Tin and zinc separation from tin, zinc bearing complex iron ores by ...

Tin and zinc separation from tin, zinc bearing complex iron ores by selective reduction process Y. B. Zhang1, T. Jiang*2, G. H. Li1, Z. C. Huang1 and Y. F. Guo1 Tin, zinc bearing complex iron ores are typically intractable and have not been efficiently utilised in China. In this investigation, the process mineralogy of the tin, zinc bearing iron ores and reduction behaviours of iron, tin and zinc oxides by CO were investigated. A selective reduction roasting process was initially developed to separate tin and zinc from the complex iron ores. Under optimum conditions, most of the tin and zinc were effectively removed from the iron ore pellets, and the roasted pellets could be used as high quality ironmaking burdens for large scale blast furnaces. Keywords: Complex iron ore, Selective reduction, Tin, Zinc, Iron ore pellet

Introduction China is rich in iron ore resources, but most of these ores are low grade, polymetallic and refractory ones.1,2 It is crucial to accelerate the exploitation of these complex iron ores in order to maintain sustainable development of the Chinese iron and steel industry. Tin, zinc bearing iron ores are typically intractable ones, great reserves of which are found in China.3,4 If tin and zinc are not effectively removed from the iron ores, they cannot be directly used as ironmaking materials because tin and zinc have negative effects on the steel production processes. Zinc oxides will be reduced and volatilised during the operation of blast furnace (BF) process. Then, zinc accretion is easily formed on the inner wall of BFs, which affects the normal production of ironmaking. If tin is residual in the molten iron and steel, mechanical performance of steel products becomes poor. In order to utilise these complex iron ores, much research has been carried out since 1970. The former literature indicated the following: (i) iron concentrates pretreated by multistage grinding separation process were still not able to be directly used as BF burdens because of the high tin and zinc in them (both of tin and zinc content . 0?1%)4 (ii) tin and zinc could be effectively removed from the ores by selective sulphurisation and chlorination roasting methods; however, there were serious environmental pollution and equipment corrosion5–7 (iii) strong reduction roasting processes to produce metallic pellets might realise effective utilisation 1 Peace Building, No. 244, School of Minerals Processing and Bioengineering, Central South University, Changsha 410083, China 2 Biology Building, No. 320, School of Minerals Processing and Bioengineering, Central South University, Changsha 410083, China *Corresponding author, email [email protected]

ß 2011 Institute of Materials, Minerals and Mining Published by Maney on behalf of the Institute Received 13 March 2011; accepted 17 June 2011 DOI 10.1179/1743281211Y.0000000036

of iron, tin and zinc, while the industrial production of coal based direct reduction process had proven that the process had low productivity and high production cost;8–10 therefore, these complex tin, zinc bearing iron ores have not been efficiently utilised hitherto. In order to develop an effective process to utilise these complex iron ores, process mineralogy of the tin, zinc bearing iron concentrates was first studied in this research. Then, reduction thermodynamics of iron, tin and zinc oxides by CO was analysed. Based on these, technological parameters, including reducer type and dosage, roasting temperature and time, were carried out. Finally, a new process for tin and zinc separation and pellet preparation for BFs from tin, zinc bearing iron concentrates by selective reduction were put forward.

Experimental Materials Iron concentrates

The tin, zinc bearing iron concentrates used were taken from a mineral processing plant in the Inner Mongolia Autonomous Region of China. The mass percentage of the concentrates with granularity below 0?075 mm was found as 63?52%. Their main chemical compositions by X-ray fluoroscopy are given in Table 1. As shown in Table 1, the iron concentrates belong to typical magnetite.11 The mass percentages of tin, zinc and sulphur were 0?245, 0?217 and 0?193% respectively, exceeding the limitation of ironmaking materials. Generally speaking, high quality BF materials require that the contents of tin, zinc and sulphur are no more than 0?08%. A former research presented that the existing shapes of the main metallic elements Fe, Sn and Zn in the iron concentrates were cassiterite (SnO2), sphalerite (ZnS) and magnetite (Fe3O4).12 Using SEM, the distribution characteristic of main metal minerals (magnetite,

Ironmaking and Steelmaking

2011

VOL

38

NO

8

613

Zhang et al.

Tin and zinc separation from tin, zinc bearing complex iron ores

a

b

a monomer form; b enwrapped and symbiotic forms 1 Occurrence forms of cassiterite [by SEM, backscattered electron image (BEI)]

Methods

cassiterite and sphalerite) was studied (shown in Figs. 1 and 2). Figure 1 shows that magnetite and cassiterite were the main minerals of Fe and Sn in the concentrates. Cassiterite existed in three forms: one was an irregularly shaped monomer (shown in Fig. 1a), with fine granularity varying between 0?01 and 0?03 mm; the second one was superfine granules embedded in magnetite (shown in Fig. 1b), and its granularity was changed from 0?005 to 0?02 mm; and the third one was that cassiterite was symbiotic to the rim of magnetite or embedded in the coarse magnetite (shown in Fig. 1b), the granularity of which ranged from 0?005 to 0?05 mm. According to the relationship between sphalerite and neighbouring minerals, the existence form of sphalerite included two forms: one was monomer granules (Fig. 2a), with grain size ranging from 0?03 to 0?09 mm; the other was that superfine magnetite was enwrapped in the coarse sphalerite, with granularity between 0?02 and 0?1 mm (Fig. 2b). The above findings presented that symbiotic relationship of iron, tin and zinc minerals was very complicated, which brought difficulties for separating tin and zinc from the iron concentrates by simple physical separation processes.

The experimental procedure included blending, balling, drying, preheating, reduction roasting, cooling, etc. The experimental flowsheet is shown in Fig. 3. The following is the detailed procedure of the bench scale experiment: first, the concentrates were blended with modified humic acid binders, balled into 12–15 mm diameter green pellets in a W1000 mm disc pelletiser and dried in a baking box for 4 h; second, dry balls were preheated at 920uC for 10 min in a horizontal electrically heated tube furnace in the air. Then, the reduction roasting tests of preheated pellets accompanied with certain quantity of reducers were performed in a vertical tube furnace, also electrically heated, with an internal and external diameter of 80 and 100 mm respectively. The temperature was measured by a Pt–Rh thermocouple and controlled by a KSY intelligent temperature controller. After being reduced, the roasted pellets were discharged from the furnace and cooled to room temperature in a sealed pot. Finally, compression strength, residual tin content, total iron (TFe) and metallic iron (MFe) content of finished pellets were detected. The compression strength of pellets was measured by a universal material testing machine (KL-WS). The loading speed was controlled at 15 mm min21. Twenty pellets were tested for each trial, and the average of all the measurements was used as final compression strength (unit: newton/pellet, N/P). Tin and zinc volatilisation rate was respectively characterised as the deprivation degree of tin and zinc from reduced pellets, which was calculated in accordance with the following expression

Reducing agents

The reducing agents used in this research included lignite, semicoke fine, anthracite and coke breeze. The analyses are listed in Table 2. Fuel ratio (the ratio of fixed carbon/per cent volatile) was an important criterion of the coal combustibility, which was closely correlated to coal reactivity.13 The greater the fuel ratio, the worse the combustibility and reactivity, and combustion was longer. In Table 2, the fuel ratio of lignite (1?23) was the lowest, whereas the fuel ratio of coke breeze (57?52) was relatively the highest; those of semicoke fine and anthracite were 6?18 and 15?48 respectively. The results indicated that the combustibility and reactivity of coke breeze were the worst, but those of the lignite were the best.

c~½1{(1{e%)|a=b|100 where c is the tin or zinc volatilisation rate (%), e is the weight loss ratio (%), a is the residual tin or zinc content in the reduced pellets (%) and b is the original tin or zinc content in the preheated pellets (%).

Table 1 Chemical composition of iron concentrates by X-ray fluoroscopy Components

TFe

FeO*

Sn

Zn

Si

Al

Ca

Mg

Pb

S

Mass-%

64.75

28.17

0.245

0.217

3.83

0.94

2.56

0.36

0.018

0.193

*FeO content is measured by chemical method.

614


2011

VOL

38

NO

8

Zhang et al.


a

b

a monomer form; b symbiotic or enwrapped form 2 Occurrence forms of sphalerite (by SEM, BEI)

Iron metallisation degree was calculated using the following expression g~ðMFe=TFeÞ|100 where g is the iron metallisation degree (%), MFe is the total weight of metallic iron in reduced pellets (%) and TFe is the total weight of iron in reduced pellets (%).

Reduction behaviour of iron, tin and zinc oxides ZnS in the ore concentrates is only reduced with difficultly by CO or coals, whereas ZnO is easily reduced into Zn(g) and volatilised. Therefore, tin, zinc bearing iron concentrate pellets should be first preheated and oxidised in the air. During the preheating, Fe3O4 was easily oxidised into Fe2O3. Most of ZnS was oxidised into ZnO and SO2. The relevant reactions were the following11,14 4Fe3 O4 zO2 2ZnSz3O2

w673 K

w753 K

a-Fe2 O3 2ZnOz2SO2 :

while SnO2 in the pellets was a kind of stable compound under high temperature oxidative atmosphere and still remained in the original form because of the low decomposition pressure of SnO2.6 Reduction roasting was the key procedure to separate tin and zinc oxides from iron ore pellets. In this investigation, solid reducers (lignite, anthracite, etc.) are used as reducers. It is well known that solid carbon is gasified into CO when the temperature is above 1000uC;15 hence, the reduction behaviour of Fe2O3, SnO2 and ZnO by CO was analysed. Generally speaking, using CO as a reducer, Fe2O3 and SnO2 are separately reduced stepwise according to Fe2O3RFe3O4RFeORFe and SnO2RSnO(s)RSn. Previous studies have shown that metallic Sn was easily

compatible with metallic Fe under strong reduction atmosphere, and then Fe–Sn alloy was formed.6,15 Thus, Sn and Fe could not be separated effectively when the roasting atmosphere was highly reducing. The possible reactions of Fe2O3 and SnO2 reduced by CO are9,16–19 3Fe2 O3 zCO~2Fe3 O4 zCO2 Fe3 O4 zCO~3FeOzCO2 (Tw5700 C) FeOzCO~FezCO2 SnO2 zCO~SnO(s) zCO2 SnO(s) zCO~SnzCO2 Joffre considered that SnO(s) had higher saturated vapour pressure (SVP) at high temperature (.1000uC); furthermore, the SVP was increased obviously with the temperature rising.20 At 1100uC, the SVP reached 2?07 kPa, so that SnO(s) could be volatilised as SnO(g) from roasted pellets during high temperature roasting. Therefore, in order to realise the effective volatilisation of tin in the form of SnO(g) by selective reduction, reduction roasting should be controlled under

Table 2 Industrial analysing results of reducers Category

Fixed carbon/% Volatile/% Ash/% Fuel ratio

Lignite Semicoke fine Anthracite Coke breeze

45.93 75.55 80.95 84.55

37.46 12.23 5.23 1.47

16.61 12.22 13.82 13.98

1.23 6.18 15.48 57.52

3 Experimental flowsheet


2011

VOL

38

NO

8

615

Zhang et al.


5 Images (SEM) of no. 3 pellets (BEI)6200

4 X-ray diffraction curve of ‘no. 3 pellets

conditions of FeO (or Fe3O4) and SnO(s) being steady existence at higher temperature. For the reduction of ZnO, when temperature was above 910uC, ZnO was easily reduced into Zn(g), then volatilised into fumes and recovered from the dusting system. The following was the main reduction reaction15,16 ZnOzCO~Zn(g) zCO2 As reported, when the temperature was higher than 1099 K, ZnO(s) was easily reduced into Zn(g).15 Moreover, with increase in reduction temperature and CO partial pressure in the gaseous phase, zinc volatilisation was improved. According to the analyses of reduction behaviour of iron, tin and zinc oxides mentioned above, SnO2 could be reduced into SnO(s), then volatilised as gaseous shape SnO(g) because of its high saturated vapour pressure; ZnO was reduced into Zn(g), then volatilised into fume, and most of Fe2O3 or Fe3O4 was reduced into FeO and residual in the roasted pellets.

Results and discussion Reductant types To begin with, the effects of reductant types were tested. Reductants used were lignite, semicoke fines, anthracite and coke breeze. The experimental conditions were fixed as reduction roasting temperature of 1075uC, reduction time of 50 min, C/Fe weight ratio of 0?3 and reducer grain size of 3–5 mm. The results are presented in Table 3. It is seen that for no. 4 and 3 pellets, the compression strength was 2488 and 2280 N/P, tin volatilisation was 72?54 and 70?98%, and zinc volatilisation was 60?84 and 61?65% respectively. Comparatively, no. 1 and 2 pellets

had lower compression strength and tin volatilisation. The residual tin and zinc percentage in no. 3 and 4 pellets were both ,0?08%, which met the requirements of BF burdens. The X-ray diffraction curve of no. 3 pellets is shown in Fig. 4, indicating that the main mineral composition in the roasted pellets was wu¨stite (greyish white, elliptical or plate-like substances in Fig. 5), a little of magnetite (fine strip, dark substances in Fig. 5) and MFe. Besides, a small quantity of fayalite (dark, black irregular substances in Fig. 5) was also found in the pellets. The results show that reduction of Fe2O3R Fe3O4RFeO occurred, but little FeO was further reduced to MFe. The SEM image of no. 3 pellets indicated that most of wu¨stite crystalline had been recrystallised and connected with one another. Furthermore, the fayalite grain was filled in the gap of wu¨stite crystalline and closely joined with them. Thus, the pellet had a compact internal structure and possessed high compression strength. As listed in Table 2, the lignite and semicoke fines had lower fuel ratio and better reactivity. At higher temperature, the gasification speed of solid carbon was rapid, and large quantities of CO were produced; therefore, CO partial pressure in the gas phase was high, which strengthened the reduction of iron and zinc oxides. Part of the iron oxides were reduced to MFe, and most of ZnO was volatilised into fumes. However, SnO2 was easily reduced to metallic Sn if the atmosphere was sufficiently reducing. Metallic Fe and Sn formed Fe–Sn alloy (shown in Fig. 6); therefore, residual tin contents in no. 1 and 2 pellets were relatively higher. From the above discussion, it was concluded that anthracite and coke breeze could be used as suitable reductants. Then, anthracite was selected as the reductant to carry out the following tests.

Table 3 Effects of different types of reducers on roasted pellets No.*

Compression strength/N/P

Tin volatilisation/%

Zinc volatilisation/%

Residual tin/%

Residual zinc/%

Metallisation rate of Fe/%

1 2 3 4

1868 1805 2280 2488

55.36 56.74 70.98 72.54

73.22 70.69 61.65 60.84

0.114 0.104 0.076 0.068

0.042 0.056 0.069 0.072

23.74 21.84 1.49 1.02

*1: roasted pellets using lignite as reducer; 2: roasted pellets using semicoke fine as reducer; 3: roasted pellets using anthracite as reducer; 4: roasted pellets using coke powder as reducer.

616


2011

VOL

38

NO

8

Zhang et al.


6 Fe–Sn alloy formed in reduced pellet by ligniteFor anthracite and coke breeze, they had high fuel ratio and a lingering combustion phenomenon; hence, the gasification speed of solid carbon was slow, which could provide persistent weak reduction atmosphere under the same roasting conditions as the lignite and semicoke fine. Thus, most of SnO2 was reduced to SnO(s), then volatilised into fumes. Comparatively, zinc volatilisation of no. 3 and 4 pellets was lower than those of no. 1 and 2 pellets because reduction atmosphere was weak.

Reducer dosage (C/Fe weight ratio) Figure 7 illustrates the effects of anthracite dosage (C/ Fe) on the reduction indices. The curves in Fig. 6a show that pellet compression strength initially increased then decreased with increase in C/Fe from 0?1 to 0?6; however, the metallisation presented a rising trend. When C/Fe was between 0?2 and 0?4, the pellet compression strength was stable at 2324–2420 N/P. As seen in Fig. 6b, changing rules of tin and zinc volatilisation were similar to the pellet strength and metallisation. The tin volatilisation increased from 51?76 to 71?86% when C/Fe was changed from 0?1 to 0?2, which reached 72?69% when C/Fe was 0?25. However, tin volatilisation, presenting an apparent descending tendency, was 60?84% when C/Fe was further increased to 0?6. With increasing anthracite dosage, residual carbon after reduction was increased. The solid carbon percentage in roasted residues was analysed and listed in Table 4. When C/Fe was low, CO percentage in the gas phase was much lower, and part of the SnO2 was not yet reduced into SnO(s) and still residual in the roasted pellets as SnO2. The data given in Table 4 indicated that

residual carbon was only 0?79% when C/Fe was 0?1, indicating that solid carbon in the anthracite was completely gasified during the roasting. When C/Fe was 0?25, the reduction atmosphere provided by the carbon gasifying into CO was appropriate to that of SnO2 reduced into SnO(s); therefore, the majority of SnO2 was reduced into SnO(s) and volatilised into fumes in the form of SnO(g). With C/Fe ratio increasing, CO percentage in the gaseous phase was enhanced under the same roasting temperature and time. Part of the FeO and SnO(s) was further reduced into metallic Fe and metallic Sn respectively, which easily formed Sn–Fe alloy. Therefore, the tin volatilisation was decreased in some degree. For zinc volatilisation, it was gradually improved with C/Fe increasing. The optimal C/Fe weight ratio was considered to be 0?25.

Roasting temperature From Fig. 8, it is seen that roasting temperature had a great effect on the reduction index. The pellet compression strength and tin and zinc volatilisation initially increased rapidly then decreased as roasting temperature increased, which almost reached the maximum at

Table 4 Solid carbon content in anthracite residues C/Fe

0.1

0.15

0.20

0.25

0.30

0.4

0.5

0.6

Residual carbon percentage/%

0.79

1.25

2.62

4.17

6.53

9.44

13.80

19.52


2011

VOL

38

NO

8

617

Zhang et al.


7 Effects of C/Fe on reduction indices (roasting temperature, 1075uC; roasting time, 50 min; anthracite grain size, 3–5 mm)

1050–1075uC. If the temperature was further increased to 1100uC, the tin and zinc volatilisation presented a declining trend. During the reduction roasting, most of iron oxides were reduced into wu¨stite (FeO), while FeO was easily combined with gangue minerals (such as SiO2, CaO, etc.) and formed low melting substances.11 The formation of liquid phases reduced the porosity of the roasted pellets, thus reducing tin and zinc volatilisation. Therefore, the optimum roasting temperature should be chosen at 1050–1075uC.

Roasting time Anthracite was used as a reductant. Effects of roasting time were also studied under conditions of roasting temperature of 1075uC, C/Fe of 0?25 and reducer grain size of 3–5 mm. The final results are plotted in Fig. 9. Figure 9a indicates that the pellet strength gradually improved with increase in roasting time. When the time was over 30 min, the compression strength was above 2045 N/P, which was increased a little with further prolongation of roasting time.

8 Effects of roasting temperature on reduction indices (roasting time, 50 min; C/Fe, 0?25; anthracite grain size, 3–5 mm)

In Fig. 9b, tin and zinc volatilisation was markedly improved and almost presented a linear relationship when roasting time increased from 10 to 50 min, which were increased from 24?78 and 15?15% to 72?69 and 59?14% respectively. When the roasting time was further increased, the volatilisation was almost unchanged, which showed that further prolonging time had little effect on the tin and zinc volatilisation.

Pilot tests The pilot experiments were performed on the dynamic analog equipment of the grate pot and rotary kiln in Central South University.10 First, the dry balls were preheated in the grate pot at 920uC for 10 min. Then, the preheated pellets with certain quantity of anthracite were put into the rotary kiln with a diameter of 1000 mm and length of 500 mm. Roasting temperature was controlled between 1060 and 1080uC, roasting time was

Table 5 Performances of roasted pellets under pilot conditions Roasted pellet strengths

618

Primary chemical composition/%

Volatilization/%

Compression strength (N/P)

Tumble index (z6.3 mm)/%

Abrasion index (20.5 mm)/%

Sn

Zn

TFe

FeO

Sn

Zn

2510

97.88

1.12

0.052

0.058

67.66

68.90

80.15

70.26


2011

VOL

38

NO

8

Zhang et al.


gaseous phase SnO(g), ZnO was reduced into Zn(g) to volatilise, and most of Fe2O3 or Fe3O4 was reduced into FeO and residual in the finished pellets. 2. The main mineral in roasted pellet was wu¨stite, coupled with a little magnetite and MFe in it. Most of the wu¨stite was recrystallised and interconnected. Fayalite grains filled in the gaps; hence, the roasted pellets had compact internal structure and possessed high compression strength. 3. Pilot scale experimental results indicated that tin and zinc were effectively separated from the roasted pellets using anthracite as a reductant. The appropriate conditions obtained were roasting temperature of 1060– 1080uC, roasting time of 40–50 min, C/Fe of 0?25 and reducer grain size of 3–5 mm. Under the above conditions, finished pellets with 2510 N/P compression strength, 97?88% tumble index and 1?12% abrasion index were achieved. Residual tin and zinc content of 0?052 and 0?058% means that the pellets could be used as high quality ironmaking burden for large scale BFs.

Acknowledgements The authors want to express their thanks to the National Science Fund for Distinguished Young Scholars (grant no. 50725416) and the National Natural Science Foundation of China (grant no. 50804059) for financial support of this research.

References

9 Effects of roasting time on reduction indices (roasting temperature, 1075uC; C/Fe, 0?25; anthracite grain size, 3–5 mm)

40 min, C/Fe was 0?25 and anthracite grain size was 3– 5 mm. After roasting, the pellets were cooled to room temperature in a sealed pot. The properties of finished pellets were measured and are presented in Table 5. The results indicate that compression strength was 2510 N/P, tumble index was 97?88% and remaining contents of tin and zinc in the finished pellets were both ,0?06%. Compared with the small scale experimental results, all the pilot test indices were improved and met with the requirements of ironmaking burdens for large scale BFs.

Conclusions A novel process of selective reduction roasting was developed to utilise complex tin, zinc containing iron ores. The main conclusions were as follows. 1. Analysis of reduction behaviour of iron, tin and zinc oxides reduced by CO showed that tin and zinc could be separated from iron when CO partial pressure in the gaseous phase was low at high temperature. SnO2 was reduced to SnO(s) then volatilised into fumes as

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

P. Li and Y. Xie: Conser. Util. Miner. Res., 2003, 4, (2), 13–16. J. Zou: Min. Eng., 2003, 2, (1), 1–4. L. Wang: Acta Petrol. Sin., 2002, 18, (4), 575–584. C. Xie: Hum. Nonferrous Met., 1996, 11, (6), 13–17. A. W. Fletcher, D. V. Jackson and A. G. Valentine: Inst. Min. Metall. Trans. C, Min. Proc. Extr. Met., 1967, 76C, (730), 145–153. W. Huang: ‘Tin’; 2001, Beijing, Metallurgical Industry Press. T. Jiang, Y. Zhang, Z. Huang and G. Li: Trans. Nonferrous Met. Soc. China, 2005, 15, (4), 902–907. Y. Chen and T. Jiang: Sintering Pelletising, 1997, 5, (3), 17–20. G. Qiu, T. Jiang, J. Xu and Y. Cai: ‘Cold-bonded pellet direct reduction’; 2001, Changsha, Central South University Press. D. Huang, Y. Zhang, G. Han, G. Li and T. Jiang: Proc. TMS Annual Meeting and Exhibition, Seattle, WA, USA, TMS, February 2010, Vol. 1, 393–401. J. Fu, T. Jiang and D. Zhu: ‘Theory of sintering and pelletising’; 1995, Changsha, Central South University Press. Y. Zhang, L. Chen, G. Li and T. Jiang: J. Cent. South Univ., 2011, 21, (6), 1501–1508. K. Xie: ‘Coal structure and its reactivity’; 2002, Beijing, Science Publishing Press. L. Bentell, L. Norrman and M. Sundgren: Scand. J. Metall., 1984, 13, (5), 308–315. G. Chen: ‘Metallurgy for heavy metals’; 1992, Beijing, Metallurgical Industry Press. Y. Liang and Y. Che: ‘Thermodynamic data of inorganic matters’; 1993, Shenyang, Northeastern University Press. T. Debroy, A. Patankar and G. Simkovich: Metall. Trans. B, 1990, 21B, (6), 449–454. J. Carbo-Nover and F. D. Richardson: Trans. Inst. Min. Metall., 1972, 81, 63–68. S. Seetharamon and Staffanson: Scand. J. Metall., 1977, 6, 143– 144. J. Joffre: Foreign Tin Industry, 1984, 23, (4), 38–46.


2011

VOL

38

NO

8

619