ABSTRACT
Solving many tasks involved with various spectroscopic methods requires the decomposition of a measured spectrum to an additive sum of elementary spectral line shapes. If individual spectral bands strongly overlap each other, it is difficult to visually specify their parameters. Various numerical algorithms were developed for solving the problem of spectral band overlap; computerized spectral profile analysis uses the following procedures or combinations: initial non-optimization analysis working with original spectral data without previous indication of band parameters and iterative optimization of band parameters obtained from the previous method. The spectral resolution achieved in practice is limited by three equipment factors: the digitization range of the experimental spectrum, the signal/noise ratio, and the instrumental resolution. Therefore, the upper boundary of spectral resolution in the deconvoluted spectrum is equal to the equipment resolution.
