MSD received an MSci from the University of Cambridge and then a PhD under Ali Alavi in the quantum modelling of H transport through crystalline materials. As a PDRA with Mats Persson at the University of Liverpool he used DFT to study the adsorption of molecules on metal surfaces. In the group of Matt Rosseinsky he computationally modelled bulk crystalline oxides with applications in photocatalysis and fuel cells. He developed the University’s first crystal structure prediction package and expanded his research to other chemistries and to materials with widely varying applications. As a lecturer, MSD continues to apply computational modelling at the atomistic scale to better understand materials and to aid material discovery. He is using machine learning alongside computational chemistry to accelerate the discovery of new functional materials.

 The Liverpool Materials Discovery Server: Background and Application

Computational tools continue to be designed and implemented to aid the discovery of new materials. The utilization of these tools is limited if they are only accessible to computational researchers with relevant technical skills. True advances in material discovery can only be made if these tools are also made easily available to experimental researchers.
In this talk I will present the recently released Liverpool Materials Discovery Server. This web-based server is accessed through a normal web-browser, enabling use of a variety of computational tools with applications in solid-state inorganic chemistry. I will demonstrate the tools, which include the ability to search databases for all entries with compositions which are similar to a queried composition; to predict lithium ion and thermal conductivity from composition alone; to predict the porosity of a metal organic framework for a given linker and metal; and to visualize and fit heat capacity data. I will also summarize some of the scientific advances behind the tools, including the use of the earth movers’ distance to quantify the similarity between two compositions and the compilation of a database of validated experimentally lithium ion conductivities.
We hope that the LMDS will be both a useful resource for material scientists and a helpful example to computational researchers of how computational advances can be made available to the wider research community.