The Royal Institution for the Advancement of Learning/McGill University

Published purpose

Organic semiconductors have been intensely investigated in the last two decades because of their wide variety of prominent commercial applications. The device fabrication process and performance heavily depend on the inherent properties of organic semiconductors. In this project the McGill team will collaborate with a team from the NRC and from NYMCTU on developing and applying machine learning (ML) methods and models to condense the available data on organic semiconductors into a ‘computational compass’ to guide the optimization of materials needed for high-performing, stable, and ‘green’ organic photovoltaic (OPV) and thermoelectric (TE) devices. This goal will be achieved by combining datadriven ML approaches with physics-based (but ML-enhanced) simulation that will fill in the details of manifestation of structural diversity and help encode the necessary morphology-function relationships into generative ML models which will be used to propose hypothetical materials. Synthesis, purification, and characterization protocols will be developed for the experimental realization of these novel materials. The work will be done in a close collaboration between the McGill and the NRC teams with HQP visiting the NRC facilities regularly and with the computational effort supported by the experts on the NRC team. The outcomes expected from this work include machine learning models for design of novel OPV and TE materials, new synthesis, purification, and characterization protocols, and ultimately new OPV and TE devices.