Microstructure, superelasticity and shape memory effect by stress – [622434]

Materials Science and Engineering
B
Manuscript Draft

Manuscript Number: MSB -D-17-01666

Title: Microstructure, superelasticity and shape memory effect by stress –
induced martens ite stabilization in Cu -Al-Mn-Ti shape memory alloys

Article Type: Research paper

Keywords: Cu -Al-Mn; precipitates; martensite stabilization;
superelasitic; shape memory effect.

Graphical Abstract (for review)

Highlights
 Cu–Al–Mn–Ti alloys consist of L2 1–Cu2AlMn, L2 1–Cu2TiAl and few 2H(γ' 1)
martensite .
 Cu2TiAl precipitates stabilize the s tress-induced 2H( γ
1) martensite from
Cu2AlMn parent.
 The same alloy not only exhibits superelasticity, but also shape memory effect
when heated.
 The largest superelasticity and shape memory strains reach to 4.6% and 2.5%,
respectively.

Highlights (for review)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
1
Microstructure , superelasticity and shape memory
effect by stress -induced martensite stabilization in
Cu–Al–Mn–Ti shape memory alloys
Xinren Chen1, Fan Zhang1, Mengyuan Chi1, Shuiyuan Yang1,2*, Cuiping Wang1,
Xingjun Liu1
1Fujian Key Laboratory of Material s Genome, College of Materials, Xiamen
University, Xiamen 361005, P. R. China
2 Shenzhen Research Institute of Xiamen University , Shenzhen, 518000, China.
* Corresponding author: [anonimizat]

Abstr act:
In this study, the microstructure, martensitic transformation, superelasitic,
stabilization of stress -induce martensite and the corresponding shape memory effect of
Ti-doped Cu–Al–Mn shape memory alloys were investigated. The results show that
the pre sent Cu–Al–Mn–Ti alloys consist of two types of L2 1 phases, including
Cu2AlMn parent and Cu2TiAl precipitat e. The dispersive L2 1–Cu2TiAl precipitat es
strengthen the stabilization of stress -induced 2H(γ' 1) martensite from L2 1–Cu2AlMn
parent . Thus, all studied alloys show not only superelasticity at deformation , but also
shape memory effect after unloading by heating. Consequently, the *Manuscript
Click here to view linked References

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
2
Cu–12.9Al–8.3Mn–0.5Ti alloy (wt.% ) possesses the best superelastic strain up to 4.6%
and a shape memory effect of 1.5% under a pre -deformation of 11% , while
Cu–12.8Al –7.7Mn –2.6Ti alloy (wt.% ) shows the best shape memory effect up to 2.5%
simultaneously showing 1.9% of super elastic strain with a pre-deformation of 10%.

Keywords: Cu–Al–Mn; Precipitates ; Martensite stabilization ; Superelasitic ; Shape
memory effect

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
3
1. Introduction
Shape memory alloy s (SMA s) are typical functional materials . They have been
widely used in biome dical and industrial applications because of their functional
properties, namely superelastic (SE) and shape memory effect (SME) [1-4]. Cu-based
SMAs , such as Cu–Al-based and Cu –Zn-based alloys with low production cost and
high conductivity , have drawn great attention s [5]. Representative alloys are Cu –Zn–Al
[6], Cu–Al–Ni [7] and Cu –Al–Mn [8]. However, the polycrystalline Cu –Al–Ni and
Cu–Zn–Al SMAs are too brittle to be cold worked due to the high degree of order and
high elastic anisotropy that exist in the parent austenitic β phase [9]. According to
previous studies, the Cu –Al–Mn SMAs with the Al content below 18 at. % show
excellent c old workability, good ductility and satisfied SM E and SE properties. They
are grateful to lower the order degree and widen the composition region of β(A2)
parent through replacing Al by Mn [10-12]. In addition, the precipitation of γ 1 (Cu 9Al4)
which lead s to poor thermal working stability and easy intergranular fracture is avoided
because of the shift of A2 phase region to lower Al range [13].
Up to now, alloying in Cu–Al–Mn SMA s mainly focus es on those elements
studied by [14-17] in order to further improve SE and SME performance s, as well as
the critical stress for indu cing martensitic tr ansformation and the fatigue strength to
benefit its further practical applications . However, in our previous investigations, we

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
4
find a kind of interesting Cu–Al–Mn-based SMAs having bcc phase separation or
liquid separation phenomenon resulted from the ad ditions of Fe [18], Cr [19] and V
[20]. Although these alloys are still L2 1 parent, but the stabilization of stress -induced
martensite occurs during deformation and unloading. The residual strain can be
recovered through following heating, exhibiting SME property. Therefore, this type of
Cu–Al–Mn–Fe/Cr/V alloys can simultaneously have SE and SME under the same
deformation temperature.
Bases on the phase equilibria of Cu –Fe/Cr/V [21], it is found that the metastable
liquid separation may exist in above systems and the solubilities of Fe/Cr/V in
Cu-rich parent are very low . Therefore , it leads to the precipitation of fine bcc second
particles. The existence of fine precipitates is beneficial to stabilize stress -induced
2H(γ
1) martensite during deformation through hinder ing the movement of habit plane
during reverse transformation [19, 22, 23]. In the present study, we consider about
another situation (Ti-doped). From the phase diagram of Cu –Ti and Ti –Al systems
[21], it can be seen that although liquid separation does not exist in Cu -Ti system, the
solubility of Ti in Cu is very small. Furthermore, Ti and Al can react to form bcc
phase with wide composition range. Therefore, this paper focuses on the effects of Ti
addition (0.5 ~ 4.3 wt.% ) on the microstructure, martensitic transformation, SE ,
stabilization of stress -induced martensite and the corresponding SME of Cu –Al–Mn

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
5
alloys.

2. Experimental procedures
Cu (99.95 wt.%), Al (99.99 wt.%) , Mn (99.7 wt.%, pickled by hydrochloric acid),
Ti (99.5 wt.%) were used as starting materials to prepare four Cu–Al–Mn–Ti alloys.
Firstly, pure raw materials were cut and ultrasonic cleaned by acetone. Then the
prepared 30g raw materials were melted under a high purity argon atmosphere using a
non-consumable tungsten electrode and cooled by water -cooled copper crucible. In
order to ensure homogeneity, each cast was turned over and re -melted for five times.
After that, all samples were wire -electrode cut into small bulks. Subsequently, the
samples were sealed by quartz ampoules with 0.5 bar argon atmosphere and anneal ed at
900 ℃ for 24 hours followed by ice -water quenching.
The microstructure was characterized by backscattered electron photographing
(BSE) and t he chemical compositions of whole alloy and each phase were further
confirmed by electron probe microanalysis (EPMA) as shown in Table 1. For the sake
of description in the following text , four studied alloys of Cu–12.9Al–8.3Mn–0.5Ti ,
Cu–12.3Al–8.6Mn–1.9Ti , Cu–12.8Al–7.7Mn–2.6Ti and Cu–13.7Al–7.4Mn–4.3Ti
were called as Ti -0.5, Ti-1.9, Ti-2.6 and Ti-4.3, respectiv ely. Then the samples were
etched by 10 g FeCl 3 + 25 ml HCl + 100 ml H 2O and observed by optical microscopy

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
6
(OM). Transmission electron microscopy (TEM), selected area diffraction pattern
(SAD P) and PANalytical X'pert PRO X -ray diffractometer with Cu Kα radiation were
used to detect the crystal structure. The martensitic transformation temperatures were
determined by differential scanning calorimetry (DSC) at a heating and cooling rate of
10 ℃·min−1.
The cylindrical specimens (3mm diameter and 5mm heig ht) for compressive tests
were cut from the quenched bulks to investigate stress -strain behaviors and shape
recovery properties. The pre -strain (εpre) targets were 7%, 8%, 9%, 10%, 11%
respectively for five samples of one alloy at room temperature. All sam ples were
unloaded to a zero stress condition after deformation. The SE (εSE) strains were
confirmed and calculated directly by stress -strain curves [17]. Thermal mechanical
analysis (TMA) was used under an argon atmosphere at a heating and cooling rate of
10 ℃·min−1 from room temperature to 600 ℃ to measure the SME (εSME). The
metallographic microstructures after 10% compressive test of each alloy were obser ved
by optical microscopy (OM) and transmission electron microscopy (TEM). The height
of the sample was measured before loading ( h0) and after unloading ( h1). The height
after TMA tests was measured and denoted as h2. The residual strain ( εr) was calculate d
as εr = (h0 – h1)/h0 × 100%. The εe implies the elastic strain when unloading to a zero
stress condition. Therefore, the εSE strains were calculated using the formulas εSE = (εpre

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
7
– εe – εr). The SME strains ( εSME) after TMA tests were calculated as εSME = (h2 – h1)/h0 ×
100% [10].

3. Results and discussion
3.1. Microstructure and martensitic transformation
The metallographic and BSE images of the four Cu –Al–Mn–Ti alloys after
quenching from 900 ℃ heat treatment for 24 hours were shown in Fig. 1. It can be seen
clearly that four studied alloys consist of two phases , i.e. the matrix (L2 1-Cu2AlMn )
and the pre cipitate (L2 1-Cu2TiAl) (see the details in the next section). The precipitates
distribute not only in the grain boundaries but also in the grains of the parent phase , and
its amount gradually increases with the increase of Ti content. The chemical
composit ions of each phase w ere determined, as shown in Table 1. From Table 1. the
solid solubility of Ti in Cu2AlMn parent is only 0.25 wt.% ~ 0.36 wt.% , whereas the Ti
content in Cu2TiAl precipitate is more than 20 wt.%.
By TEM characterization (Fig . 2a), the c ycloidal Ti -rich particles are further
conformed as L2 1-Cu2TiAl by Fig. 2b. Also, the parent phase is confirmed as L2 1
-Cu2AlMn from Fig. 2c. The result is consistent with Raman and Schubert [24] who
pointed out that the crystal structure of the Cu 2TiAl compound is the same as that of the
Cu2MnAl phase with the L2 1 Heusler type. According to the phase equilibria of

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
8
Cu–Ti–Al ternary [25], the L2 1-Cu2TiAl would precipitate when the alloys are cooled
from melted liquid. Furthermore, the L2 1-Cu2TiAl precipitates would stay stabl e even
though the alloys are annealed at 900 ℃. At this time, the matrix phase would process
ordered transformation during cooling as follow: A2 → B2 → L21-Cu2AlMn . The
calculated lattice constants from SAD P of L2 1 -Cu2TiAl and L2 1-Cu2AlMn are 6.01 Å
and 5.86 Å respectively [24, 26]. From XRD characterizations, as is shown in Fig. 3,
there are two sets of L2 1 pattern with different lattice constants respectively
representing L2 1-Cu2AlMn phase and L2 1-Cu2TiAl phase in 900 ℃ annealed samples.
Besides, trace amounts of 2H (γ' 1) martensite is also found in Fig. 3 , which was
possibly caused by quenching stress. The existence of 2H (γ' 1) martensite phase is also
confirmed by TEM photographing, see Fig. 2a. Addi tionally, the high magnification of
the parent phase (Fig. 2d) shows that there are many small er L21-Cu2TiAl particles
with rectangle or cross shape dispersedly distributing inside the parent grain (Fig. 2e).
Moreover, from XRD characterizations in Fig. 3, the diffraction peaks of L21-Cu2TiAl
and 2H (γ' 1) martensite become clearer with the increase of Ti content.
According to the DSC curves of each alloy in Fig. 4, the start temperatures ( Ms and
As) and the finish temperatures ( Mf and Af) of the forward and reverse martensitic
transformations were determined, as shown in Table 2. The transformation thermal
hysteresis is calculated using the formula: Af – Ms. From Table 2, it is found that the

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
9
reversible martensitic transformation temperatures gradually increase with the
increase of Ti addition, whi ch is similar to the report by Canbay [16]. However, the
contents of Ti in Cu 2AlMn parent in four alloys are very small, 0.25 ~ 0.36 wt.% (Table
1). Thus it is considered that the martensitic transformation temperatures of the alloys
still depend on the contents of Al and Mn in Cu2AlMn parent [12].

3.2. SE and SME characteristics
As can be seen from the compressive fracture stress -strain curves of four
Cu–Al–Mn–Ti alloys in Fig. 5, the compressive fracture strains and stresses (t he
fracture points are represented by ×) were measured to be: 12.7% and 1150 MPa for
Ti-0.5 alloy, 14.4% and 1066 MPa for Ti -1.9 alloy, 16.3% and 1182 MPa for Ti -2.6
alloy, 16.9% and 1387 MPa for Ti -4.3 alloy. And it is easy to find that the fracture strai n
ascends with more content of Ti. For Ti -0.5 alloy, there is a clear stress plateau which
represents the process of stress -induced martensitic transformation. The corresponding
stress -strain curves under the different pre-strains from 7% to 11% are shown in Fig. 6.
The SE strains ( εSE) were directly measured by stress -strain curves and presented in Fig.
6 [17]. And the SME strains ( εSME) were obtained by measuring the height changes
after the deformed sample were heated to 600 ℃ for 5 minutes by TMA tests with a
heating and cooling rate of 10 ℃·min-1. The relationships between the compressive

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
10
strain s and the εr, εSE, εSME were shown in Fig. 7a, b and c respectively.
From Fig. 6 and Fig. 7, it is interesting that all studied alloys simultaneously
exhibit SE and SME properties under the same deformation temperature. This is
consistent with our previous reports of Cu –Al–Mn–Fe/Cr/V alloys [18-20]. The SE
and SME strain s result from the certain content of L21-Cu2AlMn parent depending on
alloy composition. Thus with the increase of Ti content, the SE st rain gradually
decreases, whereas the SME increases. The SE and SME strains of the alloy s with Ti
≤ 2.6 wt.% both gradually enhance when the pre -strain increases. The largest SE
strains are 4.6%, 2.6% and 2.4% and the SME strains are 1.5%, 2.3% and 2.5%
respectively for Ti -0.5, Ti -1.9 and Ti -2.6 alloys under the pre -strains of 10% or 11%.
Additionally, it is found that the critical stress for stress -inducing martensitic
transformation gradually decreases when the content of Ti changes from 0.5 wt.% to
2.6 w t.%, it indicates that it is easier to stabilize the stress -induced martensite,
resulting in the increase in the SM E strain. However, it is strange that the critical
stress suddenly increase s in Ti -4.3 alloy, which may resulted from the obvious
increase in the content of L2 1-Cu2TiAl precipitate s. The increase of the critical stress
should cause more plastic deformation that is irreversible . So at this time, the SE and
SME strains have a degree of reduction in both .
Fig. 8 shows the shape recovery curves of the deformed Cu –Al–Mn–Ti alloys with

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
11
a pre -strain of 10% after unloading. When the deformed alloys were heated, a n obvious
increase in the height of the sample occurs (< 100 ℃ ~ < 250 ℃), implying that the
shape recovery process es due to the reverse martensitic transformation. It is found
that the As temperatures are obviously higher than those quenched alloys in Fig. 4
without deformation. It is a clear phenomenon of the stabilization of stress -induced
martensite. Additionally, during subsequent heating and cooling, some unknown
transformations happen ed as shown by the arrows in Fig. 8. In order to research this
issue, Ti -2.6 alloy was annealed at 500 ℃ for 3 hours and ai r cooling to room
temperature, the corresponding microstructure was analyzed and provide d in Fig. 9.
Except L2 1-Cu2AlMn and L2 1-Cu2TiAl phases, m ore bright precipitates are observed
at this time in Fig. 9a and 9b. The chemical composition of this precipita te is
Cu84.83 –Al8.99 –Mn5.52 –Ti0.66 (wt.%) , having more Cu content than that of
Cu2AlMn phase. Through the TEM morphology and the corresponding SADP ( Fig. 9c
and 9d). it is confirmed as fcc α-Cu phase (In Fig. 8, the arrow 1 indicates the
precipitation of α-Cu phase under the condition of deformation). This is different from
those of Cu –Al–Mn–Cr/V [19, 20], but similar to that of ternary Cu –Al–Fe alloys
[22]. Subsequently, the aged Ti-2.6 samp le at 500 ℃ again was annealed at above
600 ℃ and followed by ice -water quenching , then the sample has the same
microstructure as that of the quenched alloy in Fig. 1. It means the dissolution of α -Cu

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
12
phase back to L2 1 parent at high temperature (according to the transformation shown
by the arrow 2 in Fig. 8) . The arrow 3 in Fig. 8 implies the precipitation of α -Cu phase
again during cooling without deformation . The precipitation of α-Cu phase will lead to
serious degradation of the SE and SME properties of the alloys. Thus it is suggested
that the operation temperature for the applications of Cu –Al–Mn–Ti alloys should be
lower than 250 ℃.

3.3. Microstructur al evolution under deformation
In order to study the microstructur al evolution of Cu –Al–Mn–Ti allo ys during
deformation, the microstructures of Cu –Al–Mn–Ti alloys after compressed to a
pre-strain of 10% and unloading were studied , and as shown in Fig. 10 and 11. Clearly,
a typical martensite morphology is observed in all deformed samples in Fig. 10, in
which the martensite comes from the L2 1-Cu2AlMn parent. Fig. 11 shows the TEM
morphology and the corresponding SADP of the stress -induced martensite, which
has 2H(γ' 1) structure. The existence of 2H(γ' 1) martensite prove s the occurrence of the
stabilization of stress -induced 2H(γ' 1) during the deformation in Cu –Al–Mn–Ti alloys .
According to the previous investigations, the stabilization of stress -induced
martensi te result s from that the dislocation s by deformation restrain the movement of
the habit plane during reverse martensitic transformation , thus the stress -induced

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
13
martensite is reserve d after unloading, namely the stabilization [15, 27, 28]. In the
present study, it is believed that the L21-Cu2TiAl precipi tates play a key factor to
stabilize the stress -induced 2H(γ' 1) martensite, especially for those finer Cu 2TiAl
particles in Fig. 2d. From the result s of TEM characterization of Ti -2.6 alloy (Fig. 11) ,
the L2 1-Cu2TiAl precipitates are always inserted betwee n martensitebattens just like
nails. In this case, as is shown in Fig. 11d, high density dislocations and stacking faults
are accompanied by small and hard L2 1-Cu2TiAl precipitates, which can hinder the
reverse martensitic transformation [22, 23, 26].
From the present results, it is found that the stabilized 2H(γ' 1) martensite can be
recovered by heating, contribut ing to the SMEs of the alloys. For the non -stabilized
2H(γ' 1) martensite during deformation, it immediately recovers during unloading,
being the same as the SE property. However, there is a differenc e between
Cu–Al–Mn–Fe and Cu –Al–Mn–Ti alloys. Those Fe -rich bcc fine particles resulted
from bcc phase separation keep completely coherent with the L2 1-Cu2AlMn parent
[18], whereas the L21-Cu2TiAl precipitates do not show coherent or semi -coherent
interfac es with L2 1-Cu2AlMn parent (Fig. 2). However, the stabilization of the
stress -induced martensite still occurs during deformation. The SE and SME behaviors
of the Cu –Al–Mn–Ti alloys are similar to those of Cu –Al–Mn–Cr/V alloys, in which
the same alloy can simultaneously has the SE and SME under the same deformation.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
14
Otherwise , when the Cu –Al–Mn–Fe alloys are deformed, almost complete SE
property is obtained while applying a relatively low stress, whereas they exhibit
instantaneous and complete SME characteri stic with high applied stress [18].
Considering their microstruct ures, it is believed that the stabilization phenomenon of
the stress -induced martensite in these Cu –Al–Mn-based alloys is closely related to the
shape, size and distribution of those fine precipitates, as well as the coherency
between the precipitate s and the parent. This is a worthy question to further study .

4. Conclusions
In this paper, four Cu –Al–Mn–Ti SMAs with different Ti content s were designed
and their microstructure, phase transformation, stress -strain behaviors, SE and SME
properties were studi ed. The conclusions are summarized as follow:
(1) The addition of Ti in Cu –Al–Mn SMAs lead s to a mixed microstructure consisted of
dominant L21-Cu2AlMn parent, dispersed L2 1-Cu2TiAl phase and small amount of
2H(γ’
1) martensite. With the increase in Ti content, the amount of L21-Cu2TiAl
precipitates and the reversible martensitic transformation temperature gradually
increase. In addition, the L21-Cu2TiAl phase does not show the martensitic
transformation. The compressive fracture strain also enhances when increasing the
amounts of L21-Cu2TiAl phase .

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
15
(2) Fine L21-Cu2AlMn particles are very favorable to stabiliz e the stress -induced
2H(γ’
1) martensite , in which high density dislocations and stacking faults form around
fine L21-Cu2AlMn particles dispersively distributing within those stress -induced
martensite. This will hinder the occurrence of the reverse martensitic transformation
during unloading. As the stabilized 2H(γ' 1) martensite can be recovered by heatin g,
the SME shows . The other part of the non-stabilized 2H(γ' 1) martensite immediately
recover s during unloading, exhibiting SE property. With the increase of Ti content, the
SE strain gradually decreases, but the SME increases.
(3) When the compressive deformation increases, the SE and SME strains both
gradually enhance when Ti ≤ 2.6 wt.%. The largest SE strains are 4.6%, 2.6% and
2.4% and the SME strains are 1.5%, 2.3% and 2.5% respectively for Ti -0.5, Ti -1.9
and Ti -2.6 alloys. With further increasing Ti a ddition to 4.6 wt.%, the SE and SME
properties both start to deteriorate .

Acknowledgements
We acknowledge the financial supports from the Shenzhen science and technology
project , grant number JCYJ20170306142550151, the Fundamental Research Funds
for the Central Universities , grant number 20720160078 .

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
16
References
[1] J.M. Jani, M. Leary, A. Subic, M.A. Gibson, A review of shape memory alloy
research, applications and opportunities, Mater. Des. 56 (2014) 1078 -1113.
[2] K. Otsuka, C. Wa yman, Mechanism of shape memory effect and superelasticity,
Shape memory materials (1998) 27 -48.
[3] K. Otsuka, X. Ren, Recent developments in the research of shape memory alloys,
Intermetallics 7 (1999) 511 -528.
[4] K. Otsuka, X. Ren, Physical metallurgy of Ti –Ni-based shape memory alloys,
Progress in materials science 50 (2005) 511 -678.
[5] L. Sun, W.M. Huang, Z. Ding, Y. Zhao, C.C. Wang, H. Purnawali, C. Tang,
Stimulus -responsive shape memory materials: a review, Mater. Des. 33 (2012)
577-640.
[6] J. Per kins, W. Muesing, Martensitic transformation cycling effects in Cu -Zn-Al
shape memory alloys, Metallurgical and Materials Transactions A 14 (1983) 33 -36.
[7] K. Otsuka, C. Wayman, K. Nakai, H. Sakamoto, K. Shimizu, Superelasticity effects
and stress -induce d martensitic transformations in Cu -Al-Ni alloys, Acta Metallurgica
24 (1976) 207 -226.
[8] Y. Sutou, T. Omori, R. Kainuma, K. Ishida, Ductile Cu –Al–Mn based shape
memory alloys: general properties and applications, Materials Science and Technology

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
17
24 (2008 ) 896 -901.
[9] Y. Sutou, T. Omori, J. Wang, R. Kainuma, K. Ishida, Characteristics of
Cu–Al–Mn-based shape memory alloys and their applications, Materials Science and
Engineering: A 378 (2004) 278 -282.
[10] R. Kainuma, S. Takahashi, K. Ishida, Thermoelasti c martensite and shape memory
effect in ductile Cu -Al-Mn alloys, Metallurgical and Materials Transactions A 27
(1996) 2187 -2195.
[11] R. Kainuma, S. Takahashi, K. Ishida, Ductile shape memory alloys of the
Cu-Al-Mn system, Journal de Physique IV 5 (1995) C 8-961-C968 -966.
[12] U. Mallik, V. Sampath, Influence of aluminum and manganese concentration on
the shape memory characteristics of Cu –Al–Mn shape memory alloys, J. Alloy.
Compd. 459 (2008) 142 -147.
[13] L. Matlakhova, E. Pereira, A. Matlakhov, S. Monteir o, R. Toledo, Mechanical
behavior and fracture characterization of a monocrystalline Cu –Al–Ni subjected to
thermal cycling treatments under load, Materials Characterization 59 (2008)
1630 -1637.
[14] Y. Sutou, R. Kainuma, K. Ishida, Effect of alloying eleme nts on the shape memory
properties of ductile Cu –Al–Mn alloys, Materials Science and Engineering: A 273
(1999) 375 -379.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
18
[15] I. Hurtado, P. Ratchev, J. Van Humbeeck, L. Delaey, A fundamental study of the
χ-phase precipitation in Cu -Al-Ni-Ti(Mn) shape memor y alloys, Acta materialia 44
(1996) 3299 -3306.
[16] C.A. Canbay, Z.K. Genc, M. Sekerci, Thermal and structural characterization of
Cu–Al–Mn–X (Ti, Ni) shape memory alloys, Applied Physics A 115 (2014) 371 -377.
[17] Y. Sutou, N. Koeda, T. Omori, R. Kainuma, K. Ishida, Effects of aging on
stress -induced martensitic transformation in ductile Cu –Al–Mn-based shape memory
alloys, Acta Materialia 57 (2009) 5759 -5770.
[18] S. Yang, T. Omori, C. Wang, Y. Liu, M. Nagasako, J. Ruan, R. Kainuma, K.
Ishida, X. Liu, A ju mping shape memory alloy under heat, Scientific reports 6 (2016).
[19] S. Yang, F. Zhang, J. Wu, J. Zhang, C. Wang, X. Liu, Microstructure
characterization, stress –strain behavior, superelasticity and shape memory effect of
Cu–Al–Mn–Cr shape memory alloys, Journal of Materials Science 52 (2017)
5917 -5927.
[20] S. Yang, F. Zhang, J. Wu, Y. Lu, Z. Shi, C. Wang, X. Liu, Superelasticity and
shape memory effect in Cu –Al–Mn–V shape memory alloys, Mater. Des. 115 (2017)
17-25.
[21] W.G. Moffatt, The handbook of bi nary phase diagrams, Genium Pub. Corp.
1976.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
19
[22] S.Y. Yang, Y. Su, C.P. Wang, X.J. Liu, Microstructure and properties of Cu -Al-Fe
high-temperature shape memory alloys, Mater. Sci. Eng. B -Adv. Funct. Solid -State
Mater. 185 (2014) 67 -73.
[23] P.F. Guilem any J M, Lovey F C, et al., TEM study of β and martensite in Cu –
Al-Mn shape memory alloys, Materials Characterization 26 (1991) 23 -28.
[24] S.K. Raman A, On the construction of some alloys related to TiAl 3. III:
Investigations on the Ti –Ni–Al and T –Cu–Al system (in German), Z. Metallk 56
(1965) 99.
[25] X. Liu, C. Wang, I. Ohnuma, R. Kainuma, K. Ishida, Phase equilibria and phase
transformation of the body -centered cubic phase in the Cu -rich portion of the Cu –Ti–Al
system, Journal of Materials Research 23 (2008) 2674 -2684.
[26] S. J, X ‐ray diffraction research of the order –disorder transitions in the ternary
heusler alloys B2MnAl (B= Cu, Ni, Co, Pd, Pt), physica status solidi (a) 66 (1981)
485-491.
[27] J. Oliveira, F.B. Fernandes, N. Schell, R. Miranda, Martensite stabilization during
superelastic cycling of laser welded NiTi plates, Materials letters 171 (2016) 273 -276.
[28] S.R. Van Humbeeck J, Shape memory alloys, types and functionalities,
Encyclopedia of Smart Materials (2002).

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
20

Figure captions
Fig. 1. Optical micrographs a nd BSE images of Cu –Al–Mn–Ti alloys after quenching
from 900℃ heat treatment for 24 hours . (a) and (b) Ti -0.5; (c) and (d) Ti -1.9; (e) and (f)
Ti-2.6; (g) and (h) Ti -4.3.
Fig. 2. TEM morphology and electron diffraction patterns of Ti -2.6 alloy after
quench ing from 900 ℃. (a) TEM morphology; (b) selected area diffraction pattern
(SADP) of the Cu 2AlMn phase corresponding to the [011] zone pattern of the
L21-Cu2AlMn; (c) selected area diffraction pattern (SADP) of the Cu 2TiAl phase
corresponding to the [011] zo ne pattern of the L2 1-Cu2TiAl; (d) TEM morphology with
high magnification; (e) selected area diffraction pattern (SADP) (white dashed circle)
of the Cu 2AlMn phase corresponding to the [001] zone pattern of the L2 1 structure.
Fig.3. X-ray diffraction patter ns of Ti -0.5 (a), Ti -1.9 (b), Ti -2.6 (c) and Ti -4.3 (d) after

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
21
quench ing from 900 ℃ heat treatment for 24 hours .
Fig. 4. DSC curves of Cu –Al–Mn–Ti alloys. (a) Ti -0.5; (b) Ti -1.9; (c) Ti -2.6; (d) Ti -4.3.
Fig. 5. Compressive fracture stress -strain curves of Cu –Al–Mn–Ti alloys.
Fig. 6. Compressive stress -strain curves of the Cu –Al–Mn–Ti alloys with different
pre-strains from 7% to 11%.(a) Ti -0.5; (b) Ti -1.9; (c) Ti -2.6; (d) Ti -4.3.
Fig. 7. Residual strain (a), SE strain (b) and SME (c) of Cu –Al–Mn–Ti alloys as a
function of the pre -strain.
Fig. 8. TMA curves of the Cu –Al–Mn–Ti alloys deformed to a pre -strain strain of 10%
and unloading.
Fig. 9. The optical micrographs (a), BSE image (b), TEM morphology (c) and electron
diffraction pattern (white dashed circle) (d) of Ti -2.6 alloy annealed at 500 °C for 3 h
and air cooling to room temperature.
Fig. 10. Microstructures of Cu –Al–Mn–Ti alloys after compressed to a pre -strain of
10% and unloading. (a) and (b) Ti -0.5; (c) and (d) Ti -1.9; (e) and (f) Ti -2.6; (g) and (h)
Ti-4.3.
Fig. 11. TEM morphology and electron diffraction pattern of Ti-2.6 alloy with a 10%
pre-strain and unloading . (a) TEM morphology; (b) electron diffraction pattern of the
martensite corresponding to the [001] zone pattern of the 2H structure; (c) TEM
morphology (High magnification); (d) magnification of white dashe d box in Fig. 12(c) .

Figure
Click here to download high resolution image

Figure
Click here to download high resolution image

Figure
Click here to download high resolution image

Figure
Click here to download high resolution image

Figure
Click here to download high resolution image

Figure
Click here to download high resolution image

Figure
Click here to download high resolution image

Figure
Click here to download high resolution image

Figure
Click here to download high resolution image

Figure
Click here to download high resolution image

Figure
Click here to download high resolution image

Table 1 . The chemical compositions of Cu2AlMn phase and Cu2TiAl phase s in the Cu –
Al–Mn–Ti alloys after quenching from 900 ℃ heat treatment for 24 hours.

Alloys (wt.%) Cu2AlMn phase (wt.%) Cu2TiAl phase (wt.%)
Cu Al Mn Ti Cu Al Mn Ti
Cu–12.9Al–8.3Mn–0.5Ti 80.21 13.11 6.39 0.29 59.99 14.14 4.88 20.99
Cu–12.3Al–8.6Mn–1.9Ti 79.64 11.56 8.45 0.36 58.51 13.53 5.23 22.73
Cu–12.8Al–7.7Mn–2.6Ti 80.94 11.30 7.36 0.35 59.46 13.11 4.43 23.00
Cu–13.7Al–7.4Mn–4.3Ti 79.79 12.63 7.34 0.25 58.61 13.73 5.29 22.37

Table

Table 2. The martensitic transformation temperatures ( ℃) of Cu–Al–Mn–Ti alloys . Ms,
Mf are the start and finish temperatures of forward martensit ic transformation, As, Af
are the start and finish temperatures of reverse martensit ic transformation .

Alloys (wt.%) Ms Mf As Af Hysteresis
Cu–12.9Al–8.3Mn–0.5Ti -124.8 -133.6 -113.9 -107.8 17.0
Cu–12.3Al–8.6Mn–1.9Ti -84.5 -106.0 -60.2 -45.7 38.8
Cu–12.8Al–7.7Mn–2.6Ti -68.5 -107.8 -70.3 -37.2 31.3
Cu–13.7Al–7.4Mn–4.3Ti -74.6 -97.5 -61.0 -44.6 30.0

Table