ISIS Neutron and Muon Source Data Journal

Probing the generalisability of Machine learning interatomic potentials for Zeolites with inelastic neutron spectroscopy

Abstract: Machine learning interatomic potentials (MLIPs) have emerged as a transformative tool in molecular simulation, offering ab-initio accuracy at a fraction of the computational cost of electronic-structure based methods. This makes them compelling tools for modelling zeolites industrially critical materials whose catalytic and adsorption behaviour depends on large-scale framework flexibility and complex guest host interactions. However, agreement of MLIPs with the static first-principles calculations used for training does not guarantee physically accurate dynamics, and robust experimental validation remains essential. This proposal combines inelastic neutron spectroscopy (INS) measurements with molecular simulation to benchmark universal MLIPs fine-tuned at varying levels of electronic structure theory. Measurements of guest molecules within zeolite frameworks will yield experimental spectra directly comparable with simulated spectra, providing a stringent test of MLIP generalisability across varied conditions. Building on prior work establishing INS as an experimental probe of MLIPs at ISIS, this represents one of the first validations of MLIP dynamics for zeolite systems. Validated potentials will enable atomically resolved interpretation of dynamical features inaccessible to experiment alone, establishing confidence in MLIPs for zeolite processes, and positioning neutron scattering as a critical tool for the assessment of next-generation atomistic simulation methods.

Principal Investigator: Mr Harry Richardson
Experimenter: Dr Alexander O'Malley
Experimenter: Dr Alin Elena
Experimenter: Dr Andrew McCluskey
Local Contact: Dr Jeff Armstrong

DOI: 10.5286/ISIS.E.RB2620468

ISIS Experiment Number: RB2620468

Publisher: STFC ISIS Neutron and Muon Source

Data format: RAW/Nexus
Select the data format above to find out more about it.

Data Citation

The recommended format for citing this dataset in a research publication is as:
[author], [date], [title], [publisher], [doi]

For Example:
Mr Harry Richardson et al; (2026): Probing the generalisability of Machine learning interatomic potentials for Zeolites with inelastic neutron spectroscopy, STFC ISIS Neutron and Muon Source, https://doi.org/10.5286/ISIS.E.RB2620468

Data is released under the CC-BY-4.0 license.