Large open-source databases will play an integral role in accelerating the discovery of new polymeric materials. Historically, generation of such databases has been a time-consuming, manual process leading to databases that are either limited in size or limited in accessibility. To overcome this challenge, we have been developing information extraction pipelines to extract polymeric properties from the literature with a focus on automation while acknowledging the essential input of humans. Our initial efforts involved creating a Flory-Huggins chi parameter database relying heavily on crowdsourcing for the extraction of data. Following this work, we found that machine learning could identify the relevant papers reducing the number of papers to review and thus human input. More recently, our efforts involved using natural language processing to automatically extract glass transition temperatures significantly reducing the human input. We report on our current status and future plans. All the data is freely available and stored at the polymer property predictor and database.

Accelerating polymeric database creation

255th National Meeting (2018)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.