Are Membrane Bioreactors Really More Efficient in Removing Pharmaceutical Substances?—Variance Component Analysis Of Micropollutant Removal

Abstract This study evaluates the influence of micropollutant chemical characteristics on the removal of pharmaceutical substances through three different treatments: membrane bioreactor, full wastewater treatment with final filtration (WWTP), and secondary treatment through a conventional activated sludges system, operated in parallel at realistic sludge retention time (SRT) over three years and four sampling campaigns. Treated wastewater from the WWTP enters the local canal with a low dilution ratio. Therefore, the monitoring of water contamination is of particular interest for a reliable assessment of environmental risk. A total of 39 pharmaceutical substances were screened. While differences among the three types of treatments were found statistically not significant, data analysis performed through a generalized linear model showed that both the influent concentration and the physicochemical characteristics are strong predictors for the removal of micropollutant. SRT had no significance for the three types of treatment of this study. Finally, pharmaceuticals were divided into three major classes based on their influent concentration and removal. A canonical discriminant analysis was used to predict the removals and showed that the pharmaceuticals removal rates are strongly influenced by their hydrophobicity/hydrophilicity and enabled to predict their removal categories with high accuracy ( i.e. , 65% of correct predictions).