Ironmaking
Characterization and Properties of Scaffold in a Dissected Blast Furnace Hearth
2019 Volume 59 Issue 12 Pages 2205-2211
Details
2019 Volume 59 Issue 12 Pages 2205-2211
The blast furnace scaffold can only be obtained while the blast furnace shut down after operating for many years. Its characteristics and properties are important for the blast furnace campaign life. The key to delaying the carbon brick corrosion in blast furnace hearth is the scaffold formed between the melt and the carbon brick. In an emergency shutdown blast furnace, the scaffold in hearth is completely preserved, and the scaffold on the surface of the carbon brick above and below the taphole in hearth are sampled. The purpose of this study is to describe the characterization and properties of the scaffold in hearth. The paper presents results from investigations using electron imaging techniques such as Transmission Electron Microscopy (TEM), Optical Microscope (OM), Scanning Electron Microscope combined with Energy Dispersive Spectrometer (SEM-EDS), Raman analysis and X-ray Diffraction (XRD). The main component of the slag skull above the taphole is similar to the final slag and is rich in harmful elements. The thermal conductivity of the scaffold is about 2 W/(m·K) and the viscosity as well as the solidus temperature are higher than the final slag. The slag skull acts to isolate and contain harmful elements. The phase on the hot surface of the carbon brick below the taphole is mainly consist of graphite and the large-grained graphite phase has a random spatial network distribution in the iron matrix. The slag skull and the graphite serves to segregate the melt and harmful element, thereby protecting the carbon brick and extending hearth life.
The blast furnace (BF) is the world’s largest metallurgical reactor, with more than 95% of the molten iron produced by BF.1,2) For the economic and low carbon operation of the ironmaking BF, it is important to ensure a long campaign life without expensive and time-consuming relining.3) The service life of the BF hearth directly determines the longevity of BF.4,5) The key to delaying the carbon brick corrosion in hearth is the scaffold formed between the melt and the carbon brick.6,7) To be able to take measures to maintain a smooth operation of the BF hearth, it is necessary to get information of the scaffold. The solidification and melting of scaffold on the hearth lining depend strongly on the flow conditions in hearth. It is possible to detect changes in the internal flow patterns by monitoring the state of scaffold behavior. Therefore, it is necessary to know about the characteristics and properties of the scaffold. However, due to the hostile conditions in hearth, no direct measurements of the internal conditions such as scaffold are possible to be obtained.8) The only way to obtain the scaffold is the BF shutdown and overhaul. So the scaffold exists in the BF may exceed more than 10 years as it is the average campaign life of BF. Zhao et al.9) claim that the scaffold provides a barrier between carbon brick and flowing liquid iron or slag. As long as the scaffold exists, the hearth erosion nearly stops. The scaffold is a layer of iron or slag solidified on the inner surface of the hearth lining. But they didn’t give the compositions and the prosperities of the scaffold and the judgement is just speculation. To be able to estimate the state of the hearth, some models for calculating the remaining sound lining and the amount of scaffold on the bottom and sidewall based on thermocouple readings has been developed.10,11,12,13) For example, Keisuke et al.14) reported on the use of the scaffold thickness to monitor the hearth lining and to take control actions against erosion. But the information of the scaffold such as the thermal conductivity is hypothetical. In our previous work, the scaffold layer was classified as iron-rich layer, slag-rich layer, graphite-rich layer and titanium-rich layer according to the formation mechanism of protective layer based on a large number of BF damage surveys and dissection research.15,16,17,18) But the studies are mainly concentrated on the study of the scaffold below the center line of the BF taphole. In this work, we further analyzed the scaffold above the center line of the taphole and gave a more comprehensive analysis of scaffold below the center line of taphole. The estimates from the scaffold properties, in turn, can be used to assist in the interpretation of the internal state of the hearth. It has important guiding significance for the monitoring of the BF in the actual production process and the safety of the hearth.
The details of typical overhaul process have been described in the earlier publications.19,20) In the present study, the BF from which the samples were extracted, was characterized by a working volume of 1350 m3, 2 tap hole and 22 tuyeres. The BF was put into operation in September, 2005 and shut down in October 2015 for maintenance. The BF has been operated smoothly at a productivity of about 2.8 tHM/(m3·d). The feed iron grade was about 56.8%, and the iron-bearing materials were composed of sinter and pellet at a mass ratio of 75:25. The average coke and coal consumption rates were maintained as 380 kg/tHM and 150 kg/tHM, respectively. The average blast temperature was 1473 K (1200°C).
After the blowing-out and cooling-down processes (cooled by cooling water) of the furnace, the hearth was dissected. From the eroded hearth wall, the remaining carbon brick in hearth generally had a thickness of from 100 to 400 mm. Besides, in the front of the carbon brick, a significant amount of scaffold was observed with a thickness ranging from 100 to 300 mm. The sampling locations and the erosion profiles were shown in Fig. 1(a). In this study, the circumferential positions where the samples were collected are close to the taphole of the furnace and between tuyeres 1 and 22. The samples were taken from the hot surface of the carbon brick at different heights in hearth. The sampled areas were marked in Fig. 1(b) as sample 1 and sample 2, where sample 1 represented the location 1.0 m above the taphole centerline level, sample 2 was the thinnest area of carbon brick which is about 1.5 m below the taphole centerline.
BF hearth structure and sampling locations of hearth bottom. (a): Sampling locations; (b): Erosion profile and samples. (Online version in color.)
The samples were also selected for XRD (X-ray diffraction) examination. The analysis was conducted using a Rigaku diffractometer (DMAX-RB 12 kW; Rigaku Corporation, Tokyo, Japan). During this analysis, the scanning angles were in the range of 10 to 90 deg (2θ) at a scan rate of 10 deg/minute. The chemical compositions were analyzed via XRF (X-ray fluorescence) (Shimadzu XRF-1800, Japan). Additionally, in order to allow SEM observation, pieces were cut from the selected samples under dry conditions and then placed in a rounded plastic container with 25 mm diameter which was filled with resin. Then, the sample was ground and polished in a way similar to the previous study.15,16) The samples were coated with Au and then examined with a Quanta 250 environmental scanning electron microscope (SEM) equipped with energy-dispersive X-ray spectrometer (EDS) for chemical analysis and element mapping. The details of the microstructure of sample 2 were also detected by optical microscope (OM, Neophot-30) and transmission electron microscope (TEM, JEM-2200FS).
A large amount of slag (gray phases) are found in sample 1, as shown in Fig. 2. Figure 2(a) presents the global appearance of the sample, and reveals that several phases co-exist in the sample which has two distinct layers. The gray phase is slag, the black part is carbon and the white phase is mainly Fe. So the left side (near carbon brick side) confirms the formation of slag and the compositions of the slag skull are shown in Table 1. It is the slag skull that exists in the actual production process of the blast furnace. The compositions of slag skull is similar to the final slag while MgO content is lower and CaS is higher than that of final slag. Also, the alkali oxides and ZnO are richer in the slag skull compares with final slag. The phase of the slag skull are mainly confirmed by the distinct characteristic peaks in the XRD pattern, which are indexed as KAlSi2O6, Ca2MgSi2O7, Ca2Al2SiO7, Ca2ZnSi2O7. Figures 2(c) and 2(d) shows the SEM-EDS mapping of Sample 1. Elements of Ca, Mg, Al, Si and O are in the same phase near cold left side, and the right side is observed as carbon matrix. Na, K, Zn present at the interface of the slag phase and the carbon matrix. It can be confirmed that the slag skull is mainly formed by condensing the BF slag on the hot surface of the carbon brick. The harmful elements in the BF hearth would be accumulated near the carbon brick and permeated into the slag skull.
SEM-EDS maps showing the distribution of Ca, Si, Al, Mg, Fe and Na, K, Zn in Sample 1. (Online version in color.)
| Element | CaO | SiO2 | Al2O3 | MgO | CaS | TiO2 | Fe2O3 | K2O | ZnO | MnO | PbO |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Sample 1 | 37.95 | 32.39 | 14.32 | 6.97 | 4.30 | 1.46 | 1.041 | 0.54 | 0.52 | 0.36 | 0.15 |
| Final slag | 38.76 | 33.37 | 15.03 | 9.99 | 1.30 | 0.99 | – | 0.24 | 0.21 | – | – |
The SEM image (secondary electron) of the left side (near carbon brick side) of sample 1 is also shown in Fig. 3(a), reflecting the apparent morphology of the slag skull. The slag structure is not dense, and a large number of pores are distributed on the slag matrix. These irregular pores are disorderly distributed in the slag skull. The pore diameters vary widely, with larger pore diameters exceeding 400 μm and smaller pore diameters are not exceed 20 μm. Some of the pores are filled with white and light gray phases. Using the Photoshop software to binarize the pores and slag in the electron microscope photograph, as shown in Fig. 3(b). The ratio of the pores in the region was calculated to be 19.17%. Figure 3(c) is an SEM image under the condition of 500 times of pores, and the interface of the harmful element and the slag matrix in the figure is enlarged to 2500 times (Fig. 3(d)), and a layer of powdered layer is found at the interface. It can be seen from the distribution of the elements characterized by the mapping image (Fig. 3(e)) that in addition to the elements Ca, Mg, Al, Si, the powdered layer also aggregates Na and Zn at the interface. Element K is dispersed in the slag skull. K has a smaller ionic radius and larger mass diffusion coefficient than Na, and has a higher penetrating power at high temperatures. It is easier to infiltrate into the slag skull from the hot surface of the slag skull under the conditions of high temperature and high pressure in hearth. Reacts with a larger range of slag matrix to form potassium nepheline or leucite. XRD patterns detected Ca2ZnSi2O7 indicating that zinc can be reflected with the oxides in the slag, so that the original dense slag matrix structure is destroyed. After slag matrix pulverization, it is more likely to react with harmful elements due to the increase of surface area. Alkali and zinc vapor are circulated and enriched in hearth. Under the condition of high temperature and high pressure of the BF, the microcracks and pores along the hot surface of the slag skull are expansion, thus causing certain instability of the slag skull.
Pores distribution and microscopic morphology of slag skull. (Online version in color.)
Figure 4 shows the viscosity of the slag skull and final slag as a function of temperature. It can be seen that the viscosity decreases smoothly with increasing temperature. The viscosity of final slag is lower than that of slag skull. The dotted line indicates the solid friction in the melting slag. With the slag temperature decreasing, the solid phase precipitates as Ca2Al2SiO7 and Ca2Al2SiO7, which is confirmed with the XRD results. The solid precipitation temperature of slag skull is 1424°C while that of final slag is 1408°C. So the slag skull is much more prone to deposit on the hot surface of carbon brick. Furthermore, the solid phase of CaS in the slag skull is much higher than the final slag, which makes the slag skull much more viscous. The stable slag skull forming in brick heating surface of BF hearth, which avoided direct contact by separating molten iron with carbon brick, is a necessary condition to keep the longevity of hearth and delay the corrosion of brick.
Viscosity and solid phase friction change of the SiO2–CaO–MgO–Al2O3 slag with temperature. (Online version in color.)
The thermal conductivity of the slag skull is a vital parameter for the heat transfer calculation of the hearth, which is also an indispensable parameter for many carbon brick residual thickness prediction models. The laser method was employed to measure the thermal diffusivity of the sample with 10 mm in diameter and 5–10 mm in thickness. The thermal conductivity was calculated based on the one dimensional transient heat flow equation. The thermal conductivity of the slag skull is up to 1.9 W/(m∙K) at 200°C and 2.1 W/(m∙K) at 800°C, which is capable of meeting the requirement of the BF hearth and preventing heat from being taken away by cooling water. Figure 5 shows the hearth heat flux changes with the thickness of slag skull when the carbon brick is under different thickness conditions.15) The heat flux decrease with the thickness of slag skull as well as the thickness of carbon brick. In this investigation, the residual carbon brick is about 200 mm and the slag skull is also about 200 mm. The heat flux is calculated as 10000 W/m2 by heat transfer calculation, which is consistent with the actual measurement of heat flow intensity in BF hearth. Therefore, the experimentally measured thermal conductivity of the slag skull is predictable for the model of residual thickness of the furnace carbon brick.
Hearth heat flux changes with the thickness of slag skull at different thickness of carbon brick. (Online version in color.)
Usually, most researchers believe that the slag scaffold will be formed below the centerline of the BF hearth taphole, and its compositions and performances are consistent with the slag scaffold above the center line of the taphole.17) So the sample 2 located below the taphole was taken and it is a typical representative sample. A large amount of C (black phases) are found in the sample 2, as shown in Fig. 6(a). The white phases are Fe. Fe and C present layer by layer. The XRD pattern of the sample 2 is shown in Fig. 6(b). In the figure, the distinct characteristic peaks in the XRD pattern are indexed as Fe and graphite. Figure 6(c) shows the microscopic characterization of the graphite deposits found in the BF hearth. It is judged by mineralogical characteristics that these black phases are graphite. The graphite in the deposit layer exhibits an aggregate form formed by the accumulation of many tiny scales. There is a certain orientation between the scales, and some gullies are distributed in the graphite phase, as shown in Fig. 6(c). Figure 6(d) is an optical image of a thick strip of graphite. The strip graphite is not dense and exhibits cracks. The interface between graphite and iron can be seen that much dendritic graphite are precipitated from the iron phase.
SEM image, XRD and Optical microgragh of Sample 2. (Online version in color.)
Since the graphite is insoluble in hydrochloric acid, the three-dimensional topography of graphite in the deposited layer of the sample 2 can be observed after etching with hydrochloric acid and the secondary electron image is as shown in Fig. 7. The graphite phase is well preserved. The area covered by iron in the two-dimensional morphology becomes a gully, and the three-dimensional shape shows a more comprehensive feature of the graphite deposit. The graphite deposit is formed by the orderly arrangement of thick worm-like graphite, and the length of these structures is about 250 μm. Since the sample is etched in hydrochloric acid after being cut and flattened, the graphite present as worm-like with a cross-sectional morphology (Fig. 7(a)). There are many smaller dendritic structures between these stout graphite phases (Fig. 7(b)). The results of the partial enlargement are shown in Fig. 7(c). The length is about 50–100 μm. The Raman analysis confirmed that the phase of the dendritic structure was mainly the graphite phase (Fig. 7(d)).21,22) As we know, a peak near 1350 cm−1 is D band and a peak near 1590 cm−1 is G band. The G band present as the highly graphitic structures.
Three-dimensional morphology of the sample 2 after being eroded by hydrochloric acid.
During the pickling process, some stout worm-like graphite are easily detached from the sample. After the detached graphite phase is washed and dried, the microstructure is observed. Since these graphite are not polished, the morphology is reflected. The features may be actual topography or fracture morphology, so the structural characteristics of these worm-like graphite phases can be more realistically reflected. The three-dimensional morphology of the worm-like graphite phase is a sheet-like structure (Fig. 8(a)), because the dendritic structure is not tightly connected to the flake graphite, and most of these dendritic structures have been detached after ultrasonic cleaning. The presence of a layered structure is easily seen on a large piece of flake graphite (Fig. 8(b)). Figure 8(c) shows the fracture morphology of flaky graphite with 5000 times, and the graphite layers are arranged neatly. This structure is liable to cause slippage of flake graphite. Under the action of external force, the graphite phase in the graphite deposit layer produces relative sliding between the plane layers, and some scales appear superimposed, and some scales may be crushed. This process implies the formation of graphite deposit. The graphite phase first precipitates from the molten iron and the crystal grows continuously, and then it undergoes stress during the solidification process of the iron, resulting in fracture or slippage of the graphite sheet (Fig. 8(d)).
Three-dimensional morphology of the detached part of sample 2 after being eroded by hydrochloric acid.
Besides, the graphite phase was ground to the nanometer scale, and a microscopic topography of graphite in the deposit is observed by transmission electron microscopy (TEM). The finely ground graphite sample exhibits a collection of tiny scales in the TEM (Fig. 9(a)). Figure 9(b) is a relatively independent scale. Lattice fringe image (Fig. 9(c)) can be directly obtained through high resolution maps. The results of the electron diffraction analysis are shown in Fig. 9(d). Since the electron beam is incident perpendicular to the direction of the crystalline graphite scale layer, the diffraction pattern obtained are mainly diffraction spots. The results show that there are two hexagonal diffraction patterns, combined with XRD analysis results, indicating that the diffraction pattern is two layers of graphite. The TEM results confirm that the precipitated phase in the widely distributed sedimentary layer in BF hearth is graphite nor slag. Combined with the characterization of the three-dimensional morphology, it indicates that these graphite phases are precipitated from molten iron and grown gradually.
TEM cross section micrograph of the graphite deposit layer.
Based on the actual sampling of BF hearth, the samples above and below the taphole are analyzed separately. The main conclusions are as follows:
(1) The main component of slag skull is similar to the final slag above the taphole and is rich in harmful elements. The slag matrix component exists in the form of Ca2Mg(Si2O7) and Ca2Al(AlSi)O7, and the harmful elements are mainly in the form of KAlSi2O6 and Ca2ZnSi2O7. Harmful elements penetrate into the interface between the carbon brick and the slag skull through the pores channels, destroying the carbon brick.
(2) The thermal conductivity of the slag skull is about 2 W/(m.K) and the viscosity as well as solidus temperature are higher than the final slag. The actual survey results are consistent with the heat flux exhibited by the hearth in the production, and the slag skull acts to isolate and contain harmful elements.
(3) The phase on the hot surface of the carbon brick below the taphole is mainly graphite. The graphite in the sedimentary layer is mainly composed of a sheet-like structure, and a large number of dendritic structures are extended on the surface of the sheet-like structure, and the large-grained graphite phase has a random spatial network distribution in the iron matrix. The results of TEM and Raman once again confirm that the deposits widely distributed below the taphole in BF hearth are mainly graphite.
This work was financially supported by the National Science Foundation for Young Scientists of China (51604178, 51704019) and the Young Elite Scientists Sponsorship Program by CAST(2018QNRC001).
