Stress corrosion cracking (SCC) usually initiates at locally compromised surface regions, and ultimately at nanoscale precursor sites. The ability to identify such sites would be instrumental in predicting SCC failure and developing proactive mitigation strategies. Modern microscopy capabilities allow for the requisite micro-to-atomic scale analysis to characterize SCC and identify precursor sites at various length scales. In the latter part of his career, Roger Staehle recognized and emphasized the benefit of modern capabilities in microscopy and computational science for modeling and performing physical characterization of atomic and nanoscale processes related to SCC. Consequently, he developed the quantitative micro-nano (QMN) approach with the goal of attaining a global model of SCC on an atomistic basis. This article reviews recent studies that have applied state-of-the-art microscopy techniques to characterize SCC and associated precursors in the context of the QMN approach. Initial examples used to demonstrate characterization of nanoscale precursors include SCC of Alloy 800 in Pb-containing, caustic, and acid sulfate solutions relevant to secondary side crevice environments in nuclear power plants. In line with the QMN approach, the focus is on characterizing and understanding SCC mechanisms, leading to prediction and identification of associated precursors. Precursors to secondary side SCC of Alloy 800 are shown to include monolayer-level S or Pb at oxide-metal interfaces, the onset of dealloying, or metastable pitting corrosion. Following this, intergranular oxidation embrittlement of Alloy 600 in hydrogenated water/steam environments is explored to demonstrate the benefits of a multitechnique approach to identify SCC precursors and highlight recent advancements in in situ microscopy. Although nuclear-relevant SCC systems are used as examples, the QMN approach and benefit of identifying nanoscale precursors that correlate with SCC failure are applicable to a broad spectrum of SCC systems.
Stress corrosion cracking (SCC) is a form of localized corrosion that requires synergistic contributions from a multitude of material, environment, and mechanical variables. The complexity of the phenomenon results in difficulties with quantification and prediction, with no universal mechanism being applicable to all SCC systems. SCC usually initiates at locally corroded or otherwise compromised surface regions, many of which need to be understood at the nanoscale; some examples are: dealloyed nanoporous layers, localized corrosion (e.g., metastable pitting or intergranular corrosion), intergranular internal oxidation, or impaired surface films (e.g., Pb/S incorporation in oxides formed on Ni alloys). The ability to identify precursor sites at the micro-to-nanoscale is invaluable as a tool for SCC prediction, and for developing mitigation strategies prior to the onset of engineering-level degradation. This logic has been the impetus of several studies in recent years, examining micro-to-nanoscale precursors to known forms of industrially significant degradation, such as: metastable pitting in Cl-SCC of stainless steels1-2 and hot salt SCC of Ti alloys3-4 ; dealloying in caustic SCC of stainless steels5-6 and of intermetallics in Al alloys7-8 ; and intergranular oxidation in Ni alloys in simulated primary water.9-18
In order to identify nanoscale precursor sites on a metal surface prior to the onset of SCC, a micro-to-nano approach is beneficial. In such an approach, the first step is to obtain a fundamental understanding of the SCC propagation mechanism (at a microscale) to determine the chemical, material, and mechanical factors driving crack propagation in a given SCC system. With this knowledge, one can qualitatively interpolate the likely precursors to the SCC process, which can be identified in subsequent high-resolution microscopy and analysis. The micro-to-nanoscale chemical/material characterization required for the described SCC and precursor analysis is now possible using modern microscopy techniques. Recent advances in modern microscopy have allowed for characterization of materials from the micro-to-atomic scale with as good as monolayer-level elemental sensitivity. These technological advances now allow for application of a quantitative micro-nano (QMN) approach to understand the six segments of SCC, as described by Staehle.19-20 Elaboration on the application of modern microscopy techniques and the QMN approach is provided in Section “Approach.”
This article reviews recent developments in identification of nanoscale SCC precursor sites. Precursors to SCC in Ni-Fe-Cr alloys in nuclear-relevant environments are reviewed as examples to illustrate the approach. Alloy 600 (Ni-16Cr-9Fe), Alloy 690 (Ni-30Cr-10Fe), and Alloy 800 (Fe-32Ni-21Cr) are Ni-Fe-Cr alloys used for steam generator (SG) tubing and other major components in nuclear power plants. SCC has been induced in these materials in laboratory experiments using high-temperature (280°C to 340°C) hydrogenated water,10,21-25 and Pb26-37 and/or S38-44 contaminated water over the acidic to alkaline pH range; these environments are commonly used to simulate plausible primary- and secondary-side water, respectively. The diverse combinations of environments and materials results in a number of different SCC modes and mechanisms, each with their own associated nanoscale precursors, such as dealloying in caustic environments with and without Pb,5,26-27,35,45 metastable pitting in acidic sulfate environments,38-40,43-44 and intergranular oxidation embrittlement in hydrogenated water.9-18 The large amount of variation in SCC mechanisms and precursors, in materials with the same FCC crystal structure and major alloying elements makes these systems ideal as examples to illustrate the QMN approach to SCC. The importance of the precursor stage of SCC, in particular, for the prediction of SCC failure and development of mitigation strategies will also be explored.
In the latter part of his career, Roger Staehle recognized and emphasized the benefit of modern capabilities in microscopy and computational science for modeling and performing physical characterization of atomistic and nanoscale processes related to SCC of materials. Consequently, he developed the QMN approach19-20 ; this approach was initially applied to SCC of Ni-Fe-Cr alloys in nuclear environments by Staehle, but is generally applicable to a broad range of SCC systems. Staehle recognized that modern capabilities for atomic-scale characterization could allow for reproducible data and quantitative precision of alloy and environment variables associated with SCC, leading to truly fundamental mechanism-based models. Several QMN meetings were held with experts of multiple disciplines in corrosion science to develop the approach. Staehle described the approach as requiring a micro-to-nanoscale understanding of six critical segments of SCC, summarized in Figure 1 from his QMN meeting series as follows.19
Initial condition: including considerations related to the as-fabricated material with any applied surface preparation.
Precursor: the time associated with initial development of a locally compromised area that will lead to development of a brittle region; for example, Pb or S accumulation at oxide-metal interfaces in Ni alloys or Cr depletion in stainless steels.
Incubation: the time required for growth of a locally brittle region due to the precursor; for example, dealloyed layer growth or intergranular oxide penetration. Alternatively, this could be the time required for development of a concentrated environment to induce a dissolution-based SCC mechanism (e.g., metastable pitting).
Proto-crack: SCC penetration less than 1 μm in depth. This region provides insight into crack initiation mechanisms.
Environment at the crack tip: chemistry at the leading crack front which provides insight into the chemical factors driving crack propagation.
Metal at the crack tip: local stress-strain conditions and dislocation behavior at the crack tip that provide insight into crack propagation, direction, and mode.
In particular, the precursor stage is a critical segment in the QMN approach that ultimately leads to development of a compromised material condition. Therefore, early identification of nanoscale precursor sites could potentially allow for changing environment or material conditions to limit SCC initiation. Examples are provided in Section “The Application of the Quantitative Micro-Nano Approach and Modern Microscopy to Study Stress Corrosion Cracking Precursors,” which reviews studies exploring nanoscale precursor site identification in the context of the QMN approach. The applied approach initiates with developing an atomic-scale mechanistic understanding of crack propagation (metal and environment at the crack tip) for prediction of the likely nature of SCC precursors. Following this, the incubation and initial surface precursor stages can be characterized at the micro-to-nanoscale to evaluate hypothesized SCC precursors; for this review, segments 2 and 3 in Figure 1 are referred to as nanoscale precursor sites, interchangeably. Discussion related to hydrogen effects is not included in this review, but a similar QMN approach could be applied.
Among other areas, the QMN approach seeks to take advantage of recent technological advancements in modern electron microscopy. In particular, this paper aims to focus on the contributions of scanning electron microscopy (SEM), transmission electron microscopy (TEM), and focused-ion beam (FIB) sample preparation, covering a broad range of spatial resolution (micro-to-atomic) and elemental sensitivity (monolayer-level or even better), with the ability to characterize local plastic deformation and crystal structure/orientation.
Historically, TEM sample preparation on site-specific regions containing cracks or local precursor sites was difficult, although, the benefits of spatial resolution and elemental sensitivity were recognized decades ago. Technological advances since the mid-1990s, from electropolishing and ion-beam thinning10,46 to the eventual advent of a dual-beam FIB for TEM sample preparation,11,40,47-49 have led to the ability to prepare site-specific TEM specimens containing cracks. As a result, the use of analytical TEM techniques, such as energy dispersive x-ray spectroscopy (EDX) and electron energy loss spectroscopy (EELS), to study SCC of materials, has increased dramatically over the past decade.10-11,40,46-49 EELS is generally beneficial for chemical analysis of lighter elements and identifying variations in local bonding, while complementary EDX is more appropriate for heavier elements and spectrum quantification. Transmission Kikuchi diffraction (TKD) has been used to further characterize the local misorientation in the vicinity of crack tips, which correlates to plastic strain.50 Furthermore, state-of-the-art in situ TEM environmental testing has been applied with the generation of SCC precursor sites at the nanoscale, such as intergranular oxidation of Alloy 600.51 Also, recent advances in the FIB-SEM have allowed for advanced microscale characterization, such as 3D reconstruction of microvolumes of material16 and in situ micromechanical testing of SCC surface precursor sites.52-54 A review by the authors further expands on the benefit of using modern electron microscopy techniques for “fingerprinting” different modes of SCC in Ni alloys.35 Furthermore, a recent article by Badwe, et al.,55 effectively demonstrates how atomic scale microscopy methods can be applied to understand the synergy and decouple the role of stress and anodic dissolution in SCC mechanisms, such as film-induced cleavage.
THE APPLICATION OF THE QUANTITATIVE MICRO-NANO APPROACH AND MODERN MICROSCOPY TO STUDY STRESS CORROSION CRACKING PRECURSORS
Recent examples of characterization and identification of nanoscale precursors to SCC in Ni-Fe-Cr alloys, exposed to nuclear-relevant environments, are reviewed in this section. A wide range of SCC modes and mechanisms are known to operate in such alloys exposed to representative nuclear environments, ideal for allowing comprehensive examination of multiple associated nanoscale precursors. Sections “Caustic Stress Corrosion Cracking of Alloy 800 and the Effect of Pb” and “Acid Sulfate Stress Corrosion Cracking of Alloy 800” briefly highlight recent examples of the use of modern microscopy capabilities for elucidating mechanisms and precursors associated with secondary side SCC of Alloy 800. Having demonstrated the benefit of cutting-edge microscopy techniques for state-of-the-art SCC characterization, the review is then extended to intergranular oxidation embrittlement of Alloy 600 in Section “Primary Water Stress Corrosion Cracking of Ni-Fe-Cr Alloys.” The extensive research in this field allows for demonstrating the benefits of a multitechnique approach, highlighting recent advancements in in situ microscopy for precursor analysis, and reviewing the benefit of precursor knowledge for development of mitigation strategies.
The objective of this review is to demonstrate the application of modern microscopy to identifying SCC mechanisms and nanoscale precursor sites, in the context of the QMN approach. As such, the examples in the forthcoming sections only focus on a few recent studies that help demonstrate the benefits and capabilities of modern microscopy for SCC analysis; therefore, this paper is not an extensive review of secondary and primary side SCC in SG tubing materials, which is considered beyond the scope.
Caustic Stress Corrosion Cracking of Alloy 800 and the Effect of Pb
Impurities in secondary side water are known to accumulate and concentrate in SG heat-transfer crevices and under deposits between tube and tube supports and within tubesheet sludge piles, resulting in deleterious concentrations of species, such as Pb or S, and pH values quite different from that of the bulk water.42,56-57 SCC of Ni-Fe-Cr alloys is known to occur in both pure caustic conditions,5,35,45,58-61 and in Pb-contaminated high-temperature aqueous environments.26-37 However, because the transition from phosphate-based to all volatile treatment secondary side water chemistry, the likelihood of the high pH values required to induce pure caustic SCC has diminished significantly. Regardless, caustic SCC of Ni-Fe-Cr SG tubing materials remains of scientific interest, and several studies have investigated caustic SCC and associated surface precursors at the micro-to-nanoscale. Pb-contamination in alkaline environments has been shown to extend the domain of classical caustic cracking of Alloy 800 to lower pH regimes,26-27,33,35 considered as extreme but possible in SG secondary side crevice environments; thus, Pb-SCC remains of topical interest in the nuclear materials community.
Caustic Stress Corrosion Cracking of Alloy 800
Pure caustic SCC of Alloy 800 is usually studied in high-temperature water at pH300°C greater than 10, and can be intergranular or transgranular depending on loading conditions.5,35,45,58-61 At pH300°C greater than 10, and under the redox conditions expected in secondary side water, Cr and Fe are thermodynamically predicted to form soluble species (hydroxy complexes), while Ni metal is close to being thermodynamically stable.57 Given the thermodynamic conditions and the reactive element (Fe, Cr) content of Alloy 800, relative to the classical parting limit (50 at% to 60 at%),62 it is not surprising that dealloying (selective dissolution) of Cr and Fe has been shown to occur in several studies.5,35,45,61 Figure 2(a) shows the porous, Ni-rich film formed on Alloy 800 by dealloying after exposure to 50% caustic solution at 280°C by Bryk, et al.,61 The observation of dealloying and agreement with thermodynamic predictions has led Newman, et al.,5,45,61 to state that a brittle nanoporous film is formed on Alloy 800 that leads to a film-induced cleavage (FIC) mechanism for SCC in caustic environments; the layer is likely much finer than the gross porosity observed in Figure 2(a), as discussed further below. The FIC mechanism of Sieradzki and Newman63-64 proposes that a crack originating in a brittle film (specifically a nanoporous dealloyed layer), at a surface or crack tip, initiates a microcleavage event which can penetrate into the underlying bulk metal substrate where eventually plastic blunting arrests propagation; the open crack then undergoes further dealloying and the process repeats. It is unlikely that the size scale of porosity for FIC would ever be greater than 10 nm, as the mechanical properties would start to approach bulk values, and the interface with the substrate would deteriorate.
An SEM image of a cross section of an Alloy 800 C-ring specimen exposed to 50% caustic solution at 280°C from Bryk, et al.,61 is provided in Figure 2(b). From this image, microscale porosity is apparent, spanning the entire length of the crack. An additional STEM-EELS Ni map of a crack tip extracted from the Alloy 800 C-ring is shown in Figure 2(c); although not shown here, Cr, Fe, and O were not detected in the brightest, Ni-rich regions.61 The fine nanoscale porosity at the crack tip suggests that the gross porosity observed in Figures 2(a) and (b) are not necessary for crack propagation, and is formed upon crack opening, once Cr and Fe are completely dissolved to allow for coarsening of the porosity within the Ni-rich residue by surface diffusion. Based on the surface and crack tip analysis of Bryk, et al.,61 caustic SCC of Alloy 800 can be described in the context of the QMN approach. It can be stated that the mechanism of SCC for Alloy 800 in caustic solutions is film-induced cleavage driven by formation of a remnant brittle nanoporous film after partial selective Fe and Cr dissolution. Therefore, early-stage dealloying, even at a nanoscale, on Alloy 800 could eventually lead to cracking. With continued exposure in caustic solution, dealloyed layer growth will occur during the incubation period until a critical thickness is reached for crack initiation, dependent on the tensile stress. Rapid propagation of SCC and eventual failure would follow via successive dealloying and microcleavage events. Of course, the outer parts of the dealloyed layer may age into a more harmless form.
Pb-Stress Corrosion Cracking of Alloy 800 in Mildly Caustic Environments
In reducing mildly caustic environments, (7.5 < pH300°C < 10), dissolved Fe species and metallic Ni are thermodynamically favored, similar to pure caustic conditions in Section “Caustic Stress Corrosion Cracking of Alloy 800.” However, Cr is predicted and has been demonstrated to form a passivating chromia film on the surface of Alloy 800.42,56-57 This film is thought to inhibit the dealloying (and subsequent SCC). However, Capell, et al., have demonstrated that Pb-contamination can induce SCC of Alloy 800 in mildly caustic environments (pH330°C values as low as 8.5).33 They exposed reverse U-bend (RUB) specimens to a 330°C aqueous environment containing 3 mol/kg NaCl, NaOH for pH control, 6 ppm H2, and 500 ppm of Pb added as PbO over the pH330°C range of 8.5 to 9.5. Exposure to these environments produced through-wall cracks in RUB specimens, initiating as transgranular and transitioning to intergranular. SCC in these RUB specimens was recently characterized by Persaud, et al.,26-27 using analytical TEM techniques and atom probe tomography (APT) to garner mechanistic insight. A TEM image of an intergranular crack tip in Alloy 800 is shown in Figure 3(a),27 with the typical nanoscale chemistry observed by EELS and EDX elemental maps of Figures 3(b) and (c), respectively.
Ni enrichment (up to 70 at%) is observed along crack walls and ahead of the crack tip, indicative of a dealloying process due to selective dissolution of Fe and/or Cr. EELS results in Figure 3(b) indicate the formation of a Cr-rich oxide at the base of the crack tip (yellow in Figure 3[b]), probably normally protective, with progressive Fe incorporation in the oxide(s) moving away from the base of tip. Also, a gradient in Ni and Fe content is observed ahead of the crack tip; further results from Persaud, et al.,27 indicate that progressive Fe dealloying extends 300 nm ahead of the crack tip. Also, metallic Ni enrichment along the more mature crack was observed every 250 nm to 300 nm,27 suggesting a cyclical cracking process. Further EDX analysis in Figure 3(c) revealed the presence of Pb at the oxide-metal interfaces. Complementary APT analysis which captured a crack tip in the RUB specimen allowed for quantifying the concentration of Pb (2 at% max.);26 the 3D APT reconstruction is shown in Figure 4(a) with the composition profile across the oxide-metal interface shown in Figure 4(b). Pb is deposited in elemental form after initial oxide rupture and likely impairs the passivity of the Cr-rich oxide, allowing for selective dissolution of Fe in the region ahead. Further support for impaired oxide passivity by Pb is provided in the work of Lu, et al.,65-66 who found that Pb deposition on the bare metal surface promoted anodic dissolution in Ni-based alloys, retarded repassivation kinetics, and reduced film rupture ductility. The impairment of oxide(s) by Pb deposition may be analogous to activation of localized corrosion in Al due to low melting point metals. Sato and Newman67-68 suggest that diffusion of these metals on the surface of Al disturbs the connectivity of the reforming oxide network.
Persaud, et al.,27 suggest that SCC of Alloy 800 proceeds via a film-rupture/dealloying mechanism. Pb extends the domain of pure caustic SCC in Alloy 800 to lower pH by impairing the passivity of the Cr-rich oxide, enabling dealloying. Pb electrodeposition has been shown to be possible on Ni alloys through underpotential deposition69 or displacement plating32-33 ; readers are referred to Reference27 for further details on how Pb is suspected to impair oxide(s). The apparent ability of Pb to extend the domain of SCC modes in borderline passive conditions may not be limited to caustic conditions. This observation may be applicable to other known SCC modes of Ni alloys in high-temperature water, such as acidic, primary water, or high-potential SCC, given that each is driven by impairment or absence of a passive Cr-rich oxide. In addition, the impairment of passive oxide(s) by Pb deposition should be considered applicable to other Ni alloys, Alloy 600, and Alloy 690. For example, Thomas and Bruemmer were first to demonstrate Pb incorporation in oxides in cracks formed in mill annealed Alloy 600 exposed to Pb-contaminated high-temperature water.31 However, it should be noted that in order for Pb-induced SCC to operate, Ulaganathan, et al.,70 have recently shown that the element must surface diffuse or deposit at the oxide-metal interface in order for the passive oxide to be compromised, to enable a process such as dealloying (i.e., Pb, that is simply incorporated in the bulk surface oxide may have a limited effect). Therefore, mono-layer level Pb electrodeposition at oxide-metal interfaces on Alloy 800, and possibly other Ni alloys, should be considered a general SCC precursor, in the context of the QMN approach, given the known deleterious effect of the element in impairing passive oxides. The incubation period would likely involve a subsequent corrosion process that eventually leads to SCC initiation, such as formation of a nanoporous Ni-rich film after dealloying in mildly caustic conditions.
Acid Sulfate Stress Corrosion Cracking of Alloy 800
Minor quantities of sulfate in bulk secondary side water can accumulate at heat transfer crevices or within tubesheet sludge piles in SGs, resulting in deleterious concentrations of sulfur species and acidic pH.42,56-57 SCC of Alloy 800 has been demonstrated in laboratory experiments in acidic solutions (pH300°C 3 to 5) containing dissolved sulfate anions over the 280°C to 320°C temperature range.35,39-44 In general, the mode of degradation can be pitting, intergranular attack, or SCC, depending on the loading conditions, pH, and sulfate concentration. A decrease in pH tends to shift the degradation mode of Alloy 800 in acid sulfate solutions to pitting to eventual general corrosion. It should be noted that reduced sulfur species are well known to impair oxide passivity on Ni alloys,71-75 and reduction of sulfate has been shown to occur in high-temperature aqueous solutions,39,43 with hydrazine and magnetite both possibly playing an important role, as well as the reductive action of metal surfaces.
The deleterious effect of S on the passivity of oxides formed on Ni-Fe-Cr alloys can be considered analogous to Pb electrodeposition at oxide-metal interfaces, discussed in Section “Caustic Stress Corrosion Cracking of Alloy 800.” However, the interactions of Pb and S with the alloy surface to promote oxide impairment are entirely different. While Pb generally electrodeposits on the bare metal surface, reduced sulfur species have been shown to chemisorb on metal surfaces due to an electronegative attraction, with partial charge transfer, resulting in a partial dipole (+) charge on the metal.72-73,76 This results in reduction of the activation energy necessary for metal dissolution, accelerating the corrosion rate. Also, Ni sulfide species are thermodynamically favorable in the conditions of interest and can promote formation of a mixed oxide-sulfide layer.72-73,76 This results in more facile rupture of the surface film, enabling or accelerating a SCC mechanism driven by film rupture.
Studies examining mechanisms of sulfur-assisted degradation in Alloy 800 and associated precursors at the micro-to-nanoscale have emerged in recent years. For example, Persaud, et al.,39-40 exposed Alloy 800 C-rings (2% plastic strain) to a 0.55 M sulfate solution (NaHSO4) at 315°C with a calculated pH315°C of 4.3 and observed cracking to depths in excess of 300 μm in less than 60 h. IGSCC was reported to initiate at the base of pits on the alloy surface, with additional pits visible intermittently along the crack paths, shown in the SEM image of Figure 5(a).39-40 Crack tips were extracted from the C-rings for TEM analysis using EELS and EDX. A dual layer oxide was apparent from EELS analysis, and sulfur was detected in the crack using EDX, but could not be reliably mapped using technology available at the time. Regardless, Persaud, et al.,39-40 suggested a modified slip dissolution mechanism for SCC of Alloy 800 based on the observed crack tip chemistry, prior electrochemical work of Smith, et al.,44 and the morphology of cracks.
Acid sulfate work was later extended by Persaud, et al.,38 through exposure of actively loaded, blunt-notched Alloy 800 tensile specimens to a 0.5 m sulfate solution at 280°C at pH280°C 3, using the direct current potential drop (DCPD) technique to monitor crack growth. The lower pH relative to prior work was used to purposely generate multiple degradation mechanisms, straddling the suggested boundary between pitting and SCC. Also, controlled crack growth using DCPD ensured that an active crack was available for extraction, avoiding blunted or arrested regions. Crack tips were extracted from Alloy 800 tensile specimens for TEM analysis using EDX. The state-of-the-art EDX technology used for analysis had multiple silicon drift detectors for improved x-ray collection and element detectability, allowing for improvement on the initial analysis in prior studies.38 As a result, discerning the presence and behavior of small concentrations of sulfur was possible. In SCC, the element was found to concentrate up to 4 wt% and segregate to localized regions at oxide-metal interfaces at crack tips and along crack walls, often overlapping with Ti38 ; Figure 5(b) shows an elemental map for S along a typical crack wall. In addition, a Cr-rich oxide was visible in areas with SCC, suggesting that a mixed sulfide-oxide layer was formed; this supported the original assertion of a modified-slip dissolution mechanism. Further STEM-EDX analysis of pitting corrosion in Alloy 800 tensile specimens revealed a relative absence of oxide(s) on pit surfaces, with only a sulfur layer present at pit-metal interfaces, shown at high resolution in the oxygen and sulfur maps of Figure 5(c) along a pit-metal interface.38 This finding suggested that the change in degradation mode could be directly linked to surface coverage of sulfur, whereby near complete surface coverage by adsorbed sulfur on the bare metal resulted in absence of oxide(s) and accelerated metal dissolution, leading to pitting or general corrosion; this has been shown by Oudar and Marcus,72 Marcus, et al.,73,76 for a relative surface coverage approximately 1. However, if the surface coverage of sulfur is lower, a mixed oxide-sulfide layer can form, originally suggested by Oudar and Marcus,72 Marcus, et al.,73,76 leading to localized SCC propagation from the base of pits through a modified-slip dissolution mechanism. This description for the behavior of sulfur and transition between degradation modes agrees with established theories for sulfur-assisted uniform and localized corrosion.77-78
In the context of the QMN approach, the precursor to sulfur-assisted degradation on Alloy 800 can be directly attributed to adsorption of sulfur on the bare metal, even at monolayer levels, and the extent of surface coverage. In acid sulfate conditions, multiple degradation modes can operate in Alloy 800 depending on adsorbed sulfur coverage, pH and sulfate concentration. The incubation period prior to SCC initiation is likely metastable pit growth, depending on pH and sulfate concentration. Eventually, the adsorbed sulfur coverage may diminish with continued pit growth, allowing for partial passivation of the metal surface through formation of a mixed sulfide-oxide layer. At this point, proto-cracks can initiate from the base of pits, leading to eventual propagation of SCC.
Primary Water Stress Corrosion Cracking of Ni-Fe-Cr Alloys
In this section, the application of modern microscopy to determining the mechanism and precursors to primary water SCC (PWSCC) of Alloy 600 are explored. As one of the most studied nuclear materials degradation issues, further cutting-edge techniques, beyond those in Sections “Caustic Stress Corrosion Cracking of Alloy 800 and the Effect of Pb” and “Acid Sulfate Stress Corrosion Cracking of Alloy 800,” have been applied to characterize and test PWSCC of Alloy 600 and associated precursors in model environments.9-18,51-54 Examination of generally accepted precursors to PWSCC of Alloy 600 is briefly discussed in Section “Precursors to Primary Water Stress Corrosion Cracking of Alloy 600,” specifically underlining the benefit of characterization at multiple length scales; this is a brief overview of the current state of understanding, garnered from micro-to-nanoscale analysis, and readers should not consider this summary as a comprehensive review of PWSCC. Following this, recent studies investigating PWSCC precursors in Alloy 600 through application of in-situ environmental and mechanical testing are explored in Section “In situ Micro-nanoscale Testing of Precursors to Primary Water Stress Corrosion Cracking in Alloy 600.” Finally, the possible mitigation of PWSCC in Alloy 600 by replacement with Cr-rich Ni alloys, Alloy 690 and Alloy 800, is briefly highlighted in Section “Limiting Precursor Development using Alloy 690.”
Precursors to Primary Water Stress Corrosion Cracking of Alloy 600
Alloy 600 was originally the material of choice for SG tubing in PWRs, but was found susceptible to intergranular PWSCC in 280°C to 340°C primary water. This phenomenon has been a major issue in the nuclear materials community for decades and significant research on SCC mechanisms and associated precursors has been performed. Primary water contains added hydrogen which reduces the potential into the vicinity of the Ni/NiO equilibrium electrode potential.79 PWSCC can be partly linked to the low Cr content of the 600 alloy (16 at%) that does not allow for sufficient outward Cr flux for an external and passivating surface oxide to form. With thermodynamic conditions at or near the Ni-metal regime, the higher inward oxygen flux allows for solid state oxygen diffusion and internal oxidation within grains to shallow depths.9,15,80-81 In addition, extensive experimental evidence has suggested that short-circuit diffusion at grain boundaries results in significant penetrative intergranular oxidation of Cr, resulting in intergranular oxidation embrittlement as a precursor to PWSCC.9-18
Figure 6(a) shows STEM-EELS analysis of intergranular oxidation in solution annealed (SA) and thermally treated (TT) Alloy 600 after exposure to 480°C hydrogenated steam for 120 h under conditions where metallic Ni is stable from Langelier, et al.,16 This environment is used to produce similar, but accelerated, degradation compared to PWR primary water environments.82 In addition to penetrative intergranular oxidation, classical internal oxidation occurs within grains, which results in metallic Ni expulsion due to the compressive stress generated by the formation of discrete internal oxide particles beneath the metal surface.83-85 Results for two heat treatments of Alloy 600 are shown in Figure 6, 600TT and 600SA; 600SA was solution annealed at 1,050°C in high purity Ar gas for 1 h and water quenched, while 600TT underwent an additional thermal treatment at 704°C for 24 h to promote formation of a semicontinuous network of Cr carbides along grain boundaries.9 Intergranular oxidation is observed in 600SA and 600TT with diffusion-induced grain boundary migration (DIGM) apparent, extending well ahead of the intergranular oxidation front. Although there is still some dispute, generally the DIGM and intergranular oxidation observed in Figure 6(a) are considered precursors to PWSCC of Alloy 600.15-16,51,86 Similar oxidation phenomena have been reported in primary water conditions at 360°C, albeit more subtle due to hindered diffusion kinetics at lower temperatures.24,87
Notably, the intergranular oxidation depth and extent of DIGM are observed to be different between 600SA and 600TT in Figure 6(a). The presence of Cr carbides in the latter appears to have a beneficial effect of limiting the magnitude of oxide penetration and DIGM distance. However, this two-dimensional analysis is inconclusive without information on the variation of oxide penetration/DIGM in the plane perpendicular to Figure 6(a). Thus, Langelier, et al.,16 investigated the beneficial effect of Cr carbides by applying a 3D FIB serial sectioning technique to reconstruct microvolumes containing oxidized grain boundaries from 600SA and 600TT specimens after exposure to a 480°C hydrogenated steam environment for 120 h. A summary of their microscale 3D reconstructions is shown in Figure 6(b), with Supplementary Videos available in Langelier, et al.16 By analyzing the oxide penetration depth of Alloy 600 at multiple length scales, it is shown that intergranular oxidation of reactive elements is definitely limited by the presence of large Cr carbides at grain boundaries; this may be linked to a possible pinning effect of large carbides. Once the intergranular oxidation is pinned in this manner, the effective stress intensity at the deeper penetrations is reduced by the shielding effect of the carbide pinned regions. The results of Langelier, et al.,16 aid in successfully explaining the superior resistance of Alloy 600TT to PWSCC. Also, the data shown in Figure 6 highlights the benefits of studying SCC mechanisms and nanoscale precursor sites, taking advantage of current technology to evaluate the morphology and chemistry of precursors, in three dimensions, across the entire micro-to-nanoscale range.
In Situ Micro-Nanoscale Testing of Precursors to Primary Water Stress Corrosion Cracking in Alloy 600
Studies evaluating intergranular oxidation of Alloy 600 using in situ TEM environmental analysis and microscale mechanical testing techniques have recently emerged.51-54 Microtensile, micropillar, and microcantilever testing allow for determining the mechanical properties of microvolumes of materials under tension, compression, or shear stress.52-54 These techniques can be used to evaluate micro-to-nanoscale SCC precursor sites, allowing for better determination of the incubation period required to initiate SCC (e.g., correlate mechanical properties with intergranular oxide penetration depth).
Recent studies have applied microcantilever tests to evaluate the influence of surface oxides and oxidized grain boundaries on the mechanical response of Alloy 600.53-54 The sample chosen for testing was a mill-annealed Alloy 600 coupon that had been exposed to simulated PWR primary water at 340°C for 2,000 h. Microcantilevers containing oxidized grain boundaries were prepared using a FIB-SEM, with cantilever dimensions of approximately 20 μm to 25 μm length, 5 μm width, and 5 μm thickness. A nano-indenter was used to apply a load at the end of the microcantilever to generate a load-displacement curve that could be used to compute the stress at failure. The results clearly demonstrated that the presence of intergranular oxidation resulted in lower measured stress at failure compared to an uncompromised boundary.53-54 Also, the presence of a surface layer on the top of the oxidized boundary (e.g., simulant protective oxide) increased the required stress at failure from 527 MPa to 765 MPa, or entirely prevented microcrack initiation. In future, micromechanical testing techniques used to study brittle regions in nuclear materials should be extended to testing other precursors highlighted in the present review, such as the nanoporous, dealloyed layers in caustic conditions of Section “Caustic Stress Corrosion Cracking of Alloy 800 and the Effect of Pb.” However, while the intergranular oxidation precursor discussed here53-54 is robust to air exposure, preserving the metallic nanoporous film, or other precursor sites, for mechanical testing may present additional challenges.
Burke, et al.,51 have recently attempted to simulate preferential intergranular oxidation of Alloy 600 at the nanoscale using an in situ, heated, gas environmental TEM stage to image or perform chemical analysis while simultaneously exposing the sample to a gaseous environment. They extracted an approximately 300 nm thick TEM lamella using a FIB with an area of 25 μm × 20 μm from a virgin Alloy 600SA material that contained a single high angle grain boundary. Using the gaseous environmental holder, experiments were performed at temperatures between 320°C and 480°C in a 1 bar H2-N2-H2O environment to simulate primary water conditions and attempted to produce predicted precursors to PWSCC of Alloy 600, intergranular oxidation and DIGM.51 Their in situ imaging and chemical analysis revealed DIGM occurring across the grain boundary of interest on a micro-to-nanoscale, with minor penetrative intergranular oxidation of Ti and Al beneath a surface Cr-rich oxide. Although there were some differences in oxide formation and chemistry compared to bulk experimentation of Alloy 600 in primary water conditions, Burke, et al.,51 successfully demonstrated the potential of in situ gaseous experimentation for studying precursor corrosion reactions, inching closer to representative environments.
In situ micro-to-nanoscale SEM and TEM techniques are currently limited by the environment conditions used for experiments and the scale at which testing is being performed, which raises questions about the validity of scaling up observed nanoscale phenomena/mechanical properties to bulk samples. The best current approach for applying these techniques in corrosion studies may be to develop a statistical database with many tests (costly, time-consuming) or consider in situ analysis to be representative of the relative behavior of compromised and uncompromised regions. Regardless, the potential of in situ micro-to-nanoscale SEM and TEM techniques for studying SCC and associated precursors is very positive.
Limiting Precursor Development Using Alloy 690
Issues with PWSCC of Alloy 600 resulted in the material eventually being replaced by Alloy 690 in many PWRs. Alloy 690 was chosen as a replacement due to improved performance in primary water environments, without prior mechanistic or precursor knowledge. However, the performance of Alloy 690 can be linked to the composition of the alloy, which directly acts to limit suspected precursors to PWSCC of Alloy 600. Alloy 690 has a significantly higher Cr content (approximately 30 wt%), which aids in the formation of passivating surface oxides. Theoretically, the increase in Cr content should provide the necessary outward Cr flux to allow for external surface oxidation, limiting the penetrative intergranular oxidation observed as a precursor to PWSCC of Alloy 600. Although SCC is possible in Alloy 690 and Alloy 800 after extreme cold work,88-90 several studies have reported passivation of the alloys in primary water conditions and the in-service performance has been excellent thus far.88,91-93
Persaud, et al.,91 performed preliminary work examining the oxidation behavior of Alloy 690 by exposing flat coupon specimens to a 480°C hydrogenated steam environment with the oxygen partial pressure maintained well below the NiO dissociation pressure. Under the accelerated primary water conditions, they reported that a protective surface oxide was formed on the surface of grain boundaries on Alloy 690, which effectively inhibited grain boundary oxide penetration; optical images in Figure 7 highlight their observations. Surprisingly, intragranular internal oxidation and metallic Ni expulsion were still observed, even with the high Cr content of Alloy 690; this finding is further discussed in Moss, et al.,88 in the context of classical internal oxidation. The width of the protected regions at grain boundaries in Figure 7 are not due to Cr diffusing up the grain boundaries and then laterally—it must be due to a subsidiary stress relief mechanism wherein Ni diffuses laterally to grain boundaries as well as vertically to the surface to relieve stress from internal Cr oxidation. Over a short time, this enables Cr to passivate the grain boundary region affected by the short-term compression from the internal oxidation.83
Further preliminary 3D time-of-flight secondary ion mass spectrometry (ToF-SIMS) analysis by Zhu, et al.,94 on Alloy 690 samples exposed to a similar hydrogenated steam environment is shown in Figure 8. Their analysis was performed on a 50 μm × 50 μm area and captured ppb-level chemistry in a micro-volume of material across four grains (labeled in [a]), at the expense of spatial resolution. Internal Cr oxidation (green) is observed beneath the surface, shown in (b), penetrating grains two, three, and four with only Ni enrichment (base metal) present at grain one. Chemical analysis of the top surface in (c) revealed that oxide penetration does not occur at grain one and grain boundaries because of the formation of a passive external Cr-rich oxide, similar to observations in Figure 7. The favorable external oxidation on some grains on the Alloy 690 sample was concluded to be due to change in diffusion kinetics with grain surface orientation.94 In general, the complementary ToF-SIMS data confirmed that the original observations from Persaud, et al.,91 were consistent on a larger scale and in three-dimensions. Also, the elemental sensitivity of ToF-SIMS allowed for determining the behavior of minor elements (e.g., Ti and Al) and will be reported in a future paper.94
The replacement of Alloy 600 with Alloy 690 or Alloy 800 was originally based on superior performance in laboratory experiments, and not on the basis of a micro-to-nanoscale understanding of the SCC mechanism and/or precursor. However, the transition to high Cr-containing Ni alloys in SGs, Alloy 690 and Alloy 800, still demonstrates how fundamental understanding of known precursors to SCC can be used in future to develop strategies to mitigate precursor development. Also, mechanistic and precursor insight allows for verifying that excellent in-service performance of Alloys 690 and 800 in primary water environments can be expected in future. Therefore, the importance and potential benefit of understanding nanoscale precursor sites for combating SCC failure is clearly demonstrated.
SUMMARY, BENEFITS, AND FUTURE OUTLOOK
The QMN approach, developed by Staehle,19-20 suggested that prediction of SCC failure requires a micro-to-nanoscale understanding of six critical segments, shown in Figure 1, of which quantitative data of each could be later used to develop computational models of SCC. Staehle recognized the benefit of recent advancements in microscopy, which now allow for acquisition of the micro-to-nanoscale chemical, material, and mechanical property information required to understand each SCC segment. Using the QMN approach as a vehicle, we reviewed several recent examples from studies evaluating SCC of Ni-Fe-Cr alloys in nuclear-relevant environments which applied cutting-edge microscopy techniques to better understand SCC mechanisms and precursors. Discussion was provided in the context of the QMN approach to demonstrate the importance of nanoscale precursor sites in the prediction of SCC failure.
In general, a reverse approach to that of Figure 1 was beneficial, wherein a sound mechanistic understanding of SCC propagation was obtained first and subsequently used to hypothesize potential SCC precursors. A susceptible material in the precursor stage of degradation could then be evaluated using a variety of micro-to-nanoscale techniques. Caustic, Pb-induced, and S-induced SCC in Alloy 800 were reviewed as initial examples to demonstrate the application of modern microscopy techniques in corrosion studies. It was revealed that precursors to secondary side SCC of Alloy 800 could include monolayer levels or Pb or S at oxide-metal interfaces or the onset of dealloying. The benefit of the described QMN approach for understanding SCC mechanisms and precursors at the micro-to-nanoscale is realized in the ability to develop mitigation strategies to avoid SCC failure if precursor or incubation markers are identified on materials. For example, detection of dealloying on Alloy 800 on in-service SG tubing is now theoretically possible, offering the potential for mitigating actions.
Having demonstrated the application of modern microscopy for identification of SCC mechanisms and precursors, in the context of the QMN approach, the review was extended to intergranular oxidation embrittlement of Alloy 600. It was demonstrated that a multi-length scale analysis covering the entire breadth of spatial resolution and elemental sensitivity is often required to develop a thorough understanding of nanoscale precursors. Also, initial studies applying novel techniques still in development stages, such as in-situ environmental and mechanical testing, have shown promise for potentially expanding knowledge of SCC mechanisms and precursors. For example, it is now possible to generate and analyze intergranular oxidation embrittlement of Alloy 600 using in situ TEM in near-representative environments. Finally, the high Cr content of the 690 alloy limits the intergranular oxidation precursor to PWSCC of Alloy 600 by allowing for the formation of a passivating Cr-rich external oxide on the surface of grain boundaries. Using Alloy 690 as an example, the benefit of a fundamental understanding of SCC precursors is demonstrated in the ability to change material or environmental conditions to effectively mitigate known issues.
Identification of a nanoscale chemical SCC precursor site (e.g., dealloying, intergranular oxidation, etc.) does not necessarily mean that SCC in imminent. While concerning, additional measures should be taken when assessing SCC susceptibility because mechanical/material conditions and/or incubation period length are variables which alter the precise onset of SCC. For example, a sufficient grain boundary oxide coverage, accumulated over an incubation period, is likely required to initiate PWSCC in Alloy 600. Similarly, the pore size and thickness of a dealloyed layer is critical for the initiation of SCC in Alloy 800 exposed to alkaline conditions (mechanical/material).
This review has demonstrated the potential of using fundamental knowledge of nanoscale precursor sites for predicting SCC failure. In future, continuing technological advancements in microscopy will enable new testing methods, furthering the understanding of SCC mechanisms and precursors. For example, preliminary in situ experimental corrosion testing in liquid TEM holders has already been demonstrated.95-96 In addition, the quantitative and reproducible information acquired using advanced microscopy techniques will eventually be applicable to developing atomic scale computational models of the six segments of SCC in the QMN approach. In fact, this review demonstrates that Staehle’s original vision for modeling of SCC in the context of the QMN approach may be fully realized in the near future. This approach is not limited to SCC of nuclear materials, and should be considered applicable to a broad spectrum of SCC systems and degradation modes.