PhD Research Studentships
Machine-Learning Models for NMR Spectroscopy of Disordered Solids
| Supervisor: | Ricardo GRAU-CRESPO |
| Apply by: | 28 January 2026 |
| Start in: | September (Semester 1) |
Description
The PhD student will develop novel computational approaches for predicting solid-state nuclear magnetic resonance (NMR) spectra of disordered inorganic materials, with a particular focus on solid solutions where diverse local atomic environments contribute to the spectrum.
Solid solutions are fundamental to modern materials chemistry, from catalysts and battery electrodes to ionic conductors and optical materials. NMR spectroscopy is an excellent tool for probing their local chemical environments, which is critical for applications. However, while accurate first-principle models help in the NMR interpretation, they require sampling many atomic configurations and performing costly calculations. Traditional computational approaches, including those used in our previous work [1,2] have been able to reproduce NMR behaviour but at a high computational cost.
This PhD will push the field forward by combining density-functional theory (DFT) with machine learning (ML) to create a scalable framework for predicting NMR parameters across large configuration ensembles. The project will develop ML models at two complementary levels. First, we will use ML-interatomic potentials for structure relaxation and molecular dynamics. DFT-trained machine-learning potentials will be used to accelerate the optimisation and thermal sampling of disordered configurations. This will enable the generation of realistic structural ensembles spanning the relevant local environments. Second, we will use surrogate ML models for NMR chemical shifts and quadrupolar parameters. The student will build models that map local atomic descriptors directly to NMR observables. These models will be trained on high-quality DFT NMR calculations but generalise rapidly across thousands of environments, enabling full spectrum reconstruction at negligible cost. The focus of applications will be on energy materials including materials for thermoelectric, photocatalytic, and photovoltaic devices.
The project will be based in the School of Engineering and Materials Science at Queen Mary University of London (QMUL), within an active research environment at the interface of materials modelling, machine learning, and spectroscopy. Our group has a strong track record of combining machine learning and DFT approaches in materials science research, with a focus on energy materials [3-5]. The student will work closely with collaborators experimental NMR and will have access to high-performance computing resources and state-of-the-art software at QMUL.
This studentship offers an excellent opportunity to acquire advanced skills in electronic-structure theory, machine learning for chemistry and materials, statistical modelling of disorder, and the simulation–experiment interface.
References
[1] Moran, R.F., McKay, D., Tornstrom, P.C., Aziz, A., Fernandes, A., Grau-Crespo, R. and Ashbrook, S.E., 2019. Ensemble-based modeling of the NMR spectra of solid solutions: cation disorder in Y2(Sn, Ti)2O7. J. Am. Chem. Soc., 2019, 141, 17838-17846.
[2] Grau-Crespo, R., Hamad, S., Balestra, S.R.G., Issa, R., Sparks, T.D., Fernandes, A., Griffiths, B.L., Moran, R., McKay, D. and Ashbrook, S.E. Capturing local compositional fluctuations in NMR modelling of solid solutions. Chem. Sci., 2025,16, 19357-19369.
[3] Antunes, L.M., Butler, K.T. and Grau-Crespo, R. Crystal structure generation with autoregressive large language modeling. Nature Commun., 2024, 15, 10570.
[4] Plata, J.J., Blancas, E.J., Márquez, A.M., Posligua, V., Sanz, J.F. and Grau-Crespo, R. Harnessing the unusually strong improvement of thermoelectric performance of AgInTe2 with nanostructuring. J. Mater. Chem. A, 2023, 11, 16734-16742.
[5] Antunes, L.M., Butler, K.T. and Grau-Crespo, R. Predicting thermoelectric transport properties from composition with attention-based deep learning. Mach. Learn. Sci. Techn., 2023, 4, 015037.
Funding
Funded by: China Scholarship CouncilCandidate will need to secure a CSC scholarship.
Under the scheme, Queen Mary will provide scholarships to cover all tuition fees, whilst the CSC will provide living expenses and one return flight ticket to successful applicants.
Eligibility
- The minimum requirement for this studentship opportunity is a good honours degree (minimum 2(i) honours or equivalent) or MSc/MRes in a relevant discipline.
- If English is not your first language, you will require a valid English certificate equivalent to IELTS 6.5+ overall with a minimum score of minimum score of 6.0 in each of Writing, Listening, Reading and Speaking).
- Candidates are expected to start in September (Semester 1).
Contact
For informal enquiries about this opportunity, please contact Ricardo GRAU-CRESPO.
Apply
Start an application for this studentship and for entry onto the PhD FT Materials Science full-time programme (Semester 1 / September start):
Please be sure to quote the reference "SEMS-PHD-697" to associate your application with this studentship opportunity.
| Related website: | https://www.sems.qmul.ac.uk/staff/r.graucrespo | |
| SEMS Research Centre: | ||
| Keywords: | Computational Chemistry, Inorganic Chemistry, Physical Chemistry, Machine Learning, Energy Technologies, Ceramics, Materials Science - Other |