Probing the roles of pH and ionic strength on electrostatic binding of tetracycline by dissolved organic matters: Reevaluation of modified fitting model

The binding performance of dissolved organic matters (DOM) plays a critical role in the migration, diffusion and removal of various residual pollutants in the natural water environment. In the current study, four typical DOMs (including bovine serum proteins BSA (proteins), sodium alginate SAA (polysaccharides), humic acid HA and fulvic acid FA (humus)) are selected to investigate the binding roles in zwitterionic tetracycline (TET) antibiotic under various ionic strength (IS = 0.001–0.1 M) and pH (5.0–9.0). The dialysis equilibration technique was employed to determine the binding concentrations of TET, and the influence of IS and pH on binding performance was evaluated via UV–vis spectroscopy, total organic carbon (TOC), and Excitation-Emission-Matrix spectra (EEM), zeta potentials and molecule size distribution analysis. Our results suggested that carboxyl and phenolic hydroxyl were identified as the main contributors to TET binding based on the fourier transform infrared spectroscopy (FTIR) analysis, and the binding capability of four DOMs followed as HA > FA » BSA > SAA. The biggest binding concentrations of TET by 10 mg C/L HA, FA, BSA and SAA were 0.863 μM, 0.487 μM, 0.084 μM and 0.086 μM, respectively. The higher binding capability of HA and FA is mainly attributed to their richer functional groups, lower zeta potential (HA/FA = −15.92/-13.54 mV) and the bigger molecular size (HA/FA = 24668/27750 nm). IS significantly inhibits the binding interaction by compressing the molecular structure and the surface electric double layer, while pH had a weak effect. By combining the Donnan model and the multiple linear regression analysis, a modified Karickhoff model was established to effectively predict the binding performance of DOM under different IS (0.001–0.1 M) and pH (5.0–9.0) conditions, and the R(2) of linear fitting between experiment-measured logK(DOC) and model-calculated logK(OC) were 0.94 for HA and 0.91 for FA. This finding provides a theoretical basis for characterizing and predicting the binding performance of various DOMs to residual micropollutants in the natural water environment.