Copper complexing properties of dissolved organic matter: PARAFAC treatment of fluorescence quenching

Abstract

Fluorescence quenching of the fluorescence is a useful method to determine copper complexing properties of the dissolved organic matter. This technique provides express results often by choosing one or two excitation-emission fluorescence measurement for a straight range of concentration. That why it is not representative of the whole sample. On another hand, using total luminescence spectra gives lots of information that is difficult to manipulate in term of complexing properties. This work focus on a fluorescence quenching experiment carried out on a filtered black water from Rio Negro from north Brazil (Sao Gabriel da Cachoeira) using excitation emission matrix of fluorescence (EEMF) with a large range of copper concentration. Indeed, metal additions were performed using logarithmic increments of the total copper concentration to cover a wide range of concentration from 1.7 × 10−9 to 10−3 mol l−1. These data were treated by two different ways for comparison: on the one hand, data treatment were computed, as usually, at different fluorescence intensity positions using multi-response wavelength method with two ligand, and, on the another hand, a statistical method, parallel factor analysis (PARAFAC), was applied to extract fluorescent component. Then a fitting was done on these PARAFAC components with one or two complexing sites each one. Results show that PARAFAC enables quantitative evaluation of complexing parameters for copper as good as certain multi-response methods: L1T 16.2 × 10−6 mol l−1 (5.0 < log(K1) < 5.8) and L2T 3.0 × 10−4 mol l−1 (2.8 < log(K2) < 3.9). PARAFAC confirms that for this natural sample only two fluorescent ligands are present with two types of site for each component. No residual fluorescence was detected by the statistical treatment. Wall surface interferences were pointed-out as phenomenon to be solve to overcome the limiting efficiency of the total copper concentration range.