Effect of zinc injection on mitigating stress corrosion cracking initiation of structural materials in light water reactor primary water

2.1 Materials and specimen

Two LWR structural materials were chosen for this study: an Alloy 182 weld metal, which is known to be rather susceptible to SCC and an AISI 316L stainless steel, which has been 20% CW to decrease its SCC resistance. The chemical compositions are shown in Table 1. The specimen type used for the SCC initiation tests is the flat tapered tensile (FTT) specimen, as specified in more detail in (Bosch et al. 2020). The specimen dimensions are shown in Figure 1a. The FTT specimens were cut out of the weld center along the T direction, whereas L is the welding direction, as illustrated in Figure 1b. A block of AISI 316L stainless steel was CW by one-time forging at ambient temperature and then FTT specimens were machined along the forging direction as shown in Figure 1c. Table 2 summarizes the mechanical properties of the two materials at ambient temperature and 274 °C, respectively. The surfaces of the FTT specimens were ground with P180 (FEPA) SiC paper.

Figure 1:

Schematic and sampling locations of the FTT specimens: (a) dimensions of the FTT specimen (in mm), (b) Alloy 182 weld produced by shielded metal arc welding on/inside a low-alloy steel plate and (c) 20% CW 316L stainless steel with arrows indicating forging direction.

2.2 Experimental procedure

The SCC initiation tests were carried out by means of the constant extension rate tensile (CERT) test method using FTT specimens in both, simulated BWR hydrogen water chemistry (HWC) and PWR primary water environment (Table 3) in high-temperature water loops. The selected dissolved hydrogen content refers to the Ni/NiO phase transition boundary at the given test temperatures, which provides a maximum in SCC susceptibility for nickel-base alloys (Andresen et al. 2008; Bai et al. 2018). Even though this is not relevant for stainless steels, the hydrogen contents were kept the same for the CERT tests with the CW 316L. Nominal strain rates of 5.0× 10⁻⁷ and 1.3× 10⁻⁷ s⁻¹ were applied for the Alloy 182 and 20% CW 316L stainless steel, respectively. The tensile tests were terminated after the load dropped by approx. 150 N from its maximum. It is worth noting that the CERT test is an accelerated method, but it allows for relatively fast systematic screening of the SCC initiation behavior under environment with or without Zn injection. Large efforts have been invested in the further development and careful assessment of this test method in the framework of several collaborative research projects (e.g., MICRIN+ (Bosch et al. 2020) and MEACTOS (Que et al. 2022a, 2022b, Zimina et al. 2023)).

Each CERT test consisted of a one-week pre-oxidation phase and the straining phase, which also took approx. one week. In both, BWR and PWR environments, the CERT tests were carried out under four different Zn dosing conditions, as shown in Table 4. The first condition is considered as a reference without Zn injection. The second case refers to 40 ppb ZWC, to see the key difference and to study the Zn effect. ZWC was obtained by continuous injection of diluted Zn acetate solution into the loop water. The third and fourth conditions were designated to study the ZWC effectiveness on SCC after the materials being subjected to pre-exposure oxidation. For this purpose, the FTT specimens were exposed for approx. two weeks to high-temperature water (BWR HWC or PWR conditions) containing no or 40 ppb Zn, respectively, prior to the two-weeks lasting CERT tests with 40 ppb Zn injection.

State-of-the-art high-temperature water loops with autoclaves were used for the SCC initiation tests (Bai et al. 2019). The recirculating water loop with well-controlled water chemistry was connected to an autoclave with electromechanical tensile machine. Concentrated Zn acetate solution was continuously injected into the main flow at a very low flow rate (0.2 mL/min) just before the high-pressure pump. This way the Zn dose was diluted down to 40 ppb before the autoclave inlet. After passing through the autoclave and cooling down, the Zn was completely removed from the water through ion exchangers, so that 40 ppb Zn concentration was maintained at a constant level inside the autoclave. Water samples were taken periodically at the inlet, outlet and from the storage tank, followed by ICP-OES analysis to verify the Zn concentration. The measured concentration values were as expected, confirming the credibility of this well-controlled Zn injection technique. Since initially most of the injected Zn would be consumed by the inner surfaces of the loop and autoclave, 100 ppb Zn was continuously injected into the loop for several weeks to saturate these surfaces prior to the CERT tests with Zn injection. This enabled to sustain stable test conditions, meaning the same amount of Zn at the inlet, outlet and inside the autoclave.

2.3 Post-test analysis

The CERT tests were terminated before final fracture of the specimen. This approach simplified the characterization of the crack distribution on the whole gauge section’s surface. After the tests, both gauge surfaces of the FTT specimens were carefully characterized with a field emission gun scanning electron microscope (FEG-SEM, Zeiss ULTRA 55) equipped with a windowless energy dispersive X-ray spectroscopy (EDS, Oxford Ultim Extreme SDD) detector optimized for low acceleration voltages. Generally speaking, the crack density always decreased from the narrowest towards the wider gauge section until no cracks could be identified anymore. Towards the wider gauge section, the last intergranular crack (≥10 µm surface crack length) was determined for both specimen sides for the calculation of an apparent SCC initiation stress threshold, under which no SCC is expected to occur. The threshold was estimated by dividing the maximum load by the critical cross-sectional area, where the last crack was found. Details for the determination of SCC initiation stress thresholds from FTT specimens can be found in the literature (Bai et al. 2019; Bosch et al. 2020; Que et al. 2022a). For each test condition also the location and length of all cracks was determined on one specimen’s gauge section and the crack length density as function of applied stress was calculated.

The FTT specimens were then cut lengthwise for cross-sectional examination of the surface oxide film and cracks. The thickness of the inner oxide film was measured at two different areas along the gauge section: low and medium stress area. The thickness value for each region was obtained by averaging the inner oxide layer thickness from 45 equidistant measurements. At the high stress area (narrowest area of the gauge section) no reliable determination of the film thickness could be done due to the heavily plastically deformed (and therefore very uneven) and severely cracked surface. For the 316L stainless steel only the FTT specimens from the 0 and 40 ppb Zn CERT tests under BWR HWC conditions were cross-sectioned (so far).

3.1 SCC initiation stress thresholds

The SCC stress threshold values for crack initiation of Alloy 182 are summarized in Figure 2 for both, PWR and BWR HWC environments under the four different Zn conditions. The stress threshold for each test condition was calculated as the average value from two repetitive CERT tests i.e. from four data points (two surfaces per FTT specimen). In general, all the threshold values are slightly below or close to the yield strength (408 MPa), suggesting a relatively high SCC susceptibility of the weld metal. The high SCC susceptibility of the weld metal has also been reported from previous studies with this weld and in the literature (Bai et al. 2017, 2018; Que et al. 2021, 2022a). Under Zn-free conditions, the stress thresholds were approx. 375 MPa and they increased by 10–15% with 40 ppb Zn injection. Higher stress threshold means lower susceptibility to SCC initiation, thus this enhancement in stress thresholds is a clear evidence of the mitigation effect of ZWC. When the FTT specimens were pre-exposed without Zn for about two weeks, the stress thresholds were close to those under Zn-free condition, suggesting no mitigation effect of ZWC for the material with the mature, Zn-free oxide film. In case of the tests with a pre-exposure of about two weeks in 40 ppb Zn-containing water, the stress thresholds also increased compared to the reference tests without Zn, although to a lower extent. The trend of stress thresholds is similar for PWR and BWR HWC environments in regard to the Zn effect under the four different Zn conditions, but the stress thresholds are generally slightly lower in PWR compared to BWR HWC environment, possibly related to the higher temperature under PWR condition.

Figure 2:

Summary of the stress thresholds for SCC initiation of Alloy 182 in BWR HWC and PWR environments under the four different Zn conditions.

Figure 3 summarizes the stress thresholds of the 20% CW 316L stainless steel under all four Zn dosing conditions. The stress thresholds range from 585 to 615 MPa, which are close to the ultimate tensile strength, revealing a high SCC resistance of the stainless steel. With 40 ppb Zn injection, the stress thresholds increased, but only slightly. No clear changes in stress thresholds were observed when the specimens were pre-exposed to Zn-free high-temperature water. After pre-exposure and CERT testing in ZWC, a very small increase in threshold could only be observed in case of BWR environment. Thus, it is challenging to conclude if ZWC mitigates SCC initiation in 20% CW 316L stainless steel. Rather surprisingly, the stress thresholds are slightly higher in PWR compared to BWR HWC environment for all four Zn conditions, although the temperature in PWR environment is higher. It seems that the higher temperature does not necessarily result in a higher SCC susceptibility for the 20% CW 316L stainless steel. The PWR environment is buffered with H₃BO₃ + LiOH resulting in a pH_T of 7.46, which is higher than that of the BWR environment (pH_T = 5.63). The SCC susceptibility of metals have been widely reported to decrease in slightly alkaline environment, which may explain the reverse temperature dependence of SCC for the 20% CW 316L stainless steel.

Figure 3:

Summary of the stress thresholds for SCC initiation of 20% CW 316L stainless steel in both, BWR HWC and PWR environments under the four different Zn conditions.

3.2 Crack distributions

To obtain further evidence and get a better understanding of the Zn mitigation effect on SCC initiation, the cracking and oxide layer morphology at the gauge surfaces of the FTT specimens were carefully examined after the CERT tests. Figure 4 shows an example of a gauge surface from an Alloy 182 FTT specimen after the CERT test in PWR environment without Zn injection. Extensive intergranular cracking was identified near the smallest cross-section, where the stress and strain rates were highest. Towards the wider gauge cross-section, the crack density decreases until the last intergranular SCC crack was found for the determination of the SCC initiation stress threshold.

Figure 4:

SEM images of one gauge surface indicating the crack distribution, as well as the last crack after the CERT test (Alloy 182 in Zn-free PWR environment).

Figure 5 shows the crack distribution along the gauge length of Alloy 182 after the CERT tests under the various Zn conditions. The crack length density was calculated as a function of true stress, where the higher stress corresponds to the area with a narrower gauge width. In both, PWR and BWR HWC environments, the crack length density remains high above 550 MPa and does not change much with further increasing stress level. With decreasing stress from 550 MPa, the crack length density tends to drop gradually with some scatter and reaches the minimum value at the stress threshold for SCC initiation. Overall, no clear difference in crack length density can be identified between the four different Zn conditions.

Figure 5:

Distribution of the crack length density on the gauge surfaces of Alloy 182 after the CERT tests in (a) PWR and (b) BWR HWC environments under the four different Zn conditions.

Figure 6 shows the crack distribution along the gauge length of the 20% CW 316L stainless steel after the CERT tests under the various Zn conditions. Similar tendencies of the crack length density distribution as a function of the applied true stress are observed, but with smaller scatter than those of Alloy 182 in Figure 5. Compared to the CW 316L stainless steel, the Alloy 182 weld metal is more heterogeneous, which is likely responsible for the larger scatter in the crack distribution. With 40 ppb Zn injection, the crack length density is almost one order of magnitude lower than that under Zn-free condition at stresses below 650 MPa, suggesting a clear mitigation effect by Zn injection for the stainless steel. For the specimens pre-exposed to the high-temperature water for two weeks without Zn, 40 ppb Zn does not decrease the crack length density, confirming the stress threshold results. The crack length density results for the specimens pre-exposed to Zn-containing high-temperature water revealed only a minor mitigation effect in the BWR HWC environment and no mitigation in the PWR primary water, respectively. The reason for this is not clear (yet). At stress levels of >650 MPa, no mitigation effect was observed with or without pre-exposure. This is not surprising, because the high-stress region suffered from the highest strain rate, resulting in a dominantly mechanical-driven fracture where the impact of Zn is very limited.

Figure 6:

Distribution of the crack length density on the gauge surfaces of 20% CW 316L stainless steel after the CERT tests in (a) PWR and (b) BWR HWC environments under the four different Zn conditions.

After all, a mitigation effect by 40 ppb Zn injection was confirmed by the reduced crack length density at the gauge surfaces for the 20% CW 316L stainless steel under both, PWR and BWR HWC conditions, whereas the role of pre-exposure needs further testing and clarification.

3.3 Oxide layer morphology

Figure 7 compares the cross-sectional morphology of the gauge surface oxide film at different stress regions of Alloy 182 after the CERT tests under the four different Zn conditions in BWR HWC environment. Firstly, fracture of the oxide film was widely observed in the high stress region, regardless of the Zn condition. In most cases, the oxide film fractures reach up to 1 µm depth and do not penetrate into the bulk material. Only a few oxide fractures developed into SCC cracks, propagating into the bulk material, mostly along susceptible grain boundaries (intergranular SCC cracks). The average oxide film thickness did not vary relevantly between the medium and low stress region. In the high stress area the oxide film thickness could not be determined reliably due to the much higher plastic deformation (and resulting rougher surface) and higher crack density.

Figure 7:

The cross-sectional morphology of the Alloy 182 FTT specimens in different stress regions after CERT tests in simulated BWR HWC environment (a) without Zn injection, (b) with 40 ppb Zn injection, (c) pre-exposed without Zn followed by 40 ppb Zn injection, and (d) pre-exposed with 40 ppb Zn followed by 40 ppb Zn injection.

From Figure 7 it is already obvious that a thinner oxide film has formed during the 40 ppb Zn injection, regardless of the stress level. Figure 8 summarizes the average oxide film thickness values for the medium and low stress regions after CERT testing under different Zn conditions in simulated PWR and BWR HWC environments. Apparently, with 40 ppb Zn injection, the oxide film of both materials was clearly thinner compared to that exposed to the Zn-free high-temperature water in both, PWR and BWR HWC environments. EDS point analysis confirmed the existence of Zn in the surface oxide film, especially in the inner oxide layer, as shown in Figure 9. It seems that 40 ppb Zn injection affects the surface oxide film by creating a thinner, but more protective oxide scale, which possibly results in the increased SCC initiation resistance.

Figure 8:

Summary of the inner oxide film thickness in different stress regions of Alloy 182 after the CERT tests in (a) PWR and (b) BWR HWC environments under four different Zn conditions, and (c) 20% CW 316L SS in BWR environment under two different Zn conditions.

Figure 9:

EDS point analysis of the surface oxide film in the low stress region for Alloy 182 after the CERT tests in BWR HWC environment (a–c) without Zn injection and (d–f) with 40 ppb Zn injection.

For the specimens pre-exposed without Zn, the oxide thickness remains rather unchanged, which suggests that the exposure period of only two weeks to the ZWC during the CERT test was not sufficient to modify the existing Zn-free oxide film. This may explain the almost identical stress thresholds between the Zn-free and 0 + 40 ppb Zn conditions, as shown in Figure 2. In PWR environment (Figure 8a), the specimens which were pre-exposed to 40 ppb Zn-containing high-temperature water show low oxide film thickness values, but they are higher than expected in BWR HWC environment (Figure 8b). Although the testing temperature is higher in PWR than in BWR HWC environment, the oxide thicknesses are comparable under the other Zn conditions.

4.1 Effect of Zn injection

Above mentioned results give clear indications that 40 ppb Zn is sufficient to mitigate SCC initiation of Alloy 182 weld metal and to a smaller extend also of 20% CW 316L stainless steel in both, simulated BWR HWC and PWR environments. This is evidenced by the increased SCC initiation stress thresholds or the reduced crack length densities revealed by the accelerated SCC testing. This mitigation effect is believed to be (at least mostly) related to the oxide film morphology which is altered by the Zn. A thinner and more compact oxide film was identified after the tests with 40 ppb Zn. For most Fe–Cr–Ni binary alloys exposed to LWR primary water, a duplex-layer oxide scale is usually formed on the surface. The outer oxide layer is enriched in Fe, while the inner oxide layer is usually composed of Cr-rich spinel. The EDS results in Figure 9 confirmed the existence of Zn in the inner oxide layer for Alloy 182 with 40 ppb Zn injection. This means that Zn-rich spinel would preferentially form since it is more stable from a thermodynamic point of view (Liu et al. 2011; Miyajima and Hirano 2001), thus creating a denser and more compact surface oxide on the metal surface.

Although the Zn-treated oxide film is thinner, its protectiveness is higher than the Zn-free oxide film in at least two aspects. First, the thin Zn-treated oxide film can effectively inhibit the inward/outward diffusion of ions (Betova et al. 2011). Simulation results in literature (Bojinov et al. 2005; Betova et al. 2008, 2011) show that the diffusion rates of Fe, Cr and Ni in a Zn-containing (30 ppb) inner layer are reduced by a factor of 3–4 compared to a Zn-free oxide film. The formation of the outer oxide layer is also highly compressed with 30 ppb Zn injection (Betova et al. 2011). One of the reasons is that Zn-containing oxide films have a lower defect density, thus decreasing the diffusion rates of Fe, Cr and Ni et al. cations. Another reason is that the solubility of Zn oxide (e.g., ZnCr₂O₄) is about one third of FeCr₂O₄ at 6.9 < pH_573K < 7.2 (Miyajima and Hirano 2001), which further increases the compactness of the Zn-treated oxide film. Second, the Zn-treated oxide film exhibits enhanced mechanical properties (Angeliu and Andresen 1996). It was found that the oxide rupture strain of type 304L stainless steel increased by up to a factor of ∼2 with 20–60 ppb Zn injection at 288 °C. For the thicker and more defective Zn-free oxide scale, the oxide ductility is lower and thus is more likely to break and expose the substrate to environmental attack (Bubar and Vermilyea 1967). The Zn-treated oxide film, on the other hand, is thinner and more ductile, which makes it more difficult to rupture under external stress.

Currently, the widely accepted SCC mechanisms for metals exposed to LWR primary water are the Ford–Andresen model (Andresen and Ford 1988) and internal oxidation model (Rapp 1965; Stott and Wood 1988). According to either model, SCC occurs by the development of oxide formation and film rupture. Consequently, it is reasonable to speculate that the Zn injection mitigates SCC initiation due to the formation of a more compact and strengthened oxide film with Zn incorporation.

4.2 Effect of pre-exposure to Zn-free or Zn-containing high-temperature water

The results also suggest that pre-exposure may have an impact on the mitigation effect of the Zn injection. For both materials, the specimens pre-exposed to Zn-free water for approx. two weeks, showed no signs of SCC mitigation if 40 ppb of Zn was present for one week before as well as during the tensile test. Similar results have also been reported by other researchers (Beverskog 2004). However, recent work of Hur (Lim et al. 2021) shows that Zn could still incorporate into pre-oxidized Alloy 690 oxide films and remain there even after cessation of Zn injection. On the downside, the Zn incorporation into a mature oxide film showed little effect on corrosion reduction for both, Zn-treated and Zn-free oxidized surfaces. It seemed way more efficient in terms of SCC mitigation if Zn is incorporated into a newly formed oxide layer. A mitigation effect of Zn injection on the pre-formed oxide film was also reported for stainless steels in PWR and BWR primary water, but to a significantly reduced extent (Betova et al. 2011; Liser 1988). The somewhat contradictory results of the Zn effect on an existing Zn-free oxide scale might be related to the different experimental parameters and evaluation methods. Since Zn²⁺ has a strong tetrahedral site preference (Navrotsky 1969, Navrotsky and Kleppa 1967), it can substitute all other divalent cations from these spots to form a more stable spinel structure. This modification may not change the thickness of the existing oxide scale, but might improve the protectiveness of the oxide film in long-run exposition to Zn-containing water. It is worth noting, that the substitution by Zn takes some time depending on the experimental parameters. From this point of view, Zn injection should also result in mitigation effects on the corrosion/SCC behavior for Zn-free pre-exposed materials if given enough exposure time. In the current work, only ≤ two weeks Zn injection was applied to the Zn-free pre-exposed specimens, which seems to be too short to modify the existing oxide film. To draw further conclusions on this aspect more systematic investigations with longer exposure time are needed. Especially for LWR plants, which are considering applying ZWC in the middle of their reactor life, this topic is crucial.

The specimens which were pre-exposed and CERT-tested in Zn-containing high-temperature water showed in most cases a somewhat smaller SCC mitigation effect. This is beyond expectation and to date lacks reasonable explanation. More experimental and characterization work is needed to verify those results.

Even though the used CERT testing technique with FTT specimens has proven to be a very useful method to systematically screen the SCC initiation behavior in structural materials in a reasonable time-frame, it is also worth noting, that the CERT test used in this study is an accelerated method, which imposes massive plastic straining and different strain rates along the specimen’s gauge section. Hence, the results only provide limited quantitative information on the benefit of Zn injection on the mitigation of SCC initiation. This method may also overlook part of the time-dependent corrosion effects and hide some of the Zn mitigation capabilities. Long-term SCC initiation tests with more plant-relevant loading methods are needed to quantitatively evaluate the mitigation effect by ZWC. Finally, it also must be stated that the low SCC susceptibility of the 316L stainless steel may also hide some of the mitigation effects of the Zn.