This article undertakes an investigation of the stress corrosion crack growth behavior of a cold-rolled Alloy 690 (UNS N06690) with microstructural inhomogeneity in the primary water of a pressurized water reactor, and ascertains the relationship between this behavior and the local distribution of the residual strain. Stress corrosion cracking (SCC) was found to propagate in a transgranular mode through a banded region composed of intragranular carbides and small-sized grains, and in an inter-granular mode through a normal matrix region of normal-sized grains, with an abnormally high growth rate when aligning the cold-rolling and crack growth directions with the intragranular carbide bands. The Vickers hardness and the degree of misorientation increased as plastic deformation as a result of cold-rolling proceeded. Moreover, these parameters were consistently higher in the banded region than in the matrix region. From an analysis of the residual strain, it was suggested that higher residual strain near intragranular carbide bands in the interior of the grains leads to transgranular SCC growth when the crack tip encounters the banded region, resulting in an abnormally high crack growth rate of the heavily cold-rolled Alloy 690 with microstructural inhomogeneity in the primary water.
INTRODUCTION
Nickel-based Alloy 600 (UNS N06600)(1) was widely used as a structural material in pressurized water reactors (PWRs) owing to its high corrosion resistance at high temperature. However, many forms of corrosion have widely been reported under specific material and environmental conditions in operating PWRs.1 The most severe service failures were localized corrosion phenomena such as stress corrosion cracking (SCC) and corrosion fatigue (CF) of the thin-wall tubing2–3 or thick-wall piping4–5 in primary water of high temperature and pressure, leading to leakage of radioactive species into the atmosphere or secondary water. Under the harsh environment of a PWR, one of the major factors influencing the susceptibility to SCC is the material microstructure. A significant amount of research on the metallurgical parameters that affect the SCC of Alloy 600 has shown that Cr carbides precipitated at the grain boundaries improve the resistance, whereas intragranular carbides have the opposite effect.6–8 Another common feature of service failures is the presence of very high residual strain or stress exceeding the nominal yield strength, which may be induced by welding, cold-working, or surface machining during the fabrication or repair process of the structural components of Alloy 600.9
To replace Alloy 600, which shows susceptibility to SCC, Alloy 690 (UNS N06690) with a higher Cr concentration is used as a structural material in PWRs. Over the past few decades, Alloy 690 has been considered to be “immune” to SCC in the primary water, as there have been no reports of SCC failures of components made from Alloy 690 in operating PWRs. However, recent laboratory experiments have shown that Alloy 690 is also susceptible to SCC in primary water.10–12 As-received Alloy 690 has shown a crack growth rate (CGR) of less than 1 × 10−9 mm/s, which is much lower than Alloy 600, with a CGR of 1 × 10−7 mm/s in the primary water of a PWR when compared at the temperature of 340°C. After plastic deformation by cold-working, the CGR of the Alloy 690 with a homogeneous microstructure increased to 1 × 10−8 mm/s.
It should be noted that some heats of Alloy 690 with an inhomogeneous microstructure revealed a high CGR, which exceeded 1 × 10−7 mm/s after severe plastic deformation by cold-working.9 From these experimental findings, it can be speculated that the inhomogeneity of the microstructure of Alloy 690 results in inhomogeneous straining and thus locally enhanced susceptibility of SCC growth. Unfortunately, however, the main cause of the abnormal crack growth of the inhomogeneous Alloy 690 after cold-working and the effect of the microstructural inhomogeneity remain poorly understood. Recent attention regarding the effects of microstructural changes by cold-working has focused on investigations of the grain boundary characteristics and residual strain as possible causes of the crack growth of cold-worked Alloy 60013 and Alloy 690,14 using the new technique of electron backscatter diffraction (EBSD).15–21 However, there have been relatively few investigations on the relationship between an inhomogeneous microstructure, which is often found in forged products of commercial Alloy 690, and the local distribution of the residual strain. Considering the abnormally high CGR of cold-worked Alloy 690 with an inhomogeneous microstructure, the effect of strain on an inhomogeneous material is assumed to be much different from these effects on homogeneous materials.
In this respect, this work sought to characterize the local strain distribution of cold-rolled Alloy 690 with microstructural inhomogeneity in an effort to understand its SCC behavior in primary water. The change in the residual strain distribution induced by plastic deformation was investigated using Vickers microhardness (HV) measurement and the EBSD technique. The abnormally high SCC growth rate of severely cold-rolled Alloy 690 with an inhomogeneous microstructure is discussed here, while considering the role of intragranular carbide bands in the residual strain distribution.
EXPERIMENTAL PROCEDURES
Materials
Two heats of Alloy 690 for a control rod drive mechanism (CRDM) nozzle of a PWR were used in this study. The first is extruded tubing (Heat No. RF798, thermally treated) with a uniform distribution of inter-granular carbides and a uniform grain size, as shown in Figure 1(a). The second is a forged bar (Heat No. 135264, solution annealed) with a banded microstructure of intragranular carbides and small-sized grains in a matrix of normal-sized grains in the axial direction, as shown in Figure 1(b). The chemical compositions of these materials are summarized in Table 1. The inhomogeneous heat was cold-rolled by a one-dimensional rolling process with various final thickness reduction ratios of 10, 20, 30, and 40% along the axial direction, parallel to the carbide bands. The homogeneous heat was cold-rolled with a thickness reduction ratio of 20%. The cold-rolling process was conducted with multiple passes to achieve relatively uniform plastic deformation at room temperature.
Stress Corrosion Cracking Test
A compact tension (CT) specimen with a thickness of 12.7 mm was fabricated from the cold-rolled materials in accordance with the ASTM E647 standard.22 The surface of the specimen was ground with abrasive paper of up to 2000 grit, and then cleaned sequentially with ethanol and water. A SCC test was conducted using a servo-electric system to apply a constant load to the CT specimens; this system was equipped with a direct-current potential drop device for the in situ monitoring of the crack length increment.22 The crack direction was parallel to the direction of the intragranular carbide bands in the S-L orientation9–10 (i.e., the crack plane whose normal direction is the thickness direction of a cold-rolled plate, with the predicted direction of crack propagation coinciding with the direction of cold-rolling23).
The CT specimen was first fatigued in air to generate a precrack with a length of approximately 2 mm, in accordance with the ASTM E647 standard,22 and a CF crack growth test was then performed at a maximum stress intensity factor (K) of 30 MPa√m with a load ratio R(maximum/minimum) = 0.7 and a frequency of 1 mHz for 1 month, after which a SCC test was performed at a constant K of 30 MPa√m for 1 month in primary water. The test solution was the simulated primary water used in typical PWR operating conditions, i.e., a solution of 1,200 ppm B as H3BO3, and 2.2 ppm Li as LiOH, at 340°C. Prior to the crack growth tests, the dissolved oxygen content and dissolved hydrogen content in the solution became steady at less than 5 ppb and at 2.7 ppm, respectively. The inlet conductivity and pH of the solution were approximately 23 μS/cm and 6.4, respectively, throughout the test.
Residual Strain Analysis
Residual strain was evaluated by means of HV measurement and the EBSD technique. The HV values of the specimens were measured at a load of 0.98 N and at a holding time of 10 s. The Oxford† EBSD system attached to a field-emission scanning electron microscope (JEOL JSM 7000F†) was used for the EBSD analysis. The cross section parallel to the rolling direction of each specimen was ground with abrasive paper of up to 2000 grit and polished using progressively finer grades of diamond paste, after which it was finally polished with a cross-section polisher using an Ar ion beam.
The EBSD measurements were performed on the T-L plane23 of the specimens in an area of 300 μm × 400 μm with a step size of 1.2 μm to 1.6 μm (i.e., 256 × 256 pixels in each observed area) at a magnification of 300×, and in an area of 37 μm × 50 μm with a step size of 0.14 μm to 0.20 μm at a magnification of 2,500×. In the EBSD analysis, the kernel average misorientation (KAM) was determined as the numerical average misorientation of a given point (or pixel) with all of its neighbors in a grain.13–14,18–19 For a given point, the average misorientation of that point with all of its neighbors was calculated with the criterion that misorientation outcomes exceeding a tolerance level of 5° were excluded from the calculation.18 Because the KAM value is affected by the scanning step size,17–18 the nearest neighbors were selected to define the kernel18 and KAM values obtained at the same magnification were compared for cold-rolled specimens at various thickness reduction ratios.
KAM is commonly known to be closely related to the degree of plastic or residual strain; this can be used as an appropriate parameter to evaluate the local strain distribution.13–14,18–19 Plastic strains manifest themselves as a reduction in the contrast and sharpness of Kikuchi bands in EBSD patterns, and through the local misorientations produced by dislocations stored in the deformed crystals.15,18 From a review of recent research that undertook strain analysis using the EBSD technique by Wright, et al.,18 it was concluded that this technique is very capable of identifying areas of concentrated strain within a microstructure using local misorientation parameters including KAM, although it remains difficult to identify the actual magnitude of the strain in such areas thus far.
RESULTS AND DISCUSSION
Stress Corrosion Cracking Behavior of Cold-Rolled Alloy 690 with an Inhomogeneous Microstructure
A significant amount of research that sought to evaluate the corrosion cracking of nickel-based alloys in the primary water of PWRs showed that local strain near a growing crack tip and the strain rate played important roles in CGR.24 When Alloy 690 was used as a structural material in PWRs to replace Alloy 600, it was also reported that significant crack growth arose in heavily cold-worked specimens as compared to non-cold-worked specimens.10–12 In particular, some heats of Alloy 690 with a banded microstructure revealed the most severe crack growth behavior when the crack direction was parallel to the direction of intragranular carbide bands in the S-L orientation.9–10
As shown in Figure 2, the CGR of CF and SCC of 40% cold-rolled Alloy 690 with a banded microstructure were measured from CT specimen in the S-L orientation. This result revealed values of approximately 4.3 × 10−7 mm/s at a maximum K of 30 MPa√m and 1.2 × 10−7 mm/s at a constant K of 30 MPa√m in primary water at 340°C. The CGR of SCC was similar to the previous results obtained from cold-worked Alloy 690 with a banded microstructure,9–10 and higher than the results obtained from cold-worked Alloy 690 with a homogeneous microstructure by 10 times.10–12
Figure 3(a) shows a scanning electron microscope (SEM) micrograph of the cross section of a CT specimen in which SCC occurred under a constant K condition after the SCC growth test in this work. It was found that SCC propagated in the transgranular mode when the crack tip encountered a banded region (denoted by A in Figure 3[a]), as confirmed from the fractograph of the cracked surface shown in Figure 3(b). However, when the crack advanced through the normal matrix region (denoted by B in Figure 3[a]), SCC grew in the intergranular mode with secondary intergranular cracks, as shown in Figure 3(c). This abnormal behavior of SCC growth path was found in other regions where cracks grew further under a constant K condition.
Intergranular SCC growth is typical behavior in as-received or cold-worked Ni-based alloys1 in which cracks propagate mainly along the grain boundaries, especially random high-angle boundaries.25 Despite the intense debate concerning the mechanism of intergranular SCC in Ni-based alloys in the primary water of PWRs, that is, whether the mechanistic model considers that cracks advance by an oxidation process at the crack tip or as a result of embrittlement caused by hydrogen discharged by a matching cathodic hydrogen evolution reaction, the most recent models incorporate the idea that solid-state grain boundary diffusion is the rate-controlling step, leading to intergranular SCC.26 In this respect, it is interesting that the SCC of cold-rolled Alloy 690 with an inhomogeneous microstructure propagated in a mixed form, despite the fact that the applied load was suitably controlled to maintain a constant K condition after a sufficiently long transition time (1 month) with the advance of the crack tip from the fatigue precrack in primary water in this work.
To explain this abnormal behavior of the SCC growth path, and its relationship with the faster CGR of the cold-rolled Alloy 690 with an inhomogeneous microstructure, the interactions between SCC growth and various cold-work effects, such as the yield strength (or hardness), the residual stress and strain, the microstructural transformation, the dislocation density, and morphology,27 would all need to be clearly understood. Apart from changes in the metallurgical properties, change in corrosion properties such as oxidation, dissolution, and hydrogen segregation in plastic-deformed materials should be considered based on the proposed SCC mechanisms of Ni-based alloys.26 As the first approach to understanding these interactions, however, this work mainly focused on the effect of the residual strain distribution in an inhomogeneous heat of Alloy 690 in comparison with that in a homogeneous heat.
Relationship Between the Residual Strain of Cold-Rolled Alloy 690 and the Vickers Microhardness
Figures 4(a) and (b) show the HV measured as a function of the reduction ratio for inhomogeneous Alloy 690 and an optical micrograph of a noncold-rolled specimen after the HV measurement, respectively. The average HV value increased as the cold-rolling proceeded because of strain hardening, as can be expected. It should be noted that the HV value measured in a banded region composed of intragranular carbides and small-sized grains was higher than that in a matrix region of normal-sized grains, as confirmed in Figure 4(b). For Alloy 690 cold-rolled at a reduction ratio which exceeded 30%, it was difficult to differentiate between the banded and normal matrix regions because of a severely deformed microstructure.
The results of the HV measurement suggest that the residual strain is distributed more locally in the banded region than in the normal matrix region, but it should be noted that these results may stem partially from the grain size difference between the banded and normal matrix regions. To overcome this limitation of the conventional microhardness test, the EBSD technique was applied in this work to analyze the residual strain on the microscale. This technique is well known to be applicable to measure a local misorientation in microstructures induced by plastic deformation, which is strongly correlated with the residual strain.13–21
Relationship Between the Residual Strain of Cold-Rolled Alloy 690 and the Electron Backscatter Diffraction Parameters
Figures 5 and 6 show the EBSD results obtained at a magnification level of 300× for Alloy 690 with a banded and homogeneous microstructure after cold-rolling at various reduction ratios. The inverse pole figure maps shown in Figure 5 indicate the crystal orientations, which are parallel to the normal direction of the observed planes. As compared to a noncold-rolled specimen showing one principal orientation per grain (Figures 5[a] and [e]), the degree of color gradation within the grains increased with an increase in the thickness reduction ratio (Figures 5[b], [c], [d], and [f]). Thus, the orientations of the grains were initially uniform, but misorientation appeared within the grains as plastic deformation proceeded.
The change in the misorientation by cold-rolling is quantitatively visualized in the local (point-to-point) misorientation maps shown in Figure 6. In Figures 6(a) and (e), the KAM is uniform for both noncold-rolled materials, showing misorientation of nearly 0°. Following the cold-rolling process, however, the degree of misorientation increased. The KAM values averaged over the entire scanned area were approximately 0.5°, 1.3°, 2.0°, and 2.5° for the 0, 10, 20, and 30% cold-rolled Alloy 690 specimens with a banded microstructure, respectively; they were approximately 0.6° and 1.48° for the 0% and 20% cold-rolled Alloy 690 specimens with a homogeneous microstructure, respectively. From multiple EBSD examinations of different scanning areas with similar microstructural characteristics, the scattering of the averaged KAM value was confirmed to be less than 0.1° in this work. Therefore, it is statistically significant to state that the averaged KAM value of the 20% cold-rolled Alloy 690 specimen with a banded microstructure was higher than that of the homogeneous Alloy 690 specimen cold-rolled at the same level of plastic deformation.
As noted in the Residual Strain Analysis section, it is generally accepted that the local misorientation is a useful EBSD parameter for identifying areas of concentrated residual strain within a microstructure, although it is difficult to measure the actual magnitude of the residual strain in such areas. In this work, to evaluate the local strain distribution of the inhomogeneous Alloy 690 after cold-rolling, an EBSD analysis was conducted at a higher magnification of 2,500×, as shown in Figures 7 and 8.
Figure 7 shows the pattern quality maps obtained from inhomogeneous and homogeneous Alloy 690 materials deformed with various thickness reduction ratios. As compared to the pattern quality map of the as-received Alloy 690 with the banded microstructure, the pattern quality was lower (darker color on the gray scale) mainly near the grain boundaries as plastic deformation proceeded to 10% (Figure 7[b]), after which multiple slip bands appeared at the grain boundaries and in the interior of the grains with plastic deformation to 20% (Figure 7[c]) for the inhomogeneous Alloy 690. Upon further deformation to 30% and 40%, the pattern quality in the interior of the grains was much lower (Figures 7[d] and [e]). On the other hand, for the homogeneous Alloy 690 cold-rolled with plastic deformation to 20%, the pattern quality was lower mainly near the grain boundaries (Figure 7[f]). Figure 8 presents the KAM maps calculated from the misorientation analysis of the same areas of Figure 7. The degree of misorientation increased remarkably along the slip bands in the interior of the grains and at the grain boundaries in the cold-rolled Alloy 690 with an inhomogeneous microstructure, while in the homogeneous material, the degree of misorientation increased mainly at the grain boundaries.
Figure 9 shows the EBSD results measured at a magnification of 2,500× for the 20% cold-rolled Alloy 690 with the banded microstructure. It was clear that more slip bands and a lower pattern quality appeared in the banded region than in the normal matrix region of the same specimen. In addition, the averaged KAM value was calculated to be approximately 1.1° and 0.8° in the banded and normal matrix regions, respectively, from the KAM maps (Figures 9[c] and [d], respectively). Considering the scattering of the KAM value below 0.1°, the difference is statistically significant, suggesting that the averaged KAM value in the banded region was higher than that in the normal matrix region.
During the plastic deformation of polycrystalline metals, the geometrically necessary dislocations (GNDs) move along a specific slip plane to accommodate the strain and accumulate (or build-up) at barriers that prevent their movement, such as the grain boundaries or precipitates, leading to work hardening.28 Therefore, it is expected that the density of GND can be affected by a smaller grain size, by intragranular carbides, or by both in the banded region of the inhomogeneous Alloy 690. The density of GNDs, ρ, is known to depend on the grain size, as in earlier work:28
where ε is the tensile strain, b is the Burgers vector, d is the grain size, γ is the shear strain, and L is the geometric slip distance. In a case in which ρ dominates, the relationship between the dislocation density and the flow stress τ can be expressed by as follows:28
where τ0 is the friction stress, α is a constant, and G is the shear modulus. Equation (2) presents the d−1/2 relationship for the flow stress-grain size dependence, in accord with the well-known Hall-Petch relationship. As compared to the normal matrix region, the higher dislocation density, KAM value, and flow strength of the banded region may be attributable to the interaction between the dislocation with a smaller grain size and/or intragranular carbides during plastic deformation, which is in line with the theoretical expectations from Equations (1) and (2).
However, the relationship between the dislocation density and the strain is more complex, as a dislocation can be annihilated as a result of a recovery process according to the well-known Kocks-Mecking relationship, as follows:29
where k1 and k2 are parameters describing the rates of accumulation and recovery of dislocations, respectively. Therefore, determining a direct correlation between the plastic strain and the dislocation density would require further work on various microstructural aspects, such as the accumulation and recovery of dislocations, and their relationship to the grain size and the intragranular carbides during the plastic deformation of inhomogeneous materials. In the present work, however, the local strain distribution within the inhomogeneous Alloy 690 after cold-rolling was comparatively analyzed according to the simple prediction from Equations (1) and (2).
All of the averaged KAM values discussed are summarized in Figure 10. Although the averaged KAM value was dependent on the EBSD acquisition parameter, such as the magnification of the measurement, the deformation process, and the materials,15,17 it was clear that the averaged KAM value has a unique linear correlation with the thickness reduction ratio for a given condition in this study. All of the averaged KAM values increased monotonically with an increase in the thickness reduction ratio. The averaged KAM value for Alloy 690 with a homogeneous microstructure showed a correlation similar to that reported from tensile specimens of Alloy 690 by other researchers,30 although the deformation process was different. It should be noted that the averaged KAM value was higher in the 20% cold-rolled Alloy 690 with a banded microstructure than in the homogeneous Alloy 690, which was cold-rolled at the same reduction ratio. Moreover, from the results of an EBSD analysis performed on a specific area of a 20% cold-rolled Alloy 690 specimen with a banded microstructure, it was confirmed that the averaged KAM value was higher in the banded region than in the normal matrix region, which was also found in a plot of the HV vs. the thickness reduction ratio (Figure 4[a]).
Based on a qualitative interpretation of the EBSD parameters holding that a lower pattern quality and greater local misorientation are useful indicators of a higher amount of residual strain within a microstructure,13–21 it is suggested that the residual strain in the banded region is higher than that in the normal matrix region of the heavily cold-rolled Alloy 690 with a banded microstructure. From these results, high residual strain in the interior of the grains as a result of the intragranular carbide bands in cold-rolled Alloy 690 may be one of causes of transgranular SCC growth with a higher rate in the primary water when the crack tip encounters the banded region. Despite considerable experimental efforts, no consensus exists yet as to the nature of the rapid SCC growth of cold-worked Alloy 690 materials with inhomogeneous microstructures resulting from the complexity of the interactions between the SCC growth and various cold-work effects. Although this work does not cover all aspects of the relevant factors affecting SCC behavior, the experimental findings here provide a clue with which to begin to explain the effect of microstructural inhomogeneity and hence, to improve our understanding of the abnormally fast SCC growth of highly cold-rolled Alloy 690.
CONCLUSIONS
SCC growth behavior and its relationship with the local strain distribution of cold-rolled Alloy 690 with microstructural inhomogeneity were investigated. The following conclusions were drawn in this study:
❖ The SCC growth rate of 40% cold-rolled Alloy 690 having intragranular carbide bands was measured to be 1.2 × 10−7 mm/s at a constant K of 30 MPa√m in primary water at 340°C, which was higher by approximately 10 times than the result reported for cold-rolled Alloy 690 with a homogeneous microstructure when aligning the cold-rolling and crack growth directions with the intragranular carbide bands. Fractographs of the cracked surface revealed that SCC propagated in the transgranular mode through the banded region composed of intragranular carbides and small-sized grains, and in the intergranular mode through the normal matrix region.
❖ HV and averaged KAM values measured in the banded region were consistently higher than those in the matrix region of the inhomogeneous material after cold-rolling. In addition, more slip bands and higher residual strain appeared in the interior of the grains in the banded region than in the normal matrix region. These phenomena may be among the causes of the rapid SCC growth of the heavily cold-rolled Alloy 690 material in the primary water of PWRs.
ACKNOWLEDGMENTS
This work was financially supported by the Ministry of Science, ICT and Future Planning (MSIP), and by the Ministry of Trade, Industry and Energy of Korea (MOTIE).
REFERENCES
(1) UNS numbers are listed in Metals and Alloys in the Unified Numbering System, published by the Society of Automotive Engineers (SAE International) and cosponsored by ASTM International.
† Trade name.
Author notes
* Korea Atomic Energy Research Institute, 1045 Daedeok-Daero, Yuseong-Gu, Daejeon 305-353, Republic of Korea.