High-purity (99.99%) Al, Fe, Ni, Mo, and Nb powders with the same molar ratio of each element were invoked as raw materials to prepare an AlFeNiMoNb plate by hot pressing technology.

To improve the high-temperature chloride corrosion resistance of 904L and lower the cost of AlFeNiMoNb alloy, 904L:AlFeNiMoNb high-entropy alloy was obtained by mixing a commercial 904L plate and an AlFeNiMoNb plate with an equimolar ratio by arc melting process. At last, the as-cast 904L:AlFeNiMoNb high-entropy alloy was annealed at 900°C for 24 h to homogenize its composition. The actual chemical compositions of the 904L, AlFeNiMoNb, and 904L:AlFeNiMoNb alloys detected by energy dispersive spectroscopy (EDS) analysis are shown in Table 1. In order to investigate the corrosion performance, the tested alloys were cut into pieces with the size of 10 × 5 × 2.5 mm³ by a wire-cutting equipment (WEDM).

Corrosion tests were carried out under N₂–2.6 vol.% CO₂–1.3 vol.% O₂–2,700 vppm HCl gaseous environment at 700°C and 800°C for 55 h, respectively. Each specimen was placed into a single quartz crucible, so that any spallation of corrosion products could be collected separately after exposure. Then all the samples were placed into the silica tube of a horizontal tube furnace, which had three temperature control instruments to ensure the uniformity of temperature in the furnace as shown in Figure 1. Only N₂ and the mixture of N₂, CO₂, and O₂ were input at the temperature-rising stage, which had a heating rate of 4°C/min. The flow rate of each gas was controlled by a mass flow meter. After mixing in the chamber, gases were input into the silica tube. When the temperature reached 700°C, the mixture of HCl and N₂ was input into the corrosion system. The total flow of all gases was 75 ml/min. At the end of the gas path, excessive HCl was absorbed by NaOH aqueous solution. The corrosion kinetics of the samples were obtained by the gravimetric method. The corrosion process was stopped every 5 h. Samples were cooled naturally with the furnace, and then an electronic balance, with a precision of 1 in 100,000, was utilized to weigh the samples. After that, the corrosion process was carried on until 55 h.

In order to examine the microstructure and composition of the experimental samples, metallographic method, scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS), and x-ray diffraction (XRD) detections were used before the corrosion process. After the corrosion experiments, the morphology and composition of the surface and cross-section of the specimens were examined using SEM/EDS and XRD technology. The analysis of the cross-section morphology of the scales was undertaken by molding the corroded samples in resin, and metallographic sections were prepared. The cross sections were ground and polished dry to prevent dissolution of the formed chlorides.

The metallographic diagrams of 904L, AlFeNiMoNb plate, and 904L:AlFeNiMoNb high-entropy alloy and the EDS results of 904L:AlFeNiMoNb are presented in Figure 2. 904L had a typical austenite structure. The AlFeNiMoNb plate had several different phase compositions as shown in Figure 3, which were obtained by the analysis of its XRD and EDS results. The bright region in the metallographic diagram of the alloy was the Mo-rich phase, which contained part of Nb. The gray region contained many phases, just like FeAl₂, MoNi, and Fe₂Nb. In the black portion, AlNi was dominant. Compared with 904L and AlFeNiMoNb alloys, 904L:AlFeNiMoNb had a rather homogeneous and refined microstructure (Figure 2C). The 904L:AlFeNiMoNb alloy has a dendrite microstructure. The dendrites were mainly composed of AlNi and (Ni, Fe) solid solutions, while the composition between dendrites were mainly composed of (Cr, Ni, and Fe) and (Ni, Fe) solid solutions and other intermetallic compounds, such as Fe₂Nb, FeAl₂, and MoNi.

From the XRD patterns of the obtained alloys, it could be analyzed that the 904L:AlFeNiMoNb alloy had a similar phase composition with the AlFeNiMoNb alloy, but there was no obvious Mo diffraction peaks detected. Meanwhile, diffraction peaks of (Cr, Ni, and Fe) and (Ni, Fe) solid solutions became more obvious. These results were consistent with the metallographic diagrams of the alloys; the melting process of the 904L alloy and AlFeNiMoNb alloy promoted the homogenization of microstructure and simplification of phase composition. From the composition of the 904L:AlFeNiMoNb, it could be seen that there were more than five elements, whose concentration were in the range of 5–35 at%. Combined with the XRD analysis of the obtained alloy, 904L:AlFeNiMoNb could be considered as a kind of high-entropy alloy and might have excellent corrosion resistance.

From the perspective of thermodynamics, the driving force of chemical reaction is the change value of Gibbs free energy in the reaction process. The more negative the value, the greater the driving force of the reaction, and the more likely the reaction is to occur.

Figure 4 shows the temperature dependence of Gibbs free energy of metal oxidation and chlorination calculated by HSC6.0.

As far as the same metal was concerned, the Gibbs free energy changes in the reaction between metal and O₂ to obtain metal oxide (M_xO_y) and the reaction among metal, HCl, and O₂ to obtain chloride oxide of metal (MO_xCl_y) are similar. So, at the initial stage of the process of high-temperature chloride corrosion, the metals were oxidized to M_xO_y and MO_xCl_y easily, and the formation of MCl_x was not competitive. MCl_x and MO_xCl_y products are unstable phases under the experimental temperature, and have low melting point and high equilibrium vapor pressure. They will volatilize outward from the interface between alloy matrix and the corrosion layer to the surface of the corroded sample. With the increase in the oxygen pressure during this diffusion path, MCl_x and MO_xCl_y will be oxidized to metal oxides or leave the sample surface with the carrier gas, and the latter process could cause the mass loss of the tested sample during the corrosion process.

The difficulty of oxidation of MCl_x and MO_xCl_y to metal oxides is related to the driving force of oxidation reaction, that is, the change in Gibbs free energy of the reaction between these compounds and oxygen. Figure 5 shows the temperature dependence of Gibbs free energy of oxidation reaction of MCl_x and MO_xCl_y calculated by HSC6.0. MCl_x and MO_xCl_y will be oxidized to M_xO_y, and Cl₂ is released. These kinds of oxides, which are obtained by gas reaction, are very loose and have no effective protection of the metal matrix from the corrosive media of Cl⁻ (Li, 2000). Due to the existence of chloride concentration gradient, the generated Cl₂ will diffuse back to the corrosion layer/metal interface and continue to corrode the metal matrix. On the outer surface of the corrosion layer, the corrosive HCl will react with the formed oxides to obtain MCl_x. Therefore, what needs to be investigated next is the stability of metal oxides in the chlorine atmosphere, which could be evaluated by the driving force of the chemical reaction between metal oxide and HCl.

Figure 6 shows the temperature dependence of Gibbs free energy for the reaction of metal oxides with HCl calculated by HSC6.0. The higher the Gibbs free energy, the more stable the metal oxide is. On the basis of thermodynamic calculation, the high-temperature chloride corrosion process can be described by Figure 7; the corresponding process has been described above.

Figure 8 shows thermo-gravimetric results for the high-temperature chloride corrosion of alloys at 700°C and 800°C for 55 h under N₂–2.6 vol.% CO₂–1.3 vol.% O₂–2,700 vppm HCl gaseous environment. When the corrosion temperature was 700°C, 904L showed obvious mass loss, especially after being corroded for 10 h. On the other hand, AlFeNiMoNb and 904L:AlFeNiMoNb high-entropy alloy showed mass gain. After being corroded for 55 h, the total mass gain of 904L, AlFeNiMoNb plate, and 904L:AlFeNiMoNb high-entropy alloy was −3.09557, 0.30425, and 0.15714 mg/cm², respectively. With the corrosion temperature increasing to 800°C, 904L showed serious mass loss during the chloride corrosion process, especially at the time range of 10–25 h. After 25 h, the thermo-gravimetric curve of 904L showed a relatively slight change in mass. Both AlFeNiMoNb plate and 904L:AlFeNiMoNb high-entropy alloy showed almost straight thermo-gravimetric curves with small mass changes, especially for the latter. After being corroded for 55 h, 904L, AlFeNiMoNb plate, and 904L:AlFeNiMoNb high-entropy alloy obtained a total mass gain of −33.36776, −1.35138, and −0.0118 mg/cm², respectively.

From the thermo-gravimetric data for the high-temperature chloride corrosion process of the tested alloys, it could be concluded that the 904L alloy was not suitable for the application in oxidizing chlorine-containing atmosphere above 700°C. On the other hand, the AlFeNiMoNb plate and 904L:AlFeNiMoNb high-entropy alloy showed excellent corrosion resistance under the corrosion temperatures of 700°C and 800°C. The 904L:AlFeNiMoNb high-entropy alloy had the highest high-temperature chloride corrosion resistance among the samples and also had the most stable corrosion resistance with the change in temperature.

The corrosion properties of materials are closely related to their composition and microstructure. From the temperature dependence of Gibbs free energy of metal oxidation and chlorination as shown in Figure 4, it could be concluded that the Ni element has the highest Gibbs free energy change among the selected elements during the reaction of oxidation and chlorination. This result means that Ni has excellent high-temperature chloride corrosion resistance and could be chosen as the main composition of the matrix alloy. From Figures 5, 6, it could be analyzed that Al₂O₃ was the corrosion product with the most stable properties under the oxidizing atmosphere containing chlorine. The formation of Al₂O₃ products can effectively prevent the further corrosion of chlorine to matrix elements.

There was no Al element in the 904L sample, which had the highest corrosion resistance under the test environment. At the same time, 904L also had the highest Fe concentration, whose chlorination product FeCl₂ had very high equilibrium vapor pressure. As a result, when the corroded temperature was higher than 700°C, 904L showed significant mass loss and showed the poorest high-temperature chloride corrosion resistance.

From the temperature dependence of Gibbs free energy of metal oxidation (Figure 4), it could be seen that oxidations of Al, Nb, and Mo elements have low Gibbs free energy change during the oxidation reaction, so they would be oxidized easily, and the process will cause mass gain of the sample. The AlFeNiMoNb plate had higher concentrations of Al, Nb, and Mo than that of 904L:AlFeNiMoNb high-entropy alloy. As a result, the AlFeNiMoNb plate should obtain a higher mass gain than that of 904L:AlFeNiMoNb high-entropy alloy. In fact, when the corrosion temperature was 700°C, experimental result was consistent with the above analysis, but when the corrosion temperature was increased to 800°C, the AlFeNiMoNb plate showed mass loss. This phenomenon was caused by the fact that at a higher temperature, the volatilization of metal chlorides with high vapor pressure was dominant, especially for the AlFeNiMoNb plate, which had numerous phase boundaries. The phase boundary may still be the main position of discontinuity during the formation of corrosion layer. Therefore, a large number of phase boundaries led to the existence of many defects in the corrosion layer and became fast channels for chloride volatilization. Finally, the weight loss of the material appeared in the process of high-temperature chloride corrosion.

The 904L:AlFeNiMoNb high-entropy alloy had a rather homogeneous and refined microstructure (Figure 2C). This structure was conducive to the uniform corrosion process and the formation of continuous protective oxide film. Therefore, the high-entropy alloy showed the best corrosion resistance under experimental conditions.

Figure 9 shows the x-ray diffraction patterns of the corroded surface of alloys after high-temperature chloride corrosion at 700°C and 800°C for 55 h. Figure 10 shows the surface morphology and composition of alloys after high-temperature chloride corrosion at 700°C and 800°C for 55 h.

When the corrosion temperature was 700°C, the surface of the 904L sample was mainly composed of Cr₂O₃ (PDF#38-1479), Cr_1.3Fe_0.7O₃ (PDF#35-1112), and Fe₃O₄ (PDF#28-0491). After corrosion, the surface of the AlFeNiMoNb sample was mainly composed of Al₂O₃ (PDF#11-0517), Nb₂O₅ (PDF#12-0104), NiO (PDF#89-7390), NiAl₂O₄ (PDF#10-0339), MoO_x (PDF#21-0569, PDF#32-0671), and Fe₃O₄ (PDF#26-1136). For the 904L:AlFeNiMoNb high-entropy alloy, the surface after corrosion was mainly composed of Al₂O₃ (PDF#10-0173, PDF#13-1373), (Cr,Fe)₂O₃ (PDF#02-1357), Cr₂O₃ (PDF#38-1479), and Fe₃O₄ (PDF#26-1136). Compared with the AlFeNiMoNb plate, there was little corrosion products containing Ni on the surface of the 904L:AlFeNiMoNb high-entropy alloy. The results can be analyzed by the difference in microstructure between the two alloys. As shown in the metallographic structure picture of alloys (Figure 2), different phases of AlFeNiMoNb alloy were clearly distinguished (Figure 2B), and each phase occupied a certain area. Ni was mainly concentrated in dark and gray areas, and existed in the form of AlNi and MoNi. After the corrosion of the sample, the corrosion products of Ni elements in these areas, such as NiO, can be detected. Although the 904L:AlFeNiMoNb high-entropy alloy had a similar Ni concentration as the 904L, it has a homogenous microstructure (Figure 2C), and the dispersion of Ni was more uniform. It can be seen from the calculation results of Gibbs free energy change of element oxidation and chlorination reaction in Figure 4 that the oxidation and chlorination reaction of Ni has a smaller driving force than that of other elements, that is, during the corrosion process, Al, Cr, and Fe elements were oxidized quickly than Ni, and at last, Al₂O₃, (Cr,Fe)₂O₃, Cr₂O_3, and Fe₃O₄ products covered the corrosion surface of the sample. On the other hand, from the actual chemical composition of the tested samples, which are shown in Table 1, it could be seen that there was 11.63 at % Cr in the 904L:AlFeNiMoNb high-entropy alloy, which was absent in the AlFeNiMoNb alloy. Figure 4 shows that the driving force for the oxidation reaction of Cr is much higher than that of Ni, so Cr₂O₃ is more likely to be formed on the surface of the sample from the perspective of thermodynamics. Meanwhile, the equilibrium vapor pressure of CrCl₃ (Figure 11) is higher than that of NiCl₂, and the partial pressures of oxygen required for the oxidation reaction of CrCl₃ to Cr₂O₃ (Figure 12) are lower than that of NiCl₂ at the temperatures of 700°C and 800°C, which indicated that the Cr element in the alloy matrix could diffuse outward easily than Ni and eventually formed Cr₂O₃ on the surface of the sample, as shown in Figures 9, 10. As a result, Ni is forming oxides on the surface of AlFeNiMoNb only, but there is no sign of NiO in 904L:AlFeNiMoNb.

When the corrosion temperature was raised to 800°C, the content of Fe₃O₄ in the corrosion product of 904L increased, while the content of Cr₂O₃ decreased. At the same time, the diffraction peaks of NiO appeared. Cr could be oxidized easily to form Cr₂O₃ than Fe to form oxide of ion (Figure 4A), and Cr₂O₃ was more stable than iron oxide under the corrosion condition (Figure 6). As a result, when the corrosion temperature was 700°C, there were more Cr₂O₃ on the surface of 904L after corrosion. At the same time, from the equilibrium vapor pressures of solid chlorides at 400–800°C (Figure 11), it could be seen that the equilibrium vapor pressure of CrCl₃ was slightly higher than that of FeCl₂ at 700°C. From the oxygen pressure required for the oxidation of metal chlorides to metal oxides (Figure 12), it can be seen that the oxygen partial pressure required for the oxidation of CrCl₃ is significantly lower than that of FeCl₂ at 700–800°C. Therefore, the rapidly evaporated CrCl₃ was quickly oxidized to Cr₂O₃ on the surface, resulting in a higher content of chromium oxide on the 904L surface than iron oxide at 700°C. When the corrosion temperature was raised, the corrosion of the alloy became serious, and the thickness of the corrosion layer increased. As can be seen in Figure 12, the oxygen partial pressure required for the oxidation of FeCl₂ to oxide is higher than that for CrCl₃. Therefore, iron oxides tended to be formed on the surface of the corrosion layer closer to the gas phase where the partial pressure of oxygen was higher than the inner side of the corrosion layer. Therefore, the preferentially grown Cr₂O₃ was covered by the subsequently formed iron oxides. The theoretical analysis was in good agreement with the experimental results. On the other hand, the thick corrosion layer with poor adhesion to the substrate was prone to rupture during the cyclic corrosion process with the temperature increasing and decreasing, resulting in the exposure of corrosion products of the inner layer, such as NiO, as shown in the surface morphology and composition of 904L after corroded at 800°C for 55 h in Figure 10.

For the AlFeNiMoNb plate, the iron oxide changed from Fe₃O₄ to Fe₂O₃ with the corrosion temperature increased from 700°C to 800°C. The NiO, Mo oxides, and NiAl₂O₄ in the products were not obvious any more. Instead, the products of NiNb₂O₆ appeared. After corrosion at 700°C, the surface of the sample was composed of oxides of many elements. The driving forces of Al, Nb, and Cr that reacted with oxygen were greater than that of Fe (Figure 4A). As a result, the oxygen partial pressure left for Fe oxidation was small. It can be seen from the predominance diagram for the Fe–O–Cl system in Figure 13 that when the oxygen partial pressure is low, it is conducive to the formation of Fe₃O₄. When the corrosion temperature was raised to 800°C, the corrosion degree increased, and the corrosion layer thickened. The vapor pressure of unstable iron chloride is very high (Figure 11), and the oxygen partial pressure required for the oxidation of iron chloride is high too (Figure 12). Therefore, the surface of the corrosion layer was mainly iron oxide. Because there was no competition for oxidation of other elements at this time, the oxygen partial pressure for Fe oxidation was high, which was conducive to the formation of Fe₂O₃.

As far as the 904L:AlFeNiMoNb high-entropy alloy was concerned, the corrosion products at 800°C were mainly composed of (Cr,Fe)₂O₃, Al₂O₃, Cr₂O_3, and a small amount of Nb₂O₅. Compared with corrosion at 700°C, the content of (Cr,Fe)₂O₃ was significantly improved. The diffraction peak of Fe₃O₄ among the corrosion products was not obvious at the corrosion temperature of 800°C. The corrosion product of the high-entropy alloy had good stability in the high-temperature corrosion environment containing chlorine. Therefore, it effectively prevented the further corrosion of Cl to the alloy matrix and inhibited the formation and volatilization of unstable iron chloride. Therefore, no iron oxide was detected on the alloy surface after corrosion.

Figure 13 shows the thermodynamic stability diagrams of the systems of Fe–O–Cl, Al–O–Cl, Ni–O–Cl, Mo–O–Cl, Nb–O–Cl, and Cr–O–Cl at 800°C, which were calculated by HSC6.0. These results can help researchers to predict corrosion products and corrosion process. The partial pressure of O₂ and HCl in this study was 0.01315 and 0.0027, respectively, which had been marked with “*” in the pictures. Under these conditions, the oxide phases of the alloy elements could be predicted, which were mostly consistent with the XRD and SEM/EDS analysis as shown in Figures 9, 10.

From the surface morphology and composition of alloys after high-temperature chloride corrosion at 700°C and 800°C for 55 h (Figure 10), it can be seen that, after corrosion at 700°C, the surface of the 904L sample consisted of two kinds of regions, one of which had a loose morphology, and the other had a flaky and cracked morphology. Combined with the XRD, SEM/EDS analysis, and the thermodynamic stability diagrams of the systems (Figure 13), it could be concluded that the former was composed of Cr₂O₃ and Fe₃O₄, and the latter was composed of Cr₂O₃ and Cr_1.3Fe_0.7O₃. Under the corrosion temperature of 700°C, the surface of the corroded AlFeNiMoNb plate showed gray, dark, and white areas. The gray area was composed of FeMoO, Nb₂O₅, MoO_x, and Al₂O₃. The dark area consisted of NiAl₂O₄, NiO, and a little Fe₃O₄. MoO_x formed the principal phase of the white area. Because the driving force of Nb oxidation reaction is large, the formation of Nb₂O₅ is easier. At the same time, Nb₂O₅ has a high PBR value (2.68), and the volume of oxide is much larger than that of metal. The rapidly formed Nb₂O₅ cannot be completely spread on the surface but existed in the shape of flowering bread. For 904L:AlFeNiMoNb high-entropy alloy, the morphology of the corrosion layer after chloride corrosion at 700°C was similar to its metallographic structure. The gray area was composed of oxides of elements between dendrites, such as Al₂O₃, Cr₂O₃, (Cr,Fe)₂O₃, and Fe₃O₄. The dark area was mainly Al₂O₃, with a small amount of Cr₂O₃ and (Cr,Fe)₂O₃.

When the corrosion temperature was 800°C, the corrosion products on the surface of the 904L alloy showed obvious delamination. The outer layer was more compact, which was composed of Fe₃O₄. Deep cracks appeared in this outer layer. This cracking phenomenon led to the spallation of the corrosion products, thus, exposing the loose inner layer, which was composed of Cr₂O₃, Fe₃O_4, and NiO. For the AlFeNiMoNb plate, the thickness of the corrosion layer obtained after being corroded at 800°C was obviously larger than that at 700°C. The corrosion products were composed of a small area of dense Al₂O₃, a large area of Fe₂O₃, and protruding flower-like nodules. The nodules were mainly composed of Nb containing oxides such as NiNb₂O₆ and Nb₂O₅. Fe₂O₃ was also detected in this kind of region. When the corrosion temperature increased from 700°C to 800°C, the morphology of corrosion products on the surface of 904L:AlFeNiMoNb high-entropy alloy changed lightly, but the amount of corrosion products increased. As far as the composition was concerned, the content of Al₂O₃ in dendrites increased with the increase in temperature, while the content of Cr₂O₃ increased in inter-dendrites.

Figure 14 shows the cross-section morphology and composition of the samples corroded under N₂–2.6 vol.% CO₂–1.3 vol.% O₂–2,700 vppm HCl gaseous environment at 700°C and 800°C for 55 h. Figures 15, 16 show the element mapping of the cross-section of the corroded samples corroded under N₂–2.6 vol.% CO₂–1.3 vol.% O₂–2,700 vppm HCl gaseous environment at 700°C and 800°C for 55 h.

After chloride corrosion at 700°C, the cross-section morphology of 904L showed that the corrosion products of the sample were mainly divided into three layers. The outermost layer was mainly composed of Cr₂O₃, a small amount of Fe₃O₄, and a trace amount of FeCl₂, which had a loose structure. The middle layer was composed of Cr₂O₃, Fe₃O₄, and NiO. Oxide of Mo appeared in the corrosion product layer near the substrate. Further to the interior of the matrix, there were more serious internal corrosion areas, and the main products were Cr₂O₃ and Fe₃O₄. At the same time, the Cl ion was detected in the internal corrosion zone. The loose oxides in the outer layer were formed by active oxidation in the process of high-temperature chloride corrosion, i.e., chloride ions reacted with metal oxides to obtain metal chlorides, which have a high vapor pressure. Metal chlorides were then oxidized to metal oxides again in the process of volatilization with the increase in the oxygen partial pressure. The oxides obtained by gas phase reaction were relatively loose and unprotected. The driving force of oxidation and chlorination reaction of Ni element was lower than that of other alloy elements. Therefore, the oxide of nickel grew slowly and was mainly located in the inner corrosion layer.

After chloride corrosion at 700°C, the corrosion layer of AlFeNiMoNb sample was thick, and serious internal corrosion occurred. At the same time, more corrosion spots can be found at the phase boundary. The corrosion layer was mainly oxides and chlorine oxides of Fe, Cr, and Ni. When the corrosion layer was damaged, it was easy for the internal diffusion of corrosive element Cl to destroy the corrosion layer of metal oxides and form corresponding metal chlorides, such as CrCl₃ and NiCl₂, as shown in the figure. In the process of the outward diffusion of these chlorides, they were oxidized again to metal oxides with the increase in oxygen pressure. The metal oxides formed in this way had a loose structure, which could be seen in the middle of the corrosion layer shown in the figure. In the inner corrosion area, the main products were metal chlorides and oxides of Fe, Cr, Ni, and Al, in which the content of Fe was the highest.

For the 904L:AlFeNiMoNb sample, when being corroded under 700°C for 55 h, the sample obtained a simple corrosion layer, which was composed of (Fe,Cr,Ni,Al)_xO_y and MCl_x. Although the corrosion layer was loose and broken, there was no inner oxidation found for this sample. The inner surface between the metal matrix and the corrosion layer was very flat, which indicated that the corrosion process of the sample was uniform. This uniform corrosion process was closely related to its uniform composition distribution and microstructure. Sluggish diffusion effect of high-entropy alloy could reduce the diffusion rate of corrosion elements in the 904L:AlFeNi-MoNb alloy, such as Cl and O elements. As a result, sluggish diffusion effect weakened the internal corrosion process, which could be seen in the actual situation that, although no continuous protective oxide layer was found, the Cl element was not detected under the corroded products near the matrix. These results mean that the sluggish diffusion effect of the 904L:AlFeNiMoNb high-entropy alloy plays a positive role in the resistance to high-temperature chloride corrosion.

When the corrosion temperature was raised to 800°C, the corrosion layer on the surface of the 904L alloy was thicker than that at 700°C. The outermost layer was iron oxide and metal chlorides, which had a broken appearance. Adjacent to this layer was a corrosion product region dominated by chromium oxide and containing part of iron oxide and nickel oxide. Continuing inward was an internal corrosion area with a certain depth. The second dark corrosion product layer appeared at half of the thickness of the whole corrosion layer. Combined with the EDS results and thermodynamic analysis, it was mainly the mixture of iron oxide, nickel oxide, molybdenum, and chromium oxide. Further inward, there was another internal corrosion area with nearly the same thickness as the front inner corrosion area. At the interface between the corrosion layer and the alloy matrix, it was a corrosion product layer dominated by chromium oxide and containing part of iron oxide and nickel oxide.

After chloride corrosion at 800°C, the chloride content in the corrosion layer on the surface of AlFeNiMoNb alloy was significantly lower than that at 700°C. The corrosion products were mainly composed of surface Fe₂O₃, Nb₂O₅, intermediate Fe₂O₃, Al₂O₃, and Nb₂O₅, as well as internal Fe₂O₃ and NiNb₂O₆. The corrosion along the phase boundary in the matrix was intensified.

For the 904:AlFeNiMoNb high-entropy alloy, after chloride corrosion at 800°C, the interface between the corrosion layer and the matrix was still relatively flat, indicating that the corrosion process was relatively uniform. Compared with the corrosion products at 700°C, the content of metal chloride oxides in the corrosion layer at 800°C increased, and they had loose microstructure and dark color in the SEM picture. The corrosion products with uniform thickness had poor adhesion to the alloy matrix. During the preparation of cross-section observation samples, the corrosion layer had been stripped from the matrix. On the surface of the substrate, there was a thin layer of corrosion product, which was the mixed oxides of Al₂O₃ and Cr₂O₃. These two oxides provide good shielding and protection for the high-temperature chloride corrosion of the alloy matrix.

The composition and the distribution properties of the corrosion products for 904L, AlFeNiMoNb, and 904L:AlFeNiMoNb high-entropy alloy were consistent with the analysis of the corrosion kinetics and the theoretical corrosion process mentioned above. The component selection and preparation of 904L:AlFeNiMoNb high-entropy alloy had achieved the expected experimental results. The existence of Al and Cr elements were conducive to the improvement of high-temperature chloride corrosion resistance of the alloy due to the formation of oxides that exist stably in chlorine-containing environment. At the same time, the dendrite structure of the uniform and refined high-entropy alloy kept the uniform and slow corrosion characteristics in the corrosion process, and there was no obvious internal corrosion. Thus, the destructive effect of Cl on the alloy matrix and oxide product layer was restrained to a great extent.