Accurate thermal conductivities from optimally short molecular dynamics simulations.

Area of Science:

• Computational Materials Science
• Statistical Mechanics
• Physical Chemistry

Background:

• Calculating transport coefficients (e.g., thermal conductivity, shear viscosity) in extended systems typically requires computationally expensive, long simulations.
• Existing methods face limitations in efficiency and scalability for modern quantum simulation techniques.

Purpose of the Study:

• To introduce a novel, efficient method for computing transport coefficients from short molecular dynamics (MD) simulations.
• To leverage Green-Kubo theory and cepstral analysis for improved accuracy and reduced simulation time.

Main Methods:

• Developed a method combining Green-Kubo theory with cepstral analysis of time series data.
• Utilized information from the full sample power spectrum of relevant currents from short MD trajectories.
• Focused on reducing noise in the zero-frequency value of the power spectrum, which relates to conductivity.

Main Results:

• The proposed method is unbiased and consistent, allowing for arbitrarily small bias and statistical error in the long-time limit.
• Successfully calculated thermal conductivities for diverse systems: liquid argon (Ar), water (H2O), magnesium oxide (MgO), and amorphous silicon dioxide (a-SiO2).
• Achieved approximately 10% accuracy in thermal conductivity estimations using simulation times of only 100-300 picoseconds.

Conclusions:

• The new method offers a computationally efficient alternative for determining transport coefficients.
• Short MD simulations (picosecond timescale) are sufficient for accurate calculations, enabling the use of advanced quantum simulation methods.
• This breakthrough has significant implications for materials design and understanding transport phenomena in various states of matter.