II. EXPERIMENTAL

The metal sources (granules/sheets) of Al (99.9%), Cr (99.9%), Mo (99.9%), Nb (99.9% pure), Ti (99.9%), and Zr (99.2%) were melted using an arc melting furnace (ACM-01, DAIA-VACUUM, Tokyo, Japan) to form Al_xCr_yMo_z-NTZ. Details of the compositions of the samples are given in Table I.

TABLE I.

Detailed compositions of the HEA Al_xCr_yMo_z-NTZ.

Melting was done in an inert gas (Ar) atmosphere (10⁻⁴–10⁻⁵ mbar). A high-purity Ti-sponge was also used as an oxygen getter to minimize oxidation. Water-cooled copper molds were used as a melting pit to prevent the melting of the mold and to reduce the temperature at a cooling rate up to 100 K/s after the melting process had completed. All the samples were remelted at least five times to enhance the chemical homogeneity. Further homogenization was carried out by heat treating the samples at 1200 °C up to 24 h in an Ar atmosphere, as per previous practices.^{Reference Yurchenko, Stepanov and Salishchev25–Reference Stepanov, Shaysultanov, Salishchev and Tikhonovsky27} A microstructural analysis of the Al_xCr_yMo_z-NTZ samples was carried out by X-ray diffraction (XRD, D/MAX-2500, Rigaku, Tokyo, Japan) and scanning electron microscopy coupled with energy dispersive spectroscopy (SEM-EDS, FEI Magellan 400, Hillsboro, Oregon). The effect of varying the composition on the isothermal oxidation resistance of Al_xCr_yMo_z-NTZ (2 × 4 mm² samples) was analyzed via a thermogravimetric analysis (TGA) at 1000 °C in air up to 1 h. An oxidation test at 1000 °C for 20 h was conducted to assess the mass change and the integrity of the Al_xCr_yMo_z-NTZ HEA samples after extended oxidation.

III. RESULTS AND ANALYSIS

The SEM analysis of the homogenized Al_xCr_yMo_z-NTZ HEAs reveals a coarse granular microstructure^{Reference Stepanov, Yurchenko, Shaysultanov, Salishchev and Tikhonovsky7} with clear boundaries between the solid solution and the secondary phases as shown in Fig. 2. The secondary phases in refractory HEAs have been identified as various kinds of intermetallic phases including the hexagonal C14 Laves phase^{Reference Yurchenko, Stepanov, Shaysultanov, Tikhonovsky and Salishche2–Reference Yurchenko, Stepanov, Zherebtsov, Tikhonovsky and Salishchev4,Reference Stepanov, Yurchenko, Shaysultanov, Salishchev and Tikhonovsky7,Reference Stepanov, Yurchenko, Sokolovsky, Tikhonovsky and Salishchev19,Reference Duan, Huang, Sun, Peng and Zhou28,Reference Raghavan29} and the cubic C15 Laves phase.^{Reference Kellou, Grosdidier, Coddet and Aourag30} In the AlCrNbTiV^{Reference Stepanov, Yurchenko, Skibin, Tikhonovsky and Salishchev9} and AlCrMoNbTi^{Reference Chen, Kauffmann, Gorr, Schliephake, Seemüller, Wagner, Christ and Heilmaier31} HEAs, the Laves phase compositions of (Nb, Ti)(Al, Cr, V)₂^{Reference Stepanov, Yurchenko, Skibin, Tikhonovsky and Salishchev9} and (Nb, Ti)(Al, Cr, Mo)₂^{Reference Chen, Kauffmann, Gorr, Schliephake, Seemüller, Wagner, Christ and Heilmaier31} were described as having AB₂ stoichiometry in which A and B are elements with greater and lesser atomic sizes, respectively.^{Reference Stepanov, Yurchenko, Skibin, Tikhonovsky and Salishchev9}

FIG. 2.

A combinatorial library of SEM images of the homogenized Al_xCr_yMo_z-NTZ samples.

The coupled analyses of the SEM microstructures with XRD and EDS provided useful information about the structures and compositions of the constituent phases as illustrated using the cases of Cr₃₀Mo₁₀-NTZ and Al₁₀Cr₃₀-NTZ in Fig. 3. XRD patterns of the homogenized Cr₃₀Mo₁₀-NTZ and Al₁₀Cr₃₀-NTZ samples show major BCC structures along with secondary phases. A Cr₂Zr-type C15 Laves phase was identified in Cr₃₀Mo₁₀-NTZ and an Al₃Zr₅-type intermetallic phase was identified in Al₃₀Cr₁₀-NTZ as the major secondary phases, respectively. The chemical compositions of the BCC phases and the intermetallic phases in the Cr₃₀Mo₁₀-NTZ and Al₁₀Cr₃₀-NTZ samples measured by the EDS analysis were presented in the insets of Fig. 3, respectively. Table II shows the chemical composition of the BCC phases observed in the twelve Al_xCr_yMo_z-NTZ samples. In general, the concentrations of Nb and Ti are higher than 20 at.%, while the concentrations of Zr are lower than 20 at.% in the BCC phases in the Al_xCr_yMo_z-NTZ samples.

FIG. 3.

Microstructural analyses of (a) Cr₃₀Mo₁₀-NTZ and (b) Al₁₀Cr₃₀-NTZ by SEM, XRD, and EDS.

TABLE II.

EDS analysis of the BCC phases in the Al_xCr_yMo_z-NTZ samples.

The likelihood of intermetallic phases and solid solution phases in Al_xCr_yMo_zNbTiZr HEAs was estimated by comparing the atomic size difference (δ_r), the Allen electronegativity difference (∆X _Allen),^{Reference Yurchenko, Stepanov and Salishchev25} and the average value of the d-orbital energy level $\overline {\left( {Md} \right)}$ with the intermetallic phase formation criteria. Yurchenko et al. analyzed that Laves phase formation is observed in HEAs having an atomic size difference (δ_r) greater than 5% and an Allen electronegativity difference (∆X _Allen) exceeding 7%.^{Reference Yurchenko, Stepanov and Salishchev25}

We calculated δ_r and ∆X _Allen for Al_xCr_yMo_zNbTiZr and these values are given in Table III. For most Al_xCr_yMo_zNbTiZr HEAs, the δ_r values are greater than 5%, except for Al_xMo_z-NTZ HEAs. The ∆X _Allen values of Al_xCr_yMo_zNbTiZr exceed 7% for most compositions. An increase in the Cr content leads to an increase in ∆X _Allen. However, for Al₂₀Mo₂₀-NTZ, Al₁₀Mo₃₀-NTZ, and Cr₁₀Mo₃₀-NTZ, ∆X _Allen ≤ 7%. The values of δ_r and ∆X _Allen suggest intermetallic phases in most Al_xCr_yMo_zNbTiZr HEAs. While Al₂₀Mo₂₀-NTZ and Al₁₀Mo₃₀-NTZ have δ_r < 5 and ∆X _Allen ≤ 7%, a microstructural analysis exhibits the existence of intermetallic phases similar to the other Al_xCr_yMo_zNbTiZr HEAs as shown in Fig. 2.

TABLE III.

Atomic size mismatch (δ_r), Allen electronegativity difference (∆X _Allen), and average values of d-orbital energy level $\overline {\left( {Md} \right)}$ calculated for Al_xCr_yMo_zNbTiZr.

Morinaga et al. analyzed that superalloys having an average value of the energy level of the d-orbital $\overline {\left( {Md} \right)}$ higher than 0.915 eV exhibit the σ phase.^{Reference Morinaga, Yukawa, Adachi and Ezaki32} Lu et al.^{Reference Lu, Dong, Jiang, Wang, Li and Zhang33} applied Morinaga’s approach to HEAs and observed that the HEAs with $\overline {Md} > 1.09$ exhibit stable intermetallic phases with few exceptions. As per the methods given in the literature,^{Reference Morinaga, Yukawa, Adachi and Ezaki32,Reference Lu, Dong, Jiang, Wang, Li and Zhang33} we calculated the average value of the energy level of the d-orbital $\left( {\overline {Md} } \right)$ for Al_xCr_yMo_zNbTiZr, as shown in Table III. According to the $\overline {Md}$ values, secondary phases are expected to appear in all compositions of Al_xCr_yMo_zNbTiZr.

It has been reported that the oxidation of HEA yields a mixture of various oxides due to multiple principal elements.^{Reference Liu, Wang, Zhang, Tang and Zhang13,Reference Senkov, Senkova, Dimiduk, Woodward and Miracle21,Reference Waseem, Lee, Lee and Ryu34} The literature suggests the presence of various oxides, including TiO₂, Nb₂Zr₆O₁₇, TiNb₂O₇, ZrO₂, NbCrO₄, Cr₂O₃, NbO,^{Reference Butler, Chaput, Dietrich and Senkov20} Cr₂O₃,^{Reference Müller, Gorr, Christ, Chen, Kauffmann and Heilmaier35} Al₂O₃,^{Reference Gorr, Mueller, Christ, Mueller, Chen, Kauffmann and Heilmaier36} Cr₂Ti₅O₁₃, Nb₃Cr₂O₁₀, Nb₂Zr₈O₂₁, Ti₂ZrO₆, Ti_0.8O₂, Cr₂TiO₅, and Nb₂O₅^{Reference Senkov, Senkova, Dimiduk, Woodward and Miracle21} in refractory HEAs. To analyze the effect of various constituent elements and quantify the oxidation resistance, the TGA method can be used for HEAs.^{Reference Butler, Chaput, Dietrich and Senkov20,Reference Senkov, Senkova, Dimiduk, Woodward and Miracle21,Reference Müller, Gorr, Christ, Chen, Kauffmann and Heilmaier35,Reference Kai, Li, Cheng, Chu, Huang, Tsay and Kai37} The change in the mass of the Al_xCr_yMo_z-NTZ HEA samples during oxidation and the change in the oxidation rate were estimated.^{Reference Coats and Redfern38} The samples were exposed to air at 1000 °C and the change in the mass was measured, as presented in Fig. 4(a). The mass change per unit area after one hour of oxidation at 1000 °C is mapped on a pseudo-ternary diagram as shown in Fig. 4(b).

FIG. 4.

(a) TGA curves and (b) total mass change/unit area after 1 h of oxidation at 1000 °C in a pseudo-ternary map of homogenized Al_xCr_yMo_z-NTZ.

The HEA samples with a higher concentration of Mo showed a negative slope of the TGA curves [solid symbols in Fig. 4(a)]. For example, the masses of Al₁₀Mo₃₀-NTZ, Cr₁₀Mo₃₀-NTZ, and Al₁₀Cr₁₀Mo₂₀-NTZ decreased at a certain period during the oxidation experiment. This behavior can be attributed to the formation of a volatile oxide, as has been reported for Mo-rich refractory alloys.^{Reference Karahan, Ouyang, Ray, Kramer and Akinc39} The evaporation during the oxidation of Al₁₀Mo₃₀-NTZ, Cr₁₀Mo₃₀-NTZ, and Al₁₀Cr₁₀Mo₂₀-NTZ may also be related to the significant porosity in the oxide layers.^{Reference Meyer, Thom and Akinc40} The microstructural analysis of the oxidized samples revealed porous oxide layers for volatile oxidation in Al_xCr_yMo_z-NTZ. The representative microstructures of the porous (Cr₁₀Mo₃₀-NTZ) and dense oxide layers (Al₃₀Cr₁₀-NTZ, respectively) are shown in Figs. 5(a) and 5(b), respectively. Mixed phases were observed in the oxide layers Al₃₀Cr₁₀-NTZ, but no Al-rich layers were found in the oxide layers.

FIG. 5.

Microstructures of (a) Cr₁₀Mo₃₀-NTZ and (b) Al₃₀Cr₁₀-NTZ, and EDS analysis results of (c) Cr₁₀Mo₃₀-NTZ and (d) Al₃₀Cr₁₀-NTZ of the oxide layers of Al_xCr_yMo_z-NTZ.

Figure 6 shows the variation in the surface oxide thicknesses after the TGA. The oxide layer thickness in the Al_xCr_yMo_z-NTZ HEA system varies between 60 and 600 µm. Although a direct comparison of the oxide thicknesses of the HEA compositions is less meaningful due to volatile oxidation in some samples, it is important to note the effect of Al on reducing the oxidation thickness in Mo-free Al_xCr_y-NbTiZr as compared to that in Mo-bearing HEAs. Al₁₀Cr₃₀-NTZ shows a 107-µm-thick oxide scale that remains unchanged upon increasing the Al content to 20 at.%. A further addition of Al to yield Al₃₀Cr₁₀-NTZ shows a significant reduction in the oxide thickness to 79 µm, as shown in Fig. 6.

FIG. 6.

Variation in the thickness of the oxide layer of Al_xCr_yMo_z-NTZ.

The formation of stable oxide films reduces the contact between oxygen and the bare metal surface and controls the oxidation rate. To determine the protectiveness of the oxide scale on Al_xCr_yMo_z-NTZ HEA, the mass change versus time was plotted on a log–log scale graph [Fig. 7(a)] and the slope was determined [Fig. 7(b)]. The protective oxide layer is observed when the log–log graph of the mass change versus time shows a slope (empirical exponent, n) of 0.5.^{Reference Senkov, Senkova, Dimiduk, Woodward and Miracle21} The plot of mass gain per unit surface area of CrMo_0.5NbTa_0.5TiZr on a logarithmic scale shows two distinct features. First, the oxidation started at a higher rate and exhibited n = 0.85 (up to 25 min), after which the development of the oxide scale reduced the rate of the reaction and n became 0.5 (which remained at that value for up to 60 min).^{Reference Senkov, Senkova, Dimiduk, Woodward and Miracle21} The oxidation of NbTiVZr at 1000 °C proceeded with n = 0.9 until the HEA sample was fully oxidized within 8 h.^{Reference Butler, Chaput, Dietrich and Senkov20} By contrast, the oxidation of CrNbTiZr was controlled by an oxygen-alloy interfacial reaction in the initial stage up to 1.8 h, with n remaining at 0.74, after which the diffusion of oxygen through the oxide scale started, controlling the rate of the oxidation reaction, with n reduced to 0.34 and remaining there until 100 h of oxidation.^{Reference Butler, Chaput, Dietrich and Senkov20} To compare the oxidation behaviors of these reported HEAs with that of Al_xCr_yMo_z-NTZ, the slopes of the first 10 min and the last 10 min were compared, as shown in Fig. 7(b), while the values for the samples which presented negative slopes were excluded. During the first ten minutes of oxidation, the Al_xCr_yMo_z-NTZ HEA shows n values ranging from 0.81 to 0.91.

FIG. 7.

(a) TGA curves on a log–log scale and (b) time exponent [slopes of log₁₀(∆m/area) versus log₁₀(time) curve] during the first and last ten minutes of oxidation.

During the last ten minutes of oxidation, the n value for Al_xCr_yMo_z-NTZ changed from the values during the initial 10 min. The slopes of the Mo-containing HEAs did not decrease much, as shown in Fig. 7(b). Some of them showed increased slopes, as shown in Al₂₀Cr₁₀Mo₁₀-NTZ and Al₁₀Cr₂₀Mo₁₀-NTZ. The Mo-free HEA samples of Al₃₀Cr₁₀-NTZ, Al₂₀Cr₂₀-NTZ showed n values close to 0.5, suggesting the formation of protective oxide layers. The slope of ∼0.5 for these alloys also suggests that the degree of protectiveness of the oxide layer on these alloys is comparable to that of CrMo_0.5NbTa_0.5TiZr^{Reference Senkov, Senkova, Dimiduk, Woodward and Miracle21} and is better than those of NbTiVZr and CrNbTiZr.^{Reference Butler, Chaput, Dietrich and Senkov20} HEAs having higher Al contents (within their respective groups) showed lower n values during the last ten minutes of oxidation as well. Hence, Al is observed to be effective in forming a passive oxide layer.

Mixtures of oxides are identified by an analysis of the XRD patterns of oxidized HEA samples (Fig. 8). For example, the oxide scale on Al₂₀Mo₂₀-NTZ shows Nb₂Zr₆O₁₇, AlNbO₄, and TiO₂. Nb₂Zr₆O₁₇ and ZrTiO₄ are observed in the oxide scale of Cr₂₀Mo₂₀-NTZ and Al₃₀Cr₁₀-NTZ, respectively. In addition to these, other complex oxides may also be present. The compositions of the oxide scale are similar to those in other studies in which the formation of similar oxides after the oxidation of refractory HEAs was reported.^{Reference Butler, Chaput, Dietrich and Senkov20,Reference Senkov, Senkova, Dimiduk, Woodward and Miracle21,Reference Müller, Gorr, Christ, Chen, Kauffmann and Heilmaier35,Reference Gorr, Mueller, Christ, Mueller, Chen, Kauffmann and Heilmaier36}

FIG. 8.

XRD analysis results of the oxide layers on three Al_xCr_yMo_z-NTZ HEAs [(a) Al₂₀Mo₂₀-NTZ, (b) Cr₂₀Mo₂₀-NTZ, and (c) Al₃₀Cr₁₀-NTZ] oxidized at 1000 °C for 1 h.

To analyze the response of these HEAs to prolonged oxidation, the samples were subjected to oxidation in static air (at 1000 °C) for 20 h. After 20 h of oxidation, the Al₃₀Cr₁₀-NTZ, Al₂₀Cr₁₀Mo₁₀, and Al₁₀Cr₃₀-NTZ samples remained intact and their oxide layers remained adhered to the surface. The other samples were either fully oxidized or their oxide layers were detached. The comparison of oxidation resistance capabilities of Al₃₀Cr₁₀-NTZ, Al₂₀Cr₁₀Mo₁₀, and Al₁₀Cr₃₀-NTZ demonstrates higher oxidation resistance in these samples relative to those of several Nb-based alloys (Table IV). Furthermore, we could find only two reports of HEAs with significant oxidation resistance: AlMo_0.5NbTa_0.5TiZr^{Reference Senkov, Senkova, Dimiduk, Woodward and Miracle21} and AlCrMoNbTi.^{Reference Gorr, Mueller, Christ, Mueller, Chen, Kauffmann and Heilmaier36} The mass gain in AlCrMoNbTi^{Reference Gorr, Mueller, Christ, Mueller, Chen, Kauffmann and Heilmaier36} was quite low, suggesting quite stellar oxidation resistance. However, Al₃₀Cr₁₀-NTZ, Al₂₀Cr₁₀Mo₁₀, and Al₁₀Cr₃₀-NTZ reveal better oxidation resistance capabilities than AlMo_0.5NbTa_0.5TiZr. To quantify the volatile oxidation and evaporation of Mo-containing HEAs, TGA should be coupled with mass spectrometry.^{Reference Kamruddin, Ajikumar, Dash, Tyagi and Raj41}

TABLE IV.

Comparison of the mass gain after prolonged oxidation of Al₂₀Cr₁₀Mo₁₀-NTZ, Al₃₀Cr₁₀-NTZ, and Al₁₀Cr₃₀-NTZ with relevant HEAs and refractory alloys.

Recent publications pertaining to refractory HEAs report samples with higher mechanical strength and strain levels, but all of those alloys possess W,^{Reference Senkov, Wilks, Scott and Miracle11,Reference Zhang, Zuo, Tang, Gao, Dahmen, Liaw and Lu42,Reference Zou, Maiti, Steurer and Spolenak43} Ta,^{Reference Senkov, Wilks, Scott and Miracle11,Reference Zhang, Zuo, Tang, Gao, Dahmen, Liaw and Lu42,Reference Dirras, Gubicza, Heczel, Lilensten, Couzinié, Perrière, Guillot and Hocini44} V,^{Reference Wu, Cai, Chen, Wang, Si, Wang, Wang and Hui15,Reference Stepanov, Yurchenko, Panina, Tikhonovsky and Zherebtsov45} and/or Hf,^{Reference Guo, Wang, Luo, Li, Chen, Su, Guo and Fu12,Reference Zhang, Zuo, Tang, Gao, Dahmen, Liaw and Lu42,Reference Dirras, Gubicza, Heczel, Lilensten, Couzinié, Perrière, Guillot and Hocini44,Reference Fazakas, Zadorozhnyy, Varga, Inoue, Louzguine-Luzgin, Tian and Vitos46} with deleteriously high-density levels (due to Ta and W), high costs (due to Hf and Ta), or oxidation (due to V). Therefore, the commercial feasibility of such HEAs is uncertain. Al₂₀Cr₁₀Mo₁₀-NTZ and Al₃₀Cr₁₀-NTZ showed good oxidation resistance levels without such deleterious elements. This forecasts the potential high-temperature applications of these HEAs; therefore, their mechanical properties need to be analyzed in detail in the near future.