Materials

In this study, all the exposure experiments were conducted in nuclear grade graphite crucibles with an inner diameter of 20 mm.Each exposure experiment involved one tensile sample and one corrosion samples. The samples were EDM-cut from a hot-rolled GH3535 alloy plate, whose composition is shown in Table 1. The schematic diagrams of the samples are shown in Fig. 1. The corroded sample has a rectangular shape with a hole on one end for fixing, and the dog-bone shape was adapted for the tensile sample. The samples were first abraded sequentially with 400#, 800#, and 1500# SiC abrasive papers. Subsequently, they were polished with a nano alumina suspension containing particles of 0.05 μm diameter to achieve a smooth surface finish. After polishing, the samples were electro-polished in a solution consisting of 10% deionized water, 40% acetic acid (C₃H₈O₃), and 50% sulfuric acid (H₂SO₄) at 0 ℃ with an applied voltage of 36 V for approximately 10 s to eliminate the surface stress. Then, they were rinsed with alcohol and deionized water. Finally, the samples were dried in an oven at 150 ℃ for 12 h to remove the adsorbed impurities.

Fig. 1Full-screen View

Schematics of the corrosion and tensile samples (all values shown in the figure are in mm)

The FLiNaK salt (46.5LiF-11.5NaF-42KF, mol%) was prepared by mixing a totally 500g of LiF (99.9% purity), NaF (99.99% purity), and KF (99.9% purity) powders in a graphite crucible. Subsequently, the crucible containing the salts was baked at 350 ℃ for 5 hours and then melted at 650 ℃ for 2 hours in a muffle furnace inside an Ar-filled glovebox. The salt ingots were then polished and chunked upon cooling. EuF₃ powder (99.9% purity) was added to the produced FLiNaK salt as the redox agent, and Cr₃Te₄ powder (99.9% purity) was added as the Te source.

Corrosion tests

The corrosion technique used in this study was a static salt immersion test. Corrosion tests were carried out in four groups, each group contained 35 g of FLiNaK salt with different concentrations of Cr₃Te₄ and EuF₃ additions. Group #1: 0.1wt.% Cr₃Te₄ (35 mg); group #2: 0.1wt.% Cr₃Te₄ + 1wt.% EuF₃ (0.35 g); group #3: 0.1wt.% Cr₃Te₄ + 3wt.% EuF₃ (1.05 g); group #4: 3wt.% EuF₃ (1.05 g). The aim was to elucidate the influence of EuF₃ addition on Te-induced corrosion while isolating the direct corrosion effect by itself. All the operations were performed in an Ar glove box (the moisture and oxygen contents remained below 5ppm).

Firstly, the weighed FLiNaK salt chunks, Cr₃Te₄ powder and EuF₃ powder were placed in four crucibles in the following sequence: EuF₃ was added to the bottom of the crucible; next, 2/3 of the crucible was filled with the FLiNaK salt and Cr₃Te₄ powder, and the remaining 1/3 was filled with the FLiNaK salt. Second, the four loaded crucibles were placed into a muffle furnace and heated at 700 ℃ for 8 h to melt the salt and dissolve the powders. The salt samples before corrosion were removed from each crucible by dipping a tungsten rod. Finally, the alloys samples fixed to the crucible lids with by the GH3535 wire were placed into the molten salt and subjected to thermal exposure at 700 ℃ for 250 h. After exposure, the alloy samples were removed to cool in the glove box, and the salt samples after corrosion were dipped out using a tungsten rod, before turning off the furnace to cool the entire system.

Post corrosion tests and characterizations

The characterizations primarily encompassed the analysis of the salt, the microstructure of the alloy samples, the distribution of elements, and the mechanical properties, with a particular focus on the morphology of the cracks. The salt samples collected before and after the corrosion process underwent chemical analysis using inductively coupled plasma optical emission spectroscopy (ICP-OES) and mass spectrometry (ICP-MS). ICP-MS was utilized for the analysis of Te and Eu, while ICP-OES was employed for the analysis of all other elements. A comparison of the values obtained from these analyses before and after corrosion revealed changes in the elemental contents.

The tensile samples were subjected to room temperature (RT) testing using a mechanical testing machine (Zwick Z100) at a constant strain rate of 3.5 mm/min. Owing to the small sample size, an extensometer was not used. The data from the tensile tests were plotted as stress-strain curves to assess the impact of the different corrosion conditions. Microstructural and elemental analyses of both the corroded and tensile samples were conducted using a scanning electron microscope (SEM, Carl Zeiss Merlin Compact) equipped with an energy-dispersive X-ray spectrometer (EDS). Initially, EDS point scanning was performed at various points across the sample surface. This was followed by EDS line scanning of both the surface and cross-section of the samples, with a focus on the GBs and near-surface region. Given the difficulty in detecting Te segregation at the GBs using EDS, an electron probe microanalyzer (EPMA, SHIMADZU-1720H) was employed to ascertain the distribution of Te.

Salt composition analyses before and after corrosion

Before corrosion, the metal impurities in the salt were minimal, with Fe at a maximum of 35 ppm. After corrosion, the contents of Cr and Mn in all the salt samples increased, indicating the dissolution of the active elements in the alloy. Among them, the salt containing both Cr₃Te₄ and 3% EuF₃ increased the most, followed by the salt containing only EuF₃, and the least increase was observed in the salt containing only Cr₃Te₄, suggesting that the addition of Cr₃Te₄ and EuF₃ affected corrosion. The overall trend of Fe and Ni in the post-corrosion salt showed a decrease, but with some fluctuations. Additionally, the content of Mo increased slightly in #2, #3, and #4, with the most significant increase observed in #3.

The differences between the measured and calculated Eu values were statistically significant. Before corrosion, the measured values did not exceed 20% of the calculated values, whereas after corrosion, the measured values for #2, #3, and #4 were 51%, 88%, and 16% of the calculated values, respectively. This result can be ascribed to the limited solubility of EuF₃ in the molten FLiNaK salt or in the acid solution during ICP sample preparation. Notably, the measured Eu content after corrosion was higher than before corrosion, with Te-containing sample #3 showing a higher Eu content than the Te-free sample #4, although both samples had the same 3% EuF₃ addition.

Before corrosion, the measured Te values are close to the calculated values. The measured Te values for the 0%, 1% and 3% EuF₃ salts were 90% 99.5% and 99.7% of the calculated values, respectively. After corrosion, the measured Te values were 109%, 104%, and 88% of the calculated values, respectively. Overall, the change in Te before and after corrosion was not significant, and it is uncertain whether the differences in the measured values were due to test errors or the real changes in Te form and distribution.

Microstructure and composition characterization of corroded samples

The surface morphologies of the corroded samples under various experimental conditions are depicted in Fig 3. It is evident that all the samples exhibited GB corrosion; the overall difference between the samples was not obvious, and no tellurides were observed on the sample surface.

Fig. 2Full-screen View

The variation of elemental content in FLiNaK before and after corrosion, where (a) represents the change in Mn content and (b) represents the change in Cr content

Fig. 3Full-screen View

SEM images the samples’ surfaces after corrosion in molten FLiNaK with different additions: #1. Cr₃Te₄, #2. Cr₃Te₄ + 1%EuF₃, #3. Cr₃Te₄ + 3%EuF₃, #4. 3%EuF₃.The point is referring to the zone where the EDS point analysis was conducted

EDS point analysis was performed at the GBs and inside the grains. As shown in Table 3, compared to the original content (7.1wt.%), all the surface experienced a depletion in the element Cr. Cr depletion at the GBs was more severe than that inside the grains. Adding more EuF₃ to the salt caused an increase in Cr depletion, both inside the grains and at the GBs. When both EuF₃ and Cr₃Te₄ were present in the salt, the loss of Cr at the GBs increased faster than that inside the grains with the increasing EuF₃ addition. However, without Cr₃Te₄, the Cr depletions at the GBs and inside the grains was similar (sample #4). This indicates that the synergistic effect of EuF₃ and Cr₃Te₄ enhances the GB corrosion.

Cross-sectional analyses of the samples were performed to ascertain the depth of corrosion, and the SEM cross-sectional images of the samples under varying corrosion conditions are shown in Fig. 4(a). Overall, the grains underneath the surfaces of all tested samples remained free of corrosion defects. As is shown in Fig. 4(b), no obvious intragranular Cr depletion was found except in samples #4 that corroded in solely EuF₃ containing salt, which showed a Cr depletion of 4 μm. On the contrary, grain boundary corrosion was obvious in samples that corroded in salts containing both Cr₃Te₄ and EuF₃. After corrosion in the salt containing Cr₃Te₄ and 1% EuF₃, a small quantity of pores could be seen on grain boundaries in sample #2 up to more than 10 μm depth. But when the EuF₃ increased to 3%, sample #3 corroded in Te-containing salt showed extensive corrosion pores on the grain boundaries up to more 100 μm depth. It’s worth noting that not all grain boundaries in sample #3 suffered severe corrosion, typical corroded and un-corroded boundaries are marked with 1 and 2 in Fig. 4(a) respectively. EDS scans across grain boundaries of underneath the surface of each sample confirmed that Cr was slightly depleted in sample #2 and significantly depleted in sample #3, due to the co-existence of EuF₃ and Cr₃Te₄ in the salt; this observation was consistent with the surface analysis results.

Fig. 4Full-screen View

(Color online) SEM images the cross-sections of samples after corrosion in molten FLiNaK with different additions, along with the corresponding EDS line scans of Cr. (a) SEM images the samples’ cross-sections after corrosion in molten FLiNaK with different additions: #1.Cr₃Te₄, #2. Cr₃Te₄ + 1%EuF₃, #3. Cr₃Te₄ + 3%EuF₃, #4. 3%EuF₃.The line is referring to the zone where the EDS line scan was conducted. (b)Grain boundary and near-surface EDS line scans of Cr underneath the surfaces of samples in (a)

Because EDS does not have sufficient detection ability for trace elements, EPMA was used to further analyze the elemental distribution of the samples after corrosion in salts containing Cr₃Te₄+1%EuF₃ and Cr₃Te₄+3%EuF₃, as shown in Figs. 5(a) and (b). This again confirms that when the addition of EuF₃ in the molten salt was increased to 3%, the depth and extent of Cr depletion along the GBs of the sample increased significantly. However, none of the samples exhibited Te enrichment in the corrosion-affected regions, indicating that the Te activity in the salts was low.

Fig. 5Full-screen View

(Color online) EPMA analysis of the cross-sections of samples corroded in salt with Cr₃Te₄ and varying amounts of EuF₃. (a) EPMA analysis of the cross-section of the sample corroded in salt with Cr₃Te₄+1%EuF₃. (b)EPMA analysis of the cross-section of the sample corroded in salt with Cr₃Te₄+3%EuF₃

Tensile tests and microstructure characterization of fractured samples

After corrosion, the tensile samples under the four experimental conditions were tested at RT using a universal material-testing machine. The stress-strain curves of the samples are shown in Fig. 6, and the strength and elongation data are listed in Table 4. The sample size was small, and thus, an extensometer could not be used. Therefore, the curve of the elastic stage in Fig. 6 is not accurate; the curve after the yielding stage provides detailed information.

Fig. 6Full-screen View

(Color online) Stress-strain curves for both uncorroded and corroded alloy tensile samples under different conditions

As shown in Fig. 6 and Table 4, the yield strengths of both the corroded and uncorroded samples were similar, but the ultimate strengths and elongations varied significantly. The uncorroded samples had the greatest ultimate strength and elongation of 834.3 MPa and 63.4% respectively. The sample corroded in solely EuF₃-containing salt follow, with ultimate strength and elongation of 813 MPa and 55.7% respectively, while adding Cr₃Te₄ to the salt caused reductions in both properties. It worth noting that among the samples corroded in Cr₃Te₄-containing salts, the extent of reduction in mechanical properties increased with the increasing EuF₃ contents. The sample corroded in the salt containing Cr₃Te₄+3%EuF₃ had the lowest ultimate strength and elongation of 456.2 MPa and 17% respectively.

The surface morphologies of fractured samples at different locations from the holding end to the fractured end were observed by SEM. According to the sample geometry, the stresses borne during tensile test increased from locations I to IV. As shown in Fig. 7, all tested samples show no cracks at location I, but show clear surface cracks at location IV, indicating that the cracking was sensitive to stresses. Sample #4 corroded in the salt with solely EuF₃ addition exhibited the least severe cracking, only some short cracks on carbides were found even near the fracture and no intergranular crack was seen. On the contrary, intergranular cracks existed in all the samples corroded in Cr₃Te₄-containing salts, but the size, density and occurring location of these cracks varied. While only few intergranular cracks were seen at locations III and IV in Sample #1, more intergranular cracks existed even at location II in sample #2 and #3. It is clear that for samples corroded in Cr₃Te₄-containing salts, the severity of cracking was increased with the increasing EuF₃ additions. The order of the increasing intergranular cracking of the fractured samples was consistent with that of the decreasing tensile properties.

Fig. 7Full-screen View

SEM surface morphology images of the fractured samples corroded in molten FLiNaK with different additions: #1.Cr₃Te₄, #2. Cr₃Te₄ + 1%EuF₃, #3. Cr₃Te₄ + 3%EuF₃, #4. 3%EuF₃. I is the gripping section of the tensile specimen, II is the transition section of the tensile specimen, III is the parallel section of the tensile specimen, IV is the fracture location of the tensile specimen

The cross-sectional microstructures of the fractured samples were also observed using SEM. Samples #1 and #4 did not exhibit cracks; therefore, these images are not presented. Samples #2 and #3 exhibited increased cracking, as shown in Fig. 8(a) and (b), respectively.

Fig. 8Full-screen View

SEM Cross-sectional images of the parallel section and fracture location of the fractured samples corroded in salt with Cr₃Te₄ and varying amounts of EuF₃. (a) Cross-sectional SEM images of the parallel section and fracture location of the fractured samples corroded in FLiNaK + Cr₃Te₄ + 1%EuF₃ (a1ã3, a4ã6 approaching the fracture direction). (b)Cross-sectional SEM images of the parallel section and fracture location of the fractured samples corroded in FLiNaK + Cr₃Te₄ + 3%EuF₃ (b1~b6 approaching the fracture direction)

For the sample corroded in the salt containing Cr₃Te₄ and 1%EuF₃, 16 cracks were found within the 5mm observation range. Only one crack had a depth of 32 μm (Fig 8(a)), all other cracks were shorter than 10 μm. When the addition of EuF₃ was increased to 3%, both the density and depth of the cracks increased significantly. Further, 27 cracks were found in the 5mm observation range, and the average depth was 57.1μm. The standard deviation of cracks depth was large, with a maximum of 164μm and a minimum of 20 μm, and the depths of only nine cracks was larger than the average value. In addition to the depth, the widths of the cracks also varied. Two typical types of cracks were observed in sample #3: one had a widely opened crack mouth, such as crack 1 in Fig. 8(b), and the other had no such as crack 2 in Fig. 8(b). A possible mechanism for their formation is discussed below.

Direct corrosion induced by EuF₃

When EuF₃ was present in the molten FLiNaK salt, equilibrium between EuF₃/EuF₂ was established, and the balance concentration ratio without reacting with other elements was measured as approximately 2 at 700 ℃ [49]. This ratio may be highly oxidizing because an EuF₃/EuF₂ ratio lower than 0.05 is required to avoid the corrosion of Cr [49]. Therefore, EuF₃ is typically used as an oxidant for accelerating corrosion in many of molten fluorides corrosion tests [43, 50, 51].

The results presented in the previous section demonstrate that the contents of dissolved Cr in the salt (Table 2) and depleted Cr on the sample surface (Table 3) after corrosion both increased with the increasing EuF₃ content. This result can be ascribed to the oxidative nature of EuF₃; its reaction with Cr in the alloy is expressed by the following equation: $2{\text{EuF}}_3+\text{Cr}={\text{CrF}}_2+2{\text{EuF}}_2$ (1) However, as shown in Tables 2, 4 and Figs. 4(a), (b), 7, sample #4 corroded in the salt containing solely EuF₃ exhibited minor uniform Cr dissolution with no obvious grain boundary Cr depletion nor intergranular cracking, and the greatest strength and elongation. This indicates that the direct corrosion of EuF₃ is homogeneous and not particularly harmful to the mechanical properties of the alloy. Thus, the change in the severity of the intergranular corrosion and cracking behaviors with EuF₃ concentrations can be ascribed to the synergistic effect of EuF₃ and Cr₃Te₄.

Synergistic effect of EuF₃ and Cr₃Te₄ on corrosion and cracking

According to the early research by ORNL [22], it is the zero valence Te⁰ that diffuses into to the alloy, causes the embrittlement of the alloy GBs and results in the intergranular cracking. Therefore, they proposed to control Te⁰ activity by adjusting the UF₄/UF₃ concentration ratio, through the “complexing” reaction of Te with chromium to retain it in the salt. The reaction can be expressed as Eq. (2) which implies that increasing the concentration of UF₄ in the salt increases the content of Te⁰ released. Consequently, the redox potential of the salt has a strong influence on Te corrosion; if the UF₄/UF₃ concentration ratio in the fuel salt is kept sufficiently low, Te-induced cracking is prevented [22]. $6{\text{UF}}_4+{\text{Cr}}_3{\text{Te}}_4\rightleftharpoons 6{\text{UF}}_3+3{\text{CrF}}_2+4\text{Te}$ (2) ${\text{6EuF}}_{\text{3}}{\text{+Cr}}_{\text{3}}{\text{Te}}_{\text{4}}\rightleftharpoons {\text{6EuF}}_{\text{2}}{\text{+3CrF}}_{\text{2}}\text{+4Te}$ (3) In this investigation, the addition of EuF₃ similarly affected the Te-induced corrosion, and the reaction can be described by Eq. (3). Without EuF₃, only a small amount of Te⁰ was formed by the reaction of impurities in the salt with Cr₃Te₄, whereas with an increase in EuF₃ addition from 1% to 3%, EuF₃ reacted with Cr₃Te₄ to produce more Te⁰. The Te⁰ released from Cr₃Te₄ subsequently transferred to the alloy surface, diffused inward and aggregated at the GBs. Thus, sample #3, which was corroded in the salt containing Cr₃Te₄ + 3%EuF₃, exhibited the most Te GB segregation.

Te segregation on the GB is harmful because it not only directly embrittles the GB [20], but also causes GB expansion, which reduces the diffusion barrier of other active elements along the GB. The accelerated outward diffusion of the active element causes void formation that further damages the GB [41]. Therefore, EuF₃ in the salt induced the most Cr dissolution according to Eq. (1), whereas Cr depletion was much faster from the Te-segregated GBs in sample #3. This synergistic effect resulted in the most severe GB corrosion and the subsequent intergranular cracking in sample #3, hence greatly deteriorated its mechanical properties. To quantify the severity of Te-induced cracking, the density and length of cracks were counted from locations III and IV of cross-sectional samples, and the K values [20, 40] defined as the number of cracks per centimeter of alloy multiplied by the average depth of cracks (unit: pc × μm/cm) were calculated as follows: $\begin{array}{l}{K}_{\left(\#2\right)}=16\times (6.1\pm 7.1)\text{pc}\times \text{μm}/5\text{ mm}\approx 195\text{ pc}\times \text{μm}/\text{cm}\hfill \\ K_{\left(\#3\right)}=27\times (57.1\pm 33.5)\text{pc}\times \text{μm}/5\text{ mm}\approx 3083\text{ pc}\times \text{μm}/\text{cm}\hfill \end{array}$ (4) From Fig. 4(a), 5(b), 7, 8(b) show that all the GBs in sample #3 suffered from GB corrosion and cracking, and the morphologies of the corroded and cracked boundaries also varied. This may be ascribed to the different amounts of Te accumulated at each GB, as the Te diffusion behavior depends on the GB characteristics [23]. As shown in Fig. 9, GBs with more Te segregation may have more expansion, thus undergoing faster Cr depletion and more severe corrosion (e.g., GB 1 in Fig. 4(a)). The direct embrittlement effect of Te segregation and further damage caused by Cr depletion and void formation significantly deteriorated the GB strength. Therefore, it may crack under low stress at the early stage of tensile stress, and the opening of the crack mouth widens with the elongation of the sample during tensile stress (crack 1). In contrast, the strength of the GB with less Te segregation was less affected by direct Te embrittlement and corrosion damage (e.g., GB 2 in Fig. 4(a)). Thus, it did not crack until the late stage of tensile stress, where strain localization induced a larger stress, resulting in a narrow shape (crack 2).

Fig. 9Full-screen View

(Color online) Schematic showing the formation of different crack morphologies

Assessment of the present surrogate experiment

The major test conditions and results of the present experiment, as well as other typical Te vapor, Te-containing fuel salt and FLiNaK salt corrosion experiments for UNS N10003 alloy (alloy N) at 700 ℃, are summarized for comparison in Table 5.

When alloy N was corroded with 1.0 mg/cm² Te power sealed in a vacuum ampoule without molten salt, large amounts of surface tellurides formed, which affected the inward diffusion of Te and resulted in a GB corrosion depth of 139 μm after 3000 h [27]. In contrast, alloy N corroded in Te-containing salts did not show clear evidence of surface tellurides [22, 42-44] and often showed deeper depths even in a much shorter time [42, 44]. Therefore, the Te vapor test may underestimate the extent of Te-induced corrosion in MSRs.

In a salt corrosion test using Te powder as the Te source [44], cracks were formed without tensile stress, indicating that severe Te-induced corrosion had occurred. This result is obtained possibly because the direct introduction of Te⁰ avoids the necessity of reactions Eq. 2 and 3 to release Te⁰ from Te compounds, but may lead to high Te activity and Te release rate, which exceed those observed in real MSRs.

When alloy N was corroded in Cr₃Te₄ containing salt with and without NiF₂ for 1000 h [42], the trend was similar to that observed in the present study; however, the extent of cracking was greater. Without NiF₂, no intergranular cracking occurred even after tensile test, whilst and intergranular cracks deeper than 100 μm formed before tensile test and cracks up to 210 μm formed after tensile with NiF₂ addition. Although Te sources of the same type and concentration were used, the different redox couples contributed to the observed differences. According to the potential chart [49], the equilibrium of NiF₂/Ni resulted in a higher oxidizing potential than EuF₃/EuF₂ or UF₄/UF₃ in a similar concentration range; hence, the Te activity in the salt was higher even though the same Cr₃Te₄ was used. In addition, reduced Ni was deposited as a corrosion product on the alloy surface, which may have affected the corrosion behavior of the alloy.

For the salt corrosion test with 4.8wt.% EuF₃ but with NiTe as Te source [43], the intragranular Cr depletion depth was 23 μm after 150 h, which was much higher than that of the present study (4 μm after 250 h) probably due to the higher EuF₃ concentration. However, no GB depletion was observed in this case, which may be ascribed to the lower Te activity induced by the NiTe source and Ni crucible [46]. The probability of such a low Te activity inducing intergranular cracking is not clear because tensile tests were not performed in this study.

In the present study, the Te activity was maintained sufficiently low such that the Te concentration in the GB of the tested alloy N was below the EPMA detection limit but sufficiently high to cause intergranular cracking after the tensile tests. When 1% EuF₃ was added, the cracking behavior of alloy N was similar to that of the reducing fuel salt, and the K value was equivalent to that of the fuel salt with UF₄/UF₃ ratios of between 60 and 85. When 3% EuF₃ is added, the cracking behavior and K value resemble those of an oxidizing fuel salt. This suggests that the present experimental design may serve as a surrogate approach for avoiding the use of radioactive hazardous fuel salts in future Te corrosion studies.