Research on hydrogen-based direct reduction sticking and inhibition methods for vanadium–titanium pellets

Experimental materials

Vanadium–titanium pellets supplied by a Chinese steel enterprise were used in this study. The chemical composition of the pellets was determined, and their phase composition was characterized by XRD. The results are presented in Table 1 and Figure 1. The main phases in the vanadium–titanium pellets were identified as FeTiO₃ (ilmenite), MgTiO₃ (magnesium titanate), and Fe₂O₃ (iron oxide).

OPEN IN VIEWER

Figure 1.

X-ray diffraction (XRD) pattern of vanadium–titanium pellet raw materials.

OPEN IN VIEWER

Table 1.

Chemical composition of vanadium–titanium pellets (wt.%).

TFe	FeO	V2O5	TiO2	Fe2O3
55.06	3.05	0.62	9.39	75.27

Experimental design and methods

To investigate the sticking behavior of vanadium–titanium pellets under the complex atmosphere of a hydrogen-based shaft furnace, the effects of reduction temperature (950, 1000, 1050, 1100 °C), load-bearing time (1, 1.5, 2, 2.5 h), and coating materials (MgO, TiO₂, CaO, none) on the sticking behavior and inhibition performance of vanadium–titanium pellets were systematically studied. Pressurization was initiated 1.0 h after the start of reduction. The specific experimental conditions are listed in Table 2. The total reducing gas flow rate was fixed at 6 L/min, and the gas composition is shown in Table 3.

OPEN IN VIEWER

Table 2.

Experimental scheme for the pellet reduction-under-load test.

Operating condition	No.	Load-bearing time (h)	Reduction time (h)	Temperature (°C)	Coating material
A	1	1	2	950	None
	2	1	2	1000	None
	3	1	2	1050	None
	4	1	2	1100	None
B	5	1	2	1100	None
	6	1.5	2.5	1100	None
	7	2	3	1100	None
	8	2.5	3.5	1100	None
C	9	1	2	1100	None
	10	1	2	1100	MgO
	11	1	2	1100	TiO2
	12	1	2	1100	CaO

OPEN IN VIEWER

Table 3.

Composition of the reducing gas.

Gas composition	H2	CO	CH4	H2O	CO2	N2
Content（%）	60	25	5	1	3	6

The selected gas composition was designed to approximate the gas environment in the hydrogen-rich reduction zone of a shaft furnace. H₂ was used as the dominant reducing gas, whereas CO and CO₂ were introduced to reflect the influence of recycled top gas and the coexistence of hydrogen and carbon-bearing species in practical operation. CH₄ was included to represent the possible presence of residual methane in reformed gas and to evaluate its influence on carburization under high-temperature conditions. H₂O was added to simulate the water vapor generated during reduction, and N₂ was used as the balance gas.

The experimental samples consisted of pellets with a particle size of 10–12.5 mm. Prior to testing, the pellets were dried at 105 ± 5 °C for 4 h, and the mass of each sample was controlled at 300 ± 0.30 g. Under operating condition C, the pellets were further treated with a 0.5% coating material by spraying, as shown in Figure 2. The specific process involves mixing the coating powder with deionized water and spraying it onto the particle surface during continuous rolling to achieve the most uniform surface coverage possible. The coating dosage of 0.5 wt.% was selected to ensure effective surface coverage while minimizing changes in bed structure and pellet geometry. Because the coating was deposited by spraying, some local variation in coating thickness was unavoidable; therefore, the inhibition behavior discussed below reflects the combined effect of coating chemistry and practical coating coverage. During the experiment, the samples were heated under a nitrogen atmosphere at a flow rate of 6 L/min, with a heating rate of 10 °C/min to the target temperature, followed by a holding period of 30 min. Subsequently, the reducing gas was introduced for a reaction time of 120 min. A load pressure of 1.4 kg/cm² was applied after 60 min to simulate the static pressure exerted on pellets within the burden bed. The charging structure adopted an alumina ball-pellet-alumina ball arrangement, with alumina balls of 10–12.5 mm in diameter, as shown in Figure 3. After completion of the reaction, the system was purged with nitrogen. The samples were removed and weighed after cooling to below 100 °C.

OPEN IN VIEWER

Figure 2.

Macroscopic morphology of pellets before and after coating (a)uncoated pellets, (b) TiO₂-coated pellets.

OPEN IN VIEWER

Figure 3.

Schematic diagram of the temperature curve and load of the pellet reduction-under-load test (a) heating system, (b) load diagram.

The experimental apparatus is shown in Figure 4. This integrated system combines steam generation, gas mixing, pressurization, and off-gas treatment into a complete reduction reaction platform. The heating system consists of a high-temperature electric furnace equipped with silicon molybdenum rods as heating elements, with a maximum operating temperature of 1600 °C. The steam generation system uses a peristaltic pump to regulate the flow of deionized water. After evaporation, the steam is transported by the other reducing gases acting as carrier gases. The steam is continuously heated and delivered through stainless steel piping to the storage tank to prevent condensation. The pressure control system employs an air compressor to provide a stable pressure of 1.4 kg/cm², thereby accurately simulating the static pressure environment experienced by the burden layer in an actual shaft furnace.

OPEN IN VIEWER

Figure 4.

Schematic diagram of experimental device for pellet reduction-under-load test.

Analysis and characterization

The phase composition and morphology of the vanadium–titanium pellets and their reduction products were characterized using an X-ray diffractometer (XRD, X’Pert Pro MPD), a scanning electron microscope (SEM, Nova400 NanoSEM), and an energy-dispersive spectrometer (EDS, IE350 pentaFET X-3).

The metallization rate was determined in accordance with the Chinese national standard “Determination of Reduction Index, Final Reduction Degree, and Metallization Rate of Iron Ore for Direct Reduction Furnaces” (GB/T 24236-2009). The formula for calculating the pellet metallization rate (1) is as follows.

MR = MFe TFe × 100 %

In the formula, TFe is the total iron content in the sample, and MFe is the metallic iron content in the sample.

Since the carbide content in the pellets cannot be measured directly, the total carbon content and free carbon content in the sample are determined through component testing. The carbide content (ωFe₃C) after sample reduction and carburization is calculated using the following formula (2)

ω Fe 3 C ( % ) = ( C 0 - C 1 12 × 180 ) × 100 %

In the equation, C₀ represents the total carbon content in the sample, and C₁ represents the free carbon content in the sample.

The SI is an important indicator for evaluating the sticking behavior during the reduction process of pelletized ore. A clump of particles containing two or more particles is placed inside a drum (diameter 130 mm, length 200 mm, 30 rpm). Five rotations of the drum are counted as one drum cycle, with a total of three drum cycles performed. The percentage of remaining clusters after each drum rotation is calculated, and the drum rotation count is plotted against the percentage of clusters. The SI is calculated using formula (3).

SI = S 1 ( S 1 + S 2 ) × 100 %

In the formula, S₁ is the sum of the masses of the agglomerated pellets, and S₂ is the sum of the masses of the dispersed pellets.

The sticking strength was tested using an SH-II-500 digital push-pull tester. Two adhered particles (five sets) were selected and placed on the tensile platform, which moved at a constant speed until the bonded particles completely separated, at which point the test was stopped. This process was repeated five times. The sticking strength was calculated using formula (4).

Q = F 1 + F 2 + F 3 + F 4 + F 5 5 × 100 %

In the formula, F₁, F₂, F₃, F₄, and F₅ are the tensile forces required to stretch the pellets into a cluster until they disperse.

For sticking strength, five adhered pellet pairs were measured under each condition and the average value was reported. The SI, metallization rate, and phase/microstructural characterization were evaluated under the same experimental procedure for all operating conditions, and the conclusions were drawn from the consistency between the macroscopic indices and the corresponding characterization results. Because the differences between some coating conditions were small, the ranking of coating performance was interpreted together with the microstructural evidence rather than on the basis of a single index alone.

Effect of reduction temperature on sticking behavior

Under a reducing gas atmosphere consisting of 60% H₂, 25% CO, 5% CH₄, 1% H₂O, 3% CO₂, and 6% N₂, the effect of reduction temperature on the sticking behavior of pellets after load reduction was systematically investigated in the range of 950–1100 °C at intervals of 50 °C. The experimental parameters were fixed such that a load was applied after 1.0 h of gas reduction, and the reaction was terminated after the load had been maintained for 1.0 h.

As shown in Figure 5(a), reveals the regulatory mechanism of temperature on the SI. As the reduction temperature increased from 950 to 1100 °C, the SI showed a significant positive correlation. No obvious sticking was observed at 950 °C. The SI then increased to 9.49% at 1000 °C, further rose to 11.14% at 1050 °C, and finally exhibited a sharp increase to 32.11% at 1100 °C. The sticking strength showed a similar upward trend with increasing temperature. When the temperature increased from 1000 to 1050 °C, the sticking strength increased by 0.44 N, and at 1100 °C it reached 13.63 N. This behavior can be attributed mainly to the enhancement of interfacial diffusion and solid-state sintering of metallic iron at elevated temperatures. In addition, the sharp increase in the SI at 1100 °C suggests that localized low-melting compounds may also contribute under the present gas atmosphere, although no direct in situ evidence of liquid-phase formation was obtained in this work. Therefore, the effect of possible liquid phases is discussed here as a thermodynamically reasonable contribution rather than as a directly verified dominant mechanism. From the viewpoint of phase equilibrium, the multicomponent FeO–TiO₂–SiO₂–CaO–MgO-containing system may form locally softened phases in an iron-rich reducing environment, which would facilitate interparticle bonding once severe surface contact has already been established. By contrast, Figure 5(b) shows that the metallization rate remained highly stable over the temperature range of 950–1100 °C, varying only between 95.34% and 95.64%.

OPEN IN VIEWER

Figure 5.

Experimental results of the reduction-under-load test at different temperatures for 2 h: (a) sticking index and sticking strength, (b) metallization rate.

XRD analysis of the sticking interfaces revealed the dynamic evolution of phase composition at different reduction temperatures, as shown in Figure 6. In the lower temperature range of 950–1050 °C, the sticking interface was dominated by metallic iron (Fe), and the characteristic diffraction peak of the Fe (110) crystal plane consistently appeared at 44.674° (2θ, Fe, PDF#06-0696). The gradual narrowing of the diffraction peak with increasing temperature indicates progressive growth in metallic iron grain size. Through surface diffusion and solid-state sintering, metallic iron formed a continuous metallurgical bonding interface at the contact points between pellets. Its oriented crystallization promoted grain growth along preferred directions into fibrous or layered structures, thereby significantly enhancing both the mechanical interlocking and chemical bonding strength between particles. Meanwhile, slag-phase components such as SiO₂ and Al₂O₃ retained their solid bridging effect because of their high melting points. Their synergistic interaction with metallic iron resulted in a steady increase in sticking strength with increasing temperature. When the temperature reached 1100 °C, however, the phase composition changed markedly. A distinct carbide peak corresponding to Fe₃C, (PDF#89-2722) appeared for the first time at 43.752° in the XRD pattern. This result indicates activation of the carburization reaction under high-temperature conditions. As the temperature increased, the cracking of CH₄ was significantly intensified, resulting in localized enrichment of carbon atoms at grain boundaries. At the same time, the activation energy for carbon diffusion was reduced, thereby promoting rapid migration of carbon along iron grain boundaries and its reaction with metallic iron to form carbides.

OPEN IN VIEWER

Figure 6.

Pellet sticking interfaces at different reduction temperatures: (a) 10°–90° of 2θ, (b) 40° to 50°of 2θ.

Effect of total reduction time and load-bearing time on sticking behavior

Under a reducing atmosphere of 60% H₂, 25% CO, 5% CH₄, 1% H₂O, 3% CO₂, and 6% N₂ at 1100 °C, the combined effect of total reduction time and load-bearing time on the sticking behavior of pellets during reduction under load was investigated over the range of 2.0–3.5 h at intervals of 0.5 h. In this set of experiments, a load was applied after 1 h of gas reduction, corresponding to total reduction times of 2.0–3.5 h and load-bearing times of 1.0–2.5 h.

Figure 7(a) shows the evolution of the SI and sticking strength with reduction time during the reduction-under-load test. Both parameters exhibited a pronounced trend of first increasing and then decreasing. At a reduction time of 2.0 h, the SI was 32.11%. It then increased sharply and reached a maximum of 97.11% at 2.5 h, after which it dropped abruptly to 16.83% at 3.0 h and further declined to 13.06% at 3.5 h. The sticking strength varied synchronously, reaching a maximum value of 16.23 N at 2.5 h and decreasing to a minimum of 4.47 N at 3.5 h. Figure 7(b) further shows the relationship between reduction time and metallization rate. As the reduction time increased from 2.0 to 3.5 h, the metallization rate exhibited a continuous downward trend, decreasing from 95.56% at 2.0 h to 92.85% at 2.5 h, 91.73% at 3.0 h, and finally 90.08% at 3.5 h. The decrease in metallization rate with increasing total reduction time cannot be attributed to Fe₃C formation alone. XRD results show that prolonged reduction also promoted the formation of FeTiO₃, which consumed part of the metallic iron. Meanwhile, as sticking became severe, the contact area between pellets increased, and the pore structure at the interface became denser, which could hinder local gas transport. Under such conditions, the coexistence of H₂O and CO₂ in the reducing gas may promote partial reoxidation of freshly reduced iron in local regions. Therefore, the decline in metallization rate should be understood as the result of the combined effects of carburization, Fe–Ti–O phase formation, and possible local mass-transfer/reoxidation phenomena.

OPEN IN VIEWER

Figure 7.

Experimental results of the reduction-under-load test at different total reduction times under the present loading schedule at 1100 °C: (a) sticking index and sticking strength, (b) metallization rate.

To further clarify the changes in metallization rate, the phase composition of the reduction products obtained at different reduction times was analyzed. As shown in Figure 8(a), when the reduction time was 2.0 h, the main phase at the sticking interface was metallic iron (Fe), with its characteristic diffraction peak located at 44.674° (2θ, Fe, PDF#06-0696). When the reduction time was extended to 2.5 h, a characteristic peak of Fe₃C (Fe₃C, PDF#89-2722) appeared at 43.752°, indicating that the free carbon generated by CH₄ cracking reacted with metallic iron to form carbides. When the reduction time increased to 3.0 h, the characteristic peak of FeTiO₃ (FeTiO₃, PDF#01-075-1212) was detected at 32.588°. With further extension of the reduction time to 3.5 h, the intensities of the Fe₃C and FeTiO₃ peaks increased significantly, whereas the intensity of the metallic iron peak decreased systematically. These co-evolution characteristics indicate that prolonged reduction promotes the formation of Fe₃C and FeTiO₃. Both phases consumed part of the metallic iron and thereby contributed to the continuous decline in metallization rate. Specifically, Fe₃C formation was associated with the sustained cracking-carburization of CH₄ cracking-carburization reaction, while the appearance of FeTiO₃ is related to secondary vanadium–titanium in iron–vanadium–titanium compounds during prolonged reduction. This dynamic evolution of phase composition is mutually corroborated with the trend of decreasing metallization rate (Figure 8(b)) at the mechanistic level, revealing the fundamental mechanism by which reduction time influences pellet sticking performance by regulating the balance between the carbonization reaction and the secondary vanadium–titanium process.

OPEN IN VIEWER

Figure 8.

X-ray diffraction (XRD) results of pellet sticking surfaces at different reaction times: (a) 10°–90° of 2θ, (b) 30° to 50°of 2θ.

To further elucidate the microstructural evolution of the bonded interface at different reduction times, SEM-EDS analysis was performed on the reduction products obtained at 1100 °C without coating. The results are shown in Figure 9, where the bond strength exhibits a trend of first increasing and then decreasing with reduction time. During the initial reaction phase (2 h), the pellet porosity was low, and adjacent pellets formed “point-to-point” contacts via metallic iron clusters (Figure 9(a) and (b)). Gas-phase iron atoms diffuse on the surface to form fibrous iron whiskers, while FeO reacts with SiO₂ and Al₂O₃ to form Fe₂SiO₄ (melting point 1170 °C) and a low-melting-point slag phase of FeO-SiO₂-Al₂O₃ (<1000 °C). The liquid slag phase fills the pores via capillary action, and upon cooling, forms solid silicate bridges (Figure 9(c) and (d)), synergistically enhancing sticking strength. At this stage, metallic iron deposits in a layered, dense structure, with the contact mode transitioning from “point-to-point” to “face-to-face,” significantly increasing connection density. When the reaction is extended to 3.5 h (Figures 9(e) and (f)), over-reduction leads to the formation of FeTiO₃ and Fe₃C (confirmed by XRD in Figure 8). FeTiO₃ occupies the iron sticking points, inhibiting iron atom aggregation, causing the contact surface to contract and the interconnection distance to increase; simultaneously, silicate bridging decreases, resulting in a significant increase in the void volume between pellets. The formation of Fe₃C further weakens the continuity of metallic iron, leading to a reduction in the number of contact points and their dispersion. These mechanisms collectively weaken the synergistic sticking effect between iron whiskers and the slag phase, ultimately resulting in a decrease in the SI. These observations indicate that sticking behavior is governed by the dynamic balance between the morphological evolution of metallic iron and the formation of secondary phases.

OPEN IN VIEWER

Figure 9.

Scanning electron microscope and energy-dispersive spectrometer (SEM-EDS) analyses of the sticking interface between adjacent pellets with different reduction time: (a, c, e) morphologies of sticking interface between adjacent pellets, (a) 2 h, (c) 2.5 h, (e) 3.5 h, (b, d, f) energy spectra of sticking interface (b) 2 h, (d) 2.5 h, (f) 3.5 h.

Effect of coating materials on adhesion behavior

Under a reducing atmosphere of 60% H₂, 25% CO, 5% CH₄, 1% H₂O, 3% CO₂, and 6% N₂, the system investigated the regulatory mechanism of coating material types (uncoated, MgO, TiO₂, CaO) on the SI of pellets after load reduction. The experiment employed fixed process parameters: after 1 h of gas reduction at 1100 °C, a load was applied, and the reaction was terminated after maintaining the load for 1 h.

As shown in Figure 10(a), all coating materials exerted a significant inhibitory effect on pellet sticking. The SI of the uncoated pellets was 32.11%, whereas those of the MgO, TiO₂, and CaO-coated pellets decreased to 2.18%, 2.11%, and 16.69%, respectively. Similarly, the sticking strength of the uncoated pellets was 13.63 N, while those of the MgO, TiO₂, and CaO-coated pellets decreased to 3.04, 2.84, and 8.82 N, respectively. These results demonstrate that the coating materials markedly weakened the adhesive interaction between pellets by acting as physical barrier layers. Among the three coatings, TiO₂ exhibited the most pronounced inhibition effect, reducing the SI by 93.4% relative to the uncoated sample. By contrast, the inhibition efficiency of the CaO coating was substantially lower than that of MgO and TiO₂. Figure 10(b) shows the effect of coating materials on metallization rate. The metallization rate of the uncoated pellets was 95.56%, whereas those of the MgO-, TiO₂-, and CaO-coated pellets were 93.61%, 94.42%, and 94.51%, respectively. This decrease can be attributed to the dense isolation layer formed by the coatings, which hindered the inward diffusion and mass transfer of the reducing gases, thereby lowering the local reducing potential and suppressing the deep reduction of iron-bearing vanadium–titanium phases. Notably, while achieving the lowest adhesion index (2.11%), the TiO₂ coating exhibited a smaller decrease in metallization rate compared to MgO, demonstrating a selective influence on the reduction process. Therefore, under the present laboratory conditions, TiO₂ exhibited the best anti-sticking performance among the examined coatings.

OPEN IN VIEWER

Figure 10.

Experimental results of pellet load reduction with different coating materials at 1100 °C: (a) sticking index and sticking strength, (b) metallization rate.

The XRD results shown in Figure 11 further reveal the effect of coating materials on the phase composition of the sticking interface. At a constant temperature of 1100 °C, the sticking interface of the uncoated samples was dominated by metallic iron (Fe, PDF#06-0696), with a characteristic Fe (110) diffraction peak at 44.674° (2θ), and the peak intensity was significantly higher than that of the coated samples. This phenomenon indicates that, in the uncoated system, reducing gases can freely diffuse into the interior of the pellets, promoting the complete reduction of iron–vanadium–titanium compounds and driving the migration of iron atoms toward the particle surface. Under high-temperature thermal activation, iron atoms on the surface diffuse along grain boundaries to form a continuous metallurgical bonding interface. Meanwhile, the grains grow along preferred orientations, forming a dense layered structure that significantly enhances sticking strength through metallic bonding. In addition, slag-phase components (such as SiO₂ and Al₂O₃), which are not effectively inhibited by the coating, still contribute to the formation of low-melting-point liquid phases. This further strengthens the mechanical interlocking effect and leads to an increase in the SI. The introduction of coating materials significantly modifies the interfacial phase composition and sticking behavior through the combined effects of physical barrier formation and chemical regulation. In the CaO- and MgO-coated systems, the XRD patterns, in addition to the peaks corresponding to metallic iron, exhibit characteristic peaks of carbide phases (Fe₃C, PDF#89-2722) at 43.752° (2θ). This observation suggests that the coating materials exert a catalytic effect on the carburization reaction. However, the specific mechanism of this catalytic effect requires further research and verification. In contrast, the TiO₂-coated system exhibits a distinctly different phase composition. The XRD pattern does not show the theoretically predicted FeTiO₃ phase, whose characteristic peak is located at 32.588° (2θ) (PDF#01-075-1212). Although TiO₂ showed the best inhibitory effect among the examined coatings, the XRD pattern of the TiO₂-coated sample did not always display a distinct FeTiO₃ peak. This does not necessarily exclude the formation of FeTiO₃ or other Fe–Ti–O reaction products, because such phases may be confined to the local interfacial region, present in low content, or partially overlapped by stronger diffraction peaks from the main phases. Therefore, in this study the formation of FeTiO₃ is interpreted mainly on the basis of local interfacial evidence from SEM-EDS together with reaction plausibility. Combined with the microstructural observations, these results indicate that the TiO₂ coating achieves the most effective suppression of sticking through several mechanisms. First, its dense structure effectively impedes the penetration of reducing gases into the pellet interior; Second, the relatively low CH₄ adsorption capacity of the TiO₂ surface suppresses the generation of free carbon, thereby preventing the formation of brittle phases such as Fe₃C. According to the sticking mechanism proposed in the work of Qie et al. ²⁰ : in uncoated pellets, iron atoms migrate toward the surface during reduction and form metallic bonding at elevated temperatures, which constitutes the fundamental reason for the increase in the SI. The introduction of coating materials suppresses this process through two main mechanisms. On the one hand, the dense coating hinders the diffusion of reducing gases, thereby reducing the formation of metallic iron. On the other hand, the free carbon generated by CH₄ cracking blocks pores within the pellet, further enhancing the inhibitory effect on reduction.

OPEN IN VIEWER

Figure 11.

X-ray diffraction (XRD) results of pellet sticking surfaces at different coating (a) 2θ: 10°–90°, (b) 2θ: 40°–50°.

SEM-EDS analysis was employed to investigate the microstructural evolution of the bonded interfaces in the reduced products under different coating materials after reduction at 1100 °C for 2 h, thereby elucidating the intrinsic interfacial morphological evolution and reduction mechanisms during the direct reduction process. The results are presented in Figure 12.

OPEN IN VIEWER

Figure 12.

Scanning electron microscope and energy-dispersive spectrometer (SEM-EDS) analyses of the sticking interface between adjacent pellets with different reduction time: (a, c, e) morphologies of sticking interface between adjacent pellets (a) MgO, (c) TiO₂, (e) without coating (b, d, f) energy spectra of sticking interface (b) MgO, (d) TiO₂, (f) without coating.

After application of the MgO coating, the MgO reagent formed a continuous grid-like layer on the pellet surface (Figure 12(a)), effectively preventing direct contact between metallic iron and the underlying matrix. EDS analysis (Figure 12(b)) shows that Mg is uniformly distributed throughout the interfacial region, indicating that the MgO coating suppresses the surface migration of iron atoms and the growth of iron whiskers through physical separation. In addition, the grid-like structure restricts the penetration of reducing gases, thereby decreasing the extent of reduction and the amount of metallic iron formed. Notably, MgO does not completely suppress carburization, as small amounts of the Fe₃C phase are still detected. As an alkaline vanadium–titanium compound, MgO can react with acidic gangue components such as SiO₂ and Al₂O₃ within the pellet to form high-melting-point phases, such as calcium magnesium silicate (CaMgSiO₄, melting point approximately 1450 °C) and magnesium aluminate spinel (MgAl₂O₄, melting point 2135 °C). These high-melting-point slag phases remain solid at 1100 °C and thus effectively inhibit the formation of low-melting-point liquid phases in the FeO–SiO₂–Al₂O₃ system (below 1000 °C), thereby reducing the formation of silicate bridges through capillary action, as evidenced by the markedly decreased Si and Al signal intensities in Figure 12(b). Moreover, the combined effect of the solid slag phase and the MgO coating further suppresses surface roughening of the pellets and weakens mechanical interlocking.

The microstructural characteristics of the sticking interface in TiO₂-coated pellets are shown in Figures 12(c) and 12(d). The SEM images reveal that the interfacial region exhibits a typical porous structure, in which the originally continuous and dense distribution of metallic iron particles is disrupted and replaced by discrete, isolated island-like morphologies (Figure 12(c)). EDS elemental mapping (Figure 12(d)) reveals titanium enrichment in the interfacial fracture region, which is consistent with the presence of FeTiO₃, indicating that titanium is primarily retained in the iron–vanadium–titanium phase and contributes only minimally to slag-phase formation. The metallic iron particles are tightly encapsulated by FeTiO₃. This feature hinders the aggregation of metallic iron and the formation of interconnecting bridges at the sticking interface, thereby reducing both the number and size of contact points between metallic iron particles. Consequently, the SI of the pellets decreases.

In contrast, the microstructure of the uncoated pellets (Figure 12(e)) undergoes pronounced changes, and the sticking interface exhibits a typical point-to-point contact pattern. Fibrous iron whiskers are densely distributed in the contact regions between adjacent pellets. EDS elemental mapping shows the coexistence of metallic iron (Fe) and silicate phases containing Si, Al, and O. Under high-temperature reduction conditions, iron–vanadium–titanium compounds are fully reduced, and iron atoms diffuse via surface diffusion to form epitaxial whisker-like structures. Meanwhile, FeO reacts with SiO₂ and Al₂O₃ to form Fe₂SiO₄ and low-melting-point FeO-SiO₂-Al₂O₃ slag phases. The liquid slag phase fills pores through capillary action and subsequently forms solid bridges. These two processes act synergistically, leading to a significant increase in the SI.

Based on the experimental results concerning the effects of reduction time, reduction temperature, and coating materials on the sticking behavior of vanadium–titanium pellets in a hydrogen-based shaft furnace, Figure 13 was constructed to illustrate how these factors influence and inhibit pellet sticking during the reduction process.

OPEN IN VIEWER

Figure 13.

Mechanism of hydrogen-based reduction sticking and inhibition in vanadium–titanium pellet vertical furnaces.