PhD Studentship, Machine Learning Methods in Crystal Structure Prediction
Computational Systems Chemistry
Location: Highfield Campus
Closing Date: Tuesday 31 August 2021
Supervisory Team: Prof. Graeme Day
This project will develop computational methods to guide the discovery of new materials, combining state-of-the-art computational methods for crystal structure prediction (CSP) with machine learning approaches for interpreting the predicted crystal structure landscapes.
Advanced materials impact most aspects of our lives, including healthcare, energy production, data storage and pollution control. However, the discovery of new advanced materials is often approached by trial-and-error experimentation. Computational methods are poised to change this by allowing researchers to test hypotheses on the computer, in advance of synthesis.
The computational materials discovery research group (http://www.crystalstructureprediction.net) in Southampton, led by Professor Graeme Day, has pioneered the use of CSP for the discovery of functional molecular materials in areas such as porous materials (Nature, 2017, 543, 657), organic semiconductors (Journal of Materials Chemistry C, 2017,5, 7574-7584, Chemical Science, 2018,9, 1289-1300, Chemical Science, 2020,11, 4922-4933), photocatalysts (Journal of Materials Chemistry A, 2020,8, 7158-7170) and pharmaceuticals (Journal of the American Chemical Society 2020, 142, 39, 16668–16680). These methods rely on structure prediction methods to highlight the most likely crystal packing arrangements for a molecule, from which relevant properties can be simulated. One of the challenges in this work is the over-prediction of crystal structures: CSP methods tend to predict many more crystal structures than are ever observed experimentally.
On this project, you will develop advanced methods for identifying the most likely synthesisable crystal structures, using machine learning approaches for the analysis of computer-generated structural landscapes. The project will implement and validate the use of the generalised convex hull (Phys. Rev. Materials, 2018, 2, 103804) for the identification of synthesisable structures. This is a robust, data-driven approach to find "extremal" structures on energy landscapes that can be stabilised by application of some experimental constraint, eg. fields, doping, interaction with solvent or chemical modification of the constituent molecule.
The project is based in the computational materials discovery research group led by Prof. Graeme Day in the School of Chemistry at the University of Southampton and forms part of a collaboration with Prof. Michele Ceriotti (Ecole Polytechnique Federale de Lausanne), who will co-supervise the student. The student will be part of the Leverhulme Research Centre for Functional Materials Design (https://www.liverpool.ac.uk/leverhulme-research-centre) and will interact with other projects within the Centre, such as for experimental verification of the computational methods.
Experience with computational chemistry and/or programming is an advantage. Applicants do not need to have previous experience with crystal structure prediction, but should have a good degree (equivalent to a UK first or upper second class) in chemistry, materials science or a related discipline, and an enthusiasm for research. Applicants should thrive in a collaborative environment and expect to work closely with other computational chemists in the research group, as well as collaborators in both computational and lab-based environments.
Closing date: 31 August 2021.
Funding: full tuition fees for EU/UK students plus for UK students, an enhanced stipend of £15,285 tax-free per annum for up to 3.5 years.
How To Apply
Applications should be made online, please select the academic session 2021/22 “PhD Chemistry (Full time)” as the programme. Please enter Graeme Day under the proposed supervisor.
Applications should include:
Two reference letters
Degree Transcripts to date
For further information please contact: email@example.com