Skip to content

Dynamics of iodine geminate recombination in supercritical xenon solvent: Caging effect

Summary

The work extends Xe/I ReaxFF training (J. Chem. Phys. 2025, 162, 194309) to I₂ geminate recombination in supercritical xenon, emphasizing caging: local Xe density at recombination exceeds bulk by at least ~2.5×, coordination-number analysis shows clustered solvent, and energy correlation between Xe and recombining I supports kinetic-energy transfer to the solvent. NVT surveys at 290 K and 2–200 bar (through isochoric tuning of the sc state) find recombination rate maximized near the critical region, with low P and high P both reducing rates (diffusion-limited or cage-limited I motion).

Methods

  • Force field: Xe/I ReaxFF from prior work—fit to CASSCF(6,8) I–I, DFT Xe/Xe, Xe/I, colinear / perpendicular Xe/I₂, I–I–I data, and CCSD Xe–Xe, with single-parameter optimization against QM energies.
  • Primary production systems: 10,000 Xe + 250 I atoms (I:Xe = 1:40 default); energy minimization; multi-thermostat NVT at 290 K across pressures (2–200 bar) with Berendsen damping 1000 fs (Xe) vs 1,000,000 fs (I, near-NVE); periodic boundaries; Δt = 0.25 fs; 1 ns trajectories; 10 independent seeds averaged.
  • Concentration checks: additional 1:80 and 1:200 I:Xe ratios (125 and 50 I, respectively) to test event independence (minimum I–I separations to non-recombining atoms > bond cutoff).
  • Geminate protocols: (i) 500-Xe cells with manual I₂ placement inside the largest Xe clusters after 1 ns 290 K / 58.98 bar equilibration; (ii) 12,500-atom cells (10,000 Xe, 125 I₂) with temporary ReaxFF I–I bond dissociation energy set to zero for 100–300 fs bursts to initiate dissociation, then restored for 1 ns multi-thermostat NVT—varying dissociation interval moves outcomes between primary and secondary geminate regimes.

1 — MD application (atomistic dynamics). LAMMPS ReaxFF molecular dynamics in periodic 3D cells (10,000 Xe + 250 I at 1:40, plus 1:80/1:200); NVT at 290 K; 0.25 fs time step; 1 ns production (10 independent replicates); multi-thermostat Berendsen (damping 1000 fs on Xe vs 1,000,000 fs on I for near-NVE I behavior); 2–200 bar thermodynamic pressure for isochoric state points (NVT; N/A — no NPT Parrinello/Berendsen barostat in the main protocol as stated). N/A — no static external electric field; N/A — no replica/metadynamics/umbrella sampling. See bullets above for 1 ns (290 K, 58.98 bar) cluster-equilibration and 100–300 fs I–I bond-off “dissociation” bursts.

2 — Force-field training (prior ReaxFF line, used here in application). Xe/I ReaxFF trained to CASSCF(6,8), DFT, and CCSD reference data; single-parameter optimization to QM energies.

3 — Static QM / DFT-only (outside ReaxFF training). N/A in this work’s application section (QM serves the ReaxFF only).

4 — Review — N/A.

Findings

  • Recombination rate vs pressure peaks near supercritical conditions on the reduced phase diagram; low-pressure runs show unstable transient “recombinations” and non-geminate encounters; high-pressure Xe cages hinder I diffusion, lowering rates.
  • Local Xe density around recombination ≥ ~2.5× the global value; coordination metrics and energy pairing between cluster Xe and I support cage-accelerated geminate kinetics and vibrational energy flow to the solvent.

  • Concentration study: recombination events remain spatially isolated (Fig. 1); mechanism and average Xe coordination (~3.7) at recombination are insensitive to I loading between 1:40 and 1:80 (1:200 undersampled but consistent qualitatively).

  • Corpus honesty / outlook: Excited-state I₂ chemistry is out of scope for ground-state ReaxFF; see ## Limitations and the main text.

Limitations

The PDF is an accepted manuscript notice (AIP); copyedited VOR may differ. ReaxFF follows ground-state chemistry; excited-state I₂ photophysics is explicitly excluded (discussion cites rapid collisional deactivation rationale). Elevated I concentration vs experiment trades sampling for rarity of events.

Relevance to group

A. C. T. van Duin (Penn State) co-authors; builds on the group’s Xe/I ReaxFF line with Stanford collaborators (Majumdar, Ihme).

Citations and evidence anchors