Characterization of internal and intergranular oxidation in Alloy 690 exposed to simulated PWR primary water and its implications with regard to stress corrosion cracking

Materials Characterization 157 (2019) 109922

Contents lists available at ScienceDirect

Materials Characterization journal homepage: www.elsevier.com/locate/matchar

Characterization of internal and intergranular oxidation in Alloy 690 exposed to simulated PWR primary water and its implications with regard to stress corrosion cracking

T

⁎

Yun Soo Lim , Dong Jin Kim, Sung Woo Kim, Seong Sik Hwang, Hong Pyo Kim Safety Materials Technology Development Division, Korea Atomic Energy Research Institute, Republic of Korea 989-111 Daedeok-daero, Yuseong-gu, Daejeon 34057, Republic of Korea

A R T I C LE I N FO

A B S T R A C T

Keywords: Nickel-base alloy Alloy 690 STEM Internal oxidation Intergranular oxidation Stress corrosion cracking

Oxidation testing of Ni-base Alloy 690 was conducted in simulated PWR primary water, and the internal and intergranular (IG) oxidation layers were systematically characterized by analytical transmission electron microscopy to obtain insight into its IG stress corrosion cracking (IGSCC) resistance and behavior. The internal oxidation layer consisted of upper NixFe1-xCr2O4 and innermost Cr2O3. Each internal oxidation layer had a crystallographic orientation relationship with the Alloy 690 matrix, indicating that the layers formed through solid-state reactions between inward diffusing oxygen and the alloying elements. Strong evidence that the Cr2O3 layer formed through the outward grain boundary diffusion of Cr was found, leaving severe Cr depletion along the grain boundary with a steep composition gradient toward the surface. The innermost Cr2O3 layer had a continuous and compact band-shaped morphology around the surface grain boundary. This Cr2O3 layer was revealed to be protective of the inward grain boundary diffusion of oxygen from the surface and therefore responsible for the prevention of grain boundaries from IG oxidation. IG Cr carbides suppressed the internal and IG oxidation further. The key findings obtained in this work may present the main reason for the high resistance of Alloy 690 to IGSCC under PWR normal operating conditions.

1. Introduction Primary water stress corrosion cracking (PWSCC) of Ni-base Alloy 600 (Ni-16Cr-8Fe in wt%) has been a major concern in the primary side of pressurized water reactors (PWRs) [1]. In response to the cracking problems associated with Alloy 600, another solid-solution strengthened Ni-base Alloy 690 (Ni–30Cr–10Fe in wt%) has become the common replacement material for use in PWR service. Alloys 600 and 690 have an identical crystal structure and similar mechanical properties; however, there are noticeable differences in the corrosion resistance and cracking behavior between them owing to their different Cr contents. First, Alloy 690 [2] is well known to be more resistant to PWSCC than Alloy 600 [3]. Moreover, laboratory testing [4,5] and service experiences [6,7] have clearly shown that the predominant failure mode of Alloy 600 in PWR primary water is nearly always intergranular (IG) SCC (IGSCC), which means that grain boundaries are the preferential paths for cracking. However, PWSCC in Alloy 690 is not necessarily IG cracking under the same testing conditions; rather, it shows a mixed mode consisting of IG and transgranular cracks in many

⁎

cases [8–10] except for a few examples, the testing of which involved severely cold-rolled samples under accelerated experimental conditions [11–13]. These findings indicate that grain boundaries do not appear to be always operative for cracking in the case of Alloy 690. It is necessary, therefore, to reveal the root causes of the different cracking behaviors of Alloys 600 and 690 in PWR primary water to ensure safe service and good performance. Recently, the internal oxidation hypothesis originally proposed by Scott and Calvar [14] as a possible IGSCC mechanism for Alloy 600 has attracted much attention by virtue of the progress in analysis technologies using advanced microscopic equipment such as transmission electron microscopy (TEM) and atom probe tomography (APT). The basic concept of internal oxidation is that discrete oxides are produced from more reactive alloying elements by interacting with penetrating oxygen [15]. Two types of changes occur in Ni-base alloys owing to the inward diffusion of oxygen. Oxygen reacts selectively with the solute metal Cr beneath the surface to form discrete Cr-rich oxide particles via a process referred to as ‘internal oxidation’. The surface grain boundaries are also oxidized by oxygen penetration along the grain

Corresponding author. E-mail address: [email protected] (Y.S. Lim).

https://doi.org/10.1016/j.matchar.2019.109922 Received 26 April 2019; Received in revised form 11 August 2019; Accepted 5 September 2019 Available online 06 September 2019 1044-5803/ © 2019 Elsevier Inc. All rights reserved.

Materials Characterization 157 (2019) 109922

Y.S. Lim, et al.

direction. Detailed information pertaining to the microstructure and cracking behavior of the test alloy are described in the literature [10,32]. This alloy was specially selected for an oxidation test because IG Cr carbides were distributed at different sizes and different densities depending on the grain boundary. Therefore, the influence of IG Cr carbides on the internal and IG oxidation could be easily visualized from a single heat sample. Oxidation coupons (10 × 10 × 2 mm3) cut from the round bar were prepared by grinding with SiC paper to 2000 mesh and subsequently polishing with alumina powders down to 0.3 μm. The specimens were cleaned with methyl alcohol and subsequently rinsed with deionized water before the immersion test. The simulated PWR water was prepared prior to the test in a storage tank. 1200 ppm B (by weight) of H3BO3 and 2 ppm Li (by weight) of Li(OH) were added to pure water. After the removal of the dissolved oxygen by purging with nitrogen gas, hydrogen was added to the simulated PWR primary water by purging the gas. The dissolved oxygen concentration was maintained at < 5 ppb during the test. A 316L stainless steel autoclave with a volume of 3.8 l was used in a recirculating mode. The flow rate of the loop system was approximately 2.5 l per hour. The specimens were exposed to PWR primary water simulating normal operating conditions, that is, at a temperature of 325 °C, and at a dissolved hydrogen concentration of 30 cm3 H2/kg H2O for a period of 3600 h. The combination of the temperature and the hydrogen concentration used for the present study corresponds to the condition below the Ni/NiO equilibrium electrode potential at which the solvent metal, Ni, of this alloy is in the Ni-metal regime [33]. The conductivity, dissolved oxygen concentration, pH and hydrogen concentration were continuously monitored at room temperature. Thin-foil TEM specimens from the oxidized coupons were prepared by performing a standard lift-out procedure using a focused ion beam (FIB) milling technique. A dual-beam Hitachi FIB-2100 system was used to prepare cross-section TEM specimens with Ga ion sputtering after the deposition of protective Pt in order to characterize the internal oxidation layer and the IG penetration of oxygen at the grain boundaries. Thinning of the specimens to electron transparency was conducted using Ga ions at successively lower beam energy levels of 30 kV, 10 kV, and 5 kV. Subsequent ion milling was performed using Ar ions with an incident beam energy level of 300 V at an incidence angle of 10° for 10 to 30 min to eliminate the deformed surface layers generated by the sputtering of high-energy Ga ions. The oxidized specimens were investigated using various types of microscopic equipment. A conventional TEM analysis of the crystallography was conducted with a JEOL JEM-2100F (operating voltage 200 kV), and HRTEM images were obtained on a CCD camera and analyzed using the fast Fourier transform (FFT) technique to investigate additional crystallographic details of the oxide layer. Composition mapping around an oxidation layer was performed by STEM EELS with a GIF Quantum ER System (Model 965) in an aberration-corrected JEM ARM 200F (operating voltage 200 kV). The energy resolution was 0.3–0.6 eV. STEM EELS was combined with the high-angle annular dark-field (HAADF) technique to produce chemical contrast images formed using electrons diffracted at a high angle (> 40 mrad), with the intensity proportional to the Z atomic number of the elements composing the sample. The thickness of the local areas where the oxidation layers of the TEM specimens were analyzed was determined from the EELS by measuring the ratio of the zero loss peak intensity to that of the total spectrum [34], and the thicknesses of the locations of interest under study were estimated to range between 75 and 95 nm. An EDS analysis was carried out for line profiling and point detection around the oxidation layer with a JEOL JEM-2100F (operating voltage 200 kV) equipped with an Oxford Instruments X-max80T Silicon Drift Detector with a solid angle of 0.2361 sr and an AZTEC analysis system (Ver. 3.1b). There is intrinsic overlap between the peaks of the oxygen K line and the Cr Lα line. Therefore, deconvolution of the overlapping spectral peaks was carried out for a clear elimination of artifacts possibly

boundaries and by short-circuit outward diffusion of Cr at the grain boundaries, resulting in grain boundary oxidation, which is referred to as ‘IG oxidation’. Given that the PWSCC of Alloy 600 undoubtedly reveals IG cracking, IG oxidation theory [16–19] successfully explains that the grain boundaries of Alloy 600 become brittle when they are oxidized [20,21] and that Alloy 600 is therefore highly susceptible to IGSCC when exposed to PWR primary water. However, the cracking behavior of Alloy 690 is quite different from that of Alloy 600. Accordingly, it appears that a reasonable explanation should be made with regard to these differences so as to guarantee the appropriateness of the internal and IG oxidation as a PWSCC mechanism of Ni-base alloys. Many studies of the internal and IG oxidation of Alloy 690 have been conducted [11–13,22–31], and it is generally agreed that the internal oxidation layer of Alloy 690 is a dual layer made up of an upper spinel consisting of a mixed nickel and iron chromite (NixFe1−xCr2O4) and an innermost Cr2O3 component [27–31]. However, the detailed morphologies of the innermost Cr2O3 appear to vary depending on factors such as the specimens used and experimental conditions applied [11,27–29,31]. The innermost Cr2O3 in the bulk grain was reported to have a non-compact and discrete distribution at the oxidation layer/ matrix interface in the form of the nano-sized nodules [27,28], crystallites [29], and filaments [11,31] when precipitating away from a grain boundary. Interestingly, Olszta et al. [22] revealed that a compact thin Cr2O3 layer formed at the intersection of grain boundaries with the exposed surface, which indicates that the diffusion kinetics of Cr at the grain boundary differs from that away from the grain boundary. The internal oxidation layer is fundamental and crucial for SCC because it provides a barrier that prevents further oxidation of the matrix. In spite of the importance of internal and IG oxidation in elucidating the SCC phenomena, there remains a shortage of information in the literature about how internal and IG oxidation layers form and how they affect the SCC resistance and behavior of Alloy 690. Only limited studies on this subject have been reported [12,13,22,25]. The aim of the present study was to characterize the internal and IG oxidation phenomena of Alloy 690 microscopically in a systematic manner to obtain clear insight into the root cause of the different resistance characteristics to PWSCC and the cracking behavior of Alloy 690 compared to those of Alloy 600 when exposed to primary water simulating the normal operating conditions of a PWR. Because IG Cr carbides are known to affect IG cracking by changing the oxygen diffusion behavior [16–18], their influence on the IG oxidation is also taken into account. The nature and the structure of the internal oxides and the microchemical changes near the surface and around the grain boundaries due to oxygen diffusion were precisely characterized using various types of microscopic equipment, in this case, scanning electron microscopy (SEM), TEM, high-resolution TEM (HRTEM) imaging, fineprobe chemical analysis using energy dispersive X-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS) in the scanning TEM (STEM) mode. Finally, possible correlations between the IG oxidation and the SCC behavior of Alloy 690 are discussed on the basis of the observed results. 2. Experimental procedures A forged round bar of Alloy 690 (Heat No. 135264) for a control rod drive mechanism nozzle of a PWR was used in this study. The chemical composition of the test alloy is shown in Table 1. The as-received Alloy 690 has a banded microstructure of intragranular Cr carbides and small-sized grains in a matrix of normal-sized grains in the axial Table 1 Chemical composition of the test alloy. Cr

C

Fe

Ni

Mn

Si

Ti

Al

S

P

29.71

0.014

8.80

60.10

0.21

0.32

0.24

0.33

< 0.001

0.003

2

Materials Characterization 157 (2019) 109922

Y.S. Lim, et al.

Fig. 1. (a) STEM image, (b) corresponding HAADF, and EELS spectrum images of (c) O, (d) Cr, (e) Fe and (f) Ni around the internal oxidation layer and the surface grain boundary without IG Cr carbides of Alloy 690.

structure and properties of this Cr oxide layer appear to differ from those of the rest of the internal oxidation layer, which is a mixture of Ni, Fe and Cr oxides. The most important feature is that the Cr oxide layer appears to act as a barrier against oxygen diffusion into the matrix, and its blocking effect on oxygen diffusion is more prominent around the surface grain boundary. Oxygen diffusion along the grain boundary was interrupted after a short penetration depth of approximately 50 nm from the surface. This depth of oxygen diffusion at the grain boundary was much shallower than that at the surrounding matrix, which exceeded 100 nm on average. As a result, the Cr oxide layer suppressed the internal oxidation and more importantly the IG oxidation of this alloy by preventing oxygen from diffusing into the alloy. Another finding which should be noted is that the grain boundary below the Cr oxide layer was severely Cr-depleted, as shown in Fig. 1(d). Cr which diffuses out from the grain boundary should produce some Cr compounds in or around the grain boundary. Only the Cr oxide layer was visible around the grain boundary. Therefore, it is reasonable to consider that the formation of the Cr oxide layer originated from the diffusion of Cr in the grain boundary and resulted in Cr depletion and corresponding Ni enrichment at the grain boundary. The Cr oxide layer has a band-shaped morphology in a continuous form only around the grain boundary, whereas this layer does not show continuity away from the grain boundary. Olszta et al. [22] reported a similar finding, demonstrating that a protective Cr2O3 film formed only around a surface grain boundary, leaving a Cr-depleted grain boundary in Alloy 690 exposed to PWR primary water. A crystallographic examination was carried out to identify the crystal structures and the orientation relationships of the internal oxidation layer with the Alloy 690 matrix. Fig. 2(a) shows a TEM brightfield image and selected-area diffraction patterns (SADPs) taken from several areas in the internal oxidation layer and the surrounding

originating from the overlapping of the peaks. More detailed information on the EDS measurement conditions is available in the literature [19]. 3. Results 3.1. Internal and IG oxidation of Alloy 690 without IG Cr carbides Fig. 1 shows a STEM image (Fig. 1(a)), the corresponding HAADF image (Fig. 1(b)) and EELS composition maps of O, Cr, Fe and Ni (Fig. 1(c)–(f)) around a grain boundary just beneath the surface. The grain boundary in this figure has no IG Cr carbides near the surface; therefore the reaction of IG Cr carbide with diffused oxygen can be ignored in this case. A sulfide which appears to be an inclusion is accidently visible on the grain boundary at some distance below the surface; this is disregarded henceforth as well. From Fig. 1, it can be identified that the overall thickness of the internal oxidation layer exceeded 100 nm, much thicker than that of Alloy 600, which consists of a thin layer of about several tens of nanometers [16,19,26]. In the internal oxidation layer, the intensity of the oxygen peak was strong (Fig. 1(c)), and Cr was enriched (Fig. 1(d)); however, Fe was slightly depleted (Fig. 1(e)) and Ni was severely depleted (Fig. 1(f)). Several intriguing features are found related to the internal and IG oxidation of Alloy 690. First, there is another continuous and thin bandshaped oxide layer with a random distribution below the internal oxidation layer, roughly enclosing the upper internal oxidization layer (denoted as ‘Cr oxide layer’ in the figures). The thickness of the Cr oxide layer ranged roughly from 10 nm to 30 nm. Inside the Cr oxide layer, Cr is more enriched, whereas Fe and Ni are nearly depleted compared to those in the remaining the internal oxidation layer. All of these findings indicate that the layer consists of Cr-rich oxide. Therefore, the oxide 3

Materials Characterization 157 (2019) 109922

Y.S. Lim, et al.

Fig. 2. (a) TEM image with diffraction patterns taken from grains A and B, the upper internal oxidation layer C, and the innermost Cr oxide layer D; and (b) and (c) the indexed results of diffraction patterns C and D in (a), respectively.

reveal the orientation relationship between the innermost Cr2O3 layer and the Alloy 690 matrix more clearly. The HR lattice image was obtained from the region enclosed by the white dotted circle in Fig. 2(a), in which several different phases exist. The corresponding FFT diffractograms of the different areas, A, B, C and D denoted by dotted rectangles in the HR lattice image are also shown in Fig. 3. The FFT diffraction patterns obtained from the Alloy 690 matrix (A) and the upper NixFe1−xCr2O4 spinel (B) are identical to the SADPs with a beam direction of [011] zone axis of the Alloy 690 matrix shown in Fig. 2 and therefore clearly identify the phases from which the diffraction patterns originated. The FFT diffractogram obtained from the Cr2O3 layer (C) showed a diffraction pattern with a beam direction of [2īī0] zone axis of Cr2O3, which indicates that the direction of [011] zone axis of Alloy 690 is parallel to that of [2īī0] zone axis of the discrete Cr2O3 under examination. The FFT diffractogram D was obtained from the area containing the Cr2O3 layer and the surrounding Alloy 690 matrix. The (0006) lattice planes of Cr2O3 were parallel to the (ī1ī) lattice planes of Alloy 690 in the indexed results. Therefore, it appears that there is an orientation relationship between the Cr2O3 layer and the Alloy 690 matrix such as (ī1ī)M//(0006)C and [011]M//[2īī0]C. As a result, it is obvious from Figs. 2 and 3 that the upper NixFe1−xCr2O4 spinel and the innermost Cr2O3 have strict orientation relationships with the Alloy 690 matrix. This indicates that the internal oxides form through solid-state reactions between the diffused oxygen and the alloying elements. Identical results with regard to the crystallography of the internal oxidation layer of Alloy 690 were also obtained in the previous study [31]. An EDS point detection analysis was conducted to reveal more quantitatively the chemical compositions of the different types of oxides in the internal oxidation layer shown in Fig. 1. The EDS point detection locations are shown in Fig. 4(a) and the results are given in Table 2. In Fig. 4(a), the point detection positions of ‘P1’and ‘P2’ are located in the upper oxide layer, which was identified as the NixFe1−xCr2O4 type of spinel, while the point detection positions of ‘P3’ and ‘P4’ are situated in the innermost Cr oxide layer, which was identified as Cr2O3 from the analyses of the SADPs in Fig. 2 and of the FFT diffractogram in Fig. 3. From the chemical compositions of P1 and P2 in Table 2, the stoichiometric ratio of Cr:(Ni + Fe) in the major metallic elements was found to be approximately 2:1, indicating that the spinel is composed of NixFe1-xCr2O4. The chemical compositions of P3 and P4 were quite different from those of P1 and P2. Cr was dominant as the main metallic

regions. In Fig. 2(a), the boundary of the upper internal oxidation layer is roughly indicated by black dotted lines, and the area of the innermost Cr oxide layer below the upper internal oxidation layer is shown in light red. The letters A, B, C and D in the inserted SADPs indicate the diffraction patterns taken from two grains (A and B), the upper internal oxidation layer (C) and the innermost Cr oxide layer (D). The TEM specimen was tilted so that the [110] zone axis of grain A was aligned toward the electron beam direction. Comparing diffraction pattern A with diffraction pattern B, it is evident that the grain boundary is a random high-angle grain boundary. Diffraction pattern C contains both the diffraction spots from the upper internal oxidation layer and those from grain A. The indexed results of diffraction pattern C are shown in Fig. 2(b). In this figure, the subscripts M and S denote the Alloy 690 matrix (M) and the upper internal oxidation layer (S), respectively. From the analysis of the diffraction pattern, it was identified that the upper internal oxidation layer had a spinel structure with a lattice constant of about 0.83 nm. There are several types of spinel with similar lattice constants, such as NiFe2O4 (lattice constant of 0.8337 nm, JCPDS 86-2267) and NiCr2O4 (lattice constant of 0.8316 nm, JCPDS 23-1271). The most probable crystal structure would be the NixFe1−xCr2O4 type of spinel considering the results of a subsequent compositional analysis. From the diffraction pattern analysis, it is evident that there is a cube-cube orientation relationship between the Alloy 690 matrix and the spinel, in this case expressed as [100]M//[100]S and (100)M//(100)S. Diffraction pattern D contains the spots from the innermost Cr oxide layer, the upper internal oxidation layer and the Alloy 690 matrix as well. The indexed results of diffraction pattern D are shown in Fig. 2(c). In the figure, the faint but recognizable spots originating from the innermost Cr oxide layer are enclosed by yellow solid circles. The ring patterns originating from Cr2O3 are also overlapped in the figure. It was clearly identified that most of the spots from the innermost Cr oxide layer were on the ring pattern of Cr2O3, meaning that the innermost Cr oxide layer is Cr2O3. Cr2O3 (chromia) has a hexagonal unit cell with lattice constants of a = 0.4957 nm and c = 1.3952 nm (JCPDS 821484). Therefore, it can be concluded that the internal oxidation layer around the surface grain boundary of Alloy 690 is a dual layer consisting of a thick upper NixFe1−xCr2O4 type of spinel and innermost band-shaped Cr2O3 and that this continuous Cr2O3 layer is responsible for the prevention of oxygen diffusion into the grain boundary and the resultant suppression of IG oxidation in this alloy. Fig. 3 shows the result of the FFT diffractogram analysis conducted around the internal oxidation layer and the surrounding matrix to 4

Materials Characterization 157 (2019) 109922

Y.S. Lim, et al.

Fig. 3. HRTEM image around the internal oxidation layer and FFT diffractograms of the framed areas denoted by the letters A, B, C and D in the HRTEM image.

sized Cr2O3 nodules at the interface between the spinel and the Alloy 690 matrix. Masaki et al. [29] investigated the corrosion behavior of Alloy 690TT tubing in primary water at 325 °C maintaining a hydrogen concentration of 2.6 ppm and a flow velocity in the tubing of 1.7 m/s; they also suggested a dual internal oxidation layer consisting of NiCr2O4 spinel and innermost Cr2O3 crystallites which were both very thin at about 15 nm under the condition of a high flow velocity. Finally, Kuang et al. [31] conducted constant extension rate tensile tests in 360 °C high-purity water containing 18 cm3 H2/kg H2O, and found Cr2O3 filaments discretely spaced beneath the inner (Fe,Cr,Ni)3O4 layer. Therefore, the morphologies of the innermost Cr2O3 appear to vary drastically depending on the experimental conditions applied. EDS line profiling was conducted to reveal the microchemical variations around the internal oxidation layer and surface grain boundary due to the oxygen diffusion shown in Fig. 1. The locations and

element, whereas the amounts of Ni and Fe are negligible, meaning that the innermost Cr oxide layer is mainly composed of Cr2O3. Therefore, it can be concluded once again from the analysis of the chemical compositions that the upper spinel is NixFe1-xCr2O4 and that the innermost Cr oxide is Cr2O3. Similar crystal structural but different morphological results pertaining to the internal oxide layer of Alloy 690 have been reported [27–29,31]. Sennour et al. [27] conducted an oxidation test in primary water at 325 °C with a H2 concentration of 1.3 × 10−3 mol/L for relatively short durations of 66 to 858 h and found that the internal layer of Alloy 690 was mainly composed of a continuous upper spinel layer of mixed iron and nickel chromite in the thickness range of 1–100 nm, with an innermost layer consisting of discrete nodules of Cr2O3 approximately 5 nm in size. Lefaix-Jeuland et al. [28] conducted tests similar to those of the preceding study, finding that the defects on the alloy surface induced a large increase in the density of the nano-

Fig. 4. (a) STEM image of the internal oxidation layer around a surface grain boundary without IG Cr carbides of Alloy 690, and (b)–(d) compositional variations of O, Cr, Fe and Ni obtained from line profiling tests, denoted correspondingly by EDS1, EDS2 and EDS3 in (a). 5

Materials Characterization 157 (2019) 109922

Y.S. Lim, et al.

Table 2 EDS point detection results of the upper NixFe1−xCr2O4 type of spinel (P1,P2) and the innermost Cr oxide layer (P3,P4) of Alloy 690 exposed to PWR primary water at 325 °C. The positions used in the analysis are shown in Fig. 4(a). The measurement results are expressed in at.%. Position

P1 P2 P3 P4

(at.%) O

Cr

Fe

Ni

67.9 69.6 60.5 61.9

22.4 20.4 34.9 33.4

5.9 4.9 2.3 2.3

3.8 5.1 2.3 2.4

directions used in the line profiling process are denoted by the arrows in Fig. 4(a). Fig. 4(b) is the result of the line profiling denoted as ‘EDS1’, which covers the compositional changes across the internal oxidation layer. Fig. 4(c) and (d) is the line profiling results of ‘EDS2’ and ‘EDS3’ across the positions of the grain boundary at depths of approximately 140 nm and 230 nm from the surface, respectively. There are three clearly distinguishable regions in Fig. 4(b): the upper spinel, the innermost Cr2O3 and the Alloy 690 matrix. The innermost Cr2O3 layers were situated between the Alloy 690 matrices and between the upper spinel. In the innermost oxide regions, Cr and O are highly enriched, while the Ni and Fe contents most likely originated from the alloy matrix and spinel regions crossed by the transmitted electron beam. The innermost Cr2O3 layer is also present inside the Alloy 690 matrix in Fig. 4(c). The most important finding in Fig. 4(c) is that the grain boundary near the surface was severely Cr-depleted. The average Cr concentration of the as-received Alloy 690 was 29.7 wt%, and the grain boundaries were slightly Cr-depleted owing to the precipitation of IG Cr carbides. The minimum Cr concentrations at the grain boundaries ranged from 23.4 wt% to 27.5 wt% due to Cr depletion before the oxidation test. However, the minimum Cr concentration at this position of the grain boundary in Fig. 4(c) was 13.4 wt%, which was much lower than that measured before the oxidation test. The minimum Cr concentration measured at the position of the grain boundary in Fig. 4(d) was 15.5 wt%. This value is slightly higher than the outcome of 13.4 wt % shown in Fig. 4(c) obtained at a position closer to the surface; however, it is still much lower than that obtained before the oxidation test. Oxygen was not detected during the ‘EDS3’ line profiling process. There are no additional Cr compounds produced in and around the grain boundary related to Cr depletion, except for the innermost Cr2O3 layer, in Fig. 4(a). Therefore, it appears that the Cr depletion at this surface grain boundary is closely associated with the formation of the innermost Cr2O3 layer which arose due to the fast outward diffusion of Cr along grain boundaries. The Cr depletion phenomenon at the grain boundary was examined over the entire range of the grain boundary shown in the FIB TEM specimen. The purpose of doing this was to identify the Cr source for the Cr2O3 layer clearly and to reveal how deeply the grain boundary was affected by oxidation test. These results are presented in Fig. 5. Fig. 5(a) is a TEM image showing the entire range of the grain boundary related to Figs. 1–4 in the FIB TEM specimen. In this figure, IG Cr carbides are randomly distributed on the grain boundary and are indicated by arrows. The IG Cr carbides precipitated in Alloy 690 were identified as Cr23C6. The EDS line profiling locations across the grain boundary are indicated by the red dots. Fig. 5(b) shows the results of the EDS measurement in which the minimum Cr concentration at each point of the grain boundary depending on the depth from the surface is presented. The average Cr concentration of the as-received Alloy 690 and the range of the minimum Cr concentrations at the grain boundaries according to the precipitation of IG Cr carbides before the oxidation test are described in Fig. 5(b). The internal oxidation layer just over the grain boundary consisted of the upper NixFe1−xCr2O4 spinel with a thickness of about 50 nm and the innermost Cr2O3 layer with a thickness of about 10 nm. These are also drawn schematically in Fig. 5(b).

Cr/(Cr + Fe + Ni)

(Fe + Ni)/(Cr + Fe + Ni)

Identification

0.70 0.67 0.88 0.88

0.30 0.33 0.12 0.12

NixFe1−xCr2O4 (x < 1) ~Cr2O3

Fig. 5. (a) TEM image showing the entire range of the grain boundary in the FIB TEM specimen and (b) the minimum Cr concentrations measured at the grain boundary depending on the depth from the surface. The measurement positions are denoted in (a) by the red dots. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Cr depletion occurred up a depth of approximately 1500 nm from the surface. Beyond this depth, the minimum Cr concentrations along the grain boundary returned to approximately 30 wt%, which was the average Cr concentration of the as-received Alloy 690. Therefore, it appears that the original Cr depletion before the oxidation test was recovered in the regions unaffected by oxygen. Related to the grain boundary Cr depletion, two facts can be taken into consideration. First, grain boundary Cr depletion occurred only near the surface, which indicates that the Cr depletion is associated with the internal and IG oxidation processes. Second, the degree of Cr depletion gradually increased toward the innermost Cr2O3 layer. This implies that there should be an oxide which acts as a Cr sink in front of the grain boundary in which the diffused Cr is incorporated; Cr2O3 should be the Cr sink in this case. As a result, the innermost Cr2O3 layer was formed by the rapid diffusion of Cr through the grain boundary, leaving a Cr-depleted grain boundary.

3.2. Internal and IG oxidation of Alloy 690 with IG Cr carbides Fig. 6(a) and (b) shows a TEM image and the indexed result of a SADP, respectively, taken from Cr carbides in the grain boundary near the surface with a higher density than that of Fig. 1. From the diffraction pattern analysis, it was identified that the IG Cr carbides were 6

Materials Characterization 157 (2019) 109922

Y.S. Lim, et al.

Fig. 6. (a) TEM image and (b) the indexed result of a SADP taken around IG Cr carbides near the surface. Subscripts M and C in the diffraction pattern represent the Alloy 690 matrix and Cr23C6 carbides, respectively.

the absence of internal and IG oxidation just over the surface grain boundary in Fig. 7, compared to Fig. 1, are linked to the presence of IG Cr carbides on the grain boundary. EDS line profiling was conducted to confirm the microchemical variations around the internal oxidation layer and the surface grain boundary including IG Cr carbide. Fig. 8(a) is a STEM image showing the locations and the directions used for the EDS line profiling. The results are given in Fig. 8(b) and (). Fig. 8(b) is the result of line profiling denoted as ‘EDS4’ in Fig. 8(a), which contains the compositional changes across the internal oxidation layer including the upper spinel, the innermost Cr2O3 layer, and the Cr-depleted grain boundary. Fig. 8(c) is the result of ‘EDS5’, which covers the compositional variations across the Alloy 690 matrix, the surface grain boundary including the IG Cr carbide, the innermost Cr2O3 layer, and the upper spinel. Finally, Fig. 8(d) is a result of line profiling denoted as ‘EDS6’ across a grain boundary far below the internal oxidation layer, at which the grain boundary was not affected by oxygen diffusion. As expected, there are upper spinel and innermost Cr2O3 layers on both sides of the surface grain boundary in Fig. 8(b). The Ni and Fe contents inside the Cr2O3 layer most likely originate from the inclined boundaries of the Alloy 690 matrix or spinel regions crossed by the transmitted electron beam. An interesting fact here is that the grain boundary was severely Cr-depleted. The minimum Cr concentration at this position of the grain was measured and found to be only 4.6 wt%. This value is roughly 21 wt% lower than that shown in Fig. 8(d), which was obtained from the position of the grain boundary not affected by oxygen diffusion. This finding indicates that most Cr at the surface grain boundary diffused out to form a Cr2O3 layer around the grain boundary. In Fig. 8(c), IG Cr carbide is shown to be located between the Alloy 690 matrix and the Cr-depleted grain boundary. The measured minimum Cr concentration at the grain boundary adjacent to the IG Cr carbide was found to be 20.7 wt%, much higher than that presented in Fig. 8(b) but nevertheless lower than that of the intact grain boundary shown in Fig. 8(d). The thick upper NixFe1−xCr2O4 layer has a complicated compositional distribution. There are Ni-enriched regions inside the spinel oxide region, as denoted by the dotted circles in Fig. 8(c). This feature can also be clearly confirmed from the EELS map of Ni in Fig. 7(f). The occurrences of Ni enrichment in the oxidized regions [19,26,41] or Ni nodules on the surface [18,25] are very common in Nibase alloys during the internal and/or IG oxidation processes whenever oxygen diffuses into the alloy and Cr is internally oxidized to form an oxide.

Cr23C6. Cr23C6 has the same fcc structure as that of an Alloy 690 matrix, with a lattice constant of 1.965 nm [35]. Cr carbides precipitated during heat treatment of Alloy 690 are consistently identified as Cr23C6 because of its high Cr content [36]. It is evident from Fig. 6(b) that the IG Cr carbides are arranged crystallographically in a parallel orientation with one grain of the Alloy 690 matrix. It is well known that there is a cube-cube orientation relationship between IG Cr23C6 carbides and the Alloy 690 matrix such as {100}M//{100}Cr23C6, 〈100〉M// 〈100〉Cr23C6 [37,38]. The affirmative role of IG Cr carbides on the resistance to IGSCC in Alloy 600 is well known, despite the fact that the mechanism involved remains unclear and controversial. Panter et al. [16] and Persaud et al. [18] suggested that IG Cr carbides act as oxygen traps along the diffusion path, impeding the diffusion of oxygen. Bertali et al. [17] and Langelier et al. [39] showed that IG Cr carbides immobilized a grain boundary by a pinning effect during diffusion-induced grain boundary migration (DIGM), which finally has the beneficial effect of reducing susceptibility to IGSCC. Bertali et al. [40] found that the free Cr released from the decomposed IG Cr carbides due to the reaction with diffusing oxygen could promote the formation of a protective Cr-rich surface oxide. Sennour et al. [41] also found that the presence of IG carbides considerably reduces the cracking rate with strong limitations on crack branching and penetration. On the other hand, Bruemmer et al. [42] proposed that IG carbides could serve as effective dislocation sources which promote crack tip blunting, thus increasing the cracking resistance. In the present study, the goal was to gain insight into how IG Cr carbides affect the internal and IG oxidation of Alloy 690. Fig. 7 shows the internal and IG oxidation phenomena obtained around the surface grain boundary of Fig. 6. Fig. 7(a) is a STEM image of the area around the surface and Fig. 7(b) is an HAADF image taken from the dotted rectangle in Fig. 7(a). Fig. 7(c)–(f) presents corresponding EELS maps of O, Cr, Fe and Ni. A dense distribution of Cr carbides is clearly visible along a grain boundary in an EELS map of Cr (Fig. 7(d)). Similar to Fig. 1, there is a continuous and thin innermost Cr2O3 layer between the NixFe1−xCr2O4 spinel and the Alloy 690 matrix, roughly enclosing the internal oxidization layer. It is clear that Cr was more enriched (Fig. 7(d)), whereas Fe (Fig. 7(e)) and Ni (Fig. 7(f)) were more depleted in the Cr2O3 layer than in the rest of the internal oxidation layer. On the grain boundary, IG Cr carbides are visible in a semi-continuous distribution. The grain boundary was severely Cr-depleted, and correspondingly Ni was enriched. The noticeable finding here is that the protective innermost Cr2O3 layer extends to the surface; therefore, the upper internal NixFe1-xCr2O4 spinel is absent just over the grain boundary, as denoted by the dotted circles in Fig. 7(c) and (d). This finding clearly indicates that the protective Cr2O3 layer formed at the grain boundary with a high density of IG Cr carbides in the initial stage of the exposure to primary water and subsequently prevented oxygen from diffusing further into the material. Therefore, it can be concluded that the earlier formation of the protective Cr2O3 layer and

4. Discussion Fig. 9 summarizes the structure of the internal oxidation layer in Alloy 690 exposed to primary water at 325 °C based on the results obtained from the present study. The features of interest are highlighted in the figure with explanations. The first novel finding in the present 7

Materials Characterization 157 (2019) 109922

Y.S. Lim, et al.

Fig. 7. (a) STEM image, (b) corresponding HAADF, and EELS spectrum images of (b) O, (c) Cr, (d) Fe and (e) Ni around the internal oxidation layer and the surface grain boundary with IG Cr carbides of Alloy 690. (b)–(f) were obtained from the dotted rectangle in (a).

the prevention of oxygen diffusion into the grain boundaries. Secondly, strong evidence was found that a Cr2O3 layer had formed through the outward diffusion of Cr along a grain boundary, resulting in severe Cr depletion in the grain boundary with a steep composition gradient toward the surface (Fig. 5). Lastly, the upper NixFe1-xCr2O4 layer was

study is that the IG oxidation was significantly suppressed owing to the continuous band-shaped innermost Cr2O3 layer below the upper NixFe1−xCr2O4 (Fig. 1). In contrast, internal oxidation into the bulk grains was promoted, resulting in the formation of a relatively thick internal oxidation layer. The innermost Cr2O3 layer was responsible for

Fig. 8. (a) STEM image of the internal oxidation layer around a surface grain boundary with IG Cr carbides of Alloy 690, (b) and (c) compositional variations of O, Cr, Fe and Ni obtained from the line profiling denoted as EDS4 and EDS5 in (a), respectively, and (d) compositional variations of Cr, Fe and Ni across an intact grain boundary far below the internal oxidation layer denoted as EDS6 in (a).

8

Materials Characterization 157 (2019) 109922

Y.S. Lim, et al.

oxide–metal interfaces [44] and the incoherent oxide-metal interfaces themselves can provide a short-circuit fast diffusion paths for the continuous growth of Cr2O3 by enhancing the diffusion rates. Finally, a continuous band-shaped Cr2O3 layer forms as a protective layer against oxygen diffusion into the grain boundaries, as shown in Fig. 9. Other studies also demonstrated that grain boundaries in Alloy 690 act as fast diffusion paths for Cr to form a protective Cr2O3 layer at the surface, suppressing IG oxidation not only in high-temperature water [13,22] but also in hydrogenated steam [25]. However, the duality of the internal oxidation layer consisting of the NixFe1−xCr2O4 spinel and an innermost Cr2O3 layer around a grain boundary as found here was not explicitly shown in earlier works, possibly due to the different experimental conditions. There is a notable difference in the internal oxidation layers between around and away from the grain boundary (Figs. 1 and 7). The internal oxidation layer just over the grain boundary is very thin, though it becomes much thicker as it deviates from the grain boundary. This unique morphology of the internal oxidation layer around a grain boundary in Alloy 690 has been reported in previous studies [13,22,25]. Langelier et al. [45] studied the internal oxidation process in Alloy 600 exposed to 480 °C hydrogenated steam using APT and found preferential segregation of Cr and oxygen for fine discrete Cr oxide particles near the bottom of the internally oxidized layer. They also observed fine discrete Cr2O3 surrounded by FeCr2O4 spinel oxides. From their results, they suggested that the internal oxides preferentially precipitate as discrete Cr2O3 initially, and then start to form as FeCr2O4 spinel oxides, which is more thermodynamically favorable where the oxygen activity is decreased and when Cr was depleted locally around the surroundings of the preexisting Cr2O3. Kuang et al. [31] also suggested a similar formation mechanism for the inner oxides in a 20% cold-rolled Alloy 690 tested in 360 °C high-purity water containing 18 cm3 (STP) H2/kg H2O. They claimed that a compact Cr2O3 layer is unable to develop when inside grains exist at a low defect density and a low temperature, possibly due to the extremely low lattice diffusion rate of Cr. Discrete Cr2O3 will initially form along the widely spaced lattice planes inside the grains by virtue of fast diffusion paths, such as dislocations. In such a case, oxygen can diffuse more deeply into the grains given the non-compact and discrete Cr2O3, which cannot protect oxygen from deeply diffusing into the grains. As a result, the surface is unable to form a protective oxide layer away from the grain boundary region, leading to the thick internal oxidation layer shown in the figures. On the other hand, a continuous protective Cr2O3 layer can form around a grain boundary by virtue of the fast outward diffusion of Cr with a significant supply of Cr from the matrix adjacent to the grain boundary, causing grain boundary Cr depletion (Fig. 5). This process results in a much thinner oxidation layer just over the surface grain boundary as compared to that away from the grain boundary. One of the novel findings obtained from a thorough compositional analysis of a grain boundary in the present study is that the severe Cr depletion and corresponding Ni enrichment of the grain boundaries (Figs. 4, 5 and 8) were caused by the formation of a protective innermost Cr2O3 layer. Recently, DIGM in Ni-base alloys during exposure to simulated PWR primary water or hydrogenated steam has been reported as a commonly occurring process [12,13,17,22,39,40,46], and DIGM has been an attractive research subject in terms of IG oxidation and cracking owing to the accompanying Cr depletion. When dynamic straining is applied, successive cycles of surface oxide film rupture and repair deplete Cr at the grain boundary. Hence, an IG crack can initiate when the Cr-depleted grain boundary is no longer able to support the formation of a protective Cr oxide layer. Therefore, DIGM and IG oxidation were claimed to be two important precursors of the IGSCC process of Alloy 690 [12,13]. The migrated grain boundary has a bowed or wavy shape, and the region swept by a migrating grain boundary shows Cr depletion and corresponding Ni enrichment with a flat shape in the composition profile [12,22,46]. In the present study, the grain boundaries appear to be rather straight, and the composition profiles of

Fig. 9. STEM image showing the structures of the internal oxidation layer of Alloy 690 exposed to PWR primary water at 325 °C. The features of interest are highlighted in the figure with explanations describing the internal and IG oxidation processes.

much thinner, or even absent, just over the grain boundary when IG Cr carbides were densely distributed on the grain boundary (Fig. 7). In the subsequent discussion, how these unique findings are associated with the internal and IG oxidation will be addressed in comparison with previous studies. It was revealed from previous studies that IG Cr carbides near the surface can aid in the formation of a protective Cr2O3 layer during the early stage of internal oxidation. Bertali et al. [40] provided direct evidence of localized IG Cr carbide decomposition and the formation of an IG Cr-rich oxide embryo during the exposure of Alloy 690TT to a hydrogenated steam environment. They concluded based on their findings that IG Cr carbides in Alloy 600TT appeared to be susceptible to decomposition in the reaction with diffused oxygen and that the release of free Cr from carbide decomposition promoted the formation of a protective Cr-rich surface oxide. Sennour et al. [41] also demonstrated that oxygen which diffuses through SCC cracks in Alloy 600 caused either superficial oxidation or the entire transformation of the IG carbide into Cr2O3 oxide, leaving a Ni-enriched zone between the IG carbide and the transformed Cr2O3 oxide layer. Therefore, IG Cr carbides in relation to IG oxidation appear to act as another Cr source for the formation of a protective Cr2O3 layer and, as a result, suppress internal and IG oxidation further. Meanwhile, the theory of Bruemmer et al. [42] that IG carbides may act as effective dislocation sources is thought to still hold some merit as another possible explanation of the benefit of IG Cr carbides of reducing IGSCC susceptibility. The innermost Cr2O3 in the oxidation layer was found to form through the outward diffusion process of Cr along the grain boundary in conjunction with oxygen diffusion into the material. Many other findings pertaining to the innermost Cr2O3 were obtained around the inner surfaces of grains far from the grain boundaries [27–29,31], in which the lattice diffusion of Cr could be the major factor controlling the formation of Cr2O3. The lattice diffusion rate of Cr is negligible below 480 °C, and only fast diffusion paths such as dislocation pipes and/or other lattice defects can operate to form Cr2O3. The present study focused on the surface grain boundaries. In this case, the grain boundary diffusion of Cr is more favorable and has been reported to be several orders of magnitudes more rapid than lattice diffusion [43]. Therefore, the main reason for the formation of the continuous Cr2O3 layer around the surface grain boundary in the present study is likely the grain boundary diffusion of Cr, and the different morphologies of Cr2O3 compared to previous results [27–29,31] may have originated from the different diffusion process, that is, the lattice diffusion and grain boundary diffusion of Cr in the temperature range of 325–480 °C. Once compact Cr2O3 is nucleated just over the grain boundary through the grain boundary diffusion of Cr, the dislocations generated at the 9

Materials Characterization 157 (2019) 109922

Y.S. Lim, et al.

corresponding author on request.

Cr and Ni have an acute shape at the grain boundaries. Olszta et al. [22] also demonstrated that some of the grain boundary Cr depletion occurred at an apparently non-migrated grain boundary in response to Cr diffusion to form a protective Cr2O3 film of Alloy 690 in PWR primary water at 360 °C. Langelier et al. [39] revealed from a 3D analytical electron microscopy examination of the IG oxidation of Alloy 600 exposed to 480 °C hydrogenated steam that IG Cr carbides immobilized the grain boundary against DIGM. Therefore, the Cr depletion found at the grain boundaries in the present study does not appear to be related to grain boundary migration due to the pinning effect of IG Cr carbides against DIGM. However, it is also possible that DIGM occurred to some extent despite the lack of clear evidence of DIGM occurring. More careful and thorough studies of DIGM in the samples under study here are necessary. The IG oxidation of Alloy 690 exposed to PWR primary water in the present study was significantly suppressed by virtue of the protective innermost Cr2O3 layer. These preliminary results with regard to internal and IG oxidation can provide some insight into the cracking phenomena of Alloy 690. Previous studies also reported that oxygen scarcely penetrated into the grain boundaries of Alloy 690 in simulated PWR primary water environments [22,23]. This characteristic of the IG oxidation of Alloy 690 is noticeably different from that of Alloy 600, in which IG oxidation is dominant with very thin internal oxidation layers [16,19,26,40]. Most cracking has been shown to be the IG type [5–7,20,21], and IG oxidation consistently occurs in Alloy 600 exposed to PWR primary water [5,17–20]. This fact indicates that IG oxidation should be the main reason for IG cracking, serving as the dominant mechanism promoting IGSCC in PWR primary water for Alloy 600. Once a crack initiates at an oxidized grain boundary, the crack can propagate rapidly along the oxidized grain boundary owing to the intrinsic weakness of the boundary [20,21]. Therefore, the absence of IG oxidation in Alloy 690 partially explains why it has greater resistance to PWSCC compared to that of Alloy 600 from the standpoint of IG oxidation. As a result, the different IG oxidation behaviors in Alloys 600 and 690 appear to lead to different cracking resistance capabilities.

Acknowledgement This work was financially supported by the Korean Nuclear R&D Program organized by the National Research Foundation (NRF) in support of the Ministry of Science and ICT (2017M2A8A4015155), and by the R&D Program of Korea Atomic Energy Research Institute (KAERI). References [1] W. Bamford, J. Hall, A review of Alloy 600 cracking in operating nuclear plants: Historical experience and future trends, Proc. of the 11th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2003, pp. 1071–1079. [2] M.B. Toloczko, S.M. Bruemmer, Crack growth response of Alloy 690 in simulated PWR primary water, Proc. of the 14th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems-water Reactor, 2009, pp. 706–721. [3] G.A. White, J. Hickling, L.K. Mathews, Crack growth rates for evaluating PWSCC of thick-wall Alloy 600 material, Proc. of the 11th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2003, pp. 166–179. [4] V.Y. Gertsman, S.M. Bruemmer, Study of grain boundary character along intergranular stress corrosion crack paths in austenitic alloys, Acta Mater. 49 (2001) 1589–1598. [5] Y.S. Lim, H.P. Kim, S.S. Hwang, Microstructural characterization on intergranular stress corrosion cracking of Alloy 600 in PWR primary water environment, J. Nucl. Mater. 440 (2013) 46–54, https://doi.org/10.1016/j.jnucmat.2013.03.088. [6] L. Thomas, B.R. Johnson, J.S. Vetrano, S.M. Bruemmer, Microstructural and microchemical characterization of primary-side cracks in an Alloy 600 nozzle head penetration and its Alloy 182 J-weld from the Davis-Besse reactor vessel, Proc. of the 12th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems-water Reactor, 2005, pp. 567–577. [7] H. Xu, S. Fyfitch, Laboratory investigation of PWSCC of CRDM nozzle and its Jgroove weld on the Davis-Besse reactor vessel head, Proc. of the 12th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems-water Reactor, 2005, pp. 833–842. [8] B. Alexandreanu, Y. Chen, K. Natesan, B. Shack, Cyclic and SCC behavior of alloy 690 HAZ in a PWR environment, Proc. of the 15th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2011, pp. 109–125. [9] Q.J. Peng, J. Hou, T. Yonezawa, T. Shoji, Z.M. Zhang, F. Huang, E.-H. Han, W. Ke, Environmentally assisted crack growth in one-dimensionally cold worked Alloy 690TT in primary water, Corros. Sci. 57 (2012) 81–88, https://doi.org/10.1016/j. corsci.2011.12.031. [10] S. Kim, S. Hwang, J. Lee, Effect of local strain distribution on stress corrosion cracking of cold-rolled Alloy 690 with inhomogeneous microstructure, Corrosion 71 (2015) 1071–1081, https://doi.org/10.5006/1652. [11] S.M. Bruemmer, M.J. Olszta, M.B. Toloczko, L.E. Thomas, Linking grain boundary microstructure to stress corrosion cracking of cold-rolled Alloy 690 in pressurized water reactor primary water, Corrosion 69 (2013) 953–963, https://doi.org/10. 5006/0808. [12] W. Kuang, M. Song, G.S. Was, Insights into the stress corrosion cracking of solution annealed alloy 690 in simulated pressurized water reactor primary water under dynamic straining, Acta Mater. 151 (2018) 321–333, https://doi.org/10.1016/j. actamat.2018.04.002. [13] T. Moss, W. Kuang, G.S. Was, Stress corrosion crack initiation in Alloy 690 in high temperature water, Curr. Opin. Solid State Mater. Sci. 22 (2018) 16–25, https://doi. org/10.1016/j.cossms.2018.02.001. [14] P.M. Scott, M. Le Calvar, Some possible mechanisms of intergranular stress corrosion cracking of Alloy 600 in PWR primary water, Proc. of the 6th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems-water Reactor, 1993, pp. 657–665. [15] G.C. Wood, F.H. Stott, D.P. Whittle, Y. Shida, B.D. Bastow, The high-temperature internal oxidation and intergranular oxidation of nickel-chromium alloys, Corros. Sci. 23 (1983) 9–25. [16] J. Panter, B. Viguier, J.-M. Cloué, M. Foucault, P. Combrade, E. Andrieu, Influence of oxide films on primary water stress corrosion cracking initiation of alloy 600, J. Nucl. Mater. 348 (2006) 213–221, https://doi.org/10.1016/j.jnucmat.2005.10. 002. [17] G. Bertali, F. Scenini, M.G. Burke, Advanced microstructural characterization of the intergranular oxidation of Alloy 600, Corros. Sci. 100 (2015) 474–483, https://doi. org/10.1016/j.corsci. 2015.08.010. [18] S.Y. Persaud, A. Korinek, J. Huang, G.A. Botton, R.C. Newman, Internal oxidation of Alloy 600 exposed to hydrogenated steam and the beneficial effects of thermal treatment, Corros. Sci. 86 (2014) 108–122, https://doi.org/10.1016/j.corsci.2014. 04.041. [19] Y.S. Lim, S.W. Kim, S.S. Hwang, H.P. Kim, C. Jang, Intergranular oxidation of Nibased Alloy 690 in a simulated PWR primary water environment, Corros. Sci. 108 (2016) 125–133, https://doi.org/10.1016/j.corsci.2016.02.040. [20] K. Fujii, T. Miura, H. Nishioka, K. Fukuya, Degradation of grain boundary strength

5. Conclusions In the present work, the internal and IG oxidation phenomena of Alloy 690 exposed to high-temperature primary water to simulate the normal operating environments of a PWR were systematically characterized using microscopic equipment to obtain insight into its IGSCC resistance and behavior. The internal oxidation layer around a surface grain boundary consisted of a thick upper NixFe1−xCr2O4 spinel and continuous band-shaped innermost Cr2O3 with orientation relationships among the Alloy 690 matrix, NixFe1−xCr2O4 and Cr2O3. From a compositional analysis, it was revealed that the innermost Cr2O3 layer formed through the outward diffusion of Cr along a grain boundary in conjunction with oxygen diffusion into the material, resulting in severe Cr depletion and corresponding Ni enrichment along the grain boundary. The continuous band-shaped innermost Cr2O3 layer was responsible for the prevention of oxygen diffusion into the grain boundary and the resultant suppression of IG oxidation. The effect of IG Cr carbides in terms of IG oxidation was confirmed. The early formation of the protective Cr2O3 layer and the resultant absence of internal and IG oxidation just over the surface grain boundaries were linked to the presence of IG Cr carbides in the grain boundary. From the present results, it is believed that the different IG oxidation behaviors of Alloys 690 and 600 lead to the different cracking resistance capabilities and cracking behaviors in these alloys. However, much more research is likely necessary to verify the appropriateness of internal and IG oxidation as a PWSCC mechanism of Alloy 690. Data availability The raw data related to this paper is available from the 10

Materials Characterization 157 (2019) 109922

Y.S. Lim, et al.

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

Eng. Tech. 51 (2019) 228–237, https://doi.org/10.1016/j.net.2018.09.011. [33] S.A. Attanasio, D.S. Morton, Measurement of the nickel/nickel oxide transition in Ni-Cr-Fe alloys and updated data and correlations to quantify the effect of aqueous hydrogen on primary water SCC, Proc. of the 11th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2003, pp. 143–154. [34] R.F. Egerton, S.C. Cheng, Measurement of local thickness by electron energy-loss spectroscopy, Ultramicroscopy 21 (1987) 231–244. [35] A.L. Bowman, G.P. Arnold, E.K. Storms, N.G. Nereson, The crystal structure of Cr23C6, Acta Cryst. B 28 (1972) 3102–3103. [36] T.M. Angeliu, G.S. Was, Behavior of grain boundary chemistry and precipitates upon thermal treatment of controlled purity Alloy 690, Metall. Trans. A. 21A (1990) 2097–2107. [37] H. Li, S. Xia, B. Zhou, J. Peng, The growth mechanism of grain boundary carbide in Alloy 690, Mater. Charact. 81 (2013) 1–6, https://doi.org/10.1016/j.matchar. 2013.04.005. [38] Y.S. Lim, D.J. Kim, S.S. Hwang, H.P. Kim, S.W. Kim, M23C6 precipitation behavior and grain boundary serration in Ni-based Alloy 690, Mater. Charact. 96 (2014) 28–39, https://doi.org/10.1016/j.matchar.2014.07.008. [39] B. Langelier, S.Y. Persaud, A. Korinek, T. Casagrande, R.C. Newman, G.A. Botton, Effects of boundary migration and pinning particles on intergranular oxidation revealed by 2D and 3D analytical electron microscopy, Acta Mater. 131 (2017) 280–295, https://doi.org/10.1016/j.actamat. 2017.04.003. [40] G. Bertali, F. Scenini, M.G. Burke, The intergranular oxidation susceptibility of thermally-treated Alloy 600, Corros. Sci. 114 (2017) 112–122, https://doi.org/10. 1016/j.corsci. 2016.11.004. [41] M. Sennour, P. Laghoutaris, C. Guerre, R. Molins, Advanced TEM characterization of stress corrosion cracking of Alloy 600 in pressurized water reactor primary water environment, J. Nucl. Mater. 393 (2009) 254–266, https://doi.org/10.1016/j. jnucmat.2009.06.014. [42] S.M. Bruemmer, L.A. Charlot, C.H. Henager, Microsructure and microdeformation effects on IGSCC of Alloy 600 steam generator tubing, Corrosion 44 (1988) 782–788. [43] D.D. Pruthy, M.S. Anand, R.P. Agarwala, Diffusion of chromium in Inconel-600, J. Nucl. Mater. 64 (1977) 206–210. [44] R.C. Newman, F. Scenini, Another way to think about the critical oxide volume fraction for the internal-to-external oxidation transition? Corrosion 64 (2008) 721–726. [45] B. Langelier, S.Y. Persaud, R.C. Newman, G.A. Botton, An atom probe tomography study of internal oxidation processes in Alloy 600, Acta Mater. 109 (2016) 55–68, https://doi.org/10.1016/j.actamat.2016.02.054. [46] D.K. Schreiber, M.J. Olszta, S.M. Bruemmer, Grain boundary depletion and migration during selective oxidation of Cr in a Ni-5Cr binary alloy exposed to hightemperature hydrogenated water, Scrip. Mater. 89 (2014) 41–44, https://doi.org/ 10.1016/j.scriptamat.2014.06.022.

by oxidation in alloy 600, Proc. of the 15th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2011, pp. 1447–1458. H. Dugdale, D.E.J. Armstrong, E. Tarleton, S.G. Roberts, S. Lozano-Perez, How oxidized grain boundaries fail, Acta Mater. 61 (2013) 4707–4713, https://doi.org/ 10.1016/j.actamat.2013.05.012. M.J. Olszta, D.K. Schreiber, M.B. Toloczko, S.M. Bruemmer, Alloy 690 surface nanostructures during exposure to PWR primary water and potential influence on stress corrosion crack initiation, Proc. of the 16th Int. Conf on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2013. Z. Zhang, J. Wang, E.-H. Han, W. Ke, Analysis of surface oxide films formed in hydrogenated primary water on Alloy 690TT samples with different surface states, J. Mater. Sci. Technol. 30 (2014) 1181–1192, https://doi.org/10.1016/j.jmst.2014. 09.002. K. Kruska, P. Chou, O. Calonne, L. Fournier, S. Lozano-Perez, Atom-probe tomography of surface and grain boundary oxides in Alloy 600 and Alloy 690 exposed to simulated PWR primary water, Proc. of the 16th Int. Conf on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2013. S.Y. Persaud, S. Ramamurthy, R.C. Newman, Internal oxidation of alloy 690 in hydrogenated steam, Corros. Sci. 90 (2015) 606–613, https://doi.org/10.1016/j. corsci.2014.11.006. P. Combrade, P.M. Scott, M. Foucault, E. Andrieu, P. Marcus, Oxidation of Ni base alloys in PWR water: oxide layers and associated damage to the base metal, Proc. of the 12th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2005, pp. 883–890. M. Sennour, L. Marchetti, F. Martin, S. Perrin, R. Molins, M. Pijolat, A detailed TEM and SEM study of Ni-base alloys oxide scales formed in primary conditions of pressurized water reactor, J. Nucl. Mater. 402 (2010) 147–156, https://doi.org/10. 1016/j.jnucmat.2010.05.010. H. Lefaix-Jeuland, L. Marchetti, S. Perrin, M. Pijolat, M. Sennour, R. Molins, Oxidation kinetics and mechanisms of Ni-base alloys in pressurized water reactor primary conditions: influence of surface defects, Corros. Sci. 53 (2011) 3914–3922, https://doi.org/10.1016/j.corsci.2011.07.024. Y. Masaki, Y. Momozono, M. Kanzaki, S. Taniguchi, T. Amino, O. Miyahara, Corrosion behavior of alloy 690 in simulated PWR primary water at high flow velocity, Proc. of the 17th Int. Conf on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2015. Z. Zhang, J. Wang, E.-H. Han, W. Ke, Influence of later-dissolved oxygen on microstructural changes in oxide films formed on Alloy 690TT in hydrogenated primary water, Corros. Sci. 94 (2015) 245–254, https://doi.org/10.1016/j.corsci. 2015.02.003. W. Kuang, M. Song, P. Wang, G.S. Was, The oxidation of alloy 690 in simulated water reactor primary water, Corros. Sci. 126 (2017) 227–237, https://doi.org/10. 1016/j.corsci.2017.07.004. Y.S. Lim, D.J. Kim, S.W. Kim, H.P. Kim, Crack growth and cracking behavior of Alloy 600/182 and Alloy 690/152 welds in simulated PWR primary water, Nucl.

11