Dec 9, 2009 - Perfect pseudoelasticity was observed up to 405 K ... almost perfect shape memory effect induced by magnetic field, namely, metamagnetic ...
Materials Transactions, Vol. 51, No. 3 (2010) pp. 525 to 528 #2010 The Japan Institute of Metals
Shape Memory Response in Ni40 Co10 Mn33 Al17 Polycrystalline Alloy Wataru Ito1 , Burak Basaran2; * , Rie Y. Umetsu1 , Ibrahim Karaman2 , Ryosuke Kainuma1 and Kiyohito Ishida3 1
Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980-8577, Japan Department of Mechanical Engineering, Texas A&M University, College Station, TX 77843, USA 3 Department of Materials Science, Graduate School of Engineering, Tohoku University, Sendai 980-8579, Japan 2
Shape memory response of the polycrystalline Ni40 Co10 Mn33 Al17 alloy were investigated. In the isobaric thermal cycling experiments, the measured transformation strain levels increased with increasing stress reaching about 3.6% under 200 MPa. The slope of the stress vs. transformation temperature diagram, constructed using the data from these experiments, almost coincided with the predicted one obtained using the Clausius-Clapeyron equation, and experimental entropy data and transformation strain. Perfect pseudoelasticity was observed up to 405 K during compressive straining up to 2.5%. [doi:10.2320/matertrans.MBW200903] (Received October 1, 2009; Accepted October 20, 2009; Published December 9, 2009) Keywords: nickel-cobalt-manganese-aluminum, martensitic transformation, magnetic shape memory alloy, pseudoelasticity, metamagnetic transition
1.
Introduction
Since a large magnetic field-induced strain (MFIS) was reported for Ni2 MnGa single-crystalline alloy in 1996,1) ferromagnetic shape memory alloys (FSMAs) have received much attention as high-performance actuator materials. In 2004, Sutou et al. have discovered an unusual type of FSMAs in the Ni-Mn-X (X ¼ In, Sn and Sb) Heusler alloy systems, where the magnetization of the martensite (M) phase is considerably smaller than that of the parent (P) phase.2) Especially, the Co-doped Ni-Mn-In and -Sn quaternary alloys show a drastic change of magnetization by martensitic transformation from a ferromagnetic P phase to a very weakly magnetic M phase.3–5) The martensitic transformation temperatures of these alloys are drastically decreased by the application of a magnetic field, and magnetic field-induced reverse transformation (MFIRT) has been confirmed.3) An almost perfect shape memory effect induced by magnetic field, namely, metamagnetic shape memory effect, was found in the NiCoMnIn single crystalline and the NiCoMnSn polycrystalline alloys at room temperature.3,4) In addition, Wang et al.6) observed MFIRT in the Ni45 Co5 Mn36:6 In13:4 polycrystalline alloy under 50 MPa. Karaca et al.7) reported that a transformation strain of 5.4% was obtained in Ni45 Co5 Mn36:5 In13:5 single crystals, oriented along the [100] direction in the P phase, by thermal cycling under 125 MPa compressive stress. Karaca et al.8) have also demonstrated that actuation stress and work output levels in the NiCoMnIn alloys is at least one order magnitude higher than those in the conventional FSMAs. Furthermore, accompanying the martensitic transformation, giant magnetoresistance9–11) and large magnetocaloric effects12,13) have also been reported. Thus, these alloys show promise as new magnetic materials which are truly multifunctional. Relatively high cost of NiCoMnIn alloys due to expensive In and the brittleness of both NiCoMnIn and NiCoMnSn alloys in polycrystalline form are some of the major concerns for the insertion of these materials in practical applications. *Graduate
Student, Texas A&M University
To address this issue, Kainuma et al.14) have recently replaced In and Sn with Al and reported that the NiCoMnAl quaternary alloys show MFIRT from the paramagnetic M phase with L10 structure to the ferromagnetic P phase with B2 structure. To assess the full potential of these new alloys as both conventional and ferromagnetic SMAs with possible magnetic actuation applications, it is necessary to reveal their conventional shape memory characteristics such as transformation strain levels and stress vs. temperature phase diagram. Thus, in the present work, shape memory and pseudoelastic response of the Ni40 Co10 Mn33 Al17 metamagnetic shape memory alloy in polycrystalline form are reported for the first time. 2.
Experimental Procedures
A Ni40 Co10 Mn33 Al17 (at%) alloy specimen was prepared by induction melting under an argon atmosphere. The polycrystalline ingot was annealed at 1373 K for 168 hours in a vacuum and quenched in ice water. The quenched specimen was confirmed to have a single-phase structure with a large mean grain size of about 1 mm. The compression specimens with the dimensions of 3:0 2:5 5:5 mm3 was cut out of the annealed specimen using wire-electrical discharge machining. The thermo-mechanical experiments were conducted using an MTS servo hydraulic test frame. A capacitive displacement sensor was used to measure the strain. The heating and cooling of the specimen was achieved by conduction through the compression plates. The heating and cooling rate didn’t exceed 10 K min1 , and thus temperature variation along the sample could be kept below 2 K. The latent heat of the martensitic transformation was determined using the differential scanning calorimetry (DSC) measurements, where the heating and cooling rate was 10 K min1 . 3.
Results and Discussion
Figure 1 shows the strain vs. temperature response during thermal cycling under various compressive stresses through the phase transformation temperature interval. In these
526
W. Ito et al.
Fig. 3 Transformation strain "tr (a) and the residual strain "r (b) as a function of stress level for the Ni40 Co10 Mn33 Al17 polycrystalline specimen.
Fig. 1 Strain vs. temperature response of the Ni40 Co10 Mn33 Al17 polycrystalline alloy under various constant compressive stresses across the phase transformation temperature interval.
Fig. 2 Compressive stress vs. transformation temperatures phase diagram of the Ni40 Co10 Mn33 Al17 polycrystalline alloy. The values for each point were extracted from the experiments in Fig. 1.
experiments, the stress was applied at about 500 K in the P phase and the temperature was cycled between 283 and 500 K under constant stress levels. The experiments were started under 10 MPa and the applied stress was increased to the next level (25 MPa) after one thermal cycle, which was continued up to 200 MPa with 25 MPa stress increments. In Fig. 1(b), only some of the strain vs. temperature curves are shown for the sake of clarity. The drastic change in strain due to the transformation is observed at every stress level and the strain
Fig. 4 (a) Compressive stress-strain response of the Ni40 Co10 Mn33 Al17 polycrystalline specimen at 364, 386 and 405 K showing almost perfect pseudoelastic behaviour. (b) Temperature dependence of the critical stress c obtained from stress-strain curves in Fig. 4(a).
Shape Memory Response in Ni40 Co10 Mn33 Al17 Polycrystalline Alloy Table 1 Martensitic transformation temperatures, transformation strain and residual strain in the Ni40 Co10 Mn33 Al17 polycrystalline specimen as a function of external compressive stress. Compressive stress, /MPa
Ms /K
Mf /K
As /K
Af /K
"tr (%)
"r (%)
014Þ
351
319
349
372
—
—
10
361
320
349
381
0.27
0.008
25
361
315
353
388
0.41
0.005
50
369
316
361
396
0.80
0.011
75
371
324
370
399
1.53
0.028
100 125
373 375
330 334
376 382
404 409
2.29 2.95
0.025 0.039
150
380
347
385
418
3.31
0.067
175
389
356
391
424
3.53
0.111
200
398
361
397
428
3.66
0.174
change increases monotonically with increasing stress. In the experiments under high stress levels, it is apparent that there is a detectable residual strain after each thermal cycle. The martensitic transformation temperatures (Ms and Mf : the forward transformation start and finish temperatures, and As and Af : the reverse transformation start and finish temperatures) were determined for each stress level, as demonstrated in the curve for 200 MPa in Fig. 1(a). These transformation temperatures are plotted in Fig. 2 and listed in Table 1, where for the martensitic transformation temperatures under 0 MPa, the data acquired using a Superconducting Quantum Interference Device (SQUID) magnetometer, reported previously, are also presented.14) The plots for each transformation temperature follow almost a linear relation, although the slope of the line for the Ms is slightly smaller than the others. Similarly, the transformation strain "tr and the residual strain "r were defined at the temperature A ðAs þ Af Þ=2 using the baselines in the P and M phase regions as demonstrated in the 200 MPa curve of Fig. 1. The data evaluated for the "tr and "r are plotted in Figs. 3(a) and (b), respectively. The "tr increases with increasing stress and becomes almost constant at around 3.6% under the stress level of about 200 MPa. This strain level can be considered as the maximum "tr max obtained from this specimen. From the fact that the "tr max is not reached until the application of high stress levels, it is suggested that the martensite phase formed under low stresses is in multi-variant state, and with increasing applied stress, it should be increasingly more in the single variant state in each grain. Similarly, the "r barely appears at low stresses, however it starts to drastically increase above 150 MPa. This drastic increase suggests the introduction of plastic deformation in the stress region over 150 MPa. The slope of the T0 ¼ ðMs þ Af Þ=2 line obtained from the transformation temperature vs. compressive stress diagrams in Fig. 2 are determined to be 0.21 K MPa1 . This linear slope can be represented using the following ClausiusClapeyron equation assuming that the "tr max is independent of stress:15) @T0 Vm " S @
ð1Þ
527
where S and " ( "tr max ) are the transformation entropy change and the strain change along the corresponding direction during the transformation, respectively, and Vm is the molar volume of the specimen. For the present alloy, the @T0 =@ is calculated using eq. (1) as 0.22 K MPa1 , which is in good agreement with the experimental value. For this calculation, the S (¼ 1:21 J mol1 K1 ) is experimentally determined using S ¼ L=T where the latent heat L (¼ 463 J mol1 ) and the reverse transformation peak temperature T (¼ 383 K) are obtained from the DSC heating curve, and the experimentally determined " ("tr max ¼ 0:036) and Vm (¼ 7:42 106 m3 mol1 ) are used. Figure 4(a) shows the stress-strain curves for the NiCoMnAl polycrystalline specimen, where a compressive strain of about 2.5% was applied at 364, 386 and 405 K. At every temperature, an almost perfect pseudoelastic behaviour with full recovery of the applied strain is confirmed. The critical stress c for the stress-induced martensitic transformation was determined for each temperature as demonstrated in Fig. 4(a) and plotted in Fig. 4(b). The critical stress increases with increasing temperature, being consistent with the Ms line under the fixed stress levels shown in Fig. 2. The slope of the compressive stress vs. transformation temperature phase diagram is important in practice if the materials are to be used as magnetic actuators. As an example, if the desired actuation strain level and the temperature range of actuation are known, it would be possible to determine the magnetic field required for the onset and completion of the MFIRT under a particular stress level.8) In such a condition, the Clausius-Clapeyron equation16) is given by @c S : Vm " @T
ð2Þ
Using eq. (2), it is also possible to calculate the magnetostress level achievable in this material using the following relation:8) @c @c @T ; ¼ @H @T @H
ð3Þ
where H is the magnetic field applied to the specimen. Here, the @c =@T of 5.2 MPa K1 is directly determined from the slope of the compressive stress vs. transformation temperature diagram in Fig. 4(b). The @T=@H (¼ 3:6 K T1 ) can be estimated using the previously published data from the thermal cycling experiments in SQUID magnetometer under the magnetic fields from 0.05 to 7 T.14) From these values, the magnetostress @c =@H is estimated to be about 19 MPa T1 . This magnetostress level, although a rough estimate, is a direct indication of the achievable actuation stress per unit applied field in this material. Yet, it is significantly higher than what is reported for NiMnGa single crystalline FSMAs, i.e.
