Machine learning screening based strategy for the synthesis of a molecularly imprinted ionic liquid polymer for specific adsorption of perfluorooctanoic acid

Abstract

Perfluorooctanoic acid (PFOA) is a persistent organic pollutant that poses a significant environmental threat due to its resistance to degradation in both aquatic and terrestrial environments. Ionic liquids (ILs) are a type of salt that is liquid at or near room temperature and have been widely used in separation science, as well as showing great potential in environmental applications. However, the large number and variety of ILs make them impractical to assess the suitability of all possible ILs for PFOA adsorption using purely experimental methods. To address this issue, a high-throughput machine learning method was employed in this study to screen more than a thousand ILs and identify the most appropriate molecular for designing IL based molecularly imprinted polymer (MIP) with high affinity and selectivity of PFOA. The synthesized MIP, (Vim)C3(L-Pro)@MIP, has been proven to effectively and specifically recognize and adsorb PFOA in water. It has a maximum adsorption capacity of 568.18 mg g−1 and removal efficiency of over 99%, making it a highly efficient and selective option for PFOA removal in environmental remediation applications. This study showcases the promising potential of integrating high-throughput computer screening with experimental verification in the initial stages of MIP material design and development. By doing so, it is possible to create MIP materials that can effectively remove various pollutants, and potentially extend to other IL doped materials and applications.