Phase, microstructure and compressive properties of refractory high-entropy alloys CrHfNbTaTi and CrHfMoTaTi

1.   Introduction

With the development of the aerospace, energy generation, and petrochemical industries, the demand for structural materials that can be applied in high-temperature environments has increased. Ni-based superalloys have been proposed to cater to the corresponding rigorous operation conditions of these industries, but their incipient melting point near 1300°C greatly limits their applications to those requiring temperatures between 1160 and 1277°C [1−3]. Similar to most conventional refractory metals or alloys, the design of Ni-based superalloys is based on a simple idea: one element with attractive properties is used as the principal element, and other elements are added in small amounts to improve these properties [4]. Such a mature technology offers a finitude of modification possibilities [5].

Previous studies introduced high-entropy alloys (HEAs) consisting of at least five principal elements each with atomic concentration between 5at% and 35at% [6−7]. Then, refractory HEAs with Zr, Ti, Hf, Cr, Ta, Nb, W, Mo, and V as the most commonly used elements were first proposed by Senkov et al. [8−9]. The first two reported refractory HEAs, VNbMoTaW and NbMoTaW, possess a simple body-centered cubic (BCC) microstructure and prominent solution strengthening [8−9]. The two alloys also have high yield strengths of 477 and 405 MPa at elevated temperature of 1600°C, respectively; however, their drawbacks include a density much higher than 9.0 g/cm³, room temperature (RT) brittleness (fracture strain of about 1.7%), and a low RT yield strength of about 1000 MPa. Subsequently, many new refractory alloys have been developed [10−14]. The most well-known refractory alloy HfNbTaTiZr has a BCC structure with excellent ductility exceeding 50% but much lower strength than the first two refractory HEAs [10–11].

To improve the strength of refractory HEAs based on HfNbTaTiZr, Chen et al. [15] investigated the mechanical behavior of HfNbTaTiZr alloy with a wide range of grain size. They found that the alloy with the smallest grain size of 38 μm possesses the highest yield strength of 958 MPa, which is only slightly higher than the initial value of 929 MPa. Modifying the composition of alloying elements, such as minor element addition and element substitution, is a conventional and effective way to alter material properties. Senkov et al. [16] evidenced that the Hf in HfNbTaTiZr partially substituted by Al reduces the density of the alloy by 9%, increases the RT hardness by 29%, and enhances the yield strength by 98%. Lin et al. [1] comprehensively investigated Al addition effects on the mechanical properties and found that the yield strength of equiatomic HfNbTaTiZrAl alloy is improved to about 1500 MPa under compression. Inspired by these, Guo et al. [4] reported a new refractory HEA, HfMoNbTiZr, in which Mo substituted for Ta in comparison to HfNbTaTiZr. This new alloy shows a significantly enhanced yield strength up to 1575 MPa and excellent ductility [4]. Fazakas et al. [17] have recently studied Cr effects on the mechanical properties of Ti₂₀Zr₂₀Hf₂₀Nb₂₀Cr₂₀ refractory HEA and attributed the optimized hardness and strength to the segregation of Cr-containing Laves phases (Cr₂Hf, Cr₂Nb) during casting.

Cr tends to form a Laves phase with Hf, Nb, or Ta, which promotes strength, and Cr is also more cost-efficient than other component elements (Hf, Nb, Ta, Ti, and Zr). Thus, in the currently studied work, a new CrHfNbTaTi refractory HEA was designed by replacing Zr with Cr in HfNbTaTiZr. The modulus of Mo is three times more than that of Nb despite their similar densities [18]. Consequently, the substitution between them is expected to greatly increase the mechanical properties of alloys. Hence, another new alloy, CrHfMoTaTi, was derived from CrHfNbTaTi. The phases, microstructures, and mechanical properties of the new refractory HEAs CrHfNbTaTi and CrHfMoTaTi at RT were investigated and discussed.

2.   Experimental

CrHfNbTaTi and CrHfMoTaTi alloys were prepared using vacuum arc melting of equimolar mixtures in a Ti-getter high-purity argon atmosphere. Cr, Hf, Nb, Mo, Ta, and Ti in bulk form slugs were used with a purity of 99.99wt%, 99.9wt%, 99.95wt%, 99.95wt%, 99.95wt%, and 99.9wt%, respectively. During melting, ingots were melted for 3 min in a water-cooled copper crucible and turned over once after each melting process with electromagnetic stirring. This procedure was repeated at least four times to ensure that the alloys were in a well-mixed state, and finally a button sample with the size of 8 mm in thickness and 15 mm in diameter formed. The weight loss of each alloy during arc melting was lower than 0.2%, as indicated by the absence of volatile elements in the alloys. Ingot pieces weighing 10 g were melted by arc melting and then sucked into an RT cylinder-shaped copper mold (ϕ4 mm × 60 mm). Table 1 shows the as-prepared alloy compositions of the CrHfNbTaTi and CrHfMoTaTi cylindrical specimens, whose actual compositions were very close to the designed compositions.

The crystal structure of the as-cast cylindrical samples was identified via X-ray diffraction (XRD) using a PANalytical X’Pert Powder diffractometer with Cu K_α radiation operating at 40 kV/40 mA and a scanning rate of 3°/min. The microstructure was analyzed using a field emission scanning electron microscope (FESEM, Zeiss sigma 500, GER). Compressive tests were carried out on a computer-controlled mechanical testing machine (Instron, Norwood, MA, USA) using ϕ3.7 mm × 5.6 mm cylindrical samples at a constant ramp speed of 5.6 × 10⁻³ mm/s and an initial strain rate of 10⁻³ s⁻¹.

Table  1.  Nominal and actual compositions of the studied alloys in this work at%

Composition		Cr	Hf	Nb	Mo	Ta	Ti
CrHfNbTaTi	Nominal	20	20	20	—	20	20
	Actual	19.4	19.9	21.7	—	18.7	20.3
CrHfMoTaTi	Nominal	20	20	—	20	20	20
	Actual	20.5	21.6	—	18.2	17.7	22.0

 | Show Table

DownLoad: CSV

3.   Results and discussion

3.1   Phase analysis

The XRD patterns of the as-cast cylindrical CrHfNbTaTi and CrHfMoTaTi samples in this work and as-solidified HfNbTaTiZr alloy extracted from Ref. [19] are plotted in Fig. 1. The reference HfNbTaTiZr alloy consists of a single BCC phase (Fig. 1(c)). By comparison, the CrHfNbTaTi alloy possesses more complicated phases even though only one principal element, Zr, is replaced by Cr. This finding implies that Cr plays an important role in determining the phases of this type of alloys. In specific, the as-cast CrHfNbTaTi alloy contained three phases (i.e., BCC1, BCC2, cubic Laves phase (see Fig. 1(a)). The marginal peaks of the BCC2 phase compared with the BCC1 phase evidences that the amount of BCC1 occupied the most of the BCC phase. If all peaks are considered, then the BCC phases will be the major phases in the CrHfNbTaTi refractory HEA. In addition, according to the JCPDS cards, the BCC2 phase matched well with TiCr and the Laves phase was consistent with a Cr-rich Laves phase (Cr₂(Hf,Ti,Ta,Nb)), as well as that in the CrHfNbTiZr alloy reported in Ref. [17]. This kind of Laves phase is actually not rare, and as an example, Inoue [20] has documented this type of Laves phase and its origination.

Fig. 1.  XRD patterns of CrHfNbTaTi (a) and CrHfMoTaTi (b) alloys in the as-cast condition together with that of the as-cast HfNbTaTiZr alloy (c) reported in Ref. [19].

DownLoad: Full-Size Img PowerPoint

The largest atomic size ratio between two principal elements in an alloy larger than 1.225 is considered beneficial to form Laves phase. Thus, the presence of the Cr-rich Laves phase in the CrHfNbTaTi alloy is spontaneously anticipated because of the large atomic size difference among Cr (124.9 pm), Ti (146.15 pm), and Hf (157.8 pm). Compared with the most concerned HfNbTaTiZr alloy, the XRD diffraction peaks of the BCC1 phase in the currently studied CrHfNbTaTi alloy moved toward higher angles, implying that the lattice parameter of the primary BCC phase reduced compared with that of the BCC phase in HfNbTaTiZr. On the basis of the XRD patterns, the lattice parameters of BCC in HfNbTaTiZr and CrHfNbTaTi were determined to be a = 340.4 pm and a = 334.1 pm, respectively. The difference of the BCC lattice parameters in the two alloys might originate from the small atomic size of Cr (124.91 pm) compared with the Zr element substituted (160.3 pm) [17]. Moreover, the lattice parameters of the BCC2 and Laves phases of the studied CrHfNbTaTi alloy were determined to be 311.6 and 720.1 pm, respectively.

To unveil the effect of introducing an element with significantly different intrinsic features on the phase component, this study further replaced Nb with Mo to change CrHfNbTaTi to CrHfMoTaTi. Both elements (Nb and Mo) possess a BCC structure at RT, a relatively low density, and similar melting temperature, but the Young’s modulus of Mo is almost three times over that of Nb. The related element properties of the constituent elements in the studied alloys are listed in Table 2. After mutual substitution, despite the presence of a few unknown peaks in the XRD patterns, the number of phases dramatically increased from three to six (Fig. 1(b)). These six phases were determined to be four BCC phases (BCC1, BCC2, BCC3, and BCC4) and two laves phases (Laves1 and Laves2). According to the JCPDS cards, the BCC1, BCC2, BCC3, and BCC4 phases matched well with TaMo, Hf₉Mo, MoTi/TiCr, and MoCr, respectively. This finding suggests that the BCC1, BCC3, and BCC4 phases were mainly enriched with Mo, whereas the BCC2 phase was enriched with Hf. Their lattice parameters in the former order were determined as a = 325.3, 339.9, 316.6, and 300.3 pm, respectively, which were determined by the atomic size of the predominant and secondary elements (Hf: 157.75 pm, Ti: 146.15 pm, Ta: 143.00 pm, Mo: 136.20 pm, and Cr: 124.91 pm). As a Cr-containing alloy, the Cr-rich C15 Laves phase (Laves1, Cr₂(Hf,Ta)) was identified with a lattice parameter of a = 704.3 pm, with a same structure to the C15 Laves phase in CrHfNbTaTi but a different composition (apparently lacking Ti and Nb). This finding was in good agreement with the former BCC phase analysis, in which the incorporation of many Ti and Nb atoms formed BCC phases; however, another Laves phase (Laves2) was also interestingly matched well with a C15 Laves phase Mo₂Hf, which was unrelated to Cr. Its lattice parameter was up to 746.0 pm. The substitution of Mo for Nb facilitated the formation of Mo-containing BCC phases by affecting the formation of the Cr-containing Laves phase and by directly participating in the formation of the BCC phases.

Table  2.  Crystal structures, atomic radius (r), and Young’s modulus (E) of the elements in the studied alloys

Element	Crystal structure	r / pm	E / GPa
Cr	BCC	124.91	279
Hf	HCP	157.75	78
Nb	BCC	142.90	105
Mo	BCC	136.20	329
Ta	BCC	143.00	186
Ti	HCP	146.15	116

 | Show Table

DownLoad: CSV

3.2   Microstructure

To monitor the microstructural evolution as the compositional modification, morphology was characterized for the polished CrHfNbTaTi and CrHfMoTaTi samples along the cross-section, as shown in Fig. 2. A typical dendrite microstructure with an average primary arm size of around 5 μm was observed in the as-cast CrHfNbTaTi alloy, as shown in Fig. 2(a) and (b). The bright dendrites (marked as D) were irregularly surrounded by grey continuous-corridor interdendritic regions (marked as ID). The occupational regions of the dendrites were larger than those of the remaining interdendritic regions, indicating that the dendrites might be one of the BCC phases. By contrast, on basis of the phase analysis, the interdendritic regions likely corresponded to Cr-rich Laves phases, which showed that the BCC phases were the major phase in CrHfNbTaTi. The chemical compositions for each region were determined from the EDS analysis, as summarized in Table 3. It was found that the Cr and Hf were the main elements in the interdendritic regions, due to which it was suggested that the interdendritic regions were constituted by Cr-rich Laves phase. In addition, Cr has the lowest melting temperature and thus solidified at the last stage of solidification process, resulting in forming slim and continuous interdendritic regions. Likewise, the dendrites were mainly enriched in Ta, Ti, and Nb. On basis of the JCPDS cards, the prominent peaks of the two BCC phases suggests that both the phases approximately corresponded to Ta- and Nb-rich compounds, which is consistent with the elemental distribution of the dendrites. Combining these results, it implies that the dendrites corresponded to the BCC phases.

Fig. 2.  SEM micrographs of the microstructure of CrHfNbTaTi (a) and CrHfMoTaTi (c) in the as-cast condition. (b) and (d) are the magnifications of the local regions of (a) and (c), respectively.

DownLoad: Full-Size Img PowerPoint

Table  3.  Quantitative chemical analysis of as-cast CrHfNbTaTi and CrHfMoTaTi samples at%

Alloy	Region in Fig. 2	Phase	Cr	Hf	Nb	Mo	Ta	Ti
CrHfNbTaTi	D	BCC1, BCC2	5.2 ± 0.6	20.8 ± 0.3	22.2 ± 0.2	—	25.3 ± 1.1	26.6 ± 0.3
	ID	Laves	44.9 ± 5.4	21.4 ± 0.5	6.7 ± 1.6	—	15.8 ± 1.7	11.2 ± 2.6
CrHfMoTaTi	D	BCC1, BCC3, BCC4	11.2 ± 0.4	14.4 ± 0.3	—	35.8 ± 0.2	17 ± 0.4	21.6 ± 0.1
	ID1	BCC2 + Laves1, Laves2	30.2 ± 0.3	27.8 ± 0.1	—	19.0 ± 0.6	5.9 ± 0.2	17.1 ± 0.9
	ID2	BCC3	17.6 ± 1.9	24.8 ± 0.5	—	13.3 ± 0.4	5.3 ± 0	38.9 ± 2.9

 | Show Table

DownLoad: CSV

After replacing Nb with Mo in CrHfNbTaTi, the difference between the two typical distinct regions is recognizable in CrHfMoTaTi (Fig. 2(c) and (d)). In detail, the microstructure of the as-cast CrHfNbTaTi alloy showed a typical fishbone-like dendritic grain (bright part) with blunt edges tightly embedded into a continuous matrix (grey part) in Fig. 2(a) and (b). By contrast, the dendritic regions of CrHfMoTaTi marked as D appeared more similar to equiaxed grains, and the interdendritic regions contained two parts: a bright part marked as ID1 and a grey part marked as ID2. Moreover, the interdendritic shell regions that surrounded the dendritic regions were often incompact, and the thickness of shell was around 1 μm. The chemical compositions of all regions were determined from the EDS analysis and are summarized in Table 3. The interdendritic regions (ID1 and ID2) were rich in Cr, Hf, and Ti, whereas the dendritic regions (D) were rich in Mo and Ta. Similar to CrHfNbTaTi, the interdendritic regions in CrHfMoTaTi were also mainly ascribed to Laves phase, whereas the dendritic regions were BCC phases.

EDS mapping was also carried out to determine the distribution of each element (Fig. 3). According to mapping, 1–2 μm-thick transition layers (marked with white arrows in the upper Ti map in Fig. 3(a)) enriched with Ti were observed at the interface of the bright dendritic regions in CrHfNbTaTi. This result was also observed with the BCC crystal structures in the hot isostatically pressed NbCrMo_0.5Ta_0.5TiZr alloy [21]. Similarly, part of the ID2 regions with 2–3 μm in size (marked with red arrows in the lower Ti map in Fig. 3(b)) rich in Ti were observed in CrHfMoTaTi. Combined with the XRD analysis, they corresponded to the BCC2 TiCr phase of CrHfNbTaTi and the BCC3 TiCr phase of CrHfMoTaTi. The overall compositions measured by EDS were much close to the nominal stoichiometric compositions, as listed in Table 1, which further confirmed the reliable melting of this composition.

Fig. 3.  SEM images and EDS mapping of as-cast CrHfNbTaTi (a) and CrHfMoTaTi (b) HEAs.

DownLoad: Full-Size Img PowerPoint

3.3   Mechanical properties

Fig. 4(a) shows the engineering stress $\left(\sigma \right)$ vs. engineering strain (ε) curves for the as-cast CrHfNbTaTi and CrHfMoTaTi alloys obtained during compression testing at RT, together with that of as-solidified HfNbTaTiZr alloy studied in Ref. [19]. The compressive curve of the alloy CrHfMoTaTi was offset along the axis to display clearly. The compressive properties, such as yield strength (${\sigma }_{0.2})$, ultimate compressive strength ${(\sigma }_{\mathrm{b}})$, and plastic strain $({\varepsilon }_{\mathrm{p}})$, of these alloys are given in Table. 4. For CrHfNbTaTi, the yield strength ${\sigma }_{0.2}$ was (1258 ± 15) MPa. After yielding, the as-solidified HfNbTaTiZr and CrHfNbTaTi alloys showed obvious work-hardening behavior, which is commonly evaluated by the instantaneous work hardening index (n^*). This parameter is defined as n^*$=\mathrm{d}\left(\mathrm{l}\mathrm{n}{\sigma }_{\mathrm{T}}\right)/\mathrm{d}\left(\mathrm{l}\mathrm{n}{\varepsilon }_{\mathrm{T}}\right)$, based on the common power-law relationship ${\sigma }_{\mathrm{T}}$ = k${\varepsilon }_{\mathrm{T}}^n$, where ${\sigma }_{\mathrm{T}}$ and ${\varepsilon }_{\mathrm{T}}$ are the true stress and true strain, respectively, and n is the work hardening index [22].

Fig. 4.  (a) Compressive engineering stress–strain curves of the as-cast CrHfNbTaTi, CrHfMoTaTi, and the reported HfNbTaTiZr [19] alloys and (b) instantaneous work hardening index (n^*$={\bf d}\left({\bf {ln}}{\boldsymbol \sigma }_{{\bf T}}\right)/{\bf d}\left({\bf {ln}}{\boldsymbol \varepsilon }_{{\bf T}}\right)$ ) vs. true strain curves of CrHfNbTaTi and HfNbTaTiZr alloys. The compressive curve of the alloy CrHfMoTaTi was offset along the axis for clear illustration.

DownLoad: Full-Size Img PowerPoint

Table  4.  Compressive yield strength (${\boldsymbol \sigma }_{\boldsymbol {0.2}})$, ultimate compressive strength $({\boldsymbol \sigma }_{\bf {b}})$, and plastic strain $({\boldsymbol \varepsilon }_{\bf {p}})$ of CrHfNbTaTi, CrHfMoTaTi, and HfNbTaTiZr [19]

Alloy	${\sigma }_{0.2}$ / MPa	${\sigma }_{\mathrm{b}}$ / MPa	${\varepsilon }_{\mathrm{p}}$ / %
CrHfNbTaTi	1258	2061	15.0
CrHfMoTaTi	1051	1051	0.8
HfNbTaTiZr [19]	929	—	>50

 | Show Table

DownLoad: CSV

The instantaneous work hardening as a function of true strain was calculated for the as-cast samples of HfNbTaTiZr and CrHfNbTaTi by using compression test data. Fig. 4(b) shows that the instantaneous work hardening of HfNbTaTiZr and CrHfNbTaTi exhibited two stages. First, the instantaneous work hardening sharply decreased at the true strain ${\varepsilon }_{\mathrm{T}}$ = 0.7%–2.6% for HfNbTaTiZr and ${\varepsilon }_{\mathrm{T}}$ = 3.8%–13.8% for CrHfNbTaTi (stage I). This decrease was followed by a gradual increase stage (${\varepsilon }_{\mathrm{T}}$ ≥ 2.6%) of the HfNbTaTiZr alloy and a stable stage at true strain ${\varepsilon }_{\mathrm{T}}$ = 14.7%–21.1% of CrHfNbTaTi (stage II). The classic decrease of n^* in stage I was primarily attributed to dislocation gliding during the initial deformation stage. In stage II, the n^* value remained stable at about 0.47 for CrHfNbTaTi. The dynamic strain aging devoting to work hardening involved the suppression of microvoids/microcracks at this stage [23]. Thus, its strength eventually reached the maximum value of 2061 MPa before fracture. However, after replacing Nb with Mo, CrHfMoTaTi showed a typical brittle fracture at 1051 MPa with no work hardening behavior during compression tests.

Compared with that of the previously reported HfNbTaTiZr alloy, the yield strength of CrHfNbTaTi alloy dramatically increased from 926 to 1258 MPa, and their promising plasticity was retained. In addition, CrHfNbTaTi possessed a better strength−ductility trade-off than CrHfNbTiZr [17], CrNbTiVZr [24], HfMoNbZrTi [4], HfMoTaTiZr [25], and MoNbTaVW [9]. A close relationship exists between the microstructure and mechanical properties of an alloy. The above phase analysis indicates that the alloy changed from a single BCC phase to multiple phases, including BCC and Cr-containing Laves phases, after substituting Zr with Cr. The BCC phase is the major/only phase in the CrHfNbTaTi/HfNbTaTiZr alloys; hence, dislocation movements are sensitive to the lattice plane distance (d) of the BCC phase because a wider d is beneficial to the dislocation movement and subsequent formation of stress fields from dislocations [26]. Formation of additional stress fields increases resistance to dislocation movement and improves solid-solution strengthening behavior. The largest lattice plane distance in a BCC structure is {110}, which can be calculated from the actual lattice parameters a by d = $a/ \sqrt{{h}^{2}+{k}^{2}+{l}^{2}}$. Thereinto, the d₁₁₀ of CrHfNbTaTi (236.2 pm) is smaller than that of HfNbTaTiZr (240.7 pm) because of the substitution of Cr with a smaller atomic radius for Zr with a larger one. This result indicates that the Cr-containing Laves phase contributed to the improved yield strength of CrHfNbTaTi. Moreover, Cr-containing Laves phases are inherently brittle and strong [18], and they fill in the interdendritic regions and form a network. As a result, a mass of dislocations that take responsible for the main deformation might accumulate or stack at the boundary of the grains, significantly enhancing the strength. The interdendritic regions containing Laves phases rich in Cr actually also contain other elements. Thus, the phases in the interdendritic regions were relatively more complex, especially the thin transition layers that coated the dendritic phase, which might play an important role in the combination of strength and plasticity. This type of complexity might decrease the stiffness of the framework, resulting in a proper plasticity. For CrHfMoTaTi, the deterioration of the mechanical strength after replacing Nb with Mo in CrHfNbTaTi might be closely related to the incompact coupling between the dendritic and interdendritic regions, which facilitated the formation of poles. The poles are a type of crack initiators playing an important role in determining the brittle failure behavior at the initial deformation stage, especially in the brittle and strong Laves phase.

4.   Conclusion

In this work, the phase constitutions, microstructures, and compressive properties of the as-cast CrHfNbTaTi and CrHfMoTaTi alloy were investigated. CrHfNbTaTi was mainly composed of two BCC and one Cr-containing Laves phases, whereas CrHfMoTaTi consisted of six phases: four BCC and two Laves phases. The Laves phases in both alloys possessed a C15 structure. Compared with the reference alloy HfNbTaTiZr, replacing Zr with Cr increased the yield strength from 926 to 1258 MPa while retaining a plastic strain of around 15.0% in CrHfNbTaTi alloy. The enhanced strength might stem from the constraining effect of the strong Laves phase framework. Furthermore, the component complexity of the framework decreased the stiffness, which was responsible for the promising ductility. The CrHfMoTaTi alloy showed a typical brittle fracture with a strength of 1051 MPa, which might stem from the presence of multiple phases and the dendritic regions surrounded by an incompact interdendritic shell after the incorporation of Mo.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (No. 51604173) and the Natural Science Foundation of Jiangsu Higher Education Institution of China (No. 18KJB430012).

Conflict of Interest

The present work has no conflicts with other researchers or organizations.