Assessing sorption of fluoroquinolone antibiotics in soils from a K compilation based on pure organic and mineral components.

The presence of fluoroquinolone (FQ) antibiotics in soils may cause a threat to human health due to overexposure and the generation of antibiotic resistance genes. Understanding their sorption behavior in soils is important to predict subsequent FQ (bio) availability. Here, FQ sorption in pure soil organic (i.e., humic substances) and mineral (i.e., metal oxides; phyllosilicates) components is evaluated through a solid-liquid distribution coefficient (K (FQ)) dataset consisting of 243 entries originated from 80 different studies, to elucidate their respective contribution to the overall K (FQ) in bulk soils. First, different factors affecting FQ sorption and desorption in each of these soil phases are critically discussed. The strong role of pH in K (FQ), due to the simultaneous effect on both FQ speciation and surface charge changes, encouraged the derivation of normalized sorption coefficients for the cationic, zwitterionic and anionic FQ species in humic substances and in different phyllosilicates. K (FQ) in metal oxides revealed a key role of metal nature and material specific surface area due to complexation sorption mechanisms at neutral pH. Cumulative distribution functions (CDF) were applied to each dataset to establish a sorption affinity range for each phase and to derive best estimate K (FQ) values for those materials where normalized sorption coefficients to FQ species were unavailable. The data analysis conducted in the different soil phases set the basis for a K (FQ) prediction model, which combined the respective sorption affinity of each phase for FQ and phase abundance in soil to estimate K (FQ) in bulk soils. The model was subsequently validated with sorption data in well characterized soils compiled from the literature.