Accelerated ReaxFF simulations for describing the reactive cross-linking of polymers
This J. Phys. Chem. A paper couples barrier-targeted kinetic-energy injection in ReaxFF MD with quantum benchmarks for epoxide cure chemistry, then applies the workflow to bis-F epoxy cured with DETDA and compares simulated density, glass transition, and modulus to experiment.
Summary¶
The paper introduces an accelerated ReaxFF MD protocol for epoxy cross-linking: track reactive pairs until they reach a prereactive geometry, then inject kinetic energy comparable to (or slightly above) the lowest reaction barrier to surmount the cross-linking barrier at experimentally relevant low temperatures, while allowing rejection of events when local strain is too high. ReaxFF is benchmarked against QM for epoxide-like transition states, then applied to bis-F epoxy + DETDA curing, reporting high cross-link yield and density, glass transition temperature, and modulus consistent with experiment.
Methods¶
A — Force-field training / QM benchmarks¶
- Reactive potential: ReaxFF (CHNO organics subset consistent with CHNO-2017_weak-class training cited in related group work) with epoxide–amine chemistry exercised under bond-order reactive dynamics.
- QM validation: quantum reference calculations on three epoxide-related transition-state families discussed in the abstract (geometry/energy agreement between ReaxFF and QM reported in the article).
- Optimization context: bonded and hydrogen-bond components are adjusted as described in Methods/SI (full parameter tables on the published article).
B — Accelerated reactive MD (epoxy cure)¶
- Engine: LAMMPS ReaxFF MD; NVT/NPT stages for network growth, annealing, and property evaluation as specified in the main text and Supporting Information (timestep kept small enough for epoxide chemistry—see article).
- Acceleration protocol: track reactant pairs to a prereactive geometry; inject kinetic energy comparable to (or slightly above) the lowest reaction barrier so cross-linking occurs at low, cure-relevant temperatures; reject events when local strain is incompatible with the attempted transition state (abstract).
- Production chemistry: diglycidyl ether of bisphenol F (bis-F) cured with diethyltoluenediamine (DETDA); post-cure evaluation of mass density, glass transition temperature, and elastic modulus against dogbone experiments (values and windows in the J. Phys. Chem. A article).
C — Literature / method comparison (non-simulation)¶
- The Introduction contrasts this path-resolved acceleration with coarse-graining, distance-only polymerization scripts, and kinetic Monte Carlo schemes that may miss strain-dependent local barriers.
MD protocol (article/SI)¶
System size & composition: Periodic supercells of bis-F + DETDA mixtures spanning thousands of atoms as defined in J. Phys. Chem. A Methods for cross-link growth and property evaluation. Boundaries / periodicity: Three-dimensional periodic boundary conditions (PBC) on bulk curing cells. Thermostat: Nose–Hoover or Berendsen coupling as listed for NVT/NPT stages (damping constants in SI). Barostat / pressure: NPT with hydrostatic pressure control in bar when equilibrating density/T_g; other segments may be strict NVT at constant volume—see published protocol tables.
Findings¶
Outcomes and mechanisms¶
The accelerated dynamics scheme drives cross-linking at low temperature by injecting kinetic energy near the lowest barrier for prereactive pairs, enabling reaction pathways that brute-force MD would rarely sample. ReaxFF reproduces QM barriers for the three epoxide motif families at the level summarized in the article. For bis-F/DETDA, the authors report ~82% cross-link extent and density, glass transition temperature, and modulus consistent with dogbone experiments.
Comparisons and sensitivity¶
Simulated thermomechanical properties are compared directly to experimental dogbone data; temperature and strain around reacting sites modulate rejection of accelerated events when local strain is high (abstract).
Limitations and corpus honesty¶
Proof PDF (pdf_path); confirm all numbers against the version-of-record PDF/SI. Future work in the article discusses extending the workflow to other epoxy/amine chemistries—see Discussion in the PDF.
Limitations¶
Proof PDF; confirm all numerical benchmarks and SI protocols in the published article.
Relevance to group¶
van Duin-lab method paper extending ReaxFF to thermoset cure with an accelerated reaction driver.
Reader notes (navigation)¶
- Duplicate proof variant: 20180000-0002-4740-1296-x-accelerated-reaxff-2.
Citations and evidence anchors¶
- DOI:
10.1021/acs.jpca.8b03826.