Nonadiabatic Excited-State Molecular Dynamics Methodologies: Comparison and Convergence

Area of Science:

• Computational chemistry
• Theoretical physics
• Quantum mechanics

Background:

• Direct atomistic simulation of nonadiabatic molecular dynamics offers fundamental physical insights but faces challenges.
• Diverse computational frameworks exist, lacking standardized implementations for performance comparison.
• Accurate simulation of excited-state molecular dynamics is crucial for understanding photochemical and photophysical processes.

Purpose of the Study:

• To compare the performance of three popular nonadiabatic molecular dynamics methods: Ehrenfest, surface hopping, and multiconfigurational Ehrenfest with ab initio multiple cloning (MCE-AIMC).
• To evaluate these methods for simulating population relaxation and coherent vibronic dynamics in large chromophores.
• To investigate the numerical convergence of MCE-AIMC algorithms.

Main Methods:

• Implementation of Ehrenfest, surface hopping, and MCE-AIMC methods within the NEXMD software.
• Utilizing a common computational chemistry model for consistent comparisons.
• Analysis of population relaxation and coherent vibronic dynamics.
• Numerical convergence studies for MCE-AIMC, varying trajectory number, cloning thresholds, and wavepacket width.

Main Results:

• Comparative performance analysis of Ehrenfest, surface hopping, and MCE-AIMC methods for large chromophores.
• Identification of method-specific behaviors in population relaxation and vibronic dynamics.
• Demonstration of numerical convergence properties for MCE-AIMC algorithms.
• Provision of reference data for selecting appropriate simulation methodologies.

Conclusions:

• The study provides crucial comparative data for selecting optimal methods in excited-state molecular dynamics simulations.
• Understanding the performance nuances of different algorithms is essential for accurate theoretical predictions.
• The findings aid researchers in choosing the most suitable approach for their specific computational chemistry problems.