Effects of Nb on Creep Properties and Hot Corrosion Resistance of New Alumina-Forming Austenitic Steels at 700 °C

Abstract

Effects of Nb on the creep resistance and hot corrosion behavior of the Fe-25Cr-35Ni-2.5Al-xNb (x = 0, 0.6, 1.2) Alumina-Forming Austenitic stainless steels (AFA steels) at 700 °C were investigated. The addition of Nb promoted the precipitation of both nanoscale NbC and γ′-Ni₃(Al, Nb) phases, which exhibited very low coarsening rate constants. The nanoscale NbC and γ′-Ni₃(Al, Nb) phases effectively impeded the migration of dislocations and led to an improvement in creep performance of the Nb-addition AFA steel. The corrosion of AFA steels in Na₂SO₄-25%K₂SO₄ at 700 °C was primarily driven by an “oxidation-sulfidation” mechanism. The addition of Nb, serving as a third element, facilitated the formation of protective Cr₂O₃ and Al₂O₃ films, which improved the hot corrosion resistance performance. However, the formation Nb₂O₅ was found to compromise the compactness of the oxide film, which adversely affected the corrosion resistance.

1. Introduction

To address the challenges posed by energy consumption and environmental pollution, countries are pursuing the development of coal-fired power units that can operate at higher temperatures and steam pressures to improve power generation efficiency [1,2,3]. The development of high-temperature resistant materials for coal-fired power units is one of the critical factors [4,5,6]. The formation of protective Cr₂O₃ causes the excellent oxidation resistance of the conventional austenitic heat-resistant steels; however, Cr₂O₃ is usually unstable in the presence of water vapor or corrosive environments [7]. Aluminum-forming austenitic stainless steels (AFA steels), which can form more stable Al₂O₃ on the surface, resulting in better oxidation resistance performance, have become one of the main directions in the development of austenitic heat-resistant steels [8,9,10,11,12].

AFA steel is a typical Nb-containing precipitation strengthened heat resistant steel [13]. The nanoscale NbC precipitated in matrix offers several advantageous characteristics, including small particle size, uniform distribution, and slow growth rate, which has an important influence on the creep properties of AFA steel [14]. During the aging test, a large number of nanoscale NbC particles, pinned dislocations, were observed in the aged organization, which showed good high temperature creep properties [15]. Cold-working process is proved to enhance the creep life of the AFA steel by promoting the precipitation of NbC particles [15,16,17,18]. In AFA steels, the Laves phase (Fe₂Nb) is also one of the strengthening phases. However, the Laves phase grows faster during creep compared to NbC, leading to a lesser contribution to the creep performance, and controlling the Nb/C ratio is beneficial to the creep properties [16,19,20]. Due to the high content of Cr, Ni, and Al in AFA steel, Cr₂₃C₆, NiAl, and σ phases are also common precipitated phases, which can affect the creep properties of AFA steel [14,15,21,22]. The effect of Al, C, Cr, Nb, and Ni elements on the creep properties of AFA steels were studied and found that NiAl and Laves phases exhibited a slight effect on the creep properties of AFA steels compared to MC and M₂₃C₆ [7,16]. Precipitation of the B2-NiAl phase was reported to increase the strength at room temperature, but this strengthening effect was compromised when the temperature was increased to 750 °C due to the premature ductile-to-brittle transition temperature (DBTT) [23,24]. In Al-containing Ni-based superalloys, γ′-Ni₃Al phase is a common strengthening phase [25,26,27,28,29]. However, when the Ni content is increased to 32 wt.%, precipitation of γ′ is observed in AFA steels, which contributes to the creep life [30,31,32]. Investigations have revealed that the strengthening effect of the γ′ phase is second only to NbC, thus the study on the precipitation behavior of γ′ phase in AFA steels is attracting more and more attention [32].

Because of the formation of a stable Al₂O₃ layer on the surface at elevated temperatures, AFA steel exhibits excellent oxidation resistance [33,34]. However, in coal-fired power plants, boiler tubes mainly face sulfate corrosion on the fireside, which often leads to failures [2,3,35]. Therefore, it is necessary to investigate the resistance of AFA steel to sulfate corrosion. The corrosion behavior of 18Cr-25Ni-3Al-1.5Nb steel in 900 °C Na₂SO₄ was investigated and the initial formation of Al₂O₃ was found to facilitate the subsequent formation of Cr₂O₃ [36]. This bilayered protective oxide film effectively inhibited S penetration, resulting in a corrosion resistance that surpassed even that of K438 and K417 alloys. Furthermore, pertinent investigations have been conducted on the role of Nb in high-temperature hot corrosion. During an exploration of the hot corrosion behavior of nickel-based alloys in 900 °C Na₂SO₄ and Na₂SO₄-NaCl, it is observed that Nb promoted the formation of Cr₂O₃ film and retarded the internal oxidation of Al [37]. Furthermore, Nb was found to enhance the adhesion of the oxide film, and greatly improved the corrosion resistance of the Na₂SO₄ molten salt [37]. However, some researchers have found that the formation of Nb₂O₅ may not be beneficial for the heat resistance of Nb addition alloys [38,39]. Moreover, the effect of Nb on heat resistant corrosion is closely related to the chemical composition of the matrix.

Currently, the creep and hot corrosion properties of Aluminum-forming austenitic stainless steels (AFA steels) have been preliminarily investigated, but these studies have mainly focused on the low-Cr and low-Ni AFA steels [40,41]. To enhance the hot corrosion resistance of AFA steels under extreme conditions, the Cr and Ni contents in the matrix must be increased. Variations in Cr and Ni contents may lead to changes in precipitated phases during creep, thus affecting creep performance [8,42]. Furthermore, the corrosion resistance of high-Cr and high-Ni AFA steels requires further investigation. In this work, the creep resistance and hot corrosion resistance of high-Cr and high-Ni AFA steels in Na₂SO₄-25%K₂SO₄ environments were analyzed, and the role of Nb in the creep and hot corrosion processes were also investigated.

2. Experimental

In this study, a series of AFA steels were prepared: AFA-0Nb steel with a composition of Fe-24.92Cr-35.1Ni-2.45Al-0.06C-0.007Nb (wt.%), AFA-0.6Nb steel with a composition of Fe-25.31Cr-35.4Ni-2.47Al-0.06C-0.61Nb (wt.%), and AFA-1.2Nb steel with Fe-25.32Cr-35.02Ni-2.39Al-0.05C-1.17Nb (wt.%). These three ingots were prepared by a laboratory vacuum induction furnace with the protection of an argon atmosphere. The casting ingots were hot-rolled to 12 mm at 1100 °C, and then the hot rolled steels were solution treated at 1150 °C for 2 h. The creep specimens were machined from the 10% cold rolled plate. The specimens had a dog-bone shape with a diameter of 6 mm and a gauge length of 25 mm. The creep test was performed at 700 °C under a load of 130 MPa using an electronic creep testing machine (CN, NCS, GNCJ-100E, Beijing, China).

Before the hot corrosion test, the samples with a dimension of 15 × 10 × 5 mm were polished with SiC papers from no. 400 to 1500, and then cleaned ultrasonically in alcohol. To order to eliminate the impact of the crucible weight change during the oxidation process, all crucibles were calcined at 1100 °C until the weight stabilized. The hot corrosion experiment was carried out in a muffle furnace at 700 °C for 120 h. The specimens were preheated and coated with a Na₂SO₄-25%K₂SO₄ salt layer, which was controlled within the range of 20–30 mg/cm². During the test, the specimens were taken out at 10 h, 20 h, 40 h, 60 h, 80 h, 100 h, and 120 h. After cooling, the samples were put into boiling water to remove the remaining salt on the surface, and then dried and weighed. The corrosion rate was calculated by the weight gain method and the corrosion kinetics curve was plotted.

The surface and cross-sectional morphology of the corrosion scale were identified using a scanning electron microscope (SEM, Zeiss, Oberkochen, Germany, EVO 18) equipped with Inca energy X-ray dispersive spectroscopy (EDS, Oxford Instruments, Oxford, UK, 20 KV) and back-scattered electron (BSE) detector. The phase structures of the specimens were examined by a transmission electron microscope (TEM, FEI, Tecnai, G2 F30, Hillsboro, OR, USA) with a voltage of 200 KV. The phase present in corrosion products were analyzed by a D/Max-2550 X-ray diffractometer (XRD, Rigaku, Ultima IV, Tokyo, Japan) with Cu Kα radiation with a wavelength of 1.5406 Å and a scanning rate of 2°/min. The composition of corrosion products was also examined by X-ray photoelectron spectroscopy (XPS, K-Alpha, Thermo Fisher Scientific, Waltham, MA, USA). The Gibbs free energy of the reactions were obtained using HSC Chemistry v6.0 (Outotec, Tukkikatu, Finland).

3. Results

3.1. Microstructure

Figure 1 shows SEM photos of the microstructure of the AFA-xNb steels after solution treatment. The microstructure of the AFA-0Nb steel is shown in Figure 1a, and it can be observed that AFA-0Nb steel exhibited a twinned austenite structure. As shown in Figure 1b,c, some coarse granular particles with sizes ranging from 1–3 µm were precipitated in the austenitic matrix. The amount of precipitated particles in AFA-1.2Nb steel was even greater. The components of these particles were confirmed in Figure 1d. They were identified as primary NbC because of the higher content of Nb and C elements. Due to the high melting point, these primary NbC particles cannot completely dissolve in the matrix, which has been proven to have an effect on the refining grain size and improving strength. Figure 2 shows the XRD patterns of AFA-xNb steels after 1150 °C solution treatment. The XRD analysis results indicated that AFA steels with different Nb contents exhibited an austenitic structure after solution treatment. NbC precipitates were not observed, likely due to the presence being below the detection limit.

3.2. Creep Resistance

Figure 3 shows the creep curves of AFA-0Nb and AFA-0.6Nb steels at 700 °C with an initial stress of 130 MPa. As seen in Figure 3a, AFA-0Nb steel exhibited a creep lifetime of 331 h, while the creep lifetime of AFA-0.6Nb steel reached 5687 h, which was much higher than that of AFA-0Nb steel. Figure 3b shows the creep rate curves of AFA-0Nb and AFA-0.6Nb steels at 700 °C with an initial stress of 130 MPa. The creep strain rate curve represents the relationship between the creep strain rate with time, which reveals the deformation characteristics of the material during high-temperature creep process. Generally, the creep process can be divided into three stages: the initial transient stage with a decreasing strain rate, the secondary creep stage with a constant low strain rate, and the tertiary creep stage with a rapidly increasing strain rate. As illustrated in Figure 3b, the creep rate curve of AFA-0Nb steel exhibited distinct decelerated creep and accelerated creep stages, while the steady-state creep stage was not prominent. The minimum creep rate was 2.8 × 10⁻⁴ h⁻¹. However, three regimes can be seen in the creep rate curve of AFA-0.6Nb steel. In the primary creep regime, the creep rate gradually decreased over 200 h, transitioning into the steady creep rate regime until the 5300 h with a minimum creep rate of 4.5 × 10⁻⁶ h⁻¹, which is two orders of magnitude lower than that of AFA-0Nb steel. In the tertiary creep regime, the creep test continued to accelerate until rupture occurred. The result of the creep test indicated that the addition of Nb significantly improved the creep performance of the AFA steels.

Figure 4 shows the microstructure of AFA-0Nb and AFA-0.6Nb steels after the creep test. As shown in Figure 4a, a large number of precipitated phases were observed in the matrix of AFA-0Nb steel after 200 h of creep, including granular and needle-shaped NiAl phase (R1), granular Cr₂₃C₆ (R2), and massive sigma phase (R3), which are mainly distributed along the grain boundaries. The precipitation of Cr₂₃C₆ on the grain boundary is due to the creep temperature of 700 °C, which is in the sensitization temperature range, thus C and Cr in austenitic stainless steel will combine to form Cr₂₃C₆ distributed along the grain boundary. Corresponding EDS component results are shown in

Table 1. EDS point analysis of the areas in Figure 4.

Position	Composition (wt.%)
	C	Cr	Al	Fe	Ni	Nb
R1	0.14	7.29	26.98	12.43	53.16	-
R2	3.35	63.12	0.45	18.65	14.43
R3	1.56	46.39	1.33	43.40	7.32	-
R4	0.32	41.87	0.82	51.27	5.72	-
R5	4.65	59.45	-	12.32	23.58
R6	0.34	13.65	28.97	10.47	46.57
R7	4.17	9.76	3.53	13.94	10.68	57.92

. As illustrated in Figure 4b, except for NiAl, Cr₂₃C₆, and sigma phases, no new phase can be seen in the fractured specimen after 311 h, while the size of the precipitated phases grew larger, and among them the sigma phases were even coarser. Figure 4b displays the precipitates of the AFA-0.6Nb steels after the creep test. It is noted that NiAl (R4), Cr₂₃C₆ (R5), and sigma phases (R6) were also precipitated in the AFA-0.6Nb steel after 200 h of creep, with corresponding EDS component results shown in Table 1. Additionally, the addition of Nb led to the formation of fine NbC particles (R7) during the creep process of AFA-0.6Nb steel; the EDS component results are displayed in Table 1. After creep fracturing, there was an increase in the amount of precipitated phase, with the sizes becoming coarser, except for the NbC particles, as illustrated in Figure 4d. It is worth mentioning that numerous spherical precipitates can be observed in the magnified images of the matrix (Figure 4e), which have been identified as γ′ phase in subsequent studies. At the end of the creep process, the γ′ phases retained their fine structure, with an average size of approximately 100 nm, as depicted in Figure 4f.

To identify the precipitates of AFA-0Nb and AFA-0.6Nb steels, TEM morphology and selected area electron diffraction (SAED) patterns of the steels were analyzed. Figure 5 shows the TEM morphology of the precipitated phase in AFA-0Nb steel after high temperature creep test and corresponding diffraction spots. Several bar-shaped particles with a size of around 250 nm can be observed in AFA-0Nb steel, which were identified as Cr₂₃C₆ by the selected area electron diffraction (SAED) pattern in Figure 5b. Additionally, there were also two other massive precipitates with larger sizes, which were proved to be NiAl and sigma phases, respectively.

Figure 6 shows the TEM morphology of the precipitated phase in AFA-0.6Nb steel after creep test and corresponding diffraction spots. As illustrated in Figure 6, AFA-0.6Nb steel exhibited the precipitation of Cr₂₃C₆, NiAl, and sigma phases with dimensions exceeding 500 nm, as well as the formation of secondary NbC particles with diameters less than 100 nm, which were diffusely distributed in the matrix. Figure 6c is a dark field photo of AFA-0.6Nb steel. A large number of dispersed spherical particles with an average size of 106 nm were found, and distribution images of elements showed that the spherical particles mainly contained Ni, Al, and a little Nb, which were identified as γ′ phase based on SAED in Figure 6d. The nanoscale γ′ and NbC precipitates can inhibit the dislocation migration during creep deformation, resulting in the better creep performance of AFA-0.6Nb steel. To investigate the elemental distribution of the γ′ phase, line scanning and EDS composition analysis were conducted, and the results are shown in Figure 7. In addition to Ni and Al being higher than the matrix, Nb was also enriched in the spherical γ′ phase. As shown in Figure 7c, the content of Nb in the γ′ phase was 1.77 at.%, while the content of Nb in the matrix was 0.03 at.%, indicating that Nb tended to enrich in the γ′ phase and acted as a substitute for Al atoms, which is conducive to the precipitation of the γ′-Ni₃ (Al, Nb) phase in the low Al alloy [43,44].

The main precipitated phases of AFA-0Nb steel are Cr₂₃C₆ and NiAl phases, while the precipitation of these two phases exhibited a certain interaction during the high temperature creep process. Figure 8 displays the TEM and line scanning images of Cr₂₃C₆ and NiAl phases in AFA-0Nb steel. As seen in Figure 8a,b, the Cr₂₃C₆ and NiAl phases precipitated adjacent to each other and were tightly bound together to form a composite elliptical precipitation phase. Line scanning analysis of the precipitated phase revealed that the bright gray part is the NiAl phase and the dark gray part is Cr₂₃C₆. This is because, in the creep process, the preferential precipitation of Cr₂₃C₆ causes the Cr poverty in the surrounding matrix. This results in the relative enrichment of Ni and Al elements around Cr₂₃C₆, which promotes the generation of the NiAl phase [13].

3.3. Hot Corrosion Resistance

Figure 9 shows the corrosion kinetics curves of AFA-xNb steels after exposure to a mixed salt of Na₂SO₄ and K₂SO₄ at 700 °C for 120 h. AFA-xNb steels rapidly gained weight in the initial 10 h corrosion stage, and gradually leveled off as time increased. The corrosion rate of the AFA-0Nb steel was 0.57 mg/cm², and the corresponding surface exhibited a dense dark gray oxide film, with some loose spalling in local areas. The corrosion gain rate of AFA-0.6Nb steel was 0.38 mg/cm², which was significantly reduced. The corresponding surface of AFA-0.6Nb steel was flat and dense, without any obvious looseness and peeling. However, when the content of Nb was increased to 1.2%, the corrosion rate of AFA-1.2Nb steel slightly increased to 0.44 mg/cm², and there was significant spalling on the corrosion surface. Corrosion test results showed that the addition of Nb can improve the corrosion resistance of AFA steel to Na₂SO₄-25%K₂SO₄ mixed salt, but the Nb content must be controlled within an appropriate range.

Figure 10 shows the surface film morphology of AFA-xNb steel after corrosion at 700 °C Na₂SO₄-25%K₂SO₄ mixed salt for 120 h. As shown in Figure 10a,c,e, the surface of three AFA-xNb steels after corrosion were relatively flat, in which the surface of AFA-0Nb steel was coarser, while the corrosion surface of AFA-0.6Nb steel was fine and uniform. Figure 10b,d,f indicated that all of the three AFA steels were composed of granular corrosion products. The surface of AFA-0Nb steel exhibited poor uniformity in particle size, with agglomerations and looser particles, while the AFA-0.6Nb steel surface was uniformly fine and dense, with no obvious agglomerations, as shown in Figure 10b,d. Figure 10f revealed that the oxidized film on the surface of the AFA-1.2Nb steel was also uniform, but there were slight agglomeration and cracks, which may affect the corrosion performance. The analysis results of the corrosion morphology demonstrated that the AFA-0.6Nb steel exhibited the best corrosion resistance, which was consistent with the corrosion kinetics curve in Figure 9. Figure 11 presents the XRD spectrum of the corrosion layer on the surface of AFA-xNb steels after 120 h of hot corrosion in a mixed salt solution at 700 °C. The surface corrosion layer of AFA-xNb steels was mainly composed of Cr₂O₃, Fe₂O₃, Al₂O₃, and NiCr₂O₄, as well as a small amount of sulfide products of Fe and Ni.

Figure 12 shows the cross-section morphology of the corrosion layer. The thicknesses of the surface corrosion layers were 5.1 μm, 2.0 μm, and 3.1 μm. The thickness decreased and then increased slightly, and the oxide layer of AFA-0.6Nb steel was the thinnest, which was in agreement with the corrosion kinetic curve. Figure 12a illustrates that the corrosion layer of AFA-0Nb steel was relatively thick and loose, containing voids and pores. At the interface, some point-like and strip-like sulfides appeared in the matrix. The thickness of the corrosion layer of the AFA-0.6Nb steel, showing a dense and flat corrosion layer, was significantly reduced (Figure 12b). The corrosion layer of AFA-1.2Nb steel, as shown in Figure 12c, exhibits voids and cracks, accompanied by the appearance of Nb oxide particles within the corrosion layer. These particles may contribute to the reduction in the protective properties of the corrosion layer. Figure 12c shows the presence of holes and cracks in the corrosion layer of AFA-1.2Nb steel, as well as the appearance of grayish-white particles. EDS analysis indicated that these particles contain high levels of O and Nb, which were oxides of Nb. The XPS results confirmed that the oxide film of AFA-1.2Nb steel contained Nb⁵⁺, which proved the presence of Nb₂O₅ in the oxide layer (Figure 12d). The formation of Nb₂O₅ may cause the microcracks in the oxide film and affect the density of the oxide film [38].

Figure 13 presents the cross-sectional elemental distribution of the corrosion layer. The surface corrosion layers of the AFA steels mainly consisted of Cr-rich oxide, with a continuous and dense Al₂O₃ layer presented at the inner side of the corrosion layer. The continuous and slow-growing Al₂O₃ layer can impede the diffusion of elements, thus benefiting the corrosion resistance of AFA steels. Simultaneously, a distinct Cr-poor layer can be observed in the matrix of all three AFA steels, with thicknesses of approximately 8.3 μm, 6.0 μm, and 6.6 μm, which decreased first and then increased with the increase of Nb. This is attributed to the selective corrosion of Cr, and the corrosion depth is an important indicator for evaluating the corrosion resistance. Based on the distribution photograph of S elements, the sulfidation depth of AFA-0Nb steel was relatively deeper, while sulfidation primarily occurred in the corrosion layer with the Nb addition. This suggested that Nb addition effectively prevented the progression of sulfidation corrosion on the surface of AFA steels.

4. Discussion

4.1. Effect of Nb on Creep Resistance

The high-temperature creep mechanism of heat-resistant steel is believed to be influenced by factors such as the morphology, distribution, and high-temperature stability of the precipitated phase. It is suggested that the precipitated phase may coarsen during high-temperature creep, potentially leading to a decline in creep performance. Therefore, the study on the creep properties of AFA steel must investigate the types and coarsening trend of the precipitation phases in the creep process. The main precipitated phases in the creep process of AFA-0Nb were NiAl phase and Cr₂₃C₆, while the main precipitated phases in the creep process of AFA-0.6Nb were NiAl phase, Cr₂₃C₆, NbC, and γ′-Ni₃(Al, Nb) phase. Previous studies have shown that the fine NiAl phase precipitated during the early stage of aging can impede dislocation migration and enhance strength at room temperature. However, Chen and Bei et al. have demonstrated that the ductile-to-brittle transition temperature of NiAl phase is 500~800 °C [23,45]. Since the creep test temperature in this experiment is 700 °C, the strengthening effect of NiAl phase is not significant [16]. Therefore, the strengthening phase of AFA-0Nb was Cr₂₃C₆, and the strengthening phase of AFA-0.6Nb steel were NbC and γ′-Ni₃(Al, Nb) phase during the creep process. The influence of Cr₂₃C₆ on creep properties is related to its size. At the initial stage of creep, a large amount of Cr₂₃C₆ is precipitated with a small size, which can effectively hinder the movement of the dislocation. However, with the increase of creep time, Cr₂₃C₆ grows rapidly and the strengthening effect weakens, thus the contribution to creep performance is small.

The AFA-0.6Nb steel exhibited exceptional creep properties, primarily due to the reinforcing effects of NbC and γ′-Ni₃(Al, Nb) phases. The coarsening behavior of the precipitated phases during creep plays a significant role in determining the creep performance [25]. Although the coarsening behavior of γ′ in Ni-based alloys has been studied, the coarsening behavior in AFA steel remains to be investigated. Theoretically, the roughening of the precipitated phase follows the matrix diffusion-controlled coarsening behavior, which is explained by the classical Lifshitz–Slyozov–Wagner (LSW) coarsening model [14,15]. According to the LSW theory, the coarsening kinetics of nanoscale precipitated phases during the creep process can be expressed as follows [46]:

$$\mathrm{d}{(\mathrm{t})}^3-\mathrm{d}{(0)}^3=\mathrm{K}\mathrm{t}$$

(1)

where d(t) is the average particle size at aging time t, d(0) is the average particle size in the initial stage, and K is the coarsening rate constant. Equation (1) can be used to analyze the coarsening behavior of NbC and γ′-Ni₃(Al, Nb) phases in AFA-0.6Nb steel. Figure 14 shows the average size of the NbC and γ′-Ni₃(Al, Nb) phases as a function of the aging time, and the slopes of the fitting lines can be used as the coarsening rate constant k. As shown in Figure 14, the cubicity of the average size of the NbC and γ′-Ni₃(Al, Nb) phases was approximately linear with the aging time, indicating that the coarsening of the two precipitates was mainly controlled by the volume diffusion of the forming elements (Nb, Ni, and Al). The coarsening rate constant k_γ′ of the γ′ phase in AFA-0.6Nb steel was fitted to be 6.06 × 10⁻²⁹ m³/s, while K_NbC was fitted to be 1.25 × 10⁻²⁹ m³/s. The two precipitated phases in AFA-0.6Nb steel exhibited an exceptionally low coarsening rate constant, indicating that they are ideal strengthening phases, which is the main reason for the better creep performance of AFA-0.6Nb steel [14,47].

It is known that NbC has a very strong resistance to coarsening rate during high-temperature creep, which is mainly related to diffusion and interfacial effects [15,48]. In the coarsening process controlled by diffusion, the coarsening kinetics are often controlled by the element with the minimal diffusion coefficient. However, the diffusion of Nb in the austenite matrix is slower than that of C, making Nb the key factor in the coarsening process. On one hand, the content of Nb in AFA steel was relatively low, and the precipitation of primary NbC consumed a portion of Nb. Moreover, during the initial stage of creep, a large number of dispersed and tiny secondary NbC nucleated at dislocation sites due to the effect of cold deformation, consuming a considerable amount of Nb. As a result, the remaining Nb content was even lower. On the other hand, the Nb-poor region formed around the NbC due to the precipitation of NbC in the matrix, as illustrated in Figure 15. The formation of the Nb-poor region around the Nb resulted in a lack of supply of Nb needed for NbC coarsening, leading to a significantly low coarsening rate of NbC, which exhibited a high thermal stability. Furthermore, NbC exhibits a NaCl-type face-centered cubic structure, which typically has a cubic-cubic orientation relationship with austenite and a semi-coherent relationship with the matrix [49], resulting in a relatively low interfacial energy [45], thereby contributing to the reduced coarsening rate of NbC.

The high-temperature flow behavior of metals is usually described by the Norton equation [50], as seen in Equation (2).

$$\dot{\mathsf{\epsilon }}=\mathrm{A}\left({\displaystyle \frac{\mathrm{G}\mathrm{b}{\mathrm{D}}_{\mathrm{L}}}{\mathrm{k}\mathrm{T}}}\right){\left({\displaystyle \frac{{\mathsf{\sigma }}_{\mathrm{S}\mathrm{S}}}{\mathrm{G}}}\right)}^{\mathrm{n}}$$

(2)

where $\dot{\mathsf{\epsilon }}$ is the strain rate, n is the stress exponent, k is the Boltzmann constant, b is the Burgers vector, G is the shear modulus, the D_L is the diffusion coefficient, T is the absolute temperature, σ_ss is the steady-state stress value, and A is a constant related to material properties and temperature. Precipitated phases in alloys lead to a “threshold force” during deformation at high temperatures. During creep, the threshold force refers to the stress required for dislocations to bypass the precipitated phase by antiphase boundary shearing, stacking fault shearing, and orowan bypassing [26]. During creep, the precipitated phases of AFA-0Nb and AFA-0.6Nb steels exhibited different strengthening effects that affect the creep mechanism of AFA steels. The creep mechanism of AFA steels can be inferred from the apparent stress index, thus the strain rate and steady state stress of AFA-0Nb and AFA-0.6Nb steels after aging for 300 h were plotted and fitted to the logarithmic function curves, as shown in Figure 16. It can be concluded that the apparent stress indices of the AFA-0Nb and AFA-0.6Nb steels were 5.8 and 9.7, respectively. However, within the range of 3 to 6 for the apparent stress index, the creep process is controlled by dislocations [15,51]. It can be inferred that during the creep process, the rapid growth of the precipitation phases in AFA-0Nb steel failed to impede the movement of dislocations effectively, resulting in deteriorated creep performance. In contrast, AFA-0.6Nb steel still belongs to the precipitated phase-controlled creep deformation process, which exhibited a satisfactory hindering effect on dislocation movement.

Figure 17 shows the dislocation morphology of AFA-0Nb and AFA-0.6Nb steels after a high temperature creep test. As seen in Figure 17a, the AFA-0Nb steel exhibited coarsening Cr₂₃C₆ particles after the creep test, without any prominent interaction with dislocations, as large-size inclusions cannot effectively hinder the dislocation migration. In Figure 17b, a large number of dispersed fine NbC particles are observed to distributed along the dislocation lines, effectively pinning the dislocations, which play an important role in preventing the movement of dislocations. The interaction mode between dislocations and γ′ is primarily determined by the size of γ′ phase. As the size of γ′ phase reaches 43.4 nm or above, the interaction mechanism between dislocations and γ′ phase becomes the Orowan bypass mechanism [29]. The size of the γ′ phase after the creep test reaches more than 100 nm, thus the dislocation movement tends to adopt the Orowan bypass mechanism through the γ′ phase. Figure 17c shows the morphology of the dislocation line just contacting the γ′ phase and the dislocation ring formed by the dislocation bypassing the γ′ phase, as shown schematically in Figure 17d. Therefore, the diffuse and fine nanoscale NbC and γ′ precipitation phases in AFA-0.6Nb steel can effectively impede the movement of dislocations, which is the main reason for the excellent creep properties of AFA-0.6Nb steel.

4.2. Effect of Nb on Hot Corrosion Resistance

The melting point of the mixed salt of Na₂SO₄-25%K₂SO₄ is 837 °C, thus the corrosion at 700 °C is low temperature hot corrosion. In the initial stage of the hot corrosion test, oxygen in the air passes through the salt film and selectively oxidizes with the pro-oxygen alloying elements in the AFA steel matrix, forming protective oxide films such as A1₂O₃ and Cr₂O₃ on the alloy surface, as seen in Equation (3),

$$\mathrm{x}\mathrm{M}+{\displaystyle \frac{\mathrm{y}}{2}}{\mathrm{O}}_2(\mathrm{g})={\mathrm{M}}_{\mathrm{x}}{\mathrm{O}}_{\mathrm{y}}(\mathrm{M}=\mathrm{A}\mathrm{l},\mathrm{C}\mathrm{r},\mathrm{N}\mathrm{i},\mathrm{Fe} \mathrm{or} \mathrm{N}\mathrm{b})$$

(3)

As shown in

Table 2. Standard Gibbs free energy of various reactions of the AFA steels in mixed sulfate at 700 °C.

Reaction	∆G700°C/kJ/mol
4/3Al + O2 (g) = 2/3Al2O3	−913.3
4/3Cr + O2 (g) = 2/3Cr2O3	−585.4
2Ni + O2 (g) = 2NiO	−301.8
4/3Fe + O2 (g) = 2/3Fe2O3	−377.3
4/5Nb + O2 (g) = 2/5Nb2O5	−589.5
2Al + Na2SO4 = Al2O3 + Na2O + S(g)	−496.9
2Cr + Na2SO4 = Cr2O3 + Na2O + S(g)	−5.01
3Ni + Na2SO4 = 3NiO + Na2O + S(g)	420.4
2Fe + Na2SO4 = Fe2O3 + Na2O + S(g)	307.1
2/3Al + S = 1/3Al2S3	−196.9
Cr + S = CrS	−138.6
1/2Fe +S =1/2 FeS2	−58.4
3/2Ni + S = 1/2Ni3S2	−95.2

, the Gibbs free energies of oxide formation for Al and Cr in the alloy are −913.3 kJ/mol and −585.4 kJ/mol, respectively. These elements are the strong oxide-forming elements and are more thermodynamically prone to oxidation reactions. This leads to the formation of stable A1₂O₃ and Cr₂O₃, which inhibits the oxidation of Ni and Fe. The dense oxidation films of A1₂O₃ and Cr₂O₃ on the surface of the AFA steels can act as a barrier, preventing the reaction between the salt and the base metal during the initial stages of testing. The Gibbs free energy of the oxidation reaction of Nb is −589.5 kJ/mol, indicating its high susceptibility to oxidation. However, this is not conducive to the density of the oxide film, so the content of Nb should not be too high.

The scanning analysis of the cross section of the corrosion layer shows that there were spots of sulfide in the transition between the corrosion layer and the substrate. Due to the absence of a sulfur-containing atmosphere and free S element in the mixed salt, it is postulated that S may be derived from the products of the reaction between the sulfate mixed salt and the matrix. As illustrated in [52], under high-temperature conditions, sulfates can act as oxidants and react with metals to form oxides, concurrently generating sulfur, as demonstrated in Equations (4) and (5).

$$\mathrm{M}+{\mathrm{R}}_2\mathrm{S}{\mathrm{O}}_4\to {\mathrm{M}}_{\mathrm{x}}{\mathrm{O}}_{\mathrm{y}}+{\mathrm{R}}_2\mathrm{O}+\mathrm{S}(\mathrm{g})$$

(4)

$$\mathrm{M}+\mathrm{S}\to \mathrm{M}\mathrm{S}$$

(5)

M is the alloying elements Al, Cr, Fe, and Ni, and R is Na and K. To discuss the possibility of direct oxidation reaction between solid sulfate and substrate, the standard Gibbs free energy of reaction between substrate alloying elements and sulfate is calculated, and the results are shown in Table 2. The Gibbs free energies of the alloying elements in the matrix, such as Al and Cr, show negative values when reacting with sulfate, whereas the Gibbs free energies of Ni and Fe reactions are positive. This indicated that the alloying elements Al and Cr in the matrix can undergo spontaneous reactions with sulfate in the solid phase, confirming the earlier hypothesis regarding the source of sulfur.

However, the free energy of the oxidation reaction of the alloying elements is significantly lower, indicating a stronger driving force for oxidation. Consequently, oxidation reactions typically occurred preferentially at the onset of the experiment, leading to the formation of dense, protective Al₂O₃ and Cr₂O₃ layers. Nonetheless, due to the obstruction by the salt film and the low oxygen partial pressure at the cross-sectional area, solid-state reactions between the salt film and the substrate can simultaneously take place, as the salt film is in direct contact with the metal surface.

Figure 18 shows the corrosion process of low-temperature thermal corrosion of AFA steel in Na₂SO₄-25%K₂SO₄ mixed salt at 700 °C. In the initial stage of low-temperature corrosion, the alloying elements on the surface of the substrate reacted with the oxygen that diffused through the salt film to the interface to form a dense protective Al₂O₃ and Cr₂O₃ oxide film, which inhibited the further diffusion of the alloying elements and oxygen. At the same time, the alloying elements reacted directly with the solid sulfate to form the metal oxide MO and free S element. The free S at the interface diffused inward to the matrix and reacted with the alloying elements inside the matrix to form point sulfides (Al₂S₃ and CrS) in the region where the partial pressure of oxygen in the matrix was sufficiently low (see Equation (5)). Oxygen also diffused inward and reacted with outward-diffusing Cr and Al elements, leading to the continuous thickening of the oxide layer. As the corrosion time progresses during the steady-state stage, the progressively thickened continuous and dense oxide layer suppressed the diffusion of elements, thereby slowing down the oxidation and sulfidation reactions of alloying elements. This behavior under solid sulfate conditions was similar to that of high-temperature oxidation [53].

Hot corrosion tests on AFA-χNb steels show that the addition of Nb played an important role in enhancing the hot corrosion resistance of AFA steel. Studies on the oxidation resistance of Nb-containing steel have revealed that Nb can act as a third element [42], reducing the critical Al content required for the formation of an Al₂O₃ film and promoting the formation of an external Al₂O₃ oxide layer, thereby inhibiting elemental diffusion. Corrosion cross-section analysis indicated that the addition of 0.6% Nb resulted in the thinnest corrosion layer, while the Al₂O₃ layer was thick and dense. This effectively hindered the outward diffusion of alloying elements and the infiltration of corrosive elements. Additionally, research has been shown that Nb can also promote the formation of Cr₂O₃ and improve corrosion resistance [54,55]. Thermodynamic calculations have indicated that the activities of Al, Cr, and Fe increased with the addition of Nb, while the activity of Ni decreased. The addition of Nb increased the activities of Al and Cr [56], allowing for the rapid replenishment of Al and Cr consumed during the corrosion process. The distribution of S showed that S was mainly concentrated at the substrate interface, and the depth of sulfidation was significantly reduced, indicating a marked improvement in corrosion resistance. However, when the Nb content was increased to 1.2%, a significant amount of Nb₂O₅ was formed. The Pilling–Bedworth Ratio (PBR) of Nb₂O₅ is approximately 2.67 [57], which increased the compressive stress within the oxide layer and led to cracks and defects in the oxide film, meaning that the oxide film was unable to effectively inhibit the invasion of corrosion elements and the diffusion of alloying elements [58].

5. Conclusions

In this study, the creep resistance and hot corrosion resistance of the Fe-25Cr-35Ni-2.5Al-xNb (0, 0.6, 1.2 wt.%) AFA steels were investigated. Results show that the addition of Nb has an important effect on creep resistance and hot corrosion resistance. The conclusions can be drawn as follows:

(1)

The precipitates observed in AFA-0Nb steel after creep testing were identified as Cr₂₃C₆, NiAl, and the σ phase. In contrast, the precipitates in AFA-0.6Nb steel after creep testing included NiAl, Cr₂₃C₆, NbC, γ′-Ni₃(Al, Nb), and σ phase. The addition of Nb was found to promote the precipitation of both NbC and γ′-Ni₃(Al, Nb) phases, each exhibiting very low coarsening rate constants, with the value of K_γ′ = 6.06 × 10⁻²⁹ m³/s and K_NbC = 1.25 × 10⁻²⁹ m³/s.

(2)

The influence of precipitates on the creep performance of AFA steel was investigated. The results indicated that the creep life of AFA-0Nb steel was 331 h with a minimum creep rate of 2.8 × 10⁻⁴ h⁻¹. Conversely, AFA-0.6Nb steel demonstrated a significantly extended creep life of 5687 h and a reduced minimum creep rate of 4.5 × 10⁻⁶ h⁻¹. The addition of Nb markedly improved the creep performance of the AFA steels. The apparent stress exponent values for AFA-0Nb and AFA-0.6Nb steels were determined to be 5.8 and 9.7, respectively. The creep process in AFA-0Nb steel was dislocation-controlled, while the AFA-0.6Nb steel was governed by precipitate-controlled deformation. The precipitates NbC and γ′-Ni₃(Al, Nb) precipitates effectively hindered dislocation migration.

(3)

The corrosion of AFA steels in Na₂SO₄-25%K₂SO₄ was characterized as low-temperature hot corrosion, primarily governed by an “oxidation-sulfidation” mechanism, resulting in relatively mild surface corrosion. The addition of Nb, serving as a third element, facilitated the formation of protective Cr₂O₃ and Al₂O₃ films. Furthermore, Nb enhanced the activities of Al and Cr, ensuring rapid replenishment of these alloy elements during corrosion. However, when the Nb content reached 1.2 wt%, the formation Nb₂O₅ was found to compromise the compactness of the oxide film, which adversely affected the corrosion resistance.