2021 Volume 62 Issue 11 Pages 1639-1646
In Fe–Ni alloys, protective oxide scales are not formed in a dry atmosphere as in an atmosphere containing water vapor; however, oxide scales consisting of duplexes with outer and inner layers are formed. This oxide scale structure is similar to that formed in an atmosphere containing water vapor. This enables accurate evaluation of the effect of water vapor on the oxidation behavior of alloys. In this study, we focused on the effect of water vapor on the growth kinetics and microstructure of the inner layer, and tried to verify a model according to which the growth of the inner oxide scale is caused by the dissociative mechanism of the outer oxide layer.
Fe–10 mass%Ni alloys were oxidized at 1200°C in N2–10 O2 or N2–10 O2–20 H2O (vol%, unless state otherwise) at 1200°C for 5, 15, 30, 60, and 180 min, and the effect of water vapor on the microstructure of the oxide scale was investigated. The thickness of the outer layer was not considerably different in the N2–10 O2 atmosphere compared to that in the N2–10 O2–20 H2O atmosphere. The thickness of the inner layer was significantly greater in the N2–10 O2–20 H2O atmosphere than that in the N2–10 O2 atmosphere. The inner layer was of dense FeO in N2–10 O2–20 H2O, but contained many voids in N2–10 O2. The oxidation rate was higher in N2–10 O2–20 H2O than in N2–10 O2. This could be caused by the dissociation of FeO in the atmosphere containing water vapor, which additionally supplies Fe to the surface.
Fig. 9 Proposed model for oxidation of Fe–Ni alloy in (a) N2–10 O2 and (b) N2–10 O2–20 H2O.
Steel products require high surface quality and characteristics suitable for their intended use. The formation of oxide scales on the surface of steel during hot rolling at high temperatures in a heating furnace affects the surface quality of steel products. The scales are usually removed by high-pressure hydraulic descaling before hot rolling. If the descaling is insufficient, the scales get pushed into the steel by hot rolling, which causes surface defects and results in a poor surface appearance. Therefore, controlling the microstructure of oxide scales is important for improving the surface quality of steel products. One of the factors controlling the microstructure of oxide scales is the heating atmosphere, which has been extensively studied. In particular, the effect of atmosphere containing water vapor on the oxidation behavior of alloys has been widely investigated. Fujii et al.1,2) investigated the oxidation behavior of Fe–Cr alloys in an atmosphere containing water vapor, and proposed a model according to which the growth of the inner oxide layer is caused by the dissociative mechanism of the outer oxide layer. This model can be understood as follows: hydrogen produced by the reaction of water vapor in the atmosphere with steel diffuses in the oxide scale and reduces FeO in the outer layer, leading to the formation of water vapor. This water vapor migrates inwardly and oxidizes the alloy to form hydrogen at the scale/alloy interface. Hydrogen migrates outwardly and reduces FeO in the voids in the outer layer and further produces the water vapor again. Thus, water vapor acts as a transport medium for carrying oxygen through voids into the alloy. Therefore, the inner layer can grow inwardly via oxygen diffusion. Such formation and growth of the inner layer by the inward diffusion of oxygen through water vapor has also been reported for Fe–Cr alloys3–11) and other alloys such as Fe–Al alloys5,12–15) and Fe–Si alloys.5,16–19) However, most reports have focused on alloys in which solute elements such as Cr, Al, and Si form protective oxide scales on the surface in a dry atmosphere and the mechanism that water vapor promotes the formation of a fast-growing Fe-rich oxide scale. Thus, the dissociative mechanism proposed by Fujii et al.1,2) has not been verified.
In this study, we investigated the formation and growth mechanisms of the inner layer on Fe–Ni alloys in the atmosphere containing water vapor. In Fe–Ni alloys, protective oxide scales are not formed in a dry atmosphere as in an atmosphere containing water vapor; however, oxide scales consisting of duplexes with outer and inner layers are formed. This oxide scale structure is similar to that formed in an atmosphere containing water vapor. This enables accurate evaluation of the effect of water vapor on the oxidation behavior of alloys. Several studies have investigated the oxidation behavior of Fe–Ni alloys in an atmosphere containing water vapor. Brown et al.20) reported that the inner layer was thicker in a H2/H2O atmosphere than in pure oxygen. Fukumoto et al.21) also reported that the addition of water vapor to the atmosphere increased the oxidation rate and thickness of the inner layer. On the other hand, Kusabiraki et al.22) reported that the water vapor potential in the atmosphere has a limited effect on the oxidation rate. However, those studies did not focus on how the presence of water vapor accelerates the oxidation kinetics, in particular the growth kinetics of the inner layer. Therefore, in this study, we focused on the effect of water vapor on the growth kinetics and microstructure of the inner layer, and tried to verify the dissociative mechanism.
Fe–10 mass%Ni alloy was prepared by melting pure constituent metals, Fe (99.95% pure) and Ni (99.99% pure). The chemical composition of the material is listed in Table 1. Samples with dimensions of 30 mm × 30 mm × 3 mm were cut from the ingots, and their surfaces were grinded by #800 plate grinding.
Oxidation was performed in an infrared furnace. The samples were heated to 1200°C for 5 min in a N2 atmosphere, following which the atmosphere was changed to N2–10 O2 or N2–10 O2–20 H2O to oxidize the samples for 5, 15, 30, 60, and 180 min. After oxidation, the samples were cooled to room temperature under N2 flow. The flow rate of the gases was adjusted to 3 L/min. The oxidation mass gain during heating and isothermal oxidation was continuously monitored by thermogravimetric analysis. The cross-sectional microstructures of the samples after oxidation were observed using an optical microscopy. The thicknesses of the oxide scales were measured from their cross-sectional microstructures, including the voids. The concentration profiles of Fe, Ni, and O across the oxide scales were measured using electron-probe microanalysis (EPMA). The area fractions of Fe(Ni) precipitates and voids along the thickness of the inner layer were obtained from the cross-sectional optical micrographs by image analysis.
The oxidation kinetics i.e., the (a) oxidation time and (b) square root of oxidation time of Fe–10 mass%Ni alloys oxidized in N2–10 O2 and N2–10 O2–20 H2O at 1200°C for 180 min are shown in Fig. 1. Although deviation was observed in the plot in Fig. 1(b) after approximately 16 min of oxidation, the average oxidation kinetics in N2–10 O2–20 H2O appeared to be parabolic; on the other hand, sub-parabolic kinetics was observed in N2–10 O2. The lower oxidation kinetics of the alloy observed in N2–10 O2 was associated with this sub-parabolic behavior.
Oxidation kinetics of Fe–10 mass%Ni alloy oxidized in N2–10 O2 and in N2–10 O2–20 H2O at 1200°C for 180 min with (a) oxidation time and (b) square root of oxidation time.
The cross-sectional microstructure of the oxide scales formed on the Fe–10 mass%Ni alloy oxidized in N2–10 O2 at 1200°C for different times is shown in Fig. 2. An oxide scale consisting of a duplex structure with outer and inner layers was formed after oxidation. The outer layer consisted of a multilayered structure with Fe2O3, Fe3O4, and FeO + Fe3O4 layers (in that order, from the surface). A few voids (dark contrast) were formed in the outer layer. The outer layer was thicker than the inner layer, and notably thickened with the oxidation time. A magnified view of the inner layers in the Fe–10 mass%Ni alloy oxidized in N2–10 O2 at 1200°C for different times is shown in Fig. 3. The inner layer consisted of an Fe(Ni) metal phase (bright contrast) and FeO (dark contrast). Voids and cavities (darkest contrast) of different sizes and shapes were distributed in the inner layer. The size of the Fe(Ni) metal phase and voids increased with the oxidation time. The scale/steel interface was uneven owing to the penetration of needle-like FeO pricipitates into the steel substrate.
Cross-sectional microstructures of Fe–10 mass%Ni alloys oxidized in N2–10 O2 at 1200°C for (a) 5, (b) 15, (c) 30, (d) 60, and (e) 180 min.
High-magnification images of the inner scale of Fe–10 mass%Ni alloys oxidized in N2–10 O2 at 1200°C for (a) 5, (b) 15, (c) 30, (d) 60, and (e) 180 min.
The cross-sectional microstructure of the oxide scales formed on the Fe–10 mass%Ni alloy oxidized in N2–10 O2–20 H2O at 1200°C for different times is shown in Fig. 4. The oxide scale consisting of a duplex structure with the outer and inner layers, similar to that formed in N2–10 O2, was formed after oxidation. The outer layer tended to be thinner than that formed in N2–10 O2; however, the inner layer appeared to be much thicker than that in N2–10 O2. Many large voids were formed at the outer/inner layer interface after 5 min of oxidation. These voids grew larger and were linked with each other, creating a large gap at the interface. The outer most Fe2O3 scale became thicker after the gap was developed (Fig. 4(c)), which indicates that the formation of the gap might decrease the outward Fe flux. Metallic phase formation was also confirmed at the upper part of the gap surface (arrows in Fig. 4(b) and (c)), suggesting that the separation was caused by the reduction of FeO near the interface between outer and inner layers. With prolonged oxidation, a new FeO layer was confirmed to form within the gap from the surface of the inner layer (Fig. 4(c)). This cycle, i.e., the formation of a new oxide scale in the gap and void formation followed by gap development at the outer/inner interface, was repeated during oxidation, resulting in a layer-by-layer structure of the outer oxide scale (Fig. 4(e)). A magnified view of the inner layer formed on the Fe–10 mass%Ni alloy oxidized in N2–10 O2–20 H2O at 1200°C for different times is shown in Fig. 5. The inner layer consisted of an Fe(Ni) metal phase and FeO. Fewer voids were confirmed to have formed in the inner layer in the N2–10 O2–20 H2O atmosphere compared to those in N2–10 O2, except at the beginning of the oxidation (5 min). The size of the individual Fe(Ni) metal particles as well as voids increased with the oxidation time.
Cross-sectional microstructures of Fe–10 mass%Ni alloys oxidized in N2–10 O2–20 H2O at 1200°C for (a) 5, (b) 15, (c) 30, (d) 60, and (e) 180 min.
High-magnification images of the inner scale of Fe–10 mass%Ni alloys oxidized in N2–10 O2–20 H2O at 1200°C for (a) 5, (b) 15, (c) 30, (d) 60, and (e) 180 min.
The thickness of the oxide scales formed on the Fe–10 mass%Ni alloy as a function of oxidation time in (a) N2–10 O2 and (b) N2–10 O2–20 H2O at 1200°C for different times is shown in Fig. 6. The growth rate of the outer layer was similar in N2–10 O2 and N2–10 O2–20 H2O atmospheres up to 60 min. Thereafter, the outer layer became thicker in N2–10 O2 than in N2–10 O2–20 H2O. On the other hand, the growth kinetics of the inner layer was much faster in N2–10 O2–20 H2O than in N2–10 O2. The total thickness of the oxide scale formed in N2–10 O2–20 H2O was greater than that in N2–10 O2 because of the faster growth of the inner layer. It is apparent that water vapor greatly increases the growth rate for the inner layer but not for the outer layer, which results in an increase in the oxidation rate of the alloy in N2–10 O2–20 H2O.
Thickness of the oxide scales in Fe–10 mass%Ni alloys oxidized in (a) N2–10 O2 and (b) N2–10 O2–20 H2O at 1200°C as a function of time.
Figure 7 shows the ratio of the thickness of the inner/outer layer in the Fe–10 mass%Ni alloy oxidized in N2–10 O2 and N2–10 O2–20 H2O at 1200°C for different times. The ratio was similar in N2–10 O2 and N2–10 O2–20 H2O at the beginning of oxidation. Subsequently, the ratio decreased in N2–10 O2, suggesting that the growth rate of the inner layer in N2–10 O2 decreased. On the other hand, the ratio increased in N2–10 O2–20 H2O, suggesting that the growth rate of the inner layer in N2–10 O2–20 H2O increased.
Ratio of the thickness of the inner/outer layer in Fe–10 mass%Ni alloys oxidized in N2–10 O2 and N2–10 O2–20 H2O at 1200°C as a function of time.
The area fractions of the (a) Fe(Ni) metal phase, (b) voids, and (c) FeO in the inner layer formed on the Fe–10 mass%Ni alloy oxidized in N2–10 O2 and N2–10 O2–20 H2O at 1200°C for different times are shown in Fig. 8. The area fraction of the Fe(Ni) metal phase was comparable in N2–10 O2 and N2–10 O2–20 H2O and increased with the oxidation time. The area fraction of voids was the same only after 5 min of oxidation in N2–10 O2 and N2–10 O2–20 H2O. They tended to increase slightly in N2–10 O2 but decreased rapidly and remained small in N2–10 O2–20 H2O. The opposite trend was observed for the area fraction of FeO, which increased rapidly at the beginning of oxidation in N2–10 O2–20 H2O and remained large as oxidation proceeded. The area fraction of FeO was much smaller in N2–10 O2 than in N2–10 O2–20 H2O, which suggests that the voids formed in the inner layer were filled by the formation of FeO. Observation of the microstructure of the oxide scale shows that water vapor increases the growth rate of the inner layer, which is accompanied by a reduction in void volume due to FeO formation.
Area fraction of (a) Fe(Ni) metal phase, (b) voids, and (c) FeO in the inner layer of Fe–10 mass%Ni alloys oxidized in N2–10 O2 and N2–10 O2–20 H2O at 1200°C as a function of time.
In this study, we found the effect of water vapor on the growth rate of the outer oxide layer to be marginal; however, water vapor significantly affected the growth rate of the inner oxide layer. The inner layer formed in N2–10 O2–20 H2O was dense with a higher volume fraction of FeO, whereas that formed in N2–10 O2 contained many voids with a lower FeO fraction. This higher growth rate of the inner oxide layer in N2–10 O2–20 H2O could contribute to the faster oxidation kinetics of the alloy.
When Fe-low Cr, Fe-low Si, and Fe-low Al alloys are oxidized in atmospheres without water vapor, they generally form internal oxide precipitates but do not form the inner layer. However, the Fe–Ni alloy formed an inner layer in the N2–10 O2 atmosphere, which strongly suggests that the presence of water vapor cannot be an essential factor for the formation of the inner layer. Ni is not oxidized below the FeO scale because FeO is more stable than NiO. Since the diffusivity of Ni in Fe–Ni alloy is much lower than that of Fe,23) Ni enrichment takes place in the alloy surface region once the FeO scale is formed due to Fe consumption, as illustrated in Fig. 9(a). The injection of vacancies, which resulted from the counter diffusion of Fe in the FeO scale, made the alloy subsurface region porous. Once the subsurface region became porous, the oxygen potential in the voids increased, which resulted in the internal oxidation of Fe in the Ni-enriched subsurface region. Continuous Fe consumption by the growth of the outer layer, but limited inward oxygen supply to the inner layer in N2–10 O2, resulted in a very porous inner layer. The oxidation mechanism of the alloy in N2–10 O2 was similar in N2–10 O2–20 H2O; however, the inner layer was thicker and less porous, which indicates that the oxygen supply to the inner layer increased by the presence of water vapor in the atmosphere.
Proposed model for oxidation of Fe–Ni alloy in (a) N2–10 O2 and (b) N2–10 O2–20 H2O.
According to the dissociative mechanism proposed by Fujii et al.,1,2) H2 produced at the surface of the oxide scale diffuses inwardly in the oxide scale and reduces the outer FeO in the void surface according to reaction (1).
| \begin{equation} \text{FeO} + \text{H$_{2}$} \to \text{Fe} + \text{H$_{2}$O} \end{equation} | (1) |
Figure 10 shows the oxidation behavior of the alloy in N2–10 O2 for 30 min followed by N2–10 O2–20 H2O for 30 min (condition A) and vice versa (condition B) at 1200°C. All samples were heated at 1200°C for 5 min before oxidation and cooled to room temperature in the N2 atmosphere. Under condition A, the oxidation rate increased a few minutes after the atmosphere was changed to N2–10 O2–20 H2O. Under condition B, the oxidation rate decreased once the atmosphere was changed to N2–10 O2. The cross-sectional microstructures of the oxide scale formed under condition A shown in Fig. 11 reveal that the inner layer grew thicker compared to the inner layer formed in N2–10 O2 atmosphere oxidized for 60 min (Figs. 2, 3(d)), but was thinner than the inner layer formed in N2–10 O2–20 H2O atmosphere for 60 min (Figs. 4, 5(d)). The inner layer became denser with higher FeO content. A large gap was developed near the outer/inner interface, and the Fe(Ni) metal phase was confirmed to form along the surface of the gap. Figure 12 shows the cross-sectional microstructure of the oxide scale and a magnified view of the inner layer formed under condition B. The thickness of the inner layer was comparable to that of the oxide scale formed in N2–10 O2–20 H2O atmosphere for 60 min (Figs. 4, 5(d)); however, the layer became very porous with a decrease in the amount of FeO. These results indicate that H2O enhances inward oxygen transport, which results in faster growth of the inner layer that is dense due to the higher FeO content. These results support for the dissociative mechanism proposed by Fujii et al.,1,2) which explains the enhanced inward oxygen transport by H2O. However, this model does not explain the higher oxidation mass gain of the Fe–10 mass%Ni alloy oxidized in N2–10 O2–20 H2O because the oxygen supplied to the inner layer by dissociation of the outer FeO layer did not contribute to the oxidation mass gain.
Oxidation kinetics of Fe–10 mass%Ni alloy oxidized in N2–10 O2 and in N2–10 O2–20 H2O at 1200°C (a) for 30 min in N2–10 O2 and for 30 min in N2–10 O2–20 H2O (condition A) and (b) for 30 min in N2–10 O2–20 H2O and for 30 min in N2–10 O2 (condition B).
(a) Cross-sectional microstructures and (b) high-magnification images of the inner scale of Fe–10 mass%Ni alloys oxidized at 1200°C for 30 min in N2–10 O2 and for 30 min in N2–10 O2–20 H2O (condition A).
(a) Cross-sectional microstructures and (b) high-magnification images of the inner scale of Fe–10 mass%Ni alloys oxidized at 1200°C for 30 min in N2–10 O2–20 H2O and for 30 min in N2–10 O2 (condition B).
Figure 9(b) illustrates the proposed model in this study for oxidation in N2–10 O2–20 H2O atmosphere. As shown in Fig. 4, Fe was present at the outer part of the inner surface of the gap (interface (I)). This suggests a reduction of FeO in the outer layer, resulting in the formation of H2O by reaction (1). The reduced Fe might be oxidized to Fe2+ and diffuses toward the scale surface (JFe③) and forms a new oxide at the surface, which contributes to the oxidation mass gain. H2O generated by reaction (1) migrates within the gap and the voids in the inner layer, and reacts with the Fe2+ diffused from the substrate (JFe④) at the interface (II) and within the voids in the inner layer, to form a new oxide layer within the gap and voids, resulting in a layer-by-layer structure of the outer oxide scale and a dense inner layer. However, the FeO formation in the gap near the outer/inner interface and voids in the inner layer does not contribute to the overall oxidation mass gain, since oxygen (JO2⑤, ⑥) is supplied from the dissociation of the outer FeO by reaction (1). Thus, the outward Fe2+ diffusion (JFe③) is likely to contribute the faster oxidation kinetics in the atmosphere containing H2O.
The mass of oxygen used for additional FeO formation in the voids of the inner layer (JO2⑥) in N2–10 O2–20 H2O was estimated to be approximately 14.0 mg/cm2 after 30 min of oxidation, by comparing the amount of FeO formed in the inner layer in N2–10 O2 and N2–10 O2–20 H2O atmospheres. The mass of oxygen used for FeO formation within the gap (JO2⑤) can also be evaluated from the thickness of the FeO layers formed in the gap, and was found to be approximately 6.3 mg/cm2. Thus, the total mass of oxygen produced by the dissociation of FeO in reaction (1), which corresponds to the sum of JO2⑥ and JO2⑤, was approximately 20.3 mg/cm2. Since this mass of oxygen is supplied from reaction (1) as H2O, it must be equivalent to that of the oxygen from the reduced FeO that formed a gap in the outer layer. The total mass of oxygen in the reduced FeO can be evaluated from the width of the gap to be approximately 12.7 mg/cm2, which is lower than the total mass of oxygen, JO2⑥, and JO2⑤. This discrepancy might be caused by underestimation of the width of the gap in the outer layer, because of difficulty to evaluate the gap volume in the oxide scale due to the shape of the gap and to keep the gap width during sample preparation for cross-sectional observation.
Assume that the total Fe fluxes in the oxide scales formed in N2–10 O2 and N2–10 O2–20 H2O atmospheres, JFe① and JFe②+④, are the same regardless of void and gap formation. Then, as mentioned above, the additional Fe supply, JFe③, should contribute the total oxidation mass gain, and cause the difference in the oxidation mass gain in two atmospheres with/without H2O. The additional oxidation mass gain due to the formation of the new oxide by this additional Fe supply should be the same as the total oxygen mass in the reduced FeO by the dissociation, i.e., JO2⑥, and JO2⑤. The difference of the oxidation mass gains after 30 min of oxidation in two atmospheres was approximately 20 mg/cm2, which agrees very well with the total oxygen mass obtained by the dissociation of FeO (20.3 mg/cm2). Thus, the dissociation of FeO in the atmosphere containing H2O results in an additional supply of Fe to the surface, which is the main factor that increases the oxidation kinetics.
Fe–10 mass%Ni alloys were oxidized at 1200°C in N2–10 O2 or N2–10 O2–20 H2O at 1200°C for 5, 15, 30, 60, and 180 min, and the effect of water vapor on the microstructure of the oxide scale was investigated. The obtained results are summarized as follows:
