Machine Learning-Powered Simulations of Confinement-Driven ¹²⁹ Xe Nuclear Spin Relaxation
- 1NMR Research Unit, Faculty of Science, University of Oulu, P. O. Box 3000, Oulu, Finland, FI-90014
Abstract
The rate at which a nuclear spin returns to equilibrium encodes the dynamics of its surroundings, information the chemical shift, a time-averaged quantity, cannot provide. Building on a previous dual-model machine learning framework, combining a fine-tuned atomistic foundation model (FT-AFM) for molecular dynamics (MD) with an NMR-ML model predicting the 129 Xe magnetic shielding tensor, we compute the chemical shift anisotropy-driven 129 Xe spin-lattice ( T 1 ) and spin-spin ( T 2 ) relaxation times of xenon confined in carbon nanotubes by extending the Redfield theory treatment to arbitrary nanotube orientations relative to the external magnetic field. Across various temperatures, nanotube diameters, and Xe loadings, both relaxation times evolve non-monotonically in several cases, and for systems whose Xe densities match those of experimental free xenon gas, confinement shortens T 1 by up to fivefold. We also show that T 1 provides an indirect NMR probe of single-file Xe diffusion. This work establishes FT-AFMs and NMR-ML models as a fast, accurate, and general route to the nuclear spin relaxation of confined Xe across a broad range of porous materials.
Related
- Interpretable Machine Learning Disentangles Pore Confinement Across Multiple Zr-MOF CO2 Capture Targets
- Machine-Learning Models Bypass Charge Computation for Modeling 14N Quadrupolar Relaxation in Ionic Liquids
- Growth driven phase transitions in Zinc Oxide nanoparticles through machine-learning assisted simulations
- A Lightweight Universal Machine-Learning Interatomic Potential via Knowledge Distillation for Scalable Atomistic Simulations
- Fast State-of-Health Estimation of Lithium-Ion Batteries Using Short-Term Relaxation-Voltage Features Integrated with a Differential Evolution-Optimized Extreme Learning Machine Algorithm
BibTeX
@misc{zakary2026machine,
title = {{Machine Learning-Powered Simulations of Confinement-Driven ¹²⁹ Xe Nuclear Spin Relaxation}},
author = {Ouail Zakary and Perttu Lantto},
howpublished = {ChemRxiv},
year = {2026},
month = oct,
doi = {10.26434/chemrxiv.15009993/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009993/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Machine Learning-Powered Simulations of Confinement-Driven ¹²⁹ Xe Nuclear Spin Relaxation
AU - Ouail Zakary
AU - Perttu Lantto
PY - 2026
DA - 2026/10/06
PB - ChemRxiv
DO - 10.26434/chemrxiv.15009993/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009993/v1
AB - The rate at which a nuclear spin returns to equilibrium encodes the dynamics of its surroundings, information the chemical shift, a time-averaged quantity, cannot provide. Building on a previous dual-model machine learning framework, combining a fine-tuned atomistic foundation model (FT-AFM) for molecular dynamics (MD) with an NMR-ML model predicting the 129 Xe magnetic shielding tensor, we compute the chemical shift anisotropy-driven 129 Xe spin-lattice ( T 1 ) and spin-spin ( T 2 ) relaxation times of xenon confined in carbon nanotubes by extending the Redfield theory treatment to arbitrary nanotube orientations relative to the external magnetic field. Across various temperatures, nanotube diameters, and Xe loadings, both relaxation times evolve non-monotonically in several cases, and for systems whose Xe densities match those of experimental free xenon gas, confinement shortens T 1 by up to fivefold. We also show that T 1 provides an indirect NMR probe of single-file Xe diffusion. This work establishes FT-AFMs and NMR-ML models as a fast, accurate, and general route to the nuclear spin relaxation of confined Xe across a broad range of porous materials.
ER -