Quantitative structure-(chromatographic) retention relationships.

摘要: Since the pioneering works of Kaliszan (R. Kaliszan, Quantitative Structure-Chromatographic Retention Relationships, Wiley, New York, 1987; and R. Structure in Chromatography. A Chemometric Approach, Harwood Academic, Amsterdam, 1997) no comprehensive summary is available field. Present review covers period 1996-August 2006. The sources are grouped according to special properties kinds chromatography: structure-retention relationship gas chromatography, planar column liquid micellar affinity chromatography quantitative structure enantioselective retention relationships. General tendencies, misleading practice conclusions, validation models, suggestions for future summarized each sub-field. Some straightforward applications emphasized but standard ones. model compounds, descriptors, predicted data, modeling methods indicators their performance, stationary phases collected tables. important conclusions are: Not all physicochemical descriptors correlate with data strongly; heat formation not related chromatographic retention. It appropriate give errors Kovats indices percentages. apparently low values (1-3%) can disorient reviewers readers. Contemporary mean interlaboratory reproducibility about 5-10 i.u. non polar 10-25 phases. predictive performance QSRR models deteriorates as polarity GC phase increases. correlation coefficient alone a particularly good indicator performance. Residuals more useful than plots measured calculated values. There need form an equation if numbers compounds small. domain applicability should be given cases.