Effect of heat-treatment on the texture and anisotropy of ...

Materials Science and Engineering A273 – 275 (1999) 763 – 768 www.elsevier.com/locate/msea

Effect of heat-treatment on the texture and anisotropy of transformation strain in Ti–Ni–Fe rolled thin plates A. Khantachawana a, S. Miyazaki a,*, H. Iwai b, M. Kohl c a

Institute of Materials Science, Uni6ersity of Tsukuba, Tsukuba, Ibaraki, 305 -8573, Japan b The Furukawa Electric Co., Ltd, Nikko, Tochigi 321 -14, Japan c Forschungszentrum Karlsruhe, GmbH, IMT, Postfach 3640, 76021 Karlsruhe, Germany

Abstract Rolling and annealing processes were applied to make Ti – 50.5Ni – 0.4Fe (at.%) thin plates. Such processes create a specific texture in the plates, and the texture causes an anisotropy of transformation strain to appear. The purpose of this study is to investigate the effect of annealing on the texture and anisotropy of the transformation strain in Ti – 50.5Ni – 0.4Fe (at.%) plates with a cold-rolling reduction of 37.5% and a thickness of 100 mm. They were annealed at 673, 873 and 1273 K, respectively, for 3.6 ks. A crystallite orientation distribution function (ODF) was measured by using diffraction from {110}, {211} and {200} planes. Inverse pole figures were drawn on the basis of the ODF. Using the crystallite axis density in the inverse pole figures, the transformation strain was calculated as a function of the angle from the rolling direction (RD) in the rolling plane. In the rolling plane, the transformation strain was almost constant in a range of the angles from 0 to 30°, and decreased with a further increase of the angle. The anisotropy of the transformation strain in the rolling plane became strongest in a specimen annealed at 873 K. © 1999 Elsevier Science S.A. All rights reserved. Keywords: Ti – Ni – Fe; Shape memory effect; Superelasticity; Texture; Strain anisotropy; Martensitic transformation

1. Introduction Ti–Ni shape memory alloys have been used for fabricating various types of actuators in shapes of wire, coil and plate. Actuators made of wires and coils are characterized by one-directional actuation, while actuators made of rolled thin plates 100 mm thick or less are characterized by large bending deflection. The advantage of thin plates (compared to wires) is that they provide a high flexibility in realization of lateral shapes, which allow optimum solutions of forces and displacements for a given application. Since Ti – Ni plates of 100 mm thick or less were recently made by cold-rolling, two-dimensional shape actuators such as diaphragms and multi-directional beam-cantilever type actuators have been fabricated [1 – 6]. Rolling and annealing will affect both the deformation texture and the recrystallization texture. The usual heat-treatment temperature for shape memorization is below the recrystallization * Corresponding author. Tel./fax: +81-298-53-5283. E-mail address: [email protected] (S. Miyazaki)

temperature of about 773 K. In this case, heat-treatment only causes the recovery to occur without changing the crystallite orientation distribution appreciably, so that the deformation texture remains and affects the shape memory characteristics of the rolled plates [7]. Although the final texture is a deformation texture, the final rolled plates were subjected to cold-rolling and intermediate annealing for many times so that the annealing texture also has been involved in the rolling process. Therefore, the control of the annealing texture is also of importance for the purpose of controlling the final texture of rolled TiNi-based shape memory alloy plates. Several researchers have investigated the texture and transformation anisotropy in rolled Ti –Ni thick [8,9] and thin plates [10,11]. However, no systematic work has been conducted on the effect of annealing temperatures. The purpose of this paper is to show the effect of annealing temperature on the texture and anisotropy of the transformation strain in Ti–Ni–Fe thin plates, which have low transformation temperatures and are in the parent phase at room temperature so that the measurement of texture can be performed easily in the parent phase without using a heating stage.

0921-5093/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved. PII: S 0 9 2 1 - 5 0 9 3 ( 9 9 ) 0 0 4 1 2 - 8

764

A. Khantachawana et al. / Materials Science and Engineering A273–275 (1999) 763–768

Fig. 1. {110} pole figures for the specimens annealed at 673, 873 and 1273 K, respectively.

2. Specimens and experimental procedure

3. Results and discussion

A Ti–50.5Ni– 0.4Fe (at.%) ingot was made by melting in a high-frequency induction vacuum furnace. It was hot-forged followed by repeating coldrolling and intermediate annealing, then finally cold-rolled down to a thickness of 100 mm with a cold-rolling reduction of 37.5%. The rolled plates were cut by a spark cutting machine to make circlular plates with a diameter of 34 mm for an X-ray diffraction (XRD) pole figure measurement. The circular plates were annealed at three temperatures, 673, 873 and 1273 K, respectively, for 3.6ks. These temperatures were chosen in order to obtain different internal structures: i.e. 673 K which is below the recrystallization temperature so that the deformation texture is retained with a high density of dislocations thermally rearranged by recovery, 873 K which is a little above the recrystallization temperature so that fine grains of sub-micron sizes are formed, and 1273 K under which the recrystallized grains grow sufficiently. Oxide layers were removed by etching before and after annealing. An XRD method was applied at room temperature to investigate the texture formed at each annealing temperature. This alloy composition was designed by adding a little more Ni and Fe to an equiatomic Ti–Ni alloy in order to decrease the transformation temperatures and thus to obtain superelasticity at room temperature. Diffractions from three crystal planes, {200}, {110} and {211}, were used to measure three corresponding pole figures. Utilizing the three pole figures, a crystallite orientation distribution function (ODF) was derived. Inverse pole figures were constructed from the ODF. On the basis of the texture information, transformation strains were calculated along various directions between the rolling direction (RD) and the transverse direction (TD) in order to estimate the anisotropy of the transformation strain.

Fig. 1 shows three {011} pole figures representing three specimens annealed at 673, 873 and 1273 K, respectively. The center of the pole figures corresponds to the direction normal to the specimen surface (ND). The pole figures show how the B 011\ axis density distributes in the specimen coordinate system RD– TD–ND. Two other {200} and {211} pole figures were also measured, though they are not shown in the present paper. In the case of the specimen annealed at 673 K, the B 011\ axis density shows a dense region with two peaks symmetrically located along TD, covering a region within about 35° from ND. The two axis density peaks around 15 and − 15° from ND. This suggests that some of the crystal planes within 35° apart from {011} locate on the rolling plane. The {200} pole figure measured in the specimen reveals that the axis density peaks are located symmetrically in regions 35 –50° from ND both to RD and − RD, and 55–66° from ND both to TD and − TD. The {211} pole figure showed that the high axis density peaks are located symmetrically at 18° from ND both to RD and − RD, and at 15° from ND both to TD and − TD. According to the results of these three pole figures, any of {111}, {221}, {331}, etc. planes seem to locate preferentially on the rolling plane in the specimen annealed at 673 K. Fig. 1 also shows a {110} pole figure of the specimen annealed at 873 K. The pole figure, similarly to the above case, shows an axis density distribution with high axis density areas within 35° from ND, although the two high axis density peaks slightly shift to ND. The {200} pole figure shows two peaks at 55° from ND both to RD and − RD, while the {211} pole figure shows two peaks at 15–20° from ND both to RD and − RD. According to the three pole figures, any one of {111}, {221}, {331}, etc. planes also seems to be located preferentially on the rolling plane in the specimen annealed at 873 K.


765

Fig. 2. Sections (82 =45°) of the crystallite orientation distribution functions for the specimens annealed at 673, 873 and 1273 K, respectively.

The {110}, {200} and {211} pole figures of the specimen annealed at 1273 K do not show symmetry in high axis density peaks, the first case being shown in Fig. 1. The maximum density for each pole figure is the highest among the specimens annealed at various temperatures. Fig. 2 shows 82 =45° sections of the ODFs for the specimens annealed at 673, 873 and 1273 K, respectively. The specimen annealed at 673 K reveals a texture similar to a fiber texture with an axis around B 111\, though it is not an ideal fiber texture. The specimen annealed at 873 K shows a {221}B 11( 0\ texture. The specimen annealed at 1273 K includes two types of textures, i.e. {332}B11( 0 \ and {111}B01( 1\. The maximum orientation densities Imax for these three specimens are 5.5, 8.5 and 12.8, respectively. In order to investigate the anisotropy of the transformation strain in the rolling plane, it is necessary to know the axis density distribution as a function of angle from RD in the rolling plane. For this purpose, inverse pole figures give useful information with which we can know the transformation strain distribution qualitatively by a glance. Fig. 3 shows three sets of inverse pole figures corresponding to three angles (b) from RD in the rolling plane, 0, 45 and 90°, in the specimens annealed at 673, 873 and 1273 K, respectively. The inverse pole figures for RD of the specimen annealed at 673 K show that [011] and [1( 11] axis densities present a maximum peak and a second peak, respectively. Both orientations induce a considerably high transformation strain [7,12,13]. However, the axis density is not high for [3( 55], which is the orientation for inducing the maximum transformation strain. The inverse pole figure corresponding to 45° from RD shows a decrease in the [1( 11] axis density, while a high axis density zone expands from [011] to [1( 12]. Along the direction with an angle of 90° (TD), an axis density peak appears again at [011], though the density is lower than that for RD.

As for the specimen annealed at 873 K, the inverse pole figure corresponding to 0° (RD) shows a high axis density region covering a band region from [011] to [3( 55]. The inverse pole figure corresponding to 45° also shows an axis density distribution similar to that for RD, suggesting that the transformation strain does not change appreciably along directions between 0 and 40°. However, the inverse pole figure for 90° from RD shows that the [3( 55] axis density decreases, though an axis density peak is still located at [011]. These inverse pole figures indicate that the transformation strain will decrease with increasing angle from 45 to 90°. The inverse pole figures for the specimen annealed at 1273 K shows that the maximum axis density locates around [011] for all angles. However, the considerably high axis density area at [3( 35] for the angle of 0° shifts toward [001] with increasing angle. Some representative axis densities shown in the inverse pole figures are plotted as a function of angle from RD for the specimens annealed at 673, 873 and 1273 K, respectively, in Fig. 4. The representative axes are [001], [011], [1( 11] and [3( 55], which are characterized by the lowest transformation strain for the [001], by a considerably high transformation strain for the [011] and [1( 11], and by the highest for the [3( 55]. The [011] and [3( 55] axis densities show an opposite angle dependence in the specimen annealed at 673 K; i.e. the former and latter show their minimum and maximum points, respectively, at 45°. The combined effect of both axis densities does not appreciably contribute to the transformation strain anisotropy. The [1( 11] axis density decreases with increasing angle from RD and the variation is largest among the four axis densities, so that it affects the transformation strain anisotropy. In the case of the specimen annealed at 873 K, the [011] and [3( 55] axis densities decrease with increasing angle from RD, while the [1( 11] axis density shows a peak at 35°. The [001] axis density is negligible at angles less than 45° and shows a slight increase at higher angles in both specimens. The axis densities for


766

Fig. 3. Inverse pole figures corresponding to the directions with angles of 0, 45 and 90° from RD for the specimens annealed at 673, 873 and 1273 K, respectively.

the specimen annealed at 1273 K show a complicated angle dependence. It is conventionally assumed that the most favorable martensite variant grows to induce the maximum recoverable transformation strain o iM in each grain: o iM can be calculated by using the lattice constants of the parent phase and martensite phase. For example, the transformation strains along [001], [011], [1( 11] and [3( 35] are 2.7, 8.4, 9.8 and 10.6%, respectively. By averaging o iM for representative 36 orientations which are located in a [001]–[011]–[1( 11] stereographic standard triangle, the transformation strain for a polycrystal can be estimated as follows, if there is no specific texture and the axis density in each inverse pole figure is uniform [14].

36

o¯ 0M = % o iM

,

36

(1)

i=1

If there is texture, the axis density I i is not uniform in each inverse pole figure so that it is necessary to consider I i in the calculation of the transformation strain as follows:

36

,

o¯ H % o iMI i M=

36

30°, though the former and latter show a slight decrease and increase, respectively. Then, the strain keeps decreasing with further increasing angle. However, the strain for the specimen annealed at 1273 K decreases with increasing angle until 70°, then increases with further increasing angle. The maximum strain omax is M obtained at 0° for the specimen annealed at 673 and 1273 K, while it appears at the angle of 20° for the specimen annealed at 873 K. The minimum strain omin M appears at 90° for the specimen annealed at 673 and 873 K, while it is at 70° for the specimen annealed at 1273 K. The anisotropy of transformation strain can be represented by the following expression: {(o¯ max ¯ min ¯ max M −o M )/o M }× 100 This value is plotted as a function of annealing temperature in Fig. 6. It first increases with increasing annealing temperature, then decreases with further increasing temperature, showing that the highest transformation strain anisotropy appears at 873 K.

(2)

i=1

The transformation strain oM was calculated for three specimens as a function of the angle from RD, and the results are shown in Fig. 5. The strain is almost constant irrespective of the angle for both specimens annealed at 673 and 873 K within an angle range below

4. Conclusions The effect of heat-treatment temperature on texture and transformation strain anisotropy was investigated by means of an XRD method in Ti–Ni–Fe rolled plates. Transformation strains were calculated utilizing


767

Fig. 5. Calculated transformation strains as a function of angle from RD for the specimens annealed at 673, 873 and 1273 K, respectively.

planes seem to be located preferentially on the rolling plane in all specimens annealed at 673, 873 and 1273 K. (2) The crystallite orientation density function revealed a specific texture: i.e. a texture similar to a fiber texture with an axis around B 111\ for the specimen annealed at 673 K, a {221}B 11( 0\ texture for the specimen annealed at 873 K, and a {332}B 11( 0\ and {111}B 01( 1\ texture for the specimen annealed at 1273 K. The maximum orientation densities for these three specimens are 5.5, 8.5 and 12.8, respectively. (3) The [011] and [3( 55] axis densities show an opposite angle dependence in the specimen annealed at 673 K; i.e. the former and latter show the minimum and

Fig. 4. Axis densities of representative orientations, [001], [011], [1( 11] and [3( 35], for the specimens annealed at 673, 873 and 1273 K, respectively.

the texture information. The following results were obtained. (1) According to three {110}, {200} and {211} pole figures, any of the {111}, {332}, {221}, {331}, etc.

Fig. 6. Transformation strain anisotropy {(o¯ max ¯ min ¯ max M −o M )/o M }× 100 as a function of annealing temperature.

768


maximum, respectively, at 45°, while the [1( 11] axis density decreases with increasing angle from RD. In case of the specimen annealed at 873 K, the [011] and [3( 55] axis densities decrease with increasing angle from RD, while the [1( 11] axis density shows a peak at 35°. The [001] axis density is negligible at angles less than 45° and shows a slight increase at higher angles in both specimens. The axis densities for the specimen annealed at 1273 K show a complicated angle dependence. (4) The transformation strain anisotropy was largest in the specimen annealed at 873 K.

Acknowledgements Part of this work was supported by the research grant from the Grant-in-Aid for Fundamental Scientific Research (Kiban B (1998), A(1999) from the Ministry of Education, Science, Sports and Culture, Japan.

References [1] M. Kohl, K.I.D. Skrobanek, E. Quandt, P. Schloßmacher, A. Schu¨ßler, D. Allen, J. Phys. (France) IV 5 (1996) 1187.

.

[2] M. Kohl, K.D. Skrobanek, S. Miyazaki, Proceedings of the 5th International Conference on New Actuators (Actuator-96), AXON-tech, Bremen, 1996, p. 367. [3] K.D. Skrobanek, M. Kohl, S. Miyazaki, Proceedings of the 3rd International Conference on Intelligent Materials, Lyon, France, 1996, p. 499. [4] K.D. Skrobanek, M. Kohl, S. Miyazaki, Proceedings of the IEEE Micro Electro Mechanical Systems (MEMS-97), Nagoya, Japan, 1997, p. 256. [5] M. Kohl, D. Dittmann, E. Quandt, B. Winzek, S. Miyazaki, D.M. Allen, J. Mater. Sci. Eng. A, 273 – 275 (1999) 787–791. [6] M. Kohl, K.D. Skrobanek, S. Miyazaki, Sensors and Actuators, (1999) in press. [7] K. Kitamura, S. Miyazaki, H. Iwai, M. Kohl, J. Mater. Sci. Eng. A, 273 – 275 (1999) 761 – 765. [8] J.H. Mulder, P.E. Thoma, J. Beyer, Z. Metallkd. 84 (1993) 501. [9] H. Inoue, N. Miwa, N. Inakazu, Acta Mater. 44 (1996) 4825. [10] K. Kitamura, S. Miyazaki, M. Kohl, Proceedings of the 5th International Conference on New Actuators (Actuator-96), AXON-tech, Bremen, 1996, p. 401. [11] K. Kitamura, S. Miyazaki, H. Iwai, M. Kohl, International Conference on Shape Memory And Superelastic Technologies, SMST-97, Pacific Grove, CA, 1997, p. 47. [12] S. Miyazaki, S. Kimura, K. Otsuka, Y. Suzuki, Scripta Met. 18 (1984) 883. [13] S. Miyazaki, K. Otsuka, C.M. Wayman, Acta Met. 37 (1989) 1873. [14] S.M. Tan, S. Miyazaki, Acta Mater. 46 (1998) 2729.