Keywords

Introduction

Mineralization during sintering is a process to make part of the raw material melt after serial complex chemical reactions at high temperature. After condensation and consolidation of melt, raw mixtures are consolidated into agglomerates with good strength to satisfy the requirement of ironmaking in blast furnace [1]. The mineralization in sintering process impacts on the structure and mineral composition of sinters. Therefore the important indexes, such as production, quality and energy consumption of sinter, are depended on the high temperature mineralization behavior of raw materials [2]. Most researches of mineralization mainly focused on the ability of a single kind of iron ores to take part in mineralization, as well as the influence of chemical components [3,4,5].

Temptation to evaluate the mineralization ability of iron ores through researching their reaction behaviors has been done. Among the mineralization properties, assimilation capacity which represented the reaction ability of iron ores won more favors. Loo, C.E. from BHP Corp. Researched the assimilation through measuring the formation of calcium ferrite in the roasting process [6]. The method is that the raw materials are ground into −0.125 mm, then blended, briquetted and roasted in a heating furnace on the condition of simulating sintering temperature and atmosphere, at last, calcium ferrite in roasted agglomerate was tested to characterize assimilation properties. Wu, S.L. et al. from China researched assimilation capacity like this, the iron ore powder cake placed on the upside of CaO cake were sent into the furnace, and the temperature when the cakes appeared a cycle of product were measured to evaluate the assimilation capacity [7]. Besides, some other researchers tried to reveal the relationship between the chemical compositions and sinter strength by studying the influence of SiO2, Al2O3, MgO and basicity on mineralization [8,9,10,11,12,13]. All above research work have promoted the understanding of mineralization in the sintering process, but it’s still not easy to distinguish the mineralization ability among various grains of raw materials. This paper will study the mineralization behavior of iron ores and fluxes with different grain sizes in granules , and aims at revealing the mineralization mechanism during sintering .

Materials and Methods

Properties of Raw Materials

Iron ores and fluxes involving limestone, dolomite and quicklime were utilized to produce sinter. The chemical compositions of materials are given in Table 1. Two representative types of iron ores, hematite and limonite, were selected in this investigation. Hematite came from Brazil which had a high content of Fe, with few impurities except for 3.78% SiO2 and 0.77% Al2O3. And the limonite was from Austria, which had higher SiO2 and Al2O3 contents than that in hematite. Lime was the product of limestone after calcinations, which had a high CaO content of 89.66%. Limestone and dolomite belonged to carbonate, and limestone contained 52.31% CaO, and dolomite contained 33.08% CaO and 17.35% MgO. The size characteristics of raw materials are shown in Table 2. The contents of fine particles with −0.5 mm of hematite and limonite were 40.30 and 50.07% respectively. The sizes of limestone and dolomite were mainly −3 mm, which accounted for nearly 80%.

Table 1 Chemical compositions of materials/wt%
Table 2 Size distributions of materials

Methods

The micro-sintering was used to research the physicochemical reactions at high temperature. The micro-sintering adopted a horizontal furnace with controllable temperature-rising rate and atmosphere. In order to simulate sintering process exactly, this experiment divided the high temperature process into preheating belt, reaction belt, fusion belt, solidification belt and sinter belt. On the basis of physicochemical characteristics of each belt and temperature curve of the actual sintering layer, the micro-sintering used the heating program and atmosphere in Table 3 to simulate the sintering process.

Table 3 Heating program and atmosphere of micro-sintering

Mineralization testing process included ore matching, agglomeration, roasting and mineralogical analysis. Iron ores and fluxes (dolomite, limestone, and quicklime) were mixed to a feed, which then was used to produce sinter with SiO2 4.7%, basicity (CaO/SiO2) 2.0 and MgO 2.0%, the proportions of each material were shown in Table 4. The principle of ore matching for adhesion layer was screening out the fine powder (−0.5 mm) in iron ores and fluxes, and their proportions see Table 4. Added various kinds of nucleus particles in different size (0.5–1, 1–2, 2–3, 3–5 mm) into adhesive powder for simulating the structure of granules . After the ore matching and blending, the mixtures were compacted into cylinders with 30 × 25 mm size of the punching machine, using 300 kg/cm2 pressure for 1 min. The cylinders were heated to 1300 °C according to the parameters in Table 3. After the agglomerates were cooled down, they were made into polished sections for optical microscope (Leica DMRXP) and its image analysis software (Qwin), together with the SEM (JMS-56600LV), to observe the mineralogical characteristics of sintering mixtures. The experimental flow is shown in Fig. 1.

Table 4 Materials matching for sintering and adhesion layer/wt%
Fig. 1
figure 1

Experimental flow chart of micro-sintering

Results and Discussion

Mineralization Behavior of Adhesion Layer

The mixtures were granulated to granules before sintering. The previous study [14] of authors has confirmed that the granule consisted of adhesion layer and nucleus particles . The particles of −0.5 mm served as the adhesive fines , while the particles of +0.5 mm served as the nuclei, of which the typical structure is shown in Fig. 2.

Fig. 2
figure 2

Structural model of granule

Fine ores and fluxes in adhesion layer possessed large specific surface areas, thus the contact areas between the particles were correspondingly sufficient. Therefore the chemical reactions take place in the adhesion layer first. Mineralization behavior of adhesion layer is shown in Fig. 3.

Fig. 3
figure 3

Effect of temperature on mineralization of adhesion layer

When the temperature ranged within 1100–1150 °C, solid-phase reaction had been occurred. However, obvious phenomenon was hardly observed due to the slow reaction speed. After increasing the temperature to 1200 °C, a large number of calcium ferrite (abbreviate as CF) generated by solid-phase reactions. As the temperature increased further to 1225–1250 °C, the liquid phase generated obviously and the holes began to shrink. The liquid content was developed when the temperature increased to 1300 °C, and the main mineral constituents were CF, magnetite and hematite, without unfused fluxes, which proved that all the iron ores and fluxes in adhesion layer could participate in mineralization completely.

Mineralization Behavior of Flux Nucleus Particles

When nucleus particles of lime and dolomite (2–3, 3–5 mm) were added to the adhesion layer, the mineralization behaviors are shown in Fig. 4. From Fig. 4a, c it was found that the lime and dolomite of the fraction of 2–3 mm could mineralize completely, and there were no residual nucleus particles. From Fig. 4b, d, it could be found that most of fluxes in the fraction of 3–5 mm participate in the mineralization, and only a few tiny nucleus remained. On account of the fact that the fraction of fluxes used in sintering were all controlled under 5 mm (exactly the fraction of −3 mm accounting for more than 80%), fluxes, such as lime and dolomite, could take part in mineralization mostly.

Fig. 4
figure 4

Mineralization behavior of different size of flux particles

Mineralization Behavior of Iron Ore Nucleus Particles

Iron ore nucleus particles of hematite or limonite with the size of 0.5–1 and 1–2 mm accounting for 25% (wt%) in mixture were added into adhesion layer (75%) for the micro-sintering, and compared with the scheme without nucleus particles. The microstructure of sinter roasted at 1300 °C is shown in Fig. 5.

Fig. 5
figure 5

Mineralization behavior of different size of iron ore particles

The structure of adhesion layer without nucleus particles is shown in Fig. 5a, and it was hard to find out the unreacted ores after roasting. Figure 5b is the image of sinter with 0.5–1 mm nucleus particles of hematite, in which some remained unfused ores still existed. Figure 5c shows that the nucleus particles with a fraction of 1–2 mm could hardly be mineralized in the sintering process. In the same way, for limonite, liquid could penetrate into the interior of the unreacted core due to the developed porosity resulting from the removing of hydrate water, but the nucleus particles of limonite were still difficult to participate in mineralization (Fig. 5d). Therefore, whether hematite or limonite, when coarse iron ores (+0.5 mm) were added to mixtures, they would remain as unfused ores.

Therefore, the structure of sinter consisted of melt zone and unfused ores. It can be concluded that the unfused ores were surrounded by molten liquid. In the sintering process, all the fluxes participated in the mineralization, while the grain boundary for iron ores was 0.5 mm. Specifically, fine iron ores (−0.5 mm) took part in mineralization and formed the melting zone, while coarse iron ores (+0.5 mm) remained as unfused ores.

Condensation of the Liquid Phase and the Structures of Sinter

The initial liquid phase generated in adhesion layer first, and nucleus particles of fluxes melt into the liquid phase, which formed the melt zone of sinter after cooling and condensation. The microstructures of melt zone are shown in Fig. 6. There were mainly 3 kinds of structures, corrosion structure of magnetite and CF, eutectic structure of CF and silicate, and pilotaxitic texture of hematite and CF. The major part of the melting zone was the corrosion structure of magnetite and CF, accounting for 80–90%. Only a small area was characterized by the pilotaxitic texture of hematite and CF, the eutectic structure of CF and silicate.

Fig. 6
figure 6

Microstructures of melt zone

Surface scanning analysis by SEM was used to research the element distribution in melt zone, as shown in Fig. 7. It is shown that Fe mainly existed in iron oxide, secondly in CF, then in silicate; Ca mainly existed in CF and silicate; Si mainly existed in silicate, as well as a few in CF; the distribution of Al was relatively uniform, while its content in CF was slightly higher than that in iron oxide; Mg distributed relatively evenly.

Fig. 7
figure 7

Element distribution of melt zone in sinter

EDS was used to analyze the distribution of chemical compositions in melt zone, and the results are shown in Table 5. It can be seen that there were some difference between the compositions of each field, which proved that the melt zone was non-homogeneous. Under the assumption that all the fine iron ores (−0.5 mm) and fluxes were taken part in mineralization, the theoretical compositions of melt zone could be figured out. Compared the theoretical compositions with the average value of the testing ones, it can be found out that they were basically the same. The accuracy of sintering mineralization model could be further verified as well.

Table 5 Chemical compositions of melt zone in sinter/%

Mineralization Proceed of Granules During Sintering

From the results finding introduced, the mineralization path of granules in the sintering process can be schematically shown in Fig. 8. (1) Solid-phase (Fig. 8a, b) reaction took place in the adhesion layer of granule. Fine iron ores (−0.5 mm) with large specific surface area reacted with fine fluxes (−0.5 mm), which formed the low melting point material. Because CaO in flux had more chances for the touch with Fe2O3, the main product of solid-phase reaction was CF. (2) Primary liquid phase (Fig. 8b, c). When the temperature increased to the melting temperature of low melting point material, the CF that existed in adhesion layer began to melt to primary liquid phase. The lowest temperature for the generation of liquid phase in the CF system was the eutectic point of CaO·2Fe2O3 and 2CaO·Fe2O3, that is 1206 °C. Therefore, 1200 °C is proximately the temperature to form the primary liquid phase. (3) Extension of liquid phase (Fig. 8c, d). The fluxes which served as nucleus particles could dissolve in the primary liquid phase, which made the liquid phase get an extension. While the iron ores which served as nucleus particles could not take part in the mineralization because of the limitation of dynamical condition. As a result, the coarse iron ores (+0.5 mm) remained as unfused ores. (4) Liquid phase aggregation and crystallization (Fig. 8d, e). As temperature continued to rise, the liquidity of liquid phase got improved. And the liquid phases between the granules began to coverage, and the holes started to shrink. As temperature dropped from the peak, liquid phase began to condense gradually. As a result, crystals separated out successively and bonded the unfused ores together to form the final sinter.

Fig. 8
figure 8

Schematic of mineralization processes of granules

Conclusions

  1. 1.

    Solid-phase reaction takes place first in the adhesion layer of granules . Fine iron ores (−0.5 mm) with large specific surface area react with fine fluxes (−0.5 mm), which creates the low melting point material-CF. All the iron ores and fluxes in adhesion layer enable to participate in mineralization completely.

  2. 2.

    The flux which serves as nucleus particles can dissolve in the primary liquid phase, which makes the liquid phase extended. However, the iron ores which serve as nucleus particles cannot take part in the mineralization. As a result, the coarse iron ores (+0.5 mm) remain as unfused ores.

  3. 3.

    The sinter comprises melt zone and unfused ores. Fine iron ores (−0.5 mm) take part in mineralization and form the melt zone, and coarse iron ores (+0.5 mm) remain as unfused ores.