Regular Article
Carburization, Melting and Dripping of Iron through Coke Bed
2015 Volume 55 Issue 10 Pages 2056-2063
Details
2015 Volume 55 Issue 10 Pages 2056-2063
The carburizing, melting, and dripping behavior of iron through graphite and coke beds was investigated by an in-situ observation technique with increasing temperature from room temperature to 1773 K. An iron sample placed on the graphite bed was fully melted at 1490 K, whereas that on the coke bed was not melted at 1773 K. From the microscopic analysis of the sample after experiments, it was revealed that solid ash formed at the metal-coke interface prevented the carburization and melting of the iron sample. In order to examine the role of gangue and flux materials in sinter, iron samples mixed with slag powders at different weight fractions (15, 30, and 50 wt%) were prepared and similar experiments were carried out. The sample containing 15 wt% slag powders was not melted, but the samples containing 30 and 50 wt% slag powders were carburized and fully melted at 1773 K after holding the sample at that temperature for 342 and 37 s, respectively. It was considered that the liquid slag absorbed solid ash formed at the coke surface, yielding carburization and melting of the iron.
Coke bed in the blast furnace allows gas distribution through the furnace for heating the charged burden materials effectively. In the cohesive zone, slag and metal start to melt and flow down through the dripping zone to the hearth. As the melt formation becomes slower at the cohesive zone, gas permeability becomes worse, and the heat exchange efficiency decreases. Consequently, the formation of molten slag and molten iron should be enhanced for better energy efficiency of the blast furnace.
On the other hand, in order to decrease carbon dioxide gas emission from the blast furnace, hydrogen can be considered as an alternative reducing agent. As the hydrogen concentration increases in the blast furnace operation, the reduction of iron oxide can be accelerated from the solid state. Once iron oxide is fully reduced by hydrogen gas, the slag melting can be merely determined by the slag chemistry of gangue and flux materials. Consequently, the carburization of reduced iron by coke plays an important role in the melts formation.
Carburization and melting behavior of solid iron by solid carbon have been studied by many researchers. Murakami et al.1) investigated the reaction between solid iron and solid graphite at high temperatures using a confocal laser scanning microscopy (CLSM). Kim et al.2) reported that the carburization rate of solid iron could be increased by placing wustite near the iron due to the Marangoni flow. Ohno et al.3) showed that the initial carburization rate could be slightly increased by lowering the degree of crystallization. In our previous works, it was proved that the carbon diffusion in liquid iron is the rate-determining step.4) However, these studies were carried out with graphite as solid carbon.
Carburization of liquid iron by solid carbon has been studied by many researchers.5,6,7) Most of them used graphite as solid carbon, but several researchers investigated the carburization of liquid iron by coke. Gudenau et al. reported that the carburization of liquid iron by coke decreased with increasing the reaction time due to the formation and growth of the ash layer.5) Wu et al. asserted that the interfacial chemical reaction at the metal-coke interface would be the rate-determining step.6) Recently, Jang et al. reported that the rate-determining step was changed from the interfacial chemical reaction to liquid phase mass transfer with increasing temperature.7) In this study, the interfacial chemical reaction was the rate-determining step at 1723 K. With increasing temperature, the rate-determining step was changed to the mixed-control at 1823 K, and to the liquid phase mass transfer at 1923 K. They reported that the solid-to-liquid ratio of the ash content decreased with increasing temperature, yielding more bare carbon on the coke surface.
In the blast furnace ironmaking, flux-containing sinter is generally used in Korea and Japan. By adding flux in the sinter, the molten slag formation temperature can be controlled. Since the molten slag has a potential to absorb the solid ash formed on the coke surface,7) it is speculated that the carburization and melting rates of solid iron would increase. In this study, the carburization, melting, and dripping of solid iron placed on the carbonaceous beds were investigated by controlling the slag content in the solid iron.
Iron powders (particle size <150 μm; Wako Pure Chemical Industries Co., Ltd.) were used as the starting material. The chemical composition of the iron powders is shown in Table 1. A cylindrical iron rod (diameter: 20 mm and weight: 15 g) was prepared by applying a pressure of 125 MPa for 1 min. For the samples containing slag powders, slag content was adjusted to be 15, 30, and 50 wt%, which are respectively corresponding to 32, 53, and 77 vol% at 1773 K.8,9) The chemical composition of the slag (CAS slag) was 48wt%CaO-40wt%SiO2-12wt%Al2O3. It is calculated using the FactSage software that the liquid phase starts to be formed at 1547 K, and the slag is fully melted at 1632 K.10) In order to investigate the effect of the addition of FeO, 40.8wt%CaO-34wt%SiO2-10.2wt%Al2O3-15wt%FeO slag (CASF slag) was used. In the case of the CASF slag, the liquid phase starts to form at 1519 K, and the slag is fully melted at 1574 K. The CAS slag was obtained by melting a mixture of the regent grade oxide powder in a graphite crucible at 1773 K under an Ar atmosphere for 10 min and quenching the liquid slag in water. The CASF slag was prepared by the same method using an iron crucible. Graphite and coke granules (diameter: ~9 mm) were used for the carbonaceous bed materials. The chemical composition of the coke is shown in Table 2.
| Moisture | Fixed Carbon | Volatile Materials | Ash | ||||||
|---|---|---|---|---|---|---|---|---|---|
| SiO2 | CaO | Al2O3 | MgO | Fe2O3 | P2O5 | K2O | |||
| 0.22 | 86.36 | 1.76 | 6.36 | 0.32 | 3.19 | 0.10 | 0.07 | 0.01 | 0.01 |
Figure 1 shows the experimental sample assembly used in the in-situ observation of the carburization, melting, and dripping behavior of solid iron on the carbonaceous bed. A cylindrical sample was positioned on the carbonaceous bed, which consisted of three layers of graphite or coke granules, in a graphite crucible (inner diameter: 55 mm, height: 25 mm, 14 holes of 5 mm in diameter at the bottom). An alumina receiver was positioned just below the graphite crucible to receive the dripping liquid iron and slag.
Schematic illustration of the sample assembly used in the present experiments.
The sample assembly was positioned in the center of an alumina reaction tube (inner diameter: 60 mm, length: 1000 mm) in a horizontal electrical resistance furnace having MoSi2 heating elements. (Fig. 2) The alumina reaction tube was equipped with two water-cooled jackets with built-in quartz windows on both sides. The temperature was monitored using a B-type thermocouple contacting the surface of the alumina reaction tube. The temperature of the sample position was calibrated by placing another B-type thermocouple and measuring the temperature difference between two thermocouples before experiments. The temperature difference between two thermocouples was approximately 75 K. Argon gas (99.9999% purity) was used throughout the experiments. The carburization, melting, and dripping behavior of iron was observed using a digital camera (D5000, Nikon, 2144 × 1424 pixels) with a zoom lens (70–300 mm F4-5.6 APO DG MACRO for Nikon, Sigma CO., LTD.).
Schematic illustration of the experimental apparatus for the in-situ observations.
The sample assembly was placed in the center of the reaction tube, and the residual oxygen gas was removed using a vacuum pump. The reaction tube was then filled with a purified Ar gas. By flowing the Ar gas at a rate of 500 ml/min STP, the furnace was heated to 1273 K at a heating rate of 10 K/min, and then to 1773 K at a heating rate of 5 K/min. (Carbon diffusion in solid iron is much slower than in liquid. For example, carbon diffusion coefficient in liquid Fe–C alloy at 1523 K is 7.23×10−9 m2/s,11) whereas that in solid iron is 2.88×10−10 m2/s.4) The carbon diffusion coefficient in solid iron becomes 5.07×10−11 m2/s at 1273 K. Therefore, the carburization process of solid iron by solid graphite below 1273 K was considered ignorable, when it is compared to the entire carburization to melting.) The change of the shape and the dripping behavior of the sample were investigated using the digital camera. The image of the sample was captured at every 6 s. As the temperature reached 1773 K, it was kept at that temperature for a while. After the experiments, the sample was cooled in the furnace at a cooling rate of 20 K/min until 1273 K; thereafter the cooling-rate slightly decreased. In order to investigate the sample at a certain time, a sample assembly of the same geometry was positioned in a vertical reaction tube as shown in Fig. 3 and heated under the same condition. The sample was pulled out at the pre-determined time and temperature and then cooled by blowing N2 gas for ex-situ observation.
Schematic illustration of the experimental apparatus for the ex-situ observations.
After the experiment, the sample was mounted in polyester resin (EC-304, Aekyung Chemical Co., Ltd.). After drying the sample for 6 hours, it was cut using a diamond cutter. The cross-section of the sample was then investigated using a digital camera (D5100, Nikon, 2144 × 1424 pixels), and a field-emission scanning electron microscopy (FE-SEM, Hitachi S-4300). The composition of the slag was analyzed using scanning electron microscopy/energy dispersive X-ray spectroscopy (SEM-EDX, Hitachi S-4300). The carbon concentration in the dripped iron was analyzed using a carbon analyzer (LECO-C/S, CS230).
Figure 4 shows the sequential images of the pure iron sample placed on the graphite bed. At 1420 K, the iron sample slightly moved to the right-hand side. It is considered that this movement of the sample was caused by the formation of a liquid Fe–C alloy at the metal-graphite interface. At 1430 K, it is clearly shown that the top surface of the sample was no more flat and the sample was sinking with increasing temperature. At 1470 K, molten iron flowed through the graphite bed started to drip into the alumina receiver. The iron sample was fully melted at 1490 K, and the size of the liquid iron drop in the alumina receiver was maximized. The carbon concentration of the dripped liquid iron was 4.85 wt%, which was close to the carbon-saturation concentration (5.15 wt%) at 1773 K.
In situ observed images of the solid iron sample placed on the graphite bed.
The same experiment was performed on the coke bed; the results are shown in Fig. 5. Although there was a slight movement of the sample, change in the shape of the sample was not so considerable. After experiment, the carbon concentration of the solid iron was 0.62 wt%. It is clear that the carburization of solid iron on the coke bed was much slower than that on the graphite bed. In order to understand this difference, the cross-section of the sample on the coke bed was investigated after experiment using the digital camera and FE-SEM. Figure 6 shows macroscopic and microscopic images of the cross-section. The place labeled A, which is in contact with the iron, and the place labeled B, which is not in contact with the iron, were investigated at high magnification (12000×). A large amount of oxides were observed in A. The composition of the oxides was 87wt%Al2O3-9wt%SiO2-4wt%CaO from the EDX analysis. On the other hand, small amount of and very fine oxides were observed in B. It is speculated that the reaction might occur as follows. When the bare coke surface is contacted with solid iron, carburization may take place forming a shallow liquid Fe–C alloy layer at the interface. However, the ash content does not dissolve in iron, so that it is accumulated at the metal-coke interface. Since the liquid Fe–C alloy shows bad wettability against oxide particles, the liquid Fe–C layer can be detached from the coke’s surface and further carburization is stopped. As the liquid Fe–C alloy is contacting with solid iron, carbon diffuse to solid iron and liquid Fe–C alloy will be fully transformed to solid iron.
In situ observed images of the solid iron sample placed on the coke bed.
The cross-sectional images of the iron sample after experiments in Fig. 5.
As described in the Section 3.1, the carburization of solid iron would be suppressed by the formation of solid ash at the interface during the reaction between the solid iron and the coke. Therefore, how to remove the solid ash at the interface is critical to improve the carburization rate of iron. Since molten iron shows a good wettability with carbon, whereas molten slag a bad one,12,13,14) it can be expected that the formation of molten slag at the iron-coke interface might give chance to the iron to directly contact with carbon. Once carburization takes place, a new solid ash layer will be formed at the coke surface, and the molten slag has a chance to cover the solid ash layer again to dissolve it. Then, the reaction may occur repeatedly. Similar behavior was observed at the slag-metal-refractory interface.15) The flux content in the sinter would have a role in the molten slag formation. In this study, the samples containing 15, 30, and 50 wt% CAS slag were examined.
The sequential images of the sample containing CAS slag on the coke bed are shown in Figs. 7, 8, 9. In Fig. 7, the images of the sample containing 15 wt% CAS slag are shown. At 1623 K, the sample slightly moved probably due to the reaction of the sample at the bottom or the formation of liquid slag in the iron sample (Note that the melting temperature of the slag is 1632 K). However, there was no significant change further investigated, even by the time placing the sample at 1773 K for 391 s. After the experiment, no slag or iron was obtained in the alumina receiver. In Fig. 8, the images of the sample containing 30 wt% CAS slag are shown. The shape of the sample was changed from 1749 K. This temperature is higher than that of the sample containing 15 wt% CAS slag, but we cannot guarantee that there was no reaction below this temperature. Once the temperature reached 1773 K, the molten iron dropped into the alumina receiver. After keeping the sample at 1773 K for 342 s, it was noticed that the sample was fully melted and dripped through the coke bed. The carbon concentration of the dripped iron was 2.59 wt%. In Fig. 9, the images of the sample containing 50 wt% CAS slag are shown. A slight change in the shape was observed at 1723 K, which is close but slightly lower than that of the sample containing 30 wt% CAS slag. After keeping the sample at 1773 K for 37 s, the dripping was completed. Consequently, it is concluded that the more slag content, the faster carburization and melting. When the slag content is not so high enough, the carburization and melting of iron by coke can be prevented by the solid ash formed on the coke’s surface. The carbon concentration of the dripped iron from the experiment with the sample containing 50 wt% CAS slag was 2.72 wt%, which is not so different from the value obtained with the sample containing 30 wt% CAS slag. At 1773 K, the carbon concentration of liquid iron at graphite-saturation is 5.15 wt%. Since the carbon concentration in the dripped iron was 2.59–2.72 wt%, the chemical reaction between liquid Fe–C alloy and coke would be one of the rate–determining step. In the blast furnace, further carburization might occur when the molten iron was dripped through the coke bed and thereafter.
In situ observed images of the sample (iron-15 wt% CAS slag) placed on the coke bed.
In situ observed images of the sample (iron-30 wt% CAS slag) placed on the coke bed.
In situ observed images of the sample (iron-50 wt% CAS slag) placed on the coke bed.
Photographs of the samples taken after experiments are shown in Figs. 10(a)–10(c). In the case of the sample with 15 wt% CAS slag, no slag or iron was found in the alumina receiver. Meanwhile, tiny slag droplets were observed on the top of the coke bed after removing the sample; these were the evidence of molten slag formation. In the cases of the samples with 30 wt% and 50 wt% CAS slags, the iron and slag dripped through the coke bed were collected in the alumina receiver. Due to the bad wettability of slag with carbon, a certain amount of the slag remained on the top of the coke bed.16)
Photographs of the samples after experiments.
The cross-sections of the ex-situ samples are shown in Fig. 11. Since these samples were cooled by blowing N2 gas after pulled out from the furnace, further reactions might occur during the cooling process. Therefore it should be noted that time given below is the time when the sample was pulled out from the furnace. The cross-section of the sample with 15 wt% CAS slag after holding it at 1773 K for 390 s is shown in Fig. 11(a). The molten slag was not separated from but remained in the solid iron. The composition of the slag remaining in the solid iron (#1) was examined as 48.5wt%CaO-40.8wt%SiO2-10.7wt%Al2O3 through EDX. Although there was a slight difference in the amount of alumina present, the composition of the slag was similar to that of the prepared CAS slag, (48wt%CaO-40wt%SiO2-12wt%Al2O3). During the mounting of a sample with polyester resin, the coke could float due to the lower density. Therefore, the sample position was fixed by placing alumina flakes on the top of the iron sample. The cross-section of the sample with 30 wt% CAS slag at 1773 K is shown in Fig. 11(b). At this point, it is clearly seen that the molten slag and iron are separated, but are facing each other. Since many small slag particles are observed in the iron matrix, it is considered that the iron was not fully melted. Molten slag is mostly placed between the iron and the coke particle, but some part of the iron contacts with the coke directly. The slag composition of #2 was 45.1wt%CaO-42.9wt%SiO2-12wt%Al2O3. The SiO2 content was slightly higher than the initial composition, which might be attributed to the dissolution of ash content. After holding the sample at 1773 K for 90 s, the iron was fully melted and flowed down through the coke bed (Fig. 11(c)). It is noteworthy that the molten slag was investigated not only on the top surface of, but in the coke bed. Generally, solid oxides show bad wettability against molten iron.17) Therefore, if the coke’s surface was covered by solid ash, it would have been difficult for the molten iron to go through. Consequently, it is suspected that the molten slag removed the ash layer on the coke surface, allowing molten iron to go through. Hence, the molten slag investigated in the coke bed might have a role to remove ash layers on the coke’s surface. The slag composition of #3 was 46.4wt%CaO-41.2wt%SiO2-12.4wt%Al2O3. The concentration of SiO2 and Al2O3 slightly increased, which might be caused by the dissolution of ash content in the molten slag.
Cross-sectional images of the samples: (a) Fe-15 wt% CAS slag, t = 390 s at 1773 K, (b) Fe-30 wt% CAS slag, t = 0 s at 1773 K, (c) Fe-30 wt% CAS slag, t = 90 s at 1773 K.
From the observations in the present study, it is considered that wettability is a dominant factor determining the carburization and melting behavior. Molten slag shows good wettability with solid iron and oxides (ash), but bad wettability against solid carbon. (Molten slag also shows good wettability with solid iron. Parry and Ostrovski reported that the contact angle between CaO–SiO2–Al2O3 slag and solid iron was 44° at 1723 K.18) Molten slag shows good wettability with oxides as well. For example, the contact angle between CaO–SiO2–Al2O3 slag and alumina was reported in the range of 38–75° at 1873 K.19)) On the other hand, molten iron shows good wettability with solid carbon, but bad wettability against solid oxides. (Molten iron shows good wettability with solid carbon. It was reported that the contact angle between molten iron and graphite reached ~50° at 1873 K due to carburization reaction.12) On the other hand, molten iron shows bad wettability against solid oxide. For example, the contact angle between molten iron and alumina single crystals was reported in the range of 99–104° at 1873 K.20)) First, liquid Fe–C alloy is formed at the metal-coke interface. As the ash content is accumulated at the interface, liquid Fe–C alloy is detached from the coke’s surface. Then, molten slag is infiltrated between liquid Fe–C alloy and the ash layer. Then molten slag dissolves the ash content. Once ash content is fully dissolved by the slag, the molten slag may be removed out due to its bad wettability against carbon. Then, liquid Fe–C alloy contacts the bare coke surface again and carburization takes place successively. These processes are repeated until the iron is fully melted.
3.3. Effect of FeO in the SlagLarge slag volume in the sinter may yield high energy consumption in BF process. Therefore, we need some efforts to decrease the flux volume in the sinter. Kim et al.2) and Ohno et al.21) reported that the carburization of solid iron by carbon can be accelerated by the presence of FeO in the slag due to Marangoni flow. Therefore, it is expected that the presence of FeO in the slag might accelerate the carburization, melting, and dripping behaviors. In order to investigate the role of FeO, a sample mixed with 15 wt% CASF slag was prepared. (Note that the sample with 15 wt% CAS slag on the coke bed was not melted.)
In Fig. 12, the sequential images of the carburization and melting behaviors of the iron sample with 15 wt% CASF slag on the coke bed are shown. The change in the shape of the sample was observed from 1764 K, and the melting was initiated at 1770 K. These temperatures are not so significantly different from the previous experiments. Therefore, the melting point of the slag would not be the dominant factor. Surprisingly, the sample completely dripped into the receiver after holding the sample for 18 s at 1773 K. Figure 10(d) shows the sample after the experiment. The metal was collected in the receiver, and the carbon concentration was 3.03 wt%. Considering that the sample with 15 wt% CAS slag was not melted after holding it for 381 s at 1773 K, it is supposed that the FeO in the slag might accelerate the carburization of solid iron.
In situ observed images of the sample (iron-15 wt% CASF slag) placed on the coke bed.
The effect of slag content on the carburization, melting, and dripping of solid iron through the graphite and coke beds were investigated, and the following results were obtained.
(1) The iron sample placed on the graphite bed was carburized, melted and dripped completely at 1490 K, while the sample on the coke bed was not melted due to the interference of solid ash formed on the coke’s surface.
(2) As the slag content in the iron sample increased, the carburization, melting and dripping of solid iron was accelerated. It was considered that the molten slag dissolved the solid ash on the coke’s surface, allowing the iron to directly contact with carbon. However, when the slag content is not high enough (in the present experimental condition 15 wt% CAS slag), the melting and dripping of solid iron would not occur.
(3) The iron sample mixed with 15 wt% CASF slag was successfully carburized, melted and dripped through the coke bed. It was considered that the FeO in the slag might accelerate the carburization of solid iron.
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (No. NRF-2014R1A2A2A01007011).
