Journal of Materials Science & Technology  2020 , 40 (0): 1-14 https://doi.org/10.1016/j.jmst.2019.08.035

Microstructure and stress corrosion cracking of a SA508-309L/308L-316L dissimilar metal weld joint in primary pressurized water reactor environment

Lijin Dongab, Cheng Maa, Qunjia Penga*, En-Hou Hana, Wei Kea

a Key Laboratory of Nuclear Materials and Safety Assessment, Institute of Metal Research, Chinese Academy of Sciences, Shenyang, 110016, China
b School of Materials and Engineering, Southwest Petroleum University, Chengdu, 610500, China

Corresponding authors:   *Corresponding author.E-mail address: pengqunjia@yahoo.com (Q. Peng).*Corresponding author.E-mail address: pengqunjia@yahoo.com (Q. Peng).

Received: 2019-06-21

Revised:  2019-08-7

Accepted:  2019-08-20

Online:  2020-03-01

Copyright:  2020 Editorial board of Journal of Materials Science & Technology Copyright reserved, Editorial board of Journal of Materials Science & Technology

More

Abstract

Stress corrosion cracking (SCC) of an SA508-309 L/308 L-316 L dissimilar metal weld joint in primary pressurized water reactor environment was investigated by the interrupted slow strain rate tension tests following a microstructure characterization. The 308 L weld metal shows a higher content of δ ferrite than the 309 L weld metal. In addition, no obvious Cr-depletion but carbides precipitation at γ/δ phase boundaries was observed in both 308 L and 309 L weld metals. The slow strain rate tension tests showed that the SCC susceptibility of the base and weld metals of the dissimilar metal weld joint follows the order of SA508 < 308 L weld metal < the heat affected zone of 316 L base metal < 309 L weld metal. The higher SCC susceptibility of 309 L weld metal than that of 308 L weld metal is likely due to the lower content of δ ferrite. In addition, a preferential SCC initiation in the 309 L weld metal adjacent to 308 L weld metal is attributed to few carbides in this region.

Keywords: Dissimilar metal weld joint ; Stress corrosion cracking ; Microstructure ; Primary pressurized water reactor ; environment

0

PDF (9582KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

Lijin Dong, Cheng Ma, Qunjia Peng, En-Hou Han, Wei Ke. Microstructure and stress corrosion cracking of a SA508-309L/308L-316L dissimilar metal weld joint in primary pressurized water reactor environment[J]. Journal of Materials Science & Technology, 2020, 40(0): 1-14 https://doi.org/10.1016/j.jmst.2019.08.035

1. Introduction

Stainless steels and nickel base alloys have been widely used as key weld metals to attach the stainless steel piping to the low alloy steel (LAS) reactor pressure vessel nozzle in nuclear power plants, due to their high resistance to corrosion and appropriate intermediate thermal expansion coefficients. However, the LAS-stainless steel/nickel base alloy-stainless steel dissimilar metal weld joints were found to be susceptible to stress corrosion cracking (SCC) under light water reactor conditions [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21]]. A number of studies have been carried out to clarify the correlation of microstructure and SCC of dissimilar metal weld joints in high temperature water [[3], [4], [5], [6],[13], [14], [15], [16], [17], [18]]. The general consensus is that the fusion boundary (FB) region has a higher susceptibility to SCC than the bulk weld metal due to the change of microstructure, i.e., dilution of weld metal during welding, migration of C from the LAS to the weld metal during post-weld heat treatment, increase of residual strain in the heat affected zone (HAZ) and formation of a narrow martensite layer [[3], [4], [5], [6], [7], [8], [9],[13], [14], [15], [16], [17], [18], [19], [20]]. For example, recent investigations on SCC of LAS-Alloy 52 M metal weld in primary pressurized water reactor (PWR) environment revealed that crack could initiate at the FB and propagate into the dilution zone of weld metal [4,5]. Other studies also revealed the dilution zone of weld metal had a moderate stress corrosion crack growth rate in high temperature pure water with dissolved oxygen (DO) [19,20]. In addition, the residual strain as a result of weld shrinkage led to the occurrence of SCC in the HAZ of 316 L stainless steel in primary PWR environment [4,15,17].

Due to the difference in chemical composition and microstructure between the stainless steels and nickel base alloys, the SCC behavior of LAS-stainless steel-stainless steel weld joints may differ from that of the LAS-nickel base alloy-stainless steel weld joints. To date, a couple of works on SCC of the LAS-stainless steel-stainless steel weld joints have been carried out [[7], [8], [9], [10], [11]]. Recent research on LAS-309 L/308 L-stainless steel weld joint revealed stress corrosion crack growth occurred in the HAZ of 316 L stainless steel in off-normal water chemistry of primary PWR environment with DO, whereas the 308 L weld metal showed a high resistance to stress corrosion crack growth [9,10]. In addition, it was found that the SCC behavior of the FB region of a stainless steel-stainless steel similar metal weld was influenced by the load direction [11]. The crack propagated parallel to the FB if the angle between the neutral line of the compact tension specimen and the FB was small. Other studies investigated the SCC behavior of the LAS-309 L/308 L dissimilar metal weld by using slow strain rate tension (SSRT) tests with a tensile direction perpendicular to the FB [7,8]. It was revealed that the FB region of the LAS-309 L/308 L dissimilar metal weld had a higher susceptibility to SCC than the bulk weld metals in sulfate doped, high temperature oxygenated water [7,8]. However, the correlation of material factors and SCC behavior of various regions of the weld joint is still not fully understood. One limitation of these studies is that the designed specimen used in the SSRT tests was only suitable for the assessment of the SCC susceptibility of the FB region of the LAS-stainless steel dissimilar metal weld, whereas the SCC behavior of other regions is still unclear. On the other hand, both the axial and circumferential stress corrosion cracks were found in the safe end during the service of nuclear power plants, suggesting that different load directions could result in different SCC behavior of weld joints [4,22]. For example, since the dendrite grain boundary of 309 L and 308 L weld metals in the FB region was generally perpendicular to the FB, the tensile direction perpendicular to the FB may fail to evaluate the SCC susceptibility of the weld metals in Refs. [7,8]. Therefore, investigation on the SCC susceptibility of various regions with different load directions are needed for more accurate safety assessment of the weld joint. In addition, due to the post-weld heat treatment and dilution of 309 L weld metal, difference in chemical composition and microstructure such as δ ferrite phase and carbides between the 309 L and 308 L weld metals can be expected, both of which could affect the SCC behavior of the weld metals in primary PWR environment.

In the current study, the correlation of microstructure and SCC behavior of the SA508-308 L/309 L-316 L dissimilar metal weld joint in primary PWR environment was investigated. The microstructure of the weld joint was analyzed by using an optical microscope (OM), a Vickers microhardness tester, a scanning electron microscope (SEM), a transmission electron microscope (TEM) and electron backscatter diffraction (EBSD). The SCC behavior of different regions of the weld joint was studied by the SSRT tests. To correlate the SCC behavior with the applied strain, the SSRT tests were interrupted at five strains of 5%, 10%, 15%, 20% and 25%. After the interrupted SSRT tests, one of the specimens for each type was finally strained to failure in primary PWR environment to observe the fracture surface by SEM.

2. Experimental

2.1. Materials and specimen

The SA508-309 L/308 L-316 L dissimilar metal weld joint used for the study was cut from a mockup of the safe-end weld joint of the primary circuit of PWR, as schematically shown in Fig. 1(a). The weld joint was prepared by depositing a 309 L buttering layer on SA508 using the shielded metal arc welding. Post-weld heat treatment at 600-620 °C for about 40 h was then performed to relieve the residual stress followed by multi-layer welding between the 309 L buttering layer and the 316 L stainless steel safe-end with 308 L electrodes. No post-weld heat treatment was performed after the final welding. The chemical composition of the base and weld metals is given in Table 1.

Fig. 1.   Schematic illustration of the SA508-309 L/308 L-316 L weld joint and specimens:. (a) the location of SSRT and microstructure-analysis specimens, (b) the geometry and dimension of the SSRT specimen. All units are in mm.

Table 1   Chemical composition (wt%) of the base and weld metals used for manufacturing the SA508-309 L/308 L-316 L weld joint.

MaterialFeCrNiMnSiCMoPS
SA508Bal.0.120.751.350.180.1870.470.0090.004
316 LBal.17.9811.721.730.490.0232.370.0200.031
308 LBal.19.8810.031.330.320.016-0.0150.011
309 LBal.23.2613.431.760.380.017-0.0110.002

New window

In total five types of small-sized, round SSRT specimens extracted from the weld joint were used to investigate the SCC behavior of different regions (Fig. 1(a)). The specimens S1 to S4 are in the circumferential orientation of the weld joint with a tensile direction approximately parallel to the FB. The specimens S1 were used to investigate the SCC behavior of 308 L weld metal. However, the specimens S2 to S4 were located in the FB region of 316 L-308 L, 308 L-309 L and 309 L-SA508 metal welds, respectively. To investigate the SCC behavior of the weld joint with the tensile direction approximately perpendicular to the FB, specimens S5 were extracted across the 309 L and 308 L weld metals in the axial direction. The gauge section of the specimens S1‒S4 is 20 mm in length and 3 mm in diameter, while the gauge length of specimen S5 is 30 mm (Fig. 1(b)). These specimens were ground using SiC papers up to 3000 grit and finally polished using 1 μm diamond paste. Plate type specimens with a dimension of 10 mm × 10 mm × 1 mm were also extracted from the FB region of the weld joint for microstructure analysis (Fig. 1(a)).

2.2. Microstructure analysis

An OM (Zeiss Axio Observer. Z1) was used for the observation of the metallographic microstructure of the weld joint. Prior to the OM observation, the SA508 part was etched by 4 vol.% nital (4 ml of nitric acid and 96 ml of ethanol), while 316 L, 308 L and 309 L stainless steels were electro-etched in 10 wt% oxalic acid solution. Measurements of the content of δ ferrite phase in 309 L and 308 L weld metals were also conducted using the OM according to ASTM E1245.

Microhardness distribution of the weld joint was obtained using an AMH43 microhardness tester with a load of 100 g and a holding time of 15 s, while a load of 10 g with a holding time of 15 s was selected for the precise analysis of the FB region.

The grain/phase boundary microstructure and chemical composition of the FB region were analyzed by an FEI Quanta 650 field emission SEM equipped with an energy dispersive X-ray (EDX) spectroscopy detector.

Grain boundary character and residual strain in the FB region of 308 L/309 L-SA508 dissimilar metal weld were analyzed by using an EBSD detector attachment in the SEM, which was equipped with a camera in connection with the TSL software for analyzing the misorientation. The specimens used for EBSD analysis were ground using SiC papers up to 2000 grit, then polished by diamond paste to 1 μm, and finally polished with 0.04 μm colloidal silica polishing slurry. Due to the difference in the grain size between SA508 Cl. 3 and 309 L/308 L weld metal, the EBSD analysis was performed with a step size of 2 μm at a voltage of 25 keV to show the overall microstructure of 309 L/308 L weld metal, while a smaller step size of 0.2 μm was used for the precise analysis of FB region. Local strain distribution in the weld metals was assessed using the kernel average misorientation (KAM), which has a linear relationship with the degree of strain [23,24]. The KAM is a local misorientation defined as an average misorientation of a point with all of its neighbors in a grain. For a given point, the average misorientation of that point with all of its neighbors was calculated with a criterion that misorientations exceeding a tolerance value (5° here) were excluded from the calculation.

The microstructure of 309 L and 308 L weld metals were analyzed using a JEOL 2100 TEM equipped with an EDX spectroscopy detector. TEM analyses of 309 L weld metal were conducted in three different regions (about 50 μm, 200 μm and 500 μm to the FB of 308 L-309 L similar metal weld, respectively). Specimens with an original dimension of 10 mm × 10 mm × 0.2 mm were cut from the weld metals by wire cut electrical discharge machining. Then they were ground using SiC papers up to 3000 grit to reduce the thickness to be less than 60 μm. A drill bit with a 3 mm inner diameter was used to drill out discs through the thinned specimens. These discs were then thinned to electron transparency by a Struers Tenupol-5 Twin-jet electropolishing system, using a solution of 8 vol.% perchloric acid of 70 wt% concentration and 92 vol.% ethanol of 99.7 wt% concentration at 24 V and -25 °C.

2.3. SSRT tests in primary PWR environment

A refreshed loop equipped with a 3-L autoclave made of 316 L stainless steel was used for the SSRT tests in primary PWR environment at 320 °C and 13 MPa. The primary PWR environment was prepared by high-purity water with 1.2 g/L of B as H3BO3 and 2.3 mg/L of Li as LiOH. Before adding the chemicals into the water tank, the water in the loop was circulated and purified with a high-efficiency ion-exchanger to a resistivity of 15 MΩ/cm. Dissolved hydrogen (DH) in the influent water was controlled by bubbling H2 into the water tank until equilibration occurred. The concentration of DH and DO used for the test was 2.6 mg/L and <5 μg/L, respectively. Conductivity, DH and DO of the influent water were continuously monitored during the test.

SSRT tests were performed at a strain rate of 3 × 10-7 s-1 using a servo motor in the constant displacement rate mode. The crosshead displacement was measured using a linear variable differential transformer to examine the strain rate. Details of the setup process of the specimen and fixture were reported previously [4,5].

In order to correlate the SCC initiation behavior with the applied strain, in total five interrupted strains at of 5%, 10%, 15%, 20% and 25% were employed for the SSRT tests. Prior to strain the specimens, they were exposed to the primary PWR environment for 2 days to stabilize the environmental condition with a load of about 50 N. This small load helped to ensure that there was no slack in the load train. At each interrupted strain, all specimens were removed from the autoclave for examinations of SCC initiation behavior. Following the examination, the specimens were then pulled to the next target strain in primary PWR environment after a two-day stabilization of the environment. In addition, one of the specimens for each type was finally strained to failure in primary PWR environment to observe the fracture surface by SEM. The SCC susceptibility of different regions of the weld joint were primarily sorted by the minimum strain for crack initiation, the higher area percentage of intergranular cracking, and then the higher area percentage of brittle-like transgranular cracking.

3. Results

3.1. Microstructure of the SA508-309 L/308 L-316 L weld joint

3.1.1. OM observation

The 308 L weld metal consisted of typical dendritic grains with an amount of δ ferrite phase between the dendritic grain boundaries of γ austenite, while equiaxed grains were observed in the 316 L stainless steel base metal (Fig. 2(a)). A carbon-depleted zone was observed by OM in the HAZ of SA508 adjacent to the FB, shown in Fig. 2(b). The formation of carbon-depleted zone was due to the migration of carbon from the HAZ of SA508 to 309 L weld metal during the post-weld heat treatment and welding. Further, a martensitic layer with a width of 5-10 μm was observed between 309 L weld metal and LAS, due to the formation of intermediate compositions in the FB region [25,26]. The type-II boundary, which is parallel to the FB with a distance of < 100 μm, usually appears adjacent to the FB in austenitic alloy-LAS dissimilar metal weld. However, this boundary was only observed at some isolated locations in the 309 L weld metal with a distance of 5-30 μm to the FB (Fig. 2(c) and (d)). This is in consistence with the observation in Refs [8,24]. The microstructure of 309 L weld metal was similar to that of 308 L weld metal (Fig. 2(e) and (f)). However, the content of δ ferrite phase in 309 L weld metal was 6.7 vol.%, which was lower than the content of 10.9 vol.% in 308 L weld metal.

Fig. 2.   OM observation of the SA508-309 L/308 L-316 L weld joint: (a) the FB region of the 308 L-316 L similar metal weld, (b) the LAS part of the SA508-309 L dissimilar metal weld, (c, d) the 309 L part of the SA508-309 L dissimilar metal weld, (e, f) the FB region of the 309 L-308 L similar metal weld.

3.1.2. Microhardness and chemical composition analysis

Microhardness distribution in the FB regions of the weld joint is shown in Fig. 3(a)‒(c). The average microhardness of the HAZ of SA508 was 272 HV, which was about 50 HV higher than that of 309 L weld metal, as shown in Fig. 3(a). A hardened zone with a width of about 50 μm in the 309 L weld metal adjacent to the FB was also observed. This was likely due to the migration of carbon and formation of martensitic layer [26,27]. The microhardness of 309 L weld metal was a little higher than that of 308 L weld metal (Fig. 3(b)), while the 316 L base metal showed a lower microhardness than the HAZ of 316 L stainless steel with a width of about 6 mm and 308 L weld metal (Fig. 3(c)). Composition profiles of the FB region of the weld joint are shown in Fig. 3(d)‒(f). A steep decrease of Fe content but increase of Ni and Cr content within a distance of 20-30 μm was observed adjacent to the FB in the dilution zone of 309 L weld metal (Fig. 3(d)). In addition, the 308 L weld metal contained higher Cr but lower Fe content than the 309 L weld metal and 316 L stainless steel (Fig. 3(e) and (f)). The lower Cr content in 309 L weld metal than the nominal composition was due to the dilution by SA508 during welding.

Fig. 3.   Microhardness distribution and composition profiles of the SA508-309 L/308 L-316 L weld joint: (a-c) microhardness distribution in the FB region of the SA508-309 L, 309 L-308 L and 308 L-316 L metal welds, respectively, (d-f) composition profiles of the FB region of the SA508-309 L, 309 L-308 L and 308 L-316 L metal welds, respectively.

3.1.3. EBSD analysis of the fusion boundary region of SA508-309 L/308 L dissimilar metal weld

The inverse pole figures, KAM maps and grain boundary character distribution maps of the SA508-309 L/308 L dissimilar metal weld are shown in Fig. 4. Dendritic grains of 309 L weld metal that grew perpendicular to the FB were observed in the interdendritic region (Fig. 4(a)). Further, austenitic planar solidification zone and martensitic layer were confirmed again by the precise analysis of the FB region (Fig. 4(d)). As shown in Fig. 4(b), both low and high angle grain boundaries were observed in the weld metals. In addition, there was a fraction of coincidence site lattice boundaries in the planar solidification zone of 309 L weld metal (Fig. 4(e)). This is in consistence with the EBSD analysis in literature [26]. The residual strain corresponding to an average KAM value of 1.31 was observed in 309 L weld metal (Fig. 4(c)). However, the average KAM was only about 0.76 in 308 L weld metal, suggesting that 309 L weld metal was subjected to a higher residual strain than 308 L weld metal. Further, the residual strain in the LAS of the FB region was higher than that of the 309 L weld metal adjacent to the FB (Fig. 4(f)).

Fig. 4.   EBSD analyses of FB region of the SA508-309 L/308 L dissimilar metal weld: (a, d), inverse pole figures, (b, e), KAM maps, (c, f), grain boundary character distribution maps. In (b) and (e), the black lines represent high angle grain boundaries (∑ values bigger than 29), the white lines represent low angle grain boundaries (∑1), the red lines represent ∑3 grain boundaries, and the blue lines represent the coincidence site lattice grain boundary with ∑ values ranging from 5 to 29.

Inhomogeneous residual strain was also observed in the 309 L weld metal, as shown in Fig. 5(a). Decreasing the distance to the FB of 309 L-308 L similar metal weld led to an increase of KAM value, indicating the highest residual strain in the 309 L weld metal adjacent to 308 L weld metal. In addition, quantification of the grain boundary character distribution (GBCD) revealed that the GBCD in the 309 L and 308 L weld metals was similar, as shown in Fig. 5(b).

Fig. 5.   KAM and GBCD distribution in the 308 L and 309 L weld metals: (a) the KAM versus distance to the FB, (b) GBCD distribution.

3.1.4. SEM observation and TEM analysis of the morphology and composition in different regions of the 309 L and 308 L weld metals

Morphology and composition analysis of 308 L weld metal by SEM and TEM are shown in Fig. 6. The carbides with a width of 50-100 nm showed a semi-continuous distribution along the γ/δ boundary in 308 L weld metal, whereas the γ/γ grain boundaries were free of carbide precipitation (Fig. 6(a) and (b)). The diffraction pattern shown in Fig. 6(d) indicated that the carbides were M23C6. Precipitation of carbides led to a slight depletion of Cr at the M23C6/δ phase boundary, as displayed by the EDX mappings in Fig. 6(e). The lowest Cr content at the M23C6/δ phase boundary was about 24 at.% according to the composition profile (Fig. 6(f)). Further, no obvious Cr-depletion was observed at the γ/M23C6 phase boundary. Similar results were obtained by using the three-dimensional atom probe tomography in the previous study [28,29].

Fig. 6.   SEM observation and TEM analyses of 308 L weld metal: (a) SEM back-scattered electron image of γ/γ boundary, (b) SEM back-scattered electron image of γ/δ boundary, (c) TEM image of the γ/δ boundary in DZ, (d) diffraction pattern of the carbide, (e) EDX mappings of the carbides, (f) composition profile across the carbide.

SEM and TEM observations in different regions of 309 L weld metal are shown in Fig. 7. The distribution of carbides in the 309 L weld metal with a distance of 50 μm to the FB of 309 L-SA508 dissimilar metal weld was similar to that in 308 L weld metal (Fig. 7(a)). However, the size of carbides was bigger than that in 308 L weld metal (Fig. 7(b)). As shown in Fig. 7(c) and (d), the density and size of the carbides in 309 L weld metal decreased with increasing the distance to the FB. Further, no carbide precipitation in 309 L weld metal were observed with the distance to more than 500 μm (Fig. 7(e) and (f)). In addition, neither precipitation of carbide nor depletion of Cr was observed at the γ/γ grain boundary in both 308 L and 309 L weld metals (Fig. 7(g) and (h)). The difference in density and size of the carbides in 309 L weld metal was related to the content of carbon in the local region. Since the SA508 has a much higher content of carbon than the 309 L weld metal, the 309 L weld metal adjacent to SA508 could provide more carbon for the formation of carbides during the welding and post-weld heat treatment.

Fig. 7.   SEM observation and TEM analyses of 309 L weld metal in different distances to the FB of SA508-309 L dissimilar metal weld: (a, b), about 50 μm to the FB, (c, d) about 200 μm to the FB, (e, f) about 500 μm to the FB, (g) observation of γ/γ boundary, (h) composition profile across the γ/γ grain boundary. (a), (c) and (e) are SEM back-scattered images, (b), (d) and (f) are TEM images.

It should be noted that both the γ austenite and δ ferrite phase showed similar compositions in different regions of 308 L and 309 L weld metals, as confirmed by the TEM-EDX analyses (Fig. 8). The content of δ ferrite phase in stainless steel weld metals can be predicted by using the Schaeffler diagram [26,30], in which the nickel and chromium equivalents are defined by the following equation:(1)Creq=[Cr]+[Mo]+1.5[Si]+0.5[Nb](2)Nieq=[Ni]+30[C]+0.5[Mn]where [Cr], [Mo], [Si], [Nb], [Ni], [C] and [Mn] are weight percent of Cr, Mo, Si, Nb, Ni, C and Mn, respectively. According to the Schaeffler diagram and the nominal composition of weld metals, the calculated Nieq/Creq and ferrite content of 309 L weld metal are 0.622 and 8.7 vol%, respectively, whereas the Nieq/Creq and ferrite content of 308 L weld metal are 0.549 and 9.9 vol%, respectively. The dilution of 309 L weld metal by SA508 during welding could lead to an increase of carbon content but decrease of Cr content, and therefore, lowered the content of δ ferrite phase in 309 L weld metal. As a result, the 308 L weld metal had a higher content of δ ferrite phase than the 309 L weld metal. This is in consistence with the observation in Refs. [26,31].

Fig. 8.   TEM-EDX analysis of the chemical composition of the ferrite and austenite phase in different regions of the 308 L and 309 L weld metals: (a) ferrite phase, (b) austenite phase.

3.2. SCC behavior of the SA508-309 L/308 L-316 L weld joint in primary PWR environment

3.2.1. Cracking behavior of 308 L weld metal

SEM observations on the surface of specimen S1 following 15% and 20% interrupted strains of SSRT in primary PWR environment are shown in Fig. 9(a) and (b), respectively. No cracks were observed following 15% strain by SSRT. However, increasing the strain to 20% led to initiation of a few transgranular cracks. The transgranular cracking of 308 L weld metal in primary PWR environment is in consistence with the observation in Refs. [8,15,32,33]. In addition, no intergranular facets but only a few transgranular cracks were observed on the fracture surface of the specimen after the SSRT test, indicating a high SCC resistance of 308 L weld metal (Fig. 9(c) and (d)).

Fig. 9.   SEM observations of cracking in the 308 L weld metal on specimen S1: (a) surface morphology following 15% strain by SSRT, (b) surface morphology following 20% strain by SSRT, (c) an overall observation of fracture surface after SSRT test, (d) observation at higher magnification showing the cracking morphology, (e) schematic drawing showing the location of facture surface on specimen S1.

3.2.2. Cracking behavior of 316 L-308 L similar metal weld

No crack initiation in both the HAZ of 316 L stainless steel and 308 L weld metal was observed on the surface of specimen S2 following 15% strain by SSRT with a tensile direction approximately parallel to the FB. When the strain increased to 20%, a number of shallow transgranular cracks were observed on the surface of 308 L weld metal (Fig. 10(a)). However, both intergranular and transgranular cracks that generally perpendicular to the tensile direction initiated in the HAZ of 316 L stainless steel (Fig. 10(b)). This is in consistence with the observation of fracture surface after the SSRT test (Fig. 10(c) and (d)), suggesting the SCC susceptibility of 308 L weld metal was lower than that of the HAZ of 316 L stainless steel. Most fracture surface of specimen S2 exhibited a ductile dimple morphology, while small intergranular cracking with a depth about 50 μm were only observed near the specimen surface (Fig. 10 (d)).

Fig. 10.   SEM observations of cracking behavior of the 316 L-308 L similar metal weld on specimen S2: (a, b) surface morphology following 20% strain by SSRT, (c) an overall observation of fracture surface after SSRT test, (d) observation at higher magnification showing the cracking morphology, (e) schematic drawing showing the location of facture surface on specimen S2.

3.2.3. Cracking behavior of the 308 L-309 L similar metal weld

Backscattered electron image of the surface of specimen S3 following 10% and 20% interrupted strains of SSRT with a tensile direction approximately parallel to the FB in primary PWR environment are shown in Fig. 11(a) and (b), respectively. Following 5% strain, no crack was observed in 308 L and 309 L weld metals. However, interdendritic SCC that was vertical to the tensile direction initiated in the 309 L weld metal adjacent to 308 L weld metal following 10% strain (Fig. 11(a)). A maximum crack length of more than 200 μm was observed adjacent to FB. Increasing the strain to 20% led to the both increase of crack number and length. Further, all the cracks were restricted to 309 L weld metal (Fig. 11(b)). This was confirmed by the observation on the fracture surface of the specimen after the SSRT test (Fig. 11(c) and (d)). On the fracture surface, both intergranular and brittle-like transgranular facets were observed in 309 L weld metal (Fig. 11(c)). The intergranular crack with a max length of about 500 μm propagated in 309 L weld metal and then changed to transgranular cracking in both 309 L and 308 L weld metals (Fig. 11(c) and (d)). In addition, the depth of intergranular crack in 309 L weld metal was much shorter than that in the HAZ of 316 L shown in Fig. 10(d).

Fig. 11.   SEM observations of cracking behavior of the 309 L-308 L similar metal weld on specimen S3: (a, b) SEM back-scattered electron images following 10% and 20% strains by SSRT, respectively, (c) an overall SEM observation of fracture surface after SSRT test, (d) SEM observations at higher magnification showing the cracking morphology, (e) schematic drawing showing the location of facture surface on specimen S3.

3.2.4. Cracking behavior of the 309 L-SA508 dissimilar metal weld

SEM observations on the surface of specimen S4 following 10% and 15% interrupted strains of SSRT with a tensile direction approximately parallel to the FB in primary PWR environment are shown in Fig. 12(a) and (b), respectively. Following 10% interrupted strain, no crack was observed in the FB region (Fig. 12(a)). Then the strain was increased to 15%, resulting in shallow transgranular crack initiation in the 309 L weld metal adjacent to the FB (Fig. 12(b)). However, no cracking occurred in the SA508 after the same interrupted strain. The fracture morphology of the specimen after the SSRT tests are shown in Fig. 12(c)‒(e). Brittle-like transgranular cracking was observed in the FB region, which is in consistence with the surface observation (Fig. 12(b)). In addition, intergranular SCC was also found at some isolated locations on the fracture surface of 309 L weld metal while the fracture surface of SA508 exhibited a full ductile dimple morphology (Fig. 12(e)).

Fig. 12.   SEM observations of cracking behavior of the SA508-309 L dissimilar metal weld with a tensile direction approximately parallel to the FB on specimen S4: (a, b) surface morphology following 10% and 15% strains by SSRT, respectively, (c) an overall observation of fracture surface after SSRT test, (d) observation at higher magnification showing the transgranular cracking morphology, (e) observation at higher magnification showing the intergranular cracking morphology, (f) schematic drawing showing the location of facture surface on specimen S4.

The cracking behavior of the surface of specimen S5 following SSRT with a tensile direction approximately perpendicular to the FB in primary PWR environment was observed by SEM. Increasing the strain to 15% resulted in the initiation of intergranular crack at type II boundaries in the FB region of 309 L-SA508 dissimilar metal weld (Fig. 13(a)). However, since the type II boundary was discontinuous along the FB, transgranular cracking was also observed in 309 L weld metal (Fig. 13(b)). Further, increasing the strain to 20% resulted in the initiation of necking and transgranular cracking in 308 L weld metal (Fig. 13(c) and (d)). In addition, no cracking initiation in 309 L and SA508 was observed on the specimen surface.

Fig. 13.   SEM observations on the surface of the FB region of the SA508-309 L/308 L dissimilar metal weld with a tensile direction approximately perpendicular to the FB following 15% strain by SSRT on specimen S5: (a) the intergranular cracking morphology in the FB region, (b) the transgranular cracking morphology in the FB region, (c) necking in 308 L weld metal, (d) transgranular cracking morphology in 308 L weld metal.

4. Discussion

4.1. SCC susceptibility of various regions of the SA508-309 L/308 L-316 L weld joint

Table 2 summarized the minimum strain for crack initiation and the percentage of different failure modes obtained by SEM observations of various regions of the specimens with the tensile direction parallel to the FB. The minimum strain for crack initiation in 309 L weld metal, the HAZ of 316 L and 308 L weld metal were 5%-10%, 15%-20% and 15%-20%, respectively. The dominant failure mode was SCC for 309 L weld metal, which showed a 18.3% of intergranular area and 56.4% of transgranular area on the fracture surface (Fig. 11(c)). For the HAZ of 316 L, although most of fracture surface exhibited a ductile dimple morphology, small intergranular and transgranular cracking areas were also evident (Fig. 10(d)). However, only a small percentage of transgranular cracking area was observed on the fracture surface of 308 L weld metal (Fig. 9(d)). In addition, the SA508 showed a fully ductile failure on the fracture surface, indicating a low SCC susceptibility. Therefore, the SCC susceptibility at various regions of the dissimilar metal weld joint follows the order of SA508 < 308 L weld metal < the HAZ of 316 L base metal < 309 L weld metal by assessing the minimum strain for crack initiation in conjunction with the area percentage of different failure mode. In addition, SCC only occurred along the type II boundary of the specimen with the tensile direction perpendicular to the FB, indicating the high SCC susceptibility of 309 L weld metal as well. These results suggest that the 309 L weld metal of the safe-end weld joint is most prone to SCC under both axial and circumferential loads during the service of a nuclear power plant.

Table 2   The minimum strain for crack initiation and the area percentage of different failure modes on the fracture surface of various regions of the SA508-309 L/308 L-316 L weld joint following SSRT tests.

MaterialInitiation strainSCCDuctile fracture area percentage
Intergranular area percentageTransgranular area percentage
309 L5%-10%18.3%56.4%25.3%
HAZ of 316 L15%-20%few6.2%93.0%
308 L15%-20%0few~100%
SA508-00100%

New window

4.2. Correlation of microstructure and SCC susceptibility of different regions in the SA508-309 L/308 L-316 L weld joint

SCC is a synergistic effect of an aggressive environment, a susceptible material and the stress/strain. Due to the same environment and strain for each interrupted test, the material factor such as chemical composition, microstructure (grain boundary microstructure and chemistry, carbide precipitation, grain boundary character distribution, etc.) and residual strain varies the SCC susceptibility of different regions of the weld joint.

With regard to SA508, the chemical composition is the dominant factor causing its lower resistance to SCC than the stainless steels. As reported in literature, the oxidation rate of SA508 was very high in primary PWR environment [4,34,35]. If a microcrack initiates during the early stage of SSRT test, the crack tip will be instantly blunted due to the high oxidation rate of SA508, resulting in a lower degree of stress concentration [21,32]. Therefore, SA508 has a high SCC resistance in primary PWR environment with DH, which is consistent with Refs. [4,34]. However, increasing the stress intensity factor to more than 60 MPa m1/2 or adding an amount of Cl- (>100 μg/L) could induce SCC in high temperature water with DH [14,34].

SSRT tests indicated that the SCC susceptibility was different among the stainless steels, which is related to the variation of chemical composition and microstructure. Since the γ austenite and δ ferrite phase showed similar compositions in 308 L and 309 L weld metals, the difference in microstructure is the primary material factor influencing the SCC susceptibility. As mentioned previously, the 308 L weld metal shows a higher content of δ ferrite phase than the 309 L weld metal. The Cr-enriched δ ferrite phase could decrease the oxidation rate in high temperature water, from a more protective Cr-rich inner oxide layer and thus lower the SCC susceptibility of stainless steel [12,36,37]. As such, the higher SCC resistance of 308 L weld metal than 309 L weld metal should be caused by the higher content of δ ferrite phase, since the 308 L and 309 L weld metals were similar in grain boundary character distribution and grain boundary chemistry (Figs. 3‒8).

Further, the SCC is the most likely initiated in the 309 L weld metal adjacent to 308 L weld metal and then propagated into the bulk 309 L weld metal. Results of microstructure characterization suggests the precipitation of carbides plays a role in SCC initiation. It was proposed that the grain boundary sliding was one of the key steps for SCC in high temperature [38,39]. The carbides could act as a barrier to prevent grain boundary sliding and migrating, lowered the intergranular oxidation rate around the carbides and therefore increased the SCC resistance of stainless steel in primary PWR environment [[38], [39], [40]]. This beneficial effect in suppressing SCC were also observed in Alloy 600 and 690 [40,41]. As shown in Fig. 6, Fig. 7, there were few carbides in the 309 L weld metal adjacent to 308 L weld metal while a semi-continuous distribution of carbides was observed in the 309 L bulk weld metal. Further, the precipitation of carbides did not lead to an obvious Cr depletion, which was detrimental to SCC resistance. Therefore, the 309 L weld metal adjacent to 308 L weld metal is the most likely cracking region in primary PWR environment.

The SCC susceptibility of the HAZ of 316 L stainless steel was higher than that of 308 L weld metal, as shown in Fig. 10. This is consistent with the results of stress corrosion crack growth tests in the previous studies [9,10]. The TEM analysis of the 316 L stainless steel in the weld joint indicated the Cr content of γ phase was similar to that in 308 L weld metal [9], whereas the 308 L weld metal contained 10.9 vol.% of δ ferrite phase and a few carbides. Since the residual strain in 308 L weld meatal was similar to that in the HAZ of 316 L stainless steel [26,42], the absence of δ ferrites and carbides led to a higher SCC susceptibility of the HAZ of 316 L stainless steel than that of 308 L weld metal as discussed above.

The low SCC resistance of type-II boundary adjacent to the FB has been reported in many Refs. [6,13,18,19]. As shown in Fig. 3, the high hardness and increase of Fe but decrease of Cr, Ni concentration could reduce the SCC resistance of type-II boundary in 309 L weld metal. In addition, the type-II boundary is a long straight random high angle grain boundary, which is more susceptible to SCC than coincidence site lattice grain boundary and low angle boundary. Therefore, initiation of SCC was observed along the type II boundary in the FB region of 309 L-SA508 dissimilar metal weld with a tensile direction perpendicular to the FB.

5. Conclusions

This study investigated the microstructure and SCC behavior of the SA508-309 L/308 L-316 L dissimilar metal weld joint in simulated primary PWR environment. The following conclusions can be drawn from this investigation:

(1) The 308 L weld metal shows a higher content of δ ferrite phase but a lower residual strain than the 309 L weld metal.

(2) No obvious Cr-depletion but carbides precipitation at γ/δ phase boundaries is observed in 308 L and 309 L weld metals.

(3) The SCC susceptibility of the base and weld metals of the weld joint follows the order of SA508 ≤ 308 L weld metal < the HAZ of 316 L base metal < 309 L weld metal.

(4) The higher SCC susceptibility of 309 L weld metal is due to the lower content of δ ferrite phase.

(5) Few carbides in the 309 L weld metal adjacent to 308 L weld metal is a cause for the preferential SCC initiation in this region.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (No.51571204). One author (Lijin Dong) would acknowledge the support of the Youth Scientific and Innovation Research Team for Advanced Surface Functional Materials (Southwest Petroleum University, No. 2018CXTD06) and the Young Scholars Development Found of Southwest Petroleum University (No. 201899010040).


/