Skip to content

Reactive Molecular Dynamics Simulations of the Atomic Oxygen Impact on Epoxies with Different Chemistries

Summary

ReaxFF molecular dynamics is used to simulate hypervelocity atomic oxygen (AO) impact on cross-linked thermosetting epoxy networks built with an accelerated ReaxFF cross-linking workflow. The work compares chemistries (including aromatic vs aliphatic curatives), cross-link density, and temperature, and relates network structure to erosion resistance under AO impact relevant to low Earth orbit environments. LEO AO flux is oxidizing and energetic enough to etch organic binders in composites; the paper frames atomistic erosion as a chemistry-selective process where network topology and aromatic content modulate bond scission rates under repeated hypervelocity O impacts.

Methods

Force field and LEO AO representation. The study uses ReaxFF reactive MD to treat hypervelocity atomic oxygen (AO) impacts on cross-linked thermosetting epoxy networks. The introduction frames AO speeds near ~7–8 km s⁻¹, corresponding to roughly ~4.5–5 eV per oxygen atom at reported flux orders of ~10¹⁴–10¹⁵ atoms cm⁻² s⁻¹. The authors emphasize short-range repulsive training in the ReaxFF parameterization used here, arguing that omitting repulsion can overpredict AO damage.

Network construction (Accelerated ReaxFF cross-linking). Epoxy slabs are built using the Accelerated ReaxFF cross-linking workflow (restraint-assisted barrier crossing with geometric cutoffs), intended to better respect reaction barriers compared with naive distance-only bond-formation recipes.

MD application (LAMMPS). Engine / code: simulations are performed with LAMMPS (as stated in the article). Boundaries / periodicity: epoxy slabs use 3D periodic boundary conditions (PBC) in the plane of the slab with free surface along the AO approach direction, supplemented by the reflective wall treatment along z described in the computational details to confine atoms without unphysical loss through boundaries. Pre-bombardment thermalization: an NVT segment at 300 K for 30 ps stabilizes slab temperature after preparation. Thermostat: NVT equilibration uses a LAMMPS thermostat with coupling parameters given in the computational details (PDF). AO bombardment segment: an NVE trajectory follows, with 200 oxygen atoms initialized along −z with ~15 Å spacing and 7.4 km s⁻¹ impact velocity (sign chosen so impacts occur into the slab), yielding ~200 fs between successive impacts at the nominal spacing; the production NVE segment is 40 ps with an integration timestep of 0.00005 ps (0.05 fs) as reported in the computational details. Barostat / pressure control: N/A — the quoted bombardment protocol is NVE with fixed box considerations tied to the AO train setup rather than NPT stress control. Electric fields / enhanced sampling: N/A — not part of the summarized AO impact protocol.

Design of experiments in simulation. The authors compare epoxide/curative chemistries (including aromatic vs cycloaliphatic vs aliphatic curatives), vary percent cross-linking (including 0%, 50%, and 84% cases discussed with damage penetration depth tables), and vary slab temperature (103 K, 300 K, 396 K) for selected systems.

Analysis. Post-impact diagnostics include normalized mass loss, AO penetration depth distributions, small-molecule counts (H₂O, CO, OH, CH₂O), and mechanistic snapshots of bond rupture pathways as reported in the figures.

Findings

Epoxies cured with aromatic components show higher resistance to AO damage, attributed in part to stable aromatic moieties. Lower cross-link density and higher temperature accelerate disintegration. The study argues ReaxFF AO simulations can rank epoxy chemistries for spacecraft exposure scenarios when parametrization includes the relevant repulsive interactions.

Ranking insight: when repulsive O–substrate terms are included, relative stability orders among curatives align more closely with mechanistic expectations (ring stabilization vs aliphatic backbone vulnerability) than when only bonded ReaxFF terms from ambient chemistry training are used.

Sensitivity / levers. Cross-linking strongly affects mass loss and AO penetration depth (the paper reports representative damage penetration depths at t = 40 ps for selected bisA/DETDA cross-linking cases). Slab temperature shifts degradation kinetics: the 103 K case shows the least mass loss and a sharper near-surface AO trapping profile relative to 300 K/396 K in the plotted comparisons.

Comparisons / validation posture. The manuscript ties observed fragmentation trends to bond dissociation energy arguments and compares ReaxFF BDEs to literature values where noted; it positions the work as extending prior AO + polymer reactive MD studies by treating cross-linked networks rather than monomer slabs alone.

Authored limitations (scope). The study isolates AO chemistry and does not attempt to model concurrent UV, electrons, or micrometeorite damage channels noted as additional LEO degradation modes in the introduction.

Limitations

Computational cost limits system sizes and exposure statistics; LEO flux and fluence are represented at the simulation-impact level rather than full mission dose. UV photons, electron bombardment, and contaminant layers on flight hardware can alter erosion rates relative to pure AO impacts modeled here.

Relevance to group

van Duin group; reactive MD for polymer durability under atomic oxygen.

Citations and evidence anchors

https://doi.org/10.1021/acs.jpcc.9b03739