Predicting cost-effective carbon fiber precursors: Unraveling the functionalities of oxygen and nitrogen-containing groups during carbonization from ReaxFF simulations
Summary¶
Polyacrylonitrile (PAN) remains the dominant carbon-fiber precursor despite cost pressure for automotive-scale markets. Blending PAN with poly(\(p\)-phenylene-2,6-benzobisoxazole) (PBO) offers a potential route to lower precursor cost while preserving spinnability and carbon-yield characteristics attractive for industrial carbon fiber production. Mao et al. use ReaxFF reactive molecular dynamics to follow volatile release and all-carbon six-membered ring formation during heating and carbonization for nine PAN/PBO blend ratios, comparing blends against neat PAN, neat PBO, and pre-oxidized PAN references. The abstract frames the scientific question around how oxygen-containing versus nitrogen-containing functional groups participate differently in initiating carbonization chemistry, stabilizing radicals, and incorporating carbon into graphitic fragments.
Methods¶
2 — Force-field training (brief, as used). The study uses the published ReaxFF CHON-2019 parameterization for C/H/O/N PAN/PBO chemistry, including a triple-bond stabilization term for N₂ vs the older CHON-2010 set (see §2.1 of Carbon). In-text reoptimization of the entire ReaxFF library is not the subject here—the authors select and apply the CHON-2019 set with documented differences from CHON-2010 (per §2.1).
1 — MD application (reactive carbonization). ReaxFF reactive MD is applied to blend supercells with PAN/PBO mole ratios from 50/0 to 0/50 in the stepped series named in the article, plus pre-oxidized PAN and the pure precursors. System size / composition: each of the 50 parallel chains is built from 16 PAN and 4 PBO monomer units in the construction in §2.2 so that C content is comparable; the resulting cells are multi–10⁴-atom systems (see the table in the Carbon article for exact atom totals per blend). Monomer counts and chain count are as above. Density is set ~1.6 g cm⁻³ (within the 1.2–2.0 g cm⁻³ window cited for PAN CF processing). PBC in 3D; timestep 0.25 fs; bond-order cutoff for species ID 0.3; Berendsen thermostat (damping 100 fs). The protocol equilibrates at 300 K in NVT for 300 ps (regime I), then ramps 300 K → 2800 K over 250 ps at 10 K ps⁻¹ (regime II), and holds 2800 K in NVT for 2000 ps for carbonization (regime III). Barostat / global pressure control during carbonization: N/A — NVT at fixed cell volume and imposed mass density (no NPT mean-stress servocontrol in the quoted protocol). Static external electric field: N/A. Replica / enhanced sampling: N/A. MD engine (software package): N/A in the main-text excerpt used here; reactive ReaxFF MD is the stated tool class—consult the PDF/SI if a code name is required for reproduction.
Findings¶
Mechanisms (O vs N functionality)¶
Oxygen-containing groups act as stronger initiators of carbonization chemistry; nitrogen-containing groups persist longer in nascent graphitic regions and scavenge/couple carbon-centered radicals into extended networks. Oxygen-rich motifs correlate with earlier ring-growth onset; nitrogen retention helps stitch graphenic fragments—blend stoichiometry shifts both kinetics and porosity of the developing network.
Cost/design takeaway (1:1 blend)¶
For 1:1 PAN/PBO, the abstract highlights cost advantages (pre-oxidation avoidance, strong six-membered ring metrics on simulated horizons, ~95% of a stated ring metric—map precisely to the article’s definition).
Limitations and future process coupling¶
Atomistic cells omit melt flow, spinline stress, and furnace gradients; results are mechanistic indicators for precursor design, not factory guarantees.
Limitations¶
Atomistic cells omit melt flow, spinline tension, and furnace gradients present in industrial carbonization; results are mechanistic indicators, not process guarantees for factory-scale furnaces.
Relevance to group¶
Connects ReaxFF carbonization modeling to precursor blending strategies for carbon fibers at Penn State. The work complements broader pyrolysis and polymer-decomposition entries in the corpus by tying atomistic order parameters to economically motivated blend design.