The reactions of alcohols and alkenes with hydrogen-terminated silicon surfaces have been investigated using infrared spectroscopy and deuterium labeling of the reagents and the surface termination. Transmission FTIR spectra were obtained on samples of electrochemically grown porous silicon or mechanically abraded silicon wafers to obtain a sufficient signal-to-noise ratio. The spectral assignments are supported by ab initio calculations on small molecule models at the MP2/6-311++G(d,p) level of theory. A convenient method for the preparation of fully deuterated (D-terminated) silicon wafers is reported; however, fully deuterated porous silicon could not be prepared this way. The spectrum of partially deuterated porous silicon could be assigned on the basis of the computed harmonic vibration frequencies for Q 3 Si-SiH 2 -SiQ 3 and Q 3 Si-SiHD-SiQ 3 where Q is a pseudo-hydrogen atom with the atomic mass of Si. The reaction of O-deuterated alcohols and water on porous silicon produced Si-D stretching and Si-HD scissor modes in the infrared spectrum. The kinetics were consistent with either a dissociative adsorption or an electrochemical corrosion mechanism for this reaction. However, in all cases a net decrease of Si-H/D species on the surface was observed. The magnitude of this decrease is consistent with hydrogen evolution from a hydridic reactivity of the surface termination analogous to the formation of SiO 2 via hydrolysis of molecular hydrosilanes. The Si-H/D, O x Si-H, and Si-O vibrations could be assigned using small molecule models of the form QOSiH 2 SiQ 3 , QOSiH 2 OQ, and (QO) 3 SiH. Significant amounts of silicon alkoxide species are formed even in the presence of water, but the major process in wet solvents is hydrogen evolution and oxide formation. The currently accepted mechanism for the hydrosilylation of alkenes by hydrogen-terminated silicon surfaces involves the attack of a silyl radical on the double bond to produce a Si-C bond and a carbon-centered radical. In principle, this carbon radical may abstract a hydrogen atom from the surface and propagate a chain; however, using deuterated silicon wafers no C-D stretching vibrations could be detected. This indicates that under the conditions employed (1 M alkene solutions in refluxing toluene) the carbon radical abstracts a hydrogen atom from the solvent or another alkene molecule. Ab initio calculations on small molecule models were used to investigate theoretically the shift to low frequency in the Si-H vibrations on the formation of Si-C bonded species at the surface and this effect is attributed to the replacement of Si-H 2 with C-Si-H functionality at the surface.

Alkyl-modified silicon nanocrystallites are efficient fluorophores which are of interest for fundamental spectroscopic studies and as luminescent probes in biology because of their stability in aqueous media. In this work we have investigated these particles using scanning tunneling microscopy, synchrotron-radiation excited photoemission, and x-ray excited optical luminescence (XEOL). During the course of illumination with 145-eV photons we have monitored the evolution of the Si2p core level and, in samples which have suffered prolonged atmospheric exposure, observed in real time the growth of an extra Si2p component attributed to in situ photoinduced oxidation of the Si nanocrystallites. XEOL reveals that two emission bands are active upon soft-x-ray photon excitation and that photoluminescence intensity decreases with photon exposure, which is attributed to charge trapping within the film. (c) 2005 American Institute of Physics

The origin and stability of luminescence are critical issues for Si nanocrystals which are intended for use as biological probes. The optical luminescence of alkyl-monolayer-passivated silicon nanocrystals was studied under excitation with vacuum ultraviolet photons ͑5.1-23 eV͒. Blue and orange emission bands were observed simultaneously, but the blue band only appeared at low temperatures ͑Ͻ175 K͒ and with high excitation energies ͑Ͼ8.7 eV͒. At 8 K, the peak wavelengths of the emission bands were 430± 2 nm ͑blue͒ and 600± 2 nm ͑orange͒. The orange and blue emissions originate from unoxidized and oxidized Si atoms, respectively. © 2006 American Institute of Physics. ͓DOI: 10.1063/1.2216911͔Silicon nanostructures have attracted great interest since the observation of their efficient visible photoluminescence at room temperature in the early 1990s. 1-4 Silicon nanocrystals ͑SiNCs͒ are potential building blocks of future electronic and photonic devices. They also have promise as luminescent labels in biological applications, 5-9 because they show red-orange luminescence at small particle sizes ͑ca. 2 nm diam͒ and are not expected to show some aspects of the toxicity of cadmium-based semiconductors. However, the stability of SiNCs towards O 2 and H 2 O is a concern. Recently, several groups, including ourselves, have prepared alkyl-passivated silicon nanocrystals. [5][6][7][8][9][10][11][12][13][14][15][16] These are SiNCs whose surface is capped by a monolayer of saturated hydrocarbon molecules and anchored to the Si core via covalent Si-C bonds. Although the particles contain a little oxide from their preparation, the alkyl monolayer protects the Si core and they are not oxidized further under ambient conditions. 6 The reaction used to prepare the capping monolayer is sufficiently general to allow manipulation of the chemical functionality on the particle surface, e.g., for synthesis of DNA strands. 6 However, it is also important to characterize the physical properties, especially the luminescence, of these systems because the utility of the particles depends on their luminescence upon insertion into biological cells.Compared to the generally weak IR luminescence of bulk silicon ͑observed at low temperatures͒, the efficiency of photoluminescence ͑PL͒ from Si nanostructures is strongly increased due to the greater overlap of the electron and hole wave functions and the decrease in efficiency of nonradiative pathways. 17,18 A few electronic structure calculations on alkylated SiNCs have been made using density functional methods: these calculations include Si 29 clusters passivated with CH 3 or CH 2 ͑Ref. 19͒ and, more recently, alkyl monolayers up to C 4 H 9 on cluster sizes from Si 20 to Si 142 . 20 Theoretical work suggests that the band gap of SiNCs is little changed upon alkylation of the hydrogen-terminated particle surface, but the positions of the band edges are shifted significantly towards the vacuum level and the properties of the excited states are affected. 20 Experimental studies of the PL mechanism o...

Oligonucleotides have been synthesized on hydrogen-terminated Si(111) and porous silicon using surface hydrosilation of difunctional molecules (1,(omega)-dimethoxytritylundecenol) to produce a monolayer bearing suitable reactive groups to allow automated solid-phase DNA synthesis. The absence of an intervening oxide enables electrochemical characterisation of the surface-bound oligonucleotides. Complementary sequences to the DNA synthesized on Si(111) undergo hybridisation at the surface and a straightforward electrochemical quantitation of the amount of synthesized DNA and its hybridisation efficiency (47%) is possible using Ru(NH3)6(3+) as a redox label. In the case of DNA synthesized in porous silicon, electron transfer (ET) between DNA and the underlying bulk semiconductor can be studied by cyclic voltammetry, however the anisotropic diffusion inside the porous layer and the large resistance of the porous silicon results in voltammograms for which thin-layer behaviour is not observed and the peak currents increase with the square root of scan rate. We interpret these voltammograms in terms of charge transport limitations in the layer of metal centres bound to the DNA inside the pores. Further evidence for this interpretation has been obtained using scanning electrochemical microscopy (SECM) to study the charge transport between redox species in films of DNA synthesized on Si(111) surfaces that are in contact with an aqueous phase. As the bulk concentration of Ru(NH3)6(3+) is reduced below about 250 microM the SECM feedback indicates that the rate of charge transport between surface-bound Ru(NH3)6(3+) exceeds that due to diffusion in the liquid phase. Electrochemical quantitation of the DNA is not possible in this situation, however we have been able to obtain independent determinations using radioassay based on 32P or UV/VIS spectrophotometry of dimethoxytrityl cation cleaved from the porous layer. In the case of the former, use of labelled complementary sequences shows an inverse relationship between the current density used to prepare the porous silicon and the amount of hybridisation. This can be interpreted in terms of the specific surface area of the porous silicon layers since the hybridisation efficiencies (ca. 40%) obtained by comparing DMT+ cleaved from sequences synthesized on the surface and then from complementary sequences after hybridisation were relatively insensitive to the current density used to prepare the layers. Our recent work has also been concerned with individual Si nanocrystals generated by breaking up porous silicon during thermal hydrosilation reactions. FTIR spectroscopy shows these particles are also coated with an organic Si-C-bonded monolayer and they form stable, non-turbid and strongly luminescent (lambdamax = 600-650 nm) dispersions in apolar solvents (L. H. Lie, M. S. Duerdin, E. M. Tuite, A. Houlton and B. R. Horrocks, J. Electroanal. Chem., 2002, 538/539, 183). The effect of carrying out synthetic reactions on the porous silicon prior to breaking up the layer is to produc...

Orange luminescence attributable to a core of silicon atoms in alkyl-capped crystalline quantum dots excited at λa=355 and 405 nm is investigated as a function of applied intensity and time. The intensity of luminescence displays a linear power dependence on the intensity of the applied field, from which an exponent n=0.94±0.02 commensurate with single-photon absorption is derived. The dependence of luminescence on time is observed to be strongly nonexponential and is optimally accounted for by a probability density function which describes a continuous distribution of two decay times: the behavior is characteristic of a pair of elementary steps connected with light emission within a distribution of local environments, or a single rate process supported by two environments. Nonlinear least-squares fits to the time dependent luminescence formulated on this basis with a Gaussian, Lorentzian, or log-normal distribution of rates return most probable lifetimes T¯1=21±1 μs and T¯2=3.7±0.8 μs. The widths of the distributions vary between σ1=0.01–0.03 μs−1 and σ2=0.14–1.1 μs−1 associated with 1/T¯1 and 1/T¯2, respectively.