sample at 1 bar. The highlighted vertical band indicates the experimental temperature range applied during reduction, enabling direct comparison between predicted phase stability and experimental observation. This capability was essential for linking bulk thermodynamic predictions to experimentally relevant conditions and for guiding the interpretation of phase transformations during reduction. Figure 3 shows that iron-bearing oxides such as hematite and ilmenite are stable at relatively oxidizing conditions, with combined abundances reaching up to approximately 41 vol.% (Figure 3a). As oxygen fugacity decreases under reducing conditions, these phases progressively destabilize and are replaced by metallic iron, which becomes thermodynamically stable across a broad temperature range. At the experimental temperature (~1000 ºC), metallic iron reaches predicted abundances of approximately 25 vol.% (Figure 3e), confirming the strong thermodynamic driving force for iron reduction under hydrogen atmosphere. Magnetite appears as an intermediate phase during reduction, with stability fields reaching up to approximately 40 vol.% under moderately reducing conditions (Figure 3b). However, its stability narrows significantly at lower oxygen fugacity, indicating its transient role during the transformation from ferric iron-bearing oxides to metallic iron. Spinel-group minerals also exhibit wide stability fields across the investigated temperature range (Figure 3c, d), particularly as hercynite-rich compositions, which act as important sinks for aluminium and residual iron. Titanium is predicted to preferentially partition into perovskite, which remains stable across a wide range of redox conditions, with predicted abundances of approximately 6 vol.% at the experimental temperature (Figure 3f). This confirms the strong thermodynamic preference for titanium to concentrate in stable titanate phases rather than being reduced to metallic form. 4. DISCUSSION Semiquantitative phase analysis of the reduced material revealed metallic iron as the dominant phase (37.49 wt%), followed by perovskite (20.10 wt%) and gehlenite (13.03 wt%). Titanium was primarily concentrated in perovskite and ilmenite, while aluminium was distributed among spinel, corundum, gehlenite, and mullite. The presence of lime and larnite indicates partial calcium redistribution during reduction, confirming extensive phase transformation and elemental partitioning under the applied reducing conditions. Thermodynamic modelling using Perple_X proved to be a powerful tool for guiding the experimental design and interpretation of hydrogen reduction of BR. The calculated fO₂-T phase stability diagram (Figure 2) provided a comprehensive overview of the equilibrium relationships between iron-, titanium-, aluminium-, and calcium-bearing phases under reducing conditions relevant to hydrogen treatment. The modelling clearly demonstrated that at the investigated temperatures (1000 ºC) and low oxygen fugacities consistent with hydrogen-rich atmospheres, iron-bearing oxides such as hematite, magnetite, and ilmenite become thermodynamically unstable and are replaced by metallic iron. Simultaneously, titanium was predicted to preferentially partition into perovskite (CaTiO₃), which exhibits a wide stability field across the investigated redox conditions.
RkJQdWJsaXNoZXIy MTM0Mzk2