Researchers at the University of Wisconsin-Madison have discovered a new way to speed up the process of discovering new high-performance polymers. By using machine learning and molecular dynamics simulations, they have discovered several promising polyimides out of a field of 8 million candidates. Polyimides are commonly used in the aerospace, automobile and electronics industries for their excellent mechanical and thermal properties.
Designing New Polymers
Designing new polyimides can be a costly and time-consuming process. To improve this process, the researchers at UW-Madison collected open-source data of the chemical structures of existing dianhydride and diamine/diisocyanate molecules. This data was then used to build a library of 8 million hypothetical polyimides. To organize the enormous number of possible combinations, the researchers used a computer to combine the building blocks together.
The team created multiple machine learning models for the thermal and mechanical properties of polyimides based on experimentally reported values. By using a variety of machine learning techniques, the researchers identified chemical substructures that are most important for determining individual properties. They also incorporated techniques that explain how the machine learning model behaves, allowing human experts to understand why a certain decision was made.
The researchers obtained predictions for the properties of the 8 million hypothetical polyimides using their machine learning models. They then screened the dataset and identified the three best hypothetical polyimides with superior properties to those of existing polyimides. The researchers built all-atom models for their top-three candidates and conducted molecular dynamics simulations to calculate a key thermal property. This simulation was in good agreement with the predictions from the machine learning models which gave the researchers confidence in their predictions.
As a final validation method, the team made one of the new polyimides and performed experiments that demonstrated the material’s excellent heat resistance. The new polyimide could withstand a temperature of about 1,022 degrees Fahrenheit before it started to degrade. In contrast, existing polyimides could endure temperatures only in the range of 392 to 572 degrees F.
The research conducted by the team at UW-Madison has broad implications for the field of materials science. The use of data-driven design frameworks and machine learning predictions can dramatically speed up the discovery of new high-performance polymers. The researchers’ design strategy is much more efficient compared to the conventional trial-and-error process and can also be applied to the molecular design of other polymeric materials. They have also created a web-based application that allows users to explore the new high-performing polyimides with interactive visualization.