Low-frequency damping properties of near-stoichiometric Ni2MnGa shape memory alloys under isothermal conditions

(1)

Low-frequency damping properties of near-stoichiometric

Ni

2

MnGa shape memory alloys under isothermal conditions

S.H. Chang

a

and S.K. Wu

b,*

a

Department of Chemical and Materials Engineering, National I-Lan University, I-Lan 260, Taiwan

b

Department of Materials Science and Engineering, National Taiwan University, 1 Roosevelt Road Sec. 4, Taipei 106, Taiwan Received 9 March 2008; revised 14 June 2008; accepted 7 July 2008

Available online 16 July 2008

The low-frequency damping properties of near-stoichiometric Ni2MnGa shape memory alloys were investigated by dynamic

mechanical analysis. Ni2MnGa alloys can possess good inherent internal friction under isothermal conditions over a wide

temper-ature range from100 to 100 °C without deterioration after thermal cycling. Ni2MnGa alloys with higher martensitic

transforma-tion temperatures are candidates for high damping applicatransforma-tions since they can possess good inherent internal frictransforma-tion above room temperature.

Keywords: Ni–Mn–Ga shape memory alloys; Martensitic phase transformation; Internal friction; Dynamic mechanical analyzer

Shape memory alloys (SMAs) exhibiting a ther-moelastic martensitic transformation can show unique properties including the shape memory effect (SME) and superelasticity[1]. Meanwhile, during the martens-itic transformation or in the martensite state, SMAs usually possess good internal friction (IF) and are po-tential candidates for high damping applications [2–5]. When heating and cooling SMAs, there is an IF peak with a storage modulus (E0) minimum appearing at the martensitic transformation temperature[3]. The IF peak of SMAs can be decomposed into three individual terms: IFTr, IFPT and IFI [6,7]. The first term, IFTr, is the transitory IF which appears only at low-frequency (t) and non-zero cooling/heating rate _T . The second term, IFPT, is the inherent IF corresponding to phase transformation, and is independent of _T . The third term, IFI, is the intrinsic IF of the austenitic or martensitic phase and depends strongly on microstructural proper-ties such as dislocations, vacancies and twin boundaries. In the low-frequency range, the IF peak observed during transformation is mainly ascribed to the first term, IFTr. However, the damping capacity of IFTr usually de-creases instantly when _T is abruptly stopped and only IFPT+ IFI persists [8]. Therefore, it is more important

to consider the damping property of IFPT+ IFI since, for most engineering applications, these high-damping SMAs are used at a set temperature (generally around room temperature), instead of a constant _T . The damp-ing characteristics of IFPT+ IFI during martensitic transformation of Ti50Ni50 and Ti51Ni39Cu10 SMAs have been systematically studied by dynamic mechanical analysis (DMA)[8,9]. Both these TiNi-based SMAs ex-hibit good IFPT+ IFI(tan d > 0.02) during martensitic transformation under isothermal conditions. Near-stoi-chiometric Ni2MnGa SMAs also undergo a martensitic transformation during cooling/heating and show a fer-romagnetic transition near 100°C [10,11]. Chernenko et al. [12]indicated that the martensitic transformation temperature (Ms) of Ni–Mn–Ga SMAs is highly sensi-tive to their chemical composition. Wu and Yang [13]

showed that Ni–Mn–Ga SMAs can exhibit a wide range of Mstemperature from120 to 185 °C by controlling their chemical composition. Therefore, it can be ex-pected that Ni–Mn–Ga SMAs could possess high inher-ent IF at room temperature if their Mstemperatures are adjusted to near room temperature. In this study, three diﬀerent types of near-stoichiometric Ni2MnGa SMAs are investigated by DMA to examine their inherent IF under isothermal conditions. The damping characteris-tics of these Ni–Mn–Ga SMAs are also discussed and compared to those of the Ti50Ni50 and Ti51Ni39Cu10 SMAs examined in our previous studies.

doi:10.1016/j.scriptamat.2008.07.006

* Corresponding author. Tel.: +886 2 2363 7846; fax: +886 2 2363

4562; e-mail:[email protected]

Available online at www.sciencedirect.com

Scripta Materialia 59 (2008) 1039–1042

(2)

Near-stoichiometric Ni2MnGa SMAs ingots used in this study were prepared by vacuum arc remelting from the raw materials of Ni (purity 99.9 wt.%), Mn55–Ni45 mother alloy (in wt.%) and Ga (purity 99.9 wt.%), fol-lowed by homogenizing at 850°C for 48 h. The homog-enized ingots were then cut using a low-speed diamond saw for electron probe microanalysis (EPMA) and DMA tests. The compositions of homogenized near-stoichiometric Ni2MnGa SMAs were determined by EPMA using a JEOL JXA-8600SX microscope [13]. Specimens for the DMA experiment were also cut from the homogenized ingots to dimensions of 20.0 3.3 2.8 mm3. Tan d and E0 values for each specimen were measured by TA 2980 DMA equipment using T_ ¼ 3_{C min}1_, _t_{= 1 Hz} _and _amplitude r0= 5 lm (strain amplitude = 1.1 104). All DMA measurements in this study were conducted without applying a magnetic ﬁeld. Table 1 lists the chemical compositions measured by EPMA for the three near-stoichiometric Ni2MnGa SMAs specimens used in this study. The inherent IF (IFPT+ IFI) of each specimen was also measured using TA 2980 DMA equipment un-der isothermal conditions. The detailed procedure for the isothermal DMA test was described in Ref.[8].

Figure 1a and b show the DMA E0and tan d curves, respectively, as a function of temperature for solution-treated Ni–Mn–Ga SMAs. Only the cooling curves are shown in detail. As illustrated in Figure 1a, the Curie point, Tc, for each specimen can be determined by the change in slope of the E0curve because there is a discon-tinuous behavior of E0 curve caused by the indirect interaction between magnetization and soft phonon po-tential at this temperature[14]. Meanwhile, as shown in

Figure 1b, transformation temperatures of each speci-men can be obtained from the IF peak temperature of the tan d curve. All the characteristic temperatures noted inFigure 1, such as the Curie point temperature Tc, tem-peratures for parent to intermediate phase (P ? I, TI), intermediate to martensite or parent to martensite (I ? M or P ? M, TM), are listed inTable 1. According

to the classiﬁcation of Ni–Mn–Ga SMAs reported by Chernenko et al. [12,14], curve 1 can be classiﬁed as Group I Ni–Mn–Ga SMA because its TM temperature is well below room temperature and TC. In addition, curve 2 can be categorized as Group II SMA since its TMtemperature is close to room temperature but below TC. Curve 3 can be labeled as Group III SMA due to its high TM temperature (above TC). As shown in Figure

1b, the IF peaks for Groups I and II SMAs exhibit a high tan d value above 0.11, while that for Group III SMA shows only a value lower than about 0.08. This feature arises from the fact that, during martensitic transformation, Group III SMA exhibits a higher E0

minimum (27,600 MPa) than Group I SMA

(21300 MPa) and Group II SMA (19,100 MPa) and does not show a conspicuous softening. In addition, in

Figure 1b, the IF peak of Group I SMA at TI tempera-ture only has a low tan d value (about 0.02) during cool-ing. This is due to the fact that the parent to intermediate phase transformation (P ? I) is only asso-ciated with a soft 1/3TA2phonon mode condensation, instead of a significant martensitic transformation (P ? M or I ? M) with structural change [14,15]. Ex-cept for the aforementioned transformation peaks, an extra broad peak can also be observed in the tan d curve of Group II SMA at about75 °C. This extra peak does not accompany a significant E0drop and is known as the relaxation peak[3]. The relaxation peak is also observed in IF results of Ni–Mn–Ga SMAs reported by Seguı´ et al. [16]. Recently, Fan et al. [17] revealed that the relaxation peak of TiNi SMAs originates from the inter-action between twin boundaries and hydrogen and is also affected by dislocations. The damping mechanism of the relaxation peak appearing in Ni–Mn–Ga SMAs may be similar to that of TiNi SMAs, though further study is needed to elucidate its characteristics.

Figure 2plots the tan d values vs. isothermal interval (0–30 min) of Groups I–III Ni–Mn–Ga SMAs kept iso-thermally at each transformation peak temperature (TI

Table 1. Chemical compositions (at.%), phase transformation (TI& TM) and Curie (TC) temperatures (°C) for Groups I - III Ni-Mn-Ga SMAs. TI

and TMare the transformation temperatures corresponding to parent to intermediate (P?I) and intermediate to martensite or parent to martensite

(I?M or P?M), respectively

Alloy at.% Ni at.% Mn at.% Ga TI(°C) TM(°C) TC(°C)

Group I 50.66 26.89 22.30 22.5 55.9 93.7 Group II 52.04 26.51 21.35 — 6.9 79.3 Group III 53.72 26.80 19.32 — 129.8 64.0 0 5 10 15 20 25 30 35 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Group I (P->I) Group I (I->M) Group II (P->M) Group III (P->M) Tan δ Time (min)

(IFPT+IFI)Group II (IFTr)Group II

Figure 2. Tan d values vs. isothermal time interval for Figure 1

specimens measured at t = 1 Hz, r0= 5 lm under isothermal

condi-tions around the transformation temperature. 93.7 ˚C 79.3˚C 64.0˚C 0 10000 20000 30000 40000 50000 60000

Storage Modulus (MPa)

-150 -100 -50 50 100 150 200

Temperature (˚C)Universal V4.3A TAInstruments

0.00 0.05 0.10 0.15 Tan Delta -150-100 -50 50 100 150 200

Temperature (˚C)Universal V4.3A TA Instruments

Group I Group II Group III Group I Group II Group III TC TC TC TI TM TM TM 1 2 3 1 2 3 0 0

a

b

Figure 1. (a) Storage modulus E0 and (b) tan d curves measured at

_

T¼ 3_{C min}1_{, t = 1 Hz and r}

0= 5 lm for Groups I–III Ni–Mn–Ga

SMAs.

(3)

or TM). As shown inFigure 2, tan d values of these Ni– Mn–Ga SMAs all decrease with prolonged isothermal time and reach a steady value after 30 min. As illustrated in Figure 2, the decayed tan d value during isothermal treatment represents the IFTr which is associated with the magnitude of _T . The steady tan d value after isother-mal treatment is the IFPT+ IFI during phase transfor-mation, which is independent of _T . Note that tan d values of IFTrfor Groups I and II SMAs collapse much faster than that of Group III, i.e. 95% of IFTr for Groups I and II SMAs diminishes within 12 min but this takes 20 min for Group III SMAs. In addition, as shown inFigure 2, the IFTrfor the P ? I peak at TIin Group I SMAs is insigniﬁcant compared to the IFTr of the P ? M or I ? M transformations. This phenomenon agrees with the fact that P ? I transformation does not correspond to the thermally induced martensitic transformation which mainly contributes to the IFTr peak in the tan d curve.

In order to investigate the inherent IF, IFPT+ IFI, of Ni–Mn–Ga SMAs, DMA tan d tests under 30 min iso-thermal treatment at diﬀerent temperatures for Groups I–III SMAs were conducted, and the results are indi-cated inFigure 3a–c, respectively. As shown inFigure 3, the solid lines represent the tan d cooling curve mea-sured at _T ¼ 3_{C min}1

(same as those shown in

Fig. 1b) and the empty circles symbolize tan d values of inherent IF measured by isothermal DMA tests. In

Figure 3a, Group I SMAs show a higher (IFPT+ I-FI)I?Mpeak with a tan d value of 0.027 at TM, and have a lower (IFPT+ IFI)P?Ipeak with a tan d value of 0.011 at TI. InFigure 3b and c, Groups II and III SMAs show a (IFPT+ IFI)P?Mpeak with tan d values of 0.029 and 0.016, respectively. As shown in Figure 3, there is a small temperature deviation between the IFPT+ IFI peak and the IF peak for each specimen. This tempera-ture shift is due to the cooling rate eﬀect, which has been described in detail elsewhere[18]. When each specimen is isothermally treated (i.e. _T ¼ 0) at around the

transfor-mation temperature, the tan d value of IFTrdisappears gradually and only IFPT+ IFIterm persists. This behav-ior is similar to that observed in Ti50Ni50 SMA [8]. Therefore, the damping capacity of IFPT+ IFI for Ni– Mn–Ga SMAs is proposed to be originated from stress-assisted martensitic transformation and stress-as-sisted motions of twin boundaries generated during martensitic transformation. Meanwhile, as illustrated inFigure 3, Group II SMAs show the highest IFPT+ I-FI peak (tan d = 0.030) compared with Group I (tan d= 0.026) and Group III (tan d = 0.016) SMAs. Re-cently, Bo¨hm et al. [19] revealed that Ni50Mn30Ga20 SMA can exhibit either 5 M or 7 M martensite after plastic deformation and heat treatment. In addition, it has been reported that the twin boundaries of 5 M mar-tensite possess higher mobility and can dissipate more energy during damping [14,20]. Therefore, fromFigure 3, we suggested that 5 M martensite is more dominant in Group II Ni–Mn–Ga SMAs after martensitic trans-formation. However, more detailed investigation on the relation between martensite structure and its damp-ing capacity is needed. Furthermore, fromFigure 3, we can calculate the individual contribution of IFTr and IFPT+ IFIto the overall IF peak. In this study, the con-tribution of IFPT+ IFIto overall IF for Group I, II and III SMAs is calculated as 23.5%, 23.5% and 20.5%, respectively. These values are comparable to those of Ti50Ni50 SMA (18.5% for B2 ? R and 22.3% for R ? B190 _{martensitic transformations)} _[8] _and solu-tion-treated Ti51Ni39Cu10 SMAs (17.2% for B2 ? B19 and 20.0% for B19 ? B190_{martensitic transformations)}

[9] under the same experimental parameters ð _T Þ ¼ 3_{C min}1

, t = 1 Hz and r0= 5 lm). However, these values are lower than those of the oﬀ-stoichiome-tric Ni2MnGa SMAs reported by Seguı´ et al., i.e. 40%

[16]. This deviation may be due to the insuﬃcient iso-thermal time interval (only 15 min) and the improper step-cooling method (stoping every other 5°C in cool-ing) used by Seguı´ et al.[16,21].

Figure 4a and b plot the inherent tan d and E0curves, respectively, for Groups I–III Ni–Mn–Ga SMAs mea-sured by isothermal DMA tests as well as those for Ti50Ni50[8]and Ti51Ni39Cu10[9]SMAs. The experimen-tal parameters used for each specimen inFigure 4are all the same (t = 1 Hz and r0= 5 lm). As shown inFigure

4a, Ti51Ni39Cu10 alloy has the highest IFPT+ IFI peak height (tan d > 0.035) during the B19 ? B190 martens-itic transformation (at about 10°C) at the ﬁrst round

-150 -100 -50 50 100 150 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Tan δ Temperature(ºC) fullcycle (3ºC/min) after 30min isothermal

-150 -100 -50 100 150 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Tan δ Temperature(ºC)

full cycle (3ºC/min) after 30 min isothermal

-150 -100 -50 0 50 100 150 200 250 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Tan δ Temperature(ºC)

full cycle (3ºC/min) after 30min isothermal

0

50 0

a b

c

Figure 3. Tan d values vs. temperature for (a) Group I, (b) Group II

and (c) Group III SMAs measured at t = 1 Hz, r0= 5 lm. The solid

curves are measured at _T¼ 3_{C min}1_{and the empty symbol curves}

are the tan d values of each specimen measured after 30 min isothermal treatment. -100 -50 0 50 100 50 20 0.00 0.01 0.02 0.03 0.04 0.05 Tan δ Temperature (˚C) Ti50Ni50 [8] Ti51Ni39Cu10 1 st round [9] Ti51Ni39Cu10 3 rd round [9] Ni2MnGa Group I Ni2MnGa Group II

Ni2MnGa Group III

-100 -50 0 50 100 150 200 20000 30000 40000 50000 60000 70000 Ti50Ni50 [8] Ti51Ni39Cu10 1st round [9] Ti51Ni39Cu10 3rd round [9] Ni2MnGa Group I Ni2MnGa Group II

Ni2MnGa Group III

Storage Modulus (MPa)

Temperature (˚C)

a

b

Figure 4. (a) Tan d and (b) storage modulus E0values of Groups I–III

Ni–Mn–Ga SMAs, cold-rolled Ti50Ni50SMA annealed at 650°C for

2 min [8], solution-treated Ti51Ni39Cu10 SMA [9] measured after

30 min isotherm treatment at diﬀerent temperatures.

(4)

of 30 min isothermal treatment, but this peak height de-creases to 0.023 at the third round. The decreasing tan d value of IFPT+ IFIis associated with the increasing de-fects or dislocations introduced by repeated thermal cy-cling during the isothermal treatment at diﬀerent temperatures. The introduced defects or dislocations im-pede the stress-assisted transformation and the mobility of twin boundaries in martensite and hence lower the damping capacity [9]. In addition, Ti50Ni50 SMA also possesses good inherent IF (tan d > 0.02) but only over a very narrow temperature range during the R ? B190 transformation (from 2 to 15°C). The reason is that R ? B190_{transformation exhibits larger transformation} strain and has more twin boundaries than the B2 ? R one and thus can dissipate more damping energy.

Compared with Ti50Ni50 SMA, as shown in Figure

4a, Groups I–III Ni–Mn–Ga SMAs also process good inherent damping capacities (tan d > 0.02) since they all exhibit low inherent E0values during transformation, as shown inFigure 4b. Moreover, Groups I–III Ni–Mn– Ga SMAs can exhibit good inherent IF in a wider tem-perature range (from 100 to 100 °C) than Ti50Ni50 SMA. This feature comes from the fact that Groups I– III Ni–Mn–Ga SMAs exhibit conspicuous damping capacities not only during martensitic transformation but also in the martensite state.Figure 5plots the cool-ing tan d curves for the same specimens of Groups I–III Ni–Mn–Ga SMAs measured before (fromFig. 1b) and after isothermal tests so as to investigate the thermal cy-cling eﬀect on their damping capacities. As shown in

Figure 5, tan d curves for Groups I–III SMAs before and after isothermal treatments are very similar, apart from a deviation in the relaxation peak of Group II SMAs. This phenomenon reﬂects that, after a number of thermal cycles, the damping capacity for Ni–Mn–Ga SMAs deteriorates less seriously than for Ti51Ni39Cu10 SMA. In conclusion, Groups II and III

Ni–Mn–Ga SMAs can exhibit and maintain a high damping capacity, comparable to that of Ti50Ni50 and Ti51Ni39Cu10SMAs, even after repeating thermal cycles. Nevertheless, the undesirable brittle characteristic of Ni–Mn–Ga SMAs critically limits their workability and need to be overcome before these can serve as high-damping materials.

The authors gratefully acknowledge the ﬁnancial support for this research provided by the National Sci-ence Council (NSC), Taiwan, Republic of China, under Grants No. NSC96-2216-E002-016.

[1] C.M. Wayman, T.W. Duerig, in: T.W. Duerig, K.N. Melton, D. Sto¨ckel, C.M. Wayman (Eds.), Engineering Aspects of Shape Memory Alloys, Butterworth-Heine-mann, London, 1990, pp. 3–20.

[2] O. Mercier, K.N. Melton, Y. De Pre´ville, Acta Metall. 27 (1979) 1467.

[3] S.K. Wu, H.C. Lin, T.S. Chou, Acta Metall. 38 (1990) 95. [4] K. Sugimoto, T. Mori, K. Otsuka, K. Shimizu, Scr.

Metall. 8 (1974) 1341.

[5] B. Coluzzi, A. Biscarini, R. Campanella, L. Trotta, G. Mazzolai, A. Tuissi, F.M. Mazzolai, Acta Mater. 47 (1999) 1965.

[6] W. Dejonghe, R. De Batist, L. Delaey, Scr. Metall. 10 (1976) 1125.

[7] J. Van Humbeek, J. Stoiber, L. Delaey, R. Gotthardt, Z. Metalkd, Zeitschrift Fur Metallkunde 86 (1995) 176. [8] S.H. Chang, S.K. Wu, Scr. Mater. 55 (2006) 311. [9] S.H. Chang, S.K. Wu, Mater. Trans. 48 (2007) 2143. [10] P.J. Webster, K.R.A. Ziebeck, S.L. Town, M.S. Peak,

Philos. Mag. 49 (1984) 295.

[11] V.V. Kokorin, V.A. Chernenko, Phys. Met. Metallogr. 68 (1989) 111.

[12] V.A. Chernenko, E. Cesari, V.V. Kokorin, I.N. Vitenko, Scr. Metall. Mater. 33 (1995) 1239.

[13] S.K. Wu, S.T. Yang, Mater. Lett. 57 (2003) 4291. [14] V.A. Chernenko, J. Pons, C. Seguı´, E. Cesari, Acta

Mater. 50 (2002) 53.

[15] E. Cesari, V.A. Chernenko, V.V. Kokorin, J. Pons, C. Seguı´, Acta Mater. 45 (1997) 999.

[16] C. Seguı´, E. Cesari, J. Pons, V. Chernenko, Mater. Sci. Eng. A 370 (2004) 481.

[17] G. Fan, K. Otsuka, X. Ren, F. Yin, Acta Mater. 56 (2008) 632.

[18] S.H. Chang, S.K. Wu, Mater. Charact. 59 (2008) 987. [19] A. Bo¨hm, S. Roth, G. Naumann, W.G. Drossel, R.

Neugebauer, Mater. Sci. Eng. A 481–482 (2008) 266. [20] I. Aaltio, M. Lahelin, O. So¨derberg, O. Heczko, B.

Lo¨fgren, Y. Ge, J. Seppa¨la¨, S.P. Hannula, Mater. Sci. Eng. A 481–482 (2008) 314.

[21] S.H. Chang, S.K. Wu, J. Alloys Compd. 459 (2008) 155.

-150 -100 -50 0 50 100 150 200 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20

Group I before isothermal-treatment Group I after isothermal-treatment Group II before isothermal-treatment Group II after isothermal-treatment Group III before isothermal-treatment Group III after isothermal-treatment

Tan

δ

Temperature (o_C)

Group I Group II

Group III

Figure 5. Tan d curves for Groups I–III Ni–Mn–Ga SMAs measured

before (fromFig. 1b) and after 30 min isothermal DMA tests.