Dissolved valence state of iron fluorides and their effect on Ni-based alloy in FLiNaK salt

Highlights

• The stable valence of iron fluorides is studied by ultraviolet-visible spectroscopy and electrochemical techniques.

• Fe (II) is unstable in molten FLiNaK salts and gets converted to more stable Fe (III) through the disproportion reaction.

• The corrosion of GH 3535 alloy is aggravated by iron fluorides due to the redox reaction of Fe (III) + Cr = 2 Fe +Cr (III).

Abstract

In this study, the stable dissolved valence state of iron fluorides in molten LiF - NaF - KF (FLiNaK) salt and their effect on GH3535 alloy were systematically investigated. Results of in situ ultraviolet–visible (UV–Vis) absorption spectroscopy and electrochemical analysis indicated that Fe (II) in molten FLiNaK salt is unstable and gets converted to more stable Fe (III) through the disproportionation reaction: 3 Fe (II) = 2 Fe (III) + Fe. Thus, the corrosion of GH3535 alloy is aggravated by the addition of iron fluorides due to the redox reaction: Fe (III) + Cr = Fe + Cr (III), resulting in the formation of the Cr-depletion layer and Fe-rich layer.

Introduction

Molten salts have been widely used in numerous applications, such as heat transfer media for molten salt reactors (MSR), solar energy storage facilities, and hydrogen production [1], [2], [3], [4]. Eutectic LiF - NaF - KF (FLiNaK) was proposed for use as coolant in the secondary loop of the MSR owning to its excellent chemical stability at high temperatures [5]. However, the corrosion of structural material pose a radical challenge for the application of molten fluorides [6]. The corrosion resistance of alloy elements has a tendency of Ni > Fe > Cr > Co, thus, Ni-based alloys exhibit superior corrosion resistance than Fe-based alloys. The Hastelloy N alloy (Ni-17Mo-6Cr) is successfully developed and used for MSR by Oak Ridge National Laboratory. Up to now, many literatures have reported that the corrosion of materials in molten fluorides is mainly driven by the oxidizing impurities, such as water, oxides, acid radical ions, and metal ions [7], [8], [9]. Therefore, sparging a gas mixture of H₂ and HF to react with these highly corrosive impurities is a commonly used strategy to prevent or mitigate corrosion of molten fluorides [10], [11], [12].

nHF + Me (Me = Cr, Fe and Ni) = MeF_n + n/2H₂

However, the purification treatment in the metallic containers can introduce metallic fluorides, and the corrosion of metallic materials during the operation of MSRs also introduces metallic fluorides through Eq. (1) [12]. Therefore, it is crucial to investigate the effect of metallic fluorides on the corrosion of alloys. Studies have demonstrated that chromium fluorides [8], [13], [14] distinctly aggravated the corrosion of pure metal (Ni, Cr and Fe), Hastelloy N, 316 SS and SiC in molten FLiNaK salts. Researchers at the Oak Ridge National Laboratory found that both nickel and iron fluorides accelerated the corrosion of structural alloys [15], [16], [17]. Pavlik et al. [18] demonstrated that FeF₂ and FeF₃ (0.01–10 mol% in FLiNaK salt) can both significantly accelerate the corrosion of 800H. But they did not reveal the corrosion mechanism driven by metallic ions in fluorides.Our previous study [19] observed that relatively high concentration of Fe in FLiNaK salt resulted in the deposition of metallic Fe on the alloy surface. Because of the challenges of determining the specific valence state of iron in molten fluorides, the corrosion mechanism induced by iron fluorides has not been elucidated.

High-temperature ultraviolet–visible (UV–vis) absorption spectroscopy is an effective technique to detect the chemical valence state of iron fluorides in molten FLiNaK salt. The 3d-orbitals of Fe are partially filled and tend to coordinate with the ligands. The energy level of 3d-orbitals can be described by the action of the crystal field, and the d‐d‐orbital energy splitting depends on both the valence state of iron and the type of ligands. For instance, absorption bands for ferric nitrate in an aqueous solution are generated at ultraviolet wavelengths and begin to weaken beyond a wavelength of 550 nm [20]. By contrast, the absorption band for iron cyanide in the aqueous solution appears at 350–500 nm [21], and that for the iron divalent ion is located in the near-infrared region (>800 nm). The absorption spectra of iron ions in molten salt were mainly studied in the 1960s and 1970s. George [22] found that the absorbance peak for Fe (III) trivalent ions in LiCl - KCl salt appeared at 350–450 nm. Young et al. [23] reported that Fe (II) divalent ions in molten fluorides exhibited absorbance peaks and shoulders from 700 to 1200 nm. Therefore, the spectroscopy techniques may be useful to determine the valence state of iron ions in molten fluoride salts. Furthermore, electrochemical techniques also can be used to investigate the stable valence of electro - active species because of the quick response and in situ monitoring. Bing [1] proposed that Fe (III) trivalent ions could be converted to Fe (II) divalent ions in KCl–CaCl₂–NaCl–MgCl₂ melts by cyclic voltammetry. Our previous work also confirmed the stable existence of Cr (III) in FLiNaK salt through cyclic voltammetry and square wave voltammetry [3].

The aim of the present study was to explore the dissolution mechanism and stable valence state of iron ions in molten FLiNaK salt at 700 °C through in situ ultraviolet–visible (UV–vis) absorption spectroscopy and electrochemical techniques, which can further elucidate the corrosion mechanism of nickel-based alloy driven by iron fluorides.