Time-Reversible Implementation of MASH for Efficient Nonadiabatic Molecular Dynamics.

Area of Science:

• Computational Chemistry
• Theoretical Chemistry
• Quantum Dynamics

Background:

• Simulating nonadiabatic dynamics is crucial for understanding chemical reactions.
• Existing surface-hopping methods face challenges with accuracy and computational efficiency.
• The mapping approach to surface hopping (MASH) offers a deterministic alternative.

Purpose of the Study:

• To introduce improved implementations of the MASH method.
• To enhance the efficiency and accuracy of nonadiabatic dynamics simulations.
• To demonstrate the advantages of MASH over stochastic surface-hopping approaches.

Main Methods:

• Development of time-reversible and piecewise-continuous integrators for MASH.
• Implementation of spin propagation using wave-function overlaps or nonadiabatic coupling vectors.
• Analysis of global error scaling with time-step size (Δt).

Main Results:

• The new MASH integrators achieve a global error of $$ per time-step.
• This represents a significant improvement over standard implementations with $$ error.
• Demonstrated ability to use larger time-steps for a given error tolerance or achieve higher accuracy.

Conclusions:

• The enhanced MASH integrators provide significant advantages in efficiency and accuracy.
• MASH's deterministic nature allows for the construction of time-reversible algorithms, unlike stochastic methods.
• These advancements make MASH a more efficient and powerful tool for simulating nonadiabatic dynamics.