Skip to content

Ablative thermal protection systems: Pyrolysis modeling by scale-bridging molecular dynamics

Summary

Scale-bridging MD connects atomistic pyrolysis of a highly crosslinked phenolic formaldehyde network to a continuum thermal response model for a charring syntactic foam ablator. Curing of the thermoset uses PCFF in LAMMPS; pyrolysis stages switch to ReaxFF for bond breaking/formation. Arrhenius-like temperature dependence of bulk pyrolysis is reported from 500–2300 K MD; recession inputs feed macroscopic predictions compared to experiments. The motivating context is thermal protection systems where charring, gas blowing, and surface recession must be predicted for mission-relevant heat loads.

Methods

  • Curing (PCFF / LAMMPS): Initial 20×20×26 nm³ box of phenolic rings + CH\(_2\) (thousands of atoms once populated); 0.2 fs timestep; periodic x,y, reflecting z boundaries; NVT-style Berendsen thermostat stages over curing runs whose lengths are tabulated in ps/ns in the Carbon PDF; biaxial compression to ~1.20 g/cc; explicit C–C bond formation when ortho/para C–C distances < 0.3 nm; degree of crosslinking D tracked up to D ≈ 0.93; conjugate-gradient minimization after removing disconnected fragments.
  • Pyrolysis (ReaxFF): ReaxFF replaces PCFF (hydrocarbon oxidation–calibrated parameter set, with DFT/QM benchmarks from the prior parametrization cited in the article); N/A — new QM training set or genetic-algorithm ReaxFF optimization performed in this Carbon paper beyond adopting that file. 12–15 nm vacuum slab added; 300 K equilibration with ~10 atm biaxial stress (anisotropic pressure control); density ~1.25 g/cc; fixed bottom 0.5 nm; 5 nm heated slab with NVT-targeted Berendsen thermostat sweeps at 500–2300 K; uni-directional walls trap fragments; 0.25 fs timestep during pyrolysis segments lasting hundreds of ps as reported in the PDF.
  • Kinetics / continuum: C–C bond-breaking statistics → Arrhenius rates → effective surface recession; fed into thermal material response for AVCOAT-like foam; compared to experimental char thickness / temperature data.
  • Coupling intent: Atomistic rates are not used as standalone chemistry benchmarks alone; they are reduced to effective pyrolysis inputs that a continuum thermal-response solver can consume, closing the loop to engineering observables.

Findings

  • Bulk pyrolysis onset ~500–800 K, consistent with TGA; H\(_2\)O loss then ring fragmentation; Arrhenius behavior for bulk kinetics.
  • Mechanism vs temperature: low-T bulk fragmentation vs ~2300 K onset of surface spallation-like removal of larger clusters.
  • Continuum model predictions (char thickness, temperature fields, blowing) agree with prior experiments for the foam scenario studied.
  • The authors stress that high-temperature surface removal can involve cluster ejection pathways that differ from bulk bond-by-bond statistics, motivating separate characterization of spallation-like regimes in addition to volumetric pyrolysis.

Limitations

ReaxFF parametrization targets gas-phase hydrocarbon oxidation; time scales remain short vs flight; continuum coupling uses extracted effective kinetics.

Relevance to group

Demonstrates ReaxFF pyrolysis linked to engineering TPS response—adjacent to combustion and polymer reactive MD in the corpus.

Citations and evidence anchors

  • DOI: 10.1016/j.carbon.2017.12.099.