- Email: [email protected]
Martensite transformation and superelasticity in polycrystalline Ni–Mn–Ga–Fe microwires prepared by melt-extraction technique
Materials Science & Engineering A 636 (2015) 157–163
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Martensite transformation and superelasticity in polycrystalline Ni–Mn–Ga–Fe microwires prepared by melt-extraction technique Yanfen Liu a,b, Xuexi Zhang a, Dawei Xing a, Hongxian Shen a, Mingfang Qian a, Jingshun Liu c, Dongming Chen a, Jianfei Sun a,n a
School of Materials Science and Engineering, Harbin Institute of Technology, Harbin 150001, China Department of Physics, Qiqihar University, Qiqihar 161006, China c School of Materials Science and Engineering, Inner Mongolia University of Technology, Hohhot 010051, China b
art ic l e i nf o
a b s t r a c t
Article history: Received 14 November 2014 Received in revised form 19 March 2015 Accepted 20 March 2015 Available online 31 March 2015
The effects of Fe doping on the microstructure, martensite transformation and superelasticity in meltextracted Ni50Mn25Ga25 xFex (x ¼1–6) microwires were investigated. The unique solidification process during melt-extraction creates the micron-sized diameter wires with small grains and semicircular cross-section. At ambient temperature Ni50Mn25Ga25 xFex (x o4) microwires are austenite phases with a cubic L21 structure, while microwires with x 45 are martensitic phases with seven-layered modulated (7M) structure. The results point out that martensite transformation temperatures are strongly related to Fe content due to the change of valence electron concentration (e/a). Reversible superelastic strains of 0.92% and 0.75% are obtained in Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires, respectively. It is demonstrated that the temperature dependence of stress-induced martensite (SIM) stress follows the Clausius–Clapeyron relation. The temperature dependence of SIM stress in Fe-doped Ni–Mn–Ga microwires is 10.5 MPa/K. & 2015 Elsevier B.V. All rights reserved.
Keywords: Ni–Mn–Ga–Fe microwires Melt-extraction Superelasticity Martensite transformation
1. Introduction The interest in shape memory alloys (SMAs) has increased in recent decade alternative to conventional actuators such as hydraulic, pneumatic and motor-based systems by taking advantage of their shape memory effect (SME), superelasticity (SE) and damping capacity. SMAs, such as Ni–Ti [1], Cu-based [2–4] and Febased [5] alloys, are thermo-responsive materials where deformation can be induced and recovered through temperature change. In contrast, ferromagnetic shape memory alloys (FSMAs) such as Ni– Mn–Ga may respond with a high frequency under an external magnetic field. As a result, FSMAs have attracted many attentions because they exhibited SME and SE as well as high magnetic-fieldinduced strain (MFIS) and magnetocaloric effect (MCE), which make them potential candidate materials for sensors, actuators or magnetic refrigerant (MR) [6–12]. However, the brittleness of polycrystalline Ni–Mn–Ga ternary alloys is a vital obstacle for their practical applications [13]. Stoichiometric Ni2MnGa alloy shows high Curie point ( 376 K), saturation magnetization ( 100 emu/g at 4 K) and magnetic moment
n
Corresponding author. Tel.: þ 86 451 8641 8317; fax: þ 86 451 8641 3904. E-mail addresses: [email protected], [email protected] (J. Sun).
http://dx.doi.org/10.1016/j.msea.2015.03.092 0921-5093/& 2015 Elsevier B.V. All rights reserved.
per formula ( 4.17 μB), but low martensite transformation temperature (Ms 202 K) [14]. The Ms and ductility of Ni2MnGa alloy may be improved by doping a fourth element or grain refinement [15–21]. Among various doping elements, Fe has attracted many attentions because transformation temperatures of Ni–Mn–Ga alloys are sensitive to Fe content [18,19]. The martensite transformation (MT) temperature decreased when Fe substituted Ni in Ni52.5 xMn23Ga24.5Fex alloys, while the temperatures changed little when Fe substituted Mn in Ni51.4Mn25.2 xGa23.5Fex alloys. On the other hand, the MT temperatures increased when Fe substituted Ga in Ni51.4Mn24.5Ga24.1 xFex alloys [18]. It was pointed out that the effect of Fedoping on MT of Ni–Mn–Ga alloys was related to the valence electron concentration (e/a) [18]. SE in SMAs may be obtained by stress-induced martensite (SIM) transformation when loading and the reverse transformation upon unloading [22]. SE in single crystalline Ni–Mn–Ga alloys has been reported in Refs. [22,23], but few studies have been devoted to the SE in brittle polycrystalline Ni–Mn–Ga alloys [10]. In this paper, polycrystalline Ni50Mn25Ga25 xFex (x¼ 1 6 at%) microwires (diameter 30– 40 μm, length 100 mm) were prepared by melt-extraction technique [24]. The relationship between Fe content and MT temperatures was established. Then Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires, which experienced MT around room temperature, were selected for SE investigation. The results demonstrated strong dependence of MT
158
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
Fig. 1. SEM images of the as-extracted polycrystalline Ni50Mn25Ga20Fe5 microwire. (a) Low magnification image, (b) cross-section, (c), cellular grains, (d) columnar grains in the planar part, and (e) schematic diagram of the crystal growth.
Fig. 2. TEM micrographs showing austenite and martensite structure of the Ni–Mn–Ga–Fe microwires. (a) Austenite phase in Ni50Mn25Ga21Fe4 microwires observed at room temperature. The inset shows diffraction pattern of the austenite phase, (b) martensite structure in Ni50Mn25Ga20Fe5 microwires. The inset shows diffraction pattern of 7M martensite phase.
temperature on Fe content and the characteristic of SE in polycrystalline Ni–Mn–Ga-Fe alloys.
2. Experimental details Ni–Mn–Ga–Fe ingots were prepared by arc melting of pure elemental materials Ni (99.99%), Mn (99.99%), Ga (99.99%), and Fe (99.99%) in argon atmosphere. The microwires with a nominal composition of Ni50Mn25Ga25-xFex (x¼ 1–6 at%) were prepared using a melt-extracted facility. Details of the processing have been reported in Ref. [24] and are briefly described here. The ingot was melted by an induction coil to form a melting pool. A rolling copper wheel (diameter 320 mm, knife edge 601) was used to extract the melt out of the pool. As a result, the microwires with diameters of 30–40 μm were obtained by rapid solidification during extracted melt. The linear velocities of the wheel flange and feed rate of the ingot were 30 and 3 10 5 m/s, respectively.
The morphology of the microwires was investigated in a fieldemission scanning electron microscopy (SEM-Helios Nanolab600i) at 20 kV. Thin-foil specimens for transmission electron microscope (TEM) observations were prepared using a precision ion-polishing system. TEM observations were performed in Tecnai G2 F30. The crystal structure of the as-extracted microwires was determined using X-ray diffraction (D/max-rb with Cu Kα radiation) at ambient temperature. Transformation temperatures of the wires were examined through a differential scanning calorimeter (TA DSC Q200), with heating and cooling rates of 5 K/min. The SE test temperatures of a single microwire were determined by temperature vs. internal friction curves obtained on a dynamic mechanical analyzer (DMA Q800), with oscillation frequency of 1 Hz, strain amplitude of 5 10 4 and heating/cooling rates of 5 K/min. For superelastic tests, the microwire was heated to 350 K (higher than Af) and held at this temperature for 12 min, then it was cooled to a test temperature (higher than Ms) and subjected to a tensile loading–unloading cycle at a rate of 0.04 N/min. After a tensile cycle, the microwire was heated to 350 K for 12 min, and then
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
159
cooled to a second test temperature for a second SE cycle. The cycles were carried out at 264, 267, 270, 273, 276, 279, 282, 285 K for a Ni50Mn25Ga21Fe4 microwire, and 301, 303, 306, 309, 311, 313, 315, 318 K for a Ni50Mn25Ga20Fe5 microwire, respectively.
3. Results and discussion 3.1. Microstructure of the microwires Melt-extraction is a high efficiency, cost effective technique to produce micron-diameter metallic wires. The morphology of the asextracted, naked Ni50Mn25Ga20Fe5 microwires is displayed in Fig. 1. As shown in Fig. 1a, the length of the as-extracted microwires ranges from 5 to 10 cm. Fig. 1b shows the fracture surface of a typical microwire. The microwire consists of plane and semicircular parts. The plane part (marked as B) was formed owing to contact of the melt on the copper wheel surface, while the semicircular part
Fig. 3. (a) X-ray diffraction patterns of Ni50Mn25Ga25 xFex (x ¼1–6) microwires and (b) unit cell of the Heusler-type alloy [30]. The atomic sites in the unit cell are presented by A, B, C and D.
Fig. 4. Martensite transformation of Ni–Mn–Ga–Fe microwires. (a) DSC curves; (b) evolution of transformation temperatures as a function of Fe content; and (c) electron concentration per atom e/a dependence of transformation temperature.
Table 1 Composition, lattice parameters and unit cell volume of Ni50Mn25Ga25 xFex (x ¼1–6) microwires. Fe content x (at%)
e/a Ratio
Composition (at%) Ni
Mn
Ga
Fe
1 2 3 4 5
49.28 50.25 50.50 50.2 49.9
25.07 25.34 24.82 25.11 25.0
24.23 22.01 21.47 20.37 19.60
1.42 2.41 3.20 4.32 5.50
7.5234 7.6519 7.6875 7.7344 7.768
6
49.92
25.19
18.36
6.52
7.8277
Phase structure
A A A A A M M
Unit cell volume (Å3)
Lattice parameter a (Å)
b (Å)
c (Å)
c/a
5.829 5.826 5.826 5.828 5.824 6.123 6.159
– – – –
– – – – – 5.546 5.507
– – – – – 0.9057 0.8941
5.804 5.755
198.0 197.7 197.7 198.0 197.6 197.1 195.2
160
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
(marked as A) was formed by free solidification of the melt. On the surface of part A, fine grains (diameter 1–3 μm) created by rapid solidification during melt-extraction (Fig. 1c). The grain size is close to that in Ni52.5Mn24.5Ga23 ribbons [25] (diameters 1–2 μm) which was formed by rapid solidification during melt spinning. In contrast, fan-shaped columnar grains with 3–5 μm width were formed on the surface of part B (Fig. 1d). It is deduced from the morphology of the columnar grains that nucleus of the melt firstly appeared on the center of the plane part when the copper wheel flange contacted the melt, then the nucleus grew with increasing contact area between melt and copper wheel until the extracted
Fig. 5. Temperature dependence of the internal friction (tan δ) and storage modulus of Ni50Mn25Ga20Fe5 microwires.
Table 2 Transformation temperatures of the Ni50Mn25Ga25 xFex (x¼ 4 and 5) microwires obtained from tan δ vs. temperature curves. Fe content (at%)
As (K) Af (K) AP (K) Ms (K) Mf (K) MP (K) Hystereses AP–MP (K)
4 5
274 286 312.6 323
285.5 263 318 300
247 289
259 295
26.5 23
melt filament left the wheel surface. Finally, other part of the melt filament solidified freely under the argon atmosphere, and formed the semicircular part A. The schematic diagram of the solidification process is illustrated in Fig. 1e. The austenite and martensite microstructures of the microwires were observed by TEM, as shown in Fig. 2. At room temperature, Ni50Mn25Ga21Fe4 microwires show a characteristic “tweed” contrast in the TEM bright field image (Fig. 2a), which may be interpreted in terms of pre-transitional softening and phonon branch in this microwire, just like previous reports such as Ni–Fe–Ga, Ni–Mn–Ga and Ni–Mn–Fe–Ga alloys [26–28]. The inset shown in Fig. 2a is an electron diffraction pattern of the selected area A. The pattern was well indexed by a body-centered cubic L21 structure, corresponding to an austenite phase. Fig. 2b shows the martensite variants in the Ni50Mn25Ga20Fe5 microwires. Two variants B and C with different orientations formed a selfaccommodated state. The electron diffraction pattern in Fig. 2b displays the six periodically distributed spots (marked by arrows) between every two main spots, showing that the wire has sevenlayered modulated (7M) martensite structure [27–29]. Fig. 3a is the XRD patterns of Ni50Mn25Ga25 xFex (x¼ 1–6) microwires tested at room temperature. The lattice parameters, unit cell volume and e/a ratio, are listed in Table 1. Ni50Mn25Ga25-xFex (xo4) microwires are austenite phases with cubic L21 structure, while microwires x45 are martensitic phases with seven-layered modulated (7M) structure [9,30]. It can be seen from Table 1 that the lattice parameter a of austenite phase has a weak relationship with x. In contrast, the lattice parameter a of the martensitic phase increases with x, while b and c decrease with x. The L21 Heusler structure is conveniently considered as four interpenetrating FCC sublattices with atoms A, B, C and D at locations (0,0,0), (1/4,1/4,1/4), (1/2,1/2,1/2) and (3/ 4,3/4,3/4), respectively, as shown in Fig. 3b [31]. In Ni50Mn25Ga25 xFex microwires (xo4), Ni atoms occupy A and C sites, while Ga atoms occupy D site. Mn and Fe atoms are expected to occupy the B site randomly. It is hard to confirm the arrangement mode of Mn and Fe atoms just from XRD measurements because the atomic scattering amplitudes of Mn and Fe atoms are very close. Furthermore, for Ni50Mn25Ga25 xFex microwires (x45), excess Mn and Fe atoms are considered to substitute the Ga atoms. It is also interesting to find that the unit cell volume is smaller in martensite phases which is because the atom radius (1.27 Å) of Fe and (1.32 Å) of Mn are smaller than Ga (1.40 Å). 3.2. Martensite transformation at various Fe contents
Fig. 6. Tensile stress–strain curves obtained at various temperatures in Ni50Mn25 Ga25-xFex microwires (a) x¼4 and (b) x¼ 5.
The relationship between MT temperatures and Fe content x is shown in Fig. 4. The DSC curves of Ni50Mn25Ga24Fe1 and Ni50Mn25Ga23Fe2 microwires were not displayed because the transformation temperatures were lower than the working temperature limit of the instrument (differential scanning calorimeter TA DSC Q200). The exothermic and endothermic peaks in the DSC curves in Fig. 4a correspond to the forward and reverse martensitic transformation, respectively. Fig. 4b displays transformation temperatures as a function of Fe content. The starting (Ms, As) and finishing (Mf, Af) temperatures are determined by the intersection of tangent of the maximum gradient point with the base line of the peak. The transformation temperatures have a linear relationship with Fe content x in the range of 1 6% (slope 30.6 K/at%). The increase of the transformation temperatures is related to increase of valence electron concentration (e/a) [18,21], which is shown in Fig. 4c. When Fe (3d64s2) replaces Ga (4s24p1), the e/a value increases, thus leading to an increase of transformation temperatures. This transformation temperature dependence on e/a (681 K/(e/a)) is similar to ternary non-stoichiometric Ni–Mn–Ga alloys [11], and may also be explained by the Hume–Rothery mechanism [32]. The effect of e/a on Ms mainly arises from the change of unit cell volume which modifies the relative position of the Brillouin zone boundary and Fermi
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
161
Fig. 7. Superelastic properties of Ni–Mn–Ga–Fe microwires. (a) A typical stress–strain curve, (b) total and residual strains, (c) superelastic strain, (d) recovery ratio, and (e) absorbed energy as a function of ΔT in Ni50Mn25Ga25 xFex microwires (x ¼4 and 5).
surface. The crystal structure of Heusler alloys is stabilized when the Fermi surface just touches the (110) Brillouin zone boundary [32]. When the Fermi surface overlaps the (110) Brillouin zone boundary at different e/a, the crystal structure of the Heusler alloys becomes unstable, leading to variations of the transformation temperatures. 3.3. Superelasticity of the microwire For superelastic test of a single microwire, the transformation temperatures of a single microwire were determined from its internal
friction tan δ vs. temperature curves, as shown in Fig. 5. During cooling process, the tan δ was 0.02 in its austenite state, and then increased rapidly to 0.06 when starting at 300 K. This is related to the transformation from austenite to martensite. After the phase transformation finished, the tan δ of the martensite phase was 0.04. It can be seen that the tan δ peak corresponds to a dip of the elastic modulus, indicating that lattice vibration softens during austenite to martensite transformation, which can be explained by the localized soft mode theory (LSMT) [33]. During heating process, tan δ drops from 0.04 in the martensite state to 0.02 in the austenite state
162
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
Fig. 8. SIM critical stress as a function of ΔT in Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires.
after the reverse martensite transformation. The obtained transformation temperatures of Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires are listed in Table 2. Fig. 6 shows stress–strain curves of Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires at temperatures higher than Ms. The plateau-like yield during loading and strain recovery upon unloading suggest the occurrence of stress-induced martensite formation (SIM), i.e. SE [23]. In both microwires, the strains corresponding to the strain plateau are small at temperatures much lower than As. In addition, only elastic deformation recovered at temperatures ranging from 264 to 273 K (Ni50Mn25Ga21Fe4) and from 301 to 306 K (Ni50Mn25Ga20Fe5), respectively. By contrast, the plastic strain recovery ratio increased when the temperature increased from 276 to 285 K (Ni50Mn25Ga21Fe4) and from 309 to 318 K (Ni50Mn25Ga20Fe5), respectively. It is noted that, after initiation of SIM, the critical stresses for SIM gradually increased with strain. This is probably related to the composition inhomogeneity, internal stress and defects formed during melt-extraction. As a result, the SIM process does not proceed synchronously. The most favorably oriented austenite transforms to martensite first, and then the remaining austenite transforms to martensite at higher stress levels. Fig. 7a shows a typical stress–strain curve of a Ni50Mn25Ga21Fe4 microwire tested at 276 K. The microwire was loaded to 350 MPa to reach a strain of εt , then it was unloaded to zero stress leaving a residual strain of εr . The recoverable strain consists of elastic strain εe and SE strain εSE (εSE ¼ εt εe εr ). The ΔT (ΔT¼ T Ms, T is deformation temperature) dependence of εt and εr are displayed in Fig. 7b. The maximum elongations of Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires are 3% and 2.1%, respectively. The strains achieved are close to polycrystalline Ni–Mn–Ga microwires [10] but lower than Ni–Mn–Ga single-crystalline alloys [23]. Furthermore, the residual strains εr decrease with increase in test temperature. Fig. 7c shows the relationship between εSE and applied strain ðεt εe Þ. εSE increases with strain until the maximum strains εmax SE (0.92% in Ni50Mn25Ga21Fe4 and 0.75% in Ni50Mn25Ga20Fe5) are reached. Similar behavior has also been found in Cu-based SMAs [4]. The maximum strain recovery ratios achieved in Ni50Mn25Ga20Fe5 and Ni50Mn25Ga21Fe4 microwires are 94% and 90%, respectively, as shown in Fig. 7d. Compared to Ni49.5Mn28.6Ga21.5 microwire [10], the Fe-doped Ni–Mn–Ga microwires reported here exhibit higher recovery ratio at low temperature, and comparable recovery ratio at high temperature. Fig. 7e reveals the absorbed energy as a function of temperature difference. The inset of Fig. 7e shows scheme of the absorbed energy in a superelastic cycle. For Ni50Mn25Ga21Fe4 microwires, the maximum absorbed energy reaches 3.2 MJ/m3 under an applied strain of 3%. The absorbed energy of Ni50Mn25Ga20Fe5 is
lower compared to Ni50Mn25Ga21Fe4. The energy absorption capacities in both microwires are comparable to that in polycrystalline Cu-based SMAs [4] during superelastic cycling. This implies that Fe-doped Ni–Mn–Ga alloys are candidate materials for applications where high damping capacity is needed [34]. Fig. 8 shows ΔT dependence of critical stresses σ in Ni–Mn–Ga– Fe microwires and some other SMAs. The relationship between σ and ΔT is linear and can be described by Clausius–Clapeyron relation. The temperature dependence of the superelastic stress in the Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires are determined to be 10.45 and 10.5 MPa/K, respectively. The values are much higher than those in conventional SMAs such as Cu–13.7Al– 5Ni ( 2.6–2.8 MPa/K) [3], Ti–50.3 at% Ni alloys ( 6.6 MPa/K) [35], Ni–Mn–Ga single crystals ( 2.8 MPa/K along [001] and 5.2 MPa/ K along [110]) [22] or polycrystalline Ni49.9Mn28.6Ga21.5 microwires ( 8.7 MPa/K) [10]. In short, polycrystalline Ni–Mn–Ga–Fe microwires exhibit a higher slope compared to the Ni–Mn–Ga single crystals, which may be related to fine grains (diameter 1–3 μm) created by rapid solidification during melt-extraction. Furthermore, Ni–Mn–Ga–Fe microwires have a higher slope than Ni–Mn– Ga microwires which may be attributed to the enhanced strength by solution strengthening of Fe.
4. Conclusions The effects of Fe on microstructure, martensite transformation and superelastic effect in polycrystalline Ni50Mn25Ga25 xFex microwires were investigated. The main conclusions can be drawn as follows: (1) The 1–6% Fe addition does not change the crystal structure of Ni–Mn–Ga alloy but reduces the unit cell volume. Martensitic transformation temperature increases with increase in Fe content. (2) Ni–Mn–Ga–Fe microwires show a 0.92% recoverable superelastic strain under tensile stress. The recoverable strain increases with increasing temperature. The maximum strain recovery ratio reaches 94% in Ni50Mn25Ga20Fe5 microwires. (3) The SIM stress increases linearly with temperature and is consistent with the Clausius–Clapeyron relation. The temperature dependence of superelastic stresses in Ni–Mn–Ga–Fe microwires is larger than Ni–Mn–Ga single-crystalline alloys and polycrystalline microwires.
Acknowledgment This work was financially supported by National Natural Science Foundation of China (NSFC) under Grant no. 51001038 and Ministry of Science and Technology Bureau of Harbin under Grant no. 2011RFQXG001. References [1] H.E. Karaca, I. Kaya, H. Tobe, B. Basaran, M. Nagasako, R. Kainuma, Y. Chumlyakov, Mater. Sci. Eng. A 580 (2013) 66–70. [2] Y. Chen, X.X. Zhang, D.C. Dunand, C.A. Schuh, Appl. Phys. Lett. 95 (2009) 171906. [3] Y. Chen, C.A. Schuh, Acta Mater. 59 (2011) 537–553. [4] Y. Sutou, T. Omori, R. Kainuma, N. Ono, K. Ishida, Metall. Mater. Trans. A 33 (2002) 2817–2824. [5] J. Ma, B. Kockar, A. Evirgen, I. Karaman, Z.P. Luo, Y.I. Chumlyakov, Acta Mater. 60 (2012) 2186–2195. [6] O. Heczko, L. Straka, J. Appl. Phys. 94 (2003) 7139–7143. [7] J. Pons, E. Cesari, C. Seguí, F. Masdeu, R. Santamarta, Mater. Sci. Eng. A 481–482 (2008) 57–65.
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
[8] M. Chmielus, X.X. Zhang, C. Witherspoon, D.C. Dunand, P. Müllner, Nat. Mater. 8 (2009) 863–866. [9] N.V.R. Rao, R. Gopalan, V. Chandrasekaran, K.G. Suresh, J. Alloy. Compd. 478 (2009) 59–62. [10] M.F. Qian, X.X. Zhang, C. Witherspoon, J.F. Sun, P. Müllner, J. Alloy. Compd. (2013) S296–S299. [11] V.A. Chernenko, Scr. Mater. 40 (1999) 523–527. [12] A. Sozinov, A.A. Likhachev, N. Lanska, K. Ullakko, V.K. Lindroos, J. Phys. IV 112 (2003) 955–958. [13] Y.Q. Ma, L.H. Xu, Y. Li, C.B. Jiang, H.B. Xu, Y.K. Lee, Z. Metallkd. 96 (2005) 843–846. [14] P.J. Webster, K.R.A. Ziebeck, S.L. Town, M.S. Peak, Philos. Mag. 49 (1984) 295–310. [15] V. Sánchez-Alarcos, J.I. Pérez-Landazábal, V. Recarte, C. Gómez-Polo, J.A. Rodríguez-Velamazán, Acta Mater. 56 (2008) 5370–5376. [16] Y.Y. Li, J.M. Wang, C.B. Jiang, Mater. Sci. Eng. A 528 (2011) 6907–6911. [17] I. Glavatskyy, N. Glavatska, A. Dobrinsky, J.U. Hoffmann, O. Söderberg, S.P. Hannula, Scr. Mater. 56 (2007) 565–568. [18] D. Soto, F.A. Hernández, H. Flores-Zúñiga, X. Moya, L. Mañosa, A. Planes, S. Aksoy, M. Acet, T. Krenke, Phys. Rev. B 77 (2008) 184103. [19] D.E. Soto-Parra, E. Vives, D. González-Alonso, L. Mañ osa, A. Planes, R. Romero, J.A. Matutes-Aquino, R.A. Ochoa-Gambo, H. Flores-Zúñiga, Appl. Phys. Lett. 96 (2010) 071912. [20] W. Cai, L. Gao, A.L. Liu, J.H. Sui, Z.Y. Gao, Scr. Mater. 57 (2007) 659–662.
163
[21] I. Glavatskyy, N. Glavatska, O. Söderberg, S.P. Hannula, J.U. Hoffmann, Scr. Mater. 54 (2006) 1891–1895. [22] V.A. Chernenko, V.A. L’vov, E. Cesari, J. Pons, A.A. Rudenko, H. Date, M. Matsumoto, T. Kanomata, Mater. Sci. Eng. A 378 (2004) 349–352. [23] V.A. Chernenko, V. L’vov, J. Pons, E. Cesari, J. Appl. Phys. 93 (2003) 2394. [24] H. Wang, D.W. Xing, X.D. Wang, J.F. Sun, Metall. Mater. Trans. A 42 (2010) 1103–1108. [25] A.K. Panda, Arvind Kumar, M. Ghosh, A. Mitra, J. Magn. Magn. Mater. 320 (2008) e730–733. [26] J.Q. Li, Z.H. Liu, H.C. Yu, M. Zhang, Y.Q. Zhou, G.H. Wu, Solid State Commun. 126 (2003) 323–327. [27] J.C. Bennett, C.V. Hyatt, M.A. Gharghouri, S. Farrell, M. Robertson, I. Chen, G. Pirge, Mater. Sci. Eng. A 378 (2004) 409–414. [28] W.H. Wang, X. Ren, G.H. Wu, Phys. Rev. B 73 (2006) 092101. [29] J. Pons, V.A. Chernenko, R. Santamarta, E. Cesari, Acta Mater. 48 (2000) 3027–3038. [30] H.B. Xu, Y. Li, C.B. Jiang, Mater. Sci. Eng. A 438–440 (2006) 1065–1070. [31] T. Kanomata, S. Murakami, D. Kikuchi, O. Nashima, H. Nishihara, Int. J. Appl. Electromagn. 27 (2008) 215–224. [32] J. Smit, J. Phys. F: Met. Phys. 8 (1978) 2139–2143. [33] G. Guénin, P.F. Gobin, Metall. Trans. A 13A (1982) 1127–1134. [34] G.N. Dayananda, M. Subba Rao, Mater. Sci. Eng. A 486 (2008) 96–103. [35] S. Miyazaki, A. Ishida, Mater. Sci. Eng. A 273–275 (1999) 106–133.
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Martensite transformation and superelasticity in polycrystalline Ni–Mn–Ga–Fe microwires prepared by melt-extraction technique Yanfen Liu a,b, Xuexi Zhang a, Dawei Xing a, Hongxian Shen a, Mingfang Qian a, Jingshun Liu c, Dongming Chen a, Jianfei Sun a,n a
School of Materials Science and Engineering, Harbin Institute of Technology, Harbin 150001, China Department of Physics, Qiqihar University, Qiqihar 161006, China c School of Materials Science and Engineering, Inner Mongolia University of Technology, Hohhot 010051, China b
art ic l e i nf o
a b s t r a c t
Article history: Received 14 November 2014 Received in revised form 19 March 2015 Accepted 20 March 2015 Available online 31 March 2015
The effects of Fe doping on the microstructure, martensite transformation and superelasticity in meltextracted Ni50Mn25Ga25 xFex (x ¼1–6) microwires were investigated. The unique solidification process during melt-extraction creates the micron-sized diameter wires with small grains and semicircular cross-section. At ambient temperature Ni50Mn25Ga25 xFex (x o4) microwires are austenite phases with a cubic L21 structure, while microwires with x 45 are martensitic phases with seven-layered modulated (7M) structure. The results point out that martensite transformation temperatures are strongly related to Fe content due to the change of valence electron concentration (e/a). Reversible superelastic strains of 0.92% and 0.75% are obtained in Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires, respectively. It is demonstrated that the temperature dependence of stress-induced martensite (SIM) stress follows the Clausius–Clapeyron relation. The temperature dependence of SIM stress in Fe-doped Ni–Mn–Ga microwires is 10.5 MPa/K. & 2015 Elsevier B.V. All rights reserved.
Keywords: Ni–Mn–Ga–Fe microwires Melt-extraction Superelasticity Martensite transformation
1. Introduction The interest in shape memory alloys (SMAs) has increased in recent decade alternative to conventional actuators such as hydraulic, pneumatic and motor-based systems by taking advantage of their shape memory effect (SME), superelasticity (SE) and damping capacity. SMAs, such as Ni–Ti [1], Cu-based [2–4] and Febased [5] alloys, are thermo-responsive materials where deformation can be induced and recovered through temperature change. In contrast, ferromagnetic shape memory alloys (FSMAs) such as Ni– Mn–Ga may respond with a high frequency under an external magnetic field. As a result, FSMAs have attracted many attentions because they exhibited SME and SE as well as high magnetic-fieldinduced strain (MFIS) and magnetocaloric effect (MCE), which make them potential candidate materials for sensors, actuators or magnetic refrigerant (MR) [6–12]. However, the brittleness of polycrystalline Ni–Mn–Ga ternary alloys is a vital obstacle for their practical applications [13]. Stoichiometric Ni2MnGa alloy shows high Curie point ( 376 K), saturation magnetization ( 100 emu/g at 4 K) and magnetic moment
n
Corresponding author. Tel.: þ 86 451 8641 8317; fax: þ 86 451 8641 3904. E-mail addresses: [email protected], [email protected] (J. Sun).
http://dx.doi.org/10.1016/j.msea.2015.03.092 0921-5093/& 2015 Elsevier B.V. All rights reserved.
per formula ( 4.17 μB), but low martensite transformation temperature (Ms 202 K) [14]. The Ms and ductility of Ni2MnGa alloy may be improved by doping a fourth element or grain refinement [15–21]. Among various doping elements, Fe has attracted many attentions because transformation temperatures of Ni–Mn–Ga alloys are sensitive to Fe content [18,19]. The martensite transformation (MT) temperature decreased when Fe substituted Ni in Ni52.5 xMn23Ga24.5Fex alloys, while the temperatures changed little when Fe substituted Mn in Ni51.4Mn25.2 xGa23.5Fex alloys. On the other hand, the MT temperatures increased when Fe substituted Ga in Ni51.4Mn24.5Ga24.1 xFex alloys [18]. It was pointed out that the effect of Fedoping on MT of Ni–Mn–Ga alloys was related to the valence electron concentration (e/a) [18]. SE in SMAs may be obtained by stress-induced martensite (SIM) transformation when loading and the reverse transformation upon unloading [22]. SE in single crystalline Ni–Mn–Ga alloys has been reported in Refs. [22,23], but few studies have been devoted to the SE in brittle polycrystalline Ni–Mn–Ga alloys [10]. In this paper, polycrystalline Ni50Mn25Ga25 xFex (x¼ 1 6 at%) microwires (diameter 30– 40 μm, length 100 mm) were prepared by melt-extraction technique [24]. The relationship between Fe content and MT temperatures was established. Then Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires, which experienced MT around room temperature, were selected for SE investigation. The results demonstrated strong dependence of MT
158
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
Fig. 1. SEM images of the as-extracted polycrystalline Ni50Mn25Ga20Fe5 microwire. (a) Low magnification image, (b) cross-section, (c), cellular grains, (d) columnar grains in the planar part, and (e) schematic diagram of the crystal growth.
Fig. 2. TEM micrographs showing austenite and martensite structure of the Ni–Mn–Ga–Fe microwires. (a) Austenite phase in Ni50Mn25Ga21Fe4 microwires observed at room temperature. The inset shows diffraction pattern of the austenite phase, (b) martensite structure in Ni50Mn25Ga20Fe5 microwires. The inset shows diffraction pattern of 7M martensite phase.
temperature on Fe content and the characteristic of SE in polycrystalline Ni–Mn–Ga-Fe alloys.
2. Experimental details Ni–Mn–Ga–Fe ingots were prepared by arc melting of pure elemental materials Ni (99.99%), Mn (99.99%), Ga (99.99%), and Fe (99.99%) in argon atmosphere. The microwires with a nominal composition of Ni50Mn25Ga25-xFex (x¼ 1–6 at%) were prepared using a melt-extracted facility. Details of the processing have been reported in Ref. [24] and are briefly described here. The ingot was melted by an induction coil to form a melting pool. A rolling copper wheel (diameter 320 mm, knife edge 601) was used to extract the melt out of the pool. As a result, the microwires with diameters of 30–40 μm were obtained by rapid solidification during extracted melt. The linear velocities of the wheel flange and feed rate of the ingot were 30 and 3 10 5 m/s, respectively.
The morphology of the microwires was investigated in a fieldemission scanning electron microscopy (SEM-Helios Nanolab600i) at 20 kV. Thin-foil specimens for transmission electron microscope (TEM) observations were prepared using a precision ion-polishing system. TEM observations were performed in Tecnai G2 F30. The crystal structure of the as-extracted microwires was determined using X-ray diffraction (D/max-rb with Cu Kα radiation) at ambient temperature. Transformation temperatures of the wires were examined through a differential scanning calorimeter (TA DSC Q200), with heating and cooling rates of 5 K/min. The SE test temperatures of a single microwire were determined by temperature vs. internal friction curves obtained on a dynamic mechanical analyzer (DMA Q800), with oscillation frequency of 1 Hz, strain amplitude of 5 10 4 and heating/cooling rates of 5 K/min. For superelastic tests, the microwire was heated to 350 K (higher than Af) and held at this temperature for 12 min, then it was cooled to a test temperature (higher than Ms) and subjected to a tensile loading–unloading cycle at a rate of 0.04 N/min. After a tensile cycle, the microwire was heated to 350 K for 12 min, and then
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
159
cooled to a second test temperature for a second SE cycle. The cycles were carried out at 264, 267, 270, 273, 276, 279, 282, 285 K for a Ni50Mn25Ga21Fe4 microwire, and 301, 303, 306, 309, 311, 313, 315, 318 K for a Ni50Mn25Ga20Fe5 microwire, respectively.
3. Results and discussion 3.1. Microstructure of the microwires Melt-extraction is a high efficiency, cost effective technique to produce micron-diameter metallic wires. The morphology of the asextracted, naked Ni50Mn25Ga20Fe5 microwires is displayed in Fig. 1. As shown in Fig. 1a, the length of the as-extracted microwires ranges from 5 to 10 cm. Fig. 1b shows the fracture surface of a typical microwire. The microwire consists of plane and semicircular parts. The plane part (marked as B) was formed owing to contact of the melt on the copper wheel surface, while the semicircular part
Fig. 3. (a) X-ray diffraction patterns of Ni50Mn25Ga25 xFex (x ¼1–6) microwires and (b) unit cell of the Heusler-type alloy [30]. The atomic sites in the unit cell are presented by A, B, C and D.
Fig. 4. Martensite transformation of Ni–Mn–Ga–Fe microwires. (a) DSC curves; (b) evolution of transformation temperatures as a function of Fe content; and (c) electron concentration per atom e/a dependence of transformation temperature.
Table 1 Composition, lattice parameters and unit cell volume of Ni50Mn25Ga25 xFex (x ¼1–6) microwires. Fe content x (at%)
e/a Ratio
Composition (at%) Ni
Mn
Ga
Fe
1 2 3 4 5
49.28 50.25 50.50 50.2 49.9
25.07 25.34 24.82 25.11 25.0
24.23 22.01 21.47 20.37 19.60
1.42 2.41 3.20 4.32 5.50
7.5234 7.6519 7.6875 7.7344 7.768
6
49.92
25.19
18.36
6.52
7.8277
Phase structure
A A A A A M M
Unit cell volume (Å3)
Lattice parameter a (Å)
b (Å)
c (Å)
c/a
5.829 5.826 5.826 5.828 5.824 6.123 6.159
– – – –
– – – – – 5.546 5.507
– – – – – 0.9057 0.8941
5.804 5.755
198.0 197.7 197.7 198.0 197.6 197.1 195.2
160
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
(marked as A) was formed by free solidification of the melt. On the surface of part A, fine grains (diameter 1–3 μm) created by rapid solidification during melt-extraction (Fig. 1c). The grain size is close to that in Ni52.5Mn24.5Ga23 ribbons [25] (diameters 1–2 μm) which was formed by rapid solidification during melt spinning. In contrast, fan-shaped columnar grains with 3–5 μm width were formed on the surface of part B (Fig. 1d). It is deduced from the morphology of the columnar grains that nucleus of the melt firstly appeared on the center of the plane part when the copper wheel flange contacted the melt, then the nucleus grew with increasing contact area between melt and copper wheel until the extracted
Fig. 5. Temperature dependence of the internal friction (tan δ) and storage modulus of Ni50Mn25Ga20Fe5 microwires.
Table 2 Transformation temperatures of the Ni50Mn25Ga25 xFex (x¼ 4 and 5) microwires obtained from tan δ vs. temperature curves. Fe content (at%)
As (K) Af (K) AP (K) Ms (K) Mf (K) MP (K) Hystereses AP–MP (K)
4 5
274 286 312.6 323
285.5 263 318 300
247 289
259 295
26.5 23
melt filament left the wheel surface. Finally, other part of the melt filament solidified freely under the argon atmosphere, and formed the semicircular part A. The schematic diagram of the solidification process is illustrated in Fig. 1e. The austenite and martensite microstructures of the microwires were observed by TEM, as shown in Fig. 2. At room temperature, Ni50Mn25Ga21Fe4 microwires show a characteristic “tweed” contrast in the TEM bright field image (Fig. 2a), which may be interpreted in terms of pre-transitional softening and phonon branch in this microwire, just like previous reports such as Ni–Fe–Ga, Ni–Mn–Ga and Ni–Mn–Fe–Ga alloys [26–28]. The inset shown in Fig. 2a is an electron diffraction pattern of the selected area A. The pattern was well indexed by a body-centered cubic L21 structure, corresponding to an austenite phase. Fig. 2b shows the martensite variants in the Ni50Mn25Ga20Fe5 microwires. Two variants B and C with different orientations formed a selfaccommodated state. The electron diffraction pattern in Fig. 2b displays the six periodically distributed spots (marked by arrows) between every two main spots, showing that the wire has sevenlayered modulated (7M) martensite structure [27–29]. Fig. 3a is the XRD patterns of Ni50Mn25Ga25 xFex (x¼ 1–6) microwires tested at room temperature. The lattice parameters, unit cell volume and e/a ratio, are listed in Table 1. Ni50Mn25Ga25-xFex (xo4) microwires are austenite phases with cubic L21 structure, while microwires x45 are martensitic phases with seven-layered modulated (7M) structure [9,30]. It can be seen from Table 1 that the lattice parameter a of austenite phase has a weak relationship with x. In contrast, the lattice parameter a of the martensitic phase increases with x, while b and c decrease with x. The L21 Heusler structure is conveniently considered as four interpenetrating FCC sublattices with atoms A, B, C and D at locations (0,0,0), (1/4,1/4,1/4), (1/2,1/2,1/2) and (3/ 4,3/4,3/4), respectively, as shown in Fig. 3b [31]. In Ni50Mn25Ga25 xFex microwires (xo4), Ni atoms occupy A and C sites, while Ga atoms occupy D site. Mn and Fe atoms are expected to occupy the B site randomly. It is hard to confirm the arrangement mode of Mn and Fe atoms just from XRD measurements because the atomic scattering amplitudes of Mn and Fe atoms are very close. Furthermore, for Ni50Mn25Ga25 xFex microwires (x45), excess Mn and Fe atoms are considered to substitute the Ga atoms. It is also interesting to find that the unit cell volume is smaller in martensite phases which is because the atom radius (1.27 Å) of Fe and (1.32 Å) of Mn are smaller than Ga (1.40 Å). 3.2. Martensite transformation at various Fe contents
Fig. 6. Tensile stress–strain curves obtained at various temperatures in Ni50Mn25 Ga25-xFex microwires (a) x¼4 and (b) x¼ 5.
The relationship between MT temperatures and Fe content x is shown in Fig. 4. The DSC curves of Ni50Mn25Ga24Fe1 and Ni50Mn25Ga23Fe2 microwires were not displayed because the transformation temperatures were lower than the working temperature limit of the instrument (differential scanning calorimeter TA DSC Q200). The exothermic and endothermic peaks in the DSC curves in Fig. 4a correspond to the forward and reverse martensitic transformation, respectively. Fig. 4b displays transformation temperatures as a function of Fe content. The starting (Ms, As) and finishing (Mf, Af) temperatures are determined by the intersection of tangent of the maximum gradient point with the base line of the peak. The transformation temperatures have a linear relationship with Fe content x in the range of 1 6% (slope 30.6 K/at%). The increase of the transformation temperatures is related to increase of valence electron concentration (e/a) [18,21], which is shown in Fig. 4c. When Fe (3d64s2) replaces Ga (4s24p1), the e/a value increases, thus leading to an increase of transformation temperatures. This transformation temperature dependence on e/a (681 K/(e/a)) is similar to ternary non-stoichiometric Ni–Mn–Ga alloys [11], and may also be explained by the Hume–Rothery mechanism [32]. The effect of e/a on Ms mainly arises from the change of unit cell volume which modifies the relative position of the Brillouin zone boundary and Fermi
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
161
Fig. 7. Superelastic properties of Ni–Mn–Ga–Fe microwires. (a) A typical stress–strain curve, (b) total and residual strains, (c) superelastic strain, (d) recovery ratio, and (e) absorbed energy as a function of ΔT in Ni50Mn25Ga25 xFex microwires (x ¼4 and 5).
surface. The crystal structure of Heusler alloys is stabilized when the Fermi surface just touches the (110) Brillouin zone boundary [32]. When the Fermi surface overlaps the (110) Brillouin zone boundary at different e/a, the crystal structure of the Heusler alloys becomes unstable, leading to variations of the transformation temperatures. 3.3. Superelasticity of the microwire For superelastic test of a single microwire, the transformation temperatures of a single microwire were determined from its internal
friction tan δ vs. temperature curves, as shown in Fig. 5. During cooling process, the tan δ was 0.02 in its austenite state, and then increased rapidly to 0.06 when starting at 300 K. This is related to the transformation from austenite to martensite. After the phase transformation finished, the tan δ of the martensite phase was 0.04. It can be seen that the tan δ peak corresponds to a dip of the elastic modulus, indicating that lattice vibration softens during austenite to martensite transformation, which can be explained by the localized soft mode theory (LSMT) [33]. During heating process, tan δ drops from 0.04 in the martensite state to 0.02 in the austenite state
162
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
Fig. 8. SIM critical stress as a function of ΔT in Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires.
after the reverse martensite transformation. The obtained transformation temperatures of Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires are listed in Table 2. Fig. 6 shows stress–strain curves of Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires at temperatures higher than Ms. The plateau-like yield during loading and strain recovery upon unloading suggest the occurrence of stress-induced martensite formation (SIM), i.e. SE [23]. In both microwires, the strains corresponding to the strain plateau are small at temperatures much lower than As. In addition, only elastic deformation recovered at temperatures ranging from 264 to 273 K (Ni50Mn25Ga21Fe4) and from 301 to 306 K (Ni50Mn25Ga20Fe5), respectively. By contrast, the plastic strain recovery ratio increased when the temperature increased from 276 to 285 K (Ni50Mn25Ga21Fe4) and from 309 to 318 K (Ni50Mn25Ga20Fe5), respectively. It is noted that, after initiation of SIM, the critical stresses for SIM gradually increased with strain. This is probably related to the composition inhomogeneity, internal stress and defects formed during melt-extraction. As a result, the SIM process does not proceed synchronously. The most favorably oriented austenite transforms to martensite first, and then the remaining austenite transforms to martensite at higher stress levels. Fig. 7a shows a typical stress–strain curve of a Ni50Mn25Ga21Fe4 microwire tested at 276 K. The microwire was loaded to 350 MPa to reach a strain of εt , then it was unloaded to zero stress leaving a residual strain of εr . The recoverable strain consists of elastic strain εe and SE strain εSE (εSE ¼ εt εe εr ). The ΔT (ΔT¼ T Ms, T is deformation temperature) dependence of εt and εr are displayed in Fig. 7b. The maximum elongations of Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires are 3% and 2.1%, respectively. The strains achieved are close to polycrystalline Ni–Mn–Ga microwires [10] but lower than Ni–Mn–Ga single-crystalline alloys [23]. Furthermore, the residual strains εr decrease with increase in test temperature. Fig. 7c shows the relationship between εSE and applied strain ðεt εe Þ. εSE increases with strain until the maximum strains εmax SE (0.92% in Ni50Mn25Ga21Fe4 and 0.75% in Ni50Mn25Ga20Fe5) are reached. Similar behavior has also been found in Cu-based SMAs [4]. The maximum strain recovery ratios achieved in Ni50Mn25Ga20Fe5 and Ni50Mn25Ga21Fe4 microwires are 94% and 90%, respectively, as shown in Fig. 7d. Compared to Ni49.5Mn28.6Ga21.5 microwire [10], the Fe-doped Ni–Mn–Ga microwires reported here exhibit higher recovery ratio at low temperature, and comparable recovery ratio at high temperature. Fig. 7e reveals the absorbed energy as a function of temperature difference. The inset of Fig. 7e shows scheme of the absorbed energy in a superelastic cycle. For Ni50Mn25Ga21Fe4 microwires, the maximum absorbed energy reaches 3.2 MJ/m3 under an applied strain of 3%. The absorbed energy of Ni50Mn25Ga20Fe5 is
lower compared to Ni50Mn25Ga21Fe4. The energy absorption capacities in both microwires are comparable to that in polycrystalline Cu-based SMAs [4] during superelastic cycling. This implies that Fe-doped Ni–Mn–Ga alloys are candidate materials for applications where high damping capacity is needed [34]. Fig. 8 shows ΔT dependence of critical stresses σ in Ni–Mn–Ga– Fe microwires and some other SMAs. The relationship between σ and ΔT is linear and can be described by Clausius–Clapeyron relation. The temperature dependence of the superelastic stress in the Ni50Mn25Ga21Fe4 and Ni50Mn25Ga20Fe5 microwires are determined to be 10.45 and 10.5 MPa/K, respectively. The values are much higher than those in conventional SMAs such as Cu–13.7Al– 5Ni ( 2.6–2.8 MPa/K) [3], Ti–50.3 at% Ni alloys ( 6.6 MPa/K) [35], Ni–Mn–Ga single crystals ( 2.8 MPa/K along [001] and 5.2 MPa/ K along [110]) [22] or polycrystalline Ni49.9Mn28.6Ga21.5 microwires ( 8.7 MPa/K) [10]. In short, polycrystalline Ni–Mn–Ga–Fe microwires exhibit a higher slope compared to the Ni–Mn–Ga single crystals, which may be related to fine grains (diameter 1–3 μm) created by rapid solidification during melt-extraction. Furthermore, Ni–Mn–Ga–Fe microwires have a higher slope than Ni–Mn– Ga microwires which may be attributed to the enhanced strength by solution strengthening of Fe.
4. Conclusions The effects of Fe on microstructure, martensite transformation and superelastic effect in polycrystalline Ni50Mn25Ga25 xFex microwires were investigated. The main conclusions can be drawn as follows: (1) The 1–6% Fe addition does not change the crystal structure of Ni–Mn–Ga alloy but reduces the unit cell volume. Martensitic transformation temperature increases with increase in Fe content. (2) Ni–Mn–Ga–Fe microwires show a 0.92% recoverable superelastic strain under tensile stress. The recoverable strain increases with increasing temperature. The maximum strain recovery ratio reaches 94% in Ni50Mn25Ga20Fe5 microwires. (3) The SIM stress increases linearly with temperature and is consistent with the Clausius–Clapeyron relation. The temperature dependence of superelastic stresses in Ni–Mn–Ga–Fe microwires is larger than Ni–Mn–Ga single-crystalline alloys and polycrystalline microwires.
Acknowledgment This work was financially supported by National Natural Science Foundation of China (NSFC) under Grant no. 51001038 and Ministry of Science and Technology Bureau of Harbin under Grant no. 2011RFQXG001. References [1] H.E. Karaca, I. Kaya, H. Tobe, B. Basaran, M. Nagasako, R. Kainuma, Y. Chumlyakov, Mater. Sci. Eng. A 580 (2013) 66–70. [2] Y. Chen, X.X. Zhang, D.C. Dunand, C.A. Schuh, Appl. Phys. Lett. 95 (2009) 171906. [3] Y. Chen, C.A. Schuh, Acta Mater. 59 (2011) 537–553. [4] Y. Sutou, T. Omori, R. Kainuma, N. Ono, K. Ishida, Metall. Mater. Trans. A 33 (2002) 2817–2824. [5] J. Ma, B. Kockar, A. Evirgen, I. Karaman, Z.P. Luo, Y.I. Chumlyakov, Acta Mater. 60 (2012) 2186–2195. [6] O. Heczko, L. Straka, J. Appl. Phys. 94 (2003) 7139–7143. [7] J. Pons, E. Cesari, C. Seguí, F. Masdeu, R. Santamarta, Mater. Sci. Eng. A 481–482 (2008) 57–65.
Y. Liu et al. / Materials Science & Engineering A 636 (2015) 157–163
[8] M. Chmielus, X.X. Zhang, C. Witherspoon, D.C. Dunand, P. Müllner, Nat. Mater. 8 (2009) 863–866. [9] N.V.R. Rao, R. Gopalan, V. Chandrasekaran, K.G. Suresh, J. Alloy. Compd. 478 (2009) 59–62. [10] M.F. Qian, X.X. Zhang, C. Witherspoon, J.F. Sun, P. Müllner, J. Alloy. Compd. (2013) S296–S299. [11] V.A. Chernenko, Scr. Mater. 40 (1999) 523–527. [12] A. Sozinov, A.A. Likhachev, N. Lanska, K. Ullakko, V.K. Lindroos, J. Phys. IV 112 (2003) 955–958. [13] Y.Q. Ma, L.H. Xu, Y. Li, C.B. Jiang, H.B. Xu, Y.K. Lee, Z. Metallkd. 96 (2005) 843–846. [14] P.J. Webster, K.R.A. Ziebeck, S.L. Town, M.S. Peak, Philos. Mag. 49 (1984) 295–310. [15] V. Sánchez-Alarcos, J.I. Pérez-Landazábal, V. Recarte, C. Gómez-Polo, J.A. Rodríguez-Velamazán, Acta Mater. 56 (2008) 5370–5376. [16] Y.Y. Li, J.M. Wang, C.B. Jiang, Mater. Sci. Eng. A 528 (2011) 6907–6911. [17] I. Glavatskyy, N. Glavatska, A. Dobrinsky, J.U. Hoffmann, O. Söderberg, S.P. Hannula, Scr. Mater. 56 (2007) 565–568. [18] D. Soto, F.A. Hernández, H. Flores-Zúñiga, X. Moya, L. Mañosa, A. Planes, S. Aksoy, M. Acet, T. Krenke, Phys. Rev. B 77 (2008) 184103. [19] D.E. Soto-Parra, E. Vives, D. González-Alonso, L. Mañ osa, A. Planes, R. Romero, J.A. Matutes-Aquino, R.A. Ochoa-Gambo, H. Flores-Zúñiga, Appl. Phys. Lett. 96 (2010) 071912. [20] W. Cai, L. Gao, A.L. Liu, J.H. Sui, Z.Y. Gao, Scr. Mater. 57 (2007) 659–662.
163
[21] I. Glavatskyy, N. Glavatska, O. Söderberg, S.P. Hannula, J.U. Hoffmann, Scr. Mater. 54 (2006) 1891–1895. [22] V.A. Chernenko, V.A. L’vov, E. Cesari, J. Pons, A.A. Rudenko, H. Date, M. Matsumoto, T. Kanomata, Mater. Sci. Eng. A 378 (2004) 349–352. [23] V.A. Chernenko, V. L’vov, J. Pons, E. Cesari, J. Appl. Phys. 93 (2003) 2394. [24] H. Wang, D.W. Xing, X.D. Wang, J.F. Sun, Metall. Mater. Trans. A 42 (2010) 1103–1108. [25] A.K. Panda, Arvind Kumar, M. Ghosh, A. Mitra, J. Magn. Magn. Mater. 320 (2008) e730–733. [26] J.Q. Li, Z.H. Liu, H.C. Yu, M. Zhang, Y.Q. Zhou, G.H. Wu, Solid State Commun. 126 (2003) 323–327. [27] J.C. Bennett, C.V. Hyatt, M.A. Gharghouri, S. Farrell, M. Robertson, I. Chen, G. Pirge, Mater. Sci. Eng. A 378 (2004) 409–414. [28] W.H. Wang, X. Ren, G.H. Wu, Phys. Rev. B 73 (2006) 092101. [29] J. Pons, V.A. Chernenko, R. Santamarta, E. Cesari, Acta Mater. 48 (2000) 3027–3038. [30] H.B. Xu, Y. Li, C.B. Jiang, Mater. Sci. Eng. A 438–440 (2006) 1065–1070. [31] T. Kanomata, S. Murakami, D. Kikuchi, O. Nashima, H. Nishihara, Int. J. Appl. Electromagn. 27 (2008) 215–224. [32] J. Smit, J. Phys. F: Met. Phys. 8 (1978) 2139–2143. [33] G. Guénin, P.F. Gobin, Metall. Trans. A 13A (1982) 1127–1134. [34] G.N. Dayananda, M. Subba Rao, Mater. Sci. Eng. A 486 (2008) 96–103. [35] S. Miyazaki, A. Ishida, Mater. Sci. Eng. A 273–275 (1999) 106–133.