Molecular dynamics simulation of layered graphene clusters formation from polyimides under extreme conditions
Summary¶
Laser processing can convert polymer precursors into graphene-like carbon without traditional metal catalysts, but the pressure–temperature pathway that favors sp\(^2\) network growth over charring is difficult to probe in situ. This Carbon article uses ReaxFF molecular dynamics to mimic laser-analog high pressure–temperature spikes applied to polyimide models, contrasting them with milder isobaric protocols. The scientific question is whether layered graphene clusters can emerge from polyimide pyrolysis under extreme NVT-style conditions, and how emitted gas species and carbon-ring statistics evolve compared with lower-pressure decomposition that yields small molecules without extended graphitic layers.
Methods¶
MD application (ReaxFF in LAMMPS): ReaxFF in LAMMPS uses a 0.25 fs timestep and 3D PBC. The cell holds 32 PI monomers (C\(_{22}\)H\(_{12}\)N\(_2\)O\(_5\)) on an FCC lattice, compressed to experimental density ~1.308 g/cm³, then held at 300 K for 40 ps. Branch (i): NVT heating 300 K → \(T_p\) (2400, 2700, or 3000 K); the fixed-volume constraint yields internal pressures ~3 GPa, taken as a laser-like high \(P,T\) spike. Branch (ii): NPT at the same \(T_p\) but 27 MPa, giving small-molecule decomposition without layered graphene. Analysis tracks carbon ring populations, ring–ring PDFs, and gas releases (H\(_2\)O, N\(_2\), H\(_2\), CO, …) to infer pathways. Thermostat/barostat variants, anneal schedule after heat-up, and total production time: N/A — complete protocol tables are in the journal PDF, not the short front-matter extract.
Force-field training: N/A — application study using an existing C/H/O/N ReaxFF parameterization referenced from prior shock, pyrolysis, and polymer work (see paper introduction).
Static QM / DFT: N/A — not the focus; monomer bond lengths/angles are pre-optimized with the VAMP package in Materials Studio 7.0 as stated in the Simulation methods section, but the production study is ReaxFF MD.
Findings¶
Under NVT conditions near ~3 GPa and T > 2400 K, the simulations form layered graphene clusters from polyimide, qualitatively consistent with experimental reports on laser processing of polyimide summarized by the authors. Under NPT near ~27 MPa, pyrolysis yields small molecules without graphene formation, emphasizing strong sensitivity to the mechanical pathway. The paper interprets differences through emitted gas inventories and carbon-network metrics, proposing distinct reaction routes for high-pressure versus low-pressure protocols within the ReaxFF model. The central comparative lever is therefore the mechanical pathway (GPa-scale NVT vs MPa NPT), not temperature alone.
Limitations¶
Mapping simulation protocols to real laser heating is approximate; high-temperature reactive MD can depend on parameterization details and timestep stability.
Reader notes (MAS / retrieval)¶
Highlight GPa NVT vs MPa NPT contrast when users ask why polyimide sometimes yields graphene vs only small molecules.
Relevance to group¶
Illustrates ReaxFF carbonization chemistry tied to laser synthesis of graphene from polymers, adjacent to broader nanocarbon work in the knowledge base.