Reactive dynamics simulation study on the pyrolysis of polymer precursors to generate amorphous silicon oxycarbide structures
Summary¶
Silicon oxycarbide (SiOC) glasses combine oxidic and carbidic bonding in amorphous networks attractive for high-temperature materials, often synthesized by pyrolyzing silicon-containing polymers such as polymethylhydrosiloxane (PMHS) and hyperbranched polycarbosilane (HPCS). This Journal of Physical Chemistry C article uses ReaxFF reactive molecular dynamics to pyrolyze mixed PMHS + HPCS models toward a dense amorphous SiOC solid. A practical challenge is mimicking mass loss on simulation timescales; the authors apply a shell script to remove dominant gas molecules (H\(_2\), CH\(_4\)) during NVT segments to approximate experimental outgassing, then compress and equilibrate the residual solid. Radial distribution functions (RDFs) are used to compare Si–O, Si–C, C–C, and Si–Si correlations against experimental constraints cited in the paper.
Methods¶
Software / interactions: Reactive molecular dynamics with ReaxFF (C/Si/H/O) in LAMMPS; Section 2 documents Gaussian checks of selected Si–C/O bond and angle energies against quantum chemistry references.
System and boundaries: Initial models combine PMHS + HPCS polymer backbones in 3D PBC supercells whose atom totals and box sizes follow Section 2 / Table 1 of the JPCC PDF.
Protocol: Pyrolysis stages use NVT MD with thermostat settings, integration timestep (fractional fs), and equilibration/production run lengths in ps reported in Section 2 of the JPCC PDF; the abstract notes NVT-MD segments coupled to a script that deletes dominant H\(_2\) and CH\(_4\) to mimic experimental mass loss. Compression plus room-temperature relaxation yields dense amorphous SiOC; RDF analysis quantifies short- and medium-range order. N/A — NPT barostat and N/A — imposed hydrostatic pressure during the quoted NVT pyrolysis windows.
Findings¶
Outcomes & mechanisms: H\(_2\) and CH\(_4\) dominate gas products under the NVT + scripted removal protocol; decomposition and bond redistribution lead to a dense amorphous SiOC network after compression and relaxation.
Comparisons: radial distribution functions for Si–O, Si–C, C–C, and Si–Si correlations are reported to agree with experimental constraints summarized in the article (operators should verify peak positions against Figures in pdf_path).
Sensitivity: final oxygen balance and carbon partitioning depend on how aggressively CH\(_4\)/H\(_2\) are deleted and on the heating schedule—reproduction requires copying the script logic from Sec. 2 / SI, not only the ReaxFF file.
Limitations / outlook: ReaxFF accuracy and short simulation timescales relative to furnace pyrolysis; periodic cells omit open boundary outgassing except via the heuristic deletion scheme.
Corpus honesty: timestep/thermostat numbers must be taken from the full PDF if absent above; this page is not a substitute for Sec. 2 tables.
Limitations¶
Empirical ReaxFF accuracy; simulation timescales remain far shorter than industrial pyrolysis; gas deletion is a modeling artifact requiring sensitivity analysis.
Reader notes (MAS / retrieval)¶
Use for SiOC pyrolysis questions mentioning PMHS/HPCS mixtures and RDF validation against experiments.
Relevance to group¶
Polymer-derived SiOC ReaxFF workflow comparable to silica/silicate ceramic formation studies elsewhere in the corpus.
Citations and evidence anchors¶
- DOI:
10.1021/acs.jpcc.7b12287.