Macroscopic tensile properties of AlxCoCrCuFeNi (x = 0.3 and 1) before and after heat treatment

The relationship between morphology (grain-size order) and mechanical properties of AlxCoCrCuFeNi (x = 0.3 and 1) was evaluated. Both alloys were prepared using the arc-melting method. Crystal structures with x = 0.3, face-centered cubic (FCC), and body-centered cubic (BCC) structures were observed for x = 1, the typical crystal structures of AlxCoCrCuFeNi. The alloys prepared via arc melting exhibited two different microstructures: a Cu-rich region (area I) and an equiatomic region (area II). Microstructures of the as-fabricated specimen were homogenized through heat treatment at 600 and 1000 °C for 5 h. Although the homogenization of the microstructure proceeded, areas I and II remained after the heat treatment. In particular, the area I composition was inconsistent with the predicted phase diagram. Tensile tests for these alloys revealed that the tensile strength of x = 1 (∼310–1100 MPa) is higher than that of x = 0.3 (∼320–660 MPa), whereas the fracture strain of x = 1 (∼0.03–0.09) is lower than that of x = 0.3 (∼0.06–0.26). These results indicate that the increase in Al content caused an increase in strength and brittle fracture because it also caused an increase in the formation of the BCC and B2 phases, which required higher stresses for the movement of dislocations than the FCC phase. Because the tensile properties of AlxCoCrCuFeNi are comparable to those of conventional alloys, such as Ti alloys and steels, a design with a moderate composition for stronger and tougher AlxCoCrCuFeNi is required to apply high entropy alloys to structural materials.


Introduction
High-entropy alloys (HEAs) were proposed by Yeh et al as uniform solid-solution alloys containing five or more elements in the range of 5-35 at%. The configuration entropy of alloys (DS conf ) is expressed as follows [1,2]: where R, n, and x i denote the gas constant (8.314 J K −1 ·mol −1 ), number of elements, and atomic ratio (mol) of each component, respectively. Recently, the concept of high entropy has garnered considerable attention because alloys with ΔS conf 1.5R are considered HEAs, and the formation of intermetallic phases is allowed [3].
As a pioneering HEA material, the crystal structure and morphology of Al x CoCrCuFeNi have been extensively researched [1,2,[4][5][6][7][8][9][10][11] because the elements in these alloys are also components of conventional alloys, such as Al alloys (duralumin), Fe alloys (alloy steel), and Ni-based superalloys. The crystal structure of Al x CoCrCuFeNi significantly depended on the Al content. A face-centered cubic (FCC) structure with the L1 2 phase was observed for x = ∼0-0.6, and the ratio of body-centered cubic (BCC or B2) increased with increasing Al content from x = 0.6. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Material properties are affected by the ratio of FCC (or L1 2 ) to BCC (or B2). They are unique compared to conventional alloys because of the unpredictable interaction of elements in HEAs, which is well-known as the cocktail effect and lattice distortion owing to the difference in the atomic radii of the elements. Yeh et al [1] also reported that the morphology of Al x CoCrCuFeNi is a multiscale structure because nanostructured dendrite and inter-dendrite phases are formed in crystalline form with sizes in the order of micrometers, and the compressive yield strength of Al x CoCrCuFeNi increases with an increase in Al content and is related to the increase in BCC phases. Other researchers reported similar tendencies for hardness, yield strength, and compressive strength [9,12,13]. Shun et al [14] evaluated the tensile behavior of Al 0.3 CoCrCuFeNi, mainly composed of FCC (L1 2 ) phases, and the effect of heat treatment on the tensile behavior. They concluded that forming the nano B2 phase in the FCC phase increased the tensile strength. Thus, the relationship between the nanostructure of Al x CoCrCuFeNi and its mechanical properties has been clarified. However, the relationship between macroscopic morphologies, such as grain size, remains a challenge, and local compositional changes in the complex microstructure of Al x CoCrCuFeNi are not well-known. In addition, mechanical properties are obtained through nanoindentation, and compressive tests are often conducted because they are available for small specimens, even in recent reports [15,16]. To apply Al x CoCrCuFeNi as a structural material, evaluating its mechanical properties using tensile tests, macroscopic deformation, and fracture behavior is necessary because the relationship between the microscopic mechanical properties of grains obtained by indentation and the macroscopic mechanical properties obtained by conventional tensile tests is required. This study aims to clarify the effect of the microstructure on the macroscopic mechanical properties of Al x CoCrCuFeNi and the role of the metal phase (BCC and FCC) and intermetallic phase on the deformation behavior.

Sample preparation
In this study, we fabricated two types of HEAs with compositions of x = 0.3 and 1 for Al x CoCrCuFeNi. The composition was determined using the simulated phase diagram reported in a previous study (figure 1) [2]. In the case of x = 0.3 and 1, intermetallic phases were not formed, and this composition can simplify the evaluation of the relationship between the microstructures and tensile behaviors. Each metal chunk (Kojundo Chemical Laboratory Co., Ltd, Saitama, Japan) with a purity of 99.9% was used as raw material. Raw metals were melted in an arc-melting furnace (NEV-ADR03; Nisshin Giken Corp., Saitama, Japan). The as-melted samples were approximately 70 mm in diameter and 10 mm thick. The specimens for the tensile tests were machined from the as-melted samples, and their microstructural characteristics were graded along the thickness direction, as discussed in section 3.1. The specimens were classified as bottom, middle, or top, as shown in figures 3 and 4.
Heat treatment was performed in an inert atmosphere for 5 h using an electric furnace with a carbon heater to homogenize the microstructural heterogeneity originating from the difference in the cooling rates of the Cu hEarths. Prior to heating, Ar gas (99.999% purity) was introduced to the furnace. The temperature was set at 600 and 1000°C, increased at 50°C min −1 , and cooling was achieved through furnace cooling.

Microstructural characterization
The as-fabricated and heat-treated samples were mounted on an epoxy resin. The specimens were polished by using a standard metal procedure. The microstructures were observed by optical and scanning electron microscopy (OM, (Olympus) and SEM (TM3000, Hitachi High Technological Corp., Tokyo, Japan)). Elemental distribution was examined using energy-dispersive x-ray spectroscopy (EDS) combined with SEM. Crystalline structures were examined using an x-ray diffractometer (Miniflex-600, Rigaku, Tokyo, Japan) at a voltage of 40 kV. Measurements were performed at a rotational speed of approximately 2°min −1 . The crystal orientation of each grain was analyzed using electron backscattered diffraction (AZtecEnergy, Oxford, The U.K.).

Evaluation of mechanical properties
Tensile testing at room temperature (25°C) was conducted using a screw-driven universal testing machine (EZ-LX 5kN EZ-test, Shimadzu, Japan) through displacement control with the cross-head speed of 0.5 mm min −1 . The schematics of the tensile specimens are shown in figure 2. The length of each specimen is 25 mm. The gauge length and width were 15 and 2 mm, respectively. Three tests were conducted on both alloys. The indentation modulus and hardness were measured using a nanoindentation testing machine, and Young's modulus was determined from the slope of the unloading curve.

Results and discussion
3.1. Effect of cooling rate on the microstructural heterogeneities of HEAs The microstructures of the as-melted HEAs observed using OM are shown in figure 3. Similar to the bottom, middle, and top regions, for x = 0.3, the microstructure was composed of three different regions: a brighter minor region (area I), a darker major region (area II), and pores. Although area I was uniformly distributed at the bottom, a continuous network structure was observed in the middle. The evolution of areas I and II is attributed to the difference in the local composition. The different microstructures of areas I and II are caused by the difference in the cooling rate during arc melting. Schematic diagrams of the alloy microstructures are shown in figure 4. As the alloys were set on a water-cooled copper hEarth, the cooling rate of the bottom was the highest, and the cooling rate of the middle was the lowest because it did not contact the hEarth; cooling by radiation from the surface was negligible. The formation of non-uniform microstruture for HEAs fabricated by arc-melting was also reported [17]. For x = 1.0, similar microstructures were observed at the bottom and top. However, a continuous network was not observed in Area I. To characterize the microstructures in detail, SEM images with EDS mapping analyses of Ar, Co, Cr, Cu, Fe, and Ni are shown in figure 5(a) and (b). To clarify the differences between areas I and II in more detail, the results of the EDX point analyses are shown in figures 5(c) and (d). The distributions of Ni and Al were uniform compared to those of the other samples, and they existed in both areas I and II for x = 0.3 and 1. However, the Co, Cr, Cu, and Fe distributions were not uniform. A significant difference in the Cu content was observed: ∼50-60at% in area I and ∼10-20 at% in area II for both alloys. In addition, the amounts of Co, Cr, and Fe in area II are almost twice those in area I (similar to the composition of the starting materials). Thus, areas I and II can be defined as the Cu-rich and equiatomic regions, respectively. XRD analyses for x = 0.3 and 1 with references to each element in the alloys are shown in figure 6. The XRD patterns clearly show that the crystal structures for x = 0.3 and 1 were similar, and the XRD patterns for the FCC structure were observed. For x = 0.3, XRD patterns indicate that the crystal structure is an FCC structure and the existence of Cu-and Ni-based FCC structures because the peaks at 2θ = 42°-45°are identified as the (111) surface in the FCC structure. This was expected because the Cu-based FCC structure was attributed to the existence of area I, which is a Cu-rich region, and the crystal structure of CoCrFeNi at room temperature has been reported to have an FCC structure [18,19]. In addition to these different FCC structures, a BCC structure was observed for x = 1 because of the relative intensity of the peaks at ∼45°, indicating that the diffraction from the (111) surface was higher than that at ∼42°-43°at x = 0.3. I peaks at ∼65°and ∼82°-83°also indicated the existence of BCC structures. Thus, area II at x = 1 comprises Ni-based FCC and BCC structures. The distributions of indentation modulus and Vickers hardness (HV) are shown in figure 7. The distributions of the modulus and hardness reflect the microstructures of the area I and II alloys. The indentation modulus of area II for both alloys was higher than that of area I, which was expected because Young's modulus of Cu (128 GPa) was significantly lower than that of Co, Cr, and Fe (∼200-250 GPa) [20]. Notably, the hardness values of areas I and II were inversely correlated.

Effect of heat treatment on the microstructural heterogeneities of HEAs
Optical micrographs for x = 0.3 and 1 heat treated at 600 and 1000°C are shown in figures 8(a) and (b), respectively. Although areas I and II, defined for the as-fabricated samples, were retained after heat treatment, their distribution differed from those of the as-fabricated samples.
After heat treatment at 600°C for 5 h, significant microstructural changes are not observed for the samples, indicating that heat treatment at 600 o C for 5 h is insufficient to form uniform microstructures for both alloys.  After heat treatment at 1000°C for 5 h, the aggregation of areas I and II was observed at the top, middle, and bottom of both alloys. These results were expected because the compositions of areas I and II were different. The microstructural irregularity at the bottom compared to the top and middle still remained even after heat treatment at 1000°C. The crystal structures of the alloys were investigated to evaluate their microstructures after heat treatment, as shown in figures 9(a) and (b). BCC structures evolved in x = 0.3 after heat treatment at 600°C and disappeared after heat treatment at 1000°C. Notably, ordered crystal structure (B2 phase) is observed for x = 1 after heat treatment at 1000°C in addition to FCC and BCC structures. The formation of the B2 phase has been predicted in a previous study, and some researchers have confirmed this experimentally [2]. However, the B2 phase is typically observed when the Al content exceeds 25 at% (x > 1.5). This indicates that the concentration irregularity of Al caused the formation of the B2 phase, and heat treatment accelerated the difference in Al content between areas I and II.

Tensile test
The nominal stress-nominal strain curves of the as-melted and heat-treated HEAs are shown in figure 10. To eliminate microstructural irregularities, uniform regions (top and middle) were selected for the tensile specimens. When x was 0.3, significant plastic deformation was observed; the strain at the fracture was ∼0.06-0.26, whereas the tensile strength was ∼320-660 MPa. The effect of heat treatment on strength and fracture behavior was limited. For x = 1, the strength of the samples was 310-1100 MPa, and the fracture elongation was ∼0.03-0.09. The effect of heat treatment on tensile strength was also limited. Notably, a brittle fracture behavior was observed for x = 1 with and without heat treatment. The fracture behavior was attributed to the microstructures and crystal structures of the alloys. The FCC and BCC phases exhibited plastic deformation owing to dislocation movement. Generally, crystals undergo plastic deformation when the shear stresses in the crystal reach the critical resolved shear stress (t CRSS ). Ohashi et al proposed the following equation as an expression of t CRSS for a slip system (a) [21]: where t a T 0 ( ) ( ) and W ab ( ) denote the temperature dependence of deformation resistance caused by obstacles others than dislocation (dissolved atoms and Peierls potential) and interaction matrix between slip system a and b, respectively. The terms a, m, b, r b , S ( ) b, and d indicate numerical constant (0.3-0.6), shear modulus, length of Burgers vector, dislocation density, size-dependent constant (∼1), and grain size, respectively. The evaluation of t CRSS for crystal structures in alloys prepared in this study is challenging because it depends on several parameters, and lattice distortion occurs owing to the presence of constituent elements with different atomic radii. To briefly compare the deformability of the crystals, the Peierls-Nabarro stress (t PN ), which is required to move a straight dislocation in a slip system at 0 K, is usually considered and expressed as follows [22]: where n and h denote the Poisson's ratio and lattice spacing, respectively. A Generalized P-N plot of the estimated Peierls stresses for the FCC, BCC, and B2 structures is shown in figure 11 [22][23][24]. As a representative brittle material, α-Al 2 O 3 is also shown. The term t PN is an estimate determined using the experimental values of t CRSS and atomic molecular dynamics [22][23][24]. It clearly indicates that the normalized t PN by m for FCC is significantly smaller than that for BCC and B2. The typical fracture surfaces after the tensile tests are shown in figure 12. For x = 0.3, the FCC phase was the major crystal structure before and after the heat treatment. Because the stress required for the movement of dislocations in the FCC phase was significantly less than that in the BCC and B2 phases, a large elongation due to plastic deformation was observed. Dimples, typical microstructures that exhibit plastic deformation, were observed on the fractured surface at x = 0.3. In contrast, for x = 1, the number of BCC phases was greater than that for x = 0.3. River patterns, evidence of brittle fractures, are observed on the fracture surface at x = 1. Because the value of HV is a qualitative index for the capacity of plastic deformation, the relationship between the microstructures and hardness also shows a difference in the enlargement of both alloys because the HV for x = 1 is higher than that for x = 0.3. These results indicate that the strength and fracture behavior of Al x CoCrCuFeNi significantly depends on the Al content because the crystal structure of the alloys changes from FCC to BCC to an ordered structure (B2) with increasing Al content. To summarize the tensile properties of AlxCoCrCuFeNi obtained in this study, the relationship between the fracture elongation and tensile strength is shown in figure 13 [25][26][27]. These results indicate that Al x CoCrCuFeNi is a potential candidate for structural materials because its tensile properties are comparable to those of conventional alloys designed as structural materials. The tensile strength of Al x CoCrCuFeNi was higher than that of conventional Al and Mg alloys and reached that of Ti alloys. In particular, the strength of Al x CoCrCuFeNi was equal to those of conventional steel (x = 0.3) and martensite-phase stainless steel (x = 1). For the other HEAs, alloys with the BCC and B2 phases also showed less fracture elongation than those containing the FCC phase. The fracture elongation and tensile strength of alloys in similar systems (Al x CoCrFeNiMn) strongly depend on the Al content, which suggests that the formation of FCC (and/or BCC) is probably controllable by changing the Al content of alloys in these systems.