https://doi.org/10.1140/epjs/s11734-026-02219-9
Regular Article
Structural and dynamical properties of water–acetonitrile mixture studied by molecular dynamics and enhanced via machine learning
Faculté des Sciences de Monastir, Université de Monastir, Monastir, Tunisia
a
This email address is being protected from spambots. You need JavaScript enabled to view it.
Received:
22
October
2025
Accepted:
14
February
2026
Published online:
13
March
2026
Abstract
We investigate the structural and dynamical properties of water–acetonitrile mixtures at varying mole fractions (XACN = 0.0, 0.5, and 0.9) using classical molecular dynamics simulations, complemented by machine-learning predictions via Random Forest regression. MD simulations reveal the presence of intensified O–O and O–H peaks in the radial distribution functions (RDFs) and a sharp decrease of water coordination number from 4.57 to 1.25 with increasing XACN, indicating a progressive disruption of the hydrogen-bond network. Concurrently, acetonitrile self-organization intensifies, as shown by the increased structuring of N–N RDFs. On the dynamical side, water diffusivity rises by 60% between XACN = 0.5 and 0.9, while hydrogen-bond lifetimes decrease, suggesting the weakening of hydrogen-bond-induced cage effects. To extend the analysis, we employed a supervised machine learning framework based on Random Forest regression to predict key observables: radial distribution functions (RDFs), velocity autocorrelation functions (VACFs), and hydrogen-bond survival functions (CHB(t)). The predicted curves show excellent agreement with simulation data (R2 > 0.999) and integrals derived from them yield coordination numbers, diffusion coefficients, and hydrogen-bond lifetimes with relative errors below 5%. These results highlight the effectiveness of ensemble-based machine-learning approaches in capturing and analyzing both structural and dynamical properties in complex molecular mixtures (Belkin et al. in Proc Natl Acad Sci USA 116:15849–15854, 2019).
Copyright comment Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
© The Author(s), under exclusive licence to EDP Sciences, Springer-Verlag GmbH Germany, part of Springer Nature 2026
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
