Skip to content

ReaxFF simulations of petroleum coke sulfur removal mechanisms during pyrolysis and combustion

Summary

Petroleum coke is a carbon-rich refinery stream whose utilization and emissions profile depend strongly on heteroatom content, especially sulfur trapped in complex carbonaceous matrices. This Combustion and Flame article applies ReaxFF reactive molecular dynamics to model S removal pathways during pyrolysis-like and combustion-like high-temperature scenarios, connecting atomistic bond events to macroscopic narratives of S release during thermal processing. The work combines PSU expertise in reactive hydrocarbon chemistry (van Duin) with Mathews-group experience in coal/coke combustion modeling, targeting mechanisms that are difficult to isolate experimentally when matrices are disordered and multi-phase. The framing is explicitly chemistry-forward: which C–S cleavage channels, oxygenated attack routes, and small-molecule sulfur carriers appear under different thermal histories. From an applications perspective, the manuscript is most useful when read as a mechanism generator for S release from disordered carbon hosts: rather than committing to a single global pyrolysis scheme, ReaxFF trajectories can reveal competing channels whose prominence shifts when O\(_2\)-related attack becomes accessible versus purely pyrolytic environments. That distinction matters for emissions interpretations because SO\(_x\) precursors can emerge through different sequences depending on whether oxygen is present at the local scale modeled.

Methods

1 — MD application (petcoke pyrolysis + combustion). The authors build a large atomistic Qingdao petcoke surrogate (C₁₆₄₈H₇₇₂O₅₉N₂₄S₄₇) consistent with HRTEM, XRD, FTIR, and XPS constraints, then run ReaxFF molecular dynamics using the ReaxFF CHONSSi parameter set in the ADF software package under the constant-volume, constant-temperature (NVT) ensemble with 0.3 Å bond-order cutoffs for molecule recognition (analysis only) and 0.001 Å cutoffs for valence/torsion monitors as stated in Combust. Flame. Equilibration: 100 K damping MD with 0.25 fs timestep and Berendsen thermostat to relax geometry without chemistry. Production: pyrolysis trajectories at 3000 K for 250 ps (abstract-reported yields 44.7 wt % gas, 11.0 wt % tar at those conditions); combustion runs add an O₂ environment on the same initial structure (see article for stoichiometry/cell). PBC: three-dimensional PBC for the periodic petcoke cell. Barostat: N/A — NVT protocols without NPT barostat in the excerpted workflow. Electric fields / enhanced sampling: N/A — not used.

2 — Force-field training. N/A — applies published ReaxFF CHONSSi parameters (cited in paper).

3 — Static QM / continuum. N/A — not the primary methods beyond ReaxFF’s underlying QM training heritage.

Findings

Outcomes / mechanisms: Pyrolysis first converts thiophenic sulfur to C₁–₄S (mostly C₂S), COS, and CNS with overlapping heteroatom release at 3000 K in the 250 ps window summarized in the abstract. Combustion shows earlier COS formation from thiophenic sites; N–S motifs do not persist in the O-rich run as pyrrolic/pyridinic N oxidize to CON/NO species. S sequences during combustion are described as COS → CO₂S → CO₃S → CO₄S, while H attacks COS/C₂S/CNS gases forming HS/H₂S rather than directly stripping coke-bound S in the modeled trajectory class.

Comparisons: the atomistic picture complements industrial narratives about S retention in high-S petcoke versus calcined products.

Sensitivity: O₂ availability switches dominant S carriers between pyrolysis-like and combustion-like timelines.

Limitations: a single structural surrogate cannot capture all petcoke heterogeneity; 3000 K/250 ps windows are accelerated relative to plant reactors.

Corpus honesty: yields and temperature/time stamps are taken from the abstract on papers/Zhong_CombFlame_2018.pdf; use the PDF for species traces.

Limitations

Petroleum coke is structurally diverse; any atomistic matrix is a simplified surrogate. High-temperature MD requires scrutiny of temperature ramping and equilibration when comparing to experiment.

Relevance to group

Extends ReaxFF into heavy fuel / coke S chemistry relevant to combustion and environmental impacts.

Citations and evidence anchors