I. INTRODUCTION

Among the reactor systems proposed in the Generation-IV International Forum, the molten salt reactor (MSR) has attracted much more attention due to its unique features, such as high thermodynamic efficiency, intrinsic safety, online refueling, etc. [1-4]. In the MSR, molten salt mixtures are used as the primary coolant or even as the fuel itself. These salts are characterized by high volumetric heat capacities, low viscosities, high boiling points, and other desirable features [5]. However, the molten salts are highly corrosive, especially at high temperatures. Corrosion of the structural materials in molten salt environments is one of the main challenges that prevent the MSR from long term operation [1]. A lot of work has been done in recent years to investigate the corrosion mechanisms of materials in molten salt environments [6-13]. It has been observed that among all the candidate structure materials for the MSR, the Cr composition was selectively attacked by molten fluoride salts and the corrosion rates were correlated with the Cr content of the alloys [6, 7].

In spite of the known tendency for Cr to be selectively attacked by molten fluoride salts, Cr is still widely used in high temperature alloys because it can improve the high-temperature oxidation resistance of the materials. Pure eutectic FLiNaK salt is emerging as a leading candidate salt for use as a secondary coolant for the MSR due to its advantageous thermal and physical properties [5, 14]. Therefore, a systematic study on the corrosion mechanism of Cr in molten FLiNaK salt is needed to ensure its safe application in MSR. To do this, one needs to understand the corrosion product of Cr in molten FLiNaK salts. However, to our knowledge, most of the research considers the case of the corrosion products of Cr in fluoride fuel salts and/or nitrates [14-16], in which uranium is the key factor in the corrosion process. A good understanding of fuel salt corrosion has been developed, but few data are available for pure coolant salts (e.g. FLiNaK). Therefore, the precise corrosion products of Cr in pure FLiNaK salt are highly desirable. Since the Cr content of most high temperature alloys is no more than 25wt.%, in order to yield sufficient concentrations of the corrosion products for identification, pure metal Cr was chosen for this research. In this work, metal Cr was exposed to molten FLiNaK salt in a graphite crucible at 700 ℃ for 200 h. After the corrosion test, the FLiNaK salt was investigated using X-ray absorption near-edge structure spectroscopy (XANES), a transmission electron microscope (TEM), and X-ray diffraction (XRD). The results show that the main corrosion product of Cr in FLiNaK salt is K₂NaCrF₆.

II. EXPERIMENT

The pure FLiNaK eutectic salt was supplied by Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences. The impurities of the salt, detected by a Inductively Coupled Plasma-Optical Emission Spectrometer (ICP-OES), were as follows: 6 ppm Fe, 74 ppm Ni, 1 ppm Cr, 11 ppm Ca, 17 ppm Mg, 8 ppm Zn, and 6 ppm P. The content of H₂O was less than 100 ppm as determined using Carl-Fisher water analyzer. The high purity Cr was purchased from Beijing Cuibolin Non-ferrous Technology Developing Co., Ltd. The metal Cr used in this work are irregular metal pieces and the purity is 99.99%.

The corrosion test was performed in a graphite crucible due to the relative inertness of graphite to molten fluoride salts [6, 17]. A graphite crucible was made from purified graphite (CDI-1A). Prior to the corrosion test, the graphite crucible was ultrasonically cleaned in ethanol, and then baked at 800 ℃ for 12 h in a vacuum condition to expel contaminants, residual oxygen, and water. After the heat treatment, the crucible was removed from the furnace and immediately stored in an argon atmosphere glove box. After that, 30 g of pure metal Cr pieces and 120 g of FLiNaK salt were placed in the graphite crucible. To ensure salt purity, the graphite crucible was sealed in a 316 stainless steel capsule and welded shut in the argon atmosphere glove box. The schematic diagram of the corrosion test capsule is shown in Fig. 1.

Fig. 1.Full-screen View

(Color online) Schematic diagram of the capsule for the corrosion test at 700 ℃ for 200 h. Dimensions are in centimeters.

After the final weld, the capsule was transferred to a furnace and heated at 700 ℃ for 200 h. After the corrosion test, the capsule was cooled to room temperature and sliced open on the lathe. The Cr concentration of the FLiNaK salt after the corrosion test was analyzed by ICP-OES. The corrosion products were identified by XANES, TEM (FEI Tecnai G2 F20, 200 kV), and XRD (Rigaku D/max2500 V, Cu K_α, λ = 0.154 nm). The sample for the TEM was ultrasonically dispersed in ethanol and transferred to carbon-coated copper TEM grid. The Cr K-edge XANES spectra in the fluorescence yield mode were measured at the BL14W1 beamline of Shanghai Synchrotron Radiation Facility (SSRF) [18] which operates at an energy of up to 3.5 GeV and a stored current of 240 mA. The energy scale of the XANES spectra for the Cr K-edge (5.989 keV) was calibrated by a Cr metal foil. To avoid the impact of oxygen on the retrieved FLiNaK salt, all the salt samples were sealed in plastic bags during the XANES and XRD analysis processes.

III. RESULTS AND DISCUSSION

The cooled FLiNaK salt was retrieved after the corrosion test at 700 ℃ for 200 h. As shown in Fig. 2, the color of the FLiNaK salt changed from white to black-green after the corrosion test, which was mainly caused by the dissolution of Cr [19]. According to the results of ICP-OES analysis, the Cr and Ni contents of the FLiNaK salt after the corrosion test were 2403(190) ppm and 4.4(2) ppm, respectively. The concentration of Fe was below the quantitative detection limits of ICP-OES, as shown in Table 1. The Ni and Fe contents of the salt after the corrosion test were lower than that in the as-received salt. The main reason for this phenomenon is that Ni and Fe ions, which were introduced during the FLiNaK purification process, are consumed in the corrosion test via the following reaction [14, 19],

Fig. 2.Full-screen View

(Color online) The FLiNaK salt (a) before and (b) after the corrosion test at 700 ℃ for 200 h.

The reaction products Fe and Ni deposited on the surface of the metal Cr. Additionally, since it is a static corrosion test, the Cr corrosion product in the FLiNaK salt is not distributed evenly, especially at the contact area between the FLiNaK salt and the metal Cr. The Cr concentration of the salt is higher in the zone near the metal Cr. There are some green products at the salt-Cr interface.

Since XANES spectroscopy is sensitive to change in the Cr coordination environment, it was employed to determine the true oxidation state of Cr in FLiNaK salt after the corrosion test. Figure 3 shows the Cr K-edge XANES spectra of the post-experiment FLiNaK salt and the reference compounds (CrF₃, 99.98%, and CrF₂, 99%). Absorption edge shifts to a higher energy as the oxidation state of the material increases [20]. According to the edge position of the XANES spectra, it is clear that the edge position of the post-experiment FLiNaK salt, positioned at 6005 eV, agrees well with that in CrF₃. This indicates that the main oxidation state of Cr in FLiNaK salt is Cr³⁺. It is known that FLiNaK salt is a strong Lewis basic solvent, Cr²⁺ is not stable in molten FLiNaK salt [6, 21]. In the FLiNaK salt environment, the low oxidation ion Cr²⁺ would disproportionate in the following way [21, 22],

Fig. 3.Full-screen View

(Color online) Normalized Cr K-edge XANES spectra of the post-experiment FLiNaK salt, CrF₂ and CrF₃.

The alkali metal fluorides, i.e. FLiNaK salt, are ionic compounds, and the molten salt can easily give up their F^-. In molten FLiNaK salt, the Cr³⁺ ion would interact with F^- to form complexes, which will decrease the thermodynamic activity and stabilize the Cr³⁺ [6, 14, 22].

The morphology and crystal structure of the post-experiment FLiNaK salt are characterized by the TEM. As depicted in Fig. 4(a), some black, irregular solid particles and rod-like crystals can be clearly observed in the typical TEM image. According to the EDS analysis, the rod-like crystals in the image correspond to the FLiNaK salt, whereas the black irregular particles contain Cr. Figure 4(b) presents the EDS spectrum of the black particle, only the peak of the elements F, Na, K, Cr, C, and Cu are shown in the spectrum. Since C and Cu contributed from the carbon-coated copper TEM grid, the EDS results indicate that the black particles are rich in F, Na, K, and Cr, which correspond to the Cr corrosion products. In order to determine the structure, the Cr-rich particles (black, irregular particles) are characterized by the selected area electron diffraction (SAED), as shown in Fig. 4(a). The lattice distances calculated from the diffraction dots are 4.76 Å (111), 2.88 Å (220), 2.45 Å (311), 2.40 Å (222), 1.83 Å (420), and 1.44 Å (440), which can be indexed to the standard face centered cubic K₂NaCrF₆ structure (Table 2). Silicon single crystals with a crystalline plane of (100) are used for calibration. Combined with the EDS results, it is confirmed that the Cr-rich black particle in the FLiNaK salt is K₂NaCrF₆. The lattice constant was calculated to be 8.21 Å, which correspond closely to the standard lattice constant of K₂NaCrF₆ (a=8.27 Å, space group symmetry Fm3m, JCPDS No.: 73-0415).

Fig. 4.Full-screen View

(Color online) (a) TEM image of the post-experiment FLiNaK salt and (inset) SAED patterns of the red circle area, (b) EDS spectrum of the red circle area in (a).

To further confirm the state of the Cr corrosion product, XRD was used to determine the structure of the salt after the corrosion test. Due to the low concentration of Cr in the post-experiment salt, the Cr corrosion products are hardly detected by XRD. Fortunately, some green product precipitates out at the salt-Cr interface. The post-experiment salt containing the green product was characterized by XRD, as shown in Fig. 5. Compared to the XRD patterns of pure FLiNaK salt, some new diffraction peaks appear on the post-experimental FLiNaK salt. Although the new peaks are wide and weak because of the low concentration of the corrosion product, the new diffraction peak positions correspond closely to the diffraction peaks of K₂NaCrF₆. The calculated d-spacing values and the corresponding Miller indexes are given in Table 2. These values match well with the standard diffraction patterns of K₂NaCrF₆. The new peaks can be indexed to the face-centered cubic structure K₂NaCrF₆ (JCPDS No.: 73-0415). Comparing Fig. 5(a) with Fig. 5(b), it can be found that the diffraction peaks of the FLiNaK salt after the corrosion test are smaller than their corresponding counterparts of the salt before the corrosion test. According to the Bragg equation, this indicates that the interlayer distance of the salt increased after the corrosion test. As is known, the Cr atomic radius is larger than that of K and Na, the lattice parameters of the FLiNaK salt slightly increase when Cr dissolves into the salt. The XRD results are in good agreement with that obtained from SAED and EDS analysis, which further verified that the corrosion product of Cr in FLiNaK salt is K₂NaCrF₆.

Fig. 5.Full-screen View

XRD patterns of the FLiNaK salt (a) before and (b) after the corrosion test at 700 ℃ for 200 h. Bottom: standard XRD pattern of K₂NaCrF₆ (JCPDS No.: 73-0415).

As mentioned above, the corrosion product of Cr in cooled FLiNaK salt is confirmed to be K₂NaCrF₆. The corrosion reaction in molten FLiNaK salt can be expected to reasonably satisfy the following equation:

In addition, the FLiNaK salt used as the reactor coolant is working under a high temperature molten state, therefore, future studies will consider whether the structure of Cr in high temperature molten FLiNaK salt is consistent with that in cooled FLiNaK salt.

IV. CONCLUSION

In summary, pure metal Cr was exposed to molten FLiNaK salt in a graphite crucible at 700 ℃ for 200 h. Due to the dissolution of Cr, it was found that the color of the FLiNaK salt changed from white to black-green after the corrosion test. XANES spectra revealed that the predominant oxidation state of Cr in FLiNaK salt is Cr³⁺. The detailed structural characterizations confirm that the main corrosion product of Cr in cooled FLiNaK salt is K₂NaCrF₆. Based on the indexing in the SAED and XRD patterns, the Cr corrosion product appears to be a face centered cubic structure with a lattice constant of a 8.21 Å.