Application of a constrained regularization method to extraction of affinity distributions: proton and metal binding to humic substances.

The binding of proton and metal cations to humic substances has been analyzed with a regularized fitting procedure (using the CONTIN software package) to extract conditional affinity distributions, valid at a given ionic strength, from binding (titration) curves. The procedure was previously tested with simulated titration curves using a simple bi-Gaussian model, the NICA-Donnan model, and the Stockholm humic model. Application to literature data for proton binding shows that in several cases the affinity distribution found is bimodal (carboxylic and phenolic sites) as usually assumed; however in other cases, specially for fulvic acids, a trimodal distribution is clearly discerned, with a smaller peak between the two noted above attributed to the presence of vicinal carboxylic groups. The analysis of metal binding curves has been performed in a few cases where the available data could be reliably processed, separating the proton affinity distribution and obtaining the conditional affinity spectra. For Cd(II) and Pb(II) a bimodal distribution is found, attributed in principle to mono- and bidentate binding, based on spectroscopic data. In the case of Cu(II), a more complex affinity distribution is found showing 3-4 peaks; this is consistent with spectroscopic studies, where different binding modes, up to tetradentate, have been observed.