Liquid Methanol from DFT and DFT/MM Molecular Dynamics Simulations.

Area of Science:

• Computational Chemistry
• Physical Chemistry
• Materials Science

Background:

• Accurate simulation of liquid methanol is crucial for modeling chemical reactions in solution.
• Existing computational protocols require rigorous evaluation for reliability.

Purpose of the Study:

• To comparatively assess various computational protocols for liquid methanol simulation.
• To identify reliable methods for modeling methanol solutions in chemical reactivity studies.

Main Methods:

• Ab initio molecular dynamics simulations using Density Functional Theory (DFT) with Car-Parrinello Molecular Dynamics (CPMD) and CP2K.
• Evaluation of six DFT functionals (BLYP, BLYP-D2, BLYP-D3, BP86, BP86-D2, B97-D2) with and without dispersion corrections.
• Application of Quantum Mechanics/Molecular Mechanics (QM/MM) frameworks within CPMD and CP2K, using the OPLS model for methanol.

Main Results:

• Full DFT-MD simulations showed difficulty in simultaneously reproducing methanol's structure, dynamics, dipole moments, and IR spectra accurately.
• The B97-D2 functional demonstrated good overall performance in DFT-MD simulations.
• DFT/MM-MD simulations provided a satisfactory description of solvent-solute polarization effects across tested functionals.

Conclusions:

• No single DFT functional excels in all aspects when using full DFT-MD for liquid methanol.
• DFT/MM-MD is a viable and effective approach for modeling methanol solutions, particularly for studying chemical reactivity in explicit environments.