Effect of Y content on microstructure of CoCrNiAl0.6Yx MEAs
The morphology and microstructure of CoCrNiAl0.6Yx are shown in Fig. 1 (a), (c), (e) and (g). All alloys exhibit dendritic structure with two clear phases. All alloys consist of a dark contrast β phase and a bright contrast γ phase. The microstructure of CoCrNiAl0.6Yx shown in Fig. 1 (b), (d) and (f) has not an obvious difference, but some sub-micron γ phase precipitates are observed within the β phase in Fig. 1 (h). The chemical compositions and phase maps in the CoCrNiAl0.6Y(0.02) and CoCrNiAl0.6(Y0.11) determined by SEM-EDS are presented in Table 2, Fig. 2 (a) and (b). Generally speaking, Co and Cr can be dissolved in NiAl-based chemistry. The composition limit for Cr induced extension of the β phase field is around 15 at. % in the Ni–Al–X diagram[20]. Cr usually occupies the sites on which Ni or Al is lower in content, and it does not show special preference on Ni and Al sites in the β phase. Because of these, Cr usually has a higher solubility in γ phase than in β phase in a γ + β alloy[16]. Both EDS mapping in Fig. 2 and element distribution in Table 2 confirm that Co and Cr enrich in γ phase, Ni and Al enrich in β phase,
Table 2
Chemical composition of the phases in the CoCrNiAl0.6Y(0.02) and CoCrNiAl0.6Y(0.11) determined by EDS point analysis
Alloys | phases | Elements (at. %) |
| Co | Cr | Ni | Al | Y |
CoCrNiAl0.6Y(0.02) | γ | 32.36 ± 1.21 | 29.70 ± 0.95 | 25.90 ± 1.03 | 12.00 ± 0.78 | 0.04 ± 0.02 |
| β | 25.47 ± 1.13 | 23.32 ± 0.87 | 28.61 ± 0.98 | 22.58 ± 0.94 | 0.02 ± 0.01 |
CoCrNiAl0.6(Y0.11) | γ | 30.71 ± 1.18 | 31.38 ± 1.09 | 24.64 ± 0.95 | 13.13 ± 0.76 | 0.14 ± 0.03 |
| β | 21.50 ± 0.97 | 20.76 ± 0.89 | 27.09 ± 0.98 | 30.55 ± 1.16 | 0.10 ± 0.02 |
The white precipitates are observed at γ/β phase boundaries resulting from the extremely low solubility of Y. The higher Y content, the more precipitates at the grain boundaries, and the higher Y content, the more obvious sub-micron γ phase precipitates within the β phase. Usually, the β phase preferentially nucleates and grows during the solidification process, the γ phase forms during cooling. The β phase reacts with the residual liquid metal (L) as follow: L + β → γ, and the residual liquid metal will be rich in Y because the solubility of Y in the β phase is lower than that in γ phase. As a result, the sub-micro γ-phase precipitates forms within the primary β phase via peritectic reaction[2, 21], and the higher Y content, the more obvious sub-micron γ phase precipitates within the β phase.
Figure 3 shows XRD diffraction patterns of CoCrNiAl0.6Y(0.02) and CoCrNiAl0.6(Y0.11) MEAs, these alloys form duplex FCC plus BCC phase structure[2, 22], which is consistent with the SEM results.
Effect of Y content on oxidation behavior of CoCrNiAl0.6Yx MEAs
Figure 4 shows the surface macromorphology of CoCrNiAl0.6Yx MEAs after cyclic oxidation experiments for 60 h at 1000°C, 1100°C and 1200°C, respectively. All metals changed the color into gray after oxidation for 60 h at 1000°C, 1100°C and 1200°C, the metallic luster disappeared. The surface oxides spallation did not been found in these metals.
The mass change of each sample was recorded in the process of cyclic oxidation. After cyclic oxidation experiments for 60 h at 1000°C and 1100°C, there are little mass change in these samples. The oxidation kinetic curves after cyclic oxidation experiments for 60 h at 1200°C follow a parabolic rate law. Figure 5 is the oxidation kinetic curves of CoCrNiAl0.6Yx during cyclic oxidation experiment for 60 h at 1200°C. It is clearly observed the Y content has an obvious influence on the relative oxidation resistance. Based on the oxidation kinetic curve of CoCrNiAlxY alloy at 1200°C, we can find the obvious mass gain of CoCrNiAl0.6Yx (x = 0.02, 0.05 and 0.08) in the early oxidation stage. For CoCrNiAl0.6Y (0.11) alloy, the weight gain tends to be gentle in the first 10 h. All these samples have a relative gentle weight gain from 10 h to 60 h. The weight gain of CoCrNiAl0.6Yx with Y content of 0.02, 0.05, 0.08 and 0.11 are 0.00178, 0.00163, 0.00146 and 0.00111 g/cm2, respectively.
The growth of oxide scale follows a parabolic law at 1200°C and conforms with the classical oxidation theory during cyclic oxidation. Wagner established the parabolic kinetic law of oxide film[23], and pointed out that the diffusion of positive and negative ions through the formed oxide film was the key rate-controlling step of metal oxidation. The oxidation kinetic curves followed a parabolic rate law. The following is the equation of parabolic rate law.
(∆m/A)2=Kpt (1)
∆m, A and Kp are the quality gain, the surface area of samples and parabolic rate constants.
Based on the parabolic rate law, the calculated oxidation rate constants of CoCrNiAl0.6Yx (x = 0.02, 0.05, 0.08 and 0.11) are 1.467×10− 11, 1.230×10− 11, 9.87×10− 12 and 5.704×10− 12 g2/(cm4·s), respectively. The CoCrNiAl0.6Y (0.08) alloy have a lowest mass change and oxidation rate constant.
Because CoCrNiAl0.6Yx alloys involve volatile metal oxides during the high temperature oxidation, only the macromorphology and oxidation mass gain curve cannot reflect the real oxidation situations effectively. Therefore, the microstructure characteristics were further investigated.
Figure 6 shows XRD spectra of CoCrNiAl0.6Yx alloys after cyclic oxidation experiments for 60 h at 1200°C. The oxide scale of CoCrNiAl0.6Y (0.02) consists of γ phase, Al2O3 and Cr2O3[22]. The diffraction peaks of CoCrNiAl0.6Yx(x = 0.05, 0.08, 0.11) with higher Y content at 39.6°, 42.2°, 46° and 49° are identified to be Y oxides (JCPDS 88-1040)[24]. Besides the Y oxides, the diffraction peaks of γ phase and Al2O3 also appear on the CoCrNiAl0.6Yx (x = 0.05, 0.08, 0.11).
The microstructure of CoCrNiAl0.6Yx after cyclic oxidation experiments for 60 h at 1200°C was observed. Figures 7 and 8 display the surface microstructure and EDS maps of cyclic oxidation experiments for 60 h at 1200°C. There is an obvious difference on the surface morphology and microstructure of these metal with different Y contents. The surface morphology of CoCrNiAl0.6Y (0.02) as shown in Fig. 7 (a) that the spallation can be found on the surface. Besides the spallation, the oxide scale consists of bright phase and dark phase. The dark phase is the major phase in the CoCrNiAl0.6Y (0.02) alloy. When more Y is added in the CoCrNiAl0.6 alloy, the oxide scale morphology looks like a map, and it consists of two distinct phases. As the Y content increases, the bright phase is the major phase. As seen from the microstructure of CoCrNiAl0.6Yx in Fig. 7 (a), (b), (c) and (d), the dark phase of oxide scale consists of granular oxides, the oxide scale on the bright phase is of webbed structure. Combine the EDS maps with the microstructure in Fig. 8 (a), (b), (c) and (d), the content of Y in the bright phase is higher, the content of Al is higher in the dark phase. We can infer the bright phase majorly consists of yttrium oxide, and the dark phase majorly consists of aluminum oxide. And aluminum oxide enriched phase originates from the β phase before oxidation, yttrium oxide enriched phase originates from the γ phase. There is a large area of γ phase in the CoCrNiAl0.6(Y 0.11) alloy, which makes a larger area of yttrium oxide enriched phase after high temperature oxidation.
Figure 9 and Fig. 10 display the cross profiles and EDS maps of CoCrNiAl0.6Yx MEAs after cyclic oxidation experiments for 60 h at 1200°C. The cross profiles with a lower magnification of CoCrNiAl0.6Yx (x = 0.02, 0.05 and 0.08) in Fig. 9 (a), (c) and (e) show a discontinuous internal Al2O3 scale, the CoCrNiAl0.6Y (x = 0.11) in Fig. 9 (g) show the best continuity of the Al2O3 scale. The CoCrNiAl0.6Yx (x = 0.05, 0.08 and 0.11) with a higher Y content form the dense oxide scale with a small amount of Y oxides protrusion along the scale oxides interfaces. The thickness of Al2O3 scale in these alloys is 3.61, 3.17, 3.06 and 2.99 µm, respectively.
The cross-section EDS mappings of these samples in Fig. 10 show that Al-rich β phase is highly depleted close to the Al2O3 scale because of the surface oxidation after cyclic oxidation at 1200°C. The diffusion of Al from the alloy matrix to the surface and the enrichment of Cr at the alloy matrix-oxide scale interface are the key factors causing the microstructure changes near the alloy matrix-oxide scale interface. The main microstructure changes are the formation of β-depletion zone near the alloy matrix-oxide scale interface. For the CoCrNiAl0.6Y (0.02) alloy, an internal Al2O3 scale along with an external layer of Cr2O3 is formed on the surface of alloy. For CoCrNiAl0.6Yx (x = 0.05, 0.08, 0.11) alloys, an external layer of Al2O3 with Y oxides is formed. Y is preferably distributed along the Al2O3 scale surface and internal interfaces; it is believed that the outward segregation of Y takes place along easy diffusion paths such as phase boundaries in the process of the oxidation and participate in the oxidation. At the same time, the inward diffusion of oxygen along grain boundaries also happens. Obviously, there are a little more Y oxides along the scale oxides interfaces with the higher Y content for the CoCrNiAl0.6Yx (x = 0.08 and 0.11).
Usually, the activation energy of metal oxides is less than zero, which will be oxidized spontaneously. And the lower the activation energy is, the easier the oxidation spontaneous reaction is. In the CoCrNiAl0.6Yx MEAs, Al, Cr and Y have a relative lower activation energy. In the early oxidation stage, abundant oxygens are distributed around the alloy matrix, and the oxidation reaction is controlled by the interface metal. For the CoCrNiAl0.6Y (0.02) alloy with the lower Y content, the major oxides are Al oxides and Cr oxides. When CoCrNiAl0.6Yx (x = 0.05, 0.08, 0.11) alloys with the higher Y content, Y and Al elements distributed on the interface are easily be oxidized based on the lower activation energy of metal oxides in the early oxidation stage. The formation of Al2O3 accompanied by the oxidation of a part of the Y at the surface during initial oxidation. The further oxidation reaction is controlled by the diffusion of oxygen and metal element, and the diffusion rate of element will decide the oxidation rate of alloys. As confirmed by calculation [25], Al in Ni matrix has the largest interdiffusion coefficient in most cases, so the dense Al2O3 scale can be formed.
Since Y in the CoCrNiAlY metal is the element that forms the most stable oxide. The location of the oxidized Y after initial oxidation is also determined by two processes that occur simultaneously. The first process is the segregation of Y at the metal surface where it is oxidized. The second process that affects the initial distribution of Y in the oxide scale is the diffusion of oxygen into the metal, which causes the Y to oxidize internally in the metal. Refer to the microstructure of CoCrNiAl0.6Yx MEAs before oxidation, there are the more obvious sub-micron γ phase precipitates within the β phase and Y segregation at the phase boundaries with higher Y content. It can also explain that there is more segregation of Y at the metal surface with higher Y content. A larger amount of segregated Y towards the metal surface can be promoted by oxidation at higher temperatures. These can explain the surface microstructure of CoCrNiAl0.6Yx after cyclic oxidation at 1200°C, the area of Y oxides of CoCrNiAl0.6Yx with lower Y content is small. When the content of Y is 0.11 at. %, Y oxides almost cover the whole surface. And the sample with higher Y content, Y oxides along the inner of the oxide scale can also be observed. The depth at which the Y oxidizes internally is determined by the competition between the diffusion of Y towards the alloy surface and the diffusion of oxygen into the matrix. At high temperatures the outward segregation of Y is dominant and the extent of the internal oxidation zone is relatively small[15, 26].
Usually, when the reactive-element with low content is homogeneously distributed, it will segregate to the the oxide-matrix interface and oxide scale surface because of its solubility in the alloy[27, 28]. In this case, the oxidation rate of the alumina is reduced because of the blocked inward diffusing of oxygen along oxide grain boundaries. However, if the reactive-element with high content which is higher than the solubility limit, it is mainly heterogeneously confined within the alloy as a separate phase. Therefore, the reactive element is mostly present as oxide protrusions in the oxide scale after oxidation and the growth rate of the scale is increased.
In general, the growth rate of the oxide scale of CoCrNiAl0.6Yx (x = 0.02, 0.05, 0.08 and 0.11) is slow which originates from the sluggish diffusion and the high-entropy effect of CoCrNi MEA alloy. For the CoCrNiAl0.6Yx (x = 0.05, 0.08 and 0.11) alloy, the reactive-element mainly segregate to the oxide scale surface, only a small amount Y oxide protrusions form in the interface of alumina scale, and CoCrNiAl0.6Y (0.11) show the more Y oxide in the oxide scale surface. The growth rate of the scale of CoCrNiAl0.6Y (0.11) is slow obviously based on the oxidation kinetic curves and cross profiles compared with the alloys with lower Y content. The major reason should be the formation of a dense aluminum oxide scale accompanied by the more Y oxide in the oxide scale surface during initial oxidation, which make the very low diffusivity of both aluminum and oxygen through this scale. Therefore, the CoCrNiAl0.6Y (0.11) has the best oxidation resistance.