Shadow excited state molecular dynamics with the ΔSCF method

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Theoretical Chemistry

Background:

• Born-Oppenheimer molecular dynamics (MD) is crucial for simulating molecular behavior.
• Simulating excited-state molecular dynamics (ESMD) presents significant computational challenges.
• Existing methods often face limitations in accuracy, stability, or computational cost.

Purpose of the Study:

• To extend the shadow extended Lagrangian Born-Oppenheimer molecular dynamics (ESMD) method for excited state dynamics.
• To develop a more computationally efficient and robust approach for ESMD.
• To demonstrate the method's efficacy using self-consistent charge density functional tight binding (SCC-DFTB) theory.

Main Methods:

• Introduced a shadow ESMD approach replacing iterative solutions with a single-step solution of an approximate shadow excited-state potential.
• Linearized the energy functional about an approximate excited state density.
• Propagated the approximate excited-state (charge) density as an additional dynamical variable within an extended Lagrangian framework.

Main Results:

• The shadow ESMD method offers significant improvements in computational cost compared to direct ESMD.
• The new method demonstrates enhanced stability and robustness over its "exact" counterpart.
• The implementation in SCC-DFTB theory shows broad generalizability to other electronic structure methods.

Conclusions:

• The shadow ESMD method provides a computationally efficient, stable, and robust alternative for excited-state molecular dynamics.
• This approach is adaptable to various quantum chemistry methods, including ab initio techniques.
• The findings pave the way for more accurate and accessible simulations of excited-state phenomena.