Effect of Maximum Temperature on the Cyclic-Heating-Induced Embrittlement of High-Silicon Ferritic Spheroidal-Graphite Cast Iron

Transcription

1 Materials Transactions, Vol. 45, No. 2 (2004) pp. 569 to 576 #2004 The Japan Institute of Metals EXPRESS REGULAR ARTICLE Effect of Maximum Temperature on the Cyclic-Heating-Induced Embrittlement of High-Silicon Ferritic Spheroidal-Graphite Cast Iron Hung-Mao Lin* 1, Truan-Sheng Lui* 2 and Li-Hui Chen Department of Materials Science and Engineering, National Cheng Kung University, Tainan, Taiwan 701, R. O. China This study examines the effect of maximum temperature on cyclic-heating-induced embrittlement. Experimental results indicated a significant deterioration of tensile properties when the maximum heating temperature was 1023 K. Significant recrystallization of ferrite grains occurred when the heating temperature was raised to over 1073 K and this suppressed the embrittlement. Moreover, evidence of partial phase transformation was observed when the maximum heating temperature was raised to 1123 K. When a specimen was heated to 1023 K with a certain number of cycles, a distinct area fraction of intergranular fracture could be recognized from tensile fractography. The initiation site of these thermal cracks was at the prior-solidificational eutectic cell boundary of the ferrite matrix, which is related to magnesium-containing oxide inclusions. These oxide inclusions were analyzed and found to contain mainly magnesium, oxygen, phosphorus and cerium. It has been confirmed that the formation of magnesium-containing inclusions in the solidification process is mainly responsible for intergranular fracture, and consequently plastic deformation is promoted when the number of heating cycles is increased. From observation of the vicinity of crack initiation and propagation paths, significant slip evidence was found and was revealed using etch-pit techniques. In conclusion, the cyclic heating induced severe embrittlement at a specific maximum temperature of 1023 K. It is suggested that this resulted predominantly from periodic strain and stress accumulation of the inclusions which caused plastic deformation in the prior-solidification eutectic cell boundary region. (Received October 20, 2003; Accepted January 13, 2004) Keywords: cyclic-heating-induced embrittlement, maximum heating temperature, intergranular fracture, eutectic cell boundary region 1. Introduction Spheroidal graphite (SG) cast iron has been used to make heat-resistant components, such as guide plates for hot rolling equipment, aluminum casting related molds or exhaust manifolds etc. 1 6) High-silicon spheroidal graphite cast iron is considered to be one of the most promising candidate alloys for these components due to its castability, low oxidation, and good wear resistance. 7,8) In these cases, SG cast iron components periodically operate at elevated temperatures of up to 1123 K and, in particular, are subjected to cyclic thermal shocks from heating/cooling. 9) Thermal stresses or transformation stresses arising from a steep temperature gradient may cause fracturing or thermal fatigue cracking in these components. Some investigations agree that the thermal cracking behavior of SG cast irons may be related to the thermal stress arising from a steep temperature gradient for materials which are subjected to rapid heating and cooling cycles. 1,9,10) Our previous studies have confirmed the cyclic-heatinginduced deterioration of tensile properties which occurs in the presence of severe intergranular cracks. The cracks tend to initiate from the prior-solidification eutectic cell boundary region in which a fair amount of inclusion particles are clustered in the vicinity of the prior-eutectic cell boundary of the ferrite matrix. Weronki et al. 11) have shown how the grain boundary sulfide or oxide inclusions affect the thermal fatigue behavior of alloy steels that have suffered cyclic heating, and indicated that crack initiation occurs during these severe cyclic heating/cooling tests. However, thermal stress actually arises from a steep temperature gradient, while the hard heterogeneous phase also plays an important role in the development of thermal stress. As the maximum exposing * 1 Graduate Student, National Cheng-Kung University. * 2 Corresponding author, z @ .ncku.edu.tw Table 1 Chemical composition of the SG cast irons used in this study (mass%). C Si Mn P S Mg temperature is one of the important parameters in the investigation of thermal properties, this study focused on the frequently used heat resistant 4.0 mass% Si content SG cast iron at temperatures of up to 1123 K, to clarify the effect of maximum exposed temperature on the cyclic heating induced deterioration of tensile properties after the materials suffered a certain number of heating cycles. 2. Experimental Procedure The SG cast iron used in this study is listed in Table 1. The material was prepared by melting high-purity pig iron, ferrosilicon, and silicon steel scrap in a high frequency induction furnace. The spheroidizing treatment was performed with Fe-45 mass% Si-8 mass% Mg-2.5 mass% RE and inoculated with Fe-75 mass% Si in a ladle at 1723 K. Finally, the melts were poured into Y-shaped sand molds with dimensions of 30 mm 100 mm 150 mm in the parallel section. All specimens were fully ferritized before being machined into test specimens. The ferritization procedure followed a typical two-stage isothermal treatment, in which all specimens were maintained at 1203 K for 3 hours, then furnace cooled to 1093 K for a second isothermal holding of 5 hours and finally furnace cooled to room temperature. An image analyzer (Inspectron software) gathered quantitative analysis data, as shown in Table 2, and the optical microstructure of each specimen, as shown in Figs. 1(a) and (c). In addition, to reveal the morphology of the eutectic cell boundary, the specimens were electro-chemically etched in Morries solution (25 g CrO 3 : 133 cm 3 glacial

2 570 H.-M. Lin, T.-S. Lui and L.-H. Chen Table 2 Quantitative data of area fraction of nodule graphite (A g ), average inter-spacing (S g ), graphite nodule size (D g ), average grain size of ferrite matrix (D f ) and mean size of cell boundary region (D e ). A g (%) S g (mm) D g (mm) D f (mm) D e (mm 2 ) acetic acid: 7 cm 3 H 2 O) with a constant applied voltage of 5 volts. The electro-chemically etched microstructure of each specimen is shown in Figs. 1(b) and (d). To prevent oxidation, the specimens were periodically heated in a vacuum furnace (pressure 1:330:133 Pa), in which the maximum temperatures were selected to be 923 K, 973 K, 1023 K, 1073 K and 1123 K, respectively. Thereafter, all the specimens were then rapidly cooled in water at K. For the tensile tests, round-type specimens with 25-mm gauge length and 5-mm gauge diameter were used. After cyclic heating, tensile testing was then performed at room temperature with an initial strain rate of 3: s 1. For the purpose of identifying evidence of plastic deformation in the vicinity of thermal cracks, an electrochemical etching method also used with the Morris solution, 12 17) was applied to acquire etching pits in these specific locations. Scanning Electron Microscope (SEM), Energy Dispersive Spectroscopy (EDS), and a Scanning Auger Microprobe (SAM) were used to examine the fracture morphology, and the area fraction of intergranular fracture was measured directly from the SEM screen. Furthermore, in order to verify the degree of segregation on the intergranular fracture surface, pre-notched specimens were also periodically heated to an identical cycle number for subsequent in-situ fracture testing in the SAM chamber at 223 K to obtain an identical intergranular fracture pattern for SAM analysis. 3. Results 3.1 Effect of maximum temperature on cyclic heating induced embrittlement Figure 2(a) depicts the relationship between maximum heating temperature and load-displacement curves at 298 K, after the material received 20 heating cycles. The effect of heating cycles on the deterioration of flow stress at 1023 K is shown in Fig. 2(b). It should be noted that the tensile elongation significantly decreased to a minimum when the specimens were periodically heated to 1023 K. Our experimental data is summarized in Fig. 3, showing the relationship between tensile elongation and maximum heating temperature, which the specimens were tested at room temperature after cyclic heating. Using SEM observation (Fig. 4), an intergranular fracture surface was evident at both heating temperatures of 973 K and 1023 K. In particular, the area fraction of intergranular fracture measured from an SEM image attained Fig. 1 Optical micrographs of the test specimens: (a) after 5% nital etching; (b) after electrochemically etching; (c) SEM micrographs of the specimens which have some micro-void and inclusion in the grain boundary between graphite nodules. (d) SEM image of the electrochemically etched specimen.

3 Effect of Maximum Temperature on the Cyclic-Heating-Induced Embrittlement of High-Silicon Ferritic Spheroidal-Graphite Cast Iron 571 Load, F/ kg Flow Stress, σ/mpa 1700 (a) K K K R.T K 1023K Displacement, d/ mm (b) 15 cycles 20 cycles 10 cycles 5 cycles 0 cycle Maximum Heating Temperature: 1023K Elongation (%) Fig. 2 (a) Load-displacement curve of different maximum heating temperature after 20 heating cycles. (b) Effect of cyclic heating on the dependence of load vs. displacement at heating temperature 1023 K. All specimens were tensile tested at room temperature after cyclically heating. a maximum value at 1023 K, as shown in Fig. 4(f), which corresponds to the elongation drop shown in Fig Observation of microstructural evolution after cyclic heating with different maximum heating temperatures Figure 1 reveals the microstructural feature of samples etched by various methods. It should be noted that there is significant grain boundary segregation and a fair amount of inclusions located in the central region between graphite nodules. Figs. 1(b) and (d) display the electrochemically etched morphology showing that the inclusion particles were clustered exclusively at the prior-solidificational eutectic cell boundary regions of ferrite matrix, (denoted by the ECB arrows in Figs. 1(b) and (d), which were located at the central region of the ferritic matrix away from the graphite nodules). From the SEM microstructure shown in Fig. 1(d) we Elongation (%) D+IG D+IG Max. Heating Temperature, T/ K observed that visible heterogeneous inclusions were commonly associated with this part of the ferrite boundary region. In addition, etched pits were also present in the vicinity of cell boundary regions. IG D+IG Fig. 3 Fracture elongation at room temperature after cyclic heating plotted against the maximum cyclic heating temperature after 20 cycles of heating test. (D: dimple, C: cleavage and IG: intergranular) Area Fraction of I.G. Fracture (%) 50 (f) C Max. Heating Temperature, T/ K Fig. 4 Fractrographs of the specimens tensile tested after 20 cycles of heating at heating temperature from 923 K to 1123 K: (a) 923 K, (b) 973 K, (c) 1023 K, (d) 1073 K and (e) 1123 K. (f) The relation between maximum cyclic heating temperature and fraction of intergranular fracture.

4 572 H.-M. Lin, T.-S. Lui and L.-H. Chen Fig. 5 Optical micrographs of the specimens after cyclic heating at different heating temperatures for 20 cycles. (a) 923 K, (b) 973 K, (c) 1023 K, (d) 1073 K, (e) high magnification of micrographs at 1073 K and (f) 1123 K. Fig. 6 Electrochemically etched microstructure of the specimens after cyclic heating at different heating temperatures for 20 cycles. (a) 923 K, (b) 973 K, (c) 1023 K, (d) 1073 K, (e) high magnification of micrographs at 1073 K and (f) 1123 K. Fig. 7 (a) Electrochemically etched microstructure of the eutectic cell wall region at the high strain zone (HSZ). (b) Plastic deformation around the fatigue crack on the specimen surface after 20 cycles. Figure 5 illustrates the microstructural feature of the specimens that were periodically heated to a specific heating temperature after 20 cycles. As depicted, in the specimens periodically heated at 923 K, 973 K and 1023 K, cyclicheating-induced cracking is evident in the eutectic cell boundary region of ferrite matrix. The length of the cracks extends when the maximum temperature is raised to 1023 K. Furthermore, when raised to a maximum temperature to 1073 K, fine recrystallized ferrite grains emerged in the ferritic matrix, as shown in Fig. 5(d), causing a recovery of the tensile elongation (Fig. 3). On the other hand, when the specimen was periodically heated at 1123 K, there was a significant deterioration of tensile properties that resulted from partial phase transformation of the ferrite matrix, as shown in Fig. 5(f). Figure 6(a) shows the locus of the prior-solidification eutectic cell boundary in the fully annealed ferrite matrix as revealed by electrochemically etching. Figure 6(b) (d) show the different features of plastic deformation which could be observed in the vicinity of the eutectic cell boundary when

5 Effect of Maximum Temperature on the Cyclic-Heating-Induced Embrittlement of High-Silicon Ferritic Spheroidal-Graphite Cast Iron 573 Fig. 8 (a) Crack path (SEM image); (b) magnesium rich crack initiation site and propagation path; (c) oxygen concentration, (d) phosphorus concentration element and (e) cerium concentration. the maximum heating temperature was 973 K and 1023 K or 1073 K respectively. In particular, when the specimens were periodically heated to 1023 K, significant slip lines are observed, while fine recrystallized grains emerge when the specimen is periodically heated to 1073 K. From Fig. 7, it is confirmed that the cyclic-heating-induced cracks initiated only from inclusions. Consequently, the etching zone, as shown in Fig. 7, can be recognized to be a high strain zone (HSZ), and this indicates that the HSZ area is the main crack initiation site of ferritic SG cast iron. Based on the compositional locus of the above mentioned prior-solidification cell boundaries, Electron Probe Microanalyser Analysis (EPMA) data as shown in Fig. 8 implies that the inclusion-induced cracking behavior is probably related to the amount of magnesium, oxygen, phosphorus and cerium containing inclusions. 3.3 Intergranular fracture surfaces To investigate the characteristics of cyclic-heated-induced intergranular fracture, a sample which had been periodically heated with an identical number of heating cycles was used to conduct an in-situ fracture test at 223 K by SAM. A similar intergranular fracture pattern, as shown in Fig. 9(a), was obtained. This implies that the embrittlement also occurs if the specimens are impacted at 223 K. The fractography features correspond to the inclusion particles, the secondary

6 574 H.-M. Lin, T.-S. Lui and L.-H. Chen (b) dn/(e)/de Fig. 9 (a) The fracture surface of the specimen which suffered cyclic heating at 1023 K and fractured by impact at about 223 K (SEM image), and (b) SAM spectrum of the flat-grain boundary. dimple voids and the smooth intergranular facets, as shown in Figs. 10(a)-(c). From the SAM spectrum of the flat part of the intergranular facet, as shown in Fig. 10(b), a certain amount of oxygen, silicon and magnesium containing segregation can be detected. It should be noted that a magnesium peak was also detected on the oxidation state of magnesium containing inclusion particles, and there was a shift in the kinetic energy, as indicated in Fig. 10(d). The evidence shown in Figs. 10(e) and (f) can be further compared to Fig. 10(d). It is reasonable to suggest that the shift in the kinetic energy of magnesium is a consequence of different oxidation states. 4. Discussion When components are used in an environment where there are periodic changes in a specific heating temperature region, or where the temperature gradient is steep enough, thermal stress occurs in the components and may result in elastic and plastic strains, and ultimately in crack formation. 18) Cyclicheating-induced deformation cumulatively increases with the number of thermal cycles. Notably, the inclusion particles in the eutectic cell boundary may serve the crack initiation sites owing to the localized stress concentration, as raised the maximum temperature up to 1023 K. On the other hand, from the fact that the hydrostatic tensile stress promotes intergranular fracture at 673 K, it is deduced that the triaxial stress field built in the matrix between graphite nodules during tensile deformation will induce the onset of intergranular fracture in SG cast iron ) In ferritic SG cast iron, the graphite nodules can be considered to be voids because of their low deformation resistance (about 20 MPa tensile yield strength). The hydrostatic tensile stress m ¼ð þ r þ z Þ=3 can be developed to a maximum value at the middle point between graphite nodules. Unfortunately, as shown by electrochemical etched morphologies, this central region is also the location of an inclusion clustering area where the former eutectic cell boundary is a part of the ferrite matrix. Therefore, we suggest that the thermal stress and tensile stress will build on the eutectic cell boundary region as well. However, the development of hydrostatic tensile stress is an inevitable factor that promotes intergranular fracture under tensile loading. Yanagisawa et al. 24) pointed out that the degradation of the ferrite grain boundaries strength due to the segregation of magnesium results in 673 K embrittlement of ferrtic spheroidal graphite cast iron. On the other hand, Kobayashi et al. 25) also indicated that intermediate-temperature intergranular embrittlement must have been cause by the segregation of the metallic magnesium to the austenite grain boundaries during solidification. Consequently, another inevitable factor is that metallic magnesium tends to segregate in the abovementioned central region of ferrite grain boundary even after the annealing process is complete. With regards to the existence of magnesium in the grain boundaries, from the analysis data on the cohesion of grain boundaries with different segregants based on the results by Seah, 26) where H sub is the sublimation enthalpy of atoms in J/m 2, it is clear that the well recognized segregants, S, P, Si and Mg all with H sub values below that of iron, will all embrittle iron and promote grain boundary embrittlement, whereas Mo and C will improve the cohesion. Based on experimental evidence as shown in Fig. 9(a), when intergranular fracture was observed, a fair number of inclusions were also clearly present in the eutectic cell boundary region. This implies that the existence of inclusion particles is one of the dominant factor which cause intergranular fracture; therefore an understanding of the characteristics of inclusions is necessary. An Auger microprobe was used to further analyze the intergranular fracture facets including inclusion particles, secondary dimple voids and the smooth area that simultaneously appeared on the intergranular fracture facets after the specimen suffered a certain number of cycles as shown in Fig. 9(a) and Fig. 10(a). The inclusion particles were mainly composed of the elements of magnesium, phosphorous, oxygen and cerium. 4 6) It is well known that the presence of gaseous elements in the cast iron melt can cause casting defects such as blowholes and pinholes in the castings during the eutectic solidification process, and oxygen often concentrates in the eutectic liquid, increasing gas defects. Cast iron contains the two kinds of oxygen, which were oxides and soluble oxygen. Especially after the spheroidizing treatment by magnesium addition, the soluble oxygen decreased until ppm. 27) In addition the oxygen content is raised when the silicon content is

7 Effect of Maximum Temperature on the Cyclic-Heating-Induced Embrittlement of High-Silicon Ferritic Spheroidal-Graphite Cast Iron 575 N(E) N(E) N(E) Fig. 10 SEM image of the specimen with different regions after in-situ impact test in a vacuum: (a) inclusion particles (b) secondary dimple voids and (c) smooth facets. SAM spectra of magnesium on different regions after in-situ impact test in a vacuum: (a) inclusion particles, (b) secondary dimple voids and (c) smooth facets. increased in this study. Consequently, the formation of secondary dimple voids on the intergranular facets, as shown in Fig. 11(a), is probably correlated to the internal oxidation of metallic magnesium in the grain boundaries during the cyclic heating process. Examining the microstructural feature after cyclic heating of the specimens up to 1073 K for 20 cycles, Fig. 5(d) and Fig. 6(d) obviously reveal the emergence of a fair amount of recrystallized fine grains in the matrix. According to our previous study, 28) dynamic recrystallization occurs in the ferritic matrix when tensile tests were performed at a temperature exceeding 1023 K. Therefore, this evidence confirms that recovery and recrystallization occurred. Recrystallization can be easily observed on completion of a certain number of thermal cycles at this heating temperature. Gottstein and Chen observed that cracks are blunted during resulted from strain induced dynamic recrystallization. 29) However, sufficient strain energy will be stored to nucleate dynamic recrystallization when reached a critical strain accumulation. 30) 5. Conclusions (1) Susceptibility to intergranular embrittlement of SG cast iron after cyclic heating is severe at the specific maximum heating temperature of 1023 K. Increasing or decreasing of the heating temperature can suppress the cyclic heating induced embrittlement. Recrystallization of ferrite grains occurs on completion of a certain number of thermal cycles when the heating temperature

8 576 H.-M. Lin, T.-S. Lui and L.-H. Chen Fig. 11 (a) fractography of mass% residual magnesium content specimen after 20 cycles, (b) fractography of mass% residual magnesium content specimen after 5 cycles and (c) fracture surface of mass% residual magnesium content specimen tensile tested at 673 K. is raised to 1073 K, while there is evidence of a significant phase transformation when the cyclic heating temperature is raised to 1123 K. (2) A fair amount of magnesium-containing inclusions tends to be present in the prior-solidificational cell boundary region, which is the central region between graphite nodules, and this plays an important role in the cyclic-heating-induced embrittlement of SG cast iron. (3) Magnesium segregation can be detected at the annealed prior-solidification grain boundary and extends for a few ferrite grains from the central region of the ferrite matrix. Experimental evidence indicates that the internal oxidation phenomenon in the grain boundaries worsens when the number of heating cycles is increased. This is closely related to the degree of metallic magnesium and leads to a change of the intergranular fracture surface from smooth facet to rough facet with small dimple voids. Acknowledgements This work was financially supported by the National Science Council of Taiwan for which we are grateful (Contract No. NSC E ). REFERENCES 1) Y. J. Park, R. B. Gundlach and J. F. Janowak: AFS Trans. 95 (1987) ) C. P. Cheng, S. M. Chen, T. S. Lui and L. H. Chen: Metall. Trans. A. 28A (1997) ) C. P. Cheng, T. S. Lui and L. H. Chen: Metall. Trans. A. 30A (1999) ) H. M. Lin, T. S. Lui and L. H. Chen: Mater. Trans. 44 (2003) ) H. M. Lin, T. S. Lui and L. H. Chen: Mater. Trans. 44 (2003) ) H. M. Lin, T. S. Lui and L. H. Chen: accepted by AFS., 2003, ) J. F. Janowak, J. D. Crawford and K. Röehrig: Casting Eng./Foundry World. 14 (1982) ) W. Fairhurst and K. Röhrig: Foundry Trade J. 146 (1979) ) K. Röhrig: AFS Trans. 87 (1979) ) M. C. Rukadikar and G. P. Reddy: AFS Trans. 95 (1987) ) A. Weronski: Thermal Fatigue of Metals, (Marcel Dekker, Inc., New York, 1991) pp ) G. T. Hahn, P. N. Mincer and A. R. Rosenfield: Exp. Mech. 11 (1971) ) G. T. Hahn, P. N. Mincer and A. R. Rosenfield: Metall. Trans. 3 (1972) ) K. Tanaka, M. Hojo and Y. Nakai: Mater. Sci. Eng. 55 (1982) ) Y. Waku, T. Masumoto and T. Ogura: Trans., JIM 24 (1983) ) Y. Birol: Metallography 21 (1988) ) Y. Birol: J. Mater. Sci. 23 (1988) ) D. M. Stefanescu: Metals Handbook, 10th ed., 1 (ASM INTERNA- TIONAL Metals Park, OH, 1990) pp ) O. Yanagisawa and T. S Lui: Trans., JIM 24 (1983) ) O. Yanagisawa and T. S Lui: Metall. Trans. 16A (1985) ) C. G. Chao, T. S. Lui and M. H. Hon: Metall. Trans. 19A (1988) ) Y. F. Lin, T. S. Lui and L. H. Chen: Metall. Trans. 25A (1994) ) Y. F. Lin, T. S. Lui and L. H. Chen: Metall. Trans. 26A (1995) ) O. Yanagisawa, H. Ishii, K. Matsugi and T. Hatayama: J. JFS 72 (2000) ) T. Kobayashi, K. Nishino, Y. Kimoto, Y. Awano, Y. Hibino and H. Ueno: J. JFS 70 (1998) ) M. P. Seah: Acta Metall. 28 (1980) ) T. Kusakawa, X. Xu and S. Okimoto: Rep. Cast. Res. Lab. Waseda Univ. No. 38 (1988) ) C. P. Cheng, T. S. Lui and L. H. Chen: Cast Met. 8 (1995) ) G. Gottstein and S. Chen: International Conference on Recrystallization in Metallic Materials, (The Minerals, Metals and materials Society, 1990) pp ) C. M. Sellars and J. A. Whiteman: Met. Sci. 13 (1979)