Energy-Conserving and Thermally Corrected Neglect of Back-Reaction Approximation Method for Nonadiabatic Molecular Dynamics

Akimov, Alexey V.

doi:10.1021/acs.jpclett.3c03029

Cited by 2 publications

(4 citation statements)

References 92 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared to the experimental values, this number is somewhat overestimated; however, bringing in other effects such as back-reaction and many-body nature of excited states would decrease these numbers, bringing them closer to the experimental results. Similar to the C 60 example discussed above, the present time scales should be regarded as the upper limitsgoing beyond the NBRA may lead to a faster dynamics . Indeed, allowing the nuclei to evolve on excited state PES within the non-NBRA approach is likely to lead them to the regions of stronger nonadiabatic couplings or find conical intersections as opposed to the nuclear dynamics governed by the ground state PES.…”

Section: Resultsmentioning

confidence: 82%

“…In the original FSSH prescription of Tully, hop acceptance is decided based on the possibility to rescale nuclear velocities along the derivative coupling vector of a given pair of states. However, within the NBRA approach, this criterion is simplified by using the Boltzmann factor: P i → j normala normalc normalc ( t + Δ t 2 ) = min ( 1 , exp true( prefix− E j − E i K B T true) ) …”

Section: Theory and Methodsmentioning

confidence: 99%

“…The ML-based NA-MD approach presented in this work is formulated for systems whose nuclear dynamics is not strongly perturbed by electronic excitations, the so-called neglect of the back-reaction approximation (NBRA, vide infra). The standard NBRA NA-MD computations typically involve precomputed nuclear trajectories (guiding trajectories) that parametrize the electronic dynamics of a system. While the guiding trajectories may be obtained relatively inexpensively using adiabatic molecular dynamics with force fields, − semiempirical or xTB methods, ,, the propagation of electronic dynamics requires more expensive DFT calculations.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Machine-Learned Kohn–Sham Hamiltonian Mapping for Nonadiabatic Molecular Dynamics

Shakiba,

Akimov

2024

J. Chem. Theory Comput.

Self Cite

View full text Add to dashboard Cite

In this work, we report a simple, efficient, and scalable machinelearning (ML) approach for mapping non-self-consistent Kohn−Sham Hamiltonians constructed with one kind of density functional to the nearly selfconsistent Hamiltonians constructed with another kind of density functional. This approach is designed as a fast surrogate Hamiltonian calculator for use in long nonadiabatic dynamics simulations of large atomistic systems. In this approach, the input and output features are Hamiltonian matrices computed from different levels of theory. We demonstrate that the developed ML-based Hamiltonian mapping method (1) speeds up the calculations by several orders of magnitude, ( 2) is conceptually simpler than alternative ML approaches, (3) is applicable to different systems and sizes and can be used for mapping Hamiltonians constructed with arbitrary density functionals, (4) requires a modest training data, learns fast, and generates molecular orbitals and their energies with the accuracy nearly matching that of conventional calculations, and (5) when applied to nonadiabatic dynamics simulation of excitation energy relaxation in large systems yields the corresponding time scales within the margin of error of the conventional calculations. Using this approach, we explore the excitation energy relaxation in C 60 fullerene and Si 75 H 64 quantum dot structures and derive qualitative and quantitative insights into dynamics in these systems.

show abstract

Section: Resultsmentioning

confidence: 82%

Section: Theory and Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Machine-Learned Kohn–Sham Hamiltonian Mapping for Nonadiabatic Molecular Dynamics

Shakiba,

Akimov

2024

J. Chem. Theory Comput.

Self Cite

View full text Add to dashboard Cite

show abstract

“…To fill in the identified gap, we first implement the XF-based ITMQC methods in the Libra package, , which is already equipped with the range of traditional trajectory surface hopping (TSH) methods for NA dynamics methods as well as with fully quantum dynamics (QD) algorithms and can be used both with model Hamiltonians − and with atomistic systems. − We present a self-contained account on the theoretical grounds of the methods implemented, as well as the corresponding algorithmic details. We pay special attention to seemingly simple but practically nasty topics of state tracking and phase correction.…”

Section: Introductionmentioning

confidence: 99%

Nonadiabatic Dynamics with Exact Factorization: Implementation and Assessment

Han,

Akimov

2024

J. Chem. Theory Comput.

Self Cite

View full text Add to dashboard Cite

In this work, we report our implementation of several independent-trajectory mixed-quantum-classical (ITMQC) nonadiabatic dynamics methods based on exact factorization (XF) in the Libra package for nonadiabatic and excited-state dynamics. Namely, the exact factorization surface hopping (SHXF), mixed quantum-classical dynamics (MQCXF), and mean-field (MFXF) are introduced. Performance of these methods is compared to that of several traditional surface hopping schemes, such as the fewest-switches surface hopping (FSSH), branching-corrected surface hopping (BCSH), and the simplified decay of mixing (SDM), as well as conventional Ehrenfest (mean-field, MF) method. Based on a comprehensive set of 1D model Hamiltonians, we find the ranking SHXF ≈ MQCXF > BCSH > SDM > FSSH ≫ MF, with the BCSH sometimes outperforming the XF methods in terms of describing coherences. Although the MFXF method can yield reasonable populations and coherences for some cases, it does not conserve the total energy and is therefore not recommended. We also find that the branching correction for auxiliary trajectories is important for the XF methods to yield accurate populations and coherences. However, the branching correction can worsen the quality of the energy conservation in the MQCXF. Finally, we find that using the time-dependent Gaussian width approximation used in the XF methods for computing decoherence correction can improve the quality of energy conservation in the MQCXF dynamics. The parameter-free scheme of Subotnik for computing the Gaussian widths is found to deliver the best performance in situations where such widths are not known a priori.

show abstract

Energy-Conserving and Thermally Corrected Neglect of Back-Reaction Approximation Method for Nonadiabatic Molecular Dynamics

Cited by 2 publications

References 92 publications

Machine-Learned Kohn–Sham Hamiltonian Mapping for Nonadiabatic Molecular Dynamics

Machine-Learned Kohn–Sham Hamiltonian Mapping for Nonadiabatic Molecular Dynamics

Nonadiabatic Dynamics with Exact Factorization: Implementation and Assessment

Contact Info

Product

Resources

About