Accepted Manuscript

Phase Stability and Tensile Properties of Co-free Al0.5CrCuFeNi2 High-EntropyAlloys

Chun Ng, Sheng Guo, Junhua Luan, Qing Wang, Jian Lu, Sanqiang Shi, C.T.Liu

PII: S0925-8388(13)02268-8DOI: http://dx.doi.org/10.1016/j.jallcom.2013.09.105Reference: JALCOM 29459

To appear in:

Received Date: 22 August 2013Revised Date: 15 September 2013Accepted Date: 16 September 2013

Please cite this article as: C. Ng, S. Guo, J. Luan, Q. Wang, J. Lu, S. Shi, C.T. Liu, Phase Stability and TensileProperties of Co-free Al0.5CrCuFeNi2 High-Entropy Alloys, (2013), doi: http://dx.doi.org/10.1016/j.jallcom.2013.09.105

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customerswe are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, andreview of the resulting proof before it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

Phase Stability and Tensile Properties of Co-free Al0.5CrCuFeNi2

High-Entropy Alloys Chun Ng a, Sheng Guo b,1, Junhua Luan b,c, Qing Wang b, Jian Lu b, Sanqiang Shi a,2, and C. T.

Liu b,3 a Department of Mechanical Engineering, The Hong Kong Polytechnic University, Hung Hom,

Kowloon, Hong Kong, P. R. China b Center of Advanced Structural Materials, Department of Mechanical and Biomedical

Engineering, City University of Hong Kong, Kowloon, Hong Kong, P. R. China c School of Materials Science and Engineering, University of Science and Technology Beijing,

Beijing 10083, P. R. China

Abstract: High-entropy alloys (HEAs) are becoming new research frontiers in the metallic

materials field. The phase stability of HEAs is of critical significance, but a convincing

understanding on it has been somewhat held back by the slow diffusion kinetics, which

prevents the completion of diffusion assisted phase transformations toward the equilibrium

state. Here a unique methodology, combining both the thermomechanical treatments and

thermodynamic calculations, was employed to reveal the phase stability of HEAs, exemplified

using the newly developed Al0.5CrCuFeNi2 alloy. The metastable nature of the solid solution

phases in this high-entropy alloy was uncovered through thermomechanical treatments

induced phase transformations, and supported by the thermodynamic calculations. Meanwhile,

the tensile properties for both the as-cast and thermomechanically treated alloys were

measured, and correlated to their indentation behavior.

Keywords: High entropy alloys; Thermomechanical treatments; Phase stability;

Thermodynamics calculations; Tensile tests; Indentation behavior

1 Current address. Tel: +46 317721254; fax: +46 317721313; e-mail: sheng.guo@chalmers.se. 2 Corresponding author. Tel: +852 2766 7821; fax: +852 2365 4703; e-mail: mmsqshi@polyu.edu.hk. 3 Corresponding author. Tel: +852 3442 7213; fax: +852 3442 0172; e-mail: chainliu@cityu.edu.hk.

2

1. Introduction

High-entropy alloys (HEAs), or multi-component alloys with equiatomic or close-to-

equiatomic compositions, are becoming new hot spots in the metallic materials community [1,

2]. The configuration entropies of HEAs, assuming the alloys in a fully random state (liquid

or a random solid solution), are much higher than those of conventional alloys with one or at

most two principal elements [3], and here comes the definition of HEAs. It needs to be

emphasized here that the definition of HEAs comes simply from the compositional

considerations; the definition does not indicate that the achieved phases in HEAs would

naturally have a high-entropy state, for example, being the amorphous phase [4-8] or a single-

phase random solid solution [9, 10]. Depending on the alloy compositions, HEAs can possess

many interesting mechanical and physical properties [11], and particularly they have great

potentials to be used as high temperature materials, or coating materials requiring high

hardness and high wear resistance.

The alloying strategy in HEAs is a breakthrough in the history of physical metallurgy, as

in the past, alloys normally have only one (e.g., steels, Al alloys and Cu alloys) or two

principal elements (e.g., TiAl based alloys). It was presumed that the high concentration of

alloying elements would cause the formation of many hard intermetallic compounds, and

hence embrittle the materials. Somewhat surprisingly, in many HEAs intermetallic

compounds are essentially absent, and only simple solid solutions form, with fcc (face

centered cubic) and/or bcc (body centered cubic) crystal structures [12-16]. It is emphasized

here again that the formation of solid solution phases does not mean the formation of a single-

phase solid solution, which is rather rarely seen [9, 10]. It was shown previously that, when

the solid solutions are the sole alloying products in HEAs, the formation of fcc or bcc

structured solid solutions upon alloying from the molten liquid state can be reasonably

controlled by tuning the valence electron concentration (VEC) [17]. Based on the VEC

3

scheme, we developed a series of AlxCrCuFeNi2 alloys [18], by fully replacing the expensive

Co element with Ni in the extensively studied AlxCoCrCuFeNi alloys [19]. The target alloy

used in this work, Al0.5CrCuFeNi2, is from this series of newly developed HEAs.

The phase stability, particularly the phase stability of the solid solutions, is a critical

issue for HEAs. It has been shown previously that the solid solution phases obtained in the

cast HEAs are actually the firstly formed solid phases upon cooling from the molten liquids

[20]. Their preferred formation over intermetallic compounds is due to the large entropic

contribution to the Gibbs free energy at high temperatures [20, 21]. Furthermore, these high-

temperature stable solid solutions can be kept to the room temperature, in a way similar to the

freezing of the amorphous phase from the molten liquid [22]. The probable mechanism is that

the diffusion kinetics in HEAs is very sluggish [1, 20], which prevents the phase

transformations towards equilibrium phases within the timescale of the cooling process. The

sluggish diffusion mechanism was recently experimentally confirmed by the diffusion couple

method [23]. Bearing in mind the metastable nature of the solid solutions in the cast HEAs, it

is natural to ask questions like: What are the equilibrium phases then? And how (thermally)

stable are these metastable solid solutions? Conceivably, when the temperature decreases and

hence the entropic contribution to the Gibbs free energy becomes less significant, the solid

solutions might not be energetically stable. Intermetallic compounds are thus highly possible

to form at intermediate temperatures. However, their formation could be kinetically unfavored,

considering the sluggish diffusion kinetics in HEAs. Revealing the phase stability in HEAs

thus becomes a challenging topic, and it forms the main topic of this work.

Our methodology is to combine experiments and thermodynamic calculations.

Experimentally, the as-cast Al0.5CrCuFeNi2 alloys were cold rolled and then annealed

extensively at intermediate and high temperatures (700 to 1100 oC), to see how the metastable

solid solutions in the as-cast condition evolve with thermomechanical treatments. The cold

4

work is required to provide the driving energy for the phase transformation that occurs during

the recrystallization process. The thermodynamic calculation plays more than a

complementary role here, since the kinetic difficulty to reach the equilibrium phases in HEAs

makes it hard to tell whether the equilibrium conditions have been reached experimentally.

The predicted equilibrium phases from the thermodynamic calculations are, on the other hand,

purely based on energetic considerations, and are hence not affected by kinetic factors. This

combination of the experimental approach together with thermodynamic calculations provides

a unique and also comprehensive understanding of the phase stability in HEAs.

Another focus of this work is the tensile properties, on which the available reports have

been rare [24-31]. The tensile properties of the Al0.5CrCuFeNi2 alloys were carefully

measured. Their correlation to the indentation behavior was also studied.

2 Experimental

The target alloys, with a nominal composition of Al0.5CrCuFeNi2 (elements in atomic

ratios), were prepared by arc-melting a mixture of the constituent elements with purity better

than 99.9 %, in a Ti-gettered high-purity argon atmosphere. The melting was repeated five

times to achieve a good chemical homogeneity of the alloy. The molten alloy was drop-cast

into a 15 mm (width) x 50 mm (length) x 3 mm (thickness) copper mold. The 3 mm thick as-

cast alloys were cold rolled to 1.7 mm in thickness (a reduction of 43.3 %). Further cold

rolling caused cracking of the specimen. One sample was cold rolled to 1 mm (a thickness

reduction of 66.7%) anyway to test the phase stability of the alloy upon thermomechanical

treatments. The 1.7 mm thick cold rolled samples were annealed at 700, 900 and 1100 oC for

1 day and 5 days, respectively. All annealed samples cooled down inside the furnace. The

phase constitution was identified using the Bruker AXS D8 Discover X-ray diffractometer

(XRD) with a Co target. The microstructure of the alloys was characterized using the JEOL

JSM 6490 scanning electron microscope (SEM). For the microstructure observation, the

5

sample surface (along the length direction, also the rolling direction for the cold rolled

samples) was sequentially polished down to 0.3 m grit alumina suspension finish, then

electrochemically etched using the Cica-Reagent Electrolyte A (Kanto Chemical). The finer

microstructure was characterized using the JEOL 2100F transmission electron microscope

(TEM), operating at 200 kV. For the TEM observations, the specimens were first

mechanically thinned to < 80 m in thickness and then twin-jet electrochemically polished till

perforations, using the 10 vol. % perchloric acid - 90 vol. % ethanol solution. Vickers

hardness was measured on the polished surfaces by applying a load of 1kg for 15 s using a

Future-Tech microhardness tester. For the tensile tests, the bar-shaped specimens were cut

from the as-cast and 1-day annealed alloys, with a gauge length of 12.5 mm, a width of 3.2

mm and a thickness of ~ 1 mm. Both sides of the specimens were carefully polished. The

tension specimens were tested to failure at a strain rate of 1×10-3 s-1 on the 810 MTS

instrument. At least three specimens were tested for each condition. The melting point of the

alloy was measured using the TA Instruments SDT Q600 DSC (differential scanning

calorimeter) in the argon atmosphere and the heating rate was 20 oC / min.

3. Thermodynamic Calculation

The Thermo-Calc program, based on the CALPHAD (calculation of phase diagrams)

method, has been widely used to evaluate the phase constitution in the complex multi-

component alloy systems [32]. However, its application to multi-component alloys with

equiatomic or close-to-equiatomic compositions has been sporadic so far [20, 33-35]. In this

work, the Thermo-Calc program was used to obtain an estimation of the equilibrium phases at

various temperatures for the Al0.5CrCuFeNi2 alloy. The calculations were based on the TTNI8

database, which is developed for Ni-based alloys. The calculations based on this database

have been proved effective for a high-entropy Al0.5CoCrCuFeNi alloy in our previous work

[20].

6

4 Results

4.1 Phase Identification

The XRD patterns for as-cast, cold rolled, and 1 and 5-day annealed Al0.5CrCuFeNi2

alloys at 700 to 1100 oC are shown in Fig. 1. In both as-cast and cold-rolled alloys, there

appeared to be only one set of peaks for the fcc phase, but the following microstructural

observations would show that there actually existed two fcc phases with close lattice constants.

In 700 oC annealed alloys, bcc phase and ordered fcc phases formed, and the amount of

ordered fcc phases increased with extended annealing time. Intensities of ordered fcc phase

became weak in 900 and 1100 oC annealed alloys. Bcc phases were still observed in 900 oC

annealed alloys, but they disappeared in 1100 oC annealed ones. The lattice constants for the

fcc and ordered fcc phases were calculated to be ~ 3.59 Å, and the ordering reaction almost

did not change the lattice constant. The lattice constant for the bcc phase was ~ 2.88 Å.

4.2 Microstructural Characterization

Figure 2 shows the microstructures for as-cast, cold-rolled and annealed Al0.5CrCuFeNi2

alloys. Dendritic structures were observed in as-cast and cold-rolled alloys. The fact that the

dendritic and inter-dendritic showed distinctively different corrosion rates indicated that they

are two phases. The very close lattice constants for these two phases led to the overlapping of

their XRD peaks, and as a result only one set of peak was seen in Fig. 1.

The dendritic morphology remained in 700 oC annealed alloys, indicating that at this

temperature (T/Tm = 0.64, Tm was measured to be 1242 oC or 1515 K by DSC) the

recrystallization did not complete yet. Combining the morphological observation and the

XRD result, it is reasonable to infer that the dendritic region kept the fcc structure, while the

inter-dendritic region had a mixed fcc and bcc structure. This is further supported by the

compositional information (not given here) in that the dendritic region in the 700 oC annealed

alloy had very close compositions to those in the as-cast alloy. A closer look at the inter-

7

dendritic region revealed interesting features: prosperous needle-like microstructures

appeared to be decomposed from the inter-dendrite regions, as shown in Figs. 3 and 4. A

typical TEM image for the needles is shown in Fig. 4, with the corresponding selected area

diffraction pattern given in the inset. The needles typically had a width of ~ 60 nm, and a

length of ~ 1 m. TEM analysis revealed that these needle-like microstructures had the L12-

type ordered fcc structure, as evidenced by the indexed superlattice diffraction spots. The

microstructural observation was in agreement with the XRD result. It is highly possible that

the inter-dendritic regions decomposed into the needle-like fcc phase and also the bcc phase,

the latter being etched away.

Poly-grained microstructures were observed in 900 and 1100 oC annealed alloys,

indicating the recrystallization had completed above 900 oC (T/Tm > 0.77). Understandably,

the poly-grains in the 900 oC annealed alloys should have an fcc structure, while the inter- or

intra-granular particles had a bcc structure. A small amount of needle-like ordered fcc phases

could still be observed in the 900 oC annealed specimens, as were highlighted by the circles in

Fig. 2 (e) and (f). This observation was also in agreement with the XRD result. The extension

of annealing time from 1 to 5 days did not cause significant grain growth, as the poly-grain

size increased from ~ 7.5 to ~ 12.6 m, and the particle size increased slightly from ~ 1.4 to ~

2.0 m. No apparent secondary particles were observed in 1100 oC annealed alloys, and this

echoed the XRD result in that there was no bcc phase forming in these alloys. The small

amount of ordered fcc phases, which was detected by XRD, could not be observed

microstructurally in 1100 oC annealed alloys. The grain boundaries in 1100 oC/5-day annealed

alloy were difficult to be nicely revealed by etching, and the seemingly secondary phases

inside the grains were believed to be etching artifacts. The average grain size increased

slightly from ~ 30 to ~ 36 m when the annealing time was extended from 1 to 5 days, but the

size of larger grains increased more significantly, typically from ~ 45 to ~ 80 m.

8

4.3 Indentation Behavior

Figure 5 shows the hardness variation for as-cast, cold-rolled and annealed

Al0.5CrCuFeNi2 alloys. The hardness of the as-cast alloy was 218 Hv, and it increased to 382

Hv after being cold rolled to 1.7 mm. Annealing at 700 oC for one day increased the hardness

to 419 HV. The hardening could originate from the newly formed bcc phases and also the L12

structured phases. The extension of annealing time to 5 days at this temperature caused a

slight decrease of hardness to 410 Hv, probably due to the recovering process that released the

strain energy [20]. Compared to the cold rolled alloy, one day of annealing at 900 oC caused

almost no change in hardness: from 382 Hv to 380 Hv. The hardening due to the newly

formed bcc phases and a small amount of L12-structured phases was balanced by the hardness

reduction due to the fully released strain energy during the recrystallization process. The

hardness even remained nearly constant (to 383 Hv) when the annealing time was extended to

5 days. The annealing at a higher temperature of 1100 oC caused the hardness to drop to 305

Hv and 306 Hv, after annealing for 1 and 5 days, respectively. Apparently, the hardness

reduction due to the recrystallization could not be compensated by the hardening due to the

new phase formation. The fact that the hardness could remain almost constant after

extensively annealing the alloys at such a high temperature (T/Tm = 0.91) is indeed impressive.

The anti-softening character at elevated temperatures offers HEAs great potentials to be used

in high-temperature environments.

The indentation impressions at as-cast and 5-day annealed alloys are shown in Fig. 6.

Interestingly, slip bands around the impressions could only be observed in alloys containing

no bcc phases, i.e., as-cast and 1100 oC annealed alloy. For 700 and 900 oC annealed ones, no

slip bands could be spotted. Chang et al. suggested that the occurrence of slip bands around

the indentation impression could be used to roughly gauge the plasticity of HEAs, and those

showing slip bands should have good plasticity [36]. Considering that the formation of bcc

9

phase in HEAs often tends to embrittle the material [37], the slip banding phenomenon

observed here seemed to lend support to Chang’s argument. However, the following tension

results indicate that the slip-banding can be at most considered as an indicator of malleability

(at compression), rather than ductility (at tension).

4.4 Tensile Properties

Room temperature tensile behaviors are rarely reported for HEAs. The main reason is

due to the seemingly contradicting high strength and high tensile ductility for HEAs at the

room temperature. Fcc structured HEAs usually have some tensile ductility but their strengths

are low, while bcc phase containing HEAs can have a high strength but poor tensile ductility,

on the premise that the bcc phases significantly strengthen the material. As a result, the

reported mechanical behaviors of HEAs mainly come from compression tests, and only a

limited number of reports are available for tension test results [24-31]. Here in this work, we

carefully measured the tensile properties of both solely fcc structured and bcc phase

containing HEAs, and both as-cast and thermomechanically treated alloys.

Representative engineering stress-strain curves for as-cast and 1-day annealed alloys are

presented in Fig. 7, with the measurement distribution of the tensile stress and elongation

given in Fig. 8. The main result here is in accordance with the previous understanding on the

mechanical behavior of HEAs. As-cast alloys with disordered fcc structures had on average an

yield stress ( y) of 363 MPa, a fracture stress ( u) of 500 MPa and an elongation ( p) of 16.1 %.

The fractured surface showed a dimple-like morphology, typical of ductile deformation

behavior. The measurement scatterings on the yield stress and fracture stress were reasonable,

but the variation on tensile elongation was large. 700 oC annealed alloys had a y of 630 MPa

and u of 922 MPa, and an p of 4.2 %. The fractured surface exhibited many cracks, and this

significant brittle fracture behavior could possibly explain the anomalously low tensile stress

of these 700 oC annealed alloys: their stresses were lower than those 900 oC annealed alloys,

10

and this was against the trend of the hardness variation for these alloys (Fig. 5). The best y

and u obtained for 700 oC annealed alloy were 1085 MPa and 1326 MPa, respectively, and

these values are believed to more truly reflect their intrinsic properties. The cracks could

initiate from the remaining casting defects or the L12-structured phases, and their formation

rendered the material to fail rapidly before reaching the intrinsic strength. 900 oC annealed

alloys had a y of 704 MPa and u of 1088 MPa. The measurement scattering of u was small

and was relatively large for y, but the scattering on y was much smaller compared to that for

700 oC annealed alloys. The fractured surface exhibited mixed dimple-like and micro-

cracking features. The average tensile strain p of tested specimens was 5.6 %, close to the

widely accepted threshold tensile ductility of 5 % (marked in Fig. 8(b)) [38]. 1100 oC

annealed alloys showed a typical brittle intergranular fracture behavior, resulting in a y of 360

MPa and u of 639 MPa, and a limited p of 3.4 %. The measurement scatterings on both the

tensile stress and elongation were relatively small.

5 Discussion

5.1 Metastability of the Solid Solution Phases

One unique characteristic of HEAs is the formation of simple solid solutions in the

highly concentrated multi-component alloys. Naturally, one can raise a question on the

stability of the solid solutions, especially in the as-cast condition: are these solid solutions

equilibrium phases? There are possibly two ways to verify this issue, 1) experimentally by

using thermomechanical treatments to see whether there is any phase transformation

occurring, and 2) theoretically by calculating the equilibrium phase diagram based on

thermodynamic considerations, for example using the CALPHAD method [32]. However,

cautions are aware for both methods. For the thermomechanical treatments, it is hard to

determine whether the equilibrium condition has been reached, particularly considering the

11

advocated slow diffusion kinetics of HEAs [1, 39]. On the other hand, the reliability and

accuracy of the thermodynamic calculations depend greatly on the available databases that

have been proven to work effectively, which, however, are lacking for HEAs. There already

exist a few reports on calculating the equilibrium phase diagram for HEAs [20, 33, 35], and it

is expected that the thermodynamic calculations are going to play increasingly important roles

in understanding the phase stability of HEAs. In a previous study [20], the current authors

used the Thermo-Calc software to calculate the equilibrium phases for an Al0.5CoCrCuFeNi

high-entropy alloy at different temperatures, and compared the theoretical predictions with

experimentally detected phases after thermomechanical treatments. That work led to an

unprecedented understanding of the phase stability of the solid solution phases in HEAs.

Basically, it was proposed that the solid solution phases in as-cast alloys are actually firstly

formed solid phases upon solidification. The solid solutions are energetically stable at high

temperatures due to the significant entropic contribution to the Gibbs free energy [21], which

is a unique characteristic for HEAs. Furthermore, these high-temperature stable solid

solutions are frozen to the room temperature, enabled by the slow diffusion kinetics of HEAs.

In another word, the time for cooling is simply too short compared to the timescale that is

required to complete the phase transformations. Similarly, for annealed alloys, equilibrium

phases at the annealing temperature are also frozen to the room temperature, even when the

specimens are cooled inside the furnace and not water quenched. Following the same

methodology, the above proposed scenario was examined here using the Co-free

Al0.5CrCuFeNi2 alloy.

The calculated equilibrium phases at different temperatures for Al0.5CrCuFeNi2 are

shown in Fig. 9. It is noted here that the predicted melting point is ~ 1246 oC, almost exactly

equal to the experimentally measured value (1242 oC). This certainly gives some confidence

to the validity of the thermodynamic calculations. By comparing the predictions with

12

experimental results, and assuming equilibrium phases have been reached after

thermomechanical treatments, it can be concluded that the above proposed scenario for HEAs

was reconfirmed here. According to Fig. 9, the firstly formed solid phases are two disordered

fcc phases. Experimentally, as-cast alloys did comprise two disordered fcc phases, with close

lattice parameters. Equilibrium phases at 1100 oC are also two disordered fcc phases, while

the experimentally observed phases in 1100 oC annealed alloys were one disordered phase and

one ordered fcc phase. The discrepancy on ordering between the thermodynamic calculations

and experiments also existed for the previously studied Al0.5CoCrCuFeNi alloy [20]. This

urges the requirement on updated material databases specially designed for HEAs. At 900 oC,

the equilibrium phases are two disordered fcc phases and one ordered bcc (B2) phase, while

900 oC annealed alloys comprised one disordered fcc phase, one ordered fcc phase and one

disordered bcc phase. Once again, the agreement between predictions and experimental

results is good, except for the discrepancy on ordering. When the temperature decreased from

900 to 700 oC, other equilibrium phases existing at 900 oC remained but two new phases

formed: one ordered fcc phase ( ′ phase with the L12 structure) and one -phase.

Experimentally, 700 oC annealed alloys had a simpler phase constitution: one disordered fcc

phase, one ordered fcc phase and one disordered bcc phase, i.e., the same as those in 900 oC

annealed alloys. Apart from the ordering issue, an additional discrepancy was the existence of

the -phase. More discussions on the -phase are to be covered in the next section.

5.2 -phase

-phase is an important class of topologically close-packed (TCP) phase. It has a

tetragonal crystal structure, and is known to be hard and brittle. -phase can have the

compositions of CrFe, CoCr or CoFeNi and their compositional ranges are large, not limiting

to the stoichiometric compounds [40]. Its formation often tends to lower the ductility and

toughness of the material. It was predicted to exist in the low to intermediate temperature

13

region (~ 300 oC to 850 oC) in the Al0.5CoCrCuFeNi alloy, and was experimentally detected in

700 oC annealed alloys [20]. -phase was also experimentally observed in HEA systems like

Al0.3CrFe1.5MnNi0.5 [41], AlCoCrFeMo0.5Nix [42], CoCrFeNiTix [43], AlCoCrxFeMo0.5Ni [44],

Al0.3CoCrFeNiMo0.1 [45], AlCoxCrFeMo0.5Ni [46] and Al0.5CoCrCuFeNiVx [47]. For the

target alloy Al0.5CrCuFeNi2 here in this work, thermodynamic calculations suggested that the

-phase can form in the low temperature region up to ~ 850 oC (Fig. 9). However, it was not

experimentally observed in 700 oC annealed alloys, both from the XRD (Fig .1) and the

microstructural observation (Fig. 2).

The failure to observe the -phase could originate from the insufficient driving energy or

insufficient annealing time. For the Al0.5CoCrCuFeNi alloy, a direct annealing of the cast

alloy at 700 oC for one day, without the precursory cold rolling process, did not produce any -

phase [48]. To as much as possible exclude the driving energy factor, we did two more

experiments: 1) annealed the cold-rolled-to-1.7 mm (43.3 % reduction in thickness) alloy for

an extensive annealing time of 20 days at 700 oC, and 2) further cold rolled the alloy to 1 mm

(66.7 % reduction in thickness) and annealed it for 5 days at 700 oC. The XRD results for the

two newly treated alloys are given in Fig. 10, and no -phase was detected in either case. It is

therefore safe to conclude that the -phase indeed does not exist in 700 oC annealed alloys, or

at least their volume percentage is tiny.

The discrepancy on the σ-phase shall originate from using the non-ideal database for

thermodynamic calculations. The accuracy of Thermo-Calc calculations relies on high-quality

thermodynamic databases, developed by critical assessments and systematical evaluations of

various experimental data and theoretical information. As mentioned previously, current

calculations were based on the database that was specially designed for Ni-based alloys and

not for HEAs. Although a general agreement between thermodynamic calculations and

experimental results was achieved, the occurrence of some discrepancies is understandable,

14

like on ordering. The prediction of the temperature window for the -phase could also be

inaccurate.

5.3 Notes on the Mechanical Behavior

The measured tensile properties had some agreements with the slip banding behavior as

seen in Fig. 6 in that apparent slip banding was only observed in ductile alloys, i.e., as-cast

alloys, and not in brittle or quasi-brittle alloys, i.e., 700 oC and 900 oC annealed alloys.

However, an apparent discrepancy also existed. 1100 oC alloys exhibited distinctive slip bands

around the indentation impressions, but they were tension brittle. This result suggests that the

slip banding could not serve as an indicator of possessing the ductility (at tension), and at

most it could be related to the malleability (at compression), which still needs further

experimental verifications. Previous experimental results also showed that HEAs with good

compressive plasticity could have bad tensile ductility. For example, in Ref. [49] it was

reported that the as-cast AlCoCrCuFeNi alloy had a large compressive plasticity, but this alloy

was known to be tension brittle [24].

6 Summary

The phase stability and tensile properties of an Al0.5CrCuFeNi2 high-entropy alloy were

systematically studied. Revealing the phase stability of the solid solution phases is a critical

issue for HEAs, but a thorough understanding of it has been complicated by the slow

diffusion kinetics of HEAs, which makes the phase transformation towards the equilibrium

phases difficult to complete. The investigation of the phase stability in this work combined

both thermomechanical experiments and thermodynamic calculations, providing a unique and

comprehensive understanding towards this important topic in HEAs. Meanwhile, tensile

properties of the Al0.5CrCuFeNi2 alloy were systematically studied here, and their correlation

to the indentation behavior was also investigated. Main conclusions from this work are

summarized below.

15

(1) The metastable nature of the solid solution phases in the high-entropy

Al0.5CrCuFeNi2 alloy was confirmed. Due to the slow diffusion kinetics, high-temperature

stable phases could be easily kept to the room temperature, since the timescale required to

complete the phase transformations was much longer than that of the cooing process.

Experimental observations were in a decent agreement with thermodynamic calculation

predictions.

(2) The phase constitution in as-cast and 1100 oC annealed alloys was both fcc structured

solid solutions, with the only difference in the level of ordering. Bcc phase formed in 700 and

900 oC annealed alloys. In 700 oC annealed alloys, the -phase was absent, while needle-like

L12 structured phases were prosperously seen. Fine L12 structured phases also existed, in a

needle-like form, in 900 oC annealed alloys.

(3) The Al0.5CrCuFeNi2 alloy exhibited impressive microstructural and mechanical

stabilities against the high temperature up to 1100 oC (T/Tm = 0.91). The grain size growth

was insignificant and the hardness variation was small, even after the long time annealing at

1100 oC.

(4) Only as-cast alloys with disordered fcc structures and a dendritic morphology showed

sufficient tensile plasticity. 700 oC annealed alloys with the highest hardness exhibited an

abnormal low tensile stress, even though they possessed the highest hardness, possibly due to

the initiation of cracks at remaining casting defects or needle-like L12-structured phases. A

large variability of the measured tensile properties was recorded. The alloys showing slip

banding behavior did not necessarily have tensile ductility.

Acknowledgements

This research was financially supported by the Research Grant Council (RGC), the Hong

Kong Government, through the General Research Fund (GRF) under the project number

CityU/521411.

16

References [1] J.W. Yeh, S.K. Chen, S.J. Lin, J.Y. Gan, T.S. Chin, T.T. Shun, C.H. Tsau, S.Y. Chang, Nanostructured high-entropy alloys with multiple principal elements: Novel alloy design concepts and outcomes, Advanced Engineering Materials, 6 (2004) 299-303. [2] B. Cantor, I.T.H. Chang, P. Knight, A.J.B. Vincent, Microstructural development in equiatomic multicomponent alloys, Materials Science and Engineering A, 375-377 (2004) 213-218. [3] R.A. Swalin, Thermodynamics of solids, 2nd Ed., Wiley, New York, 1991. [4] K. Zhao, X.X. Xia, H.Y. Bai, D.Q. Zhao, W.H. Wang, Room temperature homogeneous flow in a bulk metallic glass with low glass transition temperature, Applied Physics Letters, 98 (2011) 141913. [5] A. Takeuchi, N. Chen, T. Wada, Y. Yokoyama, H. Kato, A. Inoue, J.W. Yeh, Pd20Pt20Cu20Ni20P20 high-entropy alloy as a bulk metallic glass in the centimeter, Intermetallics, 19 (2011) 1546-1554. [6] X.Q. Gao, K. Zhao, H.B. Ke, D.W. Ding, W.H. Wang, H.Y. Bai, High mixing entropy bulk metallic glasses, Journal of Non-Crystalline Solids, 357 (2011) 3557-3560 [7] L.Q. Ma, L.M. Wang, T. Zhang, A. Inoue, Bulk glass formation of Ti-Zr-Hf-Cu-M (M=Fe, Co, Ni) alloys, Materials Transactions, 43 (2002) 277-280. [8] H.Y. Ding, K.F. Yao, High entropy Ti20Zr20Cu20Ni20Be20 bulk metallic glass, Journal of Non-Crystalline Solids, 364 (2013) 9-12. [9] F. Otto, Y. Yang, H. Bei, E.P. George, Relative effects of enthalpy and entropy on the phase stability of equiatomic high-entropy alloys, Acta Materialia, 61 (2013) 2628-2638. [10] O.N. Senkov, G.B. Wilks, J.M. Scott, D.B. Miracle, Mechanical properties of Nb25Mo25Ta25W25 and V20Nb20Mo20Ta20W20 refractory high entropy alloys, Intermetallics, 19 (2011) 698-706. [11] J.W. Yeh, Y.L. Chen, S.J. Lin, S.K. Chen, High-entropy alloys - A new era of exploitation, Materials Science Forum, 560 (2007) 1-9. [12] Y. Zhang, X. Yang, P.K. Liaw, Alloy design and properties optimization of high-entropy alloys, JOM, 64 (2012) 830-838. [13] S. Guo, C.T. Liu, Phase stability in high entropy alloys: Formation of solid-solution phase or amorphous phase, Progress in Natural Science: Materials International, 21 (2011) 433-446. [14] Y. Zhang, Y.J. Zhou, J.P. Lin, G.L. Chen, P.K. Liaw, Solid-solution phase formation rules for multi-component alloys, Advanced Engineering Materials, 10 (2008) 534-538. [15] X. Yang, Y. Zhang, Prediction of high-entropy stabilized solid-solution in multi-component alloys, Materials Chemistry and Physics, 132 (2012) 233-238. [16] S. Guo, Q. Hu, C. Ng, C.T. Liu, More than entropy in high-entropy alloys: Forming solid solutions or amorphous phase, Intermetallics, 41 (2013) 96-103. [17] S. Guo, C. Ng, J. Lu, C.T. Liu, Effect of valence electron concentration on stability of fcc or bcc phase in high entropy alloys, Journal of Applied Physics, 109 (2011) 103505. [18] S. Guo, C. Ng, C.T. Liu, Anomalous solidification microstructures in Co-free AlxCrCuFeNi2 high-entropy alloys, Journal of Alloys and Compounds, 557 (2013) 77-81. [19] C.J. Tong, Y.L. Chen, S.K. Chen, J.W. Yeh, T.T. Shun, C.H. Tsau, S.J. Lin, S.Y. Chang, Microstructure characterization of AlxCoCrCuFeNi high-entropy alloy system with multiprincipal elements, Metallurgical and Materials Transactions A, 36 (2005) 881-893. [20] C. Ng, S. Guo, J.H. Luan, S.Q. Shi, C.T. Liu, Entropy-driven phase stability and slow diffusion kinetics in Al0.5CoCrCuFeNi high entropy alloy, Intermetallics, 31 (2012) 165-172. [21] P.J. Craievich, M. Weinert, J.M. Sanchez, R.E. Watson, Local stability of nonequilibrium phases, Physical Review Letters, 72 (1994) 3076-3079. [22] W. Klement, R.H. Willens, P. Duwez, Non-crystalline structure in solidified gold-silicon

17

alloys, Nature, 187 (1960) 869-870. [23] K.Y. Tsai, M.H. Tsai, J.W. Yeh, Sluggish diffusion in Co–Cr–Fe–Mn–Ni high-entropy alloys, Acta Materialia, 61 (2013) 4887–4897. [24] A.V. Kuznetsov, D.G. Shaysultanov, N.D. Stepanov, G.A. Salishchev, O.N. Senkov, Tensile properties of an AlCrCuNiFeCo high-entropy alloy in as-cast and wrought conditions, Materials Science and Engineering A, 533 (2012) 107-118. [25] M.A. Hemphill, T. Yuan, G.Y. Wang, J.W. Yeh, C.W. Tsai, A. Chuang, P.K. Liaw, Fatigue behavior of Al0.5CoCrCuFeNi high entropy alloys, Acta Materialia, 60 (2012) 5723-5734. [26] T.T. Shun, Y.C. Du, Microstructure and tensile behaviors of FCC Al0.3CoCrFeNi high entropy alloy, Journal of Alloys and Compounds, 479 (2009) 157-160. [27] F.J. Wang, Y. Zhang, G.L. Chen, H.A. Davies, Tensile and compressive mechanical behavior of a CoCrCuFeNiAl0.5 high entropy alloy, International Journal of Modern Physics B, 23 (2009) 1254-1259. [28] L. Liu, J.B. Zhu, L. Li, J.C. Li, Q. Jiang, Microstructure and tensile properties of FeMnNiCuCoSnx high entropy alloys, Materials & Design, 44 (2013) 223-227. [29] L. Liu, J.B. Zhu, C. Zhang, J.C. Li, Q. Jiang, Microstructure and the properties of FeCoCuNiSnx high entropy alloys, Materials Science and Engineering A, 548 (2012) 64-68. [30] A. Gali, E.P. George, Tensile properties of high- and medium-entropy alloys, Intermetallics, 39 (2013) 74-78. [31] F. Otto, A. Dlouhý, C. Somsen, H. Bei, G. Eggeler, E.P. George, The influences of temperature and microstructure on the tensile properties of a CoCrFeMnNi high-entropy alloy, Acta Materialia, 61 (2013) 5743–5755. [32] J.O. Andersson, T. Helander, L.H. Hoglund, P.F. Shi, B. Sundman, THERMO-CALC & DICTRA, computational tools for materials science, CALPHAD, 26 (2002) 273-312. [33] A. Manzoni, H. Daoud, S. Mondal, S. van Smaalen, R. Volkl, U. Glatzel, N. Wanderka, Investigation of phases in Al23Co15Cr23Cu8Fe15Ni16 and Al8Co17Cr17Cu8Fe17Ni33 high entropy alloys and comparison with equilibrium phases predicted by Thermo-Calc, Journal of Alloys and Compounds, 552 (2013) 430-436. [34] A. Durga, K.C. Kumar, B.S. Murty, Phase formation in equiatomic high entropy alloys: CALPHAD approach and experimental studies, Transactions of the Indian Institute of Metals, 65 (2012) 375-380. [35] C. Zhang, F. Zhang, S.L. Chen, W.S. Cao, Computational thermodynamics aided high-entropy alloy design, JOM, 64 (2012) 839-845. [36] C.W. Chang, Microstructure and Properties of As-Cast 10-component Nanostructured AlCoCrCuFeMoNiTiVZr High-Entropy Alloy, Master Thesis, in, National Tsing Hua University, Taiwan, 2004. [37] C.J. Tong, M.R. Chen, S.K. Chen, J.W. Yeh, T.T. Shun, S.J. Lin, S.Y. Chang, Mechanical performance of the AlxCoCrCuFeNi high-entropy alloy system with multiprincipal elements, Metallurgical and Materials Transactions A, 36 (2005) 1263-1271. [38] E.M. Schulson, D.R. Barker, A brittle to ductile transition in NiAl of a critical grain size, Scripta Metallurgica, 17 (1983) 519-522. [39] J.W. Yeh, Recent progress in high-entropy alloys, Annales De Chimie-Science Des Materiaux, 31 (2006) 633-648. [40] C.-C. Hsieh, W. Wu, Overview of intermetallic sigma ( ) phase precipitation in stainless steels, ISRN Metallurgy, (2012) Article ID 732471. [41] M.H. Tsai, H. Yuan, G.M. Cheng, W.Z. Xu, W.W.W. Jian, M.H. Chuang, C.C. Juan, A.C. Yeh, S.J. Lin, Y.T. Zhu, Significant hardening due to the formation of a sigma phase matrix in a high entropy alloy, Intermetallics, 33 (2013) 81-86. [42] C.C. Juan, C.Y. Hsu, C.W. Tsai, W.R. Wang, T.S. Sheu, J.W. Yeh, S.K. Chen, On microstructure and mechanical performance of AlCoCrFeMo0.5Nix high-entropy alloys,

18

Intermetallics, 32 (2013) 401-407. [43] T.T. Shun, L.Y. Chang, M.H. Shiu, Microstructures and mechanical properties of multiprincipal component CoCrFeNiTix alloys, Materials Science and Engineering A, 556 (2012) 170-174. [44] C.Y. Hsu, C.C. Juan, W.R. Wang, T.S. Sheu, J.W. Yeh, S.K. Chen, On the superior hot hardness and softening resistance of AlCoCrxFeMo0.5Ni high-entropy alloys, Materials Science and Engineering A, 528 (2011) 3581-3588. [45] T.T. Shun, C.H. Hung, C.F. Lee, The effects of secondary elemental Mo or Ti addition in Al0.3CoCrFeNi high-entropy alloy on age hardening at 700 oC, Journal of Alloys and Compounds, 495 (2010) 55-58. [46] C.Y. Hsu, W.R. Wang, W.Y. Tang, S.K. Chen, J.W. Yeh, Microstructure and Mechanical Properties of New AlCoxCrFeMo0.5Ni High-Entropy Alloys, Advanced Engineering Materials, 12 (2010) 44-49. [47] M.R. Chen, S.J. Lin, J.W. Yeh, S.K. Chen, Y.S. Huang, M.H. Chuang, Effect of vanadium addition on the microstructure, hardness, and wear resistance of Al0.5CoCrCuFeNi high-entropy alloy, Metallurgical and Materials Transactions A, 37 (2006) 1363-1369. [48] C. Ng, S. Guo, C.T. Liu, unpublished results. [49] L.H. Wen, H.C. Kou, J.S. Li, H. Chang, X.Y. Xue, L. Zhou, Effect of aging temperature on microstructure and properties of AlCoCrCuFeNi high-entropy alloy, Intermetallics, 17 (2009) 266-269.

19

Figure Captions Fig. 1 (color online) XRD patterns for as-cast (a), cold rolled (b), and 1-day and 5-day annealed alloys, at 700 oC (c) and (d), 900 oC (e) and (f) and 1100 oC (g) and (h), respectively. Fig. 2 (color online) Secondary electron images for (a) as-cast, (b) cold-rolled, and 1-day (1D) and 5-day (5D) annealed alloys, at different temperatures. (c) and (d): 700 oC for 1D and 5D; (e) and (f): 900 oC for 1D and 5D; (g) and (h): 1100 oC for 1D and 5D. Circles in (e) and (f) highlight the needle-like features. Fig. 3 (a) Needle-like microstructures in the 700 oC/1-day annealed alloy; (b) an enlarged view of needle-like microstructures Fig. 4 (color online) A typical TEM image of needle-like microstructures in the 700 oC/1-day annealed alloy, with the selected-area diffraction pattern shown in the inset. Indexed diffraction spots comprise both allowed and forbidden (or superlattice) reflections for the fcc structure. Fig. 5 (color online) The Vickers hardness for as-cast (AC), cold rolled (CR), and 1 day (1D) and 5 day (5D) annealed alloys, at different temperatures Fig. 6 Optical micrographs of the indentation impressions. Sample conditions are (a) as-cast, (b) 700 oC/5-day annealed, (c) 900 oC/5-day annealed, and (d) 1100 oC/5-day annealed. Fig. 7 (color online) Tension stress-strain curves for as-cast and 1-day (1D) annealed alloys at 700, 900 and 1100 oC. The insets show their fracture surface morphology. Fig. 8 (color online) A comparison of (a) tensile stress and (b) tensile strain for as-cast and 1-day annealed alloys at 700, 900 and 1100 oC. The 5 % tensile strain, commonly accepted as the threshold strain differentiating ductile or brittle behavior, is marked in (b) for reference. Fig. 9 (color online) Equilibrium calculation results of the phase fraction (mass) as a function of temperature for the Al0.5CrCuFeNi2 alloy. Three annealing temperature points used in this work are marked for reference. Symbols in the plot: A11 and A12 for two compositionally different disordered fcc phases, B2 for the ordered bcc phase, ′ for the ordered fcc phase, for the -phase, and L for the liquid phase. Fig. 10 (color online) XRD patterns for the cold-rolled to 1.7 mm then 700 oC/ 20-day (20D) annealed alloy, and the cold-rolled to 1 mm and 700 oC/ 5-day (5D) annealed alloy. Results for the cold-rolled to 1.7 mm then 700 oC/ 1-day (1D) and 5-day (5D) annealed alloys are also included for comparison.

20

Ng, et al., Fig. 1

21

Ng, et al., Fig. 2

22

Ng, et al., Fig. 3

23

Ng, et al., Fig. 4

24

Ng, et al., Fig. 5

25

Ng et al., Fig. 6

26

Ng, et al., Fig. 7

27

Ng, et al., Fig. 8

200

400

600

800

1000

1200

1400

Ten

sile

stre

ss (M

Pa)

σy

σu

(a)

(b)

0.0

2.5

5.0

7.5

10.0

12.5

15.0

17.5

20.0

22.5

As cast 700 oC/1D 900 oC/1D 1100 oC/1D

Ten

sile

stra

in (%

)

28

Ng, et al., Fig. 9

29

Ng, et al., Fig. 10

30

Highlights

The solid solution phase in the high-entropy alloy was confirmed to be metastable.

The alloy exhibited microstructural and mechanical stability against annealing.

Only as-cast alloys showed sufficient tensile plasticity.

A large variability of the measured tensile properties was recorded.

The alloys showing slip banding behavior did not necessarily have tensile ductility.