Skip to content

Photochemistry and Thermal Chemistry in Polymeric Ceramic Precursors

Corpus note

The ingested PDF is an ACS galley proof (Dasgupta_Mortazavi_JPCL_2025_galley.pdf).

Summary

The study compares photochemical versus thermal pathways at an early stage of silicon carbide formation from an acylsilane precursor (Si\(_{14}\)C\(_{38}\)H\(_{90}\)O\(_2\)) using nonadiabatic quantum molecular dynamics (NAQMD) and ground-state (adiabatic) quantum molecular dynamics (QMD), then extends high-temperature chemistry with ReaxFF reactive MD (RMD) on a 20-molecule assembly to reach longer times and more concerted thermal chemistry.

Methods

  • Electronic structure for QMD/NAQMD: PAW pseudopotentials, GGA exchange–correlation, 30 Ry wave-function cutoff and 300 Ry density cutoff, Γ sampling, implemented in QXMD. The precursor was placed in a 30 Å cubic cell; structures were relaxed with a quasi-Newton optimizer (600 steps) before dynamics.
  • NAQMD (photochemistry): TDDFT-based NAQMD in the NVE ensemble for 72 fs with Δt = 0.24 fs. Photoexcitation was modeled by promoting an electron from HOMO to selected unoccupied orbitals (e.g., LUMO+14, +16, +25) chosen using participation numbers and energy gaps (~4.5–4.8 eV above HOMO), mimicking UV-scale excitation and delocalized acceptor character.
  • Adiabatic QMD (thermal, short time): Same initial configuration simulated in NVT at 300–1800 K for up to 1200 fs to test whether hot ground-state dynamics alone reproduces the targeted Si–C reactivity seen in NAQMD.
  • ReaxFF RMD (thermal, longer time): Molecular dynamics in reactive mode for ~200+ atoms per twenty-molecule supercell slab with 3D PBC periodic boundaries. NVT heating in reactive MD to 1250–2000 K temperature; 0.1–0.25 fs-class timestep in the JPC PDF; 1 ns trajectory spans at the hottest K; Nose-Hoover-like or Langevin thermostat (see article for exact flavor). N/A — barostat in the RMD line quoted; N/A — isotropic 1 bar pressure control; N/A — static external electric field; N/Ametadynamics; N/Aumbrella restraints in the RMD line summarized.

Findings

  • NAQMD shows Si–C bond cleavage within tens of femtoseconds (e.g., bond overlap loss by ~48 fs for the highlighted bond), correlated with hole localization near that linkage.
  • Adiabatic QMD at up to 1800 K over 1.2 ps does not reproduce the selective early Si–C scission seen in NAQMD, supporting a nonthermal photochemical mechanism rather than a purely photothermal explanation tied to rapid heating in the NAQMD run.
  • ReaxFF RMD at high temperature shows accelerating Si–C bond loss with temperature, SiCO-type connectivity arising from carbonyl-oxygen motion toward silicon (e.g., substitution/rearrangement events illustrated in the paper), and increasing CH\(_4\) and H\(_2\) counts at higher temperature—consistent with thermally activated, many-atom pyrolysis chemistry beyond the ab initio time window.

Limitations

NAQMD/TDDFT and ground-state QMD operate at different electronic theory levels than ReaxFF; finite cells and short ab initio times restrict direct comparison. High-temperature ReaxFF explores pyrolysis chemistry beyond QM windows but inherits reactive FF errors. Local PDF may be a galley proof.

Relevance to group

Adri van Duin co-authorship on photo- vs thermal pathways in SiC precursor chemistry with USC collaborators—bridging NAQMD, QMD, and ReaxFF in one workflow narrative.

Citations and evidence anchors