Comparison and combined use of linear and non-linear fitting for the estimation of complexing parameters from metal titrations of estuarine samples by CLE/AdCSV

Despite the need to determine the concentration and conditional stability constants (K') of natural ligands, we are far from achieving a consensus about the mathematical procedure to use with metal titrations due to the complexity of the samples and the wide range of fitting procedures and problems associated with the selection of the sensitivity (S) of the method. Here, we used Competitive Ligand Exchange/Adsorptive Cathodic Stripping Voltammetry (CLE/AdCSV) empiric data from estuarine waters and computer generated titration sets to compare linear methods with iterative correction of S with non-linear fitting adding S as a parameter. We demonstrate for the first time that, independent of the fitting procedure, S cannot be retrieved if all the ligands present in the sample are not included in the speciation model. We also investigated the variables, apart from analytical noise, that can cause flawed non-linear fittings of titration data. Computer generated data under multiple combinations of analytical conditions showed that a long extension of the titration (at least twice the total ligand concentration for estuarine conditions) and an analytical window (as the side coefficient proportional to') centered below the complexing strength of the natural ligands are essential to produce reliable complexing parameters. We verified, using for the first time a combination of experimental and computer generated data, that faulty estimations of S and K' obtained in empiric titrations of estuarine samples were artifacts of non-linear fitting. Non-linear fitting flaws were caused by a combined effect of the analytical error, the analytical window and the ratio in between the copper concentration and the concentration of the strongest ligands. Here, we recommend for the study of estuarine waters to complement non-linear fitting with iterative linear fitting in order to avoid severe overestimations of Sand the conditional stability constant of strong metal ligands. (C) 2013 Elsevier B.V. All rights reserved.