An automated ultraviolet photoelectron spectroscopy method is presented in which feature localization and quantitative fitting are deliberately separated. A physical-width Savitzky-Golay filter identifies candidate secondary-electron-cutoff and Fermi-edge regions, while all reported parameters are fitted to unsmoothed samples. Noise-adaptive thresholds, persistence, directed slope, linearity, and sample support are combined to rank candidate windows. Raw-fit residuals, breakpoint order, plateau support, and metadata checks then control acceptance. The method is evaluated on one development spectrum and four additional scans with different source, bias, repeat-acquisition, and sampling conditions. Across three paired non-control scans, the maximum difference between automatic and analyst-window estimates is 20.1 meV. The evaluation therefore tests the consistency of feature localization and raw-data fitting across the available conditions.

ultraviolet photoelectron spectroscopy,automated analysis,Savitzky-Golay smoothing,feature localization,raw-data fitting