Transformation sequence and second phases in ternary Ti–Ni–W
shape memory alloys with less than 2 at.% W
S.F. Hsieh
a, S.K. Wu
b,∗, H.C. Lin
c, C.H. Yang
caDepartment of Mold and Die Engineering, National Kaohsiung University of Applied Science, Kaohsiung,
Taiwan 807, Republic of China
bDepartment of Materials Science and Engineering, National Taiwan University, Taipei,
Taiwan 106, Republic of China
cDepartment of Materials Science, Feng Chia University, Taichung, Taiwan, Republic of China
Received 10 May 2004; received in revised form 2 June 2004; accepted 2 June 2004
Abstract
Transformation sequence and second phases in Ti50−XNi50WX, Ti50Ni50−XWXand Ti51Ni49−XWXalloys with X = 1–2 at.% are investigated. Two different second-phase particles located at grain boundaries are observed. The lattice parameters of Ti50Ni49W1martensite are calculated
from the selected area diffraction patterns (SADP) and all of them are larger than those of Ti50Ni50alloy. Experimental results indicate that the
W atoms replace Ni atoms, instead of Ti ones in these alloys. In addition, Ti50−XNi50WXand Ti51Ni49−XWXalloys have a two-stage B2→
R→ B19transformation, but the Ti50Ni50−XWXalloy has a one-stage B2→ B19transformation because the former Ms temperatures are depressed significantly. The shape recovery of these alloys can be improved by the W solid-solution hardening. With higher matrix hardness in these alloys, there is better shape recovery.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Ti–Ni–W shape memory alloys; Transformation sequence; Second phases; Shape recovery; Lattice parameters
1. Introduction
TiNi alloys are known as the most important shape mem-ory alloys (SMAs) because of their superior properties in shape memory effect (SME) and pseudoelasticity (PE). This comes from the fact that TiNi alloys have very good duc-tility, strength, fatigue and corrosion resistance, recoverable strain, etc. It has been confirmed that TiNi SMA properties can be affected by various thermal-mechanical treatments, such as thermal cycling [1], aging treatment in Ni-rich al-loys[2,3]and cold rolling[4]. Furthermore, the addition of a third element to replace Ni and/or Ti has a substantial effect on phase transformation behavior in TiNi alloys. The starting temperature of the martensitic transformation, Ms, of TiNi
al-∗Corresponding author. Tel.: +886-2-23637846;
fax: +886-2-23634562.
E-mail address: [email protected] (S.K. Wu).
loys decreases following the substitution of Ni with V, Mn, Fe or Co[5–7], but increases remarkably following the sub-stitution of Ni with Au, Pd and Pt in amounts not less than 15–20 at.%[8–10]. The two-stage martensitic transformation appears for some TiNiX ternary alloys, for example, B2→ R
→ B19for Ti
50Ni47Fe3alloy[7], but B2→ B19 → B19for
Ti50Ni40Cu10alloy[11,12]. Here, B2 is the parent austenite,
B19is the monoclinic martensite, R and B19 are the rhombo-hedral and orthorhombic premartensite, respectively. On the other hand, the addition of Nb in TiNi alloy can widen the transformation hysteresis to above 130◦C and hence, extend the applications of coupling and sealing[13].
It is well known that a small deviation from stoichiom-etry in TiNi SMAs can give rise to significant precipitation of second phases [14,15]. This in turn affects both the al-loy’s strength and its shape memory effect. To our knowledge, the transformation behaviors and precipitated second phases of Ti–Ni–W ternary alloys with small amounts of W have
0925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jallcom.2004.06.026
122 S.F. Hsieh et al. / Journal of Alloys and Compounds 387 (2005) 121–127
been seldom reported[16]. The purpose of the present paper is to investigate the effect of substituted W on martensitic transformation sequence and precipitated second phases of Ti50−XNi50WX, Ti50Ni50−XWXand Ti51Ni49−XWXalloys
with X = 1–2 at.%. The shape memory effect of these alloys is also briefly discussed.
2. Experimental procedure
The conventional tungsten arc melting technique was employed to prepare Ti50−XNi50WX, Ti50Ni50−XWX and
Ti51Ni49−XWX alloys with X = 1–2 at.%, i.e. Ti49Ni50W1,
Ti48Ni50W2, Ti50Ni49W1, Ti50Ni48W2, Ti51Ni48W1 and
Ti51Ni47W2 alloys (in atomic percent). In order to
effec-tively introduce W into TiNi, a mother alloy of Ni 10 wt.%W was prepared. Titanium (purity 99.7 wt.%), nickel (purity 99.9 wt.%) and tungsten in the mother alloy, totalling about 120 g, were melted and remelted at least six times in an ar-gon atmosphere. A pure titanium button was also melted and used as a getter. The as-melted buttons were homogenized at 950◦C for 72 h and then quenched in water. The buttons were cut into several plates with a low speed diamond saw, sealed in evacuated quartz tubes, annealed at 900◦C× 2 h and quenched in water. DSC measurements were made with a Dupont 2000 thermal analyzer equipped with a quantitative scanning system 910 DSC cell for controlled heating and cooling runs on sample encapsulated in an aluminum pan. The running temperature range was from−150◦C to 250◦C with a heating and cooling rate of 10◦C/min. A quantitative analysis of the alloys, composition was performed by using a JOEL JXA-8900R electron probe microanalyzer (EPMA) equipped with a wavelength dispersive X-ray spectrometer (WDS) analysis system. The operating voltage of EPMA was 20 kV and the software for quantification was the XM-97312 quantitative analysis program. The shape memory effect was examined by a bending test, as illustrated in a previous paper
[17]. Microstructure observations were made by transmis-sion electron microscopy (TEM) with a PHILIPS-CM200 mi-croscope equipped with a conventional double-tilting stage. The TEM specimens were prepared by electropolishing at
−10◦C with an electrolyte consisting of 20% H
2SO4 and
80% CH3OH by volume. The applied voltage for
electropol-ishing was 20 V.
3. Results and discussion
3.1. Second phases and martensite in Ti–Ni–W alloys Fig. 1a–f show the EPMA back-scattering electron im-ages (BEIs) of homogenized Ti49Ni50W1, Ti48Ni50W2,
Ti50Ni49W1, Ti50Ni48W2, Ti51Ni48W1and Ti51Ni47W2
al-loys, respectively. There are three different contrast areas that can be observed in Ti50Ni49W1, Ti50Ni48W2, Ti51Ni48W1
and Ti51Ni47W2alloys, including the gray matrix, the black
particles and the white particles, as shown inFig. 1c–f. Only gray matrix and white particles located around grain bound-aries appear inFig. 1a and b of Ti49Ni50W1and Ti48Ni50W2
alloys, respectively. Chemical compositions of the matrix and second-phase particles determined by EPMA are given in
Table 1. The data shown inTable 1 are the averages taken of at least five tests for each area. The atomic ratios r = Ti/(Ni + W) for the matrix and black particles are also cal-culated inTable 1. Based on the 1227◦C Ti–Ni–W ternary phase diagram[18]and backscattering electron characteris-tics of EPMA image,Table 1indicates that all the matrix in
Fig. 1is TiNi SMA containing a little W in solid solution. The amount of W atoms in the solid solution of the TiNi matrix increases slightly with an increase in the W content in Ti–Ni–W SMAs. The black particles are the Ti2(Ni, W)
phase in Ti50Ni50−XWX, Ti51Ni50−XWXalloys with X = 1
and 2 at.%. All the white particles inFig. 1 are the W-rich solid solution (W-s), in which large amounts of Ti and Ni can be accommodated. According to the r ratios of the matrix and black particles shown inTable 1, the W atoms in Ti–Ni–W SMAs are proposed to substitute into Ni atomic sites, instead of Ti sites. This characteristic is discussed further below.
Fig. 2a shows the TEM bright field image of martensite in an as-annealed Ti50Ni49W1alloy.Fig. 2b–d are the selected
area diffraction patterns (SADPs) of this alloy, in which the foil is parallel to the [1 0 0]M, [0 1 0]Mand [0 0 1]Mdirection,
respectively. From the SADPs ofFig. 2b–c, the lattice param-eters of martensite in Ti50Ni49W1alloy can be calculated as
a monoclinic structure with a = 0.301 nm, b = 0.423 nm, c = 0.472 nm andβ = 97.5◦. These parameters are all larger than those of the binary Ti50Ni50 alloy (a = 0.2889 nm, b =
0.4120 nm, c = 0.4622 nm andβ = 96.8◦[19]). Bricknell et al.[20]reported that the factors affecting the crystal structure stability and the Ms transformation temperature in Ti–Ni–Cu SMAs are atomic size, relative ionic size, electronegativity and density of state at the Fermi level. The same phenomena may occur in Ti50Ni49W1 alloy.Table 2shows the values
of the relevant parameters for the three elements involved in this study[21]. InTable 2, W is larger than Ni but less than Ti in atomic size. W has also the electronegativity which is higher than Ni but is much higher than Ti. From the view-point of comparable electronegativity, W atoms in Ti–Ni–W SMAs should replace Ni atomic sites, instead of Ti ones. This characteristic is further confirmed by the fact that the lattice parameters of martensite in Ti50Ni49W1alloy ofFig. 2are
all larger than those in equiatomic TiNi alloy because the smaller Ni atoms are replaced by the larger W ones in this alloy.
3.2. Transformation sequence in Ti–Ni–W alloys
Fig. 3a–l show the experimental results of DSC mea-surements for the annealed Ti49Ni50W1, Ti48Ni50W2,
Ti50Ni49W1, Ti50Ni48W2, Ti51Ni48W1and Ti51Ni47W2
al-loys on cooling (Fig. 3a–f) and heating (Fig. 3g–l) cycles. The peaks M* and A* (including Ms and As temperatures)
Fig. 1. The EPMA back-scattering electron images (BEIs) of the 950◦C× 72 h homogenized (a) Ti49Ni50W1, (b) Ti48Ni50W2, (c) Ti50Ni49W1, (d) Ti50Ni48W2,
(e) Ti51Ni48W1and (f) Ti51Ni47W2alloys.
appearing inFig. 3are associated with the forward and re-verse martensitic transformation, respectively.Fig. 3c, d, i and j for Ti50Ni49W1 and Ti50Ni48W2 alloys each exhibit
a single exothermic and endothermic peak, respectively, on each cooling and heating curve.Fig. 3a, b, e–h, k and l for Ti49Ni50W1, Ti48Ni50W2, Ti51Ni48W1and Ti51Ni47W2
al-loys, respectively, show double DSC peaks appearing on the cooling curve, but only one DSC peak on the heating curve. The single DSC peak occurring inFig. 3c and d indicates that it is the B2→ B19 one-stage martensitic transformation. Double DSC peaks appearing inFig. 3a, b, e and f are asso-ciated with the B2→ R → B19two-stage martensitic trans-formation. Here R is the rhombohedral premartensite phase. The results ofFig. 3demonstrate that the transformation
se-quences of annealed Ti–Ni–W SMAs can be divided into two groups, the one-stage B2→ B19transformation appearing in Ti–Ni–W SMAs with the matrix having r≈ 1.0, and the two-stage B2→ R → B19transformation with the matrix having r > 1.0 or r < 1.0. Here r is the atomic ratio of Ti/(Ni + W), as defined inSection 3.1. The former group includes Ti50Ni49W1and Ti50Ni48W2alloys, which can be regarded
as the equivalent of Ti50Ni50SMA since W atoms occupy Ni
atomic sites in these alloys. From the same viewpoint, the lat-ter group includes the Ti-rich Ti–Ni–W SMAs (Ti51Ni48W1
and Ti51Ni47W2alloys) and Ni-rich ones (Ti49Ni50W1 and
Ti48Ni50W2alloys).
FromFig. 3, all DSC results of M*(Ms), R* and A* tem-peratures are listed in Table 3. Specimen hardnesses, Hv,
124 S.F. Hsieh et al. / Journal of Alloys and Compounds 387 (2005) 121–127 Table 1
Compositional analyses by EPMA of Ti50−XNi50WX, T50Ni50−XWXand Ti51Ni49−XWXalloys with X = 1 and 2 at.% homogenized at 950◦C for 72 h
Ti (at.%) Ni (at.%) W (at.%) Ti/(Ni + W) r ratio Remark
950◦C× 72 h as quenched Ti49Ni50W1 M 49.26 ± 0.15 50.45 ± 0.18 0.29 ± 0.08 0.971 Matrix W 10.25 ± 1.47 3.59 ± 0.72 86.16 ± 2.54 – W-s Ti48Ni50W2 M 49.13 ± 0.20 50.54 ± 0.20 0.33 ± 0.10 0.966 Matrix W 9.60 ± 1.32 3.68 ± 0.64 86.72 ± 2.36 – W-s Ti50Ni49W1 M 49.98 ± 0.24 49.71 ± 0.20 0.31 ± 0.10 0.999 Matrix W 15.17 ± 1.02 3.29 ± 0.57 81.54 ± 1.32 – W-s B 66.76 ± 0.25 32.94 ± 0.25 0.30 ± 0.07 2.008 Ti2(Ni, W) Ti50Ni48W2 M 50.07 ± 0.20 49.55 ± 0.26 0.38 ± 0.08 1.003 Matrix W 14.49 ± 1.38 3.56 ± 0.53 81.95 ± 1.21 – W-s B 67.05 ± 0.20 32.68 ± 0.26 0.27 ± 0.07 2.035 Ti2(Ni, W) Ti51Ni48W1 M 50.33 ± 0.20 49.32 ± 0.20 0.35 ± 0.07 1.013 Matrix W 14.26 ± 1.32 3.62 ± 0.45 82.12 ± 1.13 W-s B 66.97 ± 0.28 32.71 ± 0.24 0.32 ± 0.06 2.028 Ti2(Ni, W) Ti51Ni47W2 M 50.46 ± 0.18 49.17 ± 0.18 0.37 ± 0.06 1.019 Matrix W 13.94 ± 1.22 3.53 ± 0.48 82.53 ± 1.08 – W-s B 67.18 ± 0.20 32.53 ± 0.25 0.29 ± 0.07 2.047 Ti2(Ni, W)
M: matrix; B: black particles; W: white particles (a tungsun-rich solid solution, W-s). Table 2
Significant parameters of the metals species in this study[21]
Element Atomic radius (nm) Electronegativity Valence state Ionic radius (nm)
Ti 0.147 1.54 +2 0.076
+4 0.064
Ni 0.125 1.91 +2 0.078
W 0.137 2.36 +4 0.068
+6 0.065
including the matrix and second phase particles, are also shown inTable 3. FromTable 3, transformation peak temper-atures M*, R* and A* versus W content (at.%) in Ti–Ni–W SMAs are plotted inFig. 4. FromFig. 4, it is clear that the peak temperatures A*, R* and M* decrease with increasing amounts of W in Ti-rich and Ni-rich Ti–Ni–W SMAs, but not in Ti50Ni50−XWXalloys (the equivalent Ti50Ni50SMA). At
the same time, the Ms temperature can be further depressed to induce the R phase by using small amounts of W in
Ni-Table 3
Peak temperatures and specimen hardnesses, including matrix and second phase particles, for Ti50−XNi50WX, T50Ni50−XWXand Ti51Ni49−XWXalloys with X = 1 and 2 at.% M* (◦C) Ms (◦C) R* (◦C) A* (◦C) Vickers hardness (Hv) Matrix Particles 27◦C 92◦C Ti49Ni50W1 −38 −28 −17 −2 263 273 295 Ti48Ni50W2 −55 −37 −36 −16 272 280 298 Ti50Ni49W1 57 63 − 90 214 – 265 Ti50Ni48W2 51 55 − 80 223 – 269 Ti51Ni48W1 41 50 48 74 235 256 274 Ti51Ni47W2 31 43 39 64 242 262 281
rich Ti–Ni–W SMAs, rather than in Ti-rich ones.Fig. 4and
Table 3also show that transformation temperatures decrease, but the hardness increases, with increasing amounts of W in Ti-rich Ti51Ni49−XWX alloys and Ni-rich Ti50−XNi50WX
alloys. This phenomenon also occurs in Ti-rich Ti52Ni47Al1
alloy[22], and can be ascribed to W solid-solution hard-ening. It is pointed out that any strengthening mechanism which can impede the transformation shear can lower the Ms(M*) transformation temperature because the martensite
Fig. 2. (a) TEM bright-field image of the annealed Ti50Ni49W1alloy, (b) SADP of (a) with [1 0 0]Mzone axis, (c) SADP of (a) with [0 1 0]Mzone axis, (d)
SADP of (a) with [0 0 1]Mzone axis.
transformation involves a shear process[23–25]. Moreover, martensitic transformation can be retarded by the fine hard dispersed particles, which may result in transformation tem-peratures being much more depressed.
The r ratios of the matrix in Ti50Ni49W1and Ti50Ni48W2
alloys are near 1.0, and their peak temperatures are almost the same as those of the equiatomic Ti50Ni50alloy[4]. A similar
result can also be found in a reported Ti50Ni49.5W0.5alloy (As
= 97◦C, Af= 107◦C, Ms = 65◦C and Mf= 33◦C)[16].Fig. 1c
and d show that there are many second phase particles around grain boundaries in Ti50Ni49W1and Ti50Ni48W2alloys. It is
well known that M*(Ms) temperatures can be depressed due to the introduction of oxygen in the matrix of TiNi alloys[26]. But these oxygen atoms can be easily absorbed by the Ti2Ni
phase to form the Ti4Ni2O oxide[15,27]. In Ti50Ni49W1and
Ti50Ni48W2 alloys, the oxygen atoms in the matrix can be
absorbed by the Ti2(Ni, W) particles, and hence the oxygen
content in the matrix can be reduced. This feature will con-tribute to raise transformation temperatures. However, the W dissolved in the TiNi matrix increases the hardness, which in turn depresses the transformation temperatures. These two
compensation effects make that the transformation tempera-tures of Ti50Ni49W1and Ti50Ni48W2alloys are quite similar
to those of Ti50Ni50SMA.
3.3. Shape recovery on Ti–Ni–W alloys
The shape recovery was measured by bending tests. For this, the specimen is deformed in liquid nitrogen (−196◦C) for Ti49Ni50W1, Ti48Ni50W2, Ti51Ni48W1and Ti51Ni47W2
alloys, respectively, and then is heated above the Af
tem-perature to complete the reverse martensitic transformation. Both Ti50Ni49W1and Ti50Ni48W2 alloys have
microstruc-tures similar to Ti51Ni48W1and Ti51Ni47W2alloys, so
bend-ing tests are omitted for them. Table 4 presents the mea-sured shape recovery at different bending strain,εs, for the
aforementioned specimens. The volume fraction of the sec-ond phase particles in these specimens is also shown in
Table 4.Table 3indicates that the hardness of these parti-cles is greater than that of TiNi matrix.Tables 3 and 4show that the shape recovery is a function of the matrix hardness, bending strain and second phase volume. The shape recovery
126 S.F. Hsieh et al. / Journal of Alloys and Compounds 387 (2005) 121–127
Fig. 3. DSC curves for as-annealed Ti49Ni50W1, Ti48Ni50W2, Ti50Ni49W1, Ti50Ni48W2, Ti51Ni48W1and Ti51Ni47W2alloys on cooling and heating. M* and
A* are the peak temperatures of forward and reverse martensitic transformation, respectively; and R* is a peak temperature for premartensitic transformation.
Fig. 4. Transformation temperatures of M*, R*and A* vs. W content for the annealed Ti–Ni–W alloys.
increases with increasing matrix hardness, but decreases with increasing amounts of bending strain and second phase. It is reported that the shape recovery of TiNi alloys can be increased by different strengthening/hardening processes
[17,28]. In other words, if the matrix is strengthened, the slip of dislocations is more difficult than the movement of twin boundaries during the application of external stress. Hence, the permanent plastic strain is reduced and the shape recov-ery is improved. In the present study, the matrix of TiNi
al-Table 4
The measured shape recovery, RSME, at different bending strain,εs, and the
volume fraction of second phase particles for Ti–Ni–W specimens TiNiW alloys Shape recovery, RSME(%) Second phase (vol.%)
εs= 4% εs= 8%
Ti51Ni48W1 91.1 85.9 12
Ti51Ni47W2 92.1 87.2 13
Ti49Ni50W1 100 97.5 3
loys can be solid-solution strengthened by the addition of 1–2 at.% W, as shown inTable 3. Hence, the shape recov-ery is increased. Although the shape recovrecov-ery is also related to the existence of second phase particles, the effect com-ing from the volume of second phase particles is much less than that from solid-solution strengthening. For example, as can be seen fromTable 4, the volume of second phase par-ticles in Ti51Ni47W2 alloy (about 13%) and in Ti51Ni48W1
alloy (about 12%) is almost the same as in Ti51Ni49(about
10%) alloy[15], both Ti51Ni47W2(92.1%) and Ti51Ni48W1
(91.1%) have better shape recovery than Ti51Ni49 (88%) [15] since the former two have the harder matrix. Simi-lar results can also be found in annealed Ti49Ni50W1 and
Ti48Ni50W2alloys. Therefore, in Ti–Ni–W SMAs, a higher
solid-solution hardness in the matrix will provide better shape recovery.
4. Conclusion
In homogenized Ti50−XNi50WX, Ti50Ni50−XWX and
Ti51Ni49−XWX SMAs with X = 1–2 at.%, a small amount
of W can be dissolved in the matrix, forming two differ-ent second phase particles, Ti2(Ni, W) and a tungsten-rich
solid solution. DSC test results show that the martensitic transformation sequence of annealed Ti50−XNi50WX and
Ti51Ni49−XWX alloys with X = 1–2 at.% is two-stage B2 → R → B19, but that of Ti
50Ni50−XWXwith X = 1 2 at.%
is one-stage B2→ B19. The transformation peak temper-atures of the former-group alloys decrease with increasing W addition, but those of the latter group are almost identi-cal to those of Ti50Ni50alloy. The B19martensite structure
of Ti50Ni49W1alloy is calculated from the SADPs of TEM
as a = 0.301 nm, b = 0.423 nm, c = 0.472 nm andβ = 97.5◦ and these parameters are all larger than those of the binary Ti50Ni50 alloy. W is larger than Ni but smaller than Ti in
atomic size, and is much higher than Ti but not so much higher than Ni in electronegativity. Accordingly, W atoms in Ti–Ni–W SMAs are proposed to replace Ni atoms, instead of Ti ones. EPMA results for the composition of the matrix and second phase particles also confirm this characteristic. The shape recovery of Ti–Ni–W SMAs can be improved by the W solid-solution hardening of the matrix. With higher matrix hardness in Ti–Ni–W alloys, there is better shape recovery.
Acknowledgements
The authors gratefully acknowledge the financial support of this research by the National Science Council (NSC), Re-public of China under Grants NSC 91-2216-E151-009 and NSC 92-2216-E002-008.
References
[1] S. Miyazaki, Y. Igo, K. Otsuka, Acta Metall. 34 (1986) 2045. [2] S.K. Wu, H.C. Lin, T.S. Chou, Scripta Metall. 23 (1989) 2043. [3] M. Nishida, T. Honma, Scripta Metall. 18 (1984) 1293.
[4] H.C. Lin, S.K. Wu, T.S. Chou, H.P. Kao, Acta Metall. 39 (1991) 2069.
[5] K.H. Eckelmeyer, Scripta Metall. 10 (1976) 667.
[6] R. Wasilewski, in: J. Perkin (Ed.), Shape Memory Effects in Alloys, Plenum, New York, NY, 1975, p. 245.
[7] C.M. Hwang, M. Meichle, M.B. Salamon, C.M. Wayman, Phil. Mag. 47A (1983) 177.
[8] S.K. Wu, C.M. Wayman, Metallography 20 (1987) 359.
[9] Y.C. Lo, S.K. Wu, C.M. Wayman, Scripta Metall. 24 (1990) 1571. [10] K. Enami, T. Yoshida, S. Nenno, Proc. ICOMAT-86, The Japan
In-stitute of Metals, Nara, Japan, 1987.
[11] Y.C. Lo, S.K. Wu, H.E. Horng, Acta Metall. Mater. 41 (1993) 747. [12] T. Nam, T. Saburi, Y. Kawamura, K. Shimizu, Mater. Trans. JIM 31
(1990) 262.
[13] C.S. Zhang, L.C. Zhao, T.W. Duerig, C.M. Wayman, Scritpa Metall. Mater. 24 (1990) 1807.
[14] M. Nishida, C.M. Wayman, T. Honma, Metall. Trans. 17A (1986) 1505.
[15] H.C. Lin, S.K. Wu, J.C. Lin, Mater. Chem. Phys. 37 (1994) 184. [16] K. Enami, M. Hara, H. Maeda, J. de Physique IV (1995) C8-629. [17] H.C. Lin, S.K. Wu, Scripta Metall. 26 (1992) 59.
[18] P. Villars, A. Prince, H. Okamoto, Handbook of Ternary Alloy Phase Diagram, vol. 10, ASM International, Pittsburgh, USA, 1995. [19] K. Otsuka, T. Sawamura, K. Shimizu, Phys. Stat. Sol. 5 (1971) 457. [20] R.H. Bricknell, K.H. Melton, O. Mercier, Metall. Trans. 10A (1979)
693.
[21] J.P. Schaffer, A. Saxena, S.D. Antolovich, T.H. Sanders Jr., S.B. Warner, The Science and Design of Engineering Materials, second ed., McGraw-Hill, New York, 1999.
[22] S.F. Hsieh, S.K. Wu, J. Mater. Sci. 34 (1999) 1665.
[23] M. Cohen, E.S. Machlin, V.G. Paranjpe, Thermodynamics in Physi-cal Metallurgy., in: American Society for Metals, Metals Park, OH, 1950, p. 242.
[24] E. Hornbogen, Acta Metall. 33 (1985) 595.
[25] S. Eucken, E. Hornbogen, J. Mater. Sci. 19 (1984) 1343.
[26] T. Honma, in: H. Funakubo (Ed.), Shape Memory Alloys, Gordon and Breach, 1984, p. 89.
[27] M. Nishida, C.M. Wayman, Mater. Sci. Eng. 93 (1987) 191. [28] H.C. Lin, S.K. Wu, Acta Metall. 25 (1991) 1295.
