ISIJ International
Online ISSN : 1347-5460
Print ISSN : 0915-1559
ISSN-L : 0915-1559
Carburization Degree of the Iron Nugget Produced by High Al2O3 Iron Ore
Guang WangJingsong WangQingguo Xue
Author information

2017 Volume 57 Issue 3 Pages 590-592


The reduction and melting separation experiment of iron ore containing Al2O3 was performed. The results showed that Al2O3 obviously affected the melting separation behaviors and the carbon content of the iron nugget. When the Al2O3 content in iron ore was 4.59 wt%, the carbon content in the iron nugget was as low as 0.80 wt% and was much lower than the iron ores with lower Al2O3 content, and increased with the increasing of slag basicity. The melting separation and carburiz ation mechanism of high Al2O3 iron ore/coal composite pellet was deduced.

1. Introduction

Due to the limited reserves and increasing depletion of high-grade iron ore resource, the Al2O3 content of iron ore is gradually increasing, which negatively affects the blast furnace ironmaking proces.1) In direct reduced iron (DRI) processes, however, the slag composition or gas permeability is not so demanding compared with that of BF process. Consequently, some researchers believe DRI process, especially the iron nugget process, will be one of the possible choices to effectively use high Al2O3 iron ore resource.2,3,4) The iron nugget process is an environmentally friendlier, economical alternative to the traditional blast furnace process.5,6,7) It aims to produce pig iron nugget which has similar chemical and physical properties to blast furnace pig iron using fine iron ores and noncoking coals. In the iron nugget process, a series of reactions occur within about 10 minutes at 1350–1450°C due to the well enough mixing of iron oxide and carbon.8,9,10)

Carburization is one of the most important rate-determining stages in the ironmaking process. Much time was devoted to elucidating the carburization mechanism of the process.11,12,13,14,15,16,17,18) Carburization degree of the reduced iron during reduction is important for accelerating the ironmaking process and reducing energy consumption. Traditionally, the carbon content of the pig iron nugget is 2.5–3.0 wt% and is 4.5–5.4 wt% for the blast furnace iron. A relatively lower carbon content of pig iron will shorten the tap-to-tap time and the CO2 emission. In this paper, an interesting phenomenon, which the carbon content in the iron nugget can be determined by the composition of the iron ore, has been found by the authors and is presented.

2. Experimental

Three iron ores were utilized in the experiment. The chemical compositions of the ores and anthracite coal were shown in Table 1. Ore3 is magnetite ore, and Ore1 and Ore2 are the mixture of hematite and magnetite. The main gangue constituents of Ore3 is SiO2, and Ore1 and Ore2 are SiO2 and Al2O3. The Al2O3 content of Ore1 is 4.59 wt%, which is higher than Ore2. The Al2O3 content of Ore3 is only a little. The particle sizes the ores are all below 0.5 mm. The compositions of the pellets were adjusted by CaCO3 of analytical reagent grade. Pulverized anthracite coal was used as the reducing agent and the fineness is 100% passing 0.5 mm.

Table 1. Chemical composition of the raw materials/wt%.
Anthracite coal81.406.4011.100.34

The iron ore, CaCO3 and pulverized coal were completely mixed together with a mole ratio of C/O equaling to 1.2. The basicities of the slag compositions of the pellets ranged from 0.6 to 1.2 with the addition of chemical CaCO3. The pelletizing process was performed through a manual ball press under the pressure of 20 MPa. The green pellet presented in column shape with the size of Φ20 mm×20 mm. Reduction and melting process was performed in a closed MoSi2 resistance furnace. In each experiment, the dry green composite pellets were put into a graphite crucible with the inner size of 180 mm (length)×130 mm (width)×20 mm (depth) and then heated at 1400°C and 1450°C in the furnace. The depth of the carbon material bed in the crucible was about 10 mm and it can provide enough carbon for the carburization of metallic iron. The crucible had already been heated to the target temperature before the experiment. Two pellets were charged in the crucible in one reduction test. In addition, purge gases were not utilized and experiments were run at atmospheric condition. Once the reduction finished, the samples were quickly taken out and the carbon content of iron nuggets were analyzed.

The chemical compositions and melting temperature of two pellets (Ore1 pellets of natural and 1.0 basicity) slag as the progress of reduction were listed in Table 2. The slag compositions were calculated based on the chemical composition of the Ore1 and were prepared by the chemicals of CaO, SiO2, Al2O3 and FeO of analytical reagent (AR) grade. The melting temperature of slag was tested by DSC method and the temperature of endothermic peak of the DSC curve was set as the melting temperature.

Table 2. Melting property evolution of slag of the reduced pellets whose basicities were 0.15 and 1.0.

3. Results and Discussion

The carbon composite pellets of Ore2 and Ore3 can similarly separate well when reduced at 1400°C for 15 minutes. However, the pellet of Ore1 cannot melt separation completely. The pellet of Ore1 separated when the reduction temperature increasing to 1450°C. The melting state of the iron nugget separated from Ore1 is worst compared with the other two ores and the separated slag adheres closely on the surface of the nugget. The carbon content of the iron nuggets separated from the three pellets with different basicities are listed in Fig. 1. The Ore1 pellet is reduced at 1450°C, and the Ore2 and Ore3 pellets are reduced at 1400°C. The carbon contents of the Ore2 and Ore3 iron nuggets gradually decrease with the increasing of basicity in similar trend. However, the carbon content of the Ore1 iron nugget gradually increases with the increasing of basicity, which is much lower than those of Ore2 and Ore3 although the reduction temperature of Ore1 is higher. The carbon content of the iron nugget separated from Ore1 pellet with natural basicity is about 0.80 wt% and is the lowest. The experimental results obviously indicate that different Al2O3 content of the iron ores is main reason for the phenomena. The Al2O3 content in the iron ore will influence the melting and separation of the carbon composite pellets, especially when the Al2O3 content is higher than a certain value. Furthermore, slag basicity is also a significant influencing factor and the effect is different when the Al2O3 content of the slag phase is different.

Fig. 1.

Variation of carbon content of iron nuggets of three kinds of ores.

Ohno17) et al. reported that the carburization could occur through the slag and carbon content in the iron nugget could be influenced by the slag composition with graphite as the reducing and carburizing agent. Their results of relation between effective carbon mixing ratio and carbon content in the initial melting down iron nugget are given in Fig. 2(a). It can be seen that the iron nugget using slag II was higher than those using slag I by +0.4 wt%. Obviously, the basicity of slag II (R=0.92, R: basicity) is higher than slag I (R=0.38). Therefore, basicity may be the main reason for the difference in the ability of carburization of the two high Al2O3 slags. Slag composition of this work and Ohno’s is depicted in CaO–SiO2–Al2O3 phase diagram, which is shown in Fig. 2(b). The level of Al2O3 content of Ohno’s slag is similar to the slag separated from Ore1 and Ohno’s research agrees well with the results obtained in this paper.

Fig. 2.

Carburization result of Ohno’s work and slag composition comparing with this work.

Therefore, the reduction and melting separation process of the Ore1 composite pellets, without addition of CaCO3 (R=0.15) and with basicity of R=1.0, are further investigated at 1400°C. The reduction rate increases due to the addition of CaCO3. The morphology of the pellets during reduction and microstructure of the surface layer of the iron nugget are shown in Fig. 3. The pellet of natural basicity forms liquid phase at the time about 4 min during reduction and the amount of liquid gradually increases. At the time of 8 min, the slag forms a big shell covering the surface of the metallic iron, which does not melt completely. However, the pellet with basicity of R=1.0 forms liquid phase when reduced for 6 min, which is later than the natural basicity pellet. The main reason for the above phenomenon is the discrepancy in the slag composition and faster reduction rate. The iron oxide in the pellet changes in the sequence of Fe2O3→FeO→Fe during reduction. It can be seen from the data in Table 2 that the melting temperature of R=1.0 pellet slag gradually increases from 1130.8°C to 1382.0°C with the decreasing of FeO. However, the melting temperature of the R=0.15 pellet slag is relatively lower and changes little when the reduction degree is less than 0.9, and the liquidus temperature increases to higher than 1450°C if all the FeO is reduced into Fe. It is sure that the FeO in the formed slag of the pellet can’t be completely reduced into metallic iron, especially for the slag of the R=0.15 pellet due to the low basicity. The molten slag of the R=0.15 pellet has better wettability on the solid iron because of its lower melting temperature. Therefore, the molten slag will firstly occur on the pellet surface and form a big shell covering the surface of the metallic iron until the end of reduction of R=0.15 pellet. The SEM-EDS analysis shows that there is a layer of oxides on the surface of both of the two iron nuggets. For the iron nugget separated from the R=0.15 pellet, the layer is mainly composed of 43.81 wt% SiO2-34.72 wt% FeO-11.54 wt% Al2O3-9.92 wt% CaO, whose melting point is around 1100°C according to the data in Table 2. For the iron nugget separated from the R=1.0 pellet, the layer is mainly composed of iron oxide, which may formed due to the oxidation of metallic iron by the oxygen in the atmosphere during the melting separation and cooling process.

Fig. 3.

Morphology of the reduced pellets and microstructures of the surface layers of the iron nuggets reduced at 1400°C which basicities of the green pellets were 0.15 and 1.0.

It can be concluded that the low melting point slag will form at early stage during the reduction of R=0.15 pellet and will cover on the surface of the solid metallic iron until melting separation. The low melting point slag will keep fluidity due to containing a certain content of FeO and it will stop the contact between metallic iron and carbon particle and block the melting of reduced iron. Based on above knowledge, the melting separation and carburization mechanism of high Al2O3 iron ore/coal composite pellet without CaO addition is shown in Fig. 4. The low melting point high Al2O3 slag makes the carburization becoming more difficult and the effect will be weaker with increasing of melting point of the slag due to the increasing of basicity. On the other hand, the addition of CaCO3 will improve the reduction of pellet and the FeO content in slag phase will be less. The amount and fluidity of the slag will then decrease. It will be more difficult for the slag to cover the surface of the reduced pellet and the carburization will be better.

Fig. 4.

Illustration of melting separation and carburization mechanism of high Al2O3 iron ore/coal composite pellet.

4. Conclusions

The low melting point slag formed during the reduction of composite pellet of high Al2O3 iron ore (4.59 wt% Al2O3) will cover on the surface of solid metallic iron and stop the contact between metallic iron and carbon particle, which makes the carburization becoming more difficult. Addition of CaO can improve the carburization, and the carbon content increases with the increasing of its slag basicity.


The research is supported by the China Postdoctoral Science Foundation (2016M600919), Fundamental Research Funds for the Central Universities (FRF-TP-16-019A1) and National Natural Science Foundation of China (51274033, 51374024).

© 2017 by The Iron and Steel Institute of Japan