Optimization of the chromatographic separation of flavonoid compounds in camomile extracts by the simplex method and by the Monte Carlo method is described. Evaluation of the number of peaks at unit resolution (R, = 1) is used as the criterion of separation quality. The Davis-Giddings theory is applied in calculating the number of components and the results are validated by numerical simulations. Peak-purity checks for three identified compounds in commercial sample extracts are reported. Capacity factor patterns for 18 flavonoid and phenolic compounds over an extended range of methanol/aqueous buffer (pH 2.8) mobile phases are described.
Complex mixture separation together with identification and quantitation of selected components are required in many analytical methods for quality control. In the great majority of cases of biological, environmental and industrial interest, the number of components in the sample is unknown and there may be as many as several hundred. In common gradient-elution techniques by reversed-phase high-performance liquid chromatography (h.p.1.c.) with standard columns (5-10 pm), typical available peak capacities [l] are about 50-100. Because uncontrolled peak overlapping always occurs and because part of the chromatographic space may not be completely available for sample separation, complete resolution can be difficult to achieve. The resulting chromatograms of such complex mixtures often have a complex appearance and no quantitative measure of the extent of separation or of individual peak purity is available.
Recently, the many different aspects of such complex analytical problems have been widely studied. Experimental optimization methods of complex analytical procedures have been suggested [ 2-51, overlapping problems in multicomponent chromatograms have been described theoretically [ 1, 6lo], and an index of chromatographic performance, the extent of separation