July 24, 2008
Low-Temperature Aging Characteristics of Type 316L Stainless Steel Welds: Dependence on Solidification Mode
By Abe, Hiroshi Watanabe, Yutaka
Thermal aging embrittlement of light water reactor (LWR) components made of stainless steel cast has been recognized as a potential degradation issue, and careful attention has been paid to it. Although welds of austenitic stainless steels have gamma-delta duplex micro structure, which is similar to that of the stainless steel cast, examination of the thermal aging characteristics of the stainless steel welds is very limited. In this investigation, two types of type 316L stainless steel weld metal with different solidification modes were prepared using two kinds of filler metals having tailored Ni equivalent and Cr equivalent. Differences between the two weld metals in the morphology of microstructure, in the composition of delta-ferrite, and in hardening behaviors with isothermal aging at 335 [degrees]C have been investigated. The hardness of the ferrite phase has increased with aging time, while the hardness of austenite phase has stayed the same. The mottled aspect has been observed in delta-ferrite of aged samples by transmission electron microscopy (TEM) observation. These characteristics suggest that spinodal decomposition has occurred in delta-ferrite by aging at 335 [degrees]C. The age-hardening rate of delta-ferrite was faster for the primary austenite solidification mode (AF mode) sample than the primary ferrite solidification mode (FA mode) sample in the initial stage of the aging up to 2000 hours. It has been suggested that the solidification mode can affect the kinetics of spinodal decomposition. DOI: 10.1007/s11661-008-9511-8
AUSTENITIC stainless steels are widely used in the nuclear industry for core internal structures and primary loop recirculation pipes because of their good properties including corrosion resistance, strength, ductility, and weldability. The weld metal of austenitic stainless steels typically has duplex structure with a few to over 10 pct of delta-ferrite in austenite to prevent hot cracking during solidification. It is well known that stainless steel cast is susceptible to thermal aging embrittlement during long- term exposure to service temperature of LWRs. The spinodal decomposition of ferrite phase is considered to be the primary cause of the embrittlement during aging at less than 400 [degrees]C, from a number of investigations for the thermal aging embrittlement of a stainless steel cast.[1-4] Therefore, as in stainless steel cast, spinodal decomposition of delta-ferrite can occur also in austenitic stainless steel welds by thermal aging. However, the investigation about thermal aging embrittlement of austenitic stainless steel welds is very limited.
The respective region of various solidification modes on a constant iron vertical section of the Fe-Cr-Ni ternary diagram and the solidification sequence of the different primary solidification modes are shown schematically in Figure 1.[5,6] The primary ferrite solidification mode (FA mode) begins with primary ferrite and a phase transformation from delta to gamma occurs. The residual delta- ferrite can be observed at the cell or dendrite core. On the other hand, the residual delta-ferrite in primary austenite solidification mode (AF mode) can be observed at cell or dendrite boundaries. In actual components, austenitic stainless steel welds can take either type of solidification mode, FA or AF, mainly depending on the chemical compositions of the material and welding process conditions. The solidification mode of stainless steels can be predicted based on Cr and Ni equivalents using the Schaeffler equation:
Figure 1 suggests that the characteristics of delta-ferrite are different between the FA mode weld and AF mode weld, not only in morphology but also in chemical compositions. The ferrite for the FA mode weld is primary proeutectic ferrite with some eutectic ferrite. In contrast, the ferrite for the AF mode weld is formed only below the eutectic. These ferrites for FA and AF are expected to be different. However, the effect of solidification mode on thermal aging behavior is not known well.
In this study, two types of type 316L stainless steel weld metals with different solidification modes were prepared using two kinds of filler metals having tailored Ni equivalent and Cr equivalent. The differences between the two weld metals in the morphology of microstructure, in composition of delta-ferrite, and in hardening behaviors with isothermal aging at 335 [degrees]C have been investigated.
II. EXPERIMENTAL DETAILS
A. Preparation of 316L SS Weld Samples
Two types of type 316L stainless steel weld samples with different solidification modes were made by using two kinds of filler metals with different Cr/Ni equivalents. One-inch-thick type 316L stainless steel plate was butt welded using the gas tungsten arc welding process. The chemical compositions of type 316L stainless steel plate and the filler metals are shown in Tables I through III, respectively. The Cr^sub eq^/Ni^sub eq^ ratios of the type 316L stainless steel base metal and the filler metals are shown in Figure 2, where the predicted solidification mode is also indicated. The filler metal for AF mode welds contains a much higher level of nickel and nitrogen than the one for FA mode. Therefore, there is significant difference in Ni equivalence between the two filler metals. Because of dilution with the base metal at the welding process, the actual Cr^sub eq^/Ni^sub eq^ ratio of the welds is located somewhere on the line between the Cr^sub eq^/Ni^sub eq^ ratios of the base metal to those of the filler metal in Figure 2.
The actual solidification mode of the welding samples was determined from the morphology of delta-ferrite observation. The surface was mirror polished with diamond paste to 0.5-[mu]m finish, and then etched in 5 wt pct hydrochloric acid solution to reveal ferrite and austenite phases. The etched microstructures of the welding samples are shown in Figure 3. The low Ni equivalent weld is considered as an almost FA mode solidified weld, because the vermicular morphology of the delta-ferrite is dominant and partially observed lathy and acicular ferrite. In contrast, the high Ni equivalent weld considered as AF mode solidified weld, because of the island-shaped delta-ferrite, is mainly observed at cell boundaries. The ferrite fraction of an FA mode sample is evaluated to be 12.7 pct, and that of an AF mode sample is estimated to be 2.5 pct using a ferrite scope.
B. Determination of Thermal Aging Condition
Because there is no dataset available representing the spinodal decomposition rate for delta-ferrite of austenitic stainless steel welds, the aging condition was determined based on the thermal aging data of cast duplex stainless steels, CF-8 and CF-8M, as shown in Figure 4. The aging temperature was determined to be 335 [degrees]C, because Figure 4 suggests that the upper bound temperature where the spinodal decomposition rate can be extrapolated to the boiling water reactor (BWR) coolant temperature will be 335 [degrees]C. Basically, it takes about 5 years for spinodal decomposition to begin under isothermal aging at 288 [degrees]C. On the other hand, it takes only 1000 hours to begin under aging at 335 [degrees]C. The aging durations were 0 (nonaging), 500, 1000, 2000, 4000, and 8000 hours, indicated as closed symbols in Figure 4.
C. Evaluation of Thermal Aging Behavior and Chemical Composition of Welding Samples
It is widely accepted that an increase in hardness and reduction in toughness of high chromium ferritic stainless steels, so-called "475 [degrees]C embrittlement," are caused by phase separation into chromium-enriched alpha' and iron-rich alpha phases. Low- temperature embrittlement of cast duplex stainless steel was also caused mainly by spinodal decomposition of ferrite phase into alpha' and alpha, and the degree of embrittlement was commonly evaluated by measuring changes in the microhardness of ferrite.[4,9,10] Therefore, the thermal aging behavior of welding samples with aging at 335 [degrees]C was evaluated to measure the changes in hardness of the ferrite and austenite phases using a micro-Vickers indenter. The hardness values were obtained by averaging five measurements. The indentation loads are 10 mN for the FA mode material and 5 mN for the AF mode material. The hardened layer formed on the surface of the specimen during the mechanical polishing process has been removed by electropolishing using a solution of CH3COOH and HClO4 (9:1 ratio).
It has been reported that the nanoscaled mottled aspect was observed in spinodally decomposed ferrite in the duplex cast steel by TEM observation.[1,8] The changes in the nanoscaled microstructure of alpha-ferrite in the weld samples with aging at 335 [degrees]C were evaluated using a field emission gun TEM. The thin-film samples were obtained by a focused ion beam equipped with a microsampling function.
The chemical composition and compositional profile of delta- ferrite in weld metal have some variation depending on the solidification mode and welding process. Because the spinodal decomposition rate can be affected by the chemical composition of delta-ferrite, the chemical composition of δ-ferrite in weld samples was evaluated using TEM-energy-dispersive X-ray analysis (EDX). III. RESULTS AND DISCUSSION
A. Aging Behavior of Welding Samples with Aging at 335 [degrees]C
The changes in Vickers hardness of ferrite and austenite phase of the weld samples with aging at 335 [degrees]C are shown in Figures 5 and 6. The hardness of the ferrite phase has been clearly increased with isothermal aging, while the hardness of the austenite phase stayed the same. This hardening was thought to be due to the spinodal decomposition of delta-ferrite.
Figure 2 indicates that spinodal decomposition in CF-8 and CF-8M begins after 1000 hours aging at 335 [degrees]C; however, delta- ferrite in type 316L stainless steel welds, which were used in this study, was already hardened after 500 hours aging. There was a difference in the aging behavior between the cast duplex stainless steels, CF-8 and CF-8M, and the 316L SS welds.
A clear difference was observed in the age-hardening rate of delta-ferrite between the FA sample and the AF sample. In the initial stage of the isothermal aging up to 2000 hours, the hardening rate was faster for the AF sample than for the FA sample. The ferrite hardness of the FA sample continued to increase until 8000 hours. On the other hand, the ferrite hardness of the AF sample steeply increased up to 2000 hours and reached over 350 Hv, and then the hardness gradually decreased. It was reported that the spinodally decomposed chromium and iron-rich domains in the ferrite phase of duplex stainless steels apparently increased after long- term aging at 400 [degrees]C. The increase in hardness of delta- ferrite can be attributed to the formation of a modulated structure by spinodal decomposition, and the decrease in hardness of delta- ferrite for AF material beyond 2000 hours of aging may be due to a coarsening of finely dispersed alpha' and alpha.
The bright-field TEM images of delta-ferrite unaged and aged for 8000 hours at 335 [degrees]C samples are shown in Figures 7 and 8. The G phase, a nickel and silicon-rich fee phase,[11-13] and M^sub 23^C^sub 6^ carbide precipitation have not been observed in the aged samples. The mottled aspect has been observed in delta-ferrite of both FA and AF solidification mode aged samples. This aspect might be attributed to the fluctuation of composition by spinodal decomposition.[1,8,9] However, TEM of thin foils does not allow easy observation of the alpha' in the ferrite phase, because both phases have the same bcc lattice, their lattice parameters are too close, and so are their electron scattering factors.
B. Effect of Chemical Composition on Kinetics of Spinodal Decomposition of delta-Ferrite
The contrast in age-hardening behavior indicates a difference in kinetics of spinodal decomposition, which may be attributed to the chemical compositions of ferrite. It has been reported that nickel, silicon, and molybdenum affect the duplex separation of the ferrite phase,[2,14,15] and that a formation of alpha' is promoted by the presence of vanadium and nitrogen. In addition, it has been reported in the study of embrittlement of duplex stainless steels aged at 475 [degrees]C by Solomon and Levinson that nickel affects spinodal decomposition and accelerates alpha' formation. However, the similar evidence on the kinetics of spinodal decomposition at a lower temperature is not yet available.
The compositional profiles of unaged weld samples across the gamma-delta interface measured by TEM-EDX in 25-nm intervals are shown in Figures 9 and 10. The delta-ferrite phase contains higher chromium, molybdenum, and lower nickel than the austenite phase in both samples. There was no segregation of any elements observed at the gamma-delta interface. The average chemical compositions of delta-ferrite in each sample are shown in Table IV. The delta- ferrite of the AF mode sample contains higher chromium by about 4 wt pct and lower molybdenum and manganese than that of the FA mode weld. The reason for the differences in chemical compositions of delta-ferrite between the two weld samples has been described previously. It is well known that microsegregation tends to be significant for AF mode welds compared with FA mode welds. From this investigation, the possibility is suggested that minor elements microsegregated in ferrite may play a significant role in accelerating spinodal decomposition. However, to fully understand the accelerating factor of spinodal decomposition, more investigations are needed.
Two types of type 316L stainless steel weld metals with different solidification modes were prepared using two kinds of filler metals having tailored Ni equivalent and Cr equivalent. The differences between the two weld metals regarding the morphology of microstructure, the composition of delta-ferrite, and the hardening behaviors with isothermal aging at 335 [degrees]C have been investigated. The main conclusions can be summarized as follows.
1. Weld metal of type 316L stainless steel was clearly age hardened at a relatively low temperature, 335 [degrees]C. The age- hardening rate of delta-ferrite was faster for the AF sample than for the FA sample in the initial stage with aging up to 2000 hours. The ferrite hardness of the FA sample continued to increase until 8000 hours. On the other hand, the ferrite hardness of the AF sample steeply increased up to 2000 hours and reached over 350 Hv, and then the hardness gradually decreased after 2000 hours.
2. The mottled aspect has been observed in delta-ferrite of both the FA and AF solidification mode aged samples by TEM observation. This aspect might be attributed to spinodal decomposition into chromium-rich alpha' and iron-rich alpha.
3. The contrast in age-hardening behavior would indicate a difference in kinetics of spinodal decomposition, which may be due to the difference in the chemical compositions of delta-ferrite. The delta-ferrite of the AF mode sample contains higher chromium by about 4 wt pct and lower molybdenum and manganese than that of the FA mode weld. The possibility is suggested that minor elements microsegregated in ferrite may play a significant role in accelerating spinodal decomposition.
This work was performed as a part of the "SCC mechanism studies of L-grade stainless steels in hightemperature water" organized by JSCE and supported by Japanese BWR utilities. This work was supported by a grant-in-aid for JSPS fellowship (Grant No. 19008176). Help by Dr. Miyazaki, Department of Instrumental Analysis, Technical Division, School of Engineering, Tohoku University, for the TEM observation is greatly appreciated.
1. M.K. Miller and J. Bentley: Proc. 3rd Conf. on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors, TMS-AIME, Warrendale, PA, 1988, pp. 341-49.
2. H.M. Chung and T.R. Leax: Mater. Sci. Technol., 1990, vol. 6, pp. 249-62.
3. P. Auger, F. Danoix, A. Menand, S. Bonnet, J. Bourgoin, and M. Guttamann: Mater. Sci. Technol., 1990, vol. 6, pp. 301-13.
4. P.H. Pumphrey and K.H. Akhurst: Mater. Sci. Technol., 1990, vol. 6, pp. 211-19.
5. N. Suutala, T. Taklo, and T. Moisio: Metall. Trans. A, 1979, vol. 10A, pp. 512-14.
6. J.A. Brooks, M.I. Baskes, and F.A. Greulich: Metall. Trans. A, 1991, vol. 22A, pp. 915-26.
7. N. Suutala, T. Taklo, and T. Moisio: Metall Trans. A, 1980, vol. 11A, pp. 717-25.
8. H.M. Chung and O.K. Chopra: Proc. 3rd Conf. on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors, TMS-AIME, Warrendale, PA, 1988, pp. 359-70.
9. K.L. Weng, H.R. Chen, and J.R. Yang: Mater. Sci. Eng. A, 2004, vol. 379, pp. 119-32.
10. J.J. Shiao, C.H. Tsai, J.J. Kai, and J.H. Huang: J. Nucl. Mater., 1994, vol. 217, pp. 269-78.
11. M. Vrinat, R. Cozar, and Y. Meyzaud: Scripta Metall., 1986, No. 8, pp. 1101-06.
12. J.M. Vitek: Metall. Trans. A, 1987, vol. 18A, pp. 154-56.
13. A. Mateo, L. Lianes, and M. Anglada: J. Mater. Sci., 1997, No. 32, pp. 4533-40.
14. T. Yoshimura and Y. Ishikawa: ASME PED, 1990, vol. 46, pp. 77- 88.
15. H.M. Chung: ASME PVP, 1989, vol. 171, pp. 111-25.
16. P. Koutaniemi, V. Heikkinen, and A. Saarinen: Metall. Sci., 1974, vol. 8, pp. 94-96.
17. A. Hendry, Z.F. Mazur, and K.H. Jack: Metall. Sci., 1979, vol. 13, pp. 482-86.
18. H.D. Solomon and L.M. Levinson: Acta Metall., 1978, vol. 26, pp. 429-42.
HIROSHI ABE, Doctoral Student, and YUTAKA WATANABE, Associate Professor, are with the Graduate School of Engineering, Tohoku University, Sendai, Japan. Contact e-mail: [email protected] tohoku.ac.jp
Manuscript submitted June 4, 2007.
Article published online April 1, 2008
Copyright Minerals, Metals & Materials Society Jun 2008
(c) 2008 Metallurgical and Materials Transactions; A; Physical Metallurgy and Materials Science. Provided by ProQuest Information and Learning. All rights Reserved.