QSAR-based model for assessment and prediction of organic micropollutants removal during bank filtration

Recently organic micropollutants have become a concern in scientific world related to their fate and transport in the aquatic life. Pharmaceuticals are a group of organic micropollutants that have been detected as a major contaminant in environment mainly due to uncompleted process of pharmaceuticals removal of wastewater treatment plants. A significant rise at consumption and application of pharmaceuticals recently has increased concentration of pharmaceuticals in aquatic life and may contaminate groundwater even further. Due to the persistence and mobility of pharmaceuticals in environment, it is believed that pharmaceuticals may cause toxic effect in human life although not much information is available to answer this issue. Bank filtration is a robust technology used for water treatment that has high potentials for removal of different contaminants, including some pharmaceuticals and other emerging pollutants, from surface water at low cost. However, previous studies on fate of pharmaceuticals during riverbank filtration do not provide sufficient information or basis for a sustainable design and implementation of riverbank filtration systems for removing these contaminants. Furthermore, experimental data on organic micropollutants removal during soil passage is limited and doing experiment is time consuming and relatively expensive especially for increasing number of pharmaceuticals being produced recently. To address this problem, this study focussed on analysis of the available data and development of QSAR-based model using the physicochemical properties of pharmaceuticals in order to describe the fate of pharmaceuticals during riverbank filtration. Extensive literature review was conducted and removal data for different classes of pharmaceuticals were complied, along with the process and site conditions, in the form of a database. Several descriptors that represent physicochemical properties of pharmaceuticals were calculated by using different commercially available software packages. The best descriptors subsets for the models were selected using the genetic-algorithm and then the developed model was analysed further by multi-linear regression. Based on the data analysis, preliminary guidelines for the estimation of removal of six different classes of pharmaceuticals based on travel time and travel distance were developed. These guidelines can be used as a screening tool to assess removal of pharmaceuticals during bank filtration under given conditions