Uncertainty and biases on the determination of stability constants of metal complexes derived from potentiometric data. What can be expected?

The definition of reliable equilibrium constants is an essential step in speciation studies, as the uncertainty and the consistency of the estimated values concur to define the reliability of the speciation model. It is therefore necessary to evaluate the uncertainty contribution of all factors involved at each step of the data acquisition procedure and to be aware about the consequences of systematic errors on the best-estimated values of experimentally measured equilibrium constants. In this work, a series of computer-generated H+-ion selective electrode titration curves simulating the alkalimetric titration of equimolar mixtures of Zn2+-EDTA in aqueous solutions have been processed by different equilibrium data fitting software. Hence, the uncertainties of the refined stability constants of the Zn2+-EDTA complexes could be derived, while excluding the experimental variability that intrinsically affects all practical experiments. The sensitivity of a given chemical system to different data processing strategies and to possible errors in the input data was evaluated. The systematic errors considered relate to the potential reading, the concentration of the titrant and solution components, and ionic strength variations during titrations. The outcomes highlight how unsuitable decisions taken at the stage of nonlinear least squares fitting of the data can affect the results and underline that the main error contribution is related to the measurement of the glass-electrode potential. The processing of simulated data sets can be a useful tool to alert the researchers to the sensitivity of a given chemical system to different strategies and to possible errors in the input data.