Skip to main content
machine processing samples in laboratory

Revolutionising ways to find essential materials

Published: 15 February 2022

Southampton Professor Graeme Day is leading on a project which will automate the process of finding advanced materials that affect almost everything in our everyday lives. This has the potential to impact environmental outcomes, healthcare, energy regeneration, data storage, and pollution control. It can cut research times from years to weeks, and enable the discovery of materials and properties that otherwise may have never been found.

Benefits of the research

The current ‘trial and error’ method of discovering new advanced materials can take years and can be limited by existing knowledge or bias. By combining pioneering computational methods with automation and robotics it will allow fast and uninhibited breakthroughs in all kinds of research. The computer programs and robots will be open to finding things that are completely new.

A robot can work through the day and night, repeating the same experiments over and over on different molecules, or performing a series of experiments to create a material and characterise its properties. In terms of numbers of experiments, it can achieve in a matter of weeks what a student could achieve through an entire PhD.
Professor of Chemical Modelling

This is a collaborative project with Professor Andy Cooper from the University of Liverpool and Professor Kerstin Thurow from the University of Rostock in Germany. Expertise from Southampton in computational chemistry modelling and machine learning will be joined with the synthesis and characterisation of new materials at Liverpool and robotic and automation expertise at Rostock.

How it works

Robots can do a lot of experiments, but they need to be told what to do.

Southampton are developing computational methods that can propose molecules that look good and predicting how they come together in the solid state. They are working out ways to ask the computers to investigate different types of molecules without having people set up each calculation manually.

The robotics experts are developing the computational brain that will control the robots. The computational modelling will provide the data that will guide the robots decisions.

Our expertise is in predicting what the properties of molecules will be, and the properties of the material those molecules will make. Once we discover promising candidates on the computer, we can create the instructions for robots to prioritise its experiments.
Professor of Chemical Modelling

The future of the project

The €10 million research ADAM (Autonomous Discovery of Advanced Materials) project started in 2020 and is running for six years. The first members of the project team, which will reach 20 researchers at the three institutions, have been recruited over the first few months of the project.

Our partners in Liverpool and Rostock are looking at some of the initial challenges for how they are going to handle materials. Initial experiments to pass information from the computational models to the robots in the lab are due to get underway this summer.

 

Related publications

Angeles Pulido,
Linjiang Chen,
Tomasz Kaczorowski,
Daniel Holden,
Marc A. Little,
Samantha Y. Chong,
Benjamin J. Slater,
David P. Mcmahon,
Baltasar Bonillo,
Chloe J. Stackhouse,
Andrew Stephenson,
Christopher M. Kane,
Rob Clowes,
Tom Hasell,
Andrew I. Cooper,
& Graeme M. Day
, 2017 , Nature , 543 (7647) , 657--664
Type: article
Albert Hofstetter,
Martins Balodis,
Federico M. Paruzzo,
Cory M. Widdifield,
Gabriele Stevanato,
Arthur C. Pinon,
Peter Bygrave,
Graeme M. Day,
& Lyndon Emsley
, 2019 , Journal of the American Chemical Society , 141 (42) , 16624--16634
Type: article
Tom Hasell,
Jamie L. Culshaw,
Samantha Y. Chong,
Marc A. Little,
Kim E. Jelfs,
Edward O. Pyzer-Knapp,
Hilary Shepherd,
Dave J. Adams,
Graeme M. Day,
& Andrew I. Cooper
, 2014 , Journal of the American Chemical Society , 136 (4) , 1438--1448
Type: article
Edward O. Pyzer-Knapp,
Linjiang Chen,
Graeme M. Day,
& Andrew I. Cooper
, 2021 , Science Advances , 7 (33)
Type: article

Related Highlights

Back to top