Partial least squares model inversion in the chromatographic determination of triazines in water

Abstract Inside the framework of Analytical Quality by Design, a model-based approach has been developed and used to identify operating conditions (control method parameters) related to the composition and flow rate of the mobile phase for a liquid chromatographic determination with preset quality characteristics. The approach starts by defining these desired characteristics of the intended chromatogram (proper resolution for consecutive peaks and short time of analysis) and then looking for the needed control method parameters via inversion of a Partial Least Squares (PLS) prediction model. The procedure has been applied to the determination of eight triazines (simazine, simetryn, atrazine, ametryn, propazine, terbuthylazine, prometryn and terbutryn) in surface waters by means of SPE-HPLC-DAD. These triazines either are forbidden or have a maximum allowable limit due to their potential toxicity. The experimental verification of the selected parameters showed that the experimental results were significantly equal to those predicted. Besides, the validation of the developed method allowed concluding that accuracy was fulfilled for the eight triazines and there was not bias. With a probability of false positive equal to 0.05, CCβ was less than 3 µg L−1 for every triazine, except for simazine and terbutryn, which was less than 6 µg L−1 being the probability of false negative less than 10-6. No triazine was found, above their maximum allowable concentration, in any of the samples of surface water picked at fifteen different locations, mostly from streams and the Arlanzon river, near Burgos (Spain).