Fig. 3: In situ SXRD assessment of the synthesis mechanisms.

a, Schematic of the experimental set-up and the sample condition. The actual configuration of the experimental instruments is provided in Extended Data Fig. 5. b, Two-dimensional (2D) phase evolution map plotted as a function of time showing the oxide reduction pathway. c, Schematic of the multi-step reduction mechanisms. The colour scale applied in all the oxide phases quantitatively indicates the oxidation state (that is, relative oxygen content). d, Relative phase fraction evolution determined through Rietveld refinement. Note that because of the mass loss during the redox reaction (that is, phase fraction of H₂O is unmeasurable by SXRD), datum points in the left and right panels cannot be used to back-derive absolute phase fraction with respect to the reactants. In addition, because of the substantial reaction boundary condition differences between in situ SXRD and TGA, directly correlating these microscopic phase fraction evolution processes with the global conversion rate measurements may not be possible. e, Lattice constant change of the metallic fcc phase observed in the present experiment. Literature data for the pure Ni (ref. ²⁷) and the standard Fe–Ni invar alloy²⁵ are also included in the left panel as references. Inset: the peak shift of 111 and 200 peaks in the metallic fcc phase during 700 °C isothermal holding.