Abstract
Accurate prediction of gas solubility in a liquid is crucial in many areas of chemistry, and a detailed understanding of the molecular mechanism of the gas solvation continues to be an active area of research. Here, we extend the idea of the constant chemical potential molecular dynamics (CμMD) approach to the calculation of the gas solubility in the liquid under constant gas chemical potential conditions. As a representative example, we utilize this method to calculate the isothermal solubility of carbon dioxide in water. Additionally, we provide microscopic insight into the mechanism of solvation that preferentially occurs in areas of the surface where the hydrogen network is broken.
Introduction
Predicting the solubility of gases in liquids is a fundamental challenge with applications ranging from carbon sequestration to pharmaceutical design. However, standard Molecular Dynamics (MD) simulations often struggle with "finite-size" effects. In a typical NVT (constant number of particles, volume, and temperature) simulation, as gas molecules dissolve into the liquid, the vapor phase density decreases. This drop in density lowers the chemical potential, causing the simulation to underestimate the true equilibrium solubility unless a prohibitively large simulation box is used.
Methodology: Constant Chemical Potential MD
To overcome these limitations, the researchers adapted the Constant Chemical Potential Molecular Dynamics (C$\mu$MD) method. The simulation setup consists of a central liquid slab flanked by gas phases.
The core innovation lies in the use of Control Regions (CR) within the gas phase. An external force is applied to gas molecules to maintain a target density ($c_0$) within these regions. If gas molecules leave the CR to dissolve into the liquid, the C$\mu$MD algorithm "pumps" new molecules from an adjacent molecular reservoir into the CR, effectively maintaining a constant chemical potential throughout the simulation.
To analyze the liquid-gas interface, the authors employed the Willard-Chandler (WC) interface method. This approach treats the liquid surface as a dynamic, undulating "skin" rather than a flat plane, allowing for the identification of "crests" (protrusions into the vapor) and "troughs" (indentations into the liquid).
Results and Discussion
The C$\mu$MD method demonstrated high accuracy. For a $CO_2$-water system at 423 K and 100 bar, the calculated solubility ($m_{CO2}$) was 0.0140, showing excellent agreement with the experimental value of 0.0135. In contrast, standard NVT-MD simulations suffered from a 10-18% drop in gas density, leading to significant underestimation.
The Molecular Mechanism of Solvation
The most significant finding of this study is the discovery of a heterogeneous solvation mechanism. The liquid surface is not uniform; the hydrogen-bond (HB) network of water is significantly stronger and more organized in the troughs of the surface waves. Conversely, the network is weaker and more disorganized at the crests.
These structural differences lead to the formation of transient "empty patches"—tiny voids in the water's surface layer. The study found that approximately 70% of the surface's free area is localized at these crests. Consequently, $CO_2$ molecules do not enter the liquid uniformly; they preferentially "leak" through these empty patches at the crests, where the resistance from the water's hydrogen-bond network is lowest.
Conclusion
The C$\mu$MD framework provides a robust and efficient way to simulate gas solubility while avoiding common simulation artifacts. While the accuracy of the results depends on the underlying force fields used (e.g., SPC for water and EPM2 for $CO_2$), the method itself is highly effective. Future work involving more advanced machine-learning-based potentials could further refine these predictions. This approach holds promise for studying complex solvation processes, such as the transport of molecules across biological membranes or through industrial filtration systems.
Summary updated