A solvent selection framework for porous organic polymers

The relationship between solvent, morphology, and properties is crucial for developing functional materials. However, selecting suitable solvents for novel systems remains a significant challenge due to the lack of prior knowledge to guide solvent selections. In this work, we present a solvent selection toolkit for functional porous organic polymers. We have developed an interpretable machine-learning algorithm, MLoc, for the fast prediction of Hansen solubility parameters (HSPs) of novel target materials. This workflow is accessible to non-specialists, operates at negligible computing costs and relies simply on ultraviolet and visible (UV/Vis) absorbance data that can be measured using a standard laboratory setup. We demonstrate successful tuning of both morphology and carbon capture performance for target polymers using MLoc. We describe our vision for MLoc’s broader application in the sustainable development of functional materials, without the need for extensive screening, i.e., in a low data regime. Using MLoc, we report the first HSP database for porous organic polymers, which will serve as a valuable resource for future data-driven research.