Fundamentals of the Reduction of SFC in CO/CO₂ Gas at 1000°C

Abstract

SFC (Sillico-Ferrite of Calcium) is a kind of simplified SFCA (Sillico-Ferrite of Calcium and Aluminum) with no Al₂O₃. SFC and SFCA are believed to be the most desirable bonding phase in sinter. In order to better understand the fundamentals of the reduction of SFC, a series of experiments on the SFC reduction were carried out in the present work, including phase equilibria tested by XRD, morphology tested by SEM-EDS, and reduction pathway under the different CO/CO₂ mixture gas at 1000°C. The experimental results indicated, (1) in the case of CO = 20% and 40%, most of Fe₂O₃ in SFC was reduced to FeO. The equilibrium phases were FeO, CaO·Fe₂O₃, and CaO·SiO₂. (2) In the case of CO = 60%, CaO·Fe₂O₃ was reduced to generate FeO and 2CaO·Fe₂O₃. The equilibrium phases were FeO, 2CaO·Fe₂O₃, and CaO·SiO₂. (3) In the case of CO = 80% and 90%, FeO was reduced to Fe, and 2CaO·Fe₂O₃ was reduced to generate Fe and CaO. The equilibrium phases were Fe, CaO, and CaO·SiO₂. The findings from this work may provide guidelines for the improvement of sintering production and blast furnace performances.

1. Introduction

Iron ore less than 5 mm in size is too fine to direct use in a blast furnace (BF) and must be agglomerated into iron ore sinter. The quality of iron ore sinter is critical for efficient BF operation.^1,2) During the sintering process, the loose raw minerals are converted into porous but physically strong sintered ore by some Ca-rich and ferrite like bonding phase.³⁾ As we know, SFCA phase (Sillico-Ferrite of Calcium and Aluminum) and SFC (Sillico-Ferrite of Calcium, a kind of simplified SFCA with no Al₂O₃) are believed to be the most desirable bonding phases in sinter because of its high reducibility,⁴⁾ high mechanical strength,^5,6) and low reduction degradation.⁷⁾ Therefore, the quality of SFCA and SFC has been widely reported, including the chemical formula,^8,9) crystal structure,^10,11) thermal stability,^12,13) mineral phase,¹⁴⁾ properties of liquid phase,^15,16,17) and factors affecting them.^18,19,20)

Besides above properties, the reduction fundamentals are also important to improve the reducibility of iron ores. There are some reports about CaO–Fe₂O₃ binary system. Ganguly²¹⁾ reported the phase equilibria and reaction pathway of CaO·Fe₂O₃ (CF) and 2CaO·Fe₂O₃ (C2F) in CO/CO₂ mixture gas at 800°C. Schurmann et al.²²⁾ reported the phase equilibria and reaction pathway of CF and C2F in CO/CO₂ mixture gas at 1000°C. So, the fundamentals of reduction in CaO–Fe₂O₃ binary system have been investigated well. However, to the best knowledge of the authors, the fundamentals of reduction in CaO–Fe₂O₃–SiO₂ ternary system (SFC) and CaO–Fe₂O₃–SiO₂–Al₂O₃ quaternary system (SFCA) are not well understood so far because of the lack of sufficient and systematical study on single SFC or SFCA phase.

Better understanding the reduction fundamentals of SFC is a key point to explore the reduction fundamentals of SFCA. Therefore, based on the research in CaO–Fe₂O₃ binary system, a series of experiments on SFC reduction were carried out in the present work. The originalities of this investigation include: phase equilibria, morphology, and reduction pathway under the different CO/CO₂ mixture gas at 1000°C. The findings from this work may provide guidelines for the improvement of sintering production and blast furnace performances.

2. Experimental

The following experimental procedures were implemented:²³⁾

(1) Raw materials. Pure chemical reagents were used, including Fe₂O₃, CaO, and SiO₂.

(2) Chemical design of SFC. SFC has no fixed composition. It is a compound in the system of CaO·3Fe₂O₃-4CaO·3SiO₂ (shorten as CF3-C4S3). Pownceby gave a XRD pattern of SFC with 9% of C4S3.¹²⁾ Therefore, SFC with 9% of C4S3 was selected in the present work. Thus, Fe₂O₃ content is 81.49%, CaO content is 14.50%, and SiO₂ content is 4.01%.

(3) Sample preparation. The weighted chemical reagents were homogeneously mixed in an agate mortar. Then, they were pressed into briquettes (15 mm in diameter and 1 mm in height) by using a briquetting machine at a pressure of 18 MPa for 2 min. The weight of each sample was about 1 g.

(4) Calcination. The SFC sample is located on an Alumina slice. Another calcined SFC is inserted between SFC sample and Alumina slice to avoid the contamination of SFC sample by the Alumina slice (Fig. 1). Then the samples were heated at 1200°C for 6 hours. From Fig. 2, the XRD pattern of SFC sample prepared in the present work (Fig. 2(a)) is consistent well with standard SFC peak (Fig. 2(b)). Thus, the SFC samples were obtained and would be used in the reduction experiments.

Fig. 1.

Schematics of calcination for SFC sample.

Fig. 2.

XRD pattern of SFC sample prepared in the present work.

(5) Reduction experiments. In the present work, the SFC samples were reduced by CO/CO₂ mixture gas in a Tammann furnace, and the CO content in the mixture gas was set as 20%, 40%, 60%, 80%, and 90%. The reducibility indices (RI) of SFC samples can be calculated by Eq. (1).   

$$RI(\%)=\left(\frac{0.111W_{FeO}}{0.430W_{TFe}}+\frac{m_0-m_t}{m_0\times 0.430W_{TFe}}\times 100\right)\times 100$$(1)

where, W_FeO represents the FeO content in unreduced SFC sample (mass%), FeO = 0% in the present work. W_TFe represents the total Fe content in unreduced SFC sample (mass%), TFe = 57.04% in the present work. m₀ represents the weight of unreduced SFC sample (g), and m_t represents the weight of reduced SFC sample (g).

Based on the actual blast furnace, most of iron ores are reduced in the lump zone, and the temperature ranges from 800°C to 1000°C. So, 1000°C was selected as the reduction temperature in the present work. The temperature was measured by a thermocouple which fixed near to the SFC samples.

In the present study, it is quite important to make sure the reduced SFC samples reached to equilibrium states. Therefore, it is significant to determine the reducing time. Based on our preliminary tests, in the case of CO = 20%, the change in the reduced SFC sample weight and reducibility index (RI) with reducing time is shown in Fig. 3. From the figure, one can conclude that the weight of reduced SFC sample basically will not change after 30 min. This means the reduction reactions have finished, and the phases in reduced SFC sample reach to equilibrium state. Thus, 1 h was selected as the reducing time in the present work. The SFC samples were reduced by CO/CO₂ mixture gas at 1000°C for 1 h. After calcination, the SFC samples were placed in a container with cooled Ar gas flow, and then the samples were quenched.

Fig. 3.

Change in the reduced SFC sample weight loss and reducibility index (RI) with reducing time in the case of CO = 20%.

(6) Phase equilibria tests. A part of the quenched reduced SFC samples was ground into a fine powder using an agate mortar and pestle, and then sieved completely through a sieve with 45 μm pores. Then, the powder XRD analyses were carried out using a X-ray diffractometer. Cu Kα was used as the radiation source (40 kV, 400 mA) with a graphite monochromator in the diffraction beam path. The XRD data were collected by using a continuous scanning mode, of which scanning speed was maintained at 10°/min.

(7) Morphology tests. The remaining parts of the quenched reduced SFC samples were polished by setting them into an ethylenediamine-doped epoxy resin for the preparation of SEM-EDS analyses. SEM was performed using a scanning electron microscope. The accelerating voltage was 20 kV. Energy-dispersive spectroscopy (EDS) was performed on the same instrument.

3. Experimental Results

3.1. Phase Equilibria

First, in order to understand the phase equilibria in the reduced SFC, XRD analyses for the reduced SFC samples were carried out. Most of the phases in Fe₂O₃–CaO–SiO₂ system were considered in the present work, and their lattice parameters are listed in Table 1. Figures 4, 5, 6, 7, 8 show the XRD patterns of the SFC sample reduced by CO/CO₂ mixture gas with different CO content at 1000°C for 1 h. In each figure, the first curve was the X-ray diffraction pattern of each reduced SFC sample, and the others are the standard peaks of the phases.

Table 1. Lattice parameters of phases in Fe₂O₃–CaO–SiO₂ system.

Phase	Chemical formula	Index No.	Crystal structure	Lattice parameter (A)
Iron24)	Fe	01-085-1410	Cubic	a = 2.8860
Wustite25)	FeO	00-003-0968	Cubic	a = 4.294
Magnetite26)	Fe3O4	00-002-1035	Cubic	a = 8.4100
Hematite27)	Fe2O3	00-001-1053	Rhombohedral	a=5.0280 c = 13.7300
Lime23)	CaO	01-077-2376	Cubic	a = 4.8080
Calcium ferrite28)	CaO·Fe2O3	03-065-1333	Orthorhombic	a = 9.2300 b = 3.0240 c = 10.7050
Dicalcium ferrite29)	2CaO·Fe2O3	00-011-0675	Orthorhombic	a = 5.6400 b = 14.6800 c = 5.3900
Calcium iron oxides30)	CaO·FeO·Fe2O3	01-072-0890	Orthorhombic	a = 3.0210 b = 10.0090 c = 12.6430
Calcium iron oxides30)	CaO·2FeO·Fe2O3	01-072-0891	Orthorhombic	a = 3.0500 b = 9.9860 c=15.3210
Calcium iron oxides30)	CaO·3FeO·Fe2O3	01-072-0892	Orthorhombic	a = 3.0520 b = 10.0410 c = 17.9660
Calcium silicate31)	CaO·SiO2	00-034-0558	unknown	–
Dicalcium silicate23)	2CaO·SiO2	00-033-0303	Orthorhombic	a = 10.9600 b = 18.4300 c = 6.8600
Rankinite32)	3CaO·2SiO2	00-003-1105	Monoclinic	a = 10.5500 b = 8.8800 c = 7.8500

Fig. 4.

XRD pattern of SFC sample reduced by CO/CO₂ mixture gas with CO = 20%.

Fig. 5.

XRD pattern of SFC sample reduced by CO/CO₂ mixture gas with CO = 40%.

Fig. 6.

XRD pattern of SFC sample reduced by CO/CO₂ mixture gas with CO = 60%.

Fig. 7

XRD pattern of SFC sample reduced by CO/CO₂ mixture gas with CO = 80%.

Fig. 8.

XRD pattern of SFC sample reduced by CO/CO₂ mixture gas with CO = 90%.

From these figures, it can be observed that SFC was reduced to Fe (metallic iron), CaO (lime), and CaO·SiO₂ (calcium silicate). (1) In the cases of CO = 20% and 40% (Figs. 4 and 5), the major phases were FeO (wustite), CaO·SiO₂, CaO·Fe₂O₃ (calcium ferrite). (2) In the case of CO = 60% (Fig. 6), the FeO and CaO·SiO₂ were still the major phases, the other major phase was 2CaO·Fe₂O₃ (dicalcium ferrite) instead of CaO·Fe₂O₃. (3) In the case of CO = 80% and 90% (Figs. 7 and 8), only CaO·SiO₂ was still one of the major phases, the other major phases changed from FeO and 2CaO·Fe₂O₃ to Fe and CaO.

3.2. Morphology

Aimed to better understand the microstructure of reduced SFC, the SEM-EDS analyses were carried out. SEM images are shown in Figs. 9, 10, 11, 12, 13. The main elements contents tested by EDS in the selected points in these figures are listed in Table 2. From the figures and table, the following results can be obtained.

Fig. 9.

SEM-EDS analyses of SFC sample reduced by CO/CO₂ mixture gas with CO =20%.

Fig. 10.

SEM-EDS analyses of SFC sample reduced by CO/CO₂ mixture gas with CO = 40%.

Fig. 11.

SEM-EDS analyses of SFC sample reduced by CO/CO₂ mixture gas with CO = 60%.

Fig. 12.

SEM-EDS analyses of SFC sample reduced by CO/CO₂ mixture gas with CO = 80%.

Fig. 13.

SEM-EDS analyses of SFC sample reduced by CO/CO₂ mixture gas with CO = 90%.

Table 2. Elements contents in the selected points in Figs. 9, 10, 11, 12, 13 (“–”: the amount of element was too small to test).

CO, Vol%	Point	Element	Mass%	Mol%	Point	Element	Mass%	Mol%
20	A	O	20.91	47.45	B	O	30.31	57.69
		Fe	75.56	49.11		Fe	51.45	27.98
		Ca	2.90	2.63		Ca	16.90	12.87
		Si	0.63	0.81		Si	1.34	1.46
40	C	O	22.44	49.22	D	O	27.04	53.60
		Fe	71.69	45.05		Fe	54.82	31.05
		Ca	4.29	3.76		Ca	15.31	12.14
		Si	1.58	1.98		Si	2.83	3.21
60	E	O	21.09	47.21	F	O	26.70	51.83
		Fe	71.76	46.01		Fe	44.95	24.93
		Ca	6.09	5.44		Ca	24.71	19.19
		Si	1.06	1.35		Si	3.65	4.05
80	G	O	1.97	6.43	H	O	31.83	54.17
		Fe	94.60	88.22		Fe	18.79	9.14
		Ca	2.48	3.24		Ca	38.83	26.43
		Si	1.13	2.11		Si	10.55	10.26
90	I	O	–	–	J	O	35.69	58.33
		Fe	100.00	100.00		Fe	14.95	6.98
		Ca	–	–		Ca	40.71	26.61
		Si	–	–		Si	8.65	8.08

(1) In the case of CO = 20% and 40% (Figs. 9 and 10), the phases structures are similar. There are three major phases: light grey phase (represented by Point A and Point C), middle grey phase (represented by Point B and Point D), and deep grey phase. Based on the EDS tests, the main elements of Point A and Point C are Fe and O, and O:Fe ≈ 1:1 (mol:mol). Therefore, the light grey phase should be wustite phase (FeO). Based on the area scan, element Fe and Ca distribute in the middle grey phase, and Fe:Ca ≈ 2:1 (mol:mol) at Point B and Point D. Therefore, combined with the XRD results (Figs. 4 and 5), the middle grey phase should be CaO·Fe₂O₃ (calcium ferrite). Moreover, the deep grey phase is CaO·SiO₂ (calcium silicate) based on the distribution of Ca and Si in the phase.

(2) In the case of CO = 60% (Fig. 11), there still are three major phases: light grey phase (represented by Point E), middle grey phase (represented by Point F), and deep grey phase. Based on EDS analyses and element distribution, it can be observed that the light grey phase and deep grey phase are same as CO = 20% and 40%, and they are wustite and CaO·SiO₂ respectively. However, in the middle grey phase, Fe:Ca ≈ 1:1 (mol:mol). This indicated grey phase transformed from CaO·Fe₂O₃ (calcium ferrite) to 2CaO·Fe₂O₃ (dicalcium ferrite) due to the higher reducing potential in the gas.

(3) In the case of CO = 80% and 90% (Figs. 12 and 13), basically there are two major phase. One is the white phase (represented by Point G and Point I), the other one is the deep grey phase (represented by Point H and Point J). Based on the EDS analyses in Figs. 12 and 13, most of the wustite has been reduced to metallic iron, and the white phase should be metallic iron. The deep grey phase should be the mixture of CaO·SiO₂ and CaO (CaO·SiO₂ and CaO can not be distinguished in the present micrographs).

4. Discussion on the Reduction Pathway of SFC

Based on above morphology analyses, the equilibrium phases are summarized in Table 3. Also, the weight loss and reducibility index (RI) by theoretical calculation and calculated by using the experimental data are listed in Table 3 too. From the table, one can conclude that the values of weight loss and RI in both theoretical and experimental are comparable, which can verify the morphology analyses in Figs. 9, 10, 11, 12, 13.

Table 3. Weight loss and RI at different CO content and the value of Fe/(Fe + O + Ca + Si) in each phase.

CO vol%	Components	Fe/(Fe+O+Ca+Si) mass%	Theoretical		Experimental
			Weight loss mass%	RI %	Weight loss mass%	RI %
0	SFC	57.04	0	0	0	0
20	Wustite (FexO)	77.77	5.08	20.71	5.02	20.47
	CaO·SiO2	0.00
	CaO·Fe2O3	51.85
40	Wustite (FexO)	77.77	5.08	20.71	5.12	20.87
	CaO·SiO2	0.00
	CaO·Fe2O3	51.85
60	Wustite (FexO)	77.77	6.62	26.98	6.44	26.26
	CaO·SiO2	0.00
	2CaO·Fe2O3	41.18
80	Fe	100	24.45	100	23.60	96.21
	CaO·SiO2	0.00
	CaO	0.00
90	Fe	100	24.45	100	23.95	97.64
	CaO·SiO2	0.00
	CaO	0.00

Then, the reduction pathway of SFC can be systematically summarized in Fig. 14. In the figure, horizontal axis is the mass ratio of Fe to Fe + O + Ca + Si (Fe/(Fe + O + Ca + Si)), and the value of each phase is also listed in Table 3. The vertical axis is the volume percentage of CO in CO/CO₂ mixture gas. The points in the figure are the stable phase, and the red bold arrow lines are the reduction pathway of SFC. The dotted lines are the transform lines. It should be noted that the accurate CO contents for the transformations (the vertical values of transform lines) can not be determined so far, because the limited CO contents in reducing gas were experimentally investigated in the present work. Therefore, the transform lines are dotted. Also, the accurate equilibrium CO contents for the reactions (the vertical values of transform lines) will be investigated in our future work.

Fig. 14.

Schematic diagram of the reduction pathway of SFC.

(1) When CO<20%, the major iron oxide is FeO (wustite), and other SFC decomposes to CaO·SiO₂ and CaO·Fe₂O₃. This is comparable with the equilibrium phase diagram between Fe, FeO, Fe₃O₄, and CO/CO₂ mixture gas (Fig. 15). From the figure, most of Fe₂O₃ is reduced to FeO, and FeO is stable phase (Point N in Fig. 15). The un-reacted Fe₂O₃ presents as CaO·Fe₂O₃ combined with CaO.

Fig. 15.

Equilibrium phase diagram between FeO, FeO, Fe₃O₄, and CO/CO₂ mixture gas.

(2) In the case of 20%<CO<40%, the stable iron oxide is still FeO based on Fig. 15 (Point N→O), and the major phase is still FeO. Also, the CaO·Fe₂O₃ can not be reduced in this reducing potential. Therefore, no reaction occurs from CO = 20% to 40%.

(3) In the case of 40%<CO<60%, the stable iron oxide is still FeO based on Fig. 15 (Point O→P), and the major phase is still FeO. In this reducing potential, CaO·Fe₂O₃ is reduced to generate FeO and 2CaO·Fe₂O₃. Therefore, in Fig. 14, both Point D and Point G represent FeO, but the amount of FeO in Point G is more than that in Point D.

(4) In the case of 60%<CO<80%, the stable phase in Fig. 15 changed from FeO to Fe (Point P→Q). Therefore, metallic Fe is the major phase for CO = 80%. Also, in this reducing potential, the 2CaO·Fe₂O₃ is reduced to generate Fe and CaO. It should be noted that the values of Fe/(Fe + O + Ca + Si) in Point I and Point K are same (zero) in Fig. 14. But Point I is single phase (CaO·SiO₂), while Point K is mixed phase (CaO·SiO₂ and CaO).

(5) In the case of 80%<CO<90%, the stable iron oxide is still metallic Fe based on Fig. 15 (Point Q→R). All of the reduction reactions have finished, and no reaction occurs in this case. The equilibrium phases are metallic Fe, CaO and CaO·SiO₂.

It should be noted that, in the present work, the reduction experiments were carried out at the constant temperature under fixed reducing gas compositions. While, the reductions in an actual blast furnace (BF) proceed under the unsteady state conditions of temperature and gas composition. The reduction behaviors between both conditions are different. Basically, the reduction rate is slower in an actual BF, and the phases are difficult to reach equilibrium state. The reasons are, (1) as the iron ores descend from the top of furnace to the cohesive zone, the ores are heated from ambient temperature to higher temperature by the heated reducing gas. The lower temperature in the upper part of furnace is not beneficial for the reduction of iron ores. (2) Compared with lower part of BF, the CO content in reducing gas in the upper part is lower, and the reducing potential is lower too. (3) The size of iron ores in BF is bigger (5–40 mm and above), the heat transfer and mass transfer need more time. Therefore, reductions of iron ores in an actual BF are difficult to reach equilibrium state in the limited time of BF process.

5. Conclusions

In this work, the fundamentals of the reduction of SFC in CO/CO₂ gas at 1000°C were investigated. The main findings can be summarized as follows.

(1) In the case of CO = 20% and 40%, most of Fe₂O₃ in SFC was reduced to FeO. The equilibrium phases were FeO, CaO·Fe₂O₃, and CaO·SiO₂.

(2) In the case of CO = 60%, CaO·Fe₂O₃ was reduced to generate FeO and 2CaO·Fe₂O₃. The equilibrium phases were FeO, 2CaO·Fe₂O₃, and CaO·SiO₂.

(3) In the case of CO = 80% and 90%, FeO was reduced to Fe, and 2CaO·Fe₂O₃ was reduced to generate Fe and CaO. The equilibrium phases were Fe, CaO, and CaO·SiO₂.

Acknowledgement

The financial supports of National Science Foundation of China (NSFC 51874080, NSFC 51774071, and NSFC 51974073), Initiation Foundation of Doctoral Research of Liaoning (2019-BS-172) and Natural Science Foundation of Liaoning (2019-MS-132) are much appreciated.

References

• 1)   M.  Naito,  K.  Takeda and  Y.  Matsui: ISIJ Int., 55 (2015), 7.
• 2)   M.  Gan,  X. H.  Fan,  Z. Y.  Ji,  X. L.  Chen,  L.  Yin,  T.  Jiang,  G. H.  Li and  Z. Y.  Yu: ISIJ Int., 55 (2015), 742.
• 3)   S.  Nicol,  J.  Chen,  M. I.  Pownceby and  N. A. S.  Webster: ISIJ Int., 58 (2018), 2157.
• 4)   B. Y.  Cai,  T.  Watanabe,  C.  Kamijo,  M.  Susa and  M.  Hayashi: ISIJ Int., 58 (2018), 642.
• 5)   H. S.  Han,  F. M.  Shen,  X.  Jiang,  C. G.  Bi,  H. Y.  Zheng and  Q. J.  Gao: J. Iron Steel Res. Int., 26 (2019), 1171.
• 6)   X.  Jiang,  L.  Zhang,  G. S.  Li,  M. F.  Jin,  Z.  Wang,  Y. S.  Shen and  F. M.  Shen: J. Iron Steel Res. Int., 16 (2009), 253.
• 7)   M.  Mizutani,  T.  Nishimura,  T.  Orimoto,  K.  Higuchi,  S.  Nomura,  K.  Saito and  E.  Kasai: ISIJ Int., 58 (2018), 1413.
• 8)   R.  Murao,  T.  Harano,  M.  Kimura and  I. H.  Jung: ISIJ Int., 58 (2018), 259.
• 9)   J. D. G.  Hamilton,  B. F.  Hoskins,  W. G.  Mumme,  W. E.  Borbidge and  M. A.  Montague: N. Jahrb. Mineral. Abh., 161 (1989), 1.
• 10)   R.  Mežibrický and  M.  Fröhlichová: ISIJ Int., 56 (2016), 1111.
• 11)   W. G.  Mumme,  J. M. F.  Clout and  R. W.  Gable: N. Jahrb. Mineral. Abh., 173 (1998), 93.
• 12)   M. I.  Pownceby and  T. R. C.  Patrick: Eur. J. Mineral., 12 (2000), 455.
• 13)   T. R. C.  Patrick and  M. I.  Pownceby: Metall. Mater. Trans. B, 33 (2002), 79.
• 14)   T.  Takayama,  R.  Murao and  M.  Kimura: ISIJ Int., 58 (2018), 1069.
• 15)   N. A. S.  Webster,  M. I.  Pownceby,  R.  Pattel,  J. R.  Manuel and  J. A.  Kimpton: ISIJ Int., 59 (2019), 1007.
• 16)   K. O.  Jang,  X. D.  Ma,  J. M.  Zhu,  H. F.  Xu,  G.  Wang and  B. J.  Zhao: Metall. Mater. Trans. B, 48 (2017), 1547.
• 17)   N. A. S.  Webster,  M. I.  Pownceby,  I. C.  Madsen,  A. J.  Studer,  J. R.  Manuel and  J. A.  Kimpton: Metall. Mater. Trans. B, 45 (2014), 2097.
• 18)   Z.  Wang,  D.  Pinson,  S.  Chew,  B. J.  Monaghan,  M. I.  Pownceby,  N. A. S.  Webster,  H.  Rogers and  G. Q.  Zhang: ISIJ Int., 56 (2016), 1138.
• 19)   H. M.  Long,  X. J.  Wu,  T. J.  Chun,  Z. X.  Di and  B.  Yu: Metall. Mater. Trans. B, 47 (2016), 2830.
• 20)   W.  Wang,  D. W.  Yang,  Z. L.  Ou-Yang,  R. S.  Xu and  M. M.  Song: Metall. Mater. Trans. B, 50 (2019), 678.
• 21)   S.  Ganguly: Ph.D. thesis, University of Queensland, Australia, (1991).
• 22)   E.  Schürmann and  G.  Kraume: Arch. Eisenhuttenwes., 47 (1976), 267.
• 23)   T. Y.  He: Bachelor thesis, Northeastern University, China, (2015) (in Chinese).
• 24)   H. E.  Swanson,  E.  Tatge and  R. K.  Fuyat: Phys. Today, 7 (1954), 22.
• 25)   R. W. G.  Wyckoff and  E. D.  Crittenden: J. Am. Chem. Soc., 47 (1925), 2876.
• 26)   H. V.  Anderson and  K. G.  Chesley: Am. J. Sci., 22 (1931), 103.
• 27)   J. D.  Hanawalt and  H. W.  Rinn: Ind. Eng. Chem. Anal. Ed., 8 (1936), 244.
• 28)   B. F.  Decker and  J. S.  Kasper: Acta Crystallogr., 10 (1957), 332.
• 29)   E. F.  Bertaut,  P.  Blum and  A.  Sagnières: Acta Crystallogr., 12 (1959), 149.
• 30)   O.  Evrard,  B.  Malaman,  F.  Jeannot,  A.  Courtois,  H.  Alebouyeh and  R.  Gerardin: J. Solid State Chem., 35 (1980), 112.
• 31)   A. E.  Ringwood and  A.  Major: Earth Planet. Sci. Lett., 12 (1971), 411.
• 32)   C. B.  Clark: J. Am. Ceram. Soc., 29 (1946), 25.