From Spectral Interference to Overlapped Features: A Full-Spectrum LIBS Voigt Decomposition Strategy with Shoulder Detection and Residual Completion

Laser-induced breakdown spectroscopy (LIBS) provides rapid, multielement analysis of complex alloys, but dense line crowding, limited resolution, and plasma broadening cause strong spectral interference and overlapping features that degrade quantitative performance. We propose a full-spectrum Voigt decomposition strategy for LIBS that combines shoulder detection with residual-based line completion. The measured spectrum is first segmented into interference-aware regions of interest (ROIs) using line lists and instrumental resolution. Within each ROI, candidate peaks and shoulders are automatically identified and used as seeds for constrained multicomponent Voigt (or pseudo-Voigt) fitting, with bounds on line width, center drift, and amplitude derived from instrument and plasma conditions. The resulting residual spectrum is then treated as informative: localized residual structures are projected onto Voigt-shaped templates, and their amplitudes are tested statistically to decide whether additional components are required, thereby recovering missed weak or shoulder-like lines. Tests on synthetic spectra constructed from database lines and on experimental LIBS spectra of steels show improved separation of overlapped features, reduced relative standard deviation of line intensities, and enhanced correlations with elemental concentrations. The method provides a scalable, physically grounded framework for decomposing interference- and overlap-affected LIBS spectra.