Melting Behavior and Interaction of Gangue Phase of Iron-containing Burden

Abstract

The cohesive zone is an important location in the blast furnace due to decrease of permeability and sharp increase of pressure drop. Since the liquid phase has an important influence on the permeability, this study aimed to investigate the regulation of liquid phase formation in the cohesive zone. Through melting experiments of single and double slag tablets combined with thermodynamic calculations, some important phenomena were observed. The pellet slag melted prior to sinter slag, and the molten pellet slag caused the sinter slag to melt at a temperature below the melting temperature. The liquid phase of the cohesive zone originated due to the low melting point material of the gangue phase, and the interaction of different iron-containing burdens gangue phase. The production of the liquid phase was affected by the composition and temperature. The progress of the interaction seemed to be affected by the liquid phase ratio. The interaction was re-understood: interaction is physicochemical reactions of gangue phases of different iron-containing burdens which occurred during the formation of molten blast furnace slag with uniform composition. Thermodynamic calculations were consistent with the experimental results and showed that the interaction of the cohesive zone increased the liquid phase, which was disadvantageous for the permeability.

1. Introduction

Iron-containing burdens in blast furnaces are softened and melted under the upper load and high temperature to form cohesive layers with extremely small porosity, which is disadvantageous to the gas permeability of the blast furnace. As a result, permeability of the cohesive zone is significant for the production of the blast furnace. Different softening and melting properties of iron-bearing burdens due to different composition and reducibility. Softening and melting properties of iron-bearing burdens and the interaction between different burdens have a significant impact on the formation of the cohesive zone.^1,2,3,4) Therefore, it is necessary to study the properties and interaction of burdens in order to obtain a cohesive zone which is advantageous for permeability.

A significant amount of research has been directed towards the improvement of permeability of the cohesive zone. Although, low-position and thin cohesive zone have been advantageous for gas permeability,^5,6,7,8) they are determined by the characteristics of the burdens and the operating state of the blast furnace. In addition, key factors of these studies include the interaction of different iron-bearing burdens,^1,2,3,4) structure of the burden lay distribution⁹⁾ and the reduction degree.^10,11,12) Specific methods include softening and melting experiments, interaction between single-particle pellets and sinters as well as cold physical simulation of different structure experiments. However, during softening and melting experiments, the entire process cannot be observed due to the influence of the experimental equipment and the iron phase. Due to the constant liquid phase ratio, cold physical simulation methods cannot reflect the order of appearance of the liquid phase, as well as the effect of interaction on the liquid phase formation.

Although, the interaction between different burdens has been extensively studied, definition of its properties has not been explored. According to previous reports, it can be inferred that the interaction of gangue phase between different burdens promotes the formation of blast furnace slag, and this effect is more prominent at high temperatures.^1,2,3,4) The temperature seems to affect the melting process of the gangue phase and interaction. In the present study, in order to avoid the influence of iron on the observation of melting behavior, individual gangue phases and the interaction between them were investigated. In order to explain some of the phenomena in the experimental process, necessary thermodynamic calculations were carried out.

2. Experimental

2.1. Preparation of Experimental Materials

The sinter and pellet slags were gangue components of sinters and pellets, which were used in blast furnace ironmaking by a steel company in the Shandong Province of China. The main components of pellets and sinter are listed in Table 1. CaO, SiO₂, Al₂O₃, MgO and FeO were used to prepare slag of the same composition as the sinter and pellets gangue. Fe₃O₄ was reduced to produce FeO, the reduction temperature was 900°C, reducing gas was composed of CO and CO₂ with a ratio of 1:1, and the reduction time was 3 h. These oxides were mixed and retained at 1450°C for 5 h under high-purity Ar atmosphere. Pre-melted oxides were removed and cooled down to room temperature, yielding sinter (S) and pellet slags (P). The compositions of these slags are listed in Table 2. The solid pre-melted slag was broken up into 200 mesh for XRD testing and preparation of slag tablets. P slag was pressed into tablets of 10 mm in diameter, whereas S slag was pressed into tablets of 15 mm in diameter. The molding pressure was 20 MPa.

Table 1. Experimental reference to the main components of sinter and pellet.

burden	TFe	FeO	CaO	SiO2	Al2O3	MgO	R
sinter	57.51	8.44	9.55	4.92	1.82	1.54	1.94
pellets	61.4	0.87	1.16	6.63	1.27	1.97	0.17

Table 2. Compositions (mass%) of slags used in experiments.

Slag	CaO	SiO2	MgO	Al2O3	FeO	CaO/SiO2
P-1	10.51	60.11	17.86	11.52	0	0.17	1.3
S-1	51.9	29.05	10.86	8.19	0	1.78
P-2	9.46	54.1	16.07	10.37	10	0.17	1.3
S-2	46.71	26.15	9.77	7.37	10	1.78
P-3	8.93	51.09	15.18	9.79	15	0.17	1.3
S-3	44.11	24.69	9.23	6.96	15	1.78

2.2. Experimental Methods

Schematic of the experimental equipment involving a horizontal resistance furnace is shown in Fig. 1. During the experiment, the sample was placed on a molybdenum sheet in the constant temperature zone of the furnace. Double-tablet experiments were conducted: P-slag tablets were placed on top of the S-slag tablets, while the sum of the weights of the P slag and S slag tablets was maintained at 1.5 g. The basicity was 1.3. The heating rate of the furnace was 5°C/min and the temperature of the furnace was measured with a thermocouple. During the entire experiment, high purity nitrogen gas was introduced through the gas inlet at 5 L/min. The experimental temperature range was 25°C–1450°C, while the molten state of the slag tablet was recorded with a high temperature camera. Subsequently, the sample was analyzed using XRD. A single S slag experiment was performed, in order to obtain the results for comparison. In this study, the melting temperature was defined as the slag tablet was observed to deform significantly during the experiment, usually higher than solidus temperature. The melting process started with significant deformation of slag tablet and ended at the temperature of 1450°C.

Fig. 1.

Schematic diagram of experimental equipment. (Online version in color.)

2.3. Method of Thermodynamic Calculation

In order to analyze the melting mechanism, necessary thermodynamic calculations were performed by Factsage software package. The thermodynamic calculation used the function of multivariate multiphase equilibrium, and analyzed database of oxides to select specific materials based on the results of XRD and previous literature.^{13,14,15,16,17)} The calculated temperature range is from 1000°C to 1450°C, while including the solidus temperature of all calculated slag components. The calculated results yield the liquid phase ratio of each component and the material composition of the solid phase for analysis in combination with the experimental results. In addition to the P slag and S slag in the experiment, the weighted average their liquid ratio and the liquid phase ratio of the mixed slag combined with the basicity 1.3 could also be obtained.

3. Results and Disscusion

3.1. Observations of Tablet States under Different Temperature Conditions

As shown in Fig. 2, experiments of P-3 and S-3 (15% FeO content) were used as an example, while the entire experiment was recorded. Figure 2(1) shows the initial state, and Figs. 2(2) to 2(4) present the P slag melting processes. P-3 slag melting is evident in Fig. 2(2), whereas melted P-3 slag droplet can be seen in Fig. 2(3). Complete melting of P-3 slag on the tablet surface of the S-3 slag is seen in Fig. 2(4). Figure 2(5) shows the melting onset of the S-3 slag, but the temperature of Fig. 2(5) was higher than the temperature of Fig. 2(4). Figure 2(6) shows the S slag in the melting stage. As the S slag melted, the height of the tablet decreased. In Fig. 2(7), most of the S slag melted and the height of the slag tablet did not change. Figures 2(5) to (7) were the melting sections of the S slag. The temperatures of the critical states of Figs. 2(2), 2(4), 2(5) and 2(7) were plotted as a bar graph, as presented in Fig. 3.

Fig. 2.

Changes of slag tablets state during P-3 and S-3 experiments. (Online version in color.)

Fig. 3.

Status of slag at different temperatures.

In Fig. 3, P and S slags in zone 1 were in solid state. The temperature range at which the pellet slag begins to soften, completely melt and lay flat on the sinter slag is defined as the melting range of the pellet slag, i.e. zone 2. When the FeO content was 15%, 10% and 0%, temperature range of zone 2 was 53°C (1226°C–1279°C), 64°C (1228°C–1292°C) and 68°C (1252°C–1320°C), respectively. It could be observed that as the FeO content decreased, the melting zone of the P slag increased and moved towards the high temperature region. Zone 3 was the temperature zone, in which the P slag was completely melted until the S slag began to melt. When the FeO content was 15%, 10% and 0%, the temperature range of zone 3 was 10°C (1279°C–1289°C), 27°C (1292°C–1319°C) and 67°C (1320°C–1387°C), respectively. It could be observed that as the FeO content decreased, the temperature range of zone 3 was extended and zone 3 moved towards the high temperature region. Zone 4 was the temperature zone, in which, the S slag began to melt until the height of the slag did not change. When the FeO content was 15%, 10% and 0%, the temperature range of zone 4 was 50°C (1289°C–1339°C), 36°C (1319°C–1355°C) and 24°C (1387°C–1411°C), respectively. It could be observed that as the FeO content decreased, zone 4 was decreased in size and moved towards the high temperature region. Zone 5 was high liquid ratio zone.

The comparison experiments of the S slag are shown in Fig. 4. When the FeO content of the S slag was 15%, 10% and 0%, the slag tablet began to melt at 1325°C. When the FeO content of the S slag was 10%, the slag tablet melting was observed at 1346°C. When the FeO content of the slag tablet was 0, the temperature increased to 1450°C and the slag tablet did not melt. As can be seen from Figs. 3 and 4, the P slag was melted prior to the S slag, and the molten P slag caused the S slag to melt at a temperature lower than the melting temperature.

Fig. 4.

Comparison experiments of single S slag tablets. (Online version in color.)

3.2. XRD Test Results

Figure 5 shows the XRD results of the pre-melted slags and experimental samples. Figure 5(1) shows the XRD results of the pre-melted P slag, in which no substance was detected. Since the P slag was cooled down at room temperature, liquid P slag might have failed to crystallize and produced a high amount of a glass phase. As shown in Fig. 5(2), when the S slag did not contain FeO, the slag was composed four phases, namely, Ca₂SiO₄, Ca₂Al₂SiO₇, Ca₃Mg(SiO₄)₂, and MgO. When FeO was contained in the S slag, the MgO phase disappeared and the 2FeO·SiO₂ phase appeared. As shown in Fig. 5(3), only the Ca₂Al₂SiO₇, Ca₃Mg(SiO₄)₂ and Ca₂Mg(Si₂O₇) phases existed in the double-tablet experimental samples when the sample did not contain FeO. When the samples of the double tablets contained FeO, Fe₂SiO₄ was obtained.

Fig. 5.

XRD results of P slags, S slags and experimental samples. (Online version in color.)

3.3. Results of Thermodynamic Calculations

Results of the thermodynamic calculation are shown in Fig. 6. The black line, the red line, the blue line and the purple line represent the liquid phase ratio of the P slag, the liquid phase ratio of the S slag, the sum of the liquid phases of the S slag and the P slag, and liquid phase ratio of the mixed slag of the P slag and the S slag, respectively. The solidus temperatures of P1, P2 and P3 were obtained to be 1209°C, 1116.8°C and 1095°C, respectively. It is needed to indicate that the results obtained by thermodynamic calculation were based on equilibrium state and the experimental results were often based on non-equilibrium state. According to calculation results, as the FeO content increased, the solidus temperature of the P slag decreased. The solidus temperature of P slag was lower than the melting temperature of P slag in Fig. 3 from 40°C to 130°C. At the same temperature, the liquid phase ratio of P slag increased with increasing FeO content. It can be seen in conjunction with Figs. 4 and 6 that the S slag satisfies the same regulation. This reflects the consistency of experimental results and thermodynamic calculations. Additionally, the liquid ratio is about 80% when the P slag reaches equilibrium at the temperature when the melting behavior occurs. However, the equilibrium liquid phase ratio of S-2 and S-3 is about 45% when melting behavior occurs. This discrepancy can be explained by the fact that it is impossible to reach equilibrium under experimental conditions. Hence, the true liquid phase ratio is lower than the calculated value. It is found by thermodynamic calculation that the M curve is higher than Sum (LP & LS), so the liquid phase ratio of the mixed slag is not only the sum of the liquid phase rates of P and S slag, but a new liquid phase is produced. The new liquid phase leads to an increase in the liquid phase ratio, which could account for the P-slag melting promoting the melting of the S-slag.

Fig. 6.

Results of thermodynamic calculation. (Online version in color.)

In the P-1 slag, the solid phase before the liquid phase appeared as olivine phase, CaAl₂Si₂O₈ and SiO₂. The olivine phase is a solid solution composed of Mg₂SiO₄, CaMgSi₂O₆, etc. When the temperature exceeded the solidus temperature, the liquid phase began to appear. As the temperature continued to rise, the solid phase disappeared in the sequence of SiO₂, Ca₂Al₂SiO₈, and then olivine phase. In the slag of P-2 and P-3, CaFeSiO₄ and Fe₂SiO₄ phases were added to the olivine phase, which were both low melting point materials. As the temperature increased, the order in which the solid phase disappeared was SiO₂, Ca₂Al₂SiO₈, and finally olivine phase.

In the S-1 slag, the solid phase before the liquid phase appeared as Ca₂SiO₄, MgO and Ca₃MgSi₂O₈. As the temperature increased, MgO and Ca₃MgSi₂O₈ in the solid phase disappeared, and finally the solid phase in equilibrium with the liquid phase was Ca₂SiO₄. In the slag of S-2 and S-3, the solid phase before the liquid phase appeared as Ca₂SiO₄, Monoxides (FeO and MgO), Fe₂SiO₄, Mg₂SiO₄, Ca₃MgSi₂O₈ and Ca₂Al₂SiO₇. After the liquid phase appeared, Ca₃MgSi₂O₈ and MgSiO₄ disappeared, the MgO content in the monoxide increased and so did Ca₂SiO₄ content. Finally, the solid phase in equilibrium with the liquid phase was Ca₂SiO₄ and monoxide.

When the M slag did not contain FeO, the solid phase before the liquid phase appeared as olivine, CaA1₂Si2O₈ and Ca₃MgSi₂O₈. The order in which the solid phase disappeared after the liquid phase appeared was CaA1₂Si₂O₈ and then olivine. Finally, the liquid phase was in equilibrium with Ca₃MgSi₂O₈. When FeO was present in the M slag, Fe₂SiO₄ and CaFeSiO₄ appeared in the olivine phase.

3.4. Melting and Interaction Mechanism of Slag

In the P1 slag, Mg₂SiO₄, CaA1₂Si₂O₈ and SiO₂ exhibited an eutectic point at 1220°C.^13,14) Mg₂SiO₄, CaMgSi₂O₆ and CaA1₂Si₂O₈ also reacted at 1270°C to produce a liquid phase.^14,15) When FeO was added to the P slag, the above reaction occurred between 1200°C and 1220°C.¹⁶⁾ Further, in the CaO–SiO₂–FeO system, there was a low temperature eutectic point (1093°C).¹⁷⁾ P slag solidus temperatures were low and the liquid phase ratio increased with increase of FeO content at the same temperature. When the P slag melted and transferred mass to the S slag, this the high melting point substances such as Ca₂SiO₄ and Ca₃MgSi₂O₈ in the S slag were destroyed, transforming the phase of P and S slag into M slag. The S slag generated a liquid phase at a temperature lower than the solidus temperature and increased the liquid phase rate at the same temperature. As can be seen from Fig. 2, the initial interaction did not occur on the contact faces of the S slag and the P slag, but after one of the slags melted, penetrated and mixed with the other slag. Due to the phase transformation and pulverization of Ca₂SiO₄, the experimental samples could not be retained and the melting behavior could not be observed by microscopy. It can be seen from the Sum (LP & LS) and M curves of Fig. 6 that the liquid phase ratio of the fully mixed slag was high. This may be due to the high occurrence of the reaction that produced the liquid phase, since the mixing shortened the transport distance of the material at the same temperature. It can be seen from Fig. 5(3) that the phase compositions of the experimental samples of the double tablet were same, but they were different from the compositions of the pellet and the sinter slags, which indicated that the interaction between the pellet slag and the sinter slag occurred.

Through experimental and thermodynamic calculations, the interaction of different iron-containing burdens is better understood: interaction is physicochemical reactions of gangue phases of different iron-containing burdens which occurred during the formation of molten blast furnace slag with uniform composition. Additionally, the interaction is affected by both composition and temperature. The logical relationship of this effect can be represented by Fig. 7. The melting process of the double slag tablet is complicated, but this process can be described based on experimental results and thermodynamic calculations:

Fig. 7.

Different reaction mechanisms of experimental process and LM curve. (Online version in color.)

(1) As the temperature rises, the P slag melts,

(2) the molten P slag mass transfer to the S slag, changing the composition of the S slag, but the current temperature does not cause the S slag to reach the melting temperature (zone 3 in Fig. 3);

(3) As the temperature increases, the S slag after mass transfer by the P slag melts, and this temperature is lower than the melting temperature of the S slag;

(4) The increase of temperature promotes the increase of liquid phase ratio and mass transfer, and the liquid phase ratio is closer to the liquid phase ratio of the mixed slag of the two kinds of slag.

Increase in liquid phase ratio leads to poor permeability of the cohesive zone. According to Ref.,¹⁸⁾ different CaO–SiO₂–FeO slags penetrated into coke layer at the same temperature (1450°C). At the same time, the high melting point slag was allowed to remain in the coke layer, that is, when the burden entered the dripping zone, the slag may not necessarily all convert to a liquid phase. However, from the experimental and thermodynamic calculations, the temperature at which the interaction occurs was within the temperature range of the cohesive zone. This caused the cohesive zone to produce more liquid phase resulting in a deterioration in permeability.

Complete mixing of different iron-containing burdens may promote the occurrence of interactions. During the experiment, the P slag and the S slag were completely separated. After the P slag was melted, the interaction proceeded from top to bottom. The result of the thermodynamic calculation showed that the LM curve was above the sum (LP & LS) curve. The true liquid phase ratio in the experiment should be between Sum (LP & LS) and M curves, and gradually close to the M curve until the temperature increased. It can be seen that the liquid phase of the cohesive zone is generated by the burdens and the interaction. As shown in Fig. 7, the interaction of the mixed slag occurs throughout the region, which may increase the rate at which the interaction occurs and proceeds towards completion. However, this result seems to be the opposite of the conclusion of Ref. 9, because a cold-state experiment was used that could only characterize the total liquid phase of different burdens while ignoring the interaction between different burdens to the liquid phase appearing at lower temperatures.

Experimental and thermodynamic calculations indicate some methods of the improvement of the permeability of the cohesive zone. Firstly, the reducibility of the ore was improved and the atmosphere in the blast furnace was reduced along with the FeO content of the ore at the cohesive zone. For example, the utilization of magnesium-containing pellets and increasing the SFCA-1 of the sinter, the injection of oxygen-enriched with pulverized coal (PCI), the injection hot reduction gas of the blast furnace, and even the oxygen blast furnace. Secondly, the mixing of different kinds of burdens was avoided during blast furnace charging, although this was not proved by high temperature experiments or actual production.

4. Conclusion

The interaction process between different slags was observed and explained by single and double slag experiments and necessary thermodynamic calculations. P slag exhibited lower solidus temperature and higher liquid phase ratio due to different composition of the components. The melting of the P slag caused interaction, which caused the S slag to exhibit a melting behavior at the temperature lower than the melting temperature. This resulted in the higher production of the liquid phase, which was disadvantageous for the gas permeability of the blast furnace. In this study, the interaction was better understood as interaction was physicochemical reactions of gangue phases of different iron-containing burdens which occurred during the formation of molten blast furnace slag with uniform composition.

Acknowledgements

The authors gratefully acknowledge the financial support from National Key Research and Development Program (No. 2016YFB0601304).

References

• 1)   L. J.  Yan,  S. L.  Wu and  H. L.  Han: Chin. J. Process Eng., 10 (2010), Suppl. No.1, 93 (in Chinese).
• 2)   S. L.  Wu,  H. L.  Han,  H. F.  Xu and  L. J.  Yan: Chin. J. Process Eng., 10 (2010), Suppl. No.1, 37 (in Chinese).
• 3)   S. L.  Wu,  G. J.  Wang,  W. Z.  Jiang,  J. D.  Sun and  H. F.  Xu: Iron Steel, 42 (2007), No.3, 10 (in Chinese).
• 4)   S. L.  Wu,  L. X.  Wang,  Y. J.  Wang and  J. C.  Zhang: Chin. J. Eng., 38 (2016), 1546 (in Japanese).
• 5)   K.  Sunahara,  T.  Natsui,  K.  Shizawa and  Y.  Ujisawa: ISIJ Int., 51 (2011), 1322 (in Japanese).
• 6)   G. G.  Zha,  X. H.  Fan,  X. L.  Chen,  L. S.  Yuan and  M.  Gan: J. Cent. South Univ. (Sci. Technol.), 41 (2010), 2053.
• 7)   S.  Ueda,  T.  Miki,  T.  Murakami,  H.  Nogami and  T.  Sato: Tetsu-to-Hagané, 99 (2013), 1 (in Japanese).
• 8)   N.  Takeuchi,  Y.  Iwami,  T.  Higuchi,  K.  Nushiro,  N.  Oyama and  M.  Sato: Tetsu-to-Hagané, 99 (2013), 448 (in Japanese).
• 9)   J.  Ishii,  R.  Murai,  I.  Sumi,  Y. X.  Yang and  R.  Boom: ISIJ Int., 57 (2017), 1531.
• 10)   Y. L.  Liu,  W. T.  Guo,  Q. G.  Xue and  J. S.  Wang: Sci. Sin. Technol., 47 (2017), 933 (in Chinese).
• 11)   L.  Chen,  Q. G.  Xue,  W. T.  Guo,  X. F.  She and  J. S.  Wang: Ironmaking Steelmaking, 43 (2016), 458.
• 12)   X. W.  An,  J. S.  Wang,  R. Z.  Lan,  Y. H.  Han and  Q. G.  Xue: J. Iron Steel Res. Int., 20 (2013), 11.
• 13)   P. L.  Roeder and  E. F.  Osborn: Am. J. Sci., 264 (1996), 428.
• 14)   E. F.  Osborn,  R. C.  Devries,  K. H.  Gee and  H. M.  Kraner: JOM, 6 (1954), 33.
• 15)   H. L.  George,  B. J.  Monaghan,  R. J.  Longbottom,  S. J.  Chew and  P. R.  Austin: ISIJ Int., 53 (2013), 1172.
• 16)   W.  Gutt and  A. D.  Russell: J. Mater. Sci., 12 (1977), 1869.
• 17)   N. L.  Bowen,  J. F.  Schairer and  E.  Posnjak: Am. J. Sci., 26 (1933), 193.
• 18)   S.  Ueda,  T.  Kon,  T.  Miki,  S. J.  Kim and  H.  Nogami: ISIJ Int., 55 (2015), 2098.