Non-isothermal Reduction Behavior and Mechanism of Hongge Vanadium Titanomagnetite Pellet with Simulated Shaft Furnace Gases

Abstract

As an integral part of developing a novel clean smelting process for the comprehensive utilization of Hongge vanadium titanomagnetite (HVTM), the non-isothermal reduction behavior and mechanism of HVTM pellet (HVTMP) were investigated using simulated shaft furnace gases of dry pulverized coal gasification (DPCG), water-coal slurry gasification (WCSG), Midrex, and HYL-III in the current study. The results showed that the reduction degree significantly increased with the decrease of heating rate. The reduction degree was found to increase in the order of DPCG<WCSG<Midrex<HYL-III. An approximately reversed linear relation could be concluded that the compressive strength of reduced HVTMP decreased as the reduction swelling index increased. The phase transformations of valuable elements under non-isothermal reduction conditions could be described as follows: Fe₂O₃→Fe₃O₄→FeO→Fe; Fe₉TiO₁₅→Fe_2.75Ti_0.25O₄→Fe₂TiO₄; (Fe_0.6Cr_0.4)₂O₃, Fe_0.7Cr_1.3O₃→FeCr₂O₄; (Cr_0.15V_0.85)₂O₃→Fe₂VO₄. However, under non-isothermal reduction conditions, SEM results indicated that the reduced metallic iron could not be connected together to form a uniform continuous area even at 1100°C. These results could provide both theoretical and technical basis for the comprehensive utilization of HVTM.

1. Introduction

China, Russia, South Africa, New Zealand, Canada, and Australia are rich in vanadium titanomagnetite (VTM) resources, which are complex iron ores and compose by the coexistence of vanadium and titanium.^1,2,3) The Hongge mineral in Panzhihua-Xichang area is the largest VTM resource of China.⁴⁾ Hongge vanadium titanomagnetite (HVTM) is an iron ore that not only containing iron, titanium, and vanadium, but also relatively rich in chromium. The reserve of chromium is 900 Mt, accounting for 68% of the chromium reserve in China.⁵⁾ Therefore, it is meaningful to study the utilization of this special mineral resource.

At present, VTM is used as the main material for the blast furnace process. Most of the iron and partly vanadium can be reduced into the hot metal.^6,7,8,9,10) However, almost all of the titanium is concentrated into the molten slag, and it is very hard to extract titanium components through traditional separation processes so far.¹¹⁾ Over past decades, several new processes on the basis of direct reduction containing coal to improve the utilization of VTM, HVTM, or ilmenite have been developed.^{12,13,14,15,16,17,18,19,20,21,22,23)} Hu et al.¹³⁾ investigated the enhanced reduction of coal-containing titanomagnetite concentrates briquette with a three-layer bed using an electric furnace containing three heating zones under argon atmosphere. Samanta et al.¹⁷⁾ studied the mineralogy and carbothermic reduction behavior of VTM in Eastern India. Zhao et al.²⁰⁾ surveyed the reduction behavior of vanadium and chromium during coal-based direct reduction of HVTM followed by magnetic separation. However, the recovery rates of titanium, vanadium, and chromium are still low among these processes. What is more, it should be noted that the coal-based direct reduction is relatively low efficient with high energy consumption and high operating temperature.²⁴⁾ Therefore, the resource of HVTM has not put into industrial production.

In order to improve the utilization efficiency of HVTM, one novel clean approach is the gas-based shaft furnace direct reduction of HVTM pellet (HVTMP) followed by melting separation of the reduced pellet to recover the iron, titanium, vanadium, and chromium, which has been proposed by the authors’ laboratory and greatly increases the recovery rates of valuable elements.²⁵⁾ Combined with available mature coal gasification technology, the gas-based shaft furnace direct reduction process in which the coal-formed gas substitutes for natural gas as reducing agent is proposed, opening up an effective way to develop gas-based shaft furnace direct reduction process in China where natural gas resource is deficient but noncoking coal resource is plentiful.²⁶⁾ Obviously, gas-based shaft furnace direct reduction is an essential procedure in this novel clean smelting process. In conjunction with this development work, we have studied the isothermal reduction behavior of HVTMP from 900 to 1050°C, simulating the reduction zone in gas-based shaft furnace direct reduction process.²⁷⁾ However, in the full-scale gas-based shaft furnace direct reduction process, the reduction of HVTMP is not always under a constant temperature condition from the beginning to the end. Thus, it is imperative to understand the reduction behavior and mechanism of HVTMP under non-isothermal conditions.

As an integral part of developing a novel clean smelting process of HVTM, the present investigation was to reveal the reduction behavior and mechanism of HVTMP under non-isothermal conditions using simulated gas compositions of dry pulverized coal gasification (DPCG), water-coal slurry gasification (WCSG), Midrex, and HYL-III. Meanwhile, the phase compositions and morphological changes of reduced HVTMP were intensively studied. The aim of this study is to obtain scientific direction for reactor design and operational parameter determination of shaft furnace, as well as provide both theoretical and technical basis for the comprehensive utilization of HVTM.

2. Experimental Section

2.1. Experimental Materials

The green pellets were prepared from a mixture of HVTM with 100% passing 0.074 mm, 8.5% water, and 1% bentonite in an experimental balling disc pelletizer. The green pellets produced were dried at 105°C in an oven for 5 h and roasted in a muffle furnace at a roasting temperature of 1200°C for 15 min according to our previous study.²⁸⁾ The average compressive strength of HVTMP was 2893 N. The particle diameter and weight of HVTMP were about 11.3 mm and 3.1 g on average, respectively. The chemical composition and X-ray diffraction (XRD) patterns of HVTMP are shown in Table 1 and Fig. 1, respectively. The main phases are hematite (Fe₂O₃) and hematite solid solution (Fe₉TiO₁₅). Vanadium and chromium exist in the solid-solutions of (Fe_0.6Cr_0.4)₂O₃, Fe_0.7Cr_1.3O₃, and (Cr_0.15V_0.85)₂O₃. The total content of H₂ and CO studied in the present study is about 90%, but the ratio of φ(H₂) to φ(CO) differs widely. The φ(H₂) and φ(CO) are the volume fractions of H₂ and CO in the mixed gases, respectively. The compositions of the reducing gases are shown in Table 2.

Table 1. Chemical composition of HVTMP (wt%).

TFe	FeO	CaO	SiO2	MgO	Al2O3	TiO2	V2O5	Cr2O3
54.40	0.81	0.73	4.20	2.43	2.38	9.11	0.61	1.48

Fig. 1.

XRD analysis of valuable elements of HVTMP.

Table 2. Gas compositions applied for non-isothermal reduction.²⁷⁾

Atmosphere	H2 (vol%)	CO (vol%)	CO2 (vol%)	N2 (vol%)
DPCG	25.7	64.3	5	5
WCSG	45	45	5	5
Midrex	54	36	5	5
HYL-III	64.3	25.7	5	5

2.2. Experimental Apparatus and Procedure

The experimental apparatus used in this work is shown in Fig. 2. The system mainly consisted of a shaft furnace (6) connected with an electric balance (5). The output of the electric balance was connected with the computer (12) for continuous measuring of the weight of the pellets as a function of time during the reduction process. The flow rate of different gases was adjusted precisely through mass flow controller (3). The actual temperatures of the furnace and pellets were measured through temperature controller (11). In each experiment, when the furnace was first heated to 600°C, the crucible containing twenty HVTMP was placed in heating zone. After maintaining the system at this temperature for a few minutes under a pure nitrogen atmosphere, a reducing gas simulating the gas compositions of DPCG, WCSG, Midrex, and HYL-III, as given in Table 2, was introduced into the furnace at a flow rate of 4.0 L/min, and at the same time the furnace was heated at different heating rates to start the reaction. During the experiment, the weight decrease was monitored continuously by electronic balance and recorded in real-time by computer to enable a quantitative-kinetic analysis. At the end of the experiment, the pellets were immediately removed from the furnace and cooled down to the ambient temperature under a high flow rate of argon atmosphere.

Fig. 2.

Schematic diagram of experimental apparatus.

In this study, the reduction degree (R) is the mass percentage of oxygen that is reduced from Fe₂O₃ and is calculated by the following formula:   

$$R=\left[\frac{0.11w_1}{0.43w_2}+\frac{m_1-m_t}{m_0{w}_2\times 0.43}\times 100\right]\times 100\text{\%}$$(1)

where w₁ is the ferrous content in the oxidation pellet, %; m₁ is the initial mass of pellet after removal of moisture, g; m_t is the mass of pellet after each reduction time t, g; w₂ is the content of total iron before reduction, %; m₀ is the initial mass of pellet, g; 0.11 is the necessary conversion factor of corresponding amount of oxygen making FeO oxidized to Fe₂O₃; 0.43 is the conversion factor of corresponding amount of oxygen making all TFe oxidized to Fe₂O₃.

The volume swelling index is defined as follows:   

$$\text{Swelling Index}=(V_{\text{a}}-V_{\text{0}})/V_{\text{0}}\times 100\text{\%}$$(2)

where V_a is the volume of the pellets after reduction, %, and V₀ is the volume of the pellets before reduction, %.

2.3. Characterization

After the reduction experiments, the compressive strength and reduction swelling of reduced pellets were detected by ISO4700 and ISO 4689:2007, respectively. Some reduced pellets were crushed with particle sizes smaller than 0.074 mm and used for phase identification by XRD. Grinding wheel was used to cut sections of the reduced pellets, and the sections were subsequently polished. The scanning electron microscope (SEM) coupled with energy-dispersive spectroscopy (EDS) was used in order to examine the polished pellets.

3. Results and Discussion

3.1. Effect of Heating Rate

The HVTMP was reduced with the gas composition of HYL-III at three different heating rates and the results are shown in Fig. 3. The heating rates in this study are 4°C/min, 7°C/min and 10°C/min as shown in Fig. 3. The investigation of the curves indicated that as the reduction temperature increased, the reduction degree increased continuously. It was observed that the reduction degree was apparently impacted by the heating rate and it decreased with the increase of heating rate. The reduction degrees were 51.8 and 32.8% at 800°C when the heating rates were 4 and 10°C/min, respectively. In other words, at the same target reduction degree, the temperature shifted to higher value with the increase of heating rate. This could be attributed to the heat transfer lag. A higher heating rate resulted in the increase of temperature gradient inside and outside of particles, leading to temperature lag for HVTMP to reach the same temperature. The final reduction degree increased with the decrease in the heating rate, and reached 83.2, 76.1, and 72.5% at the heating rates of 4, 7, and 10°C/min, respectively. Besides, it should be noted that the reduction degree was not attained 100%. This was because of the complex chemical composition and special mineral phase structure of HVTM compared with ordinary iron ores, which was consistent with isothermal conditions in our previous study.²⁷⁾

Fig. 3.

Reduction degree versus temperature at different heating rates (gas composition of HYL-III).

Figure 4 shows the compressive strength and reduction swelling index of pellets after reduction at different heating rates with the gas composition of HYL-III. It was clearly seen that the compressive strength of reduced pellets decreased with an increase in the heating rate while the reduction swelling index of reduced pellets had the opposite tendency. This was because that when the heating rate was slow, the HVTMP was exposed in the reducing gases for longer time and its reduction was accelerated greatly. Due to this reason, the metallic iron was generated earlier, leading to the lower reduction swelling. Besides, the metallic iron grew rapidly and had high adhesive strength, resulting in the greater compressive strength of reduced pellets.

Fig. 4.

Effect of heating rate on the compressive strength and reduction swelling index of reduced pellets (gas composition of HYL-III).

3.2. Effect of Gas Composition

Figure 5 presents the reduction degree as a function of temperature ranging from 600 to 1100°C at 7°C/min with different gas compositions. As shown in Fig. 5, it was found that the gas composition had a pronounced effect on the reduction of HVTMP. The reduction with the gas composition of HYL-III showed the highest reduction degree, and the reduction with DPCG showed the lowest reduction degree. The reason was that among all gas compositions, the HYL-III which had the highest content of hydrogen owned the highest penetration capacity, as well as the highest diffusing and reducing capacities, while DPCG had the lowest content of hydrogen. Therefore, the reduction degree was increased in the order of DPCG<WCSG<Midrex<HYL-III. However, the final reduction degrees were only 61.2, 69.0, 70.8, and 76.1% with DPCG, WCSG, Midrex, and HYL-III, respectively, which indicated that the reduction of HVTMP was not 100% completed. Also, we could found that the effect of heating rate was more significant than gas composition during the reduction of HVTMP.

Fig. 5.

Reduction degree versus temperature with different gas compositions (heating rate of 7°C/min).

The compressive strength and reduction swelling index of pellets after reduction with different gas compositions were measured and the results are shown in Fig. 6. It indicated that gas composition had an evident effect on the change of the compressive strength and reduction swelling index of reduced pellets. The reduced pellets showed lowest compressive strength and highest reduction swelling index with the gas composition of DPCG. The compressive strength of reduced pellets increased in the order of DPCG<WCSG<Midrex<HYL-III and reduction swelling index showed the opposite tendency. The phenomenon could explain by the fact that the pellet swelling was dependent on the rate of metallic iron production and its diffusion. The gas composition of HYL-III with the highest content of hydrogen among all gas compositions owned the highest penetration capacity, as well as the highest diffusing and reducing capacities, which favored the diffusion rate of iron and ultimately the highest reduction swelling was observed.

Fig. 6.

Effect of gas composition on the compressive strength and reduction swelling index of reduced pellets (heating rate of 7°C/min).

3.3. Relation between Reduction Swelling Index and Compressive Strength

According to the results in Figs 4 and 6, under non-isothermal conditions, it was found that the compressive strength of reduced HVTMP differed, so did the reduction swelling index. The relationship between reduction swelling index and compressive strength of reduced pellets is shown in Fig. 7. It could be seen that the compressive strength of reduced pellets decreased as the reduction swelling index increased. They exhibited a reversed linear relationship. The experimental linear regression equation between reduction swelling index and compressive strength of reduced pellets was:   

$$C=-175.39S+1\mathrm{\hspace{0.17em} }852.50$$(3)

where C is the compressive strength and S is the reduction swelling index of reduced pellet.

Fig. 7.

Relationship between reduction swelling index and compressive strength of reduced pellets.

On the basis of the previously analysis, we could found that the reduction swelling index had relationship with the compressive strength of reduced pellets and improved the permeability in a shaft furnace. In order to obtain a higher compressive strength of reduced HVTMP for improving the permeability of shaft furnace, the reduction swelling should be controlled to a certain range. Therefore, in practical industrial HVTMP smelting in the gas-based shaft furnace, lowering the heating rate and increasing the hydrogen content of the reducing gases were suggested, which could prevent the reduced pellets from swelling and result in a higher compressive strength.

3.4. Phases Transformation during Reduction

Under non-isothermal conditions, the heating rate affects the reduction degree as well as the phase composition of reduced HVTMP. Figure 8 presents the XRD patterns of the reduced pellets at 1100°C with different heating rates when the gas composition was HYL-III. Before reduction, the XRD pattern showed that the raw pellets contained hematite, hematite solid solution, iron-chromium solid-solutions, and chromium-vanadium solid-solution. In the XRD pattern of 10°C/min, the peak intensities of metallic iron, ulvospinel, chromite, and coulsonite were first observed. The intensities of metallic iron increased from 12990 to 13684 when the heating rate decreased to 7°C/min. However, the intensities of ulvospinel became weaker (from 2038 to 1785), which was attributed to the higher exposed time in the reducing gases at 7°C/min than that at 10°C/min. As the heating rate decreased from 7 to 4°C/min, ilmenite appeared, which corresponded to a decrease in that of the ulvospinel peaks on account of the reduction of ulvospinel to ilmenite.

Fig. 8.

XRD analysis of reduced HVTMP at different heating rates (gas composition of HYL-III).

HVTMP in the progress of reduction at different temperatures (also different reduction degrees) were analyzed by XRD when the heating rate was 7°C/min with the gas composition of HYL-III. The results are shown in Fig. 9. At 650°C, the peaks of hematite solid solution, iron-chromium solid-solutions, and chromium-vanadium solid-solution were totally invisible. The intensity of hematite peaks decreased sharply (from 18032 to 4709), meanwhile, magnetite appeared which implied that the content of hematite fell and that of magnetite rose. It meant that hematite was reduced to magnetite, whereas wustite was not detected. In addition, titanomagnetite, coulsonite, and chromite phases came into being at the same time. As the temperature increased from 650 to 700°C, the peaks of hematite disappeared and magnetite became stronger. The XRD pattern at 750°C indicated that wustite and metallic iron were observed. The peaks of titanomagnetite, coulsonite, and chromite became weaker than those at 700°C and the main phases were titanomagnetite and wustite.

Fig. 9.

XRD analysis of reduced HVTMP under non-isothermal conditions at different temperatures (heating rate of 7°C/min and gas composition of HYL-III).

In the reduction stage from 750 to 800°C, no new phases appeared. The metallic iron became dominant in the XRD patterns. The main reason was that the reduction of titanomagnetite to wustite was more difficult than that of wustite to metallic iron. At 850°C, the reduction of magnetite to wustite was completed and the intensity for metallic iron peaks further increased from 5115 to 7463. As the temperature was increased to 900°C, titanomagnetite disappeared and ulvospinel was first observed. However, the peaks of wustite also disappeared at 950°C. Furthermore, no new phases appeared, the peak intensities of metallic iron kept increasing (from 10224 to 13382). For 1100°C, the reduced HVTMP mainly consisted of four phases, i.e., metallic iron, ulvospinel, chromite, and coulsonite. Comparing with our previous study,²⁵⁾ rutile was not observed in the final reduced HVTMP, which could be attributed to the shorter residence time of HVTMP during high temperature stage under non-isothermal conditions. Based on the XRD analysis, the phase transformations of valuable elements under non-isothermal conditions are shown schematically in Fig. 10.

Fig. 10.

Phase transformations of valuable elements under non-isothermal conditions (heating rate of 7°C/min and gas composition of HYL-III).

3.5. Morphology Analysis during Reduction

In order to investigate what changes occurred in the structure of reduced HVTMP, the morphological observations at a heating rate of 7°C/min with the gas composition of HYL-III were carried out using SEM and the results are shown in Fig. 11. The SEM images showed that the morphological changes at different temperatures were significantly different. In Fig. 11(a), it could be seen that the raw HVTMP crystallized well and hematite and solid solution were the predominant phases. At 700°C, it was obvious that some cracks and pores uniformly distributed widely due to the oxygen removal in iron oxides, as shown in Fig. 11(b). Besides, dense structures were damaged and became relatively loosen, but no metallic iron was developed at this temperature because of the low chemical reaction rate. When the reduction temperature was increased to 800°C, as magnetite reduced to wustite and therefore more pores were formed shown in Fig. 11(c), a small amount of metallic iron particles (Point 1) randomly distributed in the structure. Further increasing the temperature to 900°C, the structure became more porous as presented in Fig. 11(d). In addition, the distribution of small metallic iron particles was widespread and showed a sporadic distribution. At 1000°C, more and more metallic iron particles emerged gradually. Fine metallic iron particles tended to be massive and the growth of metallic iron particles enhanced the density of reduced pellets (Fig. 11(e)). The metallic iron particles were coarser and more aggregated. However, the boundary between metallic iron particles and slags was not clear.

Fig. 11.

SEM images and EDS analysis of the reduced HVTMP at different temperatures (heating rate of 7°C/min and gas composition of HYL-III). (a) raw material; (b) 700°C; (c) 800°C; (d) 900°C; (e) 1000°C; (f) 1100°C; (g) EDS analysis of point 1 and 2; (h) EDS analysis of point 3 and 4.

In the following progress of reduction, the slag phase (Point 3) was divided and distributed among substantial amounts of metallic iron phase (Point 4), metallic iron formed big particles and showed a vermicular distribution. However, comparing with our previous study,²⁵⁾ the reduced metallic iron or slag was not connected together to form a uniform continuous area even at 1100°C, as seen in Fig. 11(f). Figure 12 presents the valuable elemental distribution mapping micrograph at 1050°C. It was found that the element iron was relatively more concentrated, and the elements vanadium and chromium were more disperse and closely associated with iron.

Fig. 12.

Element distributions of reduced HVTMP at 1050°C (heating rate of 7°C/min and gas composition of HYL-III).

4. Conclusions

The non-isothermal reduction behavior and mechanism of HVTMP with simulated shaft furnace gases of DPCG, WCSG, Midrex, and HYL-III were investigated in this study. The following conclusions were obtained:

(1) The heating rate and gas composition had a significant impact on the reduction of HVTMP. With increasing heating rate, the reduction degree decreased. The reduction degree was the highest with the gas composition of HYL-III and was found to decrease in the order of HYL-III>Midrex>WCSG>DPCG.

(2) The reduction swelling index had relationship with the compressive strength of reduced HVTMP under non-isothermal conditions. The reduction swelling index and compressive strength of reduced HVTMP exhibited a reversed linear relationship.

(3) The phase transformation paths of valuable elements under non-isothermal conditions could be expressed as follows: Fe₂O₃→Fe₃O₄→FeO→Fe; Fe₉TiO₁₅→Fe_2.75Ti_0.25O₄→Fe₂TiO₄; (Fe_0.6Cr_0.4)₂O₃, Fe_0.7Cr_1.3O₃→FeCr₂O₄; (Cr_0.15V_0.85)₂O₃→Fe₂VO₄.

(4) Under non-isothermal conditions, high reduction temperatures made metallic iron particles diffused and aggregated more freely in the progress of reduction. However, the reduced metallic iron or slag phase could not be connected together to form a uniform continuous area.

Acknowledgement

This work is financially supported by National Natural Science Foundation of China (Grant No. 51574067).

References

• 1)   X. W.  Lv,  Z. G.  Lun,  J. Q.  Yin and  C. G.  Bai: ISIJ Int., 53 (2013), 1115.
• 2)   W. J.  Huang,  Y. H.  Zhao,  S.  Yu,  L. X.  Zhang,  Z. C.  Ye,  N.  Wang and  M.  Chen: ISIJ Int., 56 (2016), 594.
• 3)   H. X.  Fang,  H. Y.  Li,  T.  Zhang,  B. S.  Liu and  B.  Xie: ISIJ Int., 55 (2015), 200.
• 4)   H. G.  Du: Theory of Smelting Vanadium-bearing Titanomagnetite by Blast Furnace, Science Press, Beijing, (1996), 4.
• 5)   J. H.  Luo,  Z. Y.  Wu,  E. H.  Wu,  J. H.  Li,  X. J.  Liao,  R.  Tang and  S. L.  Yang: Iron Steel Vanadium Titan., 36 (2015), 73.
• 6)   R.  Huang,  X. W.  Lv,  C. G.  Bai,  K.  Zhang and  G. B.  Qiu: Steel Res. Int., 84 (2013), 892.
• 7)   Z. M.  Yan,  X. W.  Lv,  W. C.  He and  J.  Xu: ISIJ Int., 57 (2017), 31.
• 8)   J. J.  Shi,  L. F.  Sun,  J. Y.  Qiu,  B.  Zhang and  M. F.  Jiang: J. Alloy. Compd., 699 (2017), 193.
• 9)   G. J.  Cheng,  X. X.  Xue,  Z. X.  Gao,  T.  Jiang,  H.  Yang and  P. N.  Duan: ISIJ Int., 56 (2016), 1938.
• 10)   F. Q.  Zheng,  G. Z.  Qiu,  Y. F.  Guo,  F.  Chen,  S.  Wang and  T.  Jiang: ISIJ Int., 57 (2017), 1767.
• 11)   M. L.  Hu,  L.  Liu,  X. W.  Lv,  C. G.  Bai and  S. F.  Zhang: Metall. Mater. Trans. B, 45 (2014), 76.
• 12)   X. F.  Liao,  J. H.  Peng,  L. B.  Zhang,  T.  Hu and  J.  Li: J. Alloy. Compd., 708 (2017), 1110.
• 13)   T.  Hu,  X. W.  Lv and  C. G.  Bai: Steel Res. Int., 87 (2016), 494.
• 14)   R.  Paunova: Metall. Mater. Trans. B, 33 (2002), 633.
• 15)   M. A. R.  Dewan,  G. Q.  Zhang and  O.  Ostrovski: ISIJ Int., 50 (2010), 647.
• 16)   F.  Pan,  Z.  Du,  M. J.  Zhang and  H. Y.  Sun: ISIJ Int., 57 (2017), 413.
• 17)   S.  Samanta,  M. C.  Goswami,  T. K.  Baidya,  S.  Mukherjee and  R.  Dey: Int. J. Miner. Metall. Mater., 20 (2013), 917.
• 18)   H. F.  Yang,  L. L.  Jing and  B. G.  Zhang: J. Hazard. Mater., 185 (2011), 1405.
• 19)   T.  Hu,  X. W.  Lv,  C. G.  Bai,  Z. G.  Lun and  G. B.  Qiu: ISIJ Int., 53 (2013), 557.
• 20)   L. S.  Zhao,  L. N.  Wang,  D. S.  Chen,  H. X.  Zhao,  Y. H.  Liu and  T.  Qi: Trans. Nonferr. Met. Soc. China, 25 (2015), 1325.
• 21)   C.  Geng,  T. C.  Sun,  H. F.  Yang,  Y. W.  Ma,  E. X.  Gao and  C. Y.  Xu: ISIJ Int., 55 (2015), 2543.
• 22)   Z. C.  Huang,  K.  Wu,  B.  Hu,  P.  Hu and  T.  Jiang: J. Iron Steel Res. Int., 19 (2012), 1.
• 23)   T.  Hu,  X. W.  Lv,  C. G.  Bai and  G. B.  Qiu: Int. J. Miner. Metall. Mater., 21 (2014), 131.
• 24)   H. Y.  Sun and  A. A.  Adetoro,  Z.  Wang,  F.  Pan and  L.  Li: ISIJ Int., 56 (2016), 936.
• 25)   W.  Li,  G. Q.  Fu,  M. S.  Chu and  M. Y.  Zhu: Steel Res. Int., 88 (2017), 1600120.
• 26)   M. S.  Chu and  Q. J.  Zhao: China Metall., 18 (2008), 1.
• 27)   W.  Li,  G. Q.  Fu,  M. S.  Chu and  M. Y.  Zhu: Steel Res. Int., 88 (2017), 1600228.
• 28)   W.  Li,  G. Q.  Fu,  M. S.  Chu and  M. Y.  Zhu: Ironmaking Steelmaking, 44 (2017), 294.