Skip to content

Effects of temperature and mass conservation on the typical chemical sequences of hydrogen oxidation

Summary

PuReMD-GPU integrates gas-phase H\(_2\)/O\(_2\) chemistry with Chenoweth–van Duin / Agrawalla-type ReaxFF parameters, generating stochastic species trajectories in a 20 × 20 × 20 Å periodic cell (66 H\(_2\), 33 O\(_2\), OH seed). NVT trajectories (Nosé–Hoover, 1 ps coupling, 0.1 fs timestep) at 20 temperatures from 2400 K to 6800 K (50 runs per T) feed marginal species statistics; time series are windowed around water formation crossovers for variational typical-set and mass-conserving typical-set analyses of reaction sequences.

Elementary reaction sequences in combustion underpin ignition and extinction reduced models; typical sets summarize dominant pathways without enumerating every trajectory, and the mass-conserving extension probes how forbidden transitions distort joint probabilities when mapping ReaxFF MD to symbolic kinetics. Full trajectory lengths in ps/ns and tables are in the J. Chem. Phys. Methods / SI on pdf_path.

Methods

Reactive MD engine and force field. PuReMD-GPU integrates gas-phase H\(_2\)/O\(_2\) chemistry with ReaxFF parametrizations for hydrocarbon/hydrogen combustion (Chenoweth et al. 2008; Agrawalla & van Duin 2011; Cheng et al. adaptive variants cited in the article).

System, boundaries, and protocol. Simulations use a cubic periodic cell (20 Å per side) containing 66 H\(_2\), 33 O\(_2\), and an OH radical seed (composition as reported), with three-dimensional periodic boundary conditions (PBC). Ensemble: NVT with a Nosé–Hoover thermostat (1 ps coupling time) and a 0.1 fs timestep. Temperature is scanned across 20 set points from 2400 K to 6800 K (thermal grid as reported). Barostat / hydrostatic pressure: N/A — constant-volume NVT gas-phase box without NPT stress control in the protocol summarized here. Electric field: N/A — not used. Enhanced sampling: N/A — umbrella / metadynamics / replica exchange not used; statistics come from brute-force reactive MD trajectories analyzed by information-theoretic sequence constructions.

Sampling plan and analysis window. The study runs 50 independent trajectories at each temperature; N/A — full production segment length in ps/ns should be taken from the J. Chem. Phys. Methods (not restated in the p1–2 extract used for this page). Species time series are windowed around water formation crossovers to build joint probabilities for variational typical sets, including a mass-conserving variant compared against atom-tracked sequences from MD.

Findings

Mass-conserving typical sets capture a larger fraction of probability in the restricted sequences and overlap atom-tracked mechanistic sequences with >90% probability at intermediate sequence lengths. Topological and joint entropy rates extracted from typical-set sizes and probabilities scale approximately linearly with temperature, suggesting extrapolation of sequence statistics across temperatures without exhaustive new simulations (within the paper’s tested H2/O2 model).

Comparisons, sensitivity, and limitations. The mass-conserving construction is compared directly to atom-tracked reaction sequences from MD, while the temperature dependence of entropy rates supports extrapolation claims across the scanned thermal window. Limitations of the symbolic representation—including combinatorial cost for larger mechanisms and the approximate nature of ReaxFF kinetics—are discussed in the article and mirrored under ## Limitations below. Corpus honesty: numerical details beyond the abstract + p1–2 extract should be verified in pdf_path.

Limitations

Information-theoretic construction enumerates sequences combinatorially—expensive for larger mechanisms; reactive FF kinetics are approximate.

Relevance to group

Directly cites ReaxFF combustion parametrizations associated with van Duin-group development and applies PuReMD-GPU trajectories as data for symbolic sequence statistics.

Citations and evidence anchors