Exploring the Relationship Between Fe-rich SFCA and SFCA-III and Their Presence in Sinter Strand and Pot-grate Sinter

Abstract

Using the X-ray diffraction (XRD) patterns collected on the products of laboratory furnace experiments performed on a synthetic iron ore sinter mixture with composition 77.36% Fe₂O₃, 14.08% CaO, 3.56% SiO₂ and 5.00% Al₂O₃, it is demonstrated that the previously reported Fe-rich SFCA phase has the same crystal structure as triclinic SFCA-III. Therefore, the four principal members of the silico-ferrite of calcium and aluminium (SFCA) family are SFCA, SFCA-I, SFCA-II, and SFCA-III. In addition, using XRD patterns collected for sinter strand and pot-grate sinter samples supplied as part of a sinter analysis round robin the presence of the SFCA-III phase in industrial sinter is confirmed for the first time, with important implications for sinter characterisation and research.

1. Introduction

‘SFCA’ (Silico-Ferrite of Calcium and Aluminium) phases are key bonding materials within industrial iron ore sinter.¹⁾ Sinter is utilised extensively worldwide in the production of steel from iron ore and typically constitutes more than 60% of the ferrous burden in modern blast furnaces in Japan and most of the blast furnaces in Europe.²⁾ Nicol et al. reviewed the compositional, structural and/or textural characteristics of the ‘SFCA’ phases described in the literature – SFCA (triclinic crystal structure), SFCA-I (triclinic), SFCA-II (triclinic) and Fe-rich SFCA.³⁾ Since then, Mumme and Gable have reported the crystal structures of the monoclinic variants of SFCA-I and SFCA-II,⁴⁾ and most recently Kahlenberg et al. has reported the structural elucidation of triclinic and monoclinic SFCA-III.⁵⁾ Zoll et al. reported investigations on FCAM-III with similar structure to SFCA-III but in the Si-free system.⁶⁾

Fe-rich SFCA was first observed by Webster et al.⁷⁾ during in-situ synchrotron XRD experiments designed to reveal the formation mechanisms of SFCA and SFCA-I during heating at cooling at an oxygen partial pressure (pO₂) of 5 × 10⁻³ atm. Webster et al. reported Fe-rich SFCA in experiments where a starting sinter mixture composition SM4-5 (with bulk composition, in terms of mass% of oxides, of 77.36% Fe₂O₃, 14.08% CaO, 3.56% SiO₂ and 5.00% Al₂O₃) was heated in a furnace at pO₂ = 5 × 10⁻³ atm to 1623 K, and then quenched from 1488 K during cooling. The phase which formed contained 58.88 mass% Fe, 6.89 mass% Ca, 0.82 mass% Si, and 3.00 mass% Al, and was named Fe-rich SFCA due to its high Fe content in comparison to SFCA-I and SFCA. Subsequently the effect of basicity (CaO:SiO₂ ratio), Mg concentration and Ti concentration on the crystallisation of Fe-rich SFCA, characterised by in-situ XRD, has been reported.^8,9) Fe-rich SFCA has been observed in a pilot scale pot-grate sinter sample, prepared using a blend of natural iron ore fines with overall composition nominally 59.00 mass% T-Fe, 8.30 mass% CaO, 5.24 mass% SiO₂, 1.64 mass% Al₂O₃ and 1.16 mass% MgO,³⁾ sourced from CSIRO’s Pullenvale operations. That Fe-rich SFCA had composition 59.30 mass% Fe, 6.74 mass% CaO, 1.03 mass% SiO₂, 1.50 mass% Al₂O₃ and 1.07 mass% MgO, in pot-grate sinter with overall composition nominally 59.00 mass% T–Fe, 8.30 mass% CaO, 5.24 mass% SiO₂, 1.64 mass% Al₂O₃ and 1.16 mass% MgO.³⁾ The crystal structure of Fe-rich SFCA has not yet been determined.

The compound Ca_2.99Mg_2.67Fe³⁺_14.58Fe²⁺_0.77Al_4.56Si_0.43O₃₆ corresponds to the first Si-containing representative of the M_14+6nO_20+8n polysomatic series of SFCA phases with n = 2 (i.e., M₂₈O₃₆) and designated SFCA-III,⁵⁾ having followed the Si-free compound Ca_2.38Mg_2.09Fe³⁺_10.61Fe²⁺_1.59 Al_9.33O₃₆ designated FCAM-III. For comparison, for SFCA n = 0 (i.e., M₁₄O₂₀) and for SFCA-I n = 1 (i.e., M₂₀O₂₈). SFCA-II has general formula M₁₇O₂₄. Each structure consists of distinct ordering of spinel (S) and pyroxene (P) modules; SPSPSP for SFCA, SSPSSP for SFCA-I, SSPSP for SFCA-II, and SSSPSSSP for SFCA-III. The Fe²⁺/Fe³⁺ ratio in each of the synthesised SFCA-III and FCAM-III compounds was calculated from the total iron content based on the crystal-chemical formula obtained from EMPA measurements and charge balance considerations.^5,6) Kahlenberg et al. noted that their chemically homogenous SFCA-III sample contained both a triclinic and monoclinic polytype, with P1 and P2₁/n symmetry, respectively. They also noted that the existence of SFCA-III in industrial iron-ore sinters has yet to be confirmed,⁵⁾ and this has been more recently re-stated by Murao and Kimura.¹⁰⁾

The first aim of this paper is to compare the XRD patterns, collected for the products of heat-quench furnace experiments where the synthetic iron ore sinter mixture SM4-5 is heated at pO₂ = 5 × 10⁻³ atm to 1623 K and then quenched from high-temperature during cooling, with the published patterns for triclinic and monoclinic SFCA-III. This is done to determine whether Fe-rich SFCA and SFCA-III/FCAM-III are the same phase, at least in terms of crystal structure, if not in composition. The second aim of this paper is to examine whether SFCA-III is present in industrial, sinter; samples supplied as part of a sinter analysis round robin¹¹⁾ were used for this purpose, as well as the commercially available JSS851-5 Japanese Iron and Steel Certified Reference Material (CRM).

2. Experimental

2.1. Heat-quench and Furnace Cooling Experiments

The SM4-5 mixture used for the heat-quench experiments was prepared from synthetic hematite, Fe₂O₃ (Acros Organics, 99.999%); calcite, CaCO₃ (Thermo Fisher, 99.95%); quartz, SiO₂ (Sigma Aldrich, 99.995%); and gibbsite, Al(OH)₃ (Alcan OP25 Super White, 99.9%). 0.5 g of the starting sinter mixture was pelletised and heated in a Pt foil capsule in a vertical tube furnace in an atmosphere of pO₂ = 5×10⁻³ atm. This pO₂ was found by Hsieh and Whiteman to maximise the formation of Ca-rich ferrites whilst still producing mineral assemblages like those found in industrial sinters.¹²⁾

The heating profile was the same as that used routinely by Webster et al. during their in-situ XRD experimentation;^{7,13,14,15,16)} 20 K min⁻¹ for 298–873 K, and then 10 K min⁻¹ for 873–1623 K. The temperature adjacent to the crucible was measured using a Pt/PtRh13% thermocouple connected to an ice-point cell, and temperature measurements are considered accurate to ± 15 K. Samples were then cooled from 1623 K at 5 K min⁻¹ and (i) water quenched from 1488 K (b) water quenched from 1398 K, or, (iii) allowed to cool within the furnace, with a cooling rate below 1223 K of 20 K min⁻¹. The products were then crushed and ground for ex situ XRD analysis.

XRD patterns were collected on the ground product material in Bragg–Brentano geometry over the range 5° ≤ 2θ ≤ 80° using a Bruker D8 instrument fitted with a cobalt long-fine-focus X-ray tube operated at 40 kV and 40 mA, a programmable divergence slit with fixed beam illumination of 10 mm, and a LYNEYE XE detector used in scanning line (1D) mode with an active length of 2.94° 2θ. Rietveld refinement-based quantitative phase analysis (QPA) of the XRD data was performed using Topas (Version 7, Bruker AXS). The crystal structure data of Kahlenberg et al., Liles et al., Mumme et al., Hamilton, and Midgely were used as starting parameters in the refinement for SFCA-III, SFCA, SFCA-I, magnetite and larnite, respectively.^{5,17,18,19,20)} The unit cell parameters of each phase were allowed to refine during the Rietveld refinement, as well as the Rietveld scale factor and crystallite size parameter for each phase. The unit cell volume for the SFCA-III phase was calculated from the refined unit cell parameters.

2.2. Sinter CRM, Sinter Strand and Pot-grate Sinter XRD Characterisation

Two of the round robin samples (UN011 and UN154) were obtained from industrial sinter strands, with the third (UN016) being produced in a pilot scale sinter pot. Additional details including bulk elemental assays for these sinters have been published previously as part of a sinter analysis round robin.¹¹⁾ The JSS851-5 Japanese Iron and Steel CRM (List No. 970) was supplied by Graham B. Jackson (Aust) Pty Ltd. Powder XRD data were collected for each sample in Bragg–Brentano geometry and fixed-divergence-slot mode with a PANalytical MPD, fitted with a Co tube operated at 40 kV and 40 mA, and an X’celerator detector used in scanning line (1D) mode with an active length of 2.12° 2θ, and a post-diffraction graphite monochromator. In addition to the crystal structure details described in Section 2.1, the crystal structure data of Blake et al., Lager et al., Dollase, and Kahlenberg and Fisher, were used for hematite, quartz, cristobalite, and srebrodolskite (dicalcium ferrite, C₂F), respectively.^21,22,23,24)

3. Results

Figures 1(a) to 1(c) show the final Rietveld fits to the XRD patterns collected for the products of the heat/quench and furnace cooling experiments, and Fig. 1(d) shows the results of the QPA for each sample. SFCA-III is clearly abundant in these samples. In all cases use of the triclinic structure of SFCA-III resulted in a far superior fit to the data, compared to the use of the monoclinic structure. For example, the R_wp quality of fit parameters increased from 9.1, 7.2 and 6.6 for the 1488, 1398 and 298 K samples, respectively, when the triclinic SFCA-III structure was used in the refinements to 19.0, 11.8 and 10.8 when the monoclinic SFCA-III structure was used in the refinements. Figure 2 shows the Rietveld fit to the XRD pattern collected for the 1488 K sample when the monoclinic structure was used in the refinement. Clearly, use of this structure does not result in a calculated model which adequately accounts for the diffracted intensity. There are several peaks/regions with significant unmatched intensity, most notably at 18, 24, 37 and 43° 2θ which are annotated in Fig. 2. The key outcome of these analyses is that the Fe-rich phase described by Webster et al.⁷⁾ and triclinic SFCA-III appear to be the same phase. From a chemical and structural point of view, therefore, we confirm that there are four principal members of the SFCA-family, and they are SFCA, SFCA-I, SFCA-II, and SFCA-III.

Fig. 1. Rietveld fits to the XRD patterns collected for the products of the heat/quench experiment performed for SM4-5 where the sample was quenched from (a) 1488 K and (b) 1398 K during cooling (i.e., after melting). (c) shows the fit of the XRD pattern collected for the sample cooled to room temperature in the furnace. In each plot the experimental XRD patterns are shown as the blue line and the calculated pattern the red line. The vertical lines below are the Bragg peak markers for the mineral phases. The black curve in each plot shows the contribution of the SFCA-III (triclinic) phase to the overall calculated model. The inset in each plot shows a magnified view of the region between 6 and 13° 2θ. Annotated on each plot is the R_wp value of the fit when the triclinic SFCA-III structure is used in the refinements. (d) shows the results of the quantitative phase analysis of each XRD pattern and, therefore, the evolution of mineral phases depending on mode of cooling.

Fig. 2. Rietveld fit to the XRD pattern collected for the product of the heat/quench experiment performed for SM4-5 where the sample was quenched from 1488 K during cooling, and where the monoclinic SFCA-III structure was used in the refinement. Peaks/regions with significant unmatched intensity are annotated with *.

Figure 1(d) shows that the concentration of both SFCA-III and magnetite decrease from the 1488 K quenched sample to the 1398 K sample and the furnace cooled sample; 80 to 59 and 53 mass% (relative crystalline mass%) for SFCA-III and 15, 5 and 3 mass% for magnetite. The concentration of larnite across the three samples varies between 4 and 6 mass%. Peaks for neither SFCA nor SFCA-I were observed in the XRD pattern for the 1488 K quenched sample but were clearly present in the patterns for the 1398 K quenched sample and the furnace cooled sample. The concentration of both SFCA (20 mass%) and SFCA-I (20 mass%) was higher in the furnace-cooled sample compared to the 1398 K sample (14 and 16 mass%, respectively).

The formation of SFCA-I during cooling after melting is notable because SFCA-I has generally been regarded as forming only during the heating stage of the sintering process.²⁵⁾ Oxygen partial pressures during the sintering process vary between reducing as the coke burns and the flame front passes through the sinter bed, to oxidising during cooling.²⁶⁾ SFCA-I has been shown to accommodate more Fe²⁺ in its crystal structure than SFCA²⁷⁾ and so the use of the pO₂ = 5×10⁻³ atm throughout the entire heating and cooling regimes in these experiments is likely to have aided the formation of SFCA-I during cooling. However, we also note that in their study Kahlenberg et al. observed the formation of SFCA-I in their product material after partial melting of their synthetic pellet when synthesis was conducted in air at 1573 K.⁵⁾ Zoll et al. also noted the formation of SFCA-I in product material obtained after the onset melting for a pellet heated at 1723 K in air. Other work has described the morphology of magnetite crystals changing from idiomorphic when part of a phase assemblage consisting of magnetite + liquid only at 623 K, to irregular shapes after cooling to room temperature, indicating that the formation of SFCA phases consumed magnetite during the cooling process and with the crystal structures of the SFCA phases containing spinel structural modules.²⁸⁾ Nicol et al. also reported the formation of phase(s) with SFCA-I structure forming during cooling of an Al₂O₃ composition (i.e., in the SFC system) from a liquid phase at 1623 K.^29,30) The evidence, therefore, that the formation of SFCA-I after melt formation and during cooling does occur is becoming overwhelming, and logically this appears consistent with SFCA-I being intermediate between SFCA-III and SFCA in terms of ordering of the spinel (S) and pyroxene (P) structural modules.

Figures 3(a) to 3(d) show the final Rietveld fits to the XRD patterns collected for the JSS851-5 CRM, UN016 (pot-grate sinter), UN154 (sinter strand) and UN011(sinter strand) samples, respectively. Figure 3(d) shows the results of the QPA for each sample. The inset in each of Figs. 3(a)–3(c) show that the low-angle peaks diagnostic of SFCA-III (triclinic) are clearly present in each pattern and, therefore, SFCA-III is present in each of these samples. In Fig. 3(d) for the UN011 sample these peaks are not evident. However, it was necessary to include the SFCA-III structure in the refinements to achieve a satisfactory fit to the XRD pattern and, therefore, SFCA-III is present in this sample, but at a small amount. These results demonstrate, therefore, that SFCA-III is found in industrial sinter strand and pot-grate sinter samples; at 20, 14 and 7 mass% (relative crystalline mass%) for the UN016, UN154 and UN011 samples, respectively. SFCA-III is also abundant in the JSS851-5 CRM (20 mass%). It should be noted that SFCA-III was erroneously identified as mordenite in the original round robin paper.¹¹⁾ At that time the SFCA-III structures had not yet been published, and the FCAM-III phase/structure of Zoll et al.⁶⁾ although published had not yet been entered into the International Centre for Diffraction Data (ICDD) database (now ICDD entry no. 04-024-9709, entered 09/01/2020).

Fig. 3. Rietveld fits to the XRD patterns collected for the (a) JSS851-5, (b) UN016, (c) UN154 and (d) UN011 samples. The black curve in each plot shows the contribution of SFCA-III (triclinic) to the overall calculated model. The inset in each plot shows a magnified view of the region between 6 and 13° 2θ. (e) shows the results of the quantitative phase analysis of each XRD pattern.

4. Discussion

The relatively recent discovery of triclinic SFCA-III by Kahlenberg et al.⁵⁾ and FCAM-III by Zoll et al.,⁶⁾ and now here the confirmation that phases with this crystal structure are present in sinter strand and pot-grate sinters, is an exciting development for the iron ore sintering research and characterisation. Further fundamental understanding of SFCA-III characteristics, stability, and how it affects common sinter quality parameters such as strength, reducibility, and strength degradation during reduction, should provide significant opportunity for future research. For example, the composition of the Fe-rich SFCA/SFCA-III reported by Webster et al. (58.9 mass% Fe, 6.9 mass% Ca, 0.8 mass% Si and 3.0 mass% Al)⁷⁾ is markedly different to the composition of the SFCA-III reported by Kahlenberg et al. (48.9 mass% Fe, 6.8 mass% Ca, 0.7 mass% SiO₂, 7.0 mass% Al and 3.7 mass% Mg)⁵⁾ and the FCAM-III reported by Zoll et al. (41.16 mass% Fe, 5.76 mass% Ca, 15.21 mass% Al and 3.07 mass% Mg).⁶⁾ Furthermore, the Rietveld-refined unit cell volume of the SFCA-III phase in each of the 1488 K quenched, 1398 K quenched and furnace cooled samples (1.398, 1.395 and 1.397 nm³, respectively) reported here, and for each of the JSS851-5, UN016, UN154 and UN011 samples (1.405, 1.405, 1.402 and 1.406 nm³, respectively) reported here, differ markedly from the unit cell volume of the published SFCA-III (1.370 nm³) and FCAM-III (1.324 nm³) crystal structures.^5,6) This indicates that whilst the crystal symmetry is the same as the published structures, the elemental composition of the SFCA-III phase in each of the samples reported here differs from those of the published structures. This all suggests a wide solid solution range for the SFCA-III phase, which requires a systemic investigation to establish its boundaries. We do, however, note similarity of the Ca and Si concentrations in the Webster et al. and Kahlenberg et al. materials, and this too is worthy of further investigation.

Future work could also involve rigorous investigation into whether the SFCA-III structure accommodates more Fe²⁺ in its structure than SFCA-I and/or SFCA. Given, based on the current work and the results of various in-situ XRD investigations, this phase appears the first to crystallise from the melt during cooling, logic might suggest it would accommodate more Fe²⁺ in its structure. Consistent with this hypothesis is the theoretical analysis performed by Zoll et al. which suggests that higher oxygen fugacity favours the stability SFCA-I.⁶⁾ However, the Ca_2.99Mg_2.67Fe³⁺_14.58Fe²⁺_0.77Al_4.56Si_0.43O₃₆ SFCA-III compound synthesised by Kahlenberg et al. is calculated based on the chemical formula to contain 1.07 mass% FeO, which is lower than the Ca_3.2Al_1.3Fe³⁺_14.7Fe²⁺_0.8O₂₈ SFCA-I compound synthesised by Mumme et al. (1.37 mass% FeO),¹⁸⁾ the several new SFCA-I compounds synthesised by Webster et al. (1.83 to 2.17 mass% FeO, determined by titration),²⁷⁾ and all but one of the new SFCA-I compounds synthesised by Kahlenberg et al. (1.12 to 1.20 mass% FeO).³¹⁾

5. Conclusions

In this study we have shown that the previously reported Fe-rich SFCA and SFCA-III phases have the same crystal structure. Therefore, the four principal members of the silico-ferrite of calcium and aluminium (SFCA) family are SFCA, SFCA-I, SFCA-II, and SFCA-III. In addition, the presence of SFCA-III phase in industrial sinter is confirmed for the first time. Furthermore, we have provided additional evidence that the SFCA-I phase – which has generally been regarded as forming only during the heating stage of the sintering process – does form during cooling. This is consistent with SFCA-I being intermediate between SFCA-III and SFCA in terms of ordering of the spinel (S) and pyroxene (P) structural modules, with the SFCA-III phase appearing to be the first Ca-rich ferrite phase to crystallise from the melt during cooling.

Acknowledgements

Tom Honeyands (University of Newcastle, Australia) is thanked for provision of samples as part of the sinter analysis round robin.

References

• 1)   I.  Tonžetić and  A.  Dippenaar: Miner. Eng., 24 (2011), 1258. https://doi.org/10.1016/j.mineng.2011.04.012
• 2)   L.  Lu and  O.  Ishiyama: Iron Ore: Mineralogy, Processing and Environmental Sustainability, ed. by L. Lu, Woodhead Publishing, Cambridge, (2015), 393. ISBN978-1782421566
• 3)   S.  Nicol,  J.  Chen,  M. I.  Pownceby and  N. A. S.  Webster: ISIJ Int., 58 (2018), 2157. https://doi.org/10.2355/isijinternational.ISIJINT-2018-203
• 4)   W. G.  Mumme and  R. W.  Gable: N. Jb. Miner. Abh. (J. Min. Geochem.), 195 (2018), 89. https://doi.org/10.1127/njma/2018/0090
• 5)   V.  Kahlenberg,  H.  Krüger and  V. S.  Goettgens: Acta Crystallogr. B, 75 (2019), 1126. https://doi.org/10.1107/S2052520619014380
• 6)   K.  Zöll,  V.  Kahlenberg,  H.  Krüger and  P.  Tropper: J. Solid State Chem., 258 (2018), 307. https://doi.org/10.1016/j.jssc.2017.09.006
• 7)   N. A. S.  Webster,  M. I.  Pownceby,  I. C.  Madsen and  J. A.  Kimpton: Metall. Mater. Trans. B, 43 (2012), 1344. https://doi.org/10.1007/s11663-012-9740-5
• 8)   N. A. S.  Webster,  M. I.  Pownceby,  R.  Pattel,  J. R.  Manuel and  J. A.  Kimpton: ISIJ Int., 59 (2019), 263. https://doi.org/10.2355/isijinternational.ISIJINT-2018-573
• 9)   N. A. S.  Webster,  M. I.  Pownceby,  R.  Pattel,  J. R.  Manuel and  J. A.  Kimpton: ISIJ Int., 59 (2019), 1007. https://doi.org/10.2355/isijinternational.ISIJINT-2018-701
• 10)   R.  Murao and  M.  Kimura: ISIJ Int., 62 (2022), 1159. https://doi.org/10.2355/isijinternational.ISIJINT-2021-559
• 11)   T.  Honeyands,  J.  Manuel,  L.  Matthews,  D.  O’Dea,  D.  Pinson,  J.  Leedham,  G.  Zhang,  H.  Li,  B.  Monaghan,  X.  Liu,  E.  Donskoi,  N. A. S.  Webster and  M. I.  Pownceby: Minerals, 9 (2019), 333. https://doi.org/10.3390/min9060333
• 12)   L.-H.  Hsieh and  J. A.  Whiteman: ISIJ Int., 29 (1989), 24. https://doi.org/10.2355/isijinternational.29.24
• 13)   N. A. S.  Webster,  M. I.  Pownceby,  I. C.  Madsen,  A. J.  Studer and  J. A.  Kimpton: Metall. Mater. Trans. B, 45 (2014), 2097. https://doi.org/10.1007/s11663-014-0137-5
• 14)   N. A. S.  Webster,  M. I.  Pownceby,  I. C.  Madsen and  J. A.  Kimpton: ISIJ Int., 53 (2013), 774. https://doi.org/10.2355/isijinternational.53.774
• 15)   N. A. S.  Webster,  M. I.  Pownceby,  J. R.  Manuel,  R.  Pattel and  J. A.  Kimpton: JOM, 73 (2021), 299. https://doi.org/10.1007/s11837-020-04430-y
• 16)   N. A. S.  Webster and  M. I.  Pownceby: ISIJ Int., 62 (2022), 1624. https://doi.org/10.2355/isijinternational.ISIJINT-2021-589
• 17)   D. C.  Liles,  J. P. R.  de Villiers and  V.  Kahlenberg: Mineral. Petrol., 110 (2016), 141. https://doi.org/10.1007/s00710-015-0411-5
• 18)   W. G.  Mumme,  J. M. F.  Clout and  R. W.  Gable: N. Jb. Miner. Abh., 173 (1998), 93. https://doi.org/10.1127/njma/173/1998/93
• 19)   W. C.  Hamilton: Phys. Rev., 110 (1958), 1050. https://doi.org/10.1103/PhysRev.110.1050
• 20)   C. M.  Midgley: Acta Crystallogr., 5 (1952), 307. https://doi.org/10.1107/S0365110X52000964
• 21)   R.  Blake,  R.  Hessevick,  T.  Zoltai and  L.  Finger: Am. Mineral., 51 (1966), 123. https://pubs.geoscienceworld.org/msa/ammin/article/51/1-2/123/540239/Refinement-of-the-hematite-structure, (accessed 2023-10-01).
• 22)   G. A.  Lager,  J. D.  Jorgensen and  F. J.  Rotella: J. Appl. Phys., 53 (1982), 6751. https://doi.org/10.1063/1.330062
• 23)   W. A.  Dollase: Z Kristallogr., 121 (1965), 369. https://doi.org/10.1524/zkri.1965.121.16.369
• 24)   V.  Kahlenberg and  R. X.  Fisher: European Journal of Mineralogy, 12 (2000), 129. https://doi.org/10.1127/0935-1221/2000/0012-0129
• 25)   T.  Harvey,  M. I.  Pownceby,  J.  Chen,  N. A. S.  Webster,  T. B. T.  Nguyen,  L.  Matthews,  D.  O’Dea and  T.  Honeyands: JOM, 73 (2021), 345. https://doi.org/10.1007/s11837-020-04452-6
• 26)   L.-H.  Hsieh and  J. A.  Whiteman: ISIJ Int., 29 (1989), 625. https://doi.org/10.2355/isijinternational.29.625
• 27)   N. A. S.  Webster,  M. I.  Pownceby and  N. C.  Wilson: ISIJ Int., 62 (2022), 1618. https://doi.org/10.2355/isijinternational.ISIJINT-2021-577
• 28)   X.  Chen,  W.  Wang,  D.  Yang,  H.  Zheng and  B.  Wang: ISIJ Int., 63 (2023), 261. https://doi.org/10.2355/isijinternational.ISIJINT-2022-348
• 29)   S.  Nicol,  E.  Jak and  P. C.  Hayes: Metall. Mater. Trans. B, 50 (2019), 3027. https://doi.org/10.1007/s11663-019-01689-8
• 30)   S.  Nicol,  S.  Cheng,  E.  Jak and  P. C.  Hayes: ISIJ Int., 60 (2020), 2659. https://doi.org/10.2355/isijinternational.ISIJINT-2020-182
• 31)   V.  Kahlenberg,  H.  Krüger and  M.  Tribus: Phys. Chem. Miner., 48 (2021), 25. https://doi.org/10.1007/s00269-021-01148-4