Corrosion mechanism of K411 superalloy in sulfur-containing environment: sulfidation promoting internal nitridation

2.1 Specimens preparation

The chemical analyses were carried out with an inductively coupled plasma-optical emission spectroscopy (ICP-OES) and a combustion infrared analysis (for C and S). Table 1 shows with the detail with both nominal and analyzed values. After subjected to the standard heat treatment procedure: 1,120 °C/2 h, AC + 850 °C/24 h, AC (AC: air cooling), the specimens with a size of approximately 20 mm × 10 mm × 1.5 mm were cut by wire EDM. Then the surfaces were grinded with SiC sandpapers up to grit 800#, followed by ultrasonic cleaning in acetone for 15 min. The size and weight of each specimen was subsequently measured for the following weight gain per unit area in corrosion test.

Microstructure of the experimental alloy after standard heat treatment is shown in Figure 1. The average grain size, calculated by treating the grain cross-sectional shape as equivalent to a circle, is 1.26 mm. After the heat treatment, the typical microstructure of the alloy comprises the γ matrix, γ′ phase, carbides, and eutectic structure. The γ′ phase with an approximately cubic shape is in the γ matrix. Carbides distributed within the grains have an equivalent radius ranging from 3.06 to 3.64 μm. The eutectic structure, with an equivalent radius of 16.78 μm, forms between dendrites inside the grains.

Figure 1:

Microstructure of the experimental alloy after heat treatment: (a) grain size, (b) carbides and eutectic, and (c) γ and γ′.

2.2 Corrosion tests

The corrosion tests were conducted by exposing specimens in air-2 vol% SO₂ (99.9 % purity) gas mixture at 900 °C ± 10 °C up to 2000 h. A platinum wire in the gas stream was used to catalyze SO₃ from SO₂ and O₂ with a gas flow rate of 15 mL/min. Sodium hydroxide solution was used for the tail gas treatment (the specific image of the device is shown in the Figure 2).

Figure 2:

Schematic diagram of high temperature corrosion test equipment with SO₂.

To investigate the sulfur effect on corrosion behavior, oxidation tests were also conducted in sulfur-free atmosphere at the same temperature as corrosion tests. Parallel specimens were weighed every 20 h until 100 h, and then every 100 h up to 2000 h (Li et al. 2020). An electronic balance with a resolution of 0.1 mg was used to weigh the mass gain and the corrosion rate. Each kinetic result was the average value of the three parallel specimens.

2.3 Microstructural characterization

Surface morphology characterization and phase composition identification were carried out directly on the specimens after corrosion, which were conducted through scanning electron microscopy (SEM) and X-ray diffraction (XRD). The cross section specimens were embedded in epoxy resin, and a series of grinding and polishing processes using silicon carbide paper and 2.5 μm diamond suspensions was executed. Subsequently, to generate cross-sectional photographs, SEM with an energy dispersive spectroscopy (EDS) and electron probe microanalysis (EPMA) were employed.

3.1 Corrosion kinetics

Figure 3a shows the variation of mass gain per unit surface area versus time for various gas atmospheres at 900 °C up to 2000 h. The results illustrate that the specimens exposed in the sulfur-containing atmosphere exhibited a large final mass gain. During the first 100 h exposure, the specimen underwent a rapid increase in mass gain, whether it was kept in sulfur-containing atmosphere or air. During long-term exposure, the corrosion rate of the specimen decreased substantially. After 2000 h exposure, no spalling of corrosion products can be found on any specimen surface, suggesting the corrosion products were well coherent with the alloy substrate. In addition, the weight gain per unit area for the alloy in the sulfur-containing atmosphere and air were calculated as 2.53 mg/cm² and 2.07 mg/cm², respectively. Obviously, the mass gain within sulfur-containing atmosphere is higher than that in air.

Figure 3:

Kinetic curves of K411 alloy in different environments. (a) Variation of specific mass gain with reaction time at 900 °C; (b) fitting lgΔm−lgt after 100–2000 h at 900 °C.

According to the high temperature oxidation theory, if the oxidation process follows a single oxidation law, the oxidation process of the alloy can be expressed with Equation (1). Equation (2) is taking the logarithm based on Equation (1).

where, Δm is the oxidation weight gain per unit area of the alloy (mg/cm²), n is the number of reaction stages, k_p is the parabolic constant (mg² cm⁻⁴ h⁻¹), and t is the corrosion time (h).

Figure 3b demonstrates the results of a linear fit to the corrosion data. It appears that lgm and lgt adhere to a linear law in tandem, with corrosion in sulfur-containing displaying a reaction level of n = 2.36 and corrosion in sulfur-free representing n = 1.90. The occurrence of n values approaching to 2 suggests that the K411 alloy predominantly follows a parabolic law during both reactions (Cruchley et al. 2013). This indicates that the corrosion in a sulfur-containing atmosphere was controlled by the ion diffusion in the oxide layer.

3.2 Characterization of layers

3.2.1 Cross-sectional characterization

Figure 4 presents cross section comparison of SEM-BSE micrographs suffered corrosion at 900 °C for different times, and EDS analysis was performed to support the experimental findings. As shown in Figure 4, during the corrosion process, selective oxidation of chromium/aluminum in the alloy occurs, leading to rapid oxidation on the surface and the formation of Cr₂O₃/Al₂O₃ in the outermost layer. Initially, the oxidation scale was consisting of a discontinuous chromium oxide layer and an aluminum oxide layer underneath it.

Figure 4:

Cross-sectional morphology of K411 alloy exposed at 900 °C in sulfur-containing atmosphere for (a) 50 h, (b) 100 h, (c) 500 h, (d) 1,000 h, and (e) 2,000 h; in sulfur-free atmosphere for (f) 50 h, (g) 100 h, (h) 500 h, (i) 1,000 h, and (j) 2,000 h.

The numerous tiny needles beneath the inner oxide layer appeared from the 50 h exposure. However, the formation of such products was not observed in the alloy exposed in sulfur-free atmosphere. As the exposure extended, the thickness of oxide layer increased, dendritic Al₂O₃ formed on the substrate and expanded perpendicularly to the substrate. Within 100 h, Cr₂O₃ were found in the outer layer, but the particles were not forming a complete oxide layer.

With exposure continued over 500 h, the particles coalesced to form needles, and EDS spectroscopy verified that these needle-like phases and bulk products below the Al₂O₃ layer contained high concentrations of titanium element and nitrogen element.

Figure 5a and b displays a comparison of the cross-sectional EDS results exposed up to 2000 h. The oxidation scale can be divided into three distinct layers, featuring a continuous Cr-rich oxide layer on the exterior, a discontinuous inner Al-rich oxide layer, and Ta, Ti-rich metallic inclusions between the above oxide layers. Difference in nitride morphology beneath the oxide layer can be observed, with a paucity of block-shaped nitrides were dispersed below the Al₂O₃ layer after oxidation in sulfur-free atmosphere. Conversely, in the sulfur-containing atmosphere, considerable needle-like nitrides were present. Besides, enrichment of elemental sulfur was found beneath the nitride zone by line scanning (Figure 5c).

Figure 5:

Cross-sectional morphologies and EDS composition analysis of K411 alloy in (a) air and (b) air + 2 vol% SO₂ at 900 °C/2,000 h; (c) line scan of the red area.

The sulfides composition was analyzed by EPMA, and the results are shown in Figure 6. There are small quantities of bulk sulfides, enrich in sulfur and titanium, present between the nitride zone and substrate. Both nitrides and sulfides were observed, indicating that not only internal nitridation but also internal sulfidation occurred during corrosion in sulfur-containing atmosphere.

Figure 6:

EPMA scan for the K411 alloy cross section after corrosion at 900 °C for 2,000 h.

Figure 7 shows the cross section morphology comparisons at low magnification exposed up to 2000 h. The oxide scale of the specimens oxidized in sulfur-free atmosphere is flatter compared to the corroded specimens. Conversely, specimens exposed in sulfur-containing atmosphere exhibited numerous protrusions accompanied by several cracks.

Figure 7:

Cross-sectional morphology of K411 alloy after corrosion at 900 °C for 2,000 h: (a) low multiple in air, (b) low multiple in air + 2 vol% SO₂, and (c) high multiple in air + 2 vol% SO₂.

3.2.2 Surface characterization

Figure 8 shows the morphology of K411 alloy corrosion in sulfur-containing and sulfur-free atmospheres for different exposure times, the corrosion product compositions were examined with EDS analysis, and the results are shown in Table 2. The corrosion products exposed in air + 2 vol% SO₂ gas are mostly spherical protrusions. During the first 100 h exposure, aluminum oxide was formed on the surface of the alloy (point 1), while spherical protrusions initiated to emerge on the specimen surface, and these protrusions presented a discontinuous distribution. As the exposure extended, the size and density of spherical oxide protrusions gradually increased, and finally formed a complete and uniform layer. These spherical corrosion products analyzed by EDS (points 2 and 3) showed an abundance of Cr and O, which were identified as Cr₂O₃. It is worth noting that the surfaces of the Cr₂O₃ were subjected to crack in the later reaction period over 500 h, which is more significant in the sulfur-containing atmosphere.

Figure 8:

Surface morphology corrosion in air + 2 vol% SO₂: (a) 50 h, (b) 100 h, (c) 500 h, (d) 1,000 h, and (e) 2,000 h; oxidation in air: (f) 50 h, (g) 100 h, (h) 500 h, (i) 1,000 h, and (j) 2,000 h.

Table 2:

The EDS analysis results of each test points in Figure 8 (at%).

Point	1	2	3	4	5	6
O	49.83	60.26	75.34	40.62	77.84	74.91
Ni	18.72	1.76	0.62	20.79	0.32	0.41
Cr	5.17	28.02	7.53	7.39	20.98	23.69
Ti	2.73	8.03	15.65	1.53	0.78	0.92
Al	7.98	1.92	0.82	29.67	0.09	0.07
S	0.31	0.00	0.05	–	–	–

In the oxidation process, no complete Cr₂O₃ layer formed at the initial stage (within 100 h), which is similar to the corrosion process. The high content of Ni and Al was detected by EDS (point 4), suggesting that a substrate was detected at the beginning of the oxidation and preferentially formed Al₂O₃. Different from the corrosion in sulfur-containing atmosphere, with oxidation prolonged, short rod-shaped oxide particles formed above the surface, and these short rods were identified as TiO₂ (point 5 and point 6). In summary, the oxide layer oxidized in a sulfur-free environment for 2000 h primarily comprises extensive superimposed growth of TiO₂, resulting in a relatively smooth surface.

X-ray diffraction patterns were used to identify the phase composition of the corroded layers. The XRD results of K411 alloy exposed at 900 °C for 2000 h are shown in Figure 9. It can be found that Cr₂O₃ and TiO₂ were main products during corrosion, with small amount of AlTaO₄ and TiTaO₄. The XRD results are consistent with the elemental distribution of the cross section in Figure 4. The spinel phase with a complex crystal structure inhibits the external diffusion, resulting in a reduction of the thickness growth rate of Cr₂O₃ scale (evidenced by Figure 4). This also indicates that the Ta-rich scale plays a role of reducing the corrosion rate, consistent with previous research findings (Han et al. 2015). The diffraction peaks of TiTaO₄ were founded throughout the corrosion process and the substrate peaks were still detectable after 2,000 h. The XRD results indicated that the compositions of the reaction products after corrosion and oxidation were similar. The phase compositions after long time corrosion (more than 1,000 h) were almost the same from those after short time corrosion (within 500 h), and the exposure time mainly affected the relative intensities of the diffraction peaks.

Figure 9:

XRD patterns of the reaction products for different time: (a) in air + 2 vol% SO₂; (b) in air.

4.1 Effects of sulfur-containing atmosphere on microstructure evolution

Kinetic results demonstrate an obvious increase in the weight gain of the K411 alloy in a sulfur-containing atmosphere, which is related to the sulfur penetration through the oxide scale and consequent internal sulfidation. The whole corrosion process can be represented as the following equations: in the gaseous environment, sulfur dioxide was catalyzed at a high temperature to generate sulfur trioxide, and sulfur adsorption occurs on the alloy surface (Equations (3) and (4)). It is well known that internal sulfidation occurs when sulfur in the form of SO₂/SO₃ molecules passes through oxide scales via short-circuit paths such as microcracks and voids, even if the equilibrium partial pressure of sulfur in the gas is lower than the associated sulfide dissociation pressure (Mrowec 1995).

At the initial stage, chromium undergoes selective oxidation at elevated temperatures, resulting in the formation of Cr₂O₃ on the alloy surface. Simultaneously, the dissolution of TiO₂ in Cr₂O₃ leads to the generation of cationic vacancies. These additional cationic vacancies enhance the diffusion rate of titanium ions through chromia. Consequently, the formation of TiO₂ occurs concomitantly with the development of Cr₂O₃ on surface (see Equation (5)). As corrosion continues, lots of fine Cr₂O₃ formed at alloy surfaces and gradually coalesced to form a protective oxide layer. Following the formation of Cr₂O₃ at the alloy surface, the inward diffusion rate of oxygen markedly decreases. The oxygen activity is much lower in the oxide layer than that at the alloy surface. Since the decomposition pressure of Al₂O₃ is significantly lower than that of Cr₂O₃ (at 900 °C, the dissociation pressures for Al₂O₃ and Cr₂O₃ are 9.47 × 10⁻³⁹ atm and 5.4 × 10⁻²⁵ atm, respectively), the selective oxidation of Al₂O₃ can be promoted, leading to the formation of Al₂O₃ at the bottom of the oxide layer. Subsequent oxidation results in localized depletion of Al, slowing down the growth of the Al₂O₃ layer in the alloy. Additionally, the continuous outward diffusion of Cr and Ti forms the outermost Cr₂O₃ layer and the topmost TiO₂ layer. Duval et al. (2010) suggest that a minimum Al content of 5–7 wt% is required to form a continuous Al₂O₃ layer during the oxidation process. While that of the K411 alloy is only 3.5 wt%, which renders the Al₂O₃ layer to form discontinuously. In addition, the generated oxide scale was insufficiently dense to resist the inward diffusion of oxygen element and nitrogen element, the thickness of the oxide scale thus increased persistently.

When the concentration of SO₂ in a corrosive atmosphere is higher than 1 vol%, SO₂/SO₃ could be adsorbed on the alloy, and they react as portrayed by Equation (6) (Andersen and Kofstad. 1995), releasing S₂ into the atmosphere near the surface and concurrently producing the metal oxide. Eventually, the amalgamated effects of Equations (5) and (6) facilitated the forming a continuous oxide layer on the alloy surface.

However, voids and microcracks formed simultaneously due to the different thermal expansion coefficients among the Cr₂O₃ and TiO₂. Such defect provides a direct conduit for gaseous molecules from the corrosive environment to the alloy/oxide intersection. As the S₂ partial pressure is enough, the metal will react with S₂ in the current condition, resulting in the formation of sulfides (Equation (7)). During alternating heating and cooling cycles based on simulated gas turbine start-up and shutdown, there is a mismatch in the coefficient for thermal expansion between the sulfides and the oxides in the oxide scale. The difference in specific volume between sulfides and oxides leads to interfacial stresses in the two phases, which ultimately to microcracks. Subsequent nitrogen element diffuses through these channels led to a rise in the partial pressure of nitrogen, which results in the nitride formation.

As the exposure extended, the partial pressure of sulfur in the sulfide vicinity decreased, and the oxygen partial pressure could concomitantly increase in the area. Some metal oxides and adsorbed sulfur can thus be formed, this process has been described by Equation (8). There is only a minimal distribution of sulfur elements in the oxide-affected zone, which is not surprising considering the sulfides were oxidized and relatively low overall sulfur enrichment at the oxide layer. Metal sulfides are oxidized and then released the adsorbed sulfur, which can re-enter the corrosion process and eventually concentrated below the oxide scale. Combined with the EPMA results, the adsorbed sulfur reacts with the element Ti to form TiS as shown in Equation (9).

In this study, the continuity of the Al₂O₃ layer was improved after corrosion more than 1000 h in sulfur-containing atmosphere. It is generally accepted that the Al₂O₃ layer is more gas-tight than the Cr₂O₃ layer (Qu et al. 2018; Sand et al. 2022). Liu and Gleeson (2013) demonstrated that SO₂ promotes the formation of a continuous α-Al₂O₃ layer on the metal surface. It is considered as the chemisorbed sulfur occupied the active reaction adsorption sites, and oxygen supply thus is partly inhibited; hence, there is a “blocking effect” manifested by surface sulfur enrichment. However, in the present study, the Al₂O₃ layer appeared beneath the Cr₂O₃ layer, and its morphologies showed a dendritic with growing vertically toward the substrate, which are not effective in preventing the internal diffusion of corrosive elements such as sulfur and nitrogen. The results confirm that a trace amount of S can indeed improve the continuity in alumina layer, a similar phenomenon has also been reported previously (Task et al. 2011).

At the alloy/inner oxide interface, where the oxygen partial pressure is relatively low, the SO₂ reduction reaction is preferred to occur, leading to the formation of metal sulfides. The metal sulfides are visible below the oxidation-affected zone, only when specimens kept in the air + 2 vol% SO₂ was implemented exceeding 500 h. It implies that elemental sulfur diffused into the substrate through a repeated sulfidation–oxidation processes (Equations (6)–(8)).

4.2 Internal nitridation behavior

No nitrides were generated during the initial stage of corrosion progress, which is attributed to the thermodynamic stability and corresponding low free energy of the oxides relative to the nitrides, as shown in Figure 10. Analysis revealed that a massive chromium atom from the substrate was driven to the surface to form chromium oxides. One should note that even with a Cr₂O₃ layer on the surface, the alloy still can be nitrided, albeit at a degradation rate, since the Cr₂O₃ layer is permeable to nitrogen (Krupp and Christ 1999a,b). When the K411 alloy was exposed in air after a long period of oxidation, the limited diffusion channels only allow a small amount of nitrogen to reach inside the substrate, which results in little nitride above the substrate (Figure 5a). However, nitrides formation starts earlier when sulfur was presented in the synthesis gas. The result may be account for it that the defect concentration of sulfides is higher than that of oxides (Liu et al. 2015); therefore, the sulfides act as a rapid diffusion pathway for the gas, accelerating the entry of nitrogen into the substrate. Nitrogen partial pressure below the oxide layer consequently increases, which accelerates the formation of internal nitrides.

Figure 10:

Gibbs free energies for the formation of the metal oxides and nitrides fitted by thermodynamic software HSC6.0.

The inward expansion of the oxide and nitride zone with time was observed, and the thickness of the nitride zone was significantly larger than that of the oxide scale after 1,000 h. Figure 11 reveals the average thickness of the oxide scale for corrosion. In sulfur-containing atmosphere, the corrosion products of the experimental alloy manifest in two parts: an oxide scale and a nitride zone. The cumulative thickness of these two parts constitutes the total thickness of corrosion products. In contrast, the oxidation in sulfur-free atmosphere only engenders an oxide scale on the surface. The total thickness of the corrosion products is approximately 2.1 times thicker than that of the oxidation products, which indicates sulfur in the acceleration of the alloy corrosion. In addition, the average thickness of the nitride zone increased from 4.7 µm to 42.1 µm during the experimental tests, which implies that the corrosion of the alloy in the sulfur-containing atmosphere triggered internal nitridation, and the volume fraction of nitride was positively correlated with the reaction time.

Figure 11:

Thickness of each layer during test.

While corrosion in sulfur-containing atmosphere, the extension rate of the nitride zone was greater than that of the oxide scale, as indicated by the results in Figure 11. Some literature (Dong et al. 2001; Rubly and Douglass 1991) described that diffusion coefficient and permeability with nitrogen in γ-Ni are much larger than those of oxygen. The relatively slow diffusion of oxygen, leading to the transverse growth of existing oxides and the formation of successive oxide scale. The diffusion of nitrides was accelerated, allowing new nitride particles to nucleate in deeper regions (Sand et al. 2022). Eventually, nitride zone relatively thicker than oxide layer while corrosion in a sulfur-containing atmosphere.

4.3 Effect of internal nitridation on K411 alloy corrosion

Internal nitridation is detrimental to the alloy, as internal nitrides formed in a sulfur-containing atmosphere is responsible for the development of protrusions on the alloy surface. During the early corrosion stage (refer to the t₁ shown in Figure 12), K411 alloy underwent internal nitridation due to the high nitrogen partial pressure around the substrate. TiN with tiny size and dense dispersion formed beneath the oxidation layer indicates that its nucleation extremely easily. A recent study by Wang et al. (2022) further highlighted the precipitate easily with TiN, even at nitrogen concentrations as low as 0.022 wt%, which is attributable to its low Gibbs free energy. During the subsequent corrosion process (stage t₂ as shown in Figure 12), spherical protrusions formed on the alloy surface and the mechanism are as follows: The molar volume of the experimental alloy matrix is approximately between 6.677 and 6.890 cm³mol⁻¹, while the molar volume of TiN is 11.49 cm³mol⁻¹ (Chang et al. 2001). The increasing molar volume results in significant lattice distortion, leading to compressive stress generating in the nitride region. Where the overall stress on the oxide scale exceeds a critical value, the oxide scale protrudes locally on the surface (Dorcheh et al. 2018). Meanwhile, the establishment of an internal stress gradient leads to metal atoms release from the nitride zone to the surface by outward diffusion at elevated temperatures (Li et al. 2022; Malafaia et al. 2020). At the surface, characterized by, the diffused metal atoms undergo reactions with the oxygen at a high partial pressure. Over time, this process leads to the accumulation of oxides on the material surface. Because of the combination of stress-induced diffusion and the ongoing oxide formation, these accumulated oxides formed as protrusions on the material surface.

Figure 12:

Schematic diagram of the corrosion mechanism of K411 alloy in the air + 2 vol% SO₂ at 900 °C.

Also, part of the internal stress concentrated at the surface scale of the alloy and eventually cracks formed. The volume fraction of generated nitride exhibits a positive correlation with the coverage extent of the protrusions on the surface (Dorcheh et al. 2018). As a result, the densification of spherical protrusions on the alloy surface was observed as well as an increase in the number of cracks with the reaction proceeded.

The compressive stress generated by the nitride caused an increase in the surface cracks, which thereby accelerated the internal diffusion of elemental nitrogen until a thicker area of nitrides formed beneath the oxide layer. The formation of cracks in surface may be the reason why the kinetics cannot reach equilibrium in the post stages of corrosion. Ultimately, oxidation, internal nitridation, and internal sulfidation are responsible for the reduced corrosion resistance of K411 alloy at high temperatures. In summary, the formation of sulfide provides more short-circuit diffusion paths, and it is useful for the synthesis gas entering the interior of the substrate. This greatly increases the partial pressure of the internal nitrogen, leading to severe internal nitridation. Moreover, the nitrides produced cracking during the release of internal stress, which also accelerated the corrosion of the alloy to some extent. Nitrides are considered unfavorable for the alloy due to its deleterious effects on mechanical properties (Krupp et al. 2004). Indeed, the literature warns that the inherent brittleness of nitrides can lead to γ′ dissolution and weakens the surrounding region, thereby promoting the development of internal cracks. Furthermore, the formation of TiN consumes a large amount of Ti elements, so that interfering with the formation of γ′ (Lakhtin et al. 1984; Savva et al. 1996). We have to point out that the present study was carried out with stress-free condition, and further studies will investigate the performance of the alloys under a sulfur-containing atmosphere coupled with stress.