2025 Volume 65 Issue 11 Pages 1588-1596
The phase evolution during heating of two Al2O3-free mixtures designed to form the SFC iron ore sinter bonding phase was investigated using in-situ synchrotron X-ray diffraction over the temperature range 293–1623 K. Improved fundamental understanding of the complex Ca-rich ferrite phases which form under sintering conditions has the potential to improve the sintering process. Results reported herein confirmed that during heating at an oxygen partial pressure of 5×10−3 atm, the γ-CFF phase (nominal composition Fe14.82Ca3.0O25, crystallographic space group P321), and not the β-CFF phase (Fe14.85Ca2.95O25, P3c1), formed as a precursor phase to SFC (M14O20, where M = Fe, Ca and Si, space group P1). Results also confirmed that during heating in air an Al2O3-free analogue of triclinic SFCA-I, designated SFC-I, formed as a precursor phase to SFC, and there was no evidence for the formation of the α-CFF phase that has been reported previously to form under equilibrium conditions. The question of whether an Al2O3-free analogue of SFCA-III, SFC-III, could also be formed during crystallisation from the melt during cooling was raised and in separate laboratory in-situ XRD experiments there was no evidence for the formation of SFC-III during cooling; rather, the β-CFF phase formed along with dicalcium silicate (CaO.SiO2, C2S) as the first phases to crystallize from the melt at high temperature, and with magnetite converting to hematite as cooling continued. Any presence of β-CFF in industrial sinter would likely indicate, and be a useful marker for, localized regions of Al2O3 deficiency in the sinter mixture, where that sinter mixture has experienced melting.
SFCA and SFCA-I are key iron ore sinter bonding matrix phases, and a review of their composition, structure and formation conditions has been undertaken.1) The review was focused on a high-Fe, low-Si form called SFCA-I (general formula M20O28, e.g., Ca3.2Fe2+0.8Fe3+14.7Al1.3O28, triclinic crystal structure, space group P1), and a low-Fe form called SFCA (M14O20, e.g., Ca2.3Mg0.8Si1.1Al1.5Fe8.3O20, also triclinic P1). The review also discussed the Al2O3-free analogue of SFCA, SFC, the compositional and thermal domains of which have been probed by Pownceby and Patrick and more recently by Cheng et al.2,3) Triclinic SFC has the same structure of repeating pyroxene (P) and spinel (S) modules as SFCA, and differs from the SFCA-I structure in the ordering of the modules (SPSPSP for SFCA/SFC; SSPSSP for SFCA-I).1) Very recently, as part of a systematic investigation into the ternary system CaO–Fe2O3–SiO2, Salzmann et al. discovered a phase with composition Ca2.68Fe10.32Si1.00O20 which crystallizes in monoclinic space group I2/c.4) The crystal structure is related to that of the triclinic polytype of SFC but exhibits a higher degree of disorder due to the partial occupation of additional octahedrally and tetrahedrally coordinated sites. This results in a smaller unit cell and an increased space-group symmetry.4) Such polytypism (i.e., a monoclinic variant of the triclinic structure type) has also been proposed for SFCA-I.5)
In recent phase equilibria and in-situ XRD work that we reported,6,7) the question of whether SFCA-I forms in the alumina-free SFC system was raised after considering what appears to be conflicting evidence in the literature. It should be noted that hereafter in this manuscript SFCA-I in the alumina-free system is referred to as SFC-I. Based on phase equilibria work targeted at synthesizing new single-phase triclinic SFCA-I compounds it appears that there is a lower limit of Al2O3 concentration for SFCA-I stability, somewhere in the range ~1.4–2.7 mass%.6,8,9,10) The α-CFF phase, with nominal composition Ca3.43Fe14.39O25,11) forms at lower Al2O3 concentrations.12)
However, Scarlett et al.13,14) reported the formation of SFC-I during heating of an alumina-free composition under partial vacuum (i.e., in air). This composition – 82.36 mass% Fe2O3, 14.08 mass% CaO and 3.56 mass% SiO2 – was designated AL0 and within the SFC compositional domain.2,3)
Based on this we have previously noted that SFC-I (i.e., Al2O3-free SFCA-I) may be a metastable phase. But also, the peaks assigned to SFC-I may have been misidentified and may instead have been due to the formation of α-CFF because XRD data collected during in-situ experimentation are generally of lower quality in terms of peak resolution and signal-to-noise ratio when compared with ex-situ data collected using conventional diffractometer configurations. In addition, the available 2θ range is generally smaller during in-situ experimentation due to geometric constraints imposed by using a reaction cell. SFCA-I and α-CFF have several closely located peaks (Fig. 1), and synchrotron-based experimentation, which affords greater peak resolution and peak-to-background ratio than laboratory-based instrumentation, would provide a better opportunity to distinguish the phases.
The first aim of the current work utilizing synchrotron experimentation is to confirm whether SFC-I, α-CFF, or both, form during heating of an alumina-free SFC composition in air. Observation of the (101) peak of the triclinic SFCA-I/SFC-I phase or the (110) peak of the monoclinic phase, and/or the (0012) peak of the α-CFF phase, would provide significant insight. In addition, observation of the (311) triclinic or the (043) monoclinic SFCA-I/SFC-I peak at a 2θ lacking an α-CFF peak of significant intensity, would provide further insight.
The formation of ‘SFC-I’ has been reported to occur during cooling at 2 K s−1 in air from melt in the Fe2O3(FeO)–CaO–SiO2 system by Nicol and coworkers.15,16) The bulk chemical composition of their oxide mixture was 69.24 mass% Fe2O3, 6.16 mass% SiO2 and 24.61 mass% CaO (basicity = CaO:SiO2 mass% ratio = 4). Through detailed microstructural and microchemical characterization the SFC-I was observed to be intergrown with a phase with composition Ca7.2Fe2+0.8Fe3+30O53, giving the appearance of a single crystal. The SFC-I/Ca7.2Fe2+0.8Fe3+30O53 intergrowth was the first phase to crystallize from the melt at ~1503 K. As cooling continued other phases were observed; C2S (2CaO.SiO2, Ca2SiO4), CF2 (CaO.2Fe2O3, CaFe4O7) and CF (CaO.Fe2O3, CaFe2O4).
Figures 2(a) and 2(b) demonstrate the similarity of the XRD patterns for γ-CFF and β-CFF, with these phases having nominal compositions of Fe14.82Ca3.0O25 and Fe14.85Ca2.95O25, respectively.17,18) In our recent in-situ laboratory XRD work performed for SFC sample compositions at an oxygen partial pressure (pO2) of 5 × 10−3 atm we observed the likely formation of the γ-CFF phase and proposed the following reaction mechanism to form SFC:7)
| (1) |
We also noted, however, that S-XRD experimentation would be necessary to confirm the formation of γ-CFF instead of β-CFF by observation of the (104) and (212) γ-CFF peaks (in addition to the (210) and (211) γ-CFF peaks) rather than the (213) β-CFF peak. The second aim of the current work, therefore, was to confirm which of the γ-CFF or β-CFF phases form during heating of alumina-free SFC mixtures at pO2 = 5 × 10−3 atm, with experiments performed at this partially reducing pO2 being more meaningful to industrial iron ore sintering than experiments conducted in air. Hsieh and Whiteman found that the pO2 of 5 × 10−3 atm maximized the formation of Ca-rich ferrites while still producing mineral assemblages like those found in industrial sinters.19) For completeness, simulated patterns for triclinic and monoclinic SFC are shown in Figs. 3(a) and 3(b), respectively. It should be noted that the relative peak intensities in the simulated patterns shown in Figs. 1, 2, 3 assume sufficient particle and counting statistics in a randomly oriented powder made up of small (nominally < 10 μm) crystallites. Table 1 summarizes the crystal structure parameters used to generate, using TOPAS V7 (Bruker AXS, Karlsruhe, Germany, 2022), the simulated patterns presented in Figs. 1, 2, 3.
| Phase | Reference | Structural formula/Composition | Symmetry, space group / Lattice parameters (Å, o) |
|---|---|---|---|
| SFCA-I/SFC-I (triclinic) | Mumme et al.20) | Ca3.18Al1.34Fe15.48O28 | Triclinic, P1 |
| a = 10.43, b = 10.61, c = 11.84 | |||
| α = 94.14, β = 111.35, γ = 110.27 | |||
| SFCA-I/SFC-I (monoclinic) | Mumme and Gable5) | Ca2Fe2+8Fe3+8Al4O28 | Monoclinic, P21/n |
| a = 10.34, b = 21.44, c = 10.50 | |||
| β = 109.96 | |||
| α-CFF | Karpinskii and Arakcheeva11) | Ca3.43Fe14.39O25 | Hexagonal, R32 |
| a = 6.01, b = 6.01, c = 94.69 | |||
| γ-CFF | Arakcheeva and Karpinskii17) | Fe14.82Ca3.0O25 | Hexagonal, P321 |
| a = 5.98, b = 5.98, c = 15.748 | |||
| β-CFF | Arakcheeva and Karpinskii18) | Fe14.85Ca2.95O25 | Hexagonal, P3c1 |
| a = 5.99, b = 5.99, c = 31.39 | |||
| SFC (triclinic) | Liles et al.21) | Ca2.55Fe10.85Si0.65O20 | Triclinic, P1 |
| a = 9.12, b = 10.12, c = 10.61 | |||
| α = 63.96, β = 84.48, γ = 65.67 | |||
| SFC (monoclinic) | Salzmann et al.4) | Ca2.68Fe10.32Si1.00O20 | Monoclinic, I2/c |
| a = 10.46, b = 15.27, c = 5.31 | |||
| β = 110.02 |
Two sinter mixture samples were prepared – designated SFC-B and CF-17 – and their nominal bulk compositions are shown in Table 2. The compositions were designed to model the reactive ultrafine (< 1 mm) component of an industrial sinter mix rather than being representative of the bulk composition of a sinter blend which contains coarse nuclei (up to 6.3 mm in size), gangue and flux particles. They were chosen because they i) were shown by Pownceby and Patrick to be two of only a small number of compositions to produce single-phase SFC when equilibrated in air in the range 1473–1513 K,2) and ii) spanned the extent of the SFC compositional range. SFC-B has similar composition to the Al0 sample investigated by Scarlett et al.13,14) They were prepared by homogenising mixtures of fine-grained synthetic, dehydrated Fe2O3 (99.99%), SiO2 (99.999%) and CaCO3 (99.9%) powders. They are the same samples that were used in our previous work.7)
| Sample | Composition (mass %) | Basicity | ||
|---|---|---|---|---|
| Fe2O3 | SiO2 | CaO | ||
| SFC-B | 82.43 | 3.44 | 14.13 | 4.11 |
| CF-17 | 79.67 | 4.90 | 15.43 | 3.15 |
In-situ S-XRD experiments were performed for the SFC-B and CF-17 samples on the powder diffraction beamline at the Australian Synchrotron. The diffractometer was fitted with an Anton Paar 2000N high-temperature chamber, which incorporates a Pt heating strip. Throughout heating the Anton Paar chamber was fed by a continuous flow of either a 0.5% O2 in N2 gas mixture (pO2 = 5 × 10−3 atm), or air. A heating rate of 20 K min−1 was used for 298–873 K, then 10 K min−1 for 873–1623 K which corresponds to the region of Ca-rich ferrite phase formation and decomposition. The same gas and heating conditions were used in our previous work.7) XRD data were collected in asymmetric diffraction geometry using an incident-beam-to-sample angle of 5°, vertical and horizontal slits of 0.2 and 2 mm, respectively, and an X-ray wavelength of 1.1066 Å (0.11066 nm). Data were collected continuously during heating, with individual datasets collected for 1 min at each of the two detector positions P1 and P2. P1 and P2 datasets were merged using PDViPeR (Australian Nuclear Science and Technology Organization). The decomposition of precursor phases and the formation of new phases were visualized as a contour plot of the diffracted peak intensity versus 2θ angle and temperature. Phase identification was performed on individual datasets using Panalytical HighScore Plus V5.2 (Malvern Panalytcial B.V., Almelo, The Netherlands, 2023) incorporating the International Centre for Diffraction Data (ICDD) PDF-5+ 2025 database (ICDD, Newtown Square, USA, 2025).
2.3. Laboratory in-situ XRD ExperimentationLaboratory in-situ XRD experimentation was also performed for the SFC-B and CF-17 samples on a PANalytical Empyrean diffractometer fitted with an Anton Paar 16N high-temperature chamber incorporating a Pt heating strip. XRD data were collected in Bragg-Brentano diffraction geometry using an iCore incident beam optic incorporating a programmable divergence slit and incident beam mask for a constant X-ray beam illumination on the sample of 4 × 7 mm. The 1Der detector was operated in 1D mode with an active length of 2.1223°, and the step size of data acquisition was 0.033° 2θ. The X-ray wavelength was 1.789 Å (0.1789 nm) from a Co X-ray tube. The samples were heated at pO2 = 5 × 10−3 atm at a rate of 20 K min−1 for 298–873 K, then 10 K min−1 for 873–1623 K; they were then cooled at a rate of 2 K min−1 for 1623–1273 K and 50 K min−1 for 1273–298 K. Data were collected continuously during heating and cooling, with individual datasets collected for 2 min. The rationale behind performing these laboratory heating/cooling experiments is discussed in Section 3.4.
Figure 4 shows the plot of accumulated in-situ S-XRD data in the range 298–1623 K in air for the SFC-B mixture. During heating, the sequence of phase decomposition/formation events between 298 and 1623 K was: transformation of α-SiO2 to β-SiO2, which was complete by 863 K (dashed horizontal line); decomposition of CaCO3 (complete by 950 K) and formation of CaO; formation of dicalcium ferrite (C2F) at 1020 K; formation of monocalcium ferrite (CF) at 1178 K; formation of SFC-I at 1382 K; and formation of SFC at 1449 K. SFC-I and SFC coexist in the range 1438–1461 K but SFC-I disappears before SFC. Above 1494 K SFC decomposes and melts and the final phase assemblage consists of Fe2O3 in a Fe2O3(FeO) – CaO – SiO2 melt. Given that thermal expansion can shift peak positions significantly at elevated temperatures, typically to lower 2θ values since thermal expansion typically results in larger d-spacings, such effects needed to be considered for the phase/peak assignments. To do so, the peak offset/shift functionalities within the phase identification software were utilized to compare the quality of matches of the various phases within the ICDD PDF-5+ database (i.e. SFCA-I, SFCA, α-CFF) with the observed peaks in the experimental data.
The phase behavior above ~1360 K for SFC-B in air is shown in more detail in Fig. 5, which is a stack plot of individual in-situ S-XRD datasets over the range 19–22° 2θ. At 1432 K the peaks matched with the SFC-I phase are clearly visible in the S-XRD pattern. The peaks matched with SFC-I were also clearly observed in the S-XRD pattern collected at 1461 K, but by 1483 K only the peaks indicative of SFC were observed. Above 1483 K, the SFC phase had melted and the phase assemblage was Fe2O3 in a Fe2O3–FeO–CaO–SiO2 melt (but with no Fe2O3 peaks visible within the 2θ range shown in the plot). Like what has been observed in much of our previous work on Al2O3-containing samples,22,23,24,25) the decomposition of SFC-I above 1461 K resulted in the formation of additional SFC. What has been shown previously to affect the SFCA formation mechanism is the addition of MgO; the addition of up to 3 mass% MgO causes the formation of additional iron oxide with the decomposition of SFCA-I, rather than additional SFCA.26,27,28)
These results confirm the formation of SFC-I in the alumina-free SFC system. There was no evidence for the formation of α-CFF under the dynamic conditions in which this experiment was performed. In terms of triclinic vs monoclinic SFC-I, we consider that the triclinic SFC-I phase has formed given the presence of a (310) peak corresponding to the triclinic phase, and the absence of a strong monoclinic (200) peak (see Fig. 1). Similarly for triclinic vs monoclinic SFC, the observation of peaks corresponding to the (123) and (211) of triclinic SFC, at a region of 2θ lacking monoclinic peaks (Fig. 3) indicates the formation of triclinic SFC.
3.2. Phase Evolution during Heating at pO2 = 5 × 10−3 atmFigures 6(a) and 6(b) show the plots of accumulated in-situ S-XRD data in the range 298–1623 K at pO2 = 5 × 10−3 atm for the SFC-B and CF-17 mixtures, respectively. For SFC-B the sequence of phase decomposition/formation events between 298 and 1623 K was: transformation of α-SiO2 to β-SiO2, which was complete by 856 K; decomposition of CaCO3 (complete by 956 K) and formation of CaO; formation of C2F at 1010 K; formation of CF at 1240 K; formation of γ-Ca4Fe14O25 at 1385 K; formation of γ-CFF at 1407 K; and formation of SFC at 1440 K. Above 1532 K the phase assemblage was Fe3O4 in a Fe2O3(FeO) – CaO – SiO2 melt. Like what was observed in air, the peaks for the SFC phase appear to match more closely with those of the triclinic form.
In each of Figs. 6(a) and 6(b) the (104), (210), (211) and (212) γ-CFF peaks were clearly observed, whereas the diagnostic (213) β-CFF peak was not. These results, therefore, confirm the formation of γ-CFF, rather than β-CFF, during heating of alumina-free iron ore sinter mixtures designed to form SFC at pO2 = 5 × 10−3 atm. They confirm the reaction mechanism for the formation of SFC given by Eq. (1) reported in our previous work.7) In our previous work we had also noted that there was no evidence for the formation of SFC in the in-situ XRD experiment performed for the CF-17 composition under laboratory conditions.7) In the current work, however, the significantly improved signal-to-noise afforded by the S-XRD experimentation has revealed that the formation of a minor amount of SFC does in fact occur for the CF-17 composition (Fig. 6(b)), consistent with the phase equilibria work at pO2 = 5 × 10−3 atm of Pownceby and Clout.29)
3.3. Structure ConsiderationsNot included in the review by Nichol et al.1) as discussed in Section 1 was the SFCA-III phase, which has only been identified and described relatively recently. Kahlenberg et al. has reported the structural elucidation of triclinic and monoclinic SFCA-III,30) and Zoll et al. reported investigations into FCAM-III which has similar structure to SFCA-III but in the Si-free system.31) Webster and Pownceby subsequently confirmed that phases with the triclinic SFCA-III crystal structure are present in sinter strand and pot-grate sinters.32) Table 3 gives an overview of the different SFCA-phases whose existence has been proven by structural investigations. All SFCA-related phases involve extensive cation substitutions typically between M = Si, Fe2+, Fe3+, Al, Ca, Mg, and exhibit complex crystal structures with low symmetry and comparatively large unit-cell volumes.
| Phase | Reference | Structural formula/Composition | Symmetry, space group/Lattice parameters (Å, o) |
|---|---|---|---|
| SFCA | Hamilton et al.34) | M14O20 | Triclinic, P1 |
| Ca2.3Al1.5Fe8.3Si1.1Mg0.8O20 | a = 9.06, b = 10.02, c = 10.92 | ||
| α = 60.30, β = 73.68, γ = 65.81 | |||
| SFCA-I | Mumme et al.20) | M20O28 | Triclinic, P1 |
| Ca3.18Al1.34Fe15.48O28 | a = 10.43, b = 10.61, c = 11.84 | ||
| α = 94.14, β = 111.35, γ = 110.27 | |||
| SFCA-II | Mumme et al.5) | M17O24 | Triclinic, P1 |
| Ca2.55Al4.65Fe9.8O24 | a = 10.34, b = 10.45, c = 17.98 | ||
| α = 90.31, β = 89.81, γ = 109.45 | |||
| Mumme et al.5) | M18O24 | Monoclinic, P21/n | |
| Ca2.4Al4Fe11.6O24 | a = 10.34, b = 17.99, c = 10.49 | ||
| β = 109.50 | |||
| SFCA-III | Kahlenberg et al.30) | M26O36 | Triclinic, P1 |
| Ca2.99Al4.56Fe15.35Si0.43Mg2.67O36 | a = 10.33, b = 10.43, c = 14.38 | ||
| α = 93.49, β = 107.32, γ = 109.66 | |||
| Kahlenberg et al.30) | M26O36 | Monoclinic, P21/n | |
| Ca2.99Al4.56Fe15.35Si0.43Mg2.67O36 | a = 10.33, b = 27.01, c = 10.43 | ||
| β = 109.66 |
*M = Si, Fe2+, Fe3+, Al, Ca, Mg.
SFCA (n = 0), SFCA-I (n =1) and SFCA-III (n = 2) form a polysomatic series with general formula M14+6nO20+8n. Each structure consists of distinct ordering of spinel (S) and pyroxene (P) modules; SPSPSP for SFCA, SSPSSP for SFCA-I, and SSSPSSSP for SFCA-III (Fig. 7). Kahlenberg et al. determined the cation distribution amongst octahedral (M) and tetrahedral (T) sites in triclinic SFCA-III (composition Ca2.99Mg2.67Fe3+14.58Fe2+0.77Al4.56Si0.43O36), with Al occupying most of the M sites to a minor degree.30) The T sites of the pyroxene modules, however, displayed significant occupancy by Al, up to 70%, with the remainder of these sites being occupied by Fe3+ and Si.
Mumme et al. also noted minor Al occupancy of M sites but significant occupancy of T sites (up to 46%) in triclinic SFCA-I with composition Ca3.2Al1.3Fe3+14.7Fe2+0.8O28, with the remainder of those sites being occupied by Fe3+.20) More recently Zoll et al. reported, for an SFCA-I phase (which they designated FCAM-I due to the lack of Si and presence of Mg) with significantly higher Al content (phase composition Ca2.90Mg0.95Fe10.11Al5.99O28) occupancy of up to 32% on the M sites but up to 100% on the T sites.33) For triclinic SFCA, significant substitution of Fe3+ by Al and/or Si on T sites has also been reported,20,21,34) and in the case of the triclinic SFC structure reported by Liles et al. the Al was replaced by either Fe or Si across the sites.21) Therefore, there is precedent for significant substitution between Al, Fe3+ and Si in the T sites across the three structures in the polysomatic series, and the substitution of Al by Fe3+ and/or Si appears to occur adequately in the Al2O3-free SFCA-I system so that SFC-I is at least present during heating, and as Nicol et al. have observed can be quenched in during cooling albeit intergrown with Ca7.2Fe2+0.8Fe3+30O53.15)
It should be noted that no evidence for the formation of SFCA-III was observed in the in-situ S-XRD experiments performed during heating as part of this study. In the in-situ S-XRD experiments that have been performed to date, SFCA-III (or Fe-rich SFCA as it has been referred in those studies) has only been observed during cooling and at pO2 = 5 × 10−3 atm as the first phase to crystallize from the melt.22,35,36) SFCA-III and Fe-rich SFCA appear to be the same phase.32)
3.4. Crystallisation during CoolingTo confirm (or otherwise) the formation of SFC-III during cooling at pO2 = 5 × 10−3 atm, separate laboratory in-situ XRD experimentation was executed, and the results are shown in Fig. 8. The phase behaviour during heating matches that observed in Figs. 6(a) and 6(b) for SFC-B and CF-17, respectively. During cooling, in Fig. 8(a) for SFC-B the first peak to appear at 1462 K and at 6.4° 2θ (d-spacing = 16 Å, 1.6 nm) is a possible match for the characteristic reflection of the Ca7.2Fe2+0.8Fe3+30O53 phase observed by Nicol et al.,15) considering the effects on peak position due to thermal expansion at the high temperature. There was, however, no evidence for the formation of SFC-I during cooling, and in fact the peak at d-spacing = 16 Å (1.6 nm) is also considered to be a match for either the (001) or (002) reflections of γ- and β-CFF, respectively. Across the whole 2θ range collected, the β-CFF phase is considered a superior match to the experimental data than both γ-CFF and Ca7.2Fe2+0.8Fe3+30O53.
In terms of data quality, it should be noted that after the formation of a significant amount of melt above ~1473 K during heating, and then throughout cooling, the relative intensities of all phases are significantly adversely affected by poor particle statistics arising from an insufficient number of crystallites in the X-ray beam, and/or the preferential growth of crystallites along particular crystallographic direction(s). In other words, the criteria for accurate relative peak intensities of the phases present (discussed in Section 1) are no longer fulfilled. This effect is particularly evident in the magnetite peaks at ~34, 41 and 49° 2θ in Fig. 8(a), which are either ‘spotty’ or disappear completely during the cooling stage of the experiment. In the case of the β-CFF phase, there is a strong preferred orientation about (00l), with peaks corresponding to β-CFF (002), (004), (006), (008), (0010), (0012) and (0014) all clearly evident across the 2θ range, whereas others such as the (108), (210), (211) and (213) (i.e., those labelled in Fig. 2(b)) were not.
Between 1462 K and 1446 K in Fig. 8(a) the (002) β-CFF peak has very low intensity, but at 1441 K the intensity of this peak increases which coincides with the appearance of additional β-CFF peaks as well as peaks indicative of the α’H polymorph of C2S. As cooling continued down to 1413 K the magnetite began to oxidize to hematite, and this was complete by 1363 K. Below 900 K the C2S phase had converted to the β polymorph. In Fig. 8(b) for the CF-17 sample, peaks indicative of β-CFF and α’H-C2S are first evident in the in-situ XRD data collected at 1443 K, and as cooling continued the magnetite began to oxidise to hematite at 1422 K and this was complete by 1385 K.
Importantly, there was no evidence for the formation of SFC-III during cooling in these experiments. The characteristic SFCA-III/SFC-III peak at d = 3.3 Å (3.3 nm) which has been observed in previous in-situ S-XRD studies,22,35,36) was not observed in either Fig. 8(a) for SFC-B or Fig. 8(b) for CF-17. This suggests that there may be a lower limit to the amount of Al2O3 present in the mixture where the SFCA-III phase forms during cooling at pO2 = 5 × 10−3 atm. Given that SFCA-III formed during cooling in a mixture with 1 mass% Al2O3 in the earlier work of Webster et al.,22) this lower limit should be in the range 0 to 1 mass% for mixtures with basicity (B) of ~4 and comparable concentrations of the other constituent oxides. Future work could probe this lower limit for Al2O3 concentration for SFCA-III formation during cooling at pO2 = 5 × 10−3 atm. In terms of basicity, other earlier work by Webster et al. demonstrated a suppression of SFCA-III formation with decreasing basicity, with only a very small amount of SFCA-III phase observed to form during cooling for a mixture with B = 2.5, with SFCA being the dominant Ca-ferrite phase to form during cooling.35) In the Webster et al. mixture with B = 4, however, a significant amount of the SFCA-III phase was observed to form at 1496 K before the SFCA phase formed at 1473 K.35) Given, therefore, that the basicity of the CF-17 mixture (B = 3.15) was closer to 2.5 than to 4, this could also at least partially explain the apparent lack of SFC-III formation in the experiment performed for the CF-17 mixture.
We also note the differences in phase behaviour observed here compared to the results of Nicol et al.,15) and that the composition of SFC-B, whilst having similar basicity to the composition of the Nicol et al. mixture, was significantly richer in Fe2O3 and relatively deficient in both CaO and SiO2. Moreover, the cooling experiments performed here were at pO2 = 5 × 10−3 atm, rather than in air. Future work could involve conducting this type of in-situ XRD experimentation using these SFC-B and CF-19 Al2O3-free compositions, but in air, to assess whether SFC-III forms by crystallization from the melt under the more oxidising conditions utilised by Nicol and coworkers. Given that the structure of the SFCA, SFCA-I and SFCA-III phases is based on ordering of pyroxene and spinel modules, the formation of magnetite and then its preservation down to temperatures of ~1450 K during cooling (i.e. where the crystallization of Ca-ferrite phases is predicted to commence during cooling) would appear necessary for one or more of these phases to form. Future work could also involve performing the type of controlled cooling furnace experiments utilised by Nicol and co-workers using the SFC-B and CF-19 mixture compositions but at the slightly more reduced partial pressure of pO2 = 5 × 10−3 atm.
The Al2O3-free analogue of SFCA, SFC, is well known and well characterised. In this study, the formation of SFC-I during heating in air as a precursor phase to SFC has been confirmed. This raised the additional question of whether there is an Al2O3-free analogue of SFCA-III, SFC-III, which might also be formed. The results presented herein suggest not, however future work could involve probing the formation of SFC-III under a wider range of conditions, including equilibrium and dynamic conditions, at various values of pO2, and using sinter mixtures with Al2O3 concentrations in the range 0 to 1 mass%. SFCA-III has recently been shown to be a significant component of industrial iron ore sinter (up to ~20 mass%),32) and more detailed and extensive fundamental knowledge about its compositional limits and crystal chemistry is important for the field of iron ore sintering research and characterization. What has also been confirmed in this study through the in-situ S-XRD experimentation is the formation of γ-CFF and not β-CFF during heating of sinter mixtures in the Al2O3-free system. But then, the formation of β-CFF does occur during cooling as the first Ca-ferrite phase to crystallize from the melt in these Al2O3-free sinter mixtures. To the best of our knowledge the β-CFF phase is not typically observed in industrial or pot-grate sinter specimens. However, its presence in such would likely indicate, and be a useful marker for, localized regions of Al2O3 deficiency in the sinter mixture, where that sinter mixture has experienced melting.
There are no conflicts of interest related to the conduct of this work.
Dr Gus Mumme (CSIRO Mineral Resources, Clayton) is thanked for the crystal structure representations of SFCA, SFCA-I and SFCA-III. This research was undertaken on the powder diffraction beamline at the Australian Synchrotron, part of ANSTO.
