Damping characteristics of a Ti
40.5
Ni
49.5
Zr
10
shape memory alloy
S.F. Hsieh
a, S.K. Wu
b,∗aDepartment of Mold and Die Engineering, National Kaohsiung University of Applied Science, Kaohsiung, Taiwan 807, Republic of China bDepartment of Materials Science and Engineering, National Taiwan University, Taipei, Taiwan 106, Republic of China
Received 12 April 2005; accepted 27 April 2005 Available online 14 July 2005
Abstract
Ti40.5Ni49.5Zr10alloy undergoes B2↔ B19martensitic transformation. Damping capacities of B19and B2 phases of this alloy are lower
than those of Ti51Ni49alloy due to Zr atoms solid-soluted hardening. Transformation temperatures of this alloy decrease, but transformation
peak heightsQ−1maxincrease with increasing aging time at 300◦C due to the formation of finer (0 0 1)Mtwins for specimens aged longer. The
Q−1
maxpeaks of the slightly cold-rolled Ti40.5Ni49.5Zr10alloy are higher than those of the as-annealed alloy, which may be because the thinner
twins are induced by small deformation. © 2005 Elsevier B.V. All rights reserved.
Keywords: Shape memory alloy; Damping capacity; Martensitic transformation; Aging; Cold-rolling
1. Introduction
Near-equiatomic TiNi shape memory alloys (SMAs) are technologically important due to their superior shape memory effect and superelastic properties. However, the applications of these alloys are limited to use at temperatures lower than 100◦C because their starting temperatures of martensitic transformation, Ms, are usually lower than 60◦C. High temperature SMAs with Ms temperatures higher than 100◦C have been exhaustively researched due to their many potential applications. In particular, the TiNiZr and TiNiHf ternary high temperature SMAs have been developed with high Ms temperatures[1–3].
In addition to shape memory effect and superelastic properties, TiNi SMAs have also been found to exhibit high mechanical damping capacity [4]. Damping mecha-nisms, in general, involve the stress-induced movement of defects. For high-damping metals, the major mechanisms are either stress-induced movement of dislocations or planar defects [5], and most of these mechanisms can be phenomenologically split into three classes: dynamic
∗Corresponding author. Tel.: +886 2 2363 7846; fax: +886 2 2363 4562.
E-mail address: [email protected] (S.K. Wu).
hysteresis, static hysteresis and transformation mechanisms. Dynamic hysteresis is produced by stress-aided ordering of defects overcoming local barriers by thermal activation and it yields damping that is frequency-dependent and amplitude-independent. Static hysteresis appears due to the stress-induced ‘unpinning’ or ‘break-away’ process of the defects [5–7] and it yields damping that is frequency-independent and amplitude-dependent. Some metals exhibit a high level of damping in the transformation region, for example, thermoelastic martensitic transformation of SMAs [8,9]. Such thermoelastic damping is frequently amplitude-independent and is proportional to the transformation rate. In the present study, the characteristics of internal friction (IF) in a Ti40.5Ni49.5Zr10SMA are investigated. In addition, the effect of aging and cold-rolling on the damping capacity of this alloy is also discussed.
2. Experimental procedure
The conventional tungsten arc melting technique was employed to prepare Ti40.5Ni49.5Zr10 (in at.%) alloy. Tita-nium (purity 99.7 wt.%), nickel (purity 99.9 wt.%) and zirco-nium (purity 99.8 wt.%), totaling about 120 g, were melted and remelted at least six times in an argon atmosphere. A 0925-8388/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
pure titanium button was used as a getter during the arc melting, and the mass loss during melting was negligibly small. The as-melted ingots were homogenized at 950◦C for 72 h and then quenched in water. They were swaged at 800◦C and then annealed at 850◦C×1 h. Specimens were cut into several plates with a low speed diamond saw, and then annealed at 900◦C for 1 h and finally quenched in water (i.e. as-annealing). After the annealing, two experimental pro-cedures were conducted. First, some plates were sealed in evacuated quartz tubes and aged at 300◦C for 1–240 h and then quenched in water. Second, some plates were cold-rolled at room temperature to 5% reduction in thickness. Speci-mens for differential scanning calorimetry (DSC) measure-ment, IF test, hardness test and microstructure observation were carefully cut from plates treated by the above pro-cedures. The specimen size for the IF test was 110 mm× 4 mm× 1 mm.
Martensitic transformation temperatures were measured using TA Q10 DSC equipment with controlled heating and cooling runs on samples encapsulated in an aluminum pan. The running temperature range was from−100 to +200◦C with the heating and cooling rate of 10◦C/min. The IF test was carried out with a SINKU-RIKO 1500M/L series inverted torsion pendulum in the temperature range from −150 to +250◦C. The tests started from −150◦C, heating up to +250◦C and then cooling down to −150◦C again. Cold-rolled specimens were heated again to +250◦C. The specimens’ heating and cooling rate was precisely controlled at 2◦C/min and the test frequency was set at approximately 1 Hz. The testing strain amplitude was kept at 4.5× 10−5. The recording of the data was completely automatic; calcu-lation and plots of IF Q−1and frequency f (shear modulus) versus temperature were performed on a digital computer. Thus, experimental results with a rather good resolution were obtained. Specimens for the hardness test were first mechanically polished and then subjected to measurement in a micro-Vickers hardness tester with a 500 g load at room temperature. For each specimen, the average hardness value was taken from at least five test readings. The microstructure observations were made by transmission electron microscope (TEM) with a JEOL-4000FX microscope equipped with a conventional double-tilting stage. The TEM specimens were prepared by electropolishing at−10◦C with an electrolyte consisting of 20% H2SO4and 80% CH3OH by volume. The applied voltage for electropolishing was 20 V.
3. Experimental results
3.1. Transformation behavior in Ti40.5Ni49.5Zr10alloy
Fig. 1(a and b) show the DSC results of the Ti40.5Ni49.5Zr10 alloy as-annealed and aged at 300◦C for 240 h, respectively, in both forward and reverse trans-formations. The M* and A*peaks appearing in Fig. 1are associated with the first-order martensitic transformation of
Fig. 1. DSC curve for the annealed Ti40.5Ni49.5Zr10alloy aging at 300◦C
for (a) 0 h; S.T. and (b) 240 h.
B2↔ B19[2]. Here, B2 is the parent phase and B19is the monoclinic martensite. InFig. 1, the transformations exhibit an asymmetric shape of the heat flow peak, i.e. sharp from the high-temperature side and smooth from the low-temperature one. This behavior can be explained by the fact that the stored elastic energy associated with martensitic transfor-mation increases with increasing the volume fraction of the martensite and retards the forward transformation. During the reverse process, the stored elastic energy is released in a reversible manner. Similar behavior is also observed in a Ni42.5Mn50Ti7.5SMA[10]. The temperature differenceT between A*and M*, as shown inFig. 1(a), is approximately 62◦C for as-annealed Ti40.5Ni49.5Zr10alloy, which is larger than that for as-annealed Ti51Ni49 alloy (T = 43◦C)[11]. This feature indicates that the Zr atoms solid-soluted in Ti-rich TiNi alloy will widen the transformation hysteresis. Fig. 2(a and b) show the plots of f and IF Q−1versus temper-ature of the as-annealed Ti40.5Ni49.5Zr10alloy, respectively. The full width half-maximum, W, is the temperature range at the half maximum IF peak, as shown inFig. 2(b). In this figure, there is one peak PC1at 84◦C on the first cooling and one peak PH1− 1 at 150◦C on the first heating. Both peaks corresponding to the f minimum, as shown in Fig. 2(a), indicating that they are the minima of shear modulus. The peak PR appearing at −37◦C that does not correspond to the minimum of f is a relaxation peak, which has been observed at −70◦C in the cold-rolled TiNi binary SMAs [12,13].
Fig. 2. (a) Frequency, f, and (b) internal friction,Q−1max, vs. temperature for
the annealed Ti40.5Ni49.5Zr10 alloy. Peaks PH1− 1 and PC1 are associated
with the martensitic transformation and peak PRis a relaxation one.
3.2. Effects of aging on Ti40.5Ni49.5Zr10alloy
Fig. 3(a and b) show the same plots asFig. 2(a and b), but now for the specimen aged at 300◦C for 240 h. InFig. 3(b), there is one peak PC1 on cooling and one peak PH1− 1 on heating in the first cycle. Peaks PC1and PH1− 1are associated with B2↔ B19 martensitic transformation. A well-shaped relaxation peak PRappears at−26◦C, as shown inFig. 3(b). The plots of f and Q−1 versus temperature for specimens aged at different time are similar to those ofFigs. 2 and 3, except that the transformation temperatures PC1, PH1− 1, PR and their magnitudes ofQ−1maxand W are different, and thus are omitted here.Fig. 4shows theQ−1maxof peaks PC1and PH1− 1 versus the aging time which are measured from the IF test. FromFig. 4, theQ−1maxof peaks PC1and PH1− 1increases with increasing aging time, although transformation temperatures of PC1and PH1− 1decrease with increasing aging time[14]. The transformation temperatures also shift to lower values in aged Ti32.3Ni50Hf17.7alloy than in the as-annealed one[15]. Fig. 5(a and b) show the TEM bright-field images of the martensite in specimen as-annealed and that aged at
Fig. 3. (a) Frequency, f, and (b) internal friction,Q−1max, vs. temperature for
the annealed Ti40.5Ni49.5Zr10alloy aging at 300◦C for 240 h. Peaks PH1− 1
and PC1are associated with the martensitic transformation and peak PRis a
relaxation one.
Fig. 4. The IF peak,Q−1max, of PH1− 1and PC1vs. aging time for annealed
Fig. 5. TEM bright-field image of annealed Ti40.5Ni49.5Zr10alloy aged at 300◦C for (a) 0 h (b) 240 h, SADP taken from areas E and B in (a) and (b), showing
(1 0 0)Mand (0 0 1)Mcompound twins with (c) [0 ¯1 ¯2]Mzone axis and (d) [1 1 0]Mzone axis, respectively. 300◦C ×240 h, respectively. Fig. 5(c) shows the selected
area diffraction pattern (SADP) taken from the marten-site plate with fine striations at area E of Fig. 5(a), in which the foil normal is parallel to [0 ¯1 ¯2]M direction. No extra reflection spots can be observed in Fig. 5(c) except the (1 0 0)M twins. Therefore, the fine striations are the traces of (1 0 0)M twin plates. Based on Han et al. [16], the [0 1 1]M type II twins and (1 0 0)M twins are the main substructures of the martensite in annealed Ti36.5Ni48.5Hf15 alloy and there are many defects such as dislocations and microtwins in twin plates. FromFig. 5, (0 0 1)M compound twins are found to exist in Ti40.5Ni49.5Zr10 alloy aged at 300◦C for 240 h, as shown inFig. 5(b) with SADP shown in Fig. 5(d). Compared withFig. 5(a and b), the abundant fine (1 0 0)M twin plates are observed in the as-annealed speci-men, and fine (0 0 1)Mtwin plates are observed in long aged specimens.
3.3. The effects of light cold-rolling on Ti40.5Ni49.5Zr10
alloy
Fig. 6(a and b) show the plots of f and Q−1 versus temperature for 5% cold-rolled Ti40.5Ni49.5Zr10 alloy, respectively, and peaks PH1− 1 (191◦C), PC1 (64◦C) and PH1− 2(133◦C) are all associated with B2↔B19martensitic transformation. Peak PRappearing at−40◦C that does not correspond to the f minima ofFig. 6(a) is a relaxation peak.
Mechanically induced martensite stabilization was reported in the cold-rolled Ti50Ni50alloy and the difference between PH1− 1and PH1− 2,PH, stands for the degree of martensite stabilization[17]. Similar behavior also occurs in this alloy. ThePHof Ti40.5Ni49.5Zr10alloy shown inFig. 6is 58◦C, which is larger than that of Ti50Ni50alloy (PH= 27◦C) for the same 5% cold-rolling[17]. The increment of the hardness for the 5% cold-rolled specimen isHv = 47 for the former, which is larger than that of the latter (Hv = 25). This feature comes from the fact that the as-annealed Ti40.5Ni49.5Zr10 alloy (Hv = 273) is harder than the as-annealed Ti50Ni50 alloy (Hv = 200) due to Zr atoms solid-soluted hardening in the former. We propose that the dislocation/defect movement associated with twin plate mobility may be impeded more in Ti40.5Ni49.5Zr10 alloy than in Ti50Ni50 alloy, which causes the harder Ti40.5Ni49.5Zr10alloy to have a higher martensite stabilization under the same degree of cold-rolling.
4. Discussion
4.1. Damping capacity of B2 and B19in Ti40.5Ni49.5Zr10alloy
It is well known that there are abundant twin bound-aries in B19 martensite and R phase premartensite of TiNi
Fig. 6. (a) Frequency, f, and (b) internal friction,Q−1max, vs. temperature for the 5% thickness-reduced Ti40.5Ni49.5Zr10alloy with specimen thickness
1.05 mm. Peaks PH1, and PC1and PH1− 2are associated with the martensitic transformation and peak PRis a relaxation one.
SMAs[18,19]. These twin boundaries can be easily moved by the external stress to accommodate the strain. This phe-nomenon is the well known “accommodation/reorientation process” occurring in the martensite and R phase of the deformed TiNi alloys. No twin boundaries exist in the B2 par-ent phase of TiNi alloys[20]. FromFig. 1, Ti40.5Ni49.5Zr10 alloy exhibits the thermoelastic B2↔ B19 transformation and the aforementioned behaviors can also occur in this alloy. Hence, in Ti40.5Ni49.5Zr10 alloy, the damping capacity of the B2 parent phase is suggested to come simply from the dynamic/static hysteresis of lattice defects, which is smaller than that of B19martensite because the dynamic/static hys-teresis loop generally dissipates less energy than the accom-modation/reorientation process of twin boundaries, as shown inFig. 2(b).
Lotkov et al.[21]and Mercier et al.[22]investigated the anomalies of elastic properties of TiNi binary and ternary SMAs. They reported that the lattice-softening phenomenon promotes the shear transformation due to the thermal or
mechanical driving forces and forms a minimum yield stress around the Ms temperature. This means that, during the martensitic transformation, the movement of twin bound-aries or martensite/parent interfaces is easy and most of the energy is dissipated in the transformation region. The damp-ing capacity of B19 martensite for ternary Ti50Ni49.5Fe0.5 and Ti50Ni40Cu10alloys is higher than that for Ti50Ni50alloy because the former two alloys have the lower yield stress and shear modulus[23]. Compared withFig. 2and Ref.[11], the f (shear modulus) of B19and B2 phases of Ti40.5Ni49.5Zr10 alloy is larger than that of Ti51Ni49alloy, where the IF peak (Q−1max= 3.12 × 10−2at PCl) of the former is lower than that of the latter (Q−1max= 5.2 × 10−2at PCl) with the same spec-imen size. This feature exhibits the Zr atoms solid-soluted hardening in Ti40.5Ni49.5Zr10alloy and decreasing the mobil-ity of twin plates in martensite and/or that of interfaces between martensite and parent phase. Therefore, damping capacities of B19 and B2 phases of Ti40.5Ni49.5Zr10 alloy are lower than those of Ti51Ni49alloy.
4.2. Internal friction in aged Ti40.5Ni49.5Zr10alloy
Delorme et al.[24]indicated that all the first-order phase transformations should be accompanied by IF peaks and they deduced the Q−1as a function of temperature changing rate dT/dt, as shown in Eq.(1): Q−1=ω1dψ(Vm) dVm dVm dT dT dt (1)
where Vmis the volume fraction of the martensite,ω the angu-lar frequency of the applied stress andΨ(Vm) is a monotonic function associated with the transformation change and/or shape strain. Eq.(1) indicates that the Q−1is proportional to the heating and cooling rate, dT/dt. Dejonghe et al.[25], in order to take account of the special character of a marten-site that can be induced or reoriented by an external stressσ, introduced the stress dependence to dVm/dt as follows: dVm dt = ∂Vm ∂T ∂T ∂t + ∂Vm ∂σ ∂σ ∂t (2)
In Eq.(2), the first-term is identical to the Delorme’s model and the second-term is stress-dependent. From Delorme’s model[24], supposing dΨ(Vm)/dVmremains constant for all thermoelastic martensites and keeping the angular frequency and the cooling or heating rate as constants in this study, the Q−1max should be proportional to the mount of marten-site formed per unit temperature or time (dVm/dT or dVm/dt). FromFigs. 2 and 3, we find thatQ−1maxvalues of martensitic transformations PC1and PH1− 1of Ti40.5Ni49.5Zr10specimen aged at 300◦C for 240 h are higher than those of the as-annealed one, but the W width of the former is smaller than the latter. This result shows that the aged Ti40.5Ni49.5Zr10 alloy can enhance the martensite formation per unit temperature or time. It is well known that the mechanical energy dissipa-tion of the martensite should depend on its twin type of TiNi SMAs. The most frequently observed twinning mode in TiNi
binary alloys is the0 1 1 type II twin[26]which is also their lattice invariant shear (LIS). However, it has been proved that the (0 0 1) compound twin is the LIS of Ti42.2Ni49.8Hf8and Ti36.5Ni48.5Hf15alloys[3,27]. FromFig. 5, the (0 0 1)Mtwin planes are more dense in the 300◦C×240 h aged specimen, but (1 0 0)M twin planes are more dense in the as-annealed one. These results indicate that the0 1 1 type II and (1 0 0)M twins can evolve to (0 0 1)Mtwins in the aged Ti40.5Ni49.5Zr10 alloy. Hence, the higher peakQ−1maxand the smaller width W of the longer aged Ti40.5Ni49.5Zr10specimen are attributable to stress-assisted twin-boundary motions of the abundant (0 0 1)M twins. The behaviors of these IF peaks associ-ated with martensitic transformation agree with Delorme’s model.
The asymmetric transformation peaks of IF test shown in Figs.2(b) and3(b) are similar to those of the DSC measure-ment shown inFig. 1(a and b). We propose that the IF peaks smoothing from low-temperature side are closely related to the stored elastic energy associated with martensitic trans-formation, which is the same reason for the asymmetric DSC peaks discussed in Section3.1.
4.3. Internal friction in cold-rolled Ti40.5Ni49.5Zr10alloy
Based on the vibration theory[28], the reduction in the specimen thickness will decrease the natural frequency of the specimen under the same amplitude when the specimen’s length and width are kept constant. Therefore, the frequency f should decrease when the specimen is slightly cold-rolled (5% thickness reduction), as compared with Figs.2(a) and 6(a). At the same time, the IF peakQ−1maxof the 5% cold-rolled specimen is higher than that of the as-annealed one. Lin et al.[17]reported that the light cold-rolling (≤5% thickness reduction) of Ti50Ni50 binary alloy increases the Q−1max of transformation peaks. Dalle et al.[3]reported that (0 0 1)M twinning in the 6.7% deformed Ti42.2Ni49.8Hf8specimen is thinner than that in the undeformed one, where the small deformation is accommodated by (0 0 1)Mtwinning and the further deformation is accommodated by the detwinning of these micro-twins. We suggest that the same phenomenon can also be found in the slightly cold-rolled Ti40.5Ni49.5Zr10 alloy. Hence, the higher Q−1max of the slightly cold-rolled Ti40.5Ni49.5Zr10 alloy shown in Fig. 6(b) than that of the as-annealed alloy shown in Fig. 2(b) is attributable to the thinner (0 0 1)Mtwins induced by small deformation, which are suggested to be easier to accommodate.
5. Conclusion
1. The as-annealed Ti40.5Ni49.5Zr10 alloy undergoes one-stage B2↔ B19 martensitic transformation. Damping capacities of B19 and B2 phases of this alloy are lower than those of Ti51Ni49alloy due to Zr atoms solid-soluted hardening.
2. With increasing aging time at 300◦C, the transformation peak heightQ−1maxincreases, but transformation tempera-ture decreases. These characteristics are closely related to the aging effects on the formation of finer (0 0 1)Mtwins for specimens aged longer.
3. The degree of martensite stabilization of Ti40.5Ni49.5Zr10 alloy is larger than that of Ti50Ni50 alloy for the same 5% cold-rolling because the former has a higher inher-ent hardness than the latter. The higher peakQ−1maxof the slightly deformed Ti40.5Ni49.5Zr10alloy compared to the as-annealed one is attributable to the thinner (0 0 1)Mtwins induced by small deformation.
Acknowledgements
The authors are pleased to acknowledge the financial sup-port of this research by the National Science Council (NSC), Republic of China under Grants NSC 93-2216-E002-003 and NSC 93-2216-E151-017.
References
[1] J.H. Mulder, J.H. Mass, J. Beyer, ICOMAT (1992) 869–874. [2] S.F. Hsieh, S.K. Wu, J. Alloys Compd. 266 (1998) 276–282. [3] F. Dalle, E. Perrin, P. Vermaut, M. Masse, R. Portier, Acta Mater.
50 (2002) 3557–3565.
[4] O. Mercier, K.N. Melton, Y. De Preville, Acta Metall. 27 (1979) 1467–1475.
[5] I.G. Ritchie, Z.-L. Pan, Metall. Trans. 22A (1991) 607–616. [6] R. De Batist, Internal Friction of Structural Defects in Crystalline
Solids, North-Holland Publishing Co., Amsterdam, 1972. [7] D.W. James, Mater. Sci. Eng. 4 (1969) 1–8.
[8] R. De Batist, J. Phys. (1983) C9–C39.
[9] J. Van Humbeck, Proceedings of The ASM Materials Week and TMS/AIME Fall Meeting, Role of Interfaces on Materials Damping, ASM, Materials Park, OH, 1985, pp. 5–24.
[10] P.L. Potapov, R. Gotthardt, L. Bataillard, Phys. Stat. Sol. 165 (1998) 401–408.
[11] H.C. Lin, S.K. Wu, J.C. Lin, Mater. Chem. Phys. 37 (1994) 184– 190.
[12] K. Iwasaki, R. Hasiguti, Trans. JIM 28 (1987) 363–367.
[13] H.C. Lin, S.K. Wu, T.S. Chou, J. Alloys Compd. 355 (2003) 90– 96.
[14] S.F. Hsieh, S.K. Wu, J. Alloys Compd. 270 (1998) 237–241. [15] R. Santamarta, C. Segui, J. Pons, E. Cesari, Scr. Mater. 41 (1999)
867–872.
[16] X.D. Han, W.H. Zou, R. Wang, Z. Zhang, D.Z. Yang, K.H. Wu, J. De Phys. IV (1995), C8-73-C758.
[17] H.C. Lin, S.K. Wu, T.S. Chou, H.P. Kao, Acta Metall. 39 (1991) 2069–2080.
[18] T. Onda, Y. Bando, T. Ohba, K. Otsuka, Mater. Trans., JIM 33 (1992) 354–359.
[19] S. Miyazaki, C.M. Wayman, Acta Metall. 36 (1988) 181– 192.
[20] H.C. Lin, S.K. Wu, M.T. Yeh, Metall. Trans. 24A (1993) 2189– 2194.
[21] A.I. Lotkov, A.V. Kuznetsov, V.N. Griskov, A.A. Botaki, in: Y. Chu, T.Y. Hsu, T. Ko (Eds.), Proceedings of International Shape Memory Alloy Symposium, China Academic Publishers, Guilin, China, 1986, pp. 153–158.
[22] O. Mercier, K.N. Melton, G. Gremaud, J. Hagi, J. Appl. Phys. 41 (1980) 1833–1834.
[23] H.C. Lin, S.K. Wu, Y.C. Chang, Metall. Trans. 26A (1995) 851–858. [24] J.F. Delorme, R. Schmid, M. Robin, P. Gobin, J. Phys. C2 (1971)
101–109.
[25] W. Dejonghe, R. de Batist, L. Delaey, Scr. Metall. 10 (1976) 1125–1128.
[26] M. Nishida, H. Ohgi, I. Itai, A. Chiba, K. Yamauchi, Acta Mater. 43 (1995) 1229–1234.
[27] X.D. Han, W.H. Zou, R. Wang, Z. Zhang, D.Z. Yang, Acta Mater. 44 (1996) 3711–3721.
[28] B.H. Tongue, Principles of Vibration, second ed., Oxford University Press, Oxford, 2002, pp. 17–18.
![Fig. 5. TEM bright-field image of annealed Ti 40.5 Ni 49.5 Zr 10 alloy aged at 300 ◦ C for (a) 0 h (b) 240 h, SADP taken from areas E and B in (a) and (b), showing (1 0 0) M and (0 0 1) M compound twins with (c) [0 ¯1 ¯2] M zone axis and (d) [1 1 0] M zone](https://thumb-ap.123doks.com/thumbv2/9libinfo/8874404.249625/4.1263.261.990.146.766/bright-field-image-annealed-sadp-taken-showing-compound.webp)
