[8a-N302-1~11] 18.1 Materials Informatics

11:15 AM - 11:30 AM JST(2:15 AM - 2:30 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[8a-N302-9] Machine Learning Analysis of the Effect of the Spacer Structure on the PFAS’s aggregation

〇Shun Konno¹, Tomoya Oonuki¹, Taisuke Araki¹, Takeshi Hasegawa¹ (1.ICR, Kyoto Univ.)

Keywords:

Machine Learning,PFAS,SAR

PFAS aggregation is very important because it is closely linked to many other physicochemical properties, and thus plays a central role in governing their overall environmental behavior. The aggregation mechanism of the perfluoroalkyl chain has already been well established, but how a spacer structure inserted between the perfluoroalkyl tail and an acidic head group affects aggregation remained unclear. In this study, we used a large set of molecules bearing different spacer structures, and their aggregation strength was evaluated by first-principles-based simulations. A random forest model combined with feature selection and SHAP analysis was then used to identify the descriptors that most strongly govern aggregation strength. The resulting design guidelines show that aggregation can be controlled mainly through polarizability, a rigid, rod-like molecular shape, head-to-tail distance, and dipole moment, providing a molecular design strategy for tuning PFAS aggregation strength.