A failure analysis was performed on an Alloy C-276 pull rod which underwent unexpected brittle, intergranular fracture after exposure to 280°C to 300°C aqueous solutions designed to replicate secondary side environments in nuclear energy systems: Pb-containing alkaline (pH300°C 8.5 to 9.5), and sulfate-containing acidic solutions (pH280°C 3 to 5). The component was characterized using advanced electron microscopy methods to demonstrate the benefits of these techniques for determining the nanoscale chemical, mechanical, and material factors contributing to failure, and to provide insight into the mechanisms of stress corrosion cracking (SCC) responsible for failure. Site-specific transmission electron microscopy specimens containing crack tips were prepared using focused ion beam. Nanoscale chemical characterization methods revealed that Pb was present in some oxidized regions of cracks, suggesting that the element may be inhibiting or impairing the passivity of the Cr-rich oxide. Complementary nanoscale microstructural analysis was performed. At an intergranular to transgranular cracking mode transition, it was observed that the transgranular crack (and corrosion process) propagated along the (110) crystallographic plane. Also, the cracking mode was highly dependent on the tensile stress direction relative to grain boundary orientation, the crystallographic orientation of grains, and geometrically necessary dislocation structures. A comparison of results with proposed mechanisms for SCC of Ni alloys in similar environments are discussed; the highly directional nature of cracking is consistent with a slot-tunnel corrosion mechanism.
Recent advancements in material characterization techniques have allowed for a better understanding of mechanical and chemical phenomena by enabling characterization of materials at the nanoscale, with near-atomic level elemental sensitivity. The benefits of these characterization techniques have led to their proliferation across many scientific disciplines where chemical and physical analysis at increasingly higher magnifications is beneficial for improving fundamental mechanistic understanding, such as catalysis, microbiology, and materials science.1-3 In addition, state-of-the-art characterization techniques provide an opportunity for better understanding the cause of failure for components in engineering applications.
For the analysis of corrosion failures, the ability to evaluate nanoscale chemistry in localized regions, such as pits and crack tips, provides insight into the (electro)chemical processes taking place which initiate degradation. Energy dispersive x-ray spectroscopy (EDX) is a nondestructive elemental detection technique that, when applied using scanning transmission electron microscopy (STEM), can indicate the location of elements with angstrom-level resolution, as well as provide qualitative composition information. Additional electron microscopy characterization techniques such as microscale electron backscattered diffraction (EBSD), and nanoscale transmission Kikuchi diffraction (TKD) provide complementary mechanical and microstructural information, such as the crystal structure of secondary phases or oxide(s) within a given sample, and the change in local misorientation imparted by plastic deformation. In tandem, these chemical and microstructural characterization methods are powerful tools for developing a fundamental understanding of many nanoscale processes from multiple scientific perspectives.4 This multidisciplinary approach is particularly useful for understanding stress corrosion cracking (SCC), due to the complex dependence of this phenomenon on material, chemical, and mechanical variables.
The SCC of Ni alloys used in steam generators of nuclear power plants has been demonstrated in laboratory studies in conditions simulating the extreme end of secondary side heat-transfer crevice environments, such as Pb-contaminated alkaline environments (pH300°C 8.5 to 9.5)5-7 and sulfate-containing acidic solutions (pH280°C 3 to 5).8-10 Slow strain rate testing (SSRT) is often applied to study the SCC susceptibility of materials in general; for specific testing in the aforementioned secondary side environments, corrosion-resistant Ni superalloys, such as Alloy C-276, are often used for construction of SSRT components. However, recently, an Alloy C-276 pull rod failed unexpectedly after two years of use for SSRT tests in environments meant to replicate secondary side crevice environments. SCC of Alloy C-276 in these environments has not been extensively studied, and the mechanism governing SCC in this system is not known. However, SCC of Ni-Fe-Cr alloys in secondary side environments, such as Alloy 600 and Alloy 800, has been studied in detail;7,11-14 the following contributing mechanisms to SCC have been proposed:
In Pb-containing near-neutral or alkaline environments at 280°C to 300°C, Pb is known to impair the passivity of Cr-rich oxide films formed on Ni-Fe-Cr alloys, possibly leading to formation of SCC precursors.5-7,15 Pb has been observed to contribute to SCC of Ni alloys in secondary side water environments, specifically Alloy 600,11-12,14,16 Alloy 690,6,13,17 and Alloy 800,5,18-19 which are commonly used materials in the construction of Canada Deuterium Uranium (CANDU) heavy-water reactor steam generators. Pb is thought to impair passivity either by incorporation in the Cr-rich oxide or adsorption/deposition at the oxide/metal interface. For the latter, Pb has been shown to readily adsorb/deposit at oxide/metal interfaces, thus blocking sites where hydroxide would typically adsorb as a precursor to passive oxide film formation.20 It has been shown that this disrupted oxide film will not be protective, and the material could be susceptible to a subsequent corrosion process, such as dealloying or intergranular oxide penetration, in the underlying region. The mechanism for Pb adsorption is proposed to be via underpotential deposition or displacement plating.5,21
It is known that Ni- and Fe-based alloys undergo SCC in caustic solutions, due to selective dissolution of reactive alloying elements (e.g., Fe, Cr) in the absence of an oxide film or if a protective oxide film becomes perturbed.4-6,22 In this way, dealloying can lead to a brittle, nanoporous surface layer. Under a tensile stress, cracking through this nanoporous layer occurs at high speed such that the energy transfer allows for cleavage into the underlying bulk metal. Plastic blunting arrests the crack temporarily until further dealloying allows for the SCC process to repeat. This proposed mechanism of SCC is known as the film-induced cleavage (FIC) mechanism.8-9 In cases where a protective oxide is present, Pb or another impurity is required to impair passivity and allow access of the bulk metal to the solution.
The adsorption of S has proven to enhance H absorption within Ni-based alloys; when H is introduced to Ni systems, severe embrittlement can occur at less than 200°C in aqueous environments.23 In addition, adsorbed S leads to reduction of the activation energy required for dissolution of metals, particularly Ni. This is due to the dipole metal being strongly bonded to S, which weakens metal-metal bonds in adjacent areas. This effect of adsorbed S has been implicated in the pitting and SCC of Alloy 800 exposed to high-temperature acidic, sulfate-containing solutions. Intergranular SCC is observed to initiate at pits and a slip dissolution-type mechanism has been proposed.24-25
The objective of this study is to perform a failure analysis on the Alloy C-276 component to determine the cause of SCC, with emphasis on nanoscale mechanisms. Complementary electron microscopy methods are applied to provide insight into the, currently unknown, material and chemical mechanisms contributing to SCC of the Alloy C-276 component after extended intermittent exposure to Pb and S-containing aqueous environments at 280°C to 300°C. This research also highlights the benefits of advanced microscopy methods for application in the failure analysis of components used for engineering applications; for example, knowledge gained can aid with development of predictive and mitigation strategies for the design of future testing systems.
Experimental History of Component
An Alloy C-276 pull rod was exposed to secondary side water conditions within an autoclave while under an applied stress, the combination of which led to SCC initiation and ultimately failure after 2.5 y in service. The stress applied varied over the lifetime of the pull rod, and the maximum applied load was 48 MPa (3800N, 1 cm diameter), which is well below the yield strength of 365 MPa, as specified by the manufacturer (Table 1). The component was exposed to caustic (pH300°C 8.5 to 9.5) Pb- and Cl-containing environments, [Pb] = 500 ppm, [NaCl] = 3 M, and acidic (pH280°C 3 to 5) sulfate environments, . The composition of Alloy C-276 is shown in Table 2. The length of exposure to each separate environment is not known, but the pull rod ultimately failed in the Pb-containing caustic solution. To the authors’ knowledge, SCC of Alloy C-276 has not been studied in the described environments previously.
The failed component was sectioned using a Struers Accutom-5†. All analysis in this study was performed in the “primary oxidized zone” closest to the fracture surface, as indicated in Figure 1. Preparation for scanning electron microscopy (SEM) imaging was done by grinding and polishing to produce a flat, mirror-like surface. Grinding was performed from 120 grit to 1200 grit using SiC sandpaper, followed by polishing using a 6 μm diamond suspension. A 0.05 μm alumina suspension was used before a final finish using a 0.06 μm colloidal silica suspension. Note that although the final polishing solution is coarser than the previous alumina suspension, it was found that this produced the best final surface finish for microscope analysis. It was essential to clean the surface thoroughly with soap and water after polishing with colloidal silica to remove residue.
To study the mechanism of SCC, a micro- to nanoscale characterization approach was instituted, such that initial qualitative observations made at the microscale were used to guide regions for site-specific extraction and analysis of the specimen at the nanoscale. The benefit of this approach has been supported by Staehle.27-29
For initial microscale analysis, SEM was performed using a FEI FEG-Nano† SEM to determine the chemistry, microstructure, and texture of the material. The probe current ranged from 100 pA to 10 nA, and beam energy ranged from 5 kV to 18 kV depending on the surface finish and depth of analysis required. A Bruker QUANTAX† EDX silicon drift detector was used for chemical analysis by SEM-EDX.
A plasma-focused ion beam (FIB)-SEM was used to perform serial sectioning studies to investigate the character of transgranular cracks in the C-276 pull rod. The advantage of plasma-FIB is that it can mill material at a much faster rate than conventional FIB because it uses Xe instead of Ga as the incident ion. A Thermo Scientific Helios G4 UXe† DualBeam Plasma-FIB was used for serial sectioning.
A Bruker QUANTAX Fast e−† EBSD detector was used to analyze the crystal structure and identify secondary phases in the material. A working distance of 15 mm produced the best EBSD data in the FEI FEG-Nano SEM, and a beam energy of 15 kV was suitable for this material. The SEM-EDX detector was also used to perform TKD analysis, but a dedicated TKD detector head was used in order to collect diffraction data directly underneath the specimen. This provides the similar information as EBSD, but is applied to electron transparent transmission electron microscopy (TEM) specimens thus lowering the interaction volume of scattered electrons to achieve a nanoscale step-size unachievable by conventional EBSD. The combination of EBSD and TKD is powerful for understanding material and mechanical property changes from the micro- to nanoscale.
A Zeiss NVision 40 Gemini† dual-beam focused ion beam was used to perform in-plane TEM foil extractions (meaning the surface plane of the FIB lift out sample was parallel to the specimen surface from which it was extracted) to produce two specimens containing crack tips for further analysis, using conventional procedures described in other studies.10,30-31 Specimens were mounted to a copper grid and thinned to 100 nm for observation in TEM. TEM analysis was performed using a FEI Osiris† FEG-TEM, providing microstructural imaging and EDX at nanoscale resolution to complement SEM work. STEM-EDX analysis was used to produce nanoscale elemental maps, providing information on the qualitative composition change within cracks and at crack tips. A dwell time of 50 μs per pixel, and a live time of roughly 10 min was used to produce STEM-EDX data maps. The “bright regions” in EDX elemental maps represent regions of highest concentration and should be considered relative to the known background concentration in the base metal (see Table 2). A ChemiSTEM™† EDX detection system comprised of four symmetrically placed silicon drift detectors was used for all EDX analysis in the TEM. All imaging in the TEM was done at a voltage of 200 kV. Bruker Esprit 2.1† software was used for all post processing of EDX data.
Bulk material analysis was performed at the end of the rod which was not exposed to any corrosive environments. This initial SEM analysis provided information regarding the bulk chemistry of the material, for comparison with manufacturer specification, and the pre-exposure microstructure. Figure 2 displays the fracture surface of the as-received sample, which displays characteristics typical of brittle intergranular fracture.
Scanning Electron Microscopy-Energy Dispersive X-Ray Spectroscopy Analysis
EDX characterization indicated that there is no visible chemical segregation, and EBSD analysis suggested that the material was single phase (at this magnification). The semiquantitative EDX concentrations shown in Table 3 were similar to the manufacturer specified material composition displayed in Table 2.
Electron Backscattered Diffraction
EBSD data of the sectioned fracture surface revealed that the intergranualar cracks within the material followed high angle grain boundaries (>30°), as is shown in Figure 3. Cracks preferentially propagated along grain boundaries which were perpendicular to the applied load, which is expected given that this is the direction of greatest stress intensity.32 Straight line features through the grains are evidence of grain twinning; after reviewing literature it was determined that these are annealed twins formed during manufacturing.33-34 The grain size of this component is on the scale of hundreds of micrometers.
Serial Plasma-Focused Ion Beam Sectioning
Serial plasma-FIB sectioning and SEM imaging was used to better observe the spatially discontinuous nature of transgranular cracking in the pull rod. The advantage of plasma-FIB in this case is that it can mill a large volume of material more rapidly than conventional FIB, but at the expense of spatial resolution. For this sample, the plasma-FIB could mill a 50 μm3 trench in less than 5 min. The increased rate of milling is owing to the much heavier incident ion beam which uses Xe as opposed to the atomically lighter Ga.
Serial plasma-FIB sectioning reveals the presence of spatially discontinuous cracking or coarse “slot like” features which develop parallel to each other, as can be viewed in Figure 4. The crack morphology is not consistent with a cleavage mechanism; instead, there is evidence for dissolution and corrosion product within the crack, inferred by the relatively wide crack opening and corroded cracked walls. Plasma-FIB milling was also performed parallel to the crack plane in an effort to examine the crack wall directly. The purpose of this experiment was to search for directional corrosion tunnels which could elucidate the microstructural dependence of a possible slot-tunnel corrosion mechanism (similar to what Burke, et al., discovered in their FIB experiments evaluating SCC of Alloy 690 in caustic solutions at 307°C).35 The crack plane (Figure 5) showed no clear evidence for directional corrosion “tunnels” propagating along the crack plane, instead the bumpy and uneven surface of the crack plane is evidence for corrosion product build up within the crack region. Note the distinction between corrosion “slots,” and corrosion tunnels, in that “tunnels” refer to very fine linear corrosion pathways exhibited along a crack plane, whereas corrosion “slots” refer to the much coarser cracks which result by linkage of corrosion tunnels. It is possible that the crack plane analyzed by plasma-FIB was a more mature region of the crack relative to an actively growing crack tip, which resulted in the visual observation of corrosion tunnels being obscured. However, the directional nature of the transgranular corrosion slots on the surface and in cross-section (Figure 4) suggests the possibility of a slot-tunnel corrosion mechanism (at the microscale).
Transmission Electron Microscopy/Scanning Transmission Electron Microscopy Analysis
TEM foils produced by FIB are shown in Figure 6, as viewed in SEM. One specimen displays exclusively intergranular cracking which arrests near a triple point, and jagged features are apparent along a grain boundary. The other specimen shows mixed-mode intergranular and spatially discontinuous transgranular cracking, with the transgranular crack approximately 45° relative to the intergranular crack. Note that the cracking mode was mostly intergranular within the pull rod, with less common transgranular cracking also present throughout. The intergranular to transgranular cracking transition took place as the orientation of the grain boundary was nearly parallel to the stress direction, which results in the lowest stress concentration at the grain boundary. Additionally, the transgranular crack did not propagate perpendicular to the stress direction, but was skewed from the tensile stress direction by roughly 45°. Also, while intergranular cracking was continuous, the transgranular cracking mode was noticeably spatially discontinuous along one direction. Although the transgranular crack must be connected in three-dimensions (i.e., the thin, 2D, TEM foil does not provide information about crack continuation in the depth of the material).
Scanning Transmission Electron Microscopy-Energy Dispersive X-Ray Spectroscopy Analysis of Crack Tips
STEM-EDX analysis was performed on crack tips to reveal the local nanoscale chemistry changes that may be contributing to SCC. Figure 7 displays the crack tip for the specimen which exhibited exclusively intergranular fracture; Ni depletion and Mo enrichment were observed ahead of the crack tip, and further analysis in TEM revealed that these were grain boundary precipitates. The semi-quantitative EDX characterization of these precipitates can be found in Table 4, and the EDX spectrum taken at the precipitate is displayed in Figure 7. EDX elemental maps shown in Figure 8 were acquired from the crack tip specimen which underwent intergranular and transgranular cracking. The EDX data in Figure 8 indicate that the crack contains oxidized species incorporating some amount of Cr, Ni, and Pb (although the former two elements are less than matrix composition). Semi-quantitative EDX characterization of the crack tip oxide can be found in Table 5, and a comparison of EDX spectra from an area at the interganular crack tip, as well as in the grain adjacent to the crack tip are shown for the sake of comparison in Figure 8. Chemical segregation is observed in the grain boundary ahead of the crack tip in Figure 8, with Ni depletion in regions where Mo is enriched to levels higher than base metal concentrations. A combined EDX map for Ni, Mo, and O shown in Figure 9 further underlines this chemical segregation. Cr showed no chemical segregation within, at or ahead of the grain boundary in Figure 8, and was present within the oxide of the crack itself but not in significant quantities when compared to the base metal. The oxide present in the center of the crack was, instead, enriched in Pb with less than base metal concentrations of Ni and Cr present. EDX elemental maps shown in Figure 10 were also acquired from the crack tip specimen which underwent intergranular and transgranular cracking. The transgranular crack in Figure 10 was depleted of all constituent metallic alloying elements within the crack itself, and contained an oxide enriched with Pb. There was no chemical segregation ahead of the transgranular crack tip, but oxidation was observed at the discontinuous junctions of the crack.
Transmission Kikuchi Diffraction Analysis of Crack Tips
TKD was performed to analyze the crystal structure identification, determine the orientation of grains in the material, and quantify the misorientation angle at grain boundaries. Figure 11 shows the results of the TKD analysis for the specimen which underwent exclusively intergranular fracture, displaying the quantified grain boundary misorientations in a map: The jagged triangular features are evidence of twinning, characterized by their specific misorientation angle. TKD results shown in Figure 12 indicate that only one grain boundary is present within the specimen which displayed mixed-mode fracture, and it is a high misorientation angle boundary of approximately 50°. EBSD data indicate that the transgranular crack progressed along the <110> direction.
Using Mtex—a MATLAB®† package—geometrically necessary dislocations (GND) were calculated with diffraction data from TKD analysis. The MATLAB code used for this analysis was developed by Skippon, et al.,36 and is available open source on GitHub.37 GND results are displayed in Figures 11 and 12. The specimen which exhibited exclusively intergranular fracture displays high GND density along the grain boundary where cracking occurred, concentrated at the grain boundary precipitates, as is shown in Figure 11. Additionally, areas of both high and low dislocation density seem to be evident along linear pathways, highlighted by yellow and blue circles, respectively. The directionality of dislocation structures could not be attributed to a single crystallographic plane or direction, but the obvious linear structure of dislocation networks is noteworthy. The twin boundary noted at the top of the image displays an area of low dislocation density (blue circle). The crack tip specimen which underwent intergranular and transgranular cracking, shown in Figure 12, displays the highest local dislocation density near the grain boundary, with the high intensity locations following the (110) crystallographic plane (yellow circle). Blue areas indicative of low dislocation density also follows the (110) plane (blue circle).
TEM Nano-Beam Electron Diffraction of Grain Boundary Precipitates
TEM nano-beam electron diffraction provided unique diffraction patterns which indicate the interplanar spacing of the Mo-rich precipitate’s crystal structure. The crystal structure of the precipitates was hexagonal close-packed with lattice parameters of α = 6.27 Å, and c = 8.70 Å epitaxially grown relative to the bulk metal such that the (010) plane of the precipitate grows relative to the (100) plane of the parent grain, as is shown in Figure 13. After consulting literature,38 it was found that these features possessed the exact crystal structure and lattice parameters of MoNiP8 precipitates, though less than 1% P was detected by EDX (Table 4).
Local Chemistry at Crack Tips
Pb was a key impurity visible in the specimen which exhibited transgranular fracture, and provided possible evidence for a detrimental role of Pb on Alloy C-276 which eventually contributes to SCC (Pb-SCC). The element has been observed to incorporate in the oxide within crack tips19,35 as well as clearly deposit as metallic Pb at oxide/metal interfaces along cracks.7,13,39 In either case, the presence of Pb inhibits passivation of Ni-Fe-Cr alloys by impairing the formation of the Cr-rich oxide film. It has been suggested that Pb will readily adsorb on metal surfaces via an underpotential deposition mechanism, thereby poisoning sites where OH− would typically adsorb as a precursor to Cr oxide formation.21
EDX results for the transgranular crack demonstrate that Pb is clearly incorporated in the oxide film within the crack and there is no evidence for a passive film, which is illustrated by the regions of dissolution within the extremities of the transgranular crack shown in Figure 10. Cr, the element which usually forms a passive film, does not appear to segregate in areas where Pb is present, indicating that Cr oxide formation is either impaired at the metal/solution interface, or the Cr oxide formed in not dense enough to maintain passivity. This is further evidence for the role of Pb in inhibiting Cr-rich oxide formation at the metal/solution interface at crack tips (and possibly displace the Cr-rich oxide).14,40 Studies by Persaud, et al.,41 and Bruemmer and Thomas20 have suggested that Pb impairment of oxide(s) on Ni alloys does not, by itself, result in SCC. Usually, the impairment of the normally protective oxide film leads to a subsequent corrosion process, such as dealloying or intergranular oxidation; the nanoporous film (dealloying) or intergranular oxide formed from these processes are precursors to SCC mechanisms.
Film-induced cleavage has been proposed as a mechanism for SCC in Alloy 800 in Pb-alkaline secondary side environments.10 It should be noted that although the bulk solution is Pb-alkaline, the crack tip chemistry could vary significantly, although previous atom probe tomography characterization of Pb-alkaline SCC suggested that alkaline pH does still persist at crack tips.39 A brittle, nanoporous layer formed due to dealloying could exist at the crack tip, or from oxidation products accumulating at the crack tip. Under a tensile stress, this nanoporous layer could lead to fast fracture through the brittle layer and subsequent cleavage into the bulk metal. However, in this study, dealloying (Ni enrichment) and/or nanoporosity, similar to reported observations in Alloy 800 exposed to Pb-alkaline solutions at 330°C,39,41 was not observed. The Ni and Mo EDX maps of the grain boundary ahead of the crack tip in Figures 8 and 9 could be misinterpreted as porosity, however Cr in this area remains entirely unreactive. Cr or Fe, being highly reactive relative to Ni and Mo, would be the first element to selectively dissolve as has been observed in Alloy 800 previously,41 and so this area of Ni and Mo enrichment is likely precipitate formation, confirmed in the TEM Nano-Beam Electron Diffraction of Grain Boundary Precipitates section.
Instead of dealloying, directional corrosion slots and linkage were observed using plasma-FIB serial sectioning, as is shown in Figure 4. Additionally, diffraction data showed that transgranular cracking occurred along the (110) plane. This behavior may suggest a slot-tunnel corrosion mechanism for transgranular SCC, as originally proposed by Swann42 and supported by Burke, et al.,35 for the mechanism of SCC in Alloy 690 exposed to Pb-alkaline environments at 307°C. The slot-tunnel corrosion mechanism of SCC constitutes highly directional corrosion tunneling along chemically active regions which exhibit accelerated dissolution kinetics. In this study, Pb could be playing a role in impairing the passivating oxide, which would allow corrosion tunneling to progress via anodic dissolution in a combined Pb-SCC/slot-tunneling mechanism.
Thus far, discussion has focused on the detrimental role of Pb, and SCC mechanisms proposed in prior studies with Pb as a contributing aggressive impurity. However, it should be noted that the failed Alloy C-276 component was also exposed to acidic sulfate environments, which makes other corrosion mechanisms possible as well. SCC in acidic sulfate environments has been reported in Alloy 600 and Alloy 800,10,43-44 but Alloy 690 exhibits SCC resistance; this is attributed to the high Cr content (30%) of Alloy 690.8 Ni-Fe alloys that do undergo SCC in sulfate-containing acidic solutions have reported pitting as a precursor to SCC and clear evidence of S, either incorporated in a Cr-rich oxide or displacing the protective oxide entirely.8 Neither pitting corrosion nor the presence of S was observed at crack tips in the Alloy C-276 component. In addition, the high Mo content of the alloy should aid with inhibiting pitting corrosion as a precursor to SCC in aqueous solutions.45-46 Therefore, the nanoscale chemistry observed in this work suggests that S or pitting corrosion may not be contributing to SCC. The primary aggressive impurity is suggested to be Pb in alkaline conditions, which impairs oxide passivity; inhibition of passivation may lead to SCC by a slot-tunnel corrosion mechanism.
Material and Mechanical Interpretation
Although the majority of cracking within the C-276 pull rod was intergranular, there were a number of observed transgranular cracks which exhibited discontinuous cracking in two dimensions along parallel planes, such as in Figures 4 and 10. In the case of the intergranular/transgranular transition shown in Figure 10, a number of microstructural parameters were responsible for the change in cracking mode from intergranular to transgranular: At the location where the intergranular to transgranular crack transition occurred, the normal of the grain boundary plane was nearly perpendicular to the stress direction, resulting in a lower stress concentration along the grain boundary. This resulted in a buildup of stress at the crack tip because the crack could not progress along the grain boundary; thus, the stress was released by forming a transgranular crack. The direction of the crack, however, was not perpendicular to the stress direction, but instead along a (110) crystallographic plane (Figure 12). This indicates that SCC of the Alloy C-276 component may have a significant dependence on material microstructure. The linear nature of dislocation networks is evident from the GND data presented in Figures 11 and 12, and the transgranular cracking occurs parallel to these regions of high dislocation density, suggesting that perhaps the transgranular crack followed a linear high-density dislocation network. However, given that the crack is present in the specimen at the time of TKD collection and GND analysis, it is impossible to know what dislocation structure existed within the material prior to the crack forming. Nonetheless, it is significant that the crack formed parallel to the obviously linear dislocation structures (Figure 12) and this suggests a microstructural dependence on cracking direction. The role of Pb inhibiting passivation only enhances this mechanism of microstructurally dependent localized corrosion, as oxidation/dissolution cannot be inhibited by a passivating film.
The comparatively large degree of deformation discovered at the mixed-mode crack tip (relative to nearby areas examined using GND) indicates that local deformation could be a contributor to the transition of cracking mode, as is shown in Figure 12. Studies on cold-worked austenitic stainless steels have unveiled some of the contributions of cold work to SCC, showing that a greater degree of local misorientation creates favored pathways for corrosion.47-51 It is possible that the increased dislocation density at the mixed-mode crack tip could facilitate the same corrosion/dissolution acceleration noted for cold-worked materials, in that increased local deformation creates high energy pathways which are susceptible to corrosion. This is supported by the highly directional nature of dislocation networks at the mixed-mode crack tip, indicated by the circled regions in Figure 12, as highly established deformation bands of greater dislocation density would be more susceptible to transgranular SCC. This phenomenon would explain the highly directional corrosion slots which persisted throughout the C-276 pull rod, which is again consistent with a slot-tunnel corrosion mechanism. This mechanism could have been confirmed if there had been evidence for directional corrosion tunnels along the crack plane, displayed in Figure 5, but this was not observed. This does not discredit the slot-tunnel corrosion mechanism, because, as mentioned, corrosion processes would destroy evidence for tunneling, as Burke, et al., only observed tunneling along the crack plane at the very extremities of the crack tip.35 In addition, many cases of directional transgranular corrosion slots found elsewhere in the cracked C-276 pull rod support this mechanism, and, in combination with the serial plasma-FIB sectioning and the dislocation structures analyzed using GND, provide strong support the slot-tunnel corrosion mechanism instituting failure in this material.
It should be noted that if dealloying had occurred along the slip system then a film-induced cleavage mechanism could operate,52 but no evidence of dealloying was apparent in this study. There is the possibility that a transient dealloyed layer was formed which was removed by anodic dissolution after fracture, but further in situ work would be required to confirm this. Additional surface analysis on the pull rod is planned to determine whether a nanoporous layer is present at the surface of the sample.
To summarize, the nanoscale characterization of local chemical, mechanical and material properties at crack tips in this study suggests that a slot-tunnel corrosion mechanism may be operating for transgranular SCC of Alloy C-276. In addition, Pb likely plays a key role in allowing development of SCC precursors in Pb-alkaline environments, based on the presence of Pb at crack tips. The benefit and necessity of characterizing SCC using complementary techniques that provide insight into the nanoscale chemistry, materials science, and mechanics of degradation was apparent; this is particularly critical for mechanistic interpretation of SCC. From an engineering perspective, application of state-of-the-art electron microscopy to characterize SCC in the Alloy C-276 pull rod has allowed for determining the cause of failure, which will aid with design of future testing systems. While it is known that Alloy C-276 can undergo SCC in 50% caustic environments at 107°C,53 this study demonstrates that Pb can instigate SCC in Alloy C-276 in aqueous environments under slightly alkaline conditions (pH300°C 8.5 to 9.5), which was previously unknown. An experimental program is currently underway to identify potential surface precursors to SCC of Alloy C-276 and other Ni superalloys in Pb-alkaline environments.
An Alloy C-276 pull rod was used in a tensile testing apparatus within an autoclave and fractured unexpectedly. The Alloy C-276 component was exposed to Pb caustic, and acid sulfate solutions over its lifetime, which were meant to replicate the most extreme conditions found in secondary side heat transfer crevices within CANDU steam generators. A failure analysis was performed on the failed component using state-of-the-art electron microscopy methods. Complementary EDX, EBSD, and TKD were applied to study the nanoscale chemical and material factors contributing the SCC. This multidisciplinary approach was necessary due to the complex dependence of SCC on material, mechanical, and chemical environment variables.
Pb was observed to concentrate at two of three crack tips studied, and areas of Mo enrichment and Ni depletion were found along grain boundaries; the latter was due to formation of secondary phases along grain boundaries. O and Pb were present in some cracks, with concurrent depletion of major alloying elements (Cr, Fe, and Ni) in these regions. Therefore, Pb is suggested to play a key role in impairing the passivity of the Alloy C-276 component, and SCC is thought to occur from exposure to the Pb-alkaline environment.
It was found that the crack followed high misorientation angle grain boundaries, and the cracking mode was brittle and predominantly intergranular. SCC was found to transition from intergranular to transgranular cracking in some regions, particularly where the grain boundary was almost parallel to the tensile stress direction. The crack then transitioned to transgranular cracking along a Ni slip system—note that this slip system was not perpendicular to the tensile stress direction, and so was not the direction of greatest stress intensity.
Transgranular SCC is believed to occur via a slot-tunnel corrosion mechanism, which has been proposed previously for SCC of Ni alloys in Pb-contaminated, aqueous environments at 280°C to 300°C. Intergranular SCC may depend on the degree of local misorientation along high-angle grain boundaries, or simply because these are preferred pathways for corrosion when a passivating oxide does not form. Pb plays a key role in enabling development of SCC precursors for the suggested mechanisms by impairing the passivity of the normally protective Cr-rich oxide.
A summary of characterization techniques used in this research and the conclusions drawn from these techniques is cataloged in Table 6.
Thank you to Travis Casagrande and Hui Yuan at the Canadian Centre for Electron Microscopy at McMaster University for their help in performing FIB sample preparation. The authors would like to acknowledge Prof. Mark Daymond for providing access to the scanning electron microscope and transmission electron microscope at the Reactor Materials Testing Laboratory (RMTL) at Queen’s University. This work was partially funded by a Natural Sciences and Engineering Research Council of Canada (NSERC) Alliance grant, supported by the University Network of Excellence in Nuclear Engineering (UNENE).