Formation of Black Striped Oxide Scale on Hot-Rolled Si-Containing Carbon Steel
2018 年 59 巻 11 号 p. 1716-1722
詳細
2018 年 59 巻 11 号 p. 1716-1722
The black striped oxide scale formed on silicon (Si)-containing hot-rolled carbon steel sheets was investigated. The small black stripes of oxide scale became band shaped as the flow used in mill cooling was increased. The distance between the black oxide scale bands was the same as that of the overlapping areas in the hydraulic descaling system. The average thickness of the black striped oxide scale was larger than that of the normal scale region. No ferric oxide (Fe2O3) was observed in the scale, particularly in the fractured parts of the black striped oxide scale. Furthermore, Si did not obviously accumulate in the black striped oxide scale, indicating that the oxide scale was not induced by the residual secondary scale generated during rough milling. The cooling conditions in finish milling affected the deformability of the scale in the overlapping areas. Consequently, the difference in thickness between the overlapping areas and adjacent regions led to the formation of the black striped oxide scale.
The demand for high-strength steel for energy-saving purposes has recently been increasing because the weight of the product can be decreased. Solid solution hardening is a popular way to improve steel strength because other properties are not greatly affected by this method. Silicon (Si) is a common alloying element for solid solution hardening from not only the above-mentioned viewpoint but also the aspect of cost performance. However, the addition of Si usually leads to the formation of red scale on the hot-rolled steel, thereby deteriorating the quality and value of the final product. For Si-containing steel, Fe2SiO4 is formed at the scale/substrate interface through the reaction between FeO and SiO2.1,2) Methods to prevent red scale formation3,4) and the effect of the formation of Fe2SiO4 on descalibility5–7) have been studied.
The presence of red scale on Si-containing hot-rolled steel sheets is attributed to insufficient descaling capability during hot rolling.8–10) When reheating Si-containing steel in a furnace, the eutectic FeO/Fe2SiO4 compound formed at the FeO/substrate interface will run irregularly into the FeO layer and substrate. If the descaling is conducted at a temperature below the melting point of the eutectic FeO/Fe2SiO4 compound, the strength of the compound increases and the upper FeO scale is difficult to remove completely because of an anchoring effect. The remaining FeO scale will be fractured by the subsequent rolling and converted into Fe3O4 and then red Fe2O3.8)
Liu and coworkers studied the black striped oxide scale formed in 2.2 mass% Si hot-rolled steel sheets at temperatures ranging from 1,000 to 1,150°C.10) The formation of inner scale (composed of FeO + Fe2SiO4) increased the bonding strength between the scale and substrate, thereby leading to poor descalibility. The remaining oxide was red Fe2O3 generated by oxidation of FeO. Furthermore, accumulation of Si was observed at the outer side of the oxide scale. When the temperature is below the eutectic temperature of the FeO/Fe2SiO4 compound, Si-rich oxides will not penetrate the oxide scale because the Fe2SiO4 compound is not miscible with FeO and the Si atoms do not diffuse in the oxide scale.11) It was found that some residual Si-containing oxides were left on the surface of the steel sheet after rolling. During slow cooling after coiling, the red Fe2O3 turned black because Fe2O3 was consumed to support the oxidation of the substrate in the coil.12) In the above cases, the striped oxide scale originated from the primary and secondary oxide scale, where the primary oxide scale is that formed during reheating and the secondary oxide scale is produced during rough rolling.
In the present study, the formation of the striped oxide scale on quenched steel sheets by hot roll processing under different conditions is investigated. The fabricated steel sheets containing striped oxide scale are analyzed by scanning electron microscopy (SEM), X-ray diffraction (XRD), transmission electron microscopy (TEM), and glow discharge spectrometry (GDS). Furthermore, an alternative mechanism for the formation of the black striped oxide scale is proposed.
Carbon steel slabs with a thickness of 270 mm were used as the starting materials. A system with a one-stand rough mill and six-stand finish mill was used to conduct the hot rolling process, as shown in Fig. 1. Steel 1 (0.20 mass% Si) and Steel 2 (0.02 mass% Si) were used as the carbon steel specimens. The reheating temperature was 1,230°C. After descaling the primary oxide scale produced in the furnace, each slab was milled to a thickness of 32 mm using the rough mill. A specimen was cut from the thin slab (called the transfer bar) and quenched in water to keep the oxide scale in its original status. Before finish milling, the secondary oxide scale on the transfer bar was removed using a hydraulic descaling system with a water pressure of 22 MPa. The starting temperature of finish milling was 1,030°C. To simplify the experimental conditions, only the mill cooling system between the first and second stands was used. Mill cooling flows of 5%, 10%, and 15% were used. The finishing temperature used in finish milling ranged from about 860 to 870°C. The final thickness of the steel sheet was 2 mm. A specimen was cut immediately after coiling and quenched in water to keep the oxide scale in its original status.
Schematic diagram of the hot rolling process.
The samples were ultrasonically cleaned in acetone and then air-dried. The microstructure of the oxide scale was observed by SEM (Hitachi, 6340, Tokyo, Japan). The phase constitution of the oxide scale was identified by XRD (Bruker, D8 Advanced, Massachusetts, United States). Samples for TEM analysis were prepared using a focused ion beam (FEI, Nova-200 NanoLab, Oregon, United States). A transmission electron microscope (FEI, Titan-20, Oregon, United States) was used to conduct the TEM analysis. The depth profile of the oxide scale was analyzed using GDS (RF GD-Profiler HR, Horiba Scientific Co., Kyoto, Japan).
Figure 2 shows the surface morphologies of the steel sheets for Steel 1 obtained under different mill cooling conditions. For the sample obtained from 5% mill cooling, small black striped oxide scale with a width range of 1 to 5 mm formed on the steel surface, as shown in Fig. 2(a). The size of the black striped oxide scale increased with the intensity of mill cooling. Figure 2(c) shows the oxide scale formed under 15% mill cooling flow conditions. The black striped oxide scale became band shaped, with a width of approximately 5–20 mm. The distance between the centers of the regions with the black striped oxide scale was nearly 85 mm for all samples, as indicated by white dashed lines. Furthermore, for the sample obtained after 15% mill cooling, some small black striped oxide scale was distributed in the regions between the band-shaped oxide scales. Almost no black striped oxide scale was observed on Steel 2 exposed to the same mill cooling conditions. Thus, the subsequent analyses focused on Steel 1.
Surface morphology of hot-rolled steel sheets obtained from Steel 1 after mill cooling at (a) 5%, (b) 10%, and (c) 15%.
Figure 3 shows SEM images of the surface morphologies of the normal scale region and black striped oxide scale for Steel 1. For the normal scale region, only a few light marks were observed in the macroscopic view, as shown in Fig. 3(a). Higher magnification revealed that parts of the scale were broken into particles, whereas others remained intact, as illustrated in Fig. 3(b). Figure 3(c) shows a low-magnification image of the surface morphology of the black striped oxide scale. Over 50% of the surface was covered by large-sized light marks. The scale in the large-sized light marks was entirely broken into particles with sizes of less than 2 µm, as displayed in Fig. 3(d).
SEM images showing the surface morphology of the hot-rolled steel sheet obtained from Steel 1. (a) Low-magnification image of the normal scale region, (b) high-magnification image of the normal scale region, (c) low-magnification image of the black striped oxide scale, and (d) high-magnification image of the black striped oxide scale.
Figure 4(a) shows the cross-sectional morphology of the scale in the normal scale region. The average thickness of the scale was approximately 4.8 µm. The scale was dense and its top surface was flat. The average thickness of the small black striped oxide scale was approximately 8.2 µm, which was larger than that of the normal scale region (Fig. 4(b)). On the top surface, some parts of the scale were crushed (indicated by white dashed lines), whereas others remained intact. On the whole, the small black striped oxide scale still maintained good integrity, except that the scale/substrate interface became wavier than that observed in the normal scale region. Figure 4(c) shows the cross-section morphology of the band-shaped black striped oxide scale. Except for in the smashed parts (indicated by the white dashed line), spallation of the scale occurred. The thickness of the scale in most of this region exceeded 11 µm and in some parts was over 20 µm. The distortion of the scale/substrate interface became severe because of the non-uniformity of the scale thickness compared with that observed for the small black striped oxide scale.
Cross-sectional profiles of the oxide scale formed on Steel 1. (a) Normal scale region, (b) small black striped oxide scale, and (c) band-shaped black striped oxide scale.
Figure 5 shows the XRD patterns of the oxide scale in the normal scale region and the small and band-shaped black striped oxide scale obtained from Steel 1. The width of the individual small black striped oxide scale ranged from 1 to 5 mm. The accuracy of the analysis was not influenced by the adjacent scale. All the types of scale were composed of Fe3O4, FeO, and α-Fe. The relative diffraction intensities of the products for the normal scale region and small black striped oxide scale were similar. For the band-shaped black striped oxide scale, the diffraction intensity of α-Fe was relatively low. This finding implies that the average thickness of the band-shaped black striped oxide scale might actually be larger than that observed in Fig. 4(c), thereby weakening the signal from the substrate.
XRD patterns of the oxide scale formed on Steel 1. (a) Normal scale region, (b) small black striped oxide scale, and (c) band-shaped black striped oxide scale.
Figure 6(a) shows a morphological image of the oxide scale at the scale/substrate interface of the transfer bar. According to the SEM/energy-dispersive X-ray spectroscopy analysis, the Si concentration of the dark substance ranged from 1.5 to 7 at%, which was much higher than that in the substrate. This result implies that certain compounds might form during rough milling. TEM was used to further investigate the structure of the oxide scale, as presented in Fig. 6(b). The sampling location is indicated by the white dashed line in Fig. 6(a). The atomic percentages of iron, silicon, and oxygen for some regions of the dark gray substance were about 22.11:12.83:60.16, respectively (1.72:1:4.69 if the silicon concentration is considered to be 1), which is close to the stoichiometry of Fe2SiO4. The diffraction pattern in Fig. 6(c) indicated that a few Fe2SiO4 particles existed in the regions with the dark gray substance. The compound beside Fe2SiO4 is FeO, as identified in Fig. 6(b) and (d). The upper part of the TEM image was mainly composed of FeO, as shown in Fig. 6(d). Figure 6(e) depicts the cross-sectional morphology of the oxide scale of the hot-rolled sheet obtained from Steel 1. The sampling method for the TEM specimen was similar to that for the transfer bar. Fe2SiO4 particles were also observed, although no obvious interlayer was formed at the scale/substrate interface, as illustrated in Fig. 6(h). The particle sizes of Fe2SiO4 at the scale/substrate interface ranged from 50 to 150 nm, which were smaller than those observed for the transfer bar.
Micrographs and TEM diffraction patterns of the oxide scale formed on the transfer bar and hot-rolled steel sheet obtained from Steel 1. (a) SEM and (b) TEM images of the scale/substrate interface on the transfer bar. (c) and (d) TEM diffraction patterns for the image in (b). (e) SEM and (f) TEM images of the scale/substrate interface on the hot-rolled sheet, respectively. (g) and (h) TEM diffraction patterns for the image in (f).
Figure 7 shows the depth profile of the oxide scale on Steel 1 and 2. Si enriched at the scale/substrate interface during hot rolling. Therefore, the location of the maximum Si concentration in the depth profile could also be used to estimate the oxide scale thickness, as indicated by the dashed lines. For the normal region on Steel 1 (Fig. 7(a)), the oxide scale was approximately 4.5 µm thick, which was close to the result obtained in Fig. 4(a) (4.8 µm). The maximum Si concentration at the scale/substrate interface (0.6 mass%) was three times higher than that in the substrate. Figure 7(b) shows the depth profile of the band-shaped black striped oxide scale on Steel 1. Si enrichment also occurred at the scale/substrate interface and its maximum Si concentration was about 0.5 mass%. The thickness of the oxide scale estimated from Fig. 7(b) was about 12 µm, corresponding to that obtained in Fig. 4(c). Figure 7(c) shows the depth profile of Steel 2. Si enrichment also occurred at the scale/substrate interface but with low maximum Si concentration (0.06 mass%) because the Si concentration of Steel 2 was only 0.02 mass%. The oxide scale in Steel 2 was around 6.5 µm thick. Mn behaved as a solid solution element in FeO. Therefore, Mn was distributed throughout the entire scale and no obvious Mn enrichment was observed at the scale/substrate interface.
Depth profiles of the oxide scale for (a) normal region on Stee1 1, (b) band-shaped black striped oxide scale on Steel 1, and (c) normal region on Steel 2.
In the present study, a large amount of Si-rich oxides still formed near the base of the secondary oxide scale on the transfer bar after rough milling even though the carbon steel contained only 0.2 mass% Si (Fig. 6(a)). If the secondary oxide scale was not removed, the Si-rich oxide residue would lead to the Si enrichment in the outer part of the oxide scale after finish milling. In the result of Fig. 7(b), however, Si enrichment only occurred at the scale/substrate interface in the black striped oxide scale on the final steel sheet. It indicates that no obvious secondary oxide scale was left on the surface of the steel sheet before finish milling. The FeO and Fe2SiO4 near the scale/substrate interface existed in the form of particles (Fig. 6). Therefore, the secondary scale could be removed easily before finish milling if the wedge-shaped scale had not formed (wedge-shaped scale is the FeO/Fe2SiO4 mixture that grew along the grain boundary of the FeO layer and steel).13,14) Furthermore, the XRD results in Fig. 5 revealed that the black striped oxide scale was composed of Fe3O4 and FeO. This composition should be close to the initial status of the black striped oxide scale because the steel sheet was rapidly cooled immediately after rolling. The fractured regions in the striped oxide scale in Fig. 3(c) and (d) were not red, revealing that Fe2O3 was absent because hematite turns red when its powder grain size is less than 2 µm.2) This observation implies that these striped oxide scale might form during the latter period of the milling process; that is, during finish milling. Consequently, the duration of the oxidation was insufficient to convert FeO into Fe2O3.
The mechanism for black striped oxide scale formation on the Si-bearing carbon steel sheet proposed based on the above results is depicted schematically in Fig. 8. Hydraulic descaling is generally conducted before a steel sheet runs into the roll bite of the first stand during finish milling. To eliminate all the scale on the surface of the steel sheet, a row of high-pressure nozzles is used for descaling. Therefore, avoiding an overlapping area is difficult.15) The overlapping area is overcooled and the temperature drop of the overlapping area can reach several tens of degrees Centigrade.16) The formation of band-shaped regions with low surface temperature along the rolling direction is expected, as shown in Fig. 8(a).
Diagrams illustrating the evolution of the oxide scale in the overcooled region. (a) Formation of the low-temperature regions in overlapping areas during hydraulic descaling, (b) further cooling of the low-temperature regions during mill cooling, (c), (d), (f), (g) temperature recovery of the oxide scale, (e) and (h) oxide scale morphology after rolling. (c) to (e) Steel 1 and (f) to (h) Steel 2.
After removal of the secondary oxide scale, a new oxide scale is formed before the steel sheet runs into the finish stands, as indicated in Fig. 1. Theoretically, the thickness of the newly formed oxide scale in the band-shaped region is slightly thinner because the temperature is lower. However, the results obtained in this study are opposite. Once the new oxide scale is formed, the interactions between the oxide scale and the plant’s components must be considered because of their direct contact. The temperature of the band-shaped regions may not be recovered completely when the steel sheet runs into the finish stands. The surface temperature will further decrease because of the contact with the work rolls of each stand and cooling by other systems. Through these interactions, a surface temperature drop of 100–200°C is possible.17) As mill cooling is conducted, the band-shaped regions with low surface temperature are further cooled, as depicted in Fig. 8(b). These behaviors are related to temperature; therefore, the heat transfer in the materials must be considered.
Figure 8(c) shows a cross-section view of the overcooled region in the oxide scale for Steel 1. When the steel sheet is cooled, the surface temperature can be recovered by the heat flow from the substrate. For a hot-rolled steel sheet, the compounds on the steel surface can greatly affect heat transfer. Si has a great affinity to oxygen. It will react with the dissolved oxygen on the metal surface to form the SiO2 compound during the high temperature oxidation. When newly formed FeO meets with SiO2, the formation of Fe2SiO4 is inevitable. The Fe2SiO4 compound is not miscible with FeO and the Si atoms do not diffuse in the oxide scale.11) Therefore, FeO/Fe2SiO4 mixtures only build up at the scale/substrate interface when a new oxide scale is formed on the Si-containing steel. The amount of Fe2SiO4 increases as the Si concentration of the substrate increases. According to a previous report,18) the thermal conductivity of Fe2SiO4 is only one third that of FeO. That is, Fe2SiO4 can hinder heat transfer.
During finish milling, the scale is mainly composed of newly formed FeO if the secondary scale is completely removed by the hydraulic descaling. FeO can deform plastically at 700°C, but its steady-state deformation is observed above 1,000°C.19) The starting temperature of finish milling usually ranges from 1,000 to 1,050°C. If the flow of mill cooling is high, then the surface temperature of the low-temperature regions will sharply decrease, by even as much as up to 700°C, because the magnitude of the temperature drop is proportional to the water flow rate.20) The results in Fig. 7 revealed that the Si concentration at the scale/substrate interface of Steel 1 was about ten times higher than that of Steel 2. It is expected that higher amount of Fe2SiO4 at the scale/substrate interface in Steel 1 retards the heat transfer from the substrate. Consequently, the temperature of the overcooled region in the oxide scale may not completely recover before the subsequent rolling, as shown in Fig. 8(d). The deformability of FeO decreases as the surface temperature decreases. When a rolling force is applied, the extension of the oxide scale with low deformability in the low-temperature region along the sheet plane is smaller than that of the adjacent region. Finally, the oxide scale with low deformation is rolled into the substrate to form the scale pits, as illustrated in Fig. 8(e). The formation of scale pits leads to the difference in oxide scale thickness between the overlapping areas and adjacent regions. When the flow of mill cooling is high, a large number of scale pits spread in the overlapping areas along the rolling direction, thereby forming a black band. The distance between the black bands was approximately 85 mm, which corresponds to that between the overlapping areas of the hydraulic descaling system. Parts of the outer region of the scale were fractured through the combination of rolling and friction because the temperature of the outer region is low and the scale is brittle.
For Steel 2, less Fe2SiO4 formed at the scale/substrate interface compared with the case for Steel 1 because of the lower Si content of Steel 2. As a result, heat transfer from the substrate to the oxide scale occurs relatively easily, as does as the temperature recovery of the overcooled region in the oxide scale, as shown in Fig. 8(g). Consequently, a uniform oxide scale is formed after rolling, as depicted in Fig. 8(h). Therefore, the formation of the black striped oxide scale depends on not only the conditions of the rolling process but also the oxidation behavior of the steel sheet before rolling.
The present study investigated the formation of black striped oxide scale on Si-containing carbon steel. The outer region of the black striped oxide scale did not contain enriched Si, indicating that the scale did not originate from the remaining secondary scale on the transfer bar. The small black striped oxide scale became band-shaped as the flow of mill cooling increased. The distance between the centers of the oxide scale bands was approximately 85 mm, which was the same as the distance between adjacent nozzles of the hydraulic descaling system. Before the steel sheet ran into the finish mill, hydraulic descaling was conducted to remove the secondary scale on the transfer bar, thereby leading to a drop of the surface temperature in the overlapping area between adjacent nozzles. These areas were further cooled in the finish stands because of the contact with the work rolls of each stand and cooling by other systems. When the flow of mill cooling was high, the deformability of the scale in the overlapping area may have decreased because the surface was overcooled. When a rolling force was applied, the oxide scale with low deformability was rolled into the substrate to form scale pits. The decrease in the deformability of the oxide scale induced by overcooling was concluded to be the main reason for the formation of the black striped oxide scale. This type of oxide scale is considered to have a different thickness to that in the adjacent regions on the steel surface.
The authors acknowledge the Center for Micro/Nano Science and Technology, National Cheng Kung University, Taiwan, for their support by providing a FIB for sample preparation and microstructure observation, and the support of the Department of Materials and Optoelectronics, National Sun Yat-sen University, Taiwan, for providing a TEM for microstructure examination. We thank Natasha Lundin, PhD, from Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.
