A Review of Oxidation Phenomena in Ni Alloys Exposed to Hydrogenated Steam Below 500°C

Ni-Fe-Cr alloys are used for several components in the primary circuit of nuclear power plants, such as steam generator tubing. Currently, Alloy 690 (Ni-30Cr-10Fe, UNS N06690⁽¹⁾) and Alloy 800 (Fe-33Ni-23Cr, UNS N08800) are the preferred materials in pressurized water reactors (PWR) and Canadian deuterium uranium (CANDU) reactors. Alloy 600 (Ni-16Cr-9Fe, UNS N06600) was used previously but was found highly susceptible to primary water stress corrosion cracking (PWSCC);^1-14  thermally treated (TT) Alloy 600 is still used in some cases as a result of its added resistance through intergranular Cr carbide precipitation.^15-19 

The temperature in the primary circuit ranges from 280°C to 320°C in CANDU reactors and can reach up to 340°C in PWRs. Hydrogen is a key addition to primary water, which results in the potential being in the vicinity of the Ni/NiO equilibrium potential. The PWSCC growth rate of Alloy 600 is known to peak near the Ni/NiO equilibrium potential;^9-13  therefore, the mechanism of stress corrosion cracking (SCC) is likely related and most investigations of PWSCC in Alloy 600 are performed in environments at or near this potential. The conditions of primary water coincide with those where internal oxidation is often reported to occur at much higher temperatures, such as 800°C.

The internal oxidation phenomenon is regarded to be possible in alloys with a noble solvent element and a low concentration of a reactive solute element.^20-28  When exposed to a gas mixture, or at an oxide/metal interface, with an oxygen partial pressure within range of the dissociation pressure of the solvent metal oxide, the reactive solute element can oxidize internally rather than externally. Significant compressive stress is generated in the internally oxidized zone, which is often relieved through solvent metal expulsion to the surface. The internal oxidation mechanism was first proposed as the mechanism of PWSCC in Alloy 600 by Scott and Le Calvar.²⁹  They hypothesized that internal oxidation occurring intergranularly in Alloy 600 would result in embrittlement and eventually SCC; substantial evidence has accumulated over the past two decades supporting internal oxidation as the mechanism of PWSCC in Alloy 600.^1-7,30-51 

Hydrogenated steam between 400°C and 500°C is considered to simulate an accelerated primary water environment;¹⁴  the rationale for this environment is discussed further in the Rationale for 400°C to 500°C Hydrogenated Steam Environment section. Several studies have accumulated over the past two decades investigating oxidation phenomena and SCC in Ni-Fe-Cr alloys exposed to high-temperature hydrogenated steam environments.^{30-33,38-47,50-51}  In particular, the intergranular oxidation tendency is of particular interest when investigating embrittlement and SCC in Alloy 600 and other Ni alloys. Some critical factors which influence oxidation in high-temperature hydrogenated steam are composition, temperature, and cold work.

The following is a review of the accumulated literature for oxidation phenomena observed in Ni-Fe-Cr alloys exposed to 400°C to 500°C hydrogenated steam. Inter- and intragranular oxidation will be reviewed separately; differences in oxidation processes would be expected at grain boundaries and in the lattice as a result of differences in diffusion rates. Alloy 600 will be explored independently because of its “low” Cr content of 16 at%, which is more likely to induce internal oxidation inter- and intragranularly, possibly resulting in embrittlement. Following this, the limited studies on Ni alloys with high reactive element contents, such as Alloys 690 and 800, will be reviewed. Finally, comparisons will be drawn between oxidation phenomena reported in 300°C to 360°C primary water and 400°C to 500°C hydrogenated steam to evaluate the validity of the use of the latter type of environment as a simulated accelerated primary water environment.

As mentioned, 400°C to 500°C hydrogenated steam is considered to simulate an accelerated primary water environment. This assertion stems from a study by Economy, et al., where they reported that the time to intergranular SCC initiation in Alloy 600 U-bends exhibited an Arrhenius behavior extending from 300°C primary water to 400°C hydrogenated steam.¹⁴  In addition, lattice diffusion of Cr is essentially negligible over short laboratory timescales in hydrogenated steam at temperatures of 500°C and below, similar to lower temperature primary water environments.

The ratio of hydrogen to steam is used to control the nominal oxygen partial pressure. As PWSCC growth rate is known to peak in the vicinity of the Ni/NiO equilibrium potential,^9-13  the oxygen partial pressure is usually maintained below or near the dissociation pressure of NiO. Lindsay, et al., have recently shown that Alloy 600 may exhibit a similar peak SCC growth rate dependence at the NiO dissociation pressure in hydrogenated steam environments, providing further support for the use of the latter environment for accelerated primary water tests.³⁹  The common observation of Alloy 600 peak crack growth rates in the range of the Ni/NiO equilibrium, regardless of environment, suggests that the mechanism of SCC is somehow related.

In addition to simulating primary water, hydrogenated steam provides a viable environment to perform fundamental internal oxidation studies. Lack of significant lattice diffusion during short laboratory exposures at temperatures of 500°C and below allows for analyzing internally oxidized zones and stress relief mechanisms in simplified conditions, where lattice diffusion of metal alloying elements can be ignored. The flux of oxygen atoms diffusing inward, assuming immobile metal atoms in the lattice, can be modeled using Fick’s laws. Any observed metal atom diffusion can be assumed to happen via short-circuit diffusion pathways. For example, expulsion of the solvent element to the surface is often reported to accompany internal oxidation of reactive solute elements.^{26-28,30-33,38-40,46,51-52}  The short-circuit path used for solvent metal expulsion has been debated in literature, usually citing three possible mechanisms: extrusion through mechanical action,^26,53  Nabarro-Herring diffusional creep or diffusion via oxide/metal interfaces,^{20,25-26,53-54}  and/or dislocation pipe diffusion.^20,25,28,53  Reconstruction of an internally oxidized zone beneath the expelled solvent metal using a high-resolution technique, such as 3D atom probe tomography (APT), can potentially allow one to identify convincingly which of these mechanisms is operating.

Several studies have explored oxidation of Alloy 600 (Ni-16Cr-9Fe) in hydrogenated steam at temperatures of 500°C and below.^{30-31,38-44,46-47,50-51}  The low Cr content of 16 at% in Alloy 600 is insufficient to produce the outward flux necessary to form an external protective oxide. Instead, in conditions where the oxygen partial pressure is within range of the noble solvent metal oxide, NiO, internal oxidation can occur. Differences in oxidation tendency can be expected inter- and intragranularly as a result of differences in diffusion kinetics. Also, fundamental studies have been done which delineate the processes of internal oxide formation and stress relief in Alloy 600 and likely other Ni-Fe-Cr alloys.

The study of grain boundary oxidation phenomena is of particular importance when evaluating the SCC resistance of Alloy 600. The formation of a continuous, penetrating intergranular oxide would result in embrittlement and possibly SCC under applied stress. Several studies have explored intergranular oxidation, embrittlement, and SCC of Alloy 600 after exposure to 400°C to 500°C hydrogenated steam.^{30-31,38-44,46-47,50-51}  Figure 1 is a scanning electron microscope (SEM) image of the surface of an Alloy 600 sample after exposure to 480°C hydrogenated steam with an oxygen partial pressure maintained below the NiO dissociation pressure.³¹  The sample appears macroscopically shiny with nodular structures distributed within the grains. Grain boundaries are visible with a high density of nodular structures directly adjacent. The focus of this section will be reviewing intergranular oxidation and embrittlement in Alloy 600; a review of intragranular oxidation, internal oxidation, and stress relief in Alloy 600 is presented in the Internal Oxidation and Stress Relief section.

Gendron, et al., studied oxidation and embrittlement in Alloy 600 by exposing electropolished miniature tensile specimens to 400°C hydrogenated steam with 21 kPa D₂ and a total gas pressure of 20 MPa.^41-43  Following autoclave exposures, they fractured their samples by straining and reported that shallow intergranular embrittlement had occurred in Alloy 600 to a depth of approximately 1 μm; Cr₂O₃ was observed to penetrate along grain boundaries. Persaud, et al.,^30-31  and Scenini, et al.,^38-40,51  extended these studies by exposing fine polished and/or electropolished Alloy 600 flat samples to 480°C hydrogenated steam at an oxygen partial pressure well below the NiO dissociation pressure. In all studies, it was reported that oxygen penetration could occur to depths up to 2 μm, resulting in embrittlement and possibly fracture under stress. Figure 2(a) is a scanning transmission electron microscope (STEM) high-angle annular dark-field (HAADF) image of a cross section taken from an exposed Alloy 600 sample from Persaud, et al.^30-31  Figure 2(b) highlights the grain boundary in Figure 2(a), and intergranular penetration of oxygen is clearly visible to a depth of approximately 2 μm.

Figure 3 shows energy-dispersive x-ray spectroscopy (EDX) elemental maps taken at the grain boundary indicating that a mixed Fe and Cr oxide is formed intergranularly. The intergranular oxidation observed in Alloy 600 can be partly attributed to environment conditions.³¹  Specifically, the abundant element Ni is not oxidizing and the reactive elements, Cr and Fe, are not present in sufficient quantity to provide the outward flux necessary to form an external protective oxide. Instead, oxygen can diffuse inward and oxidize reactive solute elements along grain boundaries, resulting in embrittlement and fracture under stress. The oxygen diffusion rate after 120 h is likely still too large to be accounted for solely through grain boundary diffusion and other short-circuit paths must be available, such as oxide/metal interfaces, dislocation pipe diffusion, and/or vacancy injection at the oxide/metal interface resulting from oxidation. Definitive evidence for these mechanisms operating intragranularly have recently been acquired through 3D APT by Langelier, et al.,⁴⁶  and is discussed at length in the Internal Oxidation and Stress Relief section.

A peculiar observation from Figure 3 is the enrichment of Ti on one side of the grain boundary.³¹  In addition, Bertali, et al., have reported the presence of Al enrichment at grain boundaries in Alloy 600 exposed to similar conditions;^39-40,51  Ti and Al are both oxidized. Al and Ti are minor impurities in Alloy 600 with concentrations less than 0.5 at%. Recent APT of an oxidized grain boundary in Alloy 600 exposed to 480°C hydrogenated steam for 120 h revealed that the Ti concentration can reach concentrations up to 8 at%.⁴⁷  This significant enrichment at the grain boundary raises the issue of how minor impurity species accumulate, given that diffusion would have to occur from approximately 200 atomic spaces away. The minor impurities were either present at the grain boundaries initially or significant short-circuit diffusion is occurring;⁴⁷  the latter assertion seems more likely given that intergranular Ti and Al enrichment is not detected in solution annealed Alloy 600 prior to exposure.

Thermal treatment (TT) of Alloy 600 from 700°C to 900°C results in the precipitation of Cr carbides intergranularly.^55-58  The duration of TT is optimized to allow for recovery of initial Cr depletion adjacent to grain boundaries; usually a time period greater than 10 h is used at approximately 700°C.^55-58  Alloy 600TT is known to be more resistant to PWSCC, but the mechanism behind the added resistance is controversial. Cr carbide precipitation results in concentrating of Cr at grain boundaries, which potentially increases the outward flux of Cr to the surface. Therefore, Cr oxidation should be possible near or at the surface of grain boundaries.^4,31,41-43  Persaud, et al., have shown that intergranular oxygen penetration in Alloy 600TT is limited to approximately one quarter the depth observed in Alloy 600SA.³¹  In addition, the oxide formed in Alloy 600TT is more Cr-dense and oxidation is observed to occur favorably at carbides. The findings of Persaud, et al.,³¹  support the prediction of Gendron, et al.,^41-43  who postulated that intergranular Cr carbide precipitation would act as a sink for oxygen. Another proposed theory by Bertali, et al., suggests that diffusion-induced grain boundary migration of solute elements can result in local stress at grain boundaries as a result of volume change, which can rupture an intergranular internal oxide film leading to further internal oxidation of the underlying bare metal.⁵¹  They suggest that the formation of coarse Cr₇C₃ carbides in Alloy 600TT can “pin” grain boundaries, leading to the observed increase in PWSCC resistance.⁵¹  Other theories argue that carbides assist by blocking transmission of slip through the grain boundary or dislocation movement during grain boundary sliding,¹⁵  resulting in the slip process necessary for oxide rupture, and thus SCC, being inhibited. Rather than a single mechanism operating, the proposed mechanisms likely act in synergy increasing the SCC resistance of Alloy 600TT.

Surface finish has been demonstrated to influence oxidation, and potentially the SCC susceptibility, of Ni-Fe-Cr alloys.^39,44-45,59  Scenini, et al., evaluated the SCC susceptibility of Alloy 600 reverse U-bend samples in 400°C hydrogenated steam with conditions maintained at an oxygen partial pressure below the NiO dissociation pressure.⁴⁴  They reported that electropolishing resulted in a higher SCC susceptibility compared with mechanical polishing. This may seem counter-intuitive and, indeed, other studies have been published that indicate that surface cold work is detrimental to SCC resistance in water.^60-62  However, Scenini, et al., argue that mechanical polishing produces a shallow deformed damage layer, rich in dislocations, which aids in the diffusion of elements to the surface.⁴⁴  Thus, external rather than internal oxidation is likely. Electropolishing results in a strain-free surface which limits outward diffusion and inward diffusion of oxygen is favorable. Therefore, the absence of surface cold work permits greater depth of oxygen penetration, particularly at grain boundaries which act as a short-circuit for diffusion. It should be noted that if excessive mechanical polishing is applied, the residual stress at the surface may be sufficient to rupture the external oxide, leading to penetration of oxygen in the underlying bare metal;⁴⁴  this may explain the reported increased SCC susceptibility observed in mechanically polished Alloy 600 in other studies.^60-62  Another possibility is that in primary water the outward Cr flux is insufficient to form an external oxide in the deformed damage layer of mechanically polished Alloy 600. The conflicting results in literature suggest that the influence of surface finish on oxidation should be studied further; however, there is agreement that oxidation kinetics are heavily dependent on the degree of surface cold work.

Capell and Was performed experiments investigating surface oxidation and possible selective internal intergranular oxidation of Alloy 600 in 400°C hydrogenated steam at 1 bar (100 kPa) with the partial pressure ratio of hydrogen to steam varied between 0.001 and 0.9;⁵⁰  experiments done above the equilibrium partial pressure ratio, 0.004 at 400°C, were in the Ni-metal regime. Their results indicated the formation of an unprotective Ni(OH)₂ surface film and intergranular penetration of oxide. However, given the low chromium content of most nickel alloys tested by Capell and Was, one would expect internal oxidation of chromium to be more favorable intragranularly than the formation of a nickel hydroxide surface film. Further investigation should be done using their conditions to evaluate the possibility of Ni(OH)₂ production. However, the findings of Capell and Was do provide support for selective intergranular Cr oxidation resulting in embrittlement and likely SCC in Alloy 600 exposed to 400°C hydrogenated steam.⁵⁰ 

One possible explanation for the findings of Capell and Was could be the use of 1 bar total pressure at 400°C in their experiments. The H₂, H₂O, and nominal O₂ pressures may not be great enough to allow the Alloy 600 surface to achieve the desired equilibrium, in Equation (1), even if the ratios are correct; the Ni alloy surface may not have fast enough kinetics for one or other reactions at 400°C. In addition, the equilibrium can be disturbed by a side reaction, such as Cr oxidation, hydrogen absorption in the metal, or Ni oxidation, without attaining the desired equilibrium in Equation (1).

Other studies on Ni alloys exposed to 400°C hydrogenated steam with approximately 200 bar (20 MPa) total pressure and an O₂ nominal partial pressure maintained below the NiO dissociation pressure did not report significant Ni oxidation;^41-44  higher partial pressures of H₂ and H₂O, compared to Capell and Was, may allow the desired equilibrium, in Equation (1), to be attained faster as a result of the larger quantity of each gas in the environment available to the surface. Faster kinetics of all the involved reactions at higher temperatures, such as 480°C, likely allow for 1 bar total pressure of hydrogenated steam to be sufficient.

The accumulated literature provides strong evidence to support intergranular Cr oxidation as the mechanism of SCC for Alloy 600 exposed to 400°C to 500°C hydrogenated steam in conditions where the oxygen partial pressure is below the NiO dissociation pressure. However, the oxidation phenomenon has been classified by several descriptions, including “selective internal oxidation,”⁵⁰  “preferential oxidation,”^39-40,51  and “internal oxidation.”^{30-33,46-47,63-64}  The intergranular Cr oxide formed in the aforementioned studies does not always form as discrete particles, as defined in classical internal oxidation. However, the conditions used to induce intergranular selective Cr oxidation in Alloy 600 are similar to those in which internal oxidation would be expected to happen. Also, the cause of intergranular Cr oxidation, inhibition of solvent metal oxidation and low solute element concentration, is undoubtedly comparable to internal oxidation at higher temperatures. Therefore, the wide range of classification for the oxidation phenomenon at temperatures of 500°C and below is understandable. It is proposed that a suitable explanation for the observed intergranular oxidation would likely be selective intergranular oxidation of Cr formed through a process analogous to internal oxidation at higher temperatures, such as 800°C.

It remains unclear whether the oxide particles are initially formed as discrete particles and coalesce as the intergranular oxide develops in Alloy 600; in this case, classical intergranular internal oxidation would be occurring. A more detailed study at higher resolution is necessary to identify conclusively the initial process of intergranular oxide formation in Alloy 600; although, in 330°C representative primary water, Schreiber, et al., have reported continuous intergranular selective Cr oxidation in Alloy 600, with no evidence from APT to support discrete oxide particle formation at or ahead of the intergranular oxide.⁶⁵  However, classical internal intergranular oxidation, resulting in discrete oxide particle formation, has been shown to be possible by Schreiber, et al., in 360°C representative primary water in Ni-4Al alloys after exposure for 1,000 h.^66-67  Whether a shift from solid state oxygen diffusion in the metal (e.g., Ni-4Al) to oxygen ion diffusion in the intergranular oxide (e.g., Alloy 600) would influence the fundamental mechanism of SCC is not known and should be investigated further. In the case of oxygen diffusion in the metal, there may be an element of plastic fracture between the oxide particles, but oxygen segregation would likely occur to these areas, which would make the fracture predominantly brittle.

Scenini, et al., were first to report internal oxidation resulting in solvent metal, Ni, expulsion to the surface in Alloy 600 exposed to 480°C hydrogenated steam.³⁸  Classical internal oxidation is usually studied at temperatures where lattice diffusion rates are relevant, such as 800°C. As a result, the concentration of the solute element is usually less than the 16 at% in Alloy 600; for example, Stott, et al., performed several studies investigating internal oxidation of Ni-Cr and Ni-Al alloys at temperatures of 600°C and greater, with a solute reactive element content of 5 at% and below.^{20,25-26,53-54}  Also, by decreasing temperature to 500°C and below, where lattice diffusion is negligible over short laboratory time scales, diffusion of metallic Ni to the surface cannot occur through the lattice without significant assistance from short-circuits, such as dislocations and/or oxide/metal interfaces.

Persaud, et al., further examined the processes of internal oxide formation and solvent metal expulsion.^30-31  Initial EDX STEM analysis of a metallic Ni nodule and the internal oxide beneath an Alloy 600TT sample is shown in Figure 4; these maps were taken from a nodule and the underlying internal oxide, similar to areas identified in Figure 2(a). The nodule is confirmed to be essentially metallic Ni and has a composition consisting of greater than 95 at% Ni. Internal oxides form an interconnected network extending downward and consist of a mixture of Cr and Fe oxides. The mechanism of solvent metal expulsion is of interest and has been explored for decades; the outward solvent element flux is too great to be accounted for simply through lattice diffusion, and several theories have been proposed for possible short-circuit pathways:

• •

  Oxide/metal interfaces:^{20,25-26,53-54}  Mackert, et al.,²⁶  proposed that the volume expansion, as a result of the formation of internal oxides, generates significant compressive stress, which reduces the vacancy concentration at the internal oxide/metal interface. As a result, a vacancy gradient is created between the internal oxidation front and the stress-free alloy surface. Vacancies will then diffuse from high concentration at the alloy surface to the oxide/metal interface and a counter flow of solvent metal atoms diffuses from the oxide/metal interface to the surface.

• •

  Dislocation pipe diffusion:^20,25,28,53  Guruswamy, et al.,²⁸  proposed that the solvent metal expulsion phenomenon was better described using dislocation pipe diffusion. Dislocation pipe diffusion assumes that the regions around the oxide particles deform by slip resulting in large dislocation densities which provide a path to the surface.

• •

  Mechanical action:^26,53  The compressive stress generated in the internally oxidized zone is relieved through mechanical action extruding material to the surface. Mackert, et al., reasoned that extrusion through mechanical action was unlikely because there was no evidence of oxides being present in the nodules; minute internal oxide particles should be carried with the solvent metal to the surface with the plastically deforming alloy if such a mechanism was operating.

Studies have shown that the volume of solvent metal accumulated on the surface is equivalent to the volume increase resulting from internal oxide precipitation.²⁰  Thus, the driving force for noble solvent metal expulsion is undoubtedly the compressive stress generated during the formation of internal solute oxides. Yi, et al., demonstrated this in Ni-Al and Ni-4Al-xSi alloys exposed to Ar-yO₂ at temperatures ranging from 1,073 K to 1,273 K with an oxygen partial pressure, y, below the Ni/NiO equilibrium dissociation pressure.²⁰  They found that the measured volume of expelled Ni was comparable to the difference between the volume of Al and Si internal oxides and the volume of metals. The results of Yi, et al., are assumed to extend to lower temperature environments where lattice diffusion is negligible. It should be noted that stress relief is not always accomplished through expulsion of the solvent element to the surface. Other mechanisms of stress relief have been proposed, including vacancy injection at the oxide/metal interface as a result of external oxide formation²⁰  and grain boundary migration.⁵² 

One advantage to studying the processes of stress relief in 400°C to 500°C hydrogenated steam is retardation of lattice diffusion, which allows for clear identification of short-circuit diffusion pathways. As shown in Figure 4, Persaud, et al., performed transmission electron microscope (TEM) analysis from which it could only be concluded that metallic Ni was on the surface and the internal oxide was a mixed Fe- and Cr-oxide.³¹  Langelier, et al., extended the study to examine solvent metal expulsion and internal oxide formation in Alloy 600 exposed to similar conditions using 3D APT;⁴⁶  3D APT allows for reconstruction of a needle-like sample at a subnanometer level of resolution. Figure 5 is a 3D APT reconstruction of a metallic Ni nodule with the underlying internally oxidized zone in Alloy 600 from Langelier, et al.⁴⁶  It is apparent that the expelled metallic Ni is directly connected to the grain beneath through a Ni network at oxide/metal interfaces,⁴⁶  providing strong evidence to support the mechanism of oxide/metal interface short-circuit diffusion. However, volume increase associated with internal oxide formation was enormous, at approximately 44%.⁴⁶  With such a large volume increase, massive dislocation generation will occur because of plastic deformation around growing internal oxides. Also, further analysis by Langelier, et al., revealed that oxide particles align along planes associated with common slip planes in face-centered cubic Ni.⁴⁶  Therefore, dislocation pipe diffusion and diffusion via oxide/metal interfaces both likely contribute to increasing the diffusion rates observed during internal oxidation processes in Alloy 600 and other Ni alloys. Similar 3D APT studies should be conducted in environments where conventional internal oxidation is studied to confirm whether the findings of Langelier, et al., can be extended to higher temperatures.

Langelier, et al., reported that the process of internal oxide formation in Alloy 600 begins with the formation of a Cr₂O₃ core, which grows until the surrounding matrix becomes Cr-depleted and FeCr₂O₄ formation becomes favorable.⁴⁶  As a result, there is an interconnected skeletal structure with Cr-rich cores surrounded by FeCr₂O₄ shells. The morphology of the oxide conforms well to the percolation theory proposed by Newman, et al., to describe the transition from internal to external oxide at low temperatures, such as in 400°C to 500°C hydrogenated steam.⁴⁵  It is based on the idea that the internal oxide/metal solvent interface and adjacent dislocation networks provide good short-circuits for diffusion. At low temperatures, the inward diffusion of oxygen would be fast and would internally oxidize less noble species. There is a percolation requirement that the random internal oxide would form a connected skeletal structure rather than particles being dispersed.⁴⁵  The nanometer-scale connected skeletal oxide provides a short-circuit diffusion path for Cr to be transported to the surface; the skeletal oxide/metal interface provides higher diffusivity of the reactive alloying element to the surface compared with the lattice.

Surface finish has proven to be a critical factor in oxidation of Ni-Fe-Cr alloys in high-temperature hydrogenated steam. Lindsay, et al., studied the difference in oxidation tendency, internal or external, of Alloy 600 as a function of surface finish.³⁹  They exposed 600 grit ground and fine polished Alloy 600 flat samples to 350°C to 500°C hydrogenated steam with an oxygen partial pressure maintained below the NiO dissociation pressure. They reported that ground surfaces contained a thick deformed layer, rich in dislocations, which allowed for formation of a continuous external oxide and negligible inward oxygen penetration.³⁹  Electropolished surfaces underwent conventional internal oxidation both inter- and intragranularly. Likewise, Persaud, et al., reported that fine polished Alloy 600 samples exposed to 480°C hydrogenated steam experienced enhanced diffusion of metallic Ni to the surface in areas with minor scratches.³¹  The exact amount of applied surface cold work necessary to induce the transition from internal to external oxide remains a topic of interest; exposure of several samples with different surface status would allow for analysis of the transition from internal to external oxide. Also, the mechanism of the transition could be analyzed to determine whether the percolation threshold requirement theory of Newman, et al.,⁴⁵  is applicable.

Alloy 690 and Alloy 800 are currently used to construct many components in PWR and CANDU nuclear power plants. The alloys contain significantly greater Fe and/or Cr content compared with Alloy 600, which is believed to contribute to the formation of more passivating external oxides. Limited studies have been done on Alloys 690 and 800 in hydrogenated steam at 500°C and below, where lattice diffusion is essentially negligible over short exposure durations.^32,52  Theoretically, the reactive element contents in Alloy 690 and Alloy 800 should be far too great to promote internal oxidation as observed in Alloy 600. However, recent studies have indicated that the alloys may not behave as predicted.^32,52  In this section, the oxidation tendency, internal or external, of Ni alloys with high reactive element contents will be explored.

Persaud, et al., exposed Alloy 690 to 480°C hydrogenated steam with an oxygen partial pressure maintained well below the NiO dissociation pressure.³²  Unlike Alloy 600, Alloy 690 contains 29 at% Cr which significantly increases the outward Cr flux, especially at grain boundaries. Figure 6 shows optical (a) and SEM (b) images of the Alloy 690 surface after exposure.³²  Within the grains there are nodular structures, similar to Alloy 600 in Figure 1; nodule formation and intragranular oxidation are discussed further in the Intragranular Internal Oxidation section. The grain boundaries are covered in a greenish film, which spaces the nodular features within the grains 2 μm away on either side.³²  As mentioned in the Intergranular Oxidation and Embrittlement section, the diffusion coefficient at grain boundaries is several magnitudes higher compared to in the lattice in hydrogenated steam at 500°C and below. Therefore, it is reasonable to assert that a reactive alloying element could diffuse to the surface and oxidize externally at grain boundaries.

The external oxide was confirmed by Persaud, et al., to be a thin Cr-rich oxide, likely Cr₂O₃, using a focused ion beam (FIB) trenching technique combined with EDX analysis, shown in Figure 7. Cr can diffuse outward and oxidize on the surface of the grain boundary effectively consuming oxygen and limiting further oxygen ingress. As a result, the intergranular embrittlement observed in Alloy 600, in Figure 3, is avoided. Persaud, et al., concluded that the formation of a thin Cr-rich external oxide at grain boundaries may partly explain the SCC resistance of Alloy 690.³² 

Alloy 800 is a high alloy stainless steel that contains more Cr and significantly more Fe than Alloy 600. In fact, the combined reactive element content in Alloy 800 of approximately 60 at% far exceeds the approximately 25 at% in Alloy 600. The oxidation tendency of Alloy 800 was evaluated by Persaud, et al., in 480°C hydrogenated steam with an oxygen partial pressure maintained well below the NiO dissociation pressure.⁵²  As expected, after exposure the samples appeared to have a black film formed on the surface, likely a result of the high Fe content in Alloy 800. Further SEM imaging of the surface after exposure revealed that grain boundaries were visible with a thick external oxide visible within the grains, likely the macroscopically observable black film.⁵² 

Closer examination of the area beneath the grain boundary in Alloy 800 using TEM revealed that a thin 3 nm to 4 nm Cr-rich oxide was formed at the surface of grain boundaries, similar to Alloy 690.⁵²  It should be noted that the visible black film was indicated to be Fe-rich and was porous, allowing for inward penetration of oxygen in areas away from grain boundaries. Further work was done by Persaud, et al., on an Alloy 82 (UNS N06082) dissimilar metal weld between parent materials of carbon steel and Alloy 600.³³  Dilution of carbon steel during welding resulted in the composition of the weld becoming Fe-enriched, approximately 34 at%, and Cr-depleted, approximately 11 at%, at the point of exposure. The weld alloy was exposed to a hydrogenated steam environment, similar to Alloy 690 and Alloy 800. Study of the grain boundary chemistry after exposure at the point of exposure, the root, revealed the presence of a thick external Fe-rich oxide film at the surface of grain boundaries.⁵²  However, oxygen penetration did occur to a depth of approximately 500 nm; oxygen diffusion was hindered compared with Alloy 600, in Figure 2(b), but was still possible as a result of the formation of a porous Fe-rich oxide, likely Fe₃O₄, rather than the Cr-rich oxide observed in Alloys 690 and 800.

The accumulated literature suggests that the formation of a Cr-rich oxide film at the surface of grain boundaries is critical to limiting inward penetration of oxygen intergranularly. A thick Fe-rich oxide film is beneficial in retarding oxygen diffusion compared with Alloy 600, but likely does not entirely prevent embrittlement. Therefore, Alloy 800 and Alloy 690 are concluded to be more resistant to embrittlement and SCC than Alloy 600 in primary water-type environments resulting from their high Cr contents. Studies indicate that the added Fe in Alloy 800 does not aid in preventing intergranular inward oxygen ingress because of the formation of a porous Fe-rich external oxide;^33,52  nevertheless, the Cr content of the alloy is great enough to allow for a thin and protective external oxide at grain boundaries which prevents oxygen ingress. However, it should be noted that the increased Fe content in Alloy 800 has been shown to be beneficial for preventing embrittlement/SCC in other aqueous environments relevant to nuclear power plants.⁶⁸  The ideal Ni, Fe, and Cr ratios which limit embrittlement and SCC in aqueous nuclear power plant environments should be investigated; some important work on this topic has been done recently by Arioka, et al.⁷ 

Figure 6 shows images of an Alloy 690 surface after exposure to 480°C hydrogenated steam from Persaud, et al.³²  Nodular structures are distributed within the grains. These nodules were confirmed to be composed of metallic Ni, similar to Alloy 600. Expulsion of metallic Ni is evidence that internal oxidation was occurring in Alloy 690. A summary of oxidation phenomena reported by Persaud, et al., in Alloy 690 is presented in Figure 8.³²  There is clear evidence of internal oxidation intragranularly to a comparable depth as Alloy 600 in Figure 2(a). It is remarkable to observe internal oxidation in Alloy 690 which contains 29 at% Cr. Hindered lattice diffusion in 480°C hydrogenated steam likely prevents the outward intragranular Cr flux from being great enough to promote external oxide formation, even with 29 at% Cr present. Classical internal oxidation can then ensue with compressive stress relief accomplished through expulsion of metallic Ni to the surface.

Persaud, et al., performed further work studying possible internal oxidation in Alloy 800 exposed to a similar high-temperature hydrogenated steam environment.⁵²  TEM analysis of the external surface film and the underlying metal revealed that internal oxidation had occurred to a depth comparable to Alloy 600 and Alloy 690. In Alloy 800, the external intragranular surface film is a mixture of metallic Ni and an Fe-rich oxide, likely magnetite. Ni is in fact a solute alloying element in Alloy 800, with a content of 32 at%, making it exceptional that internal oxidation and metallic noble metal expulsion is possible. The external Fe-rich film likely injects significant vacancies at the oxide/metal interface which relieves most of the compressive stress resulting from internal oxide formation. However, metallic Ni expulsion to the surface is still possible in some regions, either as individual nodular structures or, more commonly, beneath the Fe-rich oxide.⁵²  Also, visible grain boundary migration was reported ahead of and adjacent to some internally oxidized zones, suggesting outward diffusion is not always the most favorable stress relief mechanism. Several stress relief mechanisms were observed to operate simultaneously in Alloy 800.

The observed internal oxidation in alloys with reactive element concentrations significantly higher than the usual threshold concentrations required for transition from internal to external oxide is likely a result of hindered lattice diffusion and the unique temperature range used. As mentioned, in the 400°C to 500°C temperature range the outward flux of metal alloying elements does not allow for external intragranular oxidation, even at high concentrations. However, the diffusion kinetics are still fast enough to allow for inward oxygen diffusion intragranularly and, under sufficient compressive stress, expulsion of the noble element to the surface. Studies on Alloy 800 prove that internal oxidation can be induced by maintaining an oxygen partial pressure in the range of a noble solute metal oxide, NiO. The findings of Persaud, et al., and possible exceptions to classical internal oxidation theory, are likely limited to a narrow temperature range where lattice diffusion of alloying elements is negligible but significant internal oxidation can still occur, producing high compressive stress.

Previous sections reviewed oxidation reported in Ni-Fe-Cr alloys exposed to 400°C to 500°C hydrogenated steam with an oxygen partial pressure maintained below the NiO dissociation pressure. Attention is now turned toward the practical applicability and validity of this high-temperature hydrogenated steam environment as a simulant for accelerated primary water conditions; comparisons will be made to oxidation observations commonly reported in 300°C to 360°C primary water.

In 1993, Scott and Le Calvar were first to propose internal oxidation as a possible mechanism of PWSCC in Alloy 600.²⁹  As reviewed in the Alloy 600 section and Ni Alloys with High Reactive Solute Element Contents section, internal oxidation is known to occur in Ni-Fe-Cr alloys exposed to hydrogenated steam at 500°C and below with oxygen partial pressures near or below the dissociation pressure of the noble metal oxide, NiO. The potential dependence of PWSCC suggests that the mechanism of cracking operates in a similar manner as internal oxidation.

Lozano-Perez, et al., investigated the mechanism of PWSCC in Alloy 600 exposed to 360°C representative primary water for 5,000 h with a potential maintained slightly above the Ni/NiO transition and reported that chromium oxide formed intergranularly with nanocrystalline NiO interspersed.³  While Lozano-Perez, et al., did not observe expulsion of metallic Ni to the surface, as observed in Alloy 600 exposed to 480°C hydrogenated steam, they did report segregation of metallic Ni away from the intergranular chromium oxide and enrichment of Ni ahead of the crack tip.³  Solvent metal expulsion was not observed in their work, probably because of kinetic limitations at the lower temperatures of primary water and their conditions, which were above the Ni/NiO equilibrium potential. Lozano-Perez, et al., suggest that their observations did not directly support internal oxidation, as classically defined, because of continuous, rather than discontinuous, Cr and O enrichment along grain boundaries.³  However, oxygen does penetrate grain boundaries and selectively oxidizes Cr, leading to embrittlement and PWSCC, as originally suggested by Scott and Le Calvar.²⁹  The intergranular oxidation reported by Lozano-Perez, et al., is similar to that observed in Alloy 600 exposed to 400°C to 500°C hydrogenated steam.^{30-31,39-40,51}  In addition, the oxide formed in both environments is Cr-rich and results in intergranular embrittlement and eventual fracture.

Further work by Persaud, et al., was done in 315°C representative primary water conditions with 13.7 cm³ H₂/kg H₂O added;⁶³  these conditions resulted in the potential lying well into the Ni-metal regime. The environment used by Persaud, et al., prevents Ni oxidation and allows for clearer understanding of the role of Cr and O. Also, the environment is ideal for comparison to high-temperature steam environments where the oxygen partial pressure is often maintained on the Ni-metal side of the Ni/NiO dissociation pressure. Persaud, et al., reported that intergranular oxygen penetration occurred in Alloy 600 to a depth of approximately 2 μm after only 1,190 h of exposure to 315°C representative primary water.⁶³  The diffusion rate of oxygen in these conditions is clearly accelerated compared with other studies performed in representative primary water, but the cause remains unknown. In Alloy 800, intergranular penetration of oxygen was limited compared with Alloy 600, at a maximum of 100 nm.⁶³  The conclusion drawn was that Alloy 600 is likely susceptible to PWSCC as a result of intergranular oxidation resulting in embrittlement, while Alloy 800 was more resistant to PWSCC as a result of the formation of a thin Cr-rich external oxide at the surface of grain boundaries;⁶³  once again, in spite of the large Fe content of Alloy 800, Cr oxidation is critical for limiting intergranular oxygen penetration. The intergranular oxidation and embrittlement in representative primary water reported by Persaud, et al., and others is closely analogous to those results reported in 400°C to 500°C hydrogenated steam for Alloy 600 and Alloy 800, discussed in the Intergranular Oxidation and Embrittlement section and the Oxidation at Grain Boundaries section, respectively.

Lozano-Perez, et al., proposed an internal oxidation/film rupture mechanism to describe PWSCC in Alloy 600.⁶⁴  Intergranular penetration of oxygen results in selective Cr oxidation, with the possibility of interspersed NiO if conditions are above the Ni/NiO equilibrium potential. Following this, chromia penetrates deeper along grain boundaries; diffusion may be accelerated along the oxide/metal interface or by dislocations generated as a result of intergranular oxide growth. Under applied stress, the embrittled grain boundary fractures at the free surface, opening the crack. Further intergranular oxygen penetration of the exposed bare metal occurs as primary water fills the open crack.⁶⁴  The mechanism proposed by Lozano-Perez, et al., is a result of local stresses and strains on active slip bands, whereas the slip dissolution mechanism commonly described in literature is based on a continuum model.

The internal oxidation/film rupture mechanism seems applicable to SCC observed in Alloy 600 exposed to 400°C hydrogenated steam. For example, Scenini, et al., reported that mechanically polished Alloy 600 C-ring samples were more resistant to SCC than electropolished samples after exposure to 400°C hydrogenated steam.⁴⁴  However, ground surfaces were more susceptible to SCC as a result of the stress in the deformed layer fracturing oxides and allowing for internal oxidation of the underlying material.⁴⁴  The work of Scenini, et al., conforms well to the mechanism proposed by Lozano-Perez, et al.⁶⁴  Bertali, et al., exposed Alloy 600 to 480°C hydrogenated steam and identified that extensive grain boundary migration had occurred and suggested that their findings could support an internal oxidation/film rupture mechanism.⁵¹  Bertali, et al., have proposed that the local stress increase at grain boundaries through grain boundary migration can rupture the internal oxide film leading to further internal oxidation.⁵¹ 

Alloys 690 and 800 have not been reported to undergo SCC in the absence of applied cold work. However, under severe cold work SCC has been reported in Alloy 690 in several studies.^69-73  Alloy 800 has not been studied at great depth in representative primary water conditions, but has recently been reported by Arioka, et al., to undergo shallow and localized SCC in representative primary water.⁷  The mechanism of SCC in severely cold-worked Alloy 690 and Alloy 800 is debated in literature.^69,73  Whether internal intergranular oxidation plays a role in embrittlement or SCC in cold-worked Alloys 690 or 800 remains unknown and is beyond the scope of this review; other than surface cold work, essentially no studies have investigated oxidation and/or SCC of cold-worked Ni-Fe-Cr alloys in high-temperature hydrogenated steam.

The accumulated literature suggests that there is good agreement between intergranular oxidation phenomena observed in 300°C to 360°C primary water and 400°C to 500°C hydrogenated steam for Alloys 600, 690, and 800; Economy, et al., were first to propose that the mechanism of PWSCC for Alloy 600 in both environments would be similar.¹⁴  Further studies in representative primary water should be done for Alloy 800, in general, and in conditions at potentials in the Ni-metal regime. It should be noted that the intergranular oxidation in Alloy 600 reported in both environments is not necessarily internal oxidation as classically defined. However, because the observed selective intergranular Cr oxidation is a result of conditions being in the vicinity of the Ni/NiO equilibrium and occurs through a process analogous to classical internal oxidation, relating the SCC mechanism to internal oxidation is appropriate. As mentioned, further work should be done to identify whether the initial intergranular oxides are formed as discrete oxide particles, which would confirm whether or not the mechanism of embrittlement is internal intergranular oxidation.

Olszta, et al., have reported that penetrative internal oxidation can occur intragranularly in Alloy 690TT after exposure to representative primary water at 360°C for 5,000 h with conditions at the Ni/NiO equilibrium potential.^74-75  Intragranular penetration likely occurs as a result of hindered lattice diffusion kinetics, which does not allow for sufficient outward Cr diffusion, even with the approximately 29 at% Cr content in Alloy 690. At grain boundaries Olszta, et al., reported that protective Cr₂O₃ formed externally which likely limited inward oxygen diffusion, similar to the studies discussed in the Intergranular Oxidation section.^74-75  Intragranular penetration to any extent in Alloy 690 is surprising given its high Cr content. Also, at 360°C the kinetics are sufficiently hindered and the depth of oxygen penetration in Alloy 690 is approximately 100 nm, one order of magnitude less than observed in 480°C hydrogenated steam;³²  this shallow diffusion depth in the former environment may not build the compressive stress necessary to promote Ni expulsion. However, some oxides containing Ni were reported on the surface, which could potentially be Ni that was expelled then oxidized as a result of their conditions being at the Ni/NiO equilibrium potential.

Persaud, et al., exposed Alloy 800 and Alloy 600 to 315°C primary water for 1,190 h with 13.7 cm³/kg H₂O added, on the Ni-metal side of the Ni/NiO equilibrium potential.⁶³  They reported a shallow internally oxidized zone in Alloy 800, similar to the findings of Olszta, et al., in Alloy 690.^74-75  In addition, a continuous Cr-rich external oxide was present throughout the surface. In Alloy 600, a discontinuous external Cr-rich oxide was formed with no oxygen ingress detected.⁶³  Many other studies have reported the formation of Ni/Fe spinel oxides on the surface, but the environment used in the work of Persaud, et al., was sufficiently reducing to likely inhibit formation of Ni/Fe spinel oxides.⁶³  It is interesting that internal intragranular oxidation is observed to occur, albeit to shallow depth, in Alloys 690 and 800, but not Alloy 600; the reason for this is not understood, but is likely related to the low Cr content of Alloy 600.

Intragranular internal oxidation phenomena observed in Ni-Fe-Cr alloys exposed to representative primary water conditions are subtle compared with 400°C to 500°C hydrogenated steam. These differences can be attributed largely to the slower kinetics at 360°C and below. While lattice diffusion rates are low at 500°C, the kinetics can still support significant inward atomic oxygen penetration resulting in large internally oxidized zones. Ni alloys exposed to supercritical water at higher temperatures, such as 600°C, have been shown to form internally oxidized zones, but these experiments are usually done under oxidizing conditions with respect to Ni reducing the likelihood of metallic Ni expulsion.^76-78  The internal oxides lead to a variety of stress relief mechanisms operating, such as Ni metal expulsion and grain boundary migration. In representative primary water conditions, the kinetics can only support a maximum oxygen penetration of 100 nm after 5,000 h;^63,74-75  intragranular oxidation is undeniably more subtle than in high-temperature hydrogenated steam. However, internal oxidation and expulsion of metallic Ni should not be ruled out in 300°C to 360°C primary water; it is difficult to predict whether these processes can occur over a longer period of exposure, such as after decades of exposure in nuclear power plants.

Criticisms had been raised by Staehle, et al., about the possibility of internal oxidation in Alloy 600 in primary water.⁷⁹  They argued that the depth of intergranular oxygen diffusion necessary for internal oxidation at temperatures in the range of primary water does not support experimental data of oxygen diffusion depth at temperatures ranging from 800°C to 1,300°C.⁷⁹  Staehle, et al., extrapolated oxygen diffusion experimental data in Ni alloys from high temperature to temperatures in the range of primary water.⁷⁹  They determined that high-temperature experimental evidence predicted oxygen depth penetrations up to four magnitudes lower than those predicted by Scott.²⁹  However, Staehle, et al., agree that Scott’s depth of intergranular oxygen diffusion does correlate well with the range of penetration actually observed in PWSCC-related experiments, such as those by Gendron, et al., in 400°C hydrogenated steam.^41-43  Table 1 shows bulk diffusion data for O in Ni, Ni-Cr, and Alloy 600. The data are compared with the Scott prediction for the necessary diffusion coefficient (D) required to promote embrittlement and SCC. It is clear that the bulk diffusion data do not support internal oxidation, in agreement with Staehle, et al.

Table 2 shows low-temperature intergranular D values for O in Ni, Ni-Cr alloys, and Alloy 600. Intergranular diffusion data are within range of the Scott assumption, in Table 1, required for internal oxidation at temperatures of 500°C and below. Therefore, intergranular oxygen penetration is possible and, arguably, expected. Scott further countered the criticisms of Staehle, et al., by comparing the extrapolation of high-temperature diffusion data to actual experimental evidence of the crack growth rate at intermediate temperatures, 400°C to 750°C.⁴⁸  Assuming that SCC is a result of internal oxidation, it is impossible to account for the crack growth rate at intermediate temperatures by simply extrapolating from high-temperature data.⁴⁸  Extrapolation from higher-temperature data may be inaccurate and not representative of the intergranular oxygen diffusion rate at lower temperatures.

In summary, experimental evidence suggests that the diffusion rate of oxygen, at least intergranularly, is significantly greater than proposed by Staehle, et al. Data in Table 1 suggest that the extrapolation by Staehle, et al., may be valid for intragranular diffusion, which is hindered in the temperature range of primary water, but intergranular oxygen diffusion rates, in Table 2, are several magnitudes higher, of the order of 10⁻¹⁴ cm²/s; however, even with relatively slow diffusion rate, intragranular internal oxidation has been proven to be possible in Ni-Fe-Cr alloys in the temperature range of primary water by Olszta, et al.,^74-75  and Persaud, et al.,⁶³  albeit to shallow depth. Also, the influence of stress and/or surface cold work on accelerating diffusion kinetics has been reported frequently and is discussed in the Alloy 600 section and Ni Alloys with High Reactive Solute Element Contents section. Considering short-circuit grain boundary diffusion and dislocations generated through cold work and by volume expansion, internal oxidation should be expected in Ni-Fe-Cr alloys extending from 300°C primary water to 500°C hydrogenated steam.

A review of the significant literature base accumulated over the past few decades for oxidation, embrittlement, and SCC in Ni-Fe-Cr alloys exposed to 400°C to 500°C hydrogenated steam has been conducted, with comparisons to oxidation in 300°C to 360°C representative primary water. The following general conclusions can be drawn:

• Several studies have indicated that the low Cr content of Alloy 600 results in selective intergranular Cr oxidation and embrittlement through a process analogous to internal oxidation at high temperatures.

• Classical internal oxidation is possible in Alloy 600 intragranularly resulting in solvent metal, Ni, expulsion. Oxide/metal interfaces and dislocation pipe diffusion are short-circuit diffusion pathways used to overcome slow lattice diffusion kinetics.

• Ni-Fe-Cr alloys containing high reactive element contents, such as Alloys 690 and 800, are resistant to intergranular embrittlement as a result of the increased outward intergranular Cr flux, which aids in the formation of an external and protective Cr-rich oxide.

• Remarkably, Alloys 690 and 800 are not immune to internal oxidation, which can occur intragranularly resulting in expulsion of metallic Ni. Alloy 800 is particularly exceptional because of its low noble element, Ni, concentration of approximately 32 at% and the presence of several stress relief mechanisms operating simultaneously.

• Intergranular oxidation processes reported in Ni alloys exposed to 300°C to 360°C representative primary water are similar to those observed in 400°C to 500°C hydrogenated steam. The mechanism of SCC/embrittlement in Alloy 600 is likely to be similar in both environments and related to internal oxidation.

• The hindered diffusion kinetics in 300°C to 360°C representative primary water result in intragranular oxidation being suppressed relative to 400°C to 500°C hydrogenated steam. However, there are some studies that indicate shallow internal oxidation is possible. Also, internal oxidation should not be ruled out over longer duration, such as after decades of exposure in nuclear power plants.

The authors would like to acknowledge the University Network of Excellence in Nuclear Engineering (UNENE) and the Natural Sciences and Engineering Research Council (NSERC) of Canada for funding their research.