Surface Hot Shortness of Copper Containing Steel in a Compact Strip Production Process*

Transcription

1 Materials Transactions, Vol. 52, No. 10 (2011) pp to 1911 #2011 The Japan Society for Heat Treatment Surface Hot Shortness of Copper Containing Steel in a Compact Strip Production Process* Akihiro Takemura 1, Yusuke Ugawa 2, Kazutoshi Kunishige 3, Yasuhiro Tanaka 3, Shunichi Hashimoto 4 and Kazuya Ootsuka 5 1 Department of Mechanical Engineering, Faculty of Engineering, Tokyo University of Science, Yamaguchi, Sanyo Onoda , Japan 2 Sumitomo Metal Industries, Ltd., Amagasaki , Japan 3 Department of Advance Materials Science, Faculty of Engineering, Kagawa University, Takamatsu , Japan 4 Kobelco Research Institute Inc., Kobe , Japan 5 Nippon Steel Corporation, Futtsu , Japan The compact strip production (CSP) process has received much attention because it is environment friendly and can be used for recycling resources. However, steel scrap, the main material used in the CSP process, causes surface cracks in the hot strip being manufactured as a result of surface hot shortness by Cu and Sn in the steel scrap. This study investigates the influence of heat pattern on the brittleness of Cu-containing steels prior to treatment in a reheating tunnel furnace at 1100 C. The prior austenite grain size of a sample that was reheated from room temperature before being transferred to a tunnel furnace was finer than that obtained by the direct transfer process, and the crack depth was inhibited by 50%. In contrast, the prior austenite grain size of a sample obtained by a process in which austenite is reheated from a cooling stop temperature of 650 C 850 C was almost the same as that obtained by direct transfer to the tunnel furnace. However, the crack depth in the case of reheating from the cooling stop temperature of 650 C 850 C was greater than in the case of direct transfer. This deep cracking was caused by a noninternal oxidation area at the scale/steel interface. [doi: /matertrans.h-m ] (Received April 27, 2011; Accepted July 15, 2011; Published September 25, 2011) Keywords: copper-containing steel, surface hot shortness, compact strip production process, copper-enriched alloy, oxidized scale structure, iron-selective oxidation 1. Introduction The compact strip production (CSP) process is used to make hot-rolled coils. In this process, a continuous cast slab made from refined scrap in an electric furnace is sent directly to a tunnel furnace. This process has received much attention because it is environment friendly. 1) However, tramp elements such as Cu and Sn, which are present in steel scrap, cause surface cracks during hot rolling. 2) It is extremely difficult to remove Cu from steel scrap during the steel-refining process because Fe is more oxidizable than Cu. 3 5) Fe is selectively oxidized at the slab surface during heating, and the Cu-enriched layer formed at the scale/steel interface causes surface hot shortness. This Cu-enriched layer melts and penetrates into austenite grain boundaries because the melting point of this phase is relatively low. For these reasons, surface cracks appear as a result of liquid embrittlement during hot rolling. 6 9) Many studies have shown that Ni can suppress surface hot shortness. 6 8,10 12) One study found that conducting the oxidation process at a temperature higher than 1200 C alleviated surface hot shortness. 7 11,13) The addition of Cr is effective in suppressing surface hot shortness. 14) However, these solutions have their own problems. The addition of Ni increases production costs, whereas the use of temperatures greater 1200 C in the heating process increases fuel costs. We investigated the effect of shot peening on the suppression of surface hot shortness. 15) In this paper, we focus on the *This Paper was Originally Published in Japanese in Netsu Shori 50 (2010) influence of austenite grain size on surface hot shortness. In addition, the influence of the cooling stop temperature on austenite grain size and the suppression of surface hot shortness is discussed. 2. Experimental Procedure Table 1 shows the chemical composition of steel used in this study. This steel was produced in a vacuum melting furnace and contains 0.3%Cu 0.04%Sn as tramp elements. The amount of Ni was 0.02%, which is inevitably included in normal steels. The melt was divided into two batches. The amount of Nb in one batch was adjusted to 0.03%. We produced 100-mm-thick ingots from two types of steel. The ingots were then heated at 1200 C and rolled into 40- mm-thick plates by hot rolling at 900 C 1100 C. In this paper, surface hot shortness was investigated using various specimen shapes made of hot-rolled Cu-bearing low carbon steels. The specimens were compressed by Thermecmaster (Fuji- Denpa, Japan), a thermomechanical simulator, to investigate surface hot shortness. The transformation temperature was measured by Formaster (Fuji-Denpa, Japan), a thermal expansion testing machine. Figures 1(a) and (b) show the Thermecmaster and Formaster specimens, respectively. Table 1 Chemical composition of steel (mass%). C Si Mn P S Al Cu Sn Ni N

2 1906 A. Takemura et al. (a) (b) Fig. 1 Dimensions of specimens: (a) for compression test and (b) for measuring grain size of prior austenite and transformation temperatures. Fig. 3 Experiment to measure the transformation temperatures. interface of these specimens after oxidation was closely observed using an optical microscope. Fig. 2 CSP process laboratory simulation. 2.1 Simulation of CSP process Figure 2 illustrates the experimental procedure in which the CSP process is simulated using Thermecmaster. First, the sample was heated to 1300 C at a rate of 2.5 C/s in a vacuum. Subsequently, the sample was either heated directly in a tunnel furnace or cooled to a temperature between 950 C and room temperature, to investigate the influence of the cooling stop temperature on surface hot shortness during the CSP process. Hereafter, the heating process from 1300 C to the tunnel furnace condition is called the direct transfer process, the heating process once the sample cools is called hot-charged substance (HC), and the heating process once the sample cools to room temperature is called RT. The oxidation period in the tunnel furnace was 30 min at 1100 C in air. A compression test using Thermecmaster was conducted at 1100 C to investigate cracking caused by surface hot shortness. The strain rate was 0.01/s and the amount of deformation was 50% for the compression test. Mica glass foils, which are usually laid on the top and bottom of a sample for lubrication between the sample and piston heads of the compression test machine, did not use to enhance bulging. The compression test revealed that peripheral length of specimens showed a maximum expansion of approximately 40% 45% at the center of the height. Surface crack depth was measured with an optical microscope. In addition, oxidized specimens that had the same shape and were quenched before the compression test were obtained to investigate the causes of brittleness in Cu-bearing steel during the CSP process. The microstructure at the scale/steel 2.2 Measurement of prior austenite grain size and transformation temperature The Formaster specimens shown in Fig. 1(b) were obtained using a wire electric discharge machine to investigate the grain size of prior austenite. They were quenched in He gas in a vacuum before the compression test by using the same heat pattern as in the Thermecmaster process. They were galvanically corroded at approximately 8 V for 15 s in an electrolyte of 200 ml distillated water, 53.5 g chromium acid, and 80 g phosphoric acid, to expose the prior austenite grain boundaries. Then, the microstructures were observed using an optical microscope. The transformation temperature was measured using a Formaster test machine. The heat pattern is shown in Fig. 3. The heating and cooling speeds and the 15-min holding period at 1300 C used in heat treatment were the same as those for the simulation experiment during the CSP process. The transformation temperature during cooling was measured using these test specimens. 3. Results 3.1 Results of maximum crack depth Figure 4 shows the influence of the cooling stop temperature on the maximum crack depth at the surface of the specimens. This figure also shows the grain size of prior austenite before the compression test and the transformation temperature measured with the Formaster test machine. The maximum crack depth was 111 micrometers in the direct transfer process. These cracks increased with decrease in the cooling stop temperature to 750 C. The maximum crack depths in the 950 C, 850 C, and 750 C HC processes were 128, 210, and 319 micrometers, respectively. However, the maximum crack depth decreased to 245 micrometers in the 650 C HC process and 119 micrometers in the 500 C HC process. The crack depth in the RT process was 62 micrometers. This maximum crack depth in the RT process was 20% of that of the 750 C HC process and 50% that of the 500 C HC process. Thus, the maximum crack depth is understood to change greatly depending on the cooling stop

3 Surface Hot Shortness of Copper Containing Steel in a Compact Strip Production Process 1907 Figure 6 shows the relevance of the transformation temperature to the prior austenite grain size. Comparison of the direct transfer process and the RT process reveals that surface hot shortness in the RT process was suppressed and was influenced by the grain refinement of prior austenite. However, the reason for increase in the maximum crack depth at 750 C 850 C, where the grain sizes were almost the same, cannot be explained from the prior austenite grain size. 4. Discussion Fig. 4 Relationship between maximum crack depth, grain size of prior austenite and phase region where cooling stop temperature is located. temperature prior to reheating for 30 min oxidation at 1100 C, which imitates conditions in a tunnel furnace. 3.2 Results of measurement of prior austenite grain size and transformation temperature The metallographic structure after 30 min oxidation in air at 1100 C was observed in order to measure the prior austenite grain size. Figure 5 shows optical micrographs of the Formaster specimens that were quenched in He gas before the compression test. The prior austenite grain boundaries of the specimens were visualized by galvanic corrosion. The prior austenite grain size was 90 micrometers in the direct transfer process and micrometers in the 950 C 750 C HC process. However, no influence of the cooling stop temperature was observed on the prior austenite grain size in the 950 C 750 C HC process. In contrast, the boundary of the prior austenite grain size was refined under the 750 CHC process. The grain size was refined to 92 micrometers in the 650 C HC process. The prior austenite grain size in the 650 C HC process was the same as that in the direct transfer process and the refined grain sizes were 51 and 54 micrometers in the 500 C HC and RT processes, respectively. The grain sizes of prior austenite in the 500 C HC and RT processes were 50% that of the direct transfer process. The transformation temperature of the specimens was measured using the Formaster test machine. Figure 6 shows the thermal dilatation curve during cooling from 1300 C. The transformation temperature was determined by measuring thermal dilatation of the specimen. Dilatation started at 743 C, which shows the start of transformation from austenite to ferrite. It was finished at 570 C, which shows the completion of transformation. P f can be regarded as the Ar 1 point because the cooling rate is as low as 2.5 C/s. This result suggests that the specimens had an Ar 3 point at 743 C and an Ar 1 point at 570 C. In the temperature range from 743 C to 570 C, the metallographic structure was estimated to be a two-phase structure of ferrite and austenite. When the cooling stop temperature was 650 C or less, which is lower than the transformation temperature at 743 C, the prior austenite grain size was refined by the A 3 transformation. The prior austenite grain size was refined at the start of transformation and was refined significantly at a temperature of 500 C or less. 4.1 Refinement of prior austenite grain by cooling step before transfer to tunnel furnace and suppression of surface hot shortness The maximum crack depth was suppressed in the 500 C HC and RT processes. Figure 7 shows the mechanism of the refinement of the prior austenite grain size during the cooling and reheating steps. Transformation from austenite to ferrite during cooling and the reverse transformation from ferrite to austenite during reheating create fine grains. The matrix was austenite at 1300 C, which simulated slab solidification. The metallographic structure seems to be transformed from austenite to ferrite by cooling before heating at 1100 Cina simulated tunnel furnace. For the specimens used in this study, the transformation start temperature was 743 C, the transformation finish temperature was 570 C, and the heating and the cooling rate was 2.5 C/s. Thus, the thermal treatment period of the specimens in the HC and RT processes were longer than those in the direct transfer process, because in the HC and RT processes, the samples were cooled before transfer to a tunnel furnace and were then reheated. The specimens in the 950 C 750 C HC process were not cooled to a transformation temperature. For this reason, effect of grain refining by cooling and reheating before transfer to the tunnel furnace was not observed in these specimens. Therefore, it seems that the specimens in the 950 C 750 CHC process were coarsened by the long thermal treatment time. The refinement of grain size started in the 650 C HC process because of the following reasons: The specimens used here had an Ar 3 point at 743 C and an Ar 1 point at 570 C. It seems that the cooling to 650 C before transfer to the tunnel furnace transformed the austenite specimen to a two-phase structure of ferrite and austenite. A slight refinement of the prior austenite grain size was achieved by reverse transformation from partial ferrite. Specimens cooled below 500 C transformed austenite into ferrite. At this time, the ferrite phase was refined by transformation sources, i.e., the grain boundary of austenite and the triple point of the grain boundary. Then, a reverse transformation from ferrite to austenite occurred by reheating at 1100 C in a simulated tunnel furnace. The refinement of grain size caused a reverse transformation from the grain boundary of the ferrite phase and the triple point of the grain boundary. The grain size was more refined as a result of the transformation from austenite to ferrite and the reverse transformation from ferrite to austenite. The refinement of grain size is believed to improve the suppression of surface hot shortness. 7,8) Figure 8 shows schematic figures of the influence of grain size refinement on the suppression of surface hot shortness. The area of the

4 1908 A. Takemura et al. Fig. 5 Optical microstructures of the steels were subjected to each heat treatment changing the following cooling stop temperatures: (a) direct, (b) 950 C HC, (c) 850 C HC, (d) 750 C HC, (e) 650 C HC, (f) 500 C HC and (g) RT. grain boundary is increased by the refinement of the grain size, which was achieved by cooling before transfer and reheating in a tunnel furnace. If the Cu segregation at the scale/steel interface is independent of the grain size, the area of the grain boundary increases with decrease in grain size. Hence, Cu-enriched alloy at the scale/steel interface is dispersed and the depth of Cu penetration decreases. Therefore, surface hot shortness is suppressed by the refinement of grain size. Figure 9 shows the edge of hot- rolled 40-mm-thick steel that was finished at temperatures in excess of 900 C after being heated to 1200 C. Figure 9(a) shows interstitial free steel bearing 0.3%Cu 0.04%Sn, which is called Nb free steel. Figure 9(b) shows interstitial free steel bearing 0.3%Cu 0.04%Sn 0.03%Nb, which is called 0.03%Nb bearing steel. The prior austenite grain sizes in the Nb free steel and 0.03%Nb bearing steel were 154 and 102 micrometers, respectively. In case of 0.03%Nb bearing steel used in this study, Nb was completely melted at

5 Surface Hot Shortness of Copper Containing Steel in a Compact Strip Production Process Fig Thermal contraction curve on cooling. Fig. 9 Appearance of surface hot shortness in 0.30% Cu-0.04% Sn steels: (a) Nb free and (b) 0.03% Nb bearing. Fig. 7 Schematic illustration of the structural refinement through transformation. Fig. 8 Schematic illustrations of an enriched copper phase at the scale/ steel interface: (a) coarse grain and (b) fine grain C, explaining equilibrium solubility product.16) The prior austenite grain size is refined by the solute drag effect of the dissolved Nb.17) 4.2 Change in austenite grain size and surface crack depth at 650 C 850 C Figure 5 shows the prior austenite grain size in the 650 C 850 C HC process, which shows no change compared with the direct transfer process. However, the specimens in the 650 C 850 C HC process have large cracks with more than 200 micrometers depth. This suggests that there may be another cause that enhances surface hot shortness, in addition to the coarsening of the austenite grain. The microstructure of the scale/steel interface of the oxidized specimens was investigated in detail. A characteristic phase area that had no internal oxidation was found in samples at the cooling stop temperature of 850 C or less. Large oxidized particles surrounding this area were also found. This phase area increased with decrease in the cooling stop temperature. An example of this phase area is shown in Fig. 10. Hereafter, this phase area is called the noninternal oxidation area. The chemical composition of this area was analyzed using EDX. Fe and O characteristic X-ray peaks were detected. However, no enriched Cu or Sn phase was detected. Normal interstitial free steel was used in this study, which contains small amounts of Si and Mn. Si and Mn are more easily oxidized than Fe. A solute of Si and Mn as the solid solution was oxidized by O2 ion, diffused from air. This results in the internal oxidation of Si and Mn, which are more oxidizable than Fe at the initial stage of oxidation. In addition, the austenite grain size was refined by cooling before transfer to a tunnel furnace and reheating. In this manner, the grain boundary diffusion of O2 ion was accelerated during oxidation at 1100 C. Furthermore, decarburization occurred during oxidation at the same temperature.18) The internal oxidation of Si and Mn near the decarburized area and the oxidized grain boundary massed together. It seems that this decarburized area became a noninternal oxidation area. Figure 11 shows optical micrographs of a specimen from the 650 C HC process at the scale/steel interface. Cu-enriched alloy is observed in large amounts near the noninternal oxidation area. It seems that formation of the noninternal oxidation area, which increased internal oxidation around itself, accelerated surface hot shortness. Figure 12 shows the influence of the cooling stop temperature, the prior austenite grain size, and the noninternal oxidation area on the embrittlement of Cu-bearing steel during the CSP process.

6 1910 A. Takemura et al. Fig. 10 Optical microstructures at the scale/steel interface. Fig. 11 Optical microstructure at the scale/steel interface of the 650 C HC specimen. Fig. 12 Schematic diagram for the severity of cracking against cooling stop temperatures, being influenced by two factors of grain size of prior austenite and non internal oxidation area.

7 Surface Hot Shortness of Copper Containing Steel in a Compact Strip Production Process Conclusions Interstitial free steel bearing 0.30%Cu 0.04%Sn, the constituents of which were designed to imitate a standard electric furnace steel, was melted in an experimental procedure. Processing heat treatment that imitated the CSP process was applied using a thermomechanical simulator. Before transfer to a tunnel furnace, the cooling stop temperature that affects surface hot shortness was investigated. The parameter set for the tunnel furnace was a holding period of 30 min in air at 1100 C. The amount of deformation during a compression test at 1100 C was 50%. The surface crack depth at the center part of the height, which occurs by bulging, was examined and measured for surface hot shortness. The results obtained are as follows: (1) The surface crack depth in the RT process was the smallest. This crack depth was 50% smaller that in the direct transfer process. However, it was large in the 650 C 850 C HC process. (2) The results of measuring the transformation temperature showed that the Ar 3 and the Ar 1 point were 743 C and 570 C, respectively. The prior austenite grain size was coarse in the 650 C 950 C HC process. When the specimens were cooled to 500 C or less, the prior austenite grain size was refined by 50 micrometers, suppressing surface hot shortness. (3) The results of the examination of the scale/steel interface of the He-quenched oxidized specimens revealed a peculiar phase area that had no internal oxidation at the cooling stop temperature of 850 C or less. Large oxidized particles were found around this area. (4) It seems that the formation of a noninternal oxidation area, which increased internal oxidation around itself, accelerated surface hot shortness in the 650 C 850 CHC process. Acknowledgement The authors are greatly indebted to Dr. N. Yoshinaga and D. Maeda, Nippon Steel Corporation, for their discussions and valuable advice. REFERENCES 1) O. Umezawa: Bull. Iron Steel Inst. Jpn. 7 (2002) ) H. Yamana, Y. Mizuue, M. Takeuchi and K. Noro: Effects of Cu and Other Tramp Elements on Steel Properties, (ISIJ, 1997) pp ) N. Sano, H. Katayama, M. Sasabe and S. Matsuoka: Effects of Cu and Other Tramp Elements on Steel Properties, (ISIJ, 1997) pp ) M. Tanino: Tekko Zairyo no Kagaku (Science on Steel Properties), (Uchida Rokakuho Publishing Co., Ltd., Tokyo, Japan, 2001) pp ) Y. Saitou, T. Atake and T. Maruyama: JME the Journal of Materials Education, Kinzoku no Kouon Sanka (High-Temperature Oxidation of Metals), (Uchida Rokakuho Publishing Co., Ltd., Tokyo, Japan, 1982) pp ) L. Habraken and J. Lecomte-Beckers: Copper in Iron and Steel, ed. by I. L. May et al., (John Wiley & Sons, New York, 1982) pp ) K. Shibata: Netsu Shori 40 (2000) ) K. Shibata, K. Kunishige and M. Hatano: Bull. Iron Steel Inst. Jpn. 7 (2002) ) T. Kajitani, M. Wakoh, N. Tokumitsu, S. Ogibayashi and S. Mizoguchi: Tetsu-to-Hagané 81 (1995) ) N. Imai, N. Komatsubara and K. Kunishige: ISIJ Int. 37 (1997) ) N. Imai, N. Komatsubara and K. Kunishige: ISIJ Int. 37 (1997) ) M. Hatano and K. Kunishige: Tetsu-to-Hagané 89 (2003) ) M. Hatano, K. Kunishige and Y. Komizo: Tetsu-to-Hagané 88 (2002) ) M. Hatano and K. Kunishige: Tetsu-to-Hagané 90 (2004) ) A. Takemura, K. Kunishige, S. Okaguchi and K. Fujiwara: Tetsu-to- Hagané 95 (2009) ) K. J. Irvine, F. B. Pickering and T. Gladman: J. Iron Steel Inst. 205 (1967) ) T. Tsumura, Y. Kamada, S. Tanoue and H. Ohtani: Tetsu-to-Hagané 70 (1984) ) N. Birks and G. H. Meier: Introduction to High Temperature Oxidation of Metals, ed. by K. Nishida and T. Narita (Maruzen Publishing Co., Ltd., Tokyo, Japan, 1988) pp