Catalyst Discovery by Combining Computational Chemistry with Machine Learning
Machine learning (ML) is revolutionizing the field of materials discovery. This project will fill the knowledge gap in the application of ML to homogeneous catalysis, which has been largely overlooked.
Machine learning (ML) models based on quantum mechanics data will be optimized for the discovery of catalysts enabling artificial photosynthesis and CO2 reduction. We will start the project by computing large datasets with the fundamental properties of metal complexes with high potential in catalysis. The datasets will be transformed into physically meaningful representations, including molecular graphs based on molecular orbital theory. These representations will be mapped into quantum properties by ML models, including neural networks and Gaussian processes.
We will use the ML models to discover new catalysts in the virtually infinite chemical compound space (Lilienfeld et al., Nature Rev. Chem., 2020, 4, 347). The project will build on our proof-of-concept on the machine learning of energy barriers of elementary steps involving transition metal complexes (Balcells et al., Chem. Sci., 2020, 11, 4584).
The key skills developed in the project will include:
- Input/output data processing and analysis;
- Coding catalyst representations;
- Optimizing predictive and generative ML models.
The project will benefit from the generous resources and scientific environment provided by the Hylleraas Centre of Excellence at the Department of Chemistry, and will be developed in collaboration with the Department of Mathematics. Short visits to our collaborator Alán Aspuru-Guzik (University of Toronto, Canada) may be possible.
MSc in chemistry, chemical engineering, materials science, or physics.
Candidates with documented experience in theoretical and computational chemistry, catalysis modeling, python programming, and machine learning will be prioritized.
Call 1: Project start autumn 2021
This project is in call 1, starting autumn 2021. Read about how to apply