1 Introduction

Refractory high-entropy alloys (RHEAs) have become one kind of promising structural materials for aerospace applications due to their excellent mechanical properties at high temperatures (HTs) [1,2,3,4,5]. Such alloys usually contain W, Hf, Nb, Mo and Ta, which would result in a higher density of above 10.0 g·cm−3 [6, 7] and limit their applications to some extent. In order to reduce alloy density and to achieve high strength, Al, Ti, Zr, Cr and Si are applied to substitute for partial W, Hf and Ta for the development of light-weight RHEAs [8,9,10,11,12,13]. They are based on the body-centered-cubic (BCC) structure, in which cubic B2, hexagonal Laves phase or M5Si3-type phase often appear as the second strengthening phases. It is emphasized that the B2 phase is the coherent ordered superstructure of BCC solid solution, thus the BCC/B2 coherent microstructure would have a higher thermal stability at HTs than other microstructures with non-coherent phases (Laves or M5Si3-type) precipitated in BCC matrix, leading to a better mechanical property at HTs [13,14,15,16,17,18]. For instance, the yield strength at 1273 K (σYS = 745 MPa) of AlMo0.5NbTa0.5Zr alloy is much higher than that (σYS = 240 MPa) of HfNbSi0.5TiV alloy, which is ascribed to that the former has a BCC/B2 microstructure while coarse M5Si3 phase particles are distributed into the BCC matrix in the latter [19, 20]. However, it is difficult to achieve the coherent B2/BCC microstructure in RHEAs because the B2 phase is metastable in these systems consisting of Al and early transition metals, which easily transforms to stable Laves or M5Si3-type phases [21, 22]. Thus, multi-component co-alloying is necessary to tailor the composition for obtaining coherent BCC/B2 microstructure.

It is known that the morphology, size, and distribution of second-phase particles could directly affect the strengthening effects of alloys [23,24,25]. Especially, the formation of spherical or cuboidal nanoprecipitates is closely related to the lattice misfit (ε) between coherent BCC and B2 phases [26,27,28]. Unfortunately, it is difficult to tune the lattice misfit into a moderate value due to a large composition difference between these two phases. Thus, the weave-like microstructure is common in BCC/B2-based alloys because of a relatively large lattice misfit, leading to a serious brittleness [29,30,31]. In our previous work, we investigated the morphology evolution of nanoprecipitates in BCC/B2 microstructure with alloy composition in Al–Ni–Co–Fe–Cr HEAs and found that the formation of cuboidal B2 nanoprecipitates was closely related to a moderate lattice misfit (ε = 0.2%-–1.0%), in which these HEAs were designed by the cluster formula approach of Al2M14 (M represents different combinations of transition metals) [24, 26]. Using this formula, the cuboidal coherent BCC/B2 microstructure could also be formed in light-weight Al–Ti–Zr–Nb–Ta RHEAs, as evidenced by the fact that the Al0.4Nb0.5Ta0.5TiZr0.8 alloy with M = Ti5Zr4Nb2.5Ta2.5 possesses a coherent microstructure with cuboidal B2 nanoprecipitates into BCC matrix, leading to a prominent mechanical property with σYS = 1232 MPa at room temperature [13].

Besides the mechanical property of RHEAs, the poor oxidation resistance of existing RHEAs needs to be improved. The addition of oxidation-resistant elements of Cr and Mo could form compact oxide layers to prohibit the diffusion of oxygen into the matrix [10, 32,33,34]. However, the excessive addition of Cr and Mo into RHEAs would promote the precipitation of Laves phase from the matrix and then deteriorate alloy plasticity [35,36,37,38]. Therefore, the present work will investigate the formation of coherent BCC/B2 microstructure in Al–Ti–Zr–Nb–Ta–Cr/Mo system. Alloy compositions are still designed by the composition formula of Al2M14, in which M14 are Ti6Zr2Nb3Ta2Cr1 (S1-Cr), Ti6Zr2Nb3Ta2Cr0.5Mo0.5 (S2-CrMo) and Ti6Zr2Nb3Ta2Mo1 (S3-Mo), respectively [24]. The structural stability of coherent BCC and B2 phases in these alloys with temperature increasing will be studied, in which the effect of Cr and Mo variation on the microstructure is compared. Then, the mechanical properties of these alloys will be measured and discussed in light of the precipitation strengthening mechanism.

2 Experimental

The designed alloy ingots were melted using arc-melting in a copper mold under an argon atmosphere with a weight of about 100 g. The purities of the raw elemental metals are 99.95% for Nb and Ta and 99.99% for Al, Ti, Zr, Mo and Cr, respectively. To ensure the chemical homogeneity, the ingots were remelted at least 5 times. Then, the designed alloys were solid-solutioned at 1573 K for 2 h in a muffle furnace, and finally aged at 873 and 1073 K for 24 h, respectively, in which the samples were sealed in a vacuum quartz glass tube to prevent further oxidation. Each heat treatment was followed by a water-quenching cooling. The crystalline structures of designed alloy specimens in different states were identified by a Bruker D8 X-ray diffractometer (XRD) with Cu Kα radiation (λ = 0.15406 nm). The SHIMADZU electronic probe micro-analyzer (EPMA) and JEM2100F FEG scanning transmission electron microscope (STEM) were used to examine the microstructure. The FEI Helios NanoLab 600 Dual-Beam focused ion beam (DB-FIB) instrument was used to prepare transmission electron microscopy (TEM) specimens, and the detailed description was described elsewhere [39]. The Image-Pro Plus 6.0 software was applied to analyze the particle size of precipitates and volume fraction of each phase using 5 SEM and TEM morphology images at least. EPMA and super-X energy dispersive X-ray spectrometry (EDS) detector equipped on JEM-F200 TEM operating at 200 kV were used to analyze the elemental distributions and chemical composition. The UTM5504 material test system (MTS) was used to conduct the uniaxial compressive tests at room temperature, in which three rectangular specimens with a size of 5 mm × 5 mm × 8 mm were tested for each treatment state under a strain rate of 1 × 10−3 m·s−1. The HVS-1000 Vickers hardness tester under a constant load of 4.9 N for 15 s was used to measure microhardness of alloy specimens, in which 15 indents were conducted on each sample.

3 Results

3.1 Microstructural characterization

XRD patterns of the designed S1-Cr, S2-CrMo, and S3-Mo alloys at different heat-treated states in Fig. 1 show that all these alloys in the solid-solutioned state exhibit a single BCC structure without any additional diffraction peaks of other phases. The aging treatments at different temperatures could accelerate the precipitation of second phases from BCC matrix, as evidenced by XRD patterns of 1073 K-aged alloys. These precipitates are identified as hexagonal Zr5Al3 phase according to the distinct diffraction peaks in 1073 K-aged S1-Cr and S2-CrMo alloys.

Fig. 1
figure 1

XRD patterns of designed alloys at different heat-treated states

Full size image

Although no obvious diffraction peaks of precipitated phases appear in 873 K-aged alloy samples, the back-scattered electron (BSE) observations of aged samples shown in Fig. 2 could well express the microstructural morphology of each alloy. For S1-Cr alloy, there exist three kinds of phases in 873-aged state, the bright needle-like phase, the dark flower-like phase, and the grey matrix due to the distinct Z-contrast (Fig. 2a). The elemental distributions mapped by EPMA in Fig. 3 indicate that the bright phase is enriched by Cr and Ti, and the dark phase is segregated by Al and Zr, while the grey matrix is dominated by Ti, Nb and Ta. Combined with XRD result, it could be deduced that the bright and dark phases are hexagonal Cr2Ti and Zr5Al3, respectively, which are common precipitated phases in such kind of alloy systems [40,41,42]. After aging at 1073 K for 24 h, the bright needle-like phase disappears and only the dark phase is co-existed with the grey matrix (Fig. 2b), which might be ascribed to the fast diffusion of Cr in BCC phase at HTs, as evidenced by a larger diffusion coefficient of 9.78 × 10–17 m2·s−1 at 1073 K [43,44,45], resulting in a uniform distribution in BCC matrix.

Fig. 2
figure 2

EPMA back-scattered images of designed alloys at 873 K-aged and 1073 K-aged states: a, b S1; c, d S2; e, f S3

Full size image
Fig. 3
figure 3

a BSE image and b Cr, c Ti, d Al, e Zr, f Nb and g Ta elemental distributions of S1-Cr alloy after aging at 873 K for 24 h mapped by EPMA

Full size image

When half Mo was substituted for Cr in S2-CrMo alloy, the microstructure of 873 K-aged sample is obviously different from that of aged S1-Cr alloy, as presented in Fig. 2c, which is constituted of bright nano-scaled particles and grey matrix. The phase constitution in S2 alloy was further identified by TEM. It is found that the solid-solutioned alloy exhibits a single BCC structure, as evidenced by the selected-area electron diffraction (SAED) pattern along [100]BCC direction in Fig. 4a1, a2, which is consistent with XRD result. After aging at 873 K, TEM bright-field (BF) image and corresponding SAED pattern (Fig. 4b) show that ultrafine BCC nanoparticles with a size of 10–20 nm are uniformly precipitated into B2 matrix. Moreover, there also exist the other kind of large BCC nanoparticles with a size of 30–80 nm, as shown in Fig. 4c, which corresponds to the bright particles in BSE image given in Fig. 2c. That is to say, the grey matrix consists of coherent BCC and B2 phases. Both small and large BCC nanoprecipitates exhibit a cuboidal shape although particle sizes are different. High-resolution TEM (HRTEM) image and fast Fourier transform (FFT) patterns given in Fig. 4d, d1, and d2 indicate that BCC precipitates are perfectly coherent with B2 matrix. A further super-X EDS analysis presented in Fig. 5 demonstrates that Nb, Ta, Mo and Cr are enriched in bright BCC phase and dark B2 phase is segregated by Al and Zr. The average compositions of BCC and B2 phases are Al3.93Ti31.00Zr18.36Nb28.28Ta6.16Mo3.68Cr8.89 and Al11.68Ti32.62Zr29.45Nb14.72Ta2.83Mo2.07Cr6.63 (at%), respectively. With aging temperature increasing to 1073 K, the phase could precipitate from BCC matrix, and B2 phase disappears, as shown in Figs. 1, 2d. It is easy to realize the transformation of the metastable B2 phase to stable Zr5Al3 phase thermodynamically by the collapse of the {111} plane in B2 due to the close crystalline relationship [110](001)B2∥[120](2\(\overline{1}\)1)Zr5Al3 between these two phases [13, 39].

Fig. 4
figure 4

TEM characterization of S2-CrMo alloy a1, a2 solid-solutioned and b, c 873 K-aged; d HRTEM image corresponding to SAED and d1, d2 FFT patterns along [100]BCC axis

Full size image
Fig. 5
figure 5

a STEM image and b Al, c Ti, d Zr, e Nb, f Ta, g Mo, and h Cr corresponding elemental distributions of S2-CrMo alloy at 873 K-aged state by Super-X EDS

Full size image

When Mo replaces Cr completely, a similar microstructural evolution with aging temperature occurs in S3-Mo alloy. No distinct precipitates appear in 873 K-aged sample, as presented in Fig. 2e. After aging at 1073 K, Zr5Al3 phase is precipitated on grain boundaries, and the inner part of grains still be constituted of coherent BCC and B2 phases since the microstructure is similar to that in 873 K-aged S2 alloy (Fig. 2f).

3.2 Mechanical properties

The room-temperature compressive properties of these designed alloys in both solid-solutioned and aged states were then measured, and the true stress–strain curves are shown in Fig. 6, in which the compressive yield strength (σYS) values are also marked. It is found that in the solid-solutioned state, these alloys have both high strength (σYS = 895–989 MPa) and large plasticity without fracture during compression, in which the substitution of Mo for Cr could increase the strength, as evidenced by σYS = 895 MPa for S1-Cr and σYS = 989 MPa for S3-Mo. After aging at 873 K for 24 h, the yield strength values of these alloys are significantly enhanced up to 1016–1314 MPa. In particular, the S2-CrMo alloy possesses the highest strength with σYS = 1314 MPa and good plasticity, which might be ascribed to the coherent microstructure with BCC nanoparticles precipitated into the B2 matrix (Figs. 2c, 4), while the low strength (σYS = 1016 MPa) of S1-Cr alloy is due to the precipitation of coarse needle-like Cr2Ti and flower-like Zr5Al3 phases. The further increase of aging temperature would deteriorate the mechanical property of alloys, as demonstrated by the lower yield strength (970–995 MPa) and bad plasticity in 1073 K-aged alloys. It is mainly due to the precipitation of coarse and brittle Zr5Al3 phase.

Fig. 6
figure 6

Room-temperature true compressive stress–strain curves of designed alloys at solid-solutioned, 873 K-aged and 1073 K-aged states

Full size image

4 Discussion

4.1 Microstructural evolution with Cr/Mo alloying elements and aging temperature

Combined with the above results, it can be found that the addition of Cr and Mo alloying elements has an important influence on the microstructure and mechanical properties of these current RHEAs. In 873 K-aged state, there are needle-like Cr2Ti and flower-like Zr5Al3 phases precipitated in the matrix of S1-Cr alloy, which is originated from the strong interaction between Cr and Ti that promotes the precipitation of hard and brittle Laves-Cr2Ti phase [40, 46]. The coherent BCC/B2 microstructure containing cuboidal BCC nanoprecipitates is formed in S2-CrMo alloy with half Mo substituted for Cr, which is ascribed to a moderate lattice misfit (ε = 0.94%) of BCC and B2 phases, as calculated with the equation of ε = 2 × (aB2 − aBCC)/(aB2 + aBCC), where aB2 and aBCC are the lattice constants of B2 and BCC phases, respectively [13]. When Cr was substituted by Mo completely in S3-Mo alloy, the cuboidal coherent BCC/B2 structure disappears (Fig. 2e). When aging temperature increases to 1073 K, Zr5Al3 phase particles are precipitated from the matrix and coarsened in all designed alloys due to that B2 phase can not remain stable at 1073 K and easily converts to the stable Zr5Al3 phase, which has been reported in existing RHEAs [13, 22]. Therefore, the formation of cuboidal BCC/B2 coherent microstructure is closely dependent on the alloy composition since it controls the lattice misfit between BCC and B2 phases. And such coherent microstructure is also sensitive to the aging temperature in Al–Ti–Zr–Nb–Ta–Cr/Mo system due to the metastable feature of B2 phase. According to existing reports, it is difficult to stabilize BCC/B2 coherent microstructure when the temperature is above 1073 K [13, 22].

4.2 Strengthening effect by coherent-precipitates

For 873 K-aged S2-CrMo alloy, it exhibits the highest strength with σYS = 1314 MPa without sacrificing the plasticity, which is derived from the coherent microstructure of cuboidal BCC nanoparticles coherently precipitated into B2 matrix. For the precipitation strengthening of nanoparticles, it can be generally divided into two categories, the shearing mechanism and the Orowan bypassing mechanism, in which the shearing mechanism is dominant when the particle size of nanoprecipitates is very small (< ~ 15 nm), and the Orowan bypassing mechanism will dominate when the particle size is large enough [23, 28]. Here, the current S2 alloy contains two kinds of BCC nanoprecipitates, one being ultrafine precipitates with an average radius of r = 8 nm and a volume fraction of f = 18.0 vol%, and the other with r = 29 nm and f = 10.9 vol%, in which the average radius (r) is calculated from the traced areas using a circular-equivalent, i.e., \(r = \sqrt {\left( {{\text{Area}}/{\uppi }} \right)}\). Thus, the strength increments from these two kinds of precipitates should be calculated with the shearing mechanism and the Orowan bypassing mechanism, respectively. Since the ordering strengthening effect of BCC nanoparticles is much too smaller [23], the strength increment from the shearing mechanism results from both the coherency strengthening (ΔσCS) and modulus mismatch strengthening (ΔσMS). All these contributions could be calculated with Eqs. (1–3) [23, 28]:

$$\Delta \sigma_{{{\text{CS}}}} = M\alpha_{\varepsilon } \left( {G\varepsilon_{{\text{C}}} } \right)^{\frac{3}{2}} \left( {\frac{rf}{{0.5Gb}}} \right)^{\frac{1}{2}}$$
(1)
$$\Delta \sigma_{{{\text{MS}}}} = 0.0055M\left( {\Delta G} \right)^{\frac{3}{2}} \left( \frac{2f}{G} \right)^{\frac{1}{2}} \left( \frac{r}{b} \right)^{{\frac{3m}{2} - 1}}$$
(2)
$$\Delta \sigma_{{{\text{Orowan}}}} = M\frac{0.4Gb}{{\uppi \sqrt {1 - v} }}\frac{{\ln \left( {2\sqrt{\frac{2}{3}} r/b} \right)}}{{\lambda_{{\text{P}}} }}$$
$$\lambda_{{\text{p}}} = 2\sqrt{\frac{2}{3}} r\left( {\sqrt {\frac{\uppi }{4f}} - 1} \right)$$
(3)

where M = 2.73 is the Taylor factor for BCC structure [47]; αε = 2.6; m = 0.85; εC = 2ε/3 is the constrained lattice misfit; G = 55 GPa and ΔG = 17.6 GPa are the shear modulus of the B2 matrix and the shear modulus mismatch between precipitates and matrix, respectively [24, 48]; b = \(\frac{\sqrt 3 }{2}\) aB2 = 0.279 nm is the Burgers vector; ν = 0.3 is the Poisson ratio; and λp is the inter-precipitate spacing. Thus, the strength increment from ultrafine BCC nanoparticles is calculated to be σp1 = 850 MPa, and the strength contribution from the large BCC nanoparticles is σp2 = Δ σOrowan = 411 MPa, resulting in that a total value caused by BCC nanoparticles is σp = σp1 + σp2 = 1261 MPa. It is very close to the experimental values of σYS = 1314 MPa although the solid solution strengthening from alloying elements is neglected.

5 Conclusion

In this paper, we studied the formation of coherent BCC/B2 microstructure in light-weight Al2Ti6Zr2Nb3Ta2(Cr/Mo)1 RHEAs and their mechanical properties. It is found that all these solid-solutioned RHEAs exhibit a single BCC structure, while second phases would precipitate from the matrix, as evidenced by the coarse Cr2Ti and Zr5Al3 phase particles in the alloy with Cr1 after aging at 873 K. Importantly, the substitution of Mo0.5 for Cr0.5 induces the formation of coherent microstructure with cuboidal BCC nanoprecipitates with a size of 10–80 nm uniformly distributed into B2 matrix in 873 K-aged alloy with Cr0.5Mo0.5, which is contributed to the moderate lattice misfit (ε = 0.94%) between them. But this kind of cuboidal coherent microstructure is ambiguous in aged alloy with Mo1. By further increasing the aging temperature up to 1073 K, BCC/B2 microstructure becomes unstable due to the metastable feature of B2 phase in this kind system, and the Zr5Al3 phase becomes dominant. It is due to the cuboidal BCC/B2 coherent microstructure that renders the 873 K-aged Al2Ti6Zr2Nb3Ta2Cr0.5Mo0.5 alloy with an excellent mechanical property, exhibiting the highest compressive yield strength (σYS = 1314 MPa) and good plasticity among these alloys.