The crude steel production of China reached 822.7 million tons in 2014, which accounted for 49.4 % of the world total crude steel production. The amount of iron ore imported from abroad were about 932.5 million tons, which occupied nearly 60 % of the total iron ore supply of China in 2014. Meanwhile, China imported about 70.07 million tons bauxite in 2013, which also accounted for nearly 60 % of the total bauxite supply of China in 2013. Thus, the shortage of iron ore and bauxite ore are the important bottlenecks for the development of steel industry and aluminum industry, respectively.
High iron bauxite ore is a kind of iron-rich bauxite ore, which spreads over some countries, such as China, America, Australia, Brazil, Greece, Laos, Tanzania, and so on [1–4]. The national reserve in China attains to more than 1.5 billion tons . In order to utilize high iron bauxite ore comprehensively, several treatment processes have been presented successively [6, 7]. For instance, high-intensity magnetic separation to remove hematite , the process of agglomeration – blast furnace smelting – leaching of alumina , leaching of the iron oxide [10, 11], leaching of the alumina [12–14], gas-solid reaction followed by leaching , reduction and magnetic separation the iron [16–19]. With regard to the reduction and magnetic separation processes, Jiang et al.  researched the effect of different sodium salts on Al-Fe separation in limonite. With adding 14 % (mass fraction) Na2SO4 and 2.5 % BS, the total iron grade of metallic iron product increased to 91.3 %, and the Al2O3 content decreased to 1.27 %, the iron recovery rate reached 93.64 %. Hu et al.  investigated the effect of nonmetallic additives on the grindability of iron grains. The results showed that, with adding Na2CO3, the grade of iron powder was 95.6 % and the recovery was 90.2 %. However, the alkali salt was harmful to ironmaking and the utilization of iron, which limited the application of the processes. Pickles et al. [18, 19] carried out some experiments to research the removal of iron as magnetite from high iron bauxite ore. It was shown that, under mild reducing conditions, almost half of the iron could be removed as magnetite. However, the formation of hercynite limited the iron separation as magnetite and higher iron removals could only be achieved through the formation of metallic iron under more highly reducing conditions.
Thus, in the present research, the removal of iron as metallic iron from fine particles of high iron bauxite ore was investigated by metallized reduction and magnetic separation without adding any alkali salt. First, FactSage6.4 was utilized to determine the function of reduction temperature and coal amount. Second, reduction experiments were carried out to research the effect of process parameters, including magnetic field intensity, reduction temperature, reduction time, FC/O and particle size of high iron bauxite ore, on separation indexes systematically. Third, the phase transition of iron and aluminous minerals during the process were investigated by X-ray analysis and thermodynamic consideration. Finally, the factors limiting the separation of iron were discussed basing on the thermodynamic analysis and experimental results.
The high iron bauxite ore used in this investigation is obtained from Guangxi Region in China. The chemical composition of high iron bauxite ore is listed in Table 1. The content of alumina and ferric oxide are 23.85 % and 49.21 %, respectively. This ore is a run-of-mine ore with the characteristics of high iron, high silica and low A/S. That is to say this high iron bauxite ore is not suitable to produce alumina by Bayer process .
The mineral composition of high iron bauxite ore is investigated by X-ray diffraction. Figure 1 gives the Xrd pattern of high iron bauxite ore analyzed by the search match software. As is shown in Figure 1, the main crystalline phases of this ore are goethite (FeOOH and Fe2O3·H2O), hematite (Fe2O3), gibbsite (Al(OH)3), diaspore (AlOOH), halloysite (n(Al2O3) SiO2·n(H2O)) and silica (SiO2).
The SEM micrographs of high iron bauxite ore together with EDS profiles of selected positions are presented in Figure 2. They indicate that iron minerals and aluminum minerals are superfine and conjoint with each other. Thus, it is extremely difficult to separate iron minerals and aluminum minerals from high iron bauxite ore.
The pulverized coal is used as reductant and carburizer, and its chemical composition is listed in Table 2. Before using as reductant in test, the coal is ground to the size of –0.074 mm.
The key for the reduction of ferrous oxides in present work is not only to reduce iron oxides to metallic iron but also the reduced metallic iron particles grow to big size, which is benefit for magnetic separation .
Before experiments, the high iron bauxite ore is dried at 105 °C for 5 h, and then crushed to passing the particle size set in advance using a laboratory jaw crusher and rolls crusher. Figure 3 shows a flowchart of the process. As is shown in Figure 3, the high iron bauxite ore and pulverized coal are matched with FC/O of 1.0, 1.5, 2.0, 2.5 and 3.0, respectively. Then, all the samples are mixed uniformly. Metallized reduction is performed in a closed MoSi2 muffle furnace, whose temperature control accuracy is within ±5 °C. In each test, the mixing samples are put into a closed graphite crucible and then reduced in the furnace under the preset temperature. Once the reduction experiment is finished under the preset time, the samples are taken out and cooled rapidly under isolating from air. Next, the cooled samples are ground to the size of less than 0.074 mm. Last, the ground fines are separated into magnetic product and non-magnetic product by Davis Tube with certain magnetic field intensity.
The total iron (TFe) and metallization ratio (MFe) in magnetic product, alumina content (TAl2O3) in non-magnetic product are determined by chemical analysis, and the iron recovery (ηFe) and alumina recovery (ηAl2O3) are deduced according to mass balance.
The high iron bauxite ore could be considered consisting of hematite (Fe2O3), alumina (Al2O3) and silica (SiO2). From the reaction module in FactSage 6.4 package, some iron-bearing species could be appeared in theory between the temperature of 300 K and 1,800 K, whose composition is given in Table 3.
It is now generally accepted that mechanism of carbothermic reduction of iron oxides is two stage mechanism with the participation of gaseous intermediate (CO and CO2) according to the equations FexOy+CO=FexOy–1+CO2 and CO2+C=2CO [22, 23].
In theory, the reduction process of high iron bauxite ore is extremely complex. The reduction reaction equations of iron-bearing species to metallic iron, which are based on the thermodynamic calculations with FactSage 6.4, are shown as follows: (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15)
Figure 4 gives equilibrium atmosphere for reactions occurred in the reduction of high iron bauxite ore, the data of which is calculated using the FactSage thermodynamics software package (Thermfact Ltd.-CRCT, Montreal, Canada ). It indicates that, compared with iron oxides, all iron-bearing species are more difficult to be reduced by CO. Because higher PCO/(PCO+PCO2) is required for the reduction of iron-bearing species. It is found that the reducibility from high to low of the iron-bearing species followed by Al2Fe2O6, FeSiO3, Fe2SiO4, Fe3Al2Si3O12, FeAl2O4 and Fe2Al4Si5O18. As is shown in Figure 4, the reactions (12) and (13) are occurred easily than reactions (14) and (15). That is to say the SiO2 in FeSiO3 and Fe2SiO4 would be taken place by Al2O3, and then more FeAl2O4 are generated. The FeAl2O4 and Fe2Al4Si5O18 are the most difficult minerals to be reduced by CO. They need higher reduction temperature and higher PCO/(PCO+PCO2). The values of which are no less than 1,147 K, 96.00 % and 1,166 K, 96.85 %, respectively. Maybe, at least one of the reaction (7) and reaction (10) is the limiting reaction for the reduction of high iron bauxite ore. Thus, the carbothermic reduction of high iron bauxite ore depends on high temperature and enough coal, which is used to provide high value of PCO/(PCO+PCO2).
Results and discussion
Main parameters influence on the metallized reduction and magnetic separation are magnetic field intensity, reduction temperature, reduction time, FC/O and high iron bauxite ore particle size. In order to acquire appropriate operation parameters for the metallized reduction and magnetic separation iron from fine particles of high iron bauxite ore, five group experiments are designed to investigate the effect of these factors on the separation indexes, including iron recovery (ηFe), iron content in magnetic product (TFe), metallization ratio (MFe), alumina recovery (ηAl2O3), alumina content in non-magnetic product (TAl2O3).
Effect of magnetic field intensity
In order to determine the appropriate magnetic field intensity, the reduction-grind product is separated by a Davis Tube with magnetic field intensity between 26.7 kA/m and 233.3 kA/m, while keeping other process parameters constant as follows: reduction temperature of 1,400 °C, reduction time of 180 min, FC/O ratio of 2.0, high iron bauxite ore particle size of –2.0 mm.
The effect of magnetic field intensity on separation indexes is shown in Figure 5. The data indicate that, with increasing magnetic field intensity from 26.7 to 93.3 kA/m, the ηFe increases from 86.45 % to 94.66 % sharply, while the ηAl2O3 decreases from 89.35 % to 81.90 %. However, with the magnetic field intensity is above 93.3 kA/m, the ηFe and the ηAl2O3 would level off.
When magnetic field intensity is 40 kA/m, both the ηFe and ηAl2O3 in tests exceed 85 %, which are 89.24 % and 86.09 %, respectively. Meanwhile, the TFe and TAl2O3 are 78.23 % and 53.32 %, respectively. Consequently, the mild magnetic field intensity ranges from 26.7 to 40 kA/m, and the optimal magnetic field intensity is 40 kA/m.
The morphologic features of magnetic product from SEM and EDS analysis are shown in Figure 6. It can be seen that the particles appear in a wide range and the primary element of the particles is Fe. The metallic iron particles occupy almost all image of magnetic product. Moreover, in accordance with the EDS analysis, a small amount of slag phase containing Si, O and Al is found on the surface of some metallic iron particles. In addition, a small number of slag particles are entrained by metallic iron particles. Thus, with increasing magnetic field intensity, smaller iron particles, substance contains wustite (FeO) and other minimal non-magnetic product, are entrained by iron particles, and then are dressed into the magnetic product. This may explain why the MFe and TFe decrease with increasing the magnetic field intensity from 26.7 to 93.3 kA/m. So the mild magnetic field intensity is not too high or too low, just 40 kA/m is appropriate.
Effect of reduction temperature
High iron bauxite ore is reduced at various temperatures between 1,350 °C and 1,450 °C, while keeping other process parameters constant as follows: magnetic field intensity of 40 kA/m, reduction time of 180 min, FC/O ratio of 2.0, and high iron bauxite ore particle size of –2.0 mm.
The effect of reduction temperature on separation indexes is shown in Figure 7. As is shown in Figure 7, with increasing reduction temperature from 1,350 to 1,400 °C, the ηFe, TFe, MFe and TAl2O3 increase sharply, and then level off with the reduction temperature over 1,400 °C. However, the ηAl2O3 ascends from 83.56 % to 88.06 % slowly as increasing reduction temperature from 1,350 to 1,450 °C. When reduction temperature is no less than 1,400 °C, the ηFe and ηAl2O3 in tests are 89.24 % and 86.09 % respectively, both the recovery of iron and alumina in tests are above 85.0 %.
SEM photos of the reduction-grinding products are shown in Figure 8. With increasing reduction temperature, the thermodynamic and kinetic conditions for the reduction of iron-bearing phase should be improved remarkably. The nucleation, aggregation and growth of metallic iron particles are further facilitated, and much larger iron bead are formed at 1,400 °C. Thus, the effectiveness of separating iron from other phases would be much more obvious at a higher reduction temperature. However, raising reduction temperature needs more energy consumption. Taking energy consumption into account, the recommended reduction temperature in the tests is of 1,400 °C.
Effect of reduction time
To optimize the reduction time, a series of reduction tests are performed at various reduction time, including 60, 90, 120, 150 and 180 min. Other parameters are kept as follows: magnetic field intensity of 40 kA/m, reduction temperature of 1,400 °C, FC/O ratio of 2.0, and high iron bauxite ore particle size of –2.0 mm. The effect of reduction time on the separation indexes is shown in Figure 9.
As is shown in Figure 9, with increasing reduction time from 60 to 120 min, the MFe and ηFe increase obviously. When reduction time exceeds 120 min, the MFe and ηFe increase slightly. So as to the TFe, ηAl2O3, and TAl2O3. When reduction time is 120 min, the MFe and ηFe are nearly 96.54 % and 86.79 %, respectively, both of them are above 85.0 %. As increasing reduction time to 180 min, the ηFe and ηAl2O3 are 89.24 % and 86.09 %, respectively. Because prolonging reduction time would increase the cost of production. Taking the economic cost into account, the mild reduction time of the tests is no less than 120 min, but should not exceed 180 min.
The relevant mechanisms are similar to those responsibility for the effect of the reduction temperature. With prolonging the reduction time, the nucleation, aggregation and growth of reduced iron particles are further developed, which contribute to the formation of larger iron bead and result in improving separation of the iron phases from the other phases.
Effect of FC/O
The amount of carbon addition is an important influence factor on the metallization degree of reduction product. Generally, the higher FC/O is, the faster reducing rate is, and also the metallization degree would be higher. In present work, the effect of FC/O on separation indexes is researched with the magnetic field intensity of 40 kA/m, reduction temperature of 1,400 °C, reduction time of 180 min and high iron bauxite ore particle size of –2.0 mm. The values of FC/O vary from 1.0 to 3.0.
The experiment result curves are presented in Figure 10. The data reveal that FC/O has an obvious effect on the separation indexes. With increasing FC/O from 1.0 to 3.0, the MFe and ηFe increase lightly and then come down slowly. When the FC/O is of 2.0, the TFe, TAl2O3, and ηAl2O3 come to the highest values.
As a matter of fact, increasing the content of pulverized coal, the Boudouard reaction would take place at a faster rate to provide more CO and high PCO/(PCO+PCO2). Thus, the reduction of iron minerals is accelerated and the carburization of metallic iron particles becomes easily , which is helpful to the aggregation and growth of metallic iron particles. Nevertheless, this role would be weakened with the further extension of FC/O. The carbon supply become excessive and much more graphitized carbon is formed. The graphitized carbon prevents the aggregation of metallic iron particles and results in the formation of only relatively smaller iron particles, which ultimately lead to poor separation of the Fe-containing phases from the other phases. Thus, the FC/O of 2.0 is recommended in the tests.
Effect of high iron bauxite ore particle size
Particle size of high iron bauxite ore is another influence factor on the separation indexes. Particle size of high iron bauxite ore, namely –0.074 mm, –0.5 mm, –1.25 mm, –2.0 mm and –3.2 mm, is researched with the magnetic field intensity of 40 kA/m, reduction temperature of 1,400 °C, reduction time of 180 min and carbon ratio 2.0.
Experimental results are shown in Figure 11. The results show that, with increasing particle size of high iron bauxite ore from –0.5 mm to –2.0 mm, the ηFe and MFe in magnetic product increase sharply. While the particle size above –2.0 mm, the ηFe and MFe would take down obviously. However, the TFe and ηAl2O3 increase all the time as the particle size of high iron bauxite ore increasing from –0.5 mm to –3.2 mm. There is a particular particle size of high iron bauxite ore, as the high iron bauxite ore particle size of –0.074 mm, the separation indexes is relatively higher.
The iron minerals are reduced, and then carburized, and last aggregated and grown. When iron minerals in a single particle, there is a relatively short distance between one metallic iron particle and another metallic particle, which contribute to the aggregation and growth of metallic iron particle. However, when the particle size of high iron bauxite ore become excessive big, the reduction of iron minerals become difficulty, which is negative for the aggregation and growth of metallic iron. With regard to fine powder (–0.074 mm) of high iron bauxite particle, because the particles are relatively small, the carburization, aggregation and growth for iron minerals become more easily, the separation indexes is relatively higher.
Because the smaller particle size of high iron bauxite ore is, the more grinding energy is needed. Therefore, there is a suitable particle size of high iron bauxite ore, it could be seen that when the particle size is –2.0 mm, the ηFe and ηAl2O3 are 89.24 % and 86.09 %, respectively, all of them exceed 85.0 %. Thus, mild particle size of high iron bauxite ore is –2.0 mm.
From the above discussion, the proper parameters of new process can be drawn as follows: reduction temperature of 1,400 °C, reduction time of 180 min, FC/O of 2.0, ore particle size of –2.0 mm, and magnetic field intensity of 40 kA/m. With the proper parameters of the process, the ηFe and ηAl2O3 are 89.24 % and 86.09 %, respectively, both of them are above 85 %.
Phase transition of reduction process
Some reduction products are investigated by X-ray diffraction to clarify the phase transition of reduction process of high iron bauxite ore. These reduction products are prepared with reduction time varying from 3 min to 180 min, while keep reduction temperature of 1,400 °C, FC/O of 2.0 and ore particle size of –2.0 mm constant.
The X-ray diffraction patterns are shown in Figure 12. It can be seen that, iron appears in the form of metallic iron in reduced products, which are easy to be recovered by low-intensity magnetic separation after being liberated. In comparison with the X-ray diffraction pattern of original high iron bauxite ore, the peaks of Fe2O3·H2O, FeOOH, Fe2O3, Al(OH)3, AlOOH, and n(Al2O3)·SiO2·n(H2O) are disappeared, while the peaks of Fe and FeAl2O4 are appeared with the reduction time of 3 min. At the reduction time of 5 min, the peaks of Al2O3 are appeared because of small amount of FeAl2O4 is reduced to Fe and Al2O3. When the reduction time comes to 20 min, the peaks of Al6Si2O13 are appeared, since some Al2O3 and SiO2 are reacted to generate Al6Si2O13. Increasing the reduction time to 150 min, due to the reduction of FeAlO4 and the reaction of Al2O3 and SiO2, the peaks of FeAlO4 and SiO2 are disappeared, while the peaks of Fe, Al2O3 and Al6Si2O13 are strengthened.
Comparing the X-ray diffraction pattern of reduced products with different reduction time, as increasing the reduction time, the peaks of Fe and Al2O3 are enhanced gradually, while the peaks of FeAl2O4 and SiO2 are disappeared step by step. During whole reduction progress, there is no individual wustite (FeO). Then it could be drawn that the phase transition of iron-bearing minerals are as follows: FeOOH → Fe2O3 → FeO·Al2O3 → Fe. This may explain why the metallization ratio increase as the reduction time increasing from 0 to 120 min and then level off after the reduction time is over 120 min. On the basis of the analysis on phase transition of iron minerals and the thermodynamic analysis previous statement, it could be concluded that the reaction (7) is the limiting reaction for the reduction of high iron bauxite ore.
In the tests, Al(OH)3 and AlOOH are decomposed to Al2O3. Meanwhile, the n(Al2O3)·SiO2·n(H2O) is decomposed to Al2O3 and SiO2. Then, Al2O3 would react with FeO to form FeAl2O4. Next, the FeAl2O4 is reduced to Fe and Al2O3. At last, some of the Al2O3 would react with SiO2 to generate Al6Si2O13. So, the primary aluminum mineral is Al2O3, few amount of Al6Si2O13 exist in the reduction products. Thus, the transition process of aluminum minerals are as follows:
Thus, under the proper conditions of metallized reduction, the iron compounds in the high iron bauxite ore could be reduced to metallic iron, and aluminum minerals could transform to Al2O3 almost.
However, as is shown in Figure 12, it can be seen that the time of hercynite (FeAl2O4) reduced to metallic iron is no less than 120 min. That is to say that the reduction stage takes over most of the time of the metallic iron particles forming process, while the stages of nucleating, aggregating and growing are relatively short. It also indicates that the reduction rate of hercynite (FeAl2O4) reduced to metallic iron is quite slow, even the reduction temperature is of 1,400 °C. Thus, The formation of hercynite (FeAl2O4) limits the reduction rate of iron oxide minerals to metallic iron. In order to reduce the reduction conditions and improve iron recovery, it should enhance the reduction rate of iron oxide minerals to metallic iron. Because the ore-coal composite agglomerates have the characters of high reaction rate due to the closeness of the reactants [26–29], and some catalysts could speed up the reaction rate [30–32], the lower reduction conditions and higher iron removal could be achieved with adopting ore-coal composite agglomerates and adding catalyst.
The integrated process which combines metallized reduction with magnetic separation was investigated in the paper. Particularly, the influences of magnetic field intensity, reduction temperature, reduction time, FC/O and ore particle size were discussed. Conclusions are summarized as follows:
The results show that metallized reduction and magnetic separation could separate iron as metallic iron from fine particles of high iron bauxite ore. Under appropriate reducing conditions, almost 89.24 % of iron could be removed as metallic iron, and 86.09 % of aluminum as alumina would stay in non-magnetic fraction.
In the reduction process, the phase transition of iron-bearing minerals is obtained as follows: FeOOH → Fe2O3 → FeO·Al2O3 → Fe, and also that of aluminum minerals are as follows: Al(OH)3/AlOOH/n(Al2O3)·SiO2·nH2O → Al2O3 → FeAl2O4 → Al2O3/Al6Si2O13.
The reduction stage takes over most of the time of metallic iron particles forming process. Meanwhile, the formation of hercynite (FeAl2O4) limits the reduction rate of iron oxide minerals to metallic iron.
The lower reduction conditions and higher iron removal could be achieved with adopting ore-coal composite agglomerates and adding catalyst, and the part of work would provide better reformations for the process of the rotary hearth furnace (RHF) to utilize the high iron bauxite ore proceedings from China’s actual conditions and current resource situation.
The authors would like to express their gratitude for their financial support of National Natural Science Foundation of China (No. 51374058).
 S.Y. Chen and F. Zhou, Bull. Mineral. Petrol. Geochem. S, 1 (1997) 26–27. Google Scholar
 G. Cheng, G.M. Gao and S.L. Chen, J. Cent. South Univ. (N. Sci. Edit), 39 (2008) 380–386.Google Scholar
 Z.F. Liu, Light Met., 5 (2001) 7–12.Google Scholar
 M.K.D. Mutakyahwa, J.R. Ikingura and A.H. Mruma, J. Afr. Earth Sci., 36 (2003) 357–369. Google Scholar
 P.P. Cui, Z.M. Huang and S.L. Zhou, Light Met., 2 (2008) 6–8. Google Scholar
 Y.T. Li, S.W. Bi, Z.Y. Duan, Y.H. Yang and J.D. Zhang, Light Met., 9 (1992) 6–14. Google Scholar
 Z.G. Liu, Z. Wang, J. Tang, H.T. Wang, and H.M. Long, T. Nonferr. Metal. Soc., 25 (2015) 2415–2421. Google Scholar
 P. Bolsaitis and V. Chang, IEEE Trans. Magn., 17 (1981) 3311–3313. Google Scholar
 J.D. Zhang, Y.T. Li, S.W. Bi and Y.H. Yang, Light Met., 8 (1992) 16–18. Google Scholar
 G. Patermarkakis and Y. Paspaliaris, Hydrometallurgy, 23 (1989) 77–90. Google Scholar
 Y. Paspaliaris and Y. Tsolakis, Hydrometallurgy, 19 (1987) 259–266. Google Scholar
 H.X. Xin, Y. Wu, F. Teng and Y.C. Zhai, J. Northeastern Univ. (Natural Sci.), 35 (2014) 1165–1168. Google Scholar
 H.X. Xin, Y. Wu, F. Teng, S.M. Liu and Y.C. Zhai, Chinese J. Nonferrous Metals, 24 (2014) 808–813.Google Scholar
 G.Z. Lv, T.A. Zhang, A.C. Zhao and H.X. Wang, J. Northeastern Univ. (Nat. Sci.), 34 (2013) 1442–1445. Google Scholar
 Z.L. Zhang, Q. Li and Z.S. Zou, Ironmak. Steelmak, 41 (2014) 561–567. Google Scholar
 T. Jiang, M.D. Liu, G.H. Li, N. Sun, J.H. Zeng and G.Z. Qiu, Chinese J. Nonferrous Metals., 20 (2010) 1226–1233. Google Scholar
 W.T. Hu, H.J. Wang, W.L. Xin and C.Y. Sun, Int. J. Miner. Process, 130 (2014) 108–113. Google Scholar
 C.A. Pickles, T. Lu, B. Chambers and J. Forster, Can Metall Q, 51 (2012) 424–439. Google Scholar
 T. Lu, C.A. Pickles and S. Kelebek, High Temp Mater Processes, 31 (2012) 139–148. Google Scholar
 G.H. Li, M.D. Liu, T. Jiang, T.H. Zhou and X.H. Fan, J. Cent. South Univ. (Sci. and Techno.), 40 (2009) 1165–1171. Google Scholar
 X. Jiang, F.M. Shen, L.G. Liu, X.G. Li and L. Wang, ISIJ Int., 53 (2013) 1358–1364. Google Scholar
 S.K. Wei, Thermodynamics of Metallurgical Processes, Science Press, Beijing (2010). Google Scholar
 O.M. Fortini and R.J. Fruehan, Metall. Mater. Trans. B, 36B (2005) 865–872. Google Scholar
 S. Halder and R.J. Fruehan, Metall. Mater. Trans. B, 39 (2008) 809–817. Google Scholar
 G. Wang, J.S. Wang, Y.G. Ding, S. Ma and Q.G. Xue, ISIJ Int., 52 (2012) 45–51. Google Scholar
 W. Yu, T.C. Sun, Z.Z. Liu, J. Kou and C.Y. Xu, Int. J. Miner. Metall. Mater., 21 (2014) 423–430. Google Scholar
 H.M. Ahmed, N. Viswanathan and B. Bojorkman, Steel Res. Int., 85 (2014) 293–306.Google Scholar
 M.S. Chu, Z.G. Liu, Z.C. Wang and Y. Jun-Ichiro, Steel Res. Int., 82 (2011) 521–528.Google Scholar
 S.Y. Chen and M.S. Chu, Int. J. Miner. Metall. Mater., 21 (2014) 225–233. Google Scholar
 X.M. Guo, H.F. Tang, S.B. Zhang and H.C. No, Acta Metall. Sin., 36 (2000) 638–641. Google Scholar
 T. Hu, X.W. Lv, C.G. Bai, P. Chen and G.B. Qiu, ISIJ Int., 53 (2013) 557–563.Google Scholar
 J.H. Heo, B.S. Kim and J.H. Park, Metall. Mater. Trans. B., 44B (2013) 1352–1363. Google Scholar
About the article
Published Online: 2016-04-09
Published in Print: 2017-01-01
Funding Source: National Natural Science Foundation of China
Award identifier / Grant number: 51374058
National Natural Science Foundation of China (Grant/Award Number: 51374058).