Exploring multicomponent phase space to discover new materials

Professor Brian Cantor explains the potential to discover and explore new materials with valuable properties.  

This is not a straightforward task due to existing knowledge being informed through materials with relatively few components in limited proportions, and therefore current knowledge is based on a reliance on assumptions on linear behaviour associated with component dilution and component independence that no longer apply in concentrated multicomponent materials.  An added complication is that multicomponent solid-solution phases have been shown to have trillions of different local nanostructures, and therefore far too many to be sampled effectively, even when using the most powerful computer modelling techniques.

Strategies such as trial and error, controlled substitution, parameterisation, thermodynamic modelling, atomistic modelling and machine-learning have all been employed as methods to try to explore multicomponent phase space, with varying levels of success. however none of them has proved capable of delivering consistent or guaranteed results.

In the paper 'Exploring multicomponent phase space to discover new materials' (as published in Springer Nature) Professor Cantor provides an overview of the different strategies utilised so far, and their successes and failures.