Skip to content

Multiply accelerated ReaxFF molecular dynamics: coupling parallel replica dynamics with collective variable hyper dynamics

Summary

Rare-event reactive simulations of pyrolysis and electrolyte decomposition often require enhanced sampling because bond-breaking events are infrequent on nanosecond scales accessible to brute-force MD. The published Molecular Simulation article therefore stacks collective-variable hyperdynamics with parallel replica dynamics on top of GPU-accelerated ReaxFF in LAMMPS, reporting effective speedups relative to plain MD and CVHD-only runs for hydrocarbon and battery-relevant benchmarks. papers/Ganeshan_Hossain_MolSym_2019_forNSF.pdf is an NSF-reporting PDF export of the Molecular Simulation article (DOI 10.1080/08927022.2019.1646911) on stacking collective-variable hyperdynamics (CVHD) with parallel replica dynamics (PRD) for ReaxFF simulations. The scientific content matches [[2019ganeshan-molecular-si-multiply-accelerated]] (journal PDF) and the author proof variant [[2019ganeshan-venue-eg2-pdf-author-tandf201907230943247415-g]]. Case studies include n-dodecane pyrolysis and ethylene carbonate / lithium chemistry relevant to batteries and thermal decomposition kinetics.

Methods

Canonical protocols (A/B)

PRD + CVHD with ReaxFF (LAMMPS/PuReMD workflow), n-dodecane and EC/Li benchmarks, acceleration factors vs plain MD/CVHD-only—full collective-variable, replica, and timestep tables: 2019ganeshan-molecular-si-multiply-accelerated (journal PDF). This NSF export matches science; pagination may differ.

CVHD biases bond-order dynamics using a collective variable tied to reaction progress so that rare pyrolysis or electrolyte decomposition steps are visited more often than in brute-force MD. PRD runs independent replicas of the biased dynamics to harvest first-passage times for the same rare event, then combines statistics—stacking both methods targets wall-clock gains when either accelerator alone is insufficient.

DFT (C)

Not applicable—method paper for accelerated ReaxFF MD.

Reactive MD integration. Production molecular dynamics uses LAMMPS / PuReMD with ReaxFF bond-order updates on periodic cubic cells (50 Å side for n-dodecane benchmarks) containing 24 molecules (order 10³ atoms total). Temperature: 1200 K and 1500 K thermal setpoints as reported. Timestep: sub-femtosecond timestep values appropriate to ReaxFF stability are quoted in the article’s Computational Methods (see pdf_path). Duration: CVHD bias deposits on 0.2 ps intervals with 1 ps event waits; PRD uses t_corr = 10 ps and t_dephase = 5 ps as stated above. Ensemble: NVT-style thermal control during biased production (details in article). Thermostat: specified with the CVHD implementation in Molecular Simulation (see PDF). Barostat / pressure: N/A — constant-volume supercells without GPa pressure coupling in the summarized benchmarks. External electric field: N/A. Enhanced sampling: parallel replica dynamics plus collective-variable hyperdynamics are the core accelerated dynamics workflow. Corpus note: this NSF PDF export mirrors the journal article; cite 2019ganeshan-molecular-si-multiply-accelerated for canonical pagination.

Findings

Mechanisms, limitations, outlook

PRD on CVHD helps n-dodecane; EC/Li is already fast under CVHD, so PRD adds little—diminishing returns when stacking accelerators. Prefer 2019ganeshan-molecular-si-multiply-accelerated for figure locators.

Takeaway: Wall-clock gains depend on how tight the CVHD bias is—when CVHD already dominates first bond events, extra replicas mainly duplicate work; when CVHD is weak, PRD statistics recover meaningful parallelism across independent trajectories seeded from the same initial state.

Limitations

Users reproducing acceleration factors must reproduce CVHD bias parameters and PRD replica exchange rules exactly as in the journal Methods; otherwise reported speedups are not meaningful. NSF formatting may differ from Taylor & Francis layout. Hyperdynamics biases require careful validation (reweighting, barrier estimates). This slug is provenance for a specific pdf_sha256; cite [[2019ganeshan-molecular-si-multiply-accelerated]] for bibliography unless the NSF file itself must be referenced.

Confidence rationale: med—duplicate collateral PDF; science aligned with canonical journal page.

Citations and evidence anchors

Taylor & Francis hosts Molecular Simulation content under the DOI below; use the publisher landing page for volume, issue, and page numbers when exporting BibTeX. 10.1080/08927022.2019.1646911

Reader notes (navigation)

NSF reports sometimes circulate with different figure compression; if image OCR differs between PDFs, trust equations and tables in the journal PDF for numerical benchmarks.