Recovery of Iron from Vanadium Tailings with Coal-Based Direct Reduction Followed by Magnetic Separation

Journal of Hazardous Materials 185 (2011) 1405–1411

Contents lists available at ScienceDirect

Journal of Hazardous Materials

journal homepage: www.elsevier.com/locate/jhazmat

Recovery of iron from vanadium tailings with coal-based direct reduction followed by magnetic separation

Huifen Yang a,∗, Lili Jing a, Baogang Zhang b a Key Laboratory for High-Efﬁcient Mining and Safety of Metal Mines of Ministry of Education, University of Science and Technology Beijing, Beijing 100083, China b Department of Environmental Engineering, Peking University, Key Laboratory of Water and Sediment Sciences, Ministry of Education, Beijing 100871, China article info abstract

Article history: A technique with coal-based direct reduction followed by magnetic separation is presented in this study Received 23 June 2010 for recovering and reusing iron otherwise wasted in vanadium tailings. Process parameters such as usage Received in revised form 13 October 2010 of additives, tailings/reductant/additives ratio, reduction temperature and time, as well as particle size Accepted 14 October 2010 were experimentally determined. The optimum process parameters were proposed as follows: using Available online 23 October 2010 lime as the additive, lignite as the reductant, weight ratios of vanadium tailings/lignite/lime at 100:30:10, reduction roasting at 1200 ◦C for 60 min, and particle size of 98% less than 30 ␮m in the ﬁnal roasted prod- Keywords: uct feeding to magnetic separation. Under these conditions, a magnetic concentrate containing 90.31% Vanadium tailings Coal-based direct reduction total iron and 89.76% metallization iron with a total iron recovery rate of 83.88% was obtained. In addition, Magnetic separation mineralography of vanadium tailings, coal-based reduction product and magnetic concentrate were stud- Iron recovery ied by X-ray powder diffraction technique (XRD). The microstructures of above products were analyzed Metallization iron by scanning electron microscope (SEM) to help understand the mechanism. © 2010 Elsevier B.V. All rights reserved.

1. Introduction several iron recovery techniques reported in the literature. One such example is the use of magnetizing roasting-magnetic sepa- In China, more than 1.2 million tonnes of converter vanadium- ration technique to recover iron from red mud [8,9] and iron ore bearing steel slags are produced each year [1]. Valuable vanadium tailings [10]. In other studies, sponge iron or direct reduced iron in these slags is preferentially recovered by methods such as is produced from pyrite cinder [11], red mud [12,13,9], oily hot sodium salt roasting–leaching and H2SO4–CaF2 acid leaching pro- rolling mill sludge [14] and copper matte smelting slag [15] with cess [1–4]. But the residual solid wastes, called vanadium tailings, a direct reduction-magnetic separation technique. It has been also are still discarded as waste in large quantities. Large areas of land reported that traditional mineral processing techniques being used are needed to store these vanadium tailings, which reduces the for recovering iron from ore tailings and smelting slag [16], nickel area of usable farming land. The construction and maintenance of metallurgical slag [17], blast furnace gas ash [18] and copper con- vanadium tailing disposal sites also increase the production cost verter slag [19]. But ferrous minerals in solid wastes usually have of steelmaking plants. In addition, vanadium tailings are usually small particle size and complicated compositions. It is not easy to of high alkalinity, which may increase the pH of natural water [5]. obtain high-grade iron concentrate directly using traditional min- Even more problematic, the release of heavy metals, such as Cr6+ eral processing techniques [16–19]. Magnetizing roasting or direct and V5+, from these storage sites could cause serious environmental reduction followed by magnetic separation has been demonstrated pollution in soil and water [6,7]. to be effective for iron recovery from solid wastes [8–15]. In this Although total iron content in vanadium tailings is relatively study, recovery of iron from vanadium tailings with coal-based high, its recovery has not been practiced due to high content of direct reduction followed by magnetic separation is investigated. alkaline oxides such as Na2O and K2O and the complexity of the Process parameters that affect the recovery of iron are tested and chemical and mineralogical compositions of the tailings. These fac- optimized. This coal-based direct reduction-magnetic separation tors make the recovery less attractive. However, the increasing technique is demonstrated to be effective to recover iron from demand of iron ore in China and the rapid increase of its price vanadium tailings for the ﬁrst time. in recent years have sparked renewed interests in using these iron-rich solid wastes as secondary iron ore resources. There are 2. Experimental

2.1. Materials

∗ Corresponding author. Tel.: +86 10 62332902; fax: +86 10 62332902. Vanadium tailings used in the study were the residual slags after E-mail address: [email protected] (H. Yang). extracting vanadium oxides from vanadium-bearing steel slags

Table 1 Chemical composition of received vanadium tailings.

Constituents T–Fe P MnO SiO2 Al2O3 CaO MgO Na2OK2O TiO2 V2O5 Cr2O3 Content (wt%) 36.54 0.049 4.54 13.65 1.25 0.34 1.54 3.5 1.19 9.28 0.96 1.80

200 8000 1 1.Fe O 2 3 7000 2.Fe TiO 100 2 5 6000 3.NaFe(SiO )

3 2 -1 0 5000 4.Metallic iron 2 -100 Eq.(1) 4000 3 CPS

/ kg.mol Eq.(2) 1 θ -200 3000 G Eq.(3)

3 Δ 1 1 Eq.(4) 3 2 -300 2000 2 2 Eq.(5) 3 2 2 1000 2 1 4 -400 Eq.(6) 4 Eq.(7) 0 -500 10 20 30 40 50 60 70 600 800 1000 1200 1400 1600 2Theta /Deg. Temperature / K

Fig. 1. XRD pattern of vanadium tailings as received. Fig. 2. The correlation of standard free energy (G ) with temperature for Eqs. (1)–(7). using sodium salt roasting and leaching process. The chemical com- The calculated thermodynamic results for Eqs. (1)–(7) are exhib- position is given in Table 1. ited in Fig. 2 to reflect relationship between G␪ and temperature. The total iron content was approximately 36.54%. But its Na2O Taking Eq. (1) as an example, the relationship between reduction and K2O content were relatively high in the tailings. X-ray pow- rate and pertinent parameters is shown in Eq. (8) [25]: der diffraction (XRD) pattern of vanadium tailings is shown in R r 1 1 1 / Fig. 1. Due to the mineral and compositional complexity of vana- t = + 0 0 − R − (1 − R)2 3 (8) R A −E /RT K · M D dium tailings, we only focused on iron-bearing minerals in XRD exp( 0 ) c 3 e 2 3 2 study. The main iron-bearing minerals were identified as hematite where tR is the time of Fe2O3 reduction (s), A is the reaction con- (Fe2O3) and pseudobrookite (Fe2TiO5). Some acmite (NaFe(SiO3)2) stant, R is the reduction rate defined as percentage of the start and metallic iron also appeared in the tailings. oxides reduced (%), K is the equilibrium constant of iron reduc- The coal used in this study was lignite with compositions listed tion equation, Mc is the weight ratio of coal (%), E0 is the activation in Table 2. The carbon (37.09%) was the main active reagent in the energy (J mol−1 K−1), R is the ideal gas constant (8.314 J mol−1 K−1), direct reduction process. Analytical grade lime (CaO) was used as T is the temperature (K), r0 is the grain radius when reduction rate received. −3 is R (%), 0 is the oxygen contained in Fe2O3 initially (mol m ), 2 −1 and De is the diffusion coefficient in solid phase (m s ). From 2.2. Experimental mechanism Eq. (8), one can see that the parameters influencing the coal-based direct reduction are roasting temperature, roasting time and ratio In China, coal-based direct reduction is the most common tech- of carbon to iron ore. Meanwhile, according to Fig. 2, the addition nique in direct iron reduction. Coal acts as reducing agent, and of CaO can promote the reduction of FeTiO3 generated via Eq. (2) deoxidation reactions are completed in rotary kiln or tunnel kiln by comparing G␪ for Eqs. (3) and (5). In addition, the grinding of [20]. For vanadium tailings, the purpose of coal-based direct reduc- roasted product is necessary because of the complicated dissemi- tion is to reduce ferrous minerals to metallic iron. The reaction nation characteristics of materials [26]. mechanism can be expressed with the following equations [21–24]: 2.3. Experimental method 2Fe2O3(s) + 3C(s) = 4Fe(s) + 3CO2(g) (1) Based on the analysis above, five main factors affecting Fe2TiO5(s) + TiO2(s) + C(s) = 2FeTiO3(s) + CO(g) (2) iron recovery were studied. These parameters were roasting FeTiO3(s) + C(s) = Fe(s) + TiO2(s) + CO(g) (3) temperature, roasting time, lignite ratio (weight percentage of lignite-to-vanandium tailings), CaO ratio (weight percentage of + + = + Fe2O3(s) 2TiO2(s) C(s) 2FeTiO3(s) CO(g) (4) CaO-to-vanadium tailings) and particle size of the roasted product feeding to magnetic separation. Vanadium tailings were thoroughly FeTiO (s) + CaO(s) + C(s) = Fe(s) + CaTiO (s) + CO(s) (5) 3 3 mixed with lignite and CaO at different ratios. Then the mixtures

CaO(s) + Fe2O3(s) = CaFe2O4(s) (6) were roasted at various temperatures and times in a Mufﬂe furnace. After ﬁnishing the roast process the mixtures were immediately + = 2CaO(s) Fe2O3(s) Ca2Fe2O5(s) (7) quenched with water. Next the roasted products were wet grinded

Table 2 General analysis of Lignite.

Ingredients Moisture Ash Volatiles Fixed carbon Sulfur

Content (wt%) 13.18 6.21 43.52 37.09 0.19 H. Yang et al. / Journal of Hazardous Materials 185 (2011) 1405–1411 1407

90 85

80 80 70

75 60 Total iron content Recovery rate of iron 50 70 Metallization rate of iron

40 Total iron content Seperation indexes / % Seperation indexes / % Recovery rate of iron 65 30 Metallization rate of iron

20 60 10 15 20 25 30 35 0 5 10 15 20 Lignite ratio / % CaO ratio / %

Fig. 3. Effect of lignite ratio on iron recovery. Fig. 4. Effect of CaO ratio on iron recovery.

3.1.2. Effect of CaO ratio on iron recovery in a rod mill, and then processed through magnetic separator with To optimize CaO dosage, various CaO ratios were studied while the magnetic ﬁeld intensity of 111 kA/m. The total iron content and keeping other process parameters constant as following: lignite metallized iron content in magnetic concentrate were analyzed by ratio of 30%, roasting temperature of 1100 ◦C, roasting time of chemical method. Then the metallization rate of iron, i.e., metalliza- 40 min, and particle size of roasted product after rod mill at 80% less tion iron (the ratio of metallized iron content to total iron content in than 30 ␮m. The result of CaO ratio on iron recovery is presented magnetic concentrate) was calculated according to chemical anal- in Fig. 4. ysis results, and the recovery rate of iron (the ratio of the amount As shown in Fig. 4, total iron content of magnetic concentrate of iron metal in magnetic concentrate to that in vanadium tailings) gradually increased with the increasing of CaO ratio. But maximum was calculated according to mass balance. metallization rate of iron and recovery rate of iron in magnetic The mineral and compositions of roasted product under concentrate appeared at CaO ratio of 10%. When the CaO ratio fur- optimum roasting conditions and magnetic concentrate were ther increased, metallization rate of iron and recovery rate of iron investigated by X-ray powder diffraction method (XRD) using Cu- decreased gradually. The results can be explained from the ther- K␣ radiation (40 kV, 100 mA) at the scanning rate of 8◦/min from 10◦ modynamic data in Fig. 2. From data for Eq. (3) and (5) shown in to 100◦. The microstructures of above products were also analyzed Fig. 2, the G␪ values of Eq. (5) is always less than that of Eq. (3), by scanning electron microscope (SEM). Finally, the compositional indicating the increasing possibility of reduction FeTiO3 to metallic analysis was carried out using an energy dispersion system (EDS) iron after adding CaO into the mixture roasted. But because the G␪ with the SEM. values of Eq. (6) and (7) are very negative at the temperature higher ◦ than 700 C, CaO can easily react with Fe2O3 to form Ca2Fe2O5 and CaFe2O4. Therefore, if too much CaO was added into vanadium tail- 3. Results and discussion ings, some Fe2O3 might not be reduced to metallic iron. This may explain why CaO ratio of more than 10% decreased metallization 3.1. Iron recovery rate of iron and recovery rate of iron in magnetic concentrate.

3.1.1. Effect of lignite ratio on iron recovery 3.1.3. Effect of roasting time on iron recovery To optimize the amount of lignite addition, various lignite ratios By keeping lignite ratio of 30%, CaO ratio of 10%, the mixtures were tested. Other parameters were kept constant as following: were roasted at 1100 ◦C for different time. The roasted products no CaO addition, roasting temperature of 1100 ◦C, roasting time of were ground to particle size of 80% less than 30 ␮m. The effect of 40 min, and particle size of roasted product after rod mill at 80% less roasting time on iron recovery is shown in Fig. 5. than 30 ␮m. The effect of lignite ratio on iron recovery is shown in As shown in Fig. 5, with the increasing of roasting time, total iron Fig. 3. content of magnetic concentrate slowly increased. The recovery As shown in Fig. 3, total iron content, recovery rate of iron rate of iron and metallization rate of iron in magnetic concentrate and metallization rate of iron in magnetic concentrate increased rapidly increased as the roasting time increased from 20 min to rapidly as the lignite ratio increased from 10% to 30%, but lev- 60 min, and then leveled off after roasting time passing 60 min. eled off at above 30%. As it showed in Fig. 2, the G␪ value of Eq. (1) is much more negative than that of Eq. (4) at tempera- 3.1.4. Effect of roasting temperature on iron recovery ◦ ture of 900–1300 C. Therefore, Fe2O3 is reduced to metallic iron Mixtures consisted of vanadium tailings, lignite and CaO at a more easier than FeTiO3 is formed via Eq. (4). Furthermore, by con- mixing ratio of 100:30:10, were roasted for 60 min under various sidering the G␪ values for Eq. (2) and (3), one can understand temperatures, and then ground to particle size of 80% less than that the reduction of Fe2TiO5 to FeTiO3 is relatively easy, but the 30 ␮m. The effect of roasting temperature on iron recovery is shown reduction of FeTiO3 to metallic iron is relatively difﬁcult without in Fig. 6. the presence of CaO. Therefore, reduction of Fe2TiO5 in vanadium As shown in Fig. 6, with the increasing of roasting temperature, tailings, with no CaO addition, cannot be completed using the coal- total iron content increased gradually. The recovery rate of iron and based direct reduction technique, resulting in low iron content, metallization rate of iron in magnetic concentrate increased rapidly low recovery rate and low metallization rate of iron in magnetic as roasting temperature increased from 900 ◦C to 1200 ◦C and then concentrate. decreased slightly. The highest metallization rate and recovery rate 1408 H. Yang et al. / Journal of Hazardous Materials 185 (2011) 1405–1411

95 90

80 90

70 85 60

80 50 Total iron content Recovery rate of iron Total iron content Seperation indexes / % 40 Metallization rate of iron Seperation indexes / % 75 Recovery rate of iron Metallization rate of iron 30 10 20 30 40 50 60 70 80 90 100 70 Roasting time / min. 80 85 90 95 100 Particle size (-30um) / % Fig. 5. Effect of roasting time on iron recovery. Fig. 7. Effect of particle size of roasted product on iron recovery.

of iron were obtained at the temperature of 1200 ◦C. As showed in Fig. 2, the reduction reactions (1)–(5) can be improved with the 90.31% total iron and 89.76% metallization iron with recovery rate increase of temperature. But high temperature also increases the of iron of 83.88%. So it can be used as raw material for steelmaking formation of Ca2Fe2O5 and CaFe2O4 according to Eq. (6) and (7), [27]. which decreases the portion of metallic iron. Hence, the optimal ◦ roasting temperature was determined to be 1200 C. 3.2. Analysis of roasted product

3.1.5. Effect of particle size of roasted product on iron recovery 3.2.1. XRD analysis Because of the complex dissemination characteristics of roasted The roasted product, prepared in optimized conditions as fol- products, they needed to be milled before magnetic separation. lowing: roasting temperature of 1200 ◦C, roasting time of 60 min, The effect of particle size of roasted product was studied here. The weight ratios of vanadium tailings/lignite/lime at 100:30:10 and other conditions were kept constant as following: roasting temper- particle size of 98% less than 30 ␮m in the roasted product, was ◦ ature of 1200 C, roasting time of 60 min, weight ratios of vanadium analyzed by XRD to understand the experimental results discussed tailings/lignite/lime at 100:30:10. Effect of particle size of roasted above. The XRD pattern is shown in Fig. 8. product on iron recovery is shown in Fig. 7. From the XRD pattern, it can be seen that the metallic iron was From Fig. 7, total iron content of magnetic concentrate was dra- the main iron source in the roasted product. The metallic iron is easy matically affected by particle size of roasted product. With the to be recovered by low intensity magnetic separation after grinding decreasing of particle size (more particle less than 30 ␮m), total as experimentally observed in this study. Meanwhile, in compari- iron content in the concentrate increased rapidly while recovery son with the XRD pattern of vanadium tailing, the peaks of Fe2TiO5, rate of iron decreased slightly and metallization rate of iron varied NaFe(SiO3)2 lowered in the roasted products, which indicated that very little. The decrease of particle size improved the monomeric Fe2TiO5, NaFe(SiO3)2 in the tailings were mostly decomposed after liberation of metallic iron particles in roasted product feeding to direct reduction. Moreover, the peaks of Fe2O3 disappeared com- magnetic separation. The monomeric liberation is essential for the pletely, but the peaks of intermediate products such as FeTiO3 and metallic iron particles to be effectively recovered by magnetic sep- FeO appeared in roasted product. Therefore, it can be concluded ␮ aration. At 98% of particles less than 30 m sufﬁcient monomeric that Fe2O3 was completely reduced to metallic iron. Fe2TiO5 was liberation of metallic iron particles was obtained. The magnetic ﬁrstly reduced to intermediate FeTiO3, and then FeTiO3 was further concentrate, produced under this particle size condition, contained

100 1 4000 1.Metallic iron 2.Fe TiO 90 2 5 3.FeTiO 80 3000 3 4.NaFe(SiO ) 70 3 2 5.FeO CPS 60 2000 3

50 4 1 40 1000 2 1 2 5 5 Total iron content 4 3 Seperation indexes / % 30 3 Recovery rate of iron 20 Metallization rate of iron 0 10 20 30 40 50 60 70 80 10 900 1000 1100 1200 1300 2Theta / Deg. o Roasting temperature / C Fig. 8. XRD pattern of roasted product under optimized roasting conditions: vanadium tailings/lignite ratio/CaO at 100:30:10, roasting time of 60 min and roasting Fig. 6. Effect of roasting temperature on iron recovery. temperature of 1200 ◦C. H. Yang et al. / Journal of Hazardous Materials 185 (2011) 1405–1411 1409

Fig. 9. SEM image of received vanadium tailings (a) and roasted product (b), and its corresponding EDS at area A (c) and area B (d), respectively. reduced to metallic iron with the help of CaO. The thermodynamic mechanism in Fig. 2 shows that the reduction of Fe2TiO5 to FeTiO3 is relatively easy, but the reduction of FeTiO3 to metallic iron is more difﬁcult. The XRD results also revealed that residual amount 3500 of FeTiO3 in roasted product was always higher than that of Fe2TiO5. 1 Using the coal-based direct reduction method, Fe2TiO5 and FeTiO3 3000 1.Metallic iron cannot be completely reduced to metallic iron. 2.Ca Fe (SiTiO ) 3 2 4 3 2500

3.2.2. SEM–EDS analysis 2000 Fig. 9 shows SEM image of the roasted product under optimized conditions, and the EDS results measured at different areas are also CPS 1500 presented. By comparing SEM images of received vanadium tailings (a) and 1000 2 roasted product (b), it can be found that small complicated parti- 1 500 2 1 cles in vanadium tailings were changed into larger slag particles and 2 metallic iron particles after reduction roasting. Metallic iron parti- 0 cles, with majority of particle size bigger than 5 ␮m, were loosely scattered on slag particle surface. Clear boundaries between iron 10 20 30 40 50 60 70 80 particles and slag particles were easily seen. Additionally, metallic 2Theta / Deg. iron particles were not embedded into slag particles. So the dis- sociation of metallic iron particles can be easily achieved via the Fig. 10. XRD pattern of magnetic concentrate. 1410 H. Yang et al. / Journal of Hazardous Materials 185 (2011) 1405–1411

Fig. 11. SEM image of magnetic concentrate (e) and its corresponding EDS at area A (f) and area B (g), respectively.

grinding. As shown in Fig. 9(c), the Si and Ti elements did not appear dobrookite (Fe2TiO5) and acmite (NaFe(SiO3)2). The total iron in the EDS of metallic iron particles, which indicated that residual content reaches 36.54%. ferrous minerals such as Fe2TiO5, FeTiO3 and NaFe(SiO3)2, had been (2) Lignite ratio, CaO ratio, roasting temperature, roasting time and encapsulated into slag particles. particle size of roasted product are five main factors which affect the iron recovery. Optimized process conditions are 3.3. Analysis of magnetic concentrate obtained by one-factor at a time tests for recovery iron from vanadium tailings as follows: lignite ratio of 30%, CaO ratio of 3.3.1. XRD analysis 10%, roasting temperature of 1200 ◦C, roasting time of 60 min, The magnetic concentrate was separated from roasted product and particle size of 98% less than 30 ␮m in the roasted prod- under optimized process parameters as above. Its XRD pattern is uct. Under the conditions, the magnetic concentrate obtained shown in Fig. 10. contained 90.31% total iron and 89.76% metallization iron with From Fig. 10, metallic iron was the main iron phase in the mag- recovery rate of iron of 83.88%. netic concentrate. The existence of small amount of Ca3Fe2(SiTiO4)3 (3) The results demonstrate the feasibility of the technique that indicated that the slag particles were removed incompletely. This uses coal-based direct reduction followed by magnetic separa- was the reason that total iron content of magnetic concentrate tion for recovery iron from vanadium tailings. The technique reached only 90.31%, not 100%. provides the possibility of comprehensive utilization of high- iron solid waste. 3.3.2. SEM–EDS analysis Fig. 11 shows SEM image of magnetic concentrate under optimized process parameters as above. The EDS results measured at Acknowledgments different areas are also presented. As seen in Fig. 11, metallic iron particles occupied almost all The authors thank Analytical and Testing Center of University image of magnetic concentrate. Small amount slag with fine par- of Science and Technology Beijing, which supplied us the facilities ticle sizes, containing Si, Ca, Al, Mn, etc, was observed in the to fulfill the measurement. The authors also would like to thank concentrate for intergrowth with metallic iron particles, and it professor Sun Dajun and Lu Yongqiang from the United States for was hard to be removed, which was supported by XRD shown in his generous help in the English language used in this essay. Fig. 10. References 4. Conclusions [1] Q. Qian, Review of technology on extracting vanadium pentoxide from high calcium vanadium-bearing steel slag, Chin. Res. Compr. Utiliz. 27 (2009) 15–17 (1) Major chemical compositions of vanadium tailings are Fe2O3, (in Chinese). SiO , TiO ,KO and Na O with small amount of CaO. Main 2 2 2 2 [2] Q. Qian, Extraction of V2O5 from vanadium-bearing waste slag, Hydrometall. ferrous minerals in the tailings are hematite (Fe2O3), pseu- Chin. 106 (2008) 101–102 (in Chinese). H. Yang et al. / Journal of Hazardous Materials 185 (2011) 1405–1411 1411

[3] B. Voglauer, A. Grausam, H.P. Jörgl, Reaction kinetics of the vanadium roast [15] W.P. Jin, C.A. Joong, H. Song, Reduction characteristics of oily hot rolling process using steel slag as a secondary raw material, Miner. Eng. 17 (2004) mill sludge by direct reduced iron method, Resour. Conserv. Recy. 34 (2002) 317–321. 129–140. [4] R.R. Moskalyk, A.M. Alfantazi, Processing of vanadium: a review, Miner. Eng. [16] Z.Y. Chen, Q.H. Chen, The recovery of iron from ore tailings and smelting slag 16 (2003) 793–805. using strong magnetic separation and gravity separation, Metal. Mine. 399 [5] J.Z. Hao, A.Q. Liu, Utilization of wastes produced in the production process of (2009) 182–184 (in Chinese). vanadium, Chin. Res. Compr. Utiliz. 27 (2009) 7–9 (in Chinese). [17] H.G. Dong, Y.F. Guo, T. Jiang, Study on recovering magnetite from nickel met- [6] W.M. Mayes, P.L. Younger, J. Aumônier, Hydrogeochemistry of alkaline steel allurgical residue with iron, Mine. Metall. Eng. 28 (2008) 37–39 (in Chinese). slag leachates in the UK, Water Air Soil Poll. 195 (2008) 35–50. [18] B.H. Xu, E.J. Wang, J.W. Yang, Research on recovery of iron and carbon from [7] L.H. Xu, W.C. Li, S. Volodymyr, M. Liu, H. Wang, S.M. Bi, Y.B. Bi, Metal recov- blast furnace gas ash, Conserv. Utiliz. Miner. Res. 3 (2007) 51–54 (in Chinese). ery and inorganic eco-materials from tailings by leaching-sintering processes, [19] H. Wang, Recovery of copper and iron in the converter slag from copper smelter, Mater. Manuf. Process. 23 (2008) 743–749. J. Guangdong Nonferr. Met. 13 (2003) 83–88 (in Chinese); [8] Y.F. Sun, F.Z. Dong, J.T. Liu, Technology for recovering iron from red mud by Z.J. Yu, Q.M. Feng, L.M. Ou, Y.P. Lu, G.F. Zhang, Selective leaching of a high-iron bayer process, Metal. Mine. 9 (2009) 176–178 (in Chinese). cobalt matte at atmospheric pressure, Sep. Purif. Technol. 53 (2007) 1–7. [9] L.Y. Li, A study of iron mineral transformation to reduce red mud tailing, Waste [20] Y.J. Li, Z.Y. Zhang, Z.T. Yuan, Y.X. Han, Prospect and application state of high Manage. 21 (2001) 525–534. grade iron concentrate, Metal. Mine. 11 (2006) 5–7 (in Chinese). [10] C. Li, H.H. Sun, J. Bai, L.T. Li, Innovative methodology for comprehensive uti- [21] S.K. Wei, Metallurgical Thermodynamics, Shanghai Science and Technology lization of iron ore tailings Part 1. The recovery of iron from iron ore tailings Press, Shanghai, 1980 (in Chinese). magnetic separation after magnetizing roasting, J. Hazard. Mater. 174 (2010) [22] D.L. Ye, J.H. Hu, Application Manual of Thermodynamic Data for Inorganic Mat- 71–77. ters, Metallurgy Industry Press, Beijing, 2002 (in Chinese). [11] Z.Q. Gong, S. Gong, B. Zhou, A study on techniques of producing sponge iron [23] S.L. Yang, J.F. Sheng, Technology of Ilmenite Meltied into Titanium Slag and Pig from pyrit cinder, Mine. Metall. Eng. 26 (2006) 45–48 (in Chinese). Iron, Metallurgy Industry Press, Beijing, 2006 (in Chinese). [12] W.C. Liu, J.K. Yang, B. Xiao, Application of Bayer red mud for iron recovery and [24] Y.J. Liang, Y.C. Che, Manual of Thermodynamic Data for Inorganic Matters, building material production from alumosilicate residues, J. Hazard. Mater. 161 Northeastern University Press, Shenyang, 1993 (in Chinese). (2009) 474–478. [25] Y.K. Liu, X.G. Mei, Study on coal-based direct reduction of high-iron-content [13] Z.C. Huang, L.B. Cai, Y.B. Zhang, Study on the sponge iron preparation by direct red mud, Sintering Pell. 20 (1995) 5–9 (in Chinese). reduction of high iron red mud by Bayer process, Metal. Mine. 393 (2009) [26] S. Zhou, J. Liu, Laboratory scale mineral processing research on a complex iron 173–177 (in Chinese). ore with Fe-minerals ﬁnely distributed, Nonferr. Met. (Min. Proce. Sec.) 1 (2008) [14] K. Maweja, T. Mukongo, I. Mutombo, Cleaning of a copper matte smelting slag 19–22 (in Chinese). from a water-jacket furnace by direct reduction of heavy metals, J. Hazard. [27] YB/T 4170-2008, Direct reduced iron for steelmaking [S] (in Chinese). Mater. 164 (2009) 856–862.