The role of emerging grain boundary at iron surface, temperature and hydrogen on metal dusting initiation
Summary¶
Metal dusting corrodes Fe-based alloys in carbon-supersaturated, high-temperature gases, often initiating at microstructural defects where carbon activity and surface mobility differ from terraces. Bentria et al. combine DFT with ReaxFF MD on iron models that include emerging grain boundaries and groove geometries under CO-rich environments representative of dusting exposures. The study targets CO dissociation as a gate to carbon uptake and emphasizes how boundary topography steers local reactivity. The Acta Materialia article is explicitly aimed at engineering metal dusting risk where CO and hydrocarbon feeds interact with defective Fe surfaces.
Methods¶
Static QM (DFT). The authors compare carbon adsorption site preference and relative stabilization on flat bcc-Fe terraces versus groove motifs built where an emerging grain boundary meets a free surface (including missing Fe rows to mimic local damage). The DFT block is used to argue that grooves provide preferential C binding relative to neighboring smooth regions.
MD application (ReaxFF). Reactive MD uses a 2016 Fe/C/O/H ReaxFF parameterization from the van Duin lineage (cited as fitted to DFT and experiment for C–O–H chemistry on bcc-Fe). Supercells span Σ3 and Σ5 emerging grain boundaries terminating at (111) and (210) surfaces, again with groove constructs at the GB–surface junction. Gas environments feature CO, with H\(_2\) added in some runs so the authors can separate H as a CO-dissociation promoter versus H reacting with dissociated C (hydrocarbon product channels). Temperature is scanned to count CO dissociation and recombination events and to compare groove vs non-groove segments of the same boundary models.
Force-field training is N/A: the paper adopts the published ReaxFF set rather than refitting.
Protocol numerics (MD code, ensemble, timestep, thermostat/barostat, total simulated time, full atom counts, PBC vectors, pressure control, electric fields, enhanced sampling) are specified in the Acta Mater. Computational methods; this wiki page does not transcribe those tables from the short front-matter extract—consult the PDF for reproducibility.
MD blueprint honesty (not substitute for the PDF). PBC applies to the bcc-Fe slab and grain-boundary supercells. The manuscript names the MD engine (community default for this lineage is LAMMPS with ReaxFF—verify wording in Acta Mater.). Ensemble (NVT vs NPT vs NVE), timestep, thermostat, barostat usage (if any), target pressure for NPT, equilibration/production durations (ps/ns), and electric-field/enhanced-sampling controls are N/A on this page because they are not recoverable from the short indexed excerpt—copy them from the primary PDF.
Findings¶
Grooves stabilize carbon more strongly than neighboring terraces in the DFT data, providing a geometric rationale for localized susceptibility at emerging boundaries. ReaxFF trajectories report faster CO dissociation inside groove regions than on smoother segments of the same boundary models. Hydrogen plays a dual role: it promotes CO scission under some conditions while also scavenging carbon through hydrocarbon product channels, yielding temperature-dependent trends that the authors connect—within their modeling assumptions—to measured metal dusting kinetics. The abstract further argues that, after characterizing mechanisms versus reactant content, the modeled CO dissociation response to temperature can be placed in linear correspondence with experimental metal dusting rates—an explicit attempt to bridge atomistic kinetics with engineering corrosion observables. The Introduction situates metal dusting in carbon-rich gases (including CO and hydrocarbons) at roughly 600–1100 K, where hydrocarbon dissociation, subsurface carbon uptake, and eventual cementite formation drive powdering of Fe alloys; the modeling focus on grooved grain-boundary exits is motivated as a preferential initiation locus within that broader mechanism picture.
Limitations¶
Iron ReaxFF quality is system- and parameterization-dependent; quantitative rates should be validated against the article’s uncertainty discussion. Idealized surface models omit oxide spallation, sulfur poisoning, and long-time microstructure evolution present in industrial dusting exposures. Coupling to continuum CFD carbon activity fields would be needed for plant-scale risk assessment beyond these atomistic samples.
Relevance to group¶
Demonstrates ReaxFF + DFT workflow for high-temperature metal corrosion chemistry on defective Fe surfaces.
Citations and evidence anchors¶
- DOI:
10.1016/j.actamat.2017.06.049.