Effect of Coal Type on the Reduction and Magnetic Separation of a High-phosphorus Oolitic Hematite Ore

Abstract

Coal-based direct reduction followed by magnetic separation technique was employed to produce direct reduction iron powder (DRI powder) from a high-phosphorus oolitic hematite ore, and the effects of type and particle size of coal and C/O (Fixed carbon/Oxygen) mole ratio on this process were investigated. The results showed that when using coarse-sized and medium-sized coals, bitumite and lignite presented better iron recovery as compared with anthracite, while this advantage disappeared as the particle size of coals decreased. In addition, the iron content of the DRI powder increased with improving of the coal rank depended on fixed carbon contents, while decreased with the decrease of particle size of coals. Increasing the C/O ratio resulted in a sharp rise of P content of the DRI powder. X-Ray Diffraction (XRD) analysis revealed that more liquid phase was formed in the briquettes during reduction with anthracite as reductant. SEM (scanning electron microscope) observation confirmed that the size of the metallic iron grains formed in reduced briquettes decreased with decreasing of the rank of the coal and the particle size of coal.

1. Introduction

Oolitic iron ore is one of the most important iron ore resources, which widely exists in France, Germany, America, Canada, China etc.^1,2) However, it has not been exploited around the world since its unique oolitic structure and generally contains high levels of phosphorus that resist to conventional mineral processing methods. Along with the fast depletion of easy-to-process iron ores, exploitation of this refractory iron ore becomes more and more important. Coal-based direct reduction followed by magnetic separation process has been employed to produce DRI powder from oolitic iron ore in some research.^3,4,5) This technology combines metallurgy and mineral processing technology together, in which iron oxides are primarily reduced to metallic iron with grain growth, and then the reduction roasted product was ground to liberation size followed by magnetic separation. The iron product obtained by this process generally contains more than 90 mass% Fe, which was expected to be a good substitute of steel scrap in electric arc furnaces for steelmaking. Iron recovery, which is around 90 mass%, is much greater than that achieved by conventional means. Another attractive feature of this process is that untreated non-coking coal was used as reductant instead of coke. It was shown in the concerned literatures that different types of coal, from lignite to anthracite with the fixed carbon content range from 30 mass% to 90 mass%, have been used for the reduction of low grade iron ores.^6,7)

A variety of carbonaceous materials such as graphite, coke, coal, coal char, biomass, and plastics have been employed as reductant for reducing high-grade iron ores by researchers.^8,9,10,11) It is generally recognized that the overall rate of carbothermal reduction reaction of hematite is controlled by the gasification of coal, i.e. Boudouard reaction, and the reduction can be improved by using reductant with higher reactivity or adding Boudouard reaction catalyst.^10,12,13) Although lots of studies have been done on the influence of reductant type on the reduction of iron oxides, few on the effect of reductant type on the reduction-magnetic process of the low grade iron ores. The major concerns of that process is not only reducing iron oxides to metallic phase, but also transforming the finely-divided metal into particles which are amenable to separation.³⁾ Xu⁷⁾ and Li⁶⁾ conducted the experiments of the effect of coal type on direct reduction and phosphorus removal of high-phosphorus oolitic hematite ore, but in their experiments the amount of coals was calculated on the mass ratio of ore/coal rather than the molar ratio of C/O (Fixed carbon/Oxygen). It is well known that the fixed carbon in the coal is the major content contributes to the reduction of iron ore, so in that case the effect of coal type cannot be revealed accurately. In the study of Zhang, the calculation method of coal consumption was not described.¹⁴⁾ Four types of reducing agent were employed by S. Weissberger and Y. Zimmels³⁾ for the direct reduction of Ramin oolitic iron ore, and significant differences of the iron contents in concentrates and iron recoveries were observed, however the mechanism of the effect of reductant type on this process was not studied.

The objective of this paper is to explore the influence of coal type on the reduction and magnetic separation of a high phosphorus oolitic hematite ore, the effects of particle size of coal and C/O mole ratio were studied as well.

2. Experiment

2.1. Materials

The high-phosphorus oolitic hematite ore used in this study was obtained from Hubei Province, Iron ore assayed 43.58 mass% Fe, 0.04 mass% FeO, 0.83 mass% P, 17.10 mass% SiO₂, 9.28 mass% Al₂O₃, 3.58 mass% CaO and 0.59 mass% MgO. The detail information of the iron ore has been described in a previous paper.⁴⁾ The iron ore was crushed to 100 mass% passing 1 mm.

Three kinds of coal with different rank, namely lignite, bitumite and anthracite in the ascending order, were used as reductant. These coals were crushed and screened to three size fractions: coarse-sized (−0.9~+0.6 mm), medium-sized (−0.45~+0.1 mm) and fine-sized (−0.074 mm). Each size of the three coals was analyzed, respectively, and the results are given in Table 1. It can be seen that the nature of different particle sizes of the same coal were basically the same.

Table 1. The proximate analysis of coals (mass%).

Coal type	Particle size	Cfix	Moisture	Ash	Volatile	S
Anthracite	-0.9~+0.6 mm	83.78	1.21	6.20	8.81	0.07
	-0.45~+0.1 mm	83.40	1.47	6.42	8.71	0.11
	-0.074 mm	81.86	1.22	7.56	9.36	0.12
Bitumite	-0.9~+0.6 mm	56.69	6.63	11.93	24.75	0.12
	-0.45~+0.1 mm	56.56	6.77	13.09	23.58	0.13
	-0.074 mm	54.92	6.55	14.4	24.13	0.12
Lignite	-0.9~+0.6 mm	38.52	10.31	4.82	46.35	0.20
	-0.45~+0.1 mm	38.13	10.27	5.43	46.17	0.20
	-0.074 mm	37.68	10.20	6.15	45.97	0.20

2.2. Composite Briquette Formation, Reduction and Separation

The mixture of iron ore and coal was pressed to form briquette before subjecting to reduction roasting by the following procedures: first, 20 g iron ore, a certain amount of coal, and water (8~12 mass%) were mixed together, and then the mixture was pressed to form briquette using a die with a size of 30 mm diameter with the aid of hydraulic equipment. The amount of reductant was calculated based on the mole ratio of the fixed carbon to removable oxygen of the iron oxides (C/O ratio). The mixing ratios of iron ore and each coal are shown in Table 2 at the C/O ratio of 1.0.

Table 2. Mixing ratio of iron ore and each coal in the briquette at the C/O ratio of 1.0.

Coal type	Coal size	Iron ore : Coal (mass ratio)
Anthracite	-0.9~+0.6 mm	100 : 16.72
	-0.45~+0.1 mm	100 : 16.80
	-0.074 mm	100 : 17.11
Bitumite	-0.9~+0.6 mm	100 : 24.71
	-0.45~+0.1 mm	100 : 24.77
	-0.074 mm	100 : 25.51
Lignite	-0.9~+0.6 mm	100 : 36.37
	-0.45~+0.1 mm	100 : 36.74
	-0.074 mm	100 : 37.18

The reduction roasting was carried out in a muffle furnace with a temperature control programmer. In a run two briquettes were put in a graphite crucible with a cup of 70 mm in diameter and 75 mm in height. The graphite crucibles with sample were put into the furnace after the furnace temperature reached 1200°C and held for 40 min. The holding time was determined by referring to the previous papers which revealed that 40 min of holding time was enough for the reduction of the composite briquette at the temperature of 1200°C and extending holding time may cause the reoxidation of the reduced briquette.^4,15) When the crucible was put in the furnace, the furnace temperature decreased about 100°C and it recovered in about four minutes. After 40 min, the graphite crucible were taken out of the furnace and cooled to room temperature under ambient atmosphere. The schematic of the muffle furnace, crucibles and the briquette used is given in Fig. 1.

Fig. 1.

Schematic of the muffle furnace, crucibles and the briquette used in the experiments.

After the reduced briquettes had cooled in the crucible, they were crushed to −2 mm and then treated by two-stage grinding and wet magnetic separation. The grinding experiments were conducted in a rod mill (RK/BM-1.0L, Wuhan Rock Crush & Grind Equipment Manufacture Co., Ltd, China) having ten Φ15 mm×120 mm rods at 60 mass% solid density and with a speed of 192 r/min. The first stage grinding time was 10 min. The XCGS-73 magnetic tube with a magnetic field intensity of 1120 Oe was used to recover metallic iron from the slurry. The magnetic products obtained from the first separation were reground for 40 min and separation.¹⁵⁾ The main evaluation indexes of test results were the iron content, P content and the iron recovery of DRI powder. The iron recovery was calculated as follows:   

$$\begin{array}{l}Iron\text{ recovery=}\\ \frac{WeightofDRIpowder*ironcontentofDRIpowder}{Weightofraw\text{ ore}*ironcontentofraw\text{ ore}}*100\%\end{array}$$

2.3. Analysis and Characterization

The chemical analyses were conducted by China University of Geosciences (Beijing) analysis laboratory. X-ray diffraction (XRD, Rigaku DMAX-RB, Japan) using Cu Kα radiation and a secondary monochromator were used to identify the formed phases; the samples were scanned over the 2θ range of 10° to 90°. Scanning Electron Microscope with Energy Dispersive Spectrum (Carl Zeiss EVO18) analyses were carried out on reduced briquettes mounted in epoxy resin and polished.

3. Results

3.1. Effect of Coal Types on the Reduction Followed by Magnetic Separation of Briquettes

The effect of coal type on the reduction followed by magnetic separation of high-phosphorus oolitic hematite ore was investigated. The experiments were performed at conditions of roasting temperature 1200°C, roasting time 40 min and C/O ratio 1.0. The experimental results are shown in Fig. 2.

Fig. 2.

Effects of coal type and particle size on the reduction and separation of briquettes.

Figure 2(a) shows the influence of coal type and coal size on the iron recovery. Compared with anthracite, bitumite and lignite presented an advantage on iron recovery when coarse-sized and medium-sized coals were used; however, when fine-sized coals were used, the iron recovery obtained by using anthracite as the reductant was higher than that of the others’. It means that the effect of coal size on the iron recovery varied as the coal type. When decreasing the particle size of anthracite from coarse-sized to medium-sized and then to fine-sized, the iron recovery increased significantly from 65.95 mass% to 83.10 mass% and then to 91.97 mass%; however, decreasing the particle sizes of bitumite and lignite from coarse-sized to medium-sized, iron recovery just increased slightly from 90.95 mass% to 91.81 mass% for bitumite, and 88.33 mass% to 88.41 mass% for lignite, respectively, and the further decrease of particle size of those two coals resulted in slight decrease of iron recovery.

Figure 2(b) illustrates the effects of coal type and coal particle size on the iron content of the DRI powder. It is evident that when the same size of different coals were used, the iron content of the DRI powder resulted from anthracite performing the best result, while DRI powder obtained from lignite reduction had the lowest iron content. This confirmed that iron content of the DRI powder products increased correspondingly with the rank of the coal depended on fixed carbon contents. Furthermore, for these three types of coal, decreasing particle size all led to the decrease of iron content of DRI powder. DRI powder with the highest iron content of 94.55 mass% was obtained with coarse-sized anthracite as reductant and the lowest iron content of 83.50 mass% was obtained with fine-sized lignite.

Figure 2(c) demonstrates the effects of type and particle size of coal on the P content of magnetic product. With the decrease of anthracite size, P content increased from 0.12 mass% of coarse-sized to 0.21 mass% of medium-sized, and then to 0.27 mass% of fine-sized; with the same size decrease, P contents were 0.25 mass%, 0.23 mass% and 0.25 mass% with bitumite, and 0.27 mass%, 0.20 mass% and 0.24 mass% with lignite. These results supported that the P of the DRI powder cannot be reduced to a low level without additives, and the law of the effects of type and particle size of coal on the P content of DRI powder are not obvious.

3.2. Effect of Coal Dosage on the Reduction Followed by Magnetic Separation of Briquettes

In order to evaluate the influence of coal dosage on the recovery of iron, a series of experiments were carried out by varying the C/O ratio from 0.4 to 1.0, fine-sized coals were used, and the roasting temperature and time were of 1200°C and 40 min, respectively.

Figure 3(a) shows that, as the C/O ratio varied from 0.4 to 1.0, the iron recovery exhibited a significant improvement, from 45.51 mass% to 91.97 mass%, when anthracite was used as a reductant. When bitumite was used, the iron recovery increased from 50.84 mass% of 0.4 C/O ratio distinctly to 90.87 mass% of 0.8 C/O ratio, further increasing of C/O ratio to 1.0 did not contribute to significant increase of iron recovery. Using lignite as reductant, the recovery of iron increased from 50.25 mass% of 0.4 C/O ratio to 80.26 mass% of 0.6 C/O ratio, and further increase of C/O ratio led to slight improvement of iron recovery. At the C/O ratio of 0.4 and 0.6, the iron recoveries obtained from bitumite and lignite were almost the same and considerably higher than that of anthracite; however, this advantage of lignite and bitumite was gradually lost as C/O ratio further increased with iron recoveries of bitumite kept higher than that of lignite.

Fig. 3.

Effects of coal dosage on the reduction and separation of briquettes.

Figure 3(b) shows that, (1) the iron content of the DRI powder obtained by using anthracite as reductant was higher than that of the one prepared by using bitumite, and the DRI powder achieved by using lignite as reductant performed the lowest iron content, these results are in line with those shown in Fig. 2(b); (2) when anthracite was used, the iron content increased from 94.01 mass% to 95.36 mass% along with C/O ratio rose from 0.4 to 0.8, and then dropped to 91.50 mass% as C/O ratio increased to 1.0; when bitumite was used, the iron content of the DRI powder increased from 92.75 mass% to 94.37 mass% with increasing of C/O ratio from 0.4 to 0.6, and further increase of C/O ratio to 1.0 resulted in obvious decrease of iron content of DRI powder to 85.39 mass%; (3) when using lignite as reductant, the iron content of the DRI powder decreased linearly from 92.94 mass% to 83.50 mass% with increase of C/O ratio from 0.4 to 1.0.

Figure 3(c) reveals that the P content of the DRI powder products increase with increasing of C/O ratio from 0.4 to 0.8 for all types of reductants. This phenomenon may stemmed from two aspects: 1) the reduced rate of fluorapatite was improved by increasing the C/O ratio, and the increase of reduction rate promotes P melting into the metallic iron; 2) more P-containing slag was mingled with concentrate in some case as the C/O ratio increased, such as when lignite was employed, the iron content decreased with increasing of C/O ratio throughout the range studied which indicated that more of slag was mingled with concentrate. For bitumite used product, the same issue existed as the C/O ratio excessed 0.6. As the C/O ratio increased to 1.0, slightly drop of P content of the DRI powder were observed on all these three coals. In addition, the P content of the DRI powder resulted from lignite is similar to that from the bitumite, and their P contents are higher than that from anthracite at the C/O ratio of 0.4 and 0.6. But opposite result was achieved as the C/O increased.

4. Characteristics of Roasted Briquettes

4.1. XRD Analysis

XRD was used to study the effects of coal type and coal particle size on the phase transitions in the briquettes. The C/O ratio was kept at 1.0, and the roasting temperature and time were of 1200°C and 40 min, respectively. The results are given in Fig. 4.

Fig. 4.

XRD patterns of the reduced briquettes with different coals as reductant: (a) with anthracite; (b) with bitumite; (c) with lignite.

As shown in Fig. 4(a), when anthracite was used as a reductant, a broad band indicating the presence of glass phase materials was observed between 10° and 40° in XRD patterns. Therefore, it could be inferred that there was certain amount of liquid phase formed in the briquettes during reduction process. Furthermore, as the particle size of coal decreased, the broad band reduced, while the intensity of the patterns of metallic iron and quartz increased significantly, indicating that the amount of the liquid phase decreased and the reduction of iron oxides was promoted along with the reduction of anthracite particle size. However, the broad band was almost not observed in the XRD patterns of the reduced briquettes with bitumite (Fig. 4(b)) and lignite (Fig. 4(c)) as reductants, and also the intensity of the patterns of metallic iron and quartz increased very slightly as the particle sizes of those two coals decreased. These results confirmed that, compared with briquettes using anthracite as reductant, fewer of liquid phase formed in those when bitumite and lignite were used as reductants. It also exhibited that the reduction of iron oxides were almost complete when those coals were coarse-size.

4.2. SEM Observation

The reduced briquettes studied in section 4.1 were polished and observed by SEM. The results were shown in Fig. 5.

Fig. 5.

SEM images of the reduced briquettes with different reductants:

(a)–(c) coarse-sized, medium-sized and fine-sized anthracite;

(d)–(f) coarse-sized, medium-sized and fine-sized bitumite;

(g)–(i) coarse-sized, medium-sized and fine-sized lignite;

(j) Partial enlarged view of the (i).

It can be observed that the melting degree of reduced briquettes went up significantly with increase of the rank and particle size of coal. These results were consistent with XRD results shown in Fig. 5. Moreover, the particle size of metallic iron (white) decreased as the reductants used in the order of anthracite (Fig. 5(a)–5(c)), bitumite (Fig. 5(d)–5(f)) and lignite (Fig. 5(g)–5(j)). When the same coal was used, reducing particle size of coal results in the decrease of the particle size of iron grains. These results explained why the iron content of the DRI powder decreased as the decrease of the coal rank as well as the particle sizes of coals. Additionally, it can be seen from Fig. 5(j) that, when fine-sized lignite was used as reductant, many ultrafine metallic iron grains were formed which are difficult to recover via magnetic separation process, and this may explain why using fine-sized lignite led to a decrease of iron recovery.

5. Discussion

It is well known that the reduction of hematite by coal proceeds mainly through gaseous intermediates of CO and CO₂. The reaction sequence can be represented as follows:¹⁶⁾   

$${\text{3Fe}}_2{\text{O}}_3+\text{CO}\to {\text{2Fe}}_3{\text{O}}_4{+\text{CO}}_2$$(1)

  

$${\text{Fe}}_3{\text{O}}_4+\text{CO}\to {\text{3FeO}+\text{CO}}_2$$(2)

  

$$\text{FeO}+\text{CO}\to {\text{Fe}+\text{CO}}_2$$(3)

  

$${\text{C}+\text{CO}}_2\to \text{2CO}$$(4)

The reduction of FeO to Fe, which requires a stronger reducing atmosphere, is more difficult to process than the reaction (1) and (2).¹⁷⁾ Furthermore, since the raw ore contains a large number of SiO₂, Al₂O₃, and CaO, part of the FeO may react with SiO₂ to form fayalite which has a low melting point and was harder to be reduced than FeO.⁵⁾ Moreover, the fayalite will react with other oxides contained in the ore to generate molten phase.¹⁸⁾

It is worth mentioning that the low melting point materials play an essential role in the growth of the metallic iron particles which can facilitate the transfer of iron phase. As noted above, the coarsening of the iron grains in the reduced ore is essential for the effective separation of iron and slag. That is to say, although the formation of iron-rich slag decreased the metallization rate of the reduced ore, it can promote the separation of metallic iron and slag by improving the coalescence of iron grains. When fixing the roasting conditions and the composition of the briquettes, the amount of low-melting substances formed in the briquettes during reduction process depends on the environment atmosphere. The stronger reducing atmosphere, the higher metallization rate and smaller amount of low-melting substances will form, and the coarsening of metallic iron will be inhibited, which in turn will decrease the iron content of the DRI powder. In the present study, the environmental atmosphere surrounds the briquettes and within it was determined by the gasification reactivity of the reductant, coal size, and C/O ratio.

It has been widely accepted that the gasification reactivity of coal decreases as coal rank increases, namely the reactivity of lignite, bitumite and anthracite ranks in the descending order.¹⁹⁾ Using the coal with high reactivity as reducing agent can improve the Boudouard reaction and provide a strong reducing atmosphere. Therefore, when anthracite, which owns the lowest reactivity among these three reductants, was used as a reducing agent, more FeO will react with gangue to form low melting point materials since the weak reducing atmosphere was provided. As a result, the coalescence of metallic iron particles was promoted, which improved the separation of metallic form slag consequently. However, when lignite, which presents the highest reactivity, was used as reductant, although the reduction of hematite was improved, the coalescence of metallic iron was hindered since very little of low-melting substances generated (Fig. 4(c)). The reactivity of bitumite ranks between that of lignite and anthracite, therefore the amount of low-melting substances produced in the briquettes with bitumite as reductant ranks between that with lignite and anthracite as reductants. These explained why the iron content of DRI powder increased as coal rank increased.

Additionally, Table 1 shows that the content of volatile decreased with increasing of coal rank, the volatile matters contained in the coal also can strengthen the reducing atmosphere during reduction process. Whereas the contribution of volatile matters on the reduction of iron oxides can be negligible since they almost completely released before the sample temperature reaching an active temperature of reduction reaction, and this also confirms that the reactivity of coal plays a critical role in the reduction process. Particle size reduction of coal has a great effect on increasing reactivity,²⁰⁾ and consequently promotes the reduction of iron oxides and inhibits the growth of the iron grains, and this explained why the iron content of the DRI powder decreased as the particle size of coal reduced. In addition, the iron recovery was just 65.95 mass% when coarse-sized anthracite was used at the C/O ratio of 1.0 as shown in Fig. 2(a). It is because the reduction of iron oxides was hindered in the weak reducing atmosphere.

The ash content of the bitumite was higher than that of anthracite and lignite as shown in Table 1. The coal ash mainly consists of SiO₂, Al₂O₃ and CaO which is similar to that of the gangue in the iron ore, the ash content of the coal was considered not conducive to the reduction of high grade iron concentrate since these oxides react with FeO to form some refractory substances.¹⁷⁾ However, the absolute content of ash contained in these coals is much less when compared with the gangue in low grade iron ore. Therefore, it can be inferred that the effect of coal ash on reduction of oolitic iron ore can be negligible.

Increasing the C/O ratio can also enhance the reducing atmosphere, and suppressed the production of low smelting point materials. It can be seen from Fig. 2(b) that when bitumite and anthracite were used as reductants, the iron content of the DRI powder increased with increasing of C/O ratio firstly and then decreased with further increasing of C/O ratio. This may be explained as follows: when the C/O ratio keep at low level, the particle size of iron grain increased with the C/O ratio increasing since more of metallic iron was formed. However, as the C/O ratio excess some level, the growth of iron grains will be inhibited because of the amount of low melting point materials decreased as the C/O ratio increased.

In summary, the major concerns of coal-based reduction-magnetic process for recovering DRI powder from the high-phosphorus oolitic hematite ore is reducing the vast most of iron oxides to metallic iron while promoting the coalescence of iron grains by sacrificing a few amount of FeO to form low-melting materials.

6. Conclusions

In this study, coal-based direct reduction followed by magnetic separation technique was employed to extract DRI powder from high-phosphorus oolitic hematite ore, the effects of coal type, coal size and the C/O ratio on the process were investigated at the temperature of 1200°C. From these investigations, the following conclusions can be drawn.

(1) Generally, the recovery of iron can be improved by using coal with high reactivity and smaller size. However, these measures also decreased the iron content of the DRI powder for strengthen of the reducing atmosphere inhibited the coalescence of reduced iron, which in turn deteriorates the conditions of separation of metallic iron and slag.

(2) The effect of coal size on recovering DRI powder weakens with decrease of the coal rank. Particle size decrease of anthracite has a considerable effect on increasing the iron recovery, while decreasing the particle sizes of bitumite and lignite results in slight changes of iron recovery.

(3) The major concerns of coal-based reduction-magnetic process for recovering DRI powder from the oolitic iron ore is reducing the vast most of iron oxides to metallic iron and promoting the coalescence of reduced iron by sacrificing a few amount of FeO to form low-melting materials. If the reducing atmosphere is too strong, very little of low-melting materials will be formed, and also the metallic iron grains formed will be very small, which will not only results in the decrease of iron content of DRI powder, but also reduces the iron recovery. If the reducing atmosphere is too weak, iron recovery decreased since the reduction of iron oxides was hindered.

(4) The law of the effects of type and particle size of coal on the P content of DRI powder is not obvious. Increasing the C/O ratio resulted in the increase of P content of the DRI powder since the reduction of fluorapatite was promoted.

Acknowledgement

The authors wish to express their thanks to the Natural Science Foundation of China (No. 51134002) for the finance support for this research.

References

• 1)   J. B.  Maynardand and  F. B.  Van Houten: Descriptive model of oolitic ironstones, U. S. Geological Survey, Washington, D.C., (1992), 39.
• 2)   Y.  Zhao and  C.  Bi: Mineral Deposits, 19 (2000), 350 (in Chinese).
• 3)   S.  Weissberger and  Y.  Zimmels: Int. J. Miner. Process., 11 (1983), 115.
• 4)   W.  Yu,  T.  Sun,  J.  Kou,  Y.  Wei,  C.  Xu and  Z.  Liu: ISIJ Int., 53 (2013), 427.
• 5)   Y.  Li,  T.  Sun,  J.  Kou,  Q.  Guo and  C.  Xu: Miner. Process. Extra. Metall. Rev., 35 (2014), 66.
• 6)   Y.  Li,  T.  Sun and  C.  Xu: Min. Metall. Eng., 32 (2012), 66 (in Chinese).
• 7)   C.  Xu,  T.  Sun,  C.  Qi,  Y.  Li,  X.  Mo,  D.  Yang,  Z.  Li and  B.  Xing: Chin. J. Nonferrous Met., 21 (2011), 680. (in Chinese)
• 8)   G.  Wang,  Y.  Ding,  J.  Wang,  X.  She and  Q.  Xue: Int. J. Miner. Metall. Mater., 20 (2013), 522.
• 9)   T.  Sharma: Int. J. Miner. Process., 39 (1993), 299.
• 10)   J. S. J.  Van Deventer and  P. R.  Visser: Thermochim. Acta, 111 (1987), 89.
• 11)   Y.  Ueki,  R.  Mii,  K.  Ohno,  T.  Maeda,  K.  Nishioka and  M.  Shimizu: ISIJ Int., 48 (2008), 1670.
• 12)   T.  Coetsee,  P. C.  Pistorius and  E. E.  de Villiers: Miner. Eng., 15 (2002), 919.
• 13)   J.  Moon and  S.  Veena: Metall. Mater. Trans. B, 37 (2006), 215.
• 14)   Q.  Zhang,  G.  Qiu and  Q.  Xiao: J. Cent. South Univ. Technol., 28 (1997), 126 (in Chinese).
• 15)   W.  Yu,  T.  Sun,  Z.  Liu,  J.  Kou and  C.  Xu: ISIJ Int., 54 (2014), 56.
• 16)   Y. K.  Rao: Metall. Tran., 2 (1971), 1439.
• 17)   R. J.  Fruehan: Metall. Trans. B, 8 (1977), 279.
• 18)  M. T and I. T: J. MMIJ, 116 (2000), 141.
• 19)   K. C.  Xie: Coal structure and its reactivity, Science Press, Beijing, (2002), 290 (in Chinese).
• 20)   E.  Hippo and  P. L.  Walker, Jr.: Fuel, 54 (1975), 245.