Reduction Disintegration Behavior and Mechanism of Vanadium–Titanium Magnetite Pellets During Hydrogen-Based Reduction

Abstract

Hydrogen-based reduction followed by the electric furnace smelting of vanadium–titanium magnetite pellets offers notable advantages, including high reduction efficiency, reduced energy consumption, lower CO₂ emissions, and improved titanium recovery. However, the disintegration of pellets during the reduction process presents a major barrier to industrial application. In this study, the reduction disintegration behavior and underlying mechanisms under hydrogen-based conditions were systematically investigated. The most severe disintegration was observed at 500 °C in an atmosphere of H₂/(H₂ + CO) = 0.25, where titano–magnetite [(Fe,Ti)₃O₄] was identified as the dominant phase. The complete transformation from titano–hematite [(Fe,Ti)₂O₃] to titano–magnetite occurred within 30 min of reduction. Extended reduction (60 min) further intensified disintegration (RDI_−0.5mm = 81.75%) without the formation of metallic iron. Microstructural analysis revealed that the disintegration was primarily driven by volumetric expansion resulting from the significant increase in the titanium–iron oxide unit cell volume. Raising the reduction temperature facilitated the formation of metallic iron and suppressed disintegration. These findings provide essential guidance for optimizing reduction parameters to minimize structural degradation during the hydrogen-based reduction of vanadium–titanium magnetite pellets.

1. Introduction

Vanadium–titanium magnetite is a critical strategic mineral resource. It is a primary source of iron and contains valuable co-associated elements (V, Ti, Cr, Co, Ni), offering great potential for comprehensive utilization [1,2,3]. Currently, the comprehensive utilization of vanadium–titanium magnetite mainly includes the blast furnace (BF) process and the non-blast furnace process. Although the prevailing blast furnace smelting process effectively recovered iron and vanadium, the resultant slag retains 20–25% TiO₂ [4,5,6,7,8]. The occurrence forms of the titanium compounds hinder efficient titanium recovery and utilization, leading to substantial resource waste [9,10].

In contrast, hydrogen-based reduction followed by electric furnace smelting offers several advantages. These include the enhanced iron oxide reduction of iron oxides, lower energy consumption, reduced CO₂ emissions, and better titanium recovery [11,12,13,14]. However, at low temperatures (400–700 °C), serious disintegration occurs during hydrogen-based reduction. This reduces shaft furnace permeability and significantly hinders stable production [15,16,17]. To date, the hydrogen-based reduction behavior of vanadium–titanium magnetite pellets has been predominantly studied in the medium- to high-temperature regime (>800 °C), with a focus on metallization kinetics and reduction-induced swelling. Systematic studies on low-temperature disintegration mechanisms remain limited and are mostly based on Ti-free conventional iron ore pellets [18,19,20,21].

Yi et al. investigated the disintegration behavior and mechanisms of conventional iron ore pellets under simulated shaft furnace conditions [22]. When pellets were reduced at 550 °C using a gas mixture of 31.1% H₂, 38.9% CO, 22.0% CO₂, and 8.0% N₂ (H₂/(CO + H₂) = 0.44), severe pellet disintegration occurred. Gangue minerals within the pellets were found to promote disintegration to varying extents. Hu et al. demonstrated that vanadium–titanium magnetite pellets exhibit the most severe disintegration under conditions of 11.4% H₂, 45.6% CO, 35.0% CO₂, and 8.0% N₂ (H₂/(CO + H₂) = 0.20) at 550 °C [23]. Through thermodynamic modeling, Chen et al. clarified the reduction pathways of the main phases in vanadium–titanium magnetite under H₂-CO atmospheres. Specifically, titano–hematite (Fe₂O₃·TiO₂) undergoes reduction via (Fe,Ti)₂O₃→(Fe,Ti)₃O₄→FeO→Fe, while pseudobrookite (Fe₂TiO₅) follows Fe₂TiO₅→Fe₂TiO₄→FeTiO₃→FeTi₂O₅→TiO₂→Ti₃O₅ [24]. Compared to BF conditions, vanadium–titanium magnetite pellets in hydrogen-based atmospheres (e.g., Midrex and HYL process specifications) exhibit 40–60% higher disintegration indices due to accelerated carbon deposition and phase transformation stresses [25,26,27].

Midrex and HYL processes are currently the two dominant hydrogen-based shaft furnace reduction technologies, imposing strict standardized parameters (e.g., reducing atmosphere composition, and temperature profiles) that limit investigations into the hydrogen-based reduction disintegration mechanisms of pellet formation [28,29]. Critically, prior studies on pellet disintegration have predominantly been conducted in static shaft furnaces. However, in actual hydrogen-based shaft furnace operations, pellets are in continuous motion, leading to reduction disintegration behaviors that differ from those observed under static conditions [30,31]. In this study, the dynamic reduction process of vanadium–titanium magnetite pellets was investigated using a pure reducing gas mixture of H₂ and CO. This approach facilitates a more accurate elucidation of the reduction disintegration behavior and underlying mechanisms of the pellets.

This study systematically investigated the disintegration behavior of vanadium–titanium magnetite pellets under hydrogen-based reduction. The factors influencing low-temperature disintegration were analyzed, and the relationship between reduction degree and disintegration was clarified. The reduction disintegration process was clarified, and the microstructural changes and disintegration mechanisms of vanadium–titanium magnetite pellets during hydrogen-based reduction were elucidated. These findings contribute to the development of methods aimed at suppressing reduction swelling and disintegration of vanadium–titanium magnetite pellets.

2. Materials and Methods

2.1. Materials

2.1.1. Vanadium–Titanium Magnetite Concentrate

The raw material used in this study was vanadium–titanium magnetite concentrate obtained from the Baima mining area in Panzhihua, Sichuan, China. The composition of the vanadium–titanium magnetite concentrate is presented in

Table 1. Chemical composition of vanadium–titanium magnetite concentrate.

Chemical Composition	TFe	FeO	TiO2	V2O5	CaO	MgO	Al2O3	SiO2	S
Content %	57.03	34.61	10.12	3.78	2.34	1.93	3.36	2.76	0.35

. The concentrate has a relatively high level of impurities, including aluminum, magnesium, calcium, and silicon.

The XRD analysis of the concentrate is shown in Figure 1, which reveals that the main mineral phases in the vanadium–titanium magnetite concentrate were titano–magnetite [(Fe,Ti)₃O₄] and ilmenite (FeTiO₃). The characteristic diffraction peaks for titano–magnetite appeared at approximately 2θ = 30.1°, 35.4°, and 62.6°, corresponding to the (1422), (4425), and (919) crystal planes, respectively. The phase identification was based on standard reference patterns obtained from the ICDD PDF-4+ database for iron-containing compounds.

2.1.2. Binder

The binder used in this study was bentonite, whose primary mineral component was montmorillonite. The main chemical constituents of bentonite are SiO₂ and Al₂O₃. The chemical composition of the bentonite used in this experiment is shown in

Table 2. Main chemical composition of bentonite.

Chemical Composition	SiO2	Al2O3	Na2O	CaO	Fe2O3	MgO	K2O
Content/%	62.23	12.91	1.66	4.68	2.93	3.09	1.09

.

2.2. Experimental Methods

2.2.1. Reduction Disintegration Experiment

The experimental methodology followed a previously published open-access protocol (Chen et al., 2024 CC-BY 4.0) [24]. The methodology employed in this study involved the following steps: vanadium–titanium magnetite concentrate was mixed with water and a binder and then pelletized using a disk pelletizer. After drying, the green pellets were subjected to oxidative preheating at 950 °C for 20 min and oxidative roasting at 1220 °C for 20 min in an electrically heated horizontal tube furnace under an ambient air atmosphere. The reduction disintegration index (RDI) of the pellets was tested in a rotary tube according to ISO 11257:2022 standards [32]. Approximately 250 g of vanadium–titanium magnetite oxidized pellets were placed into the rotary tube, which was connected to the gas supply and exhaust system before being inserted into the furnace. The experimental setup is shown in Figure 2. The furnace was heated at 10 °C/min, and the rotary tube was rotated at 10 r/min. Nitrogen (N₂) was introduced for protection, then replaced with a reducing gas once the reduction temperature was reached. After reduction, the rotation and gas supply were stopped, and nitrogen was reintroduced for cooling. The reduced pellets and powders were sieved using 6.3 mm, 3.15 mm, and 0.5 mm meshes to determine the reduction disintegration index (RDI), calculated by Equations (1)–(3).

$$\mathrm{RD}{\mathrm{I}}_{+6.3\mathrm{mm}}={\displaystyle \frac{m_{+6.3\mathrm{mm}}}{m_0}}\times 100\%$$

(1)

$$\mathrm{RD}{\mathrm{I}}_{+3.15\mathrm{mm}}={\displaystyle \frac{m_{+3.15\mathrm{mm}}}{m_0}}\times 100\%$$

(2)

$$\mathrm{RD}{\mathrm{I}}_{-0.5\mathrm{mm}}={\displaystyle \frac{m_{-0.5\mathrm{mm}}}{m_0}}\times 100\%$$

(3)

where m₀—the mass of vanadium–titanium magnetite pellets before dynamic reduction, m_+6.3mm—the mass of specimens greater than 6.3 mm after dynamic reduction, m_+3.15mm—the mass of specimens greater than 3.15 mm after dynamic reduction, m_−0.5mm—the mass of specimens less than 0.5 mm after dynamic reduction.

2.2.2. Analytical Methods

(1)

X-ray Diffraction Analyzer (XRD)

X-ray diffraction analyzer (XRD) was employed to determine the phase compositions of vanadium–titanium magnetite concentrate, oxidized pellets, and reduced pellets. The scanning angle range was 0–90° (2θ), with a scanning speed of 2°/min. Analyses were conducted at 25 °C using the JADE 9.0 and GSAS-II software (version 5714, Los Alamos National Laboratory, Los Alamos, NM, USA). The standard reference patterns were obtained from the ICDD PDF-4+ database for iron-containing compounds.

(2)

Scanning Electron Microscope (SEM)

The microstructures of vanadium–titanium magnetite concentrate, oxidized pellets, and reduced pellets were analyzed using backscattered electron imaging on a TESCAN MIRA3 field emission scanning electron microscope (TESCAN, Brno, Czech Republic). Additionally, the chemical composition of these materials was determined using an Oxford X-max 20 EDS spectrometer (Oxford Instruments, Abingdon, UK).

3. Results and Discussion

3.1. Reduction Disintegration Behavior

3.1.1. Effect of Reduction Atmosphere

In practical shaft furnace operations, the reduction atmosphere may contain CO, H₂, CO₂, and CH₄. However, in this study, only CO and H₂—identified as the effective reducing gases—were used to simulate the reduction environment, while CH₄ was excluded due to its endothermic decomposition and limited contribution to reduction under the tested conditions. In this section, the effect of the CO and H₂ ratio in the reduction atmosphere on the reduction disintegration behavior of vanadium–titanium magnetite pellets was investigated. The experiments were conducted at a reduction temperature of 500 °C for 60 min, and the results are presented in Figure 3. As shown in Figure 3a, the reduction disintegration index RDI_−0.5mm of vanadium–titanium magnetite pellets initially increased and subsequently decreased with increasing H₂ proportion in the reducing gas mixture. Excellent low-temperature reduction disintegration resistance (RDI_−0.5mm < 2%) was observed under both pure H₂ and pure CO atmospheres. However, severe disintegration occurred when H₂ and CO were mixed at any ratio. The RDI_+6.3mm index abruptly decreased from 97.51% to 0.51% as the H₂/(H₂ + CO) ratio was increased from 0 to 0.25, and over 80% of pellets were reduced to fine powders (−0.5 mm), leaving only minor fragmented residues (as shown in Figure 3b). Remarkably, increased H₂ proportion significantly mitigated disintegration severity. With the elevation of the H₂/(H₂ + CO) ratio to 0.75, the RDI_+6.3mm index recovered to 92.63%, indicating the near-complete elimination of disintegration. Under pure H₂ conditions (H₂/(H₂ + CO) = 1), intact pellets were retained with minimal fine powder generation (<5 vol%).

3.1.2. Effect of Reduction Temperature

The influence of reduction temperature on the disintegration behavior of vanadium–titanium magnetite pellets was systematically investigated under fixed conditions (H₂/(H₂ + CO) = 0.5, reduction time: 60 min), with the experimental results presented in Figure 4. A significant improvement in disintegration resistance was observed when the reduction temperature exceeded 500 °C. At 700 °C, the RDI_+6.3mm index reached 98.15%, indicating the complete suppression of disintegration. An analysis of the temperature-dependent disintegration curve revealed that volumetric expansion-induced disintegration predominantly occurred within the 450–550 °C range. This critical temperature interval corresponds to the preheating zone temperatures of industrial shaft furnaces, strongly suggesting that the gas-based reduction disintegration of vanadium–titanium magnetite pellets originates during the preheating stage of shaft furnace progress.

As shown in Figure 4b, the reduction products of vanadium–titanium magnetite pellets from the rotary tube reactor consisted of residual pellets and detached fines. Across all tested temperatures, the residual pellets maintained regular spherical shapes with minimal fracturing. Notably, at reduction temperatures exceeding 700 °C, only trace amounts of fines (<2 wt.%) were observed, while the remaining pellets retained structural integrity. This morphological stability ensured no operational interference in industrial gas-based reduction processes.

3.1.3. Effect of Reduction Time

The influence of reduction time on the disintegration behavior of vanadium–titanium magnetite pellets was investigated under controlled conditions (H₂/(H₂ + CO) = 0.5, reduction temperature: 500 °C), with the experimental results shown in Figure 5. A prolonged reduction duration progressively decreased the RDI_+6.3mm and RDI_+3.15mm indices while increasing RDI_−0.5mm. Minimal disintegration occurred within the initial 30 min of reduction. However, extending the reduction time beyond this threshold significantly intensified the disintegration phenomena. This behavior was attributed to the crystallographic phase transformation of titano–hematite [(Fe,Ti)₂O₃] to titano–magnetite [(Fe,Ti)₃O₄] after 30 min at 500 °C. The accompanying volumetric expansion disrupted interparticle bonding networks, thereby accelerating pellet disintegration. Notably, the RDI_+6.3mm value at 30 min (85.86%) approached the operational limit for stable shaft furnace performance. Beyond this critical duration, the substantial powder generation (>25 vol%) would render hydrogen-based reduction processes inoperable due to gas channel blockage and permeability collapse.

As shown in the morphological images (Figure 5b), disintegration was initiated by surface exfoliation of fines. After 15 min of reduction, only a small number of fines were detached from the pellet surfaces, with pellets retaining regular spherical morphology and relatively intact surfaces. Shortening the preheating duration in shaft furnaces was suggested to mitigate low-temperature reduction disintegration. After 60 min of reduction, disintegration was intensified: pellet quantity was significantly reduced, surfaces became roughened, and substantial fines were generated.

3.2. Phase Transformation

The phase transformations of vanadium–titanium magnetite pellets during reduction were investigated using XRD, with the results shown in Figure 6 and Figure 7. Figure 6 displayed the XRD pattern of oxidized vanadium–titanium magnetite pellets, where titano–hematite [(Fe,Ti)₂O₃] and pseudobrookite (Fe₂TiO₅) were identified as the dominant phases. The phase identification was performed based on standard reference patterns from the ICDD PDF-4+ database for iron-containing compounds.

Figure 7a illustrates the phase composition of pellets under different reduction atmospheres (500 °C, 60 min). It was observed that the phase compositions remained similar across atmospheres, predominantly consisting of titano–magnetite [(Fe,Ti)₃O₄] with minor metallic iron (Fe). Under these conditions, the reduction of titano–hematite to titano-magnetite was completed. Figure 7b shows the XRD patterns of vanadium–titanium magnetite pellets reduced at different temperatures (H₂/(H₂ + CO) = 0.5, 60 min). At 400 °C, the primary phases were identified as titano–magnetite [(Fe,Ti)₃O₄] and minor titano–hematite [(Fe,Ti)₂O₃]. When the temperature was increased to 500 °C, titano–magnetite [(Fe,Ti)₃O₄] remained predominant in the reduced powder, accompanied by weak metallic iron (Fe) diffraction peaks. At 800 °C, the phase composition was dominated by metallic iron (Fe), wüstite (FeO), and residual titano–magnetite [(Fe,Ti)₃O₄]. Figure 7c shows the phase composition of pellets under different reduction atmospheres (500 °C, H₂/(H₂ + CO) = 0.5). With prolonged reduction time, the diffraction peak intensity of titano–magnetite [(Fe,Ti)₃O₄] was enhanced, while that of titano–hematite [(Fe,Ti)₂O₃] was weakened. At 10 min of reduction, the pellet was dominated by titano–hematite, with minor pseudobrookite (Fe₂TiO₅) and titano–magnetite. After 30 min of reduction, titano–hematite was almost completely converted to titano–magnetite, which became the main phase. Weak metallic iron (Fe) diffraction peaks were detected, and titano–hematite peaks disappeared. At 60 min of reduction, no significant phase composition changes were observed compared to the 30 min sample. The titano–magnetite diffraction peaks were enhanced, while metallic iron peaks remained weak. The reduction of pseudobrookite (Fe₂TiO₅) was found to proceed simultaneously with the reduction of titano–hematite.

The variation in reduction disintegration rates under different reduction regimes, combined with the XRD results, indicated that the most severe disintegration of vanadium–titanium magnetite pellets occurred in the temperature range of 400–500 °C. At this stage, the main phase was identified as titano–magnetite [(Fe,Ti)₃O₄]. After 30 min of reduction, titano–hematite [(Fe,Ti)₂O₃] was completely converted to titano–magnetite. Prolonging the reduction time from 30 to 60 min at 500 °C did not result in the further transformation of titano–magnetite to metallic iron, and disintegration was exacerbated. When the reduction temperature was increased, titano–magnetite was transformed into metallic iron, resulting in a decrease in low-temperature reduction disintegration rates. These findings demonstrated that the disintegration process occurred during the transformation of titano–hematite to titano–magnetite. The formation of metallic iron was clearly not the primary cause of pellet disintegration.

3.3. Microstructural Evolution

This section examines the microstructural evolution of vanadium–titanium magnetite pellets during hydrogen-based reduction stages using scanning electron microscopy (SEM) under controlled conditions (H₂/(H₂ + CO) = 0.5, reduction temperature: 500 °C). Figure 8 shows the SEM–EDS analysis of the pellets after 10 min of reduction. No significant macroscopic cracks were observed at this stage, but microcracks had already formed within titanomagnetite particles.

Cracks first appeared at the interfaces between low-Ti iron oxides (Point 1: Ti = 3.6 wt.%) and high-Ti iron oxides (Point 2: Ti = 17.3 wt.%). Point 1 corresponded to titano–magnetite with low Ti content, while Point 2 represented pseudobrookite (Fe₂TiO₅) with high Ti content. The low-Ti titano–magnetite and high-Ti pseudobrookite formed a tightly intergrown structure, where pseudobrookite exhibited slower reduction kinetics compared to titano–magnetite. The crack formation mechanism is attributed to two factors:

• Volume expansion, which was caused by crystallographic phase transformations during the reduction of iron–titanium oxides, which generated internal stress.
• Stress amplification, which was due to the differential reduction rates between low-Ti iron oxides and high-Ti iron oxides.

The particles within the pellet were consolidated by hematite microcrystals and recrystallized bonds. However, due to the formation of a tight interwoven structure between titano–hematite and high-iron pseudobrookite, the hematite microcrystals and recrystallized bonds at the interface between these two intertwined phases were relatively weak. As a result, cracks tended to initiate at their interface. Notably, most of the iron–titanium oxides in the pellets remained as unreduced titano-hematite, with only a small fraction transformed into titano-magnetite at this reduction stage.

Figure 9 displays the microstructure and SEM–EDS analysis of vanadium–titanium magnetite pellets after 30 min of reduction. Significant crack propagation was observed both intergranularly and intragranularly. These cracks were further extended by prolonged reduction time, resulting in reduced microstructural compactness and exacerbated pellet disintegration.

A partial reduction of titano–magnetite to wüstite (Point 3: FeO) was identified through the EDS analysis of crack-adjacent particles, with microcracks retained within the wüstite phase. At this stage, titano–magnetite remained as the dominant phase, while residual titano–hematite was minimal. A transitional phase (Fe₂TiO₄) from pseudobrookite to ilmenite was detected at Point 4, exhibiting 21.3 wt.% Ti. The breakdown of vanadium–titanium magnetite microcrystalline bonds was intensified by accelerated gas diffusion through large cracks, leading to a rapid increase in disintegration rates.

Figure 10 displays the microstructure and SEM–EDS analysis of vanadium–titanium magnetite pellets after 60 min of reduction. Large penetrating cracks were formed due to volumetric expansion caused by the phase transformation of titano–hematite (Fe₂O₃·TiO₂) to titano–magnetite [(Fe,Ti)₃O₄], which led to a more dispersed distribution of titanomagnetite particles. At this stage, a significant increase in internal crack density and fine particle fraction was observed, indicating exacerbated disintegration behavior. The pellet composition was still dominated by titano–magnetite (Point 5: Ti = 5.5 wt.%) and ilmenite (Point 6: Ti = 20.0 wt.%). This phase distribution confirmed that under low-temperature reduction conditions (60 min), titano–hematite could only be partially reduced to titano–magnetite, while pseudobrookite (Fe₂TiO₅) was incompletely reduced to ilmenite.

3.4. Reduction Disintegration Mechanism

Based on the studies presented in Section 3.2 and Section 3.3, it was demonstrated that during the low-temperature hydrogen-based reduction of vanadium–titanium magnetite pellets, titano–hematite could only be reduced to titano–magnetite, while pseudobrookite was reduced to ilmenite. Significant pellet disintegration was observed during this reduction stage. When the reduction temperature was increased, the pellets were further reduced to metallic iron, resulting in structural re-densification that effectively prevented disintegration. The hydrogen-based reduction process of vanadium–titanium magnetite pellets was found to proceed through several distinct stages, as illustrated in Figure 11.

In vanadium–titanium magnetite oxidized pellets, titano–hematite and titano–magnetite were observed to form an intimately interwoven structure. During low-temperature hydrogen-based reduction, the phase transformations accompanying the reduction of iron–titanium oxides were found to generate internal stresses, which led to structural damage and subsequent pellet disintegration. In contrast, during high-temperature reduction, magnetite was successfully reduced to metallic iron through the following process:

• The reduction process was initially observed to occur at the pellet surface, where expansion stresses from phase transformation caused structural failure. Cracks were generated, and mineral particles were exfoliated from the surface. Meanwhile, the pellet core remained unreacted, with titano–hematite preserved as the main component, maintaining structural integrity.
• At low reduction temperatures (400–600 °C), crack formation on the pellet surface was found to accelerate reducing gas diffusion. The core region was subsequently reduced, resulting in complete surface disintegration and exposure of the original core material, which developed extensive cracking.
• With continued low-temperature reduction, complete pellet disintegration was ultimately achieved.
• When the reduction temperature exceeded 800 °C, surface titano–magnetite was initially reduced to metallic iron. The resulting fine metallic iron particles formed interparticle connections that effectively accommodated phase transformation stresses. As a result, interparticle cracking was minimized, and the pellet structure became re-densified, thereby significantly suppressing reduction disintegration. However, unreduced titano–magnetite still remained in the pellet core at this stage.
• Prolonged high-temperature reduction ultimately produced structurally dense metalized pellets that maintained complete integrity without disintegration.

4. Conclusions

The reduction behavior of vanadium–titanium magnetite pellets under varying reduction temperatures, durations, and atmospheres was systematically investigated. The relationship between the reduction degree and disintegration behavior during hydrogen-based reduction was established within the temperature range of 400–800 °C. The microstructural evolution and disintegration mechanisms of the pellets under hydrogen-based reduction were elucidated. The most severe disintegration was observed in the temperature range of 400–500 °C. At this stage, the main phase composition was identified as titano–magnetite. After 30 min of reduction at these temperatures, titano–hematite was completely converted to titano–magnetite. Prolonged reduction time was found to intensify disintegration without further transformation of titano–magnetite to metallic iron. When the reduction temperature was increased, titano–magnetite was transformed into metallic iron, resulting in reduced low-temperature disintegration rates. The volumetric expansion of pellets was attributed to the significant increase in the total unit cell volume of iron oxides during reduction. Internal stress was exacerbated by mismatched unit cell volume changes between iron oxides and iron–titanium oxides during reduction, leading to structural damage. The disintegration process was found to align with the reduction progression, initiating from the pellet surface (where reduction first occurred) and propagating inward as iron oxide particles were progressively detached, ultimately resulting in complete disintegration.