One-step Metallurgy: Hydrogen-Based Direct Co-Reduction Mechanisms of Multicomponent Oxides via In Situ High-Energy X‑ray Diffraction

Seeing Oxide-to-Alloy Synthesis While It Happens

In Situ Synchrotron Tracking of Hydrogen-Based Direct Reduction

Hydrogen-based direct reduction of mixed oxides offers a direct route from oxide feedstocks to alloys, bypassing the conventional sequence of metal extraction, melting, casting and homogenization. In this work, the reaction pathway is resolved in real time for a multicomponent oxide system designed to form an equiatomic CoFeMnNi alloy. The key advance is the use of in situ synchrotron high-energy X-ray diffraction during reduction in hydrogen, allowing the transient oxides, emerging metallic phases and final alloy–oxide microstructures to be followed directly rather than inferred from mass loss or post-mortem analysis.

The study compares two deliberately different precursor states with identical nominal metal chemistry: a mechanically mixed oxide powder containing Co_3O4, Fe_2O3, Mn_2O3 and NiO, and a pre-sintered chemically mixed oxide ceramic consisting of Co,Ni-rich halite and Mn,Fe-rich spinel. Both were reduced at 700 °C in H₂, with a heating rate of 10 °C min⁻¹ and a 30 min isothermal hold. This temperature was chosen because Fe, Co and Ni oxides can be reduced under these conditions, while MnO remains thermodynamically stable and is therefore retained as an oxide phase.

The central finding is clear: the oxide precursor is not a passive starting material. It controls the reduction sequence, the phase partitioning and the final microstructure.

Precursor state controls the reaction pathway

The mechanically mixed powder begins as four separate oxide phases. During heating in hydrogen, these oxides do not reduce simply and independently. Between about 250 and 420 °C, they first react into transient oxide intermediates: Mn_3O4, spinel and halite. These intermediate phases then reduce further, yielding a mixture of metallic and oxide products. At 700 °C, the reduced powder contains approximately 58.1 wt.% FCC phase, 14.5 wt.% BCC phase and 27.4 wt.% MnO. During the 30 min hold, the BCC fraction decreases from about 15.5 to 13.5 wt.%, while the FCC fraction increases from about 55.6 to 58.4 wt.%, showing that metallic interdiffusion and phase redistribution continue after the main oxygen removal step.

The pre-sintered oxide follows a different route. It starts as a dense chemically mixed ceramic with 55.3 ± 2.1 wt.% spinel and 44.7 ± 1.6 wt.% halite. No measurable reduction occurs below about 350 °C, reflecting the higher stability of the chemically mixed oxide phases. Upon further heating, the Mn,Fe-rich spinel transforms into a Mn,Fe-rich halite-type oxide, while Fe, Co and Ni partition out into metallic phases. The Co,Ni-rich halite is stable up to about 600 °C before reducing to metallic FCC phases.

At 700 °C, the pre-sintered material contains approximately 29.6 wt.% MnO, 19.2 wt.% BCC, 8.4 wt.% Fe-rich FCC I and 42.8 wt.% Fe-deficient FCC II. The two FCC variants are distinguished by their lattice parameters: FCC I, a ≈ 3.613 Å, is Fe-rich; FCC II, a ≈ 3.574 Å, is depleted in Fe and enriched in Co and Ni. During the isothermal hold, BCC decreases from 19.2 to 16.5 wt.%, while Fe-rich FCC I increases from 8.4 to 11.8 wt.%. This reveals slow solid-state equilibration of the metallic phases after reduction.

The phase evolution shows that chemical pre-mixing of oxides changes the reduction landscape. It delays reduction, stabilizes halite and spinel phases, and produces chemically partitioned metallic products. The reduction sequence cannot be predicted from the Ellingham stability of the isolated oxides alone. It is governed by the coupled evolution of oxide stability, gas transport, diffusion length, local chemistry and reactive sintering.

Microstructure formation by reduction, partitioning and exsolution

The final microstructures differ as strongly as the phase paths. The reduced mechanically mixed powder retains the character of its initial particulate state. It forms a loose, partially sintered aggregate of metallic Fe–Co–Ni-rich regions and unreduced MnO. The morphology remains coarse and powder-derived.

The pre-sintered precursor produces a much more structured material. Its final microstructure preserves the memory of the initial two-phase ceramic. Regions derived from the Co,Ni-rich halite become relatively dense, mostly metallic domains, with about 94% local density and isolated pores of roughly 200 nm. Regions derived from the Mn,Fe-rich spinel become a porous MnO-based skeleton, with about 20% porosity and finer pores of roughly 50 nm.

Most striking is the formation of 10–50 nm metallic nanoparticles on the nanoporous MnO matrix. This is an exsolution-type process: Fe, initially incorporated in the more stable Mn-containing oxide, is reduced and partitions out as metallic particles while Mn remains as MnO. Co and Ni are not confined to the halite-derived regions; EDS analysis shows substantial interdiffusion into the former spinel regions. In the spinel grain interiors, the metallic phase contains about 36 at.% Fe, 12 at.% Co and 8 at.% Ni. At former spinel grain boundaries, Co and Ni are enriched, reaching about 15.7 at.% Co and 14 at.% Ni. This indicates that reduction, exsolution and interdiffusion proceed concurrently.

The result is not simply an incompletely reduced alloy. It is a hierarchically structured metal–oxide composite generated directly by hydrogen reduction. The nanoporous MnO support decorated with metallic nanoparticles is especially relevant for catalysis and energy conversion, where high surface area, nanoscale metal dispersion and open transport pathways are essential.

Scientific novelty and innovation

The paper provides direct, phase-resolved evidence for how multicomponent oxides transform into alloy–oxide microstructures during hydrogen reduction. Its novelty lies in replacing post-reduction interpretation by real-time crystallographic tracking. The transient formation of Mn_3O4, spinel and halite in the mixed powder, the delayed reduction of chemically mixed halite and spinel in the pre-sintered ceramic, the emergence of BCC and two FCC variants, and the persistence of MnO are all resolved during the process.

The work establishes precursor-state engineering as a central design principle for one-step alloy synthesis from oxides. Mechanically mixed oxides favor early reduction, transient intermediate phases and loose particulate microstructures. Chemically mixed pre-sintered oxides delay reduction but enable inherited phase architectures, local partitioning, exsolution and nanoporous metal–oxide structures. Thus, the precursor defines both the reaction path and the microstructure.

The innovation is broader than decarbonized reduction. It shows that hydrogen-based direct reduction can be used as a microstructure synthesis tool. Oxygen removal, alloy formation, phase separation, exsolution, porosity generation and sintering occur in one thermal step. By adjusting precursor chemistry, oxide phase constitution, particle size, hydrogen activity, temperature and holding time, the process can be directed toward bulk alloys, porous alloys, alloy–oxide composites or nanoparticle-decorated oxide scaffolds.

This work therefore advances hydrogen metallurgy from a reduction technology to a platform for solid-state alloy design. It connects extractive metallurgy, phase transformation science and microstructure engineering in one process: direct conversion of designed multicomponent oxides into functional metallic and metal–oxide materials.