An approach to the spectral smoothing of Raman data applied to the specific case of thin-film carbon
- PDF / 2,053,015 Bytes
- 11 Pages / 595.276 x 790.866 pts Page_size
- 31 Downloads / 152 Views
An approach to the spectral smoothing of Raman data applied to the specific case of thin‑film carbon Jonathan Laumer1 · Stephen K. O’Leary1 Received: 19 January 2018 / Accepted: 3 April 2018 © Springer Science+Business Media, LLC, part of Springer Nature 2018
Abstract We examine how the Raman spectra associated with thin-films of carbon evolve in response to the amount of smoothing. Our analysis, performed on 93 distinct experimentally acquired Raman spectra, finds that there exists a regime over which the ID/IG evaluation is, to the extent that any experimentally derived quantity may be, independent of the amount of smoothing. We interpret this as corresponding to the regime of “optimal” smoothing and note that the corresponding evaluation of ID/IG seems to be independent of the exact details of the smoothing algorithm employed. A procedure for finding this regime of “optimal” smoothing is laid out. Our hope is that this smoothing procedure can serve as a benchmark for the field, allowing for a proper comparison of Raman spectral data of thin-films of carbon acquired by different researchers.
1 Introduction Raman spectroscopy is often employed in studies of the thinfilm carbon genome [1–4]. It allows for a non-destructive probing of the chemical structure of this material system. A metric derived from the use of Raman spectroscopy, the integrated intensity ratio of the D- and G-peaks, I D/IG, is ubiquitous in the study of thin-film carbon [5–10]. In fact, many properties of thin-film carbon, such as the optical gap, the hardness, and the elastic modulus, have been shown to be directly related to the ID/IG value [1, 7, 11]. Thus, in the characterization of thin-film carbon, there is a need to quantitatively evaluate ID/IG. Given that there are differences in the manner in which Raman spectra are processed in order to determine this aforementioned metric, and given that these interpretational differences will undoubtedly influence the corresponding value obtained, how to fairly compare results acquired by different researchers becomes a matter of critical concern [12–15].1 What is needed is the development of standardized means of processing Raman spectral data. A move towards standardization in Raman spectral processing, specific to the case of thin-film carbon, was provided for in the analysis of Laumer et al. [16]. In their
analysis, Laumer et al. [16] demonstrate how baseline correction and peak decomposition for thin-film carbon Raman spectral data can be performed. They present a means of determining the exact boundaries of the baseline and compare the traditional two Gaussian peak fitting approach with a triple Gaussian peak fitting approach, introducing an S-peak at 1490 cm−1 in order to create a more satisfying fit with the results of experiment. What still remains unclear is how to smooth the Raman spectrum prior to baseline correction and peak decomposition. As with any experimentally acquired spectrum, a Raman spectrum exhibits noise. This noise manifests itself in the wiggles and jiggles that
Data Loading...
