Application of artificial neural networks for prediction of retention factors of triazine herbicides in reversed-phase liquid chromatography.

In this paper a quantitative structure-retention relationship (QSRR) method is used to model reversed-phase high-performance liquid chromatography (HPLC) behaviour of a series of triazine herbicides and their metabolites. Accurate description of the retention factors in terms of four descriptors related to the analytes and to the mobile phase is achieved by means of an artificial neural network (ANN). For comparison, a QSRR model is derived by multilinear regression (MLR). Validation of the two models shows a better ability in prediction of the ANN as compared with the MLR method. A solid-phase extraction (SPE) procedure allowing the simultaneous determination of the five triazinic compounds in groundwater analysis is also presented. The observed recoveries from water samples range between 85 and 100% for ng/ml concentration levels of all analytes.