Machine Learning Predictions of Simulated Self-Diffusion Coefficients for Bulk and Confined Pure Liquids

Abstract

Diffusion properties of bulk fluids have been predicted using empirical expressions and machine learning (ML) models, suggesting that predictions of diffusion also should be possible for fluids in confined environments. The ability to quickly and accurately predict diffusion in porous materials would enable new discoveries and spur development in relevant technologies such as separations, catalysis, batteries, and subsurface applications. In this work, we apply artificial neural network (ANN) models to predict the simulated self-diffusion coefficients of real liquids in both bulk and pore environments. The training data sets were generated from molecular dynamics (MD) simulations of Lennard-Jones particles representing a diverse set of 14 molecules ranging from ammonia to dodecane over a range of liquid pressures and temperatures. Planar, cylindrical, and hexagonal pore models consisted of walls composed of carbon atoms. Our simple model for these liquids was primarily used to generate ANN training data, but the simulated self-diffusion coefficients of bulk liquids show excellent agreement with experimental diffusion coefficients. ANN models based on simple descriptors accurately reproduced the MD diffusion data for both bulk and confined liquids, including the trend of increased mobility in large pores relative to the corresponding bulk liquid.

Nanoconfinement of Carbon Dioxide within Interfacial Aqueous/Ionic Liquid Systems

“…This explanation can be corroborated by comparing the density profile for the [OMIM]­[TFSI]/CO 2 system (Figure S5c) to that of the [EMIM]­[TFSI]/CO 2 system (Figure S4c). Here, the density profiles can be interpreted similarly to RDFs, where particularly large peak values along with small valley values indicate highly structured or immobilized fluids . When looking at the prominent features in the density profiles, the maximum and minimum closest to the pore wall, we find a higher degree of structure for the [EMIM] + system, with an average maximum value/minimum value ratio of 6.45 for the three liquid components, compared with the [OMIM] + system ratio of only 3.78.…”

“…Here, the density profiles can be interpreted similarly to RDFs, where particularly large peak values along with small valley values indicate highly structured or immobilized fluids. 55 When looking at the prominent features in the density profiles, the maximum and minimum closest to the pore wall, we find a higher degree of structure for the [EMIM] + system, with an average maximum value/minimum value ratio of 6.45 for the three liquid components, compared with the [OMIM] + system ratio of only 3.78. This result confirms that the long [OMIM] + alkyl tail disrupts the packing of all liquid components and results in larger diffusion ratios in the 3 nm pore when compared with the [EMIM] + counterpart system.…”

Can Artificial Intelligence Accelerate Fluid Mechanics Research?

“…The supervised approaches rely on labelled data to train the ML algorithm effectively. This is the most common category, with wide applicability in science and technology [4][5][6][7]. Unsupervised ML involves extracting features from high-dimensional data sets without the need for pre-labelled training data.…”

Fluid Properties Extraction in Confined Nanochannels with Molecular Dynamics and Symbolic Regression Methods

“…In bulk systems, static and dynamic property calculation is carried out at the micro- and nano- scale with relations usually coming from statistical mechanics. In nano- and macro-devices, confinement between solid surfaces, the implied boundary conditions, as well as the interaction between fluid and wall atoms lead to fundamentally different mechanics of their mass and energy transport [ 17 ], affect fluid properties [ 18 ], and make their estimation harder [ 19 ]. Molecular dynamics (MD) simulations seem to be the most prominent solution for their investigation, which involves calculating particle positions under a given potential, incorporating Newton’s second law [ 20 , 21 , 22 , 23 , 24 ].…”