X-ray photoelectron spectroscopy (XPS) is widely used in surface and materials laboratories around the world. It is a near surface technique, providing detailed chemical information about samples in the form of survey and narrow scans. To extract the maximum amount of information about materials it is often necessary to peak fit XPS narrow scans. And while indispensable to XPS data analysis, even experienced practitioners can struggle with their peak fitting. In our previous publication, we introduced the equivalent width (EWXPS) as both a possible machine automated method, one that requires less expert judgment for characterizing XPS narrow scans, and as an approach that may be well suited for the analysis of complex spectra. The EWXPS figure of merit was applied to four different data sets. However, as previously noted, other width functions are also regularly employed for analyzing functions. Here we evaluate two other width functions for XPS narrow scan analysis: the autocorrelation width (AWXPS) and the variance (σXPS2). These widths were applied to the same four sets of spectra studied before: (a) four C 1s narrow scans of ozone-treated carbon nanotubes (CNTs) (EWXPS: ∼2.11–2.16eV, AWXPS: ∼3.9–4.1eV, σXPS2: ∼5.0–5.2eV, and a modified form of σXPS2, denoted σXPS2*: ∼6.3–6.8eV), (b) silicon wafers with different oxide thicknesses (EWXPS: ∼1.5–2.9eV, AWXPS: ∼2.28–4.9, and σXPS2: ∼0.7–4.9eV), (iii) hydrogen-terminated silicon surfaces, before and after modification with pentyl groups, and after annealing of the pentyl-terminated surface (EWXPS: ∼0.7–1.0eV, AWXPS: ∼1.2–1.6eV, and σXPS2: ∼0.12–0.19eV), and (iv) C 1s narrow scans from five different nanodiamond samples, three of which showed charging (EWXPS: ∼2.6–4.8eV, AWXPS: ∼3.8–6.9eV, and σXPS2: ∼1.6–4.2eV). All three of the width functions showed similar trends, except in the case of the C 1s spectra of the CNT samples, which were the most complex spectra evaluated, where σXPS2 showed poor correlation with the corresponding O/C ratios. Accordingly, we favor EWXPS and AWXPS. EWXPS is advantageous because it is conceptually simple, giving the most intuitive results. AWXPS has the advantage of not requiring the user to specify the height of the function at its maximum, which will be affected by noise. Because these functions are based on different mathematical operations/algorithms, best practices may involve the calculation of both widths for a set of narrow scans. The standard deviation, σXPS, i.e., the square root of the variance, was also examined. As expected, it gave results similar to σXPS2.
Read full abstract