Machine Learning for Solid Ion Conductors
Ionic conductors are materials that conduct electricity by the passage of ions. The type of material used for ionic conductors greatly affects the effectiveness and efficiency of electrochemical devices such as fuel cells and batteries. Researchers seeking to identify new ionic conductor materials utilize a combination of small-scale experimentation, high-throughput experimentation, and physics-based simulations. However, the results of these separate systems of experimental discovery can be difficult and time-consuming to integrate in real-time. While each of these methods produce valuable insights, the materials science community lacks the standardized data and analytical techniques necessary to apply advanced machine-learning and data mining for discovery of new materials. This technical gap has resulted in a relatively slow pace of new ionic conductor development, limiting the efficiency and capabilities of battery and fuel cell devices.
Project Innovation + Advantages:
The Citrine Informatics team is demonstrating a proof-of-concept for a system that would use experimental work to intelligently guide the investigation of new solid ionic conductor materials. If successful, the project will create a new approach to material discovery generally and new direction for developing promising ionic conductors specifically. The project will aggregate data (both quantitative and meta-data related to experimental conditions) relevant to ionic conductors from the published literature and build advanced, machine learning models for prediction based upon the resulting large database. The team’s system will also experimentally explore the new materials space identified and suggested by the models. The Citrine project could provide researchers near-real-time feedback as they perform experiments, allowing them to dynamically select the most promising research pathways. This would in turn unlock more rapid ionic conductor identification and development, and transform the fields of theoretical and experimental materials science at-large.