Materials Transactions, Vol. 47, No. 3 (2006) pp. 645 to 649
Special Issue on Shape Memory Alloys and Their Applications
#2006 The Japan Institute of Metals
Effect of Heat Treatment Atmosphere on Multistage R-Phase Transformation
in an Aged Ti–51.0 at%Ni Alloy
Minoru Nishida1; *1 , Kentarou Ishiuchi1; *2 , Kousuke Fujishima1; *3 and Toru Hara2
1
2
Department of Materials Science and Engineering, Kumamoto University, Kumamoto 860-8555, Japan
National Institute for Materials Science, Tsukuba 305-0047, Japan
The present study systematically investigates the effect of heat treatment atmosphere on the multistage R-phase transformation (MRT) in
an aged Ti–51.0 at%Ni alloy. No MRT occurs when the heat treatments were completed under the regulated atmosphere. On the other hand, the
MRT is observed in the specimen heat treated under the unregulated atmosphere. It is apparent from transmission electron microscope
observations that the first and the second transformations take place around the grain boundary and at the grain interior, respectively. We
conclude that the MRT is extrinsic, and is an artifact during the heat treatment, rather than intrinsic in nature.
(Received September 20, 2005; Accepted November 7, 2005; Published March 15, 2006)
Keywords: titanium nickel alloy, multistage R-phase transformation, heat treatment atmosphere, aging
1.
Introduction
The aging treatment is effective for the improvement of
shape memory and superelastic properties in Ni-rich Ti–Ni
alloys due to the precipitation strengthening with coherent
Ti3 Ni4 particles. It has been recognized that the aging
treatment induces the R-phase transformation. In addition to
the R-phase transformation, the multistage martensitic transformation (MMT) appears in the aged Ni-rich Ti–Ni alloys.
Four mechanisms of the MMT have been proposed recently.1–4) Bataillard et al. have ascribed it to coherent stress
fields around Ti3 Ni4 precipitates.1) Khalil-Allafi et al. have
explained it on the basis of evolving Ni concentration profiles
between particles and differences in nucleation barriers
between R-phase and B190 martensitic phase.2) The modified
report of Khalil-Allafi et al. has pointed out that the above
two mechanisms cannot rationalize the multistage transformation, and thus the heterogeneity in precipitation
morphology of Ti3 Ni4 phase around the grain boundary and
at the grain interior is responsible for the multistage transformation from transmission electron microscope (TEM)
observations.3) Fan et al. have ascribed it to a result of
competition between preferential grain boundary precipitation of Ti3 Ni4 particles and a tendency for homogeneous
precipitation when supersaturation of Ni is large.4) The
common feature of the microstructure in these four mechanisms is the heterogeneity in precipitation morphology of
Ti3 Ni4 phase. We do not intend to argue against those
proposed mechanisms, because they have been discussed
self-consistently within each experimental condition. However, no reports have been taken into account the heat
treatment atmosphere except for the present authors, since we
have experienced that the MMT is remarkably influenced by
heat-treatment conditions, especially heat-treatment atmosphere in the solution treatment at high temperature.5) No
MMT occurs when the evaporation of Ti and Ni and/or the
*1Corresponding
author, E-mail: [email protected]
Student, Kumamoto University, presently Graduate
Student, Tokyo Institute of Technology
*3Graduate Student, Kumamoto University
*2Undergraduate
preferential oxidation of Ti in the specimen are prevented and
the purification of heat treatment atmosphere is achieved.
The heterogeneity in precipitation morphology of Ti3 Ni4
phase, which is responsible for the multistage transformation
as mentioned above, can be suppressed with the regulation of
heat treatment atmosphere. In addition to the MMT, the
appearance of MRT in Ti–50.9 at%Ni alloy aged below
573 K has been reported recently and investigated comprehensively by Kim et al.6) They conclude that the first R-phase
transformation takes place in ‘‘affected’’ zones, i.e., around
Ti3 Ni4 precipitates and the second one occurs in ‘‘unaffected’’ matrix, i.e., away from precipitates. Consequently, the
occurrence of the MRT is attributed to precipitation-induced
heterogeneity of the matrix, both in terms of composition and
of internal stress fields. However, there is no direct evidence
of the ‘‘affected’’ zone and the ‘‘unaffected’’ matrix in their
TEM observations, although they comment that the term
‘‘unaffected’’ is only comparative to the ‘‘affected’’ zone.
Besides the mechanism of the MRT, one can easily imagine
that the appearance of the MRT also depends on the heat
treatment atmosphere as well as the MMT.
The purpose of the present study is to clarify the effect of
heat treatment atmosphere on the MRT in an aged Ni-rich Ti–
Ni alloy. TEM observations are also carried out to investigate
the correspondence between transformation regions and
multi peaks in the DSC curve.
2.
Experimental Procedure
Ti–51.0 at%Ni alloy was prepared from 99.7 mass%
sponge Ti and 99.9 mass% electrolytic Ni by a highfrequency vacuum induction furnace using a graphite
crucible, followed by casting into an iron mold. The ingot
was hot-forged and drawn to rod of 3 mm in diameter. A part
of the forged ingot was rolled to sheet of 0.5 mm in thickness.
The rod was cut into disks of about 1 mm in thickness for
DSC measurements and about 0.2 mm in thickness for TEM
observations, respectively. The sheet was used for the cover
plate of DSC and TEM specimens as described below. Four
types of heat treatment conditions were examined in the
present study are schematically illustrated in Fig. 1. In the
646
M. Nishida, K. Ishiuchi, K. Fujishima and T. Hara
Quartz tube
Vacuum
Specimen
exothermic
A
M
(a)
M
(b)
M
Ti-Ni plate
Ti foil
Heat flow
B
(c)
M
(d)
C
1 W/mol
200
250
300
Temperature, T /K
D
Fig. 2 DSC cooling curve of solution treated specimen with conditions (a)
A to (d) D.
Fig. 1
Schematic illustration of heat treatment conditions A to D.
condition A, the specimens were solely sealed in the
evacuated quartz tube of 2:5 103 Pa. The top and bottom
surfaces of specimens were covered with a pair of rolled Ti–
Ni sheets in the evacuated quartz tube in the condition B. In
the condition C, the specimens were wrapped with pure Ti
foil of 200 mm in thickness as getter material in the evacuated
quartz tube. The rest were sandwiched between the rolled
sheets and then wrapped with pure Ti foil in the evacuated
quartz tube, which is referred to as the condition D. They
were solution-treated at 1073 K for 3.6 ks, and then quenched
into ice water. Subsequently, the specimens were aged at
473 K for various periods from 36 to 1800 ks, since the Rphase appears in the Ti–50.9 at%Ni alloy aged at 473 K for
36 ks or more in the comprehensive work of Kim et al.6) For
instance, the specimen A–A described in the later section
indicates that both the solution treatment and the aging are
completed in the condition A. Differential scanning calorimetry (DSC) measurements were performed by using a
Shimadzu DSC-50 calorimeter with a cooling and heating
rate of 10 K/min. The temperature range measured was about
170 to 350 K. TEM specimens were electropolished using the
twin jet method in an electrolyte of 20% HNO3 and
80%CH3 OH about 250 K. TEM observations were carried
out in the JEOL-2000FX microscope operated at 200 kV.
3.
Results and Discussion
Figures 2(a) to (d) show DSC cooling curves of the
solution treated specimens with conditions A to D, respectively. There is no difference between the four curves
essentially. The peak temperature is about 225 K irrespective
of the heat treatment atmosphere. The single exothermic peak
is due to the martensitic transformation, since the thermal
hysteresis between heating and cooling processes was
relatively large, about 30 K in comparison to that of the Rphase transformation described below. Figure 3 shows DSC
cooling curves of each specimen aged at 473 K for various
periods. The MRT denoted as R1 and R2 takes place in the
specimens A–A and B–B aged for 72 ks or more as shown in
(a) and (b), respectively. It is most likely that both R1 and R2
peaks are attributable to the R-phase transformation, since
their thermal hysteresis is small about 5 K in the incomplete
and the complete thermal cycle experiments. The transformation behavior is essentially the same in these two
specimens. In the specimen C–C aged for 72 ks or more the
MRT is also observed. However, the peak area of R2 is
relatively smaller than that in the specimens A–A and B–B as
discussed later in Fig. 4. On the other hand, there is no MRT
in the specimen D–D for any aging periods within the present
experiment as seen in (d). It is apparent that the MRT is also
suppressible with the regulation of heat treatment atmosphere
as well as the MMT. The Ms temperature of all the specimens
aged for 36 to 1800 ks is not detected irrespective of the heat
treatment atmosphere within the temperature range measured
in the present experiment. The decrease in Ms is probably
caused by the fine and dense Ti3 Ni4 particles which obstruct
the shape change for martensitic transformation.7–9) Figure 4
shows the latent heat of transformation estimated from the
DSC cooling curve in the specimens A–A to D–D aged at
473 K for 720 ks. The latent heat of the summation of R1 and
R2 transformations in the specimens A–A to C–C and that of
single transformation in the specimen D–D are nearly the
same about 160 J/mol (3.1 J/g) which is comparable to that
in the previous reports.10) This fact suggests that the R-phase
transformation in the aged Ni-rich Ti–Ni alloys is a single
step in nature the same as that in ternary alloys such as Ti–
Ni–Fe11) and Ti–Ni–Al.12) When comparing the condition B
with the condition C, the purification of heat treatment
Effect of Heat Treatment Atmosphere on Multistage R-Phase Transformation
R1
A-A
B-B
R2
R1
R2
1800ks
1800ks
R1
exothermic
R1
R2
R2
R1
R2
720ks
360ks
exothermic
R2
720ks
R1
R2
360ks
R2
R1
R1
R1
72ks
R
180ks
Heat flow
Heat flow
180ks
R2
R2
R1
72ks
R
36ks
36ks
2 W/mol
2 W/mol
250
300
Temperature, T /K
200
C-C
200
250
300
Temperature, T /K
R
D-D
R1
R2
360ks
R1
R2
1800ks
720ks
R
360ks
R
180ks
R2
exothermic
720ks
R1
R1
72ks
R
180ks
Heat flow
exothermic
R2
R2
R
1800ks
R1
Heat flow
647
R
72ks
R
36ks
2 W/mol
200
250
300
Temperature, T /K
Fig. 3
DSC cooling curves of specimens A–A to D–D aged at 473 K for various periods.
180
Latent Heat, L /(J/mol)
160
140
R2
120
100
R
80
60
R1
40
20
0
A-A
36ks
2 W/mol
B-B
C-C
D-D
Fig. 4 Latent heat of transformation estimated from the DSC cooling curve
in the specimens A–A to D–D aged at 473 K for 720 ks.
atmosphere with the getter material is more effective to
suppress the MRT than the prevention of evaporation of Ti
and/or Ni with the Ti–Ni sheet of the same composition. In
200
250
300
Temperature, T /K
other words, it is likely that interstitial impurity elements
such as oxygen and nitrogen induce the MRT. Consequently,
a kind of compositional and/or microstructure modifications
is induced during the solution treatment with conditions A to
C and is emphasized during aging, although the detailed
mechanism has not been clarified yet.
Figure 5 shows DSC heating curves of the specimens A–A
and D–D aged at 473 K for 720 ks. As apparently from the
above results, two endothermic peaks are detected in the
curve of the specimen A–A denoted as AR1 and AR2 as
expected from the above results. On the other hand, there is a
single peak in the curve of the specimen D–D. The electropolishing for TEM specimens and the observations are
carried out around 250 K and room temperature as bounded
by solid and broken lines, respectively. One can expect that
the R-phase corresponding to the R1 peak is observed in the
specimen A–A, while the R-phase developed in the whole
area is probably seen in the specimen B–B. Figure 6 shows
M. Nishida, K. Ishiuchi, K. Fujishima and T. Hara
Heat flow
648
Jet polishing
TEM observation
AR1
AR2
D-D
Heating
endothermic
A-A
AR
2 W/mol
250
300
Temperature, T / K
Fig. 5 DSC heating curves of specimens A–A and D–D aged at 473 K for
720 ks. Electropolishing of specimens for TEM observations were
completed around 250 K bounded by solid lines. Observations were
performed around room temperature bounded by broken lines.
bright image and corresponding electron diffraction patterns
in the specimen A–A. The R-phase with dark contrast is only
observed around the grain boundary as seen in (a), which can
be confirmed by the electron diffraction patterns in (b) and (c)
taken from the areas B and C in (a), respectively. The
characteristic 1/3 h110iB2
superlattice reflections of the Rphase can be seen in (b).11) Such superlattice reflections
cannot be recognized and the spots derived from B2 structure
is only seen in (c). However, the intensity of 1/7 h321iB2
13–15)
reflections derived from Ti3 Ni4 precipitates
is too weak
to reproduce in the patterns. It is apparent that the R-phase
transformation due to the R1 peak takes place around the
grain boundary, and that corresponding to the R2 peak
probably appears at the grain interior. Figure 7(a) shows low
magnification dark field image of a crystal grain in the
specimen D–D. Bright field images in Figs. 7(b) and (c) are
enlarged micrographs at the areas B and C in (a). As expected
in Figs. 3(d) and 5, the R-phase with needle-like morphology
is recognized in the whole grain. These observations suggest
that the MRT is induced by the macroscopic heterogeneity
between around the grain boundary and inside the grain in the
specimens aged under unregulated atmosphere, rather than
the microscopic or nanoscopic heterogeneity around the
precipitates defined as the ‘‘affected’’ and the ‘‘unaffected’’
zones by Kim et al.6)
From these results it is considered that a kind of compositional fluctuation is induced during the solution treatment,
when the specimen was sealed in an evacuated quartz tube
under unregulated atmosphere. Then the fluctuation is
emphasized during the subsequent aging treatment. The
heterogeneity of internal stress fields may also occur during
the precipitation. Consequently, the MRT induces in the
specimen heat treated without the regulation of atmosphere.
Although the mechanism and process of compositional
fluctuation during heat treatment have not been clarified
yet, it may be related to the evaporation of Ti and Ni and the
contamination of interstitial impurity elements. The latter is
more effective than the former as discussed in Figs. 3 and 4.
The precise analysis around the grain boundary and inside the
grain is required to confirm this hypothesis. It is now under
study and will be reported in due course.
R
(a)
(b)
102R 030
R
000
R
R
B2
E.B. // [201]R
C
(c)
011
B
000 110
R
2µm
E.B. // [111]
Fig. 6 (a) Bright field image in specimen A–A aged at 473 K for 720 ks. (b) and (c) Electron diffraction pattern taken from areas B and C
in (a) showing R-phase and B2 phase at grain boundary and interior, respectively.
Effect of Heat Treatment Atmosphere on Multistage R-Phase Transformation
(b)
(a)
649
(c)
B
C
2µm
500nm
500nm
Fig. 7 (a) Dark field image of a crystal grain in specimen D–D aged at 473 K for 720 ks. (b) and (c) Enlarged bright field image at areas
B and C in (a) showing needle-like R-phase variants.
4.
Conclusions
In the present study we have demonstrated that the
appearance and disappearance of the MRT in aged Ni-rich
Ti–Ni alloys depend on the heat treatment atmosphere. No
MRT occurs when the evaporation of Ti and/or Ni is
prevented and the purification of heat treatment atmosphere
is achieved. The R1 and R2 transformations in the MRT take
place along the grain boundary and at the grain interior,
respectively, in the specimens aged under unregulated
atmosphere. We concluded that the MRT are extrinsic in
nature, and is an artifact during heat treatment in aged Ni-rich
Ti–Ni alloys, rather than intrinsic in nature.
Acknowledgements
This work was supported by Grant-in-Aid for Scientific
Research (B) from Japan Science Promotion Society, COEdirecting Research Program B at Kumamoto University on
Nano-space Electrochemistry and Special Program at
Kumamoto University for Promoting Research on Nanospace Electrochemistry.
REFERENCES
1) L. Bataillard, J.-E. Bidaux and R. Gotthardt: Philos. Mag. A 78 (1998)
327–344.
2) J. Khalil-Allafi, X. Ren and G. Eggeler: Acta Mater. 50 (2002) 793–
803.
3) J. Khalil-Allafi, A. Dlouhy and G. Eggeler: Acta Mater. 50 (2002)
4255–4274.
4) G. Fan, W. Chen, S. Yang, J. Zhu, X. Ren and K. Otsuka: Acta Mater.
52 (2004) 4351–4362.
5) M. Nishida, T. Hara, T. Ohba, K. Yamaguchi, K. Tanaka and K.
Yamauchi: Mater. Trans. 44 (2003) 2631–2636.
6) J. I. Kim, Yinong Liu and S. Miyazaki: Acta Mater. 52 (2004) 487–499.
7) Z. Nishiyama: Martensitic Transformation, (Academic Press, NY.,
1978), pp. 289–292.
8) M. Nishida and T. Honma: Scr. Metall. 18 (1984) 1299–1302.
9) S. Miyazaki and Otsuka: Metall. Trans. A. 17A (1986) 53–63.
10) T. Fukuda, M. Takahata, T. Kakeshita and T. Saburi: Mater. Trans. 42
(2001) 323–328.
11) C. M. Hwang, M. Meichle, M. B. Salamon and C. M. Wayman: Philos.
Mag. A 47 (1983) 31–62.
12) C. M. Hwang and C. M. Wayman: Scr. Metall. 17 (1983) 381–384.
13) T. Tadaki, Y. Nakata, K. Shimizu and K. Otsuka: Trans. JIM. 27 (1986)
731–740.
14) T. Saburi, S. Nenno and T. Fukuda: J. Less-Common Metals 12 (1986)
157–166.
15) M. Nishida, C. M. Wayman and T. Honma: Mater. Sci. Eng. 93 (1987)
191–203.