COMPARISON OF IRON ORE SINTER MINERALOGY QUANTIFICATION, WHEN XRD-PATTERNS FOR SFCA/SFCA-I AND OTHER CALCIUM FERRITES ARE CONSIDERED

At present, the powder XRD is widely used analytical method for iron ore sinter mineralogy determination. In many researches, diffraction data for SFCA and SFCA-I phases were not available and diffraction data for other calcium ferrites were used instead. Therefore, the aim of this paper is to bring a comparison between evaluation of diffractograms, when SFCA/SFCA-I, and, other calcium ferrites, respectively, are considered. As a result, the quantification of phases is quite comparable, when minerals are collected into groups, i.e. hematite, magnetite, calcium ferrites, silicates and non-assimilated non-ferrous phases. Only in the case of silicates a comparison tends to be dissatisfactory.


Introduction
Powder X-Ray diffraction (XRD) is the fastest analytical method to evaluate an iron ore sinter mineralogical compositionthe overall process takes only several minutes, just the sample must be grinded to a fine powder before the measurement [1,2]. On the other hand, this method has some disadvantages, which result from the nature of the technique. Because the XRD is based only on the physical parameters of the minerals and their structures, similarities in crystal lattice parameters of different phases can cause an incorrect identification. To avoid this, the knowledge of analyzed microstructure is necessary. Nowadays, the research in the field of iron ore sinter mineralogy is concentrated on one group of the bonding phasescomplex calcium ferrites (designated as SFCA, an acronym for silicoferrite of calcium and aluminum), which accumulate besides Fe2O3 and CaO also SiO2, Al2O3 and other minority components [3][4][5]. In industrially produced sinters, there are 2 types of SFCA-s according to the chemical composition: low-Fe SFCA and high-Fe SFCA-I [6]. Before advanced study of SFCA/SFCA-I, calcium ferrites CaFe4O7 and CaFe3O5 were considered as the main ferrite phases in lime-fluxed iron ore sinters [7,8]. Despite the fact, that diffraction patterns for complex calcium ferrites are known since 1989 and 1998 for SFCA [9] and SFCA-I [10], respectively, the diffraction databases of research centres are updated gradually, and in many cases, the data for mentioned phases are absent. For this reason, the rate of calcium ferrites in iron ore sinters is expressed through other, mostly ternary calcium ferrites, which can fit the measured diffraction pattern. Of course, this applies to older studies as well. In the current literature [11][12][13][14][15][16], it is common to substitute SFCA and/or SFCA-I DOI [17].
The main aim of the article is not to propose the replacement of diffraction patterns of mineral phases in sinters by others, but show the changes in phase quantification, which can occur when diffraction patterns of SFCA/SFCA-I are used instead of other calcium ferrites.

Materials and methods
The five sinter samples for this study were chosen from previous work [12]. Powder X-Ray diffraction analyses of iron ore sinter samples were performed using diffractometer SEIFERT XRD 3003/PTS. The parameters of measurements were as follows: generator at 35 kV and 40 mA, Coline focus, Fe filter, measuring range 10 to 120° 2θ by scan step of 0.02 θ, input slits 3 mm, 2 mm, and PSD detector Meteor1D. The measured diffraction patterns were analyzed using DIFFRAC.EVA (Search-Match) software with the database PDF2 and by program TOPAS that uses the Rietveld method. First, measured diffraction patterns were evaluated without data for SFCA and SFCA-I using patterns for other calcium ferrites -this set of evaluations is denoted here as "set A". Then, diffraction patterns for SFCA and SFCA-I replaced calcium ferrites from the set Athis set of evaluations is denoted as "set B". In all cases, the final XRD results were supported by microscopic examinations, as reported earlier [12].

Diffractogram solution
In the evaluation of diffractogram of an iron ore sinter, the aim is to fill out the vacant spaces of the measured diffraction pattern with calculated diffraction patterns of compounds present in the sinter. As an indicator of the accuracy, a difference curve is used. This curve is the more regular, the more the calculated diffraction patterns match the measured diffraction pattern. For demonstration, the evaluation of the diffractogram of sample 4 is presented in Fig 1. First, the evaluation with binary and ternary calcium ferrites was realized ( Fig. 1 a)), then the XRD patterns of SFCA/SFCA-I replaced ferrites as well as some other phases from the first case ( Fig.  1 b)).
As can be seen, the difference curve is well balanced in both cases, only on the sites of magnetite and hematite, there are some deviations. In XRD of iron ore sinters, a common problem is magnetite phase, which has a spinel structure and exists in sinters in a relatively wide range of chemical compositions because of the isomorphic admixtures. They cause slight shifts of diffraction peaks, which must be taken into account in the processing of the diffraction pattern.
Although dicalcium ferrite and dicalciumaluminoferrite are both described as the precursor phases in the SFCA phases formation process, they are quantified in much larger amounts in the sinter, when SFCA/SFCA-I calculated diffraction patterns are not present. This means that diffraction peaks of Ca2Fe2O5 fill out also a part of the non-filled measured diffraction pattern. A part of the diffraction peaks of Ca2Fe22O33 phase is often situated on the sites of SFCA, or SFCA-I, respectively. This is the reason, why it is possible to fit a part of the diffraction data for DOI  Ca2Fe22O33 to data of complex ferrites without an obvious conflict, which could be seen on the diffractogram. The remaining gap between measured diffraction pattern and missing SFCA calculated diffraction patterns is filled with calculated diffraction patterns of other calcium ferrites: Ca4Fe9O17, CaFe5O7. By replacement of calculated diffraction patterns with SFCA/SFCA-I, not only calcium ferrite types but also other phases like carbonates or wüstite were no more fittable. The choice of complex calcium ferrite patterns was based on the previous careful microscopic observation, which exactly confirmed the occurrence of SFCA and SFCA-I in studied sinters.

Phase quantification
The comparison of XRD measurements results of 5 samples without (set A) and with the use of diffraction patterns of SFCA/SFCA-I (set B) is presented in Fig. 2  When "substitute" calcium ferrites were replaced with SFCA/SFCA-I, changes in phase quantification appeared: the share of magnetite always increased, and the share of silicates always decreased. It also seems that the most affected were magnetite and silicate phases. But the magnetite phase was in all samples present in much higher amounts than silicates, so the differences cannot be taken equally. For example, in sample 1, by changing the evaluation, the amount of magnetite increased from 34.65 wt% to 39.82 wt%, so the difference is +5.17 wt%; but expressing the difference in simple %, the amount of magnetite phase has grown by 14.92 %. At the same time, the number of silicates decreased from 19.52 wt% to 12.60 wt%, so the difference is -6.92 wt%, but the ratio of silicates has dropped by 35.45 %. The least affected tend to be the ratio of hematite: the highest difference was the decrease in sample 2, only -1.15 wt% (-5.78 %). The changes in ratios of calcium ferrites and NANP were variable. The differences in the quantification of all phase groups as described above are summarized in Table 1. The very important question is if the sinters are comparable by phase ratios when the patterns for SFCA/SFCA-I are missing. As shown in Table 2, there were absolute matches in comparison of magnetite and ferrites' orders; only slight differences were in orders of hematite and NANP. However, although the same diffraction data for silicates in both sets were used, the order of phases in first and second evaluations was not comparable.

Discussion
Nowadays, the distinguishing between low-Fe and high-Fe SFCA phases becomes very important; however, when the diffraction data for these phases are not available, their quantification is practically impossible. Nevertheless, the image processing techniques based on the evaluation of the shades of grey are also limited to dividing the minerals in examined sinter samples just into main groups, i.e. hematite, magnetite, Ca-ferrites, silicates and pores [15,[18][19][20]. This means that the distinction between calcium ferrite types is also not involved. Therefore, the XRD evaluation from studies working without complete XRD database can be usable only when phase groups are considered rather than individual phases. In spite of this, the evaluation of samples in this study showed, the quantification of silicates seems to be problematic, and this must be taken into account.