Skip to content

Reactive molecular dynamics simulation for isotope-exchange reactions in H/D systems: ReaxFFHD development

Summary

The authors develop ReaxFF_HD, a ReaxFF parameterization aimed at H/D isotope-exchange chemistry in hydrogen plasma environments. The force field is trained against QM data for bond dissociation, angle distortions, and exchange reactions among triatomic hydrogenic ions (H₃⁺, D₃⁺, H₂D⁺, D₂H⁺). Reactive MD trajectories up to ~1 ns explore product formation pathways and isotope partitioning in mixed H/D plasmas. The scientific motivation ties isotope exchange in edge/divertor plasmas to fuel recycling, particle balance, and spectroscopic signatures—where classical reactive models must still capture mass-dependent rearrangements among light ions (introduction themes; abstract).

Methods

2 — Force-field training (ReaxFF\(_\mathrm{HD}\)). A ReaxFF description (ReaxFF\(_\mathrm{HD}\)) is parameterized against a QM training set covering bond dissociation, angle distortion, and an exchange reaction for the triatomic ions H\(_3^+\), D\(_3^+\), H\(_2\)D\(^+\), and D\(_2\)H\(^+\) relevant to H/D plasma media (per abstract; full fit scope and any reoptimization software in the J. Chem. Phys. / SI). QM serves as the reference for energies and structures in that training. Reference data for validation: the article compares the fitted ReaxFF\(_\mathrm{HD}\) to the same class of QM data for the triatomic reaction network.

1 — MD application (production RMD). Reactive MD (ReaxFF\(_\mathrm{HD}\)) is run on various H/D mixtures; the abstract reports analysis of reactions over ~1 ns trajectories. N/A — code (typically LAMMPS-class in this line of work), PBC setup, ensemble (NVT vs NVE), timestep, thermostat, and box composition/number of particles are not in the p1–2 extract here—confirm in the full PDF and SI. Target temperature (K): use the value(s) in the J. Chem. Phys. article and SI; not copied from the short extract to this page. Barostat / NPT / pressure: N/A if simulations are constant-volume plasma boxes as typically used—verify in the article. External electric field: N/A — not the focus in the extract. Enhanced sampling: N/A — not mentioned in the abstract; standard BOMD-style RMD is implied for the 1 ns window.

3 — Static QM / DFT-only (as a standalone report). N/A as the published result thread; DFT/QM data enter as training and validation for ReaxFF, not as the production dynamics method.

Findings

Outcomes and mechanisms. The fitted ReaxFF\(_\mathrm{HD}\) is reported to model isotope exchange of the triatomic ions and to show good transferability to related reactions in the same H/D systems. Over ~1 ns RMD, H\(_2\), D\(_2\), and HD appear as intermediates that undergo further reactions so that triatomic ions are among the most favorable products in the simulated hydrogen plasma conditions (abstract). Deuterium is enriched in some products, tied to lower zero-point energy of D-rich species—i.e. isotope effects carried in the fitted surface (not new quantum dynamics beyond ReaxFF).

Comparisons and sensitivity. The work positions H\(_3^+\)-family chemistry as central in weakly ionized H plasmas where these ions are abundant and mediate fast H/D scrambling; quantitative kinetic and branching numbers should be taken from article figures/tables rather than this note.

Limitations / outlook. Nonadiabatic plasma physics and full kinetics of the discharge are outside a classical ReaxFF RMD model; the validity region is bounded by the training chemistry (stated in the Discussion).

Limitations

Plasma chemistry is vast; the model’s domain of validity is bounded by the training set composition and nonadiabatic phenomena omitted from classical ReaxFF dynamics.

Relevance to group

A clear ReaxFF parameterization paper from the van Duin line for isotopic reactive environments—useful for benchmarking light-element reactive training practices.

Citations and evidence anchors