Dissection Investigation of Ti(C,N) Behavior in Blast Furnace Hearth during Vanadium Titano-magnetite Smelting

Abstract

This paper presents a comprehensive study on the behaviors of titanium compounds generated in the blast furnace (BF) hearth during the vanadium titano-magnetite smelting by the dissection method combined with experimental and theoretical analysis. The results show that considerable titanium compounds are formed in the eroded furnace wall below the taphole level. The phases of titanium compounds consist of Ti(C,N) crystals, slag phase and liquid iron. The phase of Ti(C,N) crystals are mainly TiC_0.3N_0.7. The formation mechanism of titanium compounds is revealed. It is found that the Ti(C,N) phase is formed within hot metal, while the slag phase originates from the interaction between the mineral in coke and final slag in the BF hearth. Furthermore, the slag is confirmed present in the low part of the BF hearth. The thermal conductivity of the titanium compounds in the hearth is determined as 11.98 W/m/K by analyzing the thermodynamic and heat transfer characteristics of the compounds. This prediction is in satisfactory agreement with the measurement.

1. Introduction

The campaign life of ironmaking blast furnaces (BF) is largely determined by the hearth erosion.^1,2) It is established that titanium-bearing ore charged into a BF for smelting ordinary iron ore forms titanium compounds in the hearth, which can provide effective protection against chemical and mechanical erosions.^3,4,5,6) However, a large quantity of titanium compounds may give rise to various operation problems in the BF practice. For example, it increases the viscosity of slag and thus the difficulty to flow, as well as deteriorates the separation process between hot metal and slag. This is particularly true for titano-magnetite smelting BF.^7,8) Therefore, a better understanding of the Ti(C,N) behaviors at various locations of the BF hearth is of great importance to achieving smooth smelting.

In recent years, various efforts have made to study the formation of Ti(C,N) at temperatures somewhat comparable to those encountered in real BFs. However, most of such studies were based on laboratory-scale tube furnace as well as numerical simulations. For example, Bergsma and Fruehan et al.⁹⁾ proposed a hypothesis on the formation mechanism of the Ti(C,N) protection layer, suggesting that the Ti(C,N) forms in hot metal. Li et al.^10,11) confirmed the formation of Ti(C,N) in the molten iron liquid under different temperatures through confocal scanning laser microscope. Through various dedicated efforts, the titanium partition between the BF slag and hot metal was reasonably established.^12,13,14,15) Recently, the transport phenomena associated with the formation and dissolution of Ti(C,N) were also revealed by CFD (Computational Fluid Dynamics) methods.^16,17,18) Generally speaking, the hearth condition is extremely complicated, featured with the intensive interactions among molten slag, metal, gas, coke in terms of flows and thermochemical behaviors companying with high temperature and high pressure. As such, various assumptions have been introduced to the previous studies on the formation of titanium carbide. The efforts dedicated to study the formation mechanism of Ti(C,N) inside real BFs are lacking.

At present, a viable method to access the internal condition of real furnaces is BF dissection after blowing-out. It provides a promising way to understand the origins of Ti(C,N) formation and the subsequent behaviors in BFs, leading to a better control of the effect of Ti(C,N) on process performance. However, the previous studies based on BF hearth dissection largely focused on the erosion profile of hearth refractory and in turn, the wear mechanisms of the hearth and bottom lining but for ordinary iron ore smelting.^19,20,21) Recently, based on dissection method, different researchers^22,23) investigated the protective layer in terms of slag and graphite formed below the BF taphole level in the smelting of ordinary iron ore, where Ti(C,N) is negligible. Liu et al.²⁴⁾ observed the significant sediment of Ti(C,N) below the tuyere and above the tap hole in the hearth and suggested that TiO₂ in the slag was reduced by coke to form Ti(C,N). To date, the Ti(C,N) behaviors in the hearth below the tap hole have not been well addressed in relation to appearance, crystallization and formation mechanism, especially for BFs operated with titano-magnetite ores.

In this study, Ti(C,N) samples were obtained from the dissection of a titano-magnetite smelting BF during the overhaul involving the renewal of the hearth. These samples were used to derive some meaningful information associated with the transformation of Ti(C,N) micrographs and their interactions with other phases. On this base, the formation mechanism of titanium compounds was elucidated. Also, by means of thermodynamic and heat transfer analysis, the thermal conductivity of the titanium compounds in the BF hearth was estimated and compared against the measurement by a thermocouple.

2. Experiment

The details of typical overhaul process have been described in the earlier publications.^20,21) In the present study, the BF, from which the samples were extracted, was characterized by a working volume of 550 m³, 1 tap hole and 14 tuyeres. The BF was put into operation in December, 2006 and shut down in August 2014 for maintenance. The BF has been operated smoothly at a productivity of about 3.8 tHM/(m³·d). The feed iron grade is about 55.8%, and the iron-bearing materials are composed of sinter and pellet at a mass ratio of 60:40. The average coke and coal consumption rates were maintained as 406 kg/tHM and 130 kg/tHM, respectively. The average blast temperature was 1423 K (1150°C). Table 1 lists the average compositions of hot metal and slag under normal operation conditions. The data are similar to other blast furnaces operated for titano-magnetite smelting.²⁵⁾

Table 1. Average compositions of hot metal and slag in normal production.

	Hot metal				Slag
Compositions	Si	S	Ti	V	SiO2	CaO	MgO	Al2O3	TiO2	V2O5	FeO	S	R
Content, wt%	0.17	0.05	0.14	0.34	27.06	32.27	10.36	14.44	9.38	0.21	1.65	0.50	1.19

After the blowing-out and cooling-down processes of the furnace, the hearth was dissected. From the eroded hearth wall, a large quantity of titanium compounds were found. The appearance of these titanium compounds are shown in Figs. 1(a) and 1(b). The remaining brickwork in the hearth generally has a thickness of from 250 to 800 mm. Besides, in the front of the brickwork, a significant amount of titanium compounds is observed with a thickness ranging from 100 to 500 mm. The sampling locations and the erosion profiles are shown in Fig. 1(c). In this study, the circumferential positions where the samples were collected are close to the tap hole of the furnace and between tuyeres 1 and 14. The samples were taken from the hot surface of the carbon brick at different heights in the hearth. The sampled areas are marked in the Fig. 1(c) as P1, P2 and P3, where P1 represents the location 1.0 m below the tap hole centerline level, P2 is the thinnest area of carbon brick which is about 2.5 m below the tap hole centerline, P3 corresponds to the bottom area where is on the top of first layer of carbon brick. Note that the thermocouple installed in the P2 area is numbered as 1239.

Fig. 1.

Erosion profile and samples location of the BF hearth bottom. (Online version in color.)

The samples were also selected for XRD (X-ray diffraction) examination. The analysis was conducted using a Rigaku diffractometer (DMAX-RB 12 kW; Rigaku Corporation, Tokyo, Japan) as well as Cu (Ka) radiation. During this analysis, the scanning angles were in the range of 10 to 90 deg (2θ) at a scan rate of 10 deg/minute. The chemical compositions were analyzed via XRF (X-ray fluorescence) (Shimadzu XRF-1800, Japan). Additionally, in order to allow SEM observation, pieces were cut from the selected samples under dry conditions and then placed in a rounded plastic container with 25 mm diameter which was filled with resin. Then, the sample was ground and polished in a way similar to the previous study.²³⁾ Finally, the samples were coated with Au and then examined with a Quanta 250 environmental scanning electron microscope (SEM) equipped with energy-dispersive X-ray spectrometer (EDS) for chemical analysis and element mapping.

3. Results and Discussion

3.1. Appearance and Morphology of Phases in Samples

A large amount of Ti(C,N) (gray phases) are found in the sample P1, as shown in Fig. 2. Figure 2(a) presents the global appearance of the sample, and reveals that several phases co-exist in the sample which has two distinct layers. Further magnification of the right side (cold side) confirms the formation of Ti(C,N) crystals (Fig. 2(b)) in a solid state, and a sharp contact in a “submersed” fashion between Ti(C,N) crystals. Also, the slag matrix is observed. The Ti(C,N) crystals with the size of about 50 μm present mainly as an oval shape. The Fe phase is combined with Ti(C,N) crystals. The existence of Fe is confirmed in the sample (Fig. 3). In other words, the FeO has been reduced completely in the selected area. This result indicates that the slag is not from the BF final slag. In the left side of sample P1, the major compositions are slag phase and Ti(C,N) crystals (Fig. 2(c)). Further magnification of the left part (Fig. 2(d)) shows that the shape of Ti(C,N) crystals are mainly square. The size of such large particles is about 50 μm. The presence of different shapes of Ti(C,N) crystals indicates the different conditions associated with the formation and the subsequent behaviors in the hearth. Here, the Fe phase can be clearly observed, which is same as that shown in Fig. 2(b). In the slag phase, some MgAlO₄ crystals are also found, as shown in Fig. 2(e).

Fig. 2.

Appearance, size, and morphology of phases observed in sample P1. (Online version in color.)

Fig. 3.

X-ray diffraction detection of extracted samples. (Online version in color.)

The XRD pattern of the sample P1 is shown in Fig. 3. In this figure, the distinct characteristic peaks in the XRD pattern are indexed as TiC_0.3N_0.7. The characteristic peaks of Fe and corundum are also detected in the XRD pattern, although they are far less intensive compared with those of TiC_0.3N_0.7. Here, the slag phase is not detected due to its relatively low content.

For the most part of the hearth, a majority of brown titanium compounds are found, e.g. P2. As seen from Fig. 4, Fe, slag and Ti(C,N) are observed in the sample. There exist many fine Ti(C,N) crystals. They are oval-shape with a size of about 50 μm and uniformly distributed (Fig. 4(b)). Moreover, Ti(C,N) particles are found to be surrounded by slag and evenly distributed in the slag matrix. This result is evidence that slag exists in the lower part of BF. The slag compositions are relatively homogenously, and CaTiO₃ is not observed in the slag phase. Additionally, the slag observed here is different from the final slag in the BF. The XRD pattern of the sample P2 (Fig. 3) shows that Ti(C,N) is also present in form of TiC_0.3N_0.7. In the XRD result, Fe is observed but not for slag phase. The phases obtained P2 are very similar to those at the right side of sample P1.

Fig. 4.

Appearance, size, and morphology of phases observed in sample P2. (Online version in color.)

The sample micrographs of the P3 are given in Fig. 5(a). Ti(C,N) crystals have grown to 500 μm in size (Fig. 5(b)) and show a square shape. These crystals should have been exposed to a high temperature over a long period. Here, the Fe phase is mainly combined with Ti(C,N) crystals in a similar way to that of P1 and P2. The XRD analysis (Fig. 3) of sample P3 shows that TiC_0.3N_0.7 is the most abundant phase. The matrix in the sample P3 is carbon instead of slag, which is different from samples P1 and P2. Furthermore, the sample P3 has some slag phase. The composition of this slag is similar to that of coke ash, where Al₂O₃ is the main ash constituents.^26,27) Notably, Al₂O₃ can be found in the slag of P3. The Al₂O₃ phase [melting point, 2045 C (2318 K)] may be from coke minerals, as indicated by Fig. 5(c). That is, the slag in the sample P3 comes from the coke ash. The presence of alkali compounds in the slag phase (Fig. 5(d)) confirms the existence of recirculating alkali in coke in the BF, which has also been reported by Gupta et al.²⁸⁾

Fig. 5.

Appearance, size, and morphology of phases observed in sample P2. (Online version in color.)

3.2. Formation Mechanism of Titanium Compounds

The chemical analysis results of the samples are listed in Table 2. Regardless of which sample is tested, abundant Ti is found. The slag compositions in the P1 and P2 are similar but significantly different from that of P3. Compared with the final slag (see Table 1), the constituents are also different, because the proportions of calcium and magnesium in the final slag are much higher than those of the samples considered here.

Table 2. Chemical compositions of the three samples.

Compositions, %	Ti	Fe	CaO	SiO2	Al2O3	MgO	K2O+Na2O
P1	37.1	9.905	10.89	13.75	8.52	4.31	0.84
P2	20.1	11.305	10.34	12.63	6.13	5.30	0.94
P3	33.8	15.365	1.54	5.58	19.59	0.23	3.50

Figure 6 shows the EDS mapping of P1. Elements of Ca, Mg, Al, Si, O are in the same phase, and Ti is in the form of Ti(C,N). This result indicates that TiO₂ is not in the slag phase, which can also be confirmed by the XRD results (Fig. 3). Furthermore, the boundary between the slag and the Ti(C,N) crystals can be demarcated, as shown in Fig. 7. Ti(C,N) does not form from the final slag. In fact, the micrographs of the slag in P1 and P2 are also different from the final slag, as shown in Fig. 8. Ti in the final slag is in the state of CaTiO₃ and thus the final slag has considerable CaTiO₃. The presence of slag phase in the titanium compounds indicates that the slag exists in the low part in the hearth. Note that the slag phase in the titanium compounds is not from the final slag. This is because in the sample P3, the SiO₂ and Al₂O₃ contents are very high, which is very different from the final BF slag.

Fig. 6.

EDS maps showing the distribution of Ca, Si, Al, Mg, O, Ti, N, C, Fe and Na, K in P1. (Online version in color.)

Fig. 7.

Appearance of the interface of Ti(C,N) and slag. (Online version in color.)

Fig. 8.

SEM micrographs showing the final BF slag. (Online version in color.)

Ten EDS points are selected to calculate the average chemical compositions of slag in P1, P2 and P3. The results are compared with the final slag compositions of the dissected BF, as listed in Table 3. From the EDS results of the phases, it can be seen that the slag mainly consists of Ca, Mg, Si, Al, and O. The compositions of slag in P1 are quite similar to those of P2, and thus the two samples should be in the same chemical state in the hearth. TiO₂ is not detected in the slag. Conversely, the compositions of sample P3 are quite different. The components of Al₂O₃ and SiO₂ are in a high level.

Table 3. Average chemical compositions of the slag in three samples.

Compositions, %	CaO	SiO2	Al2O3	MgO
P1	38.47	41.56	10.53	9.44
P2	37.98	38.54	15.27	8.21
P3	10.44	60.32	28.04	1.20

Figure 9 shows the SEM micrographs of P3. Some slag phases are present in the sample. The composition of P3 is very different form P1 and P2, as listed in Table 2. The matrix is coke. It can be seen that the pores in the coke are large and their size is about 100 μm. Most coke pores, which are originally closed, may be reopened after the consumption of coke walls and then get connected with each other. This behavior produces the channels for liquid slag to flow from the surface to the interior of the coke in the hearth zone. Figure 9(a) shows the SEM micrographs which indicate that the slag flows between coke pores through the channels. Those interconnected macro pores are filled with slag in the final slag layer of the hearth. Coke soaked by slag enters into the iron layer in the hearth. Thus, the coke contains some slag and hot metal within the pores. The original coke minerals which mainly consist of Al₂O₃ can be identified in some area (Figs. 5(c), 9(a)), indicating that only a part of coke minerals has been melted with slag. The reduction of TiO₂ from liquid slag may occur, as shown in Figs. 9(a), 9(b). On the interface between the slag in the pore and the carbon of coke, a layer of Ti(C,N) is formed. Moreover, the Ti,C,N in the hot metal can also form Ti(C,N) crystals, leading to the formation of a majority of Ti(C,N) inside the coke pores.

Fig. 9.

Appearance and morphology of slag phase observed in sample P3. (Online version in color.)

All the above results suggest that the major titanium compounds are a result of solid solution of Ti(C,N) instead of pure TiC and TiN. Ti(C,N) is not formed from slag but mainly from the liquid iron. In fact, the composition of the solid solution is thermodynamically determined by the temperature, nitrogen partial pressure and the content of carbon in hot metal. Provided that the condition nitrogen and carbon are saturated, Ti(C,N) precipitates as crystallized phases in hot metal. Notably there is some slag phase in the matrix of the samples. The multiple phases may be formed such as slag, liquid iron, and Ti(C,N), as illustrated in Fig. 10. First, the Ti(C,N) crystals are precipitated from hot metal in the cold area near the refractory. In the hearth, a considerable amount of coke is immersed in hot metal. The ash in the coke mixes with final slag and forms liquid slag. The liquid slag then spreads in the hearth at high temperatures. When the slag comes across the Ti(C,N) crystals, the liquid slag flows slowly due the relatively low temperature, and some solid Ti(C,N) particles form. This leads to that the slag has a larger viscosity. The process is more likely to occur due to the change of operation condition, the blowing down operation, or the movement of hot metal towards the lower temperature area. Then, the slag may remain in the gaps of the solid Ti(C,N) and fill the gaps. When the hearth temperature increases, the viscosity of the slag in the titanium compounds decreases. A part of the slag therefore flows away from the solid Ti(C,N), and the liquid iron in the hearth then flows into the solid Ti(C,N). With reducing temperature, Ti(C,N) accumulates and MgAl₂O₄ precipitates in the slag. Consequently, the titanium compounds gradually form as the appearance of the samples after a long period. Therefore, the slag in the titanium compounds is mainly from the interaction between the coke mineral and final slag in the BF hearth.

Fig. 10.

Model of formation of titanium protective layer in hearth (The third figure is for sample P3, The six figure is for sample P1 and P2). (Online version in color.)

3.3. Thermal Conductivity of Titanium Compounds

The equilibrium composition of Ti(C,N) can be estimated using the thermodynamic data. The precipitation of TiC and TiN is mainly determined by the solubility of [%Ti][%N] and [%Ti][%C] in hot metal²⁹⁾. On the assumption that TiC and TiN form an ideal solid solution, the reactions of formation for Ti(C,N) in hot metal and TiC proportion in Ti(C,N) are expressed as:^5,30)   

$$\begin{array}{l}\text{TiC}+{\text{1/2 N}}_{\text{2}}=\text{TiN}\mathrm{\hspace{0.17em} }\text{(dislute}\mathrm{\hspace{0.17em} }\text{in}\mathrm{\hspace{0.17em} }\text{TiC)}+\text{C}\mathrm{\hspace{0.17em} }\text{(gr)}\hspace{1em}\\ \mathrm{\Delta }{\text{G}}_{\text{Ti(CN)}}^{\theta }=-\text{35}\mathrm{\hspace{0.17em} }\text{600}+\text{18}\text{.8T}\end{array}$$(1)

  

$$n_{\text{TiC}}=\frac{1}{e^{\frac{-\mathrm{\Delta }{G}_{\text{Ti(CN)}}^{\theta }}{\text{R}T}\mathrm{\hspace{0.17em} }\cdot \mathrm{\hspace{0.17em} }\sqrt{P_{{\text{N}}_{\text{2}}}/P^{\theta }}}+1}$$(2)

Combining Eqs. (1) and (2) allows us to determine the composition of the solid solution (n_TiC) according to the liquid temperature (T) and nitrogen partial pressure (P_N2). Here, we assume that when the temperature and nitrogen partial pressure change, the equilibrium composition of Ti(C,N) crystals in contact with hot metal also changes. On this base, the local temperature to form the Ti(C,N) can be expressed as a function of TiC composition in Ti(C,N), as shown in Fig. 11. For the BF considered, the corresponding nitrogen partial pressure in the hearth is about 1.87 atm, and the composition of Ti(C,N) is in the form of Ti(C_0.3N_0.7), which is identified from Fig. 3. The crystallization temperature at the conditions can hence be calculated, which is 1696 K (1423°C).

Fig. 11.

Effect of temperature and P_N2 on TiC_0.3N_0.7 composition in hot metal. (Online version in color.)

The 1239 point location (P2) considered to study the thermal conductivity of titanium compound under the actual condition is shown in Fig. 12. The thickness of remained carbon brick and the titanium compounds formed there are about 311 mm and 120 mm, respectively. The insertion depth of the thermocouples is 280 mm and 160 mm, respectively. The schematic illustration is given in Fig. 12(b).

Fig. 12.

Actual photo and schematic diagram of 1239 point (P2). (Online version in color.)

During the late period of BF operation, the temperature of the thermocouples is quite high. The carbon brick is observed to be severely eroded. Figure 13 shows the temperature changes for the late period of point 1239. The temperature of thermocouples increases first due to the further erosion, but decreases as long as the titanium compounds are formed and become thicker. Eventually, the temperature becomes stable, indicating that the thickness of the titanium compounds is stable. Therefore, according to the following equation, the thermal conductivity of the titanium compounds can be calculated:   

$$\frac{T_2-T_1}{\mathrm{\Delta }{L}_1}{\lambda }_C=\frac{T_{\text{Ti}}-T_2}{\frac{L_{\text{Ti}}}{\lambda }+\frac{\mathrm{\Delta }{L}_2}{{\lambda }_{\text{C}}}}$$(3)

where T₂ is the temperature of hot point, K; T₁ the temperature of cold point, K; ΔL₁ is the distance between the hot point and cold point, m; T_Ti is the precipitation temperature of the TiC_0.3N_0.7, K; L_Ti is the thickness of titanium compounds, m; λ_C is the thermal conductivity of carbon brick and equal to 10 W/m/K, and λ_Ti is the thermal conductivity of titanium compounds, W/m/K.

Fig. 13.

Temperature of 1239 point in the late period of BF. (Online version in color.)

On the condition that T_Ti is determined (see above), substituting the known T₁ and T₂ and other conditions into Eq. (3) gives the average thermal conductivity of titanium compounds in the high temperature, which is 11.98 W/m/K. In this study, the laser method was employed to measure the thermal conductivity of titanium compounds with 10 mm in diameter and 5–10 mm in thickness. The results are shown in Table 4. As seen from the table, the experimental results of the conductivity measurement are close to the prediction by Eqs. (1), (2), (3). The established method and consequent results are useful for calculating the thickness of the protective layer (titanium compounds) in the hearth, especially for the actual production where the titanium compounds are formed.

Table 4. Thermal conductivity of titanium compounds.

Temperature, K	298	1073
Thermal conductivity, W/m/K	10.86	12.95

4. Conclusions

(1) Considerable titanium compounds are formed at the erosion part of the furnace wall below the taphole in a vanadium titano-magnetite smelting BF. Titanium compounds are in the form of multiple phases including Ti(C,N) crystals, slag phase and liquid iron. The composition of Ti(C,N) crystals is mainly TiC_0.3N_0.7.

(2) The formation mechanism of the titanium compounds is identified. The Ti(C,N) is formed from hot metal supersaturated with titanium at specific temperature and nitrogen partial pressure. Slag in the titanium compounds originates from the interaction coke mineral with final slag in the BF. Slag is confirmed present in the low part of the BF hearth.

(3) Crystallization temperature of Ti(C,N) was estimated based on the thermodynamic and heat transfer characteristics. The composition of the Ti(C,N) was determined by the temperature and the nitrogen partial pressure in the hearth. Accordingly, the thermal conductivity of titanium compounds in the hearth was calculated as 11.98 W/m/K, which satisfactorily matched the experimental measurement.

Acknowledgements

This work was financially supported by the National Science Foundation for Young Scientists of China (51304014), the Key Program of the National Natural Science Foundation of China (No. U1260202) and the 111 Project (No. B13004).

References

• 1)   L.  Shao and  S.  Henrik: Steel Res. Int., 83 (2012), 878.
• 2)   F. M.  Zhang: J. Iron Steel Res. Int., 20 (2013), 53.
• 3)   J. Z.  Zhang,  L. L.  Shi and  W. Z.  Ao: J. Iron Steel Res., 22 (2010), 16.
• 4)   K.  Datta,  P. K.  Sen and  S. S.  Gupta: Steel Res. Int., 64 (1993), 232.
• 5)   K.  Narita,  M.  Maekawa,  O.  Toshio,  S.  Yoshitomo and  M.  Manabu: Trans. Iron Steel Inst. Jpn., 17 (1997), 459.
• 6)   J. C.  Song: Thchnology of Titanium-bearing Material Protection in Blast Furnace, Metallurgical Industry Press, Beijing, (1994).
• 7)   K.  Datta,  P. K.  Sen,  S. S.  Gupta and  A.  Chatterjee: Steel Res. Int., 64 (1993), 232.
• 8)   D. W.  Stothart,  R. D.  Chaykowski,  R. J.  Donaldson and  D. H.  Pomeroy: Iron Steelmaker, 24 (1997), 43.
• 9)   D.  Bergsma and  R. J.  Fruehan: 60th Ironmaking Conf. Process, ISS-AIME, Warrendale, PA, (2001), 297.
• 10)   Y. Q.  Li and  R. J.  Fruehan: Metall. Mater. Trans. B, 32 (2001), 1203.
• 11)   Y. Q.  Li,  L.  Yun and  R. J.  Fruehan: ISIJ Int., 41 (2001), 1417.
• 12)   Y.  Morizane,  B.  Ozturk and  R. J.  Fruehan: Metall. Mater. Trans. B, 30 (1999), 29.
• 13)   R.  Benesch,  A.  Ledzki,  R.  Kopec and  R.  Stachura: Thermochim. Acta, 152 (1989), 447.
• 14)   K. X.  Jiao,  J. L.  Zhang,  H. B.  Zuo and  Z. J.  Liu: J. Northeast. Univ.(Nat. Sci.), 35 (2014), 1160.
• 15)   C.  Han,  H.  Yang and  X. X.  Xue: Trans, Nonferr. Met. Soc., China, 24 (2014), 3215.
• 16)   B. Y.  Guo,  Z.  Paul,  M.  Daniel and  A. B.  Yu: Metall. Mater. Trans. B, 41 (2010), 876.
• 17)   M. K.  Keisuke,  B. Y.  Guo,  Z.  Zughbi and  A. Y.  Yu: Steel Res. Int., 86 (2014), 592.
• 18)   Y.  Chen: Ph. D. thesis, Purdue University, (2012).
• 19)   Y.  Morizane,  B.  Ozturk and  J. F.  Richard: Metall. Mater. Trans. B, 30 (1999), 29.
• 20)   A.  Shinotake,  H.  Nakamura,  N.  Yadoumaru,  Y.  Morizane and  M.  Meguro: ISIJ Int., 43 (2003), 321.
• 21)   T.  Inada,  A.  Kasai,  K.  Nakano,  S.  Komatsu and  A.  Ogawa: ISIJ Int., 49 (2009), 470.
• 22)   Y.  Zhao,  F.  Dong,  L. W.  Lherbier,  Y.  Chen,  C. Q.  Zhou and  J. G.  Grindey: Steel Res. Int., 85 (2013), 891.
• 23)   K. X.  Jiao,  J. L.  Zhang,  Z. J.  Liu,  F.  Liu and  L. S.  Liang: Int. J. Miner. Met. Mater., 23 (2016), 16.
• 24)   X. J.  Liu,  Q.  Lu,  S. J.  Chen,  S. H.  Zhang and  Y. Q.  Sun: J. Iron Steel Res., Int., 22 (2015), 1009.
• 25)   Q. X.  Wang: Vanadium Titanium Magnetite Smelt in Blast Furnace, Metallurgical Industry Press, Beijing, (1994).
• 26)   K. J.  Li,  J. L.  Zhang,  Z. J.  Liu,  T.  Wang,  X. J.  Ning,  J. B.  Zhong,  R. S.  Xu,  G. W.  Wang,  S.  Ren and  T. J.  Yang: Metall. Mater. Trans. B, 45 (2014), 1581.
• 27)   D. W.  Luo,  J. L.  Zhang,  H. W.  Guo,  H. B.  Zuo and  H.  Zeng: 5th Int. Cong. on the Science and Technology of Ironmaking, CSM, Shanghai, (2009), 1154.
• 28)   S.  Gupta,  D.  French,  R.  Sakurovs,  M.  Grigore,  H.  Sun,  T.  Cham,  T.  Hilding,  M.  Hallin,  B.  Hindblom and  V.  Sahajwalla: Prog. Energy Combust. Sci., 34 (2008), 155.
• 29)   S.  Kang: Powder Metall., 40 (1997), 139.
• 30)   K. X.  Jiao,  J. L.  Zhang,  Z. J.  Liu,  M.  Xu and  F.  Liu: Int. J. Miner. Met. Mater., 22 (2015), 1017.