Unlocking the Nickel Value of Ultramafic Resources Through Innovative Thermal Treatment Process

Abstract

Nickel is a critical metal with broad applications in stainless steel and nickel-based alloys. In recent years, its rapidly growing use in secondary batteries has become particularly important, underscoring its strategic role in enabling low-carbon, green, and sustainable societal development. As high-grade nickel sulfide resources become increasingly scarce and global demand for nickel accelerates—fueled by the expansion of clean energy technologies—there is a pressing need to develop environmentally benign extraction strategies for low-grade nickel sulfide ores to safeguard the stability and sustainability of the global nickel supply chain. Here we report a low-temperature predominantly solid-state process for extracting nickel from unconventional, low-grade ultramafic ores, estimated to contain approximately 45 million tonnes of untapped nickel. The method leverages cheap metallic iron as a nickel getter offering a sustainable pathway for high-value nickel recovery aligned with decarbonized metal production. The process produces ferronickel alloys (16-24% nickel) with several key advantages, including rapid processing time ( ~ 3 hours), low operational temperatures ( < 950 °C), and the elimination of SO₂ emissions. Temperature, atmosphere, and iron addition were tailored to create favorable thermodynamic conditions within the reactor, enabling selective partitioning of nickel into metallic alloys while effectively sequestering sulfur as stable solid sulfide phases. Controlled tuning of alloy particle size and morphology facilitates efficient separation from gangue, yielding a ferronickel which can be converted to battery-grade nickel through conventional refining. This method, verified to mini-plant scale broadens the technological landscape of nickel extraction and contributes to a more equitable and resilient global nickel supply chain.

Wei Lv and colleagues report a mini-plant-scale, predominantly solid-state process for nickel recovery from low-grade ultramafic nickel resources. The process upgrades a 7.7 wt.% Ni sulfide concentrate to a ferronickel concentrate with ~16 wt.% Ni and ~89% Ni recovery while avoiding direct SO2 emissions.

Introduction

Nickel (Ni) is a critical metal for sustaining modern life, particularly as a key base element in various Ni alloys with properties such as corrosion resistance, good formability, high ductility, fracture toughness, and excellent workability¹. These include stainless steel, heat-resistant alloys, Ni-Fe-based alloys, and superalloys, which have significantly enhanced modern technology and industry². As of 2025, approximately 65%, 16%, 7%, 5% and 3% of primary nickel are used to produce stainless steel, batteries, nickel-based alloys, plating, and alloy steels, respectively³. By 2040, nickel demand is expected to reach approximately 2 million tons annually for stainless steel and 210 kilotons for non-ferrous alloys like superalloys, while electric vehicle battery production alone could require an additional 3 million tons driven by the push for transport decarbonization^4–7.

Global nickel resources are estimated to contain more than 350 million tons of nickel. Of these, approximately 54%, 35%, and 10% are found in laterites, sulfides, and hydrothermal systems, respectively⁸. Laterite and sulfide deposits are currently the two main sources of nickel ore. The primary nickel-bearing phases are nickeliferous limonite [(Fe,Ni)O(OH)] and garnierite (a hydrous nickel silicate) in laterites, and pentlandite [(Ni,Fe)₉S₈] in the sulfides⁹. Nickel sulfide deposits generally occur at great depths, while laterite deposits are a near-surface phenomenon. The high-grade nickel sulfide ores are depleting and the mining cost is increasing because of the need to mine deeper. Consequently, increases in Ni production in the last two decades have come primarily from the laterite ores. However, with the rapid advancement of low-carbon economies and clean energy technologies, the demand for nickel, particularly for applications such as batteries in electronic energy storage systems and electric vehicles, has increased significantly¹⁰. Consequently, existing high-grade ore deposits are expected to be insufficient to meet future production demands¹¹. Laterite ores are processed using both pyrometallurgical methods - such as nickel pig iron (NPI) and ferronickel production, which are primarily used in the stainless steel industry - and hydrometallurgical methods like high-pressure acid leaching (HPAL), which is increasingly used to produce battery-grade nickel¹². However, HPAL requires high capital investment, involves complex technology, and poses significant environmental challenges¹³. In contrast, sulfide nickel ores have historically been the preferred source for battery-grade nickel due to their higher grade and more straightforward extraction process.

Nickel sulfide ores are two types: massive nickel sulfides and disseminated ultramafic ores. Ultramafic nickel sulfide ores are unconventional yet abundant resources for nickel production. The nickel content in ultramafic ores ranges from 0.3% to 1%, and the primary nickel-bearing mineral is pentlandite. The other common minerals found in ultramafic ores are pyrrhotite, serpentine, pyroxene, and other magnesium silicates. The global ultramafic nickel resources are estimated to be over 11,200 million metric tonnes¹⁴, with a potential value of 863 billion USD, primarily distributed in the USA, Russia, Australia, Canada, China, and Finland. Based on current consumption rates, this resource alone could sustain global nickel demand for approximately 10–20 years¹⁵; for Europe and America, given their present consumption levels, it would correspond to roughly 100 years of supply¹⁶. For example, the Thompson Nickel Belt in Manitoba, Canada, is estimated to contain over 3 million tonnes of nickel in the form of ultramafic deposits^17,18, with a potential value of approximately 45 billion USD. The high MgO content in the ultramafic ores significantly reduces the efficiency of conventional mineral processing because serpentine particles coat the surface of pentlandite grains and increases the viscosity of flotation pulps^19–21. As a result, the upgraded ultramafic concentrate usually contains 8–10% MgO at low Ni recoveries (60–65%). The high MgO remaining in the ultramafic concentrate poses a significant challenge against the established smelting process because MgO raises the melting point of furnace slag to temperatures as high as 1500 °C and the low grade results in large quantities of slag. The high energy consumption and disposal of the massive industrial slag wastes during processing arise from the low nickel grade and limited recovery of Ni in ultramafic ores, posing major challenges to their effective utilization. With the progressive depletion of high-grade nickel sulfide resources and the rapid growth in global nickel demand driven by clean energy technologies, developing environmentally friendly extraction methods for low-grade nickel sulfide ores has become essential to ensuring the stability and sustainability of the global nickel supply chain. Advancing low-carbon and efficient nickel recovery processes will not only diversify the sources of supply but also enhance the accessibility and fairness of green nickel production worldwide, enabling broader access to the benefits of the clean energy transition.

A low-temperature, predominantly solid-state, two-stage thermal treatment method has been proposed to bypass the high-temperature smelting of ultramafic concentrate while simultaneously mitigating SO₂ emissions^22,23. Here, “predominantly solid-state” indicates that the briquettes remain macroscopically solid during thermal treatment at below its liquidus temperature, although limited local melting or interfacial liquid formation occurs in small regions under certain conditions. Metallic iron is an efficient nickel extractant and sulfur-fixing agent, enabling the direct production of ferronickel alloy (i.e., the Fe-Ni alloy formed during thermal treatment) from sulfide minerals with minimal SO₂ generation. Metallic iron can be produced through the hydrogen-based direct reduced iron (DRI) process, a virtually CO₂-free technology. One innovation is the potential use of CO₂-free DRI to extract nickel from an unconventional yet abundant nickel resource into ferronickel with minimal SO₂ emissions. The clean FeNi product serves as an essential intermediate to produce cathode materials for lithium-ion batteries that enable the clean energy transition.

The process was conceptualized based on favourable thermodynamic conditions, and then tested in bench scale using 10 g batches²⁴. To scale up this technique in a mini-plant scale, significant gaps remain. The temperature distribution within a kilogram-scale reactor has not been previously characterized, and its influence on nickel extraction efficiency under thermal non-uniformity has not been systematically investigated. To address the knowledge gap, the authors designed a kg-scale reactor that operated under an inert atmosphere and treated 100 uniform briquettes at one time, yielding around 1 kg of materials. The preliminary success of this mini-plant study provides insights into commercial utilization of the unconventional yet abundant ultramafic nickel resources. The primary contributions of the study are as follows:

1. Demonstrates the feasibility of scaling up the process from lab-scale (10 g) to kilogram-scale under industrially relevant thermal conditions.

2. Represents a critical transition from fundamental nickel extraction principles to a SO₂-free, CO₂-free, and iron-mediated process applicable to low-grade ultramafic ores.

3. Provides comprehensive system-level validation with consistent, controllable, and high-yield extraction across multiple kilogram-scale trials.

Methods

Materials

The raw materials used in this study included ultramafic nickel concentrate (UNC) supplied by Vale Base Metals, metallic iron powder (Thermo Scientific, >99% metal basis, −200 mesh), and high-purity nitrogen gas (N₂, 99.999%). The mineralogical composition and elemental chemical composition of the ultramafic nickel concentrate were determined using a Mineral Liberation Analyzer (MLA 650, FEI) and Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-OES)^25,26. Since the same batch of raw material was used in the present study, the corresponding mineralogical and elemental characterization results reported in our earlier work²⁷ are briefly summarized in Fig. 1 for completeness. The UNC primarily consists of sulfides (66.2%) and silicates (29.4%). Nickel is predominantly present in sulfides, mainly in the pentlandite (20.9%) and violarite (2.4%) phases. The elemental composition analysis indicates that the material primarily consists of Fe (29.8%), Ni (7.7%), S (23.4%), Si (8.2%), Mg (5.3%), and Ca (0.9%). The particle size distribution of the UNC was analyzed based on cumulative passing percentages, as shown in Fig. 1b. The results indicate that 81.4 wt%, 92.8 wt%, and 99.9 wt% of the particles are smaller than 38 µm, 53 µm, and 150 µm, respectively.

Fig. 1. Raw materials and process overview.

a Mineralogical composition of the ultramafic Ni concentrate determined by MLA. b Particle size distribution of the concentrate. c Bulk elemental composition of the concentrate used in this study. d Flowsheet of the thermal treatment and magnetic separation process.

Experimental

Thermal Treatment Process

The flowchart of the thermal treatment and magnetic separation process is depicted in Fig. 1d. As shown, ultramafic nickel concentrate and metallic iron powder were mixed in a 10:3 weight ratio (totaling 1 kg) using a tumbler mixer for 30 min to achieve a homogeneous blend. The mixture, without addition of any binder, was then compressed into cylindrical briquettes using a four-hole powder pressing mold under a pressure of 110 MPa, producing four briquettes per batch, each weighing approximately 10 grams. Each briquette had a diameter of 16 mm and a height of 18 mm, corresponding to an apparent density of approximately 2.76 g/cm³. A total of 100 briquettes were prepared for the experiments. The iron addition rate was selected based on thermodynamic assessment in Supplementary Fig. 1.

Thermal treatment was conducted in a custom-designed cylindrical reactor. The reactor, made of 316 stainless steel, is sealed with ten large bolts to ensure integrity under high temperatures and rapid heating or cooling. A movable thermocouple mounted on the lid enables temperature monitoring at multiple briquette positions. Inert gas was introduced uniformly from the bottom, establishing an upward flow to maintain a protective atmosphere during the reaction. The briquettes were loaded into the reactor in four layers, labeled Layer 1 (bottom) to Layer 4 (top). After loading, the reactor was sealed and placed in an open furnace. Prior to heating, the reactor was sealed and purged with N₂ at a flow rate of 200 mL/min for 10 min to eliminate residual air. The materials in the reactor were then heated under a continuous N₂ flow (200 mL/min) following a pre-programmed heating schedule. The thermal treatment process involved heating to 915 °C (measured at the midpoint between Layers 2 and 3) at a controlled rate of 3 °C/min, maintaining this temperature for 1 h, and then turning off the furnace to allow the briquettes to cool to ambient temperature under a N₂ atmosphere. It is important to note that due to the large reactor size, the temperature distribution was not uniform, simulating industrial production conditions where significant temperature variations occur. The temperature profiles across the four layers were as follows: Layer 1 (952–935 °C), Layer 2 (935–915 °C), Layer 3 (915–895 °C), and Layer 4 (895–855 °C). These variations reflect the thermal gradients typically encountered in large-scale industrial furnaces.

After cooling to ambient temperature, the thermally treated briquettes were removed from the reactor. Briquettes from different layers and positions (center, middle, and periphery) were selected for further analysis using Scanning Electron Microscopy (SEM) and an Electron Probe Micro-Analyzer (EPMA) to assess ferronickel particle size, circularity, Ni extraction efficiency, and other characteristics of the thermal treatment process. The remaining briquettes from each layer were subjected to magnetic separation to remove impurities and obtain the ferronickel concentrate.

Magnetic Separation Process

The thermally treated briquettes from each layer were separately subjected to magnetic separation. As illustrated in Fig. 1d, the briquettes were first crushed to below 0.8 mm using a jaw crusher (BB50, Retsch) and then wet-ground in an attritor mill (01-HD, Union Process Inc.) at 550 rpm. This grinding process aimed to liberate ferronickel from gangue. The initial grinding lasted 3 minutes, after which the ground materials were sieved through a 38 µm screen. Oversized particles that did not pass through the sieve were collected and subjected to a second round of grinding for 1 minute. After this secondary grinding, the materials were sieved again. The remaining oversized particles, which were too malleable for further grinding, were identified as large ferronickel alloy particles and were collected as the “alloy” fraction. The undersized particles from both grinding stages were combined and processed via three-stage magnetic separation in an agitated cell at a stirring speed of 1200 rpm. Subsequently, three-stages of magnetic separation was performed, applying magnetic field strengths of 250 gauss in the first stage, 1200 gauss in the second, and 2500 gauss in the third. The magnetic fractions obtained from all three stages were combined and designated as the “magnetic concentrate” while the non-magnetic fraction was referred to as the “tails”. The final ferronickel concentrate was obtained by combining the magnetic concentrate with the alloy fraction. The products recovered after magnetic separation were subsequently analyzed using Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES) and SEM.

Product characterization

The particle size of ferronickel (FeNi) was evaluated as the equivalent diameter of a circle with an identical exposed surface area. The size distribution of FeNi particles were examined based on around 20 SEM backscattered electron (BSE) images with over 10, 000 random alloy particles and was reported by D80 on a mass basis (80% passing D80). The shape of ferronickel particles, defined as Degree of Circularity (DoC) was used as a metric to characterize the morphology of ferronickel particles, with DoC value of 1 corresponding to a perfect circle, indicating a simple, smooth shape, while values less than 1 represent elongated or branched particles. A nickel mass balance between the resulting ferronickel phase (i.e., the Fe-Ni metallic phase formed in the reacted briquettes and identified at the microstructural level) and the iron sulfide phase was used to determine the nickel extraction and the nickel grade of the ferronickel phase. Separation efficiency was developed to evaluate the effectiveness of magnetic separation on recovering FeNi particles from gangue matrix. A nickel mass balance between the magnetic FeNi concentrate and non-magnetic tailing yields recovery.

Results

Temperature distribution in the reactor

Figure 2a presents the horizontal temperature distribution in each of the four layers within the reaction zone of the vessel. The temperature variations are small in the radial direction. As seen, the bottom layer has the highest temperature at 938–949 °C. From the bottom to the top, temperature decreases to 810–876 °C. Figure 2b shows the detailed temperature profile of briquettes in the reaction zone. At each vertical level (e.g., Layer 1/2/3/4), three concentric rings of briquettes were arranged radially from the center of the reactor outward. The maximum temperature difference across the 8 cm height is approximately 100 °C while the temperature in the radius orientation remains similar.

Fig. 2. Linking reactor conditions, phase evolution and nickel recovery during thermal treatment.

a Simulated 3D temperature distribution in the kg-scale reactor. b Spatial mapping of thermal gradients across briquette layers. c Representative microstructures of multiphase products in layer 1, light green: ferronickel, light purple: sulfide, light yellow: silicate. d Correlation between reactor temperature, Ni extraction efficiency and ferronickel grade (The dashed lines represent fitted trends of the experimental data. The shaded bands are visual aids added to emphasize the fitted trends and do not represent uncertainty, confidence intervals, standard deviation or standard error). e Layer-resolved 3D distribution of Ni grade in ferronickel (layers 1-4). f Layer-resolved 3D distribution of Ni extraction in ferronickel (layers 1-4). g Predicted equilibrium phase fractions as a function of temperature. h Thermodynamic solubility limits of Ni in competing phases. i Equilibrium partitioning of Ni among phases with increasing temperature.

Nickel extraction and grade in FeNi

Figure 2c shows representative microstructures of thermal-treatment products from layer 1, and Supplementary Fig. 2 presents additional microstructures from other layers. The products contain ferronickel, iron sulfide, and magnesium iron silicates. Nickel in the as-received ultramafic concentrate was enriched in the ferronickel phase. Figure 2d shows the variation in nickel extraction and Ni grade in FeNi as a function of temperatures in the reactor. The dashed lines represent fitted trends of the experimental data, while the shaded bands are included only as visual aids to emphasize these trends and do not represent uncertainty, confidence intervals, standard deviation or standard error. The nickel extraction demonstrates a declining trend as temperature increases. Over approximately 100 °C temperature difference, the difference in extraction is 6%. The nickel grade varies in the range of 30–35% and does not follow an obvious trend over the same temperature range. Figure 2e, f shows the nickel grade and extraction to FeNi generated at different layers. The maximum nickel grade was around 35%, where the corresponding Ni extraction was 94.8%. From bottom to the top (i.e., from Layer 1 to Layer 4), nickel extraction increases from 90% to 95.3%, while nickel grade fluctuates around 33%.

FeNi particle size and morphology

Figure 3a presents a comprehensive comparison of the DoC as a function of FeNi particle diameter across four distinct layers and their respective sample sets. As shown in the scatter plots, a clear inverse correlation is evident-smaller particles tend to exhibit larger circularity. The accompanying density plots demonstrate good consistency within each layer. Furthermore, the aggregated kernel density estimation curves exhibit unimodal, right-skewed distributions for all layers, indicating an overall tendency toward more circular particle morphologies. Notably, the density peaks for all layers converge near DoC ≈ 0.88, as indicated by the dashed orange line (Supplementary Fig. 3). Compared to the lower layers, particles in Layer 3 and Layer 4 exhibit both higher average circularity and smaller diameters, suggesting that upper deposition layers promote the formation of more isotropic and finer FeNi particles.

Fig. 3. Size and morphology of ferronickel particles.

a Distribution of degree of circularity (DoC). b 3D mapping of weighted DoC (WDoC) across briquette layers (mid-cross-sections). c 3D mapping of particle size (D80) across briquette layers (mid-cross-sections).

Figure 3b, c illustrate the variation of WDoC and its corresponding ferronickel particle size (D80) across different layers. The maximum D80, approximately 170 μm, is observed in Layer 1, where the associated WDoC is as low as 0.4. From the bottom to the top (i.e., from Layer 1 to Layer 4), D80 progressively decreases to ~17 μm, while WDoC correspondingly increases to 0.6. This inverse trend further suggests that upper layers favor the formation of smaller and more isotropic ferronickel particles. While small FeNi particles are round in nature, larger particles are primarily formed by clustering of smaller particles, resulting in uneven shapes with a low WDoC.

Ni recovery to FeNi concentrate

Supplementary Fig. 4 presents the micrographs of the ferronickel concentrate and tailings. After grinding, the liberated ferronickel particles exhibit noodle-like morphologies for smaller particles and oval shapes for larger ones. The main phases identified in the ferronickel concentrate include FeNi, sulfides, and silicates, whereas only sulfides and silicates are observed in the tailings. This indicates that FeNi phases in the thermally treated briquettes are effectively recovered into the ferronickel concentrate. Additionally, some sulfide and silicate phases are found to be encapsulated by FeNi, making them difficult to liberate. This encapsulation likely results from the agglomeration of ferronickel particles during the thermal treatment process, which causes non-metallic phases to be enclosed within the metallic matrix.

It is interesting to note that large FeNi particles are readily liberated from gangue matrix post grinding and recovered into an “alloy stream” without magnetic separation. Figure 4a shows the elemental mapping of the large FeNi particle in the “alloy stream”. As seen, the FeNi particle is dominated by Ni and Fe with Mg, Si, S, and O elements segregating from the Ni-rich area. The typical Ni grade in the large FeNi particle is around 37.9%. Figure 4b illustrates the relationship between the cumulative Ni recovery and cumulative Ni grade during the physical separation process for different layers. For each layer, the “alloy stream” (i.e., the oversize fraction prior to magnetic separation) has the highest Ni grade varying from 17% to 24% (layer 1 to layer 4). The Ni recoveries to oversize alloy for Layer 1 and Layer 2 are 43.9% and 60.7%, respectively, which are higher than those of Layer 3 and Layer 4 (17.4% and 3.3%). However, after magnetic separation, the final Ni recoveries for Layer 3 and Layer 4 reach 91.1% and 90.9%, surpassing those of Layer 1 and Layer 2, which are 85.4% and 88.1%, respectively. The Ni grade in the ferronickel concentrates across all layers ranges from 15.0% to 17.5%. Overall, the average Ni recovery achieved in the mini-plant test is 88.7%, at an average Ni grade of 15.8% in the final concentrate.

Fig. 4. Physical separation performance of ferronickel.

a Elemental mapping of a large ferronickel particle in the alloy stream. b Cumulative Ni recovery versus Ni grade during physical separation. c Impurity rejection and overall separation efficiency of thermally treated briquettes. d Effect of ferronickel particle size (D80) on separation (The dashed lines represent fitted trends of the experimental data. The shaded bands are visual aids added to emphasize the fitted trends and do not represent uncertainty, confidence intervals, standard deviation or standard error). e Effect of particle circularity (DoC) on impurity content in oversize ferronickel (The dashed lines represent fitted trends of the experimental data. The shaded bands are visual aids added to emphasize the fitted trends and do not represent uncertainty, confidence intervals, standard deviation or standard error). Oversize FeNi particles ( > 38 μm) from layer 4 were excluded owing to limited measurements (n = 2). f Elemental composition of ferronickel concentrate after magnetic separation. g Ni-to-impurity ratio in the ferronickel concentrate.

Separation efficiency of magnetic separation technique

Figure 4c illustrates the rejection rates of typical impurities (S, Si, and Mg) and the FeNi separation efficiency during magnetic separation of thermally treated briquettes from different layers. Overall, sulfur rejection decreases gradually from 85.4% in Layer 1 to 78.0% in Layer 4. A similar trend is observed for silicon and magnesium in Layers 1 and 2, where their rejection rates decline from 81.5% to 77.6% (Si), and from 79.9% to 75.3% (Mg), respectively. However, in Layers 3 and 4, Si and Mg rejection rates increase again, reaching 84.8% and 82.1%, respectively, in Layer 4. This behavior is likely related to the distribution and association of these elements with silicate phases, as well as particle encapsulation during thermal agglomeration. In contrast, the FeNi separation efficiency exhibits a consistent upward trend, increasing from 63.1% in Layer 1 to 70.2% in Layer 4.

Discussion

Principle of thermal upgrading and magnetic separation

Figure 5a shows the principle of thermal extraction of nickel from ultramafic concentrate by iron addition. The primary nickel-bearing phase in the concentrate is pentlandite with pyrrhotite (Ni bearing sulfide phase) containing the minor amounts of nickel. The formation of ferronickel relies on nickel diffusion from the sulfide phase to the metallic iron phase. At temperatures below 900 °C under an inert atmosphere, the metallic iron undergoes sulfidation, with a thin layer of iron sulfide forming on the edge of iron particles and acting as a matrix through which nickel can migrate towards the metallic iron. As a result, nickel is gradually extracted from sulfides to ferronickel, evidenced by the higher nickel concentration on the edge of the particles than their core. At temperature above 900 °C, liquid matte begins to form which facilitates the diffusion of nickel into metallic iron, thus growing large ferronickel particles. During the whole process, silicate minerals demonstrate a relatively inert behavior. This mechanism is summarized from the results out of kg-scale studies in this work and agrees with the bench-scale studies in the previous studies by the authors²⁸. Figure 5b illustrates the magnetic separation mechanism employed in this study. Magnetic separation of FeNi particles from the gangue is feasible thanks to the significant difference in specific magnetic susceptibility of the two materials: 10.5–50 × 10^-5m³/kg for FeNi compared to <1 × 10^-5m³/kg for the silicate and sulfide (FeS) phases⁹. As a result, after the ferronickel is liberated from the thermally treated briquettes through grinding, its separation is relatively straightforward.

Fig. 5. Mechanistic pathway of thermal treatment and magnetic separation.

a Phase transformations induced by thermal treatment. b Conceptual schematic of ferronickel-gangue separation under magnetic fields.

Effect of temperature on Nickel extraction

Figure 2g shows the equilibrium phases at temperatures of 500–1000 °C. As seen at 950–1000 °C, liquid matte and solid FeNi are the nickel-containing phases. As temperature decreases to below 900 °C, solid FeNi and solid iron sulfide are gradually precipitated from the liquid matte, thus, the relatively low temperature is beneficial for the nickel extraction/high nickel recovery degree. In the thermal treatment process, solid silica and solid olivine are stable across the whole temperature range. Figure 2h shows the nickel solubility in the Ni-bearing phases at 500-1000 °C. From 1000 °C to 900 °C, the nickel solubility in the liquid matte ranges from approximately 8% to 12%. Below 900 °C, the liquid is converted to solid iron sulfide phase while Ni precipitates out at FeNi. The nickel solubility in the solid iron sulfide phase reaches the maximum of around 1.2% at 880 °C and then decreases to around 0.5% as temperature decreases to 500 °C. Across the whole temperature range, the nickel solubility in FeNi increases with decreasing temperature. Figure 2i shows the nickel distribution in different phases. As seen at 950–1000 °C, most of the nickel is present in the liquid phase with FeNi being the secondary nickel-bearing phase. At temperatures lower than 900 °C, FeNi is the dominant nickel-bearing phase with solid iron sulfide being the secondary nickel-containing phase. This explains why a lower nickel extraction to FeNi was observed in Layer 1 where the temperature was 940–950 °C while higher nickel extraction was observed in Layer 4 where the temperature was 870–880 °C. In the two-stage thermal treatment, higher first-stage temperatures generate more liquid phases containing more dissolved nickel. However, reduced mobility during cooling limits nickel extraction to FeNi, resulting in a negative correlation between first-stage temperature and final extraction efficiency.

Supplementary Fig. 5 shows the changes in FeNi particle size and WDoC with the equilibrium liquid matte-to-solid FeNi ratio. The main idea of the two-stage thermal upgrading technique is to induce partial liquid formation at temperatures above its solidus temperature (900 °C) and controlled cooling to precipitate FeNi out of the liquid onto previously existing solid FeNi particles. Figure 2g shows that liquid matte and solid FeNi dominate the equilibrium at T > 900 °C. The role of liquid is to facilitate the growth of FeNi particles while the role of solid FeNi seeds is to provide surfaces for FeNi precipitation out of liquid as temperature decreases. It was found that the D80 of FeNi particles demonstrates a strong liner correlation to the liquid matte-to-solid FeNi ratio. The WDoC presents a negative correlation to the liquid-to-solid FeNi ratio, suggesting that as FeNi particles grow, the morphology of FeNi particles becomes irregular.

Effect of size and DoC on magnetic separation

Figure 4d, e illustrate the effect of D80 and average DoC on the generation of oversize FeNi and its impurity content post physical separation. The dashed lines represent fitted trends of the experimental data, while the shaded bands are included only as visual aids to emphasize these trends and do not represent uncertainty, confidence intervals, standard deviation or standard error. A D80 over 80 µm produces around 15–17% oversize FeNi containing around 20% Ni grade at a maximum recovery of 60% Ni. This highlights the significance of growing large FeNi particles and indicates the effectiveness of the thermal upgrading-physical separation process. The DoC of oversize FeNi particles presents a negative correlation to the impurity contents, suggesting that a spherical oversize FeNi particle has a better separation from the gangue.

Figure 4f, g show the Ni, Fe, S, Si, and Mg contents, as well as the Ni/Fe, Ni/S, Ni/Si, and Ni/Mg elemental ratios in the raw ultramafic Ni concentrate, ferronickel concentrate from each layer, and the combined concentrate from all layers. The raw ultramafic Ni concentrate contains 7.7 wt% Ni. After thermal treatment and magnetic separation, the Ni content in the ferronickel concentrate increases significantly, reaching up to 15.8 wt%. Among all layers, Layer 4 exhibits the highest Ni and S contents in the concentrate-16.3 wt% and 12.0 wt%, respectively—while its Si and Mg contents are the lowest. This trend is consistent with the impurity rejection behavior observed earlier for S, Si, and Mg. Interestingly, although Layer 4 was subjected to the lowest thermal treatment temperature, it resulted in the highest Ni grade. The elevated S content implies that the FeNi particles formed in Layer 4 were finer and more closely associated with sulfur phases, further confirming the role of temperature in controlling particle morphology and composition. The Ni/Si and Ni/Mg ratios in Layer 4 are the highest, reaching 5.6 and 7.4, respectively, indicating enhanced Ni enrichment alongside effective Si and Mg rejection. However, the Ni/S ratio is the lowest, pointing to reduced sulfur separation. For the overall process, the combined ferronickel concentrate contains 15.8 wt% Ni, 55.8 wt% Fe, 9.4 wt% S, 3.8 wt% Si, and 2.7 wt% Mg. The Ni/Fe ratio remains relatively stable across the raw ultramafic Ni concentrate, the FeNi concentrate from each layer, and the concentrate. In contrast, the Ni/S, Ni/Si, and Ni/Mg ratios increase substantially—from 0.3, 0.9, and 1.5 in the raw ultramafic concentrate to 1.7, 4.1, and 5.9 in the final combined ferronickel concentrate, respectively—highlighting the preferential enrichment of Ni over non-ferrous impurities.

In summary, higher thermal treatment temperatures promote FeNi particle growth and agglomeration, reducing Ni recovery, allowing only moderate rejection of Si and Mg, but enhancing S rejection. Conversely, lower temperatures result in higher S content in the concentrate, as finer FeNi particles retain more sulfur; nevertheless, they also promote Ni extraction from sulfides and improve the rejection of Si and Mg, ultimately yielding higher Ni recovery and grade.

Comparison of Process

Supplementary Fig. 6 and Table 1 show the comparison of energy consumption and CO₂ emissions between laterite smelting and this method applied to ultramafic-concentrate. Conventional RKEF smelting of lateritic ores (1500–1650 °C) produces ferronickel containing 20–40 wt% Ni with >90% Ni extraction, but at high energy intensity (3,500–8,000 kWh t^-1 FeNi) and carbon emissions of 2.8–6.4 t CO₂ t^-1 FeNi (7–32 t CO₂ t^-1 Ni)^29–32. By contrast, this thermal treatment process route employs low-temperature, predominantly solid-state extraction at 800–950 °C, delivering 89-95% Ni extraction and a 30–32 wt% Ni ferronickel. The authors conducted a comprehensive materials and energy balance for the thermal treatment process at 800 °C and 950 °C on a basis of 2000 t ultramafic concentrate as feed materials (or 435 t FeNi as products). The energy balance shows that the thermal treatment process requires 70–115 Nm³ natural gas, equivalent to 5000–9000 GJ/2000 t ultramafic concentrate. The estimated energy consumption is 3000–5300 kWh t^-1 FeNi. The energy balance assumes that the combustion of natural gas yields off-gas containing H₂O/H₂ = 1 and CO/CO₂ = 1. If heated in a natural-gas-fired kiln, the CO₂ emissions is only 0.3–0.5 t CO₂ t^-1 FeNi (1–1.7 t CO₂ t^-1 Ni) whereas electrical heating may be applied, leading to negligible direct CO₂ emission. In the laterite RKEF flowsheet, the electric-furnace step contributes ~20% to the total process energy, with drying and pre-reduction accounting for ~80%^29,30,33. In the thermal treatment route, the energy consumption is about ~80% in the extraction step and ~20% in magnetic separation. The lower operating temperature, fuel shift, and rebalanced unit-operation footprint together underpin the markedly lower energy intensity and carbon emissions of the ultramafic-concentrate pathway.

Figure 6 depicts the green nickel production pathway from mine to the market using this technology. The low-grade ultramafic nickel ore is upgraded to a concentrate and then subjected to a low-temperature predominantly solid-state thermal upgrading process. Nickel values are enriched in magnetic ferronickel that can be efficiently recovered via magnetic separation. The ferronickel concentrate ( ~ 16% Ni) is a potential intermediate to produce battery-grade nickel, nickel-based superalloys, and stainless steel. Nickel-lean tailings from both ore beneficiation and chemical extraction process are potential resources to capture CO₂. The entire process utilizes unconventional nickel resources and is clean and environmentally friendly, providing green nickel for global new energy conversion and utilization as well as a low-carbon society. This contribution aligns with the United Nations’ initiatives, including affordable and clean energy, sustainable cities and communities, responsible consumption and production, and climate action.

Fig. 6. Green nickel production pathway from ultramafic ores to ferronickel concentrate and downstream nickel products.

Schematic illustration of the proposed process route, including 1, mining of ultramafic Ni sulfide ore; 2, mineral processing to produce ultramafic Ni sulfide concentrate; 3, thermal treatment of briquettes; 4, magnetic separation to recover ferronickel concentrate; 5, further processing toward battery-grade nickel, Ni-based superalloys and stainless steel; and 6, potential carbon capture using residual tailings. The values associated with each step indicate the representative Ni grade and Ni distribution along the proposed pathway.

Conclusions

This study explores a mini-plant scale process for efficient nickel extraction from unconventional ultramafic nickel ores. The temperature distribution within a kilogram-scale reactor and its effects on nickel recovery and ferronickel particle growth were investigated. Furthermore, the roles of particle size and circularity in physical separation efficiency were elucidated. The main conclusions are as follows:

1. The mini-plant scale process, encompassing thermal upgrading and magnetic separation, produced stable and reproducible results. This process presents a pathway for the clean and efficient exploitation of unconventional, untapped, low-grade nickel resources. This method is characterized by low-temperature operation, elimination of smelting processes, absence of smelting slag, and no sulfur dioxide emissions. Compared with the conventional laterite RKEF smelting, this process achieves comparable Ni recovery and product grade while markedly lowering energy intensity and CO₂ emissions, highlighting its efficiency and sustainability.

2. Ferronickel can be directly and efficiently produced from low-grade ultramafic nickel concentrate via thermal upgrading at 810-950 °C under inert atmosphere by iron addition and magnetic separation at 250-2500 gauss at room temperature. The starting ultramafic concentrate contained 7.7% Ni, which was increased to 16% Ni in the FeNi product, with a Ni recovery of 89%.

3. Temperature has a significant impact on nickel enrichment into ferronickel, reflecting on nickel extraction and ferronickel particle size and degree of circularity. At 810–950 °C, nickel extraction decreases from 95% to 89% while nickel grade in the FeNi alloy varies between 30 and 33% (before physical separation). Higher temperatures facilitate the growth of large ferronickel particles (D80 > 100 µm) but reduce circularity (WDoC < 0.5).

4. Large ferronickel particles (D80 > 100 µm) are readily liberated from the host gangue. A D80 over 80 µm produces around 15–17 wt% oversize FeNi containing around 20% Ni grade at a maximum recovery of 60% Ni. For the overall process, the combined ferronickel concentrate contains 15.8 wt% Ni, 9.4 wt% S, 3.8 wt% Si, and 2.7 wt% Mg.

Author contributions

W.L. and F.W. conceived and designed the study, performed the experiments, analyzed the data, prepared the figures and wrote the original draft. B.M. contributed to briquetting experiments and scientific discussions. L.L. contributed to the design and execution of selected thermal treatment experiments. F.D.F. performed the MLA analysis and related mineralogical data analysis. M.X. contributed to magnetic separation discussions and analysis, and prepared the feed material through mineral processing. S.M. provided guidance on the research direction, contributed to scientific discussions and reviewed the original draft. M.B. supervised the project, provided guidance on the research direction, contributed to scientific discussions and reviewed the original draft.

Peer review

Peer review information

Communications Engineering thanks Okechukwu Vincent Dickson and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: [Philip Coatsworth]. A peer review file is available.

Funding

This research was financially supported by Vale Base Metals and by the Natural Sciences and Engineering Research Council of Canada through NSERC Grant ALLRP 592422-23.

Data availability

Data will be made available on request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s44172-026-00704-6.