Communication: Improved ab initio molecular dynamics by minimally biasing with experimental data.

Area of Science:

• Computational chemistry
• Physical chemistry
• Materials science

Background:

• Ab initio molecular dynamics (AIMD) accurately simulates electrons and nuclei simultaneously.
• AIMD struggles to reproduce properties of systems like water due to electronic density functional inaccuracies.
• Current solutions involve empirical corrections or increased simulation temperatures.

Purpose of the Study:

• To develop a maximum-entropy approach for directly incorporating limited experimental data into AIMD simulations.
• To improve the accuracy of AIMD simulations by applying a minimal bias.
• To gain physical insights into density functional theory (DFT) inaccuracies.

Main Methods:

• Developed a maximum-entropy method to introduce minimal bias using experimental data.
• Performed biased AIMD simulations for water and excess proton in water systems.
• Analyzed both biased and unbiased observables to assess simulation improvements.

Main Results:

• Biased AIMD simulations showed significantly improved properties for water and excess proton systems.
• Improvements were observed for both targeted (biased) and non-targeted (unbiased) observables.
• The approach provided new physical insights into the limitations of the underlying DFT.

Conclusions:

• The maximum-entropy approach effectively enhances AIMD accuracy by integrating experimental data.
• This method offers a direct way to correct for density functional inaccuracies without empirical adjustments.
• The technique provides valuable insights into the fundamental limitations of electronic structure methods in simulations.