The observation of a single IR and Raman active mode for H2 at a Td site is consistent with the isolated molecule being aligned along [110] and unable to rotate.
The calculated vibrational modes for H2 trapped at micro-voids in the crystal are close to the Raman line observed at 4158 cm-1 in samples exposed to hydrogen plasma. This agreement would be expected to improve further if anharmonic corrections to the frequency were included, which are ~240 cm-1 for the free molecule. The three calculated Si-H stretch modes also agree reasonably well with the broad feature observed at ~2100 cm-1. Similarly, the calculated molecular stretch frequency for H2 trapped at <111> platelets of the type favoured by Deák et al also agree with the observed line, as are the calculated band of Si-H stretch modes observed for the plasma treated material.
The effect of Fermi-level position on the diffusion barrier and molecular bond-length of H2 in silicon may explain the observation that laser illumination causes a rapid decrease in the Raman signal intensity. This observation can be explained as a result of the molecule becoming mobile even at cryogenic temperatures and travelling out of the observed region.