Abstract: Using tabular corundum as raw material, pure calcium aluminate cement and SiO2 micro powder as binding agents, we made a crucible sample of corundum blast furnace gunning material. We then conducted a blast furnace slag corrosion resistance test in a reducing atmosphere at 1550℃ for 3 hours.
We studied the effect of silicon nitride addition (0%, 5%, and 10%) on the slag corrosion resistance of corundum blast furnace gunning materials. Microscopic analysis of the corrosion samples revealed the slag corrosion mechanism. The results show that without silicon nitride, slag easily corrodes the gunning material.
Adding 5% and 10% silicon nitride improves the slag permeability resistance of the gunning material. This greatly enhances its slag resistance. Silicon nitride forms a network structure in the gunning material matrix, which is the main reason for its improved resistance to slag erosion.
Keywords: silicon nitride, slag resistance, gunning material, blast furnace
For a long time, iron-making workers and researchers have explored factors affecting blast furnace life. These factors include blast furnace design, raw materials, operations, cooling systems, construction, and refractory materials. Promising results have been achieved, continuously improving the blast furnace’s service life. Refractory materials significantly impact the blast furnace’s life.
Maintaining the furnace lining through blast furnace gunning is now an important method to extend blast furnace life. The construction method affects the blast furnace gunning material, resulting in a small volume density and high porosity of the gunning layer. Compared to masonry lining, it is more susceptible to erosion and damage by blast furnace slag. Therefore, gunning materials used below the blast furnace shaft must have good slag resistance.
To improve slag resistance, silicon nitride materials are introduced into new blast furnace gunning materials. This work studied the effect of silicon nitride on the slag resistance of corundum blast furnace gunning materials. Static slag resistance tests were conducted to understand the erosion mechanism.
Test method
The main raw materials used in the test are tabular corundum and Si3 N4 powder, with pure calcium aluminate cement and SiO2 micro powder as the binding agent. The sample proportions are shown in Table 1.
Table 1 Sample proportion (w) %
| Sample No | tabular corundum | Alumina micropowder | SiO2 micropowder | Pure aluminic acid | Calcium cement Si3N4 powder | ||||
| 5 ~3 mm | 3 ~2 mm | 2 ~1 mm | 1 ~0. 2 mm | < 0. 2 mm | |||||
| T1 | 10 | 20 | 20 | 14 | 15 | 5 | 6 | 10 | 0 |
| T2 | 10 | 20 | 20 | 14 | 10 | 5 | 6 | 10 | 5 |
| T3 | 10 | 20 | 20 | 14 | 5 | 5 | 6 | 10 | 10 |
Add an appropriate amount of water to the test materials in different proportions. Stir and mix thoroughly, then pound the mixture into the iron mold. Demould after 24 hours. Place the sample block in water at about 20°C for 24 hours. Remove it and place it in a ventilated, dry area to dry naturally for another 24 hours.
Then, put the sample block in an oven at 110°C to dry for 24 hours. Take it out and let it cool naturally. Mechanically drill holes into the specimen to form the crucible. Add 100 g of blast furnace slag to each slag-resistant crucible sample.
To simulate actual blast furnace conditions, place the crucible sample in a sagger. Bury it with graphite powder, seal it, and heat treat it in a high-temperature furnace at 1550°C for 3 hours. Remove it after cooling. The slag used in the test comes from Angang Iron and Steel Plant. The chemical composition (w) of the slag is: CaO 42.14%, MgO 7.03%, SiO2 40.06%, Al2O3 6.88%, FeO 0.53%, S 0.72%. Section the eroded specimens longitudinally for appearance and microstructural analysis.
Results and discussion
Appearance analysis of corrosion samples
Figure 1 shows the cross-sectional appearance of the sample after erosion. It can be seen from the figure that when silicon nitride is not added, the slag penetration of the sample is relatively serious, with obvious erosion zones and reaction layers. There is very little slag remaining in the crucible, and most of the slag penetrates into the sample. When the addition amount of silicon nitride is 5%, the slag penetrates the sample slightly. When the addition amount of silicon nitride is 10%, the bottom of the sample is intact, the slag hardly wets the bottom and inner wall of the sample, and there is a large amount of slag in the crucible.

Figure 1 Appearance of the sample after erosion
As we increase the amount of silicon nitride, the material’s resistance to slag penetration improves. Silicon nitride does not oxidize in a reducing atmosphere. The wetting angle between silicon nitride and slag at high temperatures ranges from 110° to 130°, which is much larger than the wetting angle between corundum refractory material and molten slag. This larger wetting angle makes it difficult for blast furnace slag to penetrate the material’s pores, thereby enhancing the material’s ability to resist slag erosion.
Microstructure analysis and corrosion mechanism analysis
Microscopic analysis of the T1, T2, and T3 etched samples was performed using SEM. Figure 2 shows the SEM image of the T1 sample after erosion. The figure reveals no clear boundary between the erosion layer and the transition layer. The entire sample has a loose structure with many holes.
Some small particles in the image have smooth edges and remain uneroded. However, the slag has spread throughout the sample. The Ca element surface distribution map shows the erosion situation. The slag penetrates the sample’s interior through its matrix. Large particles have been corroded by the slag.
The CaO in the slag reacted with the matrix part of the gunning material. This reaction generated a new mineral phase on the transition layer. Thermodynamic analysis suggests this mineral phase may be an aluminum-silicon-calcium series low-melting point compound. The diffusion of Fe element is similar to that of Ca element. The low melt formed from FeO and CaO reacting with aluminum and silicon in the gunning material quickly melts into the slag. The slag penetrates the gunning material layer along the loose interface and pores, further reacting with the aggregate and matrix, causing erosion.

Figure 2 SEM image and element surface distribution map of T1 sample after slag erosion
Figure 3 shows the SEM image of the T2 sample after erosion. The image reveals a very clear interface between the slag zone and the transition zone. The element distribution map indicates that the Ca element distribution gradually decreases from the slag zone to the transition zone, suggesting reduced slag penetration.
The silicon nitride in the raw material distributes within the matrix, hindering the invasion of slag. This network structure of silicon nitride within the matrix is the main reason for the improved penetration resistance.

Figure 3 SEM image and element surface distribution map of T2 sample after slag erosion
Figure 4 shows the morphology of the T3 sample after erosion. It can be seen from the figure that there is no transition zone, and the interface between the slag and the gunning material is very obvious. Due to the addition of 10% silicon nitride powder, the slag could not enter the matrix. Compared with the morphology of the T1 sample after erosion, it shows that the addition of silicon nitride plays a role in resisting slag erosion.

Figure 4 SEM image and element surface distribution map of T3 sample after slag erosion
Conclusion
When we do not add silicon nitride to the corundum gunning material, the blast furnace slag easily corrodes it. However, adding silicon nitride to the gunning material improves its slag resistance permeability and greatly enhances its slag resistance ability.Silicon nitride has a network structure in the sample matrix, which is the main reason for improving the slag resistance of the gunning material. As the amount of silicon nitride added increases, the ability to resist slag penetration increases. The appropriate addition amount of silicon nitride powder is more than 5%.

