ABSTRACT
We propose a procedure for estimating acid-base constants in organic solvents or mixture of solvents from the corresponding pK(a) values in aqueous medium and from certain properties of the organic solvents that characterize them. To accomplish this, we developed and validated a chemometric correlation for the calculation of the acid-base constants of different cationic acids in a broad variety of non-aqueous solvents. The parameters chosen for building the model were as follows: the acid-base constant of the compound in aqueous medium and those corresponding to the polarity-polarizability, basicity and acidity scales of the solvent. The results of the fitting were significant (p<0.01), with a root mean error in cross-validation of 18%, with no overfitting. The prediction of the acid-base constants for an external set of compounds had a mean absolute prediction error value of less than 0.8 pK(a) units.
Subject(s)
Acids/chemistry , Models, Chemical , Solvents/chemistry , Cations , Reproducibility of ResultsABSTRACT
A method of capillary electrophoresis (CE) for the determination of triazine herbicides and some of their main metabolites in water samples has been developed. The proposed CE method includes an off-line solid-phase extraction (SPE) procedure with LiChrolut EN sorbent coupled to a non-aqueous capillary electrophoresis (NACE) separation with UV detection. The target compounds were the chloro-s-triazines simazine, atrazine, propazine; the methyltio-s-triazines ametryn and prometryn and three main derivatives from the atrazine degradation products; namely, deethylatrazine, deethylhydroxyatrazine and deisopropylhydroxyatrazine. The analytical characteristics of the CE method are reported. The repeatability of the method was studied considering the different steps of the method separately in order to determine the contributions of each step to the total variability of the method. The NACE-UV results are compared with those obtained with a high performance liquid chromatography with UV detection (HPLC-UV) method. The same off-line SPE procedure was applied to both techniques. The results obtained show that both methods afford the same results in the analysis of surface and drinking water samples, with a level of significance regarding the F- and t-tests greater than 0.05 in all the cases. The detection limits in surface water samples were in the 0.04-0.32 microg l(-1) and 0.11-1.2 microg l(-1) ranges for the NACE-UV and HPLC-UV methods, respectively. The recoveries (spiked/found) were significantly 100% in all cases.
