Characterization of porous silicon by solid-state nuclear magnetic resonance

Solid-state nuclear magnetic resonance (NMR) was used to characterize porous silicon (PS) surfaces. On freshly prepared samples, the range of hydrogen content measured by H-1 NMR was equivalent to 0.5-3 monolayers, while fluorine concentrations were below the F-19 NMR detection limit. The H-1 nuclei were used to selectively cross-polarize (CP) Si-29 near the hydrogen passivation. This method was used to study the passivation of an as-prepared, thick (116 mu m), high surface area (893 m(2)/g), photoluminescent (700 nm) PS sample. CP followed by polarization inversion (CPPI) provided some spectral editing. Changes resulting from low-temperature annealing in air and an HF soak were followed by both NMR and infrared spectroscopy. The features of the Si-29 NMR spectra are assigned as (O)(2)(Si)Si-H (-50 ppm), (O)(3)Si-H (-84 ppm), (Si)(3)Si-H (-91 ppm), (Si)(2)Si-H-2 (-102 ppm), and (O)(4)Si (-109 ppm). These assignments are discussed in relationship to experimental measurements and correlations of Si-29 NMR chemical shifts for other materials. The Si-29 NMR line widths for PS fall between those for crystalline silicon and those for amorphous hydrogenated silicon (a-Si:H), suggesting that disorder near the PS surface is intermediate between these extremes. However, comparision of the isotropic chemical shift values shows that the bonding in the disordered regions of PS differs from that found in a-Si:H. In addition, the sharp Si-29 NMR resonance observed in the bulk single crystal starting material cannot be resolved in the spectra of PS. Thus, well-ordered silicon nanocrystallites in the PS are several bond lengths removed from hydrogen or comprise only a small fraction of the PS layer.