Hydrogen-based direct reduction of industrial iron ore pellets: An X-ray computed tomography analysis of porosity and crack evolution
Year: 2026
Type: article
Abstract: Hydrogen-based ironmaking is a promising pathway for steel decarbonization, the performance of which hinges on understanding gas transport mechanisms within dynamic pellet microstructures. Conventional characterization methods conflate pores and cracks into single void fraction, obscuring their distinct evolution and transport impacts. This study employs multiscale characterization combining X-ray computed tomography, scanning electron microscopy, and deep learning to separately quantify pore and crack network evolution. Two industrial pellet grades are analyzed at successive stages obtained from hydrogen reduction experiments conducted at 800 °C and 1000 °C. Results reveal substantial porosity increases through fine pore formation. Reduction at 800 °C promotes pore nucleation producing fine-scale porosity, while 1000 °C drives coarsening through pore coalescence and structural weakening. Crack development varies dramatically with temperature, where 1000 °C generates severe, interconnected networks with >70% percolating connectivity, whereas 800 °C results in much less extensive cracking and lower coarse-void connectivity (≈40%). These findings provide a quantitative microstructural basis for assessing gas-access pathways and cracking susceptibility during H 2 reduction. • Pore and crack networks are resolved independently using deep learning–assisted multiscale XCT/SEM. • Porosity increase is dominated by reduction-induced fine pores (<30 μm) from oxygen removal. • Fine pore coarsening at 1000 °C is driven by pore coalescence and binder-phase decohesion. • At 1000 °C, cracks percolate and raise coarse-void connectivity to >70%. • Binder-constrained shrinkage drives most cracking, with propagation along weak paths.
Language: English
Subfield: Mechanical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $3,730
Award: G4340104273
OpenAlex ID: W7151032437
