Skip to content
DraconiticResearch
Search Sources API
201,683 studies31 sourcesUpdated 29 min ago
PreprintCC BYChemRxiv

Machine-Learning Models Bypass Charge Computation for Modeling 14N Quadrupolar Relaxation in Ionic Liquids

  • Federico Civaia1
  • Yanan Li2,3
  • Florin Teleanu2,3,4,5
  • Simone S. Köcher1,6
  • Josef Granwehr6,7
  • Alexej Jerschow2,3
  • Christoph Scheurer1,6
7 affiliations
  1. 1Theory Department, Fritz Haber Institute of the Max Planck Society, Faradayweg 4–6, Berlin, Germany, 14195
  2. 2Department of Chemistry
  3. 3New York University
  4. 4ELI-NP
  5. 5Horia Hulubei National Institute for Physics and Nuclear Engineering
  6. 6Institute of Energy Technologies, Fundamental Electrochemistry (IET-1), Forschungszentrum Jülich GmbH, Wilhelm-Johnen-Straße, Jülich, Germany, 52425
  7. 7Institute of Technical and Macromolecular Chemistry, RWTH Aachen University, Worringerweg 2, Aachen, Germany, 52074
Open at sourceDOICite

Abstract

Modeling the electric field gradient (EFG) in molecular dynamics (MD) simulations enables the prediction of nuclear magnetic resonance (NMR) relaxation rates thus allowing to unravel the underlying relaxation-active molecular motion. Here, we compare the linear-response Sternheimer model with a novel machine-learning (ML) approach to simulate 14 N quadrupolar T 1 relaxation in the ionic liquid 1-butyl-3-methylimidazolium tetrafluoroborate ([BMIM][BF 4 ]). Improving the parametrization of the Sternheimer model at the low-symmetry nitrogen sites required an electron density-based charge-partitioning scheme. In contrast, the parameter-free ML descriptor directly captured the EFG dynamics, providing a flexible methodology for computing EFG tensors in complex ionic systems. In the extreme-narrowing regime, both models predicted relaxation rates in excellent agreement with experimental measurements. Finally, using a regularized inverse Laplace transform, we deconvolute the dynamics of the EFG tensor and of the cation inertia tensor, and identify molecular rotation as the dominant mode responsible for 14 N quadrupolar relaxation.

Related

  1. Collective Microscopic Dynamics in Aqueous LiTFSI from Dilute to Water-in-Salt Electrolytes: Insights from Quadrupolar NMR Relaxation and Molecular Simulations
    Marinella de Giovanetti, Matthieu Wolf, Iurii Chubak +2
    Journal of Physical Chemistry BPeer-reviewedCC BY 4.0Full textTheoretical & computational chemistryHAL
  2. Machine Learning-Powered Simulations of Confinement-Driven ¹²⁹ Xe Nuclear Spin Relaxation
    Ouail Zakary, Perttu Lantto
    ChemRxivPreprintCC BYFull textPhysical chemistryChemRxiv
  3. From Computation-Ready to Claim-Ready: An Evidence Framework for MOF Machine Learning
    Natalia Lydia Papadantonaki, Shayan Abaei, Joshua Emptoz +1
    ChemRxivPreprintCC BYFull textMaterials & polymersChemRxiv
  4. 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
    Zeyang Liu, Xiankui Wen, Qingxu Guo +3
    BatteriesPeer-reviewedCC BY 4.0Electrochemistry & energyCrossref
  5. Radial Kernel Attention for Long-Range Modeling in Polymer Machine Learning Force Fields
    Wenbo Fu, Chenghui Mao, Chengwen Hu +2
    ChemRxivPreprintLicence not statedGeneral chemistryChemRxiv

Cite

BibTeXRIS
BibTeX
@misc{civaia2026machinelearning,
  title = {{Machine-Learning Models Bypass Charge Computation for Modeling 14N Quadrupolar Relaxation in Ionic Liquids}},
  author = {Federico Civaia and Yanan Li and Florin Teleanu and Simone S. Köcher and Josef Granwehr and Alexej Jerschow and Christoph Scheurer},
  howpublished = {ChemRxiv},
  year = {2026},
  month = sep,
  doi = {10.26434/chemrxiv.15008848/v1},
  url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008848/v1},
  note = {CC BY},
}
RIS
TY  - UNPB
TI  - Machine-Learning Models Bypass Charge Computation for Modeling 14N Quadrupolar Relaxation in Ionic Liquids
AU  - Federico Civaia
AU  - Yanan Li
AU  - Florin Teleanu
AU  - Simone S. Köcher
AU  - Josef Granwehr
AU  - Alexej Jerschow
AU  - Christoph Scheurer
PY  - 2026
DA  - 2026/09/15
PB  - ChemRxiv
DO  - 10.26434/chemrxiv.15008848/v1
UR  - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008848/v1
AB  - Modeling the electric field gradient (EFG) in molecular dynamics (MD) simulations enables the prediction of nuclear magnetic resonance (NMR) relaxation rates thus allowing to unravel the underlying relaxation-active molecular motion. Here, we compare the linear-response Sternheimer model with a novel machine-learning (ML) approach to simulate 14 N quadrupolar T 1 relaxation in the ionic liquid 1-butyl-3-methylimidazolium tetrafluoroborate ([BMIM][BF 4 ]). Improving the parametrization of the Sternheimer model at the low-symmetry nitrogen sites required an electron density-based charge-partitioning scheme. In contrast, the parameter-free ML descriptor directly captured the EFG dynamics, providing a flexible methodology for computing EFG tensors in complex ionic systems. In the extreme-narrowing regime, both models predicted relaxation rates in excellent agreement with experimental measurements. Finally, using a regularized inverse Laplace transform, we deconvolute the dynamics of the EFG tensor and of the cation inertia tensor, and identify molecular rotation as the dominant mode responsible for 14 N quadrupolar relaxation.
ER  -