j m a t e r r e s t e c h n o l . 2 0 1 4;3(3):264–273

Available online at www.sciencedirect.com

www. jmrt .com.br

Original Article

Effect of alloying constituents on the martensiticphase formation in some Cu-based SMAs

Rupa Dasgupta ∗, Ashish Kumar Jain, Pravir Kumar,Shahadat Hussein, Abhishek Pandey

CSIR-Advanced Materials and Processes Research Institute (AMPRI), Bhopal, India

a r t i c l e i n f o

Article history:

Received 26 March 2014

Accepted 17 June 2014

Available online 30 July 2014

Keywords:

Shape memory alloys

Transition temperature

Martensite

X-ray diffraction studies

a b s t r a c t

The paper discusses the attempt made to improve upon the transition temperatures and

amount of martensite formation in Cu–Al alloys with alloying additions of Mn, Ni and Zn

in varying proportions and combinations. The alloys have been subjected to heat treat-

ment cycles to improve upon the microstructure and precipitate the required martensite

phase. The effect of these ternary and/or quaternary additions has been studied on the

phases precipitated through micro-structural analysis and X-ray diffraction and Differen-

tial Scanning Calorimetric studies. The findings confirm the possibility of improvement on

the shape memory properties like martensite formation, higher transitions temperatures

and longer retention in selected alloys based on the properties exhibited through proper

alloying additions and heat treatments.

Phase precipitation © 2014 Brazilian Metallurgical, Materials and Mining Association. Published by Elsevier

Editora Ltda.

Este é um artigo Open Access sob a licença de CC BY-NC-ND

1. Introduction

Shape memory alloys (SMAs) are metallic alloys which canrecover their shape when they are heated above a certain tem-perature. The key characteristic responsible for this behaviourof all SMAs is the occurrence of a martensitic phase transfor-mation which is a shear-dominant diffusion less solid-statephase transformation occurring by nucleation and growth ofthe martensitic phase from a parent austenitic phase. Whenan SMA undergoes a martensitic phase transformation, it

transforms from its high-symmetry, usually cubic, austeniticphase to a low-symmetry martensitic phase. Parent and prod-uct phases coexist during the phase transformation. Since the

∗ Corresponding author.E-mail addresses: rupadasgupta@ampri.res.in, dasguptarupa@gmai

http://dx.doi.org/10.1016/j.jmrt.2014.06.0042238-7854/© 2014 Brazilian Metallurgical, Materials and Mining Associa

Este é u

crystal lattice of the martensitic phase has lower symmetrythan that of the parent austenitic phase, several variants ofmartensite can be formed from the same parent phase crystal.

The martensitic transformation occurs when the freeenergy of martensite becomes less than the free energy ofaustenite at a temperature below a critical temperature T0 atwhich the free energies of the two phases are equal. How-ever, the transformation does not begin exactly at T0 but,in the absence of stress, at a temperature Ms (martensitestart), which is less than T0. The transformation continuesto evolve as the temperature is lowered until a temperature

l.com (R. Dasgupta).

denoted Mf (martensite finish) is reached. Similarly, duringthe heating cycle, the reverse transformation (martensite-to-austenite) begins at the temperature As (austenite start),

tion. Published by Elsevier Editora Ltda.

m artigo Open Access sob a licença de CC BY-NC-ND

o l . 2

aa

fcoatcvstitaii(

tomoe

mLiattsota

olCIapT

1

Cpmiotll[tp

ri

j m a t e r r e s t e c h n

nd ends at Af (austenite finish) when the material is fullyustenite.

Due to the displacive character of the martensitic trans-ormation, applied stress plays a very important role. Duringooling of the SMA material below temperature Ms in absencef applied stresses, the variants of the martensitic phasesrrange themselves in a self-accommodating manner throughwinning, resulting in no observable macroscopic shapehange; but applying mechanical loading to force martensiticariants to reorient (detwin) into a single variant, large macro-copic inelastic strain is obtained. After heating to a higheremperature, the low-symmetry martensitic phase returns tots high-symmetry austenitic phase, and the inelastic strain ishus recovered. Again, martensitic phase transformation canlso be induced by pure mechanical loading while the materials in the austenitic phase, in which case detwinned martens-te is directly produced from austenite by the applied stressStress Induced Martensite) at temperatures above Ms.

The key effects of SMAs associated with the martensiticransformation induced by temperature or stress, which arebserved according to the loading path and the thermo-echanical history of the material are: pseudoelasticity,

ne-way shape memory effect and two-way shape memoryffect.

The transformation temperatures exhibited by shapeemory alloys are highly dependent on their composition.

ittle change in their composition will result in large changesn their transformation temperatures. The advantage of beingble to change the transformation temperature by changinghe composition is that the material can be tailored to attainhe desired phase at the application temperature. There areeveral means of measuring the transformation temperaturesf SMA alloys. The two most common methods of measuringhe transformation temperatures are by electrical resistivitynd Differential Scanning Calorimetry (DSC) measurements.

There are about twenty elements in the central partf the periodic table whose alloys exhibit shape memory

ike Ag–Cd, Au–Cd, Cu–Al–Ni, Cu–Al–Mn, Cu–Au–Zn, Cu–Sn,u–Au–Sn, Cu–Zn, Cu–Zn–Al, Cu–Zn–Sn, Cu–Zn–Ga, Cu–Zn–Si,

n–Ti, Ni–Al, Ni–Ti, Fe–Pt, Fe–Pd, etc. However, mainly threelloy systems generally known to exhibit the Shape Memoryhenomena have been the focus of research and development.hey are Ni–Ti, Cu–Al and Fe based alloys.

.1. Bottom of form

u-based shape memory alloys (SMAs) are being seen asrospective alternatives due to its low cost to the more com-only used Ni–Ti ones in applications where biocompatibility

s not an issue. However, inspite of its being in the forefrontf R&D since long, Cu-based SMAs are still in its developmen-al stage and seldom commercially available due to defectsike low transition temperatures, poor mechanical properties,ow ductility and stabilisation of martensite phase or fatigue1–4]. The latter is important as the change between the highemperature austenite phase and low temperature martensite

hase is actually responsible for the shape memory properties.

The areas which still need attention in Cu–Al based SMAselate to increasing transition temperatures and improv-ng the mechanical properties which is being attempted by

0 1 4;3(3):264–273 265

researchers through synthesis and development of (i) newerclass of SMAs (ii) cost effective and more sensitive SMAs(iii) improve the transition temperature and performanceof already existing SMMs through proper composition, heattreatment, etc. (iv) fabrication of new SMAs like thin films,porous and amorphous/nano crystalline types and (v) fab-rication of SMAs with temperature memory effect (TME),characterised by a neat delay of the thermally inducedtransformation to higher temperatures where an additionalcalorimetric peak appears.

The heat treatment cycle to produce the shape memoryeffect in potential alloys of Cu–Al system is more or lessestablished and varies marginally with the alloying additions[5–10]; but the effect of alloying additions is still an open areaof research. In an attempt to improve the performance, dif-ferent alloy compositions with varying quantities of alloyingadditions are being investigated worldwide. Known elementaladditions include Ni for improved mechanical properties, Mnfor ductility, Zn for better transformation temperatures as theternary addition [11–19]. Shape memory properties are verysensitive to the percentage of alloying additions; also, minorquaternary additions could still improve upon the properties.Another method of improving the properties is through grainrefinement, which is being attempted through the additionof dispersoids and/or grain refiners to prevent grain growth.Vast variations in the ternary additions have been reported; forexample in Cu–Al–Ni system targeted for improved mechani-cal properties, Al variations have been made between 11% and23%, Ni between 3% and 5% with and without additional addi-tions of Zn, Ag, Nb, Cr, Be, Mn in varying proportions. Again,addition of Zn increases transition temperatures; Zn has beenadded between 15% and 30%; as also addition of Mn known toimprove ductility have been added between 2% and 10% withand without quaternary additions as mentioned above.

In the present paper an attempt has been made to improvethe transition temperatures and amount of martensite forma-tion in Cu–Al alloys with ternary additions of Mn, Ni and Zn invarying proportions to study the effect of each addition eitheralone or as combination with the other elements. The effectof these ternary and/or quaternary additions has been studiedon the phases precipitated through micro-structural analysisand X-ray diffraction studies in the cast, homogenised andquenched conditions. The alloys have been subjected to heattreatment cycles to improve upon the microstructure and pre-cipitate the required martensite phase. The effect of all theabove has been studied with respect to the transition temper-atures. An effort has been made to correlate the findings withthe alloying additions.

2. Experimental details

2.1. Synthesising the alloys

Seven compositions of Cu–Al base alloys (Table 1) with varyingquantities three ratios of Cu:Al were prepared in which Mn, Ni

and Zn were added in some cases. The first alloy is a knownternary alloy of Cu–Al–Zn [20,21]; in the second and thirdalloys, Ni has been additionally added and the ratio of Cu:Alvaried; the fourth alloy is again a ternary alloy of Cu–Al–Mn;

266 j m a t e r r e s t e c h n o l

Table 1 – Synthesised compositions.

Cu Al Zn Ni Mn Constituents

1 78 12 10 Cu–Al–Zn2 74 12 10 4 Cu–Al–Ni–Zn3 75 10 10 5 Cu–Al–Ni–Zn4 82 08 10 Cu–Al–Mn5 81 12 2 5 Cu–Al–Mn–Ni6 74 12 10 4 5 Cu–Al–Mn–Ni–Zn7 74 12 10 4 7 Cu–Al–Mn–Ni–Zn

Fig. 1 – Melting facility.

the fifth alloy has both Mn and Ni and the last two alloys haveall the alloying additions of Mn, Ni and Zn.

The alloys were prepared using the electric pit furnace(Fig. 1). Approximately 2 kg of each alloy from metallic rawmaterials was cast into fingers and plates (Fig. 2) using metallicmoulds.

2.2. Heat treatment

Samples from the cast alloys were subject to homogenisationtreatment at 200 ◦C for 2 h in a muffle furnace and furnacecooled. In order to precipitate the martensite phase, another

set of samples were heated at 920 ◦C held for 2 h and quenchedin iced water; the samples were left in water overnight and

Fig. 2 – Raw metal and cast fingers.

. 2 0 1 4;3(3):264–273

removed only after they were completely cooled to room tem-perature.

2.3. Microstructure

The cast, homogenised and quenched samples were polishedmetallographically using an automatic polishing machinecapable of polishing six samples together (Buehler make,Model EcoMet 3000) using standard procedures that includethe steps and procedures shown in Table 2. The microstructurewas observed in an optical microscope (LEICA make, ModelLEICA DM 6000M and Metalloplan) and FESEM (FEI make,Model Nova Nano SEM 430) were optical micrographs were notproperly resolved.

2.4. X-ray diffraction studies

X-ray diffraction studies were carried out on the cast andquenched samples (Bruker make, Model D8 Advanced) from2� 10 to 80◦ at a speed of 0.01◦/s using CuK� target; and thephases identified.

2.5. Differential Scanning Calorimetric studies

Differential Scanning Calorimetric studies were carried out onquenched samples using a Differential Scanning Calorimeter(Mettler Toledo make, Model DSC1 STARe SYSTEM) Approx-imately 8–9 mg of powder sample was used which wasscrapped from the quenched samples. Tests were carried outfrom room temperature to 550 ◦C maintaining a constant rateof 10 ◦C/min.

3. Results and discussion

3.1. Chemical composition

The chemical composition was analysed on polished sam-ples using Sparcmet. The analysis is given in Table 3 alongwith the proposed composition. The composition analysedmatches broadly the prepared composition; however in thepresence of quaternary additions, the percentage of Mn andZn is significantly less than that added; the effect being morepronounced for Zn as compared to Mn.

3.2. Microstructure

The microstructure has been observed (Fig. 3a–g) in thecast and homogenised conditions. Grain structure with � + �

phases is visible in the cast and homogenised condition of allthe alloys; which is a pre requisite for martensite formationon quenching [5–8]. It is known that single �-phase grainsare the essential requirement for the presence of pseudoelastic behaviour in Cu-based alloys, i.e., prior to quenching.Therefore, it is necessary to eliminate �-phase from themicrostructure to get a desired single-phase material (i.e.,

′

microstructure should consist only of � phase). This is donethrough homogenisation and the desired (� + �) microstruc-ture attained to facilitate the evolution of a favourable textureduring further processing. The obtained structure clearly

j m a t e r r e s t e c h n o l . 2 0 1 4;3(3):264–273 267

Table 2 – Standard procedure for polishing Cu-alloys.

Surface Abrasive/size Load lb. (N)/specimen Base speed (rpm) Relative rotationa Time (min)

CarbiMet 2 120- to 240- (P120 to P280) Grit SiC 5 (22) 300 » Until planeTexMet C 9 �m MetaDi Supreme Diamond 5 (22) 150 >< 5.00VerduTex 3 �m MetaDi Supreme Diamond 5 (22) 150 » 3.00VerduTex 1 �m MetaDi Supreme Diamond 5 (22) 150 » 2.00ChemoMet 0.02–0.06 �m MasterMet Silica 5 (22) 150 >< 1.30

a » = Complimentary (platen and specimen holder rotate in the same direction). >< = Cotra (platen and specimen holder rotate in oppositedirections).

Table 3 – Chemical composition of synthesised alloys.

Sample No. Synthesised composition Analysed composition (wt.%)

Cu Al Zn Ni Mn

1 78Cu–12Al–10Zn 77.1 11.97 10.55 – –2 74Cu–12Al–10Zn–4Ni 75.8 10.88 8.78 4.33 –3 75Cu–10Al–10Zn–5Ni 75.7 10.93 7.71 5.51 –4 82Cu–8Al–10 Mn 84.0 8.24 – – 7.325 81Cu–12Al–5Mn–2Ni 78.9 11.94 – 2.54 2.88

isot

F(

6 (74Cu–12Al–10Zn–4Ni) + 5Mn 75.37 (74Cu–12Al–4Ni–10Zn) + 7Mn 72.5

ndicates that all the alloys have potential for exhibiting the

hape memory behaviour. Precipitation in the grains typicalf shape memory alloys are also seen in the cast alloys;he most prominent being in the Sample 1. There is a large

20 um 20 um

200 um

a b

d e

g

ig. 3 – (a) Microstructure of cast samples: sample 1; (b) sample 2g) sample 7.

11.46 5.54 3.10 4.3310.90 6.80 3.49 6.23

variation in the grain size of cast alloys (Table 4). It is observed

from the table that even without adding any grain refinersor dispersoids to control the grain size, in some cases thevariation within the sample is very small (like in Samples 1,

200 um

100 um

100 um

c

f

100 um

; (c) sample 3; (d) sample 4; (e) sample 5; (f) sample 6; and

268 j m a t e r r e s t e c h n o l

Table 4 – Average grain size of cast alloys.

Sample Average grain size (�m)

1 60–702 100–1203 160–2004 50–2005 100–2506 40–150

Significantly high transformation temperatures have been

7 20–80

2 and 3) and large in some (Samples 5 and 6); the averageminimum size has been observed for Sample 7. Ternaryalloys (Samples 1 and 4) show lower grain size as comparedto quaternary alloys; however, presence of any particularelement cannot be said to affect the grain size.

Homogenisation at 200 ◦C does not affect much the grainsize as compared to the cast condition; however, homogeni-sation at 600 ◦C increases grain size significantly in mostcases. As an example, Fig. 4 illustrates the cast (marked A),homogenised at 200 ◦C (marked B) at 600 ◦C (marked C) isshown for Sample 3. Quenched samples also show increasedgrain sizes probably due to heating at 920 ◦C before quench-ing. Sudden quenching too could not restrict grain growth asall the quenched samples have shown an increase in grain sizeover the corresponding cast condition; however homogenisa-tion has brought about uniformity in grain sizes as seen fromthe micrographs which show the variation in grain size is sig-nificantly less (in most cases uniform) in the homogenisedsamples. Quenched samples also have maintained this homo-geneity.

Quenching from high temperatures was carried out expect-ing to precipitate the martensitic phases. From the quenchedmicrostructure Fig. 5, the martensite formation is not markedin all the cases, and shows different patterns. It is seenthat in Alloy 1 (Cu–Al–Zn) the precipitated martensite phaseis flower shaped typical of �-phase is observed [22]. Theplate- or spear-like martensite (the zig-zag �′ phase, mostcommonly observed in Cu-based SMAs) [20–22] has beenobserved in only the quenched Cu–Al–Mn (Sample 4) alloy.Addition of quaternary phases is seen to retain the gran-ular structure the samples. In these samples, precipitationof the martensitic phases is observed in the grains butthey are not very well resolved even in the FESEM pho-tographs. Quaternary additions tend to globulise the grainsas seen in Fig. D of Sample 7; this can be seen in low mag-nifications optical micrographs only. It may be noted thatsome of the samples are very soft and scratches just can-not be removed by polishing like is normally done. Thusa few scratches can be seen even in the best-polishedsamples.

3.3. X-ray diffraction

The X-ray diffraction pattern of the cast and quenched sam-ples of the alloys under study have been plotted in Fig. 6 and

the major peaks identified and marked in the figures. In thecast condition, Cu is expectedly the major phase precipitated.In addition to Cu, in some cases peaks of the major alloying

. 2 0 1 4;3(3):264–273

additions like Al (in Samples 1–3) and Mn (in Sample 4) hasbeen observed; however due to overlapping d values of themajor alloying element Zn its peaks could not be identifiedfor sure precipitation. The cast conditions in all the samplesrecord only the major phases of Cu with the major alloyingadditions like Al–Cu, Cu–Zn, Cu–Mn, Al2Cu. However the effectof quaternary additions was not observed in the X-ray diffrac-tion peaks, mainly because of the small percentages in whichit was added and keeping in mind the shortcoming of XRDto identify phases with less than 3% presence. However, aplethora of peaks has been observed in most of the quenchedsamples. Table 5 lists the phases precipitated and identifiedin the alloys marking the martensitic phases separately byunderlining them. From the table, a clear picture emerges asto which of the alloys have the potential of behaving as shapememory materials due the martensitic phases identified.However, only in Samples 6 and 7, no extra peak could def-initely be identified as martensite phase; no additional peaksworth recognising have been precipitated in the quenchedsamples with respect to the cast sample; however this couldbe again due to overlapping of the d values of the phases. Itmay be mentioned that due to precipitation of very few peaks,identification of individual phases remains a problem; still allthe probable phases that could be identified with the pre-cipitated d value have been marked. In Samples 3, 6 and 7only martensitic phases have been identified in the quenchedcondition signifying perhaps completion to the martensiticphase. Addition of other alloying additions was not expectedto affect the XRD peaks and the same has been reflected in thegraphs.

3.4. Differential Scanning Calorimetric (DSC) studies

The DSC studies indicate clear transformation peaks in allthe samples; however the peaks are stronger in some andrelatively blunt in the others (Fig. 7). The transformation tem-peratures are tabulated in Table 6. The temperature rangesoverlap in most of the cases. The Cu–Al–Zn series of alloystargeted for increasing the transition temperatures havebeen successful as significantly high transformation tem-peratures have been recorded over conventionally reportedvalues. Moreover, the martensite retention is over a rangeof 30–40 ◦C indicating possibility of improved shape mem-ory properties. The addition of Ni in the Cu–Al–Zn alloysas seen from comparing Samples 1 with 2 and 3 indicatedecrease in the transition temperatures when the percent-age of Cu is reduced by 4% and keeping Al and Zn constant(Samples 1 and 2); however the loss is regained by decreasingthe Al content and increasing the Cu content slightly (Sample3). This again goes to show the significance of the chemi-cal constituents for shape memory behaviour and how evenslight variations can affect the transition temperatures. Pres-ence of Zn in place of Mn is seen to marginally increasethe transformation temperatures (Samples 1 and 4). Howeverthe same cannot be said when Ni has been added in eithercase.

recorded in Cu–Al–Mn alloys than conventionally reported[10,12,23]. The range of martensite retention is the max-imum in ternary Cu–Al–Mn alloys (Sample 4); addition of

j m a t e r r e s t e c h n o l . 2 0 1 4;3(3):264–273 269

200 uma b c

200 um 200 um

Fig. 4 – (a) Cast; (b) homogenised at 200 ◦C; and (c) homogenised at 600 ◦C.

ZpaTaptCdt

n and Mn decreases this range significantly; however theirresence delays austenite formation and completion (As

nd Af) significantly as compared to the ternary alloys.his means that whereas the alloys without quaternarydditions would be better suited for its shape memoryroperties, ternary alloys would be better suited for higherransition temperatures. It may be mentioned that theu–Al–Mn alloys are predominantly targeted for improved

uctility, and Cu–Al–Zn alloys targeted for high transformationemperatures.

a b

d e

g

200 um

200 um

Fig. 5 – Microstructure of

3.5. Effect of alloying additions

Grain structure with � + � phases, a pre requisite for martens-ite formation on quenching is observed in the micrographs ofall the samples, indicating potential for exhibiting the shapememory behaviour. There is a large variation in the grain sizeof cast alloys (Table 2). Ternary alloys (Samples 1 and 4) showlower grain size as compared to quaternary alloys; however,

presence of any particular element cannot be said to affectthe grain size.

c

f

200 um 200 um

100 um

100 um

Quenched samples.

270 j m a t e r r e s t e c h n o l . 2 0 1 4;3(3):264–273

a

Inte

nsity

, arb

itrar

y va

lues

0 10 20 30 40 50 60 70 80 902θ

Quenched

Cast

Sample 2

CuZn Al

CuZn

Cu/CuZn/AlCu

AlCu AlCu

Cu CuAl/Cu AlCu

b

c

Inte

nsity

, arb

itrar

y va

lues

0 10 20 30 40 50 60 70 80 90

2θ

Sample 3

AlCu

Quenched

Cu/AlCu

Cast

AlCuCu

Al2CuCu/AlCu

Al4Cu 0

AlCu

AlCu/AlCu

Inte

nsity

, arb

itrar

y va

lues

0 10 20 30 40 50 60 70 80 90

2θ

100

Sample 4

Quenched

Cast

Mn

Mn

Alcu

Alcu3Al2Cu 2AlMn8.4

Al2Cu 2AlMn8.4

d

Inte

nsity

, arb

itrar

y va

lues

0 10 20 30 40 50 60 70 80 902θ

Cast

Quenched

Sample 5 AlCu3 Cu3Al 2, Cu

AlCu/Cu

e

Inte

nsity

, arb

itrar

y va

lues

0 20 40 60 80 100

2θ

Sample 6

Quenched

Cast200

400

600

800

1000

1200

0

AlCu, Cu-Zn

AlCu, Cu-Zn

Cu, CuAl, Cu-Zn, Al 4Cu 0

Cu, CuAl, Cu-Zn

f

g

mW

Sample 14

2

0

–2

–4

–6

0 50 100 150 200 250 300 350 400 450 500 550

°C

Inte

nsity

, arb

itrar

y va

lues

Quenched

Sample 1

0 10 20 30 40 50 60 70 80 902θ

Cast

AlCu

Al

Al

Al Cu/Al2Cu

Cu/Al/Cu

Cu/Al/CuCu

AlCu

CuCu

CuZn

CuZnAlCu

Al2Cu

Fig. 6 – (a) XRD pattern of cast and quenched samples: sample 1; (b) sample 2; (c) sample 3; (d) sample 4; (e) sample 5; and(f) sample 6.

j m a t e r r e s t e c h n o l . 2 0 1 4;3(3):264–273 271

mW

b

°C

mW

Sample 3

0 50 100 150 200 250 300 350 400 450 500 550

2.0

1.5

1.0

0.0

–0.5

–1.0

–2.0

0.5

–1.5

°C

44

2

0

–2

–4

–6

0 50 100 150 200 250 300 350 400 450 500 550

Sample 1

a

c

°C

Sample 4 1

0

–1

50 100 150 200 250 300 350 400 450 500 550

&R-2R-2, 8.2300mg

d

°C

50 100 150 200 250 300 350 400 450 500 550

1

0

–1

mW

–2

Sample 5

e

mW

Sample 6

50 100 150 200 250 300 350 400 450 500 550

–4

–2

0

2

4

6

8

RR 5–10

f

mW

Sample 7

5

0

–5

50 100 150 200 250 300 350 400 450 500 550

°C

g

°C

mW

Sample 2

0 50 100 150 200 250 300 350 400 450 500 550

1.0

0.5

0.0

–0.5

–1.0

–1.5

–2.0

mW

°C

Fig. 7 – (a) DSC of samples: sample 1; (b) sample 2; (c) sample 3; (d) sample 4; (e) sample 5; (f) sample 6; and (g) sample 7.

ita

Two typical types of martensitic phase has been precip-

tated depending on the alloying elements. Flower shapedypical of �-phase is observed in quenched Cu–Al–Znlloys plate- or spear-like martensite [the zig-zag �′ phase]

in the quenched ternary Cu–Al–Mn (Sample 4) alloy. In

alloys with quaternary additions, the granular structure isretained even in the quenched condition; however precipi-tation of the martensitic phases is observed in the grains

272 j m a t e r r e s t e c h n o l . 2 0 1 4;3(3):264–273

Table 5 – Identified phases of Samples in Cast and Quenched Conditions.

Sample Condition Phases identified by XRD

1 Cast Cu CuAl Al2Cu AlQuenched Cu CuAl Al2Cu Al Cu–Zn

2 Cast Cu CuAl AlQuenched Cu CuAl Al Cu–Zn

3 Cast Cu AlQuenched Cu CuAl Al4Cu9

4 Cast Cu CuAl Al2Cu Mn AlMn8.4

Quenched Cu CuAl Al2Cu Al4Cu9 Mn MnAl6 AlMn8.4

5 Cast Cu CuAlQuenched Cu Al3Cu MnAl6

6 Cast Cu CuAl Cu–ZnQuenched Cu CuAl Cu–Zn Al4Cu9

7 Cast Cu CuAl Cu–ZnQuenched Cu CuAl Cu–Zn Al4Cu9

Table 6 – Transformation temperatures.

Composition Sample No. Transformation temperatures (◦C)

As Af Ms Mf

78Cu–12Al–10Zn 1 530 550 510 48074Cu–12Al–10Zn–4Ni 2 485 525 450 41075Cu–10Al–10Zn–5Ni 3 530 560 525 485Cu–8Al–10Mn 4 470 520 500 430Cu–12Al–4Mn–2Ni 5 502 548 470 430

Cu–12Al–4Ni–10Zn–5Mn 6 510Cu–12Al–4Ni–10Zn–7Mn 7 402

but they are not very well resolved even in the FESEMphotographs.

Identified peaks of the X-ray diffraction patterns confirmformation of martensitic phases in most of the quenched sam-ples. Although a number of phases have been precipitated insome cases, the effect of quaternary additions could not bevery clearly estimated in the X-ray diffraction peaks, mainlybecause of the small percentages in which the additions weremade and keeping in mind that XRD peaks relate to onlyphases with more than 3% presence. In Samples 3, 6 and 7,only martensitic phases have been identified in the quenchedcondition signifying perhaps completion to the martensiticphase.

Addition of Ni in the Cu–Al–Zn reports a decrease in thetransition temperatures when the percentage is Cu is reducedby while maintaining the same Al and Zn contents (comparingSamples 1 with 2 and 3); this again is regained by decreasingthe Al content and increasing the Cu content (Sample 4 show-ing increased Af, Ms and Mf temperatures) slightly which goesto show how even slight variations can affect the transitiontemperatures. Presence of Zn in place of Mn (comparing 1 with6 and 7) is seen to marginally increase the transformation tem-peratures though the same cannot be said when Ni has beenadded in either case.

The range of martensite retention is the maximum internary Cu–Al–Mn alloys (Sample 4); addition of Zn andMn decreases this range significantly (other compositions);

however their presence delays austenite formation and com-pletion (As and Af) significantly as compared to the ternaryalloys.

535 530 520460 452 390

4. Conclusions

Grain structure with � + � phases, a pre requisite for martens-ite formation on quenching is observed in the micrographs ofall the cast and homogenised samples, indicating potential forexhibiting the shape memory behaviour. Two typical types ofmartensitic phase have been precipitated; flower shaped typ-ical of �-phase is observed in quenched Cu–Al–Zn alloys andplate- or spear-like martensite [the zig-zag �′ phase] in thequenched ternary Cu–Al–Mn alloy.

Identified peaks of the X-ray diffraction patterns confirmformation of martensitic phases in the quenched samples.

The DSC studies indicate clear transformation peaks inall the samples. Presence of alloying elements affects thetransformation temperatures. The range of martensite reten-tion is the maximum in ternary Cu–Al–Mn alloys; additionof Zn and Mn decreases this range significantly; howevertheir presence delays austenite formation and completion(As and Af) significantly as compared to the ternary alloys.This means that whereas the alloys without quaternary addi-tions would be better suited for its shape memory properties,ternary alloys would be better suited for higher transitiontemperatures.

Conflicts of interest

The authors declare no conflicts of interest.

o l . 2

A

TpMiBpPtgatt

r

j m a t e r r e s t e c h n

cknowledgments

he authors are thankful to CSIR, New Delhi for sponsoring theroject on ‘Design and Development of Thermo Responsive &agnetic Shape Memory Materials and Devices for Engineer-

ng Applications’ under its 12th Five Year Plan to CSIR, AMPRIhopal under which the related activities were carried out; aart of the findings have been reported in the present paper.roject Fellows are thankful to CSIR, New Delhi for grantingheir fellowship under the project. Authors acknowledge withratitude the contribution of Dr. V. Sampath, Professor, Met-llurgical Engineering Department, IIT Chennai for extendinghe melting facilities at his disposal for synthesising some ofhe alloys which have been dealt with in this paper.

e f e r e n c e s

[1] Funakubo H. In: Kennedy JB, editor. Shape memory alloys.New York: Gordon and Breach Science Publishers; 1987.p. 226.

[2] Wayman CM, Duerig TW. An introduction to martensite andshape memory. In: Duerig TW, Melton KN, Stöckel D,Wayman CM, editors. Engineering aspects of shape memoryalloys. Oxford: Butterworth-Heinemann; 1990. p. 3–20.

[3] Stalmans R, Van Humbeeck J, Delaey L. Training and the twoway memory effect in copper based shape memory alloys.J Phys IV 1991;1(C4):403–8.

[4] Marukawa K, Tsuchiya K. Two important aging effects on themartensite phase in Cu–Zn–Al alloys: rubber effect andstabilization of martensite. J Phys IV France 2001;11:Pr8–39.

[5] Sotou Y, Omori T, Kainuma R, Ishida K. Ductile Cu–Al–Mnbased shape memory alloys: general properties andapplications. Mater Sci Technol 2008;24(8):896–901.

[6] Ma J, Karaman I, Noebe RD. High temperature shape memoryalloys. Int Mater Rev 2010;55(5):257–65.

[7] Saule F, Ahlers M, Kropff F, Rivero EB. The martensitic phasesand their stability in Cu–Zn and Cu–Zn–Al alloys: IV. Theinfluence of lattice parameter changes and evaluation ofphase stabilities. Acta Metall 1992;40(12):3229–38.

[8] Cao L-F, Wang M-P, Li Z, Xu B, Su Y-C. Thermal cycling effect

in Cu–11.9Al–2.5Mn shape memory alloy with high Mstemperature. Trans Nonferr Soc China 2002;12(4):716–9.

[9] Qingzhou W, Fusheng H, Chunxiang C, Shaojing B, Ling B.Effect of aging on the reverse martensitic phase

0 1 4;3(3):264–273 273

transformation behaviours of Cu–Al–Mn shape memoryalloys. Mater Lett 2007;61:5185–7.

[10] Chen J, Lia Z, Zhao YY. A high-working-temperatureCuAlMnZr shape memory alloy. J Alloys Compd2009;480:481–4.

[11] Dagdelen F, Gokhan T, Aydogdu A, Aydogdu Y, Adigüzel O.Effects of thermal treatments on transformation behaviourin shape memory Cu–Al–Ni alloys. Mater Lett2003;57(5/6):1079–85.

[12] Sutou Y, Koeda N, Omori T, Kainuma R, Ishida K. Effect ofaging on stress induced martensitic transformations inductile Cu-Al-Mn based shape memory alloys. Acta Mater2009;57:5759–70.

[13] Stanciu S, Bujoreanu L-G, Özkal B, Lutfi Övecoglu B, VictorSandu A. Study of precipitate formation in Cu–Al–Ni–Mn–Feshape memory alloys. J Optoelectron Adv Mater2008;10(6):1365–9.

[14] Kayali N, Ozgen S, Adigiizel O. The influence of ageing onmartensite morphology in shape memory CuZnAl alloys.J Phys IV France 1997;7(C5):317–22.

[15] Sarı U, Kirindi T. Effects of deformation on microstructureand mechanical properties of a Cu–Al–Ni shape memoryalloy. Mater Charact 2008;59:920–9.

[16] Effects of additional elements on the super plasticity of aCu–14Al–3Ni shape memory alloy. Mater Trans JIM1989;30(12):999–1008.

[17] Shajil N, Das D, Chandrasekaran L. Effects of cycling on thepseudoelastic properties of CuAlMnNi & TiNi basedpseudoelastic alloys. Int J Struct Changes Solids Mech Appl2009;1(1):171–85.

[18] Chen Y, Zhang X, Dunand DC, Schuh CA. Shape memory andsuperelasticity in polycrystalline Cu–Al–Ni microwires. ApplPhys Lett 2009;95(17):171906.

[19] Sutou Y, Koeda N, Omori T, Kainuma R, Ishida K. Effects ofageing on bainitic and thermally induced martensitictransformations in ductile Cu–Al–Mn-based shape memoryalloys. Acta Mater 2009;57:5748–58.

[20] Cheng X, Huang F, Li N, Wu X. Microstructure and shapememory effect of Cu–26.1Zn–4.8Al Alloy. J Wuhan UnivTechnol Mater Sci Ed 2008:717–9.

[21] Asanovic V, Delujc K. The mechanical behaviour and shapememory recovery of Cu–Zn–Al alloys. Assoc Metall EngSerbia. UDC: 669.35’571.018.2:620.172.7=20.

[22] Xiaomin C, Feng H, Na LI, Xingwen WU. Microstructure andshape memory effect of Cu–26.1Zn–4.8Al alloy. J Wuhan Univ

Technol Mater Sci Ed 2008:717–9.

[23] Sutou Y, Omori T, Kainuma R, Ono N, Ishida K. Enhancementof superelasticity in Cu–Al–Mn–Ni shape-memory alloys bytexture control. Metall Mater Trans A 2002;33A:2817–24.