Communication: κ-dynamics—An exact method for accelerating rare event classical molecular dynamics

Lu, Chenggang; Makarov, Dmitrii E.; Henkelman, Graeme

doi:10.1063/1.3514030

Cited by 17 publications

(21 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The original form of time reweighing is 8 (12) where N tot is the number of integration steps, and Δt sim is the time step of MD simulation. To achieve good statistics, the simulation should be sufficiently long to sample enough boost regions.…”

Section: Theory and Methodsmentioning

confidence: 99%

Adaptive Accelerated ReaxFF Reactive Dynamics with Validation from Simulating Hydrogen Combustion

et al. 2014

View full text Add to dashboard Cite

We develop here the methodology for dramatically accelerating the ReaxFF reactive force field based reactive molecular dynamics (RMD) simulations through use of the bond boost concept (BB), which we validate here for describing hydrogen combustion. The bond order, undercoordination, and overcoordination concepts of ReaxFF ensure that the BB correctly adapts to the instantaneous configurations in the reactive system to automatically identify the reactions appropriate to receive the bond boost. We refer to this as adaptive Accelerated ReaxFF Reactive Dynamics or aARRDyn. To validate the aARRDyn methodology, we determined the detailed sequence of reactions for hydrogen combustion with and without the BB. We validate that the kinetics and reaction mechanisms (that is the detailed sequences of reactive intermediates and their subsequent transformation to others) for H 2 oxidation obtained from aARRDyn agrees well with the brute force reactive molecular dynamics (BF-RMD) at 2498 K. Using aARRDyn, we then extend our simulations to the whole range of combustion temperatures from ignition (798 K) to flame temperature (2998K), and demonstrate that, over this full temperature range, the reaction rates predicted by aARRDyn agree well with the BF-RMD values, extrapolated to lower temperatures. For the aARRDyn simulation at 798 K we find that the time period for half the H 2 to form H 2 O product is ∼538 s, whereas the computational cost was just 1289 ps, a speed increase of ∼0.42 trillion (10 12 ) over BF-RMD. In carrying out these RMD simulations we found that the ReaxFF-COH2008 version of the ReaxFF force field was not accurate for such intermediates as H 3 O. Consequently we reoptimized the fit to a quantum mechanics (QM) level, leading to the ReaxFF−OH2014 force field that was used in the simulations.

show abstract

Section: Theory and Methodsmentioning

confidence: 99%

Adaptive Accelerated ReaxFF Reactive Dynamics with Validation from Simulating Hydrogen Combustion

et al. 2014

View full text Add to dashboard Cite

show abstract

“…An optimal SVM dividing surface maximizes the transmission coefficient κ, which is calculated as in Lu et al 28 κ values for the SVM method are compared in Table III to those for dividing surfaces that are defined by (i) the adatom displacement from its equilibrium position in the reactants and (ii) the maximum fractional displacement of the adatom to neighboring atoms, as in the bond-boost method. 29 In the case of the frozen Al(100) surface, the escape pathways are well 6. For an adatom on a relaxed (100) surface, the saddle point for the exchange mechanism (a) displaces nearby surface atoms.…”

Section: A Mobile Adatom On a Relaxed (100) Surfacementioning

confidence: 99%

Optimizing transition states via kernel-based machine learning

Pozun

Hansen

Sheppard

et al. 2012

The Journal of Chemical Physics

Self Cite

107

View full text Add to dashboard Cite

We present a method for optimizing transition state theory dividing surfaces with support vector machines. The resulting dividing surfaces require no a priori information or intuition about reaction mechanisms. To generate optimal dividing surfaces, we apply a cycle of machine-learning and refinement of the surface by molecular dynamics sampling. We demonstrate that the machine-learned surfaces contain the relevant low-energy saddle points. The mechanisms of reactions may be extracted from the machine-learned surfaces in order to identify unexpected chemically relevant processes. Furthermore, we show that the machine-learned surfaces significantly increase the transmission coefficient for an adatom exchange involving many coupled degrees of freedom on a (100) surface when compared to a distance-based dividing surface.

show abstract

“…Although locating the minima on a complex PES is very important, knowledge of all the possible reaction pathways (RPs) of a MS, that is, which minima are connected by which TS, provides valuable dynamical information. Several sampling approaches have been developed to find the RPs that connect the local minima of a MS …”

Section: Introductionmentioning

confidence: 99%

An automated method to find transition states using chemical dynamics simulations

2014

View full text Add to dashboard Cite

A procedure to automatically find the transition states (TSs) of a molecular system (MS) is proposed. It has two components: high-energy chemical dynamics simulations (CDS), and an algorithm that analyzes the geometries along the trajectories to find reactive pathways. Two levels of electronic structure calculations are involved: a low level (LL) is used to integrate the trajectories and also to optimize the TSs, and a higher level (HL) is used to reoptimize the structures. The method has been tested in three MSs: formaldehyde, formic acid (FA), and vinyl cyanide (VC), using MOPAC2012 and Gaussian09 to run the LL and HL calculations, respectively. Both the efficacy and efficiency of the method are very good, with around 15 TS structures optimized every 10 trajectories, which gives a total of 7, 12, and 83 TSs for formaldehyde, FA, and VC, respectively. The use of CDS makes it a powerful tool to unveil possible nonstatistical behavior of the system under study.

show abstract

Communication: κ-dynamics—An exact method for accelerating rare event classical molecular dynamics

Cited by 17 publications

References 23 publications

Adaptive Accelerated ReaxFF Reactive Dynamics with Validation from Simulating Hydrogen Combustion

Adaptive Accelerated ReaxFF Reactive Dynamics with Validation from Simulating Hydrogen Combustion

Optimizing transition states via kernel-based machine learning

An automated method to find transition states using chemical dynamics simulations

Contact Info

Product

Resources

About