A re-examination of thin-film silicon's Raman spectrum

We note that differences in processing thin-film silicon's Raman spectral data can lead to quantitative differences in the Raman-related metrics that arise as a corollary. In order to ameliorate this unhappy state-of-affairs, we propose a number of critical steps that provide a move towards a standard approach for processing Raman spectral data associated with thin-film silicon. Seven basic steps are suggested in order to acquire the Raman spectrum associated with a given sample of thin-film silicon: (1) instrument calibration, (2) data acquisition, (3) spike removal, (4) laser-line adjustment, (5) smoothing, (6) baseline correction, and (7) peak-fitting and decomposition. In order to provide some sense as to how these steps influence the form of the Raman spectrum, we perform the post-processing steps, i.e., steps (3) through (7), inclusive, on representative thin-film silicon Raman spectral data sets. In particular, we employ data acquired from samples of ultra-high-vacuum evaporation prepared samples of thin-film silicon for the purposes of this analysis, noting that these samples cover a significant fraction of thin-film silicon's genome. Our hope is that this proposed approach can form a benchmark in the field for the thin-film silicon community.