Insight into selected emerging micropollutant interactions with wastewater colloidal organic carbon: implications for water treatment and analysis

Abstract

This study reports how adding a membrane filter (0.45-μm cellulose nitrate filter) between a glass fibre filter and the solid phase extraction (SPE) cartridge affected the GC/MS analysis of 48 emerging organic micropollutants in wastewater. Most of them are widely used as active pharmaceuticals, cosmetic and packaging material ingredients including classes of parabens, benzophenones and bisphenols among other chemicals tested. A high artificial organic carbon (OC) content in wastewater (DOC = 280 ± 14 mg/L) was investigated to gain insight into micropollutants/colloidal OC filter cake interactions. The results show that even with the use of matrix-matched calibration, the introduction of a second (membrane) filtration step can affect the analysis. Both positive, negative and no effects on the theoretical concentrations calculated from the calibration curves with and without additional filtration were observed. Positive effects on the concentration for the same analyte peak area relative to its surrogate standard were the consequence of a reduced signal for the same concentration, while the negative effects are the consequence of increasing signal for the same concentration. Effect types were dependent on the concentration and the nature of the analytes. Results show that bisphenols and parabens significantly interact with colloidal OC. Statistical analysis of molecular descriptor distribution with effect type showed that micropollutants that have a stronger interaction with colloidal OC have significantly higher ability to act as hydrogen bond donors (HBD) and have larger molar volume (MV). All compounds that experienced either positive or negative effects have a significantly higher median logD. However, further exploration within a single class of compounds (parabens, benzophenones and bisphenols) revealed that selected descriptors are unrelated to an effect type. Pearson’s correlations showed that a correlation exists for certain concentration levels and groups of compounds between a negative effect and MV and logD and a positive effect with MV, MW and rotatable bond (RB) count.