Computational Insights into Tunable Reversible Network Materials: Accelerated ReaxFF Kinetics of Furan–Maleimide Diels–Alder Reactions for Self-Healing and Recyclability
Scope
ReaxFF MD with the bond boost acceleration targets retro Diels–Alder kinetics of furan + N-methylmaleimide (endo/exo) as a model for reversible covalent adaptable networks (self-healing / recyclable polymer networks).
Summary¶
Covalent adaptable networks (CANs) use reversible bonds—here the furan–maleimide Diels–Alder (DA) motif—to enable self-healing and recyclability. Prior work mapped thermodynamics and kinetics with DFT on small models, but network environment (chains, packing, stereochemistry) alters endo vs. exo pathways. This paper presents ReaxFF molecular dynamics accelerated by the bond boost method to study the retro-DA reaction for furan and N-methylmaleimide, benchmarking endo/exo product ratios against DFT and experiment while varying bond-boost parameters, and extending to polymer-backbone models and temperature with a newly reparametrized ReaxFF description.
Methods¶
From the extract: ReaxFF reactive MD is used because bond formation and cleavage are central. Bond boost (Miron–Fichtorn-type; as applied in related epoxy–amine work) adds targeted energy along reaction coordinates to access nanosecond windows for chemistry that would otherwise require far longer real time. Simulations focus on retro-DA of endo and exo adducts, with bond breaking tracked in ensembles of product molecules (the abstract cites 40 product molecules). The endo/exo ratio from boosted trajectories is compared to DFT and experiment across bond-boost parameter sets; additional runs probe polymer backbone inclusion and temperature. Full timestep, thermostat, and boost parameter tables are in the article and Supporting Information (not fully contained in the p1–2 extract).
1 — MD application (ReaxFF + bond boost). Engine: ReaxFF molecular dynamics with Miron–Fichtorn-style bond boost to reach reactive events on ns scales (cited in relation to epoxy–amine precedent). System: furan / N-methylmaleimide adducts; ~40 molecules in the stated gas-phase batch and O(10²–10³) atoms-scale supercell-style runs in the SI (N/A for the exact atom count on this note). NVT-class sampling with PBC; N/A on this page for every thermostat constant and .fs timestep (see VOR/SI). Barostat / NPT, external E-field, umbrella/metadynamics, replica exchange: N/A in the high-level abstract summary. Accelerated dynamics — bond boost (not hyperdynamics/PRD unless stated in PDF—confirm there).
2 — Force-field training — reparametrized ReaxFF for the Diels–Alder / furan/maleimide chemistry; DFT and experiment barriers and stereoisomer ratios enter the fit and validation (abstract; SI for loss and data). 3 — Static QM — DFT retro barriers (~3 kcal/mol endo/exo gap in the introduction narrative) used as benchmarks; full level-of-theory lines on p1–2 are incomplete—use pdf_path.
Findings¶
Kinetics, stereochemistry, and CAN relevance. The abstract states ReaxFF+bond boost matches relative retro-DA kinetics for furan / N-methylmaleimide and endo/exo partitioning vs DFT and experiment across boost settings; the reparametrized FF also tracks polymer backbone and temperature effects in network-like setups (as claimed). A ~3 kcal/mol stereochemical split in retrod barriers (introduction) motivates rare-event sampling of the back reaction for reversible covalent adaptable networks (CANs).
Limitations¶
Bond boost is an acceleration approximation; parameters must be validated for each reaction class. The local PDF is a galley; cite the version-of-record for pagination. CAN kinetics in real materials spans longer times than accessible unboosted ReaxFF trajectories.
Relevance to group¶
Adri C. T. van Duin is corresponding author; the study extends ReaxFF into reversible polymer networks and accelerated reactive MD workflows.