An Efficient Molecular Simulation Methodology for Chemical Reaction Equilibria in Electrolyte Solutions: Application to CO2 Reactive Absorption

We develop a computationally efficient molecular-based simulation algorithm for chemical reaction equilibria in liquids containing neutral and ionic species, which is based on the combination of classical force field and ab initio calculations and permits calculations involving very small species concentrations. We show its application to the reactive absorption of CO2 in aqueous monoethanolamine (MEA) solvent as a benchmark case, the first time that a quantitatively accurate predictive approach requiring no experimental data has been successfully applied to calculate all solution species concentrations for this system, including the partial pressure of CO2 above the solution. The Reaction Ensemble Monte Carlo (REMC) algorithm, the only other generally applicable approach, requires special system-dependent Monte Carlo enhancements for its implementation, and to detect species with very small concentrations requires long simulation times and/or large system sizes. In contrast, the proposed algorithm can be straightforwardly implemented for systems of any molecular complexity using a standard Molecular Dynamics (MD) simulation package capable of calculating free energy changes and can calculate small species concentrations with normal simulation times and system sizes. In addition, the inherent parallelization capability of MD (which is problematic for MC-based approaches) enables the algorithm's computationally efficient implementation. The H2O-MEA-CO2 benchmark system has been the subject of many previous studies based on macroscopic thermodynamic modeling, which primarily involves fitting their parameters (of which reaction pK values are the most important) to experimental data measurements. To make contact with such approaches, we show the translation of the molecular-based quantities to the direct prediction of these parameters and calculate reaction equilibrium in the framework of a Henry Law-based chemical potential model. We consider both the ideal solution form and its extension using the Davies equation for the species activity coefficients. We study a range of temperatures and CO2 solution loadings in a 30 wt % MEA solution and incorporate an uncertainty analysis in our methodology. We find that the uncertainties of the simulated solution species compositions are comparable to those of the available experimental data. We report predictions of minor species compositions of very small magnitude, for which experimental measurements are typically extremely challenging.