In a fantastic example of cross cohort collaboration Lakshmi Shenoy (Cohort 2) and Chris Woodgate (Cohort 1) are lead authors on a new research article published in Physical Review MaterialsLink opens in a new window. The research work, which also involves international collaborators at EDF R&DLink opens in a new window in Paris and the University of LilleLink opens in a new window, combines first principles and machine learning modelling of stainless-steel alloys used in nuclear power plants. Linking theoretical insight with data-driven approaches gives a good balance of accuracy and transferability. Combining ideas from their respective PhD projects has led to the first joint publication between their research groups.  

Chris joined HetSys Cohort 1 from Warwick’s MMathPhys course for a project on the quantum physics of magnets. Lakshmi joined Cohort 2 from an MSc in applied mathematics at KCL to work on the ENTENTE Horizon 2020Link opens in a new window co-funded project applying atomistic modelling to nuclear materials. In Chris’ second year and Lakshmi’s first year they took the cross-cohort PX918 training moduleLink opens in a new window to develop their electronic structure skills. Lakshmi also took PX920 to provide continuum scale fracture modelling skills needed for her project, while Chris chose a module on the mathematics of machine learning. Lakshmi has contributed to teaching in PX914 and Chris to PX911 and PX912. They worked together on the student-initiated Multiscale Musings podcastLink opens in a new window. These activities helped to establish the scientific relationship needed for a fruitful collaboration. 

Speaking about the research Lakshmi said “Discussions with Julie [Staunton] and Chris brought a physics perspective to a project that was originally more machine learning focused, and that helped us come up the new approach to model collinear spins. It’s great to be part of a CDT where we can have such discussions with scientists working in related but different fields, as that can bring out new insights from angles we may not have considered before.” 

Chris added: “It’s been really rewarding using some of the software and computational methods developed in my own PhD project and applying them to unfamiliar problems studying a material I had not worked on before. The collaboration has given not only given me deeper insight into Lakshmi’s work, but also fed back beneficially into my own research project by providing clearer connections with industrial applications.” 

This research received funding from the HetSys CDT and the ENTENTELink opens in a new window consortium and computing resources from IRENE at TGCC via PRACE, ARCHER2 via the UKCP consortium, the Sulis tier 2 facility and Warwick HPC facilities run by the Scientific Computing Research Technology Platform. 

Article details: 

L. Shenoy, C. D. Woodgate, J. B. Staunton, A. P. Bartók, C. S. Becquart, C. Domain, and J.R. Kermode, Collinear-Spin Machine Learned Interatomic Potential for Fe₇Cr₂Ni AlloyLink opens in a new window, Phys. Rev. Mater. 8, 033804 (2024).