[Strategies for identifying compounds of biological matrices, such as extracts of medicinal plants, using liquid chromatography and UV/VIS spectroscopy without isolation].

For a satisfactory identification of known compounds of biological matrices, such as extracts of medicinal plants--using liquid chromatographic methods (HPLC, TLC) as well as on-line and in-situ off-line UV/VIS spectroscopy--not only the chromatographic, but also the spectroscopic data have to be identical. Using reversed phase (RP) high performance column liquid chromatography (HPLC) the k' values of the reference compound and the substance have to be correlated to each other. The probability of chromatographic identity is high, if the difference between the k' values are less than 0.5%. Between 0.5-2% the identification is on the medium level. If the difference is higher than 2%, the identification level is low, the identification assurance is inadequate. The normal phase (NP) thin-layer chromatographic (TLC) identification probability [IP(Chr)] is defined as the area of the triangle formed by three mobile phases characterised with different total solvent strength (ST) and total selectivity value (SV), where the hRf values are practically identical with the standard (reference) substances. The higher the value of chromatographic identification probability, the best is the probability that two compounds are chromatographically identical. Using HPLC instruments with dioden array detection (DAD) system, with modern software, the identity or the difference of the UV/VIS spectra can be determined. For the in-situ off-line spectroscopic identification probability [IP(Sp)] two criteria have to be fulfilled. First of all, every minima and maxima of the UV and/or VIS spectra have to be practically identical, secondly the ratio of the locale absorbance minima and maxima has to be identical. Therefore all values of locale minima and maxima [gamma min, gamma max], as well as the relative absorbance ratios have to be depicted, respectively. The illustrated correlations between the reference and the substance in the sample to be identified have to give linears in the applied coordinate systems. The regression coefficients (r2) show the goodness of the spectroscopic identification probability [IP(Sp)]. The chromatographic and spectroscopic identification probability data can be classified in 3 levels. If at least one of the criteria is in the low probability level, the compounds to be identified is not in agreement with the reference substances. If all criteria are placed in the medium or high identification level, the two compounds have to be identical with a sufficient probability.