Deep Learning for Nonadiabatic Excited-State Dynamics

Chen, Wenkai; Li, Xiangyang; Fang, Wei‐Hai; Dral, Pavlo O.; Cui, Ganglong

doi:10.1021/acs.jpclett.8b03026

Cited by 157 publications

(204 citation statements)

References 54 publications

Supporting

Mentioning

203

Contrasting

Unclassified

Order By: Relevance

“…Considerable efforts have been invested more recently in developing a robust algorithm for performing MCTDH on‐the‐fly . Nonadiabatic molecular dynamics combined with machine‐learning strategies for the calculation of electronic‐structure quantities has recently emerged . Quantum electrodynamics effects have been introduced in nonadiabatic dynamics to investigate, for example, molecular dynamics in an optical cavity or the description of stimulated emission processes …”

Section: Discussionmentioning

confidence: 99%

“…160 Nonadiabatic molecular dynamics combined with machine-learning strategies for the calculation of electronic-structure quantities has recently emerged. [161][162][163] Quantum electrodynamics effects have been introduced in nonadiabatic dynamics to investigate, for example, molecular dynamics in an optical cavity 164,165 or the description of stimulated emission processes. 166 The recent advancements in ultrafast spectroscopy combined with the importance of light-triggered phenomena in chemical applications will surely further amplify the interest for the development of techniques for nonadiabatic molecular dynamics.…”

Section: Applications Of Nonadiabatic Molecular Dynamicsmentioning

confidence: 99%

See 1 more Smart Citation

Different flavors of nonadiabatic molecular dynamics

Agostini

Curchod

2019

WIREs Comput Mol Sci

100

View full text Add to dashboard Cite

The Born‐Oppenheimer approximation constitutes a cornerstone of our understanding of molecules and their reactivity, partly because it introduces a somewhat simplified representation of the molecular wavefunction. However, when a molecule absorbs light containing enough energy to trigger an electronic transition, the simplistic nature of the molecular wavefunction offered by the Born‐Oppenheimer approximation breaks down as a result of the now non‐negligible coupling between nuclear and electronic motion, often coined nonadiabatic couplings. Hence, the description of nonadiabatic processes implies a change in our representation of the molecular wavefunction, leading eventually to the design of new theoretical tools to describe the fate of an electronically‐excited molecule. This Overview focuses on this quantity—the total molecular wavefunction—and the different approaches proposed to describe theoretically this complicated object in non‐Born‐Oppenheimer conditions, namely the Born‐Huang and Exact‐Factorization representations. The way each representation depicts the appearance of nonadiabatic effects is then revealed by using a model of a coupled proton–electron transfer reaction. Applying approximations to the formally exact equations of motion obtained within each representation leads to the derivation, or proposition, of different strategies to simulate the nonadiabatic dynamics of molecules. Approaches like quantum dynamics with fixed and time‐dependent grids, traveling basis functions, or mixed quantum/classical like surface hopping, Ehrenfest dynamics, or coupled‐trajectory schemes are described in this Overview. This article is categorized under: Theoretical and Physical Chemistry > Reaction Dynamics and Kinetics Software > Simulation Methods Theoretical and Physical Chemistry > Spectroscopy

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Applications Of Nonadiabatic Molecular Dynamicsmentioning

confidence: 99%

Different flavors of nonadiabatic molecular dynamics

Agostini

Curchod

2019

WIREs Comput Mol Sci

100

View full text Add to dashboard Cite

show abstract

“…Recent developments could achieve a dramatic reduction of this computational cost, for instance by combining nonadiabatic dynamics with quantum chemical calculations accelerated on graphics processing units (GPUs) (see for example Refs. [117][118][119][120]) or by employing machine (or deep-) learning strategies [121,122,123].…”

Section: Electronic Structure For Nonadiabatic Dynamicsmentioning

confidence: 99%

Excited-state dynamics of molecules with classically driven trajectories and Gaussians

2019

View full text Add to dashboard Cite

Simulating the dynamics of a molecule initiated in an excited electronic state constitutes a rather challenging task for theoretical and computational chemistry, as such dynamics leads to a strong coupling between nuclear motion and electronic states, that is, a breakdown of the Born-Oppenheimer approximation. This New Views article proposes a brief overview on recent theoretical developments aiming at simulating the excited-state dynamics of molecules-nonadiabatic molecular dynamics-focusing in particular on strategies employing travelling basis functions to portray the dynamics of nuclear wavefunctions. We start by discussing the central equations for nonadiabatic molecular dynamics in a Born-Huang representation. We then propose a comparison between two commonly employed strategies to simulate the excited-state dynamics of molecular systems in their full configuration space, Ab Initio Multiple Spawning (AIMS) and Trajectory Surface Hopping (TSH). The equations of motion for the two techniques are compared and used to contrast their respective description of phenomena involving the decoherence of nuclear wavepackets. Some recent works and developments of the AIMS method are then summarised. This New Views article ends with a highlight on the Exact Factorisation of the molecular wavefunction and how this approach contrasts with the more conventional Born-Huang picture when it comes to the description of photophysical and photochemical processes.

show abstract

“…In recent years, many efforts have been directed to the efficient improvement of force fields. In particular, machine learning combined with molecular simulation has been verified by many groups to be effective to develop force field including inferring charges based on a set of reference molecules (Botu et al, 2016;Chen et al, 2018;Inokuchi et al, 2018;Engler et al, 2019;Hu et al, 2019;Roman et al, 2019;Sanvito, 2019;Unke and Meuwly, 2019;Ye et al, 2019). Among these, the random forest regression (RFR) method has been proven to be feasible for the prediction of atomic charge without expending much effort on parameter tuning or descriptor selection.…”

Section: Introductionmentioning

confidence: 99%

Toward the Prediction of Multi-Spin State Charges of a Heme Model by Random Forest Regression

Zhao

Huang

et al. 2020

Front. Chem.

View full text Add to dashboard Cite

The random forest regression (RFR) model was introduced to predict the multiple spin state charges of a heme model, which is important for the molecular dynamic simulation of the spin crossover phenomenon. In this work, a multiple spin state structure data set with 39,368 structures of the simplified heme-oxygen binding model was built from the non-adiabatic dynamic simulation trajectories. The ESP charges of each atom were calculated and used as the real-valued response. The conformational adapted charge model (CAC) of three spin states was constructed by an RFR model using symmetry functions. The results show that our RFR model can effectively predict the on the fly atomic charges with the varying conformations as well as the atomic charge of different spin states in the same conformation, thus achieving the balance of accuracy and efficiency. The average mean absolute error of the predicted charges of each spin state is <0.02 e. The comparison studies on descriptors showed a maximum 0.06 e improvement in prediction of the charge of Fe 2+ by using 11 manually selected structural parameters. We hope that this model can not only provide variable parameters for developing the force field of the multi-spin state but also facilitate automation, thus enabling large-scale simulations of atomistic systems.

show abstract

Deep Learning for Nonadiabatic Excited-State Dynamics

Cited by 157 publications

References 54 publications

Different flavors of nonadiabatic molecular dynamics

Different flavors of nonadiabatic molecular dynamics

Excited-state dynamics of molecules with classically driven trajectories and Gaussians

Toward the Prediction of Multi-Spin State Charges of a Heme Model by Random Forest Regression

Contact Info

Product

Resources

About