Note
Silico-ferrite of Calcium and Aluminum Characterization by Crystal Morphology in Iron Ore Sinter Microstructure
2016 Volume 56 Issue 6 Pages 1111-1113
Details
2016 Volume 56 Issue 6 Pages 1111-1113
Correct identification of mineral phases in iron ore sinters is key for further research, in which individual sinter properties are derived directly from mineral phases. Nowadays widespread image analysis and techniques based on the visual quantification may not exactly evaluate mineral type, what has an unfavorable effect on the overall phase analysis and all functions derived from it. Some specifics were demonstrated by microscopic examination of sinter samples and pointing out differences in calcium ferrite identification. It was found the dendritic crystallization may cause such crystal shape of low-Fe SFCA, that it ultimately looks like plates of high-Fe SFCA-I. Aggregates of SFCA-I plates crystallized adherent are misidentifiable as prismatic SFCA crystals. The pathway for SFCA-I crystals covered with SFCA was also discussed.
A long term objective of pig iron producers is use of good quality iron ore sinters, which ensure smooth operation of blast furnace. A growing trend in research of iron ore sinters is deriving of mechanical and metallurgical properties from sinters’ microstructure and mineral phases.1,2) Because currently produced fluxed sinters have basicity CaO/SiO2 around 2, the research of mineral phases is focused on bonding phases. In sinters fluxed with limestone and dolomite these are calcium ferrites.3) There are several types of calcium ferrites in sinters, depending on mixture basicity or chemical components present in the ore gangue.4,5)
There are two different kinds of calcium ferrites in current sinters. The first category are binary calcium ferrites – monocalcium ferrite CaFe2O4 and dicalcium ferrite Ca2Fe2O5 mostly in the vicinity of reacted limestone or dolomite. Generally, they have not a regular habit shape and are present as anhedral crystals. They are not as widespread in the industrial sinters as other calcium ferrites and make only a minority of mineral phases. It should be noted, that although they also contain other elements, from the structural viewpoint they still have the structure of binary ferrites.6)
The second group are the quaternary calcium ferrites, commonly named complex, because they also contain some other elements, such as Mg, Fe2+, Na etc.. Usual name of these ferrites is silico-ferrites of calcium and aluminum with the acronym SFCA.7) After researches in the field of mineral’s structure the SFCA-s were divided into low-Fe type, designed as simple SFCA with the basic formula Ca2(Ca,Fe,Mg,Al)6(Fe,Si,Al)6O20 and M14O20 stoichiometry; and high-Fe type designed as SFCA-I with M20O28 stoichiometry.7,8) They are both very common in iron ore sinters and make a majority not only among the bonding phases but sometimes also in the whole sinter. By contrast to mentioned binary ferrites, SFCA has another structure quite different from calcium diferrite CaFe4O7, which is triclinic or monoclinic, respectively. So the differentiation is correct.8)
A lot of authors uses their own classification of ferrites. Most of researches of phase composition realized today is still based on image analysis, where only shades of grey and shape of crystal can observed.10,11) Many classify calcium ferrites according to that.
There is a number of classifications of SFCA according to its crystal shape in current scientific literature. Besides the classification into only two groups, there is also a classification with transition habits between basic shapes. Besides usual columnar and acicular shapes there are also mentioned platy and sheet.12,13,14,15)
Without closer examination of crystals, there was adopted a width limit of 10 μm between SFCA-I and SFCA. A ratio of 2.5 between longer and shorter side of crystal is also known.2,14) Other researches consider take into account mainly chemical composition determined by SEM-EDS.16,17,18)
In this research, acicular shape in cross section is denoted as platy morphology of SFCA-I and columnar as prismatic SFCA, respectively.
In this paper, there are demonstrated not only types of silico-ferrites of calcium and aluminum (SFCA) from the point of view of morphology and chemical composition in sinter samples, but also examples of such cases, in which can occur incorrect identification or confusion of SFCA types.
Sinters used for the study were prepared by sintering of raw materials, which are used in a real sintering plant – hematite ore, magnetite concentrate, fluxing agents limestone and dolomite, to achieve the basicity of final sinter (CaO+MgO)/(SiO2+Al2O3) ≈ 1.7, SiO2 level ≈ 10 wt% and MgO level ≈ 3 wt%.
The sintering was realized in a laboratory sintering pan, parameters of which and sintering conditions have been published elsewhere.19)
The samples were removed from sinter cakes and then prepared for observation under light microscopy: mounted into epoxy resin, grinded and polished. In order to expose the whole mineral habits in the microstructure a hydrochloric etching of amorphous phase was applied.
The chemical composition of mineral phases was obtained by SEM-EDS. The presence of concrete types of calcium ferrites in the sinter samples was confirmed by X-Ray diffraction analysis.
The most desirable type of sinter microstructure is bonded with SFCA-I bonding phase, i.e. the high-Fe type interpreted as acicular calcium ferrite.20) According to the long term research of authors, SFCA-I contains 74.7–85.1 wt% Fe2O3, 9–16.6 wt% CaO, 0.9–7.4 wt% SiO2, 0.5–2.8 wt% Al2O3 and up to 3.2 wt% MgO. The crystals of this ferrite are aligned closely together, have diverse directions of crystallization, and they often intersect each other. In the intercrystallic spaces are either pores or silicate phases. Fields with SFCA-I are located between coarse hematite or magnetite grains, which they bond together. An example of such microstructure is in Fig. 1(a). Chemical composition of this SFCA-I was 80.9 wt% Fe2O3, 11.6 wt% CaO, 5.1 wt% SiO2, 1.5 wt% Al2O3 and 0.9 wt% MgO. EDS analysis confirms the high-Fe type. Because this microstructure is in etched state and the EDS spot size has not reached the edges of crystal, no other minerals have affected the measurement. The width of transected crystals is about 1–2 μm, thus under the 10 μm limit.
SEM micrographs of calcium ferrites phases in iron ore sinter: a) and c) plates of SFCA-I; b) and d) dendrites of SFCA (microstructures in a), b) and d) etched).
In Fig. 1(b) are shown narrow elongated (acicular) SFCA crystals. They are oriented in similar direction but they are not entirely parallel. It is common, that the crystals diverge spokewise. Their width is also about 1–2 μm. There are also visible some intercrystallic spaces, which divide the mass of SFCA. They were filled with silicates before etching. Ordinary image analysis would classify this microstructure type as acicular calcium ferrites. But in this case, these are not plates in cross section, but an imperfectly crystallized form of prismatic SFCA. EDS proves it and detected chemical compositions as follows: 68.5 wt% Fe2O3, 17.8 wt% CaO, 8.8 wt% SiO2 and 4.9 wt% Al2O3. Despite the microstructure is on the first sight really like SFCA-I plates, it is possible to recognize this type also visually. This is dendritic shape, which ending is quite different from SFCA-I. All long parts of crystal have almost the same direction and endings like a cauliflower head (visible in Fig. 1(b)) in the upper right corner). The outlines of SFCA prismatic shape are more clear noticeable, there and the ending of crystal together with long parts forms a compact body. Moreover, the adjacent crystals to the right have a fish-fin-like ending.
A different crystallization stage of prismatic SFCA is on Fig. 1(d). Dendritic habit is more developed than that in Fig. 1(b) and is interlaced with small pores, where silicates have solidified.
3.2. Columnar ShapeThe 10 μm limit is a mistake even when SFCA-I plates are fused together and the boundaries between crystals are not visible. So in cross section these crystals look like columnar SFCA. A help for identification would be the ending of crystal, which is frayed. Individual plates are perfectly parallel, but have different length. The determined chemical composition of SFCA-I in Fig. 1(c) was 79 wt% Fe2O3, 13.1 wt% CaO, 3.7 wt% SiO2, 2 wt% Al2O3, 1.7 wt% MgO and 0.6 wt% Na2O. These crystals were captured in a pore, which is the reason why their surface is visible, and not the polished area. An example of polished (sectioned) form is given in the paper from Cores et al.2)
Few authors state that low-Fe SFCA is the only one type of SFCA, which crystallizes directly from the melt during the cooling stage.21,22) In that case, the crystals need contact areas for their initial growth (so called heterogeneous nucleation). Because only magnetite is present in solid state when all other phases are melted, magnetite crystals are the contact areas. This explains a coexistence of SFCA with magnetite in the microstructure obtained at higher temperatures.14,15,21) But the grown of crystals cannot be euhedral and they crystallize subhedraly as is shown in SEM microphotograph in Fig. 2(a). In the SFCA/glass field, there is only one side of crystals without the influence of others, so the prisms are drawn in Fig. 2(b). As visible, the endings of crystals are not uniform. There are also marked two dash lines inside the prismatic crystal, which denote the cross sections used for demonstration of different kinds of calcium ferrites’ habits in the following Fig. 3.
a) SEM micrograph of a field with prisms of SFCA; b) projected prism of SFCA.
Projected sectioned forms of calcium ferrites habits: a) prismatic SFCA; b) SFCA-I plates; c) SFCA-I plates with besieged with SFCA.
The first sketch in Fig. 3(a) is the prismatic SFCA crystal. It is homogenous by contrast to other projections. In addition, there is also displayed a growth defect. There is clearly visible the nature of SFCA, which forms an angulated “C” letter in the polished microstructures (or similar).
Figure 3(b) presents plates of SFCA-I fused together. This results in the crystal shape as in Fig. 1(c). There are not any gaps between the crystals, so other phases e.g. silicates could not crystallize here. If it would be the opposite, EDS analysis would be garbled as in analysis of dendritic hematite, which is practically impossible to analyze without signals of Si and Ca from calcium silicates in intercrystallic spaces.
Aggregation of SFCA as well as SFCA-I crystals are drawn in Fig. 3(c). The core of this aggregation are the conjoined SFCA-I crystals as in the above mentioned case (Fig. 3(b)). They are covered by SFCA crystals, which can be in plate form, too. Many authors identify these aggregations either as SFCA-I or as SFCA phase only.13,14,23) So the further derived research of the properties cannot be correct, because it examines two different mineral phases. Based on the chemical composition of ferrites, there are differences also in shades of grey. SFCA-I contains more iron, which is a heavy element, so the density of SFCA-I is higher than that of SFCA. For this reason SFCA-I reflects the light from light microscope better than SFCA and has a lighter shade in micrographs.
The previous research of such microstructure shows the possibility of formation of the intergrown crystals.6) If the proposal is right, that SFCA-I cannot crystallize from melt, there must be some SFCA-I preserved in a solid state during the melting period. After the melting period and the following cooling period, unmelted SFCA-I crystals act as contact areas for heterogeneous nucleation of newly crystallized SFCA. This theory would explain why SFCA crystals form the coverage of SFCA-I and not vice-versa. It was found, however, that SFCA and SFCA-I are together in the sinter microstructure even before melting period begins.4,21)
Current researches of iron ore sinter quality need sufficiently accurate methods for classification of mineral phases. For this purpose, observation of microstructures is mainly used. The results may not be absolutely right, because the types of calcium ferrites sometimes do not crystallize in generally adopted forms. In this paper were presented examples of sinter microstructure, which can cause incorrect identification of ferrites. Dendritic crystallization of low-Fe type SFCA can look like as so called acicular morphology in microstructure, however it creates a characteristic ending, which reveals the dendritic nature of the crystal. Crystals of high-Fe SFCA-I may form aggregations of conjoined plates. The width of such aggregations is generally bigger than 10 μm. Ultimately, they resemble columnar SFCA. Low-Fe SFCA has generally a prismatic habit, but it can probably crystallize from the melt on remaining SFCA-I crystals, which act as a contact area. As a result, SFCA-I crystals with SFCA coverage occur in sinter microstructure.
