- Email: [email protected]
Sticking behavior of iron ore–coal pellets and its inhibition
Powder Technology 262 (2014) 30–35
Contents lists available at ScienceDirect
Powder Technology journal homepage: www.elsevier.com/locate/powtec
Sticking behavior of iron ore–coal pellets and its inhibition Jiaxin Li a,c, Rufei Wei b,d,⁎, Hongming Long a, Ping Wang a, Daqiang Cang b,d a
School of Metallurgy and Resource, Anhui University of Technology, Ma'anshan 243002, China School of Metallurgical and Ecological Engineering, University of Science and Technology Beijing, Beijing 100083, China c Key Laboratory of Metallurgical Emission Reduction & Resources Recycling of Ministry of Education, Anhui University of Technology, Ma'anshan 243002, China d State Key Laboratory of Advanced Metallurgy, University of Science and Technology Beijing, Beijing 100083, China b
a r t i c l e
i n f o
Article history: Received 10 November 2013 Received in revised form 10 April 2014 Accepted 17 April 2014 Available online 26 April 2014 Keywords: Iron ore–coal pellets Sticking index Coating Reduction
a b s t r a c t Most previous studies on iron ore–coal pellets (ICPs) have focused on their reduction, but studies on their sticking behavior have not been reported. The sticking behavior of iron ore–coal pellets (ICPs) at high temperatures under a load of 0.1 MPa was studied. Temperature was determined to be an important factor that affects the sticking behavior: the sticking increased with increasing temperature; the sticking index was 89.28% at 1373 K but only 4.78% at 1223 K. The sticking mechanism results from the intergrowth of metallic iron and ferrous oxide crystals between pellet boundaries. High-melting-point materials formed by the combination of MgO and FexOy, can effectively prevent this intergrowth of iron, and the addition of a dolomite powder coating reduces the sticking index from about 90% to about 10%. The addition of limestone and iron ore powder was not found to have any beneficial effect of sticking inhibition. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Ironmaking without the use of coke has attracted significant attention due to a shortage of metallurgical coke resources, with iron ore–coal pellets (ICPs) rapidly emerging as one of the more promising noncoking coal raw materials. ICPs have already been applied to rotary hearth furnaces (RHFs), which were originally developed as a means to treat the fines and waste oxides generated in steel plants [1,2]. Currently, research into ICPs has mostly been focused on reduction, specifically the reduction mechanism and reaction rate [3–8]. However, this neglects other behaviors that occur during the reduction process, such as softening and sticking. Sticking behavior has previously been identified and studied with regard to direct-reduced iron (DRI), which is produced by iron oxide pellets (IOPs) and natural gas [9–13]. Studies performed by Lingyun Yi and co-workers indicated that sticking of IOPs was caused by fibrous iron and fresh iron with high activity. Sticking behavior relieved with the addition of H2 in reducing gas for the porous iron precipitation on the interface [13]. But ICP is different from IOP for it has high coal content and much lower H2 in its reducing gas. However, similar studies into the sticking behavior of ICPs have not been reported, even though this is of great significance to the optimization of current production techniques and reactors, as well as to the development of new reactor designs. The sticking behavior of ICPs was therefore studied through a series of reduction experiments under load. It was found that this phenomenon ⁎ Corresponding author at: School of Metallurgical and Ecological Engineering, University of Science and Technology Beijing, Beijing 100083, China. E-mail address: [email protected] (R. Wei).
http://dx.doi.org/10.1016/j.powtec.2014.04.059 0032-5910/© 2014 Elsevier B.V. All rights reserved.
does exist and that it can be inhibited by coating the ICPs. The mechanisms of sticking and its inhibition are discussed in detail, along with SEM analysis.
2. Experimental 2.1. Materials All raw materials and coatings used in this experiment were obtained from Anhui province in China. Their chemical compositions are listed in Table 1 and Table 2, respectively.
2.2. Apparatus All experiments were conducted in the device shown in Fig. 1. A SiC resistance furnace with a maximum working temperature of 1673 ± 2 K was used for heating. For each reduction experiment, a 500 g ICP sample was placed into the furnace at 473 K, with a total gas flow rate of 5 L/min (4.5 L/min N2, 0.5 L/min CO2), and the sample load was a pressure of 0.1 MPa provided by high-pressure Argon. The samples were then heated at a rate of 6 K/min and held at specified temperatures for 30 min. ICPs' reduction temperatures in previous studies were mostly set between 1173 K and 1473 K generally [15]. The reduction of ICPs is very slow before 1423 K but it becomes faster after 1423 K. So the specified temperatures in this work were 1223, 1323, 1423 and 1523 K. Finally, the reduced pellets were cooled to room temperature in a nitrogen atmosphere.
J. Li et al. / Powder Technology 262 (2014) 30–35
31
Table 1 Chemical composition of iron ore, dolomite and limestone. Composition (%)
TFe
SiO2
Al2O3
FeO
CaO
MgO
S
P
Iron ore Dolomite Limestone
68.11 – –
2.67 1.19 1.15
1.02 – –
27.16 – –
0.11 30.77 53.78
0.26 20.60 0.79
0.28 0.01 0.05
0.03 – –
2.3. Methods Clusters consisting of two or more pellets were each dropped 20 times from a height of 1 m onto a steel plate surface. The percentage of clusters remaining after each drop was calculated as the sticking percentage (SP). The number of drops was plotted against this SP value, with the area under the curve calculated as the sticking index (SI). This SI value is equal to zero when no clusters of two or more pellets remain and 100 when all pellets remain clustered and do not disintegrate during the drop tests. A schematic of the method used for coating the ICPs is shown in Fig. 2. During this coating procedure, raw materials were added to the right pan pelletizers and coatings were added to the left pan pelletizers, respectively. When the pellets attained sufficient size, they were moved out of the right pan and onto the conveyor belt, which transferred them to the left pan. In this left pan, the coating materials were bonded to the outer surface of the pellets to create a coating layer. The coating materials were added according to the quality ratio, which was 7.5% of the raw materials used in this experiment. 3. Results 3.1. Sticking behavior of ICPs and IOPs Initially, the sticking behavior of ICPs and IOPs was studied. The specified temperature for reduction was 1523 K, with the IOPs heated in air rather than in the N2/CO2 atmosphere used with the ICPs. The sticking behavior of ICPs and IOPs is illustrated in Fig. 3, which shows a difference in their respective SP and SI values at 1523 K. Initially, the SP of the IOPs gradually decreased with an increasing number of drops, reaching 7.6% by the 14th drop when almost all of the pellets became separated. However, with the ICPs there was only a very small decrease in SP as the number of drops increased, with the SP still at 99.5% after the 20th drop when almost all of the pellets were sticking. The sticking degree of the pellets is indicated by the SI value at the 20th drop [9,12,13]. For the IOPs, this value was 26.5%, which is considerably lower than the 99.7% obtained with the ICPs; thus, the sticking behavior of the ICPs is far more serious. Consequently, ICPs are far more difficult to smelt using conventional IOP producing reactors, such as grate-rotary kilns or shaft furnaces. It is therefore important to study the sticking behavior of ICPs in order to find ways in which sticking can be inhibited or accommodated by new reactor designs.
Fig. 1. Schematic diagram of the experimental apparatus used.
drop, the SP was still 5.2%, with many pellets still sticking, whereas the SI was 18.0%. The SPs also decreased with an increasing number of drops with the 1423 K and 1523 K samples, but the SPs were notably different from the 1223 K and 1323 K samples. At 1423 K, the SP was maintained at approximately 99.1% up to the 9th drop, with almost all of the pellets sticking together. The SP decreased rapidly by the 11th drop, with some separation of the pellets occurring at this time. Once the 15th drop was passed, the change in SP began to decrease, although it remained at 66.4% by the 20th drop. Similarly, the SI was 89.3% after the 20th drop. With the 1523 K sample, the SP was very high at nearly 100% after every drop due to almost all of the pellets sticking together, whereas the SI was 99.1% after the 20th drop. The SI of the pellets increased with increasing temperature, as can be seen in Fig. 5. This increase in magnitude of SI was minimal below 1323 K, but it increased rapidly beyond this temperature. At 1423 K, the SI was nearly 85%, and at 1523 K it was nearly 100%. It therefore seems likely that the move of pellets in the reactor will become very difficult when temperatures exceed 1423 K due to severe sticking. However, the reduction of pellets is very slow at temperatures below 1423 K [14], and so ICPs should ideally be treated prior to reduction production. A description of the proposed treatment is given in a subsequent section.
3.2. Influence of temperature on sticking It can be seen from Fig. 4 that the SP of the ICP reduced at 1223 K gradually decreased with an increase in the number of drops, reaching 0.0% after five drops. The SI value meanwhile was only 4.8% by the 20th drop. Fig. 4 shows a similar trend for the ICP reduced at 1323 K, although the SP was not reduced to 0.0% by the 20th drop and there was a less significant change in SP after the 4th drop. By the 20th Table 2 Chemical composition of coal. Composition (%)
Ash
Volatile
Sulfur
Fixed carbon
Coal
10.53
8.51
0.41
80.27
Fig. 2. Experimental procedure used for coating (RM: raw materials, CM: coating materials).
32
J. Li et al. / Powder Technology 262 (2014) 30–35
Fig. 5. Relationship between SI and temperature.
Fig. 3. (a) SP and (b) SI of IOPs and ICPs at 1523 K.
3.3. Influence of coating materials on sticking Three coating materials were selected for investigation: dolomite powder, limestone powder, and iron ore powder. The morphologies of these coating materials are shown in Fig. 6. As can be seen from the cross section of a pellet, the pellet consists of two layers: an interior layer consisting of the pellet itself and an outer shell layer of coating material. From Fig. 7 to Fig. 9, it can be seen that the sticking of pellets coated with iron ore or limestone powder was quite pronounced. The SP of
pellets coated with iron ore powder decreased very slowly with an increase in the number of drops. The SP remained at 91.0% and the SI at 97.2% by the 20th drop, at which point nearly all of the pellets were sticking together, as shown in Fig. 8(b). The SP of pellets coated in limestone powder also decreased with an increasing number of drops. When the number of drops was less than 13, the SP remained very high, and it decreased from the 13th drop onwards. The SP was 61.1% and the SI 90.8% after the 20th drop. Fig. 8(c) shows that most of the pellets were indeed still sticking together. The dolomite-coated pellets exhibited a very different behavior, with only very slight sticking observed. The SP decreased significantly as the number of drops increased, whereas the SP decreased to 37.6% after the 1st drop, to 16.7% after the 5th, and to 0% after the 11th drop. The SI was 12.9% at the 20th drop due to very few of the pellets sticking together, as shown in Fig. 8(d). Pellets without any coating are shown in Fig. 8(a) after the 20th drop, with both large and small agglomerations, thus producing an SI of 89.3%. Although the SI values of the samples shown in Fig. 8(a) and (c) are similar, the physical appearance of the two samples is very different, with a greater prevalence of large agglomerations shown in the uncoated pellets. Consequently, a similarity in SI between two different pellets does not inherently mean that they share the same sticking behavior. SP is also important, in which a dramatic decrease in SP indicates that large agglomerates have been dispersed into several small pieces, whereas a slight decrease indicates that small pieces have separated from a large agglomerate, which remains intact. 4. Discussion Different coating materials have a very different impact on the sticking behavior of ICPs. Influence on SI of different coating materials was shown in Fig. 9. Dolomite powder proved to be the best for inhibiting sticking, reducing the SI from 84.2% down to 9.5%. In contrast, the effect of iron ore and limestone powders proved to be unsatisfactory in the sense that they actually aggravate the problem of sticking rather than inhibiting it. The SI of the iron ore-coated pellets increased from 84.2% to 92.2%, whereas the use of limestone powder resulted in an increase from 84.2% to 85.7%. Herein, the mechanisms affecting this sticking behavior are discussed in detail. Considering the sticking mechanism of ICPs first, the fact that reduction can take place at high temperatures means that the following reactions can occur:
Fig. 4. SP of ICPs with different reduction temperatures.
Fe3 O4 ðsÞ þ CðgÞ ¼ 3FeOðsÞ þ COðgÞ
ð1Þ
FeO ðsÞ þ CðgÞ ¼ Fe þ COðgÞ
ð2Þ
J. Li et al. / Powder Technology 262 (2014) 30–35
33
Fig. 6. Morphologies of the different ICP coating materials: (a) dolomite powder, (b) limestone powder, and (c) iron ore powder.
Fig. 7. SP of different ICP coating materials at 1423 K.
Fig. 9. Influence on SI of different coating materials.
Fe3 O4 ðsÞ þ COðgÞ ¼ 3FeOðsÞ þ CO2 ðgÞ
ð3Þ
FeO ðsÞ þ COðgÞ ¼ Fe þ CO2 ðgÞ:
ð4Þ
Metallic iron and ferrous oxide are both formed throughout these reactions. From Fig. 10, it can be seen that a significant number of iron crystals are generated after 1 h at 1375 K. The white area shown in Fig. 10(a) depicts an iron crystal, which is the phase needed to ensure that ICPs have sufficient strength. Under load, plastic deformation of the ICPs occurs with increasing temperature, which increases the contact surface area between neighboring pellets and reduces the distance between them. Consequently, the likelihood of iron crystals growing between the pellet borders increases, ultimately resulting in crystals of different ICPs joining together. In addition, FeO forms and grows between the pellet boundaries during reduction. This means that the sticking mechanism is the result of intergrowth between pellet boundaries of both iron and ferrous oxide crystals. Inhibiting sticking therefore requires hindering the crystal intergrowth between pellets, but their
intergrowth cannot be inhibited by decreasing the rate of reduction. This makes the coating of materials on the outer surface of the pellets the best way to prevent crystal intergrowth. When using iron ore powder as a coating material, reactions (3) and (4) will occur in the coating at high temperatures due to the escape of CO from the internal pellet. Iron and ferrous oxide crystals therefore appear in the coating material and grow together between neighboring pellets. Since CO is only provided from the internal pellet, the percentage of CO gradually decreases over time. When this value falls below 75%, ferrous oxide is predominantly formed, as shown in Fig. 11. Because the melting point of FeO is lower than that of iron, it more readily forms low-melting-point mixtures and compounds with other elements. The SI of ICPs coated in iron oxide therefore increases, and the resulting sticking problem is greater than that with uncoated ICPs. Limestone and dolomite both decompose during the reduction process [15], with limestone starting to decompose at 1073 K. Dolomite is a double carbonate composed of CaCO3 and MgCO3, and the activity of dolomite is lower than that of each individual component. The higher stability of CaCO3 means that when dolomite is heated, MgCO3
Fig. 8. Sticking behavior of ICP with different coating materials after the 20th drop at 1423 K: (a) no coating, SI = 89.3%, (b) iron ore powder coating, SI = 97.2%, (c) limestone powder coating, SI = 90.8%, and (d) dolomite powder coating, SI = 16.7%.
34
J. Li et al. / Powder Technology 262 (2014) 30–35
Fig. 10. SEM image and energy spectrum of pellets reduced at 1375 K for 1 h: (a) SEM, (b) EDS of point 1, and (c) EDS of point 4.
decomposes first. Decomposition of dolomite is therefore divided into two phases. In stage one, it decomposes into CaCO3, MgO, and CO2, with a decomposition temperature of 773 K and a boiling temperature between 1093 and 1123 K. In stage two, the dolomite eventually breaks down into CaO, MgO, and CO2. The decomposition reaction can therefore be written as follows: CaCO3 MgCO3 ¼ CaCO3 þ MgO þ CO2
ð5Þ
CaCO3 ¼ CaO þ CO2 :
ð6Þ
These two reactions are both endothermic, and so they can reduce the temperature of the coating layer during heating. This in turn should decrease the SI of ICPs using these two coating materials, but only dolomite decreases the SI, whereas limestone increases it. The difference between these materials is the presence of magnesium compounds in dolomite, which is the reason for the differences observed in the SI. The main components are iron oxide and calcium oxide in the boundary of pellets coating limestone. Iron oxide is reduced gradually in the heating process, but either Fe2O3 or FeO will be formed liquid compounds with CaO, such as CaO·Fe2O3. From Fig. 12, it can be obtained that liquid compounds are formed at 1489 K for the binary phases
of CaO–Fe2O3, and they are formed at 1478 K for the binary phases of CaO–FeO. When Fe is produced in the pellets, liquid compounds are still formed at 1413 K for CaO–FeO–Fe phases. Otherwise, both 2FeO·SiO2–2CaO·SiO2 and 2CaO·SiO2–2FeO, whose respective melting points are 1423 and 1523 K, are formed from the CaO, FeO, and SiO2 in the pellets. It was also found by Lingyun Yi and co-workers [13] in the study of sticking of iron ore pellets. Although these liquid compounds to be present only in small quantities, they are nonetheless sufficient to cause sticking to occur between pellets; this is why the limestone powder is ineffective at inhibiting sticking. Nevertheless, liquid compounds are formed hardly for pellets coating dolomite. First, the temperatures of liquid compounds formed by MgO and Fe2O3 (or FeO) are very high (Fig. 13). They are 1976, 1583 and 1837 K for MgO–Fe2O3, MgO–FeO and MgO–FeO–Fe phases respectively. Second, the rate of iron oxide decreases greatly in the iron oxide– calcium oxide phases, because iron oxide forms high-melting-point substances with magnesium oxide, such as 2MgO·FeO. On the other hand, the high-melting-point substances can effectively prevent the intergrowth of iron and ferrous oxide crystals. So dolomite can inhibit the sticking of pellets, but limestone cannot. Moreover, if magnesite were to be used, the inhibiting effect on sticking should be even greater and the quality ratio could be reduced due to its higher MgO content. (See Fig. 13.) 5. Conclusion
Fig. 11. Equilibrium diagram of Fe–O–CO.
(1) The sticking behavior of ICPs was more serious than that of IOPs: the SI of IOPs at 1523 K of 26.5% was much lower than that of ICPs of 99.7%. The temperature proved to be an important factor; the sticking behavior became more pronounced with increasing temperature: specifically, the SI at 1373 K was 89.3% but that at 1223 K was only 4.8%. (2) Different coating materials proved to have different effects on sticking, with only the dolomite powder effectively solving the problem by reducing the SI from about 90% to about 10%. Limestone and iron ore powders had no beneficial effect. (3) The sticking mechanism is the result of intergrowth between pellet boundaries of iron and ferrous oxide crystals; this intergrowth is effectively prevented by high-melting-point substances formed from MgO and FexOy. Although the compounds formed by CaO, FeO, and SiO2 can also prevent intergrowth, but their relatively low melting points means that a liquid phase is relatively easy to generate.
J. Li et al. / Powder Technology 262 (2014) 30–35
35
Fig. 12. Phase diagrams of CaO–Fe2O3, CaO–FeO and CaO–FeO–Fe.
Acknowledgments This work was financially supported by the National Natural Science Foundation of China (Grant No. 51274003).
Fig. 13. Phase diagrams of MgO–Fe2O3, MgO–FeO and MgO–FeO–Fe.
References [1] B.B. Agrawal, K.K. Prasad, S.B. Sarkar, H.S. Ray, Cold bonded ore–coal composite pellets for sponge ironmaking — part 1 — laboratory scale development, Ironmak. Steelmak. 27 (2000) 421–425. [2] J. Pargeter, R. Hanewald, D. Dombrowski, Operating experience at INMETCO and application of the process to the production of DRI, Conserv. Recycl. 8 (1985) 363–375. [3] L.H. Zhou, F.H. Zeng, Reduction mechanisms of vanadium–titanomagnetite–noncoking coal mixed pellet, Ironmak. Steelmak. 38 (2011) 59–64. [4] B.B. Agrawal, K.K. Prasad, S.B. Sarkar, H.S. Ray, Cold bonded ore–coal composite pellets for sponge ironmaking — part 2 — plant trials in rotary kiln, Ironmak. Steelmak. 28 (2001) 23–26. [5] S. Halder, R.J. Fruehan, Reduction of iron–oxide–carbon composites: part I. Estimation of the rate constants, Metall. Mater. Trans. B. Process. Metall. Mater. Process. Sci. 39 (2008) 784–795. [6] S. Halder, R.J. Fruehan, Reduction of iron–oxide–carbon composites: part III. Shrinkage of composite pellets during reduction, Metall. Mater. Trans. B. Process. Metall. Mater. Process. Sci. 39 (2008) 809–817. [7] R.J. Fruehan Halder, Reduction of iron–oxide–carbon composites: part II. Rates of reduction of composite pellets in a rotary hearth furnace simulator, Metall. Mater. Trans. B. Process. Metall. Mater. Process. Sci. 39 (2008) 796–808. [8] S. Sun, W. Lu, Building of a mathematical model for the reduction of iron ore in ore/coal composites, ISIJ Int. 39 (1999) 130–138. [9] P. Wong, M. Kim, H. Kim, C. Choi, Sticking behaviour in direct reduction of iron ore, Ironmak. Steelmak. 26 (1999) 53–57. [10] A. Basdağ, A.I. Arol, Coating of iron oxide pellets for direct reduction, Scand. J. Metall. 31 (2002) 229–233. [11] H.A. Kortmann, V.J. Ritz, B. deMunnik, Iron, S.O.C. Steel, Improvement of DR-pellets related to sticking and process by application of specifically adjusted coating minerals, Iron & Steel Soc Aime, Warrendale, 1996. [12] J. Sterneland, P.G. Jönsson, The use of coated pellets in optimising the blast furnace operation, ISIJ Int. 43 (2003) 26–35. [13] L. Yi, Z. Huang, T. Jiang, Sticking of iron ore pellets during reduction with hydrogen and carbon monoxide mixtures: behavior and mechanism, Powder Technol. 235 (2013) 1001–1007. [14] R. Wei, J. Li, G. Tang, D. Cang, Strength and consolidation mechanism of ironore and coal pellets, Ironmak. Steelmak. (2014), http://dx.doi.org/10.1179/1743281213Y. 0000000147 in press. [15] M. Gan, X. Fan, X. Chen, L. Yuan, T. Jiang, Z. Zhao, Application of Ca, Mg-additives in oxidized pellets, J. Cent. South Univ. 41 (2010) 1645–1651.
Contents lists available at ScienceDirect
Powder Technology journal homepage: www.elsevier.com/locate/powtec
Sticking behavior of iron ore–coal pellets and its inhibition Jiaxin Li a,c, Rufei Wei b,d,⁎, Hongming Long a, Ping Wang a, Daqiang Cang b,d a
School of Metallurgy and Resource, Anhui University of Technology, Ma'anshan 243002, China School of Metallurgical and Ecological Engineering, University of Science and Technology Beijing, Beijing 100083, China c Key Laboratory of Metallurgical Emission Reduction & Resources Recycling of Ministry of Education, Anhui University of Technology, Ma'anshan 243002, China d State Key Laboratory of Advanced Metallurgy, University of Science and Technology Beijing, Beijing 100083, China b
a r t i c l e
i n f o
Article history: Received 10 November 2013 Received in revised form 10 April 2014 Accepted 17 April 2014 Available online 26 April 2014 Keywords: Iron ore–coal pellets Sticking index Coating Reduction
a b s t r a c t Most previous studies on iron ore–coal pellets (ICPs) have focused on their reduction, but studies on their sticking behavior have not been reported. The sticking behavior of iron ore–coal pellets (ICPs) at high temperatures under a load of 0.1 MPa was studied. Temperature was determined to be an important factor that affects the sticking behavior: the sticking increased with increasing temperature; the sticking index was 89.28% at 1373 K but only 4.78% at 1223 K. The sticking mechanism results from the intergrowth of metallic iron and ferrous oxide crystals between pellet boundaries. High-melting-point materials formed by the combination of MgO and FexOy, can effectively prevent this intergrowth of iron, and the addition of a dolomite powder coating reduces the sticking index from about 90% to about 10%. The addition of limestone and iron ore powder was not found to have any beneficial effect of sticking inhibition. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Ironmaking without the use of coke has attracted significant attention due to a shortage of metallurgical coke resources, with iron ore–coal pellets (ICPs) rapidly emerging as one of the more promising noncoking coal raw materials. ICPs have already been applied to rotary hearth furnaces (RHFs), which were originally developed as a means to treat the fines and waste oxides generated in steel plants [1,2]. Currently, research into ICPs has mostly been focused on reduction, specifically the reduction mechanism and reaction rate [3–8]. However, this neglects other behaviors that occur during the reduction process, such as softening and sticking. Sticking behavior has previously been identified and studied with regard to direct-reduced iron (DRI), which is produced by iron oxide pellets (IOPs) and natural gas [9–13]. Studies performed by Lingyun Yi and co-workers indicated that sticking of IOPs was caused by fibrous iron and fresh iron with high activity. Sticking behavior relieved with the addition of H2 in reducing gas for the porous iron precipitation on the interface [13]. But ICP is different from IOP for it has high coal content and much lower H2 in its reducing gas. However, similar studies into the sticking behavior of ICPs have not been reported, even though this is of great significance to the optimization of current production techniques and reactors, as well as to the development of new reactor designs. The sticking behavior of ICPs was therefore studied through a series of reduction experiments under load. It was found that this phenomenon ⁎ Corresponding author at: School of Metallurgical and Ecological Engineering, University of Science and Technology Beijing, Beijing 100083, China. E-mail address: [email protected] (R. Wei).
http://dx.doi.org/10.1016/j.powtec.2014.04.059 0032-5910/© 2014 Elsevier B.V. All rights reserved.
does exist and that it can be inhibited by coating the ICPs. The mechanisms of sticking and its inhibition are discussed in detail, along with SEM analysis.
2. Experimental 2.1. Materials All raw materials and coatings used in this experiment were obtained from Anhui province in China. Their chemical compositions are listed in Table 1 and Table 2, respectively.
2.2. Apparatus All experiments were conducted in the device shown in Fig. 1. A SiC resistance furnace with a maximum working temperature of 1673 ± 2 K was used for heating. For each reduction experiment, a 500 g ICP sample was placed into the furnace at 473 K, with a total gas flow rate of 5 L/min (4.5 L/min N2, 0.5 L/min CO2), and the sample load was a pressure of 0.1 MPa provided by high-pressure Argon. The samples were then heated at a rate of 6 K/min and held at specified temperatures for 30 min. ICPs' reduction temperatures in previous studies were mostly set between 1173 K and 1473 K generally [15]. The reduction of ICPs is very slow before 1423 K but it becomes faster after 1423 K. So the specified temperatures in this work were 1223, 1323, 1423 and 1523 K. Finally, the reduced pellets were cooled to room temperature in a nitrogen atmosphere.
J. Li et al. / Powder Technology 262 (2014) 30–35
31
Table 1 Chemical composition of iron ore, dolomite and limestone. Composition (%)
TFe
SiO2
Al2O3
FeO
CaO
MgO
S
P
Iron ore Dolomite Limestone
68.11 – –
2.67 1.19 1.15
1.02 – –
27.16 – –
0.11 30.77 53.78
0.26 20.60 0.79
0.28 0.01 0.05
0.03 – –
2.3. Methods Clusters consisting of two or more pellets were each dropped 20 times from a height of 1 m onto a steel plate surface. The percentage of clusters remaining after each drop was calculated as the sticking percentage (SP). The number of drops was plotted against this SP value, with the area under the curve calculated as the sticking index (SI). This SI value is equal to zero when no clusters of two or more pellets remain and 100 when all pellets remain clustered and do not disintegrate during the drop tests. A schematic of the method used for coating the ICPs is shown in Fig. 2. During this coating procedure, raw materials were added to the right pan pelletizers and coatings were added to the left pan pelletizers, respectively. When the pellets attained sufficient size, they were moved out of the right pan and onto the conveyor belt, which transferred them to the left pan. In this left pan, the coating materials were bonded to the outer surface of the pellets to create a coating layer. The coating materials were added according to the quality ratio, which was 7.5% of the raw materials used in this experiment. 3. Results 3.1. Sticking behavior of ICPs and IOPs Initially, the sticking behavior of ICPs and IOPs was studied. The specified temperature for reduction was 1523 K, with the IOPs heated in air rather than in the N2/CO2 atmosphere used with the ICPs. The sticking behavior of ICPs and IOPs is illustrated in Fig. 3, which shows a difference in their respective SP and SI values at 1523 K. Initially, the SP of the IOPs gradually decreased with an increasing number of drops, reaching 7.6% by the 14th drop when almost all of the pellets became separated. However, with the ICPs there was only a very small decrease in SP as the number of drops increased, with the SP still at 99.5% after the 20th drop when almost all of the pellets were sticking. The sticking degree of the pellets is indicated by the SI value at the 20th drop [9,12,13]. For the IOPs, this value was 26.5%, which is considerably lower than the 99.7% obtained with the ICPs; thus, the sticking behavior of the ICPs is far more serious. Consequently, ICPs are far more difficult to smelt using conventional IOP producing reactors, such as grate-rotary kilns or shaft furnaces. It is therefore important to study the sticking behavior of ICPs in order to find ways in which sticking can be inhibited or accommodated by new reactor designs.
Fig. 1. Schematic diagram of the experimental apparatus used.
drop, the SP was still 5.2%, with many pellets still sticking, whereas the SI was 18.0%. The SPs also decreased with an increasing number of drops with the 1423 K and 1523 K samples, but the SPs were notably different from the 1223 K and 1323 K samples. At 1423 K, the SP was maintained at approximately 99.1% up to the 9th drop, with almost all of the pellets sticking together. The SP decreased rapidly by the 11th drop, with some separation of the pellets occurring at this time. Once the 15th drop was passed, the change in SP began to decrease, although it remained at 66.4% by the 20th drop. Similarly, the SI was 89.3% after the 20th drop. With the 1523 K sample, the SP was very high at nearly 100% after every drop due to almost all of the pellets sticking together, whereas the SI was 99.1% after the 20th drop. The SI of the pellets increased with increasing temperature, as can be seen in Fig. 5. This increase in magnitude of SI was minimal below 1323 K, but it increased rapidly beyond this temperature. At 1423 K, the SI was nearly 85%, and at 1523 K it was nearly 100%. It therefore seems likely that the move of pellets in the reactor will become very difficult when temperatures exceed 1423 K due to severe sticking. However, the reduction of pellets is very slow at temperatures below 1423 K [14], and so ICPs should ideally be treated prior to reduction production. A description of the proposed treatment is given in a subsequent section.
3.2. Influence of temperature on sticking It can be seen from Fig. 4 that the SP of the ICP reduced at 1223 K gradually decreased with an increase in the number of drops, reaching 0.0% after five drops. The SI value meanwhile was only 4.8% by the 20th drop. Fig. 4 shows a similar trend for the ICP reduced at 1323 K, although the SP was not reduced to 0.0% by the 20th drop and there was a less significant change in SP after the 4th drop. By the 20th Table 2 Chemical composition of coal. Composition (%)
Ash
Volatile
Sulfur
Fixed carbon
Coal
10.53
8.51
0.41
80.27
Fig. 2. Experimental procedure used for coating (RM: raw materials, CM: coating materials).
32
J. Li et al. / Powder Technology 262 (2014) 30–35
Fig. 5. Relationship between SI and temperature.
Fig. 3. (a) SP and (b) SI of IOPs and ICPs at 1523 K.
3.3. Influence of coating materials on sticking Three coating materials were selected for investigation: dolomite powder, limestone powder, and iron ore powder. The morphologies of these coating materials are shown in Fig. 6. As can be seen from the cross section of a pellet, the pellet consists of two layers: an interior layer consisting of the pellet itself and an outer shell layer of coating material. From Fig. 7 to Fig. 9, it can be seen that the sticking of pellets coated with iron ore or limestone powder was quite pronounced. The SP of
pellets coated with iron ore powder decreased very slowly with an increase in the number of drops. The SP remained at 91.0% and the SI at 97.2% by the 20th drop, at which point nearly all of the pellets were sticking together, as shown in Fig. 8(b). The SP of pellets coated in limestone powder also decreased with an increasing number of drops. When the number of drops was less than 13, the SP remained very high, and it decreased from the 13th drop onwards. The SP was 61.1% and the SI 90.8% after the 20th drop. Fig. 8(c) shows that most of the pellets were indeed still sticking together. The dolomite-coated pellets exhibited a very different behavior, with only very slight sticking observed. The SP decreased significantly as the number of drops increased, whereas the SP decreased to 37.6% after the 1st drop, to 16.7% after the 5th, and to 0% after the 11th drop. The SI was 12.9% at the 20th drop due to very few of the pellets sticking together, as shown in Fig. 8(d). Pellets without any coating are shown in Fig. 8(a) after the 20th drop, with both large and small agglomerations, thus producing an SI of 89.3%. Although the SI values of the samples shown in Fig. 8(a) and (c) are similar, the physical appearance of the two samples is very different, with a greater prevalence of large agglomerations shown in the uncoated pellets. Consequently, a similarity in SI between two different pellets does not inherently mean that they share the same sticking behavior. SP is also important, in which a dramatic decrease in SP indicates that large agglomerates have been dispersed into several small pieces, whereas a slight decrease indicates that small pieces have separated from a large agglomerate, which remains intact. 4. Discussion Different coating materials have a very different impact on the sticking behavior of ICPs. Influence on SI of different coating materials was shown in Fig. 9. Dolomite powder proved to be the best for inhibiting sticking, reducing the SI from 84.2% down to 9.5%. In contrast, the effect of iron ore and limestone powders proved to be unsatisfactory in the sense that they actually aggravate the problem of sticking rather than inhibiting it. The SI of the iron ore-coated pellets increased from 84.2% to 92.2%, whereas the use of limestone powder resulted in an increase from 84.2% to 85.7%. Herein, the mechanisms affecting this sticking behavior are discussed in detail. Considering the sticking mechanism of ICPs first, the fact that reduction can take place at high temperatures means that the following reactions can occur:
Fig. 4. SP of ICPs with different reduction temperatures.
Fe3 O4 ðsÞ þ CðgÞ ¼ 3FeOðsÞ þ COðgÞ
ð1Þ
FeO ðsÞ þ CðgÞ ¼ Fe þ COðgÞ
ð2Þ
J. Li et al. / Powder Technology 262 (2014) 30–35
33
Fig. 6. Morphologies of the different ICP coating materials: (a) dolomite powder, (b) limestone powder, and (c) iron ore powder.
Fig. 7. SP of different ICP coating materials at 1423 K.
Fig. 9. Influence on SI of different coating materials.
Fe3 O4 ðsÞ þ COðgÞ ¼ 3FeOðsÞ þ CO2 ðgÞ
ð3Þ
FeO ðsÞ þ COðgÞ ¼ Fe þ CO2 ðgÞ:
ð4Þ
Metallic iron and ferrous oxide are both formed throughout these reactions. From Fig. 10, it can be seen that a significant number of iron crystals are generated after 1 h at 1375 K. The white area shown in Fig. 10(a) depicts an iron crystal, which is the phase needed to ensure that ICPs have sufficient strength. Under load, plastic deformation of the ICPs occurs with increasing temperature, which increases the contact surface area between neighboring pellets and reduces the distance between them. Consequently, the likelihood of iron crystals growing between the pellet borders increases, ultimately resulting in crystals of different ICPs joining together. In addition, FeO forms and grows between the pellet boundaries during reduction. This means that the sticking mechanism is the result of intergrowth between pellet boundaries of both iron and ferrous oxide crystals. Inhibiting sticking therefore requires hindering the crystal intergrowth between pellets, but their
intergrowth cannot be inhibited by decreasing the rate of reduction. This makes the coating of materials on the outer surface of the pellets the best way to prevent crystal intergrowth. When using iron ore powder as a coating material, reactions (3) and (4) will occur in the coating at high temperatures due to the escape of CO from the internal pellet. Iron and ferrous oxide crystals therefore appear in the coating material and grow together between neighboring pellets. Since CO is only provided from the internal pellet, the percentage of CO gradually decreases over time. When this value falls below 75%, ferrous oxide is predominantly formed, as shown in Fig. 11. Because the melting point of FeO is lower than that of iron, it more readily forms low-melting-point mixtures and compounds with other elements. The SI of ICPs coated in iron oxide therefore increases, and the resulting sticking problem is greater than that with uncoated ICPs. Limestone and dolomite both decompose during the reduction process [15], with limestone starting to decompose at 1073 K. Dolomite is a double carbonate composed of CaCO3 and MgCO3, and the activity of dolomite is lower than that of each individual component. The higher stability of CaCO3 means that when dolomite is heated, MgCO3
Fig. 8. Sticking behavior of ICP with different coating materials after the 20th drop at 1423 K: (a) no coating, SI = 89.3%, (b) iron ore powder coating, SI = 97.2%, (c) limestone powder coating, SI = 90.8%, and (d) dolomite powder coating, SI = 16.7%.
34
J. Li et al. / Powder Technology 262 (2014) 30–35
Fig. 10. SEM image and energy spectrum of pellets reduced at 1375 K for 1 h: (a) SEM, (b) EDS of point 1, and (c) EDS of point 4.
decomposes first. Decomposition of dolomite is therefore divided into two phases. In stage one, it decomposes into CaCO3, MgO, and CO2, with a decomposition temperature of 773 K and a boiling temperature between 1093 and 1123 K. In stage two, the dolomite eventually breaks down into CaO, MgO, and CO2. The decomposition reaction can therefore be written as follows: CaCO3 MgCO3 ¼ CaCO3 þ MgO þ CO2
ð5Þ
CaCO3 ¼ CaO þ CO2 :
ð6Þ
These two reactions are both endothermic, and so they can reduce the temperature of the coating layer during heating. This in turn should decrease the SI of ICPs using these two coating materials, but only dolomite decreases the SI, whereas limestone increases it. The difference between these materials is the presence of magnesium compounds in dolomite, which is the reason for the differences observed in the SI. The main components are iron oxide and calcium oxide in the boundary of pellets coating limestone. Iron oxide is reduced gradually in the heating process, but either Fe2O3 or FeO will be formed liquid compounds with CaO, such as CaO·Fe2O3. From Fig. 12, it can be obtained that liquid compounds are formed at 1489 K for the binary phases
of CaO–Fe2O3, and they are formed at 1478 K for the binary phases of CaO–FeO. When Fe is produced in the pellets, liquid compounds are still formed at 1413 K for CaO–FeO–Fe phases. Otherwise, both 2FeO·SiO2–2CaO·SiO2 and 2CaO·SiO2–2FeO, whose respective melting points are 1423 and 1523 K, are formed from the CaO, FeO, and SiO2 in the pellets. It was also found by Lingyun Yi and co-workers [13] in the study of sticking of iron ore pellets. Although these liquid compounds to be present only in small quantities, they are nonetheless sufficient to cause sticking to occur between pellets; this is why the limestone powder is ineffective at inhibiting sticking. Nevertheless, liquid compounds are formed hardly for pellets coating dolomite. First, the temperatures of liquid compounds formed by MgO and Fe2O3 (or FeO) are very high (Fig. 13). They are 1976, 1583 and 1837 K for MgO–Fe2O3, MgO–FeO and MgO–FeO–Fe phases respectively. Second, the rate of iron oxide decreases greatly in the iron oxide– calcium oxide phases, because iron oxide forms high-melting-point substances with magnesium oxide, such as 2MgO·FeO. On the other hand, the high-melting-point substances can effectively prevent the intergrowth of iron and ferrous oxide crystals. So dolomite can inhibit the sticking of pellets, but limestone cannot. Moreover, if magnesite were to be used, the inhibiting effect on sticking should be even greater and the quality ratio could be reduced due to its higher MgO content. (See Fig. 13.) 5. Conclusion
Fig. 11. Equilibrium diagram of Fe–O–CO.
(1) The sticking behavior of ICPs was more serious than that of IOPs: the SI of IOPs at 1523 K of 26.5% was much lower than that of ICPs of 99.7%. The temperature proved to be an important factor; the sticking behavior became more pronounced with increasing temperature: specifically, the SI at 1373 K was 89.3% but that at 1223 K was only 4.8%. (2) Different coating materials proved to have different effects on sticking, with only the dolomite powder effectively solving the problem by reducing the SI from about 90% to about 10%. Limestone and iron ore powders had no beneficial effect. (3) The sticking mechanism is the result of intergrowth between pellet boundaries of iron and ferrous oxide crystals; this intergrowth is effectively prevented by high-melting-point substances formed from MgO and FexOy. Although the compounds formed by CaO, FeO, and SiO2 can also prevent intergrowth, but their relatively low melting points means that a liquid phase is relatively easy to generate.
J. Li et al. / Powder Technology 262 (2014) 30–35
35
Fig. 12. Phase diagrams of CaO–Fe2O3, CaO–FeO and CaO–FeO–Fe.
Acknowledgments This work was financially supported by the National Natural Science Foundation of China (Grant No. 51274003).
Fig. 13. Phase diagrams of MgO–Fe2O3, MgO–FeO and MgO–FeO–Fe.
References [1] B.B. Agrawal, K.K. Prasad, S.B. Sarkar, H.S. Ray, Cold bonded ore–coal composite pellets for sponge ironmaking — part 1 — laboratory scale development, Ironmak. Steelmak. 27 (2000) 421–425. [2] J. Pargeter, R. Hanewald, D. Dombrowski, Operating experience at INMETCO and application of the process to the production of DRI, Conserv. Recycl. 8 (1985) 363–375. [3] L.H. Zhou, F.H. Zeng, Reduction mechanisms of vanadium–titanomagnetite–noncoking coal mixed pellet, Ironmak. Steelmak. 38 (2011) 59–64. [4] B.B. Agrawal, K.K. Prasad, S.B. Sarkar, H.S. Ray, Cold bonded ore–coal composite pellets for sponge ironmaking — part 2 — plant trials in rotary kiln, Ironmak. Steelmak. 28 (2001) 23–26. [5] S. Halder, R.J. Fruehan, Reduction of iron–oxide–carbon composites: part I. Estimation of the rate constants, Metall. Mater. Trans. B. Process. Metall. Mater. Process. Sci. 39 (2008) 784–795. [6] S. Halder, R.J. Fruehan, Reduction of iron–oxide–carbon composites: part III. Shrinkage of composite pellets during reduction, Metall. Mater. Trans. B. Process. Metall. Mater. Process. Sci. 39 (2008) 809–817. [7] R.J. Fruehan Halder, Reduction of iron–oxide–carbon composites: part II. Rates of reduction of composite pellets in a rotary hearth furnace simulator, Metall. Mater. Trans. B. Process. Metall. Mater. Process. Sci. 39 (2008) 796–808. [8] S. Sun, W. Lu, Building of a mathematical model for the reduction of iron ore in ore/coal composites, ISIJ Int. 39 (1999) 130–138. [9] P. Wong, M. Kim, H. Kim, C. Choi, Sticking behaviour in direct reduction of iron ore, Ironmak. Steelmak. 26 (1999) 53–57. [10] A. Basdağ, A.I. Arol, Coating of iron oxide pellets for direct reduction, Scand. J. Metall. 31 (2002) 229–233. [11] H.A. Kortmann, V.J. Ritz, B. deMunnik, Iron, S.O.C. Steel, Improvement of DR-pellets related to sticking and process by application of specifically adjusted coating minerals, Iron & Steel Soc Aime, Warrendale, 1996. [12] J. Sterneland, P.G. Jönsson, The use of coated pellets in optimising the blast furnace operation, ISIJ Int. 43 (2003) 26–35. [13] L. Yi, Z. Huang, T. Jiang, Sticking of iron ore pellets during reduction with hydrogen and carbon monoxide mixtures: behavior and mechanism, Powder Technol. 235 (2013) 1001–1007. [14] R. Wei, J. Li, G. Tang, D. Cang, Strength and consolidation mechanism of ironore and coal pellets, Ironmak. Steelmak. (2014), http://dx.doi.org/10.1179/1743281213Y. 0000000147 in press. [15] M. Gan, X. Fan, X. Chen, L. Yuan, T. Jiang, Z. Zhao, Application of Ca, Mg-additives in oxidized pellets, J. Cent. South Univ. 41 (2010) 1645–1651.