Long timescale dynamics are possible but still challenging.
In brief:
- Consistent Performance: This paper demonstrated that MCTDH, AIMS, and DC-FSSH produce qualitatively similar results for long-timescale nonadiabatic dynamics, validating their applicability beyond ultrafast regimes.
- Efficient Trajectory-Based Methods: It highlighted DC-FSSH and AIMS as computationally efficient approaches, achieving accurate simulations up to 100 ps with modest resources.
- Quantum Precision at a Cost: It showed that MCTDH and ML-MCTDH might be pushed into the long timescale, but further optimization is required.
Our latest work, coordinated by Saikat Mukherjee and published in the Journal of Chemical Theory and Computation, tackles a crucial yet underexplored frontier: the performance of nonadiabatic dynamics methods over long timescales. While these methods have been pivotal in modeling ultrafast photochemical reactions, their reliability for processes spanning tens of picoseconds has remained uncertain—until now.
Nonadiabatic phenomena, such as intersystem crossing, tunneling, and slow internal conversion, occur over extended timescales and are vital for understanding molecular processes in energy, material, and biological sciences. However, these simulations face immense challenges, from computational costs to numerical stability. To address this gap, we evaluated several established methods, including multiconfiguration time-dependent Hartree (MCTDH), ab initio multiple spawning (AIMS), and decoherence-corrected fewest-switches surface hopping (DC-FSSH), using a spin-boson Hamiltonian model designed for long decay behavior.
Our findings are both promising and eye-opening. All tested methods, despite differing theoretical underpinnings, produced qualitatively consistent results for excited-state population decay. This convergence offers confidence in their applicability to long-timescale dynamics, especially considering that none of these methods were developed for this regime.
Notably, trajectory-based approaches like DC-FSSH and AIMS showed remarkable computational efficiency while maintaining good agreement with more computationally intensive quantum methods. For instance, DC-FSSH achieved simulations up to 100 ps with modest computational resources, making it a practical choice for real-world applications.
Unsurprisingly, the quantum methods MCTDH and its multilayer variant (ML-MCTDH) had a steep computational cost. Some simulations required hundreds of days of CPU time, highlighting the trade-off between precision and feasibility. These results underscore the need for further optimization of grid-based quantum approaches to make them more accessible for extended timescale studies.
This work not only evaluates existing methodologies but also paves the way for future advancements in the field. By highlighting both the strengths and limitations of these techniques, we hope to inspire innovations that will enable researchers to tackle long-timescale nonadiabatic processes with greater confidence and efficiency.
For a detailed exploration of our methods, findings, and their implications, we invite you to read the full article in the Journal of Chemical Theory and Computation. This study marks an important step forward in bridging the gap between ultrafast and long-timescale nonadiabatic dynamics, bringing us closer to a comprehensive understanding of molecular photochemistry.
MB
Reference
S. Mukherjee, Y. Lassmann, R. S. Mattos, B. Demoulin, B. F. E. Curchod, M. Barbatti, Assessing Nonadiabatic Dynamics Methods in Long Timescales, J. Chem. Theory Comput. (2024). 10.1021/acs.jctc.4c01349
