Sustainable nickel enabled by hydrogen based reduction

A single step hydrogen plasma route for low grade laterites
U. Manzoor, L. Mujica Roncery, D. Raabe, I. R. Souza Filho · Nature, 2025 · https://doi.org/10.1038/s41586-025-08901-7

Introduction

Nickel is a critical element for the transition to sustainable energy systems. Global demand is projected to exceed 6 million tons annually by 2040, driven largely by the electrification of transport. Primary nickel production currently uses acids and carbon based reductants, emitting about 20 tons of carbon dioxide per ton of nickel produced. This is ten times the emission of steel production per ton. About 60% of annual nickel production relies on high grade sulfide ores, which contain 1.5 to 4 wt% Ni. The remaining 40% comes from low grade laterites, which average 1.5 wt% Ni. However, land based nickel reserves are inversely distributed: 60% of the total nickel available in nature is found in laterites, and only 40% in sulfide ores. The finite and declining reserves of nickel sulfides cannot meet the rapidly increasing global demand, prompting the need for sustainable nickel production from abundant, low grade laterites.

In laterite deposits, nickel is not found as discrete minerals. It is dissolved within complex magnesium silicates, such as lizardite and pyrope, or partially replaces iron in iron oxide lattices, such as goethite. This mineralogical and chemical complexity limits efficient beneficiation. Current industrial processing routes are largely dictated by the crystallographic structure of nickel hosting phases and the nickel and iron content in the ore. Limonite ores, characterized by low nickel and magnesium oxide content, are typically processed via high pressure acid leaching. Silicate rich saprolites, which are difficult to reduce, undergo pyrometallurgical processing. This involves ore drying and partial reduction in rotary kilns using coke as a reductant, followed by smelting in electric arc furnaces, a method known as RK-EF, or alternatively in blast furnaces. These routes are multi step and energy intensive. The energy demand for laterite ore processing ranges from 230 to 570 GJ per ton of nickel. Greenhouse gas emissions are substantial, with the RK-EF and blast furnace routes emitting 45 tons of carbon dioxide equivalent per ton of nickel and 14 tons of carbon dioxide per ton of nickel, respectively.

To mitigate emissions, researchers have investigated replacing carbon based reductants with hydrogen gas in the solid state direct reduction of laterites, particularly limonites. However, the complex crystallographic structure, low nickel content of 0.5 to 2 wt%, and up to 90 wt% impurity contents of oxides such as silicon, calcium and aluminum lead to sluggish kinetics and inefficient hydrogen utilization. The direct reduction of saprolites with gaseous hydrogen is scarcely reported, probably owing to the thermodynamic stability of silicates at typical direct reduction temperatures of 800 to 1000 degrees Celsius.

This study presents a method using fossil free hydrogen plasma based reduction to extract nickel from low grade ore variants known as laterites. The authors bypass the traditional multi step process and combine calcination, smelting, reduction and refining into a single metallurgical step conducted in one furnace. This approach produces high grade ferronickel alloys at fast reduction kinetics. Thermodynamic control of the atmosphere of the furnace enables selective nickel reduction, yielding an alloy with minimal impurities, eliminating the need for further refining. The proposed method has the potential to be up to about 18% more energy efficient while cutting direct carbon dioxide emissions by up to 84% compared with current practice.

Key findings at a glance
  • Single step hydrogen plasma smelting reduction of saprolitic laterite ore (1.26 wt% Ni) yields high grade ferronickel directly from dried ore.
  • Selective nickel reduction is achieved by thermodynamic control of the furnace atmosphere; Ni precipitates before Fe.
  • Impurity levels below 0.04 wt% Si, approximately 0.01 wt% P and below 0.09 wt% Ca eliminate the need for refining.
  • Up to 18% energy savings and up to 84% reduction in direct carbon dioxide emissions compared with the conventional RK-EF route.

Materials and methods

The material used in this research comprises nickel ore, specifically of saprolite type, directly sourced from the mine. The as received ore was ball milled to achieve an approximate size of 100 micrometers. The chemical and phase composition was determined using inductive coupled plasma optical emission spectroscopy and X ray diffraction. The original ore contains only 1.26 wt% Ni, mainly distributed between Ni4O4 at 0.4 wt% and Ni3Si4O10(OH)2·4H2O at 8.02 wt%. The ore also contains lizardite at 50.2 wt%, pyrope at 16.3 wt%, quartz at 14.2 wt%, pimelite at 9.01 wt%, goethite at 8.8 wt%, and bunsenite at 0.2 wt%. The ore fines were compacted into green pellets under a maximum pressure of 20 MPa. Approximately 10 grams of the pellets was placed on a water cooled copper hearth of a conventional arc melting furnace equipped with a tungsten electrode. The furnace chamber was purged with a gas mixture of Ar 10% H2 or Ar 2.5% H2 at a total pressure of 900 mbar. An arc plasma was initiated between the electrode and the input material at 200 A. Simultaneous melting and reduction occurred during a 2 minute exposure to hydrogen plasma. For 4 minute processed samples, the procedure was repeated twice. Phase quantification was performed using X ray diffraction with Rietveld refinement. Microstructural characterization was conducted using high resolution scanning electron microscopy with energy dispersive X ray spectroscopy. Equilibrium calculations were conducted using ThermoCalc software with the TCS Metal Oxide Solutions Database and the SSUSB SGTE Substances Database.

Results and discussion

Phase transformations and reduction mechanism

The as received ore was compacted into green pellets and smelted for 2 minutes in an electric arc furnace under an inert argon atmosphere. The solidified sample showed a weight loss of approximately 23 wt%, attributed to physical losses from thermal decomposition reactions occurring at temperatures above 1600 degrees Celsius and low oxygen partial pressures. X ray diffraction analysis of the solidified argon melted sample reveals solidification primarily into olivine Mg2.5Fe2.5SiO4 at 86.6 wt% and pyroxene MgSiO3 at 9.4 wt% structure types. The absence of nickel rich phases in the solidified sample indicates that nickel is uniformly dissolved in these silicate constituents, which was further confirmed by scanning electron microscopy and energy dispersive X ray spectroscopy elemental mapping.

The reduction of laterite nickel ore was performed under a hydrogen lean hydrogen plasma arc of Ar 10% H2 ignited at 200 A. Even a brief 2 minute exposure resulted in the precipitation of pure metal droplets. Rapid cooling from the water cooled copper hearth solidified these droplets into approximately 5 millimeter nodules within the unreduced silicates or smaller 1 to 3 millimeter particles near the sample bottom. The majority of the metal was concentrated at the bottom center of the solidified sample, suggesting complete mass separation between the silicate and metallic melts, driven by their differing viscosities and mass densities. A thermodynamic assessment of the system's constituent stability under increasing hydrogen exposure reveals that above 1600 degrees Celsius, olivine dissociates upon melting into ionic components such as Fe2+, Ni2+, Mg2+, SiO44-, O2- and SiO2. Oxygen in the melt exists as free ions, bridging oxygens and non bridging oxygens. Owing to their covalent bonds, bridging and non bridging oxygens are more stable than free O2-, making them less reactive with hydrogen species. When the ionic melt is exposed to the hydrogen plasma, chemical redox reactions occur at the arc melt interface. Driven by chemical potential gradients, free oxygen O2- diffuses towards the interface and reacts with hydrogen plasma species H, H+ and electrons, existing as water vapor or hydroxyl radicals depending on the temperature. This reaction leaves behind two electrons, which are preferentially consumed by metal cations with the lowest oxygen affinity, leading to metal precipitation. In the investigated system, Ni2+, which has the lowest oxygen affinity, precipitates first, followed by Fe2+. This thermodynamic behavior confirms the sequential precipitation of nickel and iron, governed by oxygen removal. Once precipitated, nickel and iron form a homogeneous molten phase, solidifying into an iron nickel alloy whose composition depends on the extent of oxygen extraction.

Figure 2: Phase transformations and reduction mechanism.
Figure 2: Unravelling phase transformations. (a) Phase evolution during melting of the original ore in argon and a lean hydrogen atmosphere. (b) Crystal structure of lizardite. (c) Schematic hypothesis of the reduction mechanism: precipitation of metals is initiated by the removal of free oxygen from the melt; hydrogen species extract free oxygen, leaving behind two electrons that reduce Ni2+ and Fe2+. (d) Solidified sample and cross section showing metallic nodules and gas entrapments.

Reduction pathway and thermodynamic control

Thermodynamic investigations revealed the mechanisms governing the sequential precipitation of nickel and iron, demonstrating the feasibility of selectively reducing nickel over iron to obtain a high grade nickel alloy under controlled conditions. Equilibrium calculations were conducted using ThermoCalc software, simulating the exposure of molten laterite ore at 1600 degrees Celsius to varying amounts of an Ar 10% H2 atmosphere. The results show that metal precipitation begins only after oxygen removal reaches about 20%. The associated oxygen loss is due to dehydroxylation and thermal transformations at high temperatures and low partial pressures. Notably, this 20% oxygen removal occurs without hydrogen consumption, as weight loss results solely from thermal processes. To trigger metal precipitation, further oxygen removal is required, necessitating hydrogen exposure. There exists a thermodynamic compositional interval, spanning a reduction degree of 20.7% to 21.3%, with the potential to yield high grade ferronickel, that is with 70 to 90 wt% nickel content, with corresponding total recovery rates ranging from 60 wt% to 90 wt%. To utilize this thermodynamic compositional interval effectively, it is essential to introduce a stoichiometric amount of hydrogen into the system, aligning with an oxygen removal degree in the range of 20.7% to 21.3%.

To validate the thermodynamic assessment, the authors focused on the smallest alloy particles obtained after 2 minutes of hydrogen plasma exposure. Although the plasma arc's direct interaction with the melt pool induces strong stirring that promotes coalescence of metallic droplets, some droplets may escape the turbulent flow and solidify independently before coalescing into larger droplets. These smaller droplets serve as a micro lab, representing instantaneous metal precipitation events. The microstructure of the particle primarily consists of a bright metallic phase and micrometer sized dendrites of unreduced nickel rich oxides. Energy dispersive spectroscopy analysis reveals a high nickel content of 97.3 wt%, with minor iron at 0.44 wt% and oxygen at 2.22 wt%. This composition aligns with the calculated selective thermodynamic compositional range, supporting the feasibility of selective nickel reduction over iron as predicted by thermodynamic theory.

After 2 minutes, 55 wt% of nickel was extracted as metallic nodules, mechanically separated from the slag, and confirmed by scanning electron microscopy and energy dispersive X ray spectroscopy to contain 52 wt% nickel. After 4 minutes, the nickel grade reached 42 wt%, with recovery increasing to 74 wt%. The decline in the nickel grade after 4 minutes indicates that continued reduction leads to iron enrichment and nickel depletion in the alloy. This was confirmed by reprocessing the unreduced silicate from the 2 minute plasma processed sample for another 2 minutes under Ar 10% H2 plasma, yielding an alloy with 9 wt% nickel and 90 wt% iron.

Figure 3: Thermodynamic compositional interval for high grade ferronickel production.
Figure 3: Evidence of optimal thermodynamic compositional interval for high grade ferronickel production. (a) Thermodynamic calculation illustrating the variation in nickel and iron recovery and nickel grade with reduction degree. The green dashed lines and arrow denote the thermodynamic compositional interval for tapping of highly enriched nickel alloy. (b) Solidified sample after 2 minutes processing. (c) Scanning electron microscopy micrograph of a small metallic droplet. (d) Local chemical composition probed via energy dispersive X ray spectroscopy. (e) Elemental mapping of the area highlighted in (c), showing a nickel rich metallic phase.

Process flexibility and atmosphere control

To assess the flexibility of hydrogen plasma smelting reduction in processing different nickel ores, a second type of nickel laterite ore containing slightly more nickel at 1.57 wt% and iron at 16 wt% was also studied. This ore variant was processed under similar conditions to the ones adopted for the first nickel ore. A 2 minute reduction process leads to the formation of high grade 65 wt% nickel metallic nodules, translating to a recovery of 45 wt% nickel. Further processing of the unreduced silicates that had been obtained after 2 minutes for an additional 2 minutes under identical conditions of Ar 10% H2 resulted in the formation of a second portion of metallic nodules with a grade of only 12.4 wt% nickel. By combining the results from these two steps, the overall grade of nickel content of all metallic droplets obtained after a cumulative processing duration of 4 minutes was about 30 wt% nickel, which translates to a global nickel recovery of 65 wt%. This result suggests that the procedure could be applied to a wider range of nickel containing silicates. However, prolonged exposure to hydrogen containing plasma reduces the ferronickel grade. These findings highlight the importance of consistently monitoring and controlling the reducing atmosphere to optimize the nickel concentration in the final alloy.

To elucidate the impact of the hydrogen containing atmosphere control on the nickel content of the recovered metallic nodules, a two step experimental protocol was adopted. This approach involved processing the nickel ore in two steps with different plasma atmospheres. In step 1, the ore was exposed to a 2 minute treatment in Ar 10% H2 followed by separation of the metallic nodules from the unreduced silicates. In step 2, the unreduced silicates obtained in step 1 were subjected to an additional 2 minute processing in Ar 2.5% H2. The resulting alloy obtained after exposure to the Ar 2.5% H2 plasma resulted in a significant enrichment in nickel, with a content of about 45 wt%. By considering the metal portions obtained in both steps 1 and 2, the overall grade of the alloy is about 55 wt% nickel, a value that can be translated into a recovery of about 64 wt% of the total nickel present in the original ore. Comparing the results of this atmosphere controlled experiment to the one processed without any control, we observe that after only an effective 4 minute processing duration, a higher grade product is obtained in controlled atmosphere without compromising the degree of metal recovery. The two step reduction process is meant to illustrate the effect of the reducing atmosphere on nickel grade and recovery. In practice, this process can run continuously, similar to conventional electric arc furnace operations, without the need for stopping the process for gas adjustments. The remaining slag can be further processed in the continuous cycle to recover the remaining nickel. The silicates obtained after 4 minutes of the reduction were further processed in Ar 2.5% H2 for 4 minutes, which allowed to tap a ferronickel of grade 10 wt% nickel, which translates to the overall recovery of 78 wt% nickel. Combining the alloys obtained from each step, we get an alloy containing 30 wt% nickel.

Figure 4: Tracking nickel grade and recovery in two step experiments.
Figure 4: Tracking both nickel grade present in the iron nickel binary product alloy and recovery degree. (a) Experimental protocol for a two step experiment with Ar 10% H2 in both steps. (b) Two step experiment with Ar 10% H2 in step 1 and Ar 2.5% H2 in step 2 to make the process as hydrogen lean as possible and increase the nickel content extracted. (c) Comparison of the weighted averages of the grade and recovery for the alloy obtained in steps 1 and 2 of the experimental protocols described in (a) and (b).

Impurities and product quality

The overall composition of the metallic phase obtained after 4 minutes of processing is 44 wt% nickel, with minimal impurities, including silicon at about 0.08 wt%, phosphorus at about 0.00 wt%, and calcium at about 0.09 wt%. Spot analysis of an inclusion free region reveals a dissolved silicon content of less than 0.04 wt%, significantly lower than typical commercial ferronickel alloys which have silicon up to 0.3 wt%. The proposed method yields an alloy with minimal impurities of less than 0.04 wt% silicon, approximately 0.01 wt% phosphorus and less than 0.09 wt% calcium, eliminating the need for further refining.

Energy efficiency and carbon dioxide emissions

Traditional nickel laterite processing via RK-EF is a multi step, energy intensive route involving ore drying in a rotary dryer, calcination and partial reduction in a rotary kiln, and smelting in an electric arc furnace, relying on carbon based fuels such as bituminous coal and natural gas. Mining and ore preparation contribute only about 4.4% and 6.6% of carbon dioxide emissions, respectively, whereas the primary extraction step, involving emissions from the rotary dryer, rotary kiln and electric arc furnace operations, accounts for a significant 87% of the overall carbon dioxide emissions. In contrast to this lengthy processing route, the one step hydrogen plasma smelting reduction process bypasses the pre and post treatments like calcination and pre reduction of ore in rotary kilns, and refining of the crude ferronickel in ladle furnaces or converters. The entire dried ore charge is processed directly in an electric arc furnace without any preceding refinement, using renewable electricity and a minimal amount of sustainably produced hydrogen as a green reductant. Preliminary calculations estimate an 84% reduction in carbon dioxide emissions with this method, assuming full reliance on renewable energy. The energy required for the pre treatments accounts for about 57% of the total energy required for the entire RK-EF process, with rotary dryers consuming about 15% and rotary kilns about 42%. Among the total energy consumed in these steps only about 27% of the energy is required to thermally run and maintain the process, such as heating of the ore, decomposition reactions of serpentine minerals, removal of free moisture and chemically bonded water, and the chemical redox reactions. From the rest, about 7% of the energy is reutilized as a heating source and the remaining fraction is lost to the environment in the form of flue gas losses, thermal losses to air during the transfer of materials and dust losses. As an example, the calcined ores leave the rotary kiln at an average temperature of 880 degrees Celsius and enter the electric arc furnace at about 400 degrees Celsius. The energy loss during this transfer step alone in today's production practice accounts for about 8% of the total energy consumed along the entire extraction process. By bypassing all these energy and carbon dioxide intensive pre treatments and instead processing the entire dried ore charge in an electric arc furnace in a single smelting reduction step, all these losses can be avoided to make the process energy efficient. On the basis of the calculation, it is estimated that energy savings of up to 18% can be achieved by using hydrogen plasma smelting reduction to produce ferronickel from laterite ores. Ore drying can be efficiently achieved by utilizing the outflow gas mixture of hydrogen and water. Hot steam can directly heat the drying vessel and unutilized hydrogen can be combusted for additional heat. Any remaining hydrogen can be recirculated to the plasma furnace after dust filtration and water condensation, a process that is well established in hydrogen based direct reduction iron reactors. Furthermore, the alloy obtained during the process is highly enriched in nickel at about 30% and it can be utilized directly for stainless steel, invar or magnet alloy making without any energy intensive refining treatments.

Figure 5: Comparative analysis of CO2 emissions and energy efficiency.
Figure 5: Comparative analysis of carbon dioxide emissions and energy efficiency in the traditional RK-EF route and the hydrogen plasma smelting reduction alternative. (a) Different steps involved in the traditional RK-EF route. (b) Hydrogen plasma smelting reduction as a one step processing alternative. (c) Comparison of carbon dioxide emissions from the traditional RK-EF route and hydrogen plasma smelting reduction. (d) Comparison of input energy consumption in different processing steps between the RK-EF and hydrogen plasma smelting reduction routes.

Upscaling considerations

When it comes to the technical perspective for using this at larger scale, one of the crucial aspects for upscaling this technology involves ensuring adequate mass transport of the unreduced melt to the reaction interface. This is significant because most of the chemical reactions during the process occur at the arc melt interface, with free oxygen consistently being removed by hydrogen plasma species. This means that for the reaction to advance, a continuous supply of free oxygen species must be maintained at such an interface. This continuous supply is sustained by strong melt stirring, driven by temperature gradients within the volume and electromagnetic forces from the electric current in direct current mode. In large scale systems, mass transfer during smelting can be enhanced by using short arcs with high currents and increasing the number of electrodes in alternating current electric arc furnaces. Alternatively, external electromagnetic stirring or bottom gas injection can create localized vortices, significantly improving melt stirring. These technologies are already well established in the industry. In addition, maintaining low viscosity is crucial for efficient mass transport. This occurs naturally as complex ore constituents such as lizardite, pyrope, goethite, pimelite and silica transform into a simpler magnesium silicate system. The conversion of polymeric silica into silicate monomers further reduces viscosity, facilitating melt flow.

Conclusion

The study demonstrates hydrogen plasma smelting reduction as a sustainable and energy efficient method for nickel extraction from low grade laterite ores. Thermodynamic control of the reduction atmosphere optimizes nickel reduction, directly forming a high grade iron nickel alloy with low impurities of less than 0.04 wt% silicon, about 0.01 wt% phosphorus and less than 0.09 wt% calcium, eliminating the need for further refining. Powered by renewable energy and using green hydrogen, the process could potentially achieve 18% energy savings and reduce carbon dioxide emissions by up to 84%. Furthermore, a detailed cost analysis highlights the positive financial outcome, reinforcing the possible commercial viability of this sustainable alternative to conventional nickel production. The process produces magnesium silicate slags, which can be used in brick, road and cement production or refined for use in the cosmetic industry. This expands feedstock options to low cost, low grade minerals and shows a sustainable approach to help resolve the contradiction between the beneficial use of nickel in sustainable energy technologies and the environmental harm caused by its production.

Key references

  1. Manzoor, U., Mujica Roncery, L., Raabe, D., Souza Filho, I. R. Sustainable nickel enabled by hydrogen based reduction. Nature (2025). https://doi.org/10.1038/s41586-025-08901-7
  2. Raabe, D. The materials science behind sustainable metals and alloys. Chem. Rev. 123, 2436–2608 (2023).
  3. Norgate, T. & Jahanshahi, S. Assessing the energy and greenhouse gas footprints of nickel laterite processing. Miner. Eng. 24, 698–707 (2011).
  4. Jovicovic-Klug, M., Souza Filho, I. R., Springer, H., Adam, C. & Raabe, D. Green steel from red mud through climate neutral hydrogen plasma reduction. Nature 625, 703–709 (2024).
  5. Souza Filho, I. R., da Silva, A. K., Buyukuslu, O. K., Raabe, D. & Springer, H. Sustainable ironmaking toward a future circular steel economy: exploiting a critical oxygen concentration for metallurgical Cu removal from scrap based melts. Steel Res. Int. 95, 2300785 (2024).

This summary was prepared for a science website serving metallurgy, alloys, microstructure and materials science experts. Figures are reproduced from the original article under a Creative Commons Attribution 4.0 International License.

Druckversion | Sitemap
© D. Raabe