Mixed Oxides Reduced Directly to Alloys by Hydrogen Reduction in a Single Step

Direct alloy synthesis from oxides by hydrogen reduction

Our latest direct co-reduction studies define a completly novel route for making alloys directly from mixed oxides, without first producing the pure metals and without melting them. The central idea is to merge extractive metallurgy and alloy synthesis into one solid-state process. Hydrogen removes oxygen from the oxide mixture, while the freshly reduced metallic species interdiffuse and form the target alloy or alloy–oxide composite during the same thermal treatment. The redox product is water, and the alloying step proceeds through solid-state diffusion, phase transformation, and sintering.

This is fundamentally different from conventional alloy production. Classical metallurgy first reduces or refines individual elements, then remelts them, casts them, homogenizes them, and finally thermomechanically processes the alloy. Here, the oxide mixture itself becomes the designed precursor. Its chemical state, phase constitution, particle size, porosity, and spatial arrangement determine the reduction path and the final microstructure.

The work covers three connected model systems: Fe–Cr from Fe_2O3–Cr_2O3, Co–Fe–Mn–Ni from Co_3O4–Fe_2O3–Mn_2O3–NiO, and NiTi from NiO–TiH₂. Together, they establish the thermodynamic and kinetic basis for one-step, hydrogen-based, solid-state alloy synthesis.

Thermodynamics: alloy formation lowers oxide reduction temperatures

A key result is that stable oxides can become reducible at lower temperatures when the reduced metal is stabilized in an alloy phase. The Fe–Cr system demonstrates this most clearly. Pure Cr_2O3 is highly stable and normally requires temperatures above about 1300 °C for hydrogen reduction. In a Fe_2O3–Cr_2O3 mixture, however, Fe_2O3 reduces first to metallic Fe. This Fe then acts as a chemical sink for Cr. Once Cr is reduced at the Fe/Cr_2O3 interface, it dissolves into the Fe matrix. This lowers the chemical activity of Cr and shifts the equilibrium toward further Cr_2O3 reduction.

The decisive reaction is therefore not the formation of pure Cr, but the formation of Fe(Cr). This activity reduction enables partial Cr_2O3 reduction already at 900–1100 °C. For Fe–10Cr oxide mixtures, reduction at 900 °C yielded about 1.3 at.% metallic Cr. At 1100 °C, the Cr content in the metallic phase increased to about 3.7–4.2 at.%, confirmed by thermogravimetry, X-ray diffraction, and atom probe tomography. Extended holding at 1100 °C for about 4 h enabled complete Cr_2O3 reduction for the Fe–10Cr composition.

This finding gives a new thermodynamic rule for sustainable alloy synthesis: the reducibility of a mixed oxide is governed not only by oxide stability, but also by the activity of the reduced element in the alloy product.

We formalize this concept through multicomponent Ellingham diagrams. Classical Ellingham diagrams assume pure metals and pure oxides. That assumption fails for alloy-forming reduction. By calculating equilibrium oxygen partial pressures with CALPHAD, the studies extend Ellingham analysis to multicomponent oxide–alloy equilibria. The resulting diagrams predict how temperature, oxide composition, and alloy chemistry control the feasibility of co-reduction. This is one of the strongest conceptual advances of these works.

Precursor architecture controls kinetics and microstructure

The Co–Fe–Mn–Ni oxide studies show that the initial oxide state is not a processing detail, but a governing design variable. Two precursor states were compared for an equiatomic 25Co–25Fe–25Mn–25Ni target composition.

The first precursor was a mechanically mixed compact of Co_3O4, Fe_2O3, Mn_2O3, and NiO. It had high porosity, about 26%, and retained the individual oxide phases before reduction. Its hydrogen reduction started at about 175 °C. The reaction proceeded through sequential reduction of the individual oxides and transient formation of spinel, halite, and Mn_3O4 phases. After reduction at 700 °C, the product consisted mainly of an Fe–Co–Ni-rich FCC metallic phase and unreduced MnO. The typical phase fraction was about 70 wt.% FCC metal and 30 wt.% MnO.

The second precursor was pre-sintered at 1100 °C in Ar before hydrogen reduction. This created a dense, chemically mixed oxide ceramic with about 55 wt.% Fe,Mn-rich spinel and 45 wt.% Co,Ni-rich halite. Its reduction onset shifted strongly upward to about 525 °C. The higher onset temperature reflects both dense morphology, which limits hydrogen access and water removal, and the higher thermodynamic stability of the spinel and halite solid solutions compared with mechanically mixed individual oxides.

Both precursor types reached a similar final reduction degree of about 78–80% at 700 °C, because MnO remained largely unreduced under these conditions. Yet their pathways and microstructures differed strongly. The mechanically mixed powder produced a loose, heterogeneous metal–oxide microstructure. The pre-sintered precursor inherited its ceramic phase architecture and transformed into a more structured composite: dense Co–Ni-rich metallic regions and nanoporous MnO-rich regions decorated with Fe-rich metallic nanoparticles.

Pulverization experiments separated kinetic and thermodynamic effects. When the dense pre-sintered oxide was crushed and milled, its reduction onset dropped from about 525 °C to 270 °C after 5 h milling and to about 175 °C after 10 h milling. Yet the main reduction event remained delayed by roughly 100 °C relative to the mechanically mixed powder. This proves that surface area and porosity control early kinetics, while the chemical stability of the mixed oxide phases controls the dominant reduction event.

In situ diffraction reveals the reaction pathway

The in situ high-energy X-ray diffraction work is critical because thermogravimetry alone cannot identify transient phases. It only records mass loss. The in situ synchrotron experiments resolve the actual phase sequence during hydrogen reduction at 700 °C.

For mechanically mixed Co–Fe–Mn–Ni oxide powders, the individual oxides first transform into transient spinel, halite, and Mn_3O4 phases between about 250 and 420 °C. These intermediates then reduce to FCC, BCC, and MnO phases. After heating to 700 °C, the phase fractions were about 58 wt.% FCC, 27 wt.% MnO, and 15 wt.% BCC. During 30 min isothermal holding, the BCC fraction decreased and the FCC fraction increased, showing slow diffusion-driven redistribution among metallic phases.

For pre-sintered oxide, no reduction occurred below about 350 °C. The Fe,Mn-rich spinel transformed into Mn,Fe-rich halite, from which Fe partitioned out into metallic phases. The Co,Ni-rich halite reduced into FCC metal. Two FCC variants appeared: a Fe-rich FCC phase with a larger lattice parameter, about 3.613 Å, and a Fe-deficient FCC phase with about 3.574 Å. This split reflects incomplete chemical homogenization and local inheritance from the original oxide architecture.

The reduced pre-sintered sample showed an exsolution-like morphology: metallic nanoparticles of about 10–50 nm on a nanoporous MnO support. This is not only relevant for alloy synthesis. It points to a pathway for producing catalytically active metal/oxide architectures directly by hydrogen reduction of designed oxide precursors.

Redox-mediated NiTi synthesis from NiO and TiH₂

The NiTi study extends the same concept to an intermetallic shape-memory alloy. Instead of using metallic Ni and Ti powders, the feedstock consists of NiO and TiH₂. TiH₂ serves as Ti source, internal hydrogen source, and pore-forming agent. The nominal reaction is:

NiO + TiH₂ → NiTi + H₂O

The process couples TiH₂ decomposition, NiO reduction, Ni–Ti interdiffusion, and intermetallic phase formation. This is scientifically demanding because Ti is highly oxygen reactive, NiTi requires near-equiatomic composition, and shape-memory functionality is highly sensitive to oxygen, hydrogen, and secondary phases.

The gas atmosphere has a decisive effect. In pure Ar, reduction relies mainly on hydrogen released by TiH₂ decomposition. NiO reduction starts at about 328 °C and finishes near 580–590 °C. The total conversion reaches only about 70%. In Ar + 5 vol.% H₂, NiO reduction starts much earlier, around 272–280 °C, and occurs mostly before major TiH₂ decomposition. The final conversion increases to about 92%.

TiH₂ decomposition follows the sequence:

δ-TiH_x → δ + α → δ + α + β → α + β → α

Hydrogen in the atmosphere stabilizes β-Ti(H). In pure Ar, β-Ti(H) disappears around 613 °C. In Ar + 5% H₂, it remains stable up to about 800 °C. Thus hydrogen improves NiO reduction but delays dehydrogenation and changes Ti availability for alloying.

At 800 °C, the main products are not yet NiTi but Ti₂Ni/Ti₄Ni₂O and Ni₃Ti. In pure Ar, these intermetallics reach about 24 wt.% Ti₂Ni and 17 wt.% Ni₃Ti after holding. In Ar + 5% H₂, their formation is suppressed to about 9 wt.% Ti₂Ni and 2 wt.% Ni₃Ti. This difference is linked to reduced metallothermic NiO reduction, lower oxygen incorporation, and stabilization of hydrogen-containing Ti phases.

A validation treatment at 1150 °C substantially increased alloying. The sample center reached about 82 vol.% NiTi/Ni₄Ti₃, with about 33.7 vol.% actual NiTi. The material showed Ti₂Ni/Ti₄Ni₂O and TiO at the surface, indicating Ti loss by oxidation. Below this shell, Ti depletion promoted Ni₃Ti and Ni₄Ti₃ formation. Importantly, retained hydrogen suppressed the martensitic transformation. Vacuum dehydrogenation at 900 °C restored the characteristic transformation, with A_f near 18 °C and M_s near −20 °C. This demonstrates that functional NiTi can be produced by the route, but only if oxygen and hydrogen are controlled tightly.

Scientific novelty of Direct Alloy Synthesis from Oxides by Hydrogen Reduction

The scientific novelty lies in treating oxide reduction as an alloy design problem. The product alloy is not a passive consequence of reduction; it actively changes the thermodynamics of reduction by lowering elemental activities. This is shown quantitatively for Cr_2O3 in Fe–Cr and conceptually extended to multicomponent alloys.

The second innovation is precursor-state engineering. Mechanically mixed oxides, chemically mixed spinel/halite ceramics, pulverized pre-sintered oxides, and oxide–hydride mixtures all follow different reaction paths. Their phase architecture controls reduction onset, intermediate phases, porosity evolution, elemental partitioning, and final microstructure.

The third innovation is the phase-resolved use of in situ synchrotron diffraction. The studies identify transient halite, spinel, Mn_3O4, MnO, BCC, FCC variants, TiH_x, α-Ti(H), β-Ti(H), Ti₂Ni, Ni₃Ti, Ni₄Ti₃, and NiTi formation during reaction. This converts hydrogen reduction from a mass-loss experiment into a mechanistic, time-resolved transformation sequence.

The fourth innovation is functional microstructure generation. The route can produce not only bulk alloys, but also alloy–oxide composites, nanoporous metal/oxide structures, metallic nanoparticles on oxide supports, and intermetallic shape-memory phases. This widens the scope from decarbonized extraction to microstructure synthesis.

Perspective of Direct Alloy Synthesis from Oxides by Hydrogen Reduction

These works establish a scientific foundation for one-step metallurgy: direct conversion of designed oxide precursors into alloys by hydrogen. The approach can reduce process complexity, avoid repeated heating and melting steps, and replace carbon-based reduction by water-forming redox chemistry. Its success will depend on controlling four coupled fields: oxygen chemical potential, hydrogen activity, diffusion length, and evolving porosity.

The remaining challenges are clear. Stable oxides such as MnO and Cr_2O3 require either stronger thermodynamic sinks, higher temperature, longer holding, or optimized precursor geometry. Premature sintering can close gas pathways and arrest reduction. Residual oxygen can stabilize unwanted oxide or oxy-intermetallic phases. Retained hydrogen can suppress functional transformations, as shown for NiTi. Local chemical heterogeneity can produce metastable BCC/FCC partitioning and incomplete alloy homogenization.

The importance of the work is that these limitations are now mechanistically mapped. The studies provide quantitative design rules for future oxide-to-alloy synthesis: choose oxide mixtures by alloy activity, design precursor phases by thermodynamic stability, control porosity for gas transport, and use thermal profiles that synchronize reduction, interdiffusion, and sintering.

The broader message is simple and powerful: alloys need not be made only from purified metals in the liquid state. They can be grown directly from oxides, in the solid state, by using hydrogen as both reductant and process variable. This brings extractive metallurgy, physical metallurgy, and microstructure design into one common processing step.