A comparative study of Ni–Ti and Ni–Ti–Cu shape memory alloy processed by plasma melting and injection molding

Materials and Design 32 (2011) 4925–4930

Contents lists available at ScienceDirect

Materials and Design journal homepage: www.elsevier.com/locate/matdes

A comparative study of Ni–Ti and Ni–Ti–Cu shape memory alloy processed by plasma melting and injection molding C.J. de Araújo a,⇑, N.J. da Silva b, M.M. da Silva a, C.H. Gonzalez c a

Universidade Federal de Campina Grande (UFCG), Department of Mechanical Engineering, Av. Aprígio Veloso, 882, Bairro Universitário, Campina Grande, PB, CEP 58429-900, Brazil Instituto Federal de Educação, Ciência e Tecnologia de Pernambuco (IFET-PE), Coordination of Mechatronics, Estrada do Alto do Moura, km 3.8, Distrito Industrial II Caruaru, PE, CEP 55040-120, Brazil c Universidade Federal de Pernambuco (UFPE), Department of Mechanical Engineering, Av. Acadêmico Hélio Ramos, s/n, Cidade Universitária, Recife, PE, CEP 50740-530, Brazil b

a r t i c l e

i n f o

Article history: Received 23 April 2011 Accepted 29 May 2011 Available online 2 June 2011 Keywords: A. Non-ferrous metals and alloys C. Melting G. Thermal analysis

a b s t r a c t Shape memory alloys (SMA) are smart materials that present potential applications in such diverse areas as aeronautics, automotive, electronics, biomedicine and others. This work aimed at comparing some physical and functional properties of a Ni–Ti–Cu and equiatomic Ni–Ti SMA. Therefore, Ni–50Ti and Ni–50Ti–5Cu (at.%) were manufactured using plasma melting followed by injection in metallic mold, named Plasma Skull Push–Pull (PSPP) process. Afterwards, samples of both Ni–Ti based SMA were annealed at 1113 K during 2400 s and water quenched. The obtained specimens were analyzed by optical microscopy, microhardness, differential scanning calorimetry, electrical resistance as a function of temperature, and force generation tests. The results showed that Ni–Ti alloy presented higher levels of hardness and lower generated recover forces during heating when compared to the Ni–Ti–Cu SMA. Moreover, the Ni–Ti alloy holds hysteresis larger than the Ni–Ti–Cu SMA as a result of the presence of the R-phase transformation. There was also a better stability under thermal cycling of NiTiCu SMA compared with the equiatomic NiTi. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Among the several materials already designed and produced, there is a relatively new class denominated smart materials, such as the shape memory alloys (SMA). SMA exhibit the capacity to recover an original shape that existed before the introduction of a plastic deformation, through a reversible martensitic transformation, when submitted to a heating up to a critical temperature. Therefore, these materials are considered thermo-mechanical actuators. However, the shape memory effect (SME) is a peculiar phenomenon presented by some alloys that exhibit reversible martensitic phase transformations [1]. These special alloys present, through SME, recoverable deformations that can reach up to 5% in uniaxial tension accompanied by a thermal hysteresis [2]. Alternatively, considerable forces can be generated if these reversible deformations are restricted in some direction during the activation of the SME. This capacity of generating forces as a function of temperature allows these SMA have a great potential application in couplings and others similar devices [3,4]. This ability has also motivated the development of composites with SMA wires embedded in polymer matrix mainly targeting attenuation of mechanical

⇑ Corresponding author. Tel.: +55 83 2101 1126; fax: +55 83 2101 1272. E-mail addresses: [email protected] (C.J. de Araújo), niedson.silva@ caruaru.ifpe.edu.br (N.J. da Silva), [email protected] (C.H. Gonzalez). 0261-3069/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.matdes.2011.05.051

vibrations and controllable shape change of the composite structure [5–7]. As demonstrated recently by da Silva et al. [8], in applications involving mechanical vibrations, high damping capacity of SMA compared with conventional metallic alloys can also be harnessed to improve the performance of any mechanical system. In general, three large families of SMA are available commercially: Ni–Ti, Cu–Zn–Al and Cu–Al–Ni [1]. Due to its excellent corrosion resistance and better shape memory properties, Ni–Ti based alloys are more frequently found in commercial applications, mainly as thermo-mechanical actuators in several sectors of the engineering in general, but as well as in the medical and dental fields [1,9]. The Ni–Ti binary equiatomic alloy is widely used in commercial devices. In this SMA the martensitic transformation can be preceded by an intermediate transformation, called pre-martensitic. This intermediate phase, designated R-phase, presents a rhombohedral structure. Furthermore, this alloy presents a relatively larger thermal hysteresis compared to copper based alloys [1]. However, to meet the needs of a given project, a third element can be added to Ni–Ti binary alloys in order to modify their phase transformation temperatures, but never interfering in the ductility of the martensitic phase and the great performance of SME verified in binary alloys. Copper is one of the elements that can be added to Ni–Ti binary alloys. In Ni–Ti–Cu alloys the intermediate R-phase does not occur and the thermal hysteresis is much smaller than those in the Ni–Ti binary alloys [1,10].

4926

C.J. de Araújo et al. / Materials and Design 32 (2011) 4925–4930

Fig. 1. Melting and casting sequence of the PSPP method. (a) Piled raw materials. (b) Plasma torch. (c) SMA button. (d) SMA layer on the crucible after mold injection. (e) Metallic mold. (f) SMA tablet.

Thus, the general aim of this work is to obtain a Ni–Ti–Cu SMA (Ti–45Ni–5Cu, at.%) through melting and casting by Plasma Skull Push–Pull (PSPP) process [11], and to perform some comparisons with Ni–Ti equiatomic SMA, obtained by the same technique. This direct comparison involves some functional properties, as the response under thermal cycling and capacity of force generation as a function of temperature, as well as mechanical properties as the microhardness of both alloys. 2. Experimental procedure The SMA nominal compositions of 50Ni–50Ti (at.%) and Ti– 45Ni–5Cu (at.%), corresponding to 55Ni–45Ti (wt.%) and 49.4Ni– 44.7Ti–5.9Cu (wt.%), were produced by the Plasma Skull Push–Pull (PSPP) method [11]. Fig. 1 shows a typical PSPP sequence of fabrication for a Ni–Ti SMA. Ni and Ti are initially piled up on a copper crucible (Fig. 1a) where a protective argon atmosphere surrounds the material. The melting of the raw elements is rapidly performed by the rotation of a plasma torch created by a tungsten electrode (Fig. 1b), originating a small button (Fig. 1c). It was verified that, in order to assure the good homogeneity of the SMA, melting process should be repeated for five times [11]. The final product is obtained when the re-melted button (Fig. 1d) is injected in a metallic mold (Fig. 1e) originating a small tablet (Fig. 1f) or others desired final formats. Finally, the Ni–Ti SMA tablets were evenly heated at 1113 K (840 °C) for 2400 s followed by water quenching. After manufacturing, SMA specimens were cut up from the SMA tablets (Fig. 1f) for metallographic observation with optical microscopy (Olimpus, BX51M). Conventional metallographic procedures (sanding and polishing with alumina) were performed and the Kroll reagent (2 mL HF, 4 mL HNO3 and 100 mL H20) was used for etching. After metallographic observations, microhardness tests were carried out using a Future-Tech equipment (FM-700). The parameters used in these tests were a 2.94 N loading, 15 s of penetration and spacing of 10 3 m (1 mm) among impressions. To determinate phase transformation temperatures, as well as thermal hysteresis of Ni–Ti and Ni–Ti–Cu alloys, electrical resistance as a function of temperature tests and Differential Scanning Calorimetry (DSC) were performed. Force generation tests for Ni–Ti SMA were performed in a mechanical press with hydraulic system. Fig. 2a shows the assembly for the force generation tests in the

SMA plate sample. For this test, the Ni–Ti SMA plate was firstly deformed in a temperature below Mf (martensite finish temperature), around 233 K ( 40 °C). The plate was deformed in central deflection (y) under three point bending loading (detail of Fig. 2a). The deflection was measured using a Linear Variable Displacement Transducer (LVDT) sensor (Solartron, DF-5 model). After deformation the plate was heated using hot air jet for activation of SME and, consequently, the force generation. For the measurement of the generated force a load cell with capacity of 500 N (Excel, MS50 model) was used. For temperature measurement during the experiment a K type microthermocouple with 100 lm in diameter was installed on the sample. Signals from the load cell and thermocouple were captured by an acquisition data system (Agilent, 34970A model). For the Ni–Ti–Cu SMA sample, the force generation tests were carried out in a universal testing machine (Instron, 5582 model) (Fig. 2b). This experiment was similar to the one performed with Ni–Ti SMA. These two different and equivalent experiments were designed to demonstrate the versatility required for the determination of some functional properties of SMA. Table 1 presents the Ni–Ti and Ni–Ti–Cu plates dimensions (L length, b width and e thickness), the central deflection (y) and the distance between the supports (d) during the three point bending loading. 3. Results Fig. 3 shows the microstructures of Ni–Ti (Fig. 3a) and Ni–Ti–Cu SMA (Fig. 3b) produced by the PSPP method. The hardness profiles for these two SMA can be observed in Fig. 4. It was confirmed that Ni–Ti SMA (profile 1) presented a higher hardness level (231.08 ± 17.23 HV) than the Ni–Ti–Cu (199.46 ± 12.53 HV). The phase transformation temperatures were measured by DSC tests, as can be seen in Fig. 5. This figure shows the DSC curves obtained for Ni–Ti (Fig. 5a) and Ni–Ti–Cu SMA (Fig. 5b), with its phase transformation temperatures measured by the tangent method. Table 2 summarizes all phase transformation temperatures and thermal hysteresis (HT) obtained directly from Fig. 5a and b. HT was defined as the difference between the peak transformation temperatures (AP MP) for the reverse and forward transformations during heating and cooling, respectively. Fig. 6 shows the electrical resistance as a function of temperature curves for Ni–Ti (Fig. 6a) and Ni–Ti–Cu SMA (Fig. 6b), as well

4927

C.J. de Araújo et al. / Materials and Design 32 (2011) 4925–4930

Detail of bend

Displacement Sensor

Plate being deformed

MechanicalPress

Detail of bend Load cell

Ni-Ti SMA

Plate being deformed

Thermocouple

Ni-Ti-Cu SMA

(a)

(b)

Fig. 2. Assemblages for tests of force generation by three point bending tests. (a) Ni–Ti plate specimen on the hydraulic press and (b) Ni–Ti–Cu plate specimen on the Instron 5582 machine.

Table 1 Dimensions of Ni–Ti and Ni–Ti–Cu plates used in force generation tests.

Distance, 10-3m

b (mm)

e (mm)

d (mm)

y (mm)

Ti–Ni Ti–Ni–Cu

23.8 24.2

5.2 5.2

1.0 1.3

18.6 18.6

1.0 1.0

as the effect of a few number of thermal cycles on the reversible phase transformation. The results of force generation tests can be observed in Fig. 7 for Ni–Ti (Fig. 7a) and Ni–Ti–Cu SMA (Fig. 7b), respectively.

4. Discussions

0 280 260 240 220 200 180 160 140 120 100 80

2

3

4

5

6

7

8

9

10 2,75

1 Ti-Ni 2,50 2 Ti-Ni-Cu 2,25

2 1

2,00 1,75 1,50 1,25 1,00

0

Comparing Fig. 3a and b it can be noted that copper addition in substitution to nickel causes a change in the microstructure morphology of Ni–Ti SMA at room temperature (about 300 K  27 °C). It can be verified the absence of martensite variants in both SMA, Ni–Ti and Ni–Ti–Cu. This is expected because martensite start and finish transformation temperatures (Ms and Mf), in both SMA, are below room temperature (Table 2). On the other hand, there is no general literature information about Ni–Ti–Cu SMA, especially with less than 5 at.% Cu. Most results shown relate to the phase transformation and thermomechanical behavior for Ti50–Ni50 x–Cux (at.%) (x = 5, 7.5, 10, 12.5, 15 and 20) [12]. For alloys with contents between 0.5 and 5.0 at.% Cu there are no results published about changes in microstructural morphology.

1

Hardness, GPa

L (mm)

Hardness, HV

Plate sample

1

2

3

4

5

6

7

8

9

10

Distance, mm Fig. 4. Hardness profiles measured for Ni–Ti and Ni–Ti–Cu SMA.

Fig. 5a shows the detection of the R-phase in the DSC curves even with the Ni–Ti equiatomic SMA produced by PSPP technique that involves no cold working followed by heat treatment, which are typical conditions for the occurrence of R-phase [1,10,13]. Then, this phase appears probably due to internal stress field around precipitates [1,10]. Even though analysis of SEM have not been performed in this study, from the literature it is known that substitution of Cu prevent the X-phase (Ti3Ni4) precipitation [12].

Fig. 3. Microstructures of the studied SMA. (a) Ni–Ti and (b) Ni–Ti–Cu.

4928

C.J. de Araújo et al. / Materials and Design 32 (2011) 4925–4930

Temperature, K

Temperature, K 2,5

252

270

306

MP= 294,1K

2,0

324

342

240

285 300 315 330 345 360 6 5 4 cooling 3 2 MS= 287,9K Mf= 277,2K 1 0 -1 Af= 306,8K AS= 290,3K -2 -3 heating -4 -5 -6 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90

360

Rf= 300,4K

Ms= 297,9K

1,5

Heat Flow, mW

288

Mf= 277,7K

1,0

Rs= 308,4K Cooling

0,5 0,0 -0,5 -1,0

As= 327,8K

heating

Af= 342,6K

-1,5 -2,0

A = 337,5K

P -2,5 -30 -20 -10 0 10 20 30 40 50 60 70 80 90

255

270

Temperature, °C

Temperature, °C

(b)

(a)

Fig. 5. DSC analysis showing phase transformation temperatures of the studied SMA. (a) Ni–Ti and (b) Ni–Ti–Cu.

induced by copper addition. In fact, the lower level of hardness in Ni–Ti–Cu is expected since it is known that the addition of 5 at.% Cu makes the material more ductile by reducing the stress to induce or reorient variants of martensite [12]. The values of phase transformation temperatures of SMA essentially depend on a severe control of its chemical composition as well as of the mechanical forming manufacture processes. For instance, in the literature it is found, for equiatomic Ni–Ti or nearequiatomic Ni–Ti SMA, values of Ms that can vary from 246 K ( 27 °C) to 327 K (54 °C) [1,13]. For Ni–Ti–Cu alloys Ms values ranging from 282 K (9 °C) to 319 K (46 °C) have been found [14,15]. Analyzing Fig. 5a and b, and Table 2, it is noted that the phase transformation temperatures obtained for the studied Ni– Ti–Cu SMA are in agreement with this literature. In Fig. 5a it can also be noted the appearance of the intermediate phase (R-phase) which occurs in equiatomic Ni–Ti SMA as a first peak together with the one of the austenite–martensite transformation during cooling. One of the effects of copper addition in substitution for nickel is the elimination of this R-phase intermediate transformation, as can be verified in Fig. 5b. Table 2 shows that the thermal hysteresis (HT) of Ni–Ti–Cu SMA is practically one half of that one observed for Ni–Ti. This reduction of HT by copper addition occurs due to the elimination of the intermediate transformation during cooling (austenite – R-phase) [1,10]. Small thermal hysteresis is an important feature of Ni–Ti– Cu SMA, making them good candidates for applications where fast

Table 2 Phase transformation temperatures and thermal hysteresis of Ni–Ti and Ni–Ti–Cu SMA. Ni–Ti SMA

Ni–Ti–Cu SMA

(°C)

(K)

Phase transformation temperatures As 54.8 327.8 Af 69.6 342.6 Ms 24.9 297.9 Mf 4.7 277.7 Rs 35.4 308.4 Rf 27.4 300.4 HT 43.4 43.4

(°C)

(K)

17.3 33.8 14.9 4.2 – – 21.1

290.3 306.8 287.9 277.2 – – 21.1

Regarding the DSC results in Fig. 5, it can be concluded that the morphologies observed in Fig. 3 corresponds almost fully to the R-phase in Ni–Ti (Rs = 35.4 °C and Rf = 27.4 °C) and austenitic phase in Ni–Ti–Cu alloy, since As = 17.3 °C and Af = 33.8 °C. Analyzing Fig. 4 it can be observed that Ni–Ti SMA (profile 1) presented higher hardness values in comparison with Ni–Ti–Cu. The average hardness measured for Ni–Ti was 2.27 ± 0.17 GPa (231.08 ± 17.23 HV), whereas Ni–Ti–Cu SMA showed a lower average hardness, of the order of 1.96 ± 0.12 GPa (199.46 ± 12.53 HV). This reduction of hardness for the Ni–Ti–Cu in relation to Ni–Ti alloy can be associated to the change of microstructure morphology

Temperature, K

Electrical resistance, 10-3 Ohm

Temperature, K 252 270 288 306 324 342 360 378

2,00

1,48 1,44

252 270 288 306 324 342 360 378 1st Cycle 2nd Cycle 3rd Cycle 4th Cycle 5th Cycle

1,95 cooling

1,40

1,90

1,36 1,32

1st Cycle 2nd Cycle 3rd Cycle 4th Cycle 5th Cycle

1,28 1,24

heating

1,20

heating

1,85 1,80

cooling

1,75 -20

0

20

40

60

80

100

120

-20

0

20

40

60

80

Temperature, °C

Temperature, °C

(a)

(b)

100 120

Fig. 6. Electrical resistance as a function of temperature during five thermal cycles. (a) Ni–Ti and (b) Ni–Ti–Cu.

4929

C.J. de Araújo et al. / Materials and Design 32 (2011) 4925–4930

Temperature, K 216

243

270

297 324 351 378 405 225 D 200 175 150 125 100 75 heating 50 B C 25 A 0 cooling -25 -60 -40 -20 0 20 40 60 80 100 120 140 mechanical loading

216 243 270 297 324 351 378 405 225 200 175 150 125 B 100 load relaxation 75 50 C heating D 25 A 0 cooling -25 -60 -40 -20 0 20 40 60 80 100 120 140 mechanical loading

Force, N

Temperature, K

Temperature, °C

Temperature, °C

(a)

(b)

Fig. 7. Force–temperature behavior for the studied SMA. (a) Ni–Ti and (b) Ni–Ti–Cu.

activation of SME is required. In this way, Ni–Ti–Cu alloys are very interesting for applications as thermomechanical sensors and/or actuators [16]. According to Cai et al. [17], the HT can yet be reduced by suitable thermomechanical training. However, SMA with more large hysteresis, such as equiatomic Ni–Ti binary alloy, are the favorite candidates for coupling applications such as bolted joints assemblies and pipe-joints, where SMA device is expanded (deformed) at low temperature, installed in a mechanical system and then heated for generating force [3,4]. Copper addition to equiatomic Ni–Ti SMA also affects the electrical resistance versus temperature (R–T) behavior. Fig. 6a shows the R–T curves obtained for Ni–Ti SMA, while Fig. 6b shows the results for Ni–Ti–Cu. It is verified that copper is able to modify the shape of the R–T curve presented by equiatomic Ni–Ti SMA, taking the classic loop form (in ‘‘S’’), as can be seen in Fig. 6b. In Fig. 6a the effect of a few thermal cycles in R–T curves of equiatomic Ni–Ti SMA was clearly verified. It can be noted that, with the increase of the number of thermal cycles, the peak corresponding to the cooling curve (R-phase) becomes more pronounced. Increasing the number of thermal cycles (N) it is expected the Ms and Mf temperatures be reduced without occurring important changes on Rs, Rf, As and Af temperatures [1]. As can be seen in Fig. 6b, the thermal cycles did not yield significant modifications on R–T curves for Ni–Ti–Cu SMA, since the curves have practically remained superimposed. Fig. 7a shows the result obtained during the force generation test for Ni–Ti SMA with the mechanical system of Fig. 2a. The point A in Fig. 7a indicates the cooling start of the plate specimen, which was deformed of a certain central deflection (y, Table 1) when the temperature has reached the value of 233 K ( 40 °C), lower than the Mf temperature. The force increases as the plate is deformed, achieving a value of approximately 115 N (point B). When the load application is stopped one notes a load relaxation up to attain a value close to 62 N (point C). The force recorded by the load cell increase again when the activation of SME is started (point D) during the specimen heating. Thus, the force generated by Ni–Ti SMA (FGTN) under three point bending loading achieved a value close to 100 N (162 N 62 N). Some studies of force generation, similar to the one in Fig. 7a, provide a force generation analysis for samples with more complex geometries than a plate. For example, preload devices of the washer kind, manufactured in another work with Ni–Ti SMA [18], demonstrate ability to generate significantly higher recovery forces, when deformed in compression and then heated, in relation to the force generated by a plate under three point bending loading. Antonios et al. [4] obtained recovery forces of the order of 1750 N for a Ni–Ti washer with 5.8 mm height, 9.6 mm inner

diameter and 10.4 mm outer diameter. As already mentioned, this force level is much higher than that one verified in Fig. 6a, since the device employed by Antonios et al. [4] is much stronger than the plate tested in this current work. Moreover, there is still the fact that the deformation is not homogeneous under three point bending loading, if compared to the compression loading. Fig. 7b presents the result of force generation test for Ni–Ti–Cu SMA carried out in the universal testing machine (Fig. 6b). The test for Ni–Ti–Cu was performed in analogous way to Ni–Ti SMA. Thus, from Fig. 6b, the force generated by Ni–Ti–Cu SMA (FGTNC), through SME, reaches a value close to 166.0 N (205.5 N 39.5 N). From the force generation results it is verified that FGTNC > FGTN (of the order of 66%). This result indicates a better ability to generate force, through SME, for Ni–Ti–Cu SMA when compared to the equiatomic Ni–Ti, even taking into account that the tests have been realized in distinct experimental apparatus. 5. Conclusions In this work it was carried out a comparative study between the Ni–50Ti (at.%) and Ti–45Ni–5Cu (at.%) SMA, both obtained by Plasma Skull Push–Pull method (PSPP). It was verified that the copper addition to equiatomic Ni–Ti alloy, in substitution for nickel, besides modifying the microstructure morphology, reduces in about 14% the average hardness of the SMA, originally equiatomic. Another feature observed in the Ni–Ti–Cu alloy was the elimination of the intermediate R-phase transformation during cooling, followed by a reduction of the thermal hysteresis. Thus, it was confirmed that the Ni–Ti–Cu SMA are promising for applications as sensors and/or actuators where a faster activation frequency is important. On the other hand, the Ni–Ti SMA presented a larger thermal hysteresis compared to Ni–Ti–Cu alloy. This behavior of the equiatomic Ni–Ti SMA is important for applications in coupling devices. Besides that, it was demonstrated through electrical resistance measurements that the thermal cycling can modify significantly the phase transformation in Ni–Ti alloy, while it did not yield relevant changes on the Ni–Ti–Cu SMA. Finally, it was also verified a better capacity of generating force in Ni–Ti–Cu SMA when compared to the equiatomic Ni–Ti alloy, what enables it to act as actuator in mechanical systems. Acknowledgment The authors thank the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) Brazilian office for sponsoring the

4930

C.J. de Araújo et al. / Materials and Design 32 (2011) 4925–4930

research (Casadinho UFCG-ITA Grant No. 620091/2008-8 and INCT-EIE Grant No. 574001/2008) during the course of these investigations. References [1] Otsuka K, Wayman CM. Shape memory materials. Cambridge (UK): Cambridge University Press; 1998. [2] Nascimento MMSF, de Araújo CJ, Almeida LAL, Rocha Neto JS, Lima AMN. A mathematical model for the strain–temperature hysteresis of shape memory alloy actuators. Mater Des 2009;30:551–6. [3] Hesse T, Ghorashi M, Inman DJ. Shape memory alloy in tension and compression and its application as clamping-force actuator in a bolted joint: Part 1 – Experimentation. J Intell Mater Syst Struct 2004;15:577–87. [4] Antonios C, Inman DJ, Smaili A. Experimental and theoretical behavior of selfhealing bolted joints. J Intell Mater Syst Struct 2004;17:499–509. [5] de Araújo CJ, Rodrigues LFA, Coutinho Neto JF, Reis RPB. Fabrication and static characterization of carbon-fiber-reinforced polymers with embedded NiTi shape memory wire actuators. Smart Mater Struct 2008;17:065004. [6] Reis RPB, Souto CR, de Araújo CJ, Silva AA, da Silva EP. Vibration attenuation in an epoxy smart composite beam with embedded NiTi shape memory wires. Mater Sci Forum 2010;643:7–13. [7] Vilar ZT, de Araújo CJ, Reis RPB. Thermal activation of CFRP with embedded NiTi shape memory wires. Mater Sci Forum 2010;643:19–28. [8] da Silva NJ, Grassi END, de Araújo CJ. Dynamic properties of NiTi shape memory alloy and classic structural materials: a comparative analysis. Mater Sci Forum 2010;643:37–41.

[9] Machado LG, Savi MA. Medical applications of shape memory alloys. Braz J Med Biol Res 2002;36:683–91. [10] Otsuka K, Ren X. Physical metallurgy of Ti–Ni-based shape memory alloys. Prog Mater Sci 2005;50:511–678. [11] de Araújo CJ, Gomes AAC, Silva JA, Cavalcanti AJT, Reis RPB, Gonzalez CH. Fabrication of shape memory alloys using the plasma skull push–pull process. J Mater Process Technol 2009;209(7):3657–64. [12] Nam HT, Saburi T, Shimizu K. Cu-content dependence of shape memory characteristics in Ti–Ni–Cu alloys. Mater Trans JIM 1990;31(11):959–67. [13] Gonzalez CH, Oliveira CAN, Pina EAC, Urtiga Filho SL, Araújo Filho OO, de Araújo CJ. Heat treatments and thermomechanical cycling influences on the Rphase in Ti–Ni shape memory alloys. Mater Res 2010;13:325–31. [14] Zu XT, Wang ZG, Dai JY, Fub P, Feng XD. Effect of thermomechanical training temperature on the two-way shape memory effect of TiNi and TiNiCu shape memory alloys springs. Mater Lett 2003;57:1501–7. [15] Fu Y, Du H. RF magnetron sputtered TiNiCu shape memory alloy thin film. Mater Sci Eng A 2003;339(1–2):10–6. [16] Colombo S, Cannizzo C, Gariboldi F, Airoldi G. Electrical resistance and deformation during the stress-assisted two-way memory effect in Ni45Ti50Cu5 alloy. J Alloy Compd 2005;422:313–20. [17] Cai W, Men XL, Zhao LC. Recent development of TiNi-based shape memory alloys. Solid State Mater Sci 2005;9:296–302. [18] Silva NJ, Silva MM, de Araújo CJ, Santos MA. Fabrication and testing of actuators with shape memory alloys for screw connections of the oil and gas sector. In: Proceedings of rio oil & gas expo and conference, Rio de Janeiro, Brazil, p. 1–7; 2008 [in portuguese].