Machine Learned Coarse Grain Water Models for Evaporation Studies

Abstract

<jats:p>Evaporation studies of water using classical molecular dynamics simulations are largely limited due to its high computational expense. We aim at addressing the computational issues by developing a coarse grain model for evaporation of water on solid surfaces by combining four water molecules into a single bead. Most commonly used mono atomic pair potentials like Lennard Jones, Morse, Mie and three body potential like Stillinger-Weber are optimized using a combination of Genetic algorithm and Nelder-Mead algorithm. Among them, Stillinger-Weber based model shows excellent agreement of density and Enthalpy of vaporization with experimental results for a wide range of temperatures. Further, the new water model is used to simulate contact angle of water and thin film evaporation from surfaces with different wettabilities.</jats:p>

Achilles: A Tool for Contact Angle Estimation from Molecular Dynamics Simulations

“…An equilibrated molecular system with 1000 coarse grained molecular dynamics (CGMD) water beads is used on the top of a platinum surface with an FCC 111 structure with an edge length of 3.912 Ang [56]. The beads follow Morse potential [57] and represents 4000 water molecules based on the CGMD water bead force fields [49], [58]. The platinum plate is modeled with dimensions of 15 nm by 15 nm sides and a single layer.…”

“…This means, the contact angle ( ) can vary between 0 deg to 180 deg. Specifically, surfaces are classified based on the wettability as superhydrophobic ( > 150°), hydrophobic ( > 90°), and hydrophilic ( < 90°) [49]. To account for these cases, the software (henceforth in this paper, this refers to our tool -Achilles) will classify the data into two categories -1) hydrophilic and 2) hydrophobic.…”

“…To account for these cases, the software (henceforth in this paper, this refers to our tool -Achilles) will classify the data into two categories -1) hydrophilic and 2) hydrophobic. The first case of hydrophilic surfaces refers to the Figure 1, where the water droplet rests on a flat surface, let us say platinum surface [41], [49]- [51]. Depending on the presence of the strong attractive forces near the surface, highly dense layer(s) can form closer to it [51].…”

The Critical Diameter for Continuous Evaporation Is between 3 and 4 nm for Hydrophilic Nanopores

“…The model is re-parameterized for higher temperatures using a hybrid ANN-GA-NM (Artificial Neural Network, Genetic Algorithm and Nelder Mead) approach and called it as MARTINI-E . In a separate study, the author of the current paper investigated the popular potential functions like LJ, Morse, Mie, and Stillinger–Weber (SW) for its suitability as a CGMD water model and found that Morse potential-based models and machine-learned SW (ML-mW) potential-based models showed good agreement with experimental values. ML-mW has an upper hand on Morse potential when it comes to accuracy and computational speed.…”

Practical Implementation of Molecular Dynamics Code for Beginners Using Python

“…However, often there arises a problem which needs customized molecular simulations. This can be due to the boundary condition, the construction of the system itself, approximate models for introducing new force fields (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) or can be coupling between different time or length scales. To account these studies, graduate students often write their own code which is a cumbersome task (20,21,28).…”