Entropy-driven phase stability and slow diffusion kinetics in an Al 0.5CoCrCuFeNi high entropy alloy
Chun Ng a ,Sheng Guo b ,Junhua Luan b ,c ,Sanqiang Shi a ,*,C.T.Liu b ,**
a
Department of Mechanical Engineering,The Hong Kong Polytechnic University,Hung Hom,Kowloon,Hong Kong,China
b
Center of Advanced Structural Materials,Department of Mechanical and Biomedical Engineering,City University of Hong Kong,Kowloon,Hong Kong,China c
School of Materials Science and Engineering,University of Science and Technology Beijing,Beijing 100083,PR China
a r t i c l e i n f o
Article history:
Received 25March 2012Received in revised form 30April 2012
Accepted 1July 2012Available online xxx
Keywords:
B.Thermal stability
C.Thermomechanical treatment E.Phase stability,prediction
a b s t r a c t
Previous work on the stability of the solid solution phases in the high entropy alloys is inconclusive.We used a series of thermo-mechanical treatments to study the stability of the solid solution phases in a high-entropy Al 0.5CoCrCuFeNi alloy.The solid solution phases were found to be stable,against the intermetallic compounds,at high temperatures >850 C and at low temperatures <300 C.At inter-mediate temperatures,however,the intermetallic s -phase co-existed with the solid solution phases.The experimental observations were veri ?ed by the thermodynamic calculation results.The mechanisms for the phase stability,both at equilibrium and after quenching-equivalent annealing treatments,were discussed,and the roles of high entropy and slow diffusion kinetics were highlighted.
ó2012Elsevier Ltd.All rights reserved.
1.Introduction
High entropy alloys (HEAs),are de ?ned as alloys made of multiple (normally !5)principle elements and these principle elements are normally mixed in an equal or nearly-equal atomic ratio [1].This is a new alloying concept,differentiating HEAs from those traditional metallic alloys like steels,Al alloys,or NiAl-and TiAl-based alloys,where one or at most two elements dominate.They are termed as high entropy alloys because the equal atomic ratio means that the con ?guration entropy is high,according to Boltzmann ’s de ?nition of the entropy of mixing [2].One interesting characteristics of HEAs is their relatively simple phase constitution:quite often only single-phase or dual-phase solid solution phases (fcc and/or bcc type)form,without the formation of intermetallic compounds,as would otherwise be expected from the traditional physical metallurgy point of view.HEAs can possess some unique mechanical and functional properties,like high hardness,high wear resistance,high temperature softening resistance,good oxidation and corrosion resistance,and low thermal conductivity [3,4].As a result,HEAs are attracting ever-increasing interests from materials scientists and engineers.
A great number of HEAs have been developed so far,however,the scienti ?c understanding toward HEAs is still at a preliminary stage,and particularly there lack scienti ?c or even empirical prin-ciples guiding the design of HEAs,to achieve the desired phase constitution and hence the mechanical properties.For example,although simple solid solution phases can form in HEAs,amor-phous phase and intermetallic compounds can also appear (note that all the phase constitutions are refereed to the as-cast state here and afterward,if not otherwise speci ?ed)[5].It is thus important for us to predict the stability of phases (solid solution,amorphous phase or intermetallic compound)from a given HEA composition.On the other hand,assuming solid solution phases are formed,can we predict whether fcc-type solid solution or bcc-type solid solu-tion will form,also for a given HEA composition?As we already know that,fully fcc-typed HEAs are soft and ductile [6],while the bcc phase containing HEAs are generally hard but tend to be brittle [7],the answer to this question is critical to design the mechanical properties of HEAs for structural applications.We have recently addressed to these important issues [5,8],based on considerations of the fundamental properties of the constituent elements,including the atomic size mismatch,mixing entropy,mixing enthalpy,electronegativity and electron concentration.Some useful information,although still not de ?nitive,has been extracted from our analyses.For example,we have revealed that the atomic size mismatch plays a decisive role in forming solid solutions or bulk amorphous alloys [5],and the valence electron concentration can critically separate the fcc or bcc solid solution formation [8].
*Corresponding author.Tel.:t852********;fax:t852********.**Corresponding author.Tel.:t852********;fax:t852********.
E-mail addresses:mmsqshi@https://www.doczj.com/doc/3e15998754.html,.hk (S.Shi),chainliu@https://www.doczj.com/doc/3e15998754.html,.hk (C.T.
Liu).
Contents lists available at SciVerse ScienceDirect
Intermetallics
jou rn al homepage:
https://www.doczj.com/doc/3e15998754.html,/locate/intermet
0966-9795/$e see front matter ó2012Elsevier Ltd.All rights reserved.https://www.doczj.com/doc/3e15998754.html,/10.1016/j.intermet.2012.07.001
Intermetallics xxx (2012)1e 8
Interested readers are recommended to?nd out more details in our previous publications[5,8]and also in Refs.[9,10].
Previously,when we made the analysis on the phase selection in HEAs[5,8],we compared the phase constitution in different alloy compositions all in the as-cast conditions.This is a fair comparison and a reasonable starting point,considering most of the reported HEAs are in the as-cast conditions.In addition,since our?rst priority was to see how stable the solid solution phases in HEAs are (fcc vs.bcc,or solid solution vs.amorphous phase and intermetallic compounds),and there do have some solid evidences showing that the solid solution phase in HEAs are quite stable[11e14],one was then tempted to think that the phase stability at the as-cast conditions has a broader indication,in terms of the range of temperature and time.However,we are also aware of the fact that in some alloy systems,the as-cast phases change after the thermo-mechanical treatments,and new solid solution phase or even intermetallic phase can appear[7,15,16].The stability of the solid solution phases in HEAs is hence inconclusive,and it becomes a critical issue for the study of this class of alloys.In this work,we used Al0.5CoCrCuFeNi,an extensively studied HE alloy[17],to exemplify the effect of thermo-mechanical treatments on the stability of the solid solution phases,with the hope to clarify the deterministic factors that control the phase stability in HEAs. Al0.5CoCrCuFeNi has the fcc structure in the as-cast condition, comprising mainly two fcc phases in the dendritic and inter-dendritic regions respectively[17,18],with some tiny amount of ordered fcc phases,which can only be detected by transmission electron microscope(TEM)but not by X-ray diffraction(XRD)[17]. Tong et al.once gave an approximate phase diagram for the Al x-CoCrCuFeNi(x?0e3.0)alloys based on the phase transition temperatures measured from the differential temperature analysis(DTA)[17]according to which,only fcc-typed phases appear in the Al0.5CoCrCuFeNi alloy at the full temperature range. The accuracy of this DTA-determined phase diagram is certainly open to question.In fact,when the alloy was cold rolled and then annealed at900 C,bcc phases started to appear[19],which is against the prediction of the phase diagram;when the alloy was cold rolled and then annealed at intermediate temperatures (500e700 C),bcc phases also appeared together with some un-identi?ed phases(at least for600 C and700 C,see Fig.10in Ref.[7]).The complicacy occurred to the Al0.5CoCrCuFeNi alloy makes it an ideal target alloy,to investigate the phase stability of the solid solution phases in HEAs.
The work we did here differentiates with previous work mainly in two senses:?rst,we used a much longer annealing time(up to20 days)at each temperature we chose,to alleviate as much as possible the kinetic effect on the achievement of the equilibrium phase;second,we did some thermodynamic calculations of the equilibrium phase diagram to verify our experimental observations. The calculated equilibrium phase diagram is critical for HEAs, characteristic of slow diffusion even at elevated temperatures[3], as otherwise there is no effective way to tell whether the experi-mentally observed phases have reached the stable states.To the best of our knowledge,this is the?rst effort to present an equilib-rium phase diagram for HEAs,based on the thermodynamic calculation.A combination of both experimental results and theo-retical prediction is expected to lead to some convincing conclu-sions on the phase stability of the high-entropy Al0.5CoCrCuFeNi alloy.
2.Experimental
The target alloy studied in this work has a nominal composition of Al0.5CoCrCuFeNi(in atomic proportion).The alloy was prepared by arc-melting a mixture of the constituent elements with purity better than99.9%,in a Ti-gettered high-purity argon atmosphere. The melting was repeated?ve times to achieve a good chemical homogeneity of the alloy.The molten alloy was suction-cast into a15mm(width)*3mm(thickness)*50mm(length)copper mold. The3mm-thick as-cast(AC)alloy was cold rolled(CR)to1mm in thickness(a reduction of66.7%)intermittently,with two high-vacuum annealing treatments?rst at900 C for1h when rolled to1.7mm,and then at900 C for2h when rolled to1.4mm.After annealing,the samples cooled down inside the furnace.One piece of the1.7mm-thick samples was water quenched(WQ)after being annealed at900 C for1h,as a test of the cooling rate effect on the phase formation.The1mm-thick cold rolled samples were then annealed in the high vacuum at700,900and1100 C for1,5and20 days,respectively.All annealed samples cooled down inside the furnace.The phase constitution was identi?ed using the Bruker AXS D8Discover X-ray diffractometer(XRD)with a Co target.The microstructure of the alloys was characterized using the JEOL JSM-6490scanning electron microscope(SEM)installed with an energy dispersive spectrometer(EDS),operating at20kV.For the micro-structure observation,the sample surface was sequentially pol-ished down to0.1m m grit alumina suspension?nish.Vickers hardness was measured on the polished surfaces by applying a load of1kg for15s using a Future-Tech microhardness tester.
3.Thermodynamic calculation
As we mentioned in the Introduction section,it is dif?cult to determine whether the existent phases after the thermo-mechanical treatments reach the equilibrium state,since the sluggish diffusion in HEAs[3]would probably necessitate a suf?-cient long time to complete the phase transformation process. Thermodynamic calculations,on the other hand,are not affected by the diffusion kinetics.They can supplement the experimental observations,and provide a prediction of the equilibrium phases and their fractions at the full temperature range.The Thermo-Calc program,with validated databases for the thermodynamic calcu-lations of the phase equilibrium,which uses the CALPHAD(calcu-lation of phase diagrams)method,has been widely used for the evaluation of the phase stability in the complex multi-component alloy systems[20].In this work the Thermo-Calc program was used to form a preliminary estimation of the equilibrium phase diagram(phase constitution vs.temperature)for the Al0.5CoCrCu-FeNi alloy.The obtained results will serve to evaluate the equilib-rium phase relationship for the multi-component alloys with equal or nearly-equal atomic ratio.In view of the fact that both Cu and Co can form continuous solid solutions with Ni at almost the full composition range[21],we approximated the Al0.5CoCrCuFeNi alloy as a Ni-based alloy,and did the preliminary calculations based on the TTNI8database for Ni-based alloys.
4.Results
4.1.Phase analysis
Fig.1shows the XRD patterns for the cast alloys fabricated through intermittent cold rolling and annealing to the1mm-thick cold rolled state.Two fcc phases(fcc1and fcc2)appear in the as-cast alloy;after cold rolling to1.7mm,the strain induced peak broadening masks the two fcc phases and only one set of fcc peaks can be observed,but apparently the peaks are not symmetric.bcc phase starts to appear after annealing at900 C for1h,which is similar to what Tsai et al.reported previously[19].Ordered bcc phase also already appears at this stage.Water quenching with a faster cooling rate from900 C does not prevent the bcc phases from being formed,although the amount of the bcc phases in the
C.Ng et al./Intermetallics xxx(2012)1e8 2
water-quenched condition appears to be less compared to that in the furnace-cooling condition,judging from their relative diffrac-tion intensities.On the other hand,Tsai et al.reported that the water quenching from 1100 C can prevent the formation of the bcc phase in the same alloy,while the furnace cooling can not.This probably indicates that the temperature,rather than the kinetic effect,affects the phase stability more signi ?cantly.We will come back to this point later in the Discussion section.No new phase appears at further cold rolling and annealing.
The 1mm-thick cold rolled samples were then annealed at 700,900and 1100 C,for 1,5and 20days,respectively,and the corre-sponding XRD patterns are given in Fig.2.As seen in Fig.2,an extended annealing at 900 C essentially does not change the phase constitution,and two fcc phases,bcc phase and ordered bcc phase remain to exist.This,together with other evidences provided afterward,actually suggests that the solid solution phases are kinetically quite stable in some particular temperature ranges.Annealing at 700 C also results in the formation of two fcc phases,bcc phase and ordered bcc phase,but with additionally the formation of the intermetallic s -phase.To our knowledge,this is the ?rst report of forming the intermetallic phase in the Al 0.5CoCrCuFeNi alloy.As mentioned in the Introduction section,
Tsai et al.did the annealing for this alloy at 600 C and 700 C for 10h,and from the XRD patterns they collected (Fig.10in Ref.[7]),the intermetallic phase also formed.However,they apparently ignored those weak peaks that are corresponding to the s https://www.doczj.com/doc/3e15998754.html,bining the crystallography information and the matching alloying elements,the s -phase could be CoCr,CrFe,CoNiCr,or more possibly the CoCr-,CrFe-,CoNiCr-base solid solutions [22].Annealing at 1100 C,however,leads to a simpler phase constitu-tion where only fcc-typed phases exist:two fcc phases and ordered fcc phase.This phase constitution is similar to that in the as-cast state,except that the amount of the fcc2phase increases,and the amount of ordered phase is now suf ?cient to be detected by XRD.4.2.Microstructure
The microstructures for the as-cast and 1mm-thick cold rolled alloys are shown in Fig.3.The as-cast alloy has a clear dendritic structure,and the interdendritic region has signi ?cant Cu segre-gation,mainly due to the positive mixing enthalpy between Cu and other alloying elements [17,23].The EDS results for the average chemical compositions in the dendritic region and interdendritic regions (Region I and II,respectively,as marked in Fig.3(a))are given in Table 1.Relating the phase analysis given in Fig.1,both the dendritic regions and interdendritic regions have the fcc-type structure,and more speci ?cally they correspond to the fcc1and fcc2phases,respectively,inferred from their relative amount.The Cu-rich fcc2phase has a larger lattice constant than that of the main fcc1phase,as revealed from their relative peak positions.
Compared to the as-cast alloy,the 1mm-thick cold rolled alloy shows an elongated structure in the dendritic regions,while the interdendritic regions become discontinuous.There exist three distinguished chemically different regions (Region I,II and III),as seen in Fig.3(b),rather than two in the as-cast state.Naturally,this is also related to the phase analysis given in Fig.1.In addition to the fcc1and fcc2phases in the as-cast state,a bcc phase is formed and it shall correspond to Region III in Fig.3(b).Seen from the EDS analysis that is given in Table 1,the bcc phase is rich in Ni and Al,and the amount of Ni and Al is almost equal.In total,Ni and Al account for w 60at.%of the total alloying elements in this bcc-type phase.The nearly equal atomic ratio of Ni and Al in the bcc phase could also be responsible for the ordered bcc phase that is detected in XRD:Tong et al.had already shown that the ordered bcc phase has a NiAl-like (B2)crystal structure [17].
The microstructures for the alloys (starting from the 1mm-thick cold rolled state)annealed at 700,900and 1100 C for 1,5and 20days,respectively,are shown in Fig.4.For the 700 C annealed conditions,as the annealing time increases,the elongated dendrites progressively transform to the poly-grained structure.There still left some long dendrites after 20days of annealing at this temperature,suggesting that the full recrystallization has not been completed.From the phase analysis given in Fig.2,we know that compared to the starting state,additional intermetallic s -phase also forms.Judging from Fig.4(a e c),the s -phase very possibly corresponds to the rod-like precipitates as indicated by the blue arrows.Although the sizes of these intermetallic compounds may prevent a meaningful EDS measurement of their chemical composition (the probe size of SEM/EDS is no better than 1m m),the darker contrast as shown in the back scattering electron images indicates that they could be CrFe-or CoCr-base solid solutions,if we recall the crystallography information that is obtained from the phase analysis.In the 900 C annealed alloys,the rod-like s -phase does not exist,in agreement with the XRD results.The micro-structure is now almost fully poly-grained,and no obvious elon-gated dendrites exist after one day of annealing at this temperature.Three regions of distinctive chemical contrast are clearly observable
30405060
708090100110120130
1112222
1666663335I n t e n s i t y
2 theta (degree)
700o
C/1D;
700o
C/5D;
700o
C/20D
900o
C/1D;
900o
C/5D;
900o
C/20D
1100o
C/1D;
1100o
C/5D;
1100o
C/20D
1 fcc1
2 fcc2
3 bcc
4 ordered bcc
5 ordered fcc
6 phase
4σFig.2.XRD patterns for the Al 0.5CoCrCuFeNi alloys after annealing at 700,900and 1100 C for 1,5and 20days,respectively.
2 theta (degree)
Fig.1.XRD patterns for the as-cast,intermittently cold rolled and annealed,and the ?nally cold rolled to 1mm-thick Al 0.5CoCrCuFeNi alloys.
C.Ng et al./Intermetallics xxx (2012)1e 83
and their compositions are all very similar to those of the corre-sponding regions in the 1mm-thick cold rolled alloy (see Tables 1and 2).The 1100 C annealed alloys have a similar microstructure to that of the as-cast alloy,which is reasonable in that both alloys have the same phase constitutions (mainly fcc1tfcc2).The difference is in the 1100 C annealed alloys,the fully poly-grained structure
replaces the dendritic structure in the as-cast alloy.The amount of the fcc2phases also increases in the 1100 C annealed alloys,as seen from the comparison between Fig.3(a)and Fig.4(g e i),as well as between the XRD patterns in Figs.1and 2.The chemical compositions for the fcc1and fcc2in the 1100 C annealed samples,no matter the annealing time,are very close to those in the as-cast alloy,as seen from the EDS results given in Table 2.In terms of the grain growth occurred in the (900 C and 1000 C)annealed alloys,although ideally a detailed linear intercept method is required to analyze the grain size,and the secondary electron images of the etched samples are preferred than the back scattering electron images as provided in Fig.4(the grain boundary can be better revealed),a rough estimate of the grain size in the 1100 C annealed samples indicates that the grain growth is not signi ?cant:the grain size changes from w 10m m in the 1day annealed sample,to w 20m m in the 20day annealed sample.It is thus expected that the contribution to the mechanical properties from the grain size variation is
limited.
Fig.3.Back scattering electron images for the (a)as-cast and (b)1mm-thick cold rolled Al 0.5CoCrCuFeNi alloys.
Table 1
EDS analysis for the as-cast and 1mm-thick cold rolled Al 0.5CoCrCuFeNi alloys.Material
Region
Element (at.%)Al
Co Cr Cu Fe Ni Nominal
9.0918.1818.1818.1818.1818.18AC 3mm I 8.0420.5621.0311.4919.7119.18II 13.33 6.18 6.3655.68 5.9312.51CR 1mm
I 6.2520.5622.4811.8622.6316.22II 10.47 6.13 6.4359.04 5.8512.08III
28.18
11.24
8.08
14.09
9.19
29.22
Fig.4.Back scattering electron images for the Al 0.5CoCrCuFeNi alloys after annealing.(a)700 C/1D;(b)700 C/5D;(c)700 C/20D;(d)900 C/1D;(e)900 C/5D;(f)900 C/20D;(g)1100 C/1D;(h)1100 C/5D;(i)1100 C/20D.
C.Ng et al./Intermetallics xxx (2012)1e 8
4
4.3.Hardness
The hardness variation from the as-cast state,to the 1mm-thick cold rolled state,through the intermittent cold rolling and annealing,is given in Fig.5.The hardness of the as-cast sample is 212Hv,and it increases to 362Hv after being cold rolled to 1.7mm (w 43.3%reduction in thickness).One hour of annealing at 900 C decreases the hardness to w 341Hv,due to the reduced degree of cold work by the cold rolling;however,this recovery is partially compensated by the strengthening effect of the newly formed bcc (and ordered bcc)phases.This can also explain why in the water quenched alloy where less amounts of the bcc phase are formed,a lower hardness of 302Hv is obtained.The further increase and decrease of hardness in the following cold-rolled and annealed alloys are understandable,as basically they are all caused by the variation of the strain stored in the materials,and no noticeable phase transformation is involved.The 1mm-thick cold rolled alloy
has a hardness of 341Hv.It is noticed that after the ?rst annealing at 900 C for 1h,the hardness variation in the further cold-rolled and annealed alloys has become relatively slow.
The hardness change for the alloys annealed at 700,900and 1100 C for 1,5and 20days is shown in Fig.6.In the 700 C annealed conditions,the hardness ?rst increases from 341Hv to 362Hv,mainly because of the contribution from the newly formed s -phases.The amount of the s -phase does not change signi ?cantly,as seen from both the XRD patterns in Fig.2and the microstruc-tures in Fig.4(a e c).On the other hand,the extended annealing at this temperature renders the progressive structural transformation from the dendrite to the poly-grained structure,and the release of the strain energy reduces the hardness to 341Hv after 5days,and to 307Hv after 20days of annealing.For the 900 C annealing conditions,the hardness decreases to 261Hv ?rst after 1day,then
150
200250300350
400H a r d n e s s (H v )Fig.5.Vickers hardness for the as-cast,intermittently cold rolled and annealed,and the ?nally cold rolled to 1mm-thick Al 0.5CoCrCuFeNi alloys.
200
220240260280300320340360380400
H a r d n e s s (H v )
Annealing time (days)
Fig.6.Vickers hardness for the Al 0.5CoCrCuFeNi alloys after annealing at 700,900and 1100 C for 1,5and 20days,respectively.The hardness for the as-cast and 1mm-thick cold rolled alloys are also given for reference.
Table 2
EDS results for the Al 0.5CoCrCuFeNi alloys annealed at 700,900and 1100 C for 1,5and 20days,respectively.Material
Element (at.%)Region Al Co Cr Cu Fe Ni Nominal
9.0918.1818.1818.1818.1818.18700C
1D
I 6.4221.4822.929.6921.8917.60II 6.607.647.6260.647.719.79III 30.1211.73 6.3210.799.3331.705D
I 5.6721.8321.9510.0622.9917.50II 7.47 5.30 4.4566.94 5.1310.70III 31.0711.36 5.327.839.1735.2520D
I 5.8521.3722.3110.5422.3617.58II 7.69 5.23 4.0569.14 4.329.57III 30.0211.31 5.859.689.3333.81900 C
1D
I 6.5321.3423.438.8823.3416.48II 9.34 5.14 4.8964.40 5.1111.13III 28.9712.057.4010.8210.0430.725D
I 5.5323.3925.26 4.9523.7717.10II 9.077.597.7055.327.6512.67III 29.2812.037.259.939.6331.8920D
I 5.1623.8324.92 4.7123.8617.52II 9.66 4.14 3.3767.62 3.6811.52III 30.7310.59 4.8311.428.4833.951100 C
1D I 8.5719.6720.4813.1919.6518.45II 12.03 6.87 5.6853.86 6.4515.115D I 8.0019.9420.6912.5719.8818.92II 12.08 5.69 4.4157.55 5.5314.7420D
I 8.2719.4420.9113.2820.3717.73II
11.38
6.14
5.03
57.76
6.04
13.64
C.Ng et al./Intermetallics xxx (2012)1e 8
5
decreases slowly to254Hv and to248Hv,respectively,after5days and20days of annealing.This is reasonable as neither signi?cant phase transformation(Fig.2)nor microstructure evolution (Fig.4(d e f))occurs during the extended annealing at this temperature.As for the1100 C annealing alloys,the hardness changes from341Hv to265Hv after1day annealing.It is noted here that both the phase transformation(bcc phases disappear)and the microstructure variation(dendrites to poly-grains)occur during this process.The hardness decreases slowly to245Hv after5 day annealing,similar to the trend in the900 C annealed alloys, but increases to284Hv after20days of annealing.This increase in hardness must be accredited to the increase of the amount of the fcc2phase(and very possibly also of the ordered fcc phase),as is seen clearly in the XRD pattern in Fig.2,although this is less sensitively re?ected in the microstructure in Fig.4(i).
5.Discussion
5.1.Equilibrium phase?
The aim of this work is to verify the stability of the solid solution phases in HEAs,and for this purpose we employed a series of thermo-mechanical treatments.As shown in Figs.1and2,the solid
solution phases(fcc1tfcc2)in the as-cast Al0.5CoCrCuFeNi alloy are not thermally stable,at least when they are annealed at the temperature range between700 C and900 C.Interestingly,when annealed at a high temperature of1100 C(referring to the melting temperature of w1279 C for this alloy[17]),the phase constitution is the same as that in the as-cast condition,although the relative amount of each phase varies.It is also noted that the extended annealing for up to20days does not change the phase constitution at a?xed annealing temperature.These phenomena naturally lead us to put forward such a scenario:the phases we obtained after annealing at700,900and1100 C are actually the equilibrium ones at these corresponding temperatures,and they are frozen to the room temperature,as the annealing of HEAs is somehow equivalent to a quenching process for conventional alloys.Apparently,there are two premises for this conceptual scenario to be real:?rst,the equilibrium phases at these relevant temperatures are in agree-ment with what we see experimentally;second,the equivalence of the annealing to quenching is valid for HEAs.
The advocated sluggish diffusion features in the HEAs[3]can be responsible for the second premise,although so far there actually lacks exact diffusion data for the HEAs to support this claim.This, however,can be conceptually understood by considering the phase transformation process which requires the cooperative diffusion of elements to achieve the equilibrium partitioning among different phases,and the severe lattice distortion occurred in the HEAs[3,24] in fact hindering the atomic movement by which the diffusion happens.In other words,although in our annealing experiments the alloys cooled down to the room temperature inside the furnace, the(cooling)time scale can still be regarded as rapid enough, compared to the time scale of diffusion that is required to complete the phase transformation,or to form new phases.The annealing is hence equivalent to a quenching process in this sense.We then come back to the?rst premise:are the phases we observed experimentally the equilibrium ones at the relevant temperatures? To answer this question,we used the Thermo-Calc software to calculate the equilibrium phase diagram(phase constitution vs. temperature)for the Al0.5CoCrCuFeNi alloy,and the results are shown in Fig.7.Seen from Fig.7,the prediction from the calcula-tions is in a good agreement with the experimental observations:at 700 C,the predicted phases are fcctordered fcctordered bccts-phase,and we observed fcc1tfcc2tbcctordered bccts-phase;at900 C,the predicted phases are fcctordered fcctordered bcc,and we observed fcc1tfcc2tbcctordered bcc; at1100 C,the predicted phases are fcctordered fcct(tiny amount of)ordered bcc,and we observed fcc1tfcc2tordered fcc. For convenience,this comparison on the phase constitution between experiments and thermodynamic calculations is shown in Fig.8.The discrepancy between the prediction and the experi-mental results is mainly on the fcc2phase and the ordering of the fcc and bcc phases.The fcc2phase is due to the Cu segregation,as explained in Sections4.1and4.2.This is not a major concern.The discrepancy on the ordering between the prediction and experi-mental observations could possibly arise from two causes.On one hand,in our experimental results,we judged the disordered or ordered solid solution phases mainly relying on the XRD patterns. However,the XRD method is inaccurate to give the quantitative information on the relative amount of disordered or ordered phases in HEAs,as the diffraction intensities decrease sharply due to the crystalline planes distorted by solid solutioning[24].As an example,Singh et al.detected not much ordered bcc phases in the cast Al1.0CoCrCuFeNi alloy from the XRD pattern;however,the transmission electron microscope(TEM)images showed that
the Fig.7.Calculated phase diagram(phase constitution vs.temperature)for the Al0.5CoCrCuFeNi alloy using the Thermo-Calc software.The dash-dotted lines indicate the three annealing temperatures used in this work.Symbols in the plot:A1for disordered fcc phase,A2for disordered bcc phase,B2for ordered bcc phase,g0for ordered fcc phase,s for s-phase,and L for liquid
phase.
Fig.8.A comparison of the phase constitution between the experimental observations and the thermodynamic calculations.
C.Ng et al./Intermetallics xxx(2012)1e8 6
bcc phases in this alloy are most ordered[25].On the other hand,it is highly possible that the slow kinetics in HEAs inhibits or at least slows down the ordering process.Certainly,the ordering infor-mation generated from our preliminary calculation needs further re?nement.
In spite of the discrepancy as acknowledged above,such a very reasonable agreement between the experimental observations and the prediction is encouraging,and it supports the scenario we proposed in that the annealing-induced phases are actually the equilibrium phases at the relevant annealing temperatures,and the cooling from annealing of HEAs can be treated as a quenching process.From Fig.7,we also know that the intermetallic s-phase will appear in the temperature range between w300 C and w850 C,and below this temperature the phase constitution is purely solid solution phases;above w1100 C and below the melting point,only fcc-type solid solution phases(disordered and ordered)exist.The calculated equilibrium phase diagram is expected to give a useful guidance on designing the thermo-mechanical treatments,to control the phase constitution in HEAs.
5.2.Entropy-driven phase stability
In the section above,we proposed that in HEAs the slow diffu-sion kinetics ensures that the equilibrium phases at elevated temperatures can be frozen to the room temperature even by conventional annealing treatments.The stability of the solid solu-tions,and particularly the random or disordered solid solutions at high temperatures,for this case above w850 C,is still of interest, considering the phase competition between the solid solution phases and potential intermetallic compounds.In another word,it is important to understand the thermodynamic nature for the phase stability of the solid solutions in HEAs at elevated tempera-tures.Assuming the random solid solution(SS)phase competes with the intermetallic compound(IM)phase during the solidi?-cation process from the liquid state,from the energy perspective which ultimately determines the phase stability,the formation energy(D G)of the SS phase has to be lower than that of the intermetallic compound,for the former to win the competition.The formation energy,D G,for a solid solution phase can be de?ned as the free energy change from the elemental state to the alloy state [3].D G is related to the enthalpy of mixing,D H,and entropy of mixing,D S,following Eq.(1)[26]:
D G?D HàT D S(1)
As is perceived from Eq.(1),the contribution from D S to D G becomes more important(lowering D G)at elevated temperatures. This effect is even signi?cant in HEAs,as D S in HEAs is much higher compared to that in the dilute multi-component alloys[9].It is thus reasonable to attribute the stability of the SS phase at high temperatures,to the high entropy effect.This claim is supported by a more re?ned thermodynamic calculation on D G of the SS phase and the IM phase in the AlCrCuFeNi alloy,using the Miedema model[27].It needs to be pointed out here that the high entropy effect to stabilize the SS phase cannot be over exaggerated:if D H is very negative,the IM phase will form anyway[5,8].
At intermediate temperatures,such as between300 C and 850 C in the Al0.5CoCrCuFeNi alloy,the D S contribution to D G is less signi?cant and as a result,D G for the SS phase is now not univer-sally lower than that of the IM phase in a wide composition range. This leads to the co-existence of SS phase and the IM phase(s-phase in this alloy).As the temperature drops to the low-temperature(<300 C here),the D S contribution to D G is limited and it is mainly D H that determines the phase stability between the SS phase and the IM phase.It is not surprised to see the SS phase winning the competition again.The similar trend of phase stability between the SS phase and the IM phase also appears in the Cr2CuFe2MnNi alloy:when annealed at high temperatures (>1100 C),only SS phase exist;while the annealing at the inter-mediate temperatures(at least in the range of600 C e950 C)leads to the co-existence of both the SS phase and IM phase(r-phase) [15].Clearly,the entropy plays a key role on the phase stability of solid solutions in HEAs,particularly at elevated temperatures.
The phase constitutions at the as-cast conditions,as we dis-cussed previously in Refs.[5,8],hence appear to be the quenched phases that are at equilibrium at temperatures closest to the solidus temperature.Currently,in most cases when the HEAs are cast,the alloy temperatures drop sharply from above the liquidus temper-ature to the ambient temperature.Considering the slow diffusion kinetics in HEAs and the relatively short time scale for the solidi-?cation,the?rst formed solid phases will be retained to the room temperature by“quenching”in the Cu mold.On the other hand,the empirical rules we proposed previously[5,8],are still useful for the design of HEAs,as they at least work effectively for the quenched-in phases in HEAs,i.e.,the directly obtained phase constitutions in the cast alloys.
6.Conclusions
The stability of the solid solution phases in the high-entropy Al0.5CoCrCuFeNi alloy was studied using a series of thermo-mechanical treatments.Although the as-cast alloy had a simple phase constitution of mainly two fcc phases,the cold rolled an annealed alloys showed different phase constitutions.The anneal-ing at the intermediate temperature of700 C led to the co-existence of two fcc phases,disordered and ordered bcc phases,together with an intermetallic s-phase;at900 C,the annealing products were two fcc phases,plus disordered and ordered bcc phases;interest-ingly,the annealing at1100 C led to the same phase constitution as in the as-cast alloy,although the relative amount of the Cu-rich fcc phase(and possibly the ordered fcc phase)was higher in the annealed condition.Referring to the calculated equilibrium phase diagram,it was found that these annealing-induced phases are actually the equilibrium ones at the relevant annealing tempera-tures.The slow diffusion kinetics in HEAs was suggested to account for the equivalent quenched-in of these phases from elevated temperatures to the room temperature.The high entropy effect was believed to play an important role in determining the phase stability between the solid solution phases and intermetallic compounds in this Al0.5CoCrCuFeNi alloy,as well as other HEAs. Acknowledgments
This research is?nancially supported by the Research Grant Council(RGC),the Hong Kong Government,through the General Research Fund(GRF)under the project number CityU/521411.The authors thank Mr.ZB Jiao for helpful discussions on the thermo-dynamic calculations.
References
[1]Yeh JW,Chen SK,Lin SJ,Gan JY,Chin TS,Shun TT,et al.Nanostructured high-
entropy alloys with multiple principal elements:novel alloy design concepts and outcomes.Advanced Engineering Materials2004;6:299e303.
[2]Swalin RA.Thermodynamics of solids.2nd ed.New York:Wiley;1991.
[3]Yeh JW.Recent progress in high-entropy alloys.Annales De Chimie-Science
Des Materiaux2006;31:633e48.
[4]Wu WH,Yang CC,Yeh JW.Industrial development of high-entropy alloys.
Annales De Chimie-Science Des Materiaux2006;31:737e47.
[5]Guo S,Liu CT.Phase stability in high entropy alloys:formation of solid-
solution phase or amorphous phase.Progress in Natural Science:Materials International2011;21:433e46.
C.Ng et al./Intermetallics xxx(2012)1e87
[6]Wang FJ,Zhang Y,Chen GL,Davies HA.Tensile and compressive mechanical
behavior of a CoCrCuFeNiAl0.5high entropy alloy.International Journal of Modern Physics B2009;23:1254e9.
[7]Tsai CW,Tsai MH,Yeh JW,Yang CC.Effect of temperature on mechanical
properties of Al0.5CoCrCuFeNi wrought alloy.Journal of Alloys and Compounds2010;490:160e5.
[8]Guo S,Ng C,Lu J,Liu CT.Effect of valence electron concentration on stability of
fcc or bcc phase in high entropy alloys.Journal of Applied Physics2011;109: 103505.
[9]Zhang Y,Zhou YJ,Lin JP,Chen GL,Liaw PK.Solid-solution phase formation
rules for multi-component alloys.Advanced Engineering Materials2008;10: 534e8.
[10]Cunliffe A,Plummer JD,Figueroa I,Todd I.Glass formation in a high entropy
alloy system by design.Intermetallics2012;23:204e7.
[11]Lin CM,Tsai HL,Bor HY.Effect of aging treatment on microstructure and prop-
erties of high-entropy Cu0.5CoCrFeNi alloy.Intermetallics2010;18:1244e50. [12]Senkov ON,Wilks GB,Scott JM,Miracle DB.Mechanical properties of
Nb25Mo25Ta25W25and V20Nb20Mo20Ta20W20refractory high entropy alloys.
Intermetallics2011;19:698e706.
[13]Wen LH,Kou HC,Li JS,Chang H,Xue XY,Zhou L.Effect of aging temperature
on microstructure and properties of AlCoCrCuFeNi high-entropy alloy.Inter-metallics2009;17:266e9.
[14]Lin CM,Tsai HL.Equilibrium phase of high-entropy FeCoNiCrCu0.5alloy at
elevated temperature.Journal of Alloys and Compounds2010;489:30e5. [15]Ren B,Liu ZX,Cai B,Wang MX,Shi L.Aging behavior of a CuCr2Fe2NiMn high-
entropy alloy.Materials and Design2012;33:121e6.
[16]Chen ST,Tang WY,Kuo YF,Chen SY,Tsau CH,Shun TT,et al.Microstructure
and properties of age-hardenable Al x CrFe1.5MnNi0.5alloys.Materials Science and Engineering A2010;527:5818e25.[17]Tong CJ,Chen YL,Chen SK,Yeh JW,Shun TT,Tsau CH,et al.Microstructure
characterization of Al x CoCrCuFeNi high-entropy alloy system with multiprincipal elements.Metallurgical and Materials Transactions A2005;36:881e93.
[18]Liu ZY,Guo S,Liu XJ,Ye JC,Yang Y,Wang XL,et al.Micromechanical char-
acterization of casting-induced inhomogeneity in an Al0.8CoCrCuFeNi high-entropy alloy.Scripta Materialia2011;64:868e71.
[19]Tsai CW,Chen YL,Tsai MH,Yeh JW,Shun TT,Chen SK.Deformation and
annealing behaviors of high-entropy alloy Al0.5CoCrCuFeNi.Journal of Alloys and Compounds2009;486:427e35.
[20]Andersson JO,Helander T,Hoglund LH,Shi PF,Sundman B.THERMO-CALC&
DICTRA,computational tools for materials science.CALPHAD2002;26:273e312.
[21]Baker H,Okamoto H.ASM handbook.In:Alloy phase diagrams,vol.3.Ohio:
American Society for Metals;1992.
[22]Hsu CY,Wang WR,Tang WY,Chen SK,Yeh JW.Microstructure and mechanical
properties of new AlCo x CrFeMo0.5Ni high-entropy alloys.Advanced Engi-neering Materials2010;12:44e9.
[23]Takeuchi A,Inoue A.Calculations of mixing enthalpy and mismatch entropy
for ternary amorphous alloys.Materials Transactions2000;41:1372e8. [24]Yeh JW,Chang SY,Hong YD,Chen SK,Lin SJ.Anomalous decrease in X-ray
diffraction intensities of Cu e Ni e Al e Co e Cr e Fe e Si alloy systems with multi-principal elements.Materials Chemistry and Physics2007;103:41e6.
[25]Singh S,Wanderka N,Murty BS,Glatzel U,Banhart J.Decomposition in multi-
component AlCoCrCuFeNi high entropy alloy.Acta Materialia2011;59: 182e90.
[26]Porter DA,Easterling KE.Phase transformations in metals and alloys.London:
Chapman&Hall;1992.
[27]Li AM,Zhang XY.Thermodynamic analysis of the simple microstructure of
AlCrFeNiCu high-entropy alloy with multi-principal elements.Acta Metal-lurgica Sinica2009;22:219e24(English Letters).
C.Ng et al./Intermetallics xxx(2012)1e8 8