Fatigue Strength of Ferritic Ductile Cast Iron Hardened by Super Rapid Induction Heating and Quenching

Transcription

1 Materials Transactions, Vol. 45, No. 9 (4) pp. 293 to 2935 #4 The Japan Institute of Metals EXPRESS REGULAR ARTICLE Fatigue Strength of Ferritic Ductile Cast Iron Hardened by Super Rapid Induction Heating and Quenching Yoshitaka Misaka 1, Kazuhiro Kawasaki 1, Jun Komotori 2 and Masao Shimizu 2 1 NETUREN Co.,Ltd., Hiratsuka , Japan 2 Department of Mechanical Engineering, Keio University, Yokohama , Japan To clarify the effects of Super Rapid Induction Heating and Quenching (SRIHQ) on fatigue properties of Ferrite Ductile Cast Iron (FDI), rotational bending fatigue tests were carried out on specimens treated with four types of heating cycle. Results showed that; (i) the SRIHQ process generated a thin dark gray area around the graphite. This dark area was composed of a martensite structure (ringed martensite). (ii) The ringed martensite generated a compressive residual stress field at the surface hardened layer. Two types of compressive residual stress generative mechanisms were observed. One was a microscopic residual stress generative process due to the formation of ringed martensite and the other was a macroscopic residual stress generative process due to the expansion of the surface hardened layer. (iii) The fatigue strength of SRIHQ treated FDI specimen was higher than that of the untreated one. This was because the compressive residual stress field generated by the ringed martensite suppressed initiation and propagation of fatigue cracks. (Received June 21, 4; Accepted July 23, 4) Keywords: rotational bending fatigue, fracture mechanism, super-rapid induction heating and quenching, ferritic ductile cast iron 1. Introduction Ductile cast iron with spherical graphite is now widely noticed as one important structural material, because it is inexpensive and can easily form relatively complicated shaped parts. One weak point of the material from a view point of strength is the existence of graphite; in Ferritic Ductile Cast Iron (FDI), fatigue cracks initiating from graphite lead to the final fracture of the specimen. It is very important to develop a new surface modification process to improve the fatigue strength of FDI. 1,2) In recent years, the development of a new device that controls heating time precisely and of high output induction heating equipment, make it possible to heat a steel work surface to a transformation temperature in less than.5 seconds. When this process is applied to small gears, the results are contour hardening, high hardness, high compressive residual stress and low distortion are achieved. 3 6) The order to put this Super Rapid Induction Heating and Quenching (SRIHQ) process to wide practical use, it is very important to understand the fatigue properties of steel modified by the process. In our previous study, 4) high cycle fatigue properties of surface hardened steel specimens, having different thickness of hardened layers, were examined. As a result, it was concluded that the SRIHQ process can induce an extremely high compressive residual stress field on the surface layer. This compressive residual stress improves fatigue properties of treated specimens. When a specimen had small notches, like low module gears; extremely high fatigue strength was observed. This implies that this process greatly affects the improvement of fatigue strength of machine parts which have sharp notches. The aims of the present study are to clarify the effects of the SRIHQ process on fatigue strength and fracture mechanism of FDI, and to discuss the applicability of SRIHQ as a surface modification process for FDI. Table 1 Chemical compositions and mechanical properties of the alloy used. mass(%) C Si Mn P S Mg Cu FCD T.S (MPa) Y.S (MPa) El. (%) FCD T.S: Tensile strength Y.S: Yield strength El.: Elongation 2. Experimental Method The material used in this study was FDI with a tensile strength of about MPa. Chemical compositions and mechanical properties are shown in Table 1. Simple cylindrical bars were taken from the lower part of the Y-shaped cast molds in order to lessen a scatter of cast defects. Then they were machined into a fatigue specimen as shown in Fig. 1. Using SRIHQ system, the specimens were quenched at four different heating cycles as shown in Fig. 2. In this study, to achieve different microstructures, we fixed φ 16 Fig. 1 R2 18 Specimen configuration. 9 φ 7

2 Fatigue Strength of Ferritic Ductile Cast Iron Hardened by Super Rapid Induction Heating and Quenching Temperature, T/K Fig. 2 Time, t/s t 1 Water Quenching Series A B C D t 1 (s) Heating cycle for Super-Rapid Induction Heating and Quenching. heating temperature at 1248 K and changed heating rate as shown in Fig. 2. We refer to these samples as series A(.13), B(.37), C(1.15) and D(7.3), respectively. The distribution of the Micro-Vickers hardness was measured with an indentation load of 2.94 N at the transverse sections of the smallest diameters. After quenching, residual stress in the radial parts of the specimen surfaces was also measured by using a micro area X-ray stress analyzer. The surfaces of the specimens were ground by emery papers and polished electrochemically. Rotational bending fatigue tests were then performed at room temperature with a cyclic applied stress frequency of 5 Hz. The maximum stress at which no fracture occurred after applying 1 7 cycles of stress, was defined as the fatigue strength in this study. To examine the fracture mechanism, successive observations of surface crack initiation and its propagation behavior were performed by a video microscope with a magnification of. The fracture surfaces were also observed using a Scanning Electron Microscope (SEM). 3. Results and Discussion 3.1 Microstructure of SRIHQ treated FDI Figure 3 shows the microstructures of the untreated (a) 5 µ m (b) Series A(.13) (c) Series B(.37) (e) Series D(7.3) Fig. 3 Optical microstructures.

3 2932 Y. Misaka, K. Kawasaki, J. Komotori and M. Shimizu Thickness of Dark Gray Area, d/µ m Heating Time,t/s Fig. 4 Relation between heating time and thickness of dark gray area. 2 Ferrite Martensite Carbon 1 µ m Martensite Ferrite specimen and the treated specimens. A thin dark gray area was observed around the graphite in all series of the SRIHQ treated specimens. The thickness of this area increased as the heating time became longer, as shown in Fig. 4. To investigate the formation mechanism of the dark gray area, the carbon distribution around the graphite was measured by a line analysis with X-ray micro analyzer. Figure 5 shows a typical example of the results. The amount of carbon in the dark gray area was greater than that of the ferrite matrix (see middle figure in Fig. 5). This is because the carbon in the graphite dissolved into the ferrite phase around the graphite. Because the dark gray area was thin, microvickers hardness of the area was also measured with an indentation load of 9.8 mn. As a result, it became clear that the hardness of the area itself was 85 to 89 HV. These results suggest that the dark gray areas observed in the treated specimens are martensite structures generated by the SRIHQ process. We will call this martensite, ringed martensite. Figure 6 shows distributions of hardness of the specimens at the transverse sections of the smallest diameters. In Figs. 6(b) (e), the dotted line represents hardness levels of the untreated specimens (Fig. 6(a)). Although no clear hardened layers were observed in the SRIHQ treated FDI, each specimen had an increase of about to HV in the hardened layer. The hardened layer became deeper as the heating time became longer. This increase of hardness is due to the formation the ringed martensite mentioned above. Next, the residual stress in the axial direction on the surface of the specimens was measured. The results are shown in Fig. 7. In our previous study using structural steel, a very high compressive residual stress of more than MPa was observed. 4) However, in the case of the FDI, a relatively lower compressive residual stress was observed. The compressive residual stress generated in series B and C was higher than that of series A, which was almost the same as that of series D. Usually, residual stress due to induction hardening is generated from anisotropy between the volume of the surface hardened layer (martensite transformed layer) and the non-transformed matrix. 7) In the case of the FDI treated by SRIHQ, there are two possible types of formation of residual stress. One is a macroscopic residual stress generative process mentioned above, and the other is a X-ray intensity, I/s 1 X-ray intensity, I/s Fig. 5 Line analysis Plane analysis (Mapping) Carbon distribution (Series B(.37)). microscopic residual stress due to the formation of ringed martensite. Figure 8 shows a schematic illustration that explains the macroscopic and microscopic residual stress generative processes. The upper figure shows that the macroscopic expansion of the surface hardened layer generated a compressive residual stress field in the surface layer; the lower figure shows that the microscopic expansion around the graphite generated a localized compressive residual stress field around the graphite. In the case of series B and C, both microscopic and macroscopic residual stress fields generated a relatively

4 Fatigue Strength of Ferritic Ductile Cast Iron Hardened by Super Rapid Induction Heating and Quenching 2933 (a) Macroscopic expansion due to martensite transformation Matrix (a) Macroscopic residual stress generative process (b) Series A(.13) (c) Series B(.37) Graphite Ringed martensite (b) Microscopic residual stress generative process (e) Series D(7.3) Fig. 6 Microhardness distributions. Fig. 8 Schematic illustration of the macroscopic and microscopic residual stress generative process. Stress Amplitude, σ /MPa 7 Series(Heating time,s) A(.13) B(.37) C(1.15) D(7.3) Compressive Residual Stress, σ /MPa (a) (b) Series A(.13) (c) Series B(.37) (e) Series D(7.3) Fig. 7 Relation between heating time and residual stress. Fig Number of cycle to Failure, Nf Result of rotational bending fatigue tests. higher compressive residual stress field. The lower residual stress in series A was caused by the fact that its ringed martensite thickness was thinner than that of series B and C. In addition, the fact that Series D had lower residual stress seems to be due to a lesser effect of the macroscopic residual stress on the compressive residual stress generative process. 3.2 Effects of SRIHQ on fatigue properties of FDI Figure 9 shows the results of rotational bending fatigue tests. The fatigue strength of the SRIHQ treated specimens was much higher than that of the untreated specimens. Series A to C showed an increase in the fatigue strength as the heating time became longer, which caused a thicker ringed martensite phase, resulting in the generation of a higher compressive residual stress field. This compressive residual

5 2934 Y. Misaka, K. Kawasaki, J. Komotori and M. Shimizu σ a=322mpa Nf= µ m µ m µ m Fig µ m Typical feature of crack propagation behavior (SeriesB(.37)). (a) µ m µ m Series B(.37) σ a=mpa Nf= µ m Fig. 1 Typical feature of crack propagation behavior. stress contributed to an increase in the resistance to the initiation and propagation of fatigue cracks. On the other hand, the fatigue strength of series D was less than that of series B and C. A possible reason is that the initiation and propagation of cracks was accelerated slightly because of a lower compressive residual stress. In order to clarify the influence of the ringed martensite formed by SRIHQ on the fatigue fracture mechanism, crack initiation and its propagation behavior were observed by a video microscope. Figure 1 shows a typical example of the results of the observation of the untreated specimen and those of series B, which had the highest fatigue strength. In the case of the untreated specimen, the crack starting point was located around the graphite and the crack propagated through the graphite. In series B, it was found that cracks generated from the ferrite phase, propagated avoiding the ringed martensite (see Fig. 11) The fracture surfaces were observed by SEM. Figure 12 shows the results of the untreated specimen and those of series B. The graphite areas on the fracture surface were different (see Fig. 13); the graphite area in the untreated specimen was much greater than that of the SRIHQ treated specimens. These results imply that the main fatigue crack propagation behavior was strongly affected by the compressive residual stress field generated by the ringed martensite. Finally, to discuss the fatigue crack propagation behavior quantitatively, crack propagation rates were measured at an (b) Series B(.37) Fig. 12 µ m µ m SEM fractographs of the fracture surface. applied stress of 1% higher than the fatigue strength. Figure 14 shows the relationship between crack propagation rate and K. Compared to the untreated one, the crack propagation rate in all SRIHQ treated series was lower. The slope of the untreated specimen was larger than that of the hardened specimens. These results suggest that SRIHQ works effectively to reduce the crack propagation rate of FDI.

6 Fatigue Strength of Ferritic Ductile Cast Iron Hardened by Super Rapid Induction Heating and Quenching 2935 Graphite percent in Fracture Surface (%) (a) (b) Series A(.13) (c) Series B(.37) (e) Series D(7.3) Crack Propagagtion,da/dN /mm cyc Fig. 14 Series(Heating time,s) A(.13) B(.37) C(1.15) D(7.3) 1 Crack Intensity Range, K/MPa mm1/2 Relation between crack propagation rate and K. Fig Conclusion Graphite percent in fracture surface. In order to clarify the effect of Super Rapid Induction Heating and Quenching (SRIHQ) on fatigue properties of Ferrite Ductile Cast Iron (FDI), rotational bending fatigue tests were carried out on specimens treated with four types of heating cycle. The results are summarized as follows: (1) The SRIHQ process generated a thin dark gray area around the graphite. This dark area was composed of martensite structure (ringed martensite). (2) The ringed martensite generated compressive residual stress fields at the surface hardened layer. Two types of compressive residual stress generative mechanism were observed. One was a microscopic residual stress generative process due to the formation of ringed martensite and the other was macroscopic residual stress generative process due to expansion of the surface hardened layer. (3) Fatigue strength of the SRIHQ treated FDI specimen was higher than that of the untreated one. This was because compressive residual stress fields generated by ringed martensite suppressed the propagation of the fatigue cracks. Acknowledgements The authors would like to thank Mr. Ishizawa at Hitachi Metals Ltd. who supplied the materials and valuable advice in preparing this paper. REFERENCES 1) L. Bartosiewicz, A. R. Krause, B. Kovacs and S. K. Putatunda: Trans. Am. Foundrymen Soc. (1992) ) P. Heuler, M. Hueck and H. Walter: Konstr Giess 17(3), (1992) ) J. Storm and M. Chaplin: Heat Treat (1987) 3. 4) Y. Misaka, Y. Kiyosawa, K. Kawasaki and T. Yamazaki: SAE Technical Paper Series (1997) ) Y. Misaka, Y. Kiyosawa, K. Kawasaki and T. Yamazaki: The 2nd Asian conference on heat treatment of material (1) ) J. Komtori, M. Shimizu, Y. Misaka and K. Kawasaki: Int. J. Fatigue 23 (1) ) George E. Totten, Maurice A. H. Howes: Steel Heat Treatment, (Marcel Dekker, Inc, 1997)