Corrosion behaviour of FeCoNiCrAl high-entropy alloy in molten NaCl–KCl

Abstract

Molten NaCl–KCl–MgCl₂ is becoming a potential heat transfer fluid (HTF) and thermal energy storage (TES) material in the third generation of concentrated solar energy power (CSP) stations due to its excellent heat storage-capacity and thermal stability. However, the strong corrosion of the molten chloride salt at high temperature limits its industrial application. In this study, the corrosion behaviour of FeCoNiCrAl HEA with BCC + FCC phase structure in molten NaCl–KCl–MgCl₂ (24.5–20.55–54.95 wt-%) at 650 °C under argon was investigated by a combination of weight loss and electrochemical methods. The results show that, the corrosion of FeCoNiCrAl HEA in molten chloride salt mainly manifests as the selective dissolution of Fe and Cr. With the further increase of immersion time, oxides gradually crack and slightly flake, resulting in large weight loss of FeCoNiCrAl HEA.

Keywords

FeCoNiCrAl HEA molten chloride salt corrosion electrochemical

Introduction

The concentrated solar power (CSP) plant system is considered as a promising and attractive energy-alternative technology to alleviate the energy problems caused by fossil-fuelled power generation due to its environmental friendliness, high efficiency and economic practicality.¹ For the second generation of CSPs system, molten salt with thermal stability and small viscosity coefficient properties at higher temperatures (≥500 °C) is usually used as heat transfer fluid (HTF) and thermal energy storage (TES) material.² The molten salts used in commercial CSPs are usually a mixture of molten nitrates, which is the commonly so-called solar salt with a melting point of about 223 °C and an operating temperature range of 290 to 565 °C. However, the high-temperature thermal decomposition of molten nitrates limits the operating temperature of CSPs from 560 to 580 °C. In order to enlarge the thermal to electrical conversion efficiency, it has become imperative to advance the operating temperature of the molten salt. Therefore, molten chloride salts with a low melting temperature, higher thermodynamic stability and better heat transfer property become the best alternative medium.^1,3–5 Compared to nitrate mixtures, the melting point of NaCl–KCl–MgCl₂ is not particularly high (383 °C is used in this article), but it has a wider operating temperature range up to 800 °C. When salt is used in third-generation CSPs, its higher operating temperature allows for more efficient power cycles, higher overall efficiency and lower costs.^5,6 Studies have shown that the thermal efficiency of the CO₂ recompression Brayton cycle system is higher when the operating temperature exceeds 600 °C.⁷

At present, the most widely used chloride salt system is eutectic NaCl–KCl–MgCl₂ molten salt mixtures. However, the impurities brought by MgCl₂, which has strong hygroscopicity, cannot be completely avoided. Therefore, it is necessary to study the corrosion behaviour of materials caused by molten chloride salts. Research results show that in molten chloride salts, the alloying elements are prone to active dissolution, especially for Cr, Fe and Si. Combined with the thermodynamic Gibbs free energy calculation results, the stability of common alloying elements augment in the order of Si < Cr < Fe < Mo < Ni. Therefore, nickel base alloys or refractory metals with low Cr content are generally considered to have better corrosion resistance to molten fluoride or chloride salts than iron base alloys. However, selective dissolution of nickel-based alloys still exists during long-term service in molten chloride salts. Grégoire et al.³ investigated the corrosion behaviour of Inconel 600 nickel base alloy in molten NaCl–KCl–MgCl₂ (24.5–20.5–55.0 wt-%) at 700 °C argon isothermal condition. The results showed that Inconel 600 nickel-base alloy was unstable in molten chloride solution at 700 °C, and a large number of Cr-rich carbides were formed in γ-(Ni) solid solution, so the alloy mainly exhibited galvanic corrosion caused by the potential difference between Cr-rich phase and Ni-rich matrix. Gomez-Vidal et al.⁸ studied the corrosion behaviour of In800H and SS310 in molten NaCl–LiCl (34.42−65.58 wt-%) at temperatures of 650 and 700 °C, respectively. It is found that the corrosion resistance of In800H was slightly better than SS310, both alloys exhibited pitting and intergranular corrosion, which were common in chloride salt corrosion, and greatly affected the corrosion resistance of the alloy. Therefore, it is urgent to develop new structural materials resistant to molten chloride salt corrosion.

The concept of ‘high-entropy alloy’ (HEA) is firstly proposed in 2004 by Yeh et al.,⁹ breaking the design concept of traditional alloys and expanding the application field of alloys. Different from traditional alloys, HEAs are characterised by four major effects: high entropy effect, the slow diffusion effect, the lattice distortion effect, and the ‘cocktail’ effect, which make the HEAs show high corrosion resistance.^10–12 Patel et al.¹³ studied the corrosion behaviour of refractory high-entropy alloys TaTiVWZr and HfTaTiVZr, 304 stainless steel (SS304) and Inconel 718 (IN718) in molten NaCl–KCl–MgCl₂ (33–22–45 wt-%) eutectic salts at 450 and 650 °C by electrochemical methods, respectively. It was found that TaTiVWZr–HEA showed the best corrosion resistance in molten chloride salt, and its corrosion rate decreased with the amplify of temperature. This was mainly ascribed to the formation of surface oxides of Ta–V and Ta–V–W and the redox couples of Ti/TiCl₂ and Zr/ZrCl₂, which inhibited the dissolution and chlorination of other constituent elements. Patel et al.¹⁴ also studied the corrosion behaviour of the duplex eutectic high-entropy alloy AlCoCrFeNi_2.1 (EHEA) and traditional duplex stainless steel 2205 (DS2205) in molten NaCl–KCl–MgCl₂ at 450 and 650 °C, respectively. The results showed that the corrosion rate of AlCoCrFeNi_2.1 eutectic high-entropy alloy in molten chloride salt was smaller than that of DS2205 at both temperatures. The FCC-L1₂ phase of AlCoCrFeNi_2.1 eutectic high-entropy alloy acted as a barrier to prevent further oxidation and protect the underlying BCC-B2 phase by enriching precious metal elements in the protective surface layer.

To further understand the corrosion mechanism of HEAs in molten chloride salts and evaluate the commercial application of HEAs in CSP system, the corrosion kinetics and electrochemical behaviour of FeCoNiCrAl HEA in molten NaCl–KCl–MgCl₂ at 650 °C were studied and the corrosion mechanism of FeCoNiCrAl HEA in molten NaCl–KCl–MgCl₂ was discussed.

Experimental procedures

Materials

The molten salt used in this paper was a ternary eutectic NaCl–KCl–MgCl₂ salt (purity >99.0 wt-%) with a composition of 54.95 wt-% MgCl₂, 20.55 wt-% KCl and 24.5 wt-% NaCl (Shanghai Macklin Biochemical Technology Co., Ltd). The salts were dried in a vacuum drying oven for at least 24 h and then placed in a beaker under air and stirred quickly to mix well. With the objective of completely immersed the sample in molten salt, the ratio of salt content to sample surface area is 1:56. The mixed salts were dried further in a vacuum drying oven for 24 h for experiment. The high-entropy alloy used in this experiment was the FeCoNiCrAl HEA, which was obtained by vacuum melting method, provided by Quark Metal Products Co., Ltd in Zhangzhou, and its chemical composition is shown in Table 1. The FeCoNiCrAl HEA plates were cut into samples with dimensions of 10.0 mm × 10.0 mm × 2.0 mm by wire and polished using 800# SiC paper. After polishing, the samples were rinsed with distilled water, cleaned with alcohol ultrasonication and dried for experiments. In order to minimise the effect of moisture, the crucibles were baked in an oven at about 150 °C for 120 min before use.

Table 1.

Chemical composition of FeCoNiCrAl HEA (wt-%).

Element	Fe	Co	Ni	Cr	Al
FeCoNiCrAl HEA	21.70	23.76	23.62	20.71	10.21

Weight loss

Weight loss test was carried out in a resistance furnace under an argon atmosphere at a temperature of 650 °C. The three specimens were fully submerged in the mixed NaCl–KCl–MgCl₂ salt in an alumina crucible with a lid at room temperature. They were heated to the target temperature under an argon flow rate of 4 mL/min, which were maintained for 10, 15, 20, 40, 60, 80 and 100 h, respectively. After experiment, the specimens were cooled down to room temperature along with the molten salt. Argon gas was continuously introduced during the cooling period to prevent further oxidation. The samples were cleaned with hot water at 80 °C and anhydrous ethanol at room temperature. After cleaning, the samples were dried and weighed. The mass loss over time can be calculated as following¹⁵:

\frac{Δ m}{S_{0}} = \frac{m_{i} - m_{f}}{S_{0}}

(1)

where m_i is the initial mass of the specimen in grams; m_f is the mass of the specimen at moment t in grams; S₀ is the initial area of the specimen in cm².

Electrochemical corrosion tests

The electrochemical corrosion tests including potentiodynamic polarisation and electrochemical impedance spectroscopy tests were conducted in a furnace under Ar at 650 °C with a Gamry Reference 3000. The stainless steel wire was spot welded to the end face of the specimen as the electrode lead. The specimen was sealed into alumina tubes with high-temperature cement and kept a working face exposed. Samples for electrochemical testing are placed in the furnace with the salt mixture at room temperature, and when the temperature rises to the target temperature for a certain period of time, the samples are then placed in the molten salt. The entire heating process is protected by argon gas with a flow rate of 4 mL/min. For potentiodynamic polarisation test, a three-electrode system with Pt wire as the reference electrode and graphite as the counter electrode was utilised. The scanning speed was set at 0.3333 mV/s. For the electrochemical impedance spectroscopy tests, two-electrode system was used as shown in Figure 1, two identical specimens are sealed into alumina tubes with high-temperature cement to form a two-electrode system, and the distance between the two electrodes is controlled at about 1 mm.

Characterisation

The phase constitutes of the samples after corrosion tests were analysed using a Japanese RIKEN SmartLab 3 kW X-ray diffractometer (XRD). The diffractometer utilised a Cu target Kɑ radiation line with a wavelength of 0.15406 nm as the radiation source. The operating voltage was set at 40 kV, with a tube current of 100 mA. The scan range was from 10° to 90°, with a scan rate of 10°/min and a step size of 0.02°. The surface and cross-sectional morphologies of the samples were observed at 10 kV using a Hitachi SU8020 field emission scanning electron microscope (SEM). Prior to SEM testing, the specimen was fixed using clamps and sealed with epoxy resin. Subsequently, the specimen was abraded with sandpaper and then polished with diamond paste. Finally, the specimen was washed and dried. Additionally, the chemical elemental composition and distribution of the sample interface were analysed at 20 kV using an Oxford energy dispersive spectrometer (EDS). The EDS scanning time is about 2 to 3 min with the energy spectrometer resolution of 129 eV@Mn Ka.

Results and discussion

Microstructural characterisation

The XRD patterns of FeCoNiCrAl HEA before and after corrosion in NaCl–KCl–MgCl₂ melt at 650 °C for different times are shown in Figure 2. It can be seen that the diffraction peaks in the (1 1 0), (1 1 1), (2 0 0), (2 1 1) and (2 2 0) planes in Figure 2(a) indicate that the structure of FeCoNiCrAl HEA belongs to a combination of FCC and BCC phases. When the corrosion time is 10 h, the corrosion products on the FeCoNiCrAl HEA surface are mainly composed of MgO, Fe₃O₄, Cr₂O₃ and MgCr₂O₄. With the raise of corrosion time, the same corrosion products exist on the FeCoNiCrAl HEA surface. Furthermore, the position of the diffraction peak in (110) plane of FeCoNiCrAl HEA after corrosion with respect to the pre-corrosion crystalline surface is slightly shifted to a higher value, which may be due to the reduce of the lattice parameters caused by the depletion of elements.¹⁶

The cross-sectional morphology of the initial FeCoNiCrAl HEA sample is shown in Figure 3. It can be seen that the morphology of FeCoNiCrAl HEA before corrosion is characterised by an alternating distribution of gray phase and white phase. Combined with the cross-sectional mapping scanning results of the specimen, it can be found that the white area is the Fe-rich phase, the gray area is the Fe-poor phase. Furthermore, the other elements are basically evenly distributed.

The surface and cross-sectional morphologies of FeCoNiCrAl HEA samples after corrosion in NaCl–KCl–MgCl₂ molten salt at 650 °C for different times are shown in Figures 4 to 10, respectively. As shown in Figure 4, after 10 h immersion, some holes appear on the surface of the matrix, leading to the weight loss of the FeCoNiCrAl HEA, which can be confirmed by the weight changes of the FeCoNiCrAl HEA after corroded in molten NaCl–KCl–MgCl₂ melt at 650 °C for 10 h in Figure 11. Furthermore, vertical bar-like projections are visible on the surface. The area between these vertical bars, known as the intraparticle attack, is also filled with some substances. Based on the EDS point scanning results, it can be determined that the vertical bar-like projections are mainly composed of Fe (points 1 and 3), while the products after intragranular attack consists mainly of Mg and O (Point 2). Additionally, traces of Fe and Cr elements are also detected. Coupled with XRD results in Figure 2, the substance in this area is identified as MgO, along with some Fe and Cr oxides. Furthermore, intergranular corrosion occurs in the matrix, and the corrosion layer depth is about 20 μm. The mapping scanning results in Figure 4(c) show that Mg, O and Cr are enriched in the corrosion layer, which also confirms the formation of the oxides.

When the corrosion time of FeCoNiCrAl HEA is prolonged to 15 h, the amount of the corrosion holes on the surface of the substrate significantly increases (Figure 5(a)), and the corrosion channels are deepened, even forming a columnar shape, the substrate sinks into the gullies. Form the cross-sectional morphology (Figure 5(c)), it can be seen that the corrosion depth is deepened and the corrosion pit contains oxides, which may be the main reason for the little difference between the weight loss for 15 and 10 h, as shown in Figure 11. According to the EDS point scanning results, the surface area of the alloy is mainly composed of a mixture of Mg, Fe and Cr oxides (point 4). Combined with the results of XRD analysis, there is no significant change in the product compared with that of the sample after 10 h corrosion. However, the formed oxide layer is discontinuous and porous.

Compared with the sample after corrosion for 15 h, larger pores and more obvious columnar gaps appear on the surface of the matrix when the immersion time is prolonged to 20 h (Figure 6), which is presumed to be caused by the continuous active dissolution of the matrix elements. In addition, dense rice-grained oxides (point 5) begin to appear on the surface with a composition of 40.22Mg–46.27O–11.81Fe–0.47Cr–0.27Al–0.17Na–0.16K (wt-%). Combined with the XRD analysis results, the main phases consist of MgO and Fe₃O₄. According to the mapping scanning results of cross-section morphology (Figure 6(c)), MgO is mainly concentrated in the outermost layer of the corrosion layer. The mixed oxides of Cr and Fe tend to form inside the corrosion layer. A Co-poor zone with a depth of about 10 μm appears on the surface of the matrix. All the results can explain the reason for the significant loss of HEA samples after 20 h of corrosion, as shown in Figure 11.

Compared with the previous surface morphologies in Figures 4 to 6, the surface morphology of FeCoNiCrAl HEA sample after 40 h immersion changes significantly with fewer corrosion pits and voids, the interior of the matrix is gradually covered with the metallic oxides (Figure 7(a) and (b)). An uneven layer of corrosion products is formed on the surface of the substrate, and the surface is covered with distributed particles of different sizes. Further analysis with the EDS point scanning results shows that the barrier is primarily composed of MgO (point 6). Additionally, the particles detected on the barrier (point 8) consist mainly of Mg–Cr oxides. Combined with XRD analysis, the oxide is MgCr₂O₄ spinel. However, the formation of MgCr₂O₄ spinel can hinder the protective effect of MgO by increasing the rate of charge transfer between the alloy and the molten salt. Therefore, it can be assumed that the formation of MgCr₂O₄ reduces the corrosion resistance of the matrix. This observation is consistent with that of Grégoire et al.³ Combined with the corrosion weight loss curve in Figure 11, it can be observed that the weight loss at 40 h is significantly less than that at 20 h, which further indicates that the formation of MgO has a protective effect on the matrix. According to the EDS results of points 7, 9 and 11 on the surface morphology and the XRD pattern analysis, point 9 mainly represents the oxide of Fe, while point 7 represents the matrix with a small amount of oxide. These manifestations indicate that the alloy experiences localised corrosion. Coupled with the cross-sectional morphology shown in Figure 7(c), the corrosion depth is further lengthened, with obvious intergranular corrosion between the Fe-rich and Fe-depleted phases. As the corrosion progressed, the alloy continues to dissolve in the molten salt, making it increasingly difficult to form an intact protective film. Furthermore, compared with corrosion products for 20 h, the oxide particles obtained on the FeCoNiCrAl HEA surface after 40 h immersion are smaller and more compact.

The surface morphology of FeCoNiCrAl HEA after 60 h immersion is similar with that for 40 h immersion, as shown in Figure 8, but with a larger size of the oxide particles (about 6–10 μm) and a deeper corrosion depth (∼45 μm). In addition, some pores are observed, which result in some weight loss of FeCoNiCrAl HEA, as is shown in Figure 11 that the mass weight loss of FeCoNiCrAl HEA at 60 h is greater than that at 40 h. In addition, since FeCl₂ and CrCl₃ volatilise significantly at 536 and 611 °C, respectively. The melting point of FeCl₃ is only 305 °C, when the temperature reaches 330 °C, the vapour pressure of FeCl₃ can reach 1 bar,^17,18 the volatilisation and diffusion of these chlorides are accelerated with the extension of corrosion time, resulting the formation of pores in the matrix. Furthermore, the formed metal chlorides can be further oxidised in the high oxygen partial pressure region during external diffusion process.

As shown in Figure 9, as the immersion time is prolonged to 80 h the corrosion degree of FeCoNiCrAl HEA is further deepened and intra-particle attack makes surface denting worse. Based on the EDS result of point 12, the main elements present are Mg and O, accounting for 54.18 and 37.37 wt-%, respectively. Additionally, small amounts of Cr and Fe are detected. Combined with the results of XRD analysis, the main substances formed on the surface are Mg–Cr oxides, MgO and a small amount of Cr₂O₃ and Fe₃O₄. The particle size of the oxides is slightly larger than those found on the surface of sample at 60 h and some particle sizes are about 10 μm. It can be also observed that the matrix not protected by the MgO film is severely damaged, and the pore channels in the matrix are denser than those of 60 h. However, the mass loss at 80 h is not significantly different from the mass loss at 60 h (Figure 11), which may be due to that the mass loss caused by the dissolution of the matrix is offset by the weight gain caused by the corrosion products formed and the residual salts in the pores, as shown in Figure 9(b).

After the corrosion of FeCoNiCrAl HEA for 100 h (Figure 10), the entire surface is characterised by uniform corrosion and the cracks in the surface oxide layer are detected. Furthermore, the oxidation particles with a size of approximately 15 μm are closely arranged to form small clusters, but there are cracks between each other. Cracks in the oxide layer and large corrosion holes in the matrix explain the mass loss in the weight loss curve (Figure 11). Combined with the results of EDS point scanning and XRD analysis, it can be seen that the corrosion products are basically unchanged. Furthermore, the surface cracks expand, which can be ascribed to that the obtained MgO layer film gradually becomes loose and loses its protective effect over immersion time. According to related reports, accumulating internal stresses as the scale grows can lead to cracks and/or spalling, which allows the electrolytes diffuse easily to the metallic substrate.³

Potentiodynamic polarisation curve

Figure 12 shows the potentiodynamic polarisation curve of FeCoNiCrAl HEA obtained in molten NaCl–KCl–MgCl₂ salt under an argon atmosphere at 650 °C. The FeCoNiCrAl HEA is actively dissolved in molten NaCl–KCl–MgCl₂. The electrochemical parameters of the corrosion of FeCoNiCrAl HEA obtained through the Tafel extrapolation method including the corrosion potential E_corr and the corrosion current density I_corr are given in Table 2. I_corr of FeCoNiCrAl HEA at 650 °C is about 1.49 × 10⁻³ A cm⁻². The corrosion current density of FeCoNiCrAl HEA is significantly smaller than that of SS304, IN718 and HfTaTiVZr HEA in similar proportions of chloride salts at the same temperature.¹³ According to the conservation of matter, the conservation of energy, the conservation of momentum, the first and second laws of thermodynamics, and the ideal gas hypothesis, the electrochemical corrosion rate of the alloy in molten salt at this time can be calculated by combining Faraday's law with the following equation¹³:

CR (mm / year) = K_{1} \times \frac{I_{corr} \times EW}{ρ}

(2)

where K₁= 3.27 × 10⁻³ [mm/(μA cm year)] is a constant, EW is equivalent in g/equivalent, ρ is the density in g/cm³, I_corr is the corrosion current density in A/cm². Based on the above equation, the corrosion rate is calculated to be about 12.8 mm/year. The higher corrosion rate may be due to the active dissolution of Fe and Cr in molten NaCl–KCl–MgCl₂.

Table 2.

Electrochemical parameters of corrosion of FeCoNiCrAl HEA obtained by Tafel extrapolation method according to Figure 12.

Sample	E_corr (mV vs. Pt)	I_corr (A cm⁻²)	EW	CR (mm/year)
FeCoNiCrAl HEA	−550.60	1.49 × 10⁻³	19.58	12.8

Electrochemical impedance spectroscopy (EIS)

To further understand the corrosion behaviour of FeCoNiCrAl HEA, the Nyquist and Bode diagrams for FeCoNiCrAl HEA in NaCl–KCl–MgCl₂ molten salt at 650 °C is given in Figure 13. The Nyquist diagram (Figure 13(a)) shows that two capacitive loops are observed for samples immersed in NaCl–KCl–MgCl₂ molten salt in the initial 40 h, i.e., a capacitive arc with a smaller radius of curvature in the high-frequency region and a capacitive arc with significantly larger radius in the low-frequency region. In Bode diagram shown in Figure 13(b), two corresponding time constants are also reflected. Based on the characteristic of the curves and considering the roughness and inhomogeneity of the FeCoNiCrAl HEA surface, the electrochemical impedance spectra can be fitted by the equivalent circuit shown in Figure 14(a), where R_s represent the solution resistance, Q_f and R_f the capacitance and resistance of the corrosion layer on the surface of the FeCoNiCrAl HEA, Q_dl the double-layer capacitance and R_ct the charge-transfer resistance. The impedance of Q is expressed as¹⁹:

Z = Y^{- 1} (j w)^{- n}

(3)

The fitting results are shown in Figure 14 and the fitting curve is in good agreement with the measured values, indicating the selected circuit is suitable. The obtained corrosion electrochemical parameters are shown in Table 3. At the beginning of the corrosion process, the value of R_ct is large and then decreases with the raise of corrosion time, which may be attributed to the active dissolution of elements in the alloy. However, when the corrosion time increases to 40 h, R_ct and R_f both intensify, which may be ascribed to the formation of oxides on the sample surface, slowing down the penetration of corrosion ions to a certain extent.²⁰ This can be also verified by the morphologies of samples after corrosion for 40 h shown in Figure 7. When the corrosion time exceeds 40 h, the Nyquist diagram consists of a capacitive arc in the high-frequency zone and a straight line in the low-frequency zone, which is represented by the Warburg impedance. The Warburg impedance Z_w is expressed as²⁰:

Z_{w} = A_{w} j w^{- 1 / 2}

(4)

Table 3.

Fitting values for EIS tests of FeCoNiCrAl HEA in molten NaCl–KCl–MgCl₂ salts at 650 °C.

Time (h)	R_s (Ω cm²)	Y_f (Ω⁻¹ cm⁻² S⁻ⁿ)	n _f	R_f (Ω cm²)	Y_dl (Ω⁻¹ cm⁻² S⁻ⁿ)	n _dl	R_ct (Ω cm²)	A_w (Ω cm² s^−0.5)
11	1.53	0.03	0.59	0.26	2.35	0.59	20.37	—
21	1.42	0.12	0.55	0.23	3.18	0.61	12.29	—
41	1.41	0.80	0.41	0.42	7.40	0.64	32.45	—
61	1.46	—	—	—	0.70	0.49	0.26	8.48
88	1.52	—	—	—	0.98	0.39	0.34	12.15
99	1.62	—	—	—	0.84	0.41	0.33	12.18

The corresponding circuit diagram is shown in Figure 14(b) and the fitting results are shown in Table 3. It can be seen that the value of R_ct is much smaller than that in the first stage. Furthermore, the exist of Warburg impedance also indicates that the corrosion process is controlled by the ion diffusion.¹⁹ At the later stage of corrosion, due to the oxidation and chlorination of alloying elements on the surface of the sample, a large number of corrosion products accumulate on the surface, which further block the diffusion of active-dissolved ions.

Discussion

Among the three salts, MgCl₂, KCl and NaCl, the alkaline earth metal salt MgCl₂ has strong hygroscopicity and the crystal water cannot be completely removed during the drying process of the salt, thus producing corrosive components such as MgOHCl and HCl after hydration decomposition.²¹ Hydrolysis reactions of MgCl₂ at different temperatures have been proposed by some researchers and some of them are shown in the following equations (5) to (8).^3,22

{MgCl}_{2} \cdot 2 H_{2} O (s) \vec{\leftarrow} {MgCl}_{2} \cdot H_{2} O (s) + H_{2} O (g) T \geq 242 \circ C

(5)

{MgCl}_{2} \cdot H_{2} O (s) \vec{\leftarrow} {MgCl}_{2} (s) + H_{2} O (g) T \geq 304 \circ C

(6)

{MgCl}_{2} \cdot H_{2} O (s) \vec{\leftarrow} MgOHCl (s) + HCl (g) T \geq 350 \circ C

(7)

MgOHCl (s) \vec{\leftarrow} MgO (s) + HCl (g) T \geq 554 \circ C

(8)

The corrosion mechanism of the FeCoNiCrAl HEA in molten NaCl–KCl–MgCl₂ at 650 °C is also shown in Figure 15. During the experiments, the temperature of the molten salt is continuously heated to 650 °C. The above-mentioned side reactions can occur throughout the process. The HCl produced by the decomposition of MgCl₂ can make the whole environment acidic and reacts with the elements in the alloy to form the corresponding metal chlorides, which has an accelerating effect on the corrosion of the sample.²³ At the same time, trace amounts of water impurities may exist inevitably in molten salt and argon (argon purity ≥ 99.999%), which provide oxygen to the molten salt environment. Thus the hydrogen chloride gas can prompt the Deacon reaction to occur and produce Cl₂, which further corrodes the substrate as following:

4 HCl (g) + O_{2} (g) \vec{\leftarrow} 2 {Cl}_{2} (g) + 2 H_{2} O (g)

(9)

M (s) + x / 2 {Cl}_{2} (g) \vec{\leftarrow} M {Cl}_{x} (s, g)

(10)

where M refers to the elements contained in the HEA such as Cr, Fe, Co, Al, Ni. The enthalpy of reaction (9) is calculated to be −117.821 kJ, which has a negative value, indicating that the reaction is exothermic and releases energy to the outside, and that the rate of reaction usually speeds up as the reaction proceeds. The Gibbs free energy for the reaction (9) was calculated by HSC 6.0 to be 7.504 kJ at 650 °C, which is positive, indicating that the reaction cannot proceed spontaneously. However, metal oxides can act as catalysts for this reaction and reduce the reaction activation energy.²⁴ The formed chlorides are partially oxidised to corresponding metallic oxides in further when encounter a region with high oxygen potential in the molten salt. However, in addition to Cl₂ and HCl, electrochemical corrosion is also one of the main driving forces. Ding et al.²⁵ experimentally verified that the corrosion of alloys in molten NaCl–KCl–MgCl₂ was mainly controlled by electrochemical reactions, and the MgOH⁺ impurity produced by the hydrolysis of MgCl₂ was the most important reductant in the reaction process. The reduction of MgOH⁺ impurity to MgO is a cathodic reaction and the corresponding anodic reactions occur to maintain charge neutrality, i.e., the dissolution reaction of the elements in FeCoNiCrAl HEA. In order to maintain the electrical neutrality in the molten salt, the Cl⁻ anion migrates towards the metal surface and form the CrCl₂ and FeCl₂, part of which are further oxidised to CrCl₃ and FeCl₃ with the progression of corrosion process. In addition, the high vapour pressure possessed by chlorides of chromium and iron causes continuous evaporation at the corrosion front, resulting in voids within the FeCoNiCrAl HEA.²⁴ The escaped chlorides are diffused from the metal surface to the molten salt through diffusion paths such as cracks and holes, and the residual metal chlorides on the sample surface are dissolved in the cleaning solution when the sample is cleaned after the experiment is completed, so the chloride cannot be detected by XRD.

The metal chlorides formed by the above reaction can react with the dissolved oxygen to generate the corresponding oxides and Cl₂ in the higher oxygen pressure region during the escape process, and Cl₂ will further corrode the substrate. Therefore, the Fe₃O₄ and Cr₂O₃ are formed. Coupled with the morphologies of the FeCoNiCrAl HEA after different corrosion time, the obtained oxides are loose, which can hardly play a protective role for the substrate, and thus accelerate the diffusion of elements inside the alloy to the surface, leading to further corrosion of the material.²⁶ In addition, MgCr₂O₄ spinel is also detected due to the reaction between the Cr₂O₃ and MgO, which brings a certain negative impact on the corrosion process of the substrate. Furthermore, according to the standard Gibbs free energy of metals forming the corresponding metal chlorides at 650 °C as shown in Table 4, it can be seen that Al is preferentially attacked due to the maximum Gibbs free energy in the FeCoNiCrAl HEA, but no substance containing Al is detected by XRD and EDS, which may be mainly due to the low content in the alloy with a value of only about 10.21 wt-% and at the same time, the formed metal chlorides may be washed away in the cleaning stage.

Table 4.

Standard Gibbs free energy of some metals forming the corresponding metal chlorides at 650 °C.

Reaction	ΔG (kJ/mol) (650 °C)
Ni + Cl₂(g) = NiCl₂	−166.805
2/3Al + Cl₂(g) = 2/3AlCl₃	−345.649
2/3Fe + Cl₂(g) = 2/3FeCl₃	−156.390
2/3Cr + Cl₂(g) = 2/3CrCl₃	−240.650
Fe + Cl₂(g) = FeCl₂	−225.978

Conclusions

The corrosion of FeCoNiCrAl HEA in NaCl–KCl–MgCl₂ melt at 650 °C is mainly manifested by selective dissolution of alloy-elements. With the extension of corrosion time, the depth of corrosion layer thickens and the corrosion hole expands, the indentation caused by intra granular attack is gradually apparent on the surface of the sample. Therefore, FeCoNiCrAl HEA exhibits significant weight loss after soaking for 100 h, but the formation of surface MgO slows down the weight loss to a certain extent. Due to the chlorination and oxidation reaction of Fe and Cr, the corrosion products on the substrate surface are mainly composed of Fe₃O₄, Cr₂O₃ and MgCr₂O₄. Furthermore, at the later stage of corrosion, the corrosion process is controlled by the ion diffusion in the corrosion product layer, which manifested as Warburg impedance in the electrochemical impedance spectrum.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This work was supported by the Natural Science Foundation of Guangxi Province (grant number 2021GXNSFAA220118).

ORCID iD

Yanli Wang

References

Gong

Shi

Chai

, et al. Molten chloride salt technology for next-generation CSP plants: compatibility of Fe-based alloys with purified molten MgCl2-KCl-NaCl salt at 700°C. Appl Energy 2022; 324: 119708.

Al Kindi

Sapin

Pantaleo

, et al. Thermo-economic analysis of steam accumulation and solid thermal energy storage in direct steam generation concentrated solar power plants. Energy Convers Manage 2022; 274: 116222.

Grégoire

Oskay

Meißner

, et al. Corrosion mechanisms of ferritic-martensitic P91 steel and inconel 600 nickel-based alloy in molten chlorides. Part II: NaCl–KCl–MgCl₂ ternary system. Sol Energy Mater Sol Cells 2020; 216: 110675.

Fernández

ÁG

Cabeza

. Molten salt corrosion mechanisms of nitrate based thermal energy storage materials for concentrated solar power plants: a review. Sol Energy Mater Sol Cells 2019; 194: 160–165.

Gong

Hanke

Kessel

, et al. Molten chloride salt technology for next-generation CSP plants: selection of cold tank structural material utilizing corrosion control at 500°C. Sol Energy Mater Sol Cells 2023; 253: 112233.

Ding

Bauer

. Progress in research and development of molten chloride salt technology for next generation concentrated solar power plants. Engineering 2021; 7: 334–347.

Dunham

Iverson

. High-efficiency thermodynamic power cycles for concentrated solar power systems. Renewable Sustainable Energy Rev 2014; 30: 758–770.

Gomez-Vidal

Tirawat

. Corrosion of alloys in a chloride molten salt (NaCl-LiCl) for solar thermal technologies. Sol Energy Mater Sol Cells 2016; 157: 234–244.

Yeh

Chen

Lin

, et al. Nanostructured high-entropy alloys with multiple principal elements: novel alloy design concepts and outcomes. Adv Eng Mater 2004; 6: 299–303.

10.

Miracle

Senkov

. A critical review of high entropy alloys and related concepts. Acta Mater 2017; 122: 448–511.

11.

Wang

, et al. High-entropy alloy: challenges and prospects. Mater Today 2016; 19: 349–362.

12.

Xiong

Guo

AXY

Zhan

, et al. Refractory high-entropy alloys: a focused review of preparation methods and properties. J Mater Sci Technol 2023; 142: 196–215.

13.

Patel

Sadeghilaridjani

Pole

, et al. Hot corrosion behavior of refractory high entropy alloys in molten chloride salt for concentrating solar power systems. Sol Energy Mater Sol Cells 2021; 230: 111222.

14.

Patel

Hasannaeimi

Sadeghilaridjani

, et al. Molten salt corrosion behavior of dual-phase high entropy alloy for concentrating solar power systems. Entropy 2023; 25: 96.

15.

Fernández

Pérez

. Improvement of the corrosion properties in ternary molten nitrate salts for direct energy storage in CSP plants. Solar Energy 2016; 134: 468–478.

16.

Guo

, et al. The high-temperature corrosion of Hastelloy N alloy (UNS N10003) in molten fluoride salts analysed by STXM, XAS, XRD, SEM, EPMA, TEM/EDS. Corros Sci 2016; 106: 249–259.

17.

Niu

. Chlorination of metallic materials at high temperature. Corros Sci Prot Technol 2000; 12: 41–44.

18.

Lambrecht

de Miguel

Lasanta

, et al. Past research and future strategies for molten chlorides application in concentrated solar power technology. Sol Energy Mater Sol Cells 2022; 237: 111557.

19.

Zhu

, et al. Corrosion of Ni–Fe based alloy in chloride molten salts for concentrating solar power containing aluminum as corrosion inhibitor. Sol Energy Mater Sol Cells 2022; 241: 111737.

20.

Zhu

Zeng

Zhang

, et al. Electrochemical study on the corrosion behaviors of 316 SS in HITEC molten salt at different temperatures. Sol Energy Mater Sol Cells 2018; 186: 200–207.

21.

Ibrahim

Peng

Riaz

, et al. Molten salts in the light of corrosion mitigation strategies and embedded with nanoparticles to enhance the thermophysical properties for CSP plants. Sol Energy Mater Sol Cells 2021; 219: 110768.

22.

Vidal

Klammer

. Molten chloride technology pathway to meet the U.S. DOE sunshot initiative with Gen3 CSP. AIP Conf Proc 2019; 2126: 080006.

23.

Kruizenga

. Corrosion mechanisms in chloride and carbonate salts. SANDIA REPORT, https://doi.org/10.2172/1051732 (2012).

24.

Zahs

Spiegel

Grabke

. Chloridation and oxidation of iron, chromium, nickel and their alloys in chloridizing and oxidizing atmospheres at 400–700°C. Corros Sci 2000; 42: 1093–1122.

25.