Damping Characteristics of TiNi Shape Memory Alloys

H.C. LIN, S.K. WU, and M.T. YEH

The damping characteristics of TiNi SMAs have been systematically studied by using techniques
of resonant-bar and low-frequency inverted torsion pendulum. Experimental results show that
both the martensite phase (M) and R phase (R) have high damping due to the movement of
twin boundaries. Because the B2 parent phase (B2) has smaller damping, it is suggested that
this may come from the dynamic ordering process of lattice defects. In the transformation re-
gions of B2 ~-~ M, B2 ~ R, and R ~ M, there are maxima of the damping capacity which
are attributed to two contributions. One arises from the plastic strain and twin-interface move-
ment during the thermal transformation, which obeys a linear variation of peak heights amax-1 vs
i/" at i/" -> 1 ~ The other originates from the stress-induced transformation formed by the
applied external stress which dominates at T < 1 ~ The elastic modulus E of martensite
and the R phase is lower than that of the B2 phase, and a modulus minimum appears in the
transformation region.

I. INTRODUCTION the intermediate rhombohedral R phase (R). The domi-

nant damping mechanisms occurring in these phases and
TiNi alloys are known as the most important shape the characteristics associated with the thermoelastic
memory alloys (SMAs) because of their many applica-
transformations of TiNi SMAs are also discussed.
tions based on the shape memory effect (SME) and
pseudoelasticity (PE). This comes from the fact that TiNi
alloys have superior properties in ductility, strength, fa-
tigue, corrosion resistance, recoverable strain, etc. It is II. EXPERIMENTAL PROCEDURES
also well known that TiNi alloys can exhibit high me- A. Materials
chanical damping. I~-41 However, our understanding in this
area is incomplete. No systematic investigation has been The conventional tungsten arc-melting technique was
reported on the damping characteristics of TiNi SMAs. employed to prepare Tia9.sNis0.2 and Ti49Nisl alloys. Ti-
Damping mechanisms, in general, involve the stress- tanium (purity, 99.7 pct) and nickel (purity, 99.98 pct),
induced movement of defects. For high-damping metals, totaling about 150 g, were melted and remelted at least
the major mechanisms are the stress-induced movement six times in argon atmosphere. Pure titanium buttons were
of dislocations or planar defects, tS1 Most of these mech- also melted and used as a getter. The mass loss during
anisms can be phenomenologically split into three classes: melting was negligibly small. The as-melted buttons were
dynamic hysteresis, static hysteresis, and transformation homogenized at 1050 ~ for 72 hours and quenched
mechanisms. Dynamic hysteresis is produced by the stress- in water, then hot-rolled to plates with 2-mm thick-
aided ordering of defects overcoming local barriers by ness. Specimens for the resonant-bar damping test
thermal activation and yields damping that is frequency (size: 200 x 15 z 2 mm) were carefully cut from these
dependent and amplitude independent. Static hysteresis plates with a low-speed diamond saw. The rest of these
appears due to the stress-induced "unpinning" or "break- plates were then hot-rolled to 1-mm thickness and sub-
away ~ process of the defects I5,6,7] and yields damping that sequently cut to be the specimens for the electrical re-
is frequency independent and amplitude dependent. Some sistivity measurement (size: 50 x 1 x 1 mm) and internal
metals exhibit a high level of damping in the region of friction test (size: 110 x 4 x 1 mm). The cut specimens
a transformation, for example, in the temperature range were sealed in evacuated quartz tubes, annealed at 800 ~
of a thermoelastic martensitic transformation, tS'm Such for 2 hours, and quenched in water. Some Tia9Nis~ spec-
thermoelastic damping is frequently amplitude indepen- imens were then aged in a 400 ~ salt bath for 5 or
dent and proportional to ITI/T, where ITI is the heating 20 hours and quenched in water. The martensitic and
or cooling rate and f is the frequency of vibration. premartensitic transformation temperatures were tested
In this study, by using both resonant-bar and low- by using four-probe electrical resistivity and internal
frequency inverted torsion pendulum techniques, the friction measurements.
damping capacity of TiNi alloys was investigated in the
high-temperature cubic B2 parent phase (B2), the low-
B. Damping Experiments
temperature monoclinic B I9' martensite phase (M), and
There are a variety of experimental methods available
H.C. LIN, Associate Professor, is with the Department of Materials to measure the damping properties of a material. Each
Science, Feng-Chia University, Taichung, Taiwan 400, Republic of technique has its own unique advantages, but they all
China. S.K. WU, Professor, is with the Institute of Materials Science yield equivalent measures of damping, lSl In this study,
and Engineering, National Taiwan University, Taipei, Taiwan 106, two techniques, the flexural resonant-bar damping sys-
Republic of China. M.T. YEH is with the Materials R&D Center,
Chung-Shan Institute of Science and Technology, Lung-Tan, Taiwan tem and the internal friction, were used to measure the
325, Republic of China. damping properties of TiNi alloys. The resonant-bar
Manuscript submitted February 15, 1993. damping system is illustrated in Figure 1. The simple

METALLURGICAL TRANSACTIONS A VOLUME 24A, OCTOBER 1993--2189

 [ l
I Accelerometer ihit ,
I II 1.
i---~[i.---I~'l. Specimen 1 Ii AJAAA,......... [J.. IAA IAAI AAALAA,, , . . . . . .]
IlrVWV'.... v'V ~'1 IlrlV vv,, . . . . . . .
I ' .~....__ i I I I !v \
I ~Thermometer \

I '\i
t I I I
Temperat ure -cont roiled chamber 80-

6, W 'I
ii L'9 IYI IVWvvvm-,

I
oscilloscope ~ ~ PC O 48 \
\,
\ o /
5" 32 o
o
Graphic Printer
16-

I 0
o
o~

-15~ -100 -50 0 50 " 180

TEMPERATURE (*C)
Fig. 2--Typical experimentalplots of both the resonant-bardamping
test and internal friction measurement for the Tiag.sNiso.2 alloy.

Fig. 1 --Schematic diagram of the resonant-bar damping system used

in this study. III. EXPERIMENTAL RESULTS
A. Experimental Results of the Resonant-Bar
Damping Test
free-free cantilever beam can be excited into its funda-
mental mode of flexural vibration I~~ by an exciting steel In order to understand the damping property of TiNi
ball. After the exciting steel ball hits the specimen, the SMAs, including the B2 and R phases and B19' mar-
amplitude of vibration gradually decreases with time as tensite, the resonant-bar damping tests were carried out
the vibrational energy is dissipated. These decayed am- at various temperatures, where TiNi alloys exhibit dif-
plitudes and frequency of vibration are transfered to an ferent phases during cooling. All the experimental re-
oscilloscope by an accelerometer attached on specimen's sults, involving testing temperatures, existing phases,
end. Measurement of successive strain amplitudes from the damping ratio ~, and resonant frequency f are
the oscilloscope will then yield the logarithmic decre- summarized in Tables I and II for Ti498Nis0.2 alloy
ment, 6 as follows: and 400 ~ • 20 h aged Ti49Nis~ alloy, respectively. In
order to further show the relationship between the damp-
6 = In (A,/A,+,) [11 ing property and existing phases, damping ratio g: and
electrical resistivity are plotted vs temperature in
Where A. and A,+I are the amplitudes of successive cycles Figures 3 and 4 for the solution-treated Ti49.8Nis0.2 and
in free decay. The relationship between 6, internal fric- 400 ~ • 20 h aged Ti49Nisx alloy, respectively. In
tion factor Q-l, and damping ratio sr is given by tSI Figure 3, the damping ratio ~ of martensite is larger than
that of the B2 phase and there is a sharp peak of ~: in
6 the temperature range of the B2 --> M transformation. In
Q-l - 2~:- - [2] Figure 4, the damping ratio ~: of martensite has nearly
'IT
the same magnitude as that of the R phase but is larger
than that of the B2 phase. The peaks of the damping
The specimen's temperature can be measured in situ by
ratio ~ appear in the temperature ranges of the B2 --> R
a thermometer and be controlled by a temperature-
and R ~ M transformations. The characteristics of the
controlled chamber, as shown in Figure 1. Internal fric-
damping shown in Figures 3 and 4 will be discussed in
tion tests were carried out using a SINKU-RIKO
Section IV.
1500-M/L series inverted torsion pendulum in the tem-
perature range from - 1 5 0 ~ to + 100 ~ The measur-
ing frequency was about 1 Hz, and the temperature
B. Experimental Results of lnternal
changing rate Ii#[ was set to be 1, 2, and 3 ~
Friction Measurement
Figure 2 shows typical experimental results of both the
resonant-bar damping test and the internal friction mea- Figures 5(a) and (b) show plots of the internal friction
surement for Ti49.sNis0.2 alloy. In Figure 2, one can find Q-~ vs temperature for the solution-treated Ti49.8Nis0.2 and
that the variation of the damping property measured by 400 ~ • 5 h aged Ti49Nis~ alloys, respectively. In
these two techniques behaves with a similar tendency. Figure 5(a), it appears that transformation peaks, Pc] and
Namely, the higher the internal friction, the faster the Pro, and relaxation peaks, PcR and PHR, on the cooling
amplitude decay (or the higher the 6 value). and heating cycles. Peaks PcJ and Pm are associated with

2190-- VOLUME 24A, OCTOBER 1993 METALLURGICAL TRANSACTIONS A

 Table I. The Existing Phases, Damping Ratio ~, and Resonant
Frequency f in Different Damping Testing Temperatures for the Ti49.sNis0.z Alloy
Temperature (~ 60 45 34 30 25 20 10 0 -9
Phases B2 B2 B2 + M B2 + M B2 + M B2 + M M M M
Damping ratio ~ 0.009 0.010 0.013 0.030 0.051 0.036 0.021 0.020 0.021
f (Hz) 167 166 158 153 150 151 155 155 156

Table II. The Existing Phases, Damping Ratio ~, and Resonant Frequency f
in Different Damping Testing Temperatures for the 400 ~ • 20 h Aged Ti49Nisl Alloy
Temperature (~ 75 60 45 39 37 30 15 -5 -15 -21 -30 -45
Phases B2 B2 B2 + R B2 + R B2 + R B2 + R R R + M R + M R + M M M
Damping ratio ~ 0.009 0.010 0.014 0.025 0.045 0.034 0.028 0.037 0.053 0.040 0.026 0.026
f(Hz) 155 155 138 130 119 125 136 137 135 141 146 147

f c7~
0 >
El "-'
0 >
~ J
J .4-,

J
LjJCr

~,~
o
0.04
/'1 005

0.04
Vl~ t e n s i t i /
/'\ ._phose/,
~ . J " i
0 0
-- 0.03. :,7, 0.03
c~ El
cr 0.02, r
0.02 - -
03
(-- B2 ph(~se cn
Mortensite c"
'~. 0.01 "~_ 0.01
E E
0.00 C~
r o 09
-10 0 10 20 30 40 50 60 70 80 90 -60 -30 0 30 60 90
Temperature (*C) Temperature (*C)

Fig. 3 - - D a m p i n g ratio ~ and electrical resistivity vs temperature curves Fig. 4 - - D a m p i n g ratio ~ and electrical resistivity vs temperature curves
for the Ti49.sNis0.z alloy. for the 400 ~ x 20 h aged TiaDNisi alloy.

the martensitic transformation, tl'31 while peaks PCR and

but some constant value higher than the background. This
P.~ are related to the relaxation of dislocation damp-
characteristic will be discussed in Section IV.
ing. TM In Figure 5(b), the Pc1 and Pm peaks are also as-
sociated with the martensitic transformation while Pc2
and PH2 peaks are associated with the premartensitic R IV. DISCUSSION
phase transformation. Comparing Figures 5(a) and (b) to
A. The Damping Capacity and Damping Mechanism
Figures 3 and 4, one can find that the Q-~ value from
of B2, B19', and R Phases in TiNi Alloys
the internal friction measurement as a function of tem-
perature behaves similarly to the damping ratio ~: ob- As mentioned in Section I, most damping mechanisms
tained from resonant-bar damping tests. Both the B19' involve the stress-induced movement of defects, tSJ Point
martensite and R phase exhibit a same-order damping defects give rise to damping in the range of low to inter-
capacity which is larger than the B2 phase, and the mediate levels, line defects give rise to damping levels
damping capacity peaks appear in the temperature ranges in the intermediate to high range, and planar defects give
of the martensitic and premartensitic transformation re- rise to damping levels in the high range. It is well known
gions in TiNi SMAs. that there are abundant twin boundaries in the B 19' mar-
The height of the transformational peaks, Pc~, pro, tensite and R phase of TiNi alloys, tl~'lzj These twin
Pc2, and PH2, of Figure 5 can be affected by the heating boundaries can be easily moved by the external stress to
or cooling rate/', as shown in Figure 6. Figures 6(a) and accommodate the strain. This phenomenon is the well-
(b) present these peak heights as a function of 7~ without known "accommodation/reorientation" process occur-
subtraction of the background for the TiaD.sNis0.2 and ring in the martensite and R phase of the deformed TiNi
400 ~ • 5 h aged Tin9Nisj alloys, respectively. A linear alloys. We postulate that this accommodation/reorientation
variation of these heights vs T is observed. When ex- phenomenon is closely related to the high damping ca-
trapolated to 7~ = 0, these heights do not approach zero pacity of TiNi alloys and propose the following damping

METALLURGICAL TRANSACTIONS A VOLUME 24A, OCTOBER 1993--2191

 Q-i X 10 s (a)
GO r = 2.32 x I0- ~
-1
Qmax E=2.32 xlO4
+ HEATING
0 COOLING f =0.90 Hz
Pcl
o
~ 0.07.
oo

~ . a 0 06
,8]
o
o o

o
~, .0<05
P~.~ o o §
32 ,,-%., o , 0.04
/&
; + § 0.03 /
o o +
l& +
0.02
Cooling Heating
0.01
i "r [*C/m in]
-150 -50 ~ 50 100 I
TEMPERATURE (*C) ' -~3 -- '2 -- 'I 0 4 2 3
(b)
(a) -I
Qmax E= 1.74x104
Q-x X I0s ,' f =0.56 Hz
40 r = 1.74 x 10 - b 0.04
= 2~
o J~c i
_~o~" 4 /
HEATING
+
'%~ '' 9~ 4;).0
32] o COOLING
0 o

"t P.i 5=
++
o ~+
v
24' i o o § +§ c2 "-
9 , Cooli ng 0.01- Heating
o o ~+§ ~P +
9 o ~ ,, ~o i" [*C/m i n]
16
I I
-~ "2 "1 o 1 2 :3 ---
o §
~ o~ Fig. 6 - - T h e internal friction peak heights (Q~x) v s temperature
o +
changing rate (T) for (a) Ti498Nis0.2 alloy and (b) 400 ~ x 5 h aged
Ti49Nisl alloy.
o*,
--" i i l # i " i
-~s* 40 -do o do ~;o
TEMPERATURE( ' C > the Mn-Cu-based alloys ll6] and some Cu-based shape
(b) memory alloys. 117,a
No twin boundaries exist in the parent B2 phase of
Fig. 5--1nternal friction v s temperature curves for (a) Tin98Niso.2 TiNi alloys, and the dislocation density in the matrix is
alloy and (b) 400 ~ x 5 h aged Ti49Nis~ alloy.
quite low. 1191Hence, the damping capacity is suggested
simply to come from the dynamic hysteresis of lattice
defects, such as vacancy or interstitial. The stress-induced
mechanism to account for this relationship. The stress- ordering process of these defects leads to an anelastic
strain diagram for the accommodation/reorientation pro- strain, and the damping mechanism is referred to as a
cess during the damping test is schematically drawn in linear reversible anelastic relaxation and opens up the
Figure 7. Also, Muller et al. had recently studied the dynamic hysteresis loop. Because the dynamic hysteresis
stress-strain curve of the pseudoelastic hysteresis on the loop generally dissipates a smaller quantity of energy,
CuZnAI single crystal, u31 However, in our proposed dia- the damping capacity in the B2 phase of TiNi alloys is
gram of Figure 7, it is shown that after an elastic re- smaller, as shown in Figures 3 through 5.
sponse to the stress, an accommodated strain e, in some
microdomains can be produced at a critical value of the
stress, O'a. This strain is due to the stress-induced move- B. The Damping Property Associated with
ment of twin boundaries between the variants of mar- Thermoelastic Phase Transformations in TiNi Alloys
tensite or R phase. The accommodated strain is retained As discussed in Section I V - A , the martensite and R
during the unloading but can be reoriented to the op- phase of TiNi alloys have a high damping capacity due
posite direction due to the movement of twin boundaries to the movement of twin boundaries, but the parent B2
induced by the following opposite-direction stress, -o'a. phase shows a low-damping capacity which is suggested
This opens up a relatively large static hysteresis loop, simply to come from the dynamic ordering process of
AW, for the cyclic movement of twin boundaries. There- lattice defects. In Figures 3 through 5, there are peaks
fore, the martensite and R phase of TiNi alloys have a of damping capacity in the transformation regions of B2
high-damping capacity, comparable to or even higher than (--) M, B2 (--) R, and R (--) M. The maximum value of
cast irons. [~uS} The high-damping properties due to the the damping capacity occurring in the temperature ranges
movement of twin boundary have also been observed in of transformation is two times or even higher that

2192--VOLUME 24A, OCTOBER 1993 METALLURGICALTRANSACTIONSA

 In Figures 6(a) and (b), an approximately linear vari-
0" ation of peak heights Qmax -1 of PHI, P c I , PH2, and Pc2 vs
J" (i" = 1,2, 3 ~ is observed. This result indicates
that both martensitic and R-phase transformations agree
with the Delorme's model at ]/" = 1 ~ to 3 ~
Besides, in Figures 6(a) and (b), as T is extrapolated to

of
zero, the peak heights are higher than the background.
At T = 0, no martensite or R phase is formed by a ther-
mal driving force, but either can be formed by the ap-
plied external stress. In TiNi alloys, the deformation
behavior shows that the stress-induced transformation
occurs before the reorientation of variants of martensite
or R phase in the temperature range of forward trans-
formations on cooling. I~9~Hence, the damping capacity

\\\\ E
v appearing at T = 0 on cooling, as shown in Figure 6, is
ascribed to the stress-induced transformation. Tadaki

t
et al.la3] have reported that the volume change and shape
strain associated with the martensitic transformation are
much larger than those associated with the R-phase
transformation. Based on this report, dqJ(Vm)/dVm for
martensitic transformation is larger than that for R-phase
transformation. Therefore, the internal friction of the
-(Ia R ~ M transformation should be much larger than that
of the B2 ~ R transformation. Therefore, the Pcl peak
(R ~ M) at ~/" = 0 is much higher than the Pc2 peak (B2
R) at the same strain amplitude. However, in the
Fig. 7--Schematic stress-strain diagram for the martensite/R phase heating process, the existing martensite or R-phase vari-
accommodation/reorientation process. AW indicates the energy loss ants should be reoriented to accommodate the applied
for the cycling movement of twin boundaries. strain. As discussed in Section I l l - A , the damping ca-
pacities due to the stress-induced movement of the twin
boundary in the accommodation/reorientation process for
occurring in martensite or the R phase. This feature of martensite and R phase have nearly the same magnitude.
high-damping capacity associated with the phase trans- Hence, the peak heights Qm~x of PH~ and PH2 at ~/" = 0
formations of TiNi alloys needs to be further understood. have nearly the same values, as shown in Figure 6(b).
Postnikov et al. 12~ had shown that all the first-order Recently, Zhu et al. 1241reported that the relationship
phase transformations should be accompanied with an between Q-~ and J ' / f for TiNi alloy is nonlinear, espe-
internal friction peak, which is explained by the differ- cially for 7~ < 1 ~ contrary to the linear relation
ence in volume between the two phases. Delorme et al. t2jl of Delorme's theory. This feature can be reasonably ex-
extended the Postnikov's model by considering the plas- plained as follows. As discussed earlier, the damping
tic strain during transformation and deduced the internal can arise from both thermal (cooling or heating) and me-
friction factor Q-~ as a function of temperature rate, chanical (external stress) driving forces. If T is higher,
dT/dt: the damping capacity is dominated by the thermal driv-
ing force, hence, Delorme's model is reasonable and the
l d~b(Vm) dVm dT relation between Q-~ and ~/'/f is approximately linear, as
Q l. . . . . . [3] shown in Figure 6 for T = 1 ~ to 3 ~ How-
w dVm dT dt ever, when T is lower (e.g., < 1 ~ the damping
capacity is dominated by the external stress and cannot
where Vm is the volume fraction of martensite, w is the be predicted by Delorme's theory.
angular frequency of applied stress, and ~O(Vm) is a
monotonic function associated with the transformation
volume change and/or shape strain. Equation [3] indi- C. The Elastic Modulus E in TiNi Alloys
cates that the internal friction factor Q-~ is proportional In addition to the variation of the damping capacity
to the heating and cooling rate, T. Dejonghe et al.,t221 in with temperature in TiNi alloys, it can also be found that
order to take account of the special character of the mar- in Tables 1 and 2, the resonant frequencies f of mar-
tensite which can be induced or reoriented by an external tensite and R phase are lower than that of the B2 phase.
stress o-, introduced the stress dependence to dVm/dt as There are frequency minima in the temperature ranges
the following: for both martensitic and R-phase transformations. This
phenomenon was also found in internal friction measure-
dVm OVm OT OVm Otr
. . . . . + --'-- [4] ments, t2"3~The resonant frequency is proportional to the
dt OT Ot Otr Ot square root of the elastic modulus E, f ~ X/E. Hence,
the experimental results indicate that the B2 phase has
In Eq. [4], the first term is identical to the Delorme's a higher elastic modulus than the martensitic and R phases,
model and the second term is the stress-dependent one. and there are modulus minima in the temperature ranges

METALLURGICAL TRANSACTIONS A VOLUME 24A, OCTOBER 1993--2193

of transformations. This result is consistent with some Y.C. Lo, Institute of Materials Science and Engineering,
reports in which the modulus is measured from stress- National Taiwan University, Taipei, Taiwan, for helpful
strain curves, t25J Comparing the resonant frequency in discussions during the research process.
Table I with that in Table II, one can find that the res-
onant frequency of Tiag.8Nis0.2 alloy is higher than that
of 400 ~ • 20 h aged Tia9Nisi alloy. This indicates that REFERENCES
the elastic modulus of the former alloy is higher than 1. O. Mercier, K.N. Melton, and Y.De. Preville: Acta Metall., 1979,
that of the latter alloy. The precipitates and/or matrix vol. 27, pp. 1467-75.
composition difference may be related to this feature. 2. K. Iwasaki and R. Hasiguti: Trans. JIM, 1987, vol. 28, pp. 363-67.
3. S.K. Wu, H.C. Lin, and T.S. Chou: Acta Metall. Mater., 1990,
vol. 38, pp. 95-102.
4. C.M. Jackson, H.J. Wagner, and R.J. Wasilewski: NASA-SP
V. CONCLUSIONS 5110, 1972, pp. 35-41.
5. I.G. Ritchie and Z.-L. Pan: Metall. Trans. A, 1991, vol. 22A,
The damping characteristics of TiNi SMAs have been pp. 607-16.
systematically studied by using resonant-bar and low- 6. R. de Batist: Internal Friction of Structural Defects in Crystalline
frequency inverted torsion pendulum techniques. Both Solids, North-Holland Publishing Co., Amsterdam, 1972.
B19' martensite (M) and R phase (R) have a high-damping 7. D.W. James: Mater. Sci. Eng., 1969, vol. 4, pp. 1-8.
capacities due to the stress-induced movement of twin 8. R. de Batist: J. Phys., 1983, pp. C9-39.
9. J. Van Humbeeck: Proc. ASM Materials Week and TMS/AIME
boundaries. On the other hand, the parent B2 phase has Fall Meeting, Role of Interfaces on Materials Damping, 1985,
a smaller damping capacity which is suggested simply pp. 5-24.
to come from the dynamic ordering process of lattice 10. A.S. Nowick and B.S. Berry: Anelastic Relaxation in Crystalline
defects. In the transformation regions of B2 ~ M, B2 Solid, Academic Press, New York, NY, 1972.
11. T. Saburi and S. Nenno: Proc. Int. Conf. on Solid to Solid Phase
R, and R o M, there are damping capacity maxima
Trans., ASM, Metals Park, OH, 1982, pp. 1455-79.
which are two or more times higher than the damping 12. S. Miyazaki and C.M. Wayman: Acta Metall., 1988, vol. 36,
capacity of martensite or R phase. Two contributions are pp. 181-92.
proposed to account for the occurrence of the maxima 13. I. Muller and H. Xu: Acta Metall. Mater., 1991, vol. 39,
of damping capacity. One arises from the plastic strain pp. 263-71.
14. M.T. Yeh, H.C. Lin, and S.K. Wu: National Taiwan University,
and twin-interface movement during the thermal-induced Taipei, Taiwan, unpublished research, 1993.
transformation, which obeys a linear variation of 15. W.J. Buehler and R.C. Wiley: Report NOLTR 61-75, U.S. Naval
peak heights amax
-1 v s heating or cooling rate l/" at J" -> Ordinance Laboratory, Silver Spring, MD, Aug. 3, 1961.
1~ The other originates from the stress-induced 16. A.F. Schwaneke and J.W. Jensen: J. Appl. Phys., 1962, vol. 33,
pp. 1350-51.
transformation caused by the applied external stress which
17. K. Sugimoto, T. Mori, K. Otsuka, and K. Shimizu: Scripta Metall.,
dominates at i/" < 1 ~ The experimental results 1974, vol. 8, pp. 1341-46.
indicate that the elastic moduli E of martensite and the 18. M. Morin, G. Guenin, S. Etienne, and P.F. Gobin: Trans. JIM,
R phase are lower than that of the B2 phase in TiNi 1981, vol. 22, pp. 1-5.
SMAs, and there are modulus minima existing in the 19. S. Miyazaki and K. Otsuka: Metall. Trans. A, 1986, vol. 17A,
pp. 53-63.
transformation regions. 20. V.S. Postnikov, S.A. Gidnev, B.M. Darinskii, and 1.M.
Sharashakov: N u o v o Cim. B, 1976, vol. 33, pp. 324.
21. J.F. Delorme, R. Schmid, M. Robin, and P. Gobin: J. Phys.,
ACKNOWLEDGMENTS 1971, vol. 32, pp. C2-101-09.
22. W. Dejonghe, R. De Batist, and L. Delaey: Scripta Metall., 1976,
The authors are pleased to acknowledge the financial vol. 10, pp. 1125-28.
support of this research by the National Science Council 23. T. Tadaki, Y. Nakata, and K. Shimizu: Trans. JIM, 1987,
(NSC), Republic of China, under Grant No NSC 82-0405- vol. 28, pp. 883-90.
24. J.S. Zhu, R. Schaller, and W. Benoit: Phys. Status Solidi A,
E002-080. The authors thank Mr. T.S. Chou, China Steel 1988, vol. 108, pp. 613-18.
Corporation, for his excellent assistance in the internal 25. K.N. Melton and O. Mercier: Acta Metall., 1980, vol. 29,
friction measurement. Also, we would like to thank Dr. pp. 393-98.

2194--VOLUME 24A, OCTOBER 1993 METALLURGICAL TRANSACTIONS A