The role of polymers in advanced engineering applications is indispensable and spans multiple disciplines including structural materials, clean energy storage, and water filtration. However, their enormous scale of production, limited recycling, and an enduring culture promoting the use of disposable items has manifested into a global ecological crisis. Current mechanical recycling practices lead to property degradation and do not provide sufficient economic incentives for large-scale plastic waste recycling. Consequently, chemical processes are being actively explored which can execute depolymerization through hydrolysis under specific conditions and yield reusable monomers in a closed-loop cycle. Despite recent advances in these circular polymers, more diverse recyclable chemistries which can solvolyze under specific pH conditions, while providing comparable in-service performance are still needed.

Subtle changes in structure like heteroatom substitutions or spacing between functional groups can radically influence the depolymerization reaction, generating a chemically intractable design space of practically infinite monomers. This means that the potential candidates amenable to hydrolysis in extreme pH cannot be feasibly explored experimentally. In this work, we have combined chemical heuristics and high throughput density functional theory (DFT) to calculate the reaction energies (ΔG) of all hydrolysis pathways for a subset of the QM9 molecular database in neutral and acidic pH. Preliminary results indicate that the reaction energy is dependent on both the reacting functional group and its neighboring chemical surroundings. The data is being used to train a Graph Neural Network (GNN) model which can predict ΔG’s of hydrolysis using just the structures of the reactants. The framework will automatically detect all the reaction centers, generate the products, and predict the ΔG’s for all possible hydrolysis routes. This can be employed to automatically screen for pH-specific stability, isolate promising synthesis targets, and eventually accelerate experimental discovery efforts towards effective next-generation circular polymers.

Samuel M. Blau

Research Scientist, Lawrence Berkeley National Laboratory

Appearances