Characterisation and classification of the solvents used in liquid chromatography, by main components analysis

Principal components analysis (PCA) presented in this paper as a distinct p art of factor analysis of nine physico-chemical properties for 50 different solvents provided a very useful two component model which explained 74% of the total variance, Each solvent was characterized on a plane by two princ ipal component scores. Scores plots on plane or in tridimensional space (us ing in that case the first three principal components) are very useful as a display tool for examining the relationships between solvents, looking for trends, groupings or outliers, A "polarity chart of solvents" obtained In this way may be successfully used, for instance, in liquid chromatography. in addition a high negative correlation was obtained between Snyder polarit y, P', and lipophilicity coefficient, log P, by the one hand and between bo th these characteristics of solvents and dipole moment and dielectric const ant by the other hand.