Beyond the silicate matrix: SQUID magnetometry and Raman spectroscopy reveal hidden hausmannite in sintered manganese ores

The depletion of high-grade manganese reserves necessitates the effective beneficiation of low-grade ferruginous manganese resources, which constitute the dominant share of Kazakhstan’s reserve base. This study investigates the structural and magnetic evolution of carbonate-rich Bogach and silicate-rich Zhaksy ores sintered between 600 °C and 1,200 °C. By integrating bulk elemental analysis with high-resolution Raman spectroscopy, Superconducting Quantum Interference Device magnetometry, and quantitative electron microscopy mapping, the research defines distinct phase transformation mechanisms. Results indicate that sintering at 1,000 °C triggers hausmannite nucleation in both ore types. In Bogach ore, carbonate decomposition facilitates phase segregation and the growth of large crystalline grains. Conversely, Zhaksy ore forms iron-substituted solid solutions where nanocrystalline hausmannite remains encapsulated within a stable silicate matrix. While these inclusions are invisible both to standard X-ray diffraction and to Raman spectroscopy, their presence is unambiguously established by a characteristic ferrimagnetic transition at 42 K. These microstructural differences dictate tailored beneficiation pathways: Bogach ore is amenable to high-intensity magnetic separation, whereas the silicate-matrix effect in Zhaksy ore requires either upstream removal of quartz ballast before sintering or a combination of ultra-fine comminution with high-gradient magnetic separation. These findings establish a notable correlation between thermal history and functional magnetic properties, providing a robust physical basis for industrial optimization.

Manganese is a key element in the global industrial landscape, serving as a critical component in steel production (Craddock, 2020), where it acts as a deoxidizer and desulfurizer and as an alloying agent for improving hardness and wear resistance (Gasik et al., 2020; Alrobei et al., 2023; Zhuniskaliyev et al., 2024). Beyond metallurgy, manganese oxides find extensive applications in the chemical industry, electrochemical energy storage, and environmental catalysis, where their redox chemistry and structural diversity support catalytic and sorption functionality (Garten et al., 2021; Sukhdev et al., 2020; Parveen et al., 2022; Ghosh, 2020). As global reserves of high-grade manganese ores are gradually depleted, the metallurgical industry has shifted toward the utilization of medium- and low-grade deposits, which are often characterized by a manganese content below 20–25 wt% and elevated silica levels exceeding 25–30 wt%, thereby increasing slag volume, coke consumption, and overall energy demand in ferroalloy production (Zhunusov et al., 2021).

Within the Republic of Kazakhstan, this challenge is structural: the average manganese content in domestic ores is approximately 19.4 wt%, substantially below the typical level of 30–50 wt% reported for the principal manganese-producing countries (Makhambetov et al., 2024). Such grades fall well below the threshold for direct industrial use, since ores with Mn < 34 wt% and Mn/Fe < 2.5 are generally considered to have limited direct applicability and are processed only after blending with high-grade material (Singh et al., 2021). The dominant share of the national reserve base therefore depends on prior beneficiation to upgrade the manganese content and to release it from the host matrix. Several beneficiation routes are under investigation for these resources, including gravity and magnetic concentration of the raw ore as well as thermal upgrading of the sintered product; the suitability of any specific route depends on the mineralogical structure of the deposit and on the response of the ore to the corresponding processing regime (Naboko et al., 2023; Samuratov et al., 2010).

This trend is particularly relevant for Central Kazakhstan, which hosts significant iron-manganese and carbonate–oxide resources belonging to the Zhezdy–Ulutau and Atasu ore regions (M. Z. Tolymbekov, 2002). Individual deposits from these groups may exhibit manganese contents as low as 15–17 wt% against silica contents of 40–49 wt% in typical run-of-mine ores, as documented for the Zhezdy–Ulutau type (Taza Metal Technologies Inc., 2024; Abishev et al., 2024). At the same time, off-balance reserves such as the Akzhar–Sarytuma area contain ores with an average manganese grade of approximately 22–23 wt% and low iron content, which makes them suitable candidates for agglomeration-based upgrading (Naboko et al., 2023). Under such conditions, the economic and environmental rationale for processing off-balance ores is determined by the potential to obtain high-grade agglomerates suitable for ferromanganese and silicomanganese smelting while reducing waste rock disposal and the footprint of tailings storage (Samuratov et al., 2010; Duka et al., 2011). The effective use of these complex ores requires technical beneficiation strategies, among which high-temperature sintering followed by magnetic separation is considered a technologically viable route for Central Kazakhstan raw materials (Naboko et al., 2023; Samuratov et al., 2010; Bhoja et al., 2021).

Thermal treatment, specifically high-temperature sintering, is a method for upgrading manganese ores by inducing mineralogical phase changes, improving mechanical strength, and increasing the manganese-to-iron ratio in the agglomerated product (Zhunusov et al., 2021; Samuratov et al., 2010). A significant body of research has been dedicated to the phase transitions occurring during the calcination of manganese-bearing minerals such as pyrolusite, psilomelane, and rhodochrosite, demonstrating that heating manganese oxides in oxidizing atmospheres typically promotes the formation of hausmannite (Mn 3 O 4). This spinel-type oxide is of particular interest due to its ferrimagnetic ordering at low temperatures, with studies utilizing SQUID magnetometry to confirm a Curie temperature (TC) of approximately 42–43 K (Schwarz et al., 2023; Thota et al., 2011). From a thermodynamic perspective, hausmannite occupies a broad stability field in the Mn–O phase diagram, yet its precise formation pathway is sensitive to the temperature-time profile and oxygen partial pressure (Fritsch et al., 1998). In situ investigations reveal that the reduction of Mn 2 O 3 to Mn 3 O 4 occurs above 900 °C in air, while the formation of Mn 5 O 8 may serve as a metastable intermediate during low-temperature oxidation processes (Sukhdev et al., 2020; Darul et al., 2013).

In complex silicate systems, the crystallization of manganese oxides is further complicated by the partitioning of Mn cations into braunite and related silicate frameworks (Litvin et al., 1984). These structures effectively compete with simple oxides for manganese, thereby shifting the thermodynamic driving forces and kinetic pathways of hausmannite nucleation during the sintering process (Tolymbekov et al., 2010; Safarov et al., 2025; Chen et al., 2018). These mineral-matrix interactions dictate the kinetic pathways of phase evolution, being influenced by the presence and reactivity of gangue minerals, particularly silicates and carbonates, whose melting and decarbonation behavior can facilitate or inhibit the growth of specific phases (Zhunusov et al., 2021). Such interactions affect the collective magnetic response of the final sinter, making it necessary to account for the microchemical distribution of elements within the slag phase (Silvestri et al., 2020; Hizi et al., 2022). A deeper understanding of these interactions is essential for optimizing industrial furnace regimes and minimizing specific energy consumption during the processing of Central Kazakhstan ores (Samuratov et al., 2010; Brusnitsyn et al., 2021). Furthermore, control over the phase assemblage ensures that the resulting agglomerate is compatible with subsequent magnetic beneficiation stages.

Despite the extensive literature on manganese mineralogy, the characterization of nanocrystalline or structurally disordered oxide phases within complex silicate matrices remains an analytical challenge. While X-ray diffraction (XRD) is the primary tool for phase identification, its sensitivity is limited when coherent scattering domains fall into the nanometer range or when their volume fraction constitutes only a minor component of the sample (Soulamidis et al., 2023; Bernardini et al., 2020). This limitation is critical in sintered ores, where partial melting and subsequent glass-ceramic formation further reduce long-range order, often encapsulating oxide inclusions within amorphous matrices (Di Genova et al., 2016; Schmid and Petra, 2020). This structural masking is consistent with the concept of “remnant metastability” (Chen et al., 2018), where surface energy contributions stabilize phases - such as specific polymorphs of the MnO 2 or hausmannite families - that remain invisible to standard diffraction despite their thermodynamic stability in the nanoscale (Soulamidis et al., 2023). For instance, Garten et al. demonstrated that in samples where XRD exclusively exhibited MnO reflections, the application of Raman spectroscopy successfully identified hausmannite (Mn 3 O 4) peaks at 650 cm -1, revealing a phase coexistence that was obscured at the powder-diffraction level (Garten et al., 2021; Bernardini et al., 2020; Tonu et al., 2024).

Recent advancements in complementary structural probes, such as Raman spectroscopy and sensitive magnetometry, provide opportunities to resolve these discrepancies by accessing local vibrational modes and collective magnetic phenomena that are largely insensitive to long-range crystallinity. In particular, low-temperature magnetometry is capable of detecting ferrimagnetic behavior associated with nanometer-scale Mn 3 O 4 domains far below the size and concentration thresholds required for reliable XRD detection (Sukhdev et al., 2020; Günay et al., 2012; Laha et al., 2014). This is supported by findings that nanostructured Mn 3 O 4 exhibits specific magnetic effects, such as spin canting on the particle surface, which leads to changes in coercivity and magnetization compared to bulk materials (Sukhdev et al., 2020; Shaif et al., 2020). Furthermore, measurements of magnetic susceptibility (χ) offer high sensitivity to the oxidation state of manganese, allowing for the tracking of phase transitions even at early stages where diffraction patterns remain insensitive to the presence of nanocrystalline or structurally disordered Mn–oxide domains (Brusnitsyn et al., 2021; Berg and Olsen, 2000).

The present research is a continuation and expansion of our previous study (Safarov et al., 2025), in which we investigated the basic structural evolution of Central Kazakhstan manganese ores using XRD, XRF, and scanning electron microscopy (SEM-EDS). In that earlier work, we identified the primary crystalline products of sintering, yet questions regarding the magnetic functionality and the microchemical distribution of elements remained. These limitations arose primarily from the spatial resolution of SEM-EDS for sub-micron inclusions and the insensitivity of XRD to nanoscale domains. Building upon these findings, the current study use a suite of advanced analytical methods, including SQUID magnetometry, high-resolution Raman spectroscopy, and quantitative pixel-wise EDS mapping, chosen to investigate the “hidden phase” problem.

By integrating these tools, this study investigates previously unexamined aspects of sintering dynamics in Central Kazakhstan manganese resources. Specifically, we analyze the emergence of ferrimagnetic ordering at cryogenic temperatures as a characteristic indication of hausmannite nucleation and employ quantitative elemental correlation analysis to distinguish between phase segregation and the formation of Mn-bearing solid solutions. Our working hypothesis is that the degree of sintering and partial vitrification of the silicate matrix controls both the spatial distribution of manganese and the blocking of magnetic moments in Mn 3 O 4 nanoparticles, influencing the efficiency of subsequent magnetic separation by preventing the entrapment of manganese within the silicate slag. The findings provided herein thus establish a physical basis for optimizing magnetic beneficiation technologies for off-balance manganese ores and contribute to the development of processing pathways for complex manganese resources of Central Kazakhstan.

The instrument identifications and the principal analytical conditions that affect data interpretation are reported in this section. The full experimental protocols, including standard treatment procedures, sample-preparation parameters and the internal parameters of the data-analysis software, are provided in Supplementary File S6 (https://doi.org/10.5281/zenodo.19642586).

Sample preparation

The study used representative samples of manganese ores collected from the Bogach and Zhaksy deposits located in Central Kazakhstan. The materials were investigated in their initial (raw) state and after isothermal annealing in an ambient air atmosphere for a duration of 4 h. Thermal treatment was conducted using a laboratory muffle furnace at temperatures of 600 °C, 800 °C, 900 °C, 1,000 °C, 1,100 °C, and 1,200 °C. After the required holding time, the samples were naturally cooled to room temperature before further physicochemical and magnetic characterization.

Physical-chemical investigation

Elemental and microstructural analysis

ICP-OES

The elemental composition of the initial and sintered ore samples was determined by Inductively Coupled Plasma Optical Emission Spectroscopy on a Thermo Scientific iCAP 7,200 series spectrometer (United States) using high-purity argon (5.0 grade, Linde Group) as the carrier gas. Sample digestion was performed in a Milestone Ethos EASY microwave reaction system (Italy) using a reversed aqua regia mixture; weighing was carried out on a KERN ALS 220-4 analytical balance (Germany). Calibration was performed against certified reference materials (Sigma-Aldrich), and all measurements were conducted in duplicate. The full digestion and dilution protocol is reported in File S6.

High-resolution EDS mapping of initial ores

High-resolution morphological characterization and elemental mapping of the initial ores were performed at the Cavendish Laboratory (University of Cambridge) on a Helios NanoLab 650 DualBeam workstation (FEI Company, United States) equipped with an integrated Oxford Instruments Energy Dispersive X-ray Spectroscopy (EDS) system (United Kingdom). Maps were acquired at 25 kV with a live time of approximately 100 s and processed using the associated Oxford Instruments software. The high-resolution map repository is hosted in the Zenodo repository (https://doi.org/10.5281/zenodo.19642586).

The elemental composites shown in Figures 1A,B were generated directly from the per-element X-ray intensity arrays exported from the AZtec project (float32, 381 × 274 pixels for the Bogach sample and 456 × 305 pixels for the Zhaksy sample). Each element map was normalized over a percentile range above the detector Poisson-noise floor and combined using a screen-blending operator with a colour-blind-aware palette assigning green to Mn, yellow to Si, red to Ca, magenta to K, and cyan to Ba (Figure 1A) or to Fe (Figure 1B), with blue for Al in Figure 1B. The colour composite was fused with the backscattered electron image via HSV blending so that surface topography is preserved under the coloured overlay. Figures were saved as lossless 600 dpi PNG. The processing pipeline (Python 3, NumPy, Pillow) is archived as Software S5 in the Zenodo repository, and the full per-element parameter set is reported in File S6.

Quantitative SEM-EDS analysis of sintered samples

Quantitative elemental microanalysis of the sintered samples was conducted at the University of Latvia on a Helios 5 UX scanning electron microscope (Thermo Fisher Scientific, United States) equipped with an integrated EDS detector. Sample preparation included pellet pressing on a Specac Atlas 15T hydraulic press (United Kingdom) and carbon coating on a Quorum Q150R sputtering system (United Kingdom); the pressing and coating parameters are reported in File S6. EDS measurements were conducted at 20 kV in spectral imaging mode with data processing in Pathfinder software (Thermo Fisher Scientific). A complete collection of micrographs and elemental maps for all sintered samples (600 °C–1,100 °C) is provided as File S3.

Quantitative elemental data were obtained by integrating the EDS spectrum over the entire field of view of the micrographs at two distinct magnifications: low (×1,000, approximately 150 × 100 μm) for the bulk composition of particle aggregates and high (×8,000, approximately 18 × 12 μm) for the local phase composition within specific morphological features. This dual-scale mapping allows a distinction between macro-homogeneity and micro-phase segregation.

Quantitative analysis of element distribution maps

Pixel-wise quantitative analysis of the EDS maps of sintered samples was performed using a custom Python analytical suite (Software S1, available at https://doi.org/10.5281/zenodo.19642586). Grayscale pixel intensities of the per-element maps (8-bit, 379 × 470 pixels) were converted into weight percent concentrations using a linear calibration model derived from the embedded colorbar scale. A pixel-wise database was constructed by spatially co-registering the element maps acquired from identical sample areas and by excluding background pixels with a total elemental sum below 1.0 wt%. Bivariate density plots of Mn versus Si (sintered Bogach ore at 1,000 °C–1,200 °C) and Mn versus Fe (sintered Zhaksy ore at 1,100 °C) were used to detect phase segregation and isomorphic Fe substitution into the Mn-oxide lattice, respectively. The full image-processing pipeline, the calibration equation and all script parameters are documented in File S6 and in Software S1.

Thermal analysis (TGA-DSC)

Thermal stability and phase transformation kinetics were investigated by simultaneous Thermogravimetry and Differential Scanning Calorimetry on a Labsys Evolution TG-DTA/DSC instrument (Setaram, France). Samples were heated from 30 °C to 1,200 °C at 10 K/min in an Al 2 O 3 crucible under a nitrogen flow of 30 mL/min. The original measurement reports, including baseline subtraction, mass-loss step calculations and peak integration for the determination of reaction enthalpies (ΔH), are provided in File S4.

Raman spectroscopy

Micro-Raman spectroscopic measurements were performed on a high-resolution TriVista 777 triple-grating spectrometer (Spectroscopy & Imaging GmbH, Germany) with 532 nm solid-state laser excitation, an integrated Olympus BX53 microscope and a Peltier-cooled CCD detector. The raw two-column ASCII spectral files are provided as Data S1. Spectral processing was performed with a custom Python script archived as Software S2 in the Zenodo repository, following the pipeline described in the Phase Evolution and Local Structure section: replicate averaging, rolling-minimum baseline correction with Savitzky–Golay smoothing, min-max normalization per spectrum and overlay of phase-diagnostic markers.

Magnetic measurements

Magnetic measurements were performed on a Quantum Design MPMS3 SQUID Magnetometer. Hysteresis loops were recorded at 10 K and 300 K with magnetic fields swept up to 10 kOe (1 T); thermomagnetic curves were recorded in applied fields of 50 Oe and 100 Oe between 10 and 350 K, in the chronological sequence Zero-Field Cooled warming followed by Field Cooled cooling and warming, so that the magnetic irreversibility is preserved.

Hysteresis data at 300 K (where the magnetic signal is comparable to the instrument noise floor) were processed by Langevin fitting combined with a linear paramagnetic term, with an automatic fallback to robust linear regression for non-saturating samples. Hysteresis data at 10 K (high signal-to-noise ratio) are reported as raw experimental points without smoothing or modelling. Temperature-dependent magnetization was processed in batch mode with explicit preservation of the chronological measurement order so that the ZFC/FC bifurcation is retained. The full algorithms, all numerical parameters and the source code are provided as Software S3 (Hysteresis_Combined_10-300K.py) and Software S4 (thermomagnetic_analysis.py) in the Zenodo repository, with the corresponding raw datasets in Data S2 and Data S3.

Quantitative phase analysis of the crystalline component of the same Bogach and Zhaksy ore series, sintered under identical conditions, was performed in our previous publication (Safarov et al., 2025) by Rietveld refinement against a corundum (Al 2 O 3) internal standard, with weight percentages of all crystalline phases and of the X-ray-amorphous component reported as a function of sintering temperature. The corresponding numerical values are referred to in the Discussion below.

Elemental composition and chemical homogeneity

Bulk elemental composition (ICP-OES)

To obtain precise quantitative data on the bulk composition, the ore samples were analyzed using ICP-OES. While previous XRF studies provided surface-averaged estimates, the acid digestion method used here allows for the determination of the bulk metal content available for phase transformations. The elemental composition of the initial ores (Bogach and Zhaksy) is summarized in Table 1.

The ICP-OES data reveals fundamental differences between the two deposits. Bogach ore is a high-grade source with a manganese content of 40.65 wt%. Zhaksy ore contains significantly less manganese (28.90 wt%).

Zhaksy ore contains an order of magnitude more iron (3.53%) compared to Bogach (0.32%). This is relevant for magnetic separation, as iron oxides can form solid solutions with manganese, potentially affecting the magnetic hardness.

The near-zero silicon content detected in the solution for both ores (despite XRF showing high Si for Zhaksy (Safarov et al., 2025)) confirms that the matrix consists of acid-resistant quartz and stable silicates which were not dissolved during sample preparation. This confirms that the high silica content (up to 20.85 wt%) identified by XRF in our previous study (Safarov et al., 2025) is associated with chemically stable quartz and braunite phases, which remain in the solid residue during acid digestion. In contrast, the high calcium content in Bogach (4.39%) points to the presence of acid-soluble carbonates (calcite), which readily decompose during sintering.

We also monitored the evolution of elemental concentration during thermal treatment. Sintering leads to a gradual increase in the relative metal content due to the removal of volatile components (loss on ignition). For instance, in Bogach ore, the Mn concentration rises from 40.7% (initial) to ∼49.0% in samples sintered at 1,100 °C (see Zenodo repository for full dataset Supplementary Table S1). This concentration increase represents a chemical enrichment effect resulting from the volatilization of unstable components during high-temperature treatment, which is a characteristic feature of the restructuring of complex manganese ores.

Morphological and elemental mapping

To establish a baseline for the microstructural study, the initial state of the ores was first investigated using high-resolution EDS spectral imaging performed at the Cavendish Laboratory (University of Cambridge) (Figure 1). High-resolution EDS spectral imaging confirms a notable textural contrast between the two deposits.

The Bogach ore (Figure 1A, File S1) exhibits a homogeneous microstructure where manganese (green) is uniformly distributed throughout the matrix. This confirms the oxide nature of the ore, where Mn-bearing minerals constitute the continuous phase with minimal segregation of gangue minerals. Moreover, the elemental maps for the initial Bogach ore, acquired at a scale of 25 μm, reveal a high degree of spatial correlation between manganese and oxygen, supporting its primary oxide nature. While manganese is uniformly distributed, calcium (red) constitutes a substantial fraction of the matrix (4.27 wt% by EDS quantification) and appears as extensive cloud-like enrichment zones distributed across the Mn-oxide framework rather than as isolated point sources. This pervasive calcium presence indicates abundant carbonate inclusions (calcite) intergrown with the oxide phase, a textural feature that has direct consequences for the subsequent sintering behaviour: thermal decarbonation of calcite during heating (CaCO 3 → CaO + CO 2) is expected to generate significant secondary porosity within the sintered product. A compact silicate inclusion, visible as a bright yellow cluster in the upper-left of the mapped area, is spatially uncorrelated with the calcium zones and represents the only resolvable Si-rich grain within the 54 × 39 μm field, consistent with the overall low silica content of the deposit. Potassium (pink) and barium (cyan) produce sparse, evenly scattered signals without distinct localisation, indicating that both elements occur as trace substituents rather than discrete mineral phases.

In contrast, the Zhaksy ore maps (acquired at a larger scale of 500 μm) (Figure 1B) demonstrate significant phase segregation. Silicon (yellow) and iron (cyan) distributions reveal a complementary architecture: Si-rich quartz grains occupy the voids within the Mn-oxide network, while the iron signal is closely associated with manganese clusters rather than the silicate matrix, providing visual evidence for the potential formation of (Mn, Fe)-oxide phases during subsequent sintering. This distinct phase segregation contrasts with the continuous Mn-distribution observed in the Bogach ore and aligns with the ICP-OES findings regarding the silicate nature of the Zhaksy deposit, illustrating a structural architecture where the manganese-bearing phases are physically dispersed and diluted within significant volumes of gangue minerals. Aluminium (blue), potassium (pink) and calcium (red) appear only as minor discrete signals distributed among the silicate and iron-rich zones.

Having identified these initial textural differences, we then investigated how thermal treatment induces microchemical transformations. Utilizing the analytical facilities at the University of Latvia, we tracked the phase evolution of these ores across a temperature range of 600 °C–1,100 °C.

Microchemical evolution and phase segregation

The microchemical evolution and phase segregation triggered by sintering were systematically investigated across the entire temperature range. The complete visual dataset of surface transformations and the corresponding area-averaged microanalysis results are available in File S3 and Supplementary Table S2, respectively.

The Bogach ore in initial state demonstrates high heterogeneity at the aggregate level (×1,000), with some regions showing significant calcium enrichment (up to 25.5 wt% Ca, File S3-Sample 001, Supplementary Table S2), corresponding to the carbonate fraction. Notably, sintering at 1,000 °C and above leads to a pronounced coarsening of the manganese phase. At high magnification (×8,000), specific scanned areas (e.g., File S3-Sample 019, Supplementary Table S2) exhibit an elemental composition of ∼94 wt% Mn with negligible impurities. The fact that the integrated signal from the entire ∼200 μm 2 scanned area corresponds to nearly pure manganese oxide indicates that the hausmannite crystallites have grown significantly, exceeding the scale of the microscopic field of view. This indicates a separation of the magnetic phase from the silicate/aluminate slag regions (e.g., File S3-Sample 023, Supplementary Table S2).

In contrast, the Zhaksy ore maintains a chemically complex signature even at the highest sintering temperatures. Regardless of the magnification (×1,000 or ×8,000) or the specific area selected, the EDS spectra consistently reveal a mixture of Manganese (40–60 wt%), Silicon (15–30 wt%), and Iron (∼5–10 wt%). We did not observe any scanned areas of pure manganese oxide in the sintered Zhaksy samples. This persistent “mixed” signal suggests that the hausmannite phase and the silicate matrix are intergrown at a scale much smaller than the resolution of the EDS scan area (i.e., sub-micron or nanoscale). This provides chemical evidence that the manganese phase in Zhaksy ore is not isolated but exists as dispersed inclusions within a Fe-doped silicate glass matrix. This microstructural architecture is expected to influence the crystallization behavior and magnetic response of the material.

Quantitative pixel-wise phase analysis

To move beyond qualitative observation, a quantitative pixel-wise correlation analysis was performed on the EDS elemental maps of the high-temperature sintered samples (≥1,000°С). This analysis was executed using the custom Python analytical suite (Software S1), with the underlying pixel-wise concentration data provided in Supplementary Table S3. This method allows for the statistical visualization of element co-localization, indicating the fundamental chemical interactions within the material matrix (Figure 2).

For the Bogach ore (Figure 2A), the joint distribution of Mn and Si exhibits a characteristic orthogonal pattern (L-shape). The data points differentiate into two mutually exclusive populations: silicon-rich/manganese-poor regions (silicate slag) and manganese-rich/silicon-poor regions (oxide grains). The absence of data points in the central region of the plot (high Mn + high Si) provides quantitative indications of phase segregation. This suggests that sintering leads to the purification of the hausmannite phase by segregating silicate impurities to the grain boundaries.

In contrast, the Zhaksy ore (Figure 2B) exhibits a positive linear correlation between Mn and Fe concentrations. The regression trend line demonstrates that iron accumulation is directly proportional to manganese content. This collinearity indicates that iron does not crystallize as a separate phase (e.g., hematite) but is chemically incorporated into the manganese oxide lattice. This provides evidence for the formation of an iron-substituted solid solution, likely (Mn,Fe) 3 O 4. Such lattice substitution is chemically distinct from the pure oxide segregation observed in the Bogach ore.

Thermal decomposition and phase stability (TGA-DSC)

To investigatethe phase transitions leading to the formation of magnetic oxides, the thermal behavior of the ores was analyzed. The TG and DSC curves (Figures 3, 4) show notable differences in the decomposition mechanisms of the two deposits. The quantitative parameters characterizing these decomposition stages are summarized in Table 2, with the full calculation reports available in File S4.

The Bogach ore exhibits a significant total mass loss of 27.5%, reflecting its high content of volatile components, carbonates, and reducible manganese oxides (Figure 3). This mass loss provides an indication of the chemical enrichment effect observed in the elemental analysis (Section 3.1.1), indicating that the increase in relative manganese concentration is driven by the removal of volatile components during thermal treatment. The observed weight decrease correlates directly with the high calcium content (4.39%) identified by ICP-OES. The decomposition proceeds in the following stages: 30 °C–400 °C (Dehydration): A mass loss step is associated with the removal of physically adsorbed water and the initial dehydroxylation of manganese hydroxides (MnO(OH) → Mn 2 O 3).

400 °C–1,200 °C (Carbonate Decomposition and Deep Reduction): This broad temperature interval accounts for the majority of the mass loss and is characterized by a notable endothermic effect with a total enthalpy of 3535 J/g. This intense energy absorption is attributed to two simultaneous high-energy processes:

The thermal decomposition of carbonates, primarily calcite (CaCO 3 → CaO + CO 2), which releases notable amounts of CO 2 gas.

The stepwise reduction of higher manganese oxides (MnO 2 → Mn 2 O 3 → Mn 3 O 4).

The magnitude of this peak suggests that the Bogach ore undergoes a structural change, transforming from a carbonate-oxide mixture into a dense hausmannite-based ceramic. >1,000 °C (Sintering): Above this temperature, the mass loss stabilizes as the material reaches a sintered state with the crystallization of the final hausmannite phase.

In contrast, the Zhaksy ore demonstrates a moderate total mass loss of 11.85%, reflecting the relative thermal stability of its silicate-rich matrix (quartz, braunite) and low carbonate content (Figure 4). The decomposition proceeds as follows: 30 °C–400 °C (Dehydration): A mass loss is observed, attributed to the evaporation of physically adsorbed moisture and partial dehydroxylation of amorphous aluminosilicates and clay minerals.

400 °C–1,000 °C (Oxide Reduction and Matrix Restructuring): This range involves the stepwise reduction of manganese oxides (MnO2 → Mn2O3 → Mn3O4). Unlike the Bogach ore, the DSC curve displays a much lower total heat effect (∼537 J/g, see Table 2). The absence of a notable endothermic peak indicates that the material does not undergo significant carbonate decomposition, but rather gradual phase transitions and structural rearrangements of the silicate matrix.

Notably, the DSC data suggest a specific endothermic event around 573 °C, characteristic of the polymorphic transition of quartz from the low-temperature α-phase to the high-temperature β-phase. This observation is consistent with the high silicon content found in the elemental analysis. Above 1,000 °C, structural densification and crystallization of the amorphous silicate matrix occur, leading to the formation of the braunite phase.

The thermal analysis indicates that for both ore types, the critical phase transformations required to form stable manganese oxides are completed by 1,000 °C. The formation of Mn 3 O 4 in this temperature range provides the crystallographic basis for the emergence of ferrimagnetic properties described in the subsequent section. The moderate mass loss in Zhaksy ore (11.85%) contrasts with the high loss in Bogach (27.5%). This suggests that the silicate matrix remains thermally stable and does not undergo significant decomposition, but rather forms a glassy host that allows for the entrapment of the dispersed manganese oxide phase within its structure.

Phase evolution and local structure

Raman spectroscopy was used to investigate the local structural order and identify the formation of the hausmannite phase. The analysis is based on a comprehensive collection of raw spectral data (wavenumber versus intensity) provided in Data S1. The comparative visualization of the Raman spectra was performed using a standardized automated procedure (Software S2): averaging of two spatial replicates, rolling-minimum baseline correction with Savitzky–Golay smoothing, min-max normalization per spectrum, and overlay of vertical markers at phase-diagnostic wavenumbers.

The evolution of Raman spectra for Bogach ores provides insight into the local structural ordering of manganese oxides (Figure 5). The spectra for the initial sample and those sintered at low temperatures (600 °C–900 °C) are characterized by broad, low-intensity bands centered around 630–640 cm -1. This distinct broadening indicates a highly disordered structure with short coherence lengths, typical of poorly crystalline or amorphous manganese oxides mixed with defects.

A significant spectral transformation occurs at 1,000 °C. The broad band sharpens into a distinct, high-intensity peak at ∼658 °cm -1, assigned to the A 1g symmetry and corresponding to the Mn–O breathing vibration of divalent manganese ions in tetrahedral coordination–a diagnostic signature of the hausmannite (Mn 3 O 4) spinel structure. In samples sintered at 1,100 °C and 1,200 °C, two additional well-resolved modes appear in the lower-frequency region at ∼318 and ∼371 °cm -1 and are assigned to the T 2g and, E g vibrations, respectively. The emergence of this three-peak pattern, together with the notable decrease in the Full Width at Half Maximum (FWHM) of the A 1g band, confirms the high crystallinity and the establishment of long-range order in the tetragonal lattice at elevated temperatures. These findings corroborate the bulk XRD data and place the local structural ordering temperature of hausmannite in Bogach at ≥ 1,000 °C.

The Raman spectra of the silicate-rich Zhaksy ore exhibit a much higher complexity compared to the Bogach ore, reflecting its multi-phase mineralogical evolution (Figure 6). The spectra of initial and low-temperature sintered samples are dominated by broad, featureless bands centered around 630 cm -1, characteristic of an amorphous or highly disordered manganese-silicate matrix.

A transition in the vibrational signature is observed starting at 900 °C, where the spectra acquire a rich set of well-resolved modes. At 1,000 °C and 1,100 °C the pattern stabilizes and includes lattice modes at ∼128, 225, 295, 335 cm -1, a prominent Mn–O triplet at 577, 620 and 690 cm -1, and a sharp band at ∼961 cm -1 assigned to the symmetric stretching vibration (ν 1) of the SiO 4 tetrahedron. Collectively, these features are consistent with the lattice vibrations of braunite (Mn 7 SiO 12). Because XRD data indicate that braunite is already present in the initial Zhaksy ore and that its fraction in the crystalline component grows monotonically with temperature (Safarov et al., 2025), the appearance of the diagnostic 961 cm -1 band and the sharpening of the 577/620/690 triplet at 900 °C should be interpreted as the moment when braunite crystallinity reaches the Raman-detection threshold, rather than the moment of its formation. No component attributable to the A 1 g breathing mode of hausmannite (expected at 655–660 cm -1) is resolved in this temperature range; the 690 cm -1 band is the high-frequency component of the braunite triplet, not a shifted hausmannite mode.

The absence of the hausmannite Raman signature, however, is not by itself sufficient to rule out the phase, since the Raman response is sensitive to crystallite size, local structural order and optical accessibility of the scattering volume. Clarifying whether Mn 3 O 4 is genuinely absent or only Raman-silent in these samples requires a probe that is not limited by these constraints. The magnetic response of Mn 3 O 4 is a direct volumetric signature of the ordered Mn sublattice, and SQUID magnetometry, described in the following section, provides the necessary independent test of the hausmannite inventory of the sintered Zhaksy ore.

Magnetic properties

Magnetic hysteresis at low temperature (10 K)

The field dependence of specific magnetization recorded at 10 K reveals a similarity in the phase evolution of both Bogach and Zhaksy ores, despite their differing mineralogical backgrounds (Figures 7, 8).

For both ore types, samples in the initial state and those annealed at moderate temperatures (up to 800 °C) exhibit a quasi-linear dependence (M ∝ H) with negligible hysteresis. The magnetization values remain low (∼2–3 emu/g at 10 kOe), characteristic of paramagnetic or antiferromagnetic manganese oxides where long-range ferromagnetic ordering is absent.

A significant phase change occurs in both ores after sintering at 1,000 °C and 1,100 °C. The M(H) curves develop wide hysteresis loops typical of hard magnetic materials, indicating the nucleation of a ferrimagnetic phase. However, a quantitative comparison shows notable differences: Bogach Ore (oxide-rich): the sintered samples at 1,000 °C and 1,100 °C exhibit a high maximum magnetization (Ms) of approximately 23 emu/g at 10 kOe, with the two temperatures giving essentially the same response within the experimental uncertainty. This ferrimagnetic response is consistent with the presence of hausmannite (Mn 3 O 4) as the principal crystalline manganese phase in the sintered Bogach ore at 1,000 °C–1,100 °C (Safarov et al., 2025).

Zhaksy Ore (silicate-rich): the sintered samples show a similar hysteretic shape but with a reduced maximum magnetization of approximately 17 emu/g at 1,100 °C, which is approximately 25% lower than the corresponding value for the Bogach sinter.

The ratio of manganese content between the two ores (40.65%/28.90% ≈ 1.41) is close to the ratio of the maximum magnetization observed in the sintered samples (approximately 23 vs. 17 emu/g, ratio ≈1.35). This supports the interpretation that the magnetic potential of the sintered product is to first order proportional to the total mass of manganese available to form the hausmannite phase, with the residual ∼4% discrepancy reflecting the fraction of manganese that is partitioned into non-magnetic silicate phases (braunite, glass) rather than into Mn 3 O 4.

Notably, although XRD analysis of the Zhaksy ore (Section 3.1) identified braunite (Mn 7 SiO 12) as the main crystalline phase, the magnetic feature at 10 K–specifically the wide hysteresis loop–is characteristic of hausmannite (Mn 3 O 4). This discrepancy suggests that in the silicate-rich Zhaksy ore, a fraction of manganese reduces to form Mn 3 O 4, likely as nanocrystalline inclusions dispersed within the silicate glassy matrix. These crystallites are too small or disordered to be clearly resolved by XRD but possess a sufficient magnetic moment to dominate the low-temperature response.

The lower M s value in Zhaksy compared to Bogach is a direct physical consequence of the microstructural architecture observed in Section 3.1.2, where the magnetic phases are effectively diluted by large volumes of diamagnetic quartz and silicate gangue. Furthermore, the reduced M s in Zhaksy ore is explained by the formation of (Mn, Fe) 3 O 4 solid solutions, as indicated by the EDS correlation analysis (Section 3.1.4), which generally possess lower magnetic moments than pure hausmannite.

Paramagnetic behavior at room temperature (300 K)

At 300 K, the ferromagnetic features observed at low temperatures are absent for both ores (Figures 9, 10). All samples exhibit strict linear dependencies of magnetization on the magnetic field with zero coercivity, indicating a transition to a paramagnetic state. This confirms that the measurement temperature notably exceeds the Curie temperature of the magnetic phases formed (T C ≈ 42 K for Mn 3 O 4).

However, a comparative analysis of magnetic susceptibility (the slope of the M(H) curves) indicates notably different evolutionary trends for the two deposits: Bogach Ore (Figure 9) in initial condition shows the highest paramagnetic response (∼0.8 emu/g at 10 kOe). Sintering leads to a notable decrease in susceptibility, with the 1,100 °C sample showing the lowest magnetization (∼0.45 emu/g). This is attributed to the recrystallization and densification of the oxide matrix, which reduces the number of surface spins and defects associated with the disordered raw ore.

Zhaksy Ore (Figure 10) exhibits a non-monotonic behavior. The susceptibility initially decreases for the sample annealed at 800 °C. However, a pronounced increase in magnetization is observed for samples sintered at 1,000 °C and 1,100 °C, reaching values of ∼1.8 emu/g, which is more than double the magnetization of the Bogach counterparts.

The notable increase in paramagnetic susceptibility for the high-temperature Zhaksy samples supports the hypothesis of hausmannite nucleation. While the initial Zhaksy ore is diluted by diamagnetic quartz and weakly magnetic braunite, the sintering at ≥1,000 °C leads to the segregation of manganese into Mn-rich clusters or nanocrystals of Mn 3 O 4 within the silicate glass. Even in the paramagnetic state (at 300 K), these Mn-concentrated regions possess a higher magnetic moment per unit mass than the initial mineral mixture, leading to the increased response observed in Figure 10.

Consequently, these divergent trends highlight a notable structural distinction: the magnetic response of sintered Bogach ore is governed by the stabilization of a lattice with reduced defects, whereas the increased susceptibility in Zhaksy ore arises from the nucleation of dispersed Mn-rich nanoclusters within the silicate matrix.

Thermomagnetic analysis (M(T))

Temperature-dependent magnetization measurements performed under weak external fields of H = 50 Oe and 100 Oe (Figures 11–14) provide information regarding the magnetic ordering temperatures and phase transitions for both ore types. Thermomagnetic analysis reveals consistent behavioral patterns for both ore types, determined primarily by the thermal treatment conditions. Specifically, samples sintered at high temperatures (1,000 °C–1,100 °C) exhibit notable magnetic ordering, whereas the initial material and samples annealed at lower temperatures display predominantly paramagnetic properties.

For both Bogach and Zhaksy ores, the samples sintered at ≥1,000 °C exhibit a substantial rise in magnetization at low temperatures, characteristic of a ferro- or ferrimagnetic ordering. A pronounced magnetic transition is observed in the temperature range of 30–50 K for both ore types (Figures 13, 14). The inflection point of this transition corresponds to a Curie temperature of T C ≈ 42 K, which is a characteristic feature of hausmannite (Mn 3 O 4). This indicates that despite the different mineralogical matrix (oxide vs. silicate), the magnetic phase formed during high-temperature sintering is identical. The clear detection of the 42 K transition confirms the nucleation of the hausmannite phase, validating the local structural ordering observed in the high-temperature Raman spectra (Section 3.3).

A notable bifurcation (splitting) between the Zero-Field Cooled (ZFC) and Field Cooled (FC) branches persists up to T C. This irreversibility indicates strong magnetic anisotropy and domain wall pinning within the formed crystallites. In the ZFC mode, several high-temperature samples (e.g., Bogach at 1,000 °C in Figure 11 and Zhaksy at 1,000 °C in Figure 12) exhibit negative magnetization values at low temperatures. This is a physical phenomenon indicating metastable magnetic domains frozen in orientations opposing the weak external field during cooling. Upon warming, thermal energy allows these domains to reorient, resulting in a rise to positive magnetization values before the paramagnetic transition.

While the qualitative behavior (T C) is similar, the magnitude of the magnetic response differs significantly between the two deposits, corroborating the hysteresis data: the magnetization of Bogach ore (Figures 11, 13) reaches peak values of ∼5.5 emu/g at 100 Oe. This high value reflects the high concentration of the magnetic hausmannite phase formed from the Mn-rich oxide precursor.

for samples of Zhaksy Ore (Figures 12, 14) the magnetic response is notably weaker, with peak magnetization reaching only ∼2.2 emu/g at 100 Oe (approx. 40% of the Bogach value). This reduction is attributed to the dilution effect of the non-magnetic silicate glass matrix formed during the sintering of the quartz-rich Zhaksy ore. Nevertheless, the clear detection of the 42 K transition in Zhaksy samples serves as sensitive indication of hausmannite nucleation, which was less evident in the bulk XRD analysis.

The precursors and low-temperature samples (Initial – 800 °C) for both ores display a weak magnetic response (visible in the insets of Figures 11–14). They exhibit notable cusps or broad peaks in the M(T) curves around 80–100 K. This behavior corresponds to the Néel transition (T N) of antiferromagnetic precursors, such as bixbyite (α–Mn 2 O 3, T N ≈ 80 K) or complex Mn(IV) oxides. The presence of these transitions confirms that below 900 °C, both ores remain in a predominantly antiferromagnetic state.

Thus, the thermomagnetic curves indicate that the ferromagnetic-like behavior observed at 10 K in both sintered ores is caused by the same phase–Mn 3 O 4. The results highlight that magnetometry is an effective technique for detecting phase transformations in complex ores, capable of identifying magnetic phases (like nanocrystalline hausmannite in Zhaksy ore) that may be obscured in standard diffraction patterns by amorphous phases.

Mineralogical control over thermal decomposition

The integration of bulk chemical analysis (ICP-OES) with thermal profiling (TGA-DSC) reveals that the sintering mechanism is governed by the initial mineralogical matrix. The demand for high-precision analytical methods to resolve the elemental distribution in complex, multi-component matrices is a well-recognized necessity in modern characterization of diverse natural systems (Zinicovscaia, 2025).

The Bogach ore, previously identified as containing distinct carbonate phases (calcite/ankerite) alongside manganese oxides (Safarov et al., 2025), exhibits notable thermal reactivity. The significant mass loss (27.5%) and the high endothermic enthalpy (∼3535 J/g) observed in this study correspond to the decomposition of the carbonate skeleton. This decomposition is suggested to create a porous, reactive intermediate state that facilitates the mobility of manganese ions and their subsequent recrystallization into the thermodynamically stable phase–hausmannite (Mn 3 O 4).

Quantitative XRD by Rietveld refinement of the same sample series (Safarov et al., 2025) confirms this scenario in concrete numbers: cryptomelane and calcite, which together account for 23 wt% of the initial mineralogy, are fully consumed by 900 °C and 800 °C, respectively, while hausmannite–absent from the initial ore–appears as a major crystalline phase from 1,000 °C onwards, reaching 26.2, 34.3 and 24.7 wt% at 1,000, 1,100 °C and 1,200 °C against an amorphous fraction of 61–74 wt% (Rwp 2.5%–3.0%).

In contrast, the Zhaksy ore behaves as a thermally stable refractory system. Its high silica content (20.85 wt% Si via ICP, confirmed as quartz/braunite by XRD in (Safarov et al., 2025)) acts as a structural backbone. The moderate thermal effects (∼537 J/g) and lower mass loss (11.85%) indicate that the phase transition is not driven by decomposition but by local phase transitions (e.g., α→β quartz) and partial softening of the silicate glass. This explains why the gross morphology of Zhaksy particles remains relatively intact compared to the complete restructuring of Bogach ore.

The reduction sequence observed in the Bogach ore (MnO 2 → Mn 2 O 3 → Mn 3 O 4) is consistent with established thermodynamic studies of manganese oxide stability. As shown by Sorensen et al. (Sorensen et al., 2010) in their study of global manganese deposits (Gabon and CVRD), the phase change to hausmannite (Mn 3 O 4) in air is characteristic end-point of thermal treatment above 900 °C. Furthermore, the high energy consumption (ΔH ≈ 3535 J/g) required for Bogach ore restructuring is typical for carbonate-rich manganese ores, where the endothermic decomposition of the calcite matrix often masks the energy signals of oxide reduction, as noted, for example, in the study by Anacleto et al. (Anacleto et al., 2004).

Resolution of the “missing phase” in silicate ores

Our previous XRD analysis with Rietveld refinement and an internal corundum standard (Safarov et al., 2025) gives the crystalline phase composition of the sintered Zhaksy ore as 23.8, 27.7 and 22.5 wt% braunite at 900, 1,000 °C and 1,100 °C, respectively, against an X-ray-amorphous content rising from 60.4 to 73.1 wt% over the same interval, and identifies no diffraction-resolvable hausmannite at any sintering temperature (Rwp 3.3%–4.0%). The SQUID measurements reported here, however, reveal a distinct ferrimagnetic transition at T C ≈ 42 K, which coincides with the paramagnetic-to-ferrimagnetic transition of pure Mn 3 O 4 established in earlier work (Du et al., 2006; Baron et al., 1998; Chardon and Vigneron, 1986) and therefore constitutes a direct volumetric signature of the hausmannite phase. The apparent contradiction between the diffraction and magnetic responses is a recognised feature of manganese mineralogy: nanocrystalline Mn oxides with small coherent-scattering domains and lattice strain are known to exhibit “X-ray amorphous” behaviour (Post, 1999), so a subordinate fraction of submicron Mn 3 O 4 dispersed within the dominant braunite matrix can produce only a weak, broadened XRD response while remaining fully detectable by the volumetric magnetic probe.

The Raman spectra of the same sintered samples likewise do not resolve the diagnostic A 1g mode of hausmannite. We attribute this to the enclosure of the subordinate Mn 3 O 4 fraction within the amorphous Mn-bearing silicate glass formed above 1,000 °C by partial melting of the Si-rich phase; a polyphasic Mn–Si precursor consistent with this scenario is documented by the macroscale compositional heterogeneity in Mn, Si, Fe and Ca seen in the EDS mapping of the raw ore (Figure 1B). The effective optical sampling depth of Raman excitation in this optically dense, Mn-absorbing glass is insufficient to recover the hausmannite signal above the braunite/glass background, whereas SQUID remains sensitive to the ordered Mn 3 O 4 sublattice regardless of such optical masking. A balanced mineralogical description of complex silicate–manganese systems therefore requires the joint deployment of a volumetric magnetic probe (SQUID), a bulk crystallographic probe (XRD), a compositional mapping probe (EDS) and a vibrational probe (Raman).

A direct analogue of the situation reported here is documented by (Rosado et al., 2018), who demonstrated that XRD analysis of Mn-containing pigments failed to detect the manganese phase owing to its amorphous character, whereas the combination of XRF, SEM-EDS and Raman spectroscopy identified pyrolusite-like MnO 2 together with the accompanying aluminosilicate, quartz and iron-oxide contaminants. A conventional characterization of the sintered Zhaksy ore relying only on XRD, still the default practice in industrial mineralogy of manganese raw materials, would have classified the material as hausmannite-free and would therefore have misestimated both its magnetic functionality and, by extension, its amenability to magnetic beneficiation.

Microchemical architecture: segregation vs. solid solution

The application of quantitative pixel-wise EDS analysis provides information about the crystallization environment. In the Bogach ore, the sintering process facilitates a purification mechanism. The “L-shaped” correlation between Mn and Si indicates that the system tends towards complete phase segregation: manganese oxides crystallize into large, chemically pure hausmannite grains (up to 94 wt% Mn), separating silicate and aluminate impurities into isolated slag pockets (Figure 2A). This segregation is likely aided by the high fluxing action of decomposing Calcium carbonates.

Conversely, the Zhaksy ore forms a nanocomposite or solid solution. The linear correlation between Mn and Fe indicates that iron does not segregate (e.g., as hematite) but is isomorphously substituted into the manganese lattice, likely forming (Mn, Fe) 3 O 4. The silicate matrix stabilizes this state, preventing the diffusion required for grain coarsening and purification. Consequently, the magnetic centers in Zhaksy ore remain chemically modified and physically encapsulated within the glass.

The linear correlation between Mn and Fe in the Zhaksy matrix suggests the formation of a disordered spinel system. This aligns with the work of Jacob et al. (Jacob, et al., 2011), who showed that iron cations can readily substitute for manganese in both tetrahedral and octahedral sites of the Mn 3 O 4 lattice. This substitution is reported to influence the long-range crystalline order, which explains the broadened Raman modes observed in our study. Similar intergrowth patterns in manganese-silicate systems were described by Miletich et al. (Miletich et al., 1998), who noted that silica-rich environments tend to stabilize complex manganese silicates like braunite, encapsulating secondary oxide phases as sub-micron inclusions.

Structural origins of magnetic performance

The difference in magnetic properties is a notable consequence of the microstructural features described above.

The Bogach ore exhibits a notable saturation magnetization (M s ≈ 25 emu/g) because the magnetic phase (Mn 3 O 4) forms a continuous, segregated bulk material with pronounced crystallinity. The Zhaksy ore exhibits a significantly reduced response (M s ≈ 18 emu/g) due to two primary factors: dilution effect–the magnetic volume fraction is physically diluted by the diamagnetic quartz/braunite matrix, which is supported by the 11.85% vs. 27.5% mass loss difference, indicating that Zhaksy retains more non-magnetic mass.

chemical doping–the incorporation of Fe into the hausmannite lattice, indicated by EDS as a solid solution, creates local lattice strain–observed as broadened Raman peaks–and may alter the net magnetic moment compared to pure manganese oxide.

Placed in the context of reference data for pure-phase hausmannite, the saturation magnetization values observed here allow an independent, mass-weighted estimate of the Mn 3 O 4 inventory of the sintered samples. The zero-field extrapolated saturation magnetization of pure bulk α-Mn 3 O 4 at low temperature is reported as 1.51 μB per formula unit, equivalent to approximately 37 emu/g (Schwarz et al., 2023), and at the applied field of 10 kOe used in the present hysteresis measurements the bulk value remains within a few percent of this figure. Taking 37 emu/g as a reference and assuming that Mn 3 O 4 is the dominant contributor to the ferrimagnetic response at 10 K, the magnetization of approximately 23 emu/g measured for the Bogach sinter at 1,000 and at 1,100 °C corresponds to an effective Mn 3 O 4 -equivalent mass fraction of the order of 63 wt%, which is essentially the same for both temperatures. The Rietveld phase analysis of the same samples (Safarov et al., 2025) instead resolves 26.2 wt% crystalline hausmannite at 1,000 °C and 34.3 wt% at 1,100 °C. The difference between the magnetometric and the crystallographic estimate of the Mn 3 O 4 mass fraction therefore decreases from approximately 37 wt% at 1,000 °C to approximately 29 wt% at 1,100 °C. This evolution is consistent with a scenario in which the hausmannite phase of the Bogach ore has nucleated in its full volumetric extent already at 1,000 °C, with a significant fraction in a nanocrystalline form that remains below the detection threshold of standard XRD, and in which further annealing at 1,100 °C progressively converts this nanocrystalline fraction into XRD-resolvable crystallites without changing the total magnetic Mn 3 O 4 content. For the Zhaksy sinter at 1,100 °C, where Rietveld refinement resolves no crystalline hausmannite, the measured magnetization of approximately 17 emu/g, obtained at an applied field of approximately five kOe, implies a Mn 3 O 4 -equivalent mass fraction of at least 45 wt%; this is a lower bound, since Fe substitution into the (Mn,Fe) 3 O 4 spinel reduces the saturation moment per formula unit relative to pure hausmannite (Jacob, et al., 2011; Shaif et al., 2020), and an upward correction for Fe-substitution would raise the estimate further. The quantitative divergence between the crystallographic and the magnetometric estimates of the Mn 3 O 4 content–resolved both across the Bogach sintering series and between the two ore types–is therefore a direct, mass-weighted measurement of the fraction of the magnetic phase that is invisible to conventional diffraction but that nevertheless determines the magnetic functionality of the sintered product.

Practical implications: Magnetic beneficiation, refractory-ore processing, and sintering regime optimization

The phase evolution and microstructural architecture revealed in this study have direct consequences for three areas of industrial practice for low-grade ferruginous manganese ores: the selection of magnetic-separation technology suited to the post-sinter microstructure, the processing of refractory silicate-rich ores in which the magnetic phase is encapsulated within a glassy matrix, and the optimization of the sintering regime itself. The unifying observation is that, for both ore types studied here, the controlling factor is not the bulk magnetic susceptibility but the degree of microstructural liberation of the manganese-bearing phase, which in turn is set by the mineralogy of the precursor and by the thermal history of the sinter.

For the Bogach ore, the sintering process induces a “self-purification” mechanism. Although the recrystallization of the lattice leads to a decrease in paramagnetic susceptibility at room temperature (as defects are annealed), this is outweighed by the morphological evolution. The quantitative EDS maps indicate the formation of notable, segregated hausmannite grains (>>10 μm) that are physically distinct from the silicate slag. Consequently, standard comminution is expected to reach a high degree of liberation, producing individual particles composed primarily of manganese oxide. Despite being paramagnetic at room temperature (T > Tc ≈ 42 K), these pure particles possess sufficient magnetic mass to be effectively recovered using High-Intensity Magnetic Separation (HIMS). The absence of intergrown silicates suggests distinct separation trajectories, predicting high concentrate grades.

In contrast, the Zhaksy ore presents a significant technological challenge due to the “composite” nature of its magnetic phase. While the formation of an Fe-doped solid solution results in a higher specific magnetic susceptibility compared to pure hausmannite (at 300 K), this advantage is negated by the microstructure. The magnetic centers are nanocrystalline and physically encapsulated within a massive diamagnetic silicate glass matrix (as indicated by Raman and Area-EDS analysis). In a magnetic separation process, the magnetic force acting on the dispersed nanoclusters must overcome the gravitational and hydrodynamic drag forces acting on the entire silica-heavy particle, referred to as the matrix effect. Standard magnetic separation is suggested to result in substantial loss of manganese into tailings. Therefore, effective beneficiation of sintered Zhaksy ore would require: Ultra-fine comminution to mechanically break the silicate layers and liberate the magnetic inclusions, though this increases energy costs.

High-Gradient Magnetic Separation (HGMS) to generate sufficient field gradients to capture weakly magnetic composite particles.

The findings on the silicate matrix effect in Zhaksy ore reflect a widespread challenge in the beneficiation of medium-grade ores. According to (Bird, 2019), the physical encapsulation of magnetic oxides within non-magnetic silica often leads to reduced recovery rates in standard magnetic separation. The proposed use of HGMS for such ‘nanocomposite’ ores is supported by the results of Liu et al. (Liu et al., 2024), who showed that increasing the field gradient is the only viable way to capture composite particles where the magnetic moment is diluted by more than 30% by the gangue matrix.

The same observations translate into specific guidance for the design of the sintering regime. The threshold temperature for the nucleation of hausmannite (and therefore for the appearance of the magnetic response on which separation depends) is identified at approximately 1,000 °C in both ore types, which sets the practical lower bound of the operating window. Below this temperature the sinter is paramagnetic and yields no useful magnetic recovery; above it the saturation magnetization plateaus, so a further increase of temperature offers little benefit while raising the specific energy demand. For the carbonate-rich Bogach ore the high decarbonation enthalpy (ΔH ≈ 3535 J/g) is unavoidable and is partially offset by the resulting porosity and self-purification, both of which improve subsequent liberation. For the silicate-rich Zhaksy ore the reaction enthalpy is lower (ΔH ≈ 537 J/g), but a substantial fraction of the input energy is spent heating refractory silica ballast that does not contribute to the magnetic concentrate; this argues for upstream pre-concentration to remove quartz before thermal treatment, or for the search of fluxing additives that would lower the critical crystallization temperature of Mn 3 O 4 below 1,000 °C and reduce the specific energy demand of the sintering stage.

Limitations of the study

Despite the detailed nature of the characterization, several limitations should be acknowledged. First, while SQUID magnetometry and Raman spectroscopy identified the hausmannite phase in the silicate-rich Zhaksy ore, the absolute quantification of this phase remains difficult. Unlike bulk crystalline phases that can be quantified via Rietveld refinement of X-ray diffraction data, nanocrystalline inclusions dispersed in an amorphous matrix provide signals that are difficult to convert into weight percentages without advanced calibration standards.

Second, the sintering experiments were conducted exclusively in an air atmosphere. This choice is dictated by the beneficiation route considered here: among the manganese oxides, Mn 3 O 4 is the only phase that carries a useful ferrimagnetic moment above the cryogenic temperature range, whereas its reduced counterpart MnO is antiferromagnetic with a Néel point near 120 K and offers no exploitable magnetic moment for separation. Under the reducing conditions typical of carbothermic processing the Mn 3 O 4 /MnO equilibrium shifts to substantially lower temperatures and Mn 3 O 4 is progressively reduced to MnO and to manganese-rich metallic phases (Sukhdev et al., 2020), eliminating the very phase on which the proposed magnetic-separation scheme is built. The phase evolution, the critical temperature of hausmannite nucleation and the magnetic ordering temperature reported in this work are therefore representative of the oxidative sintering route specifically; their extension to reducing or partially reducing atmospheres would change the phase assemblage at every temperature step and would require a separate experimental campaign with controlled oxygen partial pressure.

Finally, this study focused on powdered samples to ensure chemical and thermal homogeneity. However, industrial sintering and subsequent magnetic separation often involve large ore lumps or granules. The heat transfer kinetics and the degree of phase liberation in bulk aggregates may differ from the idealized laboratory conditions presented here.

Future research directions

Based on the findings and limitations of the current study, several avenues for future research are proposed. The theoretical expectations regarding the magnetic separation strategies (Section 4.5) require validation through experimental separation tests. Using high-gradient magnetic separators on sintered Zhaksy-type ores will provide empirical data on recovery rates and concentrate purity.

Further investigation is needed to model the kinetics of hausmannite nucleation within silicate matrices. Identifying the time-temperature-transformation (TTT) relationships will allow for the optimization of industrial furnace residence times, potentially reducing energy consumption.

Since hausmannite and braunite are valued not only for metallurgy but also as pigments and catalysts, future work should evaluate the functional properties of the obtained products. For instance, testing the catalytic activity of sintered Bogach ore in volatile organic compound (VOC) oxidation would expand its industrial utility.

The role of fluxes and of carefully metered reducing agents in the sintering process deserves further investigation. Such reducing agents must be dosed in amounts sufficient to lower the activation energy of phase transitions, yet small enough to keep the system on the Mn 3 O 4 side of the Mn 3 O 4 /MnO boundary. This route may lower the critical crystallization temperature of hausmannite below 1,000 °C and further improve the cost-effectiveness of the technology.

This study provides a physicochemical and magnetic characterization of the structural evolution of Central Kazakhstan manganese ores from the Bogach and Zhaksy deposits during high-temperature sintering. The integration of bulk elemental analysis, thermal profiling, and sensitive structural probes indicates the following conclusions: The sintering mechanism is governed by the initial mineralogical matrix of the ore. The carbonate-rich Bogach ore undergoes significant restructuring driven by the decomposition of calcite, resulting in a 27.5% mass loss and an endothermic enthalpy of approximately 3535 J/g. In contrast, the silicate-rich Zhaksy ore acts as a thermally stable framework with a significantly lower mass loss of 11.85% and an enthalpy of 537 J/g, where phase transitions are primarily localized within a glass-ceramic matrix.

A sintering temperature of 1,000 °C is identified as the critical threshold for the nucleation of the hausmannite (Mn 3 O 4) phase in both ore types. While bulk XRD analysis primarily identifies braunite in sintered Zhaksy ore, Raman spectroscopy and SQUID magnetometry were used for detecting nanocrystalline hausmannite inclusions that remain obscured in standard diffraction patterns due to their small coherent scattering domains.

Quantitative pixel-wise EDS analysis indicates two distinct crystallization regimes. The Bogach ore exhibits high-temperature phase segregation, leading to the formation of large, high-purity manganese oxide grains containing up to 94 wt% Mn. Conversely, the Zhaksy ore forms a complex solid solution where iron is isomorphously substituted into the manganese oxide lattice, creating an (Mn,Fe) 3 O 4 phase that remains physically encapsulated within a silicate glass matrix.

Thermomagnetic measurements indicate that the emergent ferrimagnetic ordering at low temperatures (Tc ≈ 42 K) is the characteristic feature of hausmannite formation. The Bogach ore exhibits a higher saturation magnetization of approximately 25 emu/g due to its bulk crystalline purity, while the Zhaksy ore shows a reduced response of roughly 18 emu/g, caused by the dilution effect of the silicate matrix and chemical doping by iron cations.

The two ore types studied here therefore require qualitatively different beneficiation strategies: the segregated, coarse-grained microstructure of sintered Bogach ore allows direct application of high-intensity magnetic separation, whereas the encapsulated nanocomposite microstructure of sintered Zhaksy ore precludes standard magnetic recovery and demands a combination of microstructural and equipment-level adaptations.

These findings translate into specific recommendations for industrial implementation. The sintering stage should be conducted in an oxidizing atmosphere at a temperature of at least 1,000 °C, since this is the threshold at which hausmannite nucleates in both ore types and the magnetic response of the sinter becomes useful for separation; further increase of the temperature offers no additional gain in saturation magnetization and only raises the specific energy demand. For carbonate-rich ores of the Bogach type, beneficiation is recommended in the form of standard comminution to conventional liberation sizes followed by high-intensity magnetic separation (HIMS); a pre-concentration step prior to sintering is not necessary, since the carbonate gangue is largely removed by decarbonation during the thermal stage itself. For silicate-rich ores of the Zhaksy type, two alternative routes are recommended: either upstream pre-concentration to remove quartz ballast prior to sintering, which improves the energy economy of the thermal stage, or, if the ore is sintered directly, a combination of ultra-fine comminution and high-gradient magnetic separation (HGMS) to overcome the matrix effect imposed by the glassy gangue. The use of fluxing additives capable of lowering the critical crystallization temperature of Mn 3 O 4 below 1,000 °C is identified as a further direction for reducing the energy intensity of the sintering stage and is the subject of ongoing work.

In summary, this research links the mineralogy of the precursor and the thermal history of the sinter to the magnetic functionality of the resulting product, defines the operating windows and separator types appropriate for each ore class, and provides a reproducible analytical template for the characterization of complex manganese-bearing systems.