Date: Tuesday 20th of December 2016, 11:00.
Location: JBL2C04 (Joseph Banks Laboratories).

‘Combining theoretical physics with machine learning for materials design’

by David Gao,
Condensed Matter & Materials Physics, University College London, London, UK.

Abstract:

Achieving control over formation of molecular films on insulating substrates is important for designing novel 2D functional materials and devices and combining experimental techniques with theoretical insight is an extremely effective way of studying these systems. We investigated the main factors governing successful control by designing, synthesizing, and depositing organic molecules with interchangeable polar functional groups, a variable length aromatic body, and variable flexibility onto the KCl (001) surface. The knowledge gained and methods developed throughout these studies was then applied towards the rational materials discovery of novel lubricating oil additives by building a database of molecule-surface interactions and shear properties. These results from these calculations can be adapted for use with a graph neural network (GNN) in order to search for novel high performance molecules and mixtures.