Microstructure evolution during hydrogen-based direct  reduction via a case study of single crystal hematite

Sustainable hydrogen-based direct reduction (HyDR) of iron oxide is an effective approach to reduce carbon emissions in steel production. As the reduction behaviour is closely related to the microstructure evolution, it is important to understand the microscopic reduction mechanisms. Industrial hematite pellets are microstructurally intricate systems with inherent porosity, defects, and impurities. Therefore, in the present study we investigated the HyDR of single crystal hematite (at 700 ◦ C) to elucidate the reduction behaviour and microstructure evolution in a model system. The reduction kinetics of the single crystal (SC) were compared to those of industrial polycrystalline porous pellets using thermogravimetric analysis. Additional SC samples were prepared such that their faces are parallel to the (0001), (1010) and (1210) crystallographic planes of hematite, and then partially reduced to 16 and 80 % reduction degree. Their microstructure was thoroughly examined by scanning electron 
microscopy and electron backscatter diffraction (EBSD). Reaction fronts were thus shown to advance into the hematite by a shrinking core model while creating a percolating pore network in the magnetite layer; this was closely followed by wüstite and iron formation, as well as pore coarsening, with the retained oxides proceeding to reduce homogenously throughout the sample abiding by the pore/grain models. Notably, a “cell-like” morphology develops in the magnetite near the hematite/magnetite interface, with finely porous “cell interiors” surrounded by coarsely porous “cell walls”. Furthermore, the hierarchal pore formation, phase transformations, texture, and orientation relationships are considered.

Microstructural Evolution and Phase Transformations

The reduction of SC hematite followed a topochemical pathway, with reaction fronts advancing inward via a shrinking core model. At 16% reduction, a ~150 μm thick magnetite layer formed around the hematite core, exhibiting crystallographic anisotropy: reduction progressed ~10% faster perpendicular to the (0001)hem basal plane than along (101‾10)hem or (121‾10)hem facets. This anisotropy reflects differences in ionic diffusion and surface reactivity. Notably, the magnetite layer developed a hierarchical "cell-like" morphology, comprising nanoporous "cell interiors" (pore diameter: ~28–35 nm, spacing: ~70–80 nm) surrounded by coarser "cell walls." These walls, enriched with dislocations and subgrains, served as nucleation sites for subsequent wüstite and iron formation.

At 80% reduction, the core-shell structure vanished, replaced by homogeneous iron-oxide mixtures. Iron nucleation occurred preferentially at magnetite cell walls, growing inward to form a 3D "pipe-like" morphology. Surprisingly, magnetite persisted even at high reduction degrees (~30% of residual oxides), contradicting the classical grain model. EBSD analysis revealed a cube-on-cube orientation relationship between magnetite and wüstite, while hematite/magnetite interfaces predominantly followed (112)mag//(0001)hem and [110]mag//[101‾10]hem orientation relationships, with rare instances of Shoji-Nishiyama OR.

Porosity Development and Crystallographic Dependencies

Porosity evolution was intricately linked to phase transformations. Near the hematite/magnetite interface, porosity peaked at ~25%, driven by vacancy accumulation during Fe₂O₃→Fe₃O₄ reduction. Pore coarsening accompanied iron formation, with maximum porosity coinciding with the onset of Fe₁₋ₓO→Fe transformation (~41% volumetric contraction). Crystallographic facets influenced pore distribution: the bottom (0001)hem facet, exposed to turbulent gas flow, exhibited larger pores (~15% wider) and higher iron content (45%) compared to the top facet (40%). In contrast, (101‾10)hem and (121‾10)hem facets showed lower iron fractions (~30%) and finer pores.

Mechanistic Insights and Implications

The study highlights the critical role of magnetite cell walls as microstructural "highways" for reduction. Their weaker texture and coarser porosity accelerated wüstite and iron nucleation, while the nanoporous cell interiors retained oxides longer. The dominance of (112)mag//(0001)hem OR over Shoji-Nishiyama OR suggests incoherent interfaces at 700 °C, favoring cellular over lamellar magnetite. These findings provide a blueprint for optimizing industrial pellets by engineering pore networks and crystallographic textures to enhance HyDR efficiency.

Advancing Hydrogen-Based Direct Reduction for Fossil-Free Ironmaking: Microstructural Determinants and Kinetic Implications

Fossil-free ironmaking represents a critical pathway for mitigating the steel industry’s substantial contribution to global anthropogenic CO₂ emissions (~7–9% of total emissions). Hydrogen-based direct reduction (HyDR) has emerged as a leading decarbonization technology due to its high technological readiness and potential to replace carbon-intensive blast furnace processes. The HyDR process involves the reduction of iron oxides (e.g., hematite, magnetite) to metallic iron via hydrogen, producing H₂O as the sole byproduct. However, the reduction kinetics and efficiency are governed by a complex interplay of coupled multiphysics phenomena:

Here are some the main chemical mechanisms: Phase transformations (e.g., Fe₂O₃ → Fe₃O₄ → FeO → Fe) mediated by gas-solid reactions, with thermodynamic and kinetic barriers influenced by local H₂ partial pressures and temperature gradients. Redox dynamics at atomic scales, including oxygen vacancy formation and cationic rearrangement. Transport Phenomena:

Multiscale diffusion processes: H₂ permeation through porous oxide layers, counter-diffusion of H₂O, and interfacial mass transfer limitations. Heat transfer constraints arising from endothermic reduction reactions (e.g., FeO + H₂ → Fe + H₂O, ΔH ≈ 25 kJ/mol).

Mechanical-Structural Evolution:

Stress accumulation due to volumetric changes during phase transitions (e.g., ~4% lattice expansion during hematite→magnetite transformation), leading to crack initiation and propagation. Plastic deformation and dislocation glide in metallic iron nuclei, altering local defect densities and diffusion pathways.

Microstructural Complexity and Defect-Mediated Kinetics.

The reduction process generates hierarchical microstructures with defect populations spanning multiple length scales:

Atomic-scale defects: Vacancies and interstitial sites acting as preferential adsorption/nucleation loci.

Mesoscale defects: Dislocation networks, grain boundaries, and subgrain structures that accelerate/delay H₂ diffusion via pipe or boundary pathways.

Macroscale defects: Interconnected pores (5–50 µm) and microfractures (≥100 µm) that modulate gas permeability and reactive surface area.

These defects create heterogeneous reaction environments, where localized chemical potentials and strain fields dominate reduction kinetics. For instance, dislocations enhance reducibility by lowering activation energies for oxide dissociation, while pore closure due to sintering can stifle H₂ access. Critically, the dynamic evolution of these microstructural features—driven by non-isothermal conditions and cyclic loading—introduces feedback loops between mechanical integrity and chemical activity.