Quantum mechanics and machine learning to understand nanoscale water from first-principles

About the Project

Link: https://www.ucl.ac.uk/condensed-matter-material-physics/phd-opportunities

Background: The properties of matter deviate dramatically from their “bulk limit” close to interfaces and when confined in cavities of nanoscale dimensions. These anomalies have widespread implications spanning everyday biological phenomena in our bodies to diverse technologically relevant processes to electronic devices, batteries, and water treatment. While the laws of quantum mechanics (QM) are sufficient to describe these anomalies, the algorithms to solve the QM equations display daunting complexity.

In the era of machine learning: Conventional so-called “first principles” simulations – that aim to treat the quantum mechanics of all electrons and nuclei – display accurate-cost limitations. Simply put, accurate simulations are too computationally expensive, while inexpensive simulations offer inadequate accuracy. However, recent developments in machine learning (ML) and artificial intelligence (AI) bypass the complexities of QM, thereby offering an unprecedented solution to simulate complex materials at the desired quantum mechanical accuracy [1,2,3].

Our recent breakthrough: Our recent study exploiting ML-enabled first-principles simulations has clarified the behaviour of water molecules within nanoconfined spaces, revealing characteristics vastly distinct from bulk states [1]. These findings include new phases like a hexatic phase, an intermediate between a solid and a liquid, and a superionic phase with a greater ionic conductivity than currently used battery materials.

Your PhD goals: Your PhD will revolve around developing methodologies at the intersection of quantum mechanics, statistical mechanics, and ML/AI and their application to model the phase behaviours of water in experimentally accessible nanoconfined cavities of reasonable complexity. This methodology development will be in collaboration with quantum chemistry and ML experts at the University of Cambridge. Furthermore, to ensure our predictions are experimentally accessible, we will team up with leading fabrication and spectroscopy experts at École Normale Supérieure, Paris, and Max Planck Institute for Polymer Research, Mainz.

If you are interested in the fundamentals of quantum and statistical mechanics, have an enthusiasm for computational methods, and the above description speaks to you, please drop me a line with a CV at v.kapil@ucl.ac.uk with the subject “Inquiry for PhD project”.

[1] Kapil, V., Schran, C., Zen, A., Chen, J., Pickard, C. J., & Michaelides, A. (2022). The first-principles phase diagram of monolayer nanoconfined water. Nature, 609(7927), 512–516. https://doi.org/10.1038/s41586-022-05036-x

[2] Kapil, V., & Engel, E. A. (2022). A complete description of thermodynamic stabilities of molecular crystals. Proceedings of the National Academy of Sciences, 119(6), e2111769119. https://doi.org/10.1073/pnas.2111769119

[3] Kapil, V., Kovács, D. P., Csányi, G., & Michaelides, A. (2023). First-principles spectroscopy of aqueous interfaces using machine-learned electronic and quantum nuclear effects. Faraday Discussions, 10.1039.D3FD00113J. https://doi.org/10.1039/D3FD00113J

---