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

Scope

Proof-stage PDF in the corpus. The methods below follow the article text on combining parallel replica dynamics (PRD) with collective-variable hyperdynamics (CVHD) in a ReaxFF setting, with application benchmarks on n-dodecane pyrolysis and a more reactive ethylene carbonate/Li example.

Summary¶

The work reviews and implements PRD + CVHD within reactive ReaxFF molecular dynamics to reach longer effective times for rare chemical events. PRD accelerates escape from a potential basin by running multiple independent replicas and summing their simulated time; CVHD applies a history-dependent bias on collective variables to lower barriers. The authors compare standard MD, CVHD, PRD, and PRD + CVHD for n-dodecane pyrolysis (ReaxFF C/H/O parameters of Chenoweth et al.) and discuss a second case (ethylene carbonate/Li), where CVHD strongly accelerates reactions so that adding PRD yields less extra benefit.

Methods¶

Theory: PRD, CVHD, and combination (A/B)¶

Section 2 defines PRD (replica workflow, dephasing, correlation time \(t_\mathrm{corr}\), event detection) and CVHD (collective variables, global distortion, bias deposition). Section 2.3 derives PRD + CVHD; PRD acceleration scales inversely with transition rate, so when CVHD already raises rates sharply, PRD adds less marginal gain.

n-Dodecane pyrolysis benchmark (B)¶

24 n-dodecane molecules, 50 Å cube, 0.054 kg dm⁻³, 1200 K and 1500 K. CVHD (Bal and Neyts): Gaussian height 0.25 kcal mol⁻¹, half-width 0.025, deposit every 0.2 ps, 1 ps event wait; ReaxFF C/H/O (Chenoweth et al.); bond-order cutoff 0.3 for intact dodecane. PRD: t_corr = 10 ps, t_dephase = 5 ps; bash replica control; primary reaction example C₄H₉ + C₈H₁₇ (Fig. 3).

Ethylene carbonate / Li (B)¶

Second benchmark where CVHD already drives very fast chemistry—PRD benefit small (details in article).

Corpus / citation note (D)¶

Proof PDF—prefer version-of-record DOI for pagination.

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: second proof PDF variant in the corpus; same science as 2019ganeshan-molecular-si-multiply-accelerated.

Findings¶

Comparisons, sensitivity, and limitations¶

Comparisons: stacked PRD + CVHD wall-clock gains are contrasted with CVHD-only and plain MD baselines in the Molecular Simulation tables (see journal PDF for exact factors). Sensitivity: PRD benefit drops when CVHD already raises reaction rates (ethylene carbonate/Li vs n-dodecane), illustrating temperature- and chemistry-dependent accelerator efficiency. Limitations / outlook: hyperdynamics bias alters kinetics; PRD assumes approximately memoryless first-passage statistics—authors discuss when secondary reactions fall outside the PRD event detector. Corpus honesty: this note tracks a proof pdf_path; cite the DOI version-of-record on 2019ganeshan-molecular-si-multiply-accelerated for authoritative figure pagination.

Mechanisms (accelerator interplay)¶

PRD + CVHD can beat CVHD alone in wall-clock for n-dodecane by harvesting rare primary events. When CVHD dominates rates, PRD adds little—consistent with inverse PRD scaling. EC/Li shows the same trade-off at higher intrinsic reactivity.

Limitations¶

Hyperdynamics bias alters kinetics; PRD assumes approximately memoryless transitions. Secondary reactions and PRD event-check targeting: PRD acceleration for follow-on chemistry can drop if secondary events are excluded from the PRD check while CVHD continues—see article. Parallel narrative with journal PDF: 2019ganeshan-molecular-si-multiply-accelerated.

Relevance to group¶

Develops accelerated reactive MD methodology in the ReaxFF ecosystem with van Duin affiliation; connects to combustion/pyrolysis and battery electrolyte-motivated chemistry.

Citations and evidence anchors¶

Full text: papers/Ganeshan_Hossain_MolSym_2019_proof_v2.pdf.