Bridging quantum mechanics to liquid properties via a universal organic force field
4 affiliations
- 1ByteDance Seed, Beijing, China. [email protected]
- 2ByteDance Seed, Beijing, China
- 3ByteDance Seed, Bellevue, WA, USA
- 4ByteDance Seed, Bellevue, WA, USA. [email protected]
Abstract
Molecular dynamics (MD) simulations are essential tools for unraveling atomistic insights into the structure and dynamics of condensed-phase systems. However, the universal and accurate prediction of macroscopic properties based on ab initio calculations remains a significant challenge, often hindered by the trade-off between computational cost and simulation accuracy. Here, we present ByteFF-Pol, a graph neural network (GNN)-parameterized polarizable force field, trained exclusively on high-level quantum mechanics (QM) data. Leveraging physically-motivated force field forms and training strategies, ByteFF-Pol exhibits exceptional performance in predicting thermodynamic and transport properties for a wide range of small-molecule liquids and electrolytes, outperforming state-of-the-art (SOTA) classical and machine learning force fields. The zero-shot prediction capability of ByteFF-Pol bridges the gap between microscopic QM calculations and macroscopic liquid properties, enabling the exploration of previously intractable chemical spaces. This advancement holds transformative potential for applications such as electrolyte design and custom-tailored solvent, representing a pivotal step toward data-driven materials discovery.
Related
- NEP89: universal neuroevolution potential for inorganic and organic materials across 89 elements
- Unified Mesoscale Framework Bridging Equilibrium Thermodynamics and Transport Properties of Ionic Liquids and Their Mixtures
- Synthesis and properties of (FeSnPbSb)S quinary sulphides: Bridging mechanochemical and high-entropy concepts
- Bridging Zeolites and Metal-Organic Frameworks: Alkali Metal Cation Modulation in Zn-based UTSA-16
- Non-bridging oxygen as the master variable linking structure to the elastic, electronic, and photonic properties of heavy, polarizable cesium silicate glasses
BibTeX
@article{zheng2026bridging,
title = {{Bridging quantum mechanics to liquid properties via a universal organic force field}},
author = {Tianze Zheng and Xingyuan Xu and Zhi Wang and Zhenze Yang and Yuanheng Wang and Xu Han and Lei Chen and Zhenliang Mu and Ziqing Zhang and Siyuan Liu and Sheng Gong and Kuang Yu and Wen Yan},
journal = {Nature Communications},
year = {2026},
month = may,
doi = {10.1038/s41467-026-73566-3},
url = {https://doi.org/10.1038/s41467-026-73566-3},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Bridging quantum mechanics to liquid properties via a universal organic force field
AU - Tianze Zheng
AU - Xingyuan Xu
AU - Zhi Wang
AU - Zhenze Yang
AU - Yuanheng Wang
AU - Xu Han
AU - Lei Chen
AU - Zhenliang Mu
AU - Ziqing Zhang
AU - Siyuan Liu
AU - Sheng Gong
AU - Kuang Yu
AU - Wen Yan
PY - 2026
DA - 2026/05/28
JO - Nature Communications
DO - 10.1038/s41467-026-73566-3
UR - https://doi.org/10.1038/s41467-026-73566-3
AB - Molecular dynamics (MD) simulations are essential tools for unraveling atomistic insights into the structure and dynamics of condensed-phase systems. However, the universal and accurate prediction of macroscopic properties based on ab initio calculations remains a significant challenge, often hindered by the trade-off between computational cost and simulation accuracy. Here, we present ByteFF-Pol, a graph neural network (GNN)-parameterized polarizable force field, trained exclusively on high-level quantum mechanics (QM) data. Leveraging physically-motivated force field forms and training strategies, ByteFF-Pol exhibits exceptional performance in predicting thermodynamic and transport properties for a wide range of small-molecule liquids and electrolytes, outperforming state-of-the-art (SOTA) classical and machine learning force fields. The zero-shot prediction capability of ByteFF-Pol bridges the gap between microscopic QM calculations and macroscopic liquid properties, enabling the exploration of previously intractable chemical spaces. This advancement holds transformative potential for applications such as electrolyte design and custom-tailored solvent, representing a pivotal step toward data-driven materials discovery.
SN - 2041-1723
ER -