Ariskina Kristina, Galliéro Guillaume, Obliger Amaël
Laboratoire des Fluides Complexes et leurs Réservoirs, University of Pau and Pays de l'Adour - E2S - TOTAL - CNRS, UMR 5150, 64000 Pau, France.
Institut des Sciences Moléculaires, University of Bordeaux - Bordeaux INP - CNRS, UMR 5255, F-33400 Talence, France.
J Chem Phys. 2024 Sep 28;161(12). doi: 10.1063/5.0225299.
We combine the use of molecular dynamics simulations and the generalized Langevin equation to study the diffusion of a fluid adsorbed within kerogen, the main organic phase of shales. As a class of microporous and amorphous materials that can exhibit significant adsorption-induced swelling, the dynamics of the kerogen's microstructure is expected to play an important role in the confined fluid dynamics. This role is investigated by conducting all-atom simulations with or without solid dynamics. Whenever the dynamics coupling between the fluid and solid is accounted for, we show that the fluid dynamics displays some qualitative differences compared to bulk fluids, which can be modulated by the amount of adsorbed fluid owing to adsorption-induced swelling. We highlight that working with the memory kernel, the central time correlation function of the generalized Langevin equation, allows the fingerprint of the dynamics of the solid to appear on that of the fluid. Interestingly, we observe that the memory kernels of fluid diffusion in kerogen qualitatively behave as those of tagged particles in supercooled liquids. We emphasize the importance of reproducing the velocity-force correlation function to validate the memory kernel numerically obtained as confinement enhances the numerical instabilities. This route is interesting as it opens the way for modeling the impact of fluid concentration on the diffusion coefficient in such ultra-confining cases.
我们结合分子动力学模拟和广义朗之万方程的使用,来研究吸附在干酪根(页岩的主要有机相)内的流体的扩散。作为一类能够表现出显著吸附诱导膨胀的微孔和无定形材料,干酪根微观结构的动力学预计在受限流体动力学中发挥重要作用。通过进行有或没有固体动力学的全原子模拟来研究这一作用。每当考虑流体与固体之间的动力学耦合时,我们表明,与本体流体相比,流体动力学表现出一些定性差异,这些差异可因吸附诱导膨胀而由吸附流体的量进行调节。我们强调,使用记忆核(广义朗之万方程的中心时间关联函数)能使固体动力学的特征出现在流体动力学特征上。有趣的是,我们观察到干酪根中流体扩散的记忆核在定性上与过冷液体中标记粒子的记忆核相似。我们强调再现速度 - 力关联函数对于验证数值获得的记忆核的重要性,因为限制会增强数值不稳定性。这条途径很有趣,因为它为模拟在这种超受限情况下流体浓度对扩散系数的影响开辟了道路。