y. Microporous Mesoporous Mater. 2009, 122, 114–120. (3) Corma, A.; Moliner, M.; Diaz-Cabanas, M. J.; Serna, P.; Femenia, B.; Primo, J.; Garcia, H. New J. Chem. 2008, 32, 1338–1345. (4) Ispas, C.; Sokolov, I.; Andreescu, S. Anal. Bioanal. Chem. 2009, 393, 543– 554. (5) Vallet-Regi, M.; Colilla, M.; Izquierdo-Barba, I. J. Biomed. Nanotechnol. 2008, 4, 1–15. (6) Slowing, I. I.; Trewyn, B. G.; Giri, S.; Lin, V. S. Y. Adv. Funct. Mater. 2007, 17, 1225–1236. (7) Vallet-Regi, M.; Balas, F.; Arcos, D. Angew. Chem., Int. Ed. 2007, 46, 7548– 7558. (8) Kilian, K. A.; Bocking, T.; Gaus, K.; King-Lacroix, J.; Gal, M.; Gooding, J. J. Chem. Commun. 2007, 1936–1938. (9) Kilian, K. A.; Boecking, T.; Gaus, K.; Gal, M.; Gooding, J. J. ACS Nano 2007, 1, 355–361. (10) Jane, A.; Dronov, R.; Hodges, A.; Voelcker, N. H. Trends Biotechnol. 2009, 27, 230–239. (11) Mayne, A. H.; Bayliss, S. C.; Barr, P.; Tobin, M.; Buckberry, L. D. Phys. Status Solidi A 2000, 182, 505–513.

328 DOI: 10.1021/la102599m

compatible with the loading and release of the therapeutic agent,12 while the exterior of the device may present antibodies or peptides aiding selective cell targeting.13,14 Toward this target, cell chips for the label-free optical sensing of proteases released by mammalian cells have been constructed from mesoporous photonic crystals characterized by selective functionalization, and therefore discrete properties, of the external and internal surfaces of the device.15 There have been a few successful strategies reported in achieving different chemistry on the internal and external surfaces of mesoporous materials. Cheng and Landry demonstrated spatial chemical selectivity on mesoporous silica by exploiting the sizeexclusion effect, which slows down the diffusion in the nanoscale pores to ensure that the exterior is modified preferentially to the interior.16 De Juan and Ruiz-Hitzky employed an alternative approach for selective functionalization of the external and internal surfaces of the mesoporous silicate material MCM-41, where the pore spaces being filled by a surfactant template enabled the modification of the exterior surface without the interior being altered.17 However, the requirement for the internal surface to be filled with surfactant template limited the range of application to which this method could be applied. Recently, Kecht et al. investigated site-selective functionalization of mesoporous silica nanoparticles by using sequential co-condensation (12) Zhao, Y.; Trewyn, B. G.; Slowing, I. I.; Lin, V. S. Y. J. Am. Chem. Soc. 2009, 131, 8398–8400. (13) Zhu, C.-L.; Song, X.-Y.; Zhou, W.-H.; Yang, H.-H.; Wen, Y.-H.; Wang, X.-R. J. Mater. Chem. 2009, 19, 7765–7770. (14) Vivero-Escoto, J. L.; Slowing, I. I.; Lin, V. S. Y. Biomaterials 2010, 31, 1325–1333. (15) Kilian, K. A.; Lai, L. M. H.; Magenau, A.; Cartland, S.; Bocking, T.; Di Girolamo, N.; Gal, M.; Gaus, K.; Gooding, J. J. Nano Lett. 2009, 9, 2021–2025. (16) Cheng, K.; Landry, C. C. J. Am. Chem. Soc. 2007, 129, 9674–9685. (17) De Juan, F.; Ruiz-Hitzky, E. Adv. Mater. 2000, 12, 430–432.

Published on Web 12/09/2010

Langmuir 2011, 27(1), 328–334

Guan et al.

of functionalized groups in the process of nanoparticle growth.18,19 Depending on when the functionalized groups are added, the nanoparticles can be modified throughout the whole structure or only in the outer regions. While a very interesting approach, it is only applicable during the formation of mesoporous materials, and thus a more general approach is required to impart desired functionalities to the mesoporous materials. In our previous report, we demonstrated an approach that gives an abrupt change in chemistry between the interior and exterior surfaces of mesoporous silicon based on surface tension and capillarity.20 In this approach, the entire structure is modified with hydrophobic monolayer terminated with a succinimide ester moiety. As water does not penetrate the mesopores, due to the relative hydrophobicity of the succinimide ester, the exterior can be modified simply by placing the material in aqueous solution. Subsequently, upon wetting the interior with an organic solvent, different chemistry can be attached inside the pores. Although a generic solution, this was particularly difficult to perform on the material we were using, porous silicon (PSi), because any oxidation of the silicon in the pores will reduce the surface tension and allow water to penetrate into the pores. A simpler and more general route to achieve different chemistry on the inside and outside of mesoporous materials that does not rely on surface tension is therefore needed. Our main interests are in the label-free biosensing applications of monolayer-modified, and unoxidized, PSi rugate filters.8,9 PSi is a promising biomaterial with advantages including large surface area,21 tunable pore size,22 biocompatibility,23 bioresorbability,24 and, importantly, a metastable and chemically versatile hydrideterminated surface for further derivatization.25 Rugate filters are a class of photonic band-gap structures with a sinusoidal refractive index profile, formed in porous silicon by sinusoidally modulating the porosity of the film on the submicrometer scale during electrochemical etching.26 The structures are characterized by a sharp spectral band of high reflectivity with the central wavelength (λ) given by λ = nd, where n is the average refractive index and d is the physical thickness of each period in rugate filters.27 As the position of this spectral band is sensitive to the average refractive index of the photonic crystal, modification of the interior surface of the porous silicon can be monitored by a change in the reflectivity spectrum.28 That is, attachment of an organic molecule to the pore walls will result in an increase in average refractive index of the material and hence a red shift in the position of the sharp reflectivity band. Conversely, desorption of any molecules from inside of the porous silicon will result in a decrease in average refractive index and a shift of the spectral band to shorter (blue) wavelength. The silicon surface on the interior and exterior of the as-prepared PSi is a hydrogenterminated silicon surface that is ideal for further modification. (18) Kecht, J.; Schlossbauer, A.; Bein, T. Chem. Mater. 2008, 20, 7207–7214. (19) Cauda, V.; Schlossbauer, A.; Kecht, J.; Zurner, A.; Bein, T. J. Am. Chem. Soc. 2009, 131, 11361–11370. (20) Kilian, K. A.; Bocking, T.; Gaus, K.; Gooding, J. J. Angew. Chem., Int. Ed. 2008, 47, 2697–2699. (21) Buriak, J. M. Philos. Trans. R. Soc. London, Ser. A 2006, 364, 217–225. (22) Foell, H.; Christophersen, M.; Carstensen, J.; Hasse, G. Mater. Sci. Eng., R 2002, R39, 93–141. (23) Kilian, K. A.; Bocking, T.; Gaus, K.; Gal, M.; Gooding, J. J. Biomaterials 2007, 28, 3055–3062. (24) Sapelkin, A. V.; Bayliss, S. C.; Unal, B.; Charalambou, A. Biomaterials 2006, 27, 842–846. (25) Stewart, M. P.; Buriak, J. M. Comments Inorg. Chem. 2002, 23, 179–203. (26) Berger, M. G.; Arens-Fischer, R.; Th€onissen, M.; Kr€uger, M.; Billat, S.; L€uth, H.; Hilbrich, S.; Theiss, W.; Grosse, P. Thin Solid Films 1997, 297, 237–240. (27) Ilyas, S.; Boecking, T.; Kilian, K.; Reece, P. J.; Gooding, J.; Gaus, K.; Gal, M. Opt. Mater. 2007, 29, 619–622. (28) Canham, L. Properties of Porous Silicon; IEE: London, 1997; Vol. 1, p 349.

Langmuir 2011, 27(1), 328–334

Article

Starting from the pioneering work of Lindford and Chidsey, reporting on the preparation of Si-C-bound alkyl chain on nonoxidized Si(111)-H surfaces,29,30 an expanding repertoire of wet chemical routes now exists for the formation of covalent Si-C-bound organic monolayers that resist a broad range of conditions31 and allow for further derivatization of the passivated porous device.25,32,33 The ability to impart precise functionalities, and thus to address the optical and biorecognition properties of the hybrid structure, are therefore key features of PSi films.34 In this paper, we report on a chemical strategy based on “click” copper(I)-catalyzed alkyne-azide cycloaddition (CuAAC) reactions35,36 of azide molecules and tethered acetylenes to selectively functionalize the external and internal pore surfaces in PSi rugate filters. Manipulation of the copper(I) environment by controlling the presence/absence of a well-characterized CuI ligand, N,N,N0 , N0 -tetramethylethane-1,2-diamine TMEDA (2; Scheme 1),37 is detailed in the context of achieving this selectivity. This selective chemical approach, in conjunction with the ability of the PSi device to promote specific cell biorecognition events, is also demonstrated.

2. Experimental Methods 2.1. Porous Silicon Rugate Filter Formation. Rugate filters with 40 sinusoidally varying refractive index layers, a porosity variation from 54.5% to 57.5%, and an average pore size of ca. 21 nm were prepared in a custom-made electrochemical cell with 25% hydrofluoric acid ethanolic solution as described in ref 27. Briefly, silicon wafers were cut into pieces (approximately 10  10 mm), cleaned in acetone and ethanol with sonication for 5 min each, and placed in the cell back-contacted with a polished steel electrode. A circular platinum electrode is immersed in the ethanolic HF solution above the wafer. A current density varying between 48 and 53 mA cm-2 was applied to the cell sinusoidally with index matching, apodization, and current breaks (to retain electrolyte concentration at the dissolution front). After etching, the wafer was rinsed with ethanol and pentane and dried with argon. 2.2. Modification of PSi Rugate Filters with 1,8-Nonadiyne (Surface 2). Assembly of the acetylenylated PSi surface by covalent attachment of 1,8-nonadiyne (1) followed a previously reported procedure.32,38 Freshly etched filters were transferred, taking extra care to completely exclude air from the reaction vessel (a custom-made Schlenk flask), to a degassed (through a minimum of five freeze-pump-thaw cycles) sample of diyne 1. The sample was kept under a stream of argon while the reaction vessel was immersed in an oil bath set to 170 °C for 3 h. The flask was then opened to the atmosphere, and the functionalized surface sample rinsed consecutively with copious amounts of dichloromethane and ethanol and then blown dry under a gentle stream of argon, before being either analyzed or further reacted with substituted azide species. (29) Linford, M. R.; Chidsey, C. E. D. J. Am. Chem. Soc. 1993, 115, 12631–12632. (30) Linford, M. R.; Fenter, P.; Eisenberger, P. M.; Chidsey, C. E. D. J. Am. Chem. Soc. 1995, 117, 3145–3155. (31) Canham, L. T.; Reeves, C. L.; Newey, J. P.; Houlton, M. R.; Cox, T. I.; Buriak, J. M.; Stewart, M. P. Adv. Mater. 1999, 11, 1505–1507. (32) Ciampi, S.; Boecking, T.; Kilian, K. A.; Harper, J. B.; Gooding, J. J. Langmuir 2008, 24, 5888–5892. (33) Ciampi, S.; Harper, J. B.; Gooding, J. J. Chem. Soc. Rev. 2010, 39, 2158-2183. (34) Kilian, K. A.; Boecking, T.; Gooding, J. J. Chem. Commun. 2009, 630–640. (35) Tornøe, C. W.; Christensen, C.; Meldal, M. J. Org. Chem. 2002, 67, 3057– 3064. (36) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. Angew. Chem., Int. Ed. 2002, 41, 2596–2599. (37) Collman, J. P.; Zhong, M.; Zhang, C.; Costanzo, S. J. Org. Chem. 2001, 66, 7892–7897. (38) Ciampi, S.; Boecking, T.; Kilian, K. A.; James, M.; Harper, J. B.; Gooding, J. J. Langmuir 2007, 23, 9320–9329.

DOI: 10.1021/la102599m

329

Article

Guan et al. Scheme 1. Selective Functionalization of the External and Internal Surfaces of PSi Rugate Filtersa

a Freshly etched porous silicon films with a metastable hydride-terminated surface (surface 1) were reacted with neat R,ω-diacetylene species 1 to afford the corresponding hydrosilylation product (surface 2). Covalently modified surface 2, favorably presenting an alkyne functionality, gave a convenient dipolarophile in the preparation of surface-bound, 1,4-disubstituted [1,2,3]-triazoles through “click” CuAAC reactions. “Click” reactions in the absence of TMEDA (2) were used to decorate the external pore surface only, with the antifouling tetra (ethylene glycol) derivative 3 (surface 3). Derivatization of the acetylene film throughout the inner pore structure was achieved under ligand (2)-assisted “click” of azide species 4 and 5 (surfaces 4 and 5, respectively).

2.3. Attachment of 11-Azido-3,6,9-trioxaundecan-1-ol to the External Surface of PSi Rugate Filters (Surface 3). In a typical “click” procedure aimed to functionalize the exterior surface only, the following chemicals were added to a reaction tube containing the alkyne-functionalized PSi sample (surface 2): (1) 11-azido-3,6,9-trioxaundecan-1-ol (3; 10 mM, ethanol/water, 1:1), (2) copper(II) sulfate pentahydrate (1 mol % relative to the azide), and (3) sodium ascorbate (25 mol % relative to the azide). Reactions were carried out at room temperature in the dark and stopped after 19 h by removal of the modified sample from the reaction vessel. The prepared surface-bound [1,2,3]-triazoles samples (surface 3) were rinsed consecutively with copious amounts of water and ethanol before being either analyzed or further reacted.

2.4. Attachment of Azide Compounds to the Internal Pore Surface of PSi Rugate Filters (Surfaces 4 and 5). Modification of the interior pore surface with azide species 1-azidooctane (4) and 4-azidophenacyl bromide (5) (surfaces 4 and 5, respectively) followed a procedure analogous to that of the previous section (surface 3). The procedure was modified by the addition of 0.5% (v/v) N,N,N0 ,N0 -tetramethylethane-1,2-diamine (2, TMEDA) to the “click” mixture. 2.5. Optical Characterization. Optical reflectivity spectra were measured in the visible and near-infrared at normal incidence using a custom-built optical arrangement. The setup incorporated a USB2000þ miniature fiber-optic spectrometer (Ocean Optics Inc.) and a fiber-coupled halogen light source (Mikropack GmbH, Germany) and had a spectral resolution of 1 nm and measurement spot size of ∼100 μm. Spectra were processed using custom spectroscopy software platform driven by LabVIEW (National Instruments, TA). 2.6. Surface Characterization. Fourier transform infrared (FTIR) spectra of PSi rugate filters were measured in transmission mode on a ThermoNicolet AVATAR 370-FTIR spectrometer, relative to unmodified Si(100), selecting a 4 cm-1 resolution and accumulating a minimum of 128 scans. X-ray photoelectron spectroscopy (XPS) data were acquired using an ESCALAB 220iXL spectrometer with a monochromated Al KR source (1486.6 eV), hemispherical analyzer, and 330 DOI: 10.1021/la102599m

multichannel detector (six detectors). Spectra were recorded in normal emission with the analyzing chamber operating below 10-9 mbar and selecting a spot size of ∼1 mm2. The incidence angle was set to 58° to the analyzer lens. The resolution of the spectrometer is ca. 0.6 eV as measured from the Ag 3d5/2 signal (full width at half-maximum, fwhm) with a 20 eV pass energy. Survey scans were carried out over 1100-0 eV range with a 1.0 eV step size, a 100 ms dwell time, and analyzer pass energy of 100 eV. High-resolution scans were run with 0.1 eV step size, dwell time of 100 ms, and the analyzer pass energy set to 20 eV. After background subtraction using the Shirley routine, spectra were fitted with a convolution of Lorentzian and Gaussian profiles as previously reported.38 All energies are reported as binding energies in eV and referenced to the Si 2p1/2 signal (corrected to 99.9 eV). Scanning electron microscopy (SEM) images were taken using a Hitachi S900 SEM with a cold field emission source (4 kV). PSi samples were cleaved in the center of the film and mounted on a brass sample base. Sessile water contact angle values were determined on a RameHart 100-00 goniometer. At least three separate spots were measured for each sample.

2.7. GRGDS Attachment and Cell Adhesion Experiments. Tetra(ethylene glycol) EO4 moiety terminated PSi samples (surface 4) were placed into a reaction tube containing (a) N,N0 disuccinimidyl carbonate (0.1 M) and (b) N,N0 -dimethylaminopyridine (0.1 M) in dry acetonitrile for 20 h in the presence of activated molecular sieves (3 A˚). The active carbonate samples were rinsed with ethyl acetate and dichloromethane and then transferred to a peptide Gly-Arg-Gly-Asp-Ser (6, GRGDS) solution (2.0 mM) in Milli-Q water for 1 h. Peptide-modified PSi samples (surface 6) were then rinsed with copious Milli-Q water and ethanol. Paxillin-GFP transfected bovine aortic endothelial cell (BAEC) were incubated in Eagle’s basal medium (EBM) with 1% fetal bovine serum (FBS) at 37 °C at 5% CO2 overnight prior to cell adhesion experiments. Cells were harvested with a trypsin solution in phosphate buffered saline (PBS), suspended in full EBM with 10% FBS, and replated on both GRGDS-terminated and EO4 moiety-terminated PSi surfaces (surfaces 6 and 4, respectively) Langmuir 2011, 27(1), 328–334

Guan et al.

Article

Figure 1. Optical reflectivity shifts of modified rugate filters: (a) hydrogen-terminated sample (surface 1); (b) acetylene-terminated sample (surface 2); (c) “click” modification of the external pore surface, azide 3-modified sample (surface 3); and (d) ligandassisted “click” immobilization of azide 4 onto the internal pore surface (surface 4).

Figure 2. XPS analysis of acetylenyl monolayers assembled from diyne 1 (surface 2): (a) survey scan; (b) C 1s narrow scan; (c) highresolution scan of the Si 2p region.

in a culture dish for 3 h at 37 °C in 5% CO2. PSi rugate filter samples were then washed twice in PBS and fixed in 4% (v/v) paraformaldehyde at room temperature for 15 min, followed by rinsing twice with PBS, and sealed with a glass coverslip using Mowiol (Calbiochem). Cell images were obtained with Nikon Eclipse TE 2000-S epifluorescence microscope fitted with a mercury arc lamp, using a 60 oil immersion objective and appropriate filters. ImageJ 1.41 software (National Institutes of Health) was employed for image processing.

3. Results and Discussion 3.1. Assembly of 1,8-Nonadiyne Monolayers on HydrogenTerminated Porous Silicon Surfaces. Figure 1 displays reflectance spectra of the PSi films at different stages of surface derivatization. Each spectrum is characterized by a sharp reflectance band, surrounded by interference fringes. As depicted in Scheme 1, the immersion of freshly etched hydrogen-terminated PSi rugate filters (surface 1) in 1,8-nonadiyne (1) at 170 °C for 3 h afforded the acetylene-functionalized surface (surface 2) which resulted in 73.9 ( 4.8 nm red shift (under 95% confidence limit) of the high reflectivity resonance position in the recorded reflectivity spectrum (Figure 1b), consistent with an increase in the composite refractive index of the porous substrate, as air is being replaced by organic material.15,32 X-ray photoelectron spectroscopy (XPS) data acquired on the acetylenyl surface 2 are shown in Figure 2. Figure 2a shows a representative survey spectrum that indicates the presence of Si, C, and O only.39 High-resolution narrow scans were collected for the C 1s and Si 2p regions to gain information on bonding configurations and on the presence of any oxidation of the silicon substrate. The narrow scan of the C 1s region (Figure 2b) shows a broad signal (1.5 eV fwhm) with a mean binding energy of 285.2 eV for the fitted function and agrees well with previous (39) The oxygen 1s emission at ca. 532 eV is ascribed to adventitiously adsorbed oxygen. (40) The observed large dispersion value of the C 1s signal is consistent with a signal arising from contributions of either carbon-, hydrogen-, and silicon-bonded carbon atoms being either sp-, sp2-, or sp3-hybridized. Given the chemical composition of the monolayer assembled from the diyne 1, a contribution ascribed to carbon atoms in a carbide carbon-silicon configuration was expected around ca. 284.1 eV but could not be assigned unambiguously.

Langmuir 2011, 27(1), 328–334

Figure 3. Transmission-mode FTIR spectra of fresh-etched PSi rugate filter (a) and acetylenyl PSi film (b) after hydrosilylation.

data.38,40 The high-resolution Si 2p scan reveals important information about the monolayer quality and its ability to prevent appreciable oxidation of the underlying Si substrate.41 The Si 2p XPS spectrum can be deconvoluted into three functions which are attributed to either Si 2p electrons in Si-C bonds (or elemental silicon bonded to hydrogen42) at 100.5 eV (1.5 eV fwhm)43 or to bulk silicon with a Si 2p1/2-Si 2p3/2 spin-orbit split doublet (99.9 and 99.3 eV, respectively). Most importantly, in Si 2p narrow scans (Figure 2c) obtained for surface 2, no significant silicon oxide or suboxide silicon was detected44 in the 102-104 eV region, indicating a well-formed monolayer effectively preventing degradation of the surface chemistry.45 Transmission-mode FTIR evidence further supported the formation of surface 2. Figure 3 shows typical transmission-mode FTIR spectra of hydrogen-terminated (Figure 3a, surface 1) and acetylene-decorated PSi samples (Figure 3b, surface 2). According to the attenuation of Si-Hx stretching modes at ∼2100 cm-1 and to the appearance of ν(CtCH), νa(CH2), νs(CH2), ν(SiCdC), (41) Sieval, A. B.; Linke, R.; Zuilhof, H.; Sudholter, E. J. R. Adv. Mater. 2000, 12, 1457–1460. (42) Uhrberg, R. I. G.; Landemark, E.; Chao, Y. C. J. Electron Spectrosc. Relat. Phenom. 1995, 75, 197–207. (43) Mizokawa, Y.; Geib, K. M.; Wilmsen, C. W. J. Vac. Sci. Technol., A 1986, 4, 1696–1700. (44) Under our XPS experimental conditions the SiOx sublayer detection limit was approximated to 0.06 monolayers. (45) Seitz, O.; Boecking, T.; Salomon, A.; Gooding, J. J.; Cahen, D. Langmuir 2006, 22, 6915–6922.

DOI: 10.1021/la102599m

331

Article

and δ(CH2) modes at 3310, 2932, 2848, 1595, and 1440 cm-1, respectively, the formation of an alkenyl (Si-CdC-R) acetyleneterminated film (surface 2) is proposed. 3.2. Selective “Click” Functionalization of the Exterior and Interior Pore Surface. The external surface of the passivated PSi structure (surface 2) was modified via ligand-free “click” reactions of 11-azido-3,6,9-trioxaundecan-1-ol (3) and surface-bound terminal acetylenes. The tetra(ethylene glycol) moiety (EO4) of azide 3 can impart resistance toward unwanted nonspecific protein adsorption events46 and has been used in the preparation of PSi rugate filters for biosensing applications.23,47 As depicted in Scheme 1, the “click” derivatization of the external surface of PSi samples to yield surface 3 was conducted in the absence of the CuI ligand species (2).48 We previously reported on the key role of compound 2 in allowing for satisfactory yields in the “click” modification of analogous PSi surfaces.32 We tentatively attributed the necessity of copper(I) ligation for achieving appreciable “click” coupling yields inside the pores of PSi to the ability of the ligand 2 to either stabilize CuI species, such that they can penetrate the pores before disproportionation occurs, or protect the CuI catalyst from complexation by tethered triazole rings (i.e., “click” products at the pore entrance). Here we seek to modulate copper ligation to allow for selective chemical functionalization of a porous structure. The reflectivity results presented in Figure 1b,c showed negligible (0.1 ( 1.1 nm, under 95% confidence limit) optical shifts for ligand-free “click” reactions between acetylenyl rugate filters (surface 2) and azide molecule 3. The lack of changes in the optical properties of the rugate filter was consistent with negligible yields for the “click” reaction over the bulk (i.e., interior pore surface) PSi structure. On the other hand, under ligand-assisted “click” reactions the coupling of azide 4 (surface 4) yielded a 12.2 ( 3.9 nm red shift (under 95% confidence limit) in the measured reflectivity spectrum (Figure 1d) and was found in good agreement with previously observed values.20 The successful coupling of alkyl azide 4 onto the interior pore surface afforded a chemically passivated PSi rugate filter with a “bulk” hydrophobic interior environment markedly distinct from the antifouling hydrophilic exterior surface. To provide further evidence that the interior and exterior of the mesoporous rugate filters were bearing chemically distinct functionalities, XPS, FTIR, and surface tension measurements were employed. High-resolution XPS spectra acquired for the externally modified EO4 surface (surface 3) and for the EO4 (exterior surface)/alkyl (pores interior) surface (surface 4) are shown in Figure 4. Evidence of a successful “click” reaction of the EO4 derivative 3 on the external surface was provided by narrow scan of the C 1s and N 1s regions (Figure 4, a and b, respectively). The C 1s core levels were deconvoluted and fitted to three functions: (1) aliphatic carbon-bonded carbon (C-C) centered at 285.0 eV49 (1.6 eV fwhm); (2) a nitrogen-bonded carbon (C-N) peak at ∼286.5 eV (1.5 eV fwhm); and (3) oxygen-bonded carbon (C-O) emission at 287.0 eV (1.6 eV fwhm). The observed binding energies were in good agreement with those reported for the corresponding carbon atoms in chemically analogous surfaces prepared on flat Si(100) surfaces.38 Importantly, representative peak area ratios, namely C-O:C-N, showed experimental values (2.3:1) in very good agreement with those predicted by the (46) Clare, T. L.; Clare, B. H.; Nichols, B. M.; Abbott, N. L.; Hamers, R. J. Langmuir 2005, 21, 6344–6355. (47) Kilian, K. A.; Bocking, T.; Ilyas, S.; Gaus, K.; Jessup, W.; Gal, M.; Gooding, J. J. Adv. Funct. Mater. 2007, 17, 2884–2890. (48) Rodionov, V. O.; Fokin, V. V.; Finn, M. G. Angew. Chem., Int. Ed. 2005, 44, 2210–2215. (49) Wallart, X.; de, V. C. H.; Allongue, P. J. Am. Chem. Soc. 2005, 127, 7871– 7878.

332 DOI: 10.1021/la102599m

Guan et al.

Figure 4. XPS narrow scans of the C 1s and N 1s regions for “click” functionalized mesoporous silicon surfaces: (a, b) distal EO4 moieties external surface/acetylenyl monolayer on the pore walls (surface 3); (c, d) EO4-modified exterior/octyl-terminated hydrophobic surface inside the pores (surface 4).

stoichiometry of the atoms on the putative surface product (2.3:1). As previously reported for immobilized triazoles,38 the best fit to the experimental N 1s signal was obtained with a linear combination of two Gaussian functions with binding energies of 400.5 eV (1.5 eV fwhm) and 402.0 eV (1.2 eV fwhm) and a 2:1 ratio of the integrated areas. A well-resolved peak at 406.0 eV (1.2 eV fwhm) was evident in the recorded XPS spectrum (Figure 4b) and corresponds to the central electron-deficient nitrogen as in the azide group.50 The XPS N 1s data therefore indicate that minor amounts of physically absorbed unreacted azide 3 are present on the “click” modified monolayer. As shown in Figure S1a,b in the Supporting Information, the selective modification of the exterior surface of the diyne samples (surface 2) to yield the “clicked” product (surface 3) did not result in appreciable changes in the measured FTIR spectra. The formation of surface 4 was further supported by C 1s narrow scans showing a decrease in the C-O:C-N ratio of the integrated areas (Figure 4c). Upon ligand-assisted “click” reactions of azide 4, the C-O:C-N ratio was reduced from 2.3:1 (surface 3) to 0.2:1 (surface 4). An increased contribution to the C-N emission (286.5 eV, 1.4 eV fwhm) was attributed to the formation of surface-bound [1,2,3]-triazole moieties decorating the internal pore surfaces of the PSi structure. On the other hand, the C-O contribution to the spectra (287.0 eV, 1.6 eV fwhm) from EO4 moiety remains the same as that of surface 3. Further, the area ratio of C-C contribution (285.0 eV, 1.6 eV fwhm) to the C-O signal increased from 16:1 as for surface 3 to 78:1 (surface 4). The latter evidence was further supported the outcome of the chemical modification as depicted in Scheme 1, suggesting the increased in carbon-bound carbons being due to the presence of aliphatic carbon atoms from the octyl chains of the grafted species 4. XPS N 1s narrow scan (Figure 4d), which shows an increased total photoelectron emission from nitrogen atoms compared to those in Figure 4b (as the narrow scans of N 1s were done under the same instrumental condition), indicated again the successful functionalization of the interior. In addition, the significant decrease of a ν(CtCH) mode at 3310 cm-1 in the FTIR spectrum of the surface 4 PSi sample (Figure S1c, Supporting Information) was also consistent with a positive outcome of (50) Devadoss, A.; Chidsey, C. E. D. J. Am. Chem. Soc. 2007, 129, 5370–5371.

Langmuir 2011, 27(1), 328–334

Guan et al.

Figure 5. FTIR spectra of PSi rugate filters on different stages of surface modification: (a) acetylene-terminated surface (surface 2); (b) EO4 moieties decorating the external surface (surface 3); (c) EO4-modified external surface and phenacyl bromide-terminated internal surface (surface 5).

ligand-assisted “click” reactions used to modify the “bulk” of the PSi structure. Transmission-mode FTIR is a creditable technique to provide complementary chemical information to verify that the surface chemistry proceeded inside the pores. In order to easily identify chemical changes occurring inside the pores of PSi rugate filters, 1-azidophenacyl bromide (5) was employed for the internal surface functionalization because the carbonyl moiety in the grafted molecule (surface 5) is a convenient FTIR marker. According to Scheme 1, the freshly etched PSi surface was first passivated by hydrosilylation of diyne 1. As shown in Figure 5a, the presence of ν(CtCH), νa(CH2), νs(CH2), ν(SiC=C), and δ(CH2) modes at 3310, 2932, 2848, 1595, and 1440 cm-1, respectively, suggests the formation of an alkyne-terminated monolayer (surface 2) via an alkenyl substitution at the hydrogen-terminated silicon surface. These stretching modes remain unchanged (Figure 5b) after exposure of surface 2 to the azide compound 3 in the absence of ligand 2. Under ligand-free CuAAC conditions the reaction occurs on the exterior of the porous structure only, resulting in surface 3 (Scheme 1). Next, the employment of compound 2 to stabilize the CuI catalyst resulted in a successful outcome of CuAAC reactions of azide 5 onto the interior pore surface (surface 5). The appearance of FTIR peaks at 1603 and 1683 cm-1 (Figure 5c), ascribed to the ν(CdO) and ν(Ar) aryl conjugated modes, supports the derivatization of the internal surface of the PSi device (surface 5, Scheme 1). Further, a ∼74 nm red shift of the high reflectivity resonance position on the spectra (Figure 6a,b) was observed and used to infer the formation of acetylene monolayers (surface 2). The “click” reaction between the azide species 3 and surface 2 was then performed in the absence of ligand 2. The fact that the “click” reaction did not occur inside the pores was supported by the absence of an appreciable reflectivity shift in the reaction product (surface 3, Figure 6b,c). Meanwhile, the reaction occurring outside the pores of PSi structure was testified by XPS narrow scans of the C 1s and N 1s regions as shown in Figure S2 of the Supporting Information. Subsequently, ligand-assisted CuAAC reactions of the azide species 5 onto the internal surface of PSi rugate filters resulted in a ∼58 nm red shift in the reflectivity spectrum, thus supporting the formation of surface 5.32 Compared to the ∼12 nm red shift observed for CuAAC reactions of azide species 4 (surface 4), the attachment of azide 5 onto the pore Langmuir 2011, 27(1), 328–334

Article

Figure 6. Optical reflectivity shifts for modified rugate filters: (a) hydrogen-terminated PSi rugate filter (surface 1); (b) acetylene-terminated PSi sample (surface 2); (c) PSi device with EO4terminated external surface and distal acetylene on the interior (surface 3); (d) internal PSi surface modified by reactions with azide 5 (surface 5).

walls surface yielded a larger reflectivity shift, i.e., a bigger increment of the composite refractive index of PSi structure, as phenacyl bromide has a higher refractive index (1.565) than octane (1.39).15 To provide further evidence that selective functionalization was achieved, contact angle and reflectivity measurements were acquired for PSi samples functionalized according to Scheme S1 of the Supporting Information. In brief, PSi rugate filters were first modified with diyne 1 to render the entire structure reasonably hydrophobic, as demonstrated by a contact angle of ∼110° (Figure S3a, Supporting Information). Diyne-modified PSi samples were then immersed in water, and importantly, almost no reflectivity spectral shift, relative to the “dry” samples, was observed. This result shows the water does not penetrate the pores. On the other hand, ligand-free “click” reactions with the EO4 species (surface b, Scheme S1) sees the contact angle of the PSi film dramatically reduced to ∼30° (Figure S3b, Supporting Information), but again, the absence of significant red shifts in the reflectivity spectrum supports the hydrophobic pores with no water penetration. This provides strong evidence that only the exterior surface is modified with the EO4 species under ligand-free conditions. Conversely, when the EO4 molecules was then reacted onto the internal surfaces via ligand-assisted CuAAC reactions (surface c, Scheme S1), the contact angle remains at ∼30° (Figure S3c, Supporting Information), but the reflectivity stop band red-shifts ∼60 nm, showing that water can now penetrate into the EO4 -bearing, and therefore hydrophilic, PSi pores (Figure S4, Supporting Information). 3.3. Cell Adhesion Experiments on RGD-Modified PSi Rugate Filters. The results presented in the previous sections of this report showed how different functionalities on the external and internal surfaces of porous silicon can be achieved via controlling the presence of CuI ligands in the “click” reaction environment. To demonstrate the applicability of this novel selective approach to cell-based biosensing, the external surface of PSi samples was modified with EO4 moieties, reaction followed by the attachment of the peptide GRGDS (6) sequence to distal hydroxyl groups (surface 6, Scheme 2). The presence of the celladhesive peptide sequence Arg-Gly-Asp (RGD) on the exterior of DOI: 10.1021/la102599m

333

Article

Guan et al.

Scheme 2. Functionalization Strategy for the Preparation of GRGDS-Terminated Monolayers on Porous Silicon Rugate Filtersa

Figure 7. Endothelial cells on two different PSi rugate filter samples: (a) PSi surface 6 with distal GRGDS peptide; (b) PSi surface 4 with distal EO4 as the antifouling layer. Scale bar is 5 μm.

a The EO4 (exterior surface)/alkyl (pores interior) modified PSi (surface 4) was activated via reaction of N,N0 -disuccinimidyl carbonate (DSC) with surface hydroxyls, followed by coupling of the peptide sequence GRGDS (6).

PSi devices was intended for the selective capture of mammalian cells via interactions with cell membrane receptors.51 Further, EO4 moieties present on the external surface are expected to act as an antifouling layer and to reduce nonspecific absorption of biomolecules during cell culturing experiments and in future biosensing applications.52 The interior remained octyl-terminated to ensure for selective tethering of the peptide on the external surface only, but the prepared PSi device may allow for future applications requiring the inclusion of lipophilic drugs or a precise control in the delivery patterns.53 Following the incubation with endothelial cells to test for cell adhesion and antifouling properties of the PSi samples, comparison was made between rugate filters modified with the RGD ligand for cell capture (surface 6) and samples presenting EO4 (51) Ruoslahti, E. Annu. Rev. Cell Dev. Biol. 1996, 12, 697–715. (52) Schilp, S.; Rosenhahn, A.; Pettitt, M. E.; Bowen, J.; Callow, M. E.; Callow, J. A.; Grunze, M. Langmuir 2009, 25, 10077–10082. (53) Doadrio, J. C.; Sousa, E. M. B.; Izquierdo-Barba, I.; Doadrio, A. L.; PerezPariente, J.; Vallet-Regi, M. J. Mater. Chem. 2006, 16, 462–466.

334 DOI: 10.1021/la102599m

moieties only (surface 4). Fluorescence micrographs showed that the cells were able to spread and form focal adhesions only on surface samples presenting the RGD ligand (Figure 7a). PSi rugate filters containing distal RGD moieties (surface 6) showed more than 10 times the number of adherent cells, if compared to surface samples derivatized with EO4 moieties only (surface 4, Figure 7b). Furthermore, endothelial cells imaged on surface 4 are predominantly round in shape, suggesting no formation of focal adhesions and thus demonstrating the antifouling properties of tethered oligoether units. Hence, the modification of PSi devices with RGD peptides on its external surface was successful as it enabled cell adhesion and spreading.

4. Conclusions In summary, a robust surface chemistry method based on “click” reactions to fabricate porous silicon with different functionalities on its external and internal surfaces has been realized. Cell adhesion on the modified external surface was also demonstrated. This unveils an important biosensing application of mesoporous silicon with different functionalities to target specific types of cells and monitor simultaneously the behaviors of the target cell. Acknowledgment. The authors thank the Australian Research Council (project no. DP1094564 and CE0384243) for support. Supporting Information Available: Experimental details including general chemicals and synthesis; additional FTIR spectrographs, XPS results, contact angle data, and optical reflectivity spectra. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2011, 27(1), 328–334