Shape Memory Alloys - J-Stage

Materials Transactions, Vol. 57, No. 3 (2016) pp. 351 to 356

© 2016 The Japan Institute of Metals and Materials

Damping Characteristics of the Inherent and Intrinsic Internal Friction of Ti50Ni50¹xFex (x = 2, 3, and 4) Shape Memory Alloys Shih-Hang Chang1, Chen Chien2 and Shyi-Kaan Wu2,3,+ 1

Department of Chemical and Materials Engineering, National I-Lan University, I-Lan 260, Taiwan Department of Materials Science and Engineering, National Taiwan University, Taipei 106, Taiwan 3 Department of Mechanical Engineering, National Taiwan University, Taipei 106, Taiwan 2

Ti50Ni50¹xFex shape memory alloys (SMAs) have higher inherent and intrinsic internal friction peaks during R ¼ B19A transformation ((IFPT + IFI)R¼19A) than during B2 ¼ R transformation ((IFPT + IFI)B2¼R). The reasons are that the former has the larger transformation strain and the abundant movable twin boundaries appearing in both R-phase and transformed B19A martensite. Ti50Ni50¹xFex SMAs also exhibit higher (IFPT + IFI)R¼19A peaks than annealed Ti50Ni50 SMA after cold-rolling because the latter needs dislocations to be introduced to form R-phase, but the former does not. For Ti50Ni48Fe2 SMA, the tan ¤ values of the (IFPT + IFI)B2¼R and (IFPT + IFI)R¼B19A peaks and that of IFI in the Rphase regime decrease after the dehydrogenation treatment at 600°C for 6 h in a vacuum furnace, due to pinning of the twin boundaries by hydrogen atoms. However, the hydrogen pinning effect is small for IFI in B2 phase and B19A martensite. Ti50Ni50¹xFex SMAs with x = 2 and 3 at% have quite good (IFPT + IFI)R¼19A tan ¤ values (> 0.035), but their low transformation temperatures may restrict their use in practical highdamping applications. [doi:10.2320/matertrans.M2015425] (Received November 16, 2015; Accepted January 6, 2016; Published February 25, 2016) Keywords: shape memory alloy (SMA), martensitic transformation, internal friction, dynamic mechanical analysis

1.

Introduction

Shape memory alloys (SMAs) that undergo thermoelastic martensitic transformation show unique properties, such as the shape memory effect and superelasticity.1) Numerous studies have revealed that SMAs also exhibit high-damping properties during martensitic transformation and are effective for energy dissipation applications.2­5) SMAs normally exhibit an internal friction peak (IFTotal peak) at the martensitic transformation temperature, and the damping capacity is closely related to experimental parameters, such as the temperature change rate (cooling/heating rate), frequency, and applied strain amplitude.2,6) The IFTotal peaks of SMAs typically consists of three individual terms; that is, IFTotal = IFTr + IFPT + IFI.6­8) The first term, IFTr , is the transient internal friction, which appears only at a low frequency and non-zero temperature change rate. The second term, IFPT, is the inherent internal friction corresponding to phase transformation, which is independent of the temperature change rate. The third term, IFI, is the intrinsic internal friction of the austenitic or martensitic phase, and it depends strongly on microstructure properties such as dislocations, vacancies, and twin boundaries. The IFTotal peak observed during martensitic transformation is mainly ascribed to the IFTr, but when the alloy is kept isothermally at a set temperature during martensitic transformation, the damping capacity of IFTr typically decreases and only the peaks of IFPT + IFI remain.9) Thus, the damping capacities of IFPT and IFI are more important than that of IFTr because highdamping SMAs are generally used at a steady temperature rather than at a constant temperature change rate. TiNi-based alloys are the most important and practical SMAs, having excellent mechanical properties.1) Ti50Ni50 SMA typically undergoes a B2 § B19A transformation sequence and exhibit good damping capacity during +

Corresponding author, E-mail: [email protected]

martensitic transformation. TiNi-based SMAs can also undergo a B2 § R § B19A two-stage martensitic transformation by cold-working or aging of a Ni-rich TiNi SMA to introduce dislocations or precipitates.1) The formation of R-phase normally greatly softens the storage modulus of TiNi-based SMAs and improves the internal friction of the alloys. Chang and Wu9­13) have systematically studied the damping characteristics of the inherent and intrinsic internal friction (IFPT + IFI) of cold-rolled and annealed Ti50Ni50 SMA, which undergoes a B2 ¼ R ¼ B19A two-stage martensitic transformation during cooling, by applying a dynamic mechanical analyzer (DMA) under isothermal conditions. According to their results, Ti50Ni50 SMA exhibits IFPT + IFI peaks with an acceptable damping capacity during martensitic transformation, and it has IFI in the regimes of R-phase and B19A martensite. The damping capacity of (IFPT + IFI) corresponds to the stress-assisted movements of martensite interfaces and the stress-assisted motions of twin boundaries in transformed martensite.9) It has been reported that replacing some of the Ni in Ti50Ni50 SMA with Fe to form Ti50Ni50¹xFex SMAs can influence the transformation sequence.14,15) Similarly to coldrolled and annealed Ti50Ni50 SMA, Ti50Ni50¹xFex SMAs also exhibit B2 ¼ R ¼ B19A two-stage martensitic transformation during cooling but do not include defects or dislocations. Yoshida et al.16) investigated the damping characteristics of Ti50Ni47Fe3 SMA and reported that the alloy exhibited a good damping capacity during B2 ¼ R and R ¼ B19A martensitic transformations. In addition, they revealed that the martensitic transformation temperatures of Ti50Ni47Fe3 SMA are related to the solution-heat treatment temperature. Ren et al.17­19) reported that Ti50Ni48Fe2 SMA possesses ultrahigh damping for the relaxation-type peak in R-phase and suggested that preventing the introduction of dislocations and precipitates is essential for obtaining high damping. They also proved that the origin of the relaxation peaks obtained in Ti50Ni48Fe2 SMA can be attributed to the interaction between

352

S.-H. Chang, C. Chien and S.-K. Wu 0.3

twin boundaries and hydrogen atoms. Recently, Zuo et al.20) stated that hydrogen is an essential ingredient for the frequency-dependent internal friction peak at around 160 K in Ti50Ni44Fe6 SMA. They also proposed that the anelastic behavior exhibited in Ti50Ni44Fe6 SMA comes from the interaction between hydrogen and nanodomains. However, most of these studies focused only on the damping characteristics of IFTotal and IFTr internal friction of Ti50Ni50¹xFex SMAs, so IFPT and IFI internal friction have not been studied yet. Therefore, this study systematically investigated the damping properties of IFPT and IFI internal friction exhibited in Ti50Ni50¹xFex (x = 2, 3, and 4) SMAs measured under isothermal conditions. The hydrogen pinning effect on the (IFPT + IFI) of B2 ¼ R and R ¼ B19A transformations and that on the IFI in the R-phase regime of Ti50Ni48Fe2 SMA were also studied.

Heat Flow (W/g)

B19'←R

0.1

Mf

cooling

Ms Rf

Rs

0.0

heating

-0.1

-0.2 -200

R→B2

B19'→R

-150

-100

-50

0

50

100

150

Temperature (oC) 0.3

(b) Ti50Ni47Fe3 R←B2

Heat Flow (W/g)

0.2

Experimental Procedures

The Ti50Ni50¹xFex (x = 2, 3, and 4, in at%) SMAs used in this study were prepared using pure raw titanium (purity 99.99 mass%), nickel (purity 99.99 mass%), and iron (purity 99.98 mass%). The raw materials were re-melted using a conventional vacuum arc remelter to form Ti50Ni50¹xFex SMAs ingots. Each as-melted ingot was hot rolled at 900°C into plates of about 2 mm thickness by a rolling machine (DBR150x200 2HI-MILL, Daito Seiki Co, Japan) and then solution heat-treated at 900°C for 1 h, followed by quenching in water. The surface oxide layer of the plate was removed using an etching solution of HF : HNO3 : H2O = 1 : 5 : 20 in volume ratio. Each plate was then cut with a diamond saw into specimens with dimensions of 30.0 mm © 4.0 mm © 2.0 mm for DMA tests, and into segments of approximately 30 mg for differential scanning calorimetry (DSC) tests. The martensitic transformation temperature and transformation enthalpy (¦H) of each Ti50Ni50¹xFex SMA were determined using TA Q10 DSC under a constant cooling/ heating rate of 10°C/min. The tan ¤ and the storage modulus (E0) values of each Ti50Ni50¹xFex specimen were determined using TA 2980 DMA equipment fitted with a single cantilever clamp and a liquid nitrogen cooling apparatus. The experimental parameters used in DMA tests were set as follows: cooling rate = 3°C/min, frequency = 1 Hz, and applied strain = 8.0 © 10¹5. The (IFPT + IFI) values of Ti50Ni50¹xFex SMAs were investigated using DMA tests under isothermal conditions. The specimen was initially cooled from 150°C at a constant cooling rate of 3°C/min and then held isothermally for 30 min at the set temperature. Afterward, the specimen was heated to 150°C to ensure that it had returned to the B2 parent phase. Subsequently, the specimen was cooled to another set temperature and held isothermally for 30 min at that temperature. Ti50Ni48Fe2 SMA was selected to investigate the hydrogen pinning effect on the damping characteristics of the intrinsic internal friction (IFI) exhibited in the regimes of B2 phase, R-phase and B19A martensite. The DMA specimen of Ti50Ni48Fe2 SMA was loaded in a vacuum furnace, which was then evacuated to a high vacuum pressure of 2.6664 © 10¹6 Pa at room temperature using a turbo-pump. Thereafter,

R←B2

0.2

B19'←R

cooling

0.1

0.0 heating

-0.1 B19'→R

-0.2 -200

-150

-100

R→B2

-50

0

50

100

150

Temperature (oC) 0.2

(c) Ti50Ni46Fe4

R←B2 cooling

Heat Flow (W/g)

2.

(a) Ti50Ni48Fe2

0.1

0.0 heating

-0.1 R→B2

-0.2 -200

-150

-100

-50

0

50

100

150

Temperature (oC) Fig. 1 The DSC curves of (a) Ti50Ni48Fe2, (b) Ti50Ni48Fe3, and (c) Ti50Ni48Fe4 SMAs.

the specimen was heated to 600°C at a constant heating rate of 15°C/min and held at 600°C for 6 h to eliminate the hydrogen atoms in the specimen. During the heating process, the vacuum pressure of the furnace was decreased to 2.6664 © 10¹4 Pa. After the dehydrogenation treatment, the specimen was furnace cooled to room temperature. 3.

Results and Discussion

3.1 DSC and DMA results of Ti50Ni50¹xFex SMAs Figures 1(a) to 1(c) display the DSC curves of Ti50Ni48Fe2, Ti50Ni47Fe3, and Ti50Ni46Fe4 SMAs, respectively. Figure 1(a) shows that Ti50Ni48Fe2 SMA exhibits B2 ¼ R and R ¼ B19A transformation peaks during cooling at approximately ¹3.7°C and ¹61.4°C, respectively. The Rs, Rf, and Ms, Mf temperatures associated with the starting and finishing temperatures of B2 ¼ R and R ¼ B19A transformations,

Damping Characteristics of the Inherent and Intrinsic Internal Friction of Ti50Ni50¹xFex (x = 2, 3, and 4) Shape Memory Alloys

Tan ¤ values of Ti50Ni48Fe2 SMA measured under isothermal conditions Figure 3 plots the tan ¤ value versus isothermal interval for Ti50Ni48Fe2 SMA isothermally treated for 30 min at the IFTotal peak temperatures of B2 ¼ R (µ¹10°C) and R ¼ B19A (µ¹70°C) shown in Fig. 2(a). Figure 3 reveals that the tan ¤ value measured at the B2 ¼ R transformation temperature decreases as the isothermal interval increases, approaching a steady value of approximately tan ¤ = 0.022 after 10 min. Figure 3 also shows that the tan ¤ value measured at R ¼ B19A transformation temperature decreases significantly as the isothermal interval increases, approaching a steady value of approximately tan ¤ = 0.018 after 20 min. As shown in Fig. 3, the decayed tan ¤ value during the isothermal treatment represents the transient internal friction of B2 ¼ R 3.2

70000

(a) Ti50Ni48Fe2

60000

B19'←R

0.15

Tan δ

50000 0.10 40000

R←B2 0.05

30000 cooling

0.00 -150

Storage Modulus E0 (MPa)

0.20

20000 -100

-50

0

50 o

100

150

Temperature ( C) 0.20

70000

60000

0.15

B19'←R

Tan δ

50000 0.10 40000

R←B2

0.05

30000 cooling

0.00 -150

Storage Modulus E0 (MPa)

(b) Ti50Ni47Fe3

20000 -100

-50

0

50 o

100

150

Temperature ( C) 70000

0.20

60000

0.15

Tan δ

50000 0.10 40000

R←B2

0.05

30000 cooling

0.00 -150

Storage Modulus E0 (MPa)

(c) Ti50Ni46Fe4

20000 -100

-50

0

50 o

100

150

Temperature ( C)

Fig. 2 The IF tan ¤ and storage modulus E0 curves of (a) Ti50Ni48Fe2, (b) Ti50Ni48Fe3, and (c) Ti50Ni48Fe4 SMAs tested by DMA. 0.20

B2→R transformation peak (-10oC) R→B19' transformation peak (-80oC)

0.15

Tan δ

respectively, are 4.5°C, ¹16.0°C, ¹47.7°C and ¹89.3°C. During heating, Ti50Ni48Fe2 SMA has B19A ¼ R and R ¼ B2 transformation peaks at approximately ¹6.3°C and 14°C, respectively. The ¦H values of B2 ¼ R ¼ B19A and B19A ¼ R ¼ B2 peaks are both determined to be approximately 18 J/g. Figure 1(b) reveals that Ti50Ni47Fe3 SMA also exhibits B2 ¼ R ¼ B19A and B19A ¼ R ¼ B2 transformation sequences during cooling and heating, respectively, and the ¦H values associated with B2 ¼ R ¼ B19A and B19A ¼ R ¼ B2 transformations are both determined to be approximate 17 J/g. However, comparing to Ti50Ni48Fe2 SMA, all transformation peak temperatures of Ti50Ni47Fe3 SMA are depressed to lower temperatures. Figure 1(c) shows that, for Ti50Ni46Fe4 SMA, only B2 ¼ R and R ¼ B2 transformations appear during cooling and heating, respectively, and the ¦H values of B2 ¼ R and R ¼ B2 transformations are determined to be approximately 4.5 J/g. In this alloy, both R ¼ B19A and B19A ¼ R transformations are depressed to below ¹150°C. Figures 2(a) to 2(c) show the DMA curves of Ti50Ni48Fe2, Ti50Ni47Fe3, and Ti50Ni46Fe4 SMAs, respectively. For clarity, only the cooling DMA curves are shown in Fig. 2, for which the cooling rate is set constantly at 3°C/min. Figure 2(a) shows that Ti50Ni48Fe2 SMA exhibits B2 ¼ R and R ¼ B19A internal friction (IFTotal) peaks in the cooling tan ¤ curve, which correspond to the B2 ¼ R and R ¼ B19A transformation peaks in the cooling DSC curve shown in Fig. 1(a). The tan ¤ values of B2 ¼ R and R ¼ B19A IFTotal peaks are determined to be approximately 0.060 and 0.157, respectively. Figure 2(a) also indicates that the cooling storage modulus E0 curve declines gently in the regime of the B2 parent phase, reaching a deeper minimum during B2 ¼ R transformation and a shallower minimum during R ¼ B19A transformation. After the R ¼ B19A transformation, the E0 value of B19A martensite increases quickly as the temperature decreases. Figure 2(b) shows that Ti50Ni47Fe3 SMA also exhibits B2 ¼ R and R ¼ B19A IFTotal peaks during cooling, and their tan ¤ values are determined to be 0.060 and 0.130, respectively. Figure 2(c) shows that Ti50Ni46Fe4 SMA exhibits only a B2 ¼ R internal friction peak at approximately ¹55°C, and its tan ¤ value is determined to be 0.054. A deep E0 minimum during B2 ¼ R transformation is also observed in Ti50Ni47Fe3 and Ti50Ni46Fe4 SMAs.

353

0.10

(IFTr)R→B19'

0.05

0.00

(IFTr)B2→R (IFPT+IFI)B2→R

(IFPT+IFI)R→B19'

0

5

10

15

20

25

30

35

Isothermal Interval (min)

Fig. 3 The tan ¤ values versus isothermal interval when Ti50Ni48Fe2 SMA was isothermally treated for 30 min at peak temperatures of B2 ¼ R (¹10°C) and R ¼ B19A (¹80°C) transformations shown in Fig. 2(a).

((IFTr)B2¼R) or R ¼ B19A ((IFTr)R¼B19A) transformation. The final steady tan ¤ value obtained after 30 min is the inherent and intrinsic internal friction associated with B2 ¼ R ((IFPT + IFI)B2¼R) or R ¼ B19A ((IFPT + IFI)B2¼R) transformation.

354

S.-H. Chang, C. Chien and S.-K. Wu 0.05

0.20

(a) Ti50Ni48Fe2 B19'←R

Tan δ

Tan δ

0.10 (IFPT+IFI)R→B19'

R←B2 (IFPT+IFI)B2→R

0.00 -150

-100

-50

0

50

100

150

Tan δ

o

cooling at 3 C/min isothermal 30 min

B19'←R

0.10 R←B2 (IFPT+IFI)B2→R

0.05

0.00 -150

(IFPT+IFI)R→B19'

-100

-50

0

50

100

150

Temperature (oC) 0.20

(c) Ti50Ni46Fe4

o

cooling at 3 C/min isothermal 30 min

0.15

0.10 R←B2 0.05

0.00 -150

(IFPT+IFI)B2→R

-100

-50

0

50

0.02

0.00 -150

-100

-50

0

50

100

150

Temperature (oC)

0.20

(b) Ti50Ni47Fe3

0.03

0.01

Temperature (oC)

Tan δ

Ti50Ni46Fe4 Ti50Ni50 [9]

0.05

100

150

Temperature (oC) Fig. 4 The tan ¤ values versus temperature of (a) Ti50Ni48Fe2, (b) Ti50Ni47Fe3, and (c) Ti50Ni46Fe4 SMAs. The solid curves were measured at a constant cooling rate of 3°C/min and the empty circles were the (IFPT + IFI) tan ¤ curves measured after 30 min isotherm treatment at different temperatures.

Inherent and intrinsic (IFPT + IFI) internal friction of Ti50Ni50¹xFex SMAs Figures 4(a) to 4(c) plot the tan ¤ curves of the inherent and intrinsic (IFPT + IFI) internal friction determined after 30-min isothermal treatment at different temperatures for Ti50Ni48Fe2, Ti50Ni47Fe3, and Ti50Ni46Fe4 SMAs, respectively. The tan ¤ curves of Ti50Ni50¹xFex SMAs, shown in Fig. 2, are measured at a constant cooling rate of 3°C/min and are also plotted in Fig. 4 for comparison. Figure 4(a) shows an (IFPT + IFI)B2¼R peak corresponding to B2 ¼ R transformation with tan ¤ = 0.019 at 0°C and an (IFPT + IFI)R¼B19A peak corresponding to R ¼ B19A transformation with tan ¤ = 0.036 at ¹50°C. There is a temperature shift 3.3

Ti50Ni47Fe3

0.04

0.15

0.15

Ti50Ni48Fe2

o

cooling at 3 C/min isothermal 30 min

Fig. 5 The overlay of the (IFPT + IFI) tan ¤ curves of Ti50Ni48Fe2, Ti50Ni47Fe3, and Ti50Ni46Fe4 SMAs shown in Fig. 4. The (IFPT + IFI) tan ¤ curve of cold-rolled and annealed Ti50Ni50 SMA is also plotted for comparison.9)

between the peak temperatures of the (IFPT + IFI) peak measured under isothermal conditions and the IFTotal peak measured at a constant cooling rate of 3°C/min. This feature is due to the cooling rate effect, which has been investigated in detail previously.13) Figure 4(b) shows that Ti50Ni47Fe3 SMA exhibits an (IFPT + IFI)B2¼R peak with tan ¤ = 0.026 at ¹20°C and an (IFPT + IFI)R¼B19A peak with tan ¤ = 0.035 at ¹65°C. In Fig. 4(c), Ti50Ni46Fe4 SMA has only a single (IFPT + IFI)B2¼R peak with tan ¤ = 0.022 at ¹50°C. Figure 5 presents the (IFPT + IFI) tan ¤ curves of Ti50Ni48Fe2, Ti50Ni47Fe3, and Ti50Ni46Fe4 SMAs shown in Fig. 4. The (IFPT + IFI) tan ¤ curve of cold-rolled and annealed Ti50Ni50 SMA reported in our previous study is also plotted in Fig. 5 for comparison.9) This figure reveals that the tan ¤ values of the (IFPT + IFI)R¼B19A peak are larger than 0.035 for both Ti50Ni48Fe2 and Ti50Ni47Fe3 SMAs, and they are also larger than that of the (IFPT + IFI)B2¼R peak (approximately 0.02) for Ti50Ni50¹xFex (x = 2, 3, and 4) SMAs. This feature comes from the fact that the transformation strain associated with R ¼ B19A transformation is much greater than that associated with B2 ¼ R transformation.21) In addition, there are abundant twin boundaries between the variants of Rphase and B19A martensite, which can easily move under the applied damping amplitude and thus dissipate the energy.9) At the same time, the (IFPT + IFI)R¼B19A peak exhibited a higher damping capacity than the (IFPT + IFI)B2¼R peak is also due to the appearance of both R-phase and transformed B19A martensite during R ¼ B19A transformation, but only transformed R-phase can be obtained during B2 ¼ R transformation. Also evident in Fig. 5 is that cold-rolled and annealed Ti50Ni50 SMA exhibits (IFPT + IFI)B2¼R and (IFPT + IFI)R¼B19A peaks at approximately 30°C and 5°C, respectively, both with tan ¤ values of approximately 0.02. This tan ¤ value is much lower than those of Ti50Ni50¹xFex SMAs (approaching tan ¤ = 0.04) for the (IFPT + IFI)R¼B19A peak shown in Fig. 5. This characteristic can be explained by the fact that the origin of the R-phase in TiNi-based SMAs is typically either the rearranged dislocation structures in cold-rolled and annealed alloys or the Ti3Ni4 precipitation formed in Ni-rich TiNi SMAs aged at the proper temperature.1) These dislocations or

Damping Characteristics of the Inherent and Intrinsic Internal Friction of Ti50Ni50¹xFex (x = 2, 3, and 4) Shape Memory Alloys 0.20

Ti3Ni4 precipitates normally impede the movements of twin boundaries during damping and therefore reduce the (IFPT + IFI) internal friction of TiNi-based SMAs. In contrast, the R-phase formation in Ti50Ni50¹xFex SMAs originated from the replacement of Ni atoms with Fe atoms, instead of the introduction of dislocations or Ti3Ni4 precipitates. As a result, the Ti50Ni50¹xFex SMAs exhibit higher (IFPT + IFI)R¼B19A peaks than Ti50Ni50 SMA does. However, the (IFPT + IFI)R¼B19A peaks of Ti50Ni50¹xFex SMAs (x = 2 and 3) are depressed to a lower temperature than that of Ti50Ni50 SMA. The Ti50Ni48Fe2 and Ti50Ni47Fe3 SMAs exhibit this peak at temperatures below ¹40°C. Therefore, although Ti50Ni50¹xFex SMAs have quite good (IFPT + IFI) internal friction, their low martensitic transformation temperatures restrict their use in practical highdamping applications.

355

(a) Ti50Ni48Fe2 B19'←R

0.15

Before dehydrogenation

Tan δ

After dehydrogenation

0.10 R←B2

0.05

0.00 -150

-100

-50

0

50

100

150

Temperature (oC) 0.06

Before dehydrogenation After dehydrogenation

(b) Ti50Ni48Fe2 0.05 (IFPT+IFI)R→B19'

3.4

0.04

Tan δ

Damping characteristics of dehydrogenated Ti50Ni48Fe2 SMA Figures 6(a) and 6(b) show, for Ti50Ni48Fe2 SMA, the (IFTotal) tan ¤ curves measured at a constant cooling rate of 3°C/min and the (IFPT + IFI) tan ¤ curves measured under isothermal conditions, respectively, before and after the dehydrogenation treatment. Figure 6(a) reveals that the tan ¤ values of B2 ¼ R and R ¼ B19A martensitic transformations both decrease after dehydrogenation. In addition, the tan ¤ value also decreases in the R-phase regime (in the temperature range of Rf to Ms) after dehydrogenation; e.g., the tan ¤ value at ¹50.0°C decreases from 0.077 to 0.053, a decrement of approximately 30%. Fan et al.17) reported that the ultrahigh damping exhibited in Ti50Ni48Fe2 SMA comes from the relaxation peaks in the R-phase regime, wherein Rphase occupies most of the volume of the specimen. The twin boundaries in R-phase are very mobile, so the hydrogen atoms can pin twin boundaries and interact with them to promote the damping capacity. Therefore, the significant decrease in the tan ¤ value in the R-phase regime is due to the elimination of hydrogen atoms after dehydrogenation. Figure 6(b) shows that the tan ¤ values of both the (IFPT + IFI)B2¼R and the (IFPT + IFI)R¼B19A peaks decrease after dehydrogenation. In addition, the IFI values in the B2 and B19A martensite regimes, i.e., in the temperature (T ) ranges of T > Rf and T < Ms, respectively, are almost identical before and after dehydrogenation, but that in the R-phase regime decreases after dehydrogenation; e.g., the tan ¤ value at ¹40.0°C decreases from 0.031 to 0.027, a decrement of approximately 12%. This feature indicates that there is no obvious hydrogen pinning effect on the damping capacities of B2 phase and B19A martensite, but the effect on that of R-phase is significant. In the B2, R-phase, and B19A martensite single-phase regimes, there is no martensitic transformation, so the IFPT is equal to zero and only the IFI exists, as indicated in Fig. 6(b). It is well known that the B2 parent phase has no twin boundaries, and thus the hydrogen pinning effect is small. However, as can be seen in Fig. 6(b), the hydrogen pinning effect in the B19A martensite regime (T < Mf ) is also small. This characteristic may arise from the fact that the storage modulus E0 value in the regime of B19A martensite is higher than those in the regimes of R-phase and B2 phase, as shown in Fig. 2(a). At the same time, for the

(IFI)R

0.03 0.02

(IFPT+IFI)B2→R

(IFI)B19'

0.01

(IFI)B2

o

− 40 C

0.00 -150

-100

-50

0

50

100

150

Temperature (oC) Fig. 6 (a) The tan ¤ curves measured at a constant cooling rate of 3°C/min, and (b) the (IFPT + IFI) tan ¤ curves of Ti50Ni48Fe2 SMA before and after dehydrogenation treated at 600°C for 6 h in a vacuum furnace.

B19A martensite, the higher the E0 value is, the lower the IFTotal value is both before and after dehydrogenation, as shown in Fig. 6(a). This feature implies that the movement of twin boundaries exhibited in B19A martensite becomes more difficult as the E0 value increases. Figure 6(b) also shows that the hydrogen pinning effect is significant in the R-phase regime. This characteristic may correspond to two factors. First, the E0 value in the R-phase regime is not as high as that in the B19A martensite regime due to the deep E0 minimum during B2 ¼ R transformation. Second, the twin boundaries could move more easily in the R-phase than in the B19A martensite because the {110}B2 twin planes in between the R-phase variants are close-packed planes.22) Nevertheless, further detailed studies are needed to ascertain the mechanism of the hydrogen pinning effect on the IFPT and IFI of Ti50Ni50¹xFex SMAs. 4.

Conclusion

This study investigated the damping properties of Ti50Ni50¹xFex (x = 2, 3, and 4 at%) SMAs using DMA under a constant cooling rate of 3°C/min and under isothermal conditions. Both Ti50Ni48Fe2 and Ti50Ni47Fe3 SMAs undergo a two-stage B2 § R § B19A transformation, but Ti50Ni46Fe4 SMA undergoes only a one-stage B2 § R transformation. For Ti50Ni48Fe2 and Ti50Ni47Fe3 SMAs, the tan ¤ values of (IFPT + IFI)R¼B19A peaks are much higher than those of (IFPT + IFI)B2¼R peaks. The reasons are that R ¼ B19A

356

S.-H. Chang, C. Chien and S.-K. Wu

transformation has greater transformation strain and that both R-phase and transformed B19A martensite appear during transformation, but B2 ¼ R transformation has only transformed R-phase during transformation. The tan ¤ values of both the (IFPT + IFI)B2¼R and the (IFPT + IFI)R¼B19A peaks and those of (IFTr + IFPT + IFI) and IFI in the R-phase regime for Ti50Ni48Fe2 SMA all decrease after the dehydrogenation treatment at 600°C for 6 h in a vacuum furnace, indicating that the hydrogen pinning effect not only influences the damping properties measured at a constant temperature change rate but also affects those measured under isothermal conditions. Experimental results show that the IFI value in the R-phase regime decreases significantly, but those in B2 phase and B19A martensite are little affected by dehydrogenation. This feature arises from the facts that B2 phase has no twin boundaries and that B19A martensite has a higher storage modulus E0 than R-phase, which may impede the movement of twin boundaries. Ti50Ni50¹xFex (x = 2, 3) SMAs exhibit higher (IFPT + IFI) peaks than Ti50Ni50 SMA because the latter needs dislocations to be introduced for the R-phase formation, but they are not necessary in the former. Although Ti50Ni50¹xFex (x = 2, 3) SMAs have quite good (IFPT + IFI)R¼B19A peaks in the isothermal condition, their low transformation temperatures may restrict their use in practical high-damping applications. Acknowledgements The authors gratefully acknowledge the financial support for this research provided by the Ministry of Science and Technology (MOST) and National Taiwan University (NTU), Taiwan, under Grant Nos. MOST 103-2221-E-197-007, MOST 104-2221-E-002-004 and NTU 104R891803.

REFERENCES 1) K. Otsuka and X. Ren: Prog. Mater. Sci. 50 (2005) 511­678. 2) O. Mercier, K. N. Melton and Y. De Préville: Acta Metall. 27 (1979) 1467­1475. 3) S. K. Wu, H. C. Lin and T. S. Chou: Acta Metall. 38 (1990) 95­102. 4) B. Coluzzi, A. Biscarini, R. Campanella, L. Trotta, G. Mazzolai, A. Tuissi and F. M. Mazzolai: Acta Mater. 47 (1999) 1965­1976. 5) S. Kustov, S. Golyandin, K. Sapozhnikov, E. Cesari, J. Van Humbeeck and R. De Batist: Acta Mater. 50 (2002) 3023­3044. 6) J. E. Bidaux, R. Schaller and W. Benoit: Acta Met. 37 (1989) 803­811. 7) J. Van Humbeeck, J. Stoiber, L. Delaey and R. Gotthardt: Z. Metalkd. 86 (1995) 176­183. 8) W. Dejonghe, R. De Batist and L. Delaey: Scr. Metall. 10 (1976) 1125­ 1128. 9) S. H. Chang and S. K. Wu: Scr. Mater. 55 (2006) 311­314. 10) S. H. Chang and S. K. Wu: Mater. Sci. Eng. A 454­455 (2007) 379­ 383. 11) S. H. Chang and S. K. Wu: J. Alloy. Compd. 437 (2007) 120­126. 12) S. H. Chang and S. K. Wu: J. Alloy. Compd. 459 (2008) 155­159. 13) S. H. Chang and S. K. Wu: Mater. Charact. 59 (2008) 987­990. 14) C. M. Hwang and C. M. Wayman: Scr. Metall. 17 (1983) 1345­1350. 15) E. Goo and R. Sinclair: Acta Metall. 33 (1985) 1717­1723. 16) I. Yoshida, D. Monma and T. Ono: J. Alloy. Compd. 448 (2008) 349­ 354. 17) G. Fan, Y. Zhou, K. Otsuka and X. Ren: Appl. Phys. Lett. 89 (2006) 161902. 18) G. Fan, Y. Zhou, K. Otsuka and X. Ren: Mater. Sci. Eng. A 521­522 (2009) 178­181. 19) D. Wang, Z. Zhang, J. Zhang, Y. Zhou, Y. Wang, X. Ding, Y. Wang and X. Ren: Acta Mater. 58 (2010) 6206­6215. 20) S. Zuo, M. Jin, D. Chen and X. Jin: Scr. Mater. 108 (2015) 113­116. 21) H. C. Lin, S. K. Wu and M. T. Yeh: Metall. Mater. Trans. A 24 (1993) 2189­2194. 22) N. T. Tsou, C. H. Chen, C. S. Chen and S. K. Wu: Acta Mater. 89 (2015) 193­204.