Selectivity of similar compounds identification using IR spectroscopy: steroids

This article focuses on investigating three methods of positive identification or discrimination of a substance when it is compared to a set of similar ones, applying the quantitative criteria and is a continuation of two recent investigations, involving groups of structurally related compounds, namely benzodiazepines and beta-lactam antibiotics. In the course of the present study a set of steroids has been explored. Identification of a substance was based on the comparison of its spectrum with that of a reference material by using several functional algorithms. The algorithm based on the calculation of correlation coefficient between the first derivatives of the spectra has been proved to have the most discriminating power. Then the results in two spectral regions were likened. Limiting values were proposed for correlation coefficients so as to allow for discrimination or qualification of a substance as identical as a reference one. Quantitative criteria are proposed for the identification instead of the qualitative ones recommended by leading pharmacopoeias.